pmc_id
stringlengths 10
11
| clean_content
stringlengths 1
3.62M
|
|---|---|
PMC10000552 | Background: The present study analyzed the impact of margin status on local control and survival, and the management of close/positive margins after transoral CO2 laser microsurgery for early glottic carcinoma. Methods: 351 patients (328 males, 23 females, mean age 65.6 years) underwent surgery. We identified the following margin statuses: negative, close superficial (CS), close deep (CD), positive single superficial (SS), positive multiple superficial (MS), and positive deep (DEEP). Results: A total of 286 patients (81.5%) had negative margins, 23 (6.5%) had close margins (8 CS, 15 CD) and 42 (12%) had positive margins (16 SS, 9 MS, 17 DEEP). Among the 65 patients with close/positive margins, 44 patients underwent enlargement, 6 radiotherapy and 15 follow-up. Twenty-two patients (6.3%) recurred. Patients with DEEP or CD margins showed a higher risk of recurrence (hazard ratios of 2.863 and 2.537, respectively), compared to patients with negative margins. Local control with laser alone, overall laryngeal preservation and disease-specific survival decreased significantly in patients with DEEP margins (57.5%, 86.9% and 92.9%, p < 0.05). Conclusions: Patients with CS or SS margins could be safely submitted to follow-up. In the case of CD and MS margins, any additional treatment should be discussed with the patient. In the case of DEEP margin, additional treatment is always recommended. laryngology early glottic cancer laser surgery margins outcomes This research received no external funding. pmc1. Introduction Transoral CO2 laser microsurgery (CO2 TOLMS) represents a well-standardized and minimally invasive approach allowing good oncological and functional results in patients with early glottic squamous cell carcinoma (SCC) . This ultraconservative approach can be burdened by an increased incidence of close and positive margins of resection, which are unanimously associated with a higher risk of local relapse . In the case of inadequate margins, different options have been suggested on the basis of the number of the involved margins (single vs. multiple) and site (deep vs. superficial): strict follow-up, revision surgery or radiotherapy (RT) . However, precise indications for additional treatment and its effective impact on local control and survival rates are still debated. There is a common consensus that further treatments are required in patients with positive deep margin of resection at definitive histology, while the management of close and/or positive superficial margins is still an ongoing matter of discussion since additional resection can potentially hinder the functional results. The high rate of false positive margins after enlargement reported in the literature (up to 80%) has driven several authors to reduce second-look procedures in recent years, implementing a watchful waiting policy . We reviewed a large homogeneous cohort of patients affected by Tis-T1-T2 glottic SCC treated using CO2 TOLMS with the aim of analyzing the impact of margin status on local control, survival and organ preservation. The decision-making process and effective indication of additional therapeutic options in cases of close and positive margins have been discussed. 2. Materials and Methods The authors performed a retrospective analysis of 351 patients with early glottic SCC (Tis-T1-T2) treated by the senior author with CO2 TOLMS from October 1993 to November 2005 and from December 2010 to December 2020 at the Department of Otorhinolaryngology of an Italian institution (ethics committee protocol number 895/2018). All the patients included in the study had no clinical involvement of the lymph node at the time of surgery. During the preoperative work up, all patients underwent fiberlaryngoscopy, while computed tomography (CT) or magnetic resonance imaging (MRI) of the neck with contrast medium were considered necessary in selected patients to rule out invasion of the paraglottic space (PGS), of the preepiglottic space and of the cartilage. Intraoperative work up was always performed using rigid 0deg and 70deg scopes. From the end of 2013, preoperative and intraoperative endoscopic work up was coupled with narrow-band imaging (NBI) (Olympus Medical Systems Corporation, Tokyo, Japan) and the IMAGE1 S SystemTM (Storz, Tuttlingen, Germany) plus enhanced contact endoscopy (ECE) . On the day of surgery, all patients received ceftriaxone (1000 mg IV) or, as a substitute if allergic, ciprofloxacin (400 mg IV), according to the antibiotic prophylaxis protocol of our institution. All patients underwent CO2 TOLMS under general anesthesia with orotracheal intubation (Mallinckrodt laser safe tube, I.D. 5.0-7.0 mm; Athlone, Ireland). Sharplan 1030 and Acupulse CO2 lasers with an Acuspot, Acublade 712 micromanipulator and Digital AcuBladeTM (Lumenis(r), Yokneam, Israel) set on the superpulsed mode (10 W, continuous, acublade 1-3 mm) were used in most of the cases. The UltraPulse/Surgitouch CO2 laser (Lumenis(r), Yokneam, Israel) was used from 2020. Adequate laryngeal exposure in microlaryngoscopy was obtained using the Kleinsasser laser laryngoscopes modified by Rudert with the Riecker-Kleinsasser suspension system (Karl Storz, Tuttlingen, Germany). Endoscopic cordectomies were classified according to the European Laryngological Society (ELS) . Resections were always performed using an en bloc procedure when the volume of the tumor allowed it. Larger tumors were removed using a piecemeal technique. When indicated, the anterior commissure was resected through a subperichondrial dissection. Resections were performed in macroscopic free margins. Specimens were sent for histology opportunely oriented by the surgeon by staining the superior edge with ink to obtain the precise mapping of the lesions, also after piecemeal removal. Frozen sections were not routinely performed because they are unrepresentative of the whole mucosal margins, and the time of execution can become excessive for organizational reasons and hospital logistics. After definitive histology, all lesions were staged and restaged according to the eighth edition of the Union for International Cancer Control-American Joint Committee on Cancer (UICC-AJCC) TNM staging system . According to Fiz et al. , the margin status was classified as follows: negative, close (tumor-margin distance < 1 mm) superficial (CS), close deep (CD), positive (presence of at least carcinoma in situ at the surgical margin) single superficial (SS), positive multiple superficial (MS) and positive deep (DEEP). In the case of close or positive margins, intraoperative recording was reviewed and discussed in a multidisciplinary team. The policy after histology was as follows: CO2 transoral enlargement or postoperative RT was always performed in the case of DEEP or MS margins; SS and CD margins were almost constantly enlarged with a laser, except selected cases who underwent a close wait-and-see policy; CS margins were managed with close follow-up. Patients scheduled for a second look with CO2 TOLMS were treated at 30 to 40 days after the first cordectomy, because by that point the scar tissue is completely healed and the glottic aspect can be better evaluated. Voice rehabilitation and regular follow-up were scheduled according to the NCCN guidelines in all cases : patients underwent fiberlaryngoscopy every month during the first year, every 2 months during the second year and every 3-4 months until the fifth year after surgery, in the absence of any recurrence and/or secondary disease. From 2013, fiberlaryngoscopy was coupled with NBI. Patients included in the present study were followed up from the date of surgery until December 2022, when possible. Statistical analysis was performed on the basis of the data reported as supplementary material (Table S1), using GraphPad Prism software (GraphPad, San Diego, CA, USA). Survival probabilities over time were estimated using the Kaplan-Meier method, considering the six different types of margins (negative, CS, CD, SS, MS, DEEP). The entry point was the date of laser cordectomy. The first studied outcome was disease-specific survival (DSS), with the end point being patient's death due to laryngeal cancer or last follow-up. The second outcome was recurrence-free survival (RFS), with the end point set at the date of recurrence or at the last available visit. The third outcome was local control with laser alone (LCL), with the end point set at the date of RT or open procedure for recurrences. Organ laryngeal preservation (OLP) was the fourth measured outcome, with the end point set at the date of total laryngectomy or at last follow-up. The log-rank (Mantel-Cox) test was applied to compare recurrence rates between patients with negative margins versus patients with close/positive margins. A p value < 0.05 was considered to be statistically significant. The influence of the routine intraoperative use of NBI, IMAGE1 S and ECE in the incidence of positive superficial margins was evaluated. 3. Results Three hundred and fifty-one patients (328 males, 23 females, mean age 65.6 years, age range 29-90 years) with early glottic SCC treated with CO2 TOLMS were included in the study. Of the 351 patients treated with CO2 TOLMS, 34 (9.7%) underwent type I cordectomy, 94 (26.8%) type II cordectomy, 77 (21.9%) type III cordectomy, 21 (6%) type IV cordectomy, 122 (34.8%) type V cordectomy and 3 (0.8%) type VI cordectomy. Patient and tumor characteristics, and the numbers and types of surgical cordectomies are detailed in Table 1. The mean hospitalization time was 2.9 days. Three patients experienced a postoperative bleeding that required an endoscopic cautery under general anesthesia. Thirty-four patients (9.7%) developed an anterior glottic web. Among them, the 12 patients with moderate to severe symptoms were managed endoscopically with the laser incision of the web and the harvesting of a mucosal microflap, while the 22 patients with mild symptoms were referred to voice therapists, as suggested in the literature . Two hundred and eighty-six patients (81.5%) had negative margins after primary surgery, while 65 patients (18.5%) had close or positive margins. Twenty-three patients (6.5%) had close margins, among whom 8 had CS and 15 had CD margins. Forty-two patients (12%) had positive margins: 16 had SS margin, 9 MS margins and 17 DEEP margin (Table 1). Seventeen out of the 229 patients (7.4%) who underwent CO2 TOLMS before the systematic intraoperative use of enhancement systems (NBI, IMAGE1 S and ECE) had positive superficial margins. After the implementation of bioendoscopy, only 8 out of 122 patients (6.6%) experienced positive superficial margins. Therefore, the use of enhancement tools reduced the incidence of positive superficial margins, although the decrease was not statically significant (p = 0.76). Forty-four (67.7%) of the 65 patients with close/positive margins underwent CO2 laser enlargement, obtaining negative margins in all cases. Among these patients, 12 had DEEP margin, 8 had MS margins, 14 had SS margin and 10 had CD margin. Definitive histology showed the presence of residual carcinoma in only 8 (18.2%) of these 44 cases: 3 patients with initially DEEP margin, one with MS margins, and 4 with SS margin. Definitively, 330 patients (94%) showed negative margins after primary surgery and subsequent enlargements. A total of 7 patients underwent postoperative RT: six (9.2%) of the 65 patients had close/positive margins (5 with DEEP margin and one with MS margins), and one patient had negative margins but lymphovascular invasion at definitive histology. Fifteen (23.1%) of the 65 patients with close/positive margins underwent a close wait-and-see policy, among whom 8 had CS margin, 5 had CD margin and 2 had SS margin. The mean follow-up was 4.97 years. Twenty-two of the 351 patients (6.3%) experienced recurrence: laryngeal recurrence in 21 cases and nodal recurrence in 1 case. No patients developed distant metastasis. Recurrence occurred in 17 patients with negative margins, in 3 patients with DEEP margin who underwent surgical enlargement without evidence of residual tumor at histology, and in 2 patients with CD margin who underwent follow-up. Twenty-one patients underwent salvage treatment; one patient refused additional treatment and died of disease. Salvage therapy of the 21 patients with recurrences included: CO2 TOLMS alone in 11 cases, CO2 TOLMS and RT in 1 case, RT alone in 1 case, type II OPHL in 1 case, type III OPHL in 1 case, type II OPHL and RT in 1 case, total laryngectomy (TL) in 3 cases, TL and RT in 1 case and radical neck dissection and RT in 1 case (Table 2). The five-year DSS, RFS, LCL and OLP of the whole series were 99.6%, 92.9%, 94.6% and 98.2%, respectively. Survival rates and Kaplan-Meier survival curves relative to the different subtypes of margins are reported in Table 3 and Figure 1. In the univariate analysis, patients with DEEP margin showed a 2.863 (p = 0.08) times increased risk for recurrence compared to patients with negative margins (Table 4). Furthermore, in the case of DEEP margin, both LCL, OLP and DSS decreased in a statistically significant way: 57.5%, 86.9% and 92.9%, respectively, in patients with DEEP margin versus 96.4%, 98.7% and 100%, respectively, in patients with negative margins, p < 0.05. Patients with CD margin experienced a 2.537 (p = 0.2) times increased risk of recurrence (RFS of 86.7% versus 93.2% in patients with negative margins) (Table 4). 4. Discussion The surgical margins required in early glottic SCC are narrower than those considered necessary in other head and neck cancers because of the scarce glottic submucosal lymphatic network. In the literature, margins >= 1 mm of healthy tissue are generally considered adequate in patients treated with CO2 TOLMS . The high magnification available with the operative microscope and modern biologic endoscopic techniques makes the surgical approach with such narrow margins easier. In our series, we found negative margins in 286 cases (81.5%) and close margins in 23 cases (6.5%), whereas positive margins were present in 42 cases (12%). Our rate of positive margins is, encouragingly, in the range reported in the literature (9.3-45.4%) . Different prognostic factors, such as stage or anterior commissure involvement, may be associated with local recurrence, but various studies have demonstrated that positive margins represent an independent risk factor for local failure . In the literature, it is reported that a local recurrence rate ranging from 3.1% to 22.8% is observed in cases of negative surgical margins, while the recurrence rate rises from 8% to 51% in cases of positive margins . This great variability of incidences of local relapse could be related to the experience of the surgeon and to the different interpretation of the margin status performed by the pathologist. According to Fiz et al. , we analyzed our oncological outcomes, classifying the margins as negative, CS, CD, SS, MS and DEEP. The intraoperative use of rigid endoscopy with bio endoscopic tools such as NBI and IMAGE1 S has been suggested as a useful tool in achieving optimal superficial margins outlines, and has been shown to potentially decrease the rate of positive superficial margins . According to the literature, in our series, the systematic use of NBI, IMAGE1 S and ECE decreased the number of positive superficial margins (6.6% vs. 7.4%), although the difference was not statistically significant (p = 0.76). In the present series, the CS and SS margins did not negatively impact RFS. Close superficial margins were always managed with close follow-up. Regarding SS margin, in the early years of the present series, our policy was for systematic enlargement, but since the majority of revision surgeries resulted in negative specimens (71.4%), in the second part of the present series, we adopted a strict endoscopic follow-up, performing a second CO2 TOLMS only when the surgeon expressed doubts concerning the resection, as suggested in the literature . Avoiding unnecessary surgical enlargements has the advantage of sparing the voice quality, because a second TOLMS clearly results in a further loss of tissue, increasing the scarring of the residual vocal cords. The low incidence of cases with residual disease at histology after the second look could be explained by the loss of the narrow healthy tissue due to the laser effect (thermal damage), and/or because of the shrinkage of the specimen . Human tissues, especially the mucosa, have an intrinsic propensity to undergo shrinkage after surgical resection because of the presence of contractile proteins in the connective tissue and their release from the surrounding structures. In the literature, a mean shrinkage of mucosal specimens after CO2 TOLMS, from intralaryngeal measurement to postresection, of 3.8 +- 0.3 mm in the anteroposterior length of the glottic plane is reported . Such important shrinkage could explain the high number of unnecessary enlargements to obtain wider or free margins. Therefore, careful harvesting and orientation of the surgical specimen is mandatory for the correct evaluation of the tumor extent. Several protocols have been described in the literature to improve margin assessment. Michel et al. and Aluffi Valletti et al. suggested the use of two different colored inks to tag superficial and deep mucosal sides before formalin fixation. In the present series, all the specimens were systematically three-dimensionally oriented by staining the superior margin with ink, and were analyzed by a dedicated pathologist. Some authors reported an increased risk of recurrence in the presence of MS margins . In our series, none of the nine cases with MS margins experienced recurrences, and postoperative RT was deemed necessary in one patient with multifocal carcinoma. In patients with multifocal SCC, it is difficult to assess the true superficial extension of the lesion and obtain a radical excision; thus, multiple CO2 TOLMS are needed, and, in some instances, RT represents an additional tool. In the eight patients with MS margins who underwent CO2 laser enlargement, the resection was carried out starting from a wider macroscopically free margin including the scar of the previous surgery. Definitive histology showed the presence of residual carcinoma in only one (12.5%) of these eight patients. These data suggest that the majority of patients with MS margins could be overtreated with revision surgery. Nowadays, the use of HD flexible endoscopy with bioendoscopy performed in the office setting during the follow-up improves the accuracy of the early detection of persistent or recurrent disease after CO2 TOLMS. This could be a reason to shift to a less aggressive attitude concerning second-look procedures in cases of doubtful or positive superficial margins, even multiple, opting for a close follow-up with NBI endoscopy . Fiz et al. reported that close deep margins were related to an increased number of relapses, with an RFS of 77.1%. This was also confirmed by our findings, since patients with CD margin experienced a lower RFS (86.7% versus 93.2% in patients with negative margins) with a 2.537 times increased risk of recurrence; however, this result was not statistically significant (p = 0.2). Two of the five patients with CD margin who underwent follow-up experienced recurrence and were managed with additional CO2 TOLMS. Among the 10 patients with CD margin who underwent the second look, none showed residual carcinoma at histology. Ultimately, LCL, OLP and DSS were not negatively impacted by the CD margin status. As a consequence, we believe that the CD margin does not represent an absolute indication for additional treatment, although a strict follow-up with imaging is mandatory to detect early recurrence. According to the literature, DEEP margins are generally associated with the highest recurrence rate . In our experience, patients with DEEP margin showed an increased risk of recurrence compared to patients with negative margins, with a hazard ratio of 2.863 (p = 0.08). Five patients required RT and one patient underwent total laryngectomy after recurrence, whereas one patient refused treatment after recurrence and died of disease. Consequently, both LCL, OLP and DSS were negatively affected in patients with DEEP margin (respectively 57.5%, 86.9% and 92.9% versus 96.4%, 98.7% and 100% in patients with negative margins, p < 0.05). In our series, none of the five patients who underwent postoperative RT after DEEP margin at histology experienced recurrence. In the case of positive margins, the role of postoperative RT is still debated. Some authors found a benefit in RFS after postoperative RT, whereas others could not demonstrate any significant difference in patients submitted for adjuvant treatment when compared with those followed up with a compulsory protocol of surveillance . Furthermore, postoperative RT results in a multimodal therapeutic approach for early tumors that could have been managed by RT alone from the beginning, with additional biological and economic costs . Moreover, the patient would lose the possibility of being treated with RT in the event of a laryngeal second primary or recurrence . A recent review pointed out that the application of postoperative RT in patients with positive margins following the first resection depends on the confidence of the surgeon with wider resection, and can range from 10% to 44% . However, if a second CO2 TOLMS is judged unlikely to result in "true negative margins", then open partial surgery or RT are the most appropriate choices . Among the 12 patients with DEEP margin who underwent additional CO2 TOLMS, 3 patients recurred, despite surgical enlargement with no residual tumor at the histology. These "false" negative margins are difficult to interpret: although missing at histology, the residual disease cannot be excluded. In two of these cases, the patients were initially treated with extended type V cordectomies for cT2 tumors, with clinical involvement of the posterior third of the vocal cord. In these cases, histology confirmed the presence of the cancer immediately lateral to the vocal process and close to the PGS. A lower local control can be observed in the case of understaging of T2/T3 tumors. The misdiagnosis of cT2 can be associated with the complexity of the assessment of the PGS. The correct identification of the involvement of the posterior PGS is essential to choose the appropriate therapeutic strategy, since patients with posterior glottic tumors have poor local control when treated with CO2 TOLMS . The limit that separates the anterior from the posterior laryngeal compartments is a virtual plane described as tangential to the vocal process and perpendicular to the ipsilateral thyroid lamina. We believe that, in the case of posterior DEEP margins after extended cordectomies, the surgeon should consider the possibility of micro infiltration of the posterior PGS and, consequently, decide on postoperative RT or open surgery. Evidently, our analysis was limited by the reduced number of each subgroup of patients, which did not allow us to perform a multivariate analysis. Therefore, to draw any definite conclusions, large controlled multi-center retrospective and prospective trials are needed. 5. Conclusions The present study confirms that, although CO2 TOLMS in early glottic cancer offers optimal oncological outcomes, the margin status impacts on local control, and it is mandatory to stratify the different types of margins for their distinctive prognostic significance. In the present series, the CS and SS margins behaved similarly, and could be followed-up after adequate counseling with the patients. We observed that the CD and MS margins do not have a statistically significant negative impact on DSS, RFS, LCL or OLP, and any additional treatment should be thoroughly discussed with the patient so as to avoid unnecessary overtreatment. Patients with DEEP margin should always undergo additional treatment, and if a second CO2 TOLMS is judged unlikely to result in "true negative margins", open surgery or RT must be considered. Supplementary Materials The following supporting information can be downloaded at: supplementary material are reported in the Methods as Table S1. Click here for additional data file. Author Contributions Conceptualization, C.M., F.C. and R.P.; data curation, C.M., M.B., V.M., M.T., V.P. and C.G.; formal analysis, C.M. and F.C.; investigation, C.M. and F.C.; methodology, C.M. and F.C.; resources, C.G. and R.P.; supervision, R.P.; validation, C.M., F.C. and R.P.; writing--original draft, C.M. and F.C.; writing--review and editing, C.M., F.C. and R.P. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement This research did not involve any animal models; this research involved human participants in accordance with the ethical standards of the institutional and/or national research committees and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The research was approved by the Ethics Committee of the University Hospital of Cagliari (protocol number 895/2018). Informed Consent Statement Written informed consent has been obtained from the patients included in the study. Data Availability Statement The data that support the findings of this study are available as electronic supplementary material. Conflicts of Interest The authors declare that this manuscript was conceived and written by the cited authors, they do not have any financial interest to disclose, and they confirm no conflict of interest concerning this manuscript. Figure 1 Kaplan-Meier curves for the entire cohort of patients showing disease-specific survival, recurrence-free survival, local control with laser alone and overall laryngeal preservation in relation to margin status. cancers-15-01490-t001_Table 1 Table 1 Patients who underwent CO2 TOLMS for early glottic cancer. Variables No. of Patients All 351 Age 65.6 (range 29-90 years) Male/female 328/23 Type of surgical cordectomies 34 type I cordectomy 94 type II cordectomy 77 type III cordectomy 21 type IV cordectomy 122 type V cordectomy (61 a, 7 ab, 4 abc, 11 abcd, 2 abd, 10 ac, 8 acd, 7 ad, 7 b, 3 bc, 2 c) 3 type VI cordectomy Clinical N classification 351 cN0 Pathological T classification 34 pTis 193 pT1a 61 pT1b 63 pT2 Margin status 286 NEG 8 CS 15 CD 16 SS 9 MS 17 DEEP NEG = negative. CS = close superficial. CD = close deep. SS = positive single superficial. MS = positive multiple superficial. DEEP = positive deep. a = extended cordectomy encompassing the contralateral vocal fold. b = extended cordectomy encompassing the arytenoid. c = extended cordectomy encompassing the ventricular fold. d = extended cordectomy encompassing the subglottis. cancers-15-01490-t002_Table 2 Table 2 Recurrences after CO2 TOLMS for early glottic cancer. Patient/Sex/Age (Years) Type of Cordectomy pT Margin Status after Primary Treatment CO2 Laser Enlargement Site of Relapse/Time of Relapse (Years) Salvage Treatment Outcome/Time of Last Follow-Up (Years) DG/M/76.3 Va 1b NEG - Larynx/1.7 CO2 TOLMS and radiotherapy DOC/2.5 FE/M/68.5 IV 1b NEG - Larynx/0.3 CO2 TOLMS NED/7 MG/M/70.1 II 1a NEG - Larynx/0.6 Type II horizontal laryngectomy and radiotherapy NED/5 VA/M/61.9 II 1a NEG - Larynx/2.4 CO2 TOLMS DOC/11.1 RB/M/82.7 Vabc 2 DEEP Yes (no residual tumor at histology) Larynx/2.4 Refused treatment DOD/2.5 DE/M/72.8 II 1a NEG - Larynx/0.7 CO2 TOLMS DOC/4.9 LS/M/76.6 Va 2 NEG - Larynx/1.7 CO2 TOLMS DOC/4.3 FG/M/64.1 Vacd 2 NEG - Larynx/1.3 Type II horizontal laryngectomy NED/11 MS/M/53.5 IV 1a NEG - Larynx/3 Total laryngectomy NED/7.8 SC/M/61.4 III 1a NEG - Larynx/0.8 CO2 TOLMS NED/6.6 CS/M/60.9 II 1a NEG - Larynx/1.8 CO2 TOLMS NED/2.1 ZF/M/71.5 Vabcd 2 NEG - Larynx/0.2 Total laryngectomy DOC/0.5 PB/M/64.1 Vac 2 DEEP Yes (no residual tumor at histology) Larynx/1.2 Total laryngectomy NED/5.2 RM/M/73.8 I 1a NEG - Larynx/1.2 Radiotherapy NED/5.7 DS/F/62.8 Vabcd 2 NEG - Larynx/0.5 Total laryngectomy and radiotherapy NED/5.1 BM/M/51.7 Vb 1a NEG - Larynx/2.1 Type III horizontal laryngectomy NED/3.2 SG/M/68.3 Vac 1b NEG - Neck node/0.9 Neck dissection and radiotherapy NED/1.5 MI/M/48.1 II 1a DEEP Yes (no residual tumor at histology) Larynx/1.2 CO2 TOLMS NED/4.8 LM/M/68.7 II 1b CD No Larynx/0.5 CO2 TOLMS NED/5 PA/M/74.1 Va 1b NEG - Larynx/0.8 CO2 TOLMS NED/5 CA/M/70.3 III 1b NEG - Larynx/0.6 CO2 TOLMS DOC/2.2 MV/M/65.7 II 1b CD No Larynx/1.3 CO2 TOLMS NED/1.6 CO2 TOLMS = transoral CO2 laser microsurgery. NEG = negative. CD = close deep. DEEP = positive deep. NED = no evidence of disease. DOD = died of disease. DOC = died of other causes. cancers-15-01490-t003_Table 3 Table 3 Survival rates of the cohort of patients who underwent CO2 TOLMS for early glottic cancer. DSS 5 Years RFS 5 Years LCL 5 Years OLP 5 Years All patients (n = 351) 99.6% 92.9% 94.6% 98.2% NEG margins (n = 286) 100% 93.2% 96.4% 98.7% CS margin (n = 8) 100% 100% 100% 100% CD margin (n = 15) 100% 86.7% 100% 100% SS margin (n = 16) 100% 100% 100% 100% MS margins (n = 9) 100% 100% 88.9% 100% DEEP margin (n = 17) 92.9% 80.9% 57.5% 86.9% DSS = disease-specific survival. RFS = recurrence-free survival. LCL = local control with laser alone. OLP = overall laryngeal preservation. NEG = negative. CS = close superficial. CD = close deep. SS = positive single superficial. MS = positive multiple superficial. DEEP = positive deep. cancers-15-01490-t004_Table 4 Table 4 Univariate analysis of different types of margins for recurrence-free survival. Recurrence 5-Year RFS Hazard Ratio (95% CI) p NEG margins (n = 286) 17 (5.9%) 93.2% 1 (Reference) NA CD margin (n = 15) 2 (13.3%) 86.7% 2.537 (0.2854-22.55) =0.2 DEEP margin (n = 17) 3 (17.6%) 80.9% 2.863 (0.4395-18.66) =0.08 NA = not applicable. CI = confidence interval. NEG = negative. CD = close deep. DEEP = positive deep. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000553 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050765 healthcare-11-00765 Article Child and Adolescent Mental Health during the COVID-19 Pandemic: Challenges of Psychiatric Outpatient Clinics Mosheva Mariela Conceptualization Methodology Investigation Data curation Writing - original draft Writing - review & editing Visualization 12 Barzilai Yael Conceptualization Methodology Software Validation Formal analysis Investigation Data curation Writing - original draft Writing - review & editing Visualization Project administration 3 Hertz-Palmor Nimrod Conceptualization Methodology Software Validation Formal analysis Investigation Data curation Writing - review & editing Visualization 14 Mekori-Domachevsky Ehud 12 Avinir Asia Conceptualization Methodology Investigation Data curation Writing - review & editing 12 Erez Galit Conceptualization Methodology Investigation Data curation Writing - review & editing 3 Vardi Noa Conceptualization Methodology Investigation Data curation Writing - review & editing 25 Schoen Gila Conceptualization Methodology Investigation Data curation Writing - review & editing 25 Lahav Tal Conceptualization Methodology Investigation Data curation Writing - review & editing 26 Sadeh Hadar Conceptualization Methodology Investigation Data curation Writing - review & editing 78 Rapaport Michal Conceptualization Methodology Investigation Data curation Writing - review & editing 29 Dror Chen Conceptualization Methodology Software Validation Formal analysis Investigation Data curation Writing - review & editing Visualization 23 Gizunterman Alex Conceptualization Methodology Investigation Data curation Writing - review & editing 10 Tsafrir Shlomit Conceptualization Methodology Investigation Data curation Writing - review & editing 11 Gothelf Doron Conceptualization Methodology Software Validation Formal analysis Investigation Resources Data curation Writing - review & editing 12 Bloch Yuval Conceptualization Methodology Software Validation Formal analysis Investigation Resources Data curation Writing - original draft Writing - review & editing Visualization Supervision 23* Hossain Liaquat Academic Editor 1 The Child Psychiatry Division, Edmond and Lily Safra Children's Hospital, Sheba Medical Center, Tel Hashomer, Ramat Gan 5262000, Israel 2 Department of Psychiatry, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv 69978, Israel 3 The Emotion-Cognition Research Unit Center, Shalvata Mental Health Center, Hod-Hasharon 45100, Israel 4 MRC Cognition and Brain Sciences Unit, University of Cambridge, Cambridge CB2 1TN, UK 5 Geha Mental Health Center, Petah Tikva 49100, Israel 6 Ness Ziona Mental Health Center, Ness Ziona 74100, Israel 7 Soroka Medical Center, Be'er-Sheva 84101, Israel 8 Goldman School of Medicine, Ben-Gurion University of the Negev, Be'er-Sheva 84105, Israel 9 Lev Hasharon Mental Health Center, Natanya 42100, Israel 10 The Jerusalem Mental Health Center, Jerusalem 91060, Israel 11 Clalit Health Services, Jerusalem 91060, Israel * Correspondence: [email protected] 06 3 2023 3 2023 11 5 76508 1 2023 25 2 2023 03 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Background: Worldwide national surveys show a rising mental health burden among children and adolescents (C&A) during COVID-19. The objective of the current study is to verify the expected rise in visits to psychiatric outpatient clinics of C&A, especially of new patients. Methods: a cross-sectional study focusing on visits as recorded in electronic medical records of eight heterogeneous C&A psychiatric outpatient clinics. The assessment was based on visits held from March to December of 2019 (before the pandemic) in comparison to visits held in 2020 (during the pandemic). Results: The number of visits was similar for both periods. However, in 2020, 17% of the visits used telepsychiatry (N = 9885). Excluding telepsychiatry reveals a monthly decrease in traditional in-person activities between 2020 and 2019 (691.6 +- 370.8 in 2020 vs. 809.1 +- 422.8 in 2019, mean difference = -117.5, t (69) = -4.07, p = 0.0002, Cohen's d = -0.30). Acceptation of new patients declined during 2020, compared to 2019 (50.0 +- 38.2 in 2020 vs. 62.8 +- 42.9 in 2019; Z = -3.12, p = 0.002, r = 0.44). Telepsychiatry was not used for new patients. Conclusions: The activity of C&A psychiatric outpatient clinics did not rise but was guarded due to the use of telepsychiatry. The decline in visits of new patients was explained by the lack of use of telepsychiatry for these patients. This calls for expanding the use of telepsychiatry, especially for new patients. outpatient clinic child and adolescents mental health COVID-19 child and adolescence psychiatry Israel National Institute for Health Policy Research5652020 Gates Cambridge TrustThis research was funded by the Israel National Institute for Health Policy Research Grant nu. 5652020. Nimrod Hertz-Palmor is supported by the Gates Cambridge Trust. pmc1. Introduction Parallel to worldwide morbidity and mortality, the COVID-19 pandemic has been causing significant emotional distress, pointing attention toward assessing and treating mental health issues. Accumulating reports address the severe and multifaceted consequences on the mental health of children and adolescents, identifying them as a particularly vulnerable group . For example, a study from China conducted at the beginning of the pandemic outbreak found that one third of 18-year-old children and adolescents were clingy, inattentive, irritable, and worried during the pandemic . Other worldwide studies have reported severe levels of psychological distress such as worries, helplessness, fear , and high anxiety and depressive symptoms among children and adolescents during the pandemic . Moreover, recent nationwide studies in the U.S. reported worsening children and adolescents' psychological well-being and behavioral health compared to before the pandemic . Additionally, in Israel, a retrospective cohort study utilized a large, computerized database and found a rise in the incidence of depression, anxiety, and eating disorders, and a rise in the use of antidepressants and antipsychotics during the pandemic years . It is of note that most of these studies were based on community surveys and questionnaires that reflected a rise in emotional distress, but not necessarily an increase in psychopathology or treatment-seeking. While hotlines, community support, and counselors relate to distress and not to psychopathology, treatment in psychiatric outpatient clinics requires a professional diagnosis. The extent to which pediatric mental health services are affected by COVID-19 and its long-term impact is still under investigation. A study that investigated hospitalization numbers in youth psychiatric wards in Israel found that the number of hospitalized patients decreased in 2020, compared to 2019 . Studies assessing ER (Emergency Room) visits in the U.S. showed a marked decrease in pediatric ER visits across a broad range of conditions; however, the proportional decline in mental health visits was less pronounced . Moreover, patients with mental health conditions presenting for ER visits since the onset of the pandemic were more likely to require admission and have had more prolonged admissions . A recent study focusing on the first year of the pandemic exemplified a decline in psychiatric ER visits of C&A, especially those suffering from stress-related anxiety and depressive disorders . In addition, C&A that did not have a previous encounter with mental health outpatient clinics were less likely to visit the ER in 2020. The burdens of lockdowns, quarantines, and social distancing, which typified the first year of the pandemic, are all suggested as major stressors for the pediatric population . The outpatient clinics are probably less intimidating than the psychiatric ER or the general ER, especially at the time of the pandemic. One would expect a rise in the activity of outpatient clinics that are central to evaluating and treating stress-related anxiety and depression disorders in the community. To overcome the challenges of physically visiting the outpatient clinics during the pandemic, the use of telemedicine and telepsychiatry (the application of telemedicine within the specialty of psychiatry) was suggested . Telepsychiatry is the delivery of psychiatric or mental health services via telecommunications technology, usually video. Studies aimed to investigate if telepsychiatry is a comparable tool for assessment and treatment show patients and providers are usually satisfied with telepsychiatry. Moreover, it was found that telepsychiatry is a comparable tool to face-to-face service in terms of reliability of assessment and treatment outcome . However, before the pandemic, outside the research setting, the use of telepsychiatry was slow and limited by clinicians' concerns about regulation and quality of care . Resulting from the worldwide COVID-19 crisis, the use and implementation of telepsychiatry has increased . A few recent papers studied outpatient clinics in different countries and displayed a decrease in referrals to psychiatry outpatient clinics compared to the year before the onset of the pandemic . Nevertheless, to the best of our knowledge, no studies have examined the actual effect of the pandemic on psychiatric outpatient mental health referrals among children and adolescents in Israel, compared to the year before the pandemic onset. The current study aims to examine the real-life activity of psychiatric outpatient clinics during the COVID-19 pandemic (March to December 2020) compared to the comparator period in the previous year (2019). Based on literature pointing to an increase in mental health burden, we hypothesized that there would be a rise in mental health needs and therapy, specifically for new patients. Beyond the general trend and numbers, it is important to evaluate the effects on different therapeutic encounters. During the pandemic, some therapeutic encounters, such as group therapy, are expected to be more challenging due to the need for social distancing. It is important to study the applicability of telepsychiatry in mending the gap between the rising needs and difficulties in accessibility to outpatient visits. Studying multiple outpatient clinics that differ in the population they serve and evaluating the condition, not just during quarantine, can aid in controlling for some of the confounders. Thus, we aimed to collect the data from eight outpatient clinics with different characteristics. 2. Materials and Methods 2.1. Sample The Israeli population is heterogeneous in ethnocultural origin, religion, and type of residency. Thus, it probably allows for better generalization compared to different populations. During the pandemic, there were more cases of infection among the Muslim minority and the Ultraorthodox Jewish community in comparison to the general Israeli population (possibly partially due to less awareness of the dangers of the pandemic and more suspiciousness about regulations set by the government) . In the Israeli health system, therapy in outpatient clinics is free of charge and provided by the four large health providers. While some patients prefer private therapy, we estimate that most C&A are treated in the public system. Approximately two percent of the general child and adolescent population is served by the public mental health system . We conducted a cross-sectional study collecting data from electronic medical records of eight public outpatient clinics in central Israel. Though we do not know the size of the population that was treated in other facilities (private or public) outside these eight outpatient clinics, we present some data on the centers and their catchment area, and we believe a considerable part of the mental health services for C&A are given by these centers. Two of the clinics are located at general tertiary hospitals. Sheba Medical Center is located in the highly populated center of the country. It was a leading center in treating COVID-19 patients, and due to its role as a leading center, can accept patients from all over the country. Soroka Medical Center is located in the south and serves a population of approximately one million inhabitants. Most of this population is considered part of Israel's "periphery" and includes both rural and non-rural habitats and a large population of Bedouins (a Muslim population with a culture and tradition of its own). Three clinics are affiliated with sizeable mental health centers (Geha Mental Health Center, Shalvata Mental Health Center, and Ness Ziona centers). The Geha Mental Health Center serves all age groups. It covers a population of approximately 800,000 inhabitants and covers a comparatively large population of Ultraorthodox Jews. The Shalvata Mental Health Center serves all age groups. It covers a population of approximately 500,000 inhabitants in the center of Israel. The catchment area includes a heterogeneous population of both Jews and Muslims. The Ness Ziona Center is the largest psychiatric center in Israel and covers an estimated population of more than a million inhabitants. Three community mental health clinics are affiliated but separated from sizeable mental health centers (community clinics affiliated with Lev Hasharon, Shalvata, and Geha). Lev Hasharon Mental Health Center serves a diverse population due to its location near both orthodox and Muslim communities, and serves approximately 500,000 inhabitants. The data analyzed covered two periods. 1 March to 31 December 2019, and 1 March to 31 December 2020. During that time in 2020, the number of COVID-19 cases in Israel escalated to 434,227 cases (4.71% percent of the population), of them 133,391 (30.71% of all COVID-19 cases) were children and adolescents. This was compared to a comparator period (1 March to 31 December 2019). Since the first case of COVID-19 in Israel was diagnosed on 27 February 2020, January and February of both years (i.e., 2019 and 2020) were removed from the analyses. 2.2. Statistical Analysis Paired t-tests were used to compare the number of monthly visits in 2020 and 2019. Each month of 2020 (March-December) was matched with its comparable pair in 2019. For example, monthly visits in March 2020 were compared to the number of visits in March 2019, April 2020 was compared to April 2019, etc. The normality of data distribution was tested using the Kolmogorov-Smirnov (K-S) test for goodness of fit. First, we compared the number of in-person visits (excluding telepsychiatry) and later conducted an additional t-test to compare the number of overall visits (in-person + telepsychiatry). We controlled the False Discovery Rate (FDR) with Hochberg and Benjamini's correction for multiple comparisons to account for type-I errors, and we presented the adjusted p-values . The number of no-shows was tested similarly to assess differences between 2020 and 2019. We calculated effect sizes for all of the above-mentioned analyses with Cohen's d and followed the accepted rule of thumb for small (d = 0.2), medium (d = 0.5), and large (d = 0.8) effect sizes . Additionally, we conducted a sensitivity analysis by stratifying the general vs. psychiatric hospital data. We compared the number of group therapy meetings and new patients' assessments (intake meetings, the first examination a patient undergoes when admitted to the clinic and held only in person) between 2020 and 2019 using the non-parametric Wilcoxon test for paired samples. We conducted non-parametric testing, instead of the parametric paired t-test, due to the data's positively skewed distribution. We compared group therapies with and without telepsychiatry and performed FDR correction as described. Effect sizes were calculated with rank-biserial r correlation. Analysis was conducted using the stats and rstatix packages in R version 4.0.3 . 3. Results 3.1. Sample Characteristics From March to December 2020, there were 56,633 visits (13,696 females, 24.2%) to outpatient child and adolescent psychiatry clinics, slightly less than the comparator period in 2019 (n = 58,294). The proportion of children younger than 12 years old decreased from 54.0% in 2019 to 46.0% in 2020, and the proportion of those older than 13 years old increased from 46.0% to 54.0% during the pandemic (Z = -27.1, p < 0.0001, Cohen's h = 0.16). The distributions of age and gender are described in Table 1. Due to technical difficulties, we were able to address diagnostic differences only for patients treated at the Shalvata Mental Health Center. 3.2. Type of Visits The central finding of the current study is that during the first year of the pandemic, a time of extreme need, in-person therapy declined in comparison to the previous year. Paired t-test revealed a significant monthly decrease in the traditional in-person activities between 2020 and 2019 (691.6 +- 370.8 in 2020 vs. 809.1 +- 422.8 in 2019, mean difference = -117.5 (-14.5% decrease), t(69) = -4.07, p = 0.0002, Cohen's d = -0.30). However, when telepsychiatry visits, accounting for 17% of the visits in 2020 (N = 9885) were included in the analysis, there was no significant difference in the number of monthly visits between 2020 and 2019 (832.8 +- 385.3 in 2020 vs. 809.1 +- 422.8 in 2019, mean difference = 23.7 (-0.03% decrease), t(69) = 1.19, p = 0.24, d = 0.06) . Monthly visits are depicted in Figure 2. There was no difference in the monthly no-show rate between 2020 and 2019 (18.9% vs. 19.1%, t(40) = -0.11, p = 0.91, d = -0.02). These trends were consistent during months when lockdowns were imposed (March-April, October-November) and non-lockdown periods. During lockdowns, there was a significant decrease in in-person visits (t(27) = -2.09, p = 0.046, d = -0.28) but no differences were present when telepsychiatry visits were taken into account (t(27) = 1.21, p = 0.24, d = 0.12) . Similarly, there was a significant decrease in in-person visits during non-lockdown months (t(41) = -3.67, p = 0.001, d = 0.31), but this decrease was eliminated when telepsychiatry was included (t(41) = 0.37, p = 0.72, d = 0.02). Sensitivity analysis results for the general vs. psychiatric data showed similar trends at the two stratified subsets . A couple of outpatient therapies were specifically more challenging during the pandemic, mainly group therapy and assessment of new patients. Paired Wilcoxon test revealed a significant monthly decrease in group therapies held in person between 2020 (Median: 282.5, IQR: 243.2) and 2019 (Median: 536, IQR: 360.25). (48.5 +- 70.3 in 2020 vs. 80.8 +- 97.5 in 2019, mean difference = -32.2 (-40.0% decrease), Z = -3.76, p = 0.0003, r = 0.49). However, also in this case, online group therapy with telepsychiatry attenuated this decline; thus, the decrease was only descriptive with no significant differences (Median: 72, IQR: 46 in 2020 VS. Median: 0, IQR: 0 in 2019) (60.7 +- 68.1 in 2020 vs. 80.8 +- 97.5 in 2019, mean difference = -20.0 (-24.9% decrease), Z = -1.65, p = 0.10, r = 0.20). We aimed to examine the differences in the encounters coded as "intake" frequencies. The paired Wilcoxon test showed a decrease in the evaluation of new patients during 2020 (Median: 258.5, IQR: 50.75), compared to 2019 (Median: 314, IQR: 53) (50.0 +- 38.2 in 2020 vs. 62.8 +- 42.9 in 2019, mean difference = -12.8 (-20.4% decrease). Z = -3.12, p = 0.002, r = 0.44). In line with this, the mean duration of being on the waiting list for an appointment was shorter in 2020 (Median: 3.7, IQR: 1.5) as compared to 2019 (Median: 5.7, IQR: 1.2) (24.6 +- 12.6 days in 2020 vs. 38.7 +- 29.9 days in 2019, mean difference = -14.1 (-36.4% decrease), Z = -2.26, p = 0.024, r = 0.50). It is worth noting that data on the waiting list was obtained from only two medical centers (Sheba Medical Center and Shalvata). 4. Discussion The present study aimed to examine the change in therapeutic services in child and adolescent psychiatric outpatient clinics during the first year of the COVID-19 pandemic in Israel. At that time, surveys and publications from non-formal services suggested an increase in the mental health burden of children and adolescents . Several challenges may have contributed to this burden, such as social distancing, fear of infection, reorganization of family routine, and possible loss of family members or friends due to the pandemic . Because of those challenges, many researchers and world health agencies (WHO, UNICEF, AACAP, and others) mentioned the need for psychological interventions and supportive care, including early identification of children's mental health deterioration by pediatric healthcare workers, differentiating symptoms indicating a need for a referral to a psychiatrist, and establishing guidelines to cope with pandemic-related mental health problems . Despite those survey results and the many challenges people had to cope with at that time, we found a similar number of psychiatric outpatient visits in 2020 compared to the comparator period in 2019. There are two reasonable explanations. First, it is possible that the stress caused was evaluated as a natural consequence and that the youngsters did not need professional help from psychiatric services. The second reasonable explanation would relate to barriers to the possibility of reaching therapy. This can be explained by the fear of visits to medical centers due to fear of being infected, leading to a decline in child and adolescent visits to mental health clinics . However, the current multicenter study included only two centers in a general hospital. The others are psychiatric clinics in psychiatric facilities or in the community, where this fear is probably less likely. In the present study, the referral trends were unchanged in both settings. The more extended period (10 months) studied in the present study included prolonged periods with fewer restrictions related to transportation and a general decline in fear of being infected. Thus, the suggested explanation relating to difficulties in accessibility is probably not the sole explanation. It has been suggested that the closure of schools, usually considered a burden on mental health, might lower the stress at least temporarily of some vulnerable youngsters, such as those who encounter school bullying or those who suffer from difficulties in academic performance and social problems. Limited direct contact with teachers and counselors, often the first to recognize mental health difficulties, is possibly an additional cause for limited referrals. The unpredictable nature of COVID-19 and significant uncertainties surrounding long-term treatment strategies resulted in the need for healthcare services to reform and explore alternative modes of service delivery such as telehealth. A recent large-scale study assessed the characteristics and barriers in the transition. The most striking difference between the two periods is that in 2020, 17% of the visits used telepsychiatry. Remote community treatment and support have long been suggested but have not previously been implemented widely because of challenges to healthcare staff and service users. These include technological barriers, legal, regulatory, and ethical issues . Clinical barriers include specific considerations in assessing emergencies such as suicidality in youth , difficulties in communication, and difficulties evaluating young children with disruptive behaviors and developmental disorders . However, in the present study, it is the telepsychiatry that "kept" the service accessible. Those therapies that telepsychiatry did not cover, precisely the encounter with new patients, had a bothering decline. Due to prudence and insufficient awareness of the possibilities of telepsychiatry, we argue that our clinics could not answer to the inclining need of youngsters for therapy. This explains the rise in the use of hotlines and not professional psychiatrists and psychologists. It is important to note that the use of telepsychiatry is also burdened by public acquaintance, comfort, and accessibility to relevant technology. For example, some of the more conservative societies in Israel (ultraorthodox Jews), do not use smartphones and the Internet to the same degree as the secular population. Evidence supporting the feasibility, acceptability, and effectiveness of telemedicine with children and adolescents in psychiatry is emerging incrementally. Telepsychiatry services in the child and adolescent population have been functioning with promising results . The COVID-19 pandemic has further highlighted the need for provisioning and setting up child and adolescent telepsychiatry services , including training medical staff, pharmacotherapies, and psychotherapies. Moreover, a recent large-scale study of eight children and adolescent mental health clinics has shown a rapid pivot from in-person services to home-based telehealth . Online group therapy is a relatively new modality for leading groups. It bears concerns such as psychotherapists' worries about being less able to communicate their empathy, build therapeutic alliances , or worry about the impact of technical barriers and confidentiality issues . As shown in our study and previous studies, with the pandemic outbreak, it became even more crucial to move groups online . 5. Limitations Our study has several limitations. First, the retrospective chart review design. It is possible that some of the cases were handled in psychiatric emergencies, outpatient clinics, or private practice. Though we used eight different centers serving a heterogeneous reasonably sized population, some of the information may be missing. The data itself relates to the therapeutic encounters. Other information, and most specifically diagnoses and outcome measures, were not available. In addition, we covered ten months of the pandemic. Thus, some of the effects might emerge as the pandemic continues. Furthermore, we did not include periods of lockdown that may affect the referrals to the clinics. 6. Conclusions There was no rise in visits to psychiatric outpatient clinics of C&A, during the first year of the COVID-19 pandemic. The therapeutic activity was salvaged by the fast incorporation of telepsychiatry. The lack of routine use of telepsychiatry for evaluating and treating new patients is probably the main reason for the decline in this type of therapeutic activity. Thus, our findings highlight the invaluable need for telepsychiatry in treating children and adolescents during crisis periods and social isolation. The results highlight the crucial need for the use of telepsychiatry in evaluating new patients. Future studies should focus on outcome measures beyond the reliability of therapy itself. Author Contributions Conceptualization: M.M., Y.B. (Yael Barzilai), N.H.-P., E.M.-D., A.A., G.E., N.V., G.S., T.L., H.S., M.R., C.D., A.G., S.T., D.G., Y.B. (Yuval Bloch). Methodology: M.M., Y.B. (Yael Barzilai), N.H.-P., E.M.-D., A.A.; G.E, N.V., G.S., T.L., H.S., M.R., C.D., A.G., S.T., D.G., Y.B. (Yuval Bloch). Software: Y.B. (Yael Barzilai), N.H.-P., C.D., D.G., Y.B. (Yuval Bloch). Validation: Y.B. (Yael Barzilai), N.H.-P., C.D., D.G., Y.B. (Yuval Bloch). Formal analysis: Y.B. (Yael Barzilai), N.H.-P., C.D., D.G., Y.B. (Yuval Bloch). Investigation: M.M., Y.B. (Yael Barzilai), N.H.-P., E.M.-D., A.A., G.E, N.V., G.S., T.L., H.S., M.R., C.D., A.G., S.T., D.G., Y.B. (Yuval Bloch). Resources: D.G., Y.B. (Yuval Bloch). Data Curation: M.M., Y.B. (Yael Barzilai), N.H.-P., E.M.-D., A.A., G.E, N.V., G.S., T.L., H.S., M.R., C.D., A.G., S.T., D.G., Y.B. (Yuval Bloch). Writing--original draft preparation: M.M., Y.B. (Yael Barzilai), Y.B. (Yuval Bloch). Writing--review and editing: M.M., Y.B. (Yael Barzilai), N.H.-P., E.M.-D., A.A., G.E, N.V., G.S., T.L., H.S., M.R., C.D., A.G., S.T., D.G., Y.B. (Yuval Bloch). Visualization: M.M., N.H.-P., Y.B. (Yael Barzilai), C.D., Y.B. (Yuval Bloch). Supervision: Y.B. (Yuval Bloch). Project administration: Y.B. (Yael Barzilai). All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement This study was reviewed and approved by the Ethics Committees 0015-20-SHA, 007-20-GEH, 7251-20-SMC, LH7/2020, 0315-20-SOR, 653-20-BN. The participants provided their written informed consent to participate in this study. Informed Consent Statement Patient consent was waived since the study is a retrospective chart review. The ethical comities approved since anonymity was granted. Data Availability Statement The raw data supporting the conclusions of this article will be made available by the authors without undue reservation. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Annual trends in visit type and use of telepsychiatry between 2019 and 2020. Y-axis represents the percent of specific visit types out of all annual visits. Data labels represent the cumulative sum of visit types during each year. No error bars are presented due to the use of cumulative sums instead of means. Figure 2 Monthly visits in 2019 as compared to 2020 (stratified with and without telepsychiatry). Figure 3 Monthly visits during the lockdown and opening periods in 2019 as compared to 2020. Bars relate to standard deviation. * p < 0.05, *** p < 0.001. Figure 4 Monthly visits to psychiatric hospitals and general hospitals in 2019 as compared to 2020. Bars relate to standard deviation. ** p < 0.01. healthcare-11-00765-t001_Table 1 Table 1 Demographic variables of patients. March-December 2019 Total = 56,663 N% March-December 2020 Total = 58,294 N% Gender Male (%) 38,095 (67.27%) 38,442 (65.95%) Female (%) 18,538 (32.73%) 19,852 (34.05%) Age groups <12 20,310 10,968 (54.0%) 9342 (46.0%) >13 13,931 6421 (46.0%) 7510 (54.0%) Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Geweniger A. Barth M. Haddad A.D. Hogl H. Insan S. Mund A. Langer T. Impact of the COVID-19 Pandemic on Mental Health Outcomes of Healthy Children, Children with Special Health Care Needs and Their Caregivers-Results of a Cross-Sectional Study Front Pediatr. 2022 10 759066 10.3389/fped.2022.759066 35223688 2. Jiao W.Y. Wang L.N. Liu J. Fang S.F. Jiao F.Y. Pettoello-Mantovani M. Somekh E. Behavioral and Emotional Disorders in Children during the COVID-19 Epidemic J. Pediatr. 2020 221 264 266.e1 10.1016/j.jpeds.2020.03.013 32248989 3. Saurabh K. Ranjan S. Compliance and Psychological Impact of Quarantine in Children and Adolescents due to COVID-19 Pandemic Indian J. Pediatr. 2020 87 532 536 10.1007/s12098-020-03347-3 32472347 4. Yeasmin S. Banik R. Hossain S. Hossain M.N. Mahumud R. Salma N. Hossain M.M. Impact of COVID-19 pandemic on the mental health of children in Bangladesh: A cross-sectional study Child. Youth Serv. Rev. 2020 117 105277 10.1016/j.childyouth.2020.105277 32834275 5. Xie X. Xue Q. Zhou Y. Zhu K. Liu Q. Zhang J. Song R. Mental health status among children in home confinement during the coronavirus disease 2019 outbreak in Hubei Province, China JAMA Pediatr. 2020 174 898 10.1001/jamapediatrics.2020.1619 32329784 6. Gadermann A.C. Thomson K.C. Richardson C.G. Gagne M. Mcauliffe C. Hirani S. Jenkins E.K. Examining the impacts of the COVID-19 pandemic on family mental health in Canada: Findings from a national cross-sectional study BMJ Open 2021 11 e042871 10.1136/bmjopen-2020-042871 33436472 7. Duan L. Shao X. Wang Y. Huang Y. Miao J. Yang X. Zhu G. An investigation of mental health status of children and adolescents in china during the outbreak of COVID-19 J. Affect Disord. 2020 275 112 118 10.1016/j.jad.2020.06.029 32658812 8. Zhou S.J. Zhang L.G. Wang L.L. Guo Z.C. Wang J.Q. Chen J.C. Liu M. Chen X. Chen J.-X. Prevalence and socio-demographic correlates of psychological health problems in Chinese adolescents during the outbreak of COVID-19 Eur. Child Adolesc. Psychiatry 2020 29 749 758 10.1007/s00787-020-01541-4 32363492 9. Ravens-Sieberer U. Kaman A. Erhart M. Devine J. Schlack R. Otto C. Impact of the COVID-19 pandemic on quality of life and mental health in children and adolescents in Germany Eur. Child Adolesc. Psychiatry 2021 31 879 889 10.1007/s00787-021-01726-5 33492480 10. Patrick S.W. Henkhaus L.E. Zickafoose J.S. Lovell K. Halvorson A. Loch S. Letterie M. Davis M.M. Well-being of Parents and Children During the COVID-19 Pandemic: A National Survey Pediatrics 2020 146 e2020016824 10.1542/peds.2020-016824 32709738 11. Gassman-Pines A. Ananat E.O. Fitz-Henley J. COVID-19 and Parent-Child Psychological Well-being Pediatrics 2020 146 e2020007294 10.1542/peds.2020-007294 32764151 12. Bilu Y. Flaks-Manov N. Bivas-Benita M. Akiva P. Kalkstein N. Yehezkelli Y. Mizrahi-Reuveni M. Ekka-Zohar A. Ben David S.S. Lerner U. Data-driven assessment of adolescents' mental health during the COVID-19 pandemic Available online: (accessed on 6 February 2023) 13. Dror C. Hertz-Palmor N. Barzilai Y. Gila S. Tali B.Z. Alex G. Tal L. Maya K.-L. Talia S. Doron G. Youth psychiatric hospitalization in Israel during COVID-19: A multi-center study Int. J. Environ. Res. Public Health 2022 19 9870 10.3390/ijerph19169870 36011509 14. DeLaroche A.M. Rodean J. Aronson P.L. Fleegler E.W. Florin T.A. Goyal M. Hirsch A.W. Jain S. Kornblith A.E. Sills M.R. Pediatric Emergency Department Visits at US Children's Hospitals during the COVID-19 Pandemic Pediatrics 2020 147 e2020039628 10.1542/peds.2020-039628 33361360 15. Krass P. Dalton E. Doupnik S.K. Esposito J. US Pediatric Emergency Department Visits for Mental Health Conditions during the COVID-19 Pandemic JAMA Netw. Open 2021 4 8 11 10.1001/jamanetworkopen.2021.8533 16. Dror C. Hertz-Palmor N. Yadan-Barzilai Y. Saker T. Kritchmann-Lupo M. Bloch Y. Increase in referrals of children and adolescents to the psychiatric emergency room is evident only in the second year of the COVID-19 pandemic--Evaluating 9156 visits from 2010 through 2021 in a single psychiatric emergency room Int. J. Environ. Res. Public Health 2022 19 8924 10.3390/ijerph19158924 35897293 17. Erez G. Yakubovich S. Sadeh H. Shoval G. Schoen G. Meiri G. Hertz-Palmor N. Bretler T. Bloch Y. Pediatric psychiatric emergency rooms during COVID-19: A multi-center study BMC Psychiatry 2021 22 828 10.1186/s12888-022-04371-7 36575441 18. Panda P.K. Gupta J. Chowdhury S.R. Kumar R. Meena A.K. Madaan P. Sharawat I.K. Gulati S. Psychological and behavioral impact of lockdown and quarantine measures for COVID-19 pandemic on children, adolescents and caregivers: A systematic review and meta-analysis J. Trop. Pediatr. 2021 67 fmaa122 10.1093/tropej/fmaa122 33367907 19. O'brien M. McNicholas F. The use of telepsychiatry during COVID-19 and beyond Ir. J. Psychol. Med. 2020 37 250 255 10.1017/ipm.2020.54 32434596 20. Hubley S. Lynch S.B. Schneck C. Thomas M. Shore J. Review of key telepsychiatry outcomes World J. Psychiatry 2016 6 269 282 10.5498/wjp.v6.i2.269 27354970 21. Cowan K.E. McKean A.J. Gentry M.T. Hilty D.M. Barriers to use of telepsychiatry: Clinicians as gatekeepers Mayo Clin. Proc. 2019 94 2510 2523 10.1016/j.mayocp.2019.04.018 31806104 22. Schulz T. Long K. Kanhutu K. Bayrak I. Johnson D. Fazio T. Telehealth during the coronavirus disease 2019 pandemic: Rapid expansion of telehealth outpatient use during a pandemic is possible if the program is previously established J. Telemed. Telecare 2022 28 445 451 10.1177/1357633X20942045 32686556 23. Greenhalgh T. Wherton J. Telepsychiatry: Learning from the pandemic Br. J. Psychiatry 2022 220 257 261 10.1192/bjp.2021.224 24. Perera S.R. Gambheera H. Williams S.S. "Telepsychiatry" in the time of COVID-19: Overcoming the challenges Indian J. Psychiatry 2020 62 (Suppl. 3) S391 S394 10.4103/psychiatry.IndianJPsychiatry_848_20 33227045 25. Saunders N.R. Kurdyak P. Stukel T.A. Strauss R. Fu L. Guan J. Fiksenbaum L. Cohen E. Toulany A. Utilization of physician-based mental health care services among children and adolescents before and during the COVID-19 pandemic in Ontario, Canada JAMA Pediatr. 2022 176 e216298 10.1001/jamapediatrics.2021.6298 35129604 26. Ayraler Taner H. Akin Sari B. Kaba D. How did the COVID-19 pandemic affect child and adolescent psychiatry outpatient clinic admissions? A single-center, retrospective study Clin. Child Psychol. Psychiatry 2022 27 824 835 10.1177/13591045221095657 35465735 27. Saban M. Myers V. Shachar T. Miron O. Wilf-Miron R.R. Effect of socioeconomic and ethnic characteristics on COVID-19 infection: The case of the ultra-orthodox and the Arab communities in Israel J. Racial Ethn. Health Disparities 2022 9 581 588 10.1007/s40615-021-00991-z 33686623 28. State Comptroller of Israel Aspects of the Reform to Transfer Insurance Liability in Mental Health Care. 2022 Available online: (accessed on 5 March 2023) 29. Benjamini Y. Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing J. R. Stat. Soc. Ser. B 1995 57 289 300 10.1111/j.2517-6161.1995.tb02031.x 30. Becker L.A. Effect Size (ES). 2000 Available online: (accessed on 5 March 2023) 31. Cohen J. Statistical Power Analysis for the Behavioral Sciences Sage Thousand Oaks, CA, USA 1997 (revised 31 December 1997) 32. Sawilowsky S.S. New effect size rules of thumb J. Mod. Appl. Stat. Methods 2009 8 26 10.22237/jmasm/1257035100 33. Kassambara A. Rstatix: Pipe-Friendly Framework for Basic Statistical Tests. R Package Version 0.7. 0. Computer Software. 2021 Available online: (accessed on 5 March 2023) 34. R Core Team R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. 2022 Available online: (accessed on 5 March 2023) 35. Meherali S. Punjani N. Louie-Poon S. Rahim K.A. Das J.K. Salam R.A. Lassi Z.S. Mental health of children and adolescents amidst COVID-19 and past pandemics: A rapid systematic review Int. J. Environ. Res. Public Health 2021 18 3432 10.3390/ijerph18073432 33810225 36. Fegert J.M. Vitiello B. Plener P.L. Clemens V. Challenges and burden of the Coronavirus 2019 (COVID-19) pandemic for child and adolescent mental health: A narrative review to highlight clinical and research needs in the acute phase and the long return to normality Child Adolesc. Psychiatry Ment. Health 2020 14 20 10.1186/s13034-020-00329-3 32419840 37. De Figueiredo C.S. Sandre P.C. Portugal L.C.L. Mazala-de-Oliveira T. da Silva Chagas L. Raony I. Ferreira E.S. Giestal-De-Araujo E. dos Santos A.A. Bomfim P.O.-S. COVID-19 pandemic impact on children and adolescents' mental health: Biological, environmental, and social factors Prog. Neuro-Psychopharmacol. Biol. Psychiatry 2021 106 110171 10.1016/j.pnpbp.2020.110171 38. Liu J.J. Bao Y. Huang X. Shi J. Lu L. Mental health considerations for children quarantined because of COVID-19 Lancet Child Adolesc. Health 2020 4 347 349 10.1016/S2352-4642(20)30096-1 32224303 39. Imran N. Zeshan M. Pervaiz Z. Mental health considerations for children & adolescents in COVID-19 Pandemic Pak. J. Med Sci. 2020 36 S67 S72 10.12669/pjms.36.COVID19-S4.2759 40. Oosterhoff B. Palmer C.A. Wilson J. Shook N. Adolescents' Motivations to Engage in Social Distancing during the COVID-19 Pandemic: Associations with Mental and Social Health J. Adolesc. Health 2020 67 179 185 Available online: (accessed on 5 March 2023) 10.1016/j.jadohealth.2020.05.004 32487491 41. Salmon G. James A. Cassidy E.L. Javaloyes M.A. Bullying a review: Presentations to an adolescent psychiatric service and within a school for emotionally and behaviourally disturbed children Clin. Child Psychol. Psychiatry 2000 5 563 579 10.1177/1359104500005004010 42. American Academy of Child and Adolescent Psychiatry (AACAP) Committee on Telepsychiatry and AACAP Committee on Quality Issues AACAP OFFICIAL ACTION Clinical Update: Telepsychiatry with Children and Adolescents American Academy of Child and Adolescent Psychiatry (AACAP) Committee on Telepsychiatry and AACAP Committee on Quality Issues J. Am. Acad. Child Adolesc. Psychiatry 2017 56 875 893 28942810 43. Myers K. Nelson E.L. Rabinowitz T. Hilty D. Baker D. Barnwell S.S. Boyce G. Bufka L.F. Cain S. Chui L. American Telemedicine Association Practice Guidelines for Telemental Health with Children and Adolescents Telemed. E-Health 2017 23 779 804 10.1089/tmj.2017.0177 44. Kommu J.V.S. Sharma E. Ramtekkar U. Telepsychiatry for Mental Health Service Delivery to Children and Adolescents Indian J. Psychol Med. 2020 42 (Suppl. 5) 46S 52S 10.1177/0253717620959256 33354063 45. Myers K.M. Roth D.E. Telepsychiatry with Children and Adolescents Lewis's Child Adolesc. Psychiatry 2018 10 885 896 46. Tandon R. COVID-19 and mental health: Preserving humanity, maintaining sanity, and promoting health Asian J. Psychiatr. 2020 51 102256 10.1016/j.ajp.2020.102256 32586624 47. Folk J.B. Schiel M.A. Oblath R. Feuer V. Sharma A. Khan S. Doan B. Kulkarni C. Ramtekkar U. Hawks J. The Transition of Academic Mental Health Clinics to Telehealth during the COVID-19 Pandemic J. Am. Acad. Child Adolesc. Psychiatry 2021 61 277 290.e2 10.1016/j.jaac.2021.06.003 34119633 48. Roesler C. Tele-analysis: The use of media technology in psychotherapy and its impact on the therapeutic relationship J. Anal. Psychol. 2017 62 372 394 10.1111/1468-5922.12317 28504335 49. Titzler I. Saruhanjan K. Berking M. Riper H. Ebert D.D. Barriers and facilitators for the implementation of blended psychotherapy for depression: A qualitative pilot study of therapists' perspective Internet Interv. 2018 12 150 164 10.1016/j.invent.2018.01.002 30135779 50. Bekes V. Aafjes-van Doorn K. Luo X. Prout T.A. Hoffman L. Psychotherapists' Challenges with Online Therapy During COVID-19: Concerns about Connectedness Predict Therapists' Negative View of Online Therapy and Its Perceived Efficacy Over Time Front. Psychol. 2021 12 705699 10.3389/fpsyg.2021.705699 34367030 |
PMC10000554 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12050952 foods-12-00952 Article Rational Food Design Targeting Micronutrient Deficiencies in Adolescents: Nutritional, Acoustic-Mechanical and Sensory Properties of Chickpea-Rice Biscuits Talens Clara Methodology Investigation Data curation Writing - original draft Writing - review & editing 1* Garcia-Fontanals Laura Writing - original draft Writing - review & editing 2 Fabregat Paula Formal analysis Data curation 2 Ibarguen Monica Conceptualization Methodology Investigation Supervision 1 Visioli Francesco Academic Editor 1 AZTI, Food Research, Basque Research and Technology Alliance (BRTA), Parque Tecnologico de Bizkaia, Astondo Bidea, Edificio 609, 48160 Derio-Bizkaia, Spain 2 Basque Culinary Center, Facultad de Ciencias Gastronomicas, Mondragon University, 20009 Donostia-San Sebastian, Spain * Correspondence: [email protected] 23 2 2023 3 2023 12 5 95227 1 2023 17 2 2023 18 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). "Hidden hunger", the deficiency of important mineral micronutrients, affects more than 2 billion people globally. Adolescence is unquestionably a period of nutritional risk, given the high nutritional requirements for growth and development, erratic or capricious diets and the increased consumption of snacks. This study applied the rational food design approach to obtain micronutrient-dense biscuits by combining chickpea and rice flours to achieve an optimal nutritional profile, crunchy texture and appealing flavour. The perception of 33 adolescents regarding the suitability of such biscuits as a mid-morning snack was examined. Four biscuits were formulated, with different ratios of chickpea and rice flours (CF:RF): G100:0, G75:25, G50:50 and G25:75. Nutritional content, baking loss, acoustic-texture and sensory analyses were carried out. On average, the mineral content of biscuits with the CF:RF ratio of 100:0 doubled compared with the 25:75 formula. The dietary reference values for iron, potassium and zinc reached 100% in the biscuits with CF:RF ratios of 50:50, 75:25 and 100:0, respectively. The analysis of mechanical properties revealed that samples G100:0 and G75:25 were harder than the others. Sample G100:0 showed the highest sound pressure level (Smax). Sensory analysis showed that increasing the proportion of CF in the formulation augments the grittiness, hardness, chewiness and crunchiness. Most of the adolescents (72.7%) were habitual snack consumers; 52% awarded scores >= 6 (out of 9) to biscuit G50:50 for its overall quality, 24% described its flavour as "biscuit" and 12% as "nutty". However, 55% of the participants could not pinpoint any dominant flavour. In conclusion, it is possible to design nutrient-dense snacks that meet the micronutrient requirements and sensory expectations of adolescents by combining flours naturally rich in micronutrients. micronutrient deficiencies rational food design adolescents biscuits texture chickpea flour targeted nutrition Centre for the Development of Industrial Technology (CDTI) of the Spanish Ministry of Science and InnovationTECNOMIFOOD project (CER-20191010). This research was financially supported by the Centre for the Development of Industrial Technology (CDTI) of the Spanish Ministry of Science and Innovation under the grant agreement: TECNOMIFOOD project (CER-20191010). This paper is a contribution no 1151 from AZTI, Food Research, Basque Research and Technology Alliance (BRTA). pmc1. Introduction The new WHO European Regional Obesity Report 2022, published on 3 May 2022 by the WHO Regional Office for Europe, reveals that overweight and obesity rates have reached epidemic proportions across the region and are still escalating. None of the 53 member states is on track to meet the WHO Global Noncommunicable Disease (NCD) target of halting the rise of obesity by 2025 . Early studies from several countries in the region indicate that the prevalence of overweight and obesity and high mean body mass index have increased in children and adolescents during the COVID-19 pandemic . Moreover, adolescence is unquestionably a period of nutritional risk given the increased nutritional requirements for growth and development, erratic or capricious diets and the increased consumption of snacks, fast food and refreshing beverages. During this period, deficiencies of specific minerals such as Ca, Fe and Zn and vitamins such as A, D, B6, B12, riboflavin, niacin, thiamine and folic acid are common. Low intake of fibre and complex carbohydrates has also been noted . The AVENA study (Alimentacion y Valoracion del Estado Nutricional en Adolescentes) conducted in 2012 with Spanish adolescents revealed that certain healthy dietary habits (i.e., mid-morning snack, afternoon snack, more than 4 meals per day, adequate eating speed) were associated with low body fat ). Moreover, public health recommendations in paediatric scientific journals suggest that most children should eat between four and six times a day . Snacks can account for up to a third of the daily energy intake. Thus, it is of great interest to the food industry to provide snacks of high nutritional quality. This could be achieved by adding protective factors and avoiding risk factors (for infants and adolescents), so the products might be recommended as meeting nutritional requirements . To improve the nutritional quality of snacks, one should reduce risk factors such as sugar, salt, refined grain flour and energy content. Likewise, the proportion of nutritionally valuable ingredients providing protein, fibre and micronutrients should be increased. Legumes are a staple food in many Asian, African and Mediterranean countries. "Hidden hunger", the deficiency of important mineral micronutrients, affects more than 2 billion people globally . Naturally rich in micronutrients (iron and zinc), legume flours could be used to formulate recipes for nutritionally rich and balanced biscuits. The cereal industry often uses rice as a substitute for wheat. Refined rice flour has nutritional disadvantages as it is high in carbohydrates but low in protein, fibre and micronutrients. However, rice flour could be combined with other ingredients, such as chickpea flour, to develop value-added snacks with a rich and well-balanced nutrient composition . Chickpeas are highly nutritious due to their low lipid content (2.7-6.48%), rich in polyunsaturated fatty acids (66%), and high content of protein (17-22%), starch (52.5%), dietary fibre (18-22%), bioactive compounds and essential vitamins and minerals. These legumes are a good source of calcium, potassium, magnesium, iron, zinc and important vitamins such as riboflavin (B2), niacin (B3), thiamine (B1), folic acid (B9), b-carotene (vitamin A precursor), vitamin E, vitamin C, pantothenic acid (B5) and pyridoxine (B6). Therefore, chickpea products can complement the vitamin pool supplied by other foods . The demand for chickpeas as a functional ingredient in food production is on the increase. Chickpea flour has been incorporated into a variety of goods, such as breads, biscuits, pasta, snacks and dairy, to improve their nutritional value . Goni and Valentin-Gamazo have shown that adding 25% of chickpea flour to wheat pasta decreases the glycaemic index and increases the mineral, fat and resistant starch content. Moreover, the consumers of chickpeas and/or hummus have a higher nutrient intake of dietary fibre, polyunsaturated fatty acids (PUFAs), vitamins A, E and C, folic acid, magnesium, potassium and iron than non-consumers . Importantly, chickpeas have high levels of the micronutrients usually lacking in diets consumed by adolescents. The term Rational Food Design (RFD) has been used in previous research to refer to the design of food products with specific functionalities to satisfy "the needs and desires" of the consumer . The official meaning of the word "rational" is "based on facts and reason", therefore, RFD should rely on any type of scientific and technological knowledge to design food formulas and food structures. Previous studies have applied RFD principles to create high-quality healthy foods by including in the RFD approach imaging technologies, such as atomic force microscopy , or microtechnology for testing ingredient functionality . However, to the best of our knowledge, the inclusion of the dietary reference values (DRV), provided by EFSA, as formula design variables, and acoustic-texture measurements, as response variables, during the RFD process has not been approached yet. The aim of this study is twofold. Firstly, the study will include EFSA's DRV for adolescents from 11 to 14 years, engaged in all types of physical activity, using the RFD approach to obtain micronutrient-dense biscuits by combining chickpea and rice flours, to achieve an optimal nutritional profile, crunchy texture and appealing flavour. Secondly, the perception of adolescents regarding the suitability of such biscuits as a mid-morning snack will be examined. 2. Materials and Methods 2.1. Rational Food Design In this study, we suggest a widening of the use of RFD for targeted nutrition (aimed at the nutritional needs of specific groups). For this purpose, RFD could be applied in two steps of the food development process:1. during the formula design step, which is also part of the process of structure design, when two important decisions are made: (1) the target nutritional composition of the new or improved formula; (2) the ingredients and raw materials used based on their micronutrients composition. 2. during the step preceding the consumer tasting, when a selection of the optimum samples has to be made based on acoustic-textural and sensory analysis. The rationale applied during the formula design was based on EFSA's DRV. For adequate energy distribution, the daily food intake should be supplied in 5 meals (Table 1). The proportions should be as follows: 20% at breakfast, 10-15% at mid-morning meal, 25-30% at lunch, 10-15% at evening snack and 25% at dinner. Eating between meals should be avoided . Many scientific groups and societies recommend distributing energy and nutrients among four to five daily meals to improve health . According to the American Heart Association , the mid-morning snack should be easily digestible and not excessive in calories to sustain the feeling of satiety and reach lunchtime with a sufficient but not unmanageable appetite. Table 2 shows the nutritional requirements for mid-morning snacks based on the European Food Safety Authority (EFSA) dietary reference values (DRV) for the EU . The target population are healthy individuals of both genders, aged 11-14, engaged in all types of physical activity. Therefore, although mid-morning snacks can reduce appetite during lunch, it is crucial to consider their nutritional quality. 2.2. Ingredients Chickpea flour (Don Pedro) (CF) was obtained from Legumbres Pedro S.L. (Cadiz, Spain). The flour contained 15.0 g of moisture, 53.5 g of carbohydrates, 4.4 g of sugar, 20.5 g of protein and 6.6 g of fat (per 100 g). Rice flour Remyflo R 200 T (RF) was purchased from BENEO-Remy N.V. (Leuven, Belgium). The composition per 100 g was 10.0 g of moisture, 81.0 g of carbohydrates, 8 g of proteins and 1 g of fat. Sunflower oil, orange blossom water, sugar and salt were bought at the local supermarket (Makro, Derio, Spain). 2.3. Biscuit-Making Procedure Four biscuits were formulated using different ratios of chickpea flour and rice flour (CF:RF). These ratios were: 100:0, 75:25, 50:50, 25:75. The biscuits were made following the formulations, in which only chickpea and rice flour percentages changed. The rest of the ingredients were kept constant (Table 3). All ingredients were weighed using high-precision (+-0.0001 g) scales AB304-S (Mettler Toledo, Greifensee, Switzerland). The powdered ingredients were mixed using a planetary mixer Sammic BM-5 (Sammic S.L., Azkoitia, Spain), for 30 s at 202 rpm. Then, water was added and the mixture was stirred for 10 s at 260 rpm. Finally, oil was mixed in for 30 s at 202 rpm. Batches of 500 g were produced in triplicate. The dough was removed from the mixer and allowed to rest for 10 min. It was processed further using a sheeting machine Sammic FMI-31 (Sammic S.L., Azkoitia, Spain). The dough sheets were flattened manually with a wooden rolling pin to an approximately 2.5-mm thickness and then cut into circles of 4.5 cm in diameter. The biscuits were baked in an electric oven (MIWE Condo type CO 2 0608, MIWE GmbH, Arnstein, Germany) at 150 degC for 25-35 min and then at 120 degC for 20-30 min. The biscuits were cooled for 30 min on a rack and stored in airtight containers until evaluation. 2.4. Physico-Chemical Analysis Powdered samples of the four biscuit types were analysed according to the ISO standards for moisture (ISO 1442:1997), protein content (ISO 937:1978), crude fat (ISO 1443:1973), fatty acids profile (by chromatography), salt content estimated by analysis of chloride by potentiometric titration with silver nitrate solution, and total sugars (ISO 22184:2021). Total dietary fibre was determined by the AOAC enzymatic gravimetric method 991.43. The carbohydrate content was determined by difference. After proximate analysis, energy values were calculated using the Atwater general factors (4 kcal/g for protein and carbohydrates, 9 kcal/g for fat and 2 kcal/g for fibre). Samples were analysed in triplicate. The mineral composition (calcium, phosphorus, iron, magnesium, manganese, zinc, potassium and copper) was examined following the standard DIN EN 16943:2017-1. The selenium content was obtained following DIN EN 15763:2010-04. Sodium levels were established using gas chromatography-mass spectrometry (LC-MS/MS). The Water activity was measured using a water activity meter (AquaLab PRE, Lab Ferrer, Cervera, Spain). The baking loss (%) was determined on three independent samples using Equation (1), 24 h after baking, where Wf is the weight of the sample after baking and W0 is the weight before baking:% Baking loss = (Wf - W0)/W0 x 100 (1) 2.5. Instrumental Texture Analysis The mechanical and acoustic properties of the biscuits were simultaneously measured employing a Texture Analyser (TA.HDplus, Stable Micro System Ltd., Surrey, UK) and a microphone in combination with a Deltatron preamplifier (Bruel Kjaer, Naerum, Denmark). A 30-kg load cell was used and the key parameters were extracted employing the Exponent software (v.6.1.16.0, Stable Micro System Ltd.). Each biscuit was placed upside down on a perforated surface. The samples were penetrated using a cylinder probe of 4 mm in diameter at a test speed of 1 mm/s. The distance was 10 mm, and the trigger force was 5 g, based on the procedure of da Quinta, Alvarez-Sabatel . Ten replicates were used for each biscuit type. During the test, the acoustic emission was registered by a microphone calibrated using a calibrator type 4231 ( 114-dB sound pressure level [SPL], 1000 Hz) (Bruel Kjaer). The distance of the microphone to the sample was 10 mm with an angle of 45deg . The filter function of the preamplifier screened out the background noise. No simultaneous activities were carried out in the laboratory to avoid noise interference. The mechanical properties of the samples were determined using the following parameters: the maximum force required to break (N) as a direct measure of the hardness and the probe distance needed to break the sample (mm) as an indicator of the fracturability . The crispness attribute was associated with several acoustic and mechanical parameters . These were the maximum loudness perceived during the break evaluated as sound pressure level (SPL) in dB, number of sound peaks (peaks higher than 1 dB), number of force peaks (peaks higher than 0.2 N) and the linear distance (N*s). This last parameter, used as an indicator of jaggedness, was calculated as the length of an imaginary line joining all force peaks in a force-time graph . 2.6. Sensory Evaluation by Trained Assessors A panel of seven assessors trained in quantitative descriptive analysis (QDA) of biscuits evaluated the samples. The panel was selected and trained following the ISO 8586:2012 procedure. Six 1-h sessions were performed for descriptor development, definition and training. Training sessions utilised reference standards. The attribute values were recorded using a 5-point scale. Finally, the four samples were evaluated in duplicate. The samples were labelled with a random three-digit numeric code and presented monadically in a randomised and balanced order. Still water was served for palate cleansing between samples. The sensory evaluations were carried out at the Sensory Science Laboratory in AZTI (Derio, Spain) in individual booths designed in accordance with ISO 8589:2007. 2.7. Consumer Tasting The sample consumer population was selected from visiting students from a local secondary school (11-12 years of age). An informed consent form was sent to the parents before the tasting session, indicating the objective of the study and listing the allergens. Thirty-three students participated; 33% were female, and 67% were male. The objective of the study was (i) to carry out a sensory evaluation of the newly developed biscuit and (ii) to establish whether they could identify chickpea flavour. The group assessed only one sample. The selection of the sample was based on the results of the texture measurements and trained panel sensory analysis. The participants assessed the samples in individual cabins illuminated by fluorescent lamps. The sample was served in individual plates. Participants were asked to record their liking intensity scores for overall appearance, overall colour, overall aroma, overall texture and overall flavour (see Supplementary Material). A 9-point hedonic scale was used (9 = like extremely and 1 = dislike extremely). The overall liking was also assessed, together with the open question, "Are you able to identify a particular flavour?". Finally, a questionnaire with two consumption habit questions was presented. The first question was, "Do you consume snacks?". The second was a multiple-choice question, "How often do you eat biscuits?". The options were: twice a week or more often, once a week, twice a month, once a month, once every 2-3 months or less frequently. 2.8. Statistical Analysis The data were processed statistically using the software package XLSTAT 2019.1.2 (Addinsoft, Boston, MA, USA). Analysis of variance (ANOVA) and Tukey's HSD test for comparison of sample means were used to identify nutritional properties and instrumental texture parameters that significantly differed between the samples. All data were expressed as means +- standard deviation (SD). The average sensory configuration obtained by the panel is displayed (as for Principal Component Analysis (PCA)) on a score plot representing the inter-product sensory distances. Descriptive analysis was performed using the liking data recorded by consumers (the overall liking data and the individual attribute liking data reported on the hedonic 9-point scale). 3. Results Results showed that, by using combinations of chickpea and rice flours, it was possible to apply a rational food design to create nutrient-dense and sensory-appealing biscuit-like structures with an improved nutritional profile (targeting EFSA's DRV for adolescents 11 to 14 years, engaged in all types of physical activity), compared to the average profile of 49 plain biscuits found in the Mintel database. 3.1. Nutritional Composition The mean values for the nutritional properties of the four different biscuits can be found in Table 4. The protein content varied from 7.71% to 13.92%, with the sample G100:0 showing the highest value and the G25:75 the lowest. Samples G75:50 and G50:50 did not significantly differ in their protein content (10.54% and 9.89%, respectively). These results indicate that the relative amount of protein in the biscuits increases with increasing chickpea flour (CF) content. This is not surprising, as rice flour (RF) contains less protein than CF. A similar trend was observed for the fat content, with values ranging between 19.92% and 21.15%; however, the differences between samples were smaller and not significant. These results reflect the lower fat content of RF compared to CF. The carbohydrate content ranged from 49.63% to 62.27%; the sample G25:75 had the highest level of carbohydrates, which was the lowest in sample G100:0 (the difference was significant). Samples G75:25 and G50:50 had intermediate profiles with 56.15% and 56.56% carbohydrate content, respectively. The carbohydrate content of the biscuits increases with rising RF content, which is consistent with the higher carbohydrate levels of RF compared to CF. However, a tendency for the sugar and fibre content to increase was observed for biscuits with larger amounts of CF. Sample G100:0 contained the most sugar and fibre (18.84% and 7.07%, respectively) and sample G25:75, the smallest amounts (11.60% and 2.59%). These differences were statistically significant. Thus, the sugar and fibre concentrations rise with increasing CF content. These observations are consistent with the differences in nutritional characteristics of these two flours . No significant differences were seen between the salt contents of the different formulations, indicating that changes in the flour ratio did not affect this nutrient. The data on the mineral content of the biscuits showed that the concentrations (in mg/100 g) of iron (1.26-3.39), potassium (233.74-576.70), zinc (1.00-1.99), phosphorus (111.67-209.73), magnesium (54.12-113.45), copper (0.25-0.62), manganese (0.93-1.10) and calcium (14.23-29.00) rise significantly with increasing CF content. On average, the mineral content increased by 2-fold when increasing the CF:RF ratio from 25:75 to 100:0. For sodium, no significant differences were observed. For selenium, the concentration increases significantly with rising RF content, varying from 5.83 mg/100 g for the sample G100:0 to 9.08 mg/100 g for the G25:75. These results are consistent with the mineral content of chickpea and rice flours . The energy content did not differ significantly between the samples. The energy values per 100 g ranged from 471.76 kcal for the sample G100:0 to 457.15 kcal for the G25:75. There is a tendency for the energy content to increase as the proportion of added CF increases. 3.2. Instrumental Texture The results of the instrumental texture analysis, as well as the baking loss, moisture content and water activity of the four different biscuits can be found in Table 5. Samples G100:0, G75:25 and G25:75 did not show significant differences between their percentages of baking loss (29.26%, 29.21% and 28.89%, respectively). The sample G50:50 showed greater baking loss (30.07%) than G25:75. The moisture content of samples G100:0, G75:25 and G50:50 did not differ significantly (3.96%, 3.59% and 3.95%, respectively); the G25:75 showed increased moisture content (6.50%). The highest water activity (aw) was observed for the sample G25:75 (0.47), followed by G100:0 (0.31), G50:50 (0.29) and G75:25 (0.27). Therefore, when the relative RF content was greater than 50%, the moisture content and aw of the biscuits increased. Changing the flour ratio did not affect the baking loss. The analysis of mechanical properties revealed greater maximum force (Fmax) for the samples G100:0 and G75:25 (18.13 N and 19.68 N) than for G50:50 and G25:75 (11.68 N and 10.97 N, respectively). When the mixture contained more CF than RF, the hardness of the biscuit increased. However, all the samples showed similar distances at break (ranging from 0.83 mm to 0.70 mm) with no significant differences; they all had similar fracturability (or fragility) despite the differences in hardness. Sample G100:0 showed a significantly greater number of force peaks (NFP) (9.80) than samples G75:25, G50:50 and G25:75 (1.17, 2.60 and 2.20, respectively; differences not significant). Accordingly, sample G100:0 probably suffered more breaking events. Moreover, the sample G100:0 showed greater linear force peak distance (LDF) (94.76 N*s) than the other samples and sample G75:25 had significantly greater linear distance (81.35 N*s) than G50:50 (68.48 N*s). This could indicate that, as the CF content increased, the LDF also tended to increase. The sample G25:75 showed an LDF of 76.40 N*s, not significantly different from G75:25 and G50:50, like the NFP. This result might reflect the relationship between the NFP and LDF; the more fluctuations in force, the longer the line joining the force points. The highest value of maximum sound pressure level (Smax) was obtained for sample G100:0 (75.74 dB), significantly higher than for samples G75:25 and G25:75 (65.36 dB and 67.99 dB). In contrast, no significant differences were detected between the numbers of sound peaks (NSP) for the four different biscuits. Since the parameters LDF, NFP, NSP and Smax are the indicators of the crispness of a product , these results suggest that, when the biscuit is made with 100% chickpea flour, its crispness will be significantly higher than when the mixture of the two flours is used. Figure 1 shows force-time and sound-time curves obtained in the texture and acoustic event analysis for each biscuit. The graph for G100:0 (as an example) shows a force increase region, starting from the first contact between the probe and the biscuit until the first major drop in the force (at around 1 s). Within this region, the compression force increased almost linearly with the displacement, while acoustically it was very quiet. This suggests that the biscuit undergoes deformation but no major structural damage. The acoustic signals recorded in this linear domain were not considered, as they resulted from surface contact between the biscuit and the probe. Then, the compression force became jagged and many acoustic events were recorded within a very short period. This is when the structural breakdown occurs, at the first crack. There was no further increase in the compression force after that break, but rises and falls in the measured force could be observed. These force reductions reflect the ongoing minor structural fracture in the biscuit. At around 2 s, a sharp drop in the compression force occurred, which corresponded to the major structural breakdown. After this significant event, the biscuit remained on the test platform and it took another push for the fractured biscuit to finally fall to the texturometer base. At this point, the compression force reached zero. The acoustic signals recorded after this major breakdown were not considered . We focused on acoustic signals in the jagged region of the force-time curve (between 1 and 2 s) , where the biscuit breaking occurred. A group of acoustic events can be observed for each major force drop. Since these events did not gradually decrease in intensity and had no periodic pattern, they were probably not due to sound echoes or resonances. It is most likely that they reflected a series of structural element fracture events captured within a major force peak. It can be hypothesised that the energy dissipated from the biscuit break will spread out, probably in the form of sound. Therefore, the Smax should correspond to the energy released by the major structural breakdown . Figure 1 shows that, for each drop in the compression force (force peak), an acoustic signal (sound peak) was detected, demonstrating the links between the acoustic events and the decrease in the force of a single break event (corresponding to the dissipated energy from the break). A correlation between the major structural breakdown (Fmax) and the Smax can also be observed. 3.3. Sensory Evaluation and Consumer Liking Sensory Profile of Biscuits The panel generated seven descriptors to describe the biscuits; five referred to the texture (hardness, grittiness, fragility, prickliness and chewiness), and two to acoustic sensations (crispness and crunchiness) (Table 6). Figure 2 presents the results of the PCA of the data generated by the sensory panel for the four biscuit formulations. Axis F1 explained 49.09% of the sensory variation between the biscuits and Axis F2, 42.16%. The results indicated that the attributes "grittiness", "fragility", "hardness", "chewiness" and "crunchiness" had discriminative power (ordered from the largest to the smallest, p-values < 0.01). However, the "crispiness" and "prickliness" attributes could not be used to distinguish the samples from each other because they did not have discriminatory power (p-values of 0.13, 0.43, respectively) and, consequently, did not appear in the PCA analysis . Figure 2 shows the vector (red line) corresponding to each attribute and the four biscuit samples (blue points). In the PCA graph, the samples close to each other have similar sensory profiles, and larger distances indicate increased sensory differences. The evaluated sensory attributes are represented by vectors. The vector resultants help to characterise the samples: the higher the resultant on an axis, the higher the discriminating power of the attribute. Table 7 presents the results of assessing the samples included in the PCA. The sample G100:0 differed from all the others; it showed the lowest fragility and the highest crunchiness and chewiness values. The G50:50 was the least crunchy and chewy and the most fragile. The G100:00 and G50:50 samples showed similar degrees of grittiness and hardness, greater than G75:25 and G25:75. In contrast, the sample G25:75 stood out from the others by showing the lowest grittiness and hardness. The sample G75:25 showed an intermediate sensory profile for grittiness, hardness, crunchiness, chewiness, and fragility, even though it had the crispiest texture. The fragility, crunchiness, and chewiness of G25:75 and G75:25 biscuits were very similar (no significant differences). Their fragility was lower than for G50:50 and greater than for G100:0 samples. The crunchiness and chewiness of G25:75 and G75:25 were lower than for the G100:00 sample and greater than for G50:50. The samples G100:00 and G25:75 were very far from each other in the graph , indicating large sensory differences. Moreover, the separation between these samples occurs on Axis F2. This means that the differences between them were explained by the attributes whose resultant vector was located on this axis (grittiness, hardness and chewiness). Therefore, it can be concluded that changing the amount of CF in the formulation directly affects the texture of the biscuit, particularly its grittiness, hardness and chewiness. The grittiness, hardness and chewiness of the biscuit increase along with the growing amount of CF. Samples G75:25 and G50:50 were closer with respect to Axis F2, so their grittiness, hardness and chewiness were similar. These results were consistent with the fact that the G75:25 and G50:50 biscuits only differed by 25% of the CF, while G100:0 had 75% more CF than G25:75. As the samples with larger amounts of CF (G100:0 and G75:25) tended to be less fragile than the others (G50:50 and G25:75), we can conclude that reducing the amount of CF increases the fragility of the biscuits. A greater crunchiness was observed as the ratio CF:RF increases from 50:50 to 100:0. 3.4. Consumer Tasting Only one sample was assessed by the adolescents participating in the study, to avoid peer pressure when comparing samples or preferences. The selection of the sample was based on the results of the acoustic-texture measurements and the sensory analysis with the trained panel. The G50:50 had the best texture for children aged between 10 and 12. It was the easiest to chew (the least effort required to chew the biscuit before swallowing) and the most fragile (the least force needed to break it into pieces) (see Table 5 and Table 7). Considering the results of texture and sensory analysis, the biscuits with a higher proportion of CF (G100:0 and G75:25) could be too hard and difficult to chew and the biscuit G25:75 contained too much moisture. Figure 3 shows the percentages for liking intensity scores (9-point hedonic scale) for overall appearance, colour, aroma, texture and flavour. The sensory attributes of overall appearance, overall colour, overall aroma and overall texture of the chickpea biscuit were evaluated positively (score >= 6) by 97%, 91%, 76% and 52% of the consumers, respectively. Only the overall flavour attribute was evaluated negatively (score <= 4) by 55% of the consumers. Regarding the overall liking results, 52% of the adolescents gave a positive sensory evaluation of the overall quality of the chickpea biscuit (sum of choice percentages with scores >= 6). However, 33% of participants gave this a negative assessment (sum of choice percentages with scores <= 4), stating that the texture seemed a bit too hard and that the biscuit did not have much flavour. These results indicate that the overall liking of the biscuit was negatively affected by unappealing flavour and texture. Adolescents participating in this study were not able to identify the main flavour of the biscuits. Some (24%) described the chickpea biscuit flavour as "biscuit flavour", and 12% of the participants perceived a "nutty flavour". However, 55% were unable to pinpoint any dominant flavour, answering "no", "I do not know" or "I cannot", and 9% described other flavours. The answers to the two consumption habit questions (the frequency of snack consumption) revealed that 72.7% of the participants consume snacks, most frequently in the form of biscuit and breadsticks. Biscuits were eaten by 14.3% of the consumers several times a week, 9.5% once a week and 14.3% once a fortnight. The remaining 61.8% indicated that they ate biscuits once a month or less frequently. 4. Discussion Table 8 shows the average nutritional content of 49 different plain biscuits, obtained by searching the Mintel Database (October 2022). Based on the average serving size from the Mintel search (Table 8), a standard 30 g serving size was assumed for the biscuits as mid-morning snack. It was possible to reach the percentages of the reference daily intake (indicated as AR or AI) of micronutrients recommended by EFSA, for adolescents from 11 to 14 years, engaged in all types of physical activity , indicated in Table 9. For none of the four biscuits did the mineral content contained in one serving exceed the Tolerable Upper Intake Level (UL). For all the biscuits, a part of the reference daily intake of the analyzed minerals is covered with the consumption of one serving of biscuits (30 g). Taking into account that mid-morning meal should represent 15% of the daily food intake, and assuming that mid-morning meal should cover approximately 15% of the reference daily intake of minerals, the biscuits should be accompanied by another food that provides minerals to complete this 15%. Among the top 20 ingredients found in the search , 93% of the biscuits contained flavourings and salt and 86%, raising agents and emulsifiers. Wheat flour and white sugar were present in 71% of the samples. Between 43% and 57% of the biscuits were supplemented with vitamin B6, riboflavin, vitamin B1, niacin, iron, vitamin A, vitamin D or other vitamins and minerals. Compared to the commercial plain biscuits found in the Mintel search , all the chickpea biscuit samples contained less carbohydrates, sugar and salt. However, they had a higher protein and fat (mainly unsaturated) content. The fibre content was augmented in all samples except for G25:75, which contained the lowest amount of CF (the main fibre source). Moreover, chickpea biscuits were rich in potassium, calcium, phosphorus, magnesium, iron, zinc, copper and manganese, and these minerals were naturally present in the ingredients (unlike the additives used in commercial biscuits). As can be seen in Figure 4, most of the vitamins and minerals in plain biscuits targeted at children and sold in Spain between 2017 and 2022 are added to the formulation rather than naturally present in the main ingredients. The moisture content of the biscuits was similar to the moisture levels reported in other studies of protein-enriched biscuits with CF . The only exception was the sample G25:75, in which the industry standard for moisture content in biscuits (1-5%) was exceeded. This high moisture content could be due to the large proportion of starch (approximately 80%) in RF, which might have increased water retention. Rababah and Al-Mahasneh replaced some wheat flour in biscuits using CF at 3%, 6%, 9% and 12%. For the biscuits enriched with 12% CF, they reported an increase in protein content (from 16.82% for the control to 19.64%) and fat content (from 14.13% to 15.31%). This was to be expected, as the CF contains more protein and fat than the wheat flour. These data are consistent with the current study, where the protein and fat content of the chickpea biscuit increased proportionally to the amounts of CF added. As the chickpea-to-wheat flour ratios increased, the values for most of the liking attributes (overall impression, overall flavour and overall colour) decreased and the hardness of the biscuits increased. As a result, the fortification ratio of 3% gave the best sensory results in descriptive analysis. In contrast, Yadav and Yadav reported a decrease in biscuit fat content with an increasing degree of wheat substitution with CF and plantain flour. However, this result seemed to be due to the low oil-holding capacity of these flours compared to wheat flour. Moreover, the authors reported an increase in protein content from 7.1% for the 100%-wheat biscuit to 9.2% for the 40%-chickpea biscuit (probably caused by the higher protein content of CF). The team also reported an increase in the amount of fibre in chickpea-enriched biscuits (again, most likely due to the high levels of fibre in CF). An increase in fracture strength, and therefore in hardness, of biscuits with the addition of plantain and CF was also observed (the highest at 40% substitution). This is in agreement with the results of the present study. Mancebo and Rodriguez added pea protein (up to 20%) to a rice flour biscuit. They found that incorporating this protein decreases the biscuit hardness compared to 100%-rice flour biscuits. Dapcevic et al. replaced 10-30% of RF with buckwheat flour, which contains twice the protein of RF and less starch. They found that increasing the relative amounts of buckwheat decreased biscuit hardness and fracturability. Similarly, Gerzhova and Mondor have demonstrated that adding canola protein to 80% rice and 20% buckwheat flour biscuits decreased hardness and increased thickness of the biscuit. Sarabhai and Prabhasankar have reported that adding soy and whey protein to a rice flour biscuit reduced its breaking strength (and, therefore, hardness). The current report does not wholly concur with these studies. Our texture and sensorial results showed that the biscuit hardness, crunchiness and chewiness increased with the rising proportion of CF (and, therefore, protein) added to the formula. This inconsistency may be due to the much higher CF percentages used here in comparison with the studies mentioned above. In those studies, the maximum RF replacement was 30% and the maximum protein concentrate addition was 20%. The Fmax value was associated with the sensory attributes of hardness, grittiness, chewiness and fragility. Samples G100:0 and G75:25 had higher Fmax values and were considered harder, grittier and more chewy and less fragile than the sample G25:75. This is in agreement with the results of Segnini and Dejmek , in whose study the fracture force for a potato chip seemed to be a good predictor of the sensory texture attributes such as hardness, chewiness, crispness (evaluated as crunchiness) and tenderness. The results of this study suggest that crunchiness was positively associated with the acoustic parameter Smax and the mechanical parameters Fmax, LDF and NFP; as the CF content of the biscuit increased, the values of Smax, Fmax, LDF, NFP and crunchiness tended to rise. No association between texture and acoustic parameters and the sensory attribute of crispiness was detected. This is in disagreement with the results of da Quinta and Alvarez-Sabatel , Gouyo and Mestres and Salvador and Varela . Their studies have reported that the LDF, NFP, NSP and Smax are positively correlated with the crispness of a product. Considering these results, one might expect that the crispness of the 100% chickpea biscuit would be significantly higher than for a product combining the CF with RF. However, the sensory panel did not detect this effect. Fillion and Kilcast have studied the perception of crispness and crunchiness in fruits and vegetables. They concluded that loudness was not considered when qualifying a product as crunchy or crispy, but it was used to assess the intensity of crunchiness or crispiness. The two attributes involve different frequencies of sound, a low frequency for crunchiness and a high frequency for crispiness. Furthermore, there was no correlation between hardness and crispiness when the hardness was very high. This suggested that a very hard texture could not be perceived as crispy and would be described as crunchy. According to that study, Smax and NSP could not be used to qualify a product as crispy or crunchy since these parameters do not reflect the frequency of the sound but its loudness. Therefore, these acoustic parameters could be related to both crispiness and crunchiness, depending on the product being evaluated. This might be the reason why the very hard chickpea biscuits were not perceived as crispy (with a minimum score of 0 and a maximum of 2 on a scale from 1 to 5) but rather as crunchy (with a minimum score of 2 and a maximum of 4 on a scale from 1 to 5). In this work, our approach has been to bring other types of knowledge into the RFD approach for targeted nutrition: the scientific knowledge, provided by EFSA's DRV, and acoustic-texture measurements to apply RFD targeted to adolescents aged 11-14 years, engaged in all types of physical activity. Increasing the addition of chickpea flour in a rice biscuit improves its nutritional pro-file, increasing by 2-fold the protein and the mineral content by increasing the CF:RF ratio from 25:75 to 100:0. However, these increase causes a rise in biscuit hardness, grittiness, chewiness and crunchiness, according to texture and sensorial analysis. 5. Conclusions Efforts to develop targeted nutrition strategies for adolescents from 11-14 years should be scrutinized beyond the nutrient contents of food, and include EFSA's DRV, as well as the acoustic-texture effect, in the case of biscuits. Increasing the addition of chickpea flour in a rice biscuit improves its nutritional profile, increasing by 2-fold the protein and the mineral content by increasing the CF:RF ratio from 25:75 to 100:0. However, these increase causes a rise in biscuit hardness, grittiness, chewiness and crunchiness, according to texture and sensorial analysis. A total of 33 adolescents participated in the study, of whom 52% scored as >=6 the overall quality of the chickpea biscuit, which was negatively affected by unappealing flavour and texture. This study shows that rational food design is a promising approach to design nutrient-dense and sensory-appealing microstructures aimed at adolescents by combining flours naturally rich in micronutrients. However, the authors acknowledge the limitations of this research in that future work should include the antinutritional factors of the plant-based ingredients used. In addition, including other disciplines, such as imaging technologies, digestion, nutrition and physiological responses, could bring multidimensional perspectives in the design of food matrices that are sensory-appealing and targeted to adolescents. Acknowledgments The authors would like to thank Jordi Talens for his help during photo editing. Supplementary Materials The following supporting information can be downloaded at: Product tasting with scholars (11-12 y). Chickpea biscuits. Click here for additional data file. Author Contributions C.T.: methodology, investigation, data curation, writing: original draft, writing: review & editing; L.G.-F.; writing: original draft, writing: review & editing; P.F.: formal analysis, data curation; M.I.: conceptualisation, methodology, investigation, supervision. All authors have read and agreed to the published version of the manuscript. Data Availability Statement The data that support the findings of this study are available as supplementary material upon request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Force-time and sound-time curves for the four chickpea biscuit formulations. Each graph corresponds to a single texture analysis measurement performed for each biscuit. It shows the force and acoustic pressure level (SPL) during the breaking of the biscuit. G100:0 = 100% CF/0% RF; G75:25 = 75% CF/25% RF; G50:50 = 50% CF/50% RF; G25:75 = 25% CF/75% RF. Figure 2 Principal Component Analysis (PCA) of the four chickpea biscuits formulations. G100:0 = 100% CF/0% RF; G75:25 = 75% CF/25% RF; G50:50 = 50% CF/50% RF; G25:75 = 25% CF/75% RF. Figure 3 Percentages (values in coloured bars) for liking intensity scores (9-point hedonic scale) for overall appearance, colour, aroma, texture and flavour obtained from consumer tasting. Figure 4 Top 20 (Food) ingredients declared in 49 plain biscuits targeted at children aged 5-13 years, sold in Spain between 2017 and 2022. Data found in Mintel GNPD Database (October 2022). foods-12-00952-t001_Table 1 Table 1 Recommended energy (% of calories) distribution per meal for adolescents. Meal Energy Requirement (Minimum) Energy Requirement (Maximum) Breakfast 20% 20% Mid-morning snack 10% 15% Lunch 25% 35% Afternoon snack 10% 15% Dinner 25% 25% TOTALTotal 90% 110% foods-12-00952-t002_Table 2 Table 2 Nutritional requirement for a mid-morning snack (15%) based on EFSA DRV guidelines. Energy Requirement (Per Day) 2.290 kcal Energy (kcal) mid-morning snack, 15% of 2.290 kcal 343.49 kcal Fat 128.63 kcal Saturated fats 36.75 kcal Proteins 23.69 kcal Carbohydrates 185.49 kcal Sugars 34.35 kcal Fibre 5.70 kcal Salt (sodium chloride) 0.47 g foods-12-00952-t003_Table 3 Table 3 Formulations for the 4 types of chickpea biscuits studied. Ingredient G100:0 G75:25 G50:50 G25:75 Chickpea flour 53.3% 40.0% 26.7% 13.3% Rice flour 0.0% 13.3% 26.7% 40.0% Water 23.5% 23.5% 23.5% 23.5% Sunflower oil 13.3% 13.3% 13.3% 13.3% Orange blossom water 2.4% 2.4% 2.4% 2.4% Sugar 7.2% 7.2% 7.2% 7.2% Salt 0.3% 0.3% 0.3% 0.3% Total 100.0% 100.0% 100.0% 100.0% foods-12-00952-t004_Table 4 Table 4 Nutritional composition of the four biscuits (mean value of 3 measurements and standard deviation are shown). G100:0 = 100% CF/0% RF; G75:25 = 75% CF/25% RF; G50:50 = 50% CF/50% RF; G25:75 = 25% CF/75% RF. Nutrient G100:0 G75:25 G50:50 G25:75 Energy (kJ/100 g) 1966.19 +- 5.16 a 1945.40 +- 6.87 b 1923.71 +- 6.80 c 1928.94 +- 2.52 bc Energy (Kcal/100 g) 471.76 +- 5.16 a 465.92 +- 6.87 a 461.60 +- 6.80 a 457.15 +- 2.52 a Protein (g/100 g) 13.92 +- 0.81 a 10.54 +- 0.31 b 9.89 +- 0.44 b 7.71 +- 0.59 c Fat (g/100 g) 21.15 +- 2.39 a 20.25 +- 0.88 a 20.09 +- 0.33 a 19.92 +- 1.26 a Saturated (g/100 g) 2.76 +- 1.01 a 2.20 +- 0.08 a 2.38 +- 0.21 a 2.10 +- 0.21 a Monounsaturated (g/100 g) 8.60 +- 0.55 a 9.05 +- 0.99 a 10.59 +- 2.36 a 8.13 +- 8.13 a Polyunsaturated (g/100 g) 8.39 +- 0.26 a 8.88 +- 0.85 a 8.32 +- 0.65 a 8.05 +- 1.06 a Carbohydrates (g/100 g) 49.63 +- 4.19 b 56.15 +- 2.70 ab 56.56 +- 0.39 a 62.27 +- 0.97 a Sugars (g/100 g) 18.84 +- 4.45 a 13.64 +- 0.97 ab 12.61 +- 2.34 ab 11.60 +- 0.97 b Fibre (g/100 g) 7.07 +- 1.55 a 4.82 +- 0.45 ab 4.84 +- 0.93 ab 2.59 +- 0.51 b Salt (g/100 g) 0.43 +- 0.077 a 0.42 +- 0.071 a 0.36 +- 0.094 a 0.44 +- 0.071 a Sodium (mg/100 g) 166.89 +- 32.42 a 176.03 +- 50.42 a 154.38 +- 27.55 a 151.08 +- 30.34 a Potassium (mg/100 g) 576.70 +- 20.10 a 397.22 +- 2.22 b 313.64 +- 5.11 c 233.74 +- 7.52 d Calcium (mg/100 g) 29.00 +- 0.77 a 24.03 +- 1.21 b 18.99 +- 0.71 c 14.23 +- 0.72 d Phosphorus (mg/100 g) 209.73 +- 8.93 a 166.29 +- 4.61 c 148.51 +- 2.28 b 111.67 +- 11.10 b Magnesium (mg/100 g) 113.45 +- 2.35 a 77.44 +- 9.50 b 67.95 +- 8.40 bc 54.12 +- 2.87 c Iron (mg/100 g) 3.39 +- 0.10 a 2.37 +- 0.096 b 1.77 +- 0.050 c 1.26 +- 0.037 d Zinc (mg/100 g) 1.99 +- 0.059 a 1.42 +- 0.056 b 1.23 +- 0.023 c 1.00 +- 0.027 d Copper (mg/100 g) 0.62 +- 0.079 a 0.45 +- 0.008 b 0.35 +- 0.020 c 0.25 +- 0.020 d Manganese (mg/100 g) 1.10 +- 0.01 a 1.01 +- 0.01 b 0.97 +- 0.014 c 0.93 +- 0.00 d Selenium (mg/100 g) 5.83 +- 0.09 c 7.56 +- 0.27 b 8.00 +- 0.25 b 9.08 +- 0.02 a abc Means with different letters in the same row are significantly different according to Tukey's test at p < 0.05. foods-12-00952-t005_Table 5 Table 5 Mechanical and acoustic properties, baking loss, moisture content and water activity (aw) of the four chickpea biscuits (mean value of 3 measurements and standard deviation are shown). Parameters for mechanical properties: maximum force required to break (N), distance at break (mm), number of force peaks (peaks higher than 0.2 N) and linear distance (N*s). Parameters for acoustic properties: maximum sound pressure level (dB) and number of sound peaks. G100:0 = 100% CF/0% RF; G75:25 = 75% CF/25% RF; G50:50 = 50% CF/50% RF; G25:75 = 25% CF/75% RF. Sample G100:0 G75:25 G50:50 G25:75 Baking loss (%) 29.26 ab +- 0.88 29.21 ab +- 0.90 30.07 a +- 1.27 28.89 b +- 0.79 Moisture content (%) 3.96 b +- 0.30 3.59 b +- 0.25 3.95 b +- 0.28 6.50 a +- 0.40 aw 0.31 b +- 0.01 0.27 c +- 0.02 0.29 bc +- 0.01 0.47 a +- 0.03 Maximum force (N) 18.13 a +- 4.83 19.68 a +- 3.49 11.68 b +- 2.69 10.97 b +- 1.11 Distance at break (mm) 0.74 a +- 0.26 0.83 a +- 0.21 0.71 a +- 0.13 0.70 a +- 0.10 Number of force peaks 9.8 a +- 6.14 1.17 b +- 0.41 2.6 b +- 0.89 2.2 b +- 1.10 Linear distance (N*s) 94.76 a +- 13.28 81.35 b +- 7.19 68.48 c +- 4.72 76.40 bc +- 2.88 Maximum SPL (dB) 75.74 a +- 6.11 65.36 b +- 3.26 69.62 ab +- 6.79 67.99 b +- 5.34 Number of sound peaks 621.60 a +- 39.14 610.17 a +- 32.77 586.40 a +- 37.36 625.40 a +- 21.35 Means with different letters in the same row are significantly different according to Tukey's test at p < 0.05. foods-12-00952-t006_Table 6 Table 6 Descriptors generated during the training of the sensory panel. Attributes Definition Technique Used Hardness The force required to deform the product or to penetrate the product with a tool (e.g., a knife). Place the sample between the incisors or between the tongue and the roof of your mouth. Crispness Sound with numerous acoustic events emitted by the product while chewing. Place the sample between the incisors and evaluate the intensity of the sound during the first bite and while chewing (e.g., potato crisps). Crunchiness The sound produced with molars while chewing the product. Place the product between the molars and evaluate the intensity of the sound emitted (e.g., nuts). Grittiness Geometric property related to the perception of the product particle size and shape. Place the sample in the mouth and evaluate the thickness/size of the particles in the sample. The grittier (sand/dust-like particles), the higher the grade. Fragility Force required to break the sample into pieces. Chew the sample and evaluate the force required to break it into pieces. Prickliness/penetration Perception of angular particles. They do not cause damage, but the edges are perceived. Chew the sample and evaluate the geometrical shape of the particles. The sharper the particles, the higher the value on the grading scale. Chewiness A mechanical attribute of texture related to the effort required to chew a solid product until it is ready to be swallowed. The number of bites needed to reduce the sample to a ready-to-swallow state. The more bites, the higher the grade value. foods-12-00952-t007_Table 7 Table 7 Adjusted means for each sample-attribute combination from the analysis of variance models. Letter "A" indicates values significantly higher than the global mean and letter "B" indicates values significantly smaller than the global mean. G100:0 = 100% CF/0% RF; G75:25 = 75% CF/25% RF; G50:50 = 50% CF/50% RF; G25:75 = 25% CF/75% RF. Prickliness Crunchiness Chewiness Crispiness Grittiness Hardness Fragility G100:0 0.6 3.4 A 3.4 A 0.6 3.4 A 3.8 A 2.4 B G75:25 0.6 2.6 3.2 1.2 A 3.2 3.4 2.6 G50:50 0.4 2.4 B 2.4 B 0.6 3.4 A 3.8 A 3.6 A G25:75 0.6 3.2 2.6 0.6 2.4 B 3.0 B 2.8 Global mean 0.55 2.9 2.9 0.75 3.1 3.5 2.85 foods-12-00952-t008_Table 8 Table 8 The average nutritional content of 49 different plain biscuits targeted at children (aged 5-12) sold in Spain between 2017 and 2022. Data found in Mintel GNPD Database (October 2022). Nutrient (n = 49) Content Energy (kcal/100 g) 455.5 Energy (kJ/100 g) 1913.1 Carbohydrates (g/100 g) 70.3 Sugars (g/100 g) 21.6 Fat (g/100 g) 16.0 Fibre (g/100 g) 3.2 Protein (g/100 g) 6.0 Salt (g/100 g) 0.9 foods-12-00952-t009_Table 9 Table 9 Micronutrient content per serving size (30 g of biscuit), % of AR covered by a serving size and DRVs for micronutrients recommended by EFSA, for adolescents from 11 to 14 years, engaged in all types of physical activity. The DRVs indicated are the UL (Tolerable Upper Intake Level), which is "the maximum level of total chronic intake of a nutrient from all sources judged to be unlikely to pose a risk of adverse health effects in humans"; and the AR (Average Requirement), which refers to "the intake of a nutrient that meets the daily needs of half the people in a typical healthy population". G100:0 = 100% CF/0% RF; G75:25 = 75% CF/25% RF; G50:50 = 50% CF/50% RF; G25:75 = 25% CF/75% RF. mg per Serving Size (30 g Biscuit) % AR Covered by 30 g of Biscuit AR (mg/day) UL (mg/day) G100:0 G75:25 G50:50 G25:75 G100:0 G75:25 G50:50 G25:75 Iron 1.02 0.71 0.531 0.378 12.71% 8.89% 6.64% 4.73% 8 - Potassium 173.01 119.17 94.092 70.122 6.41% 4.41% 3.48% 2.60% 2700 * - Zinc 0.60 0.43 0.369 0.300 6.71% 4.79% 4.15% 3.37% 8.9 18 Phosphorus 62.92 49.89 44.553 33.501 9.83% 7.79% 6.96% 5.23% 640 * - Magnesium 34.04 23.23 20.385 16.236 13.61% 9.29% 8.15% 6.49% 250 * 250 Copper 0.19 0.14 0.105 0.075 16.91% 12.27% 9.55% 6.82% 1.1 * 4 Manganese 0.33 0.30 0.291 0.279 16.50% 15.15% 14.55% 13.95% 2 * - Calcium 8.70 7.21 5.697 4.269 0.91% 0.75% 0.59% 0.44% 960 - Selenium ** 1.75 2.27 2.40 2.72 3.18% 4.12% 4.36% 4.95% 55 * 200 * DRV indicated as AI (adequate intake), which is used when there is not enough data to calculate an average requirement. An AI is "the average nutrient level, based on observations or experiments, that is assumed to be adequate for the population's needs". ** expressed in mg. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. WHO European Regional Obesity Report World Health Organization (WHO) Regional Office for Europe Copenhagen, Denmark 2022 2. Maltoni G. Zioutas M. Deiana G. Biserni G.B. Pession A. Zucchini S. Gender differences in weight gain during lockdown due to COVID-19 pandemic in adolescents with obesity Nutr. Metab. Cardiovasc. Dis. 2021 31 2181 2185 10.1016/j.numecd.2021.03.018 33994065 3. Vogel M. Geserick M. Gausche R. Beger C. Poulain T. Meigen C. Korner A. Keller E. Kiess W. Pfaffle R. weight group-specific weight gain patterns in children and adolescents during the 15 years before and during the COVID-19 pandemic Int. J. Obes. 2021 46 144 152 10.1038/s41366-021-00968-2 4. Calderon O.G. de Mena H.E. Alimentacion del nino preescolar, escolar y del adolescente Pediatr. Integral 2020 98 5. Gomez-Martinez S. Martinez-Gomez D. de Heredia F.P. Romeo J. Cuenca-Garcia M. Martin-Matillas M. Castillo M. Rey-Lopez J.P. Vicente-Rodriguez G. Moreno L. Eating Habits and Total and Abdominal Fat in Spanish Adolescents: Influence of Physical Activity. The AVENA Study J. Adolesc. Health 2012 50 403 409 10.1016/j.jadohealth.2011.08.016 22443846 6. Moreno-Villares J.M. Dalmau J. Libro blanco de la nutricion infantil en Espana Acta Pediatr. Esp. 2016 74 143 7. Villares J. Segovia M. Alimentacion del nino preescolar, escolar y del adolescente Pediatr. Integral 2015 19 268 276 8. Minana V. Correcher P. Dalmau J. La nutricion del adolescente Adolescere 2016 4 6 18 9. Joshi-Saha A. Sethy S.K. Misra G. Dixit G. Srivastava A. Sarker A. Biofortified legumes: Present scenario, possibilities and challenges Field Crops Res. 2022 279 108467 10.1016/j.fcr.2022.108467 10. Hamdani A.M. Wani I.A. Bhat N.A. Gluten free cookies from rice-chickpea composite flour using exudate gums from acacia, apricot and karaya Food Biosci. 2020 35 100541 10.1016/j.fbio.2020.100541 11. Sofi S.A. Muzaffar K. Ashraf S. Gupta I. Mir S.A. Chickpea Pulses Springer Berlin/Heidelberg, Germany 2020 55 76 12. Jukanti A.K. Gaur P.M. Gowda C.L.L. Chibbar R.N. Nutritional quality and health benefits of chickpea (Cicer arietinum L.): A review Br. J. Nutr. 2012 108 S11 S26 10.1017/S0007114512000797 22916806 13. Goni I. Valentin-Gamazo C. Chickpea flour ingredient slows glycemic response to pasta in healthy volunteers Food Chem. 2003 81 511 515 10.1016/S0308-8146(02)00480-6 14. Wallace T.C. Murray R. Zelman K. The nutritional value and health benefits of chickpeas and hummus Nutrients 2016 8 766 10.3390/nu8120766 27916819 15. Aguilera J.M. Rational food design and food microstructure Trends Food Sci. Technol. 2022 122 256 264 10.1016/j.tifs.2022.02.006 16. Morris V.J. Woodward N. Gunning A. Atomic force microscopy as a nanoscience tool in rational food design J. Sci. Food Agric. 2011 91 2117 2125 10.1002/jsfa.4501 21769874 17. Schroen K. de Ruiter J. Berton-Carabin C. Microtechnological Tools to Achieve Sustainable Food Processes, Products, and Ingredients Food Eng. Rev. 2020 12 101 120 10.1007/s12393-020-09212-5 18. Moreno Villares J.M. Dalmau Serra J. Manual de nutricion Nutr. Hosp. 2021 38 1115 19. Agostoni C. Braegger C. Decsi T. Kolacek S. Koletzko B. Mihatsch W. A Moreno L. Puntis J. Shamir R. Szajewska H. Role of dietary factors and food habits in the development of childhood obesity: A commentary by the ESPGHAN Committee on Nutrition J. Pediatr. Gastroenterol. Nutr. 2011 52 662 669 10.1097/MPG.0b013e3182169253 21593641 20. Marangoni F. Martini D. Scaglioni S. Sculati M. Donini L.M. Leonardi F. Poli A. Snacking in nutrition and health Int. J. Food Sci. Nutr. 2019 70 909 923 10.1080/09637486.2019.1595543 30969153 21. St-Onge M.-P. Ard J. Baskin M.L. Chiuve S.E. Johnson H.M. Kris-Etherton P. Varady K. Meal timing and frequency: Implications for cardiovascular disease prevention: A scientific statement from the American Heart Association Circulation 2017 135 e96 e121 10.1161/CIR.0000000000000476 28137935 22. EFSA Dietary Reference Values for the EU DRV Finder Parma, Italy 2022 23. da Quinta N. Alvarez-Sabatel S. Martinez de Maranon I. Alfaro B. Children's acceptability profiles for biscuits with different fiber content J. Texture Stud. 2022 53 41 51 10.1111/jtxs.12642 34716919 24. Chen J. Karlsson C. Povey M. Acoustic envelope detector for crispness assessment of biscuits J. Texture Stud. 2005 36 139 156 10.1111/j.1745-4603.2005.00008.x 25. Laguna L. Salvador A. Sanz T. Fiszman S.M. Performance of a resistant starch rich ingredient in the baking and eating quality of short-dough biscuits LWT-Food Sci. Technol. 2011 44 737 746 10.1016/j.lwt.2010.05.034 26. Gouyo T. Mestres C. Maraval I. Fontez B. Hofleitner C. Bohuon P. Assessment of acoustic-mechanical measurements for texture of French fries: Comparison of deep-fat frying and air frying Food Res. Int. 2020 131 108947 10.1016/j.foodres.2019.108947 32247460 27. Salvador A. Varela P. Sanz T. Fiszman S.M. Understanding potato chips crispy texture by simultaneous fracture and acoustic measurements, and sensory analysis LWT-Food Sci. Technol. 2009 42 763 767 10.1016/j.lwt.2008.09.016 28. Varela P. Chen J. Fiszman S. Povey M.J.W. Crispness assessment of roasted almonds by an integrated approach to texture description: Texture, acoustics, sensory and structure J. Chemom. A J. Chemom. Soc. 2006 20 311 320 10.1002/cem.1029 29. Haytowitz D.B. Ahuja J.K. Wu X. Somanchi M. Nickle M. Nguyen Q.A. Roseland Janet M. Williams Juhi R. Patterson Kristine Y. Li Y. USDA National Nutrient Database for Standard Reference, Legacy Release Nutrient Data Laboratory, Beltsville Human Nutrition Research Center, ARS, USDA Beltsville, MD, USA 2019 Available online: (accessed on 2 January 2023) 30. Rababah T.M. Al-Mahasneh M. Ereifej K. Effect of chickpea, broad bean, or isolated soy protein additions on the physicochemical and sensory properties of biscuits J. Food Sci. 2006 71 S438 S442 10.1111/j.1750-3841.2006.00077.x 31. Yadav R.B. Yadav B. Dhull N. Effect of incorporation of plantain and chickpea flours on the quality characteristics of biscuits J. Food Sci. Technol. 2012 49 207 213 10.1007/s13197-011-0271-x 23572843 32. Mancebo C.M. Rodriguez P. Gomez M. Assessing rice flour-starch-protein mixtures to produce gluten free sugar-snap cookies LWT-Food Sci. Technol. 2016 67 127 132 10.1016/j.lwt.2015.11.045 33. Dapcevic Hadnadev T.R. Torbica A. Hadnadev M. Influence of buckwheat flour and carboxymethyl cellulose on rheological behaviour and baking performance of gluten-free cookie dough Food Bioprocess Technol. 2013 6 1770 1781 10.1007/s11947-012-0841-6 34. Gerzhova A. Mondor M. Benali M. Aider M. Incorporation of canola proteins extracted by electroactivated solutions in gluten-free biscuit formulation of rice-buckwheat flour blend: Assessment of quality characteristics and textural properties of the product Int. J. Food Sci. Technol. 2016 51 814 827 10.1111/ijfs.13034 35. Sarabhai S. Prabhasankar P. Influence of whey protein concentrate and potato starch on rheological properties and baking performance of Indian water chestnut flour based gluten free cookie dough LWT-Food Sci. Technol. 2015 63 1301 1308 10.1016/j.lwt.2015.03.111 36. Segnini S. Dejmek P. Oste R. Relationship between instrumental and sensory analysis of texture and color of potato chips J. Texture Stud. 1999 30 677 690 10.1111/j.1745-4603.1999.tb00237.x 37. Fillion L. Kilcast D. Consumer perception of crispness and crunchiness in fruits and vegetables Food Qual. Prefer. 2002 13 23 29 10.1016/S0950-3293(01)00053-2 |
PMC10000555 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050684 healthcare-11-00684 Article Examining Factors Associated with Dynapenia/Sarcopenia in Patients with Schizophrenia: A Pilot Case-Control Study Tanioka Ryuichi Conceptualization Methodology Formal analysis Investigation Writing - original draft 1* Osaka Kyoko Investigation Writing - original draft Writing - review & editing 2 Ito Hirokazu Methodology Validation Formal analysis Writing - original draft Writing - review & editing 3 Zhao Yueren Validation 4 Tomotake Masahito Validation 3 Takase Kensaku Writing - review & editing 5 Tanioka Tetsuya Conceptualization Methodology Formal analysis Investigation Writing - original draft Writing - review & editing Project administration Funding acquisition 3 Lollgen Herbert Academic Editor 1 Faculty of Health Sciences, Hiroshima Cosmopolitan University, Hiroshima 731-3166, Japan 2 Department of Nursing, Nursing Course of Kochi Medical School, Kochi University, Kochi 783-8505, Japan 3 Graduate School of Biomedical Sciences, Tokushima University, Tokushima 770-8509, Japan 4 Department of Psychiatry, Fujita Health University, Toyoake 470-1192, Japan 5 Department of Rehabilitation, Anan Medical Center, Anan 774-0045, Japan * Correspondence: [email protected] 25 2 2023 3 2023 11 5 68430 12 2022 09 2 2023 22 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Sedentary behavior in patients with schizophrenia causes muscle weakness, is associated with a higher risk of metabolic syndrome, and contributes to mortality risk. This pilot case-control study aims to examine the associated factors for dynapenia/sarcopenia in patients with schizophrenia. The participants were 30 healthy individuals (healthy group) and 30 patients with schizophrenia (patient group), who were matched for age and sex. Descriptive statistics, Welch's t-test, cross-tabulations, adjusted residuals, Fisher's exact probability test (extended), and/or odds ratios (ORs) were calculated. In this study, dynapenia was significantly more prevalent in patients with schizophrenia than in healthy individuals. Regarding body water, Pearson's chi-square value was 4.41 (p = 0.04), and significantly more patients with dynapenia were below the normal range. In particular, body water and dynapenia showed a significant association, with an OR = 3.42 and 95% confidence interval [1.06, 11.09]. Notably, compared with participants of the healthy group, patients with schizophrenia were overweight, had less body water, and were at a higher risk for dynapenia. The impedance method and the digital grip dynamometer used in this study were simple and useful tools for evaluating muscle quality. To improve health conditions for patients with schizophrenia, additional attention should be paid to muscle weakness, nutritional status, and physical rehabilitation. dynapenia sarcopenia schizophrenia basal metabolic rate total muscle mass skeletal muscle mass body mass index Japan Society for the Promotion of ScienceJP17H01609 Center for Design-based AI Education and Research, University of TokushimaThis research was funded by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (grant number JP17H01609) and the Center for Design-based AI Education and Research, University of Tokushima. We would like to express our deepest gratitude to the sponsors. pmc1. Introduction The prevalence of schizophrenia in Japan is estimated at 0.7% . Patients with schizophrenia die on average 10 to 20 years earlier than healthy individuals . The sedentary lifestyle common among patients with schizophrenia is associated with higher metabolic syndrome (cardiovascular changes due to diabetes mellitus, hypertension, and hypercholesterolemia) and contributes to mortality risk. Lifestyles that increase the risk of such metabolic syndrome have been identified as lacking regular physical activity, poor food intake, substance use, and high rates of smoking . Strassnig et al. developed a comprehensive model to conceptualize multimodal relationships that predict impaired activities of daily living in patients with schizophrenia. According to these authors, limitations in physical abilities interfere with activities of daily living and elicit a state of physical infirmity observed in other chronic illnesses. A high prevalence of sarcopenic obesity has also been reported in patients with schizophrenia . However, little is known on the risk factors for dynapenia/sarcopenia in patients with schizophrenia. Factors that contribute to severe limitations due to the pathophysiology of schizophrenia and the effects of medications are complex and include a sedentary lifestyle as well as factors due to the effects of medication therapy, which ultimately lead to a vicious cycle of obesity and cardiovascular metabolic risk . A sedentary lifestyle and decreased functional motor skills in patients with schizophrenia reduce their quality of life . Low physical activity levels are also associated with the use of antipsychotic drugs. This implies that increasing weight is related to limitations in physical functioning and restricts activities of daily living. Physical inactivity due to obesity also adds to the burden of schizophrenia in the form of reduced physical health-related quality of life. Rehabilitation programs focusing on these risk factors should be key for physical activity for both prevention and treatment of disease and disablement in patients with schizophrenia . Moreover, antipsychotic medications have been associated with weight gain and obesity in schizophrenia. Patients with schizophrenia consume unhealthy food, and their dietary patterns identified a high consumption of saturated fat and low intake of fruit and dietary fiber . In patients with schizophrenia, the ability to supply oxygen to muscles during exercise and the ability of muscles to consume oxygen (cardiopulmonary endurance) are poorer than in healthy individuals. This means the level of cardiorespiratory fitness may be extremely low in patients with schizophrenia, amounting to a state of deconditioning and a very low capacity for sustained physical activity that is high in intensity, activities promoting low to moderate activity levels may serve the population well and lead to highly relevant improvements in health prospects . In a 12-year follow-up study of schizophrenia, the patient group had an excess of psychiatric and physical comorbidities (fractured neck of femur, parkinsonism, pneumonia, esophageal ulcer, respiratory failure, and bronchitis), including side effects of psychotropic drugs, compared to the sex-matched controls. Specifically, their finding clearly demonstrates that parkinsonism-associated complications may play a dominant role in schizophrenia-related death in general hospitals. Reducing the risk of parkinsonism-associated complications due to accurate detection and management of side effects of psychotropic and somatic medication as well as of related drug-drug interactions, continuously monitoring physical status, and accurate detection of concomitant metabolic, cardiovascular, and respiratory diseases as well as creating awareness about preventive strategies for difficulty eating and aspiration pneumonia may help in reducing parkinsonism-associated fatal consequences in general hospitals in patients with schizophrenia . In the context of socioeconomic challenges, schizophrenia leads to particularly unhealthy lifestyles that include poor diets, little exercise, marked sedentary behavior, and high rates of smoking with commensurately low physical activity levels . As a result, compared to the general healthy population, patients with schizophrenia have severe symptomatic limitations in physical capacity. Negative symptoms reduce the likelihood of patients' engagement in goal-directed behavior, including physical activity, which has been noted to increase obesity and cardiometabolic risk and induce poor physical conditions, resulting in sarcopenic obesity and muscle weakness . According to the European working group on sarcopenia in older people 2 (EWGSOP 2), sarcopenia is suspected when (1) muscle weakness is confirmed, whereas sarcopenia is confirmed when (2) muscle mass or muscle quality decline is present in addition to muscle weakness . As mentioned above, sarcopenia is defined as "loss of muscle mass or quality," whereas dynapenia is defined as "loss of muscle strength" . Sarcopenia is also associated with depressed mood, which in turn is associated with low muscle strength and physical performance . Therefore, it is problematic that patients with schizophrenia frequently have negative symptoms such as depressed mood, which is associated with low physical function and low muscle strength. Regarding schizophrenia and nutritional status, Japanese inpatients with schizophrenia are more likely to be underweight and undernourished than outpatients . Nutritional status is an issue for patients with schizophrenia in Japan. Therefore, it is important to consider the activities of daily living, dynapenia, sarcopenia/presarcopenia, and nutritional status when considering symptom management in hospitalized patients with chronic mental illness. This pilot study aimed to examine the associated factors for dynapenia/sarcopenia in patients with schizophrenia. 2. Materials and Methods 2.1. Study Participants This pilot case-control study enrolled 60 individuals in total comprising 30 healthy participants (healthy group) and 30 patients with schizophrenia (patient group), matched by age, ranging from 40 to 89 years, and sex. 2.2. Data Acquisition Period The study's data acquisition phase was from 17 August 2021 to 30 November 2021. 2.3. Target Selection Criteria Healthy group: Employees working at Hospital A and its Geriatric Health Care Facility. Patient group: Patients with schizophrenia admitted to Hospital A. Both groups were matched by sex and age. 2.4. Exclusion Criteria Healthy group: Individuals with a mental or physical disorder. Patient group: Individuals unable to understand instructions owing to a medical condition or medication status or because of a physical disorder such as a history of cerebrovascular disease such as stroke or a neurological disease. 2.5. Assessment Methods 2.5.1. Body Mass Tanita monitors use the latest bioelectrical impedance analysis technology, first developed by Tanita in 1992, to provide fast and accurate body composition results . The RD-545 InnerScan Pro provides an in-depth analysis of 26 body composition measurements. The measurements included weight, body fat, muscle mass, muscle quality score, body mass rating, bone mass, visceral fat level, basal metabolic rate, metabolic age, total body water, and body mass index (BMI). The RD-545 InnerScan PRO can perform fat and muscle analysis individually for arm, leg, and trunk segments if hand electrodes are used . The state of visceral fat accumulation is indicated as visceral fat level score measured by the RD-545 InnerScan Pro. 2.5.2. Age, Height, and Weight Healthy group: Age and height were self-reported based on the hospital's staff health examination form. Patient group: Age and height were obtained from the medical records. Weight was measured in both groups using a scale (RD-545 InnerScan Pro, TANITA Corporation. Tokyo, Japan). 2.5.3. Grip Strength of the Hands A digital grip dynamometer (T.K.K.5401; Takei Scientific Instruments, Co., Ltd., Niigata, Japan) was used to individually measure the grip strength of each hand in a stable standing posture. 2.5.4. Skeletal Muscle Mass Index (SMI) The total limb skeletal muscle mass (kg) was calculated from the information obtained from the body mass, and the data were divided by the square of the corresponding height (m2). 2.5.5. SARC-F Score The SARC-F score was presented by Morley as a screening tool for sarcopenia at the EU/US committee on sarcopenia in the frail elderly at the Conference on Sarcopenia Research (ICSR) in Orlando in 2012 . Data were self-reported by all participants using a questionnaire survey. 2.6. Sarcopenia/Dynapenia Assessment Method This study adopted the diagnostic criteria proposed by the Asian Working Group for Sarcopenia (AWGS) . The SARC-F score, grip strength, and skeletal muscle mass were used as indicators. The specific criteria were as follows: Grip strength can be used to assess muscle weakness . Peripheral quantitative computed tomography, dual X-ray energy absorptiometry, and magnetic resonance imaging techniques can be used to assess skeletal muscle mass and quality . Other than the aforementioned methods, bioelectrical impedance analysis can be used, which has the advantage of being inexpensive and portable. The cutoff values for sarcopenia in the Japanese population are 6.8 kg/m2 for men and 5.7 kg/m2 for women . The SARC-F score was used to select participants with sarcopenia; those with a score of 4 or more points were selected. Based on the two sarcopenia criteria outlined in the Section 1, muscle weakness and loss of muscle mass or muscle quality were evaluated. (1) Grip strength was used as an indicator of muscle weakness, defined for men and women as having a grip strength of less than 26 kg and less than 18 kg, respectively. In addition, (2) skeletal muscle mass (kg/m2) was used as an indicator of muscle mass or muscle quality loss, and skeletal muscle mass loss was defined as a value less than 7.0 kg/m2 for men and less than 5.7 kg/m2 for women. Presarcopenia was defined as reduced skeletal muscle mass and normal grip strength. Dynapenia was defined as a normal skeletal muscle mass and decreased grip strength. 2.7. Statistical Analysis Basic statistical parameters (mean +- standard deviation [SD], 95% confidence interval [CI]) were calculated. Welch's t-test was performed to compare the two study groups. For items that were significantly different between the two groups, cross-tabulations were performed, and adjusted residuals were calculated. Fisher's exact probability test (Extended), Pearson's chi-square test, and/or odds ratios (ORs) were calculated. All statistical analyses were performed using SPSS 21.0 (IBM Corporation). Statistical significance was set at p < 0.05. 3. Results Among the study participants, 61.7% (37/60) were women and 38.8% (23/60) were men. The healthy group comprised 63.3% (19/30) women and 36.7% (11/30) men, whereas the patient group consisted of 60.0% (18/30) women and 40.0% (12/30) men. In this study, dynapenia and sarcopenia/presarcopenia were assessed. Among the 30 participants in the patient group, 10.0% (3/30 [1/18 women, 2/12 men]) met the criteria for sarcopenia, 3.3% (1/30 [1/12 men]) for presarcopenia, and 60.0% (18/30 [14/18 women, 4/12 men]) for dynapenia. The corresponding results of the healthy group showed that sarcopenia and presarcopenia were not present (0%) and that 13.3% (4/30 [3/19 women, 1/11 men]) met the criteria for dynapenia. Table 1 shows the results of Welch's t-test. Body water content was significantly higher in the healthy group with 53.56 +- 3.94% in the healthy group and 49.77 +- 6.58% in the patient group (t = 2.71, p < 0.001). The visceral fat level score was 6.60 +- 3.71 in the healthy group and 9.12 +- 5.35 in the patient group (t = 2.11, p = 0.04). Body fat content was 24.95 +- 6.05% in the healthy group and 30.41 +- 9.00% in the patient group (t = 2.76, p < 0.01). Likewise, BMI was 21.89 +- 2.30 kg/m2 for the healthy group and 23.88 +- 4.65 kg/m2 for the patient group (t = 2.10, p = 0.04). Left grip strength was 29.16 +- 9.07 kg for the healthy group and 18.53 +- 8.38 kg for the patient group (t = 4.71, p < 0.001), whereas right grip strength was 30.05 +- 7.98 kg for the healthy group and 21.26 +- 10.92 kg for the patient group (t = 3.56, p < 0.001). These findings showed that for both sides, the grip strength of the patient group was significantly weaker than that in the healthy group. As shown in Table 2, the patient group was significantly more likely to have dynapenia or sarcopenia/presarcopenia (Fisher's exact test, p < 0.0001; OR, 17.88; 95% CI [4.74, 67.43]). The association of the study group with dynapenia, including sarcopenia and presarcopenia (hereafter referred to as dynapenia in the Section 3), was analyzed based on items with significant differences in Table 1. No significant association was found for the parameters of visceral fat level score, body fat, and BMI. In contrast, for body water, the result of Pearson's chi-square test was 4.41 (p = 0.04), and significantly more people with dynapenia were below the normal range. We also confirmed a significant association for body water (OR, 3.42, 95% CI [1.06, 11.09]). 4. Discussion As shown in Table 1, the patient group had significantly higher body fat, visceral fat level scores, and BMI. In addition, the average value of the patient group BMI is not at the obese level, and the high visceral fat level score was deemed a problem when considered overall from the cross-tabulation results in Table 2. As shown in Table 2, the patient group had a high percentage of individuals diagnosed with dynapenia, with an OR of 17.88 times the risk of developing the disease compared with healthy individuals. Thus, it was suggested that being afflicted with schizophrenia is one factor associated with dynapenia. Moreover, Table 2 shows that no significant association by the study group was found for body fat, visceral fat level score, or BMI; however, body water content was significantly associated, with the OR indicating 3.42 times higher risk of dynapenia for the patient group than for the healthy group. For these reasons, the patient group in this study may have increased fat, as well as decreased body water content and muscle mass, owing to a sedentary lifestyle . Sex differences in body fat and water content in patients with schizophrenia have been reported . The body water content was predominantly higher in the healthy group. Body water refers to water contained in various body compartments, including blood, lymphatic fluid, extracellular fluid, and intracellular fluid . These fluids play important roles in the body, such as transporting nutrients and maintaining a constant body temperature, and they tend to decrease with age. In addition, people with high body fat tend to have a lower body water content . This trend is also consistent with the previous study by Bulbul et al. Therefore, it is necessary to focus on the trends of high body fat and low water content in the patient group. Of the 307 participants in the study by Mori et al. , 60.9% were assessed as normal, and 25.7%, 8.1%, and 5.2% were found to have presarcopenia, sarcopenia, and dynapenia, respectively. Reduced grip strength is a critical indicator of dynapenia . In this study, grip strength was significantly lower in patients with schizophrenia than in healthy individuals. Because many patients with schizophrenia have dynapenia, grip strength may be a convenient screening index for dynapenia in psychiatric hospitals. The participants of the study by Kobayashi et al. were volunteers aged over 60 years who were in good general health . Their study found that in Japan, the rates of sarcopenia, presarcopenia, and dynapenia were 10%, 22%, and 8% in men, and 19%, 23%, and 13% in women. According to Neves et al., sarcopenia and dynapenia were identified in 15.3% and 38.2% of old persons . In this study, 13.3% of the healthy individuals had dynapenia, whereas 60.0% of the patient group had dynapenia, 10.0% had sarcopenia, and 3.3% had presarcopenia. Thus, our data suggest that the prevalence of dynapenia is high among patients with schizophrenia. Appetite regulation and physical activity affect energy balance and changes in body fat mass. In some patients, inflammation induces anorexia and fat loss along with sarcopenia. In others, appetite is maintained, despite the activation of systemic inflammation, leading to sarcopenia with normal or increased BMI. Inactivity contributes to sarcopenia and increased fat tissue in aging and disease . In a previous study of the BMI status of hospitalized Japanese schizophrenia patients, underweight and obesity were characteristic in schizophrenia inpatients compared with the general population. In particular, regarding the characteristics of underweight, a previous study showed that the prevalence of hypotriglyceridemia was significantly higher in the underweight group than in the normal weight group and in overweight/obese schizophrenia inpatients . Harvey and Strassnig suggested that the cognitive limitations of people with schizophrenia not only correlate with disability directly, but contribute substantially to other skills deficits (functional capacity; social competence) that exacerbate disability outcomes. Impaired cognition and negative symptoms, particularly in the domains of reasoning and problem solving and reinforcement valuation, can lead to deficits in functional capacity that then lead to poor dietary and exercise choices, contributing to poor functional outcomes. In another study, age, certification of long-term care, and malnutrition were identified as risk factors for sarcopenia . Sarcopenia is thought to primarily explain the age-related loss of muscle strength, such as dynapenia, commonly seen in older people . However, recent longitudinal data indicate that the loss of muscle strength occurs significantly faster than the accompanying loss of muscle mass . On the other hand, gains in muscle mass and strength afforded by resistive training are associated with a small but significant improvement in physical performance. It is noteworthy that lower intensity mechanical loading such as aerobic exercise, despite being considerably less effective for inducing muscle hypertrophy, has been found to promote protein synthesis and expression of growth-related genes and inhibit the expression of muscle breakdown-related genes . Muscle weakness is known to decrease physical function and increase the risk of mortality . Regarding the changes in physical function associated with aging, muscle strength declines by 30% and muscle area by 40% between 20 and 70 years of age . At the age of 75 years, muscle strength declines at a rate of 2.5-3% per year for women and 3-4% per year for men, and muscle mass is lost at a rate of 0.64-0.70% per year for women and 0.80-0.98% per year for men . Kitamura, et al. found sex-specific patterns of correlates with sarcopenia. Significant sarcopenia-related factors in addition to ageing were hypoalbuminaemia, cognitive impairment, low activity, and recent hospitalization among men and cognitive impairment and depressed mood among women. It is important to focus on these conditions. Compared to young adults, older adults have a lower limb skeletal muscle index (ASMI, kg/m2) and a significantly higher body fat percentage . It has been noted that diabetic patients with a high body fat percentage in addition to low BMI may develop sarcopenia . Moreover, the prevalence of diabetes in patients with schizophrenia in Japan has been reported to be 8.6% . Protein intake is necessary for efficient muscle growth. A person with adequate muscle mass needs 1.0-1.2 g protein per kg of body weight per day for an older person to maintain muscle mass, i.e., about 60-72 g per day if the person weighs 60 kg . However, this intake is not sufficient for those who must gain muscle mass due to sarcopenia, and they should have an intake of 1.2-1.5 g of protein per kg of body weight per day, i.e., 72-90 g per day if they weigh 60 kg . Thus, it is important to control the balance of restricted caloric intake with guaranteed protein intake for patients with dynapenia. However, if a patient has kidney problems, it is critical to pay much more attention to an appropriate protein and calorie intake during the rehabilitation process . Based on the BMI findings of our study, the patient group was not underweight. Our study subjects were inpatients; they have consumed a diet regulated by a psychiatrist and a dietitian. However, outpatients may not be eating an appropriate diet due to unbalanced diets, poverty, etc. With this in mind, we should conduct the main case-control study following this pilot study. Furthermore, inpatients may have a lower average BMI than outpatients, who are free to eat whatever they want at home because their food intake is controlled to prevent excessive weight gain. It was considered important to keep these points in mind when managing their health. Limitations and Future Research Since data on daily intakes, such as nutritional status, were not obtained in this study, it is necessary to obtain data on "official" caloric intake based on hospital diets, such as daily caloric intake, for better analyses in future studies. Additionally, it is necessary to consider "unofficial" caloric intake, such as snacks. Moreover, the patient's amount of activity needs to be considered. This pilot study was a small-scale study conducted to inform, predict, and direct an intended future full-scale study. The association of low body water and dynapenia in patient participants suggests that low body water might be a risk factor for dynapenia in these patients. Underweight is highly prevalent in Japanese inpatients with schizophrenia. Psychiatrists should be aware of underweight and their potential health risks. Treating psychiatrists should also be responsible for providing any necessary nutritional interventions . Physical health appears to be achievable in people with schizophrenia being challenged by motivational difficulties with attending regular exercise and have beneficial implications for physical function during activities of daily living, lifestyle-related diseases, and early death. Specifically, physical training is an effective countermeasure to improve the low aerobic endurance and skeletal muscle strength in these patients . Furthermore, the main study following this pilot study should include not only body composition (low body water, visceral fat level, and muscle mass), grip strength, and joint range of motion, but also medication content, heart rate variability, and motor velocity . Other factors (physical function during activities of daily living, gait and psychiatric symptoms specific to schizophrenia, age, and length of hospitalization) also must be considered in dynapenia in patients with schizophrenia. 5. Conclusions This pilot study examined the risk factors for dynapenia/sarcopenia in patients with schizophrenia. Patients with schizophrenia were overweight, had less body water than the healthy study participants, and were at a higher risk of dynapenia than participants in the healthy group. The impedance method used in this study is a simple and useful method for evaluating muscle quality in conditions such as dynapenia. To improve health conditions for patients with schizophrenia, additional attention should be paid to muscle weakness, nutritional status, and physical rehabilitation. Future research will include a larger study following on this pilot study. Acknowledgments The authors express gratitude and appreciation to the participants. The authors thank the Mifune Hospital for assistance during the conduct of the research. Author Contributions Conceptualization, R.T. and T.T.; methodology, R.T., H.I. and T.T.; validation, M.T., H.I. and Y.Z.; formal analysis, R.T., H.I. and T.T.; investigation, K.O., R.T. and T.T.; writing--original draft preparation, K.O., R.T., H.I. and T.T.; writing--review and editing, K.O., H.I., T.T. and K.T.; project administration, T.T.; funding acquisition, T.T. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Tokushima University Hospital (#3046) and the Mifune Hospital Clinical Research Ethics Review Committee (#20180502). Informed Consent Statement Informed consent was obtained from all participants involved in the study. Data Availability Statement Data presented in this study are available upon request from the corresponding author. The data are not publicly available because of privacy and ethical restrictions. Conflicts of Interest The authors declare no conflict of interest. healthcare-11-00684-t001_Table 1 Table 1 Comparison of the measured parameters between the two study groups. Items Healthy Group Patient Group Welch's t-Test 95% Confidence Interval n = 30 n = 30 Mean +-SD Mean +-SD t p Lower Upper Age 64.3 10.16 62.53 10.26 0.67 0.51 -3.51 7.04 Body height, cm 161.47 6.11 159.5 9.08 0.98 0.33 -2.04 5.97 Body weight, kg 57.25 8.49 60.55 15.52 -1.03 0.31 -9.82 3.2 BMR, kcal 1195.9 176.66 1216.97 276.33 -0.35 0.73 -141.38 99.25 Bone mass, kg 2.37 0.34 2.25 0.46 1.12 0.27 -0.09 0.33 Body water, % 53.56 3.94 49.77 6.58 2.71 <0.01 0.98 6.61 Visceral fat level score 6.6 3.71 9.12 5.35 -2.11 0.04 -4.91 -0.13 Body fat, % 24.95 6.05 30.41 9 -2.76 <0.01 -9.44 -1.49 BMI, kg/m2 21.89 2.3 23.88 4.65 -2.1 0.04 -3.91 -0.08 MM L-arm, kg 2 0.45 1.99 0.63 0.11 0.92 -0.27 0.3 MM R-arm, kg 2.02 0.46 1.88 0.54 1.09 0.28 -0.12 0.4 MM upper limb, kg 4.02 0.9 3.87 1.11 0.59 0.56 -0.37 0.68 MM L-leg, kg 7.49 1.26 7.01 1.75 1.2 0.24 -0.32 1.26 MM R-leg, kg 7.55 1.27 6.99 1.7 1.43 0.16 -0.22 1.33 MM lower limb, kg 15.04 2.52 14.01 3.4 1.33 0.19 -0.52 2.57 MM trunk, kg 21.48 3.48 21.63 4.42 -0.14 0.89 -2.21 1.91 TMM, kg 40.54 6.72 39.51 8.67 0.52 0.61 -2.98 5.04 LGS, kg 29.16 9.07 18.53 8.38 4.71 <0.001 6.12 15.15 RGS, kg 30.05 7.98 21.26 10.92 3.56 <0.001 3.83 13.74 SMI, kg/m2 7.26 0.82 6.94 1.18 1.21 0.23 -0.21 0.84 SD, standard deviation; BMR, basal metabolic rate; BMI, body mass index; MM, muscle mass; L-arm, left arm; R-arm, right arm; L-leg, left leg; R-leg, right leg; TMM, total muscle mass; LGS, left grip strength; RGS, right grip strength; SMI, skeletal muscle mass index. healthcare-11-00684-t002_Table 2 Table 2 Associations with the risk of dynapenia/sarcopenia in healthy individuals and patients with schizophrenia. Group Healthy Schizophrenia Analysis Results Normal Frequency 26 8 Fisher's exact test, p < 0.0001, OR = 17.88, 95% CI [4.74, 67.43] AR 4.7 -4.7 Dynapenia * Frequency 4 22 AR -4.7 4.7 Body Water Normal Below Standard Pearson's chi-square test = 4.413, p = 0.04, OR = 3.42, 95% CI [1.06, 11.09] Normal Frequency 28 6 AR 2.1 -2.1 Dynapenia * Frequency 15 11 AR -2.1 2.1 Visceral Fat Level Score Normal Above Standard Pearson's chi-square test = 0.184, p = 0.67, OR = 1.27, 95% CI [0.43, 3.80] Normal Frequency 24 10 AR 0.4 -0.4 Dynapenia * Frequency 17 9 AR -0.4 0.4 Body Fat Decreased Normal Increased Analysis Results Normal Frequency 6 20 8 Fisher's exact test, p = 0.159 AR 0.2 1.6 -1.8 Dynapenia * Frequency 4 10 12 AR -0.2 -1.6 1.8 BMI Level Decreased Normal Increased Analysis Results Normal Frequency 2 24 8 Fisher's exact test, p = 0.774 AR -0.8 0.7 -0.3 Dynapenia * Frequency 3 16 7 AR 0.8 -0.7 0.3 * Dynapenia includes dynapenia (n = 22), presarcopenia (n = 1), and sarcopenia (n = 3). AR, adjusted residual; BMI, body mass index; CI, confidence interval; OR, odds ratio. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Ministry of Health, Labour and Welfare Schizophrenia Available online: (accessed on 20 April 2022) 2. Ringen P.A. Engh J.A. Birkenaes A.B. Dieset I. Andreassen O.A. Increased mortality in schizophrenia due to cardiovascular disease--A non-systematic review of epidemiology, possible causes, and interventions Front. Psychiatry 2014 5 137 10.3389/fpsyt.2014.00137 25309466 3. Galletly C.A. Premature death in schizophrenia: Bridging the gap Lancet Psychiatry 2017 4 263 265 10.1016/S2215-0366(17)30079-2 28237638 4. De Hert M. Schreurs V. Vancampfort D. Van Winkel R. Metabolic syndrome in people with schizophrenia: A review World Psychiatry 2009 8 15 22 Erratum in World Psychiatry 2011, 10, 78 10.1002/j.2051-5545.2009.tb00199.x 19293950 5. Strassnig M. Signorile J. Gonzalez C. Harvey P.D. Physical performance and disability in schizophrenia Schizophr. Res. Cogn. 2014 1 112 121 10.1016/j.scog.2014.06.002 25254158 6. Sugawara N. Yasui-Furukori N. Tsuchimine S. Fujii A. Sato Y. Saito M. Matsuzaka M. Takahashi I. Kaneko S. Body composition in patients with schizophrenia: Comparison with healthy controls Ann. Gen. Psychiatry 2012 11 11 10.1186/1744-859X-11-11 22554352 7. Henderson D.C. Vincenzi B. Andrea N.V. Ulloa M. Copeland P.M. Pathophysiological mechanisms of increased cardiometabolic risk in people with schizophrenia and other severe mental illnesses Lancet Psychiatry 2015 2 452 464 10.1016/S2215-0366(15)00115-7 26360288 8. Vancampfort D. Probst M. Scheewe T. Maurissen K. Sweers K. Knapen J. De Hert M. Lack of physical activity during leisure time contributes to an impaired health related quality of life in patients with schizophrenia Schizophr. Res. 2011 129 122 127 10.1016/j.schres.2011.03.018 21493044 9. Vancampfort D. Firth J. Schuch F.B. Rosenbaum S. Mugisha J. Hallgren M. Probst M. Ward P.B. Gaughran F. De Hert M. Sedentary behavior and physical activity levels in people with schizophrenia, bipolar disorder and major depressive disorder: A global systematic review and meta-analysis World Psychiatry 2017 16 308 315 10.1002/wps.20458 28941119 10. Manu P. Dima L. Shulman M. Vancampfort D. De Hert M. Correll C.U. Weight gain and obesity in schizophrenia: Epidemiology, pathobiology, and management Acta Psychiatr. Scand. 2015 132 97 108 10.1111/acps.12445 26016380 11. Strassnig M. Brar J.S. Ganguli R. Low cardiorespiratory fitness and physical functional capacity in obese patients with schizophrenia Schizophr. Res. 2011 126 103 109 10.1016/j.schres.2010.10.025 21146958 12. Schoepf D. Uppal H. Potluri R. Heun R. Physical comorbidity and its relevance on mortality in schizophrenia: A naturalistic 12-year follow-up in general hospital admissions Eur. Arch. Psychiatry Clin. Neurosci. 2014 264 3 28 10.1007/s00406-013-0436-x 23942824 13. Strassnig M. Brar J.S. Ganguli R. Increased caffeine and nicotine consumption in community-dwelling patients with schizophrenia Schizophr. Res. 2006 86 269 275 10.1016/j.schres.2006.05.013 16859897 14. Cruz-Jentoft A.J. Bahat G. Bauer J. Boirie Y. Bruyere O. Cederholm T. Cooper C. Landi F. Rolland Y. Sayer A.A. Sarcopenia: Revised European consensus on definition and diagnosis Age Ageing 2019 48 16 31 10.1093/ageing/afy169 30312372 15. Clark B.C. Manini T.M. Functional consequences of sarcopenia and dynapenia in the elderly Curr. Opin. Clin. Nutr. Metab. Care 2010 13 271 276 10.1097/MCO.0b013e328337819e 20154609 16. Gao K. Ma W.Z. Huck S. Li B.L. Zhang L. Zhu J. Li T. Zhou D. Association between sarcopenia and depressive symptoms in Chinese older adults: Evidence from the China health and retirement longitudinal study Front. Med. 2021 8 755705 10.3389/fmed.2021.755705 17. Sugai T. Suzuki Y. Yamazaki M. Shimoda K. Mori T. Ozeki Y. Matsuda H. Sugawara N. Yasui-Furukori N. Minami Y. High prevalence of underweight and undernutrition in Japanese inpatients with schizophrenia: A nationwide survey BMJ Open 2015 5 e008720 10.1136/bmjopen-2015-008720 18. Tanita Understanding Your Measurements Available online: (accessed on 20 April 2022) 19. Tanita Basal Metabolic Rate: What Is It and How to Measure It? Available online: (accessed on 20 April 2022) 20. Vellas B. Pahor M. Manini T. Rooks D. Guralnik J.M. Morley J. Studenski S. Evans W. Asbrand C. Fariello R. Designing pharmaceutical trials for sarcopenia in frail older adults: EU/US Task Force recommendations J. Nutr. Health Aging 2013 17 612 618 10.1007/s12603-013-0362-7 23933872 21. Chen L.K. Woo J. Assantachai P. Auyeung T.W. Chou M.Y. Iijima K. Jang H.C. Kang L. Kim M. Kim S. Asian Working Group for Sarcopenia: 2019 Consensus update on sarcopenia diagnosis and treatment J. Am. Med. Dir. Assoc. 2020 21 300 307.e2 10.1016/j.jamda.2019.12.012 32033882 22. Erlandson M.C. Lorbergs A.L. Mathur S. Cheung A.M. Muscle analysis using pQCT, DXA and MRI Eur. J. Radiol. 2016 85 1505 1511 10.1016/j.ejrad.2016.03.001 27005009 23. Yamada Y. Nishizawa M. Uchiyama T. Kasahara Y. Shindo M. Miyachi M. Tanaka S. Developing and validating an age-independent equation using multi-frequency bioelectrical impedance analysis for estimation of appendicular skeletal muscle mass and establishing a cutoff for sarcopenia Int. J. Environ. Res. Public Health 2017 14 809 10.3390/ijerph14070809 28753945 24. Marthoenis M. Martina M. Alfiandi R. Dahniar D. Asnurianti R. Sari H. Nassimbwa J. Arafat S.M.Y. Investigating body mass index and body composition in patients with schizophrenia: A case-control study Schizophr. Res. Treat. 2022 2022 1381542 10.1155/2022/1381542 35265376 25. Zhou S.G. Chen W. Human body water composition measurement: Methods and clinical application Zhongguo Yi Xue Ke Xue Yuan Xue Bao 2018 40 603 609 10.3881/j.issn.1000-503X.10661 30404690 26. Hankin M.E. Munz K. Steinbeck A.W. Total body water content in normal and grossly obese women Med. J. Aust. 1976 2 533 537 10.5694/j.1326-5377.1976.tb128209.x 994955 27. Bulbul F. Tamam L. Demirkol M. Cakmak S. Namli Z. Ersahinoglu E. The prevalence of sarcopenia in patients with schizophrenia Psychiatry Clin. Psychopharmacol. 2021 31 60 66 10.5152/pcp.2021.20175 28. Mori K. Murata S. Goda A. Kikuchi Y. Shiraiwa K. Horie J. Nakano H. Gait characteristics of dynapenia, sarcopenia, and presarcopenia in community-dwelling Japanese older women: A cross-sectional study Healthcare 2022 10 1905 10.3390/healthcare10101905 36292352 29. Alino-Dies M. Sanchez-Orti J.V. Correa-Ghisays P. Balanza-Martinez V. Vila-Frances J. Selva-Vera G. Correa-Estrada P. Fores-Martos J. San-Martin Valenzuela C. Monfort-Panego M. Grip Strength, Neurocognition, and social functioning in people WithType-2 diabetes mellitus, major depressive disorder, bipolar disorder, and schizophrenia Front. Psychol. 2020 11 525231 10.3389/fpsyg.2020.525231 33324271 30. Kobayashi K. Imagama S. Ando K. Nakashima H. Machino M. Morozumi M. Kanbara S. Ito S. Inoue T. Yamaguchi H. Dynapenia and physical performance in community-dwelling elderly people in Japan Nagoya J. Med. Sci. 2020 82 415 424 10.18999/nagjms.82.3.415 33132426 31. Neves T. Ferriolli E. Lopes M.B.M. Souza M.G.C. Fett C.A. Fett W.C.R. Prevalence and factors associated with sarcopenia and dynapenia in elderly people J. Frailty Sarcopenia Falls 2018 3 194 202 10.22540/JFSF-03-194 32300708 32. Biolo G. Cederholm T. Muscaritoli M. Muscle contractile and metabolic dysfunction is a common feature of sarcopenia of aging and chronic diseases: From sarcopenic obesity to cachexia Clin. Nutr. 2014 33 737 748 10.1016/j.clnu.2014.03.007 24785098 33. Harvey P.D. Strassnig M.T. Cognition and disability in schizophrenia: Cognition-related skills deficits and decision-making challenges add to morbidity World Psychiatry 2019 18 165 167 10.1002/wps.20647 31059625 34. Strassnig M.T. Harvey P.D. Miller M.L. Depp C.A. Granholm E. Real world sedentary behavior and activity levels in patients with schizophrenia and controls: An ecological momentary assessment study Ment. Health Phys. Act. 2021 20 100364 10.1016/j.mhpa.2020.100364 35. Kurose S. Nishikawa S. Nagaoka T. Kusaka M. Kawamura J. Nishioka Y. Sato S. Tsutsumi H. Kimura Y. Prevalence and risk factors of sarcopenia in community-dwelling older adults visiting regional medical institutions from the Kadoma sarcopenia Study Sci. Rep. 2020 10 19129 10.1038/s41598-020-76185-0 33154439 36. Nilwik R. Snijders T. Leenders M. Groen B.B. van Kranenburg J. Verdijk L.B. van Loon L.J. The decline in skeletal muscle mass with aging is mainly attributed to a reduction in type II muscle fiber size Exp. Gerontol. 2013 48 492 498 10.1016/j.exger.2013.02.012 23425621 37. Mitchell W.K. Williams J. Atherton P. Larvin M. Lund J. Narici M. Sarcopenia, dynapenia, and the impact of advancing age on human skeletal muscle size and strength; a quantitative review Front. Physiol. 2012 3 260 10.3389/fphys.2012.00260 22934016 38. Santilli V. Bernetti A. Mangone M. Paoloni M. Clinical definition of sarcopenia Clin Cases Miner Bone Metab. 2014 11 177 180 10.11138/ccmbm/2014.11.3.177 25568649 39. Rogers M.A. Evans W.J. Changes in skeletal muscle with aging: Effects of exercise training Exerc. Sport Sci. Rev. 1993 21 65 102 10.1249/00003677-199301000-00003 8504850 40. Kitamura A. Seino S. Abe T. Nofuji Y. Yokoyama Y. Amano H. Nishi M. Taniguchi Y. Narita M. Fujiwara Y. Sarcopenia: Prevalence, associated factors, and the risk of mortality and disability in Japanese older adults J. Cachexia Sarcopenia Muscle 2021 12 30 38 10.1002/jcsm.12651 33241660 41. Chang C.I. Huang K.C. Chan D.C. Wu C.H. Lin C.C. Hsiung C.A. Hsu C.C. Chen C.Y. The impacts of sarcopenia and obesity on physical performance in the elderly Obes. Res. Clin. Pract. 2015 9 256 265 10.1016/j.orcp.2014.08.003 25175711 42. Fukuoka Y. Narita T. Fujita H. Morii T. Sato T. Sassa M.H. Yamada Y. Importance of physical evaluation using skeletal muscle mass index and body fat percentage to prevent sarcopenia in elderly Japanese diabetes patients J. Diabetes Investig. 2019 10 322 330 10.1111/jdi.12908 43. Okumura Y. Ito H. Kobayashi M. Mayahara K. Matsumoto Y. Hirakawa J. Prevalence of diabetes and antipsychotic prescription patterns in patients with schizophrenia: A nationwide retrospective cohort study Schizophr. Res. 2010 119 145 152 10.1016/j.schres.2010.02.1061 20304611 44. Bauer J. Biolo G. Cederholm T. Cesari M. Cruz-Jentoft A.J. Morley J.E. Phillips S. Sieber C. Stehle P. Teta D. Evidence-based recommendations for optimal dietary protein intake in older people: A position paper from the PROT-AGE Study Group J. Am. Med. Dir. Assoc. 2013 14 542 559 10.1016/j.jamda.2013.05.021 23867520 45. Verlaan S. Maier A.B. Bauer J.M. Bautmans I. Brandt K. Donini L.M. Maggio M. McMurdo M.E.T. Mets T. Seal C. Sufficient levels of 25-hydroxyvitamin D and protein intake required to increase muscle mass in sarcopenic older adults--The PROVIDE study Clin. Nutr. 2018 37 551 557 10.1016/j.clnu.2017.01.005 28132725 46. Hara H. Nakamura Y. Hatano M. Iwashita T. Shimizu T. Ogawa T. Kanozawa K. Hasegawa H. Protein Energy Wasting and Sarcopenia in Dialysis Patients Contrib. Nephrol. 2018 196 243 249 10.1159/000485729 30041234 47. Sugawara N. Maruo K. Sugai T. Suzuki Y. Ozeki Y. Shimoda K. Someya T. Yasui-Furukori N. Prevalence of underweight in patients with schizophrenia: A meta-analysis Schizophr. Res. 2018 195 67 73 10.1016/j.schres.2017.10.017 29054486 48. Brobakken M.F. Nygard M. Wang E. Physical Health Impairment and Exercise as Medicine in Severe Mental Disorders: A Narrative Review Sports Med. Open 2022 8 115 10.1186/s40798-022-00490-3 36107334 49. Tanioka R. Ito H. Takase K. Kai Y. Sugawara K. Tanioka T. Locsin R. Tomotake M. Usefulness of 2D Video Analysis for Evaluation of Shoulder Range of Motion during Upper Limb Exercise in Patients with Psychiatric Disorders J. Med. Investig. 2022 69 70 79 10.2152/jmi.69.70 35466149 |
PMC10000556 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051077 foods-12-01077 Article Validation of an LC-MS/MS Method for the Determination of Abscisic Acid Concentration in a Real-World Setting Schiano Elisabetta Conceptualization Methodology Investigation Data curation Writing - original draft Writing - review & editing Visualization 1+ Neri Ilaria Conceptualization Methodology Investigation Data curation Writing - original draft Writing - review & editing Visualization 1+ Maisto Maria Conceptualization Methodology Investigation Data curation Writing - original draft Writing - review & editing Visualization 1 Novellino Ettore Conceptualization Methodology Validation Investigation Visualization Supervision Funding acquisition 2 Iannuzzo Fortuna Investigation Visualization 1 Piccolo Vincenzo Investigation Visualization 1 Summa Vincenzo Conceptualization Methodology Investigation Visualization Supervision 1 Grumetto Lucia Conceptualization Methodology Validation Investigation Data curation Writing - original draft Writing - review & editing Visualization Supervision 1* Tenore Gian Carlo Conceptualization Methodology Validation Investigation Data curation Writing - original draft Visualization Supervision Funding acquisition 1 Gortzi Olga Academic Editor Chinou Ioanna B. Academic Editor 1 Department of Pharmacy, University of Naples Federico II, via Domenico Montesano 59, 80131 Naples, Italy 2 Department of Medicine and Surgery, Catholic University of the Sacred Heart, 00168 Rome, Italy * Correspondence: [email protected] + These authors contributed equally to this work. 03 3 2023 3 2023 12 5 107730 12 2022 20 2 2023 24 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). One of the most relevant aspects in evaluating the impact of natural bioactive compounds on human health is the assessment of their bioavailability. In this regard, abscisic acid (ABA) has attracted particular interest as a plant-derived molecule mainly involved in the regulation of plant physiology. Remarkably, ABA was also found in mammals as an endogenous hormone involved in the upstream control of glucose homeostasis, as evidenced by its increase after glucose load. The present work focused on the development and validation of a method for the determination of ABA in biological samples through liquid-liquid extraction (LLE), followed by liquid mass spectrometry (LC-MS) of the extract. To test method suitability, this optimized and validated method was applied to a pilot study on eight healthy volunteers' serum levels to evaluate ABA concentration after consumption of a standardized test meal (STM) and the administration of an ABA-rich nutraceutical product. The results obtained could meet the demands of clinical laboratories to determine the response to a glucose-containing meal in terms of ABA concentration. Interestingly, the detection of this endogenous hormone in such a real-world setting could represent a useful tool to investigate the occurrence of impaired ABA release in dysglycemic individuals and to monitor its eventual improvement in response to chronic nutraceutical supplementation. abscisic acid waste-fruit products glucose homeostasis method validation liquid chromatography mass spectrometry O.P. Giaccio Frutta--Soc. Coop. Agricola, P.O.2019-2020-2021-2022 European Union Fund1308/2013--891/2017--892/2017 This work was financed by a grant "Selezione di scarti della filiera agroalimentare per la formula-zione di un prodotto nutraceutico a base di estratto in acido abscissico (ABA) per il controllo della glicemia" O.P. Giaccio Frutta--Soc. Coop. Agricola, P.O. annualita 2019-2020-2021-2022, that is cofunded by the European Union Fund (Reg.ti U.E. n. 1308/2013--891/2017--892/2017). pmc1. Introduction 2-cis, 4-trans-abscisic acid (ABA) is a sesquiterpenoid phytohormone synthesized via an indirect pathway from the cleavage products of carotenoids . This molecule has been studied for several decades with regard to its pivotal role as a regulator of plant growth and response to abiotic and biotic stress . Due to ABA involvement as a growth regulator, immature fruits have been found to contain the highest concentration of this phytohormone in the context of vegetal matrices. In this regard, a screening of various immature fruits derived from fruit thinning has identified thinned nectarines (TN) as the richest source of this bioactive compound . Nevertheless, ABA has sparked particular interest not only as a phytohormone commonly found in vegetables and fruits, but it has also been found in mammals as an endogenous hormone involved in the upstream control of glucose homeostasis via interaction with its specific receptor lanthionine synthase C-like 2 (LANCL2) . To date, the majority of evidence for the hypoglycemic effects of ABA in vivo has addressed a role in the stimulation of peripheral glucose uptake by increasing the expression and translocation of glucose transporter 4 (GLUT4) . In addition, it is noteworthy to remark that in patients with type 2 diabetes mellitus (T2DM) or gestational diabetes, a decreased release of ABA have been found following a glucose load . This evidence further strengthens the importance of monitoring serum concentrations of ABA in individuals with altered glucose metabolism and supplementing them with plant-based exogenous sources of ABA. In this context, several studies involving both animal and human models demonstrated the significant beneficial effects of ABA-containing nutraceuticals on the glycemic profile in prediabetic and diabetic subjects, in association with an insulin-sparing mechanism of action . In virtue of its insulin-independent mechanism of action , ABA supplementation may be indicated as a useful approach to improve glucose tolerance in individuals with insulin deficiency in and/or insulin resistance. In this regard, there is a growing scientific consensus that sustained stimulation of insulin release from pancreatic b-cells under conditions of chronic hyperglycemia may ultimately contribute to their depletion . In view of this evidence, hypoglycemic molecules able to decrease glycemia without increasing insulinemia are highly desirable as they could improve the survival and function of pancreatic b-cells. On the other hand, although a wide variety of bioactive compounds of natural origin have been tested for their beneficial potential in the control of diabetic conditions , the evaluation of their bioavailability still represents a crucial aspect . Identification of ABA as a plant hormone is usually performed with various methods, mainly in plant matrices, such as gas chromatography/mass spectrometry (GC/MS) and immunological assay, i.e., enzyme-linked immunosorbent assay (ELISA) . Although these methods are able to assess ABA concentration levels, they are affected by some disadvantages. For instance, ELISA assay requires a long preparation time and has low specificity and reproducibility, while GC/MS requires derivatization of the sample . Based on such considerations, the present work focused on the development and validation of a method for the determination of ABA by liquid mass spectrometry analysis (LC-MS), through liquid-liquid extraction (LLE) in a biological matrix, i.e., serum. Subsequently, the optimized and validated method was applied to test its suitability on serum samples from eight healthy volunteers that consumed a standardized test meal (STM) with the concomitant supplementation with a nutraceutical product based on TN rich in ABA, to test the method in a real-world setting. Finally, the glycemic and insulinemic response in the above-mentioned subjects was evaluated in association with ABA serum levels at different time points of analysis. 2. Materials and Methods 2.1. Study Design 2.1.1. Participants and Standardized Test Meal Composition Briefly, healthy subjects of both sexes were recruited in May 2019 by Samnium Medical Cooperative (Sant'Agata De' Goti, Italy) as a subset of volunteers participating in a randomized clinical trial. The volunteers' letter of intent, the protocol, and the synoptic documents of the study were submitted to the Scientific Ethics Committee of AO Rummo Hospital (Benevento, Italy). The study was approved by the committee (protocol no. 28, 15 May 2017) and was conducted in accordance with the Helsinki Declaration of 1964 (as revised in 2000). The study was listed on the ISRCTN registry (www.isrctn.com, accessed on 24 June 2022) with ID ISRCTN16732651. A total of 8 healthy subjects aged 18-83 years were invited to participate. Exclusion criteria were diabetes mellitus (DM) type 1 and type 2, liver, heart, or renal disease, drug therapy or intake of dietary supplements containing ABA, underweight (body mass index < 18.5 kg/m2), pregnancy or suspected pregnancy, birch pollen allergy. All participants received oral and written information about the study before giving written informed consent. Before inclusion in the study, volunteers were subjected to self-reporting questionnaires involving the following items: residence, occupation, smoking status, alcohol consumption, drug administration, and dietary habits. The volunteers meeting the inclusion criteria (body mass index 27-35 kg/m2; waist circumference, men >= 102 cm and women >= 88 cm) were assigned to consume a standardized test meal (STM), immediately after the administration of TN (1 g, lyophilized) containing 15 mg of ABA, as reported in our previous work . TN treatment was self-administered as a tablet. The STM composition consisted of white bread (100 g) with 50 g of jam and 100 g of mozzarella and 200 mL of fruit juice. These amounts were chosen based on indications of a balanced meal, as they provided 50% of calories from carbohydrates, 20% from protein, and 30% from fat . 2.1.2. Experimental Procedures At the beginning of the study, the height, body weight, and waist circumference (WC) of all patients were measured and the Body Mass Index (BMI) was determined. Glucometabolic parameters were determined before the STM consumption as baseline, except for fasting plasma glucose (FPG) and fasting plasma insulin (FPI), which were evaluated before and after consuming the STM. After a 12-h fasting period, blood samples were collected to measure FPG, FPI, triglycerides (TG), total plasma cholesterol (TC), lipoprotein-cholesterol (HDL-C), alanine aminotransferase (ALT), aspartate aminotransferase (AST), and glycated hemoglobin (HbA1c). The concentration of the above-mentioned parameters was assayed by enzymatic colorimetric methods (Diacron International, Grosseto, Italy). The Friedewald formula was used to calculate LDL cholesterol levels. Plasma insulin levels were measured by ELISA (DIA-source ImmunoAssay S.A., Nivelles, Belgium) on a Triturus analyzer (Diagnostic Grifols S.A., Barcelona, Spain). HbA1c was measured using a commercially available kit (InterMedical s.r.l, Grassobbio, Italy). 2.2. Analytical Method 2.2.1. Chemicals and Reagents The purity of ABA as primary standard was >=98% HPLC and purchased from Sigma-Aldrich (Milan, Italy). Chromatographic-grade solvents, methanol, formic acid, and ethyl acetate were used (minimum purity 99.9%) and purchased by Sigma Aldrich, (Milan, Italy) as well as internal standard (IS), bis 4,4'- Sulfonyldiphenol, (BPS), (minimum purity 98%). Ultra-purified water Milli Q was produced in-house (conductivity 0.055 mS cm-1 at 25 degC, resistivity equals 18.2 MO*cm). 2.2.2. Real Sample Preparation and Extraction Vacu-test(r) tubes were employed to collect blood samples, collected from the antecubital vein (5 mL); the samples were immediately centrifuged at 2200 rpm for 20 min and the supernatant was frozen and stored at -80 degC until processing. Both samples, synthetic and real samples, underwent liquid-liquid extraction (LLE). Briefly, the sample preparation was performed according to the following procedures: 75 mL of serum were transferred to a 2 mL vial, spiked into 40 mL of BPS 100 ppb solution, to achieve a final concentration of 40 ppb, with the addition of 340 mL of methanol and 2 mL of 12 N HCl solution. Each sample was successively vortexed and stored in ice for 2 min. Afterwards, the samples were added to 500 mL ethyl acetate, vortexed, and finally centrifuged at 10.000 rpm for 5 min at 4 degC. The supernatant (a fixed volume of 700 mL) was transferred to a 4 mL vial, dried in SavantTM SpeedVacTM (Thermo ScientificTM, Hyannis, MA, USA) and stored until the analysis. Dried samples were dissolved in 50 mL of CH3OH:H2O 50/50 v/v, vortexed, and after 45 min to facilitate the dissolution, another 50 mL of CH3OH:H2O 50/50 v/v was added. The samples were again centrifuged at 3.500 rpm for 5 min and the supernatant was transferred to a 1.5 mL glass insert and injected into liquid mass spectrometry (LC-MS). BPS was chosen for its lipophilicity feature as an internal standard (IS) to assess the recovery of each extraction. 2.2.3. Equipment Analytical determination was performed on an Ultimate 3000 LC system (Dionex/Thermo ScientificTM, San Jose, CA, USA) coupled to a linear ion trap LTQ XLTM (Thermo ScientificTM, San Jose, CA, USA), with an electrospray ionization source. The separation was performed on Luna(r) Omega 3 mm Polar C18 column (100 x 2.1 mm) (Phenomenex Torrance USA, Torrance, CA, USA). Tuning and data acquisition were carried out using Xcalibur and quantification using Qual Broswer software 4.4 version. 2.2.4. LC-MS/MS Conditions The samples, 5 mL of each, were injected from the Autosampler (Ultimate 300) and analyzed under the following chromatographic conditions: eluent A aqueous added to 0.1% v/v formic acid and eluent B acetonitrile, added to 0.1% v/v formic acid, flow rate set to 0.4 mL min-1, at a room temperature of 35 +- 2 degC. Gradient elution was accomplished as follows: 0-2.0 min, 5% B; 2.0-9.0 min, 95% B; 9.0-12.0 min, 95% B; 12.1-16.0 min, 5% B. All mobile phases were vacuum-filtered through 0.45 mm nylon membranes (Millipore(r)(r), Burlington, MA, USA). The electrospray ionization (ESI) mass spectrometer (MS) was operated in negative ion mode using selective reaction monitoring (SRM) with nitrogen as the nebulizer, auxiliary, collision, and curtain gas. The main working source/gas parameters of mass spectrometer were optimized and maintained as follows: curtain gas, 8; nebulizer gas, 8. The instrumental parameters employed were as follows: ESI spray voltage in the negative-ion mode, 4 kV; sheath gas flow-rate, 70 arb; auxiliary gas flow-rate, 20 arb; capillary voltage, -38 V; capillary temperature, 350 degC; and tube lens, 95 V. ABA was monitored as [M-H]- ion according to its m/z values. 2.2.5. Calibration Curve and Linearity European validation guidelines were followed to validate the method . Stock solutions of ABA were obtained dissolving the reference standard in 100% methanol to obtain a final concentration of 2.000 ppm. Five solutions with different concentrations (40 ppb, 20 ppb, 10 ppb, 4 ppb, 2 ppb) were prepared by diluting this stock. The linearity ranges were tested using the average peak areas against the concentration (ppm) of ABA. Linear regression analysis and calibration curve parameters (Coefficient of Determination R2, slope, and intercept) were back-calculated from the peak areas using the regression line by the method of least squares, and mean accuracy values were determined . 2.2.6. Limits of Detection (LOD) and Quantification (LOQ) LOD and LOQ were estimated as the concentrations providing signals equal to 3 and 10 times, respectively. They were calculated based on the following equations: LOD = SD3/S and LOQ = SD10/S , where SD is the standard deviation of the intercept response with the y-axis of the calibration curves, and S is the slope of a calibration curve. The spike level was 2 ppb in the appropriate range using a concentration and was assessed by running the measurement ten times. 2.2.7. Precision and Accuracy The method's precision was evaluated by running five replicates of the sample repeated in the same day and in two different days to cover both intra-day and inter-day precision, expressed as relative standard deviation (RSD%). Repeatability was assessed using the nominal concentration of ABA (2 ppb). The accuracy of this method was determined considering samples spiked with 2-40 ppb of ABA (quality control samples, QCs) and evaluated at each level in triplicate, and reported as a percentage of the nominal value. 2.2.8. Selectivity Serum working calibration standards were prepared using sera already present in the archive of our laboratory and processed for other research, to assess the absence of ABA and that any signal interfered with the retention time of ABA. These sera, considered as blanks, were also employed to optimize the extraction process. 2.2.9. Carry-Over Carry-over effect of the method was evaluated by injecting methanol solvent after running the highest concentrated samples of ABA spiked in the serum (three times) and observing the occurrence of signals within the retention windows of the target chemicals. 2.2.10. Matrix Effect The matrix effect was investigated by calculating the ratio of the peak area in the presence of the matrix (matrix spiked with ABA post extraction) to the peak area in the absence of the matrix (ABA in methanol) . The serum matrix blank was spiked with the analyte at each concentration of the linear range (2 ppb, 4 ppb, 10 ppb, 20 ppb, and 40 ppb). The ratio was calculated as follows:(1) Matrix effect %=peak area in presence of matrixarea in absence of matrix*100 2.2.11. Recovery The recovery was assessed by evaluating the relative abundance of the BPS peak (I.S.) spiked before the extraction procedure and calculated as follows:(2) Recovery %=found concentrationstandard concentration*100 The results of the real samples were corrected for the recovery. 3. Results 3.1. Anthropometric and Glucometabolic Parameters The characteristics of the patient population at baseline are shown in Table 1. A total ofeight healthy adults (three men and five women) aged 18 to 45 years with a BMI between 18 and 25 kg/m2 met the inclusion criteria and were therefore eligible to participate in the study. The group was well balanced in terms of demographic and clinical factors. 3.2. Two-Hour Glycemic and Insulinemic Responses to Standardized Test Meal Following the STM, which was preceded by administration of the nutraceutical supplement containing ABA, mean plasma glucose levels reached a peak at 30 min and gradually decreased to pre-prandial levels by 120 min. According to the plasma glucose response curve, the post-prandial insulinemic response curve peaked at 30 min and gradually declined to the pre-prandial level by 120 min . A similar trend can be observed for serum ABA concentrations after the consumption of STM and ABA-rich nutraceutical product in volunteers under our investigation, as shown in Figure 2. 3.3. Optimization of Chromatographic Method Different "synthetic" samples with known ABA concentrations, i.e., methanolic solutions and serums spiked with ABA, were used for the method development. These samples were subjected to the above-mentioned method in order to evaluate the efficiency in isolating and detecting abscisic acid in the context of complex biological matrices. The proposed method of extraction and quantification of ABA was easy to handle and sensitive to the analysis in serum matrix, optimizing the method after several changes in operating. For the extraction procedures, there were distinct organic solvents in various percentages with water. Ethyl acetate as an extraction solvent was the most efficient solvent to extract ABA from the serum matrix (data not shown). The spike levels (40.0 ppb and 2.0 ppb) were in the recommended range, i.e., calculated LOD < spike level < 10 x calculated LOD. For LC-MS analysis, we optimized the method using different stationary reversed-phases (Luna(r)(r) Omega 3 mm Polar C18 column (100 x 2.1 mm) (Phemomenex Torrance USA) and an Inertsil ODS-3 column (2.1 mm x 100 mm, 5 mm) (Torrance, CA, USA), and by a varying gradient elution program, to achieve an adequate resolution for the two analytes from the interferents. Optimal transitions were obtained for ABA (C15H20O4, MW: 264.32 g/mol) at m/z 152.000, and for BPS (C12H10O4S, MW: 250.27 g/mol) at m/z 107.000. The linear R-squared values (r2 = 0.9981) show a good linearity in the range of the calibration curves performed in the serum matrix from 40 ppb to 2 ppb. The sensitivity of the developed method is appreciable from the listed LOD and LOQ parameters, with values of 1.59 ppb and 5.31 ppb, respectively. The RSD% of within-run precision was 2.30%, while the RSD% between-run precision was 12.01%. Repeatability was performed using the repeatedly frozen and thawed ABA samples, and we did not observe any differences in the raw data and degradation products. Recovery from the serum matrix, evaluated at high and low spiking concentrations (40 ppb and 2.0 ppb), resulted in 70.3%. Matrix effect was 39.97% and variations in the experimental parameters did not result in any appreciable change in the method performance. Table 2 summarizes all method validation parameters. These results demonstrate that the analytical method developed provides a reliable response relevant to the analysis of ABA in such a complex biological matrix. Selectivity is the ability of an analytical method to differentiate and quantify the analyte in the presence of other components in the sample. The selectivity of the method was evaluated by analyzing a blank sample, compared to a blank sample spiked with the lower limit of quantification LOQ (ABA equal to 2.00 ppb). As can be seen in Figure 3, the selectivity of this method was good. 4. Discussion In the present work, a method for the determination of ABA in biological samples by liquid-liquid extraction (LLE), followed by liquid mass spectrometry (LC-MS) of the extract, was developed and validated. The above-reported method has significant advantages, as it does not require expensive operations, in terms of procedures and amounts of solvents used, and leads to results with a good level of accuracy, reproducibility, LOD and LOQ values. These results are better in terms of sensitivity than those achieved by Reverse-Phase HPLC-DAD analysis on food and beverage matrices . Moreover, to the best of our knowledge, the scientific literature reports methods for determination of ABA by LC/MS, but in a matrix other than serum, such as in Arabidopsis thaliana , Rose Leaf Samples , and fresh Oryza sativa tissues . The scientific works analyzing ABA in the serum matrix are not focused on the validation method, and therefore, do not report validation parameters for comparison . Anyway, new strategies to detect ABA with high sensitivity are under development, as fluorescent probes, but performed on plant tissues . Moreover, the application of the optimized method on serum samples of healthy volunteers who consumed a STM together with a nutraceutical product rich in ABA allowed us to evaluate its applicability in a suitable biological model. Accordingly, the STM composition of the present work provided 50% of calories from carbohydrates, 20% from protein, and 30% from fat, in agreement with the guidelines for balanced nutrition . In this manner, the glycemic and insulinemic response, together with the increase in plasmatic ABA, was evaluated in the closest to real-life setting. The LC-MS analysis performed on the serums obtained from the eight volunteers showed different ABA levels at each time point. As observed in Figure 2, the found data confirmed the involvement of this endogenous hormone in the human response to glucose-containing foods. For all subjects, indeed, the serum ABA levels reached the highest concentration 30 min after the consumption of the STM and the nutraceutical product based on TN. In this regard, various studies carried out on human serums attempted to identify and quantify ABA levels, by performing different isolation and detection methods . Specifically, plasma ABA levels have been shown to increase in normal glucose tolerant (NGT) subjects following an oral glucose load , but not in patients with T2D or in pregnant women with gestational diabetes mellitus (GDM). On the other hand, resolution of GDM one month after childbirth is associated with a restoration of the ABA response to glucose load . Interestingly, a significant increase in ABA was observed in obese patients after biliopancreatic diversion (BPD), a bariatric surgery performed to reduce body weight and improve glucose tolerance, compared to pre-surgery levels . Another observed difference between T2D and NGT individuals was related to fasting ABA values, which were significantly higher in T2D compared to NGT subjects (1.15 vs. 0.66 as median values, respectively). Nevertheless, the distribution of ABA values was found to be normal in NGT but not in T2D patients . These alterations could be due to the heterogeneity of ABA-related dysfunction that occurs in T2D, such as the inability of ABA to increase in response to hyperglycemia or resistance to the activity of ABA. Overall, these observations suggest a role for ABA as a key hormone involved in the management of glucose homeostasis and highlight the importance of monitoring ABA levels in these groups of individuals. Notably, based on reports about daily consumption of fruits and vegetables containing ABA, epidemiological evidence indicates that the majority of the population assumes a very low intake of ABA from dietary sources . Due to the multiple positive health effects attributed to the role of ABA , interest in supplementing this bioactive molecule through the administration of nutraceutical products rich in ABA is increasing over time, also in view of the nanomolar blood concentrations of this hormone required for its efficacy. 5. Conclusions In conclusion, we herein developed and validated a method for the extraction and LC-MS/MS analysis of ABA in biological samples. Even if limited by the small sample size, requiring therefore confirmation through larger clinical evaluation, an added value is represented by the successful application of this method to real samples, which allowed the evaluation of ABA serum changes after the consumption of STM and an ABA-rich nutraceutical product. Overall, the results shown could provide a starting point for determining the response to a glucose-containing meal in clinical practice, in terms of ABA concentration. Indeed, serum detection of this endogenous hormone may be considered as a marker to assess the presence of an impaired ABA response in dysglycemic subjects. Undoubtedly, the use of this analysis would be of great interest for clinical trials involving the chronic administration of ABA-rich nutraceutical supplements with hypoglycemic potential. Acknowledgments The assistance of the staff is gratefully appreciated; a special thanks to P. Imperato for her contribution. Author Contributions Conceptualization, E.S., M.M., I.N., E.N., V.S., L.G. and G.C.T.; methodology, E.S., M.M., I.N., E.N., V.S., G.C.T. and L.G.; validation, E.N., L.G. and G.C.T.; investigation, E.S., M.M., I.N., E.N., F.I., V.P., V.S., L.G. and G.C.T.; data curation, E.S., I.N., L.G. and G.C.T.; writing--original draft preparation, E.S., M.M., I.N. and L.G; writing--review and editing, E.S., I.N., L.G. and G.C.T.; visualization, E.S., M.M., I.N., E.N., F.I., V.P., V.S., L.G. and G.C.T.; supervision, E.N., V.S., L.G. and G.C.T.; funding acquisition, E.N. and G.C.T. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of AO Rummo Hospital (Benevento, Italy) (protocol ndeg28 of 15/05/2017). This study is listed on the ISRCTN registry (www.isrctn.com, accessed on 30 November 2022) with ID ISRCTN16732651 accessed on 30 November 2022). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data used to support the findings of this study are included within the article. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Change in postprandial plasma glucose and insulin concentration in healthy adults (n = 8). Data are expressed as mean +- SEM. Figure 2 ABA concentration change after the consumption of STM and ABA-rich nutraceutical product in healthy adults (n = 8). Data are expressed as mean +- SEM. Figure 3 Chromatogram of (a) BPS in solvent, (b) ABA in solvent, (c) spiked BPS in serum (2 ppb), (d) spiked ABA in serum (2 ppb), and (e) blank serum sample. foods-12-01077-t001_Table 1 Table 1 Baseline anthropometric and glucometabolic parameters of study participants. Characteristics Study Participants (n = 8) Demographic and anthropometric parameters Male sex (No (%)) 3 (37.5%) White ethnicity (No (%)) 8 (100%) Age (years) 34 +- 3.7 Height (m) 1.6 +- 0.9 Weight (kg) 71.2 +- 6.9 BMI (kg/m2) 24.6 +- 3.3 WC (cm) 89.4 +- 5.4 Clinical parameters HbA1c (%) 6.5 +- 0.4 Triglycerides (mg/dL) 105.2 +- 14.5 Total cholesterol (mg/dL) 171.3 +- 11.9 HDL-C (mg/dL) 46.7 +- 5.4 LDL-C (mg/Dl) 89.5 +- 7.6 AST (UI/L) 25.6 +- 3.9 ALT (UI/L) 18.7 +- 2.8 Creatinine (mg/dL) 0.9 +- 0.1 Data are expressed as mean +- standard deviation. Abbreviations: AST: aspartate aminotransferase; ALT: alanine aminotransferase; BMI: Body Mass Index; Cre: creatinine; F: females; FPG: fasting plasma glucose; FPI: fasting plasma insulin; HbA1c: glycated hemoglobin; HDL-C: high density lipoprotein-cholesterol; LDL-C: low density lipoprotein-cholesterol; M: males; TG: triglycerides; WC: waist circumference. foods-12-01077-t002_Table 2 Table 2 Summary of method validation parameters. Recovery was reported for each concentration of the linear range, and also reported as mean %. Linear Range (ppb) Slope Intercept R2 Repeatability (n = 5) RSD % Intermediate Precision (n = 10) RSD% LOQ (ppb) LOD (ppb) Matrix Effect 40.0-2.0 92.84 -32.45 0.9981 2.30 12.01% 5.31 1.59 39.97% Spiking level (ppb) 2 4 10 20 40 Recovery (%) 71.8 73.3 77.7 65.5 63.5 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Finkelstein R. Abscisic Acid Synthesis and Response Arab. Book 2013 11 e0166 10.1199/tab.0166 24273463 2. Brookbank B.P. Patel J. Gazzarrini S. Nambara E. Role of basal aba in plant growth and development Genes 2021 12 1936 10.3390/genes12121936 34946886 3. Hey S.J. Byrne E. Halford N.G. The interface between metabolic and stress signalling Ann. Bot. 2010 105 197 203 10.1093/aob/mcp285 20007158 4. Leng P. Yuan B. Guo Y. Chen P. The role of abscisic acid in fruit ripening and responses to abiotic stress J. Exp. Bot. 2014 65 4577 4588 10.1093/jxb/eru204 24821949 5. Liu N. Effects of IAA and ABA on the Immature Peach Fruit Development Process Hortic. Plant J. 2019 5 145 154 10.1016/j.hpj.2019.01.005 6. Tenore G.C. Caruso D. D'avino M. Buonomo G. Caruso G. Ciampaglia R. Schiano E. Maisto M. Annunziata G. Novellino E. A pilot screening of agro-food waste products as sources of nutraceutical formulations to improve simulated postprandial glycaemia and insulinaemia in healthy subjects Nutrients 2020 12 1292 10.3390/nu12051292 32370061 7. Magnone M. Spinelli S. Begani G. Guida L. Sturla L. Emionite L. Zocchi E. Abscisic Acid Improves Insulin Action on Glycemia in Insulin-Deficient Mouse Models of Type 1 Diabetes Metabolites 2022 12 523 10.3390/metabo12060523 35736456 8. Magnone M. Sturla L. Guida L. Spinelli S. Begani G. Bruzzone S. Fresia C. Zocchi E. Abscisic acid: A conserved hormone in plants and humans and a promising aid to combat prediabetes and the metabolic syndrome Nutrients 2020 12 1724 10.3390/nu12061724 32526875 9. Bruzzone S. Bodrato N. Usai C. Guida L. Moreschi I. Nano R. Antonioli B. Fruscione F. Magnone M. Scarfi S. Abscisic acid is an endogenous stimulator of insulin release from human pancreatic islets with cyclic ADP ribose as second messenger J. Biol. Chem. 2008 283 32188 32197 10.1074/jbc.M802603200 18784081 10. Spinelli S. Begani G. Guida L. Magnone M. Galante D. D'Arrigo C. Scotti C. Iamele L. De Jonge H. Zocchi E. LANCL1 binds abscisic acid and stimulates glucose transport and mitochondrial respiration in muscle cells via the AMPK/PGC-1a/Sirt1 pathway Mol. Metab. 2021 53 101263 10.1016/j.molmet.2021.101263 34098144 11. Bruzzone S. Ameri P. Briatore L. Mannino E. Basile G. Andraghetti G. Grozio A. Magnone M. Guida L. Scarfi S. The plant hormone abscisic acid increases in human plasma after hyperglycemia and stimulates glucose consumption by adipocytes and myoblasts; The plant hormone abscisic acid increases in human plasma after hyperglycemia and stimulates glucose consumption by adipocytes and myoblasts FASEB J. Res. Commun. 2012 26 1251 1260 10.1096/fj.11-190140 12. Sturla L. Mannino E. Scarfi S. Bruzzone S. Magnone M. Sociali G. Booz V. Guida L. Vigliarolo T. Fresia C. Abscisic acid enhances glucose disposal and induces brown fat activity in adipocytes in vitro and in vivo Biochim. Biophys. Acta-Mol. Cell Biol. Lipids 2017 1862 131 144 10.1016/j.bbalip.2016.11.005 27871880 13. Magnone M. Emionite L. Guida L. Vigliarolo T. Sturla L. Spinelli S. Buschiazzo A. Marini C. Sambuceti G. De Flora A. Insulin-independent stimulation of skeletal muscle glucose uptake by low-dose abscisic acid via AMPK activation Sci. Rep. 2020 10 1454 10.1038/s41598-020-58206-0 31996711 14. Schiano E. Maisto M. Piccolo V. Novellino E. Annunziata G. Ciampaglia R. Montesano C. Croce M. Caruso G. Iannuzzo F. Beneficial Contribution to Glucose Homeostasis by an Agro-Food Waste Product Rich in Abscisic Acid: Results from a Randomized Controlled Trial Foods 2022 11 2637 10.3390/foods11172637 36076823 15. Ameri P. Bruzzone S. Mannino E. Sociali G. Andraghetti G. Salis A. Ponta M.L. Briatore L. Adami G.F. Ferraiolo A. Impaired increase of plasma abscisic acid in response to oral glucose load in type 2 diabetes and in gestational diabetes PLoS ONE 2015 10 e0115992 10.1371/journal.pone.0115992 25723556 16. Atkinson F.S. Villar A. Mul A. Zangara A. Risco E. Smidt C.R. Hontecillas R. Leber A. Bassaganya-riera J. Abscisic acid standardized fig (Ficus carica) extracts ameliorate postprandial glycemic and insulinemic responses in healthy adults Nutrients 2006 11 1757 10.3390/nu11081757 17. Guri A.J. Hontecillas R. Si H. Liu D. Bassaganya-Riera J. Dietary abscisic acid ameliorates glucose tolerance and obesity-related inflammation in db/db mice fed high-fat diets Clin. Nutr. 2007 26 107 116 10.1016/j.clnu.2006.07.008 17000034 18. Magnone M. Ameri P. Salis A. Andraghetti G. Emionite L. Murialdo G. De Flora A. Zocchi E. Microgram amounts of abscisic acid in fruit extracts improve glucose tolerance and reduce insulinemia in rats and in humans FASEB J. 2015 29 4783 4793 10.1096/fj.15-277731 26243865 19. Derosa G. Maffioli P. D'angelo A. Preti P.S. Tenore G. Novellino E. Abscisic acid treatment in patients with prediabetes Nutrients 2020 12 2931 10.3390/nu12102931 32987917 20. Muoio D.M. Newgard C.B. Mechanisms of disease: Molecular and metabolic mechanisms of insulin resistance and b-cell failure in type 2 diabetes Nat. Rev. Mol. Cell Biol. 2008 9 193 205 10.1038/nrm2327 18200017 21. Ota A. Ulrih N.P. An overview of herbal products and secondary metabolites used for management of type two diabetes Front. Pharmacol. 2017 8 436 10.3389/fphar.2017.00436 28729836 22. Egbuna C. Awuchi C.G. Kushwaha G. Rudrapal M. Patrick-Iwuanyanwu K.C. Singh O. Odoh U.E. Khan J. Jeevanandam J. Kumarasamy S. Bioactive Compounds Effective Against Type 2 Diabetes Mellitus: A Systematic Review Curr. Top. Med. Chem. 2021 21 1067 1095 10.2174/1568026621666210509161059 33966619 23. Iannuzzo F. Piccolo V. Novellino E. Schiano E. Salviati E. Summa V. Campiglia P. Tenore G.C. Maisto M. A Food-Grade Method for Enhancing the Levels of Low Molecular Weight Proanthocyanidins with Potentially High Intestinal Bioavailability Int. J. Mol. Sci. 2022 23 13557 10.3390/ijms232113557 36362344 24. Rein M.J. Renouf M. Cruz-Hernandez C. Actis-Goretta L. Thakkar S.K. da Silva Pinto M. Bioavailability of bioactive food compounds: A challenging journey to bioefficacy Br. J. Clin. Pharmacol. 2013 75 588 602 10.1111/j.1365-2125.2012.04425.x 22897361 25. Arif M.A. Alseekh S. Harb J. Fernie A. Frank W. Abscisic acid, cold and salt stimulate conserved metabolic regulation in the moss Physcomitrella patens Plant Biol. 2018 20 1014 1022 10.1111/plb.12871 29943488 26. Fang Q. Jiang T. Xu L. Liu H. Mao H. Wang X. Jiao B. Duan Y. Wang Q. Dong Q. A salt-stress-regulator from the Poplar R2R3 MYB family integrates the regulation of lateral root emergence and ABA signaling to mediate salt stress tolerance in Arabidopsis Plant Physiol. Biochem. 2017 114 100 110 10.1016/j.plaphy.2017.02.018 28285084 27. Manore M.M. Exercise and the Institute of Medicine Recommendations for Nutrition Curr. Sport. Med. Rep. 2005 4 193 198 10.1097/01.CSMR.0000306206.72186.00 28. European Medicines Agency (EMA) Status of EMEA Scientific Guidelines and European Pharmacopoeia Monographs and Chapters in the Regulatory Framework Applicable to Medicinal Products European Medicines Agency (EMA) Amsterdam, The Netherlands 2008 2 4 29. Maisto M. Schiano E. Novellino E. Piccolo V. Iannuzzo F. Salviati E. Summa V. Annunziata G. Tenore G.C. Application of a Rapid and Simple Technological Process to Increase Levels and Bioccessibility of Free Phenolic Compounds in Annurca Apple Nutraceutical Product Foods 2022 11 1453 10.3390/foods11101453 35627023 30. Maisto M. Piccolo V. Novellino E. Schiano E. Iannuzzo F. Ciampaglia R. Summa V. Tenore G.C. Optimization of Ursolic Acid Extraction in Oil from Annurca Apple to Obtain Oleolytes with Potential Cosmeceutical Application Antioxidants 2023 12 224 10.3390/antiox12020224 36829781 31. Pereira M.N. Matos B.N. Gratieri T. Cunha-Filho M. Gelfuso G.M. Development and validation of a simple chromatographic method for simultaneous determination of clindamycin phosphate and rifampicin in skin permeation studies J. Pharm. Biomed. Anal. 2018 159 331 340 10.1016/j.jpba.2018.07.007 30025298 32. Di Lorenzo R. Neri I. Russo G. Laneri S. Grumetto L. Tracking Down of a Selected Panel of Parabens: A Validated Method to Evaluate Their Occurrence in Skin Layers Cosmetics 2022 9 102 10.3390/cosmetics9050102 33. Whelan L.C. Geary M. Healy J. A Novel, Simple Rapid Reverse-Phase HPLC-DAD Analysis, for the Simultaneous Determination of Phenolic Compounds and Abscisic Acid Commonly Found in Foodstuff and Beverages J. Chromatogr. Sci. 2022 60 648 654 10.1093/chromsci/bmab114 34553229 34. Xiong D.M. Liu Z. Chen H. Xue J.T. Yang Y. Chen C. Ye L.M. Profiling the dynamics of abscisic acid and ABA-glucose ester after using the glucosyltransferase UGT71C5 to mediate abscisic acid homeostasis in Arabidopsis thaliana by HPLC-ESI-MS/MS J. Pharm. Anal. 2014 4 190 196 10.1016/j.jpha.2014.01.004 29403882 35. Bosco R. Caser M. Vanara F. Scariot V. Development of a rapid LC-DAD/FLD method for the simultaneous determination of auxins and abscisic acid in plant extracts J. Agric. Food Chem. 2013 61 10940 10947 10.1021/jf4034305 24134056 36. Bai Y.L. Cai B.D. Luo X.T. Ye T.T. Feng Y.Q. Simultaneous Determination of Abscisic Acid and Its Catabolites by Hydrophilic Solid-Phase Extraction Combined with Ultra High Performance Liquid Chromatography-Tandem Mass Spectrometry J. Agric. Food Chem. 2018 66 10906 10912 10.1021/acs.jafc.8b03820 30260222 37. Pem D. Jeewon R. Fruit and vegetable intake: Benefits and progress of nutrition education interventions-narrative review article Iran. J. Public Health 2015 44 1309 1321 26576343 38. Zocchi E. Hontecillas R. Leber A. Einerhand A. Carbo A. Bruzzone S. Tubau-Juni N. Philipson N. Zoccoli-Rodriguez V. Sturla L. Abscisic Acid: A Novel Nutraceutical for Glycemic Control Front. Nutr. 2017 4 24 10.3389/fnut.2017.00024 28660193 |
PMC10000557 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050633 healthcare-11-00633 Article Impact of the COVID-19 Pandemic on Dermatology Care in the Chilean Public Health Sector Aragon-Caqueo Diego Conceptualization Methodology Software Formal analysis Investigation Resources Data curation Writing - original draft Visualization Project administration Funding acquisition 1* Aedo Gabriel Conceptualization Methodology Validation Formal analysis Investigation Resources Data curation Writing - original draft Visualization 2 Suarez Javier Methodology Software Validation Investigation Writing - review & editing 3 Toloza Claudio Methodology Validation Investigation Data curation Writing - review & editing Supervision 3 Guglielmetti Antonio Conceptualization Formal analysis Resources Writing - review & editing Visualization Supervision Project administration 3 Granero-Molina Jose Academic Editor 1 Escuela de Medicina, Universidad de Tarapaca, Arica 1000000, Chile 2 Facultad de Ciencias Medicas, Universidad de Santiago de Chile, Santiago 8320000, Chile 3 Escuela de Medicina, Universidad de Valaparaiso, Vina del Mar 2520000, Chile * Correspondence: [email protected]; Tel.: +56-58-2205100 21 2 2023 3 2023 11 5 63324 12 2022 16 2 2023 16 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Due to the Coronavirus-19 (COVID-19) pandemic, most resources of the public health system were allocated to the increasing demand from respiratory patients. From this, it is expected that specialty consultations would decrease drastically. Access to dermatology care in the Chilean public health has been historically limited. To evaluate the impact of the pandemic on dermatology care, the total number of dermatological consultations (DCs) to the Chilean public sector in 2020 is analyzed according to sex and age range and compared with the available databases from 2017 to 2019. From this, 120,095 DCs were performed during 2020, with an incidence of 6.3 consultations per 1000 inhabitants. When compared to 2019 (n = 250,649), there was a 52.1% decrease. The regions most affected were located in the central part of Chile, which correlates with the regions most affected by the pandemic. Age and sex distributions remained similar to previous years but lower in amplitude. The month with the lowest number of consultations was April, with a gradual increase observed thereafter until December 2020. Although DCs decreased drastically in the Chilean public sector during 2020, sex and age range proportions were conserved, thus affecting all groups in a similar manner. COVID-19 pandemic dermatological consultations public health This research received no external funding. pmc1. Introduction The COVID-19 pandemic, which started in December 2019 and was declared as such by the WHO in January 2020 , was a major cause of health care system overload and resource occupation worldwide . Chile was not exempt from this situation, where despite a successful vaccination program , hospitals consistently operated close to their maximum capacity during the first year of the pandemic . Worldwide, different public health approaches were implemented to limit both the entry and the spread of the virus in the respective territories, being more or less successful, depending on the context, resources, and country preparedness . Chile opted for a mixed strategy of active screening associated with general restrictive measures, implementing mass testing, together with dynamic quarantines, which were re-evaluated weekly . The first case of COVID-19 in Chile was reported on 3 March 2020 , which gave the country an important window of time to prepare for the public health response. Then, in order to face the growing demand, all elective procedures were canceled, and new hospital facilities were promptly opened . As the pandemic progressed and hospitals reached their maximum capacity, the facilities of various medical specialties began to be occupied to cover the still-growing demand from respiratory patients, leading to the cancelation of outpatient care for multiple specialties, including dermatology. Consistent with measurements taken in other countries, for the case of dermatology, residency activities were reorganized, chronic cases were predominantly followed up using teledermatology, and in-hospital interconsultations were limited to cases where diagnostic suggestions could not be achieved using teledermatology . In addition, healthcare personnel was reassigned mostly to cover basic to intermediate wards. For outpatient consultations, similar to the experience in other countries , cases were first evaluated by a general practitioner and only urgent consultations were considered, using teledermatology as a tool to stratify the need for urgent specialist evaluation. The above-mentioned measures taken during the pandemic had important repercussions on waiting lists and access to specialized care for a population of patients who were displaced by the contingency . Moreover, Chile is a country where there has been a historical limitation in access to dermatological care in the public system, with an important, documented gap between the current supply and demand of DC . This is further deepened by the distribution of dermatologists across the country, with 64% of them concentrated in the capital city . In addition, one-third of dermatologists work in the public sector, and thus there is stronger participation of dermatologists in the private sector . Considering that the public system covers 80% of the population, and that dermatology accounts for 2.7% of the consultations to specialty care in the public health system , this distribution demonstrates wide inequities in the availability of dermatologists and access to dermatological care. The reduction of dermatological consultation in the context of the pandemic and the restrictive measures taken in 2020 further deepen these gaps in access. This study is the first country-based analysis of the consultations made to specialty care during 2020 in Chile. It is important to establish this background and to evaluate which age groups and regions of the Chilean territory were most affected. This contributes to future development strategies to mitigate this acute limitation in access to care and establishes the groundwork for targeting the most affected population. The aim of this study is to evaluate the impact of the COVID-19 pandemic on the total number of dermatology consultations (DCs) in the public health sector during 2020, evaluated according to geographic location, sex, and age range, and compared with consultations in the years before the pandemic. 2. Materials and Methods The methods used were observational, cross-sectional, and retrospective studies of the consultations performed by dermatologists at the national level, in the Public Health sector, reported by the Department of Health Statistics and Information (DEIS) of the Subsecretaria de Salud Publica of the Chilean Ministry of Health (MINSAL) during 2020, compared with the available databases from 2017, 2018, and 2019 . These consultations were contrasted with the demographic information available according to the 2017 Census and the population projections made by the National Institute of Statistics (INE) . The incidence of DCs was calculated by age group and according to the population comprising those age ranges based on the 2017 Census data and INE projections. For the rest of the data, a univariate analysis was performed with measures of central tendency, percentage, mean, and standard deviation. All analyses were performed using Stata software (Stata/SE 16.0 for macOS, Copyright 1985-2019 StataCorp LLC, College Station, TX, USA). All figures were developed using Microsoft Excel (Office 365, Microsoft Excel v16.66.1 for macOS, Copyright 1985-2022 Microsoft Crop, Redmond, DC, USA). Ethical Considerations The data analyzed were obtained from a routinely collected anonymous database from the Department of Health Statistics and Information and the National Institute of Statistics . The analyses were conducted in compliance with the Declaration of Helsinki of ethical principles for medical research. 3. Results A total of 830,724 DCs were recorded during the period 2017-2020. Of these, 120,095 were performed in 2020, with a reduction of 52.1% in 2020 as compared to 2019 . Regarding the consultations according to sex, 59% (n = 71,387) corresponded to female patients, 41% (n = 48,700) corresponded to male patients, and 8 patients (0.0067% of the sample) had no information regarding sex. Finally, the incidence of DCs in 2020 was 6.3 per 1000 population at the national level. 3.1. Consultations by Region during 2020 The Metropolitan Region, where the capital of Chile, Santiago, is located, accounted for 46% (n = 55,255) of the total number of DCs performed in 2020 (n = 120,095). Note that this region encompasses 42% of the total population of Chile . When establishing the total number of consultations in terms of the population from each region of the territory, the region with the highest number of consultations per 1000 inhabitants is the Aysen region . On the other hand, when establishing the incidence of consultations per 1000 inhabitants according to the geographical area (North, Center, or South), it is clear that the southern regions consult more than the northern and central regions of Chile . 3.2. Consultations by Age Range in 2019 Versus 2020 During 2019, the age range with the highest number of DCs was 0-4 years (n = 18,300) and 15-19 (n = 18,197), while the lowest was 40-44 years (n = 11,142). In 2020, the age range with the highest number of DCs was 20-24 years (n = 8455), 15-19 years (n = 8274), and 25-29 years (n = 8256), accounting for 21% of the total number of DCs between 15 and 29 years. At the same time, the age range with the lowest number of DCs was 5-9 years (n = 5424) . The average reduction in DCs between 2019 and 2020 for each age range was 52 +- 5.1%, with a maximum reduction of 62.2% for age ranges 5-9 years and a minimum reduction of 42.5% for age ranges 30-34 years. In terms of DCs by age range according to the population of the range, a maximum incidence was observed in the 75-79 age range, with 14.4 consultations per 1000 inhabitants. On the other hand, the age range with the lowest number of DCs as a function of its population was 40-44 years, with 4.2 consultations per 1000 inhabitants in the range . Furthermore, when analyzing the proportions from the total consultations for a specific age group, it can be found that those proportions remain the same for the ranges from 50 to 54 (5.9% of the total sample), 55 to 59 (6.2% of the total sample), and 60 to 64 (6.3% of the total sample), with small variations of +-0.41% along the other age groups . Finally, when analyzing the incidence of DCs per 1000 inhabitants of a given age group for a specific year and dividing it by the average incidence of DCs per 1000 inhabitants of that year, the distribution shown in Figure 3d was observed. 3.3. Temporal Sequence of Consultations in 2020 When analyzing DCs month by month during 2020, a decreasing trend was observed after January, reaching a maximum minimum in April, one month after the first COVID-19 case was reported in Chile. From April onwards, a sustained growth trend was observed, even when, in June, the first peak of cases of the first wave of the pandemic occurred (13 June 2020) and when hospital capacity reached the peak occupation of ICU beds at the national level (2 July 2020) . Finally, towards the end of the year, the growing trend of consultations tended to stabilize . 4. Discussion The COVID-19 pandemic has had a notable impact on access to specialized health care worldwide. The closure of specialized ambulatory care, the need to allocate both economic and human resources, as well as the occupation of hospital units of different specialties to meet the growing demand for respiratory patients, had an important impact on diagnosis, treatment, procedures performed, and dermatology patient follow-up . In 2020 in Chile, dermatology care in the public health system was reduced by more than half as compared to 2019 (52.1% reduction). This reduction can be correlated to the COVID-19 pandemic and the sanitary control measures, which restricted the movement of non-essential personnel and reallocated human resources and occupied available health care facilities to face the contingency, as previously mentioned. In contrast, it is reported that such factors are associated with the reduction in DCs by up to 80-90% . Regarding the incidence of DCs, it can be seen that in 2020 it was 6.3 DCs per 1000 inhabitants. Comparing to 2019, with 14.1 DC per 1000 inhabitants , a reduction of more than half in the incidence of DCs between both years is observed. This is also consistent with the reduction observed in the total number of consultations in absolute terms. This documented reduction also has a notable impact on public health in the Chilean scenario. Note that, before the pandemic, waiting times for DCs in the public healthcare system were on average 341 days . From this, it is expected that these waiting times have been extended even further in the context of the acute reduction in access to specialized care. Concerning consultations by region, the Metropolitan Region accounts for the vast majority of DCs. However, when these consultations are contrasted according to the population of each region, it can be observed that the northern and southern regions had a higher incidence of DCs than the central regions . This may be because the gradual implementation of restrictive measures initially affected Santiago de Chile and the central regions of the country, which experienced the highest contagion rates at the beginning of the pandemic . Furthermore, on 13 May 2020, a total quarantine was decreed for the Metropolitan Region. As cases increased in the central regions, dynamic quarantines were enforced based on the Plan Paso a paso implemented by the government at the time . Nevertheless, when the cases started to increase in the northern and southern regions, the aforementioned restrictive measures were then extended to the rest of the country . According to the results obtained, when analyzing the proportions of the consulting population by age range, these proportions remained relatively similar between 2019 and 2020, both in absolute numbers and concerning consultations per 1000 inhabitants , with the difference that in 2020 the consulting population decreased significantly . This is consistent with other studies, where, when comparing DC in 2019 to 2020, it has been reported that there were no significant differences concerning the mean ages of patients (46.26 +- 23.58 years vs. 47.06 +- 23.08), but a higher number of patients did not show up for DCs in 2020 . As reported in the results section, the population aged 15 to 29 years led the total number of DCs, accounting for 21% of the sample . However, when analyzing the incidence of DCs according to the population belonging to the age range, it was observed that patients aged 65 years and older led the dermatological consultations as a function of their population . This is important to note, since at the time, attending a DC could risk exposure to COVID-19, and older adults are a known risk group for mortality due to COVID-19 . In contrast, several studies have shown that worldwide, the average age of the consulting population was 40-50 years . Regarding the distribution by sex, there was a predominance of consultations by the female sex. The results obtained in this area are consistent with the trend widely reported before the pandemic, where the approximate proportions were between 60-70% for females and 30-40% for males . Regarding the temporal progression of the decrease in DCs , it can be observed that the month of April 2020 recorded the lowest number of DCs. This was possibly reactive to the confirmation of the first case of SARS-CoV-2 recorded in Chile on 3rd March of the same year and the subsequent declaration of a state of catastrophe and constitutional exception, decreed by the government on 18 March 2020 . Given the implications that such measures carry for both the general population and the health sector, this sharp drop in the total number of consultations was expected. In addition, this reduction is well documented in most studies evaluating the impact of the COVID-19 pandemic on DCs and could be due to the risk of SARS-CoV-2 transmission associated with outpatient care, and the postponement of elective consultations during the pandemic . Nevertheless, as the pandemic progressed, dermatology consultations increased progressively and consistently. Even though June of 2020 recorded the highest number of cases up to that time, and during July of the same year ICU bed occupation reached the maximum nationwide capacity , such events did not seem to affect the growing trend of dermatology consultations in the following months. Moreover, this consistent increase could have been further fueled by the onset of new dermatological pathologies related to sanitary measures , and the stress caused by the confinement. Frequent handwashing and personal protective equipment usage generated an acute increase in contact dermatitis and other related dermatoses . Furthermore, stress-related skin pathology could have started or worsened given the epidemiological scenario to which the population was exposed at the time. All of this could have impacted this dissociation between the epidemiological events and the increasing trend in consultations as the pandemic progressed. Finally, it is worth highlighting the role that telemedicine could play in this context . Teledermatology in Chile has been exponentially growing over the last decade. In December 2018, a unified teledermatology platform was implemented for the public health system, covering all primary care centers throughout the territory . Chilean studies before the implementation of this platform have shown promising results in waiting times, training of general practitioners, improving resolution, and promoting timely access to care . Thus, it can be implemented as an effective measure to provide dermatologic care to communities with limited access . Along with this, international studies have shown encouraging results in delivering dermatological care , and its implementation in hospitals has been widely recommended due to the limitations of face-to-face care during the pandemic . Therefore, teledermatology plays a central role in supplying the population's demand remotely in the context of the pandemic , and also to optimize access and referral in the following years . Finally, it is worth mentioning that this study has some limitations. First, it does not include diagnoses or the reason for consultation. From this, is not possible to determine whether the patients who consulted were following up previously diagnosed dermatologic pathologies or were first-time consultations. This has important implications in policy-making, as loss of follow-up or delay in diagnosis of potentially severe or malignant dermatologic pathologies could negatively affect the health of the consulting population. In addition, there is no information regarding the severity of the pathology or the resolution of the DC. Furthermore, there are no data regarding the consultations in the private sector at the time. The latter could not be determined, as the private sector is not required to have a centralized database of all the consultations performed by specialty care services. Moreover, this study was based on results obtained during the initial period of the pandemic (2020), when restrictive measures prevailed as a public health strategy to mitigate COVID-19 infections. It does not include the later years (2021-2022) when there was a greater knowledge of the virus and a considerable proportion of the population was vaccinated. In view of this, it would be prudent to propose new studies on the delay in diagnosis and treatment of chronic or malignant dermatologic pathologies secondary to the acute deficit in access to dermatologic care that occurred during 2020, as well as to compare these data with the distribution of DCs during 2021 and 2022 and consider the impacts that it might have. 5. Conclusions The COVID-19 pandemic has had a great impact on the total number of dermatology consultations in the Chilean public health sector, drastically modifying the discrete increasing trend in the total number of consultations observed from 2017 to 2019. Thus, during 2020, these consultations were drastically reduced to more than half, consistent with reports from other studies. Finally, despite the restrictions and difficulties in access during the pandemic months in 2020 in Chile, the consulting population that accessed dermatologic care did so in similar proportions in terms of sex and age range. Author Contributions Conceptualization, D.A.-C., G.A. and A.G.; methodology, D.A.-C., G.A., J.S. and C.T.; software, D.A.-C. and J.S.; validation, G.A., J.S., C.T. and A.G; formal analysis, D.A.-C., G.A. and A.G.; investigation, D.A.-C., G.A., J.S. and C.T.; resources, D.A.-C., G.A. and A.G.; data curation, D.A.-C., G.A. and C.T.; writing--original draft preparation, D.A.-C., G.A. and A.G.; writing--review and editing, J.S. and C.T.; visualization, D.A.-C., G.A. and A.G.; supervision, C.T. and A.G.; project administration, D.A-C. and A.G.; funding acquisition, D.A.-C. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement All data from this study is in the public domain and can be accessed at the DEIS and INE websites of the Chilean government. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Total dermatology consultations in the public health sector from 2017 to 2020. Figure 2 (a) Dermatology consultations in Chile's public sector per 1000 inhabitants in each region, ordered from north to south from left to right. (b) Consultations to dermatology according to the average number of consultations per 1000 inhabitants by geographic zone (North, Center, and South) North zone covers the regions of Arica and Parinacota to Coquimbo. The central zone covers the regions of Valparaiso to Bio Bio. The South zone covers La Araucania to Magallanes and Antarctica regions. Figure 3 Dermatology consultations by age range according to total consultations and (a) according to the number of consultations per 1000 inhabitants in the age range; (b) their respective average trend lines, for the years 2019 and 2020; (c) dermatology consultations by age range with respect to the total consultations for the year; and (d) the ratio between the incidence of dermatology consultations per 1000 inhabitants of a specific age range and the average incidence of dermatology consultations per 1000 inhabitants. Figure 4 Monthly temporal progression of dermatology consultations during 2020 and its correlation with the epidemiological situation of the country at the time . Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Zhou P. Yang X.-L. Wang X.-G. Hu B. Zhang L. Zhang W. Si H.-R. Zhu Y. Li B. Huang C.-L. A pneumonia outbreak associated with a new coronavirus of probable bat origin Nature 2020 579 270 273 10.1038/s41586-020-2012-7 32015507 2. World Health Organization Statement on the Second Meeting of the International Health Regulations (2005) Emergency Committee Regarding the Outbreak of Novel Coronavirus (2019-NCoV) Available online: -emergency-committee-regarding-the-outbreak-of-novel-coronavirus-(2019-ncov) (accessed on 21 August 2021) 3. World Health Organization COVID-19 Weekly Epidemiological Update. Edition 93. Geneva; WHO 25 May 2022 Available online: (accessed on 20 June 2022) 4. Jara A. Undurraga E.A. Gonzalez C. Paredes F. Fontecilla T. Jara G. Pizarro A. Acevedo J. Leo K. Leon F. Effectiveness of an Inactivated SARS-CoV-2 Vaccine in Chile N. Engl. J. Med. 2021 385 875 884 10.1056/NEJMoa2107715 34233097 5. iCOVID Chile Informe #80: Se Deterioran Sustantivamente los Indicadores de la Pandemia Durante Mayo, Aunque Casos Severos se Mantienen Bajos Available online: (accessed on 20 June 2022) 6. World Health Organization COVID-19 Strategic Preparedness and Response Plan: Country Preparedness and Response Status for COVID-19. Geneva: WHO; March 16 2020 Available online: (accessed on 21 August 2021) 7. Grebe G. Velez J.A. Tiutiunnyk A. Aragon-Caqueo D. Fernandez-Salinas J. Navarrete M. Laroze D. Dynamic quarantine: A comparative analysis of the Chilean public health response to COVID-19 Epidemiol. Infect. 2020 148 e270 10.1017/S0950268820002678 33143799 8. Cuadrado C. Monsalves M.J. Gajardo J. Bertoglia M.P. Najera M. Alfaro T. Canals M. Kaufman J.S. Pena S. Impact of Small-Area Lockdowns for the Control of the COVID-19 Pandemic medRxiv 2020 2020.05.05.20092106 Available online: (accessed on 21 August 2021) 9. Ministerio de Salud de la republica de Chile (MINSAL) Ministerio de Salud Confirma Primer caso de Coronavirus en Chile. Santiago de Chile, Marzo 3 del 2020. MINSAL Available online: (accessed on 21 August 2021) 10. Ministerio de Salud de la Republica de Chile (MINSAL) : Subsecretario Alberto Dougnac Anuncio Postergacion de Cirugias Electivas Santiago de Chile, Marzo 16 del 2021 Available online: (accessed on 2 September 2021) 11. Nazzaro G. Marzano A.V. Berti E. What is the role of a dermatologist in the battle against COVID-19? The experience from a hospital on the frontline in Milan Int. J. Dermatol. 2020 59 e238 e239 10.1111/ijd.14926 32378731 12. Giacalone S. Bortoluzzi P. Nazzaro G. Which are the "emergent" dermatologic practices during COVID-19 pandemic? Report from the lockdown in Milan, Italy Int. J. Dermatol. 2020 59 e269 e270 10.1111/ijd.15005 32501536 13. Naderi-Azad S. Manion R. Alhusayen R. Impact of COVID-19 on Access to Care: A National Patient Survey J. Cutan. Med. Surg. 2021 25 453 455 10.1177/12034754211013035 33934612 14. Ministerio de Salud de Chile Subsecretaria de Redes Asistenciales: Division de Gestion y Desarrollo de las Personas y Departamento de Estudios Planificacion y Control de Gestion del Area de Personas. Informe Sobre Brechas de Personal de Salud en los Servicios de Salud y Estado de Situacion de los Recursos Humanos de Salud en Chile. Santiago de Chile: Ministerio de Salud, Abril 2019 Available online: (accessed on 27 August 2021) 15. Mena Vergara L. Carrasco Cancino C. Rojas Pizarro H. Saavedra W. Demografia y distribucion de dermatologos en Chile Piel 2022 37 392 400 10.1016/j.piel.2021.10.001 16. Aragon-Caqueo D. Parrao F. Palacios J. Mora F. Correa J. Guglielmetti A. Consultas a Dermatologia en el sector publico de salud en Chile en 2019 Rev. Chil. Derm. 2020 36 172 177 10.31879/rcderm.v36i4.348 17. Ministerio de Salud Departamento de Estadisticas e Informacion de Salud. Sistema de Reportes REM: Atencion de Especialidades Periodo 2020 Available online: (accessed on 15 December 2021) 18. Instituto Nacional de Estadisticas Sintesis de Resultados CENSO 2017, Junio 2018 Available online: (accessed on 15 December 2021) 19. Ministerio de Salud de Chile COVID-19 en Chile. Pandemia 2020-2022 Hitos de la primera ola|COVID-19 Departamento de Comunicaciones y Relaciones Publicas del Ministerio de Salud Santiago, Chile 2022 96 98 Available online: (accessed on 16 July 2022) 20. Karadag A.S. Aslan Kayiran M. Wollina U. How dermatology has changed in the COVID-19 pandemic Clin. Dermatol. 2021 39 457 460 10.1016/j.clindermatol.2021.01.015 34518004 21. Gisondi P. Piaserico S. Conti A. Naldi L. Dermatologists and SARS-CoV-2: The impact of the pandemic on daily practice J. Eur. Acad. Dermatol. Venereol. 2020 34 1196 1201 10.1111/jdv.16515 32320091 22. Aragon-Caqueo D. Arceu Ojeda M. Aragon-Caqueo G. Zamora Aragon K. Tom Montalva D. Gatica Monsalve J.L. Comparacion del tiempo de espera de atencion dermatologica mediante el uso de teledermatologia y derivacion presencial Piel 2020 35 220 224 10.1016/j.piel.2019.07.001 23. Wang R. Helf C. Tizek L. Neuhauser R. Eyerich K. Zink A. Eberlein B. Biedermann T. Brockow K. Boehner A. The impact and consequences of SARS-CoV-2 pandemic on a single university dermatology outpatient clinic in Germany Int. J. Environ. Res. Public Health 2020 17 6182 10.3390/ijerph17176182 32858870 24. Marin B.G. Aghagoli G. Lavine K. Yang L. Siff E.J. Chiang S.S. Salazar-Mather T.P. Dumenco L. Savaria M.C. Aung S.N. Predictors of COVID-19 severity: A literature review Rev. Med. Virol. 2021 31 1 10 10.1002/rmv.2146 32845042 25. Temiz S.A. Dursun R. Daye M. Ataseven A. Evaluation of dermatology consultations in the era of COVID-19 Dermatol. Ther. 2020 33 e13642 10.1111/dth.13642 32441360 26. Ministerio de Salud de la Republica de Chile (MINSAL) Presidente Declara Estado de Excepcion Constitucional de Catastrofe en Todo el Territorio Nacional. Santiago de Chile, Marzo 18 del 2020 Available online: (accessed on 1 September 2021) 27. Kwatra S.G. Sweren R.J. Grossberg A.L. Dermatology Practices as vectors for COVID-19 transmission: A call for immediate cessation of nonemergent dermatology visits J. Am. Acad. Dermatol. 2020 82 e179 e180 10.1016/j.jaad.2020.03.037 32213307 28. Ministerio de Salud de la Republica de Chile (MINSAL) Departamento de Epidemiologia. Informe Epidemiologico No25 Enfermedad por SARS-CoV-2 (COVID-19). Santiago 2020 Available online: (accessed on 4 September 2022) 29. Goh C.F. Ming L.C. Wong L.C. Dermatologic reactions to disinfectant use during the COVID-19 pandemic Clin. Dermatol. 2021 39 314 322 10.1016/j.clindermatol.2020.09.005 34272029 30. Shanshal M. Ahmed H.S. Asfoor H. Salih R.I. Ali S.A. Aldabouni Y.K. Impact of COVID-19 on medical practice: A nationwide survey of dermatologists and health care providers in Iraq Clin. Dermatol. 2021 39 500 509 10.1016/j.clindermatol.2020.11.010 34518011 31. Gronbeck C. Grant-Kels J.M. Lu J. Feng H. Increased Utilization of Teledermatology Among Medicare Part B Beneficiaries During the COVID-19 Pandemic Clin. Dermatol. 2022 40 760 763 10.1016/j.clindermatol.2022.09.004 36075538 32. Subsecretaria de Redes Asistenciales Programa Nacional de Telesalud 2018 Available online: (accessed on 4 September 2021) 33. Fuenzalida Cruz H. Jimeno Ortega I. Toso Diaz de la Vega S. Sepulveda Munoz A. Loubies Munoz R. Teledermatologia: Impacto de una herramienta de gestion informatica para zonas remotas de Chile Piel 2017 32 257 262 10.1016/j.piel.2016.11.016 34. Gatica J.L. Bertolo S. Morales E. Espinoza M. Contreras C. Teledermatologia en Chile, un aporte a la atencion primaria de salud Piel 2015 30 148 154 10.1016/j.piel.2014.07.007 35. Aragon-Caqueo D. Arceu M. Aragon-Caqueo G. Zamora K. Tom D. Gatica J.L. Teledermatologia en Chile: Experiencia de su implementacion temprana Piel 2022 37 1 6 10.1016/j.piel.2020.12.009 36. Lopez-Liria R. Lopez-Villegas A. Valverde-Martinez M.A. Perez-Heredia M. Vega-Ramirez F.A. Peiro S. Leal-Costa C. Comparative Analysis of Quality of Life of Patients with Dermatological Problems: Teledermatology Versus Face-to-Face Dermatology Healthcare 2022 10 2172 10.3390/healthcare10112172 36360514 37. Trondsen M.V. Managing Everyday Life: A Qualitative Study of Patients' Experiences of a Web-Based Ulcer Record for Home-Based Treatment Healthcare 2014 2 492 504 10.3390/healthcare2040492 27429289 38. Rismiller K. Cartron A.M. Trinidad J.C.L. Inpatient Teledermatology During the COVID-19 Pandemic J. Dermatolog. Treat. 2020 31 441 443 10.1080/09546634.2020.1762843 32364809 39. Elsner P. Teledermatology in the times of COVID-19--A systematic review J. Dtsch. Dermatol. Ges. 2020 18 841 845 10.1111/ddg.14180 40. Gatica J.L. Aedo Inostroza G. Letelier M. Sabando V. Loubies R. Fuenzalida H. Aragon-Caqueo D. Martinez G. Teledermatology in Chile: Implementation of a Single National System J. Clin. Exp. Dermatol. Res. 2022 13 605 10.35248/2155-9554.22.13.605 |
PMC10000558 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050723 cells-12-00723 Communication Blocking Store-Operated Ca2+ Entry to Protect HL-1 Cardiomyocytes from Epirubicin-Induced Cardiotoxicity Liu Xian Conceptualization Methodology Software Validation Formal analysis Data curation Writing - original draft 12 Chang Yan Methodology Software Validation Data curation 23 Choi Sangyong Methodology Validation Formal analysis Investigation Resources 2 Cai Chuanxi Conceptualization Validation Investigation Resources Writing - review & editing 4 Zhang Xiaoli Methodology Software Formal analysis Resources Data curation 5* Pan Zui Conceptualization Methodology Writing - original draft Writing - review & editing Visualization Supervision Project administration Funding acquisition 123* Nagareddy Prabhakara Academic Editor El-Refaey Mona M. Academic Editor 1 Department of Kinesiology, College of Nursing and Health Innovation, The University of Texas at Arlington, Arlington, TX 76010, USA 2 Department of Graduate Nursing, College of Nursing and Health Innovation, The University of Texas at Arlington, Arlington, TX 76010, USA 3 Bone and Muscle Research Center, College of Nursing and Health Innovation, The University of Texas at Arlington, Arlington, TX 76010, USA 4 Department of Surgery, Division of Surgical Sciences, School of Medicine, University of Virginia, Charlottesville, VA 22903, USA 5 Department of Biomedical Informatics, The Ohio State University, Columbus, OH 43210, USA * Correspondence: [email protected] (X.Z.); [email protected] (Z.P.) 24 2 2023 3 2023 12 5 72301 12 2022 17 2 2023 21 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Epirubicin (EPI) is one of the most widely used anthracycline chemotherapy drugs, yet its cardiotoxicity severely limits its clinical application. Altered intracellular Ca2+ homeostasis has been shown to contribute to EPI-induced cell death and hypertrophy in the heart. While store-operated Ca2+ entry (SOCE) has recently been linked with cardiac hypertrophy and heart failure, its role in EPI-induced cardiotoxicity remains unknown. Using a publicly available RNA-seq dataset of human iPSC-derived cardiomyocytes, gene analysis showed that cells treated with 2 mM EPI for 48 h had significantly reduced expression of SOCE machinery genes, e.g., Orai1, Orai3, TRPC3, TRPC4, Stim1, and Stim2. Using HL-1, a cardiomyocyte cell line derived from adult mouse atria, and Fura-2, a ratiometric Ca2+ fluorescent dye, this study confirmed that SOCE was indeed significantly reduced in HL-1 cells treated with EPI for 6 h or longer. However, HL-1 cells presented increased SOCE as well as increased reactive oxygen species (ROS) production at 30 min after EPI treatment. EPI-induced apoptosis was evidenced by disruption of F-actin and increased cleavage of caspase-3 protein. The HL-1 cells that survived to 24 h after EPI treatment demonstrated enlarged cell sizes, up-regulated expression of brain natriuretic peptide (a hypertrophy marker), and increased NFAT4 nuclear translocation. Treatment by BTP2, a known SOCE blocker, decreased the initial EPI-enhanced SOCE, rescued HL-1 cells from EPI-induced apoptosis, and reduced NFAT4 nuclear translocation and hypertrophy. This study suggests that EPI may affect SOCE in two phases: the initial enhancement phase and the following cell compensatory reduction phase. Administration of a SOCE blocker at the initial enhancement phase may protect cardiomyocytes from EPI-induced toxicity and hypertrophy. anthracycline chemotherapy store-operated Ca2+ entry (SOCE) apoptosis cardiac hypertrophy NFAT4 reactive oxygen species (ROS) National Institutes of HealthOD025230 the UTA College of Nursing and Health InnovationThis research was partially supported by the National Institutes of Health (NIH) Grant S10 OD025230 (to Z.P.), and by the Pilot Research Grant from the UTA College of Nursing and Health Innovation (to X.L.). pmc1. Introduction Anthracyclines listed in the 22nd (the latest) version of the World Health Organization (WHO) model list of essential medicines are among the most efficacious and widely used chemotherapy drugs since the late 1960s . Epirubicin (EPI) belongs to the anthracycline family; it is often used together with new generation targeted drugs and play a major role in the modern era of cancer treatment. EPI kills cancer cells likely via multiple mechanisms, including DNA adduct formation, reactive oxygen species (ROS) production, and lipid peroxidation. While EPI makes a great contribution to the improvement of treatment outcomes, dose-limiting cardiotoxicity hinders its clinical application and often leads to requirements for regimen modification or even discontinuation . The anthracycline-induced cardiotoxicity can be manifested either acutely during the treatment period or chronically, from several weeks to even years after treatment has stopped . The associated cardiac dysfunction has a broad range of symptoms including cardiac hypertrophy, cardiomyopathy, and ultimately congestive heart failure . Cardiac hypertrophy is the enlargement of the heart, which can be divided into two categories: physiological and pathological, both of which develop as an adaptive response to cardiac stress, but their underlying molecular mechanisms, cardiac phenotype and prognosis are distinctly different. For example, Ca2+ signaling-related genes are only changed in pathological hypertrophy but not in physiological hypertrophy . Studies have revealed that intracellular Ca2+ regulates the calcineurin-NFAT signaling pathway and thus initiating hypertrophy-related gene transcription . An increase in intracellular Ca2+ leads to the activation of the phosphatase activity of calcineurin, the dephosphorylation of NFAT family members, and their translocation to the nucleus to initiate gene transcription . Store-operated Ca2+ entry (SOCE) is a ubiquitous Ca2+ entry pathway activated in response to the depletion of sarcoplasmic or endoplasmic reticulum (SR/ER) Ca2+ stores. Although SOCE has been well-studied in non-excitable cells and skeletal muscles, the understanding of its important role in cardiomyocytes is emerging . SOCE machinery components, including stromal interaction molecule 1 (Stim1) as an ER Ca2+ sensor and Orais and transient receptor potential channels (TRPCs) as plasma membrane Ca2+ channels, have been shown to be essential for heart development and to regulate heart remodeling after stress . Accumulating evidence also shows enhanced SOCE in cardiac hypertrophy and heart failure . While dysregulated Ca2+ signaling has been reported to contribute to EPI-induced cardiotoxicity , whether SOCE plays a role in this process and in the consequent cardiac remodeling remains unknown. Thus, the objective of the present study is to determine the specific role of SOCE in EPI-induced cell apoptosis and hypertrophy in cardiomyocytes. 2. Materials and Methods 2.1. Chemicals and Reagents Claycomb cell culture medium was purchased from Sigma-Aldrich. FBS (fetal bovine serum), PBS (phosphate-buffered saline), HBSS (Hanks' balanced salt solution), and penicillin/streptomycin antibiotic were purchased from Invitrogen/Thermofisher Scientific Pittsburgh, PA, USA. Other reagents used include BTP2 and ML204 (Millipore Sigma, St. Louis, MO, USA), EPI (Alfa Aesar, Haverhill, MA, USA), thapsigargin (TG, Adipogen, San Diego, CA, USA), fura-2 AM (Biotium 50033, Fremont, CA, USA), DAPI (Invitrogen D357, Carlsbad, CA, USA), and phalloidin (Enzo BML-T111, New York, NY, USA). 2.2. Cell Culture HL-1 cardiomyocytes were maintained in Claycomb medium supplemented with 10% FBS, 100 U/mL penicillin, 100 ug/mL streptomycin, 0.1 mM norepinephrine, and 2 mM L-glutamine . HL-1 cells were cultured at 37 degC in a humidified 5% CO2 incubator. 2.3. Measurement of Intracellular Ca2+ Concentration Intracellular Ca2+ concentrations in the HL-1 cell line was measured following previously published procedures . In brief, the intracellular Ca2+ was measured using a fluorescence microscope with a SuperFluo 40x objective (N.A. 1.3) connected to a dual-wavelength spectrofluorometer (Horiba Photon Technology International, Piscataway, NJ, USA). The excitation wavelengths were set at 350 nm and 385 nm and the emission wavelength was set at 510 nm. Cells were loaded with 5 mM fura-2 acetoxymethyl ester (Biotium, Fremont, CA, USA) for 30 min at 37 degC in the dark. Cellular endoplasmic reticulum (ER) Ca2+ stores were depleted by 10 mM TG in 0.5 mM EGTA dissolved in balanced salt solution (140 mM NaCl, 2.8 mM KCl, 2 mM MgCl2, 10 mM HEPES, pH 7.2). SOCE was observed upon the rapid exchange of extracellular solution to bath saline containing 2 mM CaCl2 at indicated time. The intracellular Ca2+ elevation was presented as DF350 nm/F385 nm. 2.4. Cytotoxicity Assay HL-1 cells were seeded at 1.5 x 105 cells per well in a 29 mm glass-bottom dish. The cells were treated with vehicle, 20 mM BTP2, 1 mM EPI, or 20 mM BTP2 plus 1 mM EPI for 5 h. Then, the culture medium was removed, and cells were fixed with 4% paraformaldehyde for 10 min at room temperature. The paraformaldehyde was removed and then cells were immersed in 0.1% Triton X-100 in PBS for 10 min, washed with PBS twice, followed by incubation with PBS containing 6.6 mM phalloidin (Enzo, New York, NY, USA) for 15 min. The cells were washed with PBS three times, and then counter staining with PBS containing 1 mg/mL DAPI (1:500) for 5 min at room temperature in the dark. The cells were then washed with PBS twice and immersed in ProlongTM Gold antifade reagent (Life Technologies Corporate, Eugene, OR, USA). The fluorescence signals were observed using a DMi8 inverted microscope (Leica, Wetzlar, Germany) with a 40x objective (NA 1.3). The excitation/emission wavelengths set for DAPI and phalloidin were 405/430 nm and 547/572 nm, respectively. The imaging was performed at room temperature. 2.5. Western Blotting Analysis HL-1 cardiomyocytes were lysed in modified RIPA buffer (150 mM NaCl, 50 mM Tris-Cl, 1 mM EGTA, 1% Triton X-100, 0.1% SDS, and 1% sodium deoxycholate, pH 8.0) containing protease inhibitors cocktail (Sigma-Aldrich, US) as previously described . Protein concentration was quantified using a BCA kit (ThermoFisher, Pittsburgh, PA, USA). Equal amounts of proteins were loaded onto SDS polyacrylamide gels, and the separated proteins were transferred to PVDF membranes (Bio-Rad, Hercules, CA, USA). The blot was incubated with 5% non-fat dry milk blocking buffer (Bio-Rad, Hercules, CA, USA) for 1 h at room temperature and probed with specific primary antibodies in blocking buffer at 4 degC overnight. The primary antibodies used in this study included anti-caspase-3 (1:1000, catalog #9662, Cell Signaling Technology, Massachusetts, MA, USA) and anti-GAPDH (1:1000, GeneTex, Irvine, CA, USA). The next day, the blots were washed with PBST three times followed by incubation with secondary antibodies including the appropriate horse radish peroxidase (HRP)-conjugated goat anti-rabbit IgG (1:5000, Cell Signaling Technology, Massachusetts, USA) and anti-mouse IgG (1:5000, Cell Signaling Technology, USA). Signals were detected using the ECL detection method on a ChemiDoc instrument. 2.6. Cell Size Measurement HL-1 cells seeded at 1 x 106 cells per well in a 6-well plate were treated with vehicle or 20 mM BTP2, 1 mM EPI, or 20 mM BTP2 plus 1 mM EPI for 5 h followed by switching to normal culture media for 24 h. The cells were then observed and phase contrast imaging was conducted using a DMi8 inverted microscope (Leica, Wetzlar, Germany). The cell surface area was quantified using ImageJ and Graphpad 6 software. 2.7. Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR) Total RNAs were extracted from HL-1 cells using Illustra RNAspin MiniRNA Isolation Kit and the quality and concentration of RNA were evaluated by photometrical measurement of 260/280 nm. The primers were obtained from Sigma Aldrich. Four hundred nanograms of total RNA was applied for reverse transcription using the qScript microRNA Synthesis Kit (QuantaBio, Beverly, MA, USA) following the manufacturer's protocol. cDNA was diluted 1:5 in DNase-, RNase-, and protease-free water and 2 mL template was used for PCR. The primer pairs for BNP and GAPDH were used. The sequences for the BNP primers are forward (5'-3') GCCAGTCTCCAGAGCAATTC and reverse (5'-3') TCTTTTGTGAGGCCTTGGTC. The sequences for the GAPDH primers are forward (5'-3') AGGTCGGTGTGAACGGATTTG and reverse (5'-3') TGTAGACCATGTAGTTGAGGTCA. For qRT-PCR, QuantaBio PerfecTa SYBR Green FastMix ROX was used according to the manufacturer's procedure. The signals generated by integration of SYBR Green into the amplified DNA were detected in a real-time machine (StepOne Plus Real-Time PCR System, ThermoFisher Scientific, USA). Data were expressed as 2-DDCT relative to GAPDH gene expression. 2.8. Immunofluorescence Staining Cells were seeded into 29 mm glass-bottom dishes. The cells were fixed with 4% paraformaldehyde for 10 min at room temperature. The paraformaldehyde was then removed and the cells were immersed in PBS containing 0.1% Triton X-100 for 10 min. After washing with PBS three times, the cells were blocked in PBS containing 0.1% Triton X-100 supplemented with 10% horse serum for 30 min at room temperature. Then, the cells were incubated with rabbit anti-NFAT4 primary antibody (1:100, ProteinTech, Rosemont, IL, USA) in blocking solution at 4 degC overnight. The next day, the cells were taken out and washed with PBS three times, then incubated with Alexa Fluor 488-labelled secondary antibody (1:500, Abcam, Cambridge, MA, USA) at room temperature in the dark for 1 h to visualize the expression and localization of NFAT4. The cells were counter-stained with PBS containing 1 mg/mL DAPI (1:500) for 5 min at room temperature in the dark and then immersed in ProlongTM Gold antifade reagent (Life Technologies Corporate, Eugene, OR, USA). Images were taken using a Nikon A1R HD25 LSM confocal microscope with a 40x oil immersion objective (NA 1.3) using GFP and DAPI filters (Ex: 488/405; Em: 509/430 nm). 2.9. RNA-Seq Data Analysis The RNA-Seq dataset GSE217421 was used . Different human induced pluripotent stem cell (iPSC)-derived cardiomyocyte cell lines were treated with 2 mM EPI or DMSP for 48 h, followed by bulk RNA-seq analysis. Differentially expressed gene were identified between control treated cell lines. A total 17 EPI samples and 56 control samples covering five different cell types were used with each cell type having a different number of replicates as shown in Table S2. Table S3 shows all the control cell lines and replicate numbers. Two-way ANOVA of the effects of treatment (EPI vs. Con) and cell lines (five cell lines) was used for analysis. 2.10. Statistical Analysis Data were analyzed using Graphpad Prism 6 software (Boston, MA, USA) unless indicated otherwise. The results were presented as mean +- standard deviation (SD) or as otherwise indicated. Comparisons between two groups were analyzed using a Student's t-test. Comparisons among more than two groups were analyzed using one-way analysis of variance (ANOVA) followed by Bonferroni post hoc analysis. A p value of <0.05 was considered statistically significant in all experiments except the RNA-seq data analysis. 3. Results 3.1. SOCE Machinery Genes Were Downregulated by EPI Treatment in Human iPSC-Derived Cardiomyocytes RNA-seq data analysis of human iPSC-derived cardiomyocytes showed that cells treated with 2 mM EPI for 48 h had significantly reduced expression of SOCE machinery genes, i.e., Orai1, Orai3, TRPC3, TRPC4, Stim1, and Stim2, and increased expression of TRPC2 . The expression of Orai2, TRPC1, TRPC5, and TRPC6 were similar between the EPI and control groups. To confirm the changes in SOCE in live cells, HL-1, a cardiomyocyte cell line derived from adult mouse atria was used for its easy culture and well-characterized cardiomyocyte properties. After being treated with 1 mM EPI or vehicle control (0.1% DMSO) for 6 h, HL-1 cells were loaded with 5 mM fluorescent Ca2+ indicator fura-2 AM at 37 degC in the dark for 30 min. The ER Ca2+ stores were depleted by 10 mM TG in BSS containing 0.5 mM EGTA. When re-introducing BSS containing 2 mM CaCl2, the intracellular Ca2+ level (presented as F350 nm/F385 nm) was monitored using live cell imaging and the SOCE was calculated as the difference (DF350/F385) between the peak and baseline before the addition of 2 mM Ca2+. Compared to vehicle control (black curve), SOCE was significantly reduced in the EPI-treated HL-1 cells (red curve) . 3.2. Acute Treatment of EPI Increased SOCE in HL-1 Cardiomyocytes In addition to transcriptional regulation, EPI can increase ROS production and lipid peroxidation. Oxidative stress has been shown to promote STIM1 oligomerization and alter channel activity . We next examined whether acute treatment of EPI and its resulting oxidative stress can affect SOCE in HL-1 cells. Administration of BTP2, a SOCE inhibitor, significantly decreased DF350/F385 (0.142 +- 0.064) compared with that of the vehicle-treated control cells (0.188 +- 0.058, n = 35; ** p = 0.0051). This data confirmed the presence of BTP2-sensitive SOCE in HL-1 cardiomyocytes . Contrary to prolonged treatment, acute treatment of EPI for only 30 min resulted in significantly enhanced SOCE in HL-1 cells (0.254 +- 0.069, n = 37) compared to those treated with the vehicle control (0.188 +- 0.058, n = 37; **** p < 0.0001). Addition of BTP2 could significantly decrease SOCE (0.045 +- 0.027, n = 36) in EPI-treated HL-1 cells compared to those treated with EPI alone (0.254 +- 0.069, n = 36; **** p < 0.0001), indicating that pharmacologically inhibiting SOCE with BTP2 can reduce the EPI-enhanced SOCE in HL-1 cells. Furthermore, ML204, a relative specific TRPC4 inhibitor could significantly reduce SOCE in HL-1 cells as well . 3.3. BTP2 Diminished EPI-Induced ROS Production in HL-1 Cardiomyocytes The reciprocal regulation between mitochondria and intracellular Ca2+ suggests that SOCE may regulate mitochondrial ROS production. Thus, ROS were measured by using DHE dye in HL-1 cells treated with EPI . In the HL-1 cells treated with 1 mM EPI for 30 min, ROS levels were significantly increased compared to that in vehicle control cells. Interestingly, BTP2 was able to significantly inhibit ROS production in HL-1 cells even in the presence of EPI . 3.4. BTP2 Inhibited EPI-Induced Apoptosis in HL-1 Cardiomyocytes Disruption of F-actin is a hallmark for apoptosis . We next examined the expression of F-actin in HL-1 cells using phalloidin staining. Reduced expression of F-actin was observed in cells treated with 1 mM EPI for 5 h compared to that of vehicle-treated control cells , suggesting that EPI induced apoptosis in HL-1 cardiomyocytes. When co-treated with 20 mM BTP2, the EPI-induced degradation of F-actin was partially rescued , indicating that BTP2 inhibited EPI-induced F-actin disruption. Anthracyclines have been shown to induce apoptosis in HL-1 cardiomyocytes through caspase-3 . We then examined the levels of cleaved caspase-3 in HL-1 cells. EPI induced abundant amounts of cleaved caspase-3, evidenced by the Western blot analysis . The EPI-increased level of cleaved caspase-3 was significantly diminished by co-treatment with 20 mM BTP2. Consistent with the data from F-actin degradation, the cleaved caspase-3 analysis again indicated that EPI induced apoptosis in HL-1 cardiomyocytes, which could be alleviated by BTP2. 3.5. BTP2 Inhibited EPI-Induced Hypertrophy in HL-1 Cardiomyocytes EPI-induced cardiac remodeling includes hypertrophy. SOCE plays a major role in pathophysiological hypertrophy. We thus examined whether BTP2 can inhibit EPI-induced hypertrophy in HL-1 cardiomyocytes. HL-1 cells were treated with vehicle control, 1 mM EPI, or co-treated with 1 mM EPI and 20 mM BTP2 for 5 h, followed by drug withdrawal and then growth in normal culture medium for 24 h. Phase contrast images were then taken of these cells and the surface area of the HL-1 cardiomyocytes was measured and quantified. EPI treatment increased the size of cardiomyocytes to almost twice that of vehicle-treated control cardiomyocytes . In the BTP2 and EPI co-treatment group, the size of HL-1 cells was significantly reduced compared to that in the EPI group. The expression of brain natriuretic peptide (BNP), a specific marker of cardiac hypertrophy , was also examined. As shown in Figure 5C, the mRNA level of BNP was significantly increased upon the treatment with 1 mM EPI (4.861 +- 0.697, n = 9) compared to that of vehicle-treated cells (control, 1.010 +- 0.155, n = 9). Consistent with the cell size analysis, BTP2 could significantly alleviate EPI-induced BNP expression (3.054 +- 0.260) in HL-1 cells. These data indicate that blocking SOCE by BTP2 can reduce EPI-induced hypertrophy in HL-1 cardiomyocytes. 3.6. BTP2 Inhibited EPI-Induced NFAT4 Nuclear Translocation in HL-1 Cardiomyocytes Nuclear factor of activated T cells (NFAT) was reported to be a critical nuclear transcriptional factor regulating cardiac hypertrophy . We lastly examined whether SOCE contributes to EPI-induced hypertrophy through the NFAT pathway in HL-1 cells. Since NFAT4 is the most abundant one out of the five subtypes of NFAT expressed in cardiomyocytes , we focused on NFAT4 in this study. After being treated with vehicle, 1 mM EPI, 20 mM BTP2, or 1 mM EPI combined with 20 mM BTP2 for 5 h, HL-1 cells were cultured in growth media for another 24 h until fixation and immunostaining with anti-NFAT4 antibody. HL-1 cells treated with 10 mM ionomycin for 15 min were used as a positive control for NFAT4 immunostaining since ionomycin is a strong activator for NFAT signaling . The nuclear translocation of NFAT4 was examined by confocal microscopy imaging. As shown in Figure 6, EPI treatment induced NFAT4 nuclear translocation (indicated by the white arrows), whereas co-treatment with BTP2 showed minimal NFAT4 nuclear translocation. This data suggested that the EPI-induced nuclear translocation of NFAT4 was inhibited by BTP2 in HL-1 cells. 4. Discussion EPI is a widely used anthracycline chemotherapy drug, but it also causes cardiotoxicity and results in heart remodeling and even failure. This study confirmed that EPI can induce ROS production, cell apoptosis, and hypertrophy in cardiomyocytes. Furthermore, this study showed that acute treatment of EPI can increase SOCE in HL-1 cells and blocking SOCE by BTP-2 not only reduced EPI-enhanced SOCE , but also alleviated EPI-induced apoptosis and hypertrophy . Although SOCE has been associated with hypertrophy in cardiomyocytes and heart failure, this study provides the first evidence, to our knowledge, that SOCE plays a key role in EPI-induced cardiotoxicity and hypertrophy. More importantly, this study may shed light on developing an approach to alleviate EPI-induced cardiotoxicity by targeting SOCE in the initial phase of EPI treatment . It is well-known that SOCE has a complex nature and co-exists with other Ca2+ influx mechanisms, such as receptor-operated Ca2+ entry (ROCE). SOCE machinery may contain several molecules as channel complexes at the plasma membrane interacting with STIMs at the SR/ER. Previous reports showed that Orai1 is expressed in HL-1 cells and knockdown of Orai1 could abolish SOCE in HL-1 cells . In addition, TRPC1, 3/6, and 4 may also form SOCE channel complexes in hypertrophic cardiomyocytes and STIM1 can bind and regulate TRPC1, TRPC4, and TRPC5 . The current study showed evidence for a bona fide, BTP2-sensitive SOCE in HL-1 cells. Since BTP2 can block both Orai and TRPC channels , our current data cannot exactly pinpoint whether Orais or TRPCs mediate SOCE in hypertrophic cardiomyocytes. RNA-seq analysis showed that treatment of EPI significantly reduced Orai1, Orai3, TRPC3, and TRPC4 expression in human iPSC-derived cardiomyocytes, which is consistent with reduced SOCE . Interestingly, ML204, a relatively selective blocker of TRPC4 could significantly reduce SOCE in HL-1 cells . These data suggest that these Orais and TRPCs may contribute to SOCE in cardiomyocytes. Future investigation is required to dissect the exact components in the SOCE channel complex in cardiomyocytes, which contribute to EPI-induced cardiotoxicity. After cardiomyocytes survived the cardiotoxicity after EPI treatment, they may undergo cell remodeling which leads to hypertrophy, cardiac remodeling, and eventual heart failure. SOCE plays a major role in the pathogenesis of heart hypertrophy. Numerous studies suggest that pathological stimuli activate SOCE and further trigger the NFAT signaling cascade, which is critical for the regulation of growth gene expression and promotion of cardiomyocyte hypertrophy. Suppression of SOCE machinery genes, such as STIM1 and Orai1, attenuates the hypertrophic responses to pressure overload or agonists . Our current findings are in line with these previous reports, indicating that EPI-induced cardiomyocyte hypertrophy could also be inhibited by SOCE blocker. Interestingly, RNA-seq data analysis of human iPSC-derived cardiomyocytes showed that cells treated with 2 mM EPI for 48 h had significantly reduced expression of SOCE machinery genes, e.g., Orai1, Orai3, TRPC3, TRPC4, Stim1, and Stim2 . Intracellular Ca2+ measurement in live HL-1 cells confirmed that SOCE was indeed reduced in cardiomyocytes treated with EPI for 6 h or longer. The apparent discrepancy suggests that EPI may affect SOCE in two phases: the initial enhancement phase followed by cell compensatory reduction phase. The initial enhancement phase is likely due to immediately increased ROS production and lipid peroxidation right after administration of EPI. The rapid generation of ROS has been best studied in myocardial ischemia-reperfusion models . In addition to the regulatory roles of ROS in many cellular events , oxidative stress is also able to promote STIM1 oligomerization, deplete ER Ca2+, and active SOCE . Since there is a reciprocal regulation between mitochondria and SOCE, EPI-triggered initial mitochondrial ROS production could be further amplified by enhanced SOCE, which is supported by the evidence that blocking SOCE by BTP2 attenuated EPI-triggered ROS production . Additionally, ROS is also known to directly activate TRPCs channels . During the initial enhancement phase, EPI triggers the apoptotic pathway in cardiomyocytes. The surviving cardiomyocytes from the initial phase may develop compensatory mechanisms at the transcription level. This may explain why prolonged treatment of EPI (at 48 h) resulted in a reduction in the expression of SOCE machinery genes. Chemotherapeutic agents (anthracycline therapy in particular) have been reported to damage the F-actin of cells. In cardiac H9c2 cells, doxorubicin reduces number of F-actin filaments, especially at higher concentrations . The reorganization of F-actin filaments and characteristic features of apoptosis have also been reported in Chinese hamster ovary cells, pancreatic b cells, breast cancer cells, and other cells upon doxorubicin treatment . Others and our previous studies suggest that SOCE is an effective chemotherapy drug target . The findings of the present study have shown that SOCE contributes to EPI-induced cardiotoxicity, indicating that SOCE blockers may be able to protect cardiomyocytes from the side effects of anthracycline chemotherapy drugs. Together, the results suggest that SOCE blockers may be dual-function drugs for both chemotherapy and cardio-protection. Acknowledgments We thank Frank Yi for his help in editing the manuscript. Supplementary Materials The following supporting information can be downloaded at: Table S1. Comparison of genes of interest between EPI-treated and control cells; Table S2. Table of trt by cell-final used; Table S3. Table of state by sell (all EPI and control samples); Figure S1. ML204 inhibited SOCE in HL-1 cells; Figure S2. EPI failed to increase SOCE in HL-1 cells treated with ROS scavenger; Figure S3. TG-sensitive SR/ER Ca2+ storage analysis in HL-1 cells; Figure S4. Orai1 expression was not affected by BTP2 treatment in HL-1 cells. Click here for additional data file. Author Contributions Conceptualization, Z.P., X.Z. and X.L.; methodology, X.L. and X.Z.; software, X.L. and X.Z.; validation, X.L.; formal analysis, X.L. and Y.C.; investigation, S.C., X.L. and Y.C; data curation, X.L. and X.Z; writing--original draft preparation, X.L.; writing--review and editing, X.L., Z.P., X.Z. and C.C.; visualization, X.L. and Z.P.; supervision, Z.P.; project administration, Z.P.; funding acquisition, Z.P. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Publicly available datasets were analyzed in this study. This data can be found here: GSE217421 accessed on 30 November 2022.) Conflicts of Interest The authors declare no conflict of interest. Figure 1 SOCE machinery genes were downregulated by EPI in cardiomyocytes. (A) RNA-seq data analysis of human iPSC-derived cardiomyocytes treated with 2 mM EPI for 48 h. SOCE machinery genes include Orais, TRPCs, and Stims. (B) Representative traces of intracellular Ca2+ in HL-1 cells treated with vehicle control or 1 mM EPI for 6 h. (C) Statistics of changes in intracellular Ca2+ in HL-1 cells without (0.28 +- 0.085) or with the treatment of EPI (0.02897 +- 0.021). Mean +- SD, n = 10, ****: p < 0.0001 (based on t-test). Figure 2 Acute treatment of EPI increased SOCE in HL-1 cardiomyocytes. Ratio of fluorescence of fura-2 AM at two excitation wavelengths (F350 nm/F385 nm) was used to monitor changes in intracellular Ca2+ concentration. (A-D) Representative traces of intracellular Ca2+ in HL-1 cells treated with vehicle control, 20 mM BTP2, 1 mM EPI, or 20 mM BTP2 plus 1 mM EPI for 30 min. (E) Statistical analysis of SOCE in HL-1 cells. Mean +- SD, n = 35. **: p = 0.0056, ****: p < 0.0001 (based on one-way ANOVA and Bonferroni post hoc analysis). Figure 3 EPI induced ROS production in HL-1 cardiomyocytes. ROS was measured in HL-1 cells by using DHE dye in control group with only resting cells; 1 mM EPI treatment for 30 min (EPI); 20 mM BTP2 (BTP2) treatment; 1 mM EPI together with 20 mM BTP2 treatment for 30 min (EPI + BTP-2). Quantitative fluorescent intensity (a.u.) from each independent well was showed. n = 6, Mean +- SD. ***: p < 0.001; ****: p < 0.0001. (Based on One Way ANOVA and Bonferroni post hoc analysis). Figure 4 BTP2 inhibited EPI-induced apoptosis in HL-1 cardiomyocytes. (A) Cell nuclei were stained with DAPI (blue) and F-actin was stained with phalloidin (red). F-actin degradation was observed with 1 mM EPI treatment for 5 h. With 20 mM BTP2 co-treatment, the EPI-induced F-actin disruption was partially rescued. (B) The F-actin fluorescence intensity was quantified. ***: p < 0.001; ****: p < 0.0001 (C) Western blotting analysis of cleaved caspase 3 expression in HL-1 cardiomyocytes. Cells were treated by vehicle (control), 1 mM EPI, 20 mM BTP2, or 1 mM EPI plus 20 mM BTP2 for 5 h followed by normal culture conditions for 24 h. GAPDH was used as loading control. (D) The quantification of cleaved caspase 3 normalized to the expression of GAPDH. Three independent biological replicates were carried out and used for the quantification. ***: EPI vs. Control, p = 0.0009; *: EPI + BTP2 vs. EPI, p = 0.0218 (based on one-way ANOVA and Bonferroni post hoc analysis). Figure 5 BTP2 inhibited EPI-induced hypertrophy in HL-1 cardiomyocytes. HL-1 cells treated with 1 mM EPI for 5 h followed by fresh media culture for 24 h. (A) Phase contrast images of HL-1 cells treated with 1 mM EPI (EPI), 1 mM EPI together with 20 mM BTP2 (EPI + BTP-2), or vehicle-treated control (control). Scale bar, 50 mm. (B) Quantification of the cell surface area of HL-1 cardiomyocytes. Mean +- SEM, n >= 581 per group. ****: p < 0.0001 (based on one-way ANOVA and Bonferroni post hoc analysis). (C) BTP2 reduced hypertrophy marker BNP transcript in HL-1 cardiomyocytes. Quantitative reverse transcription PCR expression levels of BNP were normalized to GAPDH and plotted relative to the level in the vehicle-treated control cells. n = 9, triplicates from three independent experiment. Mean +- SD. ****: p < 0.0001 (based on one-way ANOVA and Bonferroni post hoc analysis). Figure 6 BTP2 inhibited the EPI-induced nuclear translocation of NFAT4 in HL-1 cardiomyocytes. HL-1 cells were treated with vehicle, 1 mM EPI, 20 mM BTP2, or 1 mM EPI combined with 20 mM BTP2 for 5 h, followed by drug withdrawal and culture for 24 h and stained for NFAT4. Cells treated with 10 mM ionomycin for 15 min were used as positive control. (A) Representative confocal images of NFAT. Scale bar, 10 mm. white arrows indicate the co-localization of NFAT4 and Nuclei (B) Statistical analysis of the percentage of cells with NFAT nuclear translocation. Each dot is an averaged datapoint from a single confocal image. Each image represents one biological replicate. Mean +- SD. **: p < 0.01 (based on one-way ANOVA and Bonferroni post hoc analysis). Figure 7 Working model. Acute treatment of EPI enhanced SOCE, triggering NFAT4-mediated hypertrophy and caspase3-mediated apoptosis in cardiomyocytes. Both the EPI-induced apoptosis and hypertrophy can be inhibited by blocking SOCE. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. World Health Organization WHO Model List of Essential Medicines--22nd List 2021 Available online: (accessed on 30 September 2021) 2. Zamorano J.L. Lancellotti P. Rodriguez Munoz D. Aboyans V. Asteggiano R. Galderisi M. Habib G. Lenihan D.J. Lip G.Y.H. Lyon A.R. 2016 ESC Position Paper on cancer treatments and cardiovascular toxicity developed under the auspices of the ESC Committee for Practice Guidelines: The Task Force for cancer treatments and cardiovascular toxicity of the European Society of Cardiology (ESC) Eur. Heart J. 2016 37 2768 2801 10.1093/eurheartj/ehw211 27567406 3. Mitry M.A. Edwards J.G. Doxorubicin induced heart failure: Phenotype and molecular mechanisms Int. J. Cardiol. Heart Vasc. 2016 10 17 24 10.1016/j.ijcha.2015.11.004 27213178 4. Cardinale D. Colombo A. Lamantia G. Colombo N. Civelli M. De Giacomi G. Rubino M. Veglia F. Fiorentini C. Cipolla C.M. Anthracycline-induced cardiomyopathy: Clinical relevance and response to pharmacologic therapy J. Am. Coll. Cardiol. 2010 55 213 220 10.1016/j.jacc.2009.03.095 20117401 5. Nakamura M. Sadoshima J. Mechanisms of physiological and pathological cardiac hypertrophy Nat. Rev. Cardiol. 2018 15 387 407 10.1038/s41569-018-0007-y 29674714 6. Hunton D.L. Lucchesi P.A. Pang Y. Cheng X. Dell'Italia L.J. Marchase R.B. Capacitative calcium entry contributes to nuclear factor of activated T-cells nuclear translocation and hypertrophy in cardiomyocytes J. Biol. Chem. 2002 277 14266 14273 10.1074/jbc.M107167200 11827959 7. Ohba T. Watanabe H. Murakami M. Sato T. Ono K. Ito H. Essential role of STIM1 in the development of cardiomyocyte hypertrophy Biochem. Biophys. Res. Commun. 2009 389 172 176 10.1016/j.bbrc.2009.08.117 19715666 8. Voelkers M. Salz M. Herzog N. Frank D. Dolatabadi N. Frey N. Gude N. Friedrich O. Koch W.J. Katus H.A. Orai1 and Stim1 regulate normal and hypertrophic growth in cardiomyocytes J. Mol. Cell. Cardiol. 2010 48 1329 1334 10.1016/j.yjmcc.2010.01.020 20138887 9. Hulot J.S. Fauconnier J. Ramanujam D. Chaanine A. Aubart F. Sassi Y. Merkle S. Cazorla O. Ouille A. Dupuis M. Critical role for stromal interaction molecule 1 in cardiac hypertrophy Circulation 2011 124 796 805 10.1161/CIRCULATIONAHA.111.031229 21810664 10. Pan Z. Brotto M. Ma J. Store-operated Ca2+ entry in muscle physiology and diseases BMB Rep. 2014 47 69 79 10.5483/BMBRep.2014.47.2.015 24411466 11. Luo R. Gomez A.M. Benitah J.P. Sabourin J. Targeting Orai1-Mediated Store-Operated Ca2+ Entry in Heart Failure Front. Cell Dev. Biol. 2020 8 586109 10.3389/fcell.2020.586109 33117812 12. Nan J. Li J. Lin Y. Saif Ur Rahman M. Li Z. Zhu L. The interplay between mitochondria and store-operated Ca2+ entry: Emerging insights into cardiac diseases J. Cell. Mol. Med. 2021 25 9496 9512 10.1111/jcmm.16941 34564947 13. Bozza W.P. Takeda K. Alterovitz W.L. Chou C.K. Shen R.F. Zhang B. Anthracycline-Induced Cardiotoxicity: Molecular Insights Obtained from Human-Induced Pluripotent Stem Cell-Derived Cardiomyocytes (hiPSC-CMs) AAPS J. 2021 23 44 10.1208/s12248-021-00576-y 33719006 14. Chan E.M. Thomas M.J. Bandy B. Tibbits G.F. Effects of doxorubicin, 4'-epirubicin, and antioxidant enzymes on the contractility of isolated cardiomyocytes Can. J. Physiol. Pharmacol. 1996 74 904 910 10.1139/y96-091 8960379 15. Swift L. McHowat J. Sarvazyan N. Inhibition of membrane-associated calcium-independent phospholipase A2 as a potential culprit of anthracycline cardiotoxicity Cancer Res. 2003 63 5992 5998 14522926 16. Claycomb W.C. Lanson N.A. Jr. Stallworth B.S. Egeland D.B. Delcarpio J.B. Bahinski A. Izzo N.J. Jr. HL-1 cells: A cardiac muscle cell line that contracts and retains phenotypic characteristics of the adult cardiomyocyte Proc. Natl. Acad. Sci. USA 1998 95 2979 2984 10.1073/pnas.95.6.2979 9501201 17. Luo Y. Liu X. Ma R. Wang Y. Zimering M. Pan Z. Circulating IgGs in Type 2 Diabetes with Atrial Fibrillation Induce IP3-Mediated Calcium Elevation in Cardiomyocytes iScience 2020 23 101036 10.1016/j.isci.2020.101036 32315831 18. Cui C. Chang Y. Zhang X. Choi S. Tran H. Penmetsa K.V. Viswanadha S. Fu L. Pan Z. Targeting Orai1-mediated store-operated calcium entry by RP4010 for anti-tumor activity in esophagus squamous cell carcinoma Cancer Lett. 2018 432 169 179 10.1016/j.canlet.2018.06.006 29908962 19. Choi S. Cui C. Luo Y. Kim S.H. Ko J.K. Huo X. Ma J. Fu L.W. Souza R.F. Korichneva I. Selective inhibitory effects of zinc on cell proliferation in esophageal squamous cell carcinoma through Orai1 FASEB J. Off. Publ. Fed. Am. Soc. Exp. Biol. 2018 32 404 416 10.1096/fj.201700227RRR 20. Hansen J. Xiong Y. Dhanan P. Hu B. Jayaraman G. Tolentino R. Chen Y. Beaumont K.G. Sebra R. Vidovic D. Transcriptomic Profiling of Drug-Treated Human Induced Pluripotent Stem Cells (iPSCs) 2022 Available online: (accessed on 3 May 2021) 21. Hawkins B.J. Irrinki K.M. Mallilankaraman K. Lien Y.C. Wang Y. Bhanumathy C.D. Subbiah R. Ritchie M.F. Soboloff J. Baba Y. S-glutathionylation activates STIM1 and alters mitochondrial homeostasis J. Cell Biol. 2010 190 391 405 10.1083/jcb.201004152 20679432 22. Suarez-Huerta N. Mosselmans R. Dumont J.E. Robaye B. Actin depolymerization and polymerization are required during apoptosis in endothelial cells J. Cell. Physiol. 2000 184 239 245 10.1002/1097-4652(200008)184:2<239::AID-JCP12>3.0.CO;2-R 10867649 23. Toldo S. Goehe R.W. Lotrionte M. Mezzaroma E. Sumner E.T. Biondi-Zoccai G.G. Seropian I.M. Van Tassell B.W. Loperfido F. Palazzoni G. Comparative cardiac toxicity of anthracyclines in vitro and in vivo in the mouse PLoS ONE 2013 8 e58421 10.1371/journal.pone.0058421 23516478 24. Eder P. Molkentin J.D. TRPC channels as effectors of cardiac hypertrophy Circ. Res. 2011 108 265 272 10.1161/CIRCRESAHA.110.225888 21252153 25. Pu W.T. Ma Q. Izumo S. NFAT transcription factors are critical survival factors that inhibit cardiomyocyte apoptosis during phenylephrine stimulation in vitro Circ. Res. 2003 92 725 731 10.1161/01.RES.0000069211.82346.46 12663489 26. Kar P. Parekh A.B. Distinct spatial Ca2+ signatures selectively activate different NFAT transcription factor isoforms Mol. Cell 2015 58 232 243 10.1016/j.molcel.2015.02.027 25818645 27. Touchberry C.D. Elmore C.J. Nguyen T.M. Andresen J.J. Zhao X. Orange M. Weisleder N. Brotto M. Claycomb W.C. Wacker M.J. Store-operated calcium entry is present in HL-1 cardiomyocytes and contributes to resting calcium Biochem. Biophys. Res. Commun. 2011 416 45 50 10.1016/j.bbrc.2011.10.133 22079292 28. Yuan J.P. Zeng W. Huang G.N. Worley P.F. Muallem S. STIM1 heteromultimerizes TRPC channels to determine their function as store-operated channels Nat. Cell Biol. 2007 9 636 645 10.1038/ncb1590 17486119 29. Zhang X. Xin P. Yoast R.E. Emrich S.M. Johnson M.T. Pathak T. Benson J.C. Azimi I. Gill D.L. Monteith G.R. Distinct pharmacological profiles of ORAI1, ORAI2, and ORAI3 channels Cell Calcium 2020 91 102281 10.1016/j.ceca.2020.102281 32896813 30. He L.P. Hewavitharana T. Soboloff J. Spassova M.A. Gill D.L. A functional link between store-operated and TRPC channels revealed by the 3,5-bis(trifluoromethyl)pyrazole derivative, BTP2 J. Biol. Chem. 2005 280 10997 11006 10.1074/jbc.M411797200 15647288 31. Benard L. Oh J.G. Cacheux M. Lee A. Nonnenmacher M. Matasic D.S. Kohlbrenner E. Kho C. Pavoine C. Hajjar R.J. Cardiac Stim1 Silencing Impairs Adaptive Hypertrophy and Promotes Heart Failure Through Inactivation of mTORC2/Akt Signaling Circulation 2016 133 1458 1471 1458-1471; discussion 1471 10.1161/CIRCULATIONAHA.115.020678 26936863 32. Luo X. Hojayev B. Jiang N. Wang Z.V. Tandan S. Rakalin A. Rothermel B.A. Gillette T.G. Hill J.A. STIM1-dependent store-operated Ca2+ entry is required for pathological cardiac hypertrophy J. Mol. Cell. Cardiol. 2012 52 136 147 10.1016/j.yjmcc.2011.11.003 22108056 33. Garlick P.B. Davies M.J. Hearse D.J. Slater T.F. Direct detection of free radicals in the reperfused rat heart using electron spin resonance spectroscopy Circ. Res. 1987 61 757 760 10.1161/01.RES.61.5.757 2822281 34. Haines D.D. Juhasz B. Tosaki A. Management of multicellular senescence and oxidative stress J. Cell. Mol. Med. 2013 17 936 957 10.1111/jcmm.12074 23789967 35. Miller B.A. The role of TRP channels in oxidative stress-induced cell death J. Membr. Biol. 2006 209 31 41 10.1007/s00232-005-0839-3 16685599 36. Poteser M. Graziani A. Rosker C. Eder P. Derler I. Kahr H. Zhu M.X. Romanin C. Groschner K. TRPC3 and TRPC4 associate to form a redox-sensitive cation channel. Evidence for expression of native TRPC3-TRPC4 heteromeric channels in endothelial cells J. Biol. Chem. 2006 281 13588 13595 10.1074/jbc.M512205200 16537542 37. Venkatakrishnan C.D. Tewari A.K. Moldovan L. Cardounel A.J. Zweier J.L. Kuppusamy P. Ilangovan G. Heat shock protects cardiac cells from doxorubicin-induced toxicity by activating p38 MAPK and phosphorylation of small heat shock protein 27 Am. J. Physiol. Heart Circ. Physiol. 2006 291 H2680 H2691 10.1152/ajpheart.00395.2006 16782845 38. Grzanka D. Domaniewski J. Grzanka A. Effect of doxorubicin on actin reorganization in Chinese hamster ovary cells Neoplasma 2005 52 46 51 15739026 39. Taskin E. Guven C. Kaya S.T. Sahin L. Kocahan S. Degirmencioglu A.Z. Gur F.M. Sevgiler Y. The role of toll-like receptors in the protective effect of melatonin against doxorubicin-induced pancreatic beta cell toxicity Life Sci. 2019 233 116704 10.1016/j.lfs.2019.116704 31369761 40. Mustafa E.H. Mahmoud H.T. Al-Hudhud M.Y. Abdalla M.Y. Ahmad I.M. Yasin S.R. Elkarmi A.Z. Tahtamouni L.H. 2-deoxy-D-Glucose Synergizes with Doxorubicin or L-Buthionine Sulfoximine to Reduce Adhesion and Migration of Breast Cancer Cells Asian Pac. J. Cancer Prev. 2015 16 3213 3222 10.7314/APJCP.2015.16.8.3213 25921122 41. Zhu H. Zhang H. Jin F. Fang M. Huang M. Yang C.S. Chen T. Fu L. Pan Z. Elevated Orai1 expression mediates tumor-promoting intracellular Ca2+ oscillations in human esophageal squamous cell carcinoma Oncotarget 2014 5 3455 3471 10.18632/oncotarget.1903 24797725 42. Pan Z. Ma J. Open Sesame: Treasure in store-operated calcium entry pathway for cancer therapy Sci. China. Life Sci. 2015 58 48 53 10.1007/s11427-014-4774-3 25481035 43. Cui C. Merritt R. Fu L. Pan Z. Targeting calcium signaling in cancer therapy Acta Pharm. Sin. B 2017 7 3 17 10.1016/j.apsb.2016.11.001 28119804 |
PMC10000559 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050762 cells-12-00762 Article Application of Digital Holographic Microscopy to Analyze Changes in T-Cell Morphology in Response to Bacterial Challenge vom Werth Kari Lavinia 1 Kemper Bjorn Methodology Software Writing - review & editing 2 Kampmeier Stefanie Conceptualization Investigation Data curation Writing - review & editing Supervision Project administration Funding acquisition 1 Mellmann Alexander Conceptualization Validation Investigation Resources Writing - review & editing Supervision Project administration Funding acquisition 1* Gonzalez Cruz Jazmina Libertad Academic Editor Leggatt Graham Academic Editor 1 Institute of Hygiene, University Hospital Munster, 48149 Munster, Germany 2 Biomedical Technology Center of the Medical Faculty, University of Munster, 48149 Munster, Germany * Correspondence: [email protected]; Tel.: +49-251-83-55361 27 2 2023 3 2023 12 5 76225 1 2023 16 2 2023 24 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Quantitative phase imaging (QPI) is a non-invasive, label-free technique used to detect aberrant cell morphologies caused by disease, thus providing a useful diagnostic approach. Here, we evaluated the potential of QPI to differentiate specific morphological changes in human primary T-cells exposed to various bacterial species and strains. Cells were challenged with sterile bacterial determinants, i.e., membrane vesicles or culture supernatants, derived from different Gram-positive and Gram-negative bacteria. Timelapse QPI by digital holographic microscopy (DHM) was applied to capture changes in T-cell morphology over time. After numerical reconstruction and image segmentation, we calculated single cell area, circularity and mean phase contrast. Upon bacterial challenge, T-cells underwent rapid morphological changes such as cell shrinkage, alterations of mean phase contrast and loss of cell integrity. Time course and intensity of this response varied between both different species and strains. The strongest effect was observed for treatment with S. aureus-derived culture supernatants that led to complete lysis of the cells. Furthermore, cell shrinkage and loss of circular shape was stronger in Gram-negative than in Gram-positive bacteria. Additionally, T-cell response to bacterial virulence factors was concentration-dependent, as decreases in cellular area and circularity were enhanced with increasing concentrations of bacterial determinants. Our findings clearly indicate that T-cell response to bacterial stress depends on the causative pathogen, and specific morphological alterations can be detected using DHM. digital holographic microscopy sepsis bacteria T-cells cell morphology Deutsche Forschungsgemeinschaft (DFG, German Research Foundation), project P4 of the CRU342427775125 This research was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation), project P4 of the CRU342 (project number 427775125). pmc1. Introduction Bacterial infections can quickly progress into sepsis, a life-threatening condition caused by a dysregulated host response . Sepsis incidence and mortality remain high with 49 million cases and 11 million deaths per year worldwide . Late diagnosis and delayed or inappropriate treatment are major contributors to these fatality rates. Therefore, different biomarkers, e.g., cytokines, chemokines or acute-phase proteins, have been proposed to facilitate identification of patients at high risk for subsequent organ failure . However, sepsis is a highly complex and heterogeneous syndrome; hence, no single biomarker had sufficient sensitivity and specificity to reliably detect septic patients . In addition to singular molecules as biomarker, analysis of immune cell morphology came into focus as a potential diagnostic approach. Numerous studies using automated hematology analyzers to measure leucocyte volume, conductivity and scatter (VCS) reported changes in monocyte and neutrophil morphology in septic patients compared to non-infected or healthy controls . Variations in cell size, shape and composition correlated with sepsis severity and prognosis and allowed for monitoring of treatment efficacy . Quantitative phase imaging (QPI) is an emerging method in the context of clinical diagnostics . Digital holographic microscopy (DHM) , which is a variant of QPI, enables the determination of morphological and physical parameters of living cells. It requires neither fixation nor staining or labeling procedures, making it a simple, fast and non-invasive imaging technique that is suitable for long-term measurements. Thereby, DHM provides a promising tool for the diagnosis of diseases that are associated with altered cell morphologies. Potential applications that have been reported thus far include identification of cancer cells , sickle cell disease and infectious diseases such as malaria or COVID-19 . Since sepsis is also associated with morphological changes in circulating immune cells , we aimed to investigate the potential of DHM in sepsis diagnosis. Previous studies showed that DHM is a suitable method to monitor changes in cell morphology in response to bacterial toxins over time . In this study, we focused on T-cells as part of the adaptive immune system. T-cell immunity strongly depends on the invading pathogen with great variations not only between different bacterial species but between also strains within the same species . Based on this context, we hypothesize that T-cells under bacterial stress exhibit changes in cell morphology that are specific for the causative bacteria. Therefore, we chose different Gram-negative (Escherichia coli, Klebsiella pneumoniae) and Gram-positive (Staphylococcus aureus, Streptococcus pneumoniae, Enterococcus faecium and E. faecalis) strains that are frequently detected in septic patients . T-cells were either exposed to sterile culture supernatants, bacterial membrane vesicles (MVs) or living bacteria. Both Gram-positive and Gram-negative bacteria produce a number of virulence factors that enable them to interfere with the immune response. The systemic spread of such microbial toxins is a key driver of disease progression and contributes to the outcome of sepsis patients . Additionally, MVs are used by many pathogenic bacteria to deliver toxins and other biomolecules into host cells . By entering the systemic circulation, MVs are able to spread throughout the body and interfere with the immune system, thereby contributing to disease progression . Ultimately, we want to find whether the analysis of T-cell morphology using DHM offers a new approach to support early sepsis diagnosis and prediction of the causative agent. Starting an appropriate treatment as early as possible is essential to improve the patient's outcome and decrease mortality among septic patients. 2. Materials and Methods 2.1. T-Cell Isolation Whole blood samples were collected from healthy volunteers after written informed consent was obtained. T-cells were isolated by negative selection using the RosetteSep Human T-cell Enrichment Cocktail (StemCell Technologies, Cologne, Germany) according to the manufacturer's instructions. Following a density gradient centrifugation with Lymphocyte Separation Medium (Promocell, Heidelberg, Germany) at 800 g for 30 min, the cells were washed twice with PBS (Sigma-Aldrich, Darmstadt, Germany) supplemented with 2% bovine serum albumin (BSA; Serva, Heidelberg, Germany). Finally, the cells were cultured in Roswell Park Memorial Institute 1640 medium (RPMI; Sigma-Aldrich, Darmstadt, Germany) supplemented with 10% fetal calf serum (FCS; PAA, Pasching, Austria) and 2 mM ultraglutamine (Lonza, Cologne, Germany) in a humidified atmosphere at 37 degC and 5% CO2. 2.2. Bacterial Strains and Culture Conditions The bacterial strains used in this study are listed in Table 1. E. coli, S. aureus, and K. pneumoniae strains were grown in lysogeny broth (LB; Roth, Karlsruhe, Germany) while E. faecium/faecalis and S. pneumoniae strains were grown in brain heart infusion (BHI) medium (Roth, Karlsruhe, Germany). 2.3. Preparation of Bacterial MVs Bacterial MVs were isolated from overnight cultures as described previously . E. coli, S. aureus, E. faecium/faecalis and K. pneumoniae strains were grown at 37 degC with shaking at 180 rpm. S. pneumoniae strains were grown stationary at 37 degC and 5% CO2. Bacteria were removed from the overnight cultures by centrifugation (5600x g, 20 min, 4 degC) and sterile filtration through 0.22 mm pore-size filters. Subsequently, MVs were collected by ultracentrifugation (235,000x g, 2 h, 4 degC) in a 45 Ti rotor (Beckman Coulter, Krefeld, Germany), and the pellet was resuspended in 20 mM TRIS-HCl (pH 8.0). The MV suspensions were stored at 4 degC and used for experiments within three months after preparation. Nanoparticle tracking analysis (NTA) was performed with a NanoSight NS300 instrument (Malvern Panalytical, Kassel, Germany) to determine the MV size and concentration. For every sample, five videos of 60 s were recorded, and particle size distribution and concentration were analyzed by NanoSight NTA 3.4 software (version 3.4.4; Malvern Panalytical, Kassel, Germany). All measurements were performed under constant flow with temperature control at 25 degC. When T-cells were exposed to bacterial MVs, the MV suspension was diluted in cell culture medium (RPMI + 10% FCS + 2 mM ultraglutamine) to a final concentration of 2 x 109 particle/mL. 2.4. Preparation of Sterile Culture Supernatants For preparation of sterile culture supernatants, bacteria were grown in cell culture medium (RPMI + 10% FCS + 2 mM ultraglutamine) overnight at 37 degC and 180 rpm (E. coli, S. aureus, E. faecium/faecalis, K. pneumoniae) or 37 degC and 5% CO2 (S. pneumoniae). Bacteria were sedimented by centrifugation (5600x g, 5 min, RT), and the supernatant was passed through 0.22 mm pore-size filters (Corning, Wiesbaden, Germany). Sterile supernatants were freshly prepared prior to every experiment, and T-cells were treated with different concentrations between 5% and 100% (v/v). 2.5. In Vitro Infection Bacteria were grown in overnight cultures at 37 degC with shaking at 180 rpm. To prepare the infection, the cultures were centrifuged at 5600x g and RT for 5 min. The pellet was resuspended in PBS and adjusted to an optical density at 600 nm (OD600) of 1 in PBS (corresponding to 5 x 108 colony forming units (CFU)/mL). T-cells were seeded at a density of 4 x 105 cells/mL and infected with different multiplicities of infection (MOI) ranging from 1 to 10. Four hours post infection (p.i.), 2 mg/mL Lysostaphin (Sigma-Aldrich, Darmstadt, Germany) was added to prevent overgrowth of the bacteria, and timelapse DHM was started. As control, cells were treated in the same way but without addition of bacteria. 2.6. Digital Holographic Microscopy (DHM) Timelapse QPI with DHM was performed with a fiber optic Mach-Zehnder DHM setup via a 20x microscope objective (Zeiss, Jena, Germany) as previously described with minor modifications. Briefly, off-axis holograms were captured automatically every three minutes using a single longitudinal mode laser (Cobolt 06-DPL, l = 532 nm, 25 mW; Cobolt AB, Solna, Sweden) with an exposure time of 0.15 ms. During acquisition, the sample illumination light was modulated by an electrically focus tunable lens (ETL; Optotune, Dietikon, Switzerland) while series of digital off-axis holograms were recorded to reduce image disturbances caused by the coherence properties of the applied laser light . Quantitative phase contrast images were numerically reconstructed as previously reported using custom-built software implemented in python 3.7. For DHM measurements, T-cells were seeded in 4-well Ph+ m-slides (ibidi, Grafelfing, Germany) at a density of 4 x 105 cells/mL, and the slides were sealed with anti-evaporation oil (ibidi, Grafelfing, Germany). Holograms were captured every three minutes at three different positions for every treatment to record at least 100 cells per time point for a reliable evaluation . Subsequently, quantitative phase images were reconstructed, and further analysis including segmentation and determination of morphological parameters was performed with FIJI software (version 2.3.0/1.53f51) . Morphological parameters such as single cell area, perimeter, circularity and phase contrast values were calculated for every measuring point as mean across all detected cells. Three or more independent biological replicates were performed for every treatment using T-cells obtained from different donors. Results are presented as mean +- SD. 3. Results 3.1. Species-Dependent Morphological Changes in Reaction to Bacterial Culture Supernatants To investigate the effect of secreted virulence factors on T-cell morphology, human primary T-cells were exposed to sterile culture supernatants of bacterial overnight cultures, and the cellular response was monitored with DHM. Ten hours after the addition of culture supernatants, we observed distinct T-cell shapes, depending on the causative bacteria . In addition to pseudo-colored phase contrast images , cross-sections through single cells are shown to clearly outline the different cell morphologies. S. aureus supernatant elicited the strongest effect with entire lysis of the cells. S. pneumoniae was the only species leading to an increase in cellular phase contrast while cells treated with K. pneumoniae supernatant showed overall decreased phase contrast values. For E. faecalis and E. coli supernatants, we observed similar morphological changes. Small compartments with higher phase contrast formed inside the cells, whereas in non-treated control cells, the phase contrast values were evenly distributed. Applying timelapse DHM with image acquisition every three minutes, we were able to quantify changes in cell morphology over time. We included two to three different strains for every species to analyze not only species-specific but also strain-specific differences. The overall pattern of morphological changes varied between the different species but not the strains . Both strains of S. pneumoniae, for instance, led to an increase in mean phase contrast and reduction of cell area while the circular shape, quantified by the parameter circularity, was retained. In contrast, supernatant derived from E. coli induced a decrease in the cell area and circularity, whereas the mean phase contrast remained constant. The strongest effects were observed for S. aureus supernatants that led to rapid lysis of the cells. For that reason, the graphs of strain 6850 and USA300 were cut after 2.5 and 3 h, respectively. Even with dilution down to 5% (v/v), S. aureus supernatants induced a strong decrease in T-cell area . S. pneumoniae supernatants caused minor morphological changes at low concentrations, while we did not observe any effect of the remaining species. Although the response pattern within one species was similar, different strains could be distinguished by the intensity and time course of the morphological changes . The pathogenic E. coli strain IHE3034, for example, led to a stronger decrease in circularity than the commensal strain MG1655. Taken together, all strains used in this study induced specific morphological changes and differed from each other in at least one of the analyzed parameters. 3.2. T-Cell Response to S. aureus MVs Depends on the Strain In addition to soluble virulence factors, bacteria release MVs during their normal growth that contain a range of biological cargo including nucleic acids, proteins, enzymes and toxins. MVs can stimulate the innate as well as the adaptive immune response, thereby contributing to pathogenesis within the host . We therefore exposed T-cells to bacterial MVs and monitored changes in single-cell area, circularity and mean phase contrast with timelapse DHM. Only MVs derived from S. aureus induced a cellular response . Pseudo-colored phase contrast images clearly show a loss of cell integrity . This observation is quantified by a decrease in single cell area and circularity score . In line with the supernatant treatment, the intensity of the cellular response to MVs varied between the different strains. S. aureus strain USA300-derived MVs induced the most rapid and strongest effects with lysis of the cells after 3 and 7 h, respectively . In contrast, we did not observe any effects on cell morphology when T-cells were treated with MVs derived from S. pneumoniae, E. faecium, E. faecalis, E. coli or K. pneumoniae . 3.3. Cellular Changes in Response to Bacterial Stress Are Concentration-Dependent It was shown that alterations of neutrophil morphology and motility in sepsis patients correlate with disease severity . Thus, we hypothesized that the T-cell response to living bacteria or bacterial virulence factors might also be concentration-dependent. To test this hypothesis, T-cells were treated with different concentrations of sterile culture supernatant or were infected with increasing MOIs of living S. aureus strain 6850 . The cellular response to sterile culture supernatant was considerably stronger compared to live bacteria. Even at low supernatant concentrations of 5% (v/v), T-cell area as well as circularity were rapidly reduced, and increasing the concentration enhanced this effect . After infection with living bacteria, cell area and circularity also decreased but to a lesser extent . In line with the previous results, these changes were also concentration-dependent, and a higher bacterial load led to a stronger T-cell response. 4. Discussion DHM is a minimally invasive imaging technique that can give information about the cell number and different morphological parameters including area, thickness, volume or shape . In contrast to other staining-based methods, DHM is cost-saving and overcomes several limitations including phototoxicity or photobleaching. The samples are only exposed to low laser light intensities, allowing for long-term measurements without affecting cell viability. Various diseases are associated with altered cell morphology and motility. Analysis of cellular morphological changes might therefore provide a useful diagnostic tool in the clinical context . In our study, we aimed to investigate the potential of DHM to monitor changes in cell morphology in response to bacterial stress. Furthermore, we were interested in whether differences between various causative agents can be captured. Here, we focused on adaptive immune cells, i.e., T-lymphocytes, as the T-cell response to bacterial pathogens greatly varies depending on the causative agent . Primary T-cells were isolated from healthy volunteers and were exposed to bacterial determinants derived from different Gram-negative and Gram-positive species that are frequently detected in septic patients. Additionally, we used multiple strains of the same species to analyze not only also strain-dependent differences. Timelapse DHM was applied to monitor the changes in cell morphology over time. Pathogenic bacteria produce and secrete a broad range of virulence factors that enable them to evade the immune response, disseminate within the host and cause disease . Bacterial toxins bind to specific receptors on the cell surface and are able to trigger a dysregulated host response, thereby contributing to the progression of an infection to sepsis . We therefore treated primary T-cells with sterile supernatants from bacterial overnight cultures to analyze the effect of secreted virulence factors on the cell morphology. We observed large variations between the different species and strains . Representative images 10 h after exposure to culture supernatants show clearly different cell shapes and phase contrast distribution at this time point . Cross-sections through single cells are depicted to visualize the differences in cell morphology. Using numerically reconstructed phase contrast images, we quantified changes in single-cell area, circularity and mean phase contrast per cell over time . S. aureus strain 6850 and USA300 caused the strongest effects, with entire loss of cell integrity. Cell area and circularity rapidly decreased, and already a few hours after exposure, we were not able to detect any cells. In contrast, all other strains used in this study also induced a reduction of the cell area, but the cells were not completely lysed. The extent of cell shrinkage varied between different bacterial species but not between strains. Strain-specific differences could be observed regarding the cell shape. The extent to which the circularity of single cells declined differed not only between different species but also between strains of the same species. Supernatants from both S. pneumoniae strains induced an increase in the mean phase contrast, whereas this parameter was not affected or did not decrease in all other species investigated. Taken together, we observed specific patterns of cellular changes that depended on both the bacterial species and strain. Bacteria are not equally pathogenic since every strain expresses a specific set of virulence factors leading to diverse cellular reactions . S. aureus, as an example, produces various toxins that are able to induce T-cell death, and especially, the a-toxin has a strong cytotoxic effect on T-lymphocytes . The expression of these toxins varies greatly between different S. aureus strains, thereby influencing the intensity of the cytotoxic effect . Furthermore, different bacteria trigger different cell death pathways that are characterized by specific morphological features . In a recent study, it was shown that DHM in combination with a deep learning algorithm is able to distinguish apoptotic from necrotic cells . However, it is also known that the immune response not only depends on the invading pathogen but also on different host factors . To address this issue, every treatment was performed at least three times, with T-cells obtained from different donors. Therefore, we can assume that the different observed effects of bacterial virulence factors on T-cell morphology do not result from donor-specific responses. Based on this context, changes in the cell morphology analyzed by DHM might give information about the causative agent. Gram-negative as well as Gram-positive bacteria produce MVs that contain a broad range of virulence factors and are able to interfere with the immune system . We therefore aimed to investigate if we can confirm our previous results of a strain-specific T-cell response by using MVs isolated from overnight cultures. Previous studies with MVs derived from different bacterial species already showed a dose-dependent cytotoxic effect on various cell types . The physiological concentration of MVs during bacterial infection has not been reported thus farm although MVs were detected inside the bloodstream . For our experiments, we therefore chose a concentration that is in the range of other studies in this research field. Only MVs derived from S. aureus strains induced changes in T-cell morphology . In line with our previous observations, strain 6850 and USA300 caused a strong and fast reduction of cell area and entire loss of cell integrity while strain ST398 had a weaker impact on cell morphology . S. aureus MVs contain biologically active a-toxin that strongly contributes to the cytotoxic effect . In addition, MVs are equipped with numerous other toxins, nucleic acids, proteins and enzymes . Differences regarding the extent of cytotoxicity might therefore result from variations in the MV proteome . In contrast, neither S. pneumoniae nor E. faecalis/faecium, E. coli or K. pneumoniae derived MVs were able to induce alterations of T-cell morphology , which is in line with previous studies . Changes in neutrophil motility, morphology and mechanics in septic patients correlate with disease severity . We therefore hypothesized that the T-cell response to bacterial virulence factors or living bacteria is also concentration-dependent. To test this hypothesis, primary T-cells were exposed to different concentrations of culture supernatant derived from S. aureus strain 6850 or infected with different amounts of living bacteria . As expected, the changes in T-cell morphology strongly depended on the concentration under both conditions. However, this effect was more pronounced when the cells were treated with sterile culture supernatant compared to living bacteria. This might result from our experimental procedure. In order to prevent bacterial overgrowth and interference with the DHM measurement, we removed the bacteria four hours after infection of the cells. The cellular response under these conditions might therefore be weaker compared to the supernatant treatment. Taken together, we were able to show that the T-cell response upon bacterial challenge depends on the causative pathogen and the concentration of the bacterial determinant. Alterations of morphological parameters such as single-cell area, cell shape or in quantitative phase contrast can be captured using DHM. However, our study has some limitations. Despite the use of primary T-cells, our cell culture approach can only partly reflect the in vivo situation. The immune response to infection and development of sepsis is based on a complex interaction of different cell types, including direct cell contacts and production of various cytokines . However, T-cells themselves also express pattern recognition receptors, e.g., toll-like receptors, and can be directly activated by virulence factors . We therefore assume that our results relate to the in vivo setting to a certain degree, and T-cells from sepsis patients are morphologically different compared to healthy controls. Previous studies using hematology analyzers described significant alterations of monocyte and neutrophil cell sizes under septic conditions compared to healthy controls . In our future studies, we aim to discriminate not only between infection and sepsis but also between different causative agents. Thereby, we want to investigate the potential of DHM in the context of sepsis diagnosis and identification of the underlying bacteria. A correct and rapid diagnosis along with early administration of pathogen-specific antibiotics is crucial to improve patient outcome . For conventional blood culture procedures, it is necessary to grow the bacteria, which is time-consuming and can take up to 72 h. Moreover, blood cultures are often negative, impeding a targeted treatment of the patient. If we are able to transfer our findings into the clinical setting, application of QPI by DHM could rapidly provide useful information about the infecting agent and guide the antibiotic treatment. DHMs are commercially available and could be easily implemented into the clinical workflow. No staining or labeling procedures are applied, making it is a fast and easy-to-use method that does not require extensive training. Moreover, QPI provides quantitative information, allowing for the integration of machine-learning models to improve and facilitate data interpretation. DHM could rather complement established diagnostic tools than replace them. Evaluation of immune cell morphology cannot reach the same accuracy such as bacteriology techniques or provide information about antibiotic resistances. Additional tests, such as antibiotic susceptibility testing, are still needed for a precise diagnosis. However, QPI could provide first information to support an early treatment initiation that is followed by the more time-consuming methods for further specification. Our study indicates that analysis of T-cell morphology with DHM might be a promising approach to support such an early identification of sepsis patients. Acknowledgments We thank the Institute of Medical Microbiology, University Hospital of Munster, Munster, Germany for providing the S. aureus, S. pneumoniae and K. pneumoniae isolates. Supplementary Materials The following supporting information can be downloaded at: Figure S1: Morphological changes in T-cells treated with 5% culture supernatant; Figure S2: Morphological changes in T-cells treated with bacterial MVs. Click here for additional data file. Author Contributions Conceptualization, S.K. and A.M.; methodology, K.L.v.W. and B.K.; software, B.K.; investigation, K.L.v.W., S.K. and A.M.; resources, S.K. and A.M.; writing--original draft preparation, K.L.v.W.; writing--review and editing, K.L.v.W., B.K., S.K. and A.M.; visualization, K.L.v.W. and B.K.; supervision, A.M.; project administration, A.M.; funding acquisition, S.K. and A.M. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and was approved by the ethics committee of the Medical Association Westphalia-Lippe (Germany) and the University of Muenster (registration number: 2022-274-f-S). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Distinct cell morphologies after exposure to bacterial culture supernatant derived from different species. Human primary T-cells were treated with sterile bacterial culture supernatants derived from various species or left untreated as control. DHM was applied to visualize morphological changes. Representative images illustrating the cell morphology 10 h after the addition of supernatants are shown. (A) Pseudo-colored phase contrast images. The scale corresponds to 10 mm. (B) Distribution of phase contrast of cross-sections through single cells (black dashed line in A). Figure 2 T-cell response to bacterial culture supernatant depends on both the species and strain. Human primary T-cells were exposed to sterile bacterial culture supernatants from various species and strains (colored lines) or left untreated as control (black lines). Timelapse DHM was applied to monitor the cellular response over time, and the resulting quantitative phase contrast images were analyzed for average single cell area, average circularity and mean phase contrast per cell. Results are presented as mean (solid lines) +- SD (shading) of at least three independent experiments. The curves were smoothed using moving averages with a window size of 15 measuring points. Figure 3 S. aureus MVs induce rapid changes in T-cell morphology that vary between different strains. Human primary T-cells were treated with MVs (2 x 109 particle/mL) derived from different S. aureus strains or left untreated as control. The phenotypical cell response was recorded over 24 h by timelapse DHM. (A) Average cell area, average circularity and mean phase contrast per cell were analyzed using quantitative phase contrast images. Results represent mean (solid lines) +- SD (shading) of at least three independent biological replicates. The curves were smoothed using moving averages with a window size of 15 measuring points. (B) Representative pseudo-colored phase contrast images at indicated measurement time points. The scale bar corresponds to 10 mm. Figure 4 Concentration-dependent effect of bacterial culture supernatant and living bacteria on T-cell morphology. (A) Human primary T-cells were exposed to different concentrations of S. aureus 6850 culture supernatant. (B) T-cells were infected with different bacterial loads of S. aureus 6850. Timelapse DHM was started 5 h after the addition of bacteria. (A,B) Quantitative phase contrast images were analyzed for average single cell area, average circularity and mean phase contrast per cell. Results are presented as mean (solid lines) +- SD (shading) of at least three independent measurements. The graphs were smoothed using moving averages with a window size of 15 measuring points. (C) Representative pseudo-colored phase contrast images of T-cells treated with S. aureus 6850 supernatant at indicated measurement time points. The scale bar corresponds to 10 mm. cells-12-00762-t001_Table 1 Table 1 Bacterial strains used in this study. Species Strain Reference S. aureus 6850 USA300 ST398 S. pneumoniae ATCC 49619 ATCC 49619 ATCC 6305 ATCC 6305 E. faecium ATCC 6057 ATCC 6057 E. faecalis ATCC 29212 ATCC 29212 E. coli 536 IHE3034 MG1655 K. pneumoniae ATCC 13883 ATCC 13883 ATCC 700603 ATCC 700603 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Singer M. Deutschman C.S. Seymour C. Shankar-Hari M. Annane D. Bauer M. Bellomo R. Bernard G.R. Chiche J.D. Coopersmith C.M. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) JAMA--J. Am. Med. Assoc. 2016 315 801 810 10.1001/jama.2016.0287 2. Rudd K.E. Johnson S.C. Agesa K.M. Shackelford K.A. Tsoi D. Kievlan D.R. Colombara D.V. Ikuta K.S. Kissoon N. Finfer S. Global, Regional, and National Sepsis Incidence and Mortality, 1990-2017: Analysis for the Global Burden of Disease Study Lancet 2020 395 200 211 10.1016/S0140-6736(19)32989-7 31954465 3. Barichello T. Generoso J.S. Singer M. Dal-Pizzol F. Biomarkers for Sepsis: More than Just Fever and Leukocytosis--A Narrative Review Crit. Care 2022 26 14 10.1186/s13054-021-03862-5 34991675 4. Pierrakos C. Velissaris D. Bisdorff M. Marshall J.C. Vincent J.-L. Biomarkers of Sepsis: Time for a Reappraisal Crit. Care 2020 24 287 10.1186/s13054-020-02993-5 32503670 5. Mammen J. Choudhuri J. Paul J. Sudarsan T.I. Josephine T. Mahasampath G. Jeyaseelan V. Nair S.C. Peter J.V. Cytomorphometric Neutrophil and Monocyte Markers May Strengthen the Diagnosis of Sepsis J. Intensive Care Med. 2018 33 656 662 10.1177/0885066616682940 30411670 6. Urrechaga E. Boveda O. Aguirre U. Improvement in Detecting Sepsis Using Leukocyte Cell Population Data (CPD) Clin. Chem. Lab. Med. 2019 57 918 926 10.1515/cclm-2018-0979 30838839 7. Arora P. Gupta P.K. Lingaiah R. Mukhopadhyay A.K. Volume, Conductivity, and Scatter Parameters of Leukocytes as Early Markers of Sepsis and Treatment Response J. Lab. Physicians 2019 11 29 33 10.4103/JLP.JLP_102_18 30983799 8. Shalini P. Rao P.S. Rao S. Anil M. Benny A. Diagnostic Utility of Cell Population Data (CPD) in Sepsis Using Automated Hematology Analysers Ann. Pathol. Lab. Med. 2019 6 284 10.21276/apalm.2395 9. Zonneveld R. Molema G. Plotz F.B. Analyzing Neutrophil Morphology, Mechanics, and Motility in Sepsis Crit. Care Med. 2016 44 218 228 10.1097/CCM.0000000000001266 26296143 10. Park Y.K. Depeursinge C. Popescu G. Quantitative Phase Imaging in Biomedicine Nat. Photonics 2018 12 578 589 10.1038/s41566-018-0253-x 11. Lee K. Kim K. Jung J. Heo J. Cho S. Lee S. Chang G. Jo Y. Park H. Park Y. Quantitative Phase Imaging Techniques for the Study of Cell Pathophysiology: From Principles to Applications Sensors 2013 13 4170 4191 10.3390/s130404170 23539026 12. Kemper B. Von Bally G. Digital Holographic Microscopy for Live Cell Applications and Technical Inspection Appl. Opt. 2008 47 A52 A61 10.1364/AO.47.000A52 18239699 13. El-Schich Z. Janicke B. Alm K. Dizeyi N. Persson L.J. Gjorloff Wingren A. Discrimination between Breast Cancer Cells and White Blood Cells by Non-Invasive Measurements: Implications for a Novel in Vitro-Based Circulating Tumor Cell Model Using Digital Holographic Cytometry Appl. Sci. 2020 10 4854 10.3390/app10144854 14. Benzerdjeb N. Garbar C. Camparo P. Sevestre H. Digital Holographic Microscopy as Screening Tool for Cervical Cancer Preliminary Study Cancer Cytopathol. 2016 124 573 580 10.1002/cncy.21727 27136615 15. O'Connor T. Anand A. Andemariam B. Javidi B. Overview of Cell Motility-Based Sickle Cell Disease Diagnostic System in Shearing Digital Holographic Microscopy JPhys Photonics 2020 2 031002 10.1088/2515-7647/ab8a58 16. Ugele M. Weniger M. Leidenberger M. Huang Y. Bassler M. Friedrich O. Kappes B. Hayden O. Richter L. Label-Free, High-Throughput Detection of P. Falciparum Infection in Sphered Erythrocytes with Digital Holographic Microscopy Lab Chip 2018 18 1704 1712 10.1039/C8LC00350E 29796511 17. O'Connor T. Santaniello S. Javidi B. COVID-19 Detection from Red Blood Cells Using Highly Comparative Time-Series Analysis (HCTSA) in Digital Holographic Microscopy Opt. Express 2022 30 1723 1736 10.1364/OE.442321 35209327 18. Urrechaga E. Reviewing the Value of Leukocytes Cell Population Data (CPD) in the Management of Sepsis Ann. Transl. Med. 2020 8 953 10.21037/atm-19-3173 32953753 19. Kunsmann L. Ruter C. Bauwens A. Greune L. Gluder M. Kemper B. Fruth A. Wai S.N. He X. Lloubes R. Virulence from Vesicles: Novel Mechanisms of Host Cell Injury by Escherichia Coli O104:H4 Outbreak Strain Sci. Rep. 2015 5 13252 10.1038/srep13252 26283502 20. Bauwens A. Bielaszewska M. Kemper B. Langehanenberg P. Von Bally G. Reichelt R. Mulac D. Humpf H.U. Friedrich A.W. Kim K.S. Differential Cytotoxic Actions of Shiga Toxin 1 and Shiga Toxin 2 on Microvascular and Macrovascular Endothelial Cells Thromb. Haemost. 2011 105 515 528 10.1160/TH10-02-0140 21136010 21. Sela U. Euler C.W. Correa da Rosa J. Fischetti V.A. Strains of Bacterial Species Induce a Greatly Varied Acute Adaptive Immune Response: The Contribution of the Accessory Genome PLoS Pathog. 2018 14 e1006726 10.1371/journal.ppat.1006726 29324905 22. Shepherd F.R. McLaren J.E. T Cell Immunity to Bacterial Pathogens: Mechanisms of Immune Control and Bacterial Evasion Int. J. Mol. Sci. 2020 21 6144 10.3390/ijms21176144 32858901 23. vom Werth K.L. Wormann T. Kemper B. Kumpers P. Kampmeier S. Mellmann A. Investigating Morphological Changes of T-Lymphocytes after Exposure with Bacterial Determinants for Early Detection of Septic Conditions Microorganisms 2022 10 391 10.3390/microorganisms10020391 35208846 24. Sakr Y. Jaschinski U. Wittebole X. Szakmany T. Lipman J. Namendys-Silva S.A. Martin-Loeches I. Leone M. Lupu M.N. Vincent J.-L. Sepsis in Intensive Care Unit Patients: Worldwide Data from the Intensive Care over Nations Audit Open Forum Infect. Dis. 2018 5 ofy313 10.1093/ofid/ofy313 30555852 25. GBD 2019 Antimicrobial Resistance Collaborators Global Mortality Associated with 33 Bacterial Pathogens in 2019: A Systematic Analysis for the Global Burden of Disease Study 2019 Lancet 2022 400 2248 10.1016/S0140-6736(22)02185-7 26. Ramachandran G. Gram-Positive and Gram-Negative Bacterial Toxins in Sepsis: A Brief Review Virulence 2014 5 213 218 10.4161/viru.27024 24193365 27. Schmitt C.K. Meysick K.C. O'Brien A.D. Bacterial Toxins: Friends or Foes? Emerg. Infect. Dis. 1999 5 224 234 10.3201/eid0502.990206 10221874 28. Macion A. Wyszynska A. Godlewska R. Delivery of Toxins and Effectors by Bacterial Membrane Vesicles Toxins (Basel). 2021 13 845 10.3390/toxins13120845 34941684 29. Kaparakis-Liaskos M. Ferrero R.L. Immune Modulation by Bacterial Outer Membrane Vesicles Nat. Rev. Immunol. 2015 15 375 387 10.1038/nri3837 25976515 30. Tulkens J. Vergauwen G. Van Deun J. Geeurickx E. Dhondt B. Lippens L. De Scheerder M.A. Miinalainen I. Rappu P. De Geest B.G. Increased Levels of Systemic LPS-Positive Bacterial Extracellular Vesicles in Patients with Intestinal Barrier Dysfunction Gut 2020 69 191 193 10.1136/gutjnl-2018-317726 30518529 31. Peng Y. Gao M. Liu Y. Qiu X. Cheng X. Yang X. Chen F. Wang E. Bacterial Outer Membrane Vesicles Induce Disseminated Intravascular Coagulation through the Caspase-11-Gasdermin D Pathway Thromb. Res. 2020 196 159 166 10.1016/j.thromres.2020.08.013 32882448 32. Svennerholm K. Park K.S. Wikstrom J. Lasser C. Crescitelli R. Shelke G.V. Jang S.C. Suzuki S. Bandeira E. Olofsson C.S. Escherichia Coli Outer Membrane Vesicles Can Contribute to Sepsis Induced Cardiac Dysfunction Sci. Rep. 2017 7 17434 10.1038/s41598-017-16363-9 29234030 33. Vann J.M. Proctor R.A. Ingestion of Staphylococcus Aureus by Bovine Endothelial Cells Results in Inoculum-Dependent Damage to Endothelial Cell Monolayers Infect. Immun. 1987 55 2155 2163 10.1128/iai.55.9.2155-2163.1987 3623696 34. Tenover F.C. Goering R.V. Methicillin-Resistant Staphylococcus Aureus Strain USA300: Origin and Epidemiology J. Antimicrob. Chemother. 2009 64 441 446 10.1093/jac/dkp241 19608582 35. Planet P.J. Life after USA300: The Rise and Fall of a Superbug J. Infect. Dis. 2017 215 S71 S77 10.1093/infdis/jiw444 28375517 36. Witte W. Strommenger B. Stanek C. Cuny C. Methicillin-Resistant Staphylococcus Aureus ST398 in Humans and Animals, Central Europe Emerg. Infect. Dis. 2007 13 255 258 10.3201/eid1302.060924 17479888 37. Berger H. Hacker J. Juarez A. Hughes C. Goebel W. Cloning of the Chromosomal Determinants Encoding Hemolysin Production and Mannose-Resistant Hemagglutination in Escherichia Coli J. Bacteriol. 1982 152 1241 1247 10.1128/jb.152.3.1241-1247.1982 6754701 38. Korhonen T.K. Valtonen M.V. Parkkinen J. Vaisanen-Rhen V. Finne J. Orskov F. Orskov I. Svenson S.B. Helena Makela P. Serotypes, Hemolysin Production, and Receptor Recognition of Escherichia Coli Strains Associated with Neonatal Sepsis and Meningitis Infect. Immun. 1985 48 486 491 10.1128/iai.48.2.486-491.1985 2580792 39. Bachmann B.J. Derivations and genotypes of some mutant derivatives of Escherichia coli K12 Escherichia coli and Salmonella: Cellular and Molecular Biology 2nd ed. Neidhardt FCCurtiss RIngraham JLLin ECCLow KBMagasanik BReznikoff WSRiley MSchaechter MUmbarger H.E. ASM Press Washington, DC, USA 1996 2460 2488 40. Bielaszewska M. Ruter C. Kunsmann L. Greune L. Bauwens A. Zhang W. Kuczius T. Kim K.S. Mellmann A. Schmidt M.A. Enterohemorrhagic Escherichia Coli Hemolysin Employs Outer Membrane Vesicles to Target Mitochondria and Cause Endothelial and Epithelial Apoptosis PLoS Pathog. 2013 9 e1003797 10.1371/journal.ppat.1003797 24348251 41. Kemper B. Barroso A. Eder K. Marzi A. Ritz S. Ntovas A. Schnekenburger J. Ketelhut S. Enhanced Quantitative Phase Imaging in Mach-Zehnder Interferometer-Based Digital Holographic Microscopy by Modulation of the Object Illumination with an Electrically Focus Tunable Lens Proc. SPIE 2021 11786 117860I 10.1117/12.2592534 42. Kemper B. Carl D. Schnekenburger J. Bredebusch I. Schafer M. Domschke W. von Bally G. Investigation of Living Pancreas Tumor Cells by Digital Holographic Microscopy J. Biomed. Opt. 2006 11 034005 10.1117/1.2204609 16822055 43. Min J. Yao B. Ketelhut S. Engwer C. Greve B. Kemper B. Simple and Fast Spectral Domain Algorithm for Quantitative Phase Imaging of Living Cells with Digital Holographic Microscopy Opt. Lett. 2017 42 230 10.1364/OL.42.000227 28081079 44. Kastl L. Isbach M. Dirksen D. Schnekenburger J. Kemper B. Quantitative Phase Imaging for Cell Culture Quality Control Cytom. Part A 2017 91 470 481 10.1002/cyto.a.23082 45. Schindelin J. Arganda-Carreras I. Frise E. Kaynig V. Longair M. Pietzsch T. Preibisch S. Rueden C. Saalfeld S. Schmid B. Fiji: An Open-Source Platform for Biological-Image Analysis Nat. Methods 2012 9 676 682 10.1038/nmeth.2019 22743772 46. Gilmore W.J. Johnston E.L. Zavan L. Bitto N.J. Kaparakis-Liaskos M. Immunomodulatory Roles and Novel Applications of Bacterial Membrane Vesicles Mol. Immunol. 2021 134 72 85 10.1016/j.molimm.2021.02.027 33725501 47. Chen L. Yang J. Yu J. Yao Z. Sun L. Shen Y. Jin Q. VFDB: A Reference Database for Bacterial Virulence Factors Nucleic Acids Res. 2005 33 D325 D328 10.1093/nar/gki008 15608208 48. van der Poll T. Opal S.M. Host-pathogen interactions in sepsis The Lancet Infectious Diseases Elsevier Amsterdam, The Netherlands 2008 32 43 10.1016/S1473-3099(07)70265-7 49. Strobel M. Pfortner H. Tuchscherr L. Volker U. Schmidt F. Kramko N. Schnittler H.J. Fraunholz M.J. Loffler B. Peters G. Post-Invasion Events after Infection with Staphylococcus Aureus Are Strongly Dependent on Both the Host Cell Type and the Infecting S. Aureus Strain Clin. Microbiol. Infect. 2016 22 799 809 10.1016/j.cmi.2016.06.020 27393124 50. Zhang X. Hu X. Rao X. Apoptosis Induced by Staphylococcus Aureus Toxins Microbiol. Res. 2017 205 19 24 10.1016/j.micres.2017.08.006 28942840 51. Nygaard T.K. Pallister K.B. DuMont A.L. DeWald M. Watkins R.L. Pallister E.Q. Malone C. Griffith S. Horswill A.R. Torres V.J. Alpha-Toxin Induces Programmed Cell Death of Human T Cells, B Cells, and Monocytes during USA 300 Infection PLoS ONE 2012 7 e36532 10.1371/journal.pone.0036532 22574180 52. King J.M. Kulhankova K. Stach C.S. Vu B.G. Salgado-Pabon W. Phenotypes and Virulence among Staphylococcus Aureus USA100, USA200, USA300, USA400, and USA600 Clonal Lineages mSphere 2016 1 e00071-16 10.1128/mSphere.00071-16 27303750 53. Jorgensen I. Rayamajhi M. Miao E.A. Programmed Cell Death as a Defence against Infection Nat. Rev. Immunol. 2017 17 151 164 10.1038/nri.2016.147 28138137 54. Ziegler U. Groscurth P. Morphological Features of Cell Death Physiology 2004 19 124 128 10.1152/nips.01519.2004 15143207 55. Verduijn J. Van der Meeren L. Krysko D.V. Skirtach A.G. Deep Learning with Digital Holographic Microscopy Discriminates Apoptosis and Necroptosis Cell Death Discov. 2021 7 229 10.1038/s41420-021-00616-8 34475384 56. Piasecka B. Duffy D. Urrutia A. Quach H. Patin E. Posseme C. Bergstedt J. Charbit B. Rouilly V. MacPherson C.R. Distinctive Roles of Age, Sex, and Genetics in Shaping Transcriptional Variation of Human Immune Responses to Microbial Challenges Proc. Natl. Acad. Sci. USA 2018 115 E488 E497 10.1073/pnas.1714765115 29282317 57. Macia L. Nanan R. Hosseini-Beheshti E. Grau G.E. Host-and Microbiota-Derived Extracellular Vesicles, Immune Function, and Disease Development Int. J. Mol. Sci. 2020 21 107 10.3390/ijms21010107 58. Thay B. Wai S.N. Oscarsson J. Staphylococcus Aureus a-Toxin-Dependent Induction of Host Cell Death by Membrane-Derived Vesicles PLoS ONE 2013 8 e54661 10.1371/journal.pone.0054661 23382935 59. Jin J.S. Kwon S.-O. Moon D.C. Gurung M. Lee J.H. Kim S.I. Lee J.C. Acinetobacter Baumannii Secretes Cytotoxic Outer Membrane Protein A via Outer Membrane Vesicles PLoS ONE 2011 6 e17027 10.1371/journal.pone.0017027 21386968 60. Jones E.J. Booth C. Fonseca S. Parker A. Cross K. Miquel-Clopes A. Hautefort I. Mayer U. Wileman T. Stentz R. The Uptake, Trafficking, and Biodistribution of Bacteroides Thetaiotaomicron Generated Outer Membrane Vesicles Front. Microbiol. 2020 11 57 10.3389/fmicb.2020.00057 32117106 61. Nagakubo T. Nomura N. Toyofuku M. Cracking Open Bacterial Membrane Vesicles Front. Microbiol. 2020 10 3026 10.3389/fmicb.2019.03026 32038523 62. Jeon H. Oh M.H. Jun S.H. Kim S.I. Choi C.W. Kwon H.I. Na S.H. Kim Y.J. Nicholas A. Selasi G.N. Variation among Staphylococcus Aureus Membrane Vesicle Proteomes Affects Cytotoxicity of Host Cells Microb. Pathog. 2016 93 185 193 10.1016/j.micpath.2016.02.014 26924795 63. Mehanny M. Koch M. Lehr C.-M. Fuhrmann G. Streptococcal Extracellular Membrane Vesicles Are Rapidly Internalized by Immune Cells and Alter Their Cytokine Release Front. Immunol. 2020 11 80 10.3389/fimmu.2020.00080 32117243 64. Lee J.C. Lee E.J. Lee J.H. Jun S.H. Choi C.W. Kim S.I. Kang S.S. Hyun S. Klebsiella Pneumoniae Secretes Outer Membrane Vesicles That Induce the Innate Immune Response FEMS Microbiol. Lett. 2012 331 17 24 10.1111/j.1574-6968.2012.02549.x 22428779 65. van der Poll T. Shankar-Hari M. Wiersinga W.J. The Immunology of Sepsis Immunity 2021 54 2450 2464 10.1016/j.immuni.2021.10.012 34758337 66. Caramalho I. Lopes-Carvalho T. Ostler D. Zelenay S. Haury M. Demengeot J. Regulatory T Cells Selectively Express Toll-like Receptors and Are Activated by Lipopolysaccharide J. Exp. Med. 2003 197 403 411 10.1084/jem.20021633 12591899 67. Liu G. Zhang L. Zhao Y. Modulation of Immune Responses through Direct Activation of Toll-like Receptors to T Cells Clin. Exp. Immunol. 2010 160 168 175 10.1111/j.1365-2249.2010.04091.x 20128825 68. Piva E. Zuin J. Pelloso M. Tosato F. Fogar P. Plebani M. Monocyte Distribution Width (MDW) Parameter as a Sepsis Indicator in Intensive Care Units Clin. Chem. Lab. Med. 2021 59 1307 1314 10.1515/cclm-2021-0192 33675202 69. Paoli C.J. Reynolds M.A. Sinha M. Gitlin M. Crouser E. Epidemiology and Costs of Sepsis in the United States-an Analysis Based on Timing of Diagnosis and Severity Level Crit. Care Med. 2018 46 1889 1897 10.1097/CCM.0000000000003342 30048332 70. Liu V.X. Fielding-Singh V. Greene J.D. Baker J.M. Iwashyna T.J. Bhattacharya J. Escobar G.J. The Timing of Early Antibiotics and Hospital Mortality in Sepsis Am. J. Respir. Crit. Care Med. 2017 196 856 863 10.1164/rccm.201609-1848OC 28345952 |
PMC10000560 | Mutational signature analysis promises to reveal the processes that shape cancer genomes for applications in diagnosis and therapy. However, most current methods are geared toward rich mutation data that has been extracted from whole-genome or whole-exome sequencing. Methods that process sparse mutation data typically found in practice are only in the earliest stages of development. In particular, we previously developed the Mix model that clusters samples to handle data sparsity. However, the Mix model had two hyper-parameters, including the number of signatures and the number of clusters, that were very costly to learn. Therefore, we devised a new method that was several orders-of-magnitude more efficient for handling sparse data, was based on mutation co-occurrences, and imitated word co-occurrence analyses of Twitter texts. We showed that the model produced significantly improved hyper-parameter estimates that led to higher likelihoods of discovering overlooked data and had better correspondence with known signatures. mutational signature panel sequencing data biterm topic model United States-Israel Binational Science FoundationThis research was supported by a grant from the United States-Israel Binational Science Foundation (BSF) in Jerusalem, Israel. IS was supported, in part, by a fellowship from the Edmond J. Safra Center for Bioinformatics at Tel-Aviv University. pmc1. Introduction Statistical models for discovering and characterizing mutational signatures are crucial for revealing biomarkers for practical applications. Mutational signatures reveal the mutational processes that transform a "normal" genome into a cancerous genome. The activity of these processes have provided insights into the development of tumorigenesis, and they also have led to new and expanded potential applications for personalized data . Consequently, as more and more cancer data become available, significant efforts have been made to introduce statistical models that can accurately and effectively capture these signatures. Most models of mutational signatures of cancer represent each N patient with cancer as having mutations that were generated from a linear combination of K mutational signatures. Therefore, each signature is represented as a probability distribution over a set of mutational categories, which are typically the 96 categories given by the 6 single base substitution types and the 5' and 3' flanking bases . Each patient's mutations are represented as exposures to the mutational signatures, in addition to some noise. Alexandrov et al. were the first to use non-negative matrix factorization (NMF) to perform a census of mutation signatures across thousands of tumors. Subsequent methods have used different forms of NMF , or have focused on inferring the exposures (also known as refitting) based on the signatures and mutation counts . More recent approaches have borrowed from the world of topic modeling in order to provide a probabilistic model of the data so as to maximize the model's success . The Catalogue of Somatic Mutations in Cancer (COSMIC) now includes a census of dozens of validated mutational signatures , accessed on 1 January 2022), and there have been many efforts to investigate using these signatures as biomarkers for diagnosis and known cancer therapies (e.g., ). Developing methods for analyzing mutational signatures in targeted sequencing datasets have presented new opportunities in the research. To date, most efforts to model mutational signatures have focused on data-rich scenarios, such as whole-exome or whole-genome sequences, where there are from dozens to even thousands of mutations per patient. The most popular targeted sequencing panels have only included several hundred genes and, in general, have had fewer than 10 mutations per patient . The standard topic modeling and non-negative matrix factorization frameworks are not capable of generalizing according to such cases , even though targeted sequencing has been more common in clinical practice. Methods that could accurately infer exposures from targeted sequencing data were thus critical for demonstrating the potential of mutational signatures-based precision medicine in real applications . At the same time, the largest targeted sequencing datasets have included data from many more samples (e.g., see ). Therefore, along with scaling and sparsity challenges, there is also an opportunity for discovering novel and rare signatures. To partially address this challenge, SigMA relied on whole-genome training data to interpret sparse samples and predict their homologous recombination deficiency status. However, SigMA still suffered from the fact that not all cancer types have available whole-genome sequencing data. The Mix model simultaneously clustered the samples and learned the mutational landscape of each cluster, thereby overcoming the sparsity problem. However, it still suffered from high computational costs when learning its hyper-parameters. Therefore, we developed a new topic model for sparse data that borrowed from similar works in the natural language processing (NLP) domain. Specifically, the advent of Twitter has produced an explosion of much shorter (sparser) documents that researchers have to analyzed, where "documents" have a mean length of <35 characters . One of the main insights for handling sparse documents has been to model word co-occurrence directly , under the assumption that words that co-occur frequently were likely from the same topic. While computationally much more intensive than the standard topic model, co-occurrence has shown greater sensitivity on sparse datasets. Following the biterm topic model , we proposed modeling mutation co-occurrence in a similar way. In detail, the generation of each mutation pair was modeled as a two-step process. First, a signature was chosen from a global, cohort-level exposure vector th, and then a pair of mutations was drawn from that signature. The rationale was that, in the case of mutational signatures, the "vocabulary" (mutational categories) was much smaller than that of Twitter. In a targeted sequencing setting, only approximately 0.1% of a patient's mutations can be observed. Therefore, modeling the co-occurrence of mutations could provide additional signals as the number of data points (i.e., mutation pairs) would be quadratic for the number of mutations. Furthermore, because the number of mutational categories was low, it would also be computationally feasible. In the next section, we formally described the model and provided an expectation-maximization (EM) framework for learning the model parameters and estimating the number of signatures in the data. Then, we applied it to various simulated and real targeted sequencing datasets and showed that the model was significantly more efficient and outperformed other hyper-parameter estimation methods. This method was used as a pre-processing step for the Mix method, which improved the training time by an order of magnitude and led to higher likelihoods of discovering overlooked data and improving the correspondence with known signatures. 2. Materials and Methods 2.1. Preliminaries We followed previously published research and assumed that the somatic mutations in cancer fell into M=96 categories (denoting the mutation identity and its flanking bases). These mutations were assumed to be the result of the activity of K (a hyper-parameter) mutational processes, each of which was associated with a signature Si=(ei(1)ei(M)) of probabilities to represent each of the mutation categories.For a given genome n, we denoted its mutation categories as On=(o1noTnn) and assumed that this sequence was represented by the (hidden) signature sequence Zn=(z1nzTnn). We denoted the exposures of the signatures across all patients as p=(p1,,pK). Note that, as compared to most previous works, this was a single "global" exposure vector, rather than a per-patient vector. 2.2. Btm: A Biterm Topic Model To enrich the input data, we adapted a method previously used to analyze short texts in . Instead of viewing mutations as individuals, we examined their co-occurrence patterns with other mutations. Let a biterm be a pair of mutations that co-occur in the same patient. The assumption in Btm was that each biterm was the product of a single mutational process. Formally, patient n was determined by a sequence of biterms (b1nbTnn), where btn=bt1n,bt2n and the corresponding mutations are represented by the hidden signature sequence Zn=(z1nzTnn), as described in Figure 1. Where BnN>=0MxM is the biterm matrix for patient n and Bijn={t1t2|bt1=i,bt2=j} is the number of times words i and j co-occur in the patient. Given the count vector Vn of a patient, we constructed the biterm matrix as Bn=VnTVn-diag(Vn). Given a high number of patients, we constructed the biterm matrix B as the summation of all the biterm matrices together:B=n=1NBn=n=1NVnTVn-diag(Vn) Note that building B, at worst, cost O(NM2), but it could also be calculated as O(|B|) if that was more efficient.We could also perform any combinations as required by the situation, i.e., for fewer patients with more than M mutations, we could use the matrix multiplication option, and for the rest, we computed biterms, one by one. We searched for p=(p1pK) and signatures e, as they could maximize the model's success: PrB|p,e=i=1Mj=1MPrb=(i,j)|p,eBij=i=1Mj=1Mk=1NPrb=(i,j),z=k|p,eBij=i=1Mj=1Mk=1Npkek(i)ek(j)Bij We optimized the model using the following EM algorithm: E-step: Compute for every i,j,k: pk|ij=Prz=k|b=(i,j),p,e=pkek(i)ek(j)k'=1Kpk'ek'(i)ek'(j) Ek(i)=j=1MBijpk|ji+Bjipk|ij Ak=i=1MEk(i) M-step: Compute for every i,k: pk=Akk'=1KAk' ek(i)=Ek(i)i'=1MEk(i') Each EM iteration could be completed in O(KM2) time for K signatures and M mutation categories. To avoid bad local minima, Btm was trained for 100 iterations from 10 random seeds, and then the best one was chosen and further trained for 500 additional iterations. 2.3. Mix: A Mixture of MMMs For completeness, we briefly present the Mix method, which was previously developed in . In order to handle sparse data, the Mix approach clustered the samples and learned the exposures per cluster, rather than per sample. To this end, we proposed a mixture model, which led to simultaneous optimizations of sample (soft) clustering, exposures, and signatures . Given the hyper-parameter L, which indicated the number of clusters, denoted by cn{1L}, the hidden variables representing the true cluster identity of each sample. Our goal was to learn the cluster a priori probabilities w=(w1wL), cluster exposures p=(p1pL), and shared signatures e, so as to maximize the model's success:PrV|w,p,e=n=1NPrVn|w,p,e=n=1Nl=1LPrcn=l,Vn|w,p,e=n=1Nl=1LPrcn=lPrVn|pl,e=n=1Nl=1Lwlj=1Mi=1Kpiei(j)Vj Similarly to Btm, the Mix model was optimized with an EM algorithm. Each iteration could be completed in O(NLKM). To avoid bad local minima, the Mix method was trained for 100 iterations from 10 random seeds, then the best one was chosen and further trained for 500 additional iterations. 2.4. Btm2K-Learning the Number of Signatures in a Dataset Using Btm We present below a method to learn the hyper-parameter K, which was the number of signatures that underpinned a highly sparse dataset. Given a mutation matrix V, we applied a 2-fold cross-validation, training Btm with a varying number of signatures on one-half and testing the overlooked log-likelihood on the other, and vice versa. We repeated this process T times and chose the number of signatures with the best median overlooked log-likelihoods.Following the previous work , we further applied a rollback mechanism to choose the more concise solution in cases where the differences in log-likelihood were not significant. Because the number of biterms was quadratic in the number of mutations in a given patient, small changes in the number of mutations could lead to larger changes in the number of biterms. To avoid this balancing problem in the cross-validation, we defined "big patients" as patients with more than 5 times the average number of biterms in the data. On all the datasets we tested, there were 1-3% big patients, containing 75-85% of the biterms. This phenomenon affected the cross-validation more than the number of signatures, and thus, we applied the cross-validation to the other patients only and used the big patients in addition to the training fold (i.e., they were used only for training alongside the training fold). The algorithm is summarized below. The method was summarized in the pseudo-code Algorithm 1. Algorithm 1Btm2K(V,Kmin,Kmax). 1: Input: VR>=0NxM,1<=Kmin<Kmax<=min{N,M} 2: Parameters: T=numberofrunsforeachK 3: Vbig=Sampleswithmorethan5timestheaveragebitermsinV 4: V=Therest 5: fort=1,,Tdo 6: V1,V2=splitVrandomlytotwoequalsizedsets 7: for k=Kmin,,Kmax do 8: btm=BTM(k,V1Vbig) 9: S[k,t]=btm.log-likelihood(V2) 10: btm=BTM(k,V2Vbig) 11: S[k,t]=S[k,t]+btm.log-likelihood(V1) 12: K =arg mink(median(S[k,:])) 13: repeat 14: K*=K 15: K =min{K<K*|Wilcoxons-rank-sum(S[K,:],S[K*,:])>0.05} 16: untilK <K* 17: returnK 2.5. Previous Hyper-Parameter Selection Algorithms There were several previous algorithms for selecting the number of signatures in a dataset. For rich data, one of the leading methods, CV2K , was based on testing the ability of NMF to reconstruct overlooked data when varying its number of components (which corresponded to signatures). For sparse data, the only previous method that was used in Mix was based on the Bayesian information criterion (BIC), which combines model likelihood with its number of parameters. In the case of Mix, the BIC was applied to select the number of signatures as well as the number of model clusters, thus requiring the model likelihood evaluation in settings with many parameters. 2.6. Running-Time Estimation For simplicity, we assumed that each model required the same number of iterations R to converge and that BIC was iterated over all options for the number of clusters, from 1 to Lmax. To train a model, we used 10 random seeds and improved them for 100 iterations, and then chose the best one and trained it for 500 more iterations, so R=1500. We also assumed that Btm2K and CV2K were both processed T=30 times. Last, we iterated through all the options for K=1Kmax signatures, denoted by N,M the number of samples and mutation categories (96), respectively. Then, the algorithms' complexities were as follows (Table 1):Btm2K: For a given k, we needed to train Btm 2T times (T repetitions of 2 folds). To train Btm, we needed to create biterms with NM2 time and RkM2 training time. In total the cost for k was 2TNM2+2TRkM2. Note that we created biterms one time for all k in each run, so in total, the run time was ~2TNM2+TRKmax2M2. BIC: For a given k, we considered all possible L=1Lmax. For a given pair, we trained Mix once for a cost of RNkLM. In total, for all Ls, we needed ~RkNMLmax2/2. Overall, ~RKmax2Lmax2NM/4 was needed. CV2K: For a given k, we needed to train NMF T times, and each iterations cost NkM time, for a total of TRNkM time. In total, for all k, we spent TRNKmax2M. Note that for Btm2K and CV2K, the cost did not include learning the number of clusters; thus, if we want to train Mix, we needed to use BIC and find the number of clusters. This added ~RKmaxNMLmax2/2 more time to the process. Figure 3 shows that Btm2K was order of magnitudes faster than the other methods. 2.7. Data We present below both real panel datasets, as well as down-sampled and simulated datasets on which we tested our model. MSK-IMPACT Pan-Cancer. We downloaded mutations for a cohort of patients from Memorial Sloan Kettering Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT), which was targeted sequencing data from (accesed on 1 January 2022). The MSK-IMPACT dataset contained 11,369 pan-cancer patients' sequencing samples across 410 target genes. We restricted our analysis to 18 cancer types with more than 100 samples, which resulted in a dataset of 5931 samples and about 7 mutations per sample. Whole genome/exome (WGS/WXS) data. We combined mutations from different sources and cancer types of whole-genome-sequencing and whole-exome-sequencing (WGS/WXS): ovarian cancer (OV), chronic lymphocytic leukemia (CLL), malignant lymphoma (MALY), and colon adenocarcinoma (COAD). We downloaded the OV samples from the Cancer Genome Atlas . For CLL and MALY, we used ICGC release 27, analyzed the sample with the most mutations per patient, and restricted those to samples annotated as "study = PCAWG" . For evaluation purposes, we down-sampled the data to target regions of MSK-IMPACT . The data characteristics are summarized in Table 2. Simulated data. The simulated data were generated and described in detail in to evaluate SigProfiler (SP) and SignatureAnalyzer (SA). Here, for each of the 12 datasets, we evaluated our method on two sets of realistic synthetic data: SP-realistic, based on SP's reference signatures and attributes, and SA-realistic, based on SA's reference signatures and attributes. For each of the (i)-(x) tests, the synthetic datasets were generated based on observed statistics for each signature of each cancer type. Different datasets could differ by the number of signatures, the number of active signatures per samples (sparsity), the number of mutations per sample (whole exome/genome sequencing), whether they reflected a single cancer type or multiple types, and the similarity between signatures. All these factors affected the difficulty of determining the number of components. For each simulated sample, we sampled an MSK-IMPACT patient and down-sampled the simulated sample, so it had the same number of mutations. We removed datasets with missing mutation categories. 2.8. Implementation Details Btm was implemented in Python 3 using numpy . For NMF, we used the scikit-learn implementation . The code for Mix was available at (accesed on 1 January 2022), and the code for CV2K was sourced from (accesed on 1 January 2022). 3. Results 3.1. Evaluating the Number of Signatures from Simulated Data We applied Btm2K to a range of datasets to test its performance and compare the results to current methods. In our first set of results, we used a down-sampled version of the simulated data from . While each dataset was generated by a known set of signatures, due to the down-sampling, this true number may not be reflected in the remaining mutations, which was potentially a result of having only a subset of the true signatures. To mitigate this difficulty, we matched each mutation to a signature with maximum a priori probability (using the known exposures and known signatures).Next, we counted the occurrences of each signature in the down-sampled sample and summed all samples in the dataset. We reported the number of signatures that appeared in more than 5% of the mutations in the down-sampled data. We omitted datasets where all the methods inferred a single signature. The results are summarized in Table 3 and show the superiority of Btm2K over the other approaches. 3.2. Evaluating the Number of Signatures from MSK-IMPACT Data Next, we applied Btm to analyze 5931 samples from the MSK-IMPACT dataset. In Figure 4, the performances of the three estimation methods on this dataset are shown. BIC, Btm2K, and CV2K estimated 6, 7, and 3 signatures, respectively. BIC took around 100 hours to learn both parameters while Btm2K took 1 hour to learn the number of signatures (BIC required 8 additional hours to learn the number of clusters). Complexity-wise, Btm2K was 10-100-fold faster than BIC. To evaluate the quality of their estimations, we trained Mix, Btm, and NMF models on 3, 6, and 7 signatures, respectively, and then we assessed the quality of signatures and log-likelihood of the resulting model on unseen data. We presented the results in the range of 3-9 signatures. To evaluate the quality of the learned signatures, we compared them to the COSMIC signatures. We matched each learned signature to the most similar COSMIC signatures (cosine similarity). We used 0.7 and 0.8 thresholds to determine if a signature was similar to a COSMIC signature. If two signatures were similar to the same COSMIC signature, we determined that the signature with the lower similarity was a duplicate. The results are summarized in Figure 5 and showed that for both thresholds, the maximum number of high quality signatures that had been learned was 7, supporting the estimate of Btm2K and suggesting the other methods underestimated the true number. A more detailed view of the learned signatures appears in Figure 6. Evidently, Btm learned high-quality signatures at a fraction of the time Mix used, supporting its improvements. To further show the advantage of the Btm-inspired method Btm2K, we used Mix to compute the likelihood of yet unseen down-sampled WGS/WXS data, with the different numbers of signatures. For each number of signatures chosen, we used BIC to learn the best number of clusters. The results appear in Figure 7 (left panel) and show that seven signatures, as suggested by Btm2K outperformed the other choices. Of interest, eight signatures performed worse than seven signatures, supporting the use of the rollback mechanism in Btm2Kto avoid over-fitting. Last, we used the three methods to estimate the number of signatures on the down-sampled data. The methods estimated 2 (BIC) and 3 (Btm2K and CV2K) signatures. We trained Mix with these parameters and estimated the performance on the full WXS/WGS mutation catalogs. As shown in Figure 7 (right panel), although the five signatures performed better than three, the latter outperformed the BIC choice of two signatures. 4. Conclusions We adapted Btm, which was developed for the task of handling short texts, and showed it to be useful on panel mutation data. We then developed Btm2K, a method that used Btm to select the number of components on sparse data, such as panel mutations. Our method performed well on several real and simulated datasets, with considerable computational benefits, as compared to current methods. A particularly interesting use case for this method was as a pre-processing step for Mix serving as a better and faster way to choose hyper-parameters. Future work should harness this approach to learn improved topic models for sparse mutation data. Author Contributions Conceptualization, I.S., M.D.M.L. and R.S.; methodology, I.S. and Y.C.; software, I.S.; writing--original draft preparation, I.S.; writing--review and editing, all authors; supervision, M.D.M.L. and R.S.; funding acquisition, M.D.M.L. and R.S. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The data presented in this study are available on request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 A plate diagram for Btm. Figure 2 A plate diagram for Mix. Figure 3 Log running time estimation for Btm2Kand BIC with a maximum of 10-15 clusters and CV2K as a function of the number of samples. Here, Kmax=10 was used. For CV2K and Btm2K, Lmax=15 was used. Figure 4 Performance evaluation on MSK-IMPACT data for varying number of signatures. Left: BIC scores of Mix with varying parameters. Middle: Btm2K log of minus log-likelihoods. Right: CV2K log reconstruction errors. For the middle and right panels, dots represent runs with their median denoted by a dashed line. The minimum median is denoted in the figure legend and the final chosen K is marked in gray. Figure 5 Summed cosine similarities of de novo signatures and COSMIC signatures. In the summation, only unique signatures with similarity above 0.7 (left) or 0.8 (right) were considered. Figure 6 De novo signature discovery of MSK-IMPACT panel data. Shown are the sorted cosine similarities between learned signatures and most similar COSMIC signature (denoted next to the plot) for Mix, Btm, and NMF, across a range of number of signatures (6, 7 corresponding to (left) and (right), respectively). Repeating signatures of the same model are not shown. Figure 7 Log-likelihood of Mix on unseen data as a function of the number of signatures. Left: Mix was trained on MSK-IMPACT data and tested on the down-sampled WGS/WXS data. Right: Mix was trained on down-sampled WGS/WXS data and tested on the original data. cancers-15-01601-t001_Table 1 Table 1 Summary of time complexity for BIC, Btm2K, and CV2K. Here, R,N, and M denotes number of iterations to train a model (1500), samples, and categories (96). T denotes the number of repetitions of Btm2K and CV2K (30), and Kmax Lmax denotes the maximum number of signatures and clusters used when the methods iterated. Method ~Learning Number of Signatures Complexity ~Learning Number of Clusters Complexity (BIC) BIC RKmax2Lmax2NM/4 Btm2K 2TNM2+TRKmax2M2 RKmaxNMLmax2/2 CV2K TRNKmax2M RKmaxNMLmax2/2 cancers-15-01601-t002_Table 2 Table 2 Summary of WGS/WXS down-sampled datasets. Cancer #Samples #Mutations #Panel Mutations OV 411 46,299 1812 Maly 100 1,220,526 1770 CLL 100 270,870 278 COAD 44 52,827 1789 Combined 653 1,590,520 5604 cancers-15-01601-t003_Table 3 Table 3 Estimation of number of signatures in simulated data. For Btm2K and CV2K, the numbers of the best run and the numbers after rollback are shown. In the last column, the number of signatures were present in more than 5% of the mutations as an estimate for the true solution. In bold are the methods that performed best with regard to this estimate. Data Set BIC Btm2K CV2K # Signatures with >=5% Down-sampled Mutations ii-sa 3 4->4 4->2 8 ii-sp 3 10->7 4->2 6 v-sa 2 3->3 3->2 6 v-sp 2 3->2 6->2 5 vii.a(pri.)-sp 1 2->2 3->1 2 vii.b(sec.)-sa 1 1->1 5->2 3 viii-sp 1 2->2 5->1 7 ix-sa 2 4->4 4->2 8 ix-sp 4 6->6 4->3 6 x-sa 1 3->3 5->1 8 x-sp 1 6->6 5->4 6 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000561 | We examined differences in HER2 expression between primary tumors and distant metastases, particularly within the HER2-negative primary breast cancer cohort (HER2-low and HER2-zero). The retrospective study included 191 consecutive paired samples of primary breast cancer and distant metastases diagnosed between 1995 and 2019. HER2-negative samples were divided into HER2-zero (immunohistochemistry [IHC] score 0) and HER2-low (IHC score 1+ or 2+/in situ hybridization [ISH]-negative). The main objective was to analyze the discordance rate between matched primary and metastatic samples, focusing on the site of distant metastasis, molecular subtype, and de novo metastatic breast cancer. The relationship was determined by cross-tabulation and calculation of Cohen's Kappa coefficient. The final study cohort included 148 paired samples. The largest proportion in the HER2-negative cohort was HER2-low [primary tumor 61.4% (n = 78), metastatic samples 73.5% (n = 86)]. The discordance rate between the HER2 status of primary tumors and corresponding distant metastases was 49.6% (n = 63) (Kappa -0.003, 95%CI -0.15-0.15). Development of a HER2-low phenotype occurred most frequently (n = 52, 40.9%), mostly with a switch from HER2-zero to HER2-low (n = 34, 26.8%). Relevant HER2 discordance rates were observed between different metastatic sites and molecular subtypes. Primary metastatic breast cancer had a significantly lower HER2 discordance rate than secondary metastatic breast cancer [30.2% (Kappa 0.48, 95%CI 0.27-0.69) versus 50.5% (Kappa 0.14, 95% CI -0.03-0.32)]. This highlights the importance of evaluating potentially therapy-relevant discordance rates between a primary tumor and corresponding distant metastases. HER2-low HER2-zero HER2 overexpression HER2 dynamics de-novo metastasis antibody-drug conjugate This research received no external funding. pmc1. Introduction Overexpression or amplification of human epidermal growth factor receptor 2 (HER2) characterizes a molecular subtype of breast cancer that progresses rapidly and has a poor prognosis . However, with the advent of targeted therapies against HER2 such as the monoclonal antibody trastuzumab, the original prognostic disadvantage of HER2 positivity has been transformed into a clinically relevant predictive advantage . In advanced HER2-positive breast carcinoma, survival was further prolonged by pertuzumab . In the event of progression, tyrosine kinase inhibitors such as lapatinib or tucatinib showed efficacy in HER2-positive breast carcinoma . In addition, antibody-drug conjugates (ADC) provided a further improvement not only in progression-free survival (PFS) but also in overall survival (OS) in advanced HER2-positive breast carcinoma . Therefore, it was obvious to use anti-HER2 therapies also in early breast carcinoma. Indeed, the use of trastuzumab resulted in a statistically significant prolongation of overall survival . Surprisingly, some of the patients who participated in the original trastuzumab trials and were ultimately HER2-negative at central HER2 reassessment benefited from trastuzumab . Based on these findings, a large phase III trial was conducted in 3270 women, but it clearly showed that trastuzumab was not beneficial for patients without IHC 3+ or ISH-enhanced breast cancer . In HER2-positive early breast cancer, the addition of pertuzumab led to a relevant increase in pathologic complete response rates (pCR) and an improvement in disease-free survival (DFS) . The prolongation of DFS was further increased using the tyrosine kinase inhibitor neratinib after the completion of trastuzumab-based therapy in HER2-positive patients . Furthermore, the ADC T-DM1 improved DFS in early HER2-positive breast cancer with residual disease after neoadjuvant trastuzumab-based treatment . Overall, these HER2-targeted therapies represent tremendous progress for the 15-20% HER2-positive patients. Meanwhile, several retrospective studies have taken a closer look at the large group of HER2-negative breast cancer. Breast cancer showing HER2 protein expression without HER2 gene amplification could be divided into two separate groups (HER2-low [IHC 1+ or 2+ and ISH-negative] and HER2-zero [IHC 0]) with different prognosis or pCR after neoadjuvant chemotherapy . However, these results could not be confirmed in several other studies . Nevertheless, interest in HER2-low tumors has increased greatly due to the results of the DESTINY-Breast04 trial, which demonstrated the superiority of trastuzumab-deruxtecan (T-DXd) over physician's choice chemotherapy in patients with advanced HER2-low breast cancer . The prolongation of PFS (10.1 months vs. 5.4 months; hazard ratio [HR] 0.51; p < 0.001) and OS (23.9 months vs. 17.5 months; HR 0.64; p = 0.003) was both statistically significant and clinically relevant. These compelling results led to a rapid update of the American Society of Clinical Oncology (ASCO) guideline and a positive opinion of the Committee for Medicinal Products for Human Use (CHMP) of the European Medicines Agency (EMA) recommending the use of T-DXd in patients with HER2-low metastatic BC . Until now, the only question regarding HER2 status was whether the tumor was HER2-positive or HER2-negative. However, the impressive data from T-DXd in HER2-low breast cancer highlights the importance of dividing the large group of HER2-negative patients. In principle, the well-known ASCO/College of American Pathologists (CAP) clinical practice guideline allows such a distinction . However, potential difficulties such as tumor heterogeneity (clustered or mosaic type) or unusual staining patterns (moderate to intense but incomplete staining or carcinomas with limited strong HER2 overexpression) must be considered . To address difficulties in distinguishing between HER2-low and HER2-zero, pathologists have already pointed out possible solutions for the assessment of immunohistochemical staining such as (1) application of the "magnification rule", (2) staining pattern-circularity of membrane staining, and (3) percentage of tumor cells with HER2 expression . In addition to these briefly outlined challenges for pathologists in distinguishing between HER2-low and HER2-zero, another fundamental problem is discordance between the primary tumor and corresponding metastases, since whenever possible, a recent metastatic biopsy is encouraged to guide therapy in advanced breast cancer. The problem of discordance between primary tumors and distant metastases arises when treating patients with metastatic disease with targeted therapies. The discordance of traditional HER2 dichotomy (positive or negative) between primary breast cancer and distant metastases is well established. Among others, Grassini et al. reviewed the phenomenon of HER2 conversion between primary breast tumors and relapsed/distant metastatic . While early studies described a wide variability in HER2 discordance rates (0-44%), several meta-analyses showed discordance rates ranging from 7.8% to 13.7% . Most commonly, conversion from HER2-positive to HER2-negative was observed, which is clinically important in both advanced and early breast cancer. In neoadjuvant studies, a loss of HER2 expression from a therapy-naive primary tumor and the post-neoadjuvant residual tumor was described with a prognostic disadvantage . However, less is known about the discordance of HER2-low between primary tumors and distant metastases. Thus, Tarantino and coworkers demonstrated a relevant discordance in HER2 expression between PTs and their associated metastases: 44% of HER2-zero PTs had an elevated HER2 score on biopsy, and 22% of HER2-low PTs became HER2-zero tumors . Miglietta et al. reported an overall rate of HER2 discordance of 38.0%, with most transitioning from HER2-zero to HER2-low (15%) and from HER2-low to HER2-zero (14%) . This discordance rate is clinically relevant to the use of ADCs and prompted us to investigate the discordance rate in 148 paired samples (primary breast tumor and distant metastasis), focusing on (i) molecular subtype, (ii) distant metastasis site, and (iii) differences between primary metastatic breast cancer (PMBC) and secondary metastatic breast cancer (SMBC). 2. Materials and Methods 2.1. Study Cohort The certified breast cancer center of the University Medical Center Mainz has been prospectively documenting clinic-pathological characteristics as well as therapies of all treated breast cancer patients. This database was searched for patients with metastatic breast cancer between 15.06.1995 and 10.10.2019. We studied 191 consecutive paired samples of primary breast tumor and distant metastasis. Only solid distant metastases were considered. Paired samples without complete HER2 status (n = 31), with HER2 status equivocal (n = 4), or bilateral BC or secondary malignancy with different HER2 status (n = 8) were not eligible for this study . PMBC was defined as the presence of metastasis at the time of diagnosis of the PT . The median age at the time of initial breast cancer diagnosis was 53 years (range, 31-86 years). Table 1 provides an overview of the established clinico-pathologic prognostic factors in the final study cohort (n = 148). 2.2. Immunohistochemistry (IHC) and In Situ Hybridization (ISH) Immunohistochemical analyses and in situ hybridization were performed on 3 mm thick sections of paraffin-embedded formalin-fixed tissues according to standard procedures. HER2 was scored from 0 to 3+ . HER2 2+ cases (n = 22) were further classified as amplified or non-amplified by either fluorescence in situ hybridization (FISH) (Her2 FISH pharmDX kit, Dako) or chromogenic in situ hybridization (CISH) (Ventana Her2 dual ISH assay, Roche). HER2 2+ tumors with amplification of HER2 and 3+ tumors were classified as HER2-positive. The HER2-negative cohort was defined as 0, 1+, and 2+ without amplification of HER2. HER2-low tumors included all 1+ and 2+ tumors without amplification of HER2. Tumors with a HER2 score 0 were classified as HER2-zero . Hormone receptor status was positive if tumor cells showed nuclear expression of either the estrogen receptor (ER) and/or the progesterone receptor (PR), the cut-off being defined as 1% of tumor cells . The study was approved by the Ethics Committee of the Rhineland-Palatinate Medical Association, Germany (2021-15657). Written informed consent was obtained from all patients, and all clinical investigations were performed according to ethical and legal standards and in compliance with the Declaration of Helsinki. This study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline . 2.3. Statistical Analysis The main objective was to evaluate HER2 expression differences between primary tumor and distant metastasis, particularly in the HER2-negative (HER2-low and HER2-zero) primary breast cancer collective. Secondary objectives of our analysis were (i) that discordance rates differ according to the molecular subtype of the primary tumor, (ii) that discordance rates differ according to the site of distant metastasis, and (iii) the discordance rate in PMBC is lower than in SMBC. The relationship between the different categorical variables was determined by cross-tabulation. Comparisons between different HER2 status (primary tumor and metastasis) were calculated by Cohen's Kappa coefficient. Relationships between HER2 status for primary tumor and clinico-pathological parameters were assessed by cross-tabulation and using Pearson's chi-squared test (kh2 test). Statistical analyses were performed using the SPSS statistical software program, version 27.0 (SPSS Inc., Chicago, IL, USA), and the statistical language R, version 4.1.2 . Patients' characteristics were given in absolute and relative numbers. All p-values are two-sided. Because no correction was made for multiple testing due to the exploratory nature of the study, these are descriptive measures that should be interpreted with caution. 3. Results 3.1. Patient Population The final study cohort included 148 paired samples. Primary tumors were divided into 127 (85.8%) HER2-negative samples [49 HER2-zero (38.6%) and 78 HER2-low (61.4%)] and 21 (14.2%) HER2-positive samples . One-hundred and seven (72.3%) primary tumors showed a luminal-like phenotype, 21 (14.2%) were HER2-positive, and 20 (13.5%) had a triple-negative phenotype. PMBC occurred in 35.8% (n = 53), and more frequently in the HER2-low (58.5%, n = 31) than in the HER2-zero cohort (15.1%, n = 8). The median time to first metastasis was 25 months (range 0-150). The median time between diagnosis of metastatic disease and biopsy was one month (range 0-131). Metastases were located in the liver (n = 50, 33.8%), bone (n = 38, 25.7%), skin/soft tissue (n = 18, 12.2%), central nervous system (CNS) (n = 15, 10.1%), other sites (n = 13, 8.8%), lung/pleura (n = 9, 6.1%), and lymph nodes (n = 5, 3.4%). Seventy-nine (53.4%) patients received adjuvant endocrine therapy and 72 (48.6%) neo-/adjuvant chemotherapy. A small proportion of patients were treated with adjuvant anti-HER2 therapy (n = 9, 6.1%). At the time of metastatic biopsy, 35.1% (n = 52) of patients were receiving endocrine therapy, 30.4% (n = 45) chemotherapy, and/or 8.8% (n = 13) anti-HER2 therapy for metastatic disease. Compared with HER2-zero and HER2-positive phenotype, HER2-low was significantly less frequently diagnosed in larger tumors (>T2) (HER2-low 66.0% vs. HER2-zero 77.1% and HER2-positive 85.6%, p = 0.032) and poorly differentiated tumors (G3) (HER2-low 33.3% vs. HER2-zero 49.0% and HER2-positive 60.0%, p = 0.049). HER2-low status was more common in tumors with higher Ki-67 (>20%) compared with HER2-zero (82.7% vs. 64.3%). However, higher Ki-67 levels were most frequently found in the HER2-positive cohort (100%) (p = 0.022). Low HER2 was significantly more common in luminal-like tumors than in triple-negative tumors (88.5% vs. 11.5%), while conversely, HER2-zero was more common in triple-negative tumors (22.4% vs. 11.5%) (p < 0.001). Additional tumor and patient characteristics are listed in Table 1. 3.2. Change of HER2 Status between Primary Breast Cancer and Metastasis In the HER2-negative cohort, the HER2-low phenotype represented the largest group [primary tumor 61.4% (n = 78), metastatic samples 73.5% (n = 86)]. Discordance in HER2 status between the primary tumor and the matched metastatic biopsy was 49.6% (n = 63) (Kappa -0.003, 95%CI -0.15-0.15). Development of the HER2-low phenotype (HER2-zero to HER2-low or HER2-low to HER2-zero) was most common (n = 52, 40.9%), especially with enrichment to HER2-low (n = 34, 26.8%) . In the entire cohort (n = 148), HER2 discordance was 43.2% (n = 64) (Kappa 0.270, 95%CI 0.14-0.41). Most frequently, an evolution from HER2-zero to HER2-low phenotype was observed (n = 34, 23.0%). Within the HER2-zero cohort, this represented a switch of 69.4% from HER2-zero to HER2-low. A change from HER2-low to HER2-zero occurred in 12.2% (n = 18). Considered for the HER2-low cohort alone, a switch from HER2-low to HER2-zero resulted in 23.1%. The HER2-positive cohort showed the greatest stability, with a discordance of 4.8% (n = 1 of 21) . When additional metastatic biopsies were performed, a discordance rate of 57.9% was observed compared with the previous biopsy (Table 1). Again, the most common finding was a change from HER2-zero to HER2-low (15.8%; within the HER2-zero cohort: 60.0%). 3.3. Change of HER2 Status in Different Metastatic Sites The proportion of HER2-low did not differ significantly between the different metastatic sites (p = 0.349) (Table 1). In the HER2-negative population, a relevant HER2 discordance rate was observed between the different metastatic sites (bone: Kappa 0.022, -0.230-0.273; liver: Kappa 0.048, -0.195-0.291; skin/soft tissue: Kappa 0.082, -0.363-0.527, lymph node: Kappa 0.000; CNS: Kappa -0.250, -0.606-0.106; others: Kappa -0.467, -0.980-0.047). Only pulmonary/pleural metastases showed absolute concordance with the primary breast tumor (Kappa 1.0, 95%CI 1.0-1.0). An increase from HER2-zero to HER2-low at metastatic biopsy was most frequently detected, excluding bone metastases. A switch to HER2 positive has been observed especially in CNS metastases (n = 2, 20.0%) . Similarly, there was a relevant change in HER2 expression in the entire cohort . 3.4. Change of HER2 Status in Different Molecular Subtypes HER2 discordance was observed according to the molecular subtype, in the Luminal A/B cohort (Kappa -0.044, -0.202-0.114) and in triple-negative breast cancer (Kappa 0.107, -0.247-0.461). In both subcohorts, a switch from HER2-zero to HER2-low was most frequently detected (Luminal A/B n = 28, 26.2%; Triple-negative n = 6, 30.0%) . 3.5. Change of HER2 Status in Primary vs. Secondary Metastatic Breast Cancer PMBC was evidenced in 35.8% (n = 53). HER2-low represented the largest proportion in both the primary tumor (58.5%, n = 31) and matched de-novo metastases (54.7%, n = 29). The de-novo cohort showed a higher prevalence of HER2-low expression in PT (58.5%, n = 31) than in the relapsed collective (49.5%, n = 47). The HER2-low phenotype was represented more frequently in secondary metastases than in de-novo metastases (60.0% vs. 54.7%) (Tables S6 and S7). The discordance rate was lower in the PMBC than in the SMBC cohort [30.2% (Kappa 0.48, 95%CI 0.27-0.69) vs. 50.5% (Kappa 0.14, 95% CI -0.03-0.32)]. In the de-novo cohort, the increase almost corresponded to the loss of HER2 expression (11.3% vs. 13.2%), whereas in the SMBC cohort, the change from HER2-zero to HER2-low clearly predominated (29.5% vs. 11.6%) . 4. Discussion In our retrospective analysis, we showed a relevant discordance rate of HER2 status between primary breast cancer and distant metastases, with the conversion from HER2-zero to HER2-low observed most frequently. Recently published studies have addressed the heterogeneity of HER2-negative breast cancer, focusing on the HER2-negative cohort . The need to relativize the traditional dichotomization between HER2-positive and HER2-negative appeared at the latest with the results of the DESTINY-Breast04 trial, which demonstrated the superiority of trastuzumab-deruxtecan (T-DXd) vs. chemotherapy of the physician's choice in patients with advanced HER2-low breast cancer . In addition, other HER2-targeting ADCs like trastuzumab duocarmazine showed promising activity in early studies . In our previous study, we reported a rate of 48.3% HER2-low tumors, which was within the range of approximately half of HER2-negative breast cancer patients (31.0% to 60.6%) reported by Prat and coworkers in a recent review article . Since 80-85% of all breast cancer tumors have a HER2-negative phenotype, which was 85.2% in our study, a better understanding of this cohort has potential therapeutic implications for the majority of breast cancer patients. In this context, the aspect of the evolution of HER2 expression from early to advanced breast cancer is important. In the present study, we demonstrated a discordance rate between HER2 status of primarytumors and associated distant metastases within the HER2-negative cohort of 49.6% (Kappa -0.003, 95%CI -0.15-0.15). The development of HER2-low occurred frequently (40.9%), particularly with a switch from HER2-zero to HER2-low (26.8%). Discordance rates from our study were slightly higher than in other studies (38.0% and 40.9%, respectively . However, all studies showed an increase in HER2 expression from HER2-zero to HER2-low during disease progression. Our results of additional metastatic biopsies compared with initial metastatic biopsies point in the same direction, with a discordance rate of 57.9%. There are several hypotheses for low HER2 stability (e.g., genetic drift and clonal evolution during tumor progression, intratumoral heterogeneity, and the selective effect of administered therapies) leading to the enrichment of HER2 expression . Our study cohort showed heterogeneity in terms of time to metastasis, with a significantly longer time in the HER2-zero cohort than in the HER2-low cohort [median 44 months (0-150) vs. 14 months (0-121)] and in terms of systemic treatments given. Neo-/adjuvant chemotherapy was significantly more common in the HER2-zero than in the HER2-low cohort (71.4% vs. 38.5%, p < 0.001). Both aspects may have an impact on the increase of HER2-low from a primary tumor to distant metastases. In our study, there were no significant differences in HER2-low expression depending on the metastatic site (p = 0.349). Comparable to our results, Tarantino and coworkers also found no significant difference in HER2-low expression at different metastatic sites (p = 0.88), even when they divided biopsy sites into visceral (liver, lung, and pleura) and nonvisceral (skin and soft tissues, lymph nodes, bone, other) (p = 0.56) . Miglietta et al. examined locoregional recurrences in addition to HER2-low prevalence in various distant metastases, also with similar results for HER2-low . However, in Miglietta's cohort, a significant difference in discordance rates was observed between the different metastases (p = 0.001), with the greatest HER2 instability in liver and bone metastases and the greatest concordance in lung and CNS metastases. Lung/pleural metastases also had the strongest concordance in our study. In the HER2-negative cohort, discordance rates ranged from 40.9% (liver) to 80.0% (CNS). The HER2 score changed from HER2-zero to HER2-low most frequently at metastatic biopsy, except for bone metastases, which is in contrast to the results of Lin et al. . Overall, the aspects of similar HER2 expression levels at different metastatic sites and the different discordance rates are of clinical importance when a metastatic site has to be selected for biopsy to decide on targeted therapies that are also effective in HER2-low tumors. Another objective of our study was to analyze whether discordance rates depend on the molecular subtype of the primary BC. Recently published studies that examined discordance rates as a function of the molecular subtype of the primary tumor showed that HER2-low was more common in HR-positive tumors than in triple-negative tumors . Similarly, the HR-positive subtype was associated with a higher discordance rate, most frequently switching from HER2-zero to HER2-low . In our study, HER2-low was also significantly more common in luminal-like tumors than in triple-negative tumors, although the difference was much greater than in the above studies. A particular aspect of our study was the evaluation of PMBC, which we assessed for HER2-low prevalence and discordance rates. Compared to SMBC, PMBC had a significantly lower discordance rate between the PT and matched metastases [30.2% (Kappa 0.48, 95%CI 0.27-0.69) versus 50.5% (Kappa 0.14, 95% CI -0.03-0.32)]. While an increase in HER2 expression between primary tumor and distant metastases was observed across the cohort in our study, this trend was not observed in PMBC, supporting the hypothesis of HER2 enrichment during tumor progression. Overall, not all studies clearly addressed PMBC , excluded de-novo tumors, or applied various de-novo definitions (e.g., <6 months) . Therefore, a comparison between our results and those of other studies is difficult. Although the de novo cohort of our study was small (n = 53), the aspect of low HER2 instability should also receive attention in de novo metastatic BC. In this regard, further studies with larger study cohorts are needed. Our study has several limitations, such as the retrospective and unicenter setting. Another limitation is the lack of a central pathology assessment given the low interobserver reproducibility, especially in HER2-low and HER2-zero . In addition, we did not investigate a possible prognostic impact of HER2 discordance between the primary tumor and the corresponding distant metastases. However, the consecutive inclusion of matched pairs is a strength. In addition, we examined the impact of both the site of distant metastases and molecular subtypes and took a closer look at the PMBC group. 5. Conclusions In summary, we have shown in our study that there is a significant discordance rate between HER2-negative primary breast cancer and the corresponding metastases, with HER2-zero tumors more likely to become HER2-low tumors at advanced stages. In contrast, there is less discordance in PMBC. This evolution of HER2 expression is a clinically relevant rationale for metastatic biopsy and offers patients the opportunity to receive an effective treatment such as trastuzumab-deruxtecan. Acknowledgments This work contains parts of the doctoral thesis of Lisa Krauthauser. Supplementary Materials The following supporting information can be downloaded at: Table S1: Change of HER2 status between primary tumor and metastasis in the HER2-negative cohort (n = 127); Table S2: Change of HER2 status between primary tumor and metastasis in the entire cohort (n = 148); Table S3: Change of HER2 status in different metastatic sites in the HER2-negative cohort (n = 127); Table S4: Change of HER2 status in different metastatic sites (n = 148); Table S5: Change of HER-2 in different molecular subtypes (luminal-like and triple-negative) (n = 127); Table S6: Change of HER2 status between primary tumor and metastasis in the de-novo cohort (n = 53); Table S7: Change of HER2 status between primary tumor and metastasis in the secondary metastastic breast cancer cohort (n = 95); Figure S1: Change of HER2 status between primary tumor and metastasis in the entire cohort (n = 148); Figure S2: Change of HER2 status in different metastatic sites (n = 148); Figure S3: Change of HER-2 in different molecular subtypes (luminal-like and triple-negative) (n = 127). Click here for additional data file. Author Contributions Conceptualization, K.A. (Katrin Almstedt) and M.S.; methodology, K.A. (Katrin Almstedt), F.K., J.R. and M.S.; validation, K.A. (Katrin Almstedt) and M.S.; formal analysis, K.A. (Katrin Almstedt), L.K., F.K., D.-C.W., J.R., W.R. and M.S.; investigation, D.-C.W., W.R. and M.S.; resources, W.B., A.H. and M.S.; data curation, K.A. (Katrin Almstedt), F.K., J.R., K.S. and M.S.; writing--original draft preparation, K.A. (Katrin Almstedt) and M.S.; writing--review and editing, K.A. (Katrin Almstedt), L.K., F.K., D.-C.W., A.-S.H., M.J.B., K.A. (Katharina Anic), S.K., A.L., R.S., W.B., W.W., J.R., J.G.H., W.R., A.H., K.S. and M.S.; visualization, K.A. (Katrin Almstedt) and F.K.; supervision, M.S.; project administration, K.A. (Katrin Almstedt). All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of the Rhineland-Palatinate Medical Association, Germany (2021-15657, ethical approval date 08.04.2021). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The dataset analyzed during the current study is available from the corresponding author upon reasonable request. Conflicts of Interest K.A. received speaker honoraria from Roche Pharma AG, Pfizer Pharma GmbH, Seagen, Med publico GmbH, and Astra Zeneca. D.-C.W. received speaker honoraria from Roche, BMS, MSD, and Janssen. He received funded research from Roche and BMS. A.-S.H. received speaker honoraria from Pfizer Pharma GmbH and honoraria from Medupdate GmbH. M.J.B. received honoraria and expenses from Astra Zeneca, Clovis Oncology, GSK, MSD, Pharma Mar, Roche, and Tesaro Bio Germany GmbH. He is a consultant to Eisai, GSK, MSD, Pharma Mar, Roche Pharma AG, and Tesaro Bio Germany GmbH. He received funded research from Astra Zeneca, Clovis Oncology, MSD, and Novartis. S.K. received speaker honoraria from Roche Pharma AG and Novartis Pharma GmbH Germany, research funding from Novartis Pharma GmbH Germany, and travel reimbursement from Pharma Mar and Novartis Pharma GmbH Germany. A.H. received honoraria from Astra Zeneca, Celgene, MedConcept Gm, Med update GmbH, Medicultus, Pfizer, Promedicis GmbH, Pierre Fabre, Softconsult, Roche Pharma AG, Streamedup!GmbH and Tesaro Bio Germany GmbH. She is a member of the advisory board of Pharma Mar, Promedicis GmbH, Pierre Fabre Pharma GmbH, Roche Pharma AG, and Tesaro Bio Germany GmbH. She received research funding from Celgene. M.S. reports personal fees from Astra Zeneca, BioNTech, Daiichi Sankyo, Eisai, Lilly, MSD, Novartis, Pantarhei Bioscience, Pfizer, Roche, and SeaGen outside the submitted work. Institutional research funding from Astra Zeneca, BioNTech, Eisai, Genentech, German Breast Group, Novartis, Palleos, Pantarhei Bioscience, Pierre Fabre, and SeaGen. In addition, Marcus Schmidt has a patent for EP 2390370 B1 issued and a patent for EP 2951317 B1 issued. All other authors declare that they have no conflict of interest. Figure 1 Patient enrolment. Figure 2 Change of HER2 status between primary tumor and metastasis in the HER2-negative cohort (n = 127). Figure 3 Change of HER2 status in different metastatic sites in the HER2-negative cohort (n = 127). Figure 4 Change of HER2 status between primary tumor and metastasis in the de-novo cohort (n = 53). cancers-15-01413-t001_Table 1 Table 1 Clinico-pathological parameters at diagnosis (n = 148). Total Number of Patients (n = 148) HER2-Negative (n = 127) HER2-Positive (n = 21) p-Value HER2-Zero (n = 49) HER2-Low (n = 78) Age at primary breast surgery 0.087 Median [years] 53 48 54 52 <50 years 66 (44.6%) 28 (57.1%) 29 (37.2%) 9 (42.9%) >50 years 82 (55.4%) 21 (42.9%) 49 (62.8%) 12 (57.1%) Histological subtype 0.221 Invasive carcinoma of no special type (NST) 115 (77.7%) 33 (67.3%) 63 (80.8%) 19 (90.5%) Invasive lobular carcinoma 24 (16.2%) 12 (24.5%) 11 (14.1%) 1 (4.8%) other 9 (6.1%) 4 (8.2%) 4 (5.1%) 1 (4.8%) Tumor size 0.032 pT1 32 (22.2%) 11 (22.9%) 18 (24.0%) 3 (14.3%) pT2 67 (46.5%) 30 (62.5%) 30 (40.0%) 7 (33.3%) pT3/4 45 (31.3%) 7 (14.6%) 27 (36.0%) 11 (52.3%) missing 4 (2.7%) Nodal status 0.710 Negative 44 (29.7%) 13 (26.5%) 25 (33.3%) 6 (28.6%) positive 101 (68.2%) 36 (73.5%) 50 (66.7%) 15 (71.4%) missing 3 (2.0%) Histological grade 0.049 G1 11 (7.4%) 6 (12.2%) 5 (6.7%) 0 G2 72 (48.6%) 19 (38.8%) 45 (60.0%) 8 (40.0%) G3 61 (41.2%) 24 (49.0%) 25 (33.3%) 12 (60.0%) missing 4 (2.7%) Hormone receptor status 0.252 negative 23 (15.5%) 11 (22.4%) 9 (11.5%) 3 (14.3%) positive 125 (84.5%) 38 (77.6%) 69 (88.5%) 18 (85.7%) HER2 status Negative 127 (85.8%) 49 (100.0%) 78 (100.0%) Positive 21 (14.2%) 21 (14.2%) 0 49 (33.1%) 49 (100.0%) 1+ 59 (39.9%) 59 (75.6%) 2+ 22 (14.9%) 19 (24.4%) 3 (14.3%) 2+/ISH negative 19 (12.8%) 2+/ISH positive 3 (2.0%) 3+ 18 (12.2%) 18 (85.7%) Ki-67 0.022 <20% 19 (12.8%) 10 (35.7%) 9 (17.3%) 0 >20% 74 (50.0%) 18 (64.3%) 43 (82.7%) 13 (100.0%) Missing 55 (37.2%) Molecular subtype <0.001 Luminal-like 107 (72.3%) 38 (77.6%) 69 (88.5%) 0 Luminal-A-like 18 (12.2%) 9 9 0 Lumina-B-like 46 (31.1%) 10 36 0 Missing Ki-67 43 (29.1%) HER2 positive 21 (14.2%) 21 (100.0%) Triple-negative 20 (13.5%) 11 (22.4%) 9 (11.5%) 0 Metastatic site 0.349 Liver 50 (33.8%) 16 (32.7%) 28 (35.9%) 6 (28.6%) Bone 38 (25.7%) 13 (26.5%) 23 (29.5%) 2 (9.5%) Skin/Soft tissue 18 (12.2%) 6 (12.2%) 10 (12.8%) 2 (9.5%) Central nervous system 15 (10.1%) 6 (12.2%) 4 (5.1%) 5 (23.8%) others 13 (8.8%) 5 (10.2%) 6 (7.7%) 2 (9.5%) Lung/Pleura 9 (6.1%) 1 (2.0%) 5 (6.4%) 3 (14.3%) Lymph node 5 (3.4%) 2 (4.1%) 2 (2.6%) 1 (4.8%) Additional metastatic biopsy Yes 19 (12.8%) 5 (10.2%) 11 (14.1%) 3 (14.3%) 0.797 HER2 concordance with previous biopsy 8 (42.1%) 1 (20.0%) 6 (54.5%) 1 (33.3%) HER2 discordance with previous biopsy 11 (57.9%) 4 (80.0%) 5 (45.4%) 2 (66.7%) No 129 (87.2%) 44 (89.8%) 67 (85.9%) 18 (85.7%) Treatment for early breast cancer Neo-/Adjuvant chemotherapy 72 (48.6%) 35 (71.4%) 30 (38.5%) 7 (33.3%) <0.001 Neo-/Adjuvant Anti-HER2-therapy 9 (6.1%) 0 0 9 (42.9%) <0.001 Adjuvant endocrine therapy 79 (53.4%) 31 (63.3%) 42 (56.0%) 6 (28.6%) 0.083 Treatment for metastatic breast cancer chemotherapy 45 (30.4%) 8 (16.3%) 26 (34.2%) 11 (52.4%) 0.007 Anti-HER2-therapy 13 (8.8%) 2 (4.1) 0 11 (52.4%) <0.001 Endocrine therapy 52 (35.1%) 18 (36.7%) 29 (38.2%) 5 (23.8%) 0.468 Tumor progression Time to metastasis, Median [month] 25 (0-150) 44 (0-150) 14 (0-121) 0 (0-111) Time to metastasis biopsy, Median [month] 39 (0-165) 51 (0-150) 36 (0-165) 17 (0-37) Time from metastasis diagnosis to metastasis biopsy, Median [month] 1 (0-131) 0 (0-55) 1 (0-131) 0 (0-94) Primary metastatic breast cancer (PMBC) Yes 53 (35.8%) 8 (16.3%) 31 (39.7%) 14 (66.7%) No 95 (64.9%) 41 (83.7%) 47 (60.3%) 7 (33.3%) Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000562 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050828 diagnostics-13-00828 Editorial Editorial on Special Issue "Artificial Intelligence in Pathological Image Analysis" Tsuneki Masayuki Medmain Research, Medmain Inc., 2-4-5-104, Akasaka, Chuo-ku, Fukuoka 810-0042, Japan; [email protected]; Tel.: +81-92-707-1977 21 2 2023 3 2023 13 5 82817 2 2023 20 2 2023 (c) 2023 by the author. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). This research received no external funding. pmcThe artificial intelligence (AI), especially deep learning models, is highly compatible with medical images and natural language processing and is expected to be applied to pathological image analysis and other medical fields. In routine pathological diagnosis, the histopathological and cytopathological examination of specimens is conventionally performed under a light microscope. Whole slide images (WSIs) are the digitized counterparts of the conventional glass slides obtained using specialized scanning devices. In recent years, digital pathology has been steadily introduced into clinical workflows, such as intraoperative consultations and secondary consultations. Pathology diagnosis support systems (computer-aided detection/diagnosis: CAD) using AI are useful for various classification tasks, such as histopathological subtyping, tumor grading, immunohistochemical scoring, and predictions of genetic mutation and protein expression profiles . It is becoming possible to develop AI that can not only perform image classification and detection tasks, but also infer histopathological findings from images by combining pathological images with natural language. In a time of distinct paradigm shifts and novel technological innovations, it is necessary for us to establish a unified comprehension(s) of AI approaches in experimental and clinical pathology. In this Special Issue "Artificial Intelligence in Pathological Image Analysis", we collected a review and thirteen research articles in the areas of AI models in clinical and experimental pathology and computer vision in pathological image analysis. The published studies in this Special Issue provide great insights into the latest knowledge about the application of AI for pathological image analysis. Kim et al. summarized the current trends and challenges to the application of AI in pathology . In this review article, the authors described the development of computational pathology (CPATH), its applicability to AI development, and the challenges it faces, such as algorithm validation and interpretability, computing systems, reimbursement, ethics, and regulations. Further, the authors presented an overview of novel AI-based approaches that could be integrated into the pathology laboratory workflow. As the authors described, explainable AI and ethics and security issues are important topics in CPATH. To develop safe and reliable AI, the pathology community needs more clinical research and laboratory practices. This review paper provides the current research status of AI in pathology and future perspectives for successful applications. Our research article demonstrated a deep learning model for prostate adenocarcinoma classification in core needle biopsy WSIs using transfer learning . In routine clinical practice, diagnosing 12 core biopsy specimens using a microscope is time-consuming, manual process, and it is limited in terms of human resources. The authors trained deep learning models capable of classifying core needle biopsy WSIs into adenocarcinoma and benign (non-neoplastic) lesions and achieved an ROC-AUC of up to 0.978 in the core needle biopsy test sets for adenocarcinoma. Deep learning-based computational algorithms might be useful as routine histopathological diagnostic aids for prostate adenocarcinoma classification in core needle biopsy specimens. Rakovic et al. conducted a survey of prostate cancer UK supporters for the use of digital pathology and AI in the histopathological diagnosis of prostate cancer . A total of 1276 responses to the online survey were analyzed. It was revealed that most of the respondents were in favor of advances in prostate cancer diagnosis by means of digital pathology and AI-assisted diagnostics as adjuncts to current clinical workflows. However, a small minority of them were not in favor of the use of AI in histopathology for reasons which are not easily addressed. Importantly, the patients are more comfortable with the overall responsibility for a histopathology report remaining with the histopathologist and relying on their decision making to use AI and integrate its findings into the final report. Baek et al. demonstrated the AI-assisted image analysis of acetaminophen-induced acute hepatic injury in Sprague-Dawley rats . The aim of this study was to apply deep learning models for the assessment of toxicological pathology in a non-clinical study. Authors trained the model for various hepatic lesions, including necrosis, inflammation, infiltration, and portal triad at the WSI level. The deep learning model achieved an overall model accuracy of 96.44%. Importantly, the model predicted lesions of portal triad, necrosis, and inflammation with high correlations with annotated lesions by toxicologic pathologists. This study suggested that the deep learning algorithm (Mask R-CNN algorithm) can be applied to implement diagnosis and prediction of hepatic lesions in toxicological pathology. Zurac et al. developed the AI-based method for identifying mycobacterium tuberculosis in Ziehl-Neelsen-stained tissue specimen WSIs . In routine histopathological diagnosis, detecting mycobacterium tuberculosis in Ziehl-Neelsen-stained slides is difficult and time consuming because of the bacillus size. The developed deep learning model achieved an ROC-AUC of 0.977, an accuracy of 98.33%, a sensitivity of 95.65%, and a specificity of 100% for identifying mycobacterium tuberculosis bacilli on WSIs, which were better than or similar to those data of a team of pathologists who manually inspected slides and WSIs. By using the developed deep learning algorithm, the pathologists saved at least one-third of the total examining time. Park et al. proposed a new training method called MixPatch, which was designed to improve a CNN-based classifier by specifically addressing the prediction uncertainty problem and examine its effectiveness at improving the diagnosis performance in the context of histopathological image analysis . MixPatch generates and uses a new sub-training dataset, which consists of mixed patches and their pre-defined ground-truth labels. Importantly, by specifically considering the mixed region variation characteristics of the histopathology images, MixPatch augments the extant mixed image methods for medical image analysis, in which the prediction uncertainty is a crucial issue. MixPatch provides a new way to systematically alleviate the overconfidence problem of CNN-based classifiers and improve their prediction accuracy, contributing toward more calibrated and reliable histopathology image analysis. Serbanescu et al. demonstrated the morphological difference between nodular (low-risk subtype) and micronodular (high-risk subtype) basal cell carcinomas using a classical morphometric approach (a gray-level co-occurrence matrix and histogram analysis) and a deep learning semantic segmentation approach . The authors identified distinct pathological patterns of the tumor component in random fields of the tumor island that did not contain peripheral palisading. They demonstrated that the most significant difference between the morphology of the two (nodular and micronodular) subtypes was represented by the peritumoral cleft component. Importantly, the deep learning semantic segmentation approach provided new insight into the morphologies of nodular and micronodular subtypes of basal cell carcinoma. Nofallah et al. demonstrated the potential application of the semantic segmentation of clinically important tissue structures for improving the diagnosis of skin biopsy WSIs . It has been revealed that including a clinically important tissue structure along with WSIs improves the learning of the model, especially in challenging diagnostic classes, such as melanoma in situ and invasive melanoma (T1a). The model showed a 6% improvement in the F-score when whole slide images were used along with epidermal nests and cancerous dermal nest segmentation masks compared to that which was achieved using WSIs alone in training and testing the diagnosis pipeline. Importantly, comparing scores with 187 pathologists' performance on the same test set showed that the model can outperform or have comparable performance in the cases with the aforementioned diagnostic classes. Legnar et al. investigated the possibility to predict a final diagnosis based on a written neuropathological description using various natural language processing (NLP) methods . Certain diagnoses or groups of diagnoses (e.g., amyloid-deposition-associated diseases) could be predicted very well; however, in several cases, the morphological description was apparently not sufficient to make accurate predictions. This is because some diagnoses are associated with one pattern, but for others, there is a complex pattern combination which makes the prediction difficult without patho-physiological knowledge. Overall, it has been revealed that the morphological description texts, used as a surrogate for image analysis, enable the correct diagnosis to be achieved for some entities. Cazzato et al. trained the fast random forest (FRF) algorithm to be able to support the specialist to automatically highlight the anomalous pixel regions and to estimate a possible risk by quantifying the percentage of these regions with atypical morphological features starting from routine histopathological images . An important tool for melanoma diagnosis is the probability image estimated by the processed FRF output image. The probability image is useful to discriminate between information about ambiguous lesions. The FRF algorithm proved to be successful, with a discordance of 17% with respect to the results of the dermatopathologist, meaning that this type of supervised algorithm to can help the dermatopathologist in achieving the challenging diagnosis of malignant melanoma. VanBerlo et al. developed a deep learning solution for automatic lung ultrasound view annotation that effectively improves the efficiency of downstream annotation tasks, which can distinguish between parenchymal and pleural lung ultrasound views with 92.5% accuracy . The automatic partitioning of a 780 clip lung ultrasound dataset by view led to a 42 min reduction of the downstream manual annotation time and resulted in the production of 55 +- 6 extra relevant labels per hour. This deep learning-based automated tool considerably improved the annotation efficiency, resulting in a higher throughput relevant to the annotating task at hand, which can be applied to other unannotated datasets to save considerable manual annotation time and effort. Kawazoe et al. demonstrated an automated computational pipeline to detect glomeruli and to segment the histopathological regions inside of the glomerulus in a WSI . The computational pipeline automatically detects glomeruli on PAS-stained WSIs, followed by segmenting the Bowman's space, the glomerular tuft, the crescentic, and the sclerotic region inside of the glomeruli. To predict the estimated glomerular filtration rate (eGFR) in cases of immunoglobulin A nephropathy (IgAN), it is important to quantify the sclerotic region using the developed pipeline. Importantly, the developed automated computational pipeline could aid in diagnosing renal pathology by visualizing and quantifying the histopathological feature of the glomerulus and potentially accelerate the research in order to better understand the prognosis of IgAN. Fauzi et al. demonstrated a cell detection and classification system based on a deep learning model for use with the Allred scoring system for breast carcinoma hormone receptor status testing . The computational pipeline first detects all of the cells within the specific regions and classifies them into negatively, weakly, moderately, and strongly stained ones, followed by Allred scoring for the estrogen receptor (ER) status evaluation on WSIs. The automated Allred scores matches well with pathologists' scores for both the actual Allred score and hormonal treatment cases. The proposed system can automate the exhaustive exercise to provide fast and reliable assistance to pathologists and medical personnel. Palm et al. examined the performance of a digitalized and artificial intelligence (AI)-assisted workflow for HER2 status determination in accordance with the American Society of Clinical Oncology (ASCO)/College of Pathologists (CAP) guidelines . The HER2 4B5 algorithm in the uPath enterprise software and the HER2 Dual ISH image analysis algorithm (Roche Diagnostic International, Rotkreuz, Switzerland) were used in this study. The authors demonstrated the feasibility of a combined HER2 IHC and ISH AI workflow in the primary and metastatic breast cancers, with a Cohen's k of 0.94 when it was assessed in accordance with the ASCO/CAP recommendations. In summary, in this Special Issue, there are wide varieties of valuable scientific papers including a review article and papers on deep learning models in pathological applications, human and toxicological pathology, and various methodologies. AI-based computational algorithms, including deep learning models, are taking digital pathology beyond mere digitization and telepathology . The incorporation of AI-based computer vision and natural language processing algorithms in routine clinical workflows is on the horizon, reducing processing time and increasing the detection rate of anomalies . It is necessary to continue to share the latest findings and updated methodologies in "Artificial Intelligence in Pathological Image Analysis" and continue to conduct valuable research. Conflicts of Interest M.T. is the employee of Medmain Inc. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Tsuneki M. Deep learning models in medical image analysis J. Oral Biosci. 2022 64 312 320 10.1016/j.job.2022.03.003 35306172 2. Kim I. Kang K. Song Y. Kim T.J. Application of Artificial Intelligence in Pathology: Trends and Challenges Diagnostics 2022 12 2794 10.3390/diagnostics12112794 36428854 3. Tsuneki M. Abe M. Kanavati F. A Deep Learning Model for Prostate Adenocarcinoma Classification in Needle Biopsy Whole-Slide Images Using Transfer Learning Diagnostics 2022 12 768 10.3390/diagnostics12030768 35328321 4. Rakovic K. Colling R. Browing L. Dolton M. Horton M.R. Protheroe A. Lamb A.D. Bryant R.J. Scheffer R. Crofts J. The Use of Digital Pathology and Artificial Intelligence in Histopathological Diagnostic Assessment of Prostate Cancer: A Survey of Prostate Cancer UK Supporters Diagnostics 2022 12 1225 10.3390/diagnostics12051225 35626380 5. Baek E.B. Hwang J.H. Park H. Lee B.S. Son H.Y. Kim Y.B. Jun S.Y. Her J. Lee J. Cho J.W. Artificial Intelligence-Assisted Image Analysis of Acetaminophen-Induced Acute Hepatic Injury in Sprague-Dawley Rats Diagnostics 2022 12 1478 10.3390/diagnostics12061478 35741291 6. Zurac S. Mogodici C. Poncu T. Trascau M. Popp C. Nichita L. Cioplea M. Ceachi B. Sticlaru L. Cioroianu A. A New Artificial Intelligence-Based Method for Identifying Mycobacterium Tuberculosis in Ziehl-Neelsen Stain on Tissue Diagnostics 2022 12 1484 10.3390/diagnostics12061484 35741294 7. Park Y. Kim M. Ashraf M. Ko Y.S. Yi M.Y. MixPatch: A New Method for Training Histopathology Image Classifiers Diagnostics 2022 12 1493 10.3390/diagnostics12061493 35741303 8. Serbanescu M.S. Bungardean R.M. Georgiu C. Crisan M. Nodular and Micronodular Basal Cell Carcinoma Subtypes Are Different Tumors Based on Their Morphological Architecture and Their Interaction with the Surrounding Stroma Diagnostics 2022 12 1636 10.3390/diagnostics12071636 35885545 9. Nofallah S. Li B. Mokhtari M. Wu W. Knezevich S. May C.J. Chang O.H. Elmore J.G. Shapiro L.G. Improving the Diagnosis of Skin Biopsies Using Tissue Segmentation Diagnostics 2022 12 1713 10.3390/diagnostics12071713 35885617 10. Legnar M. Daumke P. Hesser J. Porubsky S. Popovic Z. Bindzus J.N. Siemoneit J.H.H. Weis C.A. Natural Language Processing in Diagnostic Texts from Nephropathology Diagnostics 2022 12 1726 10.3390/diagnostics12071726 35885630 11. Cazzato G. Massaro A. Colagrande A. Lettini T. Cicco S. Parente P. Nacchiero E. Lospalluti L. Cascardi E. Giudice G. Dermatopathology of Malignant Melanoma in the Era of Artificial Intelligence: A Single Institutional Experience Diagnostics 2022 12 1972 10.3390/diagnostics12081972 36010322 12. VanBerlo B. Smith D. Tschirhart J. VanBerlo B. Wu D. Ford A. McCauley J. Wu B. Chaudhary R. Dave C. Enhancing Annotation Efficiency with Machine Learning: Automated Partitioning of a Lung Ultrasound Dataset by View Diagnostics 2022 12 2351 10.3390/diagnostics12102351 36292042 13. Kawazoe Y. Shimamoto K. Yamaguchi R. Nakamura I. Yoneda K. Shinohara E. SHintani-Domoto Y. Ushiku T. Tsukamoto T. Ohe K. Computational Pipeline for Glomerular Segmentation and Association of the Quantified Regions with Prognosis of Kidney Function in IgA Nephropathy Diagnostics 2022 12 2955 10.3390/diagnostics12122955 36552963 14. Fauzi M.F.A. Ahmad W.S.H.M.W. Jamaluddin M.F. Lee J.T.H. Khor S.Y. Looi L.M. Abas F.S. Aldahoul N. Allred Scoring of ER-IHC Stained Whole-Slide Images for Hormone Receptor Status in Breast Carcinoma Diagnostics 2022 12 3093 10.3390/diagnostics12123093 36553102 15. Palm C. Connolly C.E. Masser R. Sgier B.P. Karamitopoulou E. Simon Q. Bode B. Tinguely M. Determining HER2 Status by Artificial Intelligence: An Investigation of Primary, Metastatic, and HER2 Low Breast Tumors Diagnostics 2023 13 168 10.3390/diagnostics13010168 36611460 |
PMC10000563 | Pediatric renal cell carcinoma (RCC) is a rare malignancy. Magnetic resonance imaging (MRI) is the preferred imaging modality for assessment of these tumors. The previous literature has suggested that cross-sectional-imaging findings differ between RCC and other pediatric renal tumors and between RCC subtypes. However, studies focusing on MRI characteristics are limited. Therefore, this study aims to identify MRI characteristics of pediatric and young-adult RCC, through a single-center case series and literature review. Six identified diagnostic MRI scans were retrospectively assessed, and an extensive literature review was conducted. The included patients had a median age of 12 years (63-193 months). Among other subtypes, 2/6 (33%) were translocation-type RCC (MiT-RCC) and 2/6 (33%) were clear-cell RCC. Median tumor volume was 393 cm3 (29-2191 cm3). Five tumors had a hypo-intense appearance on T2-weighted imaging, whereas 4/6 were iso-intense on T1-weighted imaging. Four/six tumors showed well-defined margins. The median apparent diffusion coefficient (ADC) values ranged from 0.70 to 1.20 x 10-3 mm2/s. In thirteen identified articles focusing on MRI characteristics of MiT-RCC, the majority of the patients also showed T2-weighted hypo-intensity. T1-weighted hyper-intensity, irregular growth pattern and limited diffusion-restriction were also often described. Discrimination of RCC subtypes and differentiation from other pediatric renal tumors based on MRI remains difficult. Nevertheless, T2-weighted hypo-intensity of the tumor seems a potential distinctive characteristic. renal cell carcinoma magnetic resonance imaging pediatrics radiology pathology Stichting Kinderen Kankervrij (KiKa)341 This work was supported by a grant (grant number 341) from the Stichting Kinderen Kankervrij (KiKa). pmc1. Introduction Pediatric renal cell carcinoma (RCC) is a rare renal malignancy . Although Wilms tumors (WTs) show the highest prevalence in young children, the incidence of RCC increases in the second decade of life . Whereas in the Renal Tumor Study Group of the International Society of Pediatric Oncology (SIOP-RTSG), pre-operative chemotherapy is the standard of care for WTs, upfront surgery is recommended for localized RCC . Invasive procedures to determine histology before the start of therapy in young children are discouraged . Age and size of the tumor are important factors in the consideration of the diagnosis of pediatric renal tumors as well as in the consideration of performing a biopsy, indicating age >7 years as a criterion to consider tumor biopsy . Thus far, no specific imaging characteristics discriminating RCC from WTs and other non-WTs have been identified . Magnetic resonance imaging (MRI) is currently the preferred modality for the assessment of pediatric renal tumors within the SIOP-RTSG given its lack of ionizing radiation and excellent soft-tissue contrast. Furthermore, MRI is subject to continuous technical developments, such as the possibility of calculating the apparent diffusion coefficient (ADC) value using diffusion-weighted imaging (DWI) . MRI could, therefore, play a potential role in the non-invasive discrimination of pediatric renal tumors . Contrary to the rarity of RCC in children, this tumor type is the most common renal tumor in adolescents and adults . Nevertheless, childhood RCC shows distinct histological characteristics, possibly related to the different distribution of RCC subtypes. Whereas translocation-type RCC (MiT-RCC), which has been officially recognized since 2004 by the World Health Organization, is the most frequent subtype in children, clear-cell RCC (ccRCC) is the predominant histological subtype in adults . MiT-RCC is diagnosed based on translocations including transcription factor E3 (TFE3) and EB (TFEB), which are members of the family of microphthalmia transcription factors (MiT) . Interestingly, the previous literature has suggested that cross-sectional imaging findings differ between RCC subtypes . Until today, studies focusing on the MRI characteristics of pediatric RCC are limited in number, although identification of potential specific MRI characteristics of WTs and non-WTs is important for future validation studies . Therefore, this study aims to retrospectively identify MRI characteristics of pediatric RCC patients at diagnosis through a case series in our center, including a literature review focusing on this topic. 2. Materials and Methods 2.1. Patients Institutional Review Board approval was obtained. For this retrospective study, obtaining further formal consent was waived. All diagnostic MRI scans included were clinically indicated and were performed as the standard of care. Between 2014 and 2019, we identified 6 children with RCC that underwent an MRI scan at diagnosis. The standard of care for localized pediatric RCC is upfront total nephrectomy . Only in case of doubt of a WT diagnosis, based on predefined clinical and imaging characteristics, a core needle biopsy was performed. If there was no suspicion of a non-WT, the patients were pre-operatively treated with 4 weeks of vincristine/actinomycin-D (stage I-III) or 6 weeks of vincristine/actinomycin-D/doxorubicin (stage IV/V), according to the SIOP-RTSG protocol. 2.2. Magnetic Resonance Imaging Acquisition Abdominal MRI for pediatric renal tumors in this study was performed using a 1.5T MRI system (Achieva, Philips Healthcare, Eindhoven, The Netherlands and Ingenia, Philips Healthcare, Eindhoven, The Netherlands). Two patients were scanned in external hospitals at diagnosis before referral to our center (Signa HDxt; GE Healthcare, Boston, USA and Magnetom Avanto; Siemens, Erlangen, Germany). Scan protocols slightly varied but at least consisted of coronal and axial T2-weighted imaging, axial T1-weighted turbo spin-echo and axial DWI with automatically generated ADC maps. Five patients underwent post-contrast T1-weighted imaging, whereas for one patient, contrast-enhanced MRI was not available (Table 1). Children were awake, sedated or under general anesthesia depending on their ability to cooperate, according to the standard-of-care procedures. Gadobutrol (Gadovist; Bayer B.V., Leverkusen, Germany) was administered intravenously at a dose of 0.1 mL/kg body weight. Hyoscine butylbromide (Buscopan; Sanofi, Paris, France) was administered intravenously at a dose of 0.4 mg/kg body weight to reduce peristaltic artifacts, with a maximum of 10 mg in children >=6 years and a maximum of 5 mg in children <6 years. All children were screened for contraindications for MRI and those concerning intravenous agents. For the two patients scanned at local hospitals, specifications of gadobutrol and hyoscine butylbromide were not available. 2.3. Image Analysis The anonymized MRI datasets were transferred to DICOM software Osirix v. 5.5.2 (Pixmero, SARL, Bernex, Switzerland). Two pediatric radiologists (ASL with 13 years of experience and RAJN with 26 years of experience in body MRI, respectively), who were blinded to the histopathological subtype and clinical characteristics but were aware of the pediatric RCC diagnosis, reviewed the diagnostic MRI scans. All diagnostic scans were assessed using a case report form based on previous studies identifying potential specific imaging characteristics of different pediatric renal tumors . The pediatric RCC cases were analyzed focusing on tumor presentation, growth pattern, characteristics of solid components and enhancement pattern, if available. Tumor volume was calculated based on the three dimensions of the tumor times 0.523. Moreover, up to four round-shaped ROIs containing solid areas of the tumor, mainly based on enhancement, were drawn in order to measure the ADC value of the most representative parts of the tumor. To limit inter-observer variability, an instruction form accompanying the case report form was provided. 2.4. Histopathological Review Our national coordinating SIOP-RTSG histopathologist (RRK with 23 years of experience with pediatric renal tumor histopathology) reviewed the available macroscopy and microscopy from the surgically resected tumors and biopsies of all patients following the most recent WHO classification system . Following protocol, the dorsal and ventral side and hilar region of the resected specimen were marked with varying color dyes following the instructions of the involved surgeons. The specimens were sliced free-handed in successive 10 to 20 mm transverse macroscopic slices in a cranial to caudal sequence or through longitudinal incision to bivalve the specimen. 2.5. Statistical Analyses Due to the small number of patients, inter-observer agreement between the two pediatric radiologists was difficult to assess because Cohen's kappa is affected by the prevalence of the finding under observation. Only six patients were included in this study, potentially resulting in low values or even an impossible calculation of kappa when focusing on separate characteristics . Therefore, the inter-observer agreement was assessed using percentages of observed agreement, including the intra-class correlation coefficient (ICC) for the median ADC values including the regions of interest and for median tumor volumes. ICC values were interpreted as satisfactory >0.75 . 2.6. Literature Review A literature review was performed following PRISMA guidelines to reflect on the case series and elaborate on the current knowledge about the MRI appearance of RCC by focusing on the predominant histological subtypes in the pediatric and young-adult population. For this purpose, PubMed, Embase/Medline and Cochrane libraries were searched in November 2021, using the main search terms 'renal cell carcinoma' and 'magnetic resonance imaging' (Table S1). The study has not been registered. Cross-referencing and a citation check of the included papers were executed using Scopus. Articles were included when they (1) included MRI characteristics of patients with proven RCC; (2) were prospective or retrospective cohort studies, randomized controlled trials or case reports; (3) were written in the English language; and (4) were available in the full-text form. Subsequently, articles focusing on children (<19 years), potentially also including adolescents or young adults (<=35 years) and articles focusing on MiT-RCC were separated to serve as the focus of this literature review. Given the rarity of studies focusing on the MRI characteristics of MiT-RCC, articles focusing on adults were also included for this purpose. With this approach, we guaranteed identification of all relevant articles while subdividing their relevance for our study based on their full-text content. After removal of duplicates, 7012 articles were screened based on title and abstract, leaving 363 articles for full-text screening, resulting in the inclusion of 95 articles. Of these, 13 articles focused on pediatric, adolescent and young-adult RCC, and 13 articles focused on MiT-RCC, with an overlap of 6 articles . In November 2022, the search was updated, with no additional results for articles focusing on children and/or MiT-RCC. 3. Results 3.1. Case Presentation 3.1.1. Patient Characteristics The six identified patients in our center had a median age of 12 years (range 63-193 months) (Table 2). Four patients were female, and half of the patients presented with a right-sided tumor. Two/six patients received pre-operative chemotherapy following suspicion of a WT, whereas 4/6 underwent upfront surgery. In one case, RCC was pre-operatively confirmed through tumor biopsy. Three patients had stage 1 disease, whereas the other patients had stage 2 (1/6) and stage 3 (2/6) disease (Table 2). 3.1.2. Histopathology The average post-operative specimen weight was 898.6 g (range 210-2100 g), whereas the maximum post-operative tumor diameter ranged from 2.4 to 12.9 cm (median 6.8 cm). The post-operative weight of the specimen was missing for one patient, of which the largest tumor diameter was 9.5 cm (Table 2). Five patients were tested for MiT-RCC, resulting in 2/5 MiT-RCC cases . In 4/5 cases, FISH was used, whereas in the two most recent cases, also RNA sequencing was performed, resulting in a rearrangement of TFE3 and SFPQ in the sixth patient. Two patients were diagnosed with ccRCC, and in one patient, the subtype could not be specified. The first patient, who was not tested for MiT-RCC, showed an FH mutation in the context of a hereditary leiomyomatosis and RCC cancer syndrome (Table 2) . For the 5-year-old patient diagnosed with ccRCC, the FISH for MiT-RCC was not conclusive, and RNA sequencing for further analysis of TFEB was not available. 3.1.3. Imaging Characteristics at Diagnosis The median observed agreement between the two observers was 83% (range 33.3%-100%). The few imaging characteristics with low observed agreement were discussed between the two radiologists, and mismatching concepts were resolved (Table S2). Furthermore, the inter-reader agreement for median tumor volume was excellent, with an ICC of 0.991 (95% 0.941-0.999). Therefore, the imaging characteristics found by the first reader (ASL) were reported (Table 2). Tumor volume ranged from 29 to 2191 cm3, with varying locations. The shape of the tumors was predominantly lobulated (4/6), and margins were well-defined in a majority of the patients (4/6). Capsule rupture was seen in only 2/6 cases, which was defined as an interruption of the hypo-intense capsule of the tumor. None of the cases presented with a tumor thrombus. Concerning hemorrhage and necrosis, these components were present in 3/6 and 1/6 cases, respectively. Cysts were present in 2/6 cases, whereas fatty tissue and subcapsular fluid were not observed (Table 2). The tumors presented mainly homogeneously (4/6), with a predominant hypo-intense appearance on T2-weighted imaging and iso-intense appearance on T1-weighted imaging. Almost all cases showed a homogeneous enhancement pattern, varying from mild to strong enhancement (Table 2). There was no obvious consistency concerning MRI characteristics within patients based on histological subtype . 3.1.4. Diffusion-Weighted Imaging Inter-reader agreement was excellent for median ADC values with an ICC of 0.942 (95% CI 0.639-0.992) (Table S3). Therefore, only the median surfaces of ROIs and median ADC values measured by the first reader (ASL) were reported (Table 2). The median ADC values ranged from 0.70 to 1.20 x 10-3 mm2/s. The MiT-RCC cases and the case diagnosed as ccRCC but with inconclusive TFE results showed the lowest ADC values, ranging from 0.70 to 0.98 x 10-3 mm2/s . 3.2. Literature Review 3.2.1. Pediatric and Young-Adult RCC We identified thirteen studies focusing on MRI findings of pediatric RCC, with a total of 25 patients . Ages ranged from 4 to 33 years, with four studies also including young adults <=35 years . Six studies focused on MiT-RCC, whereas other histological subtypes represented ccRCC, papillary type RCC (pRCC), chromophobe RCC (chrRCC), renal medullary carcinoma (RMC) and other rare RCC types. The location of all reported pediatric RCC tumors in the identified articles varied from central to peripheral (Table 3). On T1-weighted imaging and T2-weighted imaging, tumors appeared predominantly heterogeneously, whereas no clear predominant intensity was seen for one of these sequences. Accordingly, enhancement pattern on contrast-enhanced MRI was reported mostly as heterogeneous. Cysts, when specified, were found in only three cases, whereas the presence of necrosis and/or hemorrhage was often not specified . Regional lymph node involvement and/or metastases to lymph nodes were reported in five studies (Table 3) . In a study of seven patients, Wang et al. reported positive regional lymph node status in four patients and positive cervical lymph node status in one patient . Blitman et al. reported two patients with vascular tumor involvement of the renal vein and three patients with encasement of the vascular pedicle out of a total of six patients, all with infiltrative tumors with ill-defined margins (Table 3) . Only one study specified findings of DWI, reporting the iso-intense appearance of the tumor on the b500 DWI sequence compared to the renal parenchyma . Concerning MRI characteristics of RCC subtypes other than MiT-RCC in children, Zou et al. reported a case of a 17-year-old male with von Hippel-Lindau disease with bilateral renal cysts and ccRCC (Table 3) . This patient showed T2-weighted hyper-intensity and T1-weighted hypo-intensity, whereas enhancement was limited on contrast-enhanced imaging. Koetter et al. described a 16-year-old female at 31 weeks' gestation presenting with a large, heterogeneous cystic-solid mass, which was histologically diagnosed as pRCC . Another reported pRCC that presented as a complex cyst containing bloody elements, whereas a pediatric chrRCC showed a well-defined T1-weighted hypo-intense and T2-weighted hyper-intense tumor with necrosis (Table 3) . Finally, RMC has been described as a very rare and malignant renal tumor, especially in children and young adults, and is often seen in RCC patients with sickle-cell traits . Norena-Rengifo et al. described a 12-year-old male with an intermediate enhancing mass on T1-weighted imaging with evident retroperitoneal lymphadenopathies, similar to the reported regional adenopathy identified on MRI in a retrospective study by Blitman et al. (Table 3) . 3.2.2. MiT-RCC Thirteen studies focusing on MRI characteristics of MiT-RCC were identified, including the six identified studies focusing on pediatric patients with MiT-RCC . There was a total of 46 patients, who were aged 4-76 years old, with MiT-RCC included in the identified articles. Whereas the tumor location was again highly variable among patients, overall, there was a majority showing hyper-intensity on T1-weighted imaging and hypo-intensity on T2-weighted imaging, with a heterogeneous enhancement pattern. Wang et al. reported 8/9 patients with necrosis, and 7/9 patients with hemorrhage, whereas in other studies, these characteristics were often not specified . The tumor composition and growth pattern of MiT-RCC was very heterogeneous, although a substantial part of the cases seems to present with an infiltrative and/or irregular growth pattern. Fifteen patients presented with lymph node involvement; however, four studies lacked information concerning this characteristic. Reported metastatic sites were liver and/or lungs in a total of three patients . DWI characteristics were reported in 5 studies for a total of 23 patients . Overall, diffusion restriction seemed limited in these cases, with, for instance, Tohi et al. reporting no restriction and Chen et al. reporting a relatively high signal on the ADC map . Razek et al. showed a mean ADC value of 1.50 +- 0.97 for four patients . In our case series, the 14-year-old female patient in particular showed a typical presentation of MiT-RCC based on these findings in the previous literature. The tumor showed an ill-defined tumor with capsule invasion and an infiltrative growth pattern, appearing hypo-intense on T2-weighted imaging with a relatively high median ADC value . The presentation of the 16-year-old male patient with MiT-RCC seemed less typical . 3.2.3. Other Subtypes The RCC subtypes most frequently occurring in children and adolescents besides MiT-RCC are ccRCC, pRCC and chrRCC (Table 3) . Knowledge of MRI characteristics of these subtypes is based mainly on adult studies. A retrospective study of Wang et al. focused on the MRI characteristics of 57 adult RCC patients, in which ccRCC and pRCC showed hemorrhage in 20-25% of the cases compared to no evidence of hemorrhage for chrRCC . Moreover, a very high percentage of cystic necrosis was seen in ccRCC and pRCC, resulting in a significant difference of this characteristic compared to chrRCC, for which no cases were seen. Compared to ccRCC, other RCC subtypes often show a less aggressive growth pattern on MRI, which is illustrated by a higher numbers of cases with well-defined margins, less peripheral invasion and less extension of the tumor . Oliva et al. described the MRI-features of 21 pRCCs and 28 ccRCCs, concluding that pRCC typically presents with T2 hypo-intensity, whereas ccRCC typically shows T2 hyper-intensity . This finding, as well as the occurrence of increased enhancement in ccRCC compared to pRCC and chrRCC, has often been reported in the previous literature . Furthermore, ccRCC seems to show significantly higher ADC values than pRCC and chrRCC . 4. Discussion There seems to be a lack of specific imaging characteristics for discrimination of pediatric RCC and its subtypes based on MRI characteristics alone . Nevertheless, imaging plays an increasingly important role in the diagnosis and follow-up of pediatric renal tumors and in the discrimination of different renal tumor types . The heterogeneous diagnostic appearance of our patients was in line with findings in the identified literature and with previous studies stating that RCC is often indistinguishable from WTs based on MRI characteristics alone . Part of the included patients showed cysts, necrosis and hemorrhage; however, none of these characteristics were explicitly found in all patients . Calcifications have often been reported as common findings in pediatric RCC; however, MRI does not allow for a trustworthy assessment of calcifications and was, therefore, not included as an imaging characteristic in our case report form . Despite the recommendation of the SIOP-RTSG to use MRI for cross-sectional imaging of renal tumors, various countries still perform abdominal CT scans in these patients. One of the largest studies focusing on CT characteristics of pediatric RCC to date also reported a widely variable radiological appearance, often with the presence of calcifications . Nevertheless, calcifications can also be seen in WTs, making discrimination based on this imaging characteristic difficult given the rarity of pediatric RCC and other non-WTs . Finally, the findings in our case series were in concordance with the frequently reported localized presentation and small size of pediatric RCC . Whereas MiT-RCC is the most frequent histological subtype in children, we reported only two out of six patients with a proven TFE translocation. The MRI characteristics of these two patients were quite different from one another. MiT-RCC, similar to ccRCC, is often described as a relatively aggressive tumor in terms of growth pattern and tumor extension as well as prognosis . Nevertheless, only one MiT-RCC case showed an infiltrative growth pattern with capsule rupture, whereas the second MiT-RCC case and both ccRCC cases had well-defined margins with the presence of a pseudocapsule, without any signs of aggressive growth. In general, capsule rupture remains difficult to assess. Concerning the discrimination between histological RCC subtypes, the predominantly reported T2-weighted hypo-intensity in MiT-RCC is also often described for pRCC and chrRCC, whereas ccRCC classically demonstrates high intrinsic T2-weighted signal intensity . Nonetheless, knowledge of specific MRI characteristics of MiT-RCC remains limited, given the rarity of MiT-RCC in adult patients and its relatively recent recognition as an official subtype by the WHO . Whereas in adult RCC, the main focus is often the discrimination of histological subtypes, in pediatric RCC, discrimination from the much more frequently occurring WTs in the early diagnostic stages is of great importance . WTs have a very heterogeneous presentation at diagnosis and are, most often, large intra-renal tumors with a pseudocapsule . Whereas an irregular growth pattern and absence of a capsule are often described as common for RCC in the previous literature, we observed a majority of well-defined margins and the presence of a pseudocapsule in our case series. Nonetheless, an enhancing capsule has also been reported as a characteristic of MiT-RCC . MRI characteristics reported to be typical for RCC will still not be discriminative given the heterogeneous appearance of WTs. Nevertheless, WTs often appear hyper-intense on T2-weighted imaging, which is opposite to the T2-weighted hypo-intensity in a majority of our cases with RCC, as substantiated by the findings in the previous literature . Finally, RCC is often reported to be smaller than WTs . Following SIOP-RTSG protocols, based on the suspicion of a non-WT, a biopsy is recommended for children >=10 years of age and for children between 7 and 10 years old with a tumor volume <200 mL . In our case series, tumor volume was relatively low, except for the expected large FH-RCC case (case nr. 1). In the previous literature, tumor volume ranged widely; however, often only the largest diameter was reported . Overall, there seems to remain a lack of pathognomonic MRI characteristics for the discrimination of pediatric RCC from other renal malignancies in children, as well as for the differentiation of histological subtypes . Nevertheless, DWI has shown an increasing potential reliability for the non-invasive discrimination of renal lesions . Whereas only one included pediatric study focused on the diffusion restriction of pediatric RCCs, our literature review confirmed results from previous overviews stating adult clear-cell RCC has shown significantly higher ADC values compared to non-clear-cell RCC . In contrast, our case series showed the three lowest median ADC values in the ccRCC and MiT-RCC cases, whereas also relatively high ADC values were reported. In WTs, relatively low ADC values can be observed, varying among histological WT subtypes . In children, discrimination of common histological RCC subtypes, as well as discrimination from WTs based on DWI, therefore, remains difficult. Nonetheless, the female patient with MiT-RCC in our case series appeared to have a typical presentation in the light of previous reports, showing potential discriminative MRI characteristics for TFE-positive tumors. Future studies may focus on validating adult findings in the pediatric population and explore the relationship between ADC values and common pediatric RCC subtypes combined with other typical MRI characteristics. Over the past decades, differences between adult and pediatric RCC have increasingly been appreciated. Concerning imaging studies, the direct comparison of the pediatric and adult population has become even more complicated by the preference of CT in the adult population, whereas MRI has developed as the preferred imaging modality within the SIOP-RTSG . Nevertheless, MRI also plays an increasingly important role in the adult population, mainly due to its ability to perform quantitative measurements . Therefore, when searching the literature databases for MRI characteristics of pediatric RCC and MiT-RCC, the literature about the adolescent and adult population cannot be ignored. Not only because knowledge of MR imaging of these cases is scarce, but also because they are often embedded in studies focusing on adolescents and/or adults as well. Concerning cut-off values for age classification, we focused on the predefined range of 18-35 years for the 'adolescents and young adults' often used in Europe. However, this classification varies around the world . Our study has a few limitations, mainly based on its retrospective nature and small study population. The limited number of patients did not allow any statistical analysis or strong conclusions. Furthermore, scan parameters were inconsistent due to not as yet centralized care. Nevertheless, these cases served mainly as an illustration accompanying the literature review in this developing field of research. In this way, this descriptive study contributes to the increasing knowledge of pediatric RCC and its diagnostic presentation on MRI. Concerning the reported imaging characteristics by two independent observers, there was excellent inter-observer agreement . The small number of patients in this study does not allow for strong conclusions concerning validity of the use of the CRF in other populations. 5. Conclusions For a few years, MRI has been the preferred imaging modality for imaging pediatric renal tumors within the SIOP-RTSG protocol. This case series represents one of the largest retrospective reports so far, including an extensive review focusing on MRI characteristics of RCC in the pediatric and young-adult population. The reported cases showed a varying presentation of different pediatric RCC subtypes on MRI, in line with the published literature. Nevertheless, based on this study, T2-weighted hypo-intensity of the tumor has been shown to be a potential distinctive characteristic for the discrimination of RCC from other renal tumors that are prevalent at this age, especially WTs. Future studies should focus on larger study populations through international collaboration, also exploring innovative techniques such as DWI as a non-invasive biomarker. Supplementary Materials The following supporting information can be downloaded at Table S1: Search strategy focusing on MRI characteristics of RCC; Table S2: Observed percentage agreement for dichotomous and categorical characteristics in the case report form for the two observers; Table S3: Median surface of ROI and median ADC values per patient for the two observers. Click here for additional data file. Author Contributions Conceptualization, J.N.v.d.B., R.R.d.K., M.M.v.d.H.-E. and A.S.L.; methodology, J.N.v.d.B., R.R.d.K., R.A.J.N., A.B., A.J.K., M.M.v.d.H.-E. and A.S.L.; formal analysis, J.N.v.d.B., A.S.L.; investigation, J.N.v.d.B., R.R.d.K., R.A.J.N. and A.S.L.; resources, J.N.v.d.B., R.R.d.K., M.M.v.d.H.-E. and A.S.L.; writing--original draft preparation, J.N.v.d.B.; writing--review and editing, J.N.v.d.B., R.R.d.K., R.A.J.N., A.B., A.J.K., M.M.v.d.H.-E. and A.S.L.; visualization, J.N.v.d.B., R.R.d.K. and A.S.L.; supervision; R.R.d.K., M.M.v.d.H.-E. and A.S.L. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approval from the Institutional Review Board of the University Medical Center Utrecht (WAG/mb/20/019804 20-332, 26-05-2020) was obtained. For this retrospective study, formal consent was waived. Informed Consent Statement Additional patient consent was waived due to the retrospective nature of this study. All diagnostic MRI scans included were clinically indicated and were performed as the standard of care. Data Availability Statement Restrictions apply to the availability of these data. The data that support the findings of this study are available in the Supplementary Materials and from the International Society of Pediatric Oncology-Renal Tumor Study Group office following standard access procedures upon reasonable request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Flow chart of the literature review. Figure 2 Imaging and histopathology of a 14-year-old female patient with a right-sided translocation-type RCC (MiT-RCC). On T2-weighted imaging (A,B) the tumor appears hypo-intense with ill-defined margins, compared to the iso-intense homogeneous appearance on T1-weighted imaging (E) with relatively strong homogeneous enhancement on T1-weighted contrast-enhanced imaging (F). DWI showed restricted diffusion on the b500 scan (C), with a relatively high median ADC value of 0.98x10-3 mm2/s calculated based on the b0/b500s map (D). The macroscopic (G) and microscopic histopathology (H) showed an infiltrating tumor, detail showing tumor cells with hyperchromatic nuclei and papillary growth pattern (I). Figure 3 Imaging and histopathology of a 16-year-old male patient with a left-sided translocation-type RCC (MiT-RCC). On T2-weighted imaging (A,D) the tumor appears hypo-intense and heterogeneous with well-defined margins, similar to a hypo-intense appearance on T1-weighted imaging (B) with mild, heterogeneous enhancement on T1-weighted contrast-enhanced imaging (C). DWI showed restricted diffusion on the b1000 scan (E), with a median ADC value of 0.80 x10-3 mm2/s on the b0/b1000 map (F). The macroscopic histopathology (G) shows a large, round tumor, with little remaining normal renal tissue. The microscopic HE image (H) shows a capsule around the tumor, with a predominantly epithelial growth pattern in nests, often with cells with clear cytoplasm and mildly atypical nuclei (I). cancers-15-01401-t001_Table 1 Table 1 Scan parameters at 1.5-T MRI of the scanned sequences. Patient nr. 1 2 3 4 5 6 T2-weighted imaging Repetition time (ms) 7500 447 1400 454 2457 2457 Echo time (ms) 123 90 92 90 100 100 Slice thickness (mm) 5.5 1.15 4 1.15 5 5 Echo train length 17 85 256 85 39 39 Slicing gap 6.5 1.15 4.4 1.15 5 5 Acquisition matrix 320 x 224 348 x 348 384 x 194 348 x 348 452 x 78 452 x 78 T1-weighted imaging Repetition time (ms) 6.3 5.5 4.7 5.4 5.5 5.5 Echo time (ms) 3.1 2.7 2.4 2.7 2.7 2.7 Slice thickness (mm) 5 3 3 3 3 3 Echo train length 1 60 1 60 60 60 Slicing gap 2.5 1.5 NS 1.5 1.5 1.5 Acquisition matrix 288 x 192 232 x 233 320 x 170 232 x 233 232 x 233 260 x 261 Diffusion weighted imaging Repetition time (ms) 13333 2084 5300 2084 2398 2398 Echo time (ms) 634 72 75 72 73 73 Slice thickness (mm) 6 5 6 5 5 5 Echo train length 1 35 1 35 35 35 Slicing gap 7.2 5 7.2 5 5 5 Acquisition matrix NS 88 x 70 192 x 153 88 x 70 88 x 70 88 x 70 b values 0/50/600/1000 0/50/200/400/800 0/500 0/50/200/400/800 0/100/1000 0/100/1000 ms = milliseconds; mm = millimeters; NS = not specified. cancers-15-01401-t002_Table 2 Table 2 Characteristics of the included pediatric patients with RCC. Patient nr. 1 2 3 4 5 6 Clinical characteristics Age (months) 184 63 179 109 63 193 Sex Female Female Female Male Female Male Tumor side Right Left Right Left Right Left Pre-operative chemotherapy No Yes No No Yes No Surgical approach TN TN TN TN TN TN Tumor stage 1 1 2 3 1 3 Biopsy performed No No Yes No No No Pathology findings Weight of the specimen (gram) 2100 NS 210 610 753 820 Tested for MiT-RCC (test) No Yes (FISH) Yes (FISH) Yes (FISH) Yes (FISH, RNA-seq) Yes (RNA-seq) Histopathological subtype FH-RCC ccRCC MiT-RCC NOS ccRCC MiT-RCC Genetic analysis FH-mutation d NS NS NS None NS General tumor characteristics on MRI Tumor volume (cm3) 2191 110 29 353 433 554 Location of the tumor Indist Central Peripheral Peripheral Central Indist Regional lymph nodes No No No No No No Shape Lobulated Round Lobulated Lobulated Lobulated Round Margins Well-def Well-def Ill-def Ill-def Well-def Well-def Pseudocapsule Yes Yes No No Yes Yes Growth pattern on MRI Capsule rupture/invasion No No Yes Yes No No Infiltrative growth pattern No No Yes No No No Venous invasion/Tumor thrombus No No No No No No MRI characteristics of solid components of the tumor T2W imaging Pattern Hetero Homo Homo Homo Homo Hetero Intensity Hypo, Iso Iso Hypo Hypo Hypo Hypo T1W imaging Pattern Hetero Homo Homo Homo Homo Hetero Intensity Iso Iso Iso Hypo Iso Hypo Enhancement, degree and pattern Strong, homo Mild, homo Strong, homo Strong, homo NA a Mild, hetero Hemorrhage, degree No Yes, ext b No Yes, minimal No Yes, min c Necrosis No No No No No Yes Cysts Yes Yes b No No No Yes Septation No No No No No No Fatty tissue No No No No No No Subcapsular fluid No No No No No Yes c Increased vascularity No No Yes Yes No Yes Median surface ROIs (cm2) 4.29 0.45 2.66 9.06 18.14 2.61 Median ADC valued (x10-3 mm2/s) 1.20 1.05 0.98 1.20 0.70 0.80 TN = total nephrectomy; NS = not specified; FISH = fluorescence in situ hybridization; RNA-seq = RNA sequencing; RCC = renal cell carcinoma; FH-RCC = fumarate-hydratase-deficient RCC; ccRCC = clear cell type RCC; MiT-RCC = translocation-type RCC; NOS = not otherwise specified; Indist = indistinguishable; def = defined; Hetero = heterogeneous; Homo = homogeneous; ext = extensive; min = minimal. a No contrast-enhanced diagnostic MRI scan available. b Hemorrhage of/in the cystic lesion. c Subcapsular fluid suspected of hemorrhage. d Hereditary leiomyomatosis and renal cell cancer. cancers-15-01401-t003_Table 3 Table 3 Review of the literature focusing on MRI characteristics of pediatric and young-adolescent renal cell carcinoma. Author (Year) Country Nr. of Patients Age (Years) Sex (M:F) Histological Subtype Study Design Tumor Side (L:R) Tumor Size (largest Diameter in cm) Tumor Location T1-Weighted Imaging Appearance T2-Weighted Imaging Appearance Contrast-Enhanced Imaging Appearance Tumor Composition and Growth Pattern Necrosis (nr. of Total) Hemorrhage (nr. of Total) Vascular Involvement (nr. of Total) Intra-Tumoral Fat Regional Lymph Node Involvement/Lymph node Metastases (nr. of Total) (Distant) Metastases Other Than Lymph Nodes Norena-Rengifo (2021) Col 1 12 1:0 RMC CR 1:0 NS central inter hetero, hypo hypovascular solid, infiltrative 1 NS absent absent renal hilum, para-aortic absent Koetter (2020) USA 1 16 0:1 P1 CR 1:0 17.3 exophytic NS NS hetero cystic-solid 1 NS absent NS peri-aortic, peri-caval absent Schaefer (2017) USA 1 14 1:0 MiT CR 0:1 5.2 upper pole homo hetero NS solid NS NS absent NS absent absent Okabe (2016) Japan 1 4 1:0 CHR CR 0:1 2.5 NS hypo hetero, hyper NS well defined 1 NS NS NS NS NS Zhou (2016) China 1 17 1:0 CC a CR B 0.2-2.0 a B hypo hypo strong multiple B a NS NS absent NS absent synchronous CNS hemangioblastoma and pancreatic neuroendocrine tumor Liu (2014) China 3 15-33 1:2 MiT CR 1:2 18; 6; 11 cortical hyper hetero, hypo hetero hypo solid (2); cystic (1); infiltrative (3) focal (2), central (1) inter-tumor (3) absent NS regional (2) absent Wang (2014) USA 7 b 13-33 3:4 MiT RS 4:3 3.5-22 medullary (2); medullary cortical (4); exophytic (1) iso (1); hyper (1); hetero (5) hypo (1); hyper (1); hetero (5) hetero: mild (1); moderate (4); marked rim/capsule (2) irregular (6); not irregular (1); well defined (4); ill defined (3) 7 6 3 NS regional (4), cervical (1) absent Koo (2013) South Korea 1 28 0:1 MiT RS 0:1 2.7 NS NS hetero, hyper NS well defined NS NS NS absent NS absent Dang (2012) USA 2 18; 31 1:1 MiT RS 0:1 B 8.9; 4.9 NS hetero, hyper NS limited hetero (1); NS (1) NS 1 2 absent NS absent absent Downey (2012) USA 2 c NS NS NS RS NS NS NS hetero, hyper (1); NS (1) NS hetero NS NS intra-tumoral (1) NS NS NS NS Kato (2011) Japan 1 18 1:0 MiT CR 0:1 4.1 peripheral NS hetero, hypo rim, central hyper delayed peripheral hyper, rim hyper well demarcated NS NS NS absent (hemosiderin) NS NS Blitman (2005) USA 6 (3) d 15-27 3:3 RMC RS 0:6 NS central NS NS hetero infiltrative, ill-defined margins 4 intra-tumoral (4); sub-capsular (1) ipsilateral renal vein (2); encasement vascular pedicle (3) NS cervical (6); retroperitoneal (5) e liver (2); lung (3) Adachi (2003) Japan 1 4 1:0 CCP CR 1:0 NS NS NS NS hyper walls complicated cyst NS cystic (1) NS NS absent absent a Multiple (cystic and) bilateral lesions in patient with von Hippel-Lindau disease; b MRI findings were not specified for each patient separately, so two adult patients (36 and 46 years old) could not be excluded from the overall MRI-data but were not included in the clinical characteristics data; c Total of nine children but only 2 with MRI scan and no specific details for separate patients (5:4 sex, mean age 12.9 years with a range of 7-17 years, mean maximum diameter 6.2cm (1.5-12.6)); d Imaging characteristics were not reported separately per patient, leaving no opportunity to extract MRI-specific information. Information displayed is for all 6 patients, based on CT and MRI; e Retroperitoneal adenopathy was heterogeneous and ranging in volume from small (n = 1) or moderate (n = 2) to extensive (n = 2).; M = male; F = female; L = left; R = right; Col = Colombia; RMC = renal medullary carcinoma; CHR = chromophobe RCC; CC = clear-cell RCC; P1 = papillary type 1 RCC; MiT = translocation-type RCC; CCP = clear-cell papillary type RCC; CR = case report; RS = retrospective cohort study; B = bilateral; homo = homogeneous; hetero = heterogeneous; hypo = hypo-intense; iso = iso-intense; hyper = hyper-intense; inter = intermediate; incr = increase; CNS = central nervous system; NS = not specified. cancers-15-01401-t004_Table 4 Table 4 Review of the literature focusing on MRI characteristics of translocation-type renal cell carcinoma (MiT-RCC). Author (Year) Country Nr. of Patients Age (Median Years, Range) Sex (M:F) Study Design Tumor Side (L:R) Tumor Size (Largest Diameter in cm) Tumor Location T1-Weighted Imaging Appearance T2-Weighted Imaging Appearance Contrast-Enhanced Imaging Appearance Diffusion Restriction (ADC value x10-3 mm2/s) Tumor Composition and Growth Pattern Necrosis (nr. of Total) Hemorrhage (nr. of Total) Vascular Involvement (nr. of Total) Intra-Tumoral Fat Regional Lymph Node involvement/Lymph Node Metastases (nr. of total) (Distant) Metastases Other than Lymph Nodes Tohi (2021) Japan 1 78 1:0 CR R 2.0 posterior iso hypo NS no restriction a well circumscribed, no capsule NS NS NS absent absent absent Dai (2019) China 16 47.4 (20-76) 9:7 RS 9:7 1.7-14.6 endophytic epicenter (14) hypo (2), iso (5), hyper (9) hetero (14); hypo (13), iso (6), hyper (2) hetero (7) hyper on DWI (b0/500) ((16) irregular (9), regular (7); complete capsule (11), incomplete capsule (5); solid (11), cystic (2), mixed (3) NS 5 2 absent 3 retroperitoneal space and liver (1); lung (1) Gong (2018) China 2 50; 45 1:1 CR 1:1 10.6; 5.2 upper pole (1); lower pole (1) iso (1), hypo (1) hypo (2) hetero (1) NS irregular (1) 1 NS absent NS 1 absent Chen (2017) China 2 46; 30 0:2 RS 0:2 7.8; NS NS hetero iso (2) hetero (2); hyper (1), hypo (1) hetero (2) relatively high signal on DWI (b0/800) (1) oval (17), irregular (4); solid (4), cystic (1), mixed (16) b NS NS v. renalis (1) NS 1 liver (1) Schaefer (2017) USA 1 14 1:0 CR 0:1 5.2 upper pole homo hetero NS NS solid NS NS absent NS absent absent Yu (2016) China 1 40 1:0 CR 0:1 12 NS iso hetero hypo-hyper NS NS well defined, irregular 1 patchy (1) absent NS 1 absent D'Antonio (2016) Italy 1 71 0:1 CR B c 12.0 NS hetero hyper NS NS poorly circumscribed (1) 1 1 NS NS NS NS Liu (2014) China 4 15-45 1:3 RS 1:3 4-18 cortical (4) hyper (4) hypo (3), hyper (1) Hypo NS infiltrative (4); solid (3); cystic (1) focal (3), center (1) inter-tumor (4) absent absent lymphadenopathy (3) absent Wang (2014) USA 9 13-46 3:6 RS 4:5 2-22 medullary (3), medullary cortical (4), exophytic (1), pelvis (1) iso (1), hyper (3), hetero (5) hypo (1), hyper (2), hetero (60) hetero: mild (1), moderate (6), marked rim/capsule (2) NS capsule (3); irregular (8); oval (1); well defined (5); ill defined (4) 8 7 4 NS regional (5), cervical (1) absent Koo (2013) South Korea 2 28; 71 0:2 RS 0:2 2.7; 4.6 NS NS hetero, hypo (2) NS NS well defined (2) NS intra-tumoral (1) NS absent NS absent Dang (2012) USA 2 18; 31 RS 0:1 B 8.9; 4.9 NS hetero, hyper NS limited hetero (1); NS (1) NS NS 1 2 absent NS absent absent Razek (2011) Egypt 4 5-67 d NS PS NS NS NS NS NS NS mean 1.50 +-0.97 (1.37-1.62) (b0/800) NS NS NS NS NS NS NS Kato (2011) Japan 1 18 1:0 CR 0:1 4.1 peripheral NS hetero, hypo rim, central hyper delayed peripheral hyper, rim hyper NS well demarcated NS NS NS absent (hemosiderin) NS NS a Tumor showed no restricted diffusion with a low signal; a fat-poor angiomyolipoma was in the differential diagnosis; b Total study consisted of 21 patients, of which MRI characteristics were reported for only 2 patients. The tumor composition, shape and growth pattern are, therefore, reported for the total population, mainly based on CT; c Bilateral tumor, with a right conventional RCC and a left MiT-RCC. Therefore, the characteristics of the MiT-RCC are presented in the table; d Study with 55 patients, of which 4 had an MiT-RCC. Age was presented for all patients. M = male; F = female; L = left; R = right; CR = case report; RS = retrospective cohort study; PS = prospective cohort study; B = bilateral; homo = homogeneous; hetero = heterogeneous; hypo = hypo-intense; iso = iso-intense; hyper = hyper-intense; inter = intermediate; ADC = apparent diffusion coefficient; DWI = diffusion weighted imaging; incr = increase; NS = not specified. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000564 | (1) Background: Transient increase in volume of vestibular schwannomas (VS) after stereotactic radiosurgery (SRS) is common and complicates differentiation between treatment-related changes (pseudoprogression, PP) and tumor recurrence (progressive disease, PD). (2) Methods: Patients with unilateral VS (n = 63) underwent single fraction robotic-guided SRS. Volume changes were classified according to existing RANO criteria. A new response type, PP, with a >20% transient increase in volume was defined and divided into early (within the first 12 months) and late (>12 months) occurrence. (3) Results: The median age was 56 (range: 20-82) years, the median initial tumor volume was 1.5 (range: 0.1-8.6) cm3. The median radiological and clinical follow-up time was 66 (range: 24-103) months. Partial response was observed in 36% (n = 23), stable disease in 35% (n = 22) and PP in 29% (n = 18) of patients. The latter occurred early (16%, n = 10) or late (13%, n = 8). Using these criteria, no case of PD was observed. (4) Conclusion: Any volume increase after SRS for vs. assumed to be PD turned out to be early or late PP. Therefore, we propose modifying RANO criteria for SRS of VS, which may affect the management of vs. during follow-up in favor of further observation. radiosurgery vestibular schwannoma Cyberknife(r) skull base tumors RANO criteria This research received no external funding. pmc1. Introduction Over the past decades, stereotactic radiosurgery (SRS) for small vestibular schwannoma (VS) has been established as an efficient and well-tolerated treatment, achieving high rates of tumor control and low risks for complications , and is therefore recommended in current guidelines . However, in post-treatment MR images, a transient volume expansion (TVE) occasionally combined with central low signal intensity is frequently observed. So far, no clear criteria exist which define this phenomenon with respect to the percentage of volume increase or the observed time course. This lack of a clear definition can also be seen in the different terms used in the literature, e.g., "temporary enlargement" , "tumor expansion" , "transient expansion" or simply "pseudoprogression" . Additionally, in many clinical situations it is unclear whether TVE reflects a treatment-related reaction or a true tumor recurrence. Particularly, in the case of Cyberknife(r) SRS, only scarce data exist compared to the well-examined gamma-knife series. Thus, we prospectively analyzed the time course and extent of the tumor volume changes after Cyberknife(r) SRS. 2. Methods 2.1. Study Population Due to the retrospective nature of this investigation, approval from the ethical committee of the University of Cologne was waived (reference number 16-476). In this single center retrospective cohort study, we evaluated all patients undergoing single session robotic-guided SRS by Cyberknife(r) for progressive vestibular schwannomas. Inclusion criteria were a minimum of at least two years of follow-up MR images after SRS. Clinical data were obtained through a review of the patients' electronic medical record, the Cyberknife(r) database, and available imaging studies. Data included patient demographic information, clinical history of neurological symptoms and parameters of the radiosurgical dose plan. In general, follow-up imaging was scheduled at 6 and 12 months after SRS and annually thereafter. Earlier imaging was occasionally obtained based on clinical judgment and course. 2.2. Radiosurgery Technique The radiosurgical technique for Cyberknife(r)-SRS has been described in detail in previous reports . In brief, the tumor and the adjacent critical structures (e.g., brainstem, cerebellum, trigeminal nerve) were outlined by an experienced neurosurgeon on contrast enhanced, T1-weighted MR images (Phillips, MR-Scanner 1.5 or 3 Tesla), which were obtained prior to SRS and registered to a stereotactic planning CT (1 mm slice thickness, Toshiba 16-slice multidetector CT). The software Multiplan v4.5 was used for treatment planning. The final irradiation plan was evaluated in an interdisciplinary consensus meeting between the stereotactic neurosurgeon, a radiation oncologist experienced in SRS and the medical physicist. For radiosurgery the patient was immobilized on the Cyberknife(r) treatment table (Accuray, Sunnyvale, California) by means of a custom-made aquaplast mask. 2.3. Tumor Imaging and Volumetric Analysis For volumetric analysis, we used axial T1-weighted, gadolinium-enhanced MRIs. Patients with follow-up MR images with slice thicknesses exceeding 3 mm were excluded from the study to minimize inaccuracy. On each image slice, the tumor cross-sectional area was calculated by delineation of the contrast-enhancing lesion via Osirix software (vs. 6, pixmeo). Tumor volume was subsequently calculated as the sum of the cross-sectional areas multiplied by the slice thickness, as previously described . Tumor volume on the T1-weighted, gadolinium-enhanced MRIs performed for the SRS planning was defined as baseline and served in each case as 100%. Each follow-up MRI until the last available scan was used for the volumetric analysis. At each follow-up time point, the percentage volume change (%DV) from the baseline was also calculated. For the purpose of the present analysis, we applied the RANO volumetric response criteria for meningioma to the vestibular schwannomas (Table 1). Progressive disease (PD) was assumed if %DV exceeded 40% with additional continuous growth beyond 48 months of observation according to previous research . Stable disease (SD) was stated if the tumor volume decreased by less than 65% DV or increased by less than 20% DV. A decrease of more than 65% DV was classified as partial response (PR). Complete response (CR) meant a complete remission of the entire tumor. Additionally, pseudoprogression (PP) was defined as a tumor volume increase of more than 20% DV, followed by a decrease, finally resulting in SD or PR. We differentiated between early (volume peak within the first 12 months after SRS) and late PP (volume peak more than 12 months after SRS) (Table 1). Furthermore, morphological changes in the tumor in terms of loss of central contrast enhancement were documented (Table 2). 2.4. Statistical Analysis Statistical analyses were performed using PRISM Ver. 8 (GraphPad software) and SPSS Vers. 25 (IBM). Descriptive statistics were used to characterize the patient population and treatment variables. Statistical significance was assumed if p < 0.05. The paired t-test was used to analyze volume changes over the observation period. A LogRank test or chi-square test was used to correlate onset of PP with new onset or deterioration of symptoms after SRS. In order to model the time course %DV(t) of the different types of volume changes in vs. after SRS, the following exponential equation was fitted to the data: %DV(t) = exp(-(A x t + B x t2)). 3. Results 3.1. Patients and Treatments We identified 106 patients who underwent Cyberknife(r) SRS for a progressive vestibular schwannoma between 2013 and 2016. Thirty-two patients were excluded due to a follow-up comprising less than two MR images. Eleven patients were excluded due to MR images with a slice thickness of more than 3 mm. Finally, we were able to include 63 patients with a median tumor volume of 1.5 cm3 (range 0.1-8.6 cm3, see Table 2). The median clinical and radiological follow-up period was 66 months (range 24-104 months). Three patients had subtotal surgical resection prior to radiosurgery. 3.2. Clinical Outcome Clinical data regarding symptoms prior to and after treatment were available for all 63 patients . Twenty patients suffered worsening of one or more symptoms after Cyberknife(r) SRS. In contrast, in thirteen patients symptoms improved. We did not find a correlation between the development of PP and aggravation of existing, or occurrence of new, symptoms after SRS (chi-square test, p = 0.339). Similarly, this was also true for new onset or deterioration of vertigo and balance disorders (LogRank, p = 0.636). Since only two patients in the collective suffered a loss of functional hearing after SRS, a correlation was statistically not feasible. 3.3. Volumetric and Tumor Characteristics Overall, 391 MR scans were volumetrically analyzed. The scans had a median slice thickness of 2.1 mm (range 1-3 mm). The mean tumor volume showed a significant decrease at all time points beyond 12 months after SRS . According to the predefined tumor response criteria regarding percentage volume change (%DV), four types of response were observed: partial response in 36% (n = 23) of cases; stable disease in 35% (n = 22) and pseudoprogression (PP) in 29% (n = 18) of cases. The latter was further divided into early PP and late PP . The median time to onset of early PP was six months (range: 4-10) after SRS and of late PP, 15.5 months (range: 4-35). The overall median time to peak of tumor volume enlargement after SRS was 18 months (range: 4-61) in the case of early PP and 36 months (range: 20-61) in the case of late PP. The median DV% at the peak of transient enlargement was 57% (range: 20-225%). Eight out of ten patients in the early PP group and all patients in the late PP group had a transient enlargement exceeding 40% DV. The median duration of tumor enlargement (time to complete resolution) until regression was 12.5 months (range: 5-82). The mean time to resolve from peak was 8.9 months +- 4.8 (range 5-20) in the case of early PP, and 19.7 months +- 10.4 (range: 6-33) in the case of late PP. The difference between both groups was statistically significant (p = 0.011). The tumor enlargement resolved completely after 60 months in 94% of patients. An illustrative case with late PP is shown in Figure 4. In 65% of the patients (n = 41), morphological changes with loss of contrast enhancement in the internal structure of the tumor were observed (Table 2). However, a correlation with early or late PP was not observed in the chi-square test (p = 0.378). Possible influencing factors such as patient characteristics or radiation parameters are compared in Table 3. There were no statistical differences between the cohort of patients with PP or the cohort with PR and SD. 4. Discussion Pseudoprogression is a well-known phenomenon after SRS of VS. Nevertheless, the definition and reported duration or frequency of this phenomenon vary in the literature. To shed some light on this topic, we conducted an extensive literature search that included studies that met the following criteria: (1) single fraction SRS; (2) median follow-up of at least 24 months; and (3) volumetric response analysis. Eleven series with a total of 982 patients (Table 4) matched these criteria. While most of the available series reported results from the Gammaknife (GK), our series is one of the first retrospective surveys with detailed analyses of volumetric responses in vs. after robotic-guided SRS using the Cyberknife(r). The incidence of pseudoprogression in our study was within the range of the other current studies (Table 4). Accordingly, pseudoprogression (PP) varied with an incidence of 4.7% to 77% and five-year tumor control varied between 87% and 100% . Some authors explain this disparity as being due to the large variation in observation periods in these studies . Apart from this factor, different types of tumor volume measurements (2D or 3D) and different MRI protocols may hinder the general comparability of the results. Comparable to our results, most patients seemed to develop a peak of volume increase between 6 months and 1 year after SRS (Table 4). However, in our series, seven patients had no early volume increase but developed a late transient tumor swelling, peaking at 3 or 4 years and then resolving very slowly. Other authors also observed this late peak around 36 months after SRS (Table 4). Of note, this is the crucial time point for some authors to define loss of tumor control. For instance, Delsanti et al. defined treatment failure as "a continuous growth for more than 3 years after radiosurgery". Mindermann et al. were even stricter and defined tumor progression in the case of a tumor volume increase of more than 20% after 24 months. Taking into account the present data, these definitions should be regarded with care, since all tumors still regressed spontaneously, even after 36 months or later. The case illustrated in Figure 4 demonstrates this observation and is in line with the results of Breshears et al. . In about 90% of cases with transient enlargement, they observe a transient tumor enlargement at 3.2 years after SRS. Furthermore, both Breshears et al. and Fouard et al. reported that the transient enlargement resolved completely in 90% of the patients after up to 6.9 years, which is quite similar to our results, where a remission was observed in 94% of the patients after five years. 4.1. Morphological Changes after SRS of Vestibular Schwannoma Besides volume changes, we also observed morphological changes such as transient loss of central contrast enhancement in 65% of the patients, which is within the range of 45 to 83% reported in the literature . The tumors showed loss of contrast only within the first year after SRS, but this phenomenon was not related to early or late PP, since in the patients with late PP, loss of contrast was also only apparent within the first year. These findings are similar to those of several other authors . For instance, in the study of Fouard et al. , loss of contrast enhancement was also found in progressive tumors. This observation suggests that these morphologic changes may only represent an early radiation effect and are not necessarily predictive for early or late PP or PD. So far, hardly any data are available about the mechanisms underlying morphological changes and PP. Iwai et al. investigated the pathological changes in a group of patients who had salvage surgery after a median interval to radiosurgery of 28 months. Half of the patients had histologic findings of intratumoral hemorrhage, presence of macrophages, myxoid degeneration or necrosis; all changes were regarded as radiation-induced. Several authors have assumed that tumor enlargement after SRS is also due to radiation-induced tumor necrosis or chronic intratumoral hemorrhage . Accordingly, here we suggest that during follow-up, the natural time course of tumor regression should be taken into account and surgical approaches should be limited to subtotal removal for functional preservation . 4.2. Risk Factors and Complications of Pseudoprogression Similar to the morphological changes after SRS, the pathological explanation for PP is not really known yet. However, vs. cells react with a combination of acute inflammation and vascular occlusion to radiation . Therefore, it is, on the one hand, not surprising that morphological changes and also temporary enlargement occur. On the other hand, it is all the more incomprehensible that this phenomenon only affects a subset of treated VS. To date, possible risk factors for the development of PP are largely unknown or discussed controversially, although prognostic assessment for PP occurrence would be important for post-interventional management. Regarding patient-, treatment-related factors such as age, tumor volume and radiation dose, we did not find any influencing factors. This goes in line with the majority of the available studies , with the exception of the study of Kim et al. , which showed that solid-type tumors had a higher probability of developing PP. This might be due to the fact that cystic tumors have less tissue that can respond to radiation. New insights can probably be gained from MRI-based radiomics. A study by Langenhuizen et al. detected PP in 38 out of 99 patients. A correlation between treatment-related factors and PP was not evident. However, textural features from MRI scans derived from the three-dimensional gray-level co-occurrence matrices (GLCM) showed a prognostic value for PP with a sensitivity of 0.82 and specificity of 0.69. These findings suggest that MRI-based tumor texture analysis provides information that could be used to predict PP and serve as a basis for individualized vs. treatment and FU strategy, especially in patients with large VS in which the phenomenon of PP is most relevant. In our series, we did not find a significant association between postradiosurgical clinical deterioration (cranial neuropathies, ataxia and hydrocephalus) and PP. This is in line with a majority of studies . Nevertheless, a relationship between onset of new cranial neuropathies or hydrocephalus was present in some studies within the first year after SRS . For example, Pollock et al. reported adverse effects in 20% of patients with tumor growth. However, the number of patients with AEs without tumor growth was not shown in that study. Nagano et al. saw an association with tumor volume increase of more than 30% associated with a significantly higher rate of cranial nerve impairment (e.g., facial hemispasm, facial weakness and facial dysesthesia). Aoyama et al. reported a rate of post-SRS hydrocephalus, with 11% of patients and a median of 11.3 months after the treatment . Tumor expansion and tumor size measuring 30 mm or greater are considered as risk factor for the new onset of hydrocephalus. In general, it is not so much the size but rather a higher protein concentration of cerebrospinal fluid after SRS which leads to a blockage of arachnoid granulations and is therefore seen as the main risk factor. This hypothesis also explains better why smaller tumors and cases without PP can also develop hydrocephalus. However, in many other series , the development of hydrocephalus plays no or only a minor role. 4.3. Development of RANO Criteria for vs. after SRS The phenomenon of pseudoprogression seems to be a quite typical phenomenon for vestibular schwannoma and is almost absent in other benign tumors such as meningiomas, hemangiopericytomas, pituitary adenomas or glomus tumors. However, in meningiomas, PP after SRS occurs in 5-11% and is typically associated with morphological changes indicating intratumoral necrosis . In VS, the situation is probably more complex, as in many tumors, PP occurs without morphological changes. Regarding benign brain tumors, RANO criteria are, so far, available only for meningioma , but not for VS. Thus, apart from arbitrarily defined criteria for distinguishing progression from pseudoprogression in VS, clear recommendations for response assessment would be helpful (Tab 3.). Even the well-established guidelines for the treatment of treatment of VS do not provide a consistent and clear recommendation, only the guideline of the International Society for Radiosurgery (ISRS, ) mentions that PP should be considered within the first three years after SRS. In order to fill this gap, we here propose additional response criteria to existing RANO criteria to overcome the ambiguities and different results of the various studies. These criteria use the well-established RANO terms of "stable disease" (SD), "partial response" (PR) and "progressive disease" (PD), as these allow a clearer understanding of the tumor response after treatment. The percentage volume changes of the respective category are taken from the RANO criteria for meningioma . In addition, the time interval in which volume changes occur should be considered. Therefore, we suggest the new category, "pseudoprogression" (PP), divided into early and late occurrence. Some studies on PP in vs. suggest >20% for the definition of PP. We follow this definition. On the other hand, the RANO criteria for meningiomas define PD as >25% volume increase . Since 20% and 25% are very close to each other and late PP can still be present after 36 months, it is difficult to separate both definitions. Especially, the temporal overlap makes it difficult to define a valid threshold. In our study, the best discriminatory power is 40%, because at >48 months all but three patients with late PP fell below this threshold. An increase of the threshold value, on the other hand, bears the risk of overlooking true progression. Furthermore, we follow the results of the study by Fouard et al. . They found that all patients with true progression had a volume increase of 37-43% at 36-48 months, so a value of 40% seems plausible. Further studies with a larger cohort might show that the current thresholds should be set differently after all. Furthermore, our proposed RANO criteria take the duration of tumor enlargement into account, and as such, are in line with our results and results from the existing literature (Table 4). Nevertheless, any response assessment criteria have the problem that they must tell stable from progressive disease also from the clinical viewpoint. As there were only mild clinical deteriorations associated with pseudoprogression in our cohort, salvage treatment was never necessary. If a clinical deterioration correlates with the imaging changes and appears to be uncontrollable, a new intervention may be required. However, surgery as a salvage treatment should always be considered carefully to avoid treating too early and to allow the tumors to regress spontaneously. 4.4. Conclusion and Practical Implications of Proposed RANO Criteria for VS The criteria proposed here are primarily intended to provide guidance for post-interventional management. Patients with vestibular schwannomas scheduled for radiosurgery should be informed about the phenomenon of early and late pseudoprogression. High-resolution contrast-enhanced T1-weighted MRI should be used as gold-standard for follow-up imaging, as this sequence allows the best visualization of the tumor. Annually, MRI exams are recommended during the first 5 years and can be reduced to an interval of 2-3 years if the tumor volume is stable or smaller than before treatment . A slice thickness of 2 mm or less should be used because volume measurement errors increase exponentially with the slice thickness . In clinical routine, the measurement of transverse tumor diameter may be adequate in most of the cases . Any case with suspected progressive disease in MRI controls should be examined in short intervals (e.g., 3-6 months) by thin-slice MRI and 3D volumetry of the tumor, since the latter can reduce measurement errors . Even in large tumors with or without onset of new symptoms, caution regarding salvage surgery or radiotherapy is required. 5. Conclusions In light of the newly developed response criteria and the inconsistency of the actual recommendations, the criteria for tumor progression often proposed in the literature (e.g., tumor volume increase of more than 10-20%, tumor growth after 3 years, no return to pretreatment volume after transient swelling) are, in our opinion, no longer valid and should not be used. Additionally, longer observation periods after SRS are strongly recommended based on these findings. Author Contributions Conceptualization, D.R. and M.I.R.; methodology, D.R. and A.H.; software, V.N. and A.H.; validation, V.N. and M.E.; formal analysis, D.R., M.E. and M.K.; investigation, D.R., B.S. and S.T.J.; data curation, M.K.; writing--original draft preparation, D.R.; writing--review and editing, E.C., C.B., S.T.J.,V.N., M.K. and M.I.R.; supervision, M.I.R. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Ethics Committee of the University hospital of Cologne (16-476, date of approval: 20 November 2016). Informed Consent Statement The ethic committee of the University Hospital of Cologne approved the study protocol (Identity: Az 16-476). Due to the retrospective character, the ethic committee waived the need for informed consent. Nevertheless, informed consent was obtained from all subjects involved in the study. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Symptoms prior to and post Cyberknife(r) SRS in a collective of 63 patients with progressive VS. Figure 2 Mean tumor volume during the follow-up period. Twelve months after SRS, the mean tumor volume of the collective was significantly decreased (marked with *) compared to the baseline (6 months: p = 0.91, 12 months: p < 0.0001, 24 months: p < 0.0001, 36 months: p = 0.018, 48 months: p = 0.021, 60 months: p < 0.0001, >72 months: p < 0.0001). Figure 3 Comparison of mean %DV during follow-up after SRS of VS. (A): Partial response; (B): Stable disease; (C): Early pseudoprogression; (D): Late pseudoprogression. Figure 4 Illustrative case of a 56-year-old woman with a vestibular schwannoma that showed an increase in volume and was classified as late pseudoprogression. At the baseline, Cyberknife(r) SRS with a radiation dose of 13 Gy was applied. During follow-up, a continuous volume increase could be observed with a maximum of %DV = 225 after 48 months. Afterwards, until 72 months, the tumor volume regressed almost completely to baseline. cancers-15-01496-t001_Table 1 Table 1 Proposed RANO criteria for assessing tumor response after SRS of VS. Criterion CR Complete Response PR Partial Response SD Stable Disease PP Pseudoprogression PD Progressive Disease (1) Target lesion None >=65% decrease in volume relative to baseline <65% decrease relative to baseline but <20% increase in volume relative to nadir >20% increase in volume relative to baseline followed by decrease within 24-36 months >=40% increase in volume relative to baseline and/or change of Koos grade I or II to III or IV (2) Onset of volume increase after SRS n.a. n.a. n.a. Early peak: <12 months Late peak: >12 months >36-48 months (3) Clinical status (besides hearing deterioration and vertigo) Stable or improved Stable or improved Stable or improved Stable or improved Stable or deteriorated Requirement for response All or (1) alone All or (1) alone All All All cancers-15-01496-t002_Table 2 Table 2 Patient and treatment characteristics, clinical data and response criteria of the observed cohort after Cyberknife(r) SRS. Unless mentioned otherwise, data are presented as mean with SD and range in brackets. Patient Characteristics Total no. of patients 63 Gender (m:f) 25:38 No. of Koos-Grade KoosI: 6 KoosII: 45 KoosIII: 11 KoosIV: 1 Age (years) 56 +- 14 (range: 20-82) Tumor volume (cm3) 1.5 +- 1.4 (range: 0.1-8.6) Median radiological and clinical FU (months) 66 (range: 24-103) Radiation Parameters Marginal dose (Gy) 13.0 +- 0.2 (range: 12-13) Dose prescription, isodose (%) 80 +- 3.5 (range: 65-81) Coverage (%) 99.6 +- 0.9 (range: 94.7-100) Dmax 16.25 +- 0.7 (range: 15-18.6) Dmean 14.89 +- 0.3 (range: 14-15.8) Dmin 12.94 +- 0.4 (range: 11-13.2) nCI 1.19 +- 0.07 (range: 1.08-1.44) Response/Pseudoresponse Criteria Loss of central contrast enhancement 41 (65%) CR--complete response 0 ePP--early pseudoprogression 10 (16%) lPP--late pseudoprogression 8 (13%) SD--stable disease 22 (35%) PR--partial response 23 (36%) PD--progressive disease 0 cancers-15-01496-t003_Table 3 Table 3 The comparison of pseudoprogression (PP) and partial response/stable disease (PR+SD) of the observed cohort after Cyberknife(r) SRS showed no significant difference (I > 0.05) with regard to patient and treatment characteristics. Patient Characteristics PR + SD PP p-Value No. of patients 45 18 Age (years) 57 +- 14.3 54 +- 13.2 p = 0.731 Tumor volume (cm3) 1.62 +- 1.5 1.15 +- 0.7 p = 0.057 follow-up (months) 61.5 +- 21.6 64.3 +- 21.1 p = 0.083 Radiation Parameters Marginal dose (Gy) 13 13 Dose prescription, isodose (%) 78.02 +- 3.7 78.9 +- 2.7 p = 0.082 Coverage (%) 99.36 +- 1.1 98.8 +- 1.4 p = 0.087 Dmax 16.6 +- 0.7 16.5 +- 0.6 p = 0.223 Dmean 14.9 +- 0.2 14.8 +- 0.3 p = 0.111 Dmin 12.8 +- 0.4 12.8 +- 0.2 p = 0.158 nCI 1.17 +- 0.63 1.22 +- 0.09 p = 0.051 cancers-15-01496-t004_Table 4 Table 4 Summary of pseudoprogression after single fraction SRS of vs. evaluated by volumetric analysis in former studies. Incidence and time course of pseudoprogression. Study n Mean FU (months) Radiation Technique Tumor Volume (ml) Incidence of PP (%) Definition of PP Mean Volume Increase (%) Median Yime to Peak for PP (months) Late Peak (% of Collective @ Median Ttime to Peak) Duration of PP Yu, 2000 91 22 GK 63 NA 20 6 NA Nakamura, 2000 78 34 GK 0.6 41 Volume increase of more than 20% NA 12 6.4% @ 24-36 months 24 mo Nagano, 2008, 2010 87 65 GK 2.5 77 Volume increase > 10% 58 8.6 NA 90% resolved after 5 yr Van de Langenberg, 2011 17 40 LINAC 2.09 54 Volume increase >19.7% NA 5 (3-17) NA 15 (8-27) mo Hayhurst, 2012 75 29 GK 1.7 23 Volume increase > 10% 23 NA NA NA Kim, 2013 60 42 GK 0.34 47 Volume increase > 10% within a year NA NA NA NA Mindermann, 2014 235 62 GK 1.85 NA Volume increase > 20% within 24 months NA 6-18 NA 12-18 Matsuo, 2015 44 165 LINAC 2.38 54.5 Volume increase > 20% within 24 months 88 9 7.1% @ 39.6months Kim, 2017 235 34 GK 2.2 18 Volume increase > 20% within a year NA 7 (4-55) NA NA Breshears, 2019 18 49 GK 0.74 42 Volume increase at any time during FU followed by reduction 49 12.5 10% @36-48 months 28.8 mo, 90% resolved at 6.9 yr Fouard, 2021 42 83 GK 0.69 63.5 Volume increase > 13% at any time during FU followed by reduction 64 6-12 mo 17% @ 36-48 months 24 mo, Our study 63 66 CK 1.5 29 Volume increase > 20% at any time during FU followed by reduction 57 18 12.7% @ 36 months 12.5 mo, 90% resolved after 4 yr Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000565 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12050926 foods-12-00926 Article Antimelanogenic Effects of Curcumin and Its Dimethoxy Derivatives: Mechanistic Investigation Using B16F10 Melanoma Cells and Zebrafish (Danio rerio) Embryos Jeon Hwang-Ju Conceptualization Methodology Software Validation Formal analysis Investigation Data curation Writing - original draft Visualization 1+ Kim Kyeongnam Validation Formal analysis Investigation Data curation 23+ Kim Chaeeun Validation Formal analysis Investigation Data curation 34 Lee Sung-Eun Resources Writing - original draft Writing - review & editing Supervision 234* Wang Yun Academic Editor 1 Red River Research Station, Louisiana State University Agricultural Center, Bossier City, LA 71112, USA 2 Institute of Quality and Safety Evaluation of Agricultural Products, Kyungpook National University, Daegu 41566, Republic of Korea 3 Department of Applied Biosciences, Kyungpook National University, Daegu 41566, Republic of Korea 4 Department of Integrative Biology, Kyungpook National University, Daegu 41566, Republic of Korea * Correspondence: [email protected]; Tel.: +82-53-950-7768 + These authors contributed equally to this work. 22 2 2023 3 2023 12 5 92616 1 2023 09 2 2023 15 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Regulation of melanin production via the MC1R signaling pathway is a protective mechanism of the skin of living organisms against exposure to ultraviolet rays. The discovery of human skin-whitening agents has been one of the most intense pursuits of the cosmetic industry. The MC1R signaling pathway is activated by its agonist, alpha-melanocyte stimulating hormone (a-MSH), and mainly regulates melanogenesis. Here, we evaluated the antimelanogenic activities of curcumin (CUR) and its two derivatives, dimethoxycurcumin (DMC) and bisdemethoxycurcumin (BDMC), in B16F10 mouse melanoma cells and zebrafish embryos. CUR and BDMC reduced the a-MSH-induced melanin production in B16F10 cells and also downregulated the expression of the melanin-production-related genes Tyr, Mitf, Trp-1, and Trp-2. Moreover, the biological activity of these two compounds against melanogenesis was confirmed in in vivo experiments using zebrafish embryos. However, the highest concentration of CUR (5 mM) resulted in slight malformations in zebrafish embryos, as indicated by acute toxicity tests. In contrast, DMC did not show any biological activity in vitro or in vivo. Conclusively, BDMC is a strong candidate as a skin-whitening agent. curcumin derivative melanogenesis MC1R signaling pathway zebrafish melanin production whitening agent Ministry of EnvironmentThis research was supported by a project to train professional personnel in biological materials by the Ministry of Environment. pmc1. Introduction Natural products have been widely used in various fields, including biological, pharmaceutical, and nutraceutical areas, with the intention of obtaining and exploiting their beneficial effects, such as anticancer, anti-inflammatory, antifungal, antibacterial, and antimelanogenesis functions. These natural products are found in various organisms in Nature, including microorganisms, animals, and plants . Plants are the main source of natural products, and as such the usage of plant extracts in the pharmaceutical and cosmetic industries has been increasing, partly owing to the negative perception associated with animal-origin extracts . Numerous research groups are exploring the beneficial effects of new bioactive plant extracts or isolated plant-derived chemical compounds. Their studies attempt to demonstrate the biological mechanisms of valuable candidates, and suggest their potential usage. Investigating the underlying biological mechanisms of the active compounds of natural products is crucial for confirming the lack of harmful side effects in humans. In order to investigate the molecular mechanisms and potential toxicity of natural products, both in vitro and in vivo methods are employed, including cell cultures and animal experiments . However, recently introduced ethics regarding the use of animals in research have enforced the pursuit of alternative animal models, such as small fish . The zebrafish is an animal that adheres well to the requirements of alternative experimental animal models. Their genome is very similar to that of humans, they have a high reproduction rate, and their early development is achieved within 96 h after fertilization. Moreover, because of the clear chorion of the embryo, the entire developmental process can be easily observed . As a result of these advantages, zebrafish have been used in various fields of biological research, including those of natural products. Recently, the global cosmetic market has seen large growth . In addition, whitening products are one of the most popular products in the cosmetics industry. Based on this, the exploration of new candidates for whitening agents is spotlighted in this field. Whitening agents play an important role in the cosmetic industry. Many studies have attempted to discover new agents, such as arbutin and kojic acid, with improved skin-whitening effects . The main focus of such studies is to suppress the activation of the melanocyte-specific melanocortin-1 receptor-tyrosinase (MC1R-TYR) signaling pathway in melanocytes. The MC1R signaling pathway is the primary pathway regulating melanin production . MC1R is activated by the alpha-melanocyte-stimulating hormone (a-MSH), adrenocorticotropic hormone (ACTH), and ACTH-secreting phaeochromocytoma (ASP) ligands . Once activated, the a-MSH-induced cAMP pathway stimulates the activation of TYR . The binding of a-MSH to MC1R on the cell membrane activates adenylate cyclase and leads to an increase in the levels of intracellular cAMP . Subsequently, cAMP-dependent protein kinase A (PKA) phosphorylates cAMP response element-binding protein (CREB), leading in turn to the phosphorylation of MITF, which acts as a DNA-binding transcription factor of melanin-production-related genes, including TYR: a rate-limiting enzyme of melanogenesis. In particular, MITF regulates the expression of TYRP-1 and TYRP-2 proteins by binding to M-box in the tyrosinase distal element of these genes . Curcumin (CUR), dimethoxycurcumin (DMC), and bisdemethoxycurcumin (BDMC) are active polyphenolic compounds of turmeric (Curcuma longa), a member of the ginger family, called curcuminoids. Although turmeric originates from India, it has been grown in many other places, including China and Southeast Asia . Owing to its special taste and flavor, turmeric has been used over the past few centuries as a coloring agent, insect repellent, and antimicrobial agent. Because of the increased interest in natural products in the past several decades, turmeric has also been used as a nutraceutical, dietary supplement, and functional food . Curcuminoids have been well known to exhibit a broad spectrum of bioactivities, including antioxidant, anti-inflammatory, antimicrobial, antifungal, and anticancer effects . These beneficial effects have been established both in vitro and in vivo by many studies using cell lines and mouse models, respectively . In this study, we evaluated the antimelanogenic effects of curcuminoids in vitro and in vivo using a mouse melanoma cell line (B16F10) and zebrafish embryo model, respectively, to explore new whitening agent candidates for use in cosmetics. Concomitantly, we performed acute toxicity tests to ensure the safety of using these compounds in vertebrates. 2. Materials and Methods 2.1. Chemicals and Reagents CUR, DMC, BDMC, a-melanocyte stimulating hormone (a-MSH), and kojic acid were purchased from Sigma-Aldrich (St. Louis, MO, USA). All other reagents used in this study were of molecular biology grade. 2.2. Cell Culture The B16F10 mouse melanocarcinoma cell line was purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). Culturing cells were performed as described before . Briefly, cells were cultured in Dulbecco's modified Eagle's medium (DMEM, GE Healthcare, Chicago, IL, USA) containing 10% fetal bovine serum (Corning, Corning, NY, USA) and 1 % penicillin/streptomycin (GE Healthcare, Chicago, IL, USA). Cells were cultured in a CO2 incubator (BINDER, Tuttlingen, Germany) under conditions of humidified atmosphere with 5% CO2 at 37 degC. Cells used in experiments were kept at as low a passage number as possible. 2.3. Cell Viability Assay The proliferation of B16F10 cells was evaluated using the CellTiter 96 Aqueous One Solution Cell Proliferation Assay Kit (Promega, Madison, WI, USA) according to the manufacturer's instructions . Specifically, 1 x 103 cells were seeded on a 96-well plate and cultured for 24 h to recover. After recovery, the medium was changed with fresh medium containing CUR (5, 10, 20, 40, and 80 mM), DMC (5, 10, 20, 40, and 80 mM), or BDMC (5, 10, 20, 40, and 80 mM). Treated cells were incubated for an additional 48 h and then 20 mL of MTS solution per 100 mL media was added to each well. After an additional incubation for 4 h, the plate was placed at 25 degC for 30 min, followed by determining absorbance at 490 nm using the Multiskan GO microplate spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Data were normalized to the absorbance value of the control and represented as percentage ratio. 2.4. Determination of Melanin Contents Total melanin contents were determined according to the method previously reported . Briefly, B16F10 cells were pretreated with 200 nM a-MSH to induce melanogenesis, and then incubated for 72 h with phenol-red-free DMEM supplemented with 10% FBS and antibiotics. Kojic acid (200 mM), CUR (5 and 10 mM), DMC (5 and 10 mM), or BDMC (5 and 10 mM) was added to the media. After incubation, media were collected for measuring extracellular melanin contents, whereas 200 mL RIPA lysis buffer was added to harvested cells for determination of intracellular melanin contents. To isolate intracellular melanin, lysed cells were centrifuged at 4 degC and 13,000x g for 15 min and the pellet was resolved in 10% DMSO solution containing 1 N NaOH at 95 degC for 2 h. Melanin contents were determined by measuring absorbance at 470 nm using a spectrophotometer. For normalization of samples, protein concentration was determined by the BCA method as previously described . 2.5. RNA Isolation and qRT-PCR Isolation of total RNA and determination of the level of mRNA was performed using the QuantStudio 3 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) according to a previously described method . Briefly, total RNA was extracted from samples using the Trizol reagent (Ambion, Austin, TX, USA), and the absorbance at 260 nm, 260/280 nm, and 260/230 nm was measured for quantity and quality check of extracted RNA using the Multiskan GO microplate reader. Subsequently, 5 mg cDNA was synthesized from extracted total RNA using the Maxima first strand cDNA synthesis kit (Thermo Fisher Scientific, Waltham, MA, USA). The obtained cDNA was used as the template for the determination of the levels of Mitf, Tyr, Trp-1, and Trp-2 mRNAs. Accordingly, qRT-PCR was performed using the Luna Universal qPCR Master Mix (New England Biolabs, Ipswich, MA, USA), according to the manufacturer's instructions. The primers used in qRT-PCR are listed in Table S1. All data were normalized to the level of Gapdh mRNA. 2.6. Intracellular Tyrosinase Activity Assay The activity of intracellular tyrosinase was determined according to a previously reported method . Briefly, a-MSH-pretreated B16F10 cells were incubated with various concentrations of test compounds (200 nM kojic acid; 5 and 10 mM CUR, DCM, or BDCM) for 72 h. Cells were then harvested in a lysis buffer containing proteinase inhibitors and centrifuged at 13,000x g and 4 degC for 15 min to collect the supernatant. For measuring the oxidation rate of L-DOPA, 20 mL of collected sample and 80 mL of 40 mM L-DOPA were mixed and placed in each well of a 96-well plate. Samples were incubated at 37 degC for 2 h and their absorbance was measured at 475 nm using a microplate reader. Absorbance values were normalized to the protein concentration of each sample. 2.7. Zebrafish Embryo Test AB strain zebrafish were generously provided by Prof. Tae-Lin Huh from the Korean Zebrafish Resource Bank, Kyungpook National University (Daegu, Republic of Korea) and kept in the laboratory in an automatic flow-through system under conditions of 26 +- 1 degC and 8 h:16 h day and night ratio for more than eight generations. Fish were cared for according to the modified ZFIN general fish care method previously reported . The night before the experiment, 10 pairs of fish were mated in order to obtain embryos. Early in the morning, embryos were collected, and healthy embryos (>80% fertilization ratio) were transferred to E3 media. Fifteen healthy embryos were placed in each well of a 6-well plate, followed by the addition at the shield stage of E3 media containing kojic acid (8 mM) and test compounds (CUR, DMC, or BDMC; 1.25, 2.5, and 5 mM each). After treatment, the developmental stage, abnormal development, and phenotype were checked every 24 h until 72 hpf. The phenotype of embryos was photographed at 72 hpf using an Olympus BX53 microscope equipped with a DP80 CCD camera (Olympus, Waltham, MA, USA). After documentation, embryos were homogenized with CETi lysis buffer (Translab, Daejeon, Republic of Korea), and melanin concentration was determined by measuring absorbance at 475 nm using a microplate reader. 2.8. Statistical Analysis All statistical analyses were performed using GraphPad Prism 8.0 (GraphPad Software, San Diego, CA, USA). One-way ANOVA with Tukey's post hoc test was performed for multiple comparisons. All data are presented as the mean +- standard deviation (SD). A p < 0.05 was considered statistically significant. 3. Results 3.1. Cytotoxic Effect of Curcumin Derivatives in B16F10 Cells To set the dose range of CUR, DCM, and BDCM, we performed a Tetrazolium Assay (MTS assay). The inhibition ratio was calculated and represented as a percent compared with the control . We found that the inhibition ratio of the proliferation of B16F10 cells treated with 10 mM of CUR, DCM, and BDCM was 12.31 (+-7.90)%, 18.03 (+-16.70)%, and 15.14 (+-6.74)%, respectively. We also observed that at concentrations of curcuminoids above 20 mM, the percent inhibition ratios of treated cells were above 20% (28.13, 23.26, and 22.38% for cells treated with 20 mM CUR, DCM, and BDCM, respectively). Considering this result, we set the highest dose of curcuminoids to 10 mM when we evaluated their antimelanogenic effect in B16F10 mouse melanoma cells. 3.2. Curcumin and Bisdimethoxycurcumin Reduced Total Melanin Contents in B16F10 Mouse Melanoma Cells We detected that both the extracellular and intracellular melanin contents were elevated in a-MSH-induced B16F10 cells, exhibiting a 1. 2.75-fold and 2. 3.29-fold increase for extracellular and intracellular melanin content, respectively, compared with those in the control . Overall, treatment with a-MSH induced a 2. 3.04-fold increase in total melanin contents in B16F10 cells . We also observed that the stimulation of melanin production in B16F10 cells was suppressed by 1. 1.98-fold after the addition of 5 mM CUR or BDMC, respectively, in comparison to that in the control . We further noticed that at its highest concentration (10 mM), CUR reduced both the extracellular and intracellular melanin contents by 1. 1.31-fold in comparison to those in the control . Likewise, the amount of extracellular and intracellular melanin was 1. 1.28-fold higher after the administration of 10 mM BDMC than that in the control . In contrast, we observed that the addition of 5 and 10 mM DMC increased the production of total melanin in a dose-dependent manner, resulting in 2. 3.43-fold higher total melanin contents, respectively, than those in the control. Moreover, we found that both 10 mM CUR and DMC inhibited the a-MSH-induced increase in tyrosinase activity, keeping it to 1. 1.27-fold higher than that in the control . 3.3. Curcumin Derivatives Downregulated the mRNA Levels of Melanin-Production-Related Genes We subsequently evaluated the antimelanogenic properties of CUR, DMC, and BDMC at the molecular level. We detected that the a-MSH-induced increases in the mRNA levels of Mitf, Tyr, Trp-1, and Trp-2 in B16F10 cells were dramatically reduced following treatment with CUR, DMC, or BDMC . We specifically found that the a-MSH-induced increase in the level of Mitf mRNA was reduced 0.10-, 0.19-, and 0.08-fold following administration of 10 mM CUR, DMC, or BDMC, respectively, compared with that in the control . Likewise, we observed that the expression of Tyr and Trp-1 was reduced in B16F10 cells following administration of CUR, DMC, or BDMC (5 and 10 mM) in a dose-dependent manner . In contrast, the expression of Trp-2 was unaltered or slightly reduced following treatment with a low or high dose of curcuminoids, respectively . In particular, the expression of Trp-2 was reduced by 0.28-, 0.53-, and 0.43-fold following 10 mM CUR, DMC, or BDMC administration, respectively, in comparison to that in the control . 3.4. Toxicity of Curcumin Derivatives in Early-Stage Zebrafish Embryos Before evaluating the antimelanogenic effect of curcuminoids in early-stage zebrafish embryos, we performed acute toxicity tests to set the range of treatment dose. We accordingly observed various malformations in zebrafish embryos treated with 5 mM CUR. More specifically, we found that the spine was curved, and the tail tip was also curved to one side in embryos at 72 hpf . The survival rate was also slightly decreased at 72 hpf . However, these effects disappeared at concentrations of curcuminoids below 2.5 mM. Interestingly, we observed neither abnormal development nor death in zebrafish treated with DMC or BDMC . 3.5. Curcumin Derivatives Reduced Melanin Contents in Zebrafish Embryos We observed the pigmentation of zebrafish as dark dots in the overhead view of zebrafish embryos at 72 h post fertilization (hpf) . These dots are located between the top of the embryo head and its trunk. Kojic acid, a known antimelanogenic compound, reduces the synthesis of melanin in developing zebrafish embryos, and was thus used here at a concentration of 8 mM as positive control . We determined that kojic acid inhibited the production of melanin by 41% compared with that in the control. We also found that CUR and BDMC inhibited the formation of dark dots on the surface of zebrafish skin in a dose-dependent manner . More specifically, we detected that administration of 5 mM CUR (highest tested concentration) inhibited the production of melanin in zebrafish by 50% . Finally, we noticed that although BDMC showed the highest bioactivity in in vitro experiments, it was not the most effective compound in in vivo experiments. In particular, we observed that the relative amount of melanin in zebrafish embryos exposed to 1.25, 2.5, and 5 mM BDMC was 0.92-, 0.84-, and 0.53-fold lower, respectively, than that in the control . 4. Discussion In this study, we evaluated the potential antimelanogenic effects of CUR, DMC, and BDMC. Among the tested curcuminoids, CUR and BDMC tended to weaken the a-MSH-induced production of melanin. These two compounds reduced not only the intracellular, but also the extracellular melanin contents. Moreover, CUR suppressed the a-MSH-induced increase in tyrosinase activity in B16F10 cells . Notably, the inhibitory effect of CUR was stronger than that of another well-known whitening agent, kojic acid, which has already been reported in many previous studies . Although the pharmacokinetics study of curcuminoids reported poor bioavailability of these chemicals , skin-whitening reagents were not designed for oral administration, but topical cosmetics. In order to overcome this limitation, researchers tried to use curcumin in encapsulated form, and this advanced formulation was very promising . In a previous study, 10 mM CUR was reported to suppress tyrosinase activity by 47.85% . This study's reported suppression ratio was slightly lower than that in our study; however, their experimental standard deviation was greater. Despite small differences in the inhibition ratio, other studies have also reported the inhibitory effect of CUR against tyrosinase activity and melanin production in B16F10 cells and human melanocytes . Based on these studies, CUR has strong antimelanogenic activity and can thus be considered a strong candidate whitening agent. In contrast to its antityrosinase activity , the antimelanogenic effect of BDMC has not been reported previously. Nonetheless, our in vitro results suggested that BDMC has great potential as a skin-whitening agent. Our quantitative analysis of the mRNA levels of melanin-production-related genes in B16F10 cells revealed that the a-MSH-induced upregulation in the expression of these genes was ameliorated following administration of these three curcumin derivatives. CUR-induced reduction in the level of key proteins in melanogenesis, including MITF, tyrosinase, TRP-1, and TRP-2 was previously reported , but changes in the levels of mRNA were not reported. As reported in a previous study, CUR downregulated the expression of members of the MITF-tyrosinase signaling pathway at the protein level . Likewise, our study suggested that the mRNA levels of melanogenesis-related genes were also reduced following treatment with CUR. Combining these findings and those of previous studies, we concluded that CUR inhibits melanogenesis by suppressing the expression of melanogenesis-related genes regulated by the MC1R signaling pathway, including MITF, TYR, TRP-1, and TRP-2 . Administration of DMC downregulated the expression of melanogenesis-related genes; however, the degree of inhibition was lower than that of CUR and BDMC. In addition, DMC did not lead to a reduction in the total melanin contents in B16F10 cells. To explain these results, further studies are needed. Similar to CUR, BDMC also ameliorated the a-MSH-induced activation of the melanogenesis signaling pathway. To date, BDMC was only known to inhibit tyrosinase , which is obviously an enzyme that plays a very important role in the synthesis of melanin; however, melanogenesis is much more tightly regulated by the melanogenesis signaling pathway . Hence, we approached it here from a diverse perspective and found that BDMC reduced not only the production of melanin in B16F10 melanoma cells, but also the mRNA level of melanin-production-regulating genes including Mitf, Tyr, Trp-1, and Trp-2. This finding suggested that administration of CUR or BDMC ameliorated the MC1R signaling pathway-induced upregulated expression of melanin-production-related genes . In vivo evaluation of the biological activity of a natural compound is one of the most important steps in exploring the potential medicinal use of natural products, because sometimes the biological activity of a compound evaluated in an in vitro system might be different from that exhibited in an in vivo system . In addition, the rapid exploration of new candidates is important in the screening phase to reduce time and cost. Therefore, many research groups have been trying to develop faster and more effective in vivo evaluation systems . Many studies, including our own, have shown the effectiveness of the zebrafish embryo model for screening compounds with inhibitory activity against melanogenesis . The main biosynthetic pathways in zebrafish are very similar to those in humans . In addition, the targets of well-known inhibitors of melanogenesis were shown to be similar, and their efficacy has already been confirmed in both the zebrafish embryo model and humans . In this study, we screened three curcumin derivatives for their antimelanogenic properties using zebrafish embryos. None of the tested curcuminoids showed acute toxicity at concentrations below 5 mM in 96 hpf zebrafish. However, at a concentration of 5 mM, both abnormal development of zebrafish embryos and lethality were confirmed. Notably, 5 mM CUR was shown to result in the tail bending of zebrafish embryos. A previous study also reported the malformation of zebrafish embryos in the presence of CUR . In that study, the toxic effects of CUR were reported to occur in a dose-dependent manner , with a minimum effective concentration of 5 mM, which is consistent with our findings . Considering the malformations observed in zebrafish embryos, we concluded that CUR is cytotoxic in zebrafish embryos during early development. Both CUR and BDMC exhibited outstanding inhibition properties in the in vivo test. Importantly, the inhibitory effect of both curcuminoids was stronger than that of kojic acid, a compound used as positive control in our study and one of the most popular whitening agents in the cosmetic industry . In addition, the concentration of kojic acid was 8 mM, which was much higher than that of CUR (5 mM) or BDMC (5 mM). Another compound frequently used as positive control in antimelanogenic studies, arbutin, has also been used at a concentration range of 10 to 20 mM , with both compounds, arbutin and kojic acid, being reported to have similar activity and working concentrations . Compared with these positive control compounds, the working concentration of our two candidate agents was much lower (5 mM for curcuminoids vs. 10 mM for the two positive controls) in the in vivo investigation of melanin production. Hence, these findings suggested that our curcuminoid candidate compounds have great potential to be used as whitening agents in cosmetics. 5. Conclusions In conclusion, as all these results demonstrate, CUR and BDMC have outstanding inhibition effects on melanin production. Considering the toxic effect of CUR on zebrafish embryos during early-stage development, BDMC may be used as a whitening agent for melanogenesis by inhibiting melanin-production-related proteins TYR, TRP-1, and TRP-2 in both in vitro and in vivo. Supplementary Materials The following supporting information can be downloaded at: Table S1: Primer list. Click here for additional data file. Author Contributions H.-J.J.: conceptualization, methodology, software, validation, formal analysis, investigation, data curation, writing--original draft preparation, writing--review and editing; K.K.: validation, formal analysis, investigation, data curation, writing--original draft preparation; C.K.: validation, formal analysis, investigation, data curation; S.-E.L.: resources, supervision, writing--original draft preparation, writing--review and editing. All authors have read and agreed to the published version of the manuscript. Data Availability Statement The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. Conflicts of Interest The authors have no conflict of interest to disclose. Figure 1 Molecular structures and cytotoxicity of curcumin (CUR), dimethoxycurcumin (DMC), and bisdemethoxycurcumin (BDMC). Viability of B16F10 cells with the treatment of 0, 5, 10, 20, 40, 80 mM of CUR, DMC, and BDMC. Data are presented as mean +- standard deviation (SD). Significant differences are expressed by the symbols **, ***, and **** as ** p < 0.01, *** p < 0.001, and **** p < 0.0001 compared with control. Figure 2 Inhibitory effect of curcumin (CUR), dimethoxycurcumin (DMC), and bisdemethoxycurcumin (BDMC) on alpha-melanocyte stimulating hormone (a-MSH) inducing melanogenesis in B16F10 mouse melanoma cells. The color changes of phenol-free media by accumulation of melanin at 48 h and determined intracellular, extracellular, total melanin contents, and tyrosinase activity after treatment of stimuli (a-MSH), positive control (kojic acid), CUR, DMC, and BDMC, respectively. Data are presented as mean +- standard deviation (SD). Significant differences are expressed by the symbols ****, #, ###, and #### as **** p < 0.0001 compared with control and # p < 0.05, ### p < 0.001, and #### p < 0.0001 compared with the a-MSH-treated group, respectively. Figure 3 Effect of curcumin, dimethoxycurcumin, and bisdemethoxycurcumin on mRNA level of melanin-biosynthesis-related genes in B16F10 cell line using RT-qPCR. a-MSH, a-melanocyte stimulating hormone; Mitf, microphthalmia-associated transcription factor; Tyr, tyrosinase; Trp-1, tyrosinase-related protein 1; Trp-2, tyrosinase-related protein 2. Data are shown as means +- standard deviations (SDs). Significant differences are expressed by the symbols **, ***, ****, #, ##, ###, and #### as ** p < 0.01, *** p < 0.001, and *** p < 0.0001 compared with control, respectively. # p < 0.05, ## p < 0.01, ### p < 0.001, and #### p < 0.0001 compared with the a-MSH-treated group, respectively. Figure 4 Acute toxicity of curcumin, dimethoxycurcumin and bisdemethoxycurcumin on early-stage development of zebrafish embryo. (a) Images of lateral and dorsal view of zebrafish embryo at 72 h after exposure to 8 mM of kojic acid (KA), various concentrations of curcumin (CUR), dimethoxycurcumin (DMC), and bisdemethoxycurcumin (BDMC). (b) Survival rate of zebrafish embryos after 72 h of treatment of 8 mM kojic acid or CUR, DMC, and BDMC. Hpf, hours post fertilization. Figure 5 Effect of curcumin (CUR), dimethoxycurcumin (DMC), and bisdemethoxycurcumin (BDMC) on melanin production in zebrafish embryos. (a) Images of overhead view of zebrafish embryo at 72 h after exposure of various concentrations of CUR, DMC, and BDMC. (b) Color of whole-body lysate with 1 M NaOH. (c) Determined absorbance at 490 nm for melanin contents. Data are presented as mean +- standard deviation (SD) of triplicates. Significant differences are expressed by the symbols *, **, ***, and **** as * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001, respectively. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Aburjai T. Natsheh F.M. Plants used in cosmetics Phytotherapy Res. 2003 17 987 1000 10.1002/ptr.1363 2. Freires I.A. Sardi J.D.C.O. de Castro R.D. Rosalen P.L. Alternative Animal and Non-Animal Models for Drug Discovery and Development: Bonus or Burden? Pharm. Res. 2017 34 681 686 10.1007/s11095-016-2069-z 27858217 3. Choi T.-Y. Kim J.-H. Ko D.H. Kim C.-H. Hwang J.-S. Ahn S. Kim S.Y. Kim C.D. Lee J.-H. Yoon T.-J. Zebrafish as a new model for phenotype-based screening of melanogenic regulatory compounds Pigment. Cell Res. 2007 20 120 127 10.1111/j.1600-0749.2007.00365.x 17371438 4. Kalueff A.V. Stewart A.M. Gerlai R. Zebrafish as an emerging model for studying complex brain disorders Trends Pharmacol. Sci. 2014 35 63 75 10.1016/j.tips.2013.12.002 24412421 5. MacRae C.A. Peterson R.T. Zebrafish as tools for drug discovery Nat. Rev. Drug Discov. 2015 14 721 731 10.1038/nrd4627 26361349 6. Kumari S. Thng S.T.G. Verma N.K. Gautam H.K. Melanogenesis Inhibitors Acta Derm. Venereol. 2018 98 924 931 10.2340/00015555-3002 29972222 7. D'Mello S.A.N. Finlay G.J. Baguley B.C. Askarian-Amiri M.E. Signaling Pathways in Melanogenesis Int. J. Mol. Sci. 2016 17 1144 10.3390/ijms17071144 27428965 8. Rzepka Z. Buszman E. Beberok A. Wrzesniok D. From tyrosine to melanin: Signaling pathways and factors regulating melanogenesis Postep. Hig. Med. Dosw. 2016 70 695 708 10.5604/17322693.1208033 9. Yasumoto K. Yokoyama K. Shibata K. Tomita Y. Shibahara S. Microphthalmia-associated transcription factor as a regulator for melanocyte-specific transcription of the human tyrosinase gene Mol. Cell. Biol. 1994 14 8058 8070 10.1128/mcb.14.12.8058-8070.1994 7969144 10. Bertolotto C. Abbe P. Hemesath T.J. Bille K. Fisher D.E. Ortonne J.-P. Ballotti R. Microphthalmia Gene Product as a Signal Transducer in cAMP-Induced Differentiation of Melanocytes J. Cell Biol. 1998 142 827 835 10.1083/jcb.142.3.827 9700169 11. Amalraj A. Pius A. Gopi S. Gopi S. Biological activities of curcuminoids, other biomolecules from turmeric and their derivatives--A review J. Tradit. Complement. Med. 2017 7 205 233 10.1016/j.jtcme.2016.05.005 28417091 12. Carolina Alves R. Perosa Fernandes R. Fonseca-Santos B. Damiani Victorelli F. Chorilli M. A critical review of the properties and analytical methods for the determination of curcumin in biological and pharmaceutical matrices Crit. Rev. Anal. Chem. 2019 49 138 149 10.1080/10408347.2018.1489216 30252504 13. Kotha R.R. Luthria D.L. Curcumin: Biological, Pharmaceutical, Nutraceutical, and Analytical Aspects Molecules 2019 24 2930 10.3390/molecules24162930 31412624 14. Kumari M. Nanda D.K. Potential of Curcumin nanoemulsion as antimicrobial and wound healing agent in burn wound infection Burns 2022 in press 10.1016/j.burns.2022.10.008 15. Wang S. Gao X. Li J. Wei S. Shao Y. Yin Y. Zhang D. Tang M. The anticancer effects of curcumin and clinical research progress on its effects on esophageal cancer Front. Pharmacol. 2022 13 1058070 10.3389/fphar.2022.1058070 36386215 16. Zhu C. Fang Z. Peng L. Gao F. Peng W. Song F. Curcumin Suppresses the Progression of Colorectal Cancer by Improving Immunogenic Cell Death Caused by Irinotecan Chemotherapy 2022 67 211 222 10.1159/000518121 36380644 17. Jeon H.-J. Kim K. Kim C. Kim M.-J. Kim T.-O. Lee S.-E. Molecular Mechanisms of Anti-Melanogenic Gedunin Derived from Neem Tree (Azadirachta indica) Using B16F10 Mouse Melanoma Cells and Early-Stage Zebrafish Plants 2021 10 330 10.3390/plants10020330 33572083 18. Garcia-Contreras R. Vos P. Westerhoff H.V. Boogerd F.C. Why in vivo may not equal in effectors revealed by measurement of enzymatic activities under the same in vivo-like assay conditions FEBS J. 2012 279 4145 4159 10.1111/febs.12007 22978366 19. Lai K.P. Gong Z. Tse W.K.F. Zebrafish as the toxicant screening model: Transgenic and omics approaches Aquat. Toxicol. 2021 234 105813 10.1016/j.aquatox.2021.105813 33812311 20. Hosoya T. Nakata A. Yamasaki F. Abas F. Shaari K. Lajis N.H. Morita H. Curcumin-like diarylpentanoid analogues as melanogenesis inhibitors J. Nat. Med. 2012 66 166 176 10.1007/s11418-011-0568-0 21830091 21. Lee J.-H. Jang J.-Y. Park C. Kim B.-W. Curcumin suppresses a-melanocyte stimulating hormone-stimulated melanogenesis in B16F10 cells Int. J. Mol. Med. 2010 26 101 106 Available online: (accessed on 9 January 2023) 20514428 22. Tu C.-X. Lin M. Lu S.-S. Qi X.-Y. Zhang R.-X. Zhang Y.-Y. Curcumin Inhibits Melanogenesis in Human Melanocytes Phytotherapy Res. 2012 26 174 179 10.1002/ptr.3517 23. Sharma R.A. Steward W.P. Gescher A.J. Pharmacokinetics and pharmacodynamics of curcumin Adv. Exp. Med. Biol. 2007 595 453 470 10.1007/978-0-387-46401-5_20 17569224 24. Yerneni S.S. Yalcintas E.P. Smith J.D. Averick S. Campbell P.G. Ozdoganlar O.B. Skin-targeted delivery of extracellular vesicle-encapsulated curcumin using dissolvable microneedle arrays Acta Biomater. 2022 149 198 212 10.1016/j.actbio.2022.06.046 35809788 25. Goenka S. Johnson F. Simon S. Novel Chemically Modified Curcumin (CMC) Derivatives Inhibit Tyrosinase Activity and Melanin Synthesis in B16F10 Mouse Melanoma Cells Biomolecules 2021 11 674 10.3390/biom11050674 33946371 26. Goenka S. Simon S. Novel Chemically Modified Curcumin (CMC) Analogs Exhibit Anti-Melanogenic Activity in Primary Human Melanocytes Int. J. Mol. Sci. 2021 22 6043 10.3390/ijms22116043 34205035 27. Lo C.-Y. Liu P.-L. Lin L.-C. Chen Y.-T. Hseu Y.-C. Wen Z.-H. Wang H.-M. Antimelanoma and Antityrosinase from Alpinia galangal Constituents Sci. World J. 2013 2013 186505 10.1155/2013/186505 28. D'Ischia M. Wakamatsu K. Napolitano A. Briganti S. Garcia-Borron J.-C. Kovacs D. Meredith P. Pezzella A. Picardo M. Sarna T. Melanins and melanogenesis: Methods, standards, protocols Pigment. Cell Melanoma Res. 2013 26 616 633 10.1111/pcmr.12121 23710556 29. Feng D. Fang Z. Zhang P. The melanin inhibitory effect of plants and phytochemicals: A systematic review Phytomedicine 2022 107 154449 10.1016/j.phymed.2022.154449 36126406 30. Davies S.N. Martin D. Millar J.D. Aram J.A. Church J. Lodge D. Differences in results from in vivo and in vitro studies on the use-dependency of N-methylaspartate antagonism by MK-801 and other phencyclidine receptor ligands Eur. J. Pharmacol. 1988 145 141 151 10.1016/0014-2999(88)90225-7 2832187 31. Kim C. Choe H. Park J. Kim G. Kim K. Jeon H.-J. Moon J.-K. Kim M.-J. Lee S.-E. Molecular mechanisms of developmental toxicities of azoxystrobin and pyraclostrobin toward zebrafish (Danio rerio) embryos: Visualization of abnormal development using two transgenic lines Environ. Pollut. 2021 270 116087 10.1016/j.envpol.2020.116087 33234374 32. Lajis A.F.B. A Zebrafish Embryo as an Animal Model for the Treatment of Hyperpigmentation in Cosmetic Dermatology Medicine Medicina 2018 54 35 10.3390/medicina54030035 30344266 33. Hu S. Huang J. Pei S. Ouyang Y. Ding Y. Jiang L. Lu J. Kang L. Huang L. Xiang H. Ganoderma lucidum polysaccharide inhibits UVB-induced melanogenesis by antagonizing cAMP/PKA and ROS/MAPK signaling pathways J. Cell. Physiol. 2019 234 7330 7340 10.1002/jcp.27492 30362532 34. Wu J.-Y. Lin C.-Y. Lin T.-W. Ken C.-F. Wen Y.-D. Curcumin Affects Development of Zebrafish Embryo Biol. Pharm. Bull. 2007 30 1336 1339 10.1248/bpb.30.1336 17603177 35. Cheng M.-C. Lee T.-H. Chu Y.-T. Syu L.-L. Hsu S.-J. Cheng C.-H. Wu J. Lee C.-K. Melanogenesis Inhibitors from the Rhizoma of Ligusticum Sinense in B16-F10 Melanoma Cells In Vitro and Zebrafish In Vivo Int. J. Mol. Sci. 2018 19 3994 10.3390/ijms19123994 30545008 36. Li Y. Xiang H. Xue X. Chen Y. He Z. Yu Z. Zhang L. Miao X. Dual Antimelanogenic Effect of Nicotinamide-Stabilized Phloretin Nanocrystals in Larval Zebrafish Pharmaceutics 2022 14 1825 10.3390/pharmaceutics14091825 36145574 |
PMC10000566 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050913 diagnostics-13-00913 Article Whale Optimization Algorithm with a Hybrid Relation Vector Machine: A Highly Robust Respiratory Rate Prediction Model Using Photoplethysmography Signals Dong Xuhao Conceptualization Methodology Software Validation Formal analysis Investigation Resources Data curation Writing - original draft 1 Wang Ziyi Visualization 1 Cao Liangli Project administration 1 Chen Zhencheng Project administration Funding acquisition 123* Liang Yongbo Conceptualization Writing - review & editing Supervision Project administration 123* Aguirre Juan Academic Editor 1 School of Life and Environmental Sciences, Guilin University of Electronic Technology, Guilin 541004, China 2 Guangxi Colleges and Universities Key Laboratory of Biomedical Sensors and Intelligent Instruments, Guilin 541004, China 3 Guangxi Engineering Technology Research Center of Human Physiological Information Noninvasive Detection, Guilin 541004, China * Correspondence: [email protected] (Z.C.); [email protected] (Y.L.) 28 2 2023 3 2023 13 5 91315 1 2023 18 2 2023 25 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Due to the simplicity and convenience of PPG signal acquisition, the detection of the respiration rate based on the PPG signal is more suitable for dynamic monitoring than the impedance spirometry method, but it is challenging to achieve accurate predictions from low-signal-quality PPG signals, especially in intensive-care patients with weak PPG signals. The goal of this study was to construct a simple model for respiration rate estimation based on PPG signals using a machine-learning approach fusing signal quality metrics to improve the accuracy of estimation despite the low-signal-quality PPG signals. In this study, we propose a method based on the whale optimization algorithm (WOA) with a hybrid relation vector machine (HRVM) to construct a highly robust model considering signal quality factors to estimate RR from PPG signals in real time. To detect the performance of the proposed model, we simultaneously recorded PPG signals and impedance respiratory rates obtained from the BIDMC dataset. The results of the respiration rate prediction model proposed in this study showed that the MAE and RMSE were 0.71 and 0.99 breaths/min, respectively, in the training set, and 1.24 and 1.79 breaths/min, respectively, in the test set. Compared without taking signal quality factors into account, MAE and RMSE are reduced by 1.28 and 1.67 breaths/min, respectively, in the training set, and reduced by 0.62 and 0.65 breaths/min in the test set. Even in the nonnormal breathing range below 12 bpm and above 24 bpm, the MAE reached 2.68 and 4.28 breaths/min, respectively, and the RMSE reached 3.52 and 5.01 breaths/min, respectively. The results show that the model that considers the PPG signal quality and respiratory quality proposed in this study has obvious advantages and application potential in predicting the respiration rate to cope with the problem of low signal quality. respiratory rate photoplethysmography signal hybrid relation vector machine (HRVM) whale optimization algorithm (WOA) ensemble empirical mode decomposition with principal component analysis (EEMD-PCA) National Natural Science Foundation of China62101148 Natural Science Foundation of Guangxi2020GXNSFBA297156 2021GXNSFBA220051 Guangxi Innovation Driven Development ProjectGuike AA19254003 This work was supported in part by the National Natural Science Foundation of China (62101148), Natural Science Foundation of Guangxi (2020GXNSFBA297156, 2021GXNSFBA220051) and Guangxi Innovation Driven Development Project (Guike AA19254003). pmc1. Introduction Impedance spirometry is a clinically important method for measuring the respiratory rates of patients; however, it is not very convenient and comfortable for patients and is especially not suitable for dynamic monitoring . To overcome these restrictions, studies have given more attention to physiological signals, such as electrocardiogram (ECG) and photoplethysmography (PPG) signals. However, PPG signals are more attractive than ECG signals because of their simplicity, portability, and low number of sensors . Due to the low cost and portability of PPG signal acquisition, the continuous, non-invasive, and reliable monitoring of patients' respiratory rate based on PPG signals has attracted many researchers in recent years , which will help monitor primary health status and contribute to the diagnosis of cardiorespiratory diseases, including pneumonia and obstructive sleep apnoea (OSA) , and physiological situations, such as hypercarbia and pulmonary embolism . However, for critically ill patients in hospital care situations, where human respiration is relatively weaker and venous blood is at a very low pressure, flowing with the frequency of movement , the pulsatile alternating current (AC) component is also easily affected by physiological activity and movement, causing the PPG signals to contain more noise than physiological information , which would make estimating the respiratory rate based on PPG signals specifically challenging. In the 1990s, the method of estimating the respiration rate (RR) based on PPG signals was investigated by Nakajima et al. , who developed a digital filter technology to estimate the RR from PPG signals. This method proved the feasibility of RR estimation based on PPG, but lacked universality. To solve this problem, Karlen et al. used an intelligent fusion algorithm to fuse three respiration-modulated changes, including pulse amplitude modulation (AM), frequency modulation (FM), and baseline wander (BW), from PPG, which estimated a relatively accurate RR. Similarly, Meredith et al. explained that respiratory components are reflected by the AM, FM, and BW of PPG signals. To reduce the effect of poor respiratory modulation on the accuracy of the final estimated respiration rate, Birrenkott et al. proposed three respiratory quality indices (RQIs), which set an adjustable threshold to fuse the respiratory rate estimated by three respiratory modulations, on the dataset of elective surgery or routine anaesthesia (CapnoBase Dataset), which showed good results with a mean absolute error (MAE) of only 0.71 +- 0.89 bpm. However, for critically ill patients during hospital care (BIDMC dataset), the MAE reached 3.12 +- 4.39 bpm. Selvakumar et al. also showed that the BIDMC dataset was more challenging. Faint physiological conditions and the more complex pulse wave morphology in critically ill patients lead to the inaccurate extraction of respiratory modulation . To enhance the robustness of the respiration rate estimation algorithm. Ambekar el al. used a data-driven algorithm, ensemble empirical mode decomposition (EEMD), to obtain the RR from PPG signals. Compared to the empirical mode decomposition algorithm (EMD), EEMD overcomes the problem of modal aliasing . Adami et al. and Mohammod et al. compared the performance of EMD and its derivative algorithms, such as EEMD, CEEMD, CEEMDAN, and ICEEMDAN, to decompose the respiratory component from PPG signals and demonstrated that both EEMD and CEEMDAN have better performance but that the EEMD method has a lower computational cost than CEEMDAN. To further isolate the respiratory components from the selected inherent modal functions (IMF), a Kalman filter (KF) was used to exclude other components . Considering the nonlinearity of the PPG signal, Adami et al. introduced the PPG signal quality indicator as a way to adjust the Kalman gain to implement extended Kalman filtering (EKF). However, these methods are dependent on the stability of the previous window, and if the respiratory rate error in the first window is large, it will significantly increase the overall estimation error level. The principal component analysis (PCA) method avoids this problem and does not require parameter adjustment . Mohammod et al. proposed an EEMD-PCA method, demonstrated its advantages for respiration rate estimation based on PPG signals , and explained the failure as high-amplitude noise in the RR band range . Machine-learning methods can be used to correct respiration rate estimation errors due to these high-amplitude noises , For example, the RRWaveNet and CAGAB methods demonstrate the feasibility of using machine-learning methods to improve the accuracy of PPG-based respiratory rate detection. In summary, both PPG signal quality and respiratory signal quality influence the accuracy of the PPG-based estimation of respiration rate to further reduce their effect on respiration rate estimation. In this paper, we use machine-learning methods to construct a respiration rate estimation model by fusing PPG signal quality indices and respiratory quality indices (RQIs) to limit the influences of other physiological activities and noise on respiration rate estimation based on PPG, developing a respiration rate estimation model with high accuracy and robustness. The remainder of this study is summarized as follows: Section 2 introduces the dataset and respiration rate prediction model construction process with model performance evaluation metrics. In Section 3, the performance of the model proposed in this study on the BIDMC dataset is reported, followed by a discussion of the advantages and disadvantages of the model in Section 4. Section 5 concludes the study and provides recommendations for future work. 2. Materials and Methods Figure 1 shows a clear block diagram of the respiration rate model construction process. The whole process is divided into three stages: (a) pre-extracted respiratory wave and respiratory rate, (b) signal quality index calculation, and (c) respiratory rate prediction model construction. In the first stage, the EEMD-PCA method was selected to pre-extract the respiration wave and respiration rate. In the second stage, the PPG signal quality indices and RQIs were calculated for further processing. The HRVM method was employed to fuse the respiration rate estimated by the EEMD-PCA method (RRP) and the signal quality indicators that have an impact on the estimation of RRP for RR prediction model construction, and the kernel parameters were optimized by the WOA algorithm to prevent falling into a local optimum. To evaluate the performance of the constructed model, it is tested with the BIDMC dataset. 2.1. Database The BIDMC dataset was collected from 53 subjects (20 males and 33 females; age range: 19-90 years old) and acquired from critically ill patients during hospital care at the Beth Israel Deaconess Medical Centre (Boston, MA, USA). For each subject, over an 8 min duration, each subject contains physiological signals that are sampled at 125 Hz, such as PPG signals, impedance respiratory signals, and electrocardiogram (ECG) signals; simultaneously, reference physiological parameters such as respiratory rate (RR) and heart rate (HR) are sampled at 1 Hz. Two annotators manually annotated the start and end time points of each single respiration in all recordings via the impedance respiration signal, and the corresponding PPG signal segment with a 1-s difference in the duration of single respiration annotated by the two annotators was removed. Due to a severe loss of the reference respiration rate in the 13th subject, the remaining 52 subjects were retained with the PPG signal split into 8-s nonoverlapping windows with a 32-s length. This process resulted in 2719 (93.4%) windows being retained. The distribution of the impedance respiration rate (RRI) values for all windows is shown in Figure 2, which reveals that the distribution of the reference respiration rate ranges from 3 to 30 bpm, mainly between 16 and 20 bpm, follows a regular distribution, and reflects well the real-world respiration rate distribution. In this study, the dataset was randomly divided into a training set with a validation part (70%) and a test set (30%). 2.2. Pre-Extracted Respiratory Wave and Respiratory Rate by the EEMD-PCA Method EEMD-PCA is a novel, data-driven method for estimating the respiration rate based on PPG signals that was proposed by Mohammod et al. . In this study, the pre-extracted respiratory waves and RRP will be extracted by this method for use in the next stage. The pre-extraction process is subdivided into four steps: (a) EEMD is applied to PPG signals to separate the respiratory components and other components, (b) intrinsic mode functions (IMFs) dominated by respiratory components are selected for further processing, (c) the selected IMFs are used to reconstruct the respiratory waves and are further denoised with principal component analysis (PCA), and (d) fast Fourier transform (FFT) is applied to the pre-extracted respiratory waves from the previous step to calculate the RRP. Figure 3 shows the time domain and frequency domain of the impedance respiratory signals and the pre-extracted respiratory waves for the first 32-s window of subject BIDMC 10. They have strong consistency in the waveform period in the time domain, and the main frequency components are similar in the frequency domain. FFT was applied to the pre-extracted respiratory signals, which are dominated by respiration, and the frequency corresponding to the maximum peak of the spectrogram is expressed as the frequency corresponding to the respiration rate and then converted to the RRP using Formula (1). (1) RRP=fRRP*60breaths/min 2.3. Signal Quality Index Calculation Both PPG signal quality indices and respiration quality indices (RQIs) affect the accuracy of RR estimation based on PPG signals. An optimal PPG signal quality index (SSQI) and three typical RQIs (QR1, QR2, and QR3) are calculated in this section to fuse the RRP to reduce the error of RR estimation and enhance the robustness of the algorithm proposed in this paper. 2.3.1. PPG Signal Quality Index (SQI) Calculation Skewness is a measure of the symmetry of the probability distribution. Mohamed et al. discovered that the skewness value of a 2-s PPG signal significantly varies with the change in the quality of the PPG signals, with an accuracy of 82.86% in determining between high-quality PPG signals and damaged unusable PPG signals, which is calculated by Formula (2). (2) SSQI=1Ni=1Nxi-m^x/s3 where xi is the ith sample point value of PPG, m^x and s are the empirical estimates of the mean and standard deviation of xi, respectively, and N is the number of samples in the PPG signals. For each 32-s PPG signal with a two-second nonoverlapping sliding window, a total of 16 skewness values are calculated, and the average of 16 SSQI represents the overall quality level of the 32-s PPG signal for that segment. The specific process is expressed as follows:(3) SSQI =1nw=1nSSQIw where SSQIw denotes the SSQI of the PPG signals for the wth 2-s window and n is the number of windows. SSQI denotes the quality level of the PPG signal for each 32-s window. 2.3.2. Respiratory Quality Index (RQI) Calculation The autocorrelation RQI, FFT RQI, and autoregression RQI (QR1, QR2, and QR3) were proposed by Birrenkott et al. , who directly calculated their RQIs on the PPG signal after filter processing with a fixed cut-off frequency and down-sampled to 4 Hz, which still contains much low-frequency motion noise and cardiac components. In contrast, this paper will calculate the three RQIs on the pre-extracted respiration waves down-sampled to 4 Hz, which are dominated by respiratory components that are more reflective of respiratory signal quality. 2.4. WOA-HRVM Model In the previous stage, the RRP, SSQI , and three RQIs were obtained and used as features in this stage, and the RRI corresponding to each window was applied as labels. The WOA-HRVM method was applied to the training set to construct an RR prediction model, and the testing set was employed to evaluate the performance of the model. As a highly sparse model that provides probabilistic predictions by Bayesian inference, the central idea of related vector machines (RVMs) is to obtain the correlation vectors and weights by maximizing the marginal likelihood . RVMs are often utilized as a machine-learning method for regression prediction, and its kernel function and kernel parameters are adjusted according to the requirements, which are also important parameters affecting the final regression performance. To improve the performance of the regression, a hybrid relation vector machine (HRVM) was employed in this paper. Since convex combinations of finitely many elementary kernel functions can always generate optimal kernels, hybrid kernel learning methods are more efficient than single kernel learning methods . The multiple heterogeneous kernel learning method is defined as (4) Kxi,x=m=1MdmKmxi,x where dm is the weight of the mth kernel function with dm >= 0, and Kxi,x denotes the mth kernel function, Gaussian kernel function, sigmoid kernel function, polynomial kernel function, and Laplacian kernel function as common kernel functions used in this study. In addition, the initial values of the kernel parameters are highly random; the convergence of the regression model constructed based on HRVM will be greatly affected as a result, and it is easy to fall into the local optimum. Aimed at the limitations of the HRVM algorithm, the whale optimization algorithm (WOA) has the advantages of few adjustment parameters, simple operation, and strong ability for a global search. The optimal parameters and weights of the kernel function are obtained by continuous iteration of the WOA algorithm to prevent local optimality, so the respiration rate model proposed in this study based on the WOA-HRVM algorithm can be represented by Formula (4), it is a hybrid function consisting of Gaussian kernel function, sigmoid kernel function, polynomial kernel function and Laplace kernel function. 2.5. Performance Measurement The ability of our RR prediction model was assessed using three methods: (i) Bland-Altman plot: the plot visualizes the consistency of the predicted respiration rate by the model proposed in this study (RRM) with the RRI; (ii) mean absolute error (MAE): the accuracy of the model was demonstrated by averaging the absolute value of the difference between RRM and RRI over all windows; and (iii) root-mean-square error (RMSE): RMSE is used to reflect the precision of the model proposed in this study; it is very sensitive to the very large or very small errors of the RRM compared to RRI. 3. Results The model in this study was constructed and tested based on MATLAB 2020a (MathWorks, Natick, MA, USA). The Bland-Altman plot visualizes the relationship among RRP (RR estimated by the EEMD-PCA method), RRM (RR estimated by the prediction model proposed in this study), and RRI in Figure 4. In the training set, the difference in RRP and RRI was 0.07 bpm, with limits of agreement from -5.138 to 5.278 bpm, and the difference in RRM and RRI was almost 0 bpm, with smaller limits of agreement from -1.930 to 1.930 bpm. In the test set, the difference in RRP and RRI was 0.121 bpm, with limits of agreement from -4.660 to 4.906 bpm. The difference in RRM and RRI is only -0.015 bpm, with narrowed limits of agreement from -3.564 to 3.533 bpm. The figure includes a total of three lines for each method, with the middle line indicating the mean of the differences, and the upper and lower lines showing the upper and lower limits, respectively, of the 95% consistency limits (mean +- 1.96SD). The closer the line showing the mean of the differences is to 0 bpm, the higher the agreement between the two measurement methods and the smaller the 95% confidence interval. The closer the method is to the impedance respiration test, the higher the clinical acceptability. Therefore, Figure 4 and Table 1 show that, compared to the EEMD-PCA method, the respiration rate estimated by the proposed method is more consistent with the respiration rate measured by the impedance spirometry. Table 2 shows the MAE and RMSE of the PPG-derived RR (RRP and RRM) with RRI in the training set and test set. Even in the test set, the MAE and RMSE of the respiration rate prediction model proposed in this study are only 1.24 and 1.79 bpm, respectively, which are 0.62 lower and 0.65 bpm lower, respectively, than those of the EEMD-PCA method. Figure 5 illustrates the performance of the proposed method for the continuous monitoring of RR based on PPG signal segments at different reference respiration rates in different people. The top half shows the training set results, and the bottom half shows the test set results, both for an 8 min duration. The PPG signal segments from different people are mixed and have many sudden changes in RR, so we can check the capability of the respiration rate prediction model proposed in this study in tracking sudden changes and adaptability among different people. As shown in Figure 5, the proposed method shows good performance for the continuous detection of the respiration rate with mixed PPG signals at different respiration rates in different people. The proposed method is capable of estimating the sharp change in RR better than the EEMD-PCA algorithm. Considering the performance in different respiratory rate ranges, Table 3 shows the performance in different respiratory rate ranges in the training set and testing set of the proposed model. According to Table 3, both the training set and the test set show good performance in the normal respiratory rate ranges of 12-16 bpm, 17-20 bpm, and 21-24 bpm. Even in the test set, the MAE is less than 2 bpm for the respiration rate prediction model proposed in this study, especially in the ranges of 17-20 bpm and 21-24 bpm, and the MAE decreases nearly twofold. In the range of human respiratory rates that are too fast or too slow (<12 bpm and >24 bpm), the MAE on the test set reaches 2.68 bpm and 4.28 bpm, respectively. However, compared to the EEMD-PCA method, the MAE still decreased by 3.57 bpm and 2.34 bpm. 4. Discussion In this study, we consider both the PPG signal quality and respiration signal quality to estimate the respiration rate based on PPG signals and validate its accuracy and robustness on the BIDMC dataset. This proposed model is developed based on the HRVM and WOA algorithm. The use of hybrid kernel functions allows an exploration of the relationship between the error in the respiration rate estimated by the EEMD-PCA method and the signal quality indicators in a wider range of dimensions, and the WOA algorithm avoids falling into the local optimum by continuously iterating to identify the most suitable kernel function width and weight parameters. The respiration rate prediction model proposed in this study has the advantages of both a local kernel function and nonlocal kernel function. Therefore, the method has a higher accuracy in RR detection compared with other methods. Previous studies used fixed threshold filters to pre-process PPG signals, which will inevitably filter out some respiration information. Therefore, we did not use any filters to pre-process the PPG signal. Instead, the data-driven EEMD-PCA method was directly utilized to exclude motion and cardiac noise and to pre-extract respiratory waves. In addition, the EEMD method is robust to noise, and PCA further reduces noise and cardiovascular signal interference based on variance, making the initially extracted respiratory signal highly reliable. Therefore, RQIs are calculated based on the pre-extracted respiratory waves better than the pre-processed PPG signal. The complexity of the physiological condition of critically ill patients and the uncertainty of external noise produce complex changes in the error rate of the RR estimated by the EEMD-PCA method. Due to the differences in the human body, for different subjects, there is a significantly different Pearson correlation coefficient between the error rate of the RR estimated by the EEMD-PCA method and the four signal quality indices. Table 4 presents the R1, R2, R3, and R4 for 52 subjects, from which we observe that the R1, R2, R3 and R4 are significantly different for different subjects. For example, the correlation of four signal quality indicators of Subject 01 with the error in the respiration rate estimated based on the EEMD-PCA method is clearly more relevant than that of Subject 02. In addition, the sensitivity of different signal quality indicators differed for the same subject; for example, for Subject 04, compared to other signal quality indices, QR2 was not as relevant, while for Subject 01, it was QR1 that was less relevant. However, the mean values of these four correlation coefficients show that none of these four signal quality indicators is better than the other three signal quality indicators. To reduce the influence of signal quality indicators with a low correlation with the respiration rate estimation, the current state-of-the-art approach is to improve the accuracy of the estimates at the expense of discarding unusable data by 'intelligent fusion' methods. RQI Fusion reduces the percentage of discards by setting an adjustable signal quality indicator threshold. In this paper, we use the sparsity of the HRVM algorithm to select a certain percentage of data from the training set for the model construction and optimization by the WOA algorithm. This method is data-driven to determine the percentage of discards without setting a threshold parameter, and it is only necessary to give an objective function that yields an estimated respiration rate that is closest to the impedance respiration rate, which provides better robustness than other RR estimation methods. In this paper, an end-to-end respiration rate prediction model is constructed. The advantages of the model proposed in this paper, in comparison with the end-to-end respiration rate prediction methods based on PPG signals proposed by other researchers, are shown in Table 5. According to our results and those of other authors on the BIDMC dataset in recent years, the MAE and RMSE of the model proposed in this paper on the test set are only 1.24 bpm and 1.79 bpm, respectively, which are much lower than those of other methods. As shown in Table 6, although the framework proposed in and the EEMD + KF method both show better results, the framework is too complicated to calculate; each time, it needs to use the EMD and DWT methods to calculate seven different predicted respiratory waves to fuse, and it takes 30 s to update the RR, 22 s more than the model proposed in this paper. The reason for choosing 8 s to update the breathing rate in this paper is based on matlab2020a with the EEMD-PCA method to decompose a 32-s PPG signal to extract the predicted respiratory rate and respiratory waves, plus the time to calculate four signal quality indicators is close to 8 s. Although the method in is simpler to calculate, it relies on the signal quality of the first PPG window, and the error in estimating the respiration rate in the first window can lead to large errors in all subsequent windows. Neither the conventional respiratory-modulation-based methods in in the paper showed good results, which is caused by the challenging respiratory modulation extraction when the signal quality is poor. For some advanced machine-learning methods or deep learning methods , they are affected by the accuracy of feature extraction or feature selection, resulting in their outcomes not being better than this paper's. This paper uses the EEMD-PCA method to pre-extract the respiration rate and respiration wave, which not only avoids the challenging and inaccurate extraction based on the traditional respiration modulation, but also improves the accuracy and robustness of the respiration rate prediction by incorporating signal quality factors into the respiration prediction model using machine-learning methods. The RMSE of the CAGBA method was much lower than that of the other methods as only 20 subjects were selected. To balance the continuity of the respiration rate detection and the accuracy of the respiration rate estimation, the appropriate PPG signal length is also important. It is evident from recent literature that the performance of respiration rate detection algorithms decreases as the PPG signal data length decreases. It is well-known that short data lengths are important for real-time respiratory rate detection in critical care or wearable devices, but a PPG signal that is too short is not conducive to accurate respiratory rate detection. In , the authors concluded that a length of 32 s is the most stable and shortest length for extracting respiratory signals based on PPG signals. Table 6 compared with other respiration rate estimation methods in recent years, the model proposed in this study showed better robustness and accuracy in estimating RR than other existing methods. A limitation of this study is the method for calculating PPG signal quality indicators and respiratory signal quality indicators. We calculated four signal quality indicators and tested them on different people. For some subjects, the sensitivity was poor, and a more sensitive signal quality index should be investigated. The choice of kernel function and the number of iterations of the WOA algorithm are also key factors affecting the accuracy of the final respiration rate prediction model. Other kinds of kernel functions and larger numbers of iterations should continue to be explored. In addition, the model was only tested on the BIDMC dataset; other datasets or an autonomous collection of real-world data should be collected by the latter to further validate the stability of the proposed method. The advantage of the proposed method is the end-to-end estimation of the respiration rate based only on PPG signals without a complex parameter adjustment, and the performance is significantly improved compared to other respiration rate estimation methods. diagnostics-13-00913-t006_Table 6 Table 6 Performance comparison of different PPG-signal-based respiration rate detection methods on BIDMC dataset. References Method Length (sec) Subjects Overlap (sec) MAE RMSE Dis (%) This study WOA-HRVM 32 53 24 1.24 1.79 6.6% Adami EMD and DWT + EKF 60 53 30 0.73 - - Pongpanut RRWaveNet 32 53 - 1.62 - 1.9% Sharma EEMD + KF 32 53 29 1.90 - - Aqajari CycleGAN 30 53 - 1.90 - - Lee CAGBA 32 20 0 1.94 0.61 62.26% Bian Deep learning 60 53 59 2.50 - - Karlen SmartQualityFusion method 60 53 - 2.60 - - Birrenkott RQI calculation and fusion 32 53 17 3.12 4.39 23.2% Notes: Dis (%) indicates the percentage of discarded data in the entire dataset. 5. Conclusions In this paper, we used the WOA-HRVM method to fuse the PPG signal quality and respiratory signal quality with the respiratory rate estimated based on the EEMD-PCA method to construct a highly accurate and robust respiratory rate prediction model based on the PPG signal. The method is data-driven and does not require complex parameter tuning, which affects the stability of the respiratory rate prediction model, and overcomes the problem of difficult and inaccurate extraction when the signal quality is poor with traditional respiratory modulation methods. It also does not require the extraction of PPG morphological features and screening with feature selection methods as with other machine-learning methods or deep learning methods, and the final performance is affected by the accuracy of feature recognition and the performance of feature selection methods. After comparing the performance of the PPG-signal-based estimation of the respiration rate on the BIDMC dataset with that of previous investigators, the proposed methods showed more accurate results in estimating the RR than other existing methods for subjects from the BIDMC dataset with a short data length and a 32-s PPG signal. In future studies, we will validate the method using other datasets or an autonomous collection of real-world data in large cohorts of short data lengths while exploring more effective PPG and respiratory signal quality metrics to further improve the accuracy of the respiratory rate prediction model so that it can eventually be applied to the real-time detection of the respiratory rate on wearable devices or be utilized instead of impedance detection for the real-time detection of the respiratory rate in patients under intensive care. It promotes the development of portability for the real-time respiratory detection of patients under intensive care, and has a very important theoretical value for realizing the real-time detection of the respiratory rate in wearable devices and telemedicine, and improving the accuracy of respiratory rate measurement. Author Contributions X.D. designed the study. Z.W., L.C., Y.L. and Z.C. conceived the study, provided directions, feedback, and/or revised the manuscript. X.D. led the investigation and drafted the manuscript for submission with revisions and feedback from the contributing authors. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The data used in this manuscript can be downloaded from this link (accessed on 14 January 2023). Conflicts of Interest The authors declare that there are no competing interest. Figure 1 Flow chart of respiration rate prediction model construction. Figure 2 Distribution of each 32-s PPG window corresponding to the RRI. The horizontal axis indicates the value of the RRI for each 32-s window, and the vertical axis indicates the number of windows at the corresponding RRI. Figure 3 (a) Time domain comparison of impedance respiratory signals with pre-extracted respiratory signals based on the EEMD-PCA method. (b) Frequency domain comparison of impedance respiratory signals with pre-extracted respiratory signals based on the EEMD-PCA method. The blue line represents the pre-extracted respiratory signals in the time-frequency domain, and the red line represents the impedance respiratory signals in the time-frequency domain. Figure 4 Bland-Altman plot of the RRM (RR estimated by the prediction model proposed in this study), RRP (RR estimated by the EEMD-PCA method), and RRI for the training set and test set. The figure includes a total of 3 lines for each method, with the middle line indicating the mean of the differences, and the upper and lower lines showing the upper and lower limits, respectively, of the 95% consistency limits. The horizontal axis represents the average of RRP and RRI or the average of RRM and RRI, and the vertical axis represents the difference, RRP-RRI or RRM-RRI: (a) Bland-Altman plot of the RRM and RRI in the training set, (b) Bland-Altman plot of the RRp and RRI in the training set, (c) Bland-Altman plot of the RRM and RRI in the test set, and (d) Bland-Altman plot of the RRp and RRI in the test set. Figure 5 The EEMD-PCA method and respiration rate prediction model proposed in this study with different subjects and time windows for the estimation of mutant respiration rates were assessed for consistency with impedance respiration rates. (The red dotted line represents the referenced impedance respiration rate (RRI), the blue dotted line denotes the respiration rate predicted by the respiration rate prediction model proposed in this study (RRM), and the green dotted line indicates the respiration rate estimated by the EEMD-PCA method (RRP)). diagnostics-13-00913-t001_Table 1 Table 1 A literature review of respiratory rate detection based on PPG. References Database Subjects Methodology Innovation Drawbacks Nakajima Self-collection 11 Digital filtering technique Real-time PPG-based respiration rate detection Lack of universal applicability Karlen CapnoBase 94 Smart fusion method Set discard thresholds with PPG signal quality metrics Discard 45% of data Birrenkott CapnoBase 42 RQI calculation and fusion Adjustable threshold value to change accuracy Inaccurate estimation of low-quality PPG signals Selvakumar CapnoBase 42 RIAV based on FFT Respiration rate detection on low-cost hardware Low accuracy in detecting too-fast breathing Sharma BIDMC 53 EEMD + KF Kalman filtering is applied to the reconstructed signal KF is not suitable for non-linear PPG signals Adami BIDMC 53 CEEMDAN + DWT + EKF Leverage time and frequency domain information Framework calculation is too complicated Mohammad MIMIC 121 EMD family and PCA Free from parameter selection Sensitivity to high-amplitude noise in the respiratory range Shuzan VORTAL 39 Machine-learning model Hyperparameter optimization Tested only on resting young people Pongpanut BIDMC 53 RRWaveNet Improve model robustness using transfer learning Discarded low-quality signals by SQI metric diagnostics-13-00913-t002_Table 2 Table 2 Mean of the differences, and the upper and lower limits of the 95% consistency limits for the RRM and RRP with the RRI in the training set and test set. Dataset Method Mean Mean +1.96SD Mean -1.96SD training set This study 0 1.930 -1.930 EEMD-PCA 0.070 5.278 -5.138 test set This study -0.015 3.533 -3.564 EEMD-PCA 0.121 4.906 -4.660 diagnostics-13-00913-t003_Table 3 Table 3 Mean absolute error (MAE) and root mean square error (RMSE) for the RRM (RR estimated by the prediction model proposed in this study) and RRP (RR estimated by the EEMD-PCA method) with the RRI in the training set and testing set. Dataset Method MAE RMSE training set This study 0.71 0.99 EEMD-PCA 1.99 2.66 test set This study 1.24 1.79 EEMD-PCA 1.86 2.44 diagnostics-13-00913-t004_Table 4 Table 4 Prediction performance in different ranges of respiratory rates. RRI (bpm) Training Set Test Set N MAEthis study [MAEEEMD-PCA] RMSEthis study [RMSEEEMD-PCA] N MAEthis study [MAEEEMD-PCA] RMSEthis study [RMSEEEMD-PCA] below 12 61 1.08 [6.40] 1.56 [6.57] 16 2.68 [6.25] 3.52 [6.34] 12-16 563 0.79 [2.34] 1.08 [2.65] 202 1.45 [2.21] 1.90 [2.50] 17-20 1118 0.63 [1.05] 0.82 [1.36] 407 0.91 [1.03] 1.21 [1.32] 21-24 211 0.75 [3.72] 1.16 [3.89] 88 1.80 [3.41] 2.45 [3.56] above 24 40 0.98 [7.17] 2.14 [7.47] 13 4.28 [6.62] 5.01 [6.77] Notes: MAEthis study and RMSEthis study denote the MAE and RMSE of RRM and RRI, MAEEEMD-PCA and RMSEEEMD-PCA denote the MAE and RMSE of RRP and RRI, and N denotes the number of windows. diagnostics-13-00913-t005_Table 5 Table 5 Pearson correlation coefficient (PCC) between signal quality indices and error rate of respiration rate estimated based on the EEMD-PCA method for different subjects. Subject R1 R2 R3 R4 Subject 01 0.14 0.70 -0.24 -0.70 Subject 02 0.30 -0.10 0.07 0.08 Subject 03 -0.22 0.07 -0.27 -0.54 Subject 04 -0.07 0.43 -0.16 -0.43 ... ... ... ... ... Subject 53 -0.32 -0.29 -0.29 -0.16 Average 0.28 0.24 0.27 0.28 Notes: R1 denotes the Pearson correlation coefficient between QR1 and the error rate of respiration rate estimated based on the EEMD-PCA method; R2 denotes the Pearson correlation coefficient between QR2 and the error rate of respiration rate estimated based on the EEMD-PCA method; R3 denotes the Pearson correlation coefficient between QR3 and the error rate of respiration rate estimated based on the EEMD-PCA method; R4 denotes the Pearson correlation coefficient between SSQI and the error rate of respiration rate estimated based on the EEMD-PCA method. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. AL-Khalidi F.Q. Saatchi R. Burke D. Elphick H. Tan S. Respiration rate monitoring methods: A review Pediatr. Pulmonol. 2011 46 523 529 10.1002/ppul.21416 21560260 2. Orphanidou C. Derivation of respiration rate from ambulatory ECG and PPG using ensemble empirical mode decomposition: Comparison and fusion Comput. Biol. Med. 2017 81 45 54 10.1016/j.compbiomed.2016.12.005 28012294 3. Selvakumar K. Kumar E.V. Sailesh M. Varun M. Allan A. Biswajit N. Namrata P. Upasana S. Realtime PPG based respiration rate estimation for remote health monitoring applications Biomed. Signal Process. Control. 2022 77 103746 10.1016/j.bspc.2022.103746 4. Alian A.A. Shelley K.H. Photoplethysmography Best Pract. Res. Clin. Anaesthesiol. 2014 28 395 406 10.1016/j.bpa.2014.08.006 25480769 5. Ginsburg A.S. Lenahan J.L. Izadnegahdar R. Ansermino J.M. A systematic review of tools to measure respiratory rate in order to identify childhood pneumonia Am. J. Respir. Crit. Care Med. 2018 197 1116 1127 10.1164/rccm.201711-2233CI 29474107 6. Zhang X. Kassem MA M. Zhou Y. Shabsigh M. Wang Q. Xu X. A brief review of non-invasive monitoring of respiratory condition for extubated patients with or at risk for obstructive sleep apnea after surgery Front. Med. 2017 4 26 10.3389/fmed.2017.00026 7. Rolfe S. The importance of respiratory rate monitoring Br. J. Nurs. 2019 28 504 508 10.12968/bjon.2019.28.8.504 31002547 8. Ruangsetakit C. Nuntiwattana N. Hahtapornsawan S. Hongku K. Wangwanit C. Chinsakchai K. Sermthanasawadi N. Mutirangura P. Digital photoplethysmography in the diagnosis of deep vein thrombosis in Thai patients J. Med. Assoc. Thai. 2017 100 S155 S161 9. Li D. Wang X. Liu D. Association between elevated central venous pressure and outcomes in critically ill patients Ann. Intensive Care 2017 7 1 7 10.1186/s13613-017-0306-1 28050894 10. Park S.H. Choi S.J. Park K.S. Advance continuous monitoring of blood pressure and respiration rate using denoising auto encoder and LSTM Microsyst. Technol. 2022 28 2181 2190 10.1007/s00542-022-05249-0 11. Nakajima K. Tamura T. Miike H. Monitoring of heart and respiratory rates by photoplethysmography using a digital filtering technique Med. Eng. Phys. 1996 18 365 372 10.1016/1350-4533(95)00066-6 8818134 12. Karlen W. Raman S. Ansermino J.M. Dumont G.A. Multiparameter respiratory rate estimation from the photoplethysmogram IEEE Trans. Biomed. Eng. 2013 60 1946 1953 10.1109/TBME.2013.2246160 23399950 13. Meredith D.J. Clifton D. Charlton P. Brooks J. Pugh C.W. Tarassenko L. Photoplethysmographic derivation of respiratory rate: A review of relevant physiology J. Med. Eng. Technol. 2012 36 1 7 10.3109/03091902.2011.638965 22185462 14. Birrenkott D.A. Pimentel M.A. Watkinson P.J. Clifton D.A. A robust fusion model for estimating respiratory rate from photoplethysmography and electrocardiography IEEE Trans. Biomed. Eng. 2017 65 2033 2041 10.1109/TBME.2017.2778265 29989939 15. Charlton P.H. Birrenkott D.A. Bonnici T. Pimentel M.A.F. Johnson A.E.W. Alastruey J. Tarassenko L. Watkinson P.J. Beale R. Clifton D.A. Breathing rate estimation from the electrocardiogram and photoplethysmogram: A review IEEE Rev. Biomed. Eng. 2017 11 2 20 10.1109/RBME.2017.2763681 29990026 16. Ambekar M.R. Prabhu S. A novel algorithm to obtain respiratory rate from the PPG signal Int. J. Comput. Appl. 2015 126 9 12 Available online: (accessed on 14 January 2023) 17. Colominas M.A. Schlotthauer G. Torres M.E. Improved complete ensemble EMD: A suitable tool for biomedical signal processing Biomed. Signal Process. Control. 2014 14 19 29 10.1016/j.bspc.2014.06.009 18. Adami A. Boostani R. Marzbanrad F. Charlton P.H. A new framework to estimate breathing rate from electrocardiogram, photoplethysmogram, and blood pressure signals IEEE Access 2021 9 45832 45844 10.1109/ACCESS.2021.3066166 19. Motin M.A. Karmakar C.K. Palaniswami M. Selection of empirical mode decomposition techniques for extracting breathing rate from PPG IEEE Signal Process. Lett. 2019 26 592 596 10.1109/LSP.2019.2900923 20. Sharma H. Extraction of respiratory rate from PPG using ensemble empirical mode decomposition with Kalman filter Electron. Lett. 2020 56 650 653 10.1049/el.2020.0566 21. Madhav K.V. Raghuram M. Krishna E.H. Komalla N.R. Reddy K.A. Komalla A.R. Use of multi scale PCA for extraction of respiratory activity from photoplethysmographic signals Proceedings of the 2012 IEEE International Instrumentation and Measurement Technology Conference Proceedings Graz, Austria 13-16 May 2012 1784 1787 22. Motin M.A. Karmakar C.K. Palaniswami M. Ensemble empirical mode decomposition with principal component analysis: A novel approach for extracting respiratory rate and heart rate from photoplethysmographic signal IEEE J. Biomed. Health Inform. 2017 22 766 774 10.1109/JBHI.2017.2679108 28287994 23. Prathyusha B. Rao T.S. Asha D. Extraction of respiratory rate from PPG signals using PCA and EMD Int. J. Res. Eng. Technol. 2012 1 164 184 24. Shuzan M.N.I. Chowdhury M.H. Hossain M.S. Chowdhury M.E.H. Reaz M.B.I. Monir Uddin M. Khandakar A. Mahbub Z.B. Hamid M. Ali S. A novel non-invasive estimation of respiration rate from motion corrupted photoplethysmograph signal using machine learning model IEEE Access 2021 9 96775 96790 10.1109/ACCESS.2021.3095380 25. Osathitporn P. Sawadwuthikul G. Thuwajit P. Ueafuea K. Mateepithaktham T. Kunaseth N. Choxtchawathi T. Punyabukkana P. Mignot E. Wilaiprasitporn T. RRWaveNet: A Compact End-to-End Multi-Scale Residual CNN for Robust PPG Respiratory Rate Estimation arXiv 2022 2208.08672 26. Lee S. Moon H. Son C.-H. Lee G. Respiratory Rate Estimation Combining Autocorrelation Function-Based Power Spectral Feature Extraction with Gradient Boosting Algorithm Appl. Sci. 2022 12 8355 10.3390/app12168355 27. Elgendi M. Optimal signal quality index for photoplethysmogram signals Bioengineering 2016 3 21 10.3390/bioengineering3040021 28952584 28. Gao P. Yue M. Wu Z. A novel intrusion detection method based on WOA optimized hybrid kernel RVM[C]//2021 IEEE 6th International Conference on Computer and Communication Systems (ICCCS) Proceedings of the 2021 IEEE 6th International Conference on Computer and Communication Systems (ICCCS) Chengdu, China 23-26 April 2021 1063 1069 29. Rafi M. Shaikh M.S. A comparison of SVM and RVM for Document Classification arXiv 2013 1301.2785 30. Zhang C. He Y. Yuan L. Xiang S. A multiple heterogeneous kernel RVM approach for analog circuit fault prognostic Clust. Comput. 2019 22 3849 3861 10.1007/s10586-018-2437-z 31. Mirjalili S. Lewis A. The whale optimization algorithm Adv. Eng. Softw. 2016 95 51 67 10.1016/j.advengsoft.2016.01.008 32. Bian D. Mehta P. Selvaraj N. Respiratory rate estimation using PPG: A deep learning approach Proceedings of the 42nd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC) Montreal, QC, Canada 20-24 July 2020 5948 5952 33. Aqajari S.A.H. Cao R. Zargari AH A. Rahmani A.M. An end-to-end and accurate ppg-based respiratory rate estimation approach using cycle generative adversarial networks Proceedings of the 2021 43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC) Mexico 1-5 November 2021 744 747 34. Kyriacou P.A. Allen J. Photoplethysmography: Technology, Signal Analysis and Applications Academic Press Cambridge, MA, USA 2021 |
PMC10000567 | In myelofibrosis, the C-reactive protein (CRP)/albumin ratio (CAR) and the Glasgow Prognostic Score (GPS) add prognostic information independently of the Dynamic International Prognostic Scoring System (DIPSS). Their prognostic impact, if molecular aberrations are considered, is currently unknown. We performed a retrospective chart review of 108 MF patients (prefibrotic MF n = 30; primary MF n = 56; secondary MF n = 22; median follow-up 42 months). In MF, both a CAR > 0.347 and a GPS > 0 were associated with a shorter median overall survival (21 [95% CI 0-62] vs. 80 months [95% CI 57-103], p < 0.001 and 32 [95% CI 1-63] vs. 89 months [95% CI 65-113], p < 0.001). Both parameters retained their prognostic value after inclusion into a bivariate Cox regression model together with the dichotomized Mutation-Enhanced International Prognostic Scoring System (MIPSS)-70: CAR > 0.374 HR 3.53 [95% CI 1.36-9.17], p = 0.0095 and GPS > 0 HR 4.63 [95% CI 1.76-12.1], p = 0.0019. An analysis of serum samples from an independent cohort revealed a correlation of CRP with levels of interleukin-1b and albumin with TNF-a, and demonstrated that CRP was correlated to the variant allele frequency of the driver mutation, but not albumin. Albumin and CRP as parameters readily available in clinical routine at low costs deserve further evaluation as prognostic markers in MF, ideally by analyzing data from prospective and multi-institutional registries. Since both albumin and CRP levels reflect different aspects of MF-associated inflammation and metabolic changes, our study further highlights that combining both parameters seems potentially useful to improve prognostication in MF. myelofibrosis C-reactive protein albumin CRP/albumin ratio Glasgow Prognostic Score MIPSS70 prognostication Research fund of the Cantonal Hospital St. Gallen21.35 Swiss National Science Foundation320030_189090/1 Novartis Foundation for medical-biological Research20C223 Swiss Cancer LeagueKLS-5158-08-2020 This study was supported by the research fund of the Cantonal Hospital St. Gallen, project ID 21.35. TNR. acknowledges the support from the Swiss National Science Foundation (320030_189090/1); The Novartis Foundation for medical-biological Research (20C223); Fond'Action contre le cancer (Lausanne); Swiss Cancer League (KLS-5158-08-2020). pmc1. Introduction Both primary and secondary myelofibrosis (PMF/SMF) are caused by a complex interplay of (epi)genetic alterations in hematopoietic stem cells and inflammatory changes, which affect hematopoiesis and impact patient survival . The Dynamic International Prognostic Scoring System (DIPSS) as a standard tool for prognostication considers age, anemia, leukocytosis, peripheral blast counts and constitutional symptoms . It can be refined by incorporating information about cytogenetic aberrations, mutational profile or both . In addition to these complex and expensive parameters, some routine laboratory markers add prognostic information, such as C-reactive protein (CRP) and albumin. Elevated CRP levels have been associated with several adverse disease features and a shorter leukemia-free survival , and albumin has been consistently shown to add additional prognostic information independently of DIPSS and several DIPSS-based prognostic scoring systems . Furthermore, indices combining CRP and albumin such as the CRP/albumin ratio (CAR) or the Glasgow Prognostic Score (GPS) provide DIPSS-independent prognostic information in MF. With regard to both CAR and GPS, it remains elusive as to whether they still add prognostic value if the molecular risk profile is considered. We therefore examined the prognostic impact of CAR and GPS in relation to the Mutation-Enhanced International Prognostic Scoring System (MIPSS)70, which includes the mutational profile without needing conventional metaphase cytogenetics . 2. Patient Population and Methods We performed a retrospective chart review of patients diagnosed with MF at the Cantonal Hospital St. Gallen between 2000 and 2020 (Cohort A). One hundred and eight patients were identified (47 female and 61 male, median age 72; pre-fibrotic MF: 30/108 (28%), PMF 56/108 (52%) and SMF 22/108 (20%)), and clinical and laboratory data were collected at the time of diagnosis and before commencement of treatment. All of the cases were reviewed individually, to ensure correct classification according to WHO2016 . If the diagnostic work-up did not include next-generation sequencing (NGS), we performed mutational profiling using material from the diagnostic samples (see the Supplementary Materials "Supplementary Methods"). Detailed patient characteristics of the cases with MF in cohort A (PMF and SMF) are shown in Table 1. The CAR was calculated by dividing the CRP concentration (mg/L) by the albumin concentration (g/L). The GPS was determined according to (GPS 0: albumin >= 35 g/L and CRP <= 10 mg/L; GPS 1: either albumin < 35 g/L or CRP > 10 mg/L; GPS 2: both albumin < 35 g/L and CRP > 10 mg/L). For CRP, we used the upper limit of normal from our local laboratory for dichotomization (<=/>8 mg/L), and for albumin, the median of our population was used (</>=40 g/L). For the CAR, we used a cut-off of </>=0.204, as proposed by and a CAR of </>=0.374, representing the fourth quartile of our cohort. The methods applied for the statistical analysis are described in detail in the Supplementary Materials. Plasma probes from an independent Canadian cohort (Cohort B) of 64 MPN patients (MF n = 28, PV n = 18, ET n = 18; Supplementary Table S1) and healthy controls (n = 16) were available to assess the correlation of high-sensitivity (hs)CRP and albumin levels with pro-inflammatory cytokines, which were measured as described in detail in the Supplementary Materials. 3. Results 3.1. Levels of CRP and Albumin, the CAR in Different MF Subgroups and Their Association with Disease Characteristics Within Cohort A, we found higher levels of conventional CRP in patients with MF (PMF: n = 56, median 5 mg/L, [IQR 2-18], SMF: n = 22, median 5 mg/L [IQR 3-9]) compared to pre-fibrotic MF (n = 30, median 1 mg/L, [IQR 1-8], p = 0.034). With regard to the albumin concentration, we found no difference (PMF median 40.5 g/L [IQR 37-42.6], SMF median 39 g/L [IQR 36.4-42.7], pre-fibrotic MF median 42 g/L [IQR 38-43.6], p = 0.253). In MF, a CRP-elevation > 8 mg/L was associated with lower levels of hemoglobin and platelets, a higher percentage of peripheral blasts, higher LDH-levels, transfusion-dependency and the presence of constitutional symptoms, whereas levels of albumin < 40 g/L were associated only with the degree of anemia and with a lower body mass index (BMI), as shown in Table 1. An additional comparison of disease characteristics following the cut-offs used within the GPS (CRP <=/> 10 mg/L and albumin </>= 35 g/L) is provided in Supplementary Table S2. There was no difference in CRP, albumin and the CAR between JAK2-mutated cases and CALR-mutated cases. With regard to JAK2-V617F variant allele frequency (VAF), we observed a significantly higher CAR in patients with a VAF > 50% (median 0.243 vs. 0.095, p = 0.035) and a trend towards higher CRP values (median 7.5 vs. 4.5 mg/L, p = 0.071). No difference was noted for albumin (median 39 vs. 38 g/L, p = 0.158). Patients with high-risk mutations according to MIPSS70 showed a tendency towards a higher CAR (median 0.579 vs. 0.115, p = 0.051) but did not differ significantly with regard to the single parameters. Further details are shown in Supplementary Table S3. MIPSS70 was available for 59/78 patients (76%): intermediate risk 43/59 (72.9%), high risk 14/59 (23.7%) and low-risk 2/59 (3.4%). Overall survival (OS) different significantly among these groups . Compared to the MIPSS70-intermediate patients, the MIPSS70-high-risk patients had significantly higher CRP levels (median 14 mg/L [IQR 5-30] vs. 5 mg/L [IQR 1-10], p = 0.012), but not lower albumin levels (median 38 vs. 39 g/L, p = 0.224). Accordingly, the CAR was higher in MIPSS70-intermediate patients (median 0.504 [95% CI 0.95-0.739] vs. 0.116 [95% CI 0.026-0.255,], p = 0.025). Given their low number, we did not include the MIPSS70-low risk group in this analysis. 3.2. Prognostic Impact of CRP, Albumin and Derived Indices (CAR and GPS) in MF The probability of death rose continuously with lower albumin levels even in the range determined as normal , and an albumin concentration below the population median was associated with a significantly shorter survival (albumin </>= 40 g/L, median OS 50 [95% CI 38-62] vs. 101 [95% CI 51-151] months, p = 0.026). CRP > 8 mg/L (n = 24) was associated with shorter survival compared to CRP within the normal limits (<=8 mg/L, n = 47): median OS 44 [95% CI 0-89] vs. 89 [95% CI 56-122] months, p < 0.001. Correspondingly, a higher CAR was associated with inferior survival (median OS CAR <=/> 0.204: 89 [95% CI 67-111] vs. 44 [95% CI 3-85] months, p = 0.001; and CAR <=/> 0.374: 80 [95% CI 57-103] vs. 21 [95% CI 0-62] months, p < 0.001). Similar results were obtained for patients with a GPS of 1 or 2 (n = 18) compared to patients with a GPS of 0 (n = 39): median OS 32 [95% CI 1-63] vs. 89 [95% CI 65-113] months, p < 0.001. Kaplan-Meier curves for the patients for whom both CRP and albumin were available (n = 57) are shown in Figure 1A-D. For all of the factors, a higher HR for mortality was observed in univariate Cox regression models (Table 2). Given the low number of MIPSS70-low-risk patients (n = 2), we dichotomized the cohort into a "MIPSS70dichlow/intermediate" risk group (n = 45) and a "MIPSSdichhigh" risk group (n = 14) for analyses in bivariate models. Here, CRP > 8 mg/L, albumin < 40 g/L, and both a CAR > 0.374 and a GPS > 0 retained their prognostic value together with MIPSS70dich, whereas a CAR > 0.204 did not (Table 2). In a separate analysis considering only the PMF patients (n = 35) and applying the same threshold for CAR (>0.374) and GPS (>0), the results remained significant, albeit with large 95% confidence intervals (Table 3). Of note, for SMF, the very low number of cases (n = 12) for whom both CRP and albumin were available precluded a separate analysis. For MIPSS70-intermediate patients with both CRP and albumin available (n = 35), OS was significantly shorter for albumin < 40 g/L, CAR > 0.374 and GPS > 0, whereas CRP <=/> 8 mg/L was not associated with an adverse prognosis . 3.3. Association of Levels of CRP and Albumin with Inflammatory Cytokines Analysis of cohort B showed higher levels of hsCRP (median 10.07 vs. 7.02 mg/L; p < 0.0004) and lower levels of albumin (median 31.4 vs. 25.87 g/L; p = 0.0012) in MF versus MPN without fibrosis and/or the healthy controls. The VAF of the driver mutation was correlated only to levels of hsCRP (p = 0.008) . The levels of interleukin-1b, interferon-g, CCL17, I-TAC and ENA-78/CXCL-5 correlated positively with hsCRP, while no significant correlation was observed for IL-6, TNFa, IFNa, IL-8, IL-18, IL-10, IL-33, IL-17a, IL-23 and MCP-1 . Albumin levels were inversely correlated to TNFa and MCP-1 . 4. Discussion CRP and albumin resemble surrogate markers for the extent of inflammation, a key element of MPN pathophysiology ( ). Higher CRP levels are known to be associated with shortened leukemia-free and overall survival in univariate analyses , whereas for albumin, a prognostic value independent of several DIPPS-based scoring systems has been described previously . As expected, we therefore found a significant impact of both parameters on survival in our cohort. Levels of CRP were more closely related to the established adverse features of MF, which are in part or indirectly taken into account by current models, e.g., peripheral blasts, more severe anemia and/or transfusion-dependency or thrombocytopenia < 100 x 109/L, whereas only lower albumin levels were associated with a lower BMI as a measure of MF-induced cachexia. In addition, both factors were associated with levels of different cytokines, namely CRP with interleukin-1b, a driver of MF pathogenesis , and albumin with TNF-a, a key mediator of cachexia . This implies that CRP and albumin probably reflect different aspects of MF pathophysiology. It is therefore of interest to combine them in the CAR or the GPS. For both parameters, a DIPSS-independent prognostic value has already been described in MF . A recent report on acute myeloid leukemia patients not eligible for stem-cell transplantation illustrates that a combined assessment of CRP and albumin is of interest in myeloid malignancies in general . We found a MIPSS70-independent prognostic value for both a CAR > 0.347 and GPS > 0. Hence, both parameters add prognostic information, even in the context of a molecular prognostic score. However, the relevant cut-off for the CAR used within our MIPSS70-based model was higher than that published for DIPSS-based prognostication . This might be due to different composition of the patient populations, different access to potentially disease-modifying drugs such as ruxolitinib or the influence of age, which is part of the DIPSS but not the MIPSS70. Further studies are needed to define the optimal cut-off of the CAR to be used in the context of the single different scoring systems and/or to decide whether CAR or the GPS provides better prognostic information. Malnutrition and/or activation of catabolic pathways leading to hypoalbuminemia are probably not sufficient to explain the prognostic impact of albumin, since levels still in the lower range of normal represent an adverse risk factor not only in our cohort, but also according to all of the reports currently available on the prognostic role of albumin in MF . Several pleiotropic effects of albumin have been described . Amongst others, it represents the main anti-oxidant in the extracellular space , and higher levels could be associated with an increased capability to counteract ROS-mediated inflammation, which is linked to disease progression in MF. This would indicate a vicious cycle, if inflammation has reached a point where albumin synthesis is limited. However, this hypothesis warrants confirmation in further studies. Considering albumin and CRP in clinical practice evidently helps to identify a more vulnerable population of MF patients who elude current prognostic models and could benefit from multimodal interventions. Both markers are associated with cardiovascular risk ; therefore, modifiable risk factors should be aggressively managed in MF patients with low albumin and elevated CRP levels and/or a higher CAR. The JAK2 inhibitor ruxolitinib controls not only constitutional symptoms and splenomegaly, but also lowers CRP levels and increases albumin concentration . This may justify its use even in low-risk patients harboring one of the risk factors based on CRP and albumin, especially if splenomegaly is already present. Non-pharmacological interventions, such as physical exercise and nutritional interventions, can positively affect both parameters . In this context, the Mediterranean Diet is currently under investigation in MF . As this was a monocentric and retrospective study, the interpretation of our observations is subject to several limitations. Apart from a potential selection bias, the limited number of patients is most relevant, since it precludes defining the cut-off of the CAR that is best suited for prognostication or adjusting for possibly confounding factors such as age and treatment with disease-modifying drugs such as ruxolitinib. Due to the low number of patients, we had to combine cases of primary and secondary MF. Whether prognostic scores established for PMF are of value for patients with SMF is still a matter of debate , and the Myelofibrosis Secondary to PV and ET-Prognostic Model (MYSEC-PM) was developed especially for this population . However, the MYSEC-PM does not consider the presence and type of additional non-driver mutations; hence, the MIPSS70 represents one of the currently suggested tools for prognostication in both PMF and SMF, if the mutational profile has to be considered . A further limitation is the fact that conventional metaphase cytogenetics were available only for a minority of patients, precluding the assessment of the factors studied in the context of scoring systems, which consider chromosomal aberrations in addition to the mutational profile, such as the MIPSS70+ Version 2.0 . 5. Conclusions Our data have shown for the first time that CAR and GPS add prognostic information independently of the MIPSS70-based molecular risk profile in MF. Albumin and CRP are easily available in clinical routine at low cost and represent potential biomarkers to faithfully identify a more vulnerable population of MF patients not identified by current prognostic model systems. Moreover, since CRP and albumin probably reflect different aspects of MF pathophysiology, including inflammation and metabolic aspects, combining both parameters seems particularly useful for MF prognostication. However, further studies involving multi-center registries with larger cohorts are necessary to validate the prognostic impact of albumin and CRP within the context of prognostic scoring systems considering both cytogenetics and the mutational status. In addition, it remains to be determined as to whether improving levels of CRP and albumin during therapy are associated with a better prognosis. Despite all limitations, our observations fit well into the emerging data and support the prognostic role of albumin and CRP in MF. Supplementary Materials The following supporting information can be downloaded at: Figure S1: Overall survival of MF patients from Cohort A according to MIPSS70; Figure S2: Probability of death according to albumin concentration as a continuous variable; Figure S3: Elevated CRP-levels in MPN and its association with MPN-subtypes and disease progression; Figure S4: Reduced albumin levels in MPN and its association with MPN-subtypes and disease progression; Figure S5: Correlation plots of inflammatory cytokine panel and CRP levels in MPN patients; Figure S6: Correlation plots of inflammatory cytokine panel and CRP levels in MPN patients; Figure S7: Correlation plots of inflammatory cytokine panel and albumin levels in MPN patients; Figure S8: Correlation plots of inflammatory cytokine panel and CRP levels in MPN patients; Table S1: Patient characteristics of Cohort B ; Table S2: Patient characteristics of patients with MF from Cohort A including PMF (56/78) and MF post ET/PV (22/78) according to levels of CRP and albumin as used in the Glagow prognostic score; Table S3: Levels of albumin, C-reactive protein (CRP) and CRP/albumin ratio (CAR) according to molecular characteristics in patients with MF from Cohort A. Click here for additional data file. Author Contributions Conceptualization, T.N.R. and T.S.; Formal analysis, N.-M.M., N.R.U., T.V., T.N.R. and T.S.; Funding acquisition, T.N.R. and T.S.; Investigation, N.-M.M. and N.R.U.; Methodology, N.R.U., T.V., S.C., W.J. and T.S.; Project administration, T.S.; Resources, N.-M.M., S.C., T.L., A.H., R.B., L.G., V.G., I.D., W.J., T.N.R. and T.S.; Visualization, N.-M.M., N.R.U., T.V., T.N.R. and T.S.; Writing--original draft, N.-M.M.; Writing--review & editing, N.R.U., T.V., S.C., T.L., A.H., R.B., L.G., V.G., I.D., W.J., T.N.R. and T.S. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the Ethics Committee of Eastern Switzerland (Ethikkommission Ostschweiz, Project ID 2020-01371) and by the Research Ethics Board of the Princess Margaret Cancer Center, University of Toronto, Toronto, Canada (REB 01-0573). Informed Consent Statement Written informed consent was obtained from each patient alive before inclusion into the study. In accordance with the vote of the ethics committee, data of patients deceased or lost to follow-up were included into the dataset, in accordance with Paragraph 34 of the Swiss Law regulating research with human subjects (Humanforschungsgesetz). Data Availability Statement The data presented in this study are available on request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Prognostic significance of CRP, albumin and the CRP/albumin ratio in MF as assessed by univariate analyses. Survival of patients with MF, for whom both albumin and CRP were available (n = 57), stratified by albumin </>= 40 g/L (A), CRP <=/> 8 mg/L (B), CRP/albumin ratio <=/> 0.374 (C) and GPS </> 0 (D). Figure 2 Albumin and CRP provide MIPSS70-independent prognostic information in MF. Survival of MIPSS70-intermediate-risk patients, for whom both albumin and CRP were available (n = 35), stratified by albumin </>= 40 g/L (A), CRP <=/> 8 mg/L (B), CRP/albumin ratio <=/> 0.374 (C) and GPS </> 0 (D). cancers-15-01479-t001_Table 1 Table 1 Characteristics of patients with MF from in Cohort A including PMF (56/78) and MF post ET/PV (22/78). Data are shown for the whole population and according to the levels of CRP and albumin. Whole Population CRP <=8 mg/L CRP >8 mg/L p Albumin >= 40 g/L Albumin < 40 g/L p n 78 47 24 31 30 Age [years], median, (IQR) 72 (60-78) 70 (60-77) 76 (63-80) 0.068 69 (60-77) 76 (67-79) 0.075 Female n, (%) 37 (46.2) 24/47 (51) 10/24 (41.7) 0.616 14/31 (45.2) 16/30 (53.3) 0.612 Bone marrow fibrosis grade 2, n (%) 54/78 (70) 37/47 (78.7) 14/24 (58.3) 25/31 (80.6%) 20/30 (67) Bone marrow fibrosis grade 3, n (%) 24/78 (30) 10/47 (21.3) 10/24 (41.7) 0.096 6/31 (19.4) 10/30 (33) 0.255 Hemoglobin [g/L], median (IQR) 107 (88-122) 117 (102-131) 86 (77-103 <0.001 115 (103-130) 96 (80-113) 0.003 Platelet count (x109/L) median (IQR) 410 (197-663) 550 (340-773) 255 (110-442) 0.001 541 (200-770) 391 (231-576) 0.329 Leukocytes (x109/L), median (IQR) 8.9 (6.0-15.6) 9.5 (6.8-16) 8.7 (5.5-20) 0.551 9.4 (6.8-16) 10 (6.3-21) 0.751 Neutrophils (x109/L), median (IQR) 6.15 (3.8-12.8) 6.8 (4.3-13.4) 6.15 (2.7-1.4) 0.378 6.1 (4.3-13) 7.25 (3.3-14.9) 0.902 Monocytes (x109/L), median (IQR) 0.56 (0.33-0.84) 0.65 (0.38-0.83) 0.44 (0.28-0.87) 0.397 0.54 (0.36-0.82) 0.56 (0.29-0.85) 0.813 Blasts PB (%), median (IQR) 0 (0-1) 0 (0-1) 1 (0-2) 0.017 0 (0-1) 0 (0-1) 0.605 Constitutional symptoms, n (%) 36/78 (46) 16/47 (34) 16/24 (66.7) 0.012 12/31 (38.7) 19/30 (63.3) 0.074 LDH available, n (%) Median [U/L] (IQR) 69/78 (88) 525 (347-700) 41/47 (87) 457 (329-606) 23/24 (96) 609 (463-932) 0.043 30/31 (97) 541 (365-829) 26/30 (87) 541 (321-686) 0.730 CRP available, n (%) Median [mg/L] (IQR) 71/78 (91) 5 (2-12) 47/47 (100) 3 (1-5) 24/24 (100) 21 (11-35) 28/31 (90) 4 (1.25-6) 29/30 (97) 10 (3.5-24.5) 0.005 Albumin available, n (%) Median [g/L] (IQR) 61/78 (78) 40 (37-43) 36/47 (76) 42 (39-43) 21/24 (88) 37 (35-38) <0.001 31/31 43 (42-44) 30/30 37 (35-38) CAR available, n (%) Median (IQR) 57/78 (73) 0.128 (0.051-0.374) 36/57 (63) 0.073 (0.0263-0.125) 21/24 (87) 0.579 (0.315-0.808) <0.001 28/31 (90) 0.093 (0.029-0.142) 29/30 (97) 0.263 (0.094-0.727) 0.001 Need of transfusion, n (%) 12/78 (15) 2/47 (4.3) 8/24 (33.3) 0.002 3/31 (9.7) 6/30 (20) 0.301 Platelets < 100 x 109/L, n (%) 5/77 (6.5) 0/47 (0) 4/24 (17) 0.011 1/30 (3) 2/30 (17) 1.00 Splenomegaly (clinically or imaging), n (%) 63/78 (81) 37/47 (79) 20/24 (83) 0.759 25/31 (80.6) 26/30 (86.7) 0.731 BMI, available, n (%) Median (kg/m2) (IQR) 72/78 (92) 24.5 (21-28) 42/47 (89) 24.4 (21.1-28.3) 23/24 (96) 26 (22.0-28.2) 0.484 27/31 (87) 26.4 (22.9-29.2) 30/30 22.1 (20.4-26.1) 0.008 Driver Mutations JAK2-V617F (n, %) 46/78 (59) CALR (n, %) 16/78, (20.5) MPL (n, %) 3/78, (3.8) Triple negative (n, %) 5/78, (6.4) Unknown * (n, %) 8/78, (10.3) IQR, interquartile range; CAR, CRP/albumin ratio; BMI, body mass index. * cases diagnosed between 2000 and 2009 and no DNA available for retrospective analyses. cancers-15-01479-t002_Table 2 Table 2 bivariate Cox regression models including MIPSS70dich, albumin, CRP, CRP/albumin ratio [CAR] and Glasgow Prognostic Score [GPS] for all cases including primary and secondary myelofibrosis. Univariate Bivariate n HR 95% CI p n HR 95% CI p MIPSS70dich 59 4.90 1.99-12.0 0.001 56 3.45 1.28-9.32 0.0148 CRP > 8 mg/L 71 3.85 1.85-8.0 <0.001 2.50 1.13-5.52 0.0236 MIPSS70dich 50 8.65 2.87-26.07 <0.001 Albumin < 40 g/L 61 2.49 1.13-5.49 0.024 5.49 1.89-15.96 0.0018 MIPSS70dich 47 4.86 1.99-11.88 0.0005 CAR > 0.204 57 1.84 1.01-3.34 0.046 1.37 0.66-2.84 0.4026 MIPSS70dich 47 5.98 1.84-19.46 0.0030 CAR > 0.374 57 4.25 1.75-10.32 0.001 3.53 1.36-9.17 0.0095 MIPSS70dich 47 6.35 1.95-20.73 0.0022 GPS > 0 57 5.38 2.17-13.37 <0.001 4.63 1.76-12.1 0.0019 cancers-15-01479-t003_Table 3 Table 3 bivariate Cox regression models including MIPSS70dich, albumin, CRP, CRP/albumin ratio [CAR] and Glasgow Prognostic Score [GPS] for primary myelofibrosis only. Univariate Bivariate n HR 95% CI p n HR 95% CI p MIPSS70dich 45 6.26 2.19-17.89 0.0006 42 4.21 1.40-12.65 0.0104 CRP > 8 mg/L 52 3.86 1.52-9.76 0.0044 2.12 0.77-5.87 0.146 MIPSS70dich 37 9.92 2.46-40.0 0.0013 Albumin < 40 g/L 48 2.12 0.91-4.92 0.0823 4.89 1.15-20.78 0.0317 MIPSS70dich 35 10.15 2.61-39.44 0.0008 CAR > 0.204 44 4.06 1.41-11.66 0.0093 2.71 0.85-8.64 0.0923 MIPSS70dich 35 8.33 2.09-33.18 0.0026 CAR > 0.374 44 3.88 1.39-10.80 0.0094 3.38 1.09-10.50 0.0353 MIPSS70dich 35 9.83 2.22-43.60 0.0026 GPS > 0 44 4.60 1.61-13.18 0.0044 4.32 1.33-14.02 0.0148 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000568 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050657 healthcare-11-00657 Article Choosing Alternative Career Pathways after Immigration: Aspects Internationally Educated Physicians Consider when Narrowing down Non-Physician Career Choices Chowdhury Nashit 123 Lake Deidre 3 Turin Tanvir C. 12* Giansanti Daniele Academic Editor 1 Department of Family Medicine, Cumming School of Medicine, University of Calgary, Calgary, AB T2N 4N1, Canada 2 Department of Community Health Sciences, Cumming School of Medicine, University of Calgary, Calgary, AB T2N 4N1, Canada 3 Alberta International Medical Graduates Association, Calgary, AB T2E 3K8, Canada * Correspondence: [email protected] 23 2 2023 3 2023 11 5 65727 12 2022 19 2 2023 21 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Many developed countries admit internationally educated physicians (IEPs) as highly skilled migrants. The majority of IEPs arrive with the intention of becoming licensed physicians to no avail, resulting in underemployment and underutilization of this highly skilled group of people. Alternative careers in the health and wellness sector provide IEPs opportunities to use their skills and reclaim their lost professional identity; however, this path also includes great challenges. In this study, we determined factors that affect IEPs' decisions regarding their choice of alternative jobs. We conducted eight focus groups with 42 IEPs in Canada. Factors affecting IEPs' career decisions were related to their individual situations and tangible aspects of career exploration, including resources and skills. A number of factors were associated with IEPs' personal interests and goals, such as a passion for a particular career, which also varied across participants. Overall, IEPs interested in alternative careers took an adaptive approach, largely influenced by the need to earn a living in a foreign country and accommodate family needs and responsibilities. internationally educated physicians alternative career pathway career choice decision factors international medical graduates Social Sciences and Humanities Research Council, Canada892-2019-2004 This study was supported by a grant from the Social Sciences and Humanities Research Council, Canada (892-2019-2004). pmc1. Introduction Internationally educated physicians (IEPs) are individuals who have graduated from medical schools located outside of Canada . They are also known as foreign medical graduates (FMGs), international medical graduates (IMGs), or internationally trained physicians (ITPs) . Most IEPs are immigrants, refugees, or temporary migrants who moved to Canada after completing their medical graduation. Others are Canadian-born citizens who studied medicine abroad . While these IEPs may have had successful medical careers in the countries they trained in, they often become underemployed in Canada due to demanding and resource-intensive licensing procedures. Most IEPs cannot re-enter their intended career to become practicing physicians in Canada because of numerous individual and systemic barriers . Generally, only 20-30% of IEPs are able to obtain a residency training spot, which restricts and hinders IEPs from obtaining physician licensure in Canada . Many IEPs, thus, have become a symbol of the deskilling of highly skilled migrants in developed countries . The low success rate of becoming a practicing physician in Canada causes frustration about life and career prospects in Canada, as well as financial strain leading, to being unable to bear family responsibilities . This drives IEPs to look for viable alternate employment options based on their background . In general, alternative careers are the "career options that immigrants pursue other than but related to the regulatory profession in which they were originally trained, that make use of and relate to an immigrant's skills and experience" . For IEPs, alternative careers may be defined as those jobs in the health and wellness sectors that utilize their medical skills and knowledge. A Canada-wide survey of 1740 participants found that 68.78% of employed unlicensed IEPs worked in health-related alternative professions, with the majority (50.45% of total employed) working in non-regulated professions . However, a notable proportion (31.22%) were working in non-health-related fields . As such, another nationwide survey of 356 unlicensed IEPs reported that the majority of the IEPs were dissatisfied (61%) with their current alternative professions and wished to have greater government and community support (93%) for the development of alternative careers for IEPs . The IEPs also reported that their years of effort and exhausting of resources after licensing processes did not help in their current alternative profession as well . The intent of an alternative career is not to start a career from scratch, but rather to find a position where IEPs can build on the education and training they already have. The non-recognition of all the qualifications and experiences of the IEPs and the lack of systemic support make the alternative career pursuit very challenging . Less than 10% of IEPs currently employed in alternative careers accessed government-supported career-related training or education resources . Having spent a good portion of their life training to become a physician, many IEPs have never considered doing anything else, are unsure of how their skills and knowledge base might be applied to other areas, and are uncertain how to prepare for alternative careers, including obtaining information and accessing resources . Further, employers are unaware of this situation and IEPs' potential for non-physician health and wellness-related jobs; they do not understand why a person trained to become a physician would want to do anything else . The lack of information, organizational support, job counseling, and coaching further hinders IEPs' pursuit of alternative careers . The pathways to the professional integration of IEPs in Canada are already very complex, with multiple routes, requirements, and obstacles . Learning about, deciding on, and ultimately pursuing potential alternative careers adds to that complexity. As such, studies have reinforced the urgent need to develop alternative career guides and supports . While previous studies provided certain hints of the determining factors for IEPs (e.g., salary and further course requirements) in terms of choosing an alternative career came up as a part of the discussion focused on a different objective , none of these studies specifically examined the decision-making process of IEPs around alternative careers. Therefore, through this exploratory study, we planned to investigate the decision-making process of IEPs regarding their pursuit of alternative careers and aimed to identify factors IEPs need to consider when seeking alternative careers. 1.1. Prior Research and Significance of the Study Canada continues to bring skilled migrants to health and other professional sectors to compensate for the aging population and low population growth . Existing literature has discussed various issues of skilled migrants in general that are primarily driven by immigrant-serving organizations in collaboration with some academics, often those with immigrant or refugee backgrounds . This includes un/underemployment of skilled migrants and the economic losses from this phenomenon , barriers and facilitators of credential recognition , work environment , perceived discrimination , and policies around skilled migration . People with foreign-sounding names were 20-40% less likely to receive a call from human resources for job interviews , which forces them to be employed in low-skilled jobs for survival. Reitz et al. (2014) showed that deskilling causes $11 billion per year in economic losses in Canada. Moreover, this exploitation of skilled migrants affects their physical and mental health and adds more to the loss of human capital . The issue is more prominent for regulated jobs. Augustine compared regulations and policies for alternative routes to the primary careers of skilled migrants of three professions between Canada and Australia: accounting, engineering, and medicine. There are some acceptable alternative routes; however, they are not sufficient for the IEPs. This study reported that certain occupations such as physicians have more need for suitable alternative careers because their training is highly specific. While the use of supervised practice as physicians reduced the bottleneck for IEPs in Australia as opposed to Canada, however that may not result desirable outcomes in a competitive labour market. Alternative careers provide flexibility in integration of skilled immigrants into the labour market by allowing different ways to demonstrate competency or fill in gaps in education without redoing the whole degree or training. Professions like IEPs this is very crucial as in Canada most of the IEPs has to go through the full residency training to enter their primary profession . There is scarce literature regarding alternative career pathways for IEPs in the health and wellness sector . As one study noted, IEPs are on average 10 years older than Canadian medical school graduates and may have more obligations and responsibilities to their families . They have smaller professional and social networks and greater financial constraints, which may affect their ability to successfully pursue a new career . IEPs often wait too long to take steps toward alternative careers due to a lack of information, uncertainty, and fear of losing professional status . While an alternative career is not what they have dreamed of and are prepared for, it offers them a way to reclaim their professional identity and earn a respectable living . A study that interviewed IEPs and seven other professionals reported that their IEP participants suggested a toolbox that would help IEPs understand alternative job options and pursue them . By identifying the thought processes of IEPs around alternative career pursuits, this study's findings could be used to develop a tool that will enable IEPs to find suitable alternative careers and regain their professional identity and economic losses to a certain extent. 1.2. Theoretical Underpinnings Human Capital Theory (HCT) has been the foundation of Canadian immigration policies in recent decades, which welcomes highly qualified migrant professionals based on their skill, education, and experience and intends to assimilate them back into their field of expertise following migration . During the immigration process, most of them come in skill level category A or B (post-secondary advanced degree and experiences), determined by the Immigration, Refugees, and Citizenship Canada (IRCC). Unfortunately, contrary to the intent, the majority of them are forced to work in skill-level C or D jobs that do not need advanced degrees or expertise, such as taxi driving, retail and service industries, and so on. Moreover, employers of potential alternative careers often do not recognize IEPs' educational qualifications and experience, preferring local Canadian credentials and experience. . Paul and colleagues identify that this scenario has been happening due to the disconnection between the policies of the federal, provincial, and local levels of governance. While at the federal level, policies recognize the educational training and experience of IEPs and welcome them into the country as skilled migrants, the provincial level often does not recognize their education and experiences . This disconnection becomes worse at local levels where some employers fail to understand the level of skill and knowledge of the IEPs. Others, on the other hand, may reject them for non-physician health and wellness career (i.e., alternative) positions, considering them overqualified recognizing as IEPs, which is attributable to unawareness of the long, complex, and very uncertain physician licensure process for IEPs compelling IEPs to look for alternative careers. This complicated process results in the loss of recognition of the human capital of IEPs in the post-migration period . It is important to work for the re-recognition of IEPs' human capital in the post-migration period at all levels. In this study, we look into the IEP perspectives on pursuing alternative careers, which we believe is crucial to the development of any support programs for the facilitation of alternative career pursuits for IEPs. We also draw on intersectionality theory in understanding the alternative career decision-making factors among IEPs. Intersectionality allows the understanding of phenomena through multiple and overlapping lenses of intersecting identities such as gender, race, socioeconomic, and immigrant origins . The IEPs come from diverse socio-economic backgrounds and may have experienced a range of discrimination during their attempt to professionally integrate into Canada. Their socio-economic and individual determinants may influence their mechanism of the decision-making process around alternative careers. For example, women IEPs from certain cultural backgrounds may prioritize taking care of their families over a career and look for an alternative career with a more flexible work schedule or only part-time positions . 2. Methods As this study focused on alternative career pathways for IEPs where there has been very little previous empirical work, we chose to adopt an exploratory qualitative descriptive approach to the methodology . This study employed focus group discussions (FGDs) as a tool for data collection from IEPs. Due to COVID-19 restrictions at the time of data collection, the FGDs were conducted via Zoom, an online meeting software (Zoom Video Communications, Inc., San Jose, CA, USA), using a secured account provided by the University of Calgary. This also offered us the benefit of recruiting participants from across Canada. All of the selected IEPs were either already working in an alternative career or had expressed an interest in pursuing an alternative career and had similar backgrounds in terms of being an IEP and having experience of facing myriad barriers to their career pursuit, which for the purposes of our study constituted a homogenous group. The barriers were explored by a different research question in the same setting, which has been published elsewhere . 2.1. Recruitment and Participants Purposive sampling was used to recruit participants for our study. This is a non-probabilistic sampling where participants are selected by the researcher based on their knowledge, experience, and ability to expand on a certain topic, theme, or phenomenon . Despite there being a chance of researcher bias, this sampling method is used in qualitative research to maximize the relevance of collected data to the research objectives . Despite the chance of selection bias, we chose this method to ensure that only the correct participants (i.e., those who were really into an alternative career, not just having it as a thought) were recruited. Ethical approval was granted by the Conjoint Health Research Ethics Board (CHREB) at the University of Calgary. Our study included a total of 42 participants across eight focus groups. Most participants were either Canadian citizens (naturalized) or permanent residents, and only two were temporary residents (Table 1). A total of 25 participants were employed and most of them (20 participants) worked in health-related alternative careers. Among the 17 unemployed participants, 14 were actively seeking work at the time of the focus group discussion. Most of the participants were female (31 vs. 11), and most belonged to the age groups 30-39 and 40-49 years (17 and 14 participants, respectively). 2.2. Data Collection Each of the eight focus groups comprised four to seven participants. The FGDs were facilitated using a semi-structured questionnaire that was developed by the research team and reviewed by two IEP citizen researchers. Each FGD was audio-recorded using Zoom recording options and transcribed verbatim. All FGDs were conducted in English. Each session lasted from one hour to one and a half hours. 2.3. Data Analysis We adopted an inductive thematic analysis approach to analyze the data . We exported the transcriptions to NVivo qualitative data analysis software (QSR International, Version 12, Melbourne, VIC, Australia), which was used to generate codes and themes. At first, NC coded the data from the transcription of the first three FGDs and then met with the other members of the team to examine the coding for appropriateness and biases. Following a discussion with the team, the initial coding scheme was rectified. NC then continued coding the rest of the focus group transcripts, and the team reconvened after completion of the coding of all eight focus groups (which yielded 21 initial codes). The team discussions led to the finalization of all codes, followed by sub-themes and themes. The validity of the data was determined by several procedures, including through the lens of the researcher, the study participants, and IEPs external to the study. In addition, a representative sample of the participants member-checked the quotes and the findings. 3. Results The analysis of the focus group discussions revealed five themes regarding determining factors IEPs consider while exploring alternative career options. The first theme, "Qualification and experience requirement", included the requirement of a certain number of hours of Canadian voluntary or paid work experience, certification, and certain undergraduate or postgraduate degrees/diplomas. The second theme was "Personal resource requirement for capacity building", which entailed the time and cost required to be competitive for a particular job. The third theme, "Possibility of the utilization of transferable skills", included factors related to the extent of transferability of the skills earned in medical schools and experience as physicians. The fourth theme, "Employment-related factors", represented those factors that arise from being employed in alternative careers, such as financial compensation, opportunities for further growth within the job, and others. The final theme, "Personal factors", addressed intangible factors arising from personal preferences and constraints. 3.1. Theme 1: Qualification and Experience Requirement 3.1.1. Sub-Theme: Qualification Requirement Participants indicated that one of the first things they would consider in choosing an alternative job was the qualifications required for that job. "Yeah, sure. I'm actually very, um, enthusiastic about learning. So I am doing the clinical research certificate and I want to work and I don't mind going back to school. I really want to be settled here, so I don't mind going back to school." (FGD4P5), FGD: Focus Group Discussion; P5: Participant No. 5. However, their views differed regarding the need for the qualification requirement. While some were happy to take up further training/institutional education for an alternative career, others were not interested in further schooling. Some would consider additional training, but they would prefer programs that are quick and easy and that complement their backgrounds. "Uh, the thing that I am not open to is, as I mentioned before that I am at this point. Yeah. I'm not looking into going to school or going to upgrade myself. Uh, I think that it's, it's not in me. I lost that desire to do it. So I'm looking for something that can give me an opportunity to work now, instead of asking me to go back to school for one or two years, um, the otherwise I'm, I'm, uh, I can't think of anything in particular. Uh, other than that." (FGD4P2) A few participants also mentioned that they would look into whether an English language proficiency test (such as the International English Language Testing System (IELTS)) and/or other credential evaluation tests (such as the Graduate Record Examinations (GRE)) was required as prerequisites to pursue further education for alternative careers. "So, so first off for the preparatory phase, I mean to say that certain courses require extensive qualification, uh, for example, um, like IELTS score GRE score experiences. Uh, so just to get into that course, I think that, yeah." (FGD8P3). 3.1.2. Sub-Theme: Experience Requirement Participants also said that they tried to narrow down their career choices based on the probable importance of Canadian volunteer or paid work experience for that specific type of job. Many participants had the perception that employers preferred the Canadian experience and did not recognize their extensive experience from back home or in other countries. "You actually have the skills, you have the certification, but they still want you to have had the Canadian experience, which still makes it tedious. Even with this medical office administration. The fact that I even went to school here, they still want me to have had two years experience." (FGD3P3). 3.2. Theme 2: Personal Resource Requirement for Capacity Building 3.2.1. Sub-Theme: Time Requirement The time required to become prepared for a job by earning a degree, going through training, and accumulating volunteer experience was voiced by many participants. Participants were keen to develop and be eligible or competitive candidates for a career, either by taking short-term training (e.g., 6 months to 1 year) or enrolling in programs that were often two or more years in length. "I will think about the time factor. So I, um, I can take any program or any, uh, on any course for a short time period. Um, maybe up to a year or two years maximum, uh, until I finished my studying and exams on so much, but more than that, I think it." (FGD4P6). Length of time varied according to their personal choices and interest and was influenced by their family responsibilities, financial circumstances, and career goals. 3.2.2. Sub-Theme: Cost of Required Courses Participants also recognized that certain career pathways might require them to undergo additional courses or retraining. The cost related to this capacity was also considered a factor in their decision of an alternative career. "...you know, first and foremost, what I think is, uh, what is the cost of that course? Have to do, if I have to do a course and, uh, if I need, you know." (FGD1P1) However, the cost might be compared to the potential outcome of the job in terms of job certainty and financial outcome. "That's not a problem for me, if it is needed, I'm also. Ready to invest for a course or whatever, but really the ever-changing Canadian and [muffled] landscape will be able to provide me with a job once I come out of it." (FGD1P1) 3.3. Theme 3: Possibility of the Utilization of Transferable Skills 3.3.1. Sub-Theme: Skills from Previous Specialty or Additional Professional Experience IEPs in general showed interest in finding an alternative job that would utilize their clinical skills and knowledge. IEPs looked for these jobs because they felt it would require less effort and fewer additional resources for them to obtain the required knowledge and skills for those alternative jobs that suited the medical field. "Third thing, uh, trying to select something, that near to the medical field. Okay. So that it will not take lots of money from the person who is trying, for example, to study or to prepare to get in that field. And it will not be time-consuming the same time." (FGD8P4) Another purpose of participants looking for careers utilizing clinical skills and knowledge was to ensure continuous use of such skills, which would ultimately benefit participants when pursuing medical licensure in Canada in the long term, while pursuing alternative careers to meet immediate goals and needs. This was reiterated by those participants who had specialty training or additional qualifications such as a certification in Diagnostic Sonography. Participants intended to use these qualifications and experiences to guide their career search to a position where they could employ such experiences. "...since I have done the ARDMS (American Registry for Diagnostic Medical Sonography). So, my first priority will be, when I think about the alternative career, it will be something related to this field. Like radiology or something, I know that's really hard, but that will be my first priority and something related to patient." (FGD2P6) 3.3.2. Sub-Theme: Non-Specific Transferable Skills Some participants pointed out that as a physician they also had experience with certain essential skills such as managing, interpersonal skills, communication skills, and administrative skills. They would also consider these when shortlisting their alternative career choices. "So I have been working with people, managing people, a part of work that I have done before. Also, it has given me those skills to work with. So. People rather than just dealing with patients." (FGD4P2) 3.4. Theme 4: Employment-Related Factors 3.4.1. Sub-Theme: Wages or Financial Outcome The salary/wages of jobs were a crucial factor for the pursuit of alternative careers, which was a common theme that arose from the discussions. "So financial outcome. Yes, remuneration. Actually, what you're going to get, it should be sufficient enough to maintain dignity as well as your life." (FGD1P3) Most participants wanted a career that paid a decent amount; however, depending on other factors, such as growth opportunity, they might waive the priority of financial gain. "Yeah, top three for me would be one would be income, um, location and three would be, um, how close is it to my actual work? So income, location, and the third one, the closeness to my role as a doctor." (FGD4P3) 3.4.2. Sub-Theme: Flexibility in Working Hours Many participants were concerned about the working hours for the job. It was often expressed as a surrogate for flexibility. Some participants were interested in part-time jobs to give more time to their families or have time to study for Canadian medical licensure and other long-term career options. "That I might hold some kind of job, which is not too hectic to have a family and, um, to, to, to give time to the husband, to the kids. And you know, it's not like, uh, a difficult, you know, like, well, flexible hours relaxing, but it's still some, you know, some kind of income, basic income. (FGD2P5) "...and I consider that is, working hours. Do I have to work on weekends? I have to do nights, or it's just weekdays?" (FGD4P3) 3.4.3. Sub-Theme: Opportunity for Growth Another important employment-related factor was the opportunity for growth and career progression within an alternative career. Some participants expressed not wanting a job with no opportunity for promotion to higher positions accompanied by salary raises. "I needed to be in a, in a profession or something that I would like, can you going on and what I can grow and just continue and feel satisfied with it." (FGD3P2). When asked to elaborate on how participants would know if the job had growth opportunities, participants mentioned they would explore to see whether the position had mentorship or job coaching programs, which may be indicated in the job description. "Yep. Um, so for the growth opportunity, I would say, you know, it might not be even again, a tangible thing that a lot of time actually come around with assistant development opportunities are provided or, you know, a mentorship is provided or coaching job coaching is private or something like that." (FGD8P5). One participant, however, suggested that information on growth opportunities was not usually found to be documented, and participants had to communicate with individuals already employed by the organization to explore this aspect of the position. "Initially they said that there was no opportunity for growth, but then as I was in the job, I saw a lot of opportunities as I went through it. Cause like, um, I had the opportunity to be part of research studies as well as our medical lab technician. So I was in contact with a lot of physicians, like in Alberta Children's Hospital. And then at the same time now at the South health campus, we do a lot of research studies. And then there's also this opportunity for more of the administrative part, like knowing the ins and outs of the laboratory, how it works." (FGD5P5). 3.4.4. Sub-Theme: Job Demand and Availability Participants reported that they often explored the market demand and long-term outlook of a career before pursuing it. This was especially the case when educational training was required, where participants were more concerned about the availability of jobs within and the sustainability of a career once they had invested their time, effort, and money to become qualified. "Cause whenever I want to train on something, I want to train it good, like a hundred percent perfection, but the money that is required to do the training, if it has to come from me and I would want to have some kind of, um, reassurance that I'm going to get a job at the end." (FGD2P5). 3.4.5. Sub-Theme: Networking Opportunity Some participants indicated that they would investigate whether there was a networking opportunity within an alternative career. To elaborate, participants mentioned that they would look to see whether there was an opportunity for collaborative work with other institutions or organizations so that the professional network could be widened. "Um, yeah, many of the research groups that I've known I've worked with or know about they work in collaborations. So there's a lot of opportunity to work with different stakeholders and different collaborators, even though they are different than universities that are collaborating together on a project." (FGD6P1). One participant mentioned that working in a team environment might be a good opportunity to widen their personal professional network. "So if you're working in a team environment, you're working with like, you know, 10 or 15 different peoples, you will get to know them. There is a networking opportunity, for sure." (FGD8P2) However, participants did feel that the ability to successfully network was largely due to the personal competencies of the IEPs to build and maintain professional relationships. "Yeah. You can't give it right. Like you can't dictate it. Uh it's because honestly, um, but from person to person, the networking ability differs." (FGD8P1) 3.4.6. Sub-Theme: Work Environment When seeking jobs, participants prioritized workplaces that possessed a collaborative attitude, gave importance to stress management, had effective leadership from management, and had good staff benefits and support. (From the Zoom chat feature) "Collaborative work environment, amount of stress and what do the employer (do) to manage staffs' stress. One of the proxy could be what proportion of people are working in the specific workplace for what duration. How much they appreciate their staffs for the work they do. Paid vacation, sick leaves, family leaves, paid holidays, etc." (FGD8P5). 3.5. Theme 5: Personal Level Factors 3.5.1. Sub-Theme: Passion or Interest for Certain Jobs or Work Area Certain participants mentioned that the alternative career had to be something that was meaningful to them and satisfied them as an individual. Participants elaborated to say that careers should be something that involves patient engagement and care and decision-making capacity regarding patient care, something that is an important attribute in practicing physicians. "Like radiology, or something, I know that's really hard, but that will be my first priority and something related to patient. Actually, since I, I worked solely through the clinical side. I don't know. Is it possible or what, but if I got chance to work with the patient, uh, I will, I will definitely go through that, that being my first priority." (FGD2P6) 3.5.2. Sub-Theme: Alignment with Future Goals Some participants mentioned that the job must relate to their personal future goals, which varied among participants. For example, some IEPs who had the intention of pursuing Canadian medical licensure focused on careers that would ultimately benefit their chances of gaining residency spots when going through the licensure process. "...either that or that particular job is going to be a bridging job. Uh, or a position like a stepping stone, uh, for me to, uh, work for a little bit and then move on to something that is going to be my destination." (FGD4P4). 4. Discussion 4.1. Expositions Our study's findings indicate that IEPs interested in alternative careers took an adaptive approach in general, perhaps influenced by the need to earn in a foreign country and accommodate family needs and responsibilities. Many interrelated factors and barriers, from very individualized (e.g., passion for a certain career) to very general (e.g., wages/salary) to systemic (e.g., recognition of foreign education/experience), are associated with IEPs choosing alternative careers in combination with the resettlement challenges they face in Canada. The IEPs go through a critical process of finding a career that balances all these factors and constraints. Non-recognition of immigrants' education and work experience has been well documented in the literature, and it results in limited access to opportunities IEPs can pursue with interest, passion, and dignity . These experiences were also shared by our participants and other studies and deemed as systemic discrimination, including ageism and racism . The results of our study especially shed light on the significant training and skills required to enter certain alternative careers for IEPs, causing IEPs to spend resources to update their skillset to match job trends within Canada . These decisions to update skills, including meeting certain score requirements in English proficiency tests often require immense resources, such as time and money, leading to a disproportionate rate of unemployment, underemployment, and lower earning for IEPs as compared to the native Canadian professional population . Further, many IEPs present to Canada with significant international experience; however, studies have shown the systemic disadvantage skilled immigrants face when compared to other professionals with Canadian experience . A study that explored alternative careers for certain internationally trained professionals, including IEPs, concluded that IEPs are the driver in making the decision as to which alternative careers to pursue based on their backgrounds, which was reflected in this study . Moreover, identifying one's transferable skills was noted as a key element in successful transitions to alternative professions. We also observed that the participants were seeking careers where they could employ their transferable skills. Likewise, Peters et al. (2011) in Ontario reported that IEPs' experiences mimic the responses of IEP participants in our study . Further, a study by Sood et al. (2020) pointed out that IEPs may not necessarily secure an alternative job even after receiving training in Canada and investing their financial capital in it and may need to switch careers again . Worries regarding this sort of outcome were also frequently mentioned by the participants in our study. The results of this study also highlight the importance of both objective and subjective career success as important determinants of choosing alternative careers. Apart from IEPs utilizing their human capital, their personal feelings of self-accomplishment and pride, or subjective career success, are an important facet of transitioning into a successful contributing member of society . Participants in this study commonly linked job satisfaction with not only growth opportunities but also passion and interest in the job. Participants focused on certain jobs where they can be satisfied intellectually and have an interest and passion in the area. However, passion often may not coincide with the transferable skills and experience they require to achieve an alternative career, causing a mismatch in expectations and reality . This study presents an exploration of important factors in career exploration for IEPs and may assist in the development of a career decision-making tool specific to IEPs and assist them in finding alternative careers. Prior studies that discussed the career decision-making model match the findings of our study . Scholars have discussed models of career decision-making that include consideration of rational factors focused on maximizing individual gains, such as financial outcome, career growth, and future goals, as well as other less rational factors that include a passion for certain jobs, individual satisfaction, and perspectives . Our study findings revealed both kinds of factors in the alternative career decision-making process, which is drawn from a multicultural perspective and has high potential to be effective . Likewise, another study compared three career decision-making strategies for employees going through a career change that included rational (based on careful thought), intuitive (based on emotional satisfaction), and dependent (assistance/approval from others) strategies, finding that a combination of all three, especially using the head (rationality) and heart (intuition), makes the most effective career choices . Figure 1 shows how the decision-making factors identified in our current study can be integrated into these three strategies for adaptive career decision-making by IEPs. 4.2. Limitations The moderator of the focus groups ensured that all participants contribute to each idea to reduce bias in our findings due to dominant participants. Despite the lack of interpersonal interactions, conducting focus groups online came out as rather beneficial for this study as the participants could express themselves more freely and could join from their homes anywhere in Canada . However, most of the participants in this study were from Alberta (71.4%) due to our purposive sampling technique. We wanted to make sure that the participants were genuinely working in or considering alternative careers, which was more feasible for us from Alberta compared to other provinces. Further, perhaps participants from Alberta felt more interested than those from other provinces in participating in research conducted by the local and familiar university and community organizations. As IEPs across Canada encounter similar struggles in achieving their primary career and have similar educational and socio-cultural backgrounds, we believe the findings of our study can be applicable to a great extent in other provinces. We also observed a higher proportion of female participants in our study. We acknowledge this might be a limitation due to our chosen non-random sampling technique; however, other population-based studies also found a higher proportion of female participants in their studies . 4.3. Implications This study extracted the thoughts and views of IEPs on alternative career choices, which can help inform future research and professional integration. The findings of the study can be used to develop a decision-support tool for IEPs who are considering alternative career options . The research team aims to develop a web-based and/or mobile device application tool to help guide IEPs in choosing an alternative career according to their interest, skills, and other individual factors related to entry into and outcomes of the job. Moreover, the perspectives of potential employers, institutions that offer training or courses for alternative careers, and other stakeholders need to be captured and integrated to facilitate IEPs' pursuit of an alternative career and remove unconscious bias. A concept note outlining the potential strategies and future research recommendations was published to inform policymakers, researchers, service providers, and other stakeholders . Furthermore, IEPs in different professional roles (trans-professional adaptation) need to be evaluated . The integration of IEPs and other skilled immigrants into the Canadian economy is a major policy issue within Canada and can have a significant impact on local economies. Utilizing the human capital that IEPs bring to their host countries can have major implications not only for the career success of economies, but also for major workflows, policies, and labour shortages, especially during a time of globalization when economies welcome international expertise. As seen through this study, IEPs use a series of contextual and personal factors to determine alternative career choices, making them self-reliant and proactive in their career behaviours. This career self-management is an excellent opportunity for labour policymakers and employers to assist IEPs through supportive human resource policies that focus on development, training, teamwork, and experiential learning that can assist IEPs to transition successfully into the labour market. This type of career self-management should be supported by labour policy to increase both the objective and subjective career success IEPs may gain from their alternative careers. 5. Conclusions This qualitative study explored the decision-making process of IEPs pursuing alternative careers. A variety of influencing factors was identified that arise from educational qualification and professional experience, capacity-building resources, transferable skill utilization, employment, and individual-level concerns and priorities. These can be used to develop a decision-making support tool for IEPs considering alternative careers. Considering the difficult challenges IEPs encounter after moving to Canada and failing to succeed in their natural order of career--to become a physician--it is imperative to work toward supporting them to find and grow in suitable alternative careers and redress systemic barriers. While these factors and the hypothesized decision-support tool will help facilitate alternative career choices for IEPs, research and engagement initiatives to inform employers and other stakeholders about the IEPs' situation and potential strategies to enhance their professional integration through alternative careers should be recognized as a pressing need. Author Contributions Conceptualization, T.C.T. and D.L.; Methodology, T.C.T. and N.C.; Validation, D.L.; Formal analysis, N.C. and T.C.T.; Data curation, T.C.T. and N.C.; Writing--original draft, N.C. and T.C.T.; Writing--review & editing, D.L.; Supervision, T.C.T.; Project administration, N.C.; Funding acquisition, T.C.T. and D.L. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of University of Calgary (protocol code REB19-1198 and date of approval 10 October 2019). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The datasets generated and analyzed during the current study are not publicly available due unavailability of the participants' permission, privacy considerations, and ethical restrictions. Code-level deidentified data might be available from the corresponding author on reasonable request after all the planned publications are completed. Conflicts of Interest The authors declare no conflict of interests. T.C.T. and N.C. declare their lived experience as IEPs. Figure 1 The alternative career decision-making process for IEPs. healthcare-11-00657-t001_Table 1 Table 1 Characteristics of study participants. Traits % Count Age 29 or younger 9.5 4 30-39 40.5 17 40-49 33.3 14 50 or over 16.8 7 Sex Male 26.2 11 Female 73.9 31 Country of origin Armenia 2.4 1 Bangladesh 12.0 5 Canada 4.8 2 China 2.4 1 Colombia 2.4 1 Egypt 2.4 1 India 14.3 6 Iraq 2.4 1 Mexico 2.4 1 Nepal 2.4 1 Nigeria 14.5 6 Pakistan 21.4 9 Philippines 7.1 3 Somalia 2.4 1 Spain 2.4 1 Sudan 2.4 1 United Kingdom 2.4 1 Province currently living in Alberta 71.4 30 British Columbia 7.1 3 Manitoba 4.8 2 Ontario 14.3 6 Quebec 2.4 1 Immigration status Citizen 47.7 20 Permanent resident 47.7 20 Refugee 0.0 0 Temporary migrant (on a student visa, work visa, or visitor visa) 4.8 2 Specialty before coming to Canada Emergency medicine specialist 4.8 2 Family/general physician 38.1 16 Nephrologist 2.4 1 Neurological surgeon 2.4 1 Obstetrician 7.1 3 Occupational medicine specialist 2.4 1 Ophthalmologist 4.8 2 Paediatrician 4.8 2 Radiologist 4.8 2 Surgeon 2.4 1 Other 26.2 11 Others included: MPH, MD, FCPS, FRCS, or other post-graduate training in various specialty Current work position Employed (full-time) 33.3 14 Employed (part-time) 26.2 11 Unemployed; seeking work 33.3 14 Unemployed; not seeking work 7.1 3 Current area of work (among employed 25 participants) Health-related (regulated alternative career, i.e., requires licensure procedure, e.g., nursing, pharmacy technician, EMS tech, sonography, or laboratory technician) 20.0 5 Health-related (non-regulated alternative career, i.e., does not require licensure, e.g., health educator, health administrative officer, researcher, health policy analyst) 60.0 15 Non-health-related professional job (non-medical career build-up, e.g., engineering, business, or life sciences) 8.0 2 Non-health-related non-professional job (i.e., survival job, e.g., Uber/taxi driving, store jobs, or business owner) 12.0 3 Years spent preparing for alternative careers Less than a year 35.7 15 1-3 years 47.6 20 4-5 years 9.5 4 More than 5 years 7.1 3 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Sumalinog R. Zacharias A. Rana A. Backgrounder: International Medical Graduates (IMGs) and the Canadian Health Care System Can. Fed. Med. Stud. 2015 1 12 2. Motala M.I. Van Wyk J.M. Experiences of Foreign Medical Graduates (FMGs), International Medical Graduates (IMGs) and Overseas Trained Graduates (OTGs) on Entering Developing or Middle-Income Countries like South Africa: A Scoping Review Hum. Resour. Health 2019 17 7 10.1186/s12960-019-0343-y 30665452 3. Health Force Ontario Registration Requirements Available online: (accessed on 1 September 2020) 4. Chowdhury N. Ekpekurede M. Lake D. Chowdhury T.T. The Alternative Career Pathways for International Medical Graduates in Health and Wellness Sector International Medical Graduates in the United States Springer International Publishing Cham, Switzerland 2021 293 325 5. Lim Consulting Associates Foreign Qualifications Recognition and Alternative Careers The Best Practices and Thematic Task Team of the Foreign Qualifications Recognition Working Group Available online: (accessed on 10 September 2022) 6. Independent Review of Access to Postgraduate Programs by International Medical Graduates in Ontario; Volume I: Findings and Recommendations Available online: (accessed on 12 June 2022) 7. The Canadian Resident Matching Service Available online: (accessed on 12 June 2022) 8. Turin T.C. Chowdhury N. Ekpekurede M. Lake D. Lasker M. O'Brien M. Goopy S. Alternative Career Pathways for International Medical Graduates towards Job Market Integration: A Literature Review Int. J. Med. Educ. 2021 12 45 63 10.5116/ijme.606a.e83d 33839694 9. Bourgeault I.L. Neiterman E. Lebrun J. Viers K. Winkup J. Brain Gain, Drain & Waste: The Experiences of Internationally Educated Health Professionals in Canada University of Ottawa Ottawa, ON, Canada 2010 10. Sood A. A Study of Immigrant International Medical Graduates' Re-Licensing in Ontario: Their Experiences, Reflections, and Recommendations Ph.D. Thesis Queen's University Kingston, ON, Canada 2020 11. Wang Y. Das R.L.V. Lapa T. Marosan P. Pawliuk R. Chable H.D. Lake D. Lofters A. Career Development of International Medical Graduates in Canada: Status of the Unmatched Humanit. Soc. Sci. Commun. 2023 10 10 10.1057/s41599-023-01534-z 36628103 12. Turin T.C. Chowdhury N. Lake D. Chowdhury M.Z.I. Labor Market Integration of High-Skilled Immigrants in Canada: Employment Patterns of International Medical Graduates in Alternative Jobs Healthcare 2022 10 1705 10.3390/healthcare10091705 36141317 13. Beran T.N. Violato E. Faremo S. Violato C. Watt D. Lake D. Ego Identity Development in Physicians: A Cross-Cultural Comparison Using a Mixed Method Approach BMC Res. Notes 2012 5 249 10.1186/1756-0500-5-249 22620975 14. Alberta International Medical Graduates Association (AIMGA) Available online: (accessed on 10 September 2022) 15. Blain M.J. Fortin S. Alvarez F. Professional Journeys of International Medical Graduates in Quebec: Recognition, Uphill Battles, or Career Change J. Int. Migr. Integr. 2017 18 223 247 10.1007/s12134-016-0475-z 16. Newaz S. Career Options and Professional Integration of Internationally Trained Physicians in Canada Can. Health Policy 2020 10 1 19 17. Al Ariss A. Koall I. Ozbilgin M. Suutari V. Careers of Skilled Migrants: Towards a Theoretical and Methodological Expansion J. Manag. Dev. 2012 31 92 101 10.1108/02621711211199511 18. Murillo K.S. Underemployment and Labor Market Incorporation of Highly Skilled Immigrants With Professional Skills Ph.D. Thesis University of Oregon Eugene, OR, USA 2018 19. Employment Gaps and Underemployment for Racialized Groups and Immigrants in Canada: Current Findings and Future Directions Available online: (accessed on 10 September 2022) 20. Cheng L. Spaling M. Song X. Barriers and Facilitators to Professional Licensure and Certification Testing in Canada: Perspectives of Internationally Educated Professionals J. Int. Migr. Integr. 2013 14 733 750 10.1007/s12134-012-0263-3 21. Lai D.W.L. Shankar J. Khalema E. Unspoken Skills and Tactics: Essentials for Immigrant Professionals in Integration to Workplace Culture J. Int. Migr. Integr. 2017 18 937 959 10.1007/s12134-017-0513-5 22. Block S. Galabuzi G.-E. Canada's Colour Coded Labour Market Can. Cent. Policy Altern. 2011 1 20 23. Raza K. Chua C. Linguistic Outcomes of Language Accountability and Points-Based System for Multilingual Skilled Immigrants in Canada: A Critical Language-in-Immigration Policy Analysis J. Multiling. Multicult. Dev. 2022 1 15 10.1080/01434632.2022.2060242 24. Reitz J.G. Curtis J. Elrick J. Immigrant Skill Utilization: Trends and Policy Issues J. Int. Migr. Integr. 2014 15 1 26 10.1007/s12134-012-0265-1 25. Siar S.V. From Highly Skilled to Low Skilled: Revisiting the Deskilling of Migrant Labor Philippine Institute for Development Studies (PIDS) Quezon, Philippines 2013 26. Augustine H.J. Immigrant Professionals and Alternative Routes to Licensing: Policy Implications for Regulators and Government Can. Public Policy 2015 41 S14 S27 10.3138/cpp.2014-022 27. Neiterman E. Bourgeault I.L. Covell C.L. What Do We Know and Not Know about the Professional Integration of International Medical Graduates (IMGs) in Canada? Healthc. Policy 2017 12 18 32 10.12927/hcpol.2017.25101 28617235 28. International Medical Graduates-Current Issues Available online: (accessed on 10 September 2022) 29. Shuval J.T. The Reconstruction of Professional Identity Among Immigrant Physicians in Three Societies J. Immigr. Health 2000 2 191 202 10.1023/A:1009588229071 16228740 30. Becker G.S. Human Capital: A Theoretical and Empirical Analysis, with Special Reference to Education University of Chicago Press Chicago, IL, USA 2009 0226041220 31. Buzdugan R. Halli S.S. Labor Market Experiences of Canadian Immigrants with Focus on Foreign Education and Experience Int. Migr. Rev. 2009 43 366 386 10.1111/j.1747-7379.2009.00768.x 32. Internationally Educated Health Professionals in Canada: Navigating Three Policy Subsystems along the Pathway to Practice Available online: (accessed on 15 July 2022) 33. Creese G. Wiebe B. 'Survival Employment: Gender and Deskilling among African Immigrants in Canada Int. Migr. 2012 50 56 76 10.1111/j.1468-2435.2009.00531.x 34. Occupational Dimensions of Immigrant Credential Assessment: Trends in Professional, Managerial, and Other Occupations Available online: (accessed on 15 September 2022) 35. Mapping the Margins: Intersectionality, Identity Politics, and Violence against Women of Color Available online: (accessed on 16 June 2022) 36. Hassan F. Intersectionality and the Role of Service Providers: A Step Towards Improving the Employment Outcomes of Immigrant Women Ph.D. Thesis McMaster University Hamilton, ON, Canada 2020 37. Krueger R.A. Focus Groups: A Practical Guide for Applied Research Sage Publications, Inc. Thousand Oaks, CA, USA 1988 0-8039-3186-7 38. Turin T.C. Chowdhury N. Lake D. Alternative Careers toward Job Market Integration: Barriers Faced by International Medical Graduates in Canada Int. J. Environ. Res. Public Health 2023 20 2311 10.3390/ijerph20032311 36767681 39. Robinson R.S. Purposive Sampling Encycl. Qual. Life Well-Being Res. 2014 5243 5245 10.1007/978-94-007-0753-5_2337 40. Creswell J.W. Miller D.L. Determining Validity in Qualitative Inquiry Theory Pract. 2000 39 124 130 10.1207/s15430421tip3903_2 41. Braun V. Clarke V. Using Thematic Analysis in Psychology Qual. Res. Psychol. 2006 3 77 101 10.1191/1478088706qp063oa 42. Who Is Succeeding in the Canadian Labour Market?: Predictors of Career Success for Skilled Immigrants Available online: (accessed on 5 August 2022) 43. Racism, Discrimination and Migrant Workers in Canada: Evidence from the Literature Available online: (accessed on 15 August 2022) 44. Awareness of Ageism While Researching Multiple Minority Discrimination: A Discourse and Grounded Theory Analysis Revisiting Own Qualitative Research Available online: (accessed on 2 August 2022) 45. Determinants and Effects of Post-Migration Education Among New Immigrants in Canada Available online: (accessed on 16 August 2022) 46. Osaze E.D. The Non-Recognition or Devaluation of Foreign Professional Immigrants Credentials in Canada: The Impact on the Receiving Country (Canada) and the Immigrants York Space Denver, CO, USA 2018 47. Augustine J. Employment Match Rates in the Regulated Professions: Trends and Policy Implications Can. Public Policy 2015 41 s28 s47 10.3138/cpp.2014-085 48. Peters C. The Bridging Education and Licensure of International Medical Doctors in Ontario: A Call for Commitment, Consistency, and Transparency Ph.D. Thesis University of Toronto Toronto, ON, Canada 2011 49. Fang T. Novicevic M.M. Zikic J. Career Success of Immigrant Professionals: Stock and Flow of Their Career Capital Int. J. Manpow. 2009 30 472 488 10.1108/01437720910977661 50. Novak L. Chen C.P. Career Development of Foreign Trained Immigrants from Regulated Professions Int. J. Educ. Vocat. Guid. 2012 131 5 24 10.1007/s10775-012-9235-6 51. Phillips S.D. Choice and Change: Convergence from the Decision-Making Perspective Convergence in Career Development Theories: Implications for Science and Practice Savickas M.L. Lent R.W. Davies-Black Palo Alto, CA, USA 1994 155 163 52. Phillips S.D. Toward an Expanded Definition of Adaptive Decision Making Career Dev. Q. 1997 45 275 287 10.1002/j.2161-0045.1997.tb00471.x 53. Hartung P.J. Blustein D.L. Reason, Intuition, and Social Justice: Elaborating on Parsons's Career Decision-Making Model J. Couns. Dev. 2002 80 41 47 10.1002/j.1556-6678.2002.tb00164.x 54. Singh R. Greenhaus J.H. The Relation between Career Decision-Making Strategies and Person-Job Fit: A Study of Job Changers J. Vocat. Behav. 2004 64 198 221 10.1016/S0001-8791(03)00034-4 55. Oringderff J. "My Way": Piloting an Online Focus Group Int. J. Qual. Methods 2004 3 69 75 10.1177/160940690400300305 56. Turin T.C. Chowdhury N. Ekpekurede M. Lake D. Lasker M.A.A. O'Brien M. Goopy S. Professional Integration of Immigrant Medical Professionals through Alternative Career Pathways: An Internet Scan to Synthesize the Current Landscape Hum. Resour. Health 2021 19 51 10.1186/s12960-021-00599-8 33865402 57. Turin T.C. Chowdhury N. Lake D. Lost in transition: The need for a strategic approach to facilitate job market integration of internationally educated physicians through alternative careers Int. J. Environ. Res. Public Health 2022 19 3503 10.3390/ijerph19063503 35329190 58. Vapor V.R. Xu Y. Double Whammy for a New Breed of Foreign-Educated Nurses: Lived Experiences of Filipino Physician-Turned Nurses in the United States Res. Theory Nurs. Pract. 2011 25 210 226 10.1891/1541-6577.25.3.210 22216694 |
PMC10000569 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12050942 foods-12-00942 Review Composition of Nuts and Their Potential Health Benefits--An Overview Goncalves Berta 1* Pinto Teresa 1 Aires Alfredo 1 Morais Maria Cristina 1 Bacelar Eunice 1 Anjos Rosario 1 Ferreira-Cardoso Jorge 1 Oliveira Ivo 1 Vilela Alice 2 Cosme Fernanda 2 Contreras Maria del Mar Academic Editor Alvarruiz Andres Academic Editor 1 CITAB, Centre for the Research and Technology of Agro-Environmental and Biological Sciences, Inov4Agro, Institute for Innovation, Capacity Building and Sustainability of Agri-Food Production, University of Tras-of-Montes and Alto Douro, Quinta de Prados, 5000-801 Vila Real, Portugal 2 CQ-VR, Chemistry Research Centre--Vila Real, University of Tras-os-Montes and Alto Douro, Quinta de Prados, 5000-801 Vila Real, Portugal * Correspondence: [email protected] 23 2 2023 3 2023 12 5 94217 11 2022 15 2 2023 16 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). The possibility that nut intake may defend human health is an interesting point of view and has been investigated worldwide. Consequently, nuts are commonly promoted as healthy. In recent decades, the number of investigations proposing a correlation between nut consumption and a decrease in the risk of key chronic diseases has continued to increase. Nuts are a source of intake of fiber, and dietary fiber is associated with a reduced occurrence of obesity and cardiovascular diseases. Nuts likewise provide minerals and vitamins to the diet and supply phytochemicals that function as antioxidant, anti-inflammatory, and phytoestrogens agents and other protective mechanisms. Therefore, the main goal of this overview is to summarize current information and to describe the utmost new investigation concerning the health benefits of certain nuts. antioxidant activity consumer perception fatty acids fiber health benefits minerals phenolic compounds vitamins volatile compounds National Funds FCT Portuguese Foundation for Science and Technology--Portugal and COMPETEUIDB/00616/2020 UIDP/00616/2020 (CQ-VR) UIDB/04033/2020 (CITAB) This research was funded by the National Funds FCT Portuguese Foundation for Science and Technology--Portugal and COMPETE under the projects UIDB/00616/2020 and UIDP/00616/2020 (CQ-VR) and UIDB/04033/2020 (CITAB). pmc1. Introduction Currently, consumers are concerned about making a diversified and well-balanced diet. Therefore, the inclusion of nuts in the diet has undergone significant increases due to a growing recognition of their unique nutritional value, distinctive taste, flavor, nutraceutical properties, and healthy bioactive compounds, including high-quality proteins, fibers, minerals, tocopherols, phytosterols, and phenolic compounds . Nuts are usually described as dry fruits with an edible seed and a hard shell, with cashews (Anacardium occidentale), walnuts (Juglans regia), almonds (Prunus dulcis), chestnuts (Castanea sativa), pistachios (Pistacia vera), and hazelnuts (Corylus avellana) as the ones with higher production worldwide . There is the recognition that nuts are a good source of many nutrients, including monounsaturated and polyunsaturated fatty acid profile, vitamins E and K, selected minerals such as magnesium, copper, potassium, and selenium, dietary fibers, carotenoids, and phytosterols with potential antioxidant action . In addition, the ease of transport due to their size makes them even more recommended to be consumed in all situations. In addition, the ingesting of nuts is often related to reducing risk factors for chronic diseases, due to the fatty acid profiles, squalene, fibers, vegetable proteins, minerals, vitamins, carotenoids, and phytosterols with potential antioxidant action . Curiously, in all nuts, most of the antioxidants are located in the pellicle, as shown for almonds and peanuts , and they are lost when the skin is removed . In addition, in pistachios, most of the antioxidants are destroyed when the hard shells are cracked . This review article aims to synthesize the current state of knowledge on the nutritional composition and health outcomes of some selected nuts. 2. Nuts 2.1. Proteins Nuts are a rich source of proteins and essential amino acids as indicated by the USDA National Nutrient Database for Standard Reference and as presented in Table 1. The major sources of proteins are peanuts, almonds, and pistachios, while chestnuts are the poorest in proteins. Chung et al. reported higher protein content for some of these nuts, which can be ascribed to different geographic regions. The protein content also varied within the same nut species, denoting a significant effect of cultivar . Other factors, such as the harvest year, post-harvest storage, and even the processing method can affect the content of proteins in nuts. For example, Dodevbka et al. reported differences between raw, boiled, and roasted nut samples from Serbia. The seed storage proteins are the main type of proteins present in nuts and are responsible for nut allergies . Except for chestnuts, the other nine nuts referred to in Table 1 are the most common nuts capable of triggering adverse allergic reactions in some people. The proteins involved in nut allergy belong to different families, especially 2S albumins, globulins (legumins and vicilins), non-specific lipid transfer proteins (nsLTP), plant pathogenesis-related proteins (PR-10), profilins and oleosins . Amandin, a legumin-type protein, is the most abundant protein in almonds, while the PR-10 Cor a 1 is the main allergen in hazelnuts, and the legumin Ju r 4 is the prevalent allergen in walnuts. The 2S albumins Car I 1 and Pis v 1 are dominant in pecan nuts and pistachios, respectively. The vicilins Ana o 1, Ara h 1, and Mac Ii1 are the most common allergens in cashew nuts, peanuts, and macadamia nuts, respectively . Regarding the amino acid profile of each nut (Table 1), there is a considerable variation in the content of essential and non-essential amino acids. The nut protein composition is dominated by hydrophobic amino acids, followed by acidic, basic, and hydrophilic amino acids . Among the non-essential amino acids, glutamic acid is the most important, ranging from 0.02 g/100 g in chestnuts to 6.21 g/100 g in almonds. The second major non-essential amino acid is arginine ranging from 0.12 g/100 g in chestnuts to 3.08 g/100 g in peanuts, followed by aspartic acid that ranges between 0.03 g/100 g in chestnuts and 3.15 g/100 g in peanuts. Leucine is the most essential amino acid, followed by phenylalanine and valine. Chestnuts present the lowest values of these essential amino acids (0.10, 0.07, and 0.09 g/100 g for leucine, phenylalanine and valine, respectively), while peanuts are the richest source of leucine and phenylalanine, and pistachios are the richest source of valine. Although the amino acid profile can differ significantly with variety and location, studies with 23 hazelnuts in northeast China revealed the dominance of the same non-essential and essential amino acids (glutamic acid, arginine, aspartic acid, and leucine) described in Table 1 for hazelnut. The composition and dominance of essential and non-essential amino acids can influence several attributes of nuts, including the taste, aroma, or color being used, for example, for the characterization of almond cultivars . The essential amino acid contents and their digestibility determine the nutritional value of a food protein. Although nut proteins are often recognized as incomplete proteins (i.e., do not contain all essential amino acids) when compared to animal proteins, their consumption is strongly associated with cardiovascular health . Moreover, the presence of large quantities of arginine in all tree nuts has positive effects on immune response and inflammation, and cardiovascular function, including its key role in reducing the risk of cardiovascular disease and reproductive performance . The health benefits of nut consumption can be enhanced by combining different protein sources to provide adequate levels of all essential amino acids. 2.2. Vitamins Vitamins are essential for a balanced and healthy diet. Nuts contain fat-soluble vitamins (ascorbic acid, B1, B2, B3, B6) and antioxidants such as a-tocopherol (vitamin E), promoting better health, playing an important role against the aging process, improving brain function, and helping consumers to have healthy skin . According to studies carried out by several researchers, the existence of vitamin C (ascorbic acid) is an important antioxidant for human colon cells . The nut's nutritional value depends on its chemical composition, and this is the result of the interaction of the cultivar (genotype), meteorological factors such as temperature and radiation, and production practices . As Table 2 shows, walnuts, almonds, pine nuts, and hazelnuts are especially rich in vitamin E. Almonds, cashews, pistachios, walnuts, and peanuts are abundant sources of B vitamins. The concentration of folic acid was higher in pistachios and chestnuts. It is also the chestnuts that reveal the highest amount of vitamin C. 2.3. Minerals Nuts are also rich sources of minerals such as magnesium and potassium (Table 3). In recent years, increased consumption of nuts has been considered good for human health to increase the intake of certain minerals, and they are considered a heart-healthy snacks when eaten in moderation . Nuts are an important food source of minerals such as copper and magnesium. These two minerals may be protective against coronary heart disease. Nuts are also fairly high in potassium, particularly pistachio and cashew nuts (Table 1). Most nuts have a decent amount of zinc and iron, but pine nuts, cashews, and almonds stand above the rest. In contrast, nuts do not have a high content of calcium, still, some nuts such as almonds are better in terms of calcium content. 2.4. Fiber Fiber is a health-promoting nut ingredient. The intake of dietary fiber is inversely related to obesity, type two diabetes, cancer, and cardiovascular disease according to epidemiological and clinical studies . Among nuts, almonds present the highest content of fiber (Table 4), with a clear effect of genotype influencing its amount recorded . Some works have highlighted the influence of genotype on the fiber content found in almonds , ranging from 6.88% to 9.74% in blanched almonds , showing that almond skin is also responsible for the fiber content of this nut, as it is composed of around 60% of fiber . However, not all available data follow the same trend, with similar values of fiber recorded for different cultivars . Cashews have the lowest fiber content among the referred nuts, with recent works pointing at values always around 3% to 4% , with no apparent significant effect of the cultivar on its content, although comprehensive studies are lacking for this specific nut. Chestnuts are considered to be a good source of dietary fiber , with similar values to those of cashews. The content of fiber in chestnuts has been the subject of studies that cannot find a trend on the factors behind their variation. Some authors point out the clear effect of cultivar on fiber content or area of production or year . However, other works clearly state the similar content of fiber, regardless of cultivar . The hazelnut fiber content is usually referred to as ranging from 6.5 g/100 g to 9.7 g/100 g . Researchers have found higher amounts of fiber in some cultivars, such as the Turkish tombul hazelnut (12.9 g/100 g) , or other cultivars, with fiber values ranging from 9.8 g/100 g to 13.2 g/100 g (dry weight basis), being lowest in Tonda di Giffoni and highest in Campanica . This also shows the variation of fiber content among cultivars, also recorded in the comparison of sixteen hazelnut cultivars . For pistachio, the available works dealing with fiber content are few. However, early data indicate a content of 1.1-2.0% although more recent works show considerably higher values. Dreher , Bullo et al. , and Terzo et al. refer to values of fiber as around 10%, with Rabadan et al. suggesting that the major factor between variations is the crop year and related weather conditions. Finally, walnut presents an intermediate amount of fiber when compared to other nuts. Although the majority of available works indicate values ranging from 4% to 6% , some authors have found considerably different amounts of fiber, namely Ozcan , which indicates 1.8%, and Ozcan et al. that reports values between 3.8% and 3.9%. Again, the major factor affecting the fiber content of walnut is the genotype, with a slight effect also found to be caused by the crop year and related weather conditions . 2.5. Lipids and Fatty Acids Nuts are rich in several nutrients, although with great differences between them and minor but sometimes still significant variations within cultivars. Lipid content and fatty acid profile are two of the parameters that can change considerably when discussing nut composition (Table 4 and Table 5). Besides these great variations between nut species, changes in lipid content and profile can also occur due to several other factors, with genotype as one of the most important that influences nut composition. Recent works show that genotype and the environment are key factors behind changes in several compositional parameters of some nuts, namely fat content . There are some very good examples in the available literature, and to illustrate this fact, we will refer only to some for each nut. For almonds, Summo et al. , working with samples from a germplasm collection under the same growing condition, recorded variations of lipid content, depending on the cultivar, from 42.4% to 56.2% (fresh weight). Barreca et al. also reported a significant cultivar effect on the content of lipids in almonds. Almonds are also known for their interesting fatty acid profile, which is mainly composed of monounsaturated (MUFA) (60%) and polyunsaturated (PUFA) (30%) fatty acids, with a predominance of oleic, linoleic, palmitic, or stearic acids . The work of Summo et al. also shows the effect of the genotype on the fatty acid profile. Although major fatty acids are the same across the studied cultivars, changes can be observed in the individual amount of each fatty acid, as well as for the sum of unsaturated ( polyunsaturated) and saturated (SFA) fractions. For cashew nuts, recent studies show great variability in fat content and associated fatty acid profiles when comparing different production regions. The work of Rico et al. , analyzing 11 cashew origins, shows that fat content can vary from 45.05 g/100 g in Vietnamese samples to 50.40 g/100 g in samples from Kenya. In the fatty acid profiles, oleic, linoleic, and palmitic acids are the three major ones. Although monounsaturated fatty acids represent the major fraction in all samples, followed by saturated fatty acids, at least in one sample, the second most important fraction is polyunsaturated fatty acids. Chestnuts are featured with low-fat content and compared to other nuts, such as hazelnut, macadamia, pecan, or almond, chestnuts, exhibit the lowest fat content . However, in this minor chestnut fraction, fat-soluble bioactive compounds, such as tocols and phytosterols, are present in higher quantities when compared to fat-rich nuts. They contain a high quantity of essential fatty acids (those that must be provided by food intake, as they are not synthesized in the body but are necessary for health) , either saturated or unsaturated, linked to several processes involved in health and chronic diseases . Among them, the most important unsaturated fatty acids are linoleic and linolenic acids . Fat content and fatty acid profiles can, as for other nuts, change significantly among cultivars. A thorough study of 17 chestnut cultivars produced in Portugal shows significant variations ranging in fat content from 1.67% to 3.50% . Chestnut fat is primarily composed of three fatty acids, namely linoleic, oleic, and palmitic acids, with a predominance of polyunsaturated fatty acids. However, when comparing samples, significant variations of these fractions can be seen, with some presenting almost the same amount of polyunsaturated fatty acids. Similarly, the amount of saturated fatty acid also recorded significant variations across cultivars. Among nuts, hazelnut presents one of the highest contents of fat, above 60% (Table 5), with some authors indicating the amount of fat above 70%, depending on the cultivar or even on the canopy position of the fruits . The fat present in hazelnuts is mainly composed of MUFA, representing around 80% of the total fatty acid content, and oleic acid is the major individual monounsaturated fatty acid . Polyunsaturated fatty acids represent the second major fraction in hazelnut fat, almost exclusively due to the content of linoleic acid . However, some works have found that SFA can represent the second major group of fatty acids , influenced by the higher content of palmitic acid. Like most other nuts, pistachio is rich in fat, the available works indicating values around 50% , although some cultivars can have increased fat content, reaching values as high as 74.15% . Following the trend of other nuts, pistachio fat is rich in unsaturated fatty acids, namely MUFA. This fraction is mainly composed of oleic acid, with a contribution from palmitoleic acid, while the second most important fraction, PUFA, is mainly composed of linoleic acid . Regarding SFA, the minor fatty acid fraction is made almost entirely of palmitic acid . The fat content of walnut is very high, with average values that can be surpassed only by hazelnuts . Although the fat content is in the 60% range, considerable variations have been observed when comparing cultivars. Values varied between 49% and 82% . However, as referred before, most of the works show values of fat around 60%, with some variations associated with the studied cultivar . Walnut fat is mostly composed of unsaturated fatty acid, namely PUFA, while MUFA is the second most important type of fatty acid . Linoleic and linolenic acids are the ones responsible for the high amount of PUFA, with oleic as the major MUFA. Regarding SFA content, palmitic and stearic are the ones present in higher amounts (Table 5). 2.6. Phenolic Compounds Like in numerous other crops, phenolics are present in nuts. Many studies are reporting the beneficial effects of nut consumption on human health, including cardioprotective, neuroprotective, antidiabetic, anti-inflammatory, and antioxidant properties . Studies have shown that the consumption of nuts improves flood lipoprotein profile and gut microbiota . These health effects are mainly due to the presence of several type of compounds, including phenolics, as reported by Lamuel-Raventos and Onge . Each nut species presents its typical phenolic profile and content. For example, Liu et al. found a high content of phenolics, such as vanillic acid, catechin, naringin, quercetin, and ellagic acid, in chestnuts, while Smeriglio et al. , in almonds, found a high content of phenolics, such as quercetin, kaempferol, and isorhamnetin. Instead, Tas and Gokmen reported high levels of procyanidins A and B, trimers and tetramers, and prodelphinidin in peanuts. Table 6 shows several examples of phenolics found in the most common species of edible nuts. Similar to other crops, the variation in both profile and content of phenolic of nuts is highly related to genotype, cultural practices, climate conditions, fruit ripeness stage, storage, and post-harvest settings . In addition, differences in the methods used to extract and quantify phenolic compounds (e.g., microwave-assisted extraction--MAE; supercritical CO2 extraction--SC-CO2; enzyme-assisted extraction--EAE; pressurized liquid extraction--PLE) by researchers may interfere with the number of phenolic compounds identified. However, based on the literature, it is possible to find a more or less common pattern. The most abundant phenolics in almonds are catechin, epicatechin, protocatechuic acid, ferulic acid, kaempferol, and isorhamnetin ; in chestnuts are gallic acid, vanillic acid, syringic acid, catechin, and ellagic acid ; while in hazelnuts, the preponderance is for the catechin, epicatechin gallate, and gallic acid ; in peanuts, p-hydroxybenzoic acid, p-coumaric acid, ferulic acid, and epicatechin dominate ; in pistachios, gallic acid, syringic acid, catechin, and epicatechin ; while pecans and walnuts have in common high contents of chlorogenic, caffeic, p-coumaric, ferulic, ellagic and syringic acids . In general, all nuts have in common the presence of high amounts of phenolic acids and flavonoids. The anthocyanins are present in vestigial amounts and are therefore not considered. All these compounds are highly important because they have been associated with important beneficial effects on human health, as reported in the review of Lamuel-Raventos and Onge and De Souza et al. . Consumer perception of their beneficial effects has increased the intake of nuts. Different important findings from researchers have also contributed to the increment of such products in the human diet. For example, Brown et al. found that higher nut consumption was associated with a reduced prevalence of high cholesterol and blood pressure, diabetes, and gallstones, due to the richness of phenolic compounds. In addition, Musarra-Pizzo et al. tested a mix of phenolics present in natural almond skin and found that epicatechin and catechin were able to stop the growth of Staphylococcus aureus, suggesting that extracts from almond skins can be used to develop novel products for topical use. Neuroprotective effects against Alzheimer's disease were found in almonds, hazelnuts, and walnuts due to their richness in tocopherols and phenolics . 2.7. Aroma and Flavor Compounds The aroma compound profile of nuts is dependent on geographical origin and thermal processing and the presence of microorganisms. In almonds, several studies indicate aldehydes as the major volatiles, namely benzaldehyde with a characteristic bitter-almond flavor, although this compound might not be found in several cultivars . Besides terpenoids and substances derived from amino acids, volatiles are usually present as a result of the oxidation of fatty acids . Processing causes several modifications, either in the number of compounds, but also in the chemical classes present . In the work of Elmore et al. , they verified that walnuts from China and Ukraine contained high levels of lipid-derived volatiles from the linoleic acid breakdown (hexanal, pentanal, 1-hexanol, and 1-pentanol) and a-linolenic acid breakdown (1-penten-3-ol), whereas Chilean walnuts contained high levels of alkylbenzenes. Pyrazines are the major group of aromatic compounds in peanuts. They are formed by the thermally induced the Maillard reaction. The same applies to other nuts, such as pistachio and hazelnut. It is the roasting process that makes the fruit commercially viable and valuable, improving the nut's sales and sensory characteristics . Two pyrazines represent peanut flavor: 2,5-dimethyl pyrazine (with a characteristic nutty aroma) and 2-methoxy-5-methyl pyrazine (roasted nutty aroma) . In hazelnuts, the results from Kiefl and Schieberle showed that the aroma-active compounds 2-acetyl-1-pyrroline, 2-propionyl-1-pyrroline, 5-methyl-(E)-2-hepten-4-one (fibertone), 2,3-diethyl-5-methyl pyrazine, 3,5-dimethyl-2-ethyl pyrazine, and 2-furfurylthiol are appropriate odorant indicators to distinguish the several nut aromas. Specifically, the roasted or nutty aroma of roasted hazelnuts was developed if both 5-methyl-(E)-2-hepten-4-one and 3-methyl-4-heptanone were higher than 450 mg/kg, whereas the sum of the two 2-acyl-1-pyrrolines and two pyrazines should not exceed 400 mg/kg to avoid an over-roasted odor. A favored aroma can be obtained for each cultivar if specific temperatures, roasting techniques, and roasting times can be applied. One major quality concern related to nuts is the development of off-flavors due to the formation of oxidative degradation products . Various volatiles are involved in off-flavor; 1-Pentanol, 1-hexanol, and hexanal are the most important volatiles involved in off-flavor, and their presence at the highest levels is a synonym of nut degradation. 3. Impact of Nuts Processing on Nutrients and Phytochemicals The phytochemicals in tree nuts have been linked to various health benefits, but processing steps can affect their bioavailability. Nuts can be processed in various ways to create different final products. For example, nuts that are consumed are often dehulled, peeled, blanched, and roasted . Roasting is a common processing method used to preserve the quality and storability of nuts. It improves the flavor, aroma, color, texture, and appearance of the nuts through non-enzymatic reactions, such as Maillard browning. Roasting also inactivates enzymes that accelerate nutrient deterioration, remove microorganisms and food contaminants, and reduce degradative reactions such as lipid oxidation and rancidity, which are major factors that limit the shelf life of nuts. Additionally, the roasting process alters the microstructure and chemical composition of nuts, resulting in changes such as moisture reduction, modifications to lipids, changes in color, and the development of unique roasted flavors through the Maillard reaction . Thus, the roasting process improves the nuts' sensory characteristics such as flavor, color, taste, texture, appearance, and crispiness . This improves the overall sensory characteristics of the nuts, making them more appealing to consume. The antioxidant activity, nutritional content, and total phenolic compounds in nuts may decrease after blanching and peeling, but roasting can improve these factors by releasing bound phenolic compounds and forming Maillard reaction products such as melanins . However, the research on the effect of roasting on the phenolic compounds in nuts is limited. Based on available studies, the impact of roasting on the phenolic compounds in nuts can vary depending on the roasting temperature and duration. Some studies indicate that lower temperatures or shorter heating times may increase phenolic compounds, but higher temperatures or longer heating periods may decrease phenolic compounds . For example, in hazelnuts, the content of flavan-3-ols (catechin and epicatechin) decreases significantly when roasted, with significant differences observed between raw nuts with skin and roasted nuts without skin. Thermal treatment also negatively impacts the content of procyanidin dimers and trimers in hazelnuts. Studies have shown that polyphenols in hazelnuts are mostly present in the skin and that roasting reduces the levels of phenolic compounds in most nuts, not only because of the removal of the skin but also due to the chemical degradation of many phenolic compounds . These compounds are highly unstable and may be lost during processing, particularly when heat treatment is involved. Roasting can also alter the levels of antioxidants in the nuts, as the level of individual phenolics is higher in whole unroasted nuts and alters the protein profile and allergenic properties . Previous studies suggest that roasting enhances the allergenicity of roasted peanuts compared to raw peanuts , but the same was not observed in almonds . The antioxidant activity of raw and roasted nuts depends on the type of nut and the roasting conditions. According to Schlormann et al. , roasting can lead to a decrease in antioxidant activity in some nuts (hazelnut and walnut), but in others (almond and pistachio), the activity remains stable or is slightly enhanced. This decrease in activity is due to the loss of polyphenols due to thermal treatment, but the formation of antioxidant-active compounds due to Maillard reactions can counter this effect. The impact of roasting on bioavailability is still uncertain and requires further research. It is also important to evaluate the necessity of thermal processing by proving that nutritional and other properties are of great value, with antinutrients considerably decreased . Thermal processing significantly reduces the protein, ash, and fiber content. The decrease in protein content may be caused by high-temperature denaturation and/or solubilization . Additionally, the precipitation of mineral components leads to a decrease in ash content. In terms of carbohydrate content, roasting can increase it from 4.17% to 5.5%. This may be due to the hydrolysis of carbohydrates and to the reduction of other compounds in nuts due to thermal processing, making them easier to capture . Concerning the effect of hot water blanching on protein composition, the results also depend on species and conditions of thermal processing . Tian et al. demonstrated that subjecting peanuts to 100 degC for 20 min reduced their allergenicity, due to the denaturation of allergenic proteins and to the transition of low molecular weight to the boiling water . In turn, boiling almonds for 10 min or cashews and pistachios for 60 min did not affect their properties . 4. Nut Consumer Perceptions of Health Benefits Plant science research has been primarily focused on increasing production, with health benefits as a minor concern. The food industry is currently adapting its market trends to accommodate sustainability values, especially those related to health benefits, as they are increasingly researched by consumers , based, on nuts, on the phytonutrients present in these foods . The current use of phytonutrients by food producers and the knowledge of their effect on the prevention of chronic disease points out the need for a careful look at crop production strategies (fertilization, season, soil fertility, and irrigation) affecting the quantitative and qualitative profiles of these compounds, but also to post-harvest techniques (processing or packaging) that can modify phytonutrients . There is mounting evidence of the potential health benefits of a nut-rich diet. The ingestion of phytochemicals from nuts and their positive influence on several diseases (cancer, heart disease, stroke, hypertension, birth defects, cataracts, diabetes, diverticulosis, and obesity) are established . There are many phytochemicals present in nuts that can be responsible for their health-promoting activities. Of those, one must refer to the vitamins, carotenoids, phenolic acids, or flavonoids, and their role in the prevention of certain cancers and cardiovascular diseases, but also to phytoestrogens, organosulfur compounds, fiber, or isothiocyanates (reviewed by several authors ). Nuts have been traditionally looked at as a high-fat and high-calorie food that should be consumed in moderation, which may be part of the reason why their intake is still below the recommended amount . Although the link between weight gain and nut intake has been disproven , the usual high cost of nuts is another barrier to the increase in daily intake by consumers. The intake of nuts has been linked to several benefits to health, including favorable plasma lipid profiles, reduced risk of coronary heart disease, certain types of cancer, stroke, atherosclerosis, type-2 diabetes, inflammation, and several other chronic diseases . However, it appears that consumers are not fully aware of the potential benefits of the intake of nuts. Recent works have shown that consumers link nuts to the high content of fat and proteins and that they are healthy. Nevertheless, a large percentage of consumers are still not aware of the link between nuts and the effects on blood cholesterol, cardiovascular disease risk, obesity, cancer, or diabetes . Recently, there has been a huge effort to emphasize the beneficial action for the health by changing consumers' eating habits, leading them to increase the consumption of certain foods such as nuts. There is no doubt that an informed consumer makes better decisions when choosing certain foods. In the long term, a higher intake of nuts will lead to clear benefits in the health sector, but other sectors will also benefit, such as producers and sellers. Major concerns of the food industry related to the production and commercialization of nuts are the effects of processing and storage on the quality of nuts. Both temperature and humidity after harvesting can influence the appearance, moisture content, texture, and sensory characteristics of nuts . Specifically, higher post-harvest temperature conditions can reduce crispness, increase moisture content and change oiliness and sweetness, resulting in the development of rancidity . According to Mexis et al. , the alteration of sensorial characteristics leads to the formation of unpleasant flavors in pistachios, almonds, peanuts, and walnuts, as a result of alterations in the oxidation rate caused by high storage temperatures. It is mentioned in this study that storage temperatures of 30, 36, and 40 degC showed that nuts are more rancid compared to those stored at 8, 10, 20, or 25 degC. Another possible alternative to increase the shelf life of nuts is the use of suitable packaging to reduce the problems mentioned above. Food can be packaged properly using modified atmosphere packaging or vacuum packaging to control the oxidation reaction . The packaging material will be an important aspect to take into account as it will influence the shelf life of the nuts, will affect the respiration and transpiration rates of the fruit, as well as, the development of microorganisms. Fernandes et al. , in their comparative studies of chestnut conservation packages, concluded that chestnut conservation through the use of a specific packaging can have a substantial impact on preserving the color and texture of the fruit, preventing loss of weight, microbial growth, and in maintaining the water content of the fruit. Consumer demand for eco-friendly and sustainable product packaging has proven to be remarkably stable and robust in recent years, including willingness to pay more for eco-friendly packaging. Consumers also recognize the value of reuse. The refillable packaging is proving to be a versatile and valuable solution for consumer products. Therefore, a holistic view of these issues is a growing requirement for everyone involved, from production, conservation, and marketing of this type of food product. Food choice is one of the most frequent human decisions and is determined by a complex set of factors and interrelated determinants . Although several models attempting to explain that process have been proposed, one of the most accepted is the Total Food Quality Model . This model can be divided into three parameters: 'search', 'experience', and 'credence' attributes. The first two (search attributes, such as appearance or price, and experience characteristics, such as flavor or taste) are those more easily observed by consumers and can be straightforwardly experienced by them. For credence properties, such as health and nutritional benefits, the consumer cannot validate those claims . This is even more important in the current society, where the available fast food supply is large and more easily responds to the fast-paced life of consumers, with nutritionally poor foods taking place of a healthier diet. 5. Conclusions and Final Remarks Nuts are a good source of many bioactive compounds with recognized health benefits, such as tocopherols, vitamins, and phenolic compounds. However, acquiring knowledge about the variation of bioactive compounds during fruit development and the ripening stage is crucial. How global environmental change and innovative crop production technology affect tree physiology and thus yield and fruit quality is at the moment mostly unknown. The development of species-specific strategies that improve both fruit quality and nutritional properties without significantly affecting yield should be aimed at by future research studies. The selection of high-yielding nut species and cultivars well-adapted to the different growing regions and future climatic conditions, with improved fruit traits, are needed to produce fruits with excellent quality and high consumer acceptability. Acknowledgments The authors would like to thank the CITAB/Inov4Agro Center for the Research and Technology of Agro-Environmental and Biological Sciences/Institute for Innovation, Capacity Building, and Sustainability of Agri-Food Production and Chemistry Research Center--Vila Real (CQ-VR) for their financial support. Author Contributions All authors have contributed equally to this work. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Some representative flavor compounds in almonds, peanuts, and hazelnuts. foods-12-00942-t001_Table 1 Table 1 Protein and amino acid contents of selected nuts (compiled from ). Nut Protein (g/100 g) Amino Acids (g/100 g of Portion) Trp Thr Ile Leu Lys Met Cys Phe Tyr Val Arg His Ala Asp Glu Gly Pro Ser Almond 16.8-25.4 0.211 0.601 0.751 1.47 0.568 0.157 0.215 1.13 0.45 0.855 2.46 0.539 0.999 2.64 6.21 1.43 0.969 0.912 Cashew nut 17.5-19.0 0.287 0.688 0.789 1.47 0.928 0.362 0.393 0.951 0.508 1.09 2.12 0.456 0.837 1.80 4.51 0.937 0.812 1.08 Chestnut 1.63 0.018 0.058 0.064 0.096 0.096 0.038 0.052 0.069 0.045 0.091 0.116 0.045 0.109 0.028 0.021 0,084 0,086 0.081 Hazelnut 14.5-15.2 0.193 0.497 0.545 1.06 0.42 0.221 0.277 0.663 0.362 0.701 2.21 0.432 0.73 1.68 3.71 0.724 0.561 0.735 Macadamia nut 7.55-8.58 0.067 0.37 0.314 0.602 0.018 0.023 0.006 0.665 0.511 0.363 1.40 0.195 0.388 1.10 2.27 0.454 0.468 0.419 Peanut 25.8 0.25 0.883 0.907 1.67 0.926 0.317 0.331 1.38 1.05 1.08 3.08 0.652 1.02 3.15 5.39 1.55 1.14 1.27 Pecan nut 9.0-9.3 0.093 0.306 0.336 0.598 0.287 0.183 0.152 0.426 0.215 0.411 1.18 0.262 0.397 0.929 1.83 0.453 0.363 0.474 Pine nut 13.7 0.107 0.37 0.542 0.991 0.54 0.259 0.289 0.524 0.509 0.687 2.41 0.341 0.684 1.3 2.93 0.691 0.673 0.835 Pistachio 19.4-22.1 0.251 0.684 0.917 1.60 1.14 0.36 0.292 1.09 0.509 1.25 2.13 0.512 0.973 1.88 4.3 1.01 0.938 1.28 Walnut 14.4-16.0 0.17 0.596 0.625 1.17 0.424 0.236 0.208 0.711 0.406 0.753 2.28 0.391 0.696 1.83 2.82 0.816 0.706 0.934 Trp--Tryptophan, Thr--Threonine, Ile--Isoleucine, Leu--Leucine, Lys--Lysine, Met--Methionine, Cys--Cystine, Phe--Phenylalanine, Tyr--Tyrosine, Val--Valine, Arg--Arginine, His--Histidine, Ala--Alanine, Asp--Aspartic Acid, Glu--Glutamic Acid, Gly--Glycine, Pro--Proline, Ser--Serine. foods-12-00942-t002_Table 2 Table 2 Vitamin contents (mg/100 g) of selected nuts (source: ). Nut Ascorbic Acid (C) Vit A (IU) Niacin (B3) Thiamine (B1) Riboflavin (B2) Pyridoxine (B6) Folic Acid (B9) Pantothenic Acid (B5) a-Tocopherol (E) Almond 3.62-3.90 0.06 3.62-3.90 0.21 0.80-1.14 0.1 0.04 0.3 2.4-25.9 Cashew nut 1.06-1.10 - 1.06-1.10 0.42 0.06-0.10 0.4 0.25 0.9 0.0-0.9 Hazelnut 1.81 20 1.81 0.30 0.10 0.2-0.6 ND 0.9 3.5-15.0 Peanut 5.75-12.10 - 5.75-12.10 0.60 0.04-0.10 0.1-0.3 0.24 0.6 0.4 Pine nut 4.40 29 4.40 0.20 0.1 ND 0.3 2.5-9.3 Pistachio 1.30 415 1.30 0.87 0.16-0.20 1.7 51.00 0.5 0.3-2.3 Walnut 0.47-1.13 20 0.47-1.13 0.34 0.15-0.20 0.5-0.6 0.98 0.6 0.1-13.0 Chestnut 40.2 26 1.1 0.14 0.02 ND 58 0.48 - Not detected--ND. foods-12-00942-t003_Table 3 Table 3 Mineral contents (mg/100 g) of selected nuts (source: ). Nut Na Mg K Ca Cu Zn Fe Almond 1.00 275 728 248 0.90-1.03 1.91-3.12 3.71-6.21 Cashew nut 12.00 292 660 37 0.56 0.96-5.78 3.82-6.68 Chestnut 2.00 30.00 484.00 19.00 0.418 0.49 0.94 Hazelnut 0.70-0.98 140-163 514-680 84-114 0.65-0.99 1.95-2.96 0.56-4.70 Peanut 1.30-18.00 168-173 558-705 67-92 0.75-0.83 0.44-3.27 0.58-4.58 Pine nut 2.00 251-265 597 16 1.32-1.60 3.08-6.45 5.53-6.64 Pistachio 1.00-9.36 117 -121 642-1025 107-171 0.75-1.70 2.77-6.72 0.41-8.86 Walnut 2.00 158-201 441-523 61-98 2.54 1.52-3.37 2.91-5.74 foods-12-00942-t004_Table 4 Table 4 Fiber and lipid contents (%) of selected nuts (adapted from Amarowicz et al. ). Nut Fiber (%) Lipid (%) Almond 11.8-13.0 43.3-50.6 Cashew nut 1.4-3.3 42.8-43.9 Chestnut 2.3-3.7 1.6-7.4 Hazelnut 3.4-9.7 59.8-61.5 Pistachio 10.3 44.4-45.4 Walnut 6.7 64.5-65.2 foods-12-00942-t005_Table 5 Table 5 Fatty acid composition of selected nuts (g/100 g nut) (Source: ). Nut SFA MUFA PUFA Total Palmitic 16:0 Stearic 18:0 Total Oleic 18:1 Palmitoleic 16:1 Total Linoleic 18:2 Linolenic 18:3 Almond 3.802 3.083 0.704 31.551 31.294 0.227 12.329 12.324 0.003 Cashew 7.783 3.916 3.223 23.797 23.523 0.136 7.845 7.782 0.062 Chestnut 0.425 0.384 0.021 0.780 0.749 0.021 0.894 0.798 0.095 Hazelnut 4.464 3.097 1.265 45.652 45.405 0.116 7.920 7.833 0.087 Pistachio 5.907 5.265 0.478 23.257 22.674 0.495 14.380 14.091 0.289 Walnut 6.126 4.404 1.659 8.933 8.799 0.134 (C20:1) 47.174 38.093 9.080 foods-12-00942-t006_Table 6 Table 6 Main phenolic compounds found in the most common nuts (skins + kernels). Nut Phenolic Compound Reference Almond Catechin, epicatechin, naringenin, eriodictyol, gallic acid, caffeic acid, chlorogenic acid, o-coumaric acid, p-coumaric acid ferulic acid, hydroxybenzoic acid, protocatechuic, vanillic acid, quercetin, kaempferol, isorhamnetin Chestnut Gallic acid, syringic acid, chlorogenic acid, ferulic acid, vanillic acid, catechin, naringin, quercetin, ellagic acid Hazelnut Gallic acid, protocatechuic acid, caffeic acid, o-coumaric acid, p-coumaric acid, ferullic acid, catechin, epicatechin, epicatechin gallate, rutin Peanut Catechin, epicatechin, quercetin, isorhamnetin, gallic acid, protocatechuic, caffeic acid, p-coumaric acid, procyanidins A and B, trimers and tetramers, prodelphinidin Pecan nut Ellagic acid, catechin, gallic acid, hydroxybenzoic acid, trans-cinnamic acid, syringic acid, caffeic acid, p-coumaric acid, ferulic acid, naringenin, apigenin, quercetin, rutin, kaempferol, isorhamnetin, resveratrol Pistachio Cyanidin, gallic acid, protocatechuic, eriodictyol, catechin, epicatechin, epicatechin gallate, luteolin, quercetin, myricetin, procyanidin B1, trimers, and tetramers Walnut Vanillic acid, catechin, pyrocatechin, protocatechuic acid, epicatechin, syringic acid, gallic acid, juglone and cinnamic acid, ellagic acid, rutin Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Alasalvar C. Salvado J.S. Ros E. Bioactives and health benefits of nuts and dried fruits Food Chem. 2020 314 126192 10.1016/j.foodchem.2020.126192 31958750 2. INC 2022 Annual Report 2020/2021 Available online: (accessed on 18 September 2022) 3. de Souza R.G.M. Schincaglia R.M. Pimentel G.D. Mota J.F. Nuts and human health outcomes: A systematic review Nutrients 2017 9 1311 10.3390/nu9121311 29207471 4. Maguire L.S. O'Sullivan S.M. Galvin K. O'Connor T.P. O'Brien N.M. Fatty acid profile, tocopherol, squalene and phytosterol content of walnuts, almonds, peanuts, hazelnuts and the macadamia nut Int. J. Food Sci. Nutr. 2004 55 171 178 10.1080/09637480410001725175 15223592 5. Chen C.-Y. Milbury P.E. Lapsley K. Blumberg J.B. Flavonoids from almond skins are bioavailable and act synergistically with vitamins C and E to enhance hamster and human LDL resistance to oxidation J. Nutr. 2005 135 1366 1373 10.1093/jn/135.6.1366 15930439 6. Milbury P.E. Chen C.Y. Dolnikowski G.G. Blumberg J.B. Determination of flavonoids and phenolics and their distribution in almonds J. Agric. Food Chem. 2006 54 5027 5033 10.1021/jf0603937 16819912 7. Lou H. Yuan H. Ma B. Ren D. Ji M. Oka S. Polyphenols from peanut skins and their free radical-scavenging effects Phytochemistry 2004 65 2391 2399 10.1016/j.phytochem.2004.06.026 15381013 8. Blomhoff R. Carlsen M.H. Frost Andersen L. Jacobs D.R. Jr. Health benefits of nuts, potential role of antioxidants Br. J. Nutr. 2006 96 S52 S60 10.1017/BJN20061864 17125534 9. Seeram N.P. Zhang Y. Henning S.M. Lee R. Niu Y. Lin G. Heber D. Pistachio skin phenolics are destroyed by bleaching resulting in reduced antioxidative capacities J. Agric. Food Chem. 2006 54 7036 7040 10.1021/jf0614948 16968060 10. USDA, Agricultural Research Service, Nutrient Data Laboratory USDA National Nutrient Database for Standard Reference, Legacy. Version Current: April 2018 Available online: (accessed on 18 September 2022) 11. Chung K.H. Shin K.O. Hwang H.J. Choi K.S. Chemical composition of nuts and seeds sold in Korea Nutr. Res. Pract. 2013 7 82 88 10.4162/nrp.2013.7.2.82 23610599 12. Kafkas E. Attar S.H. Gundesli M.A. Ozcan A. Ergun M. Phenolic and fatty acid profile, and protein content of different walnut cultivars and genotypes (Juglans regia L.) grown in the USA Int. J. Fruit Sci. 2020 20 S1711 S1720 10.1080/15538362.2020.1830014 13. Muller A.K. Helms U. Rohrer C. Mohler M. Hellwig F. Glei M. Schwerdtle T. Lorkowski S. Dawczynski C. Nutrient composition of different hazelnut cultivars grown in Germany Foods 2020 9 1596 10.3390/foods9111596 33153116 14. Nankya R. Mulumba J.W. Lwandasa H. Matovu M. Isabirye B. De Santis P. Jarvis D.I. Diversity in nutrient content and consumer preferences of sensory attributes of peanut (Arachis hypogaea L.) varieties in Ugandan agroecosystems Sustainability 2021 13 2658 10.3390/su13052658 15. Ferrari V. Gil G. Heinzen H. Zoppolo R. Ibanez F. Influence of cultivar on nutritional composition and nutraceutical potential of pecan growing in Uruguay Front. Nut. 2022 9 868054 10.3389/fnut.2022.868054 16. Bolling B. Chen C. McKay D. Blumberg J. Tree nut phytochemicals: Composition, antioxidant capacity, bioactivity, impact factors. A systematic review of almonds, Brazils, cashews, hazelnuts, macadamias, pecans, pine nuts, pistachios and walnuts Nutr. Res. Rev. 2011 24 244 275 10.1017/S095442241100014X 22153059 17. Dodevska M. Kukic M.J. Sofrenic I. Tesevic V. Jankovic M. Djordjevic B. Ivanovic N.D. Similarities and differences in the nutritional composition of nuts and seeds in Serbia Front. Nutr. 2022 9 1003125 10.3389/fnut.2022.1003125 36185649 18. Luparelli A. Losito I. De Angelis E. Pilolli R. Lambertini F. Monaci L. Tree nuts and peanuts as a source of beneficial compounds and a threat for allergic consumers: Overview on methods for their detection in complex food products Foods 2022 11 728 10.3390/foods11050728 35267361 19. Venkatachalam M. Sathe S.K. Chemical composition of selected edible nut seeds J. Agric. Food Chem. 2006 54 4705 4714 10.1021/jf0606959 16787018 20. Fan L. Ren J. Yang Y. Zhang L. Comparative analysis on essential nutrient compositions of 23 wild Hazelnuts (Corylus heterophylla) grown in Northeast China J. Food Qual. 2020 2020 9475961 10.1155/2020/9475961 21. Seron L.A. Garrigos E.S. Berenguer V. Grane-Teruel N. Characterisation of 19 almond cultivars on the basis of their free amino acids composition Food Chem. 1998 61 455 459 10.1016/S0308-8146(97)00083-6 22. Hertzler S.R. Lieblein-Boff J.C. Weiler M. Allgeier C. Plant proteins: Assessing their nutritional quality and effects on health and physical function Nutrients 2020 12 3704 10.3390/nu12123704 33266120 23. Arya S.S. Salve A.R. Chauhan S. Peanuts as functional food: A review J Food Sci Technol. 2016 53 31 41 10.1007/s13197-015-2007-9 26787930 24. Suna S. Avsar B. Kocer S. Copur O.U. Effects of different pretreatments on the physicochemical characteristics and quality criteria of chestnut (Castanea sativa Mill.) pickle: A new value-added product J. Food Process. Preserv. 2021 45 e15669 10.1111/jfpp.15669 25. Pinto D. Rodrigues F. Braga N. Santos J. Pimentel F.B. Palmeira-De-Oliveira A. Oliveira M.B.P.P. The Castanea sativa bur as a new potential ingredient for nutraceutical and cosmetic outcomes: Preliminary studies Food Funct. 2016 8 201 208 10.1039/C6FO01469K 26. Beton K. Brozek-Pluska B. Vitamin C-Protective role in oxidative stress conditions induced in human normal colon cells by Label-Free Raman Spectroscopy and Imaging Int. J. Mol. Sci. 2021 22 6928 10.3390/ijms22136928 34203225 27. Choupina A.B. Nutritional and health potential of European chestnut Rev. Cienc. Agrar. 2019 42 801 807 28. Cevriye M.E.R.T. Erturk U. Chemical compositions and sugar profiles of consumed chestnut cultivars in the Marmara Region, Turkey Not. Bot. Horti Agrobot. 2017 45 203 207 10.15835/nbha45110729 29. Mota M. Pinto T. Vilela A. Marques T. Borges A. Caco J. Ferreira-Cardoso J. Raimundo F. Gomes-Laranjo J. Irrigation positively affects the chestnut's quality: The chemical composition, fruit size and sensory attributes Sci. Hortic. 2018 238 177 186 10.1016/j.scienta.2018.04.047 30. Santos M.J. Pinto T. Vilela A. Sweet chestnut (Castanea sativa Mill.) nutritional and phenolic composition interactions with chestnut flavor physiology Foods 2022 11 4052 10.3390/foods11244052 36553794 31. Ackurt F. Ozdemir M. Biringen G. Loker M. Effects of geographical origin and variety on vitamin and mineral composition of hazelnut (Corylus avellana L.) varieties cultivated in Turkey Food Chem. 1999 65 309 313 10.1016/S0308-8146(98)00201-5 32. Sathe S. Monaghan E. Kshiesagar H. Venkatachalam M. Chemical composition of edible nut seeds and its implications in human health Tree Nuts Composition, Phytochemicals and Health Effects Alsalvar C. Shahidi F. Taylor & Francis Group Boca Raton, FL, USA 2008 12 29 10.1201/9781420019391.ch2 33. Camara C.R.S. Schlegel V. A Review on the potential human health benefits of the black walnut: A comparison with the english walnuts and other tree nuts Int. J. Food Prop. 2016 19 2175 2189 10.1080/10942912.2015.1114951 34. Pradhan C. Peter N. Dileep N. Nuts as dietary source of fatty acids and micro nutrients in human health Nuts and Nut Products in Human Health and Nutrition Rao V. Rao L. Ahiduzzaman M. Aminul A.K.M.A. IntechOpen London, UK 2021 10.5772/intechopen.94327 978-1-78985-511-1 35. Alhassan K. Agbenorhevi J.K. Asibuo J.Y. Sampson G.O. Proximate composition and functional properties of some new groundnut accessions J. Food Secur. 2017 5 9 12 10.12691/jfs-5-1-2 36. Caglarirmak N. Batkan A.C. Nutrients and biochemistry of nuts in different consumption types in Turkey J. Food Proc. Preserv. 2005 29 407 423 10.1111/j.1745-4549.2005.00037.x 37. Ros E. Health benefits of nut consumption Nutrients 2010 2 652 682 10.3390/nu2070652 22254047 38. Lattimer J.M. Haub M.D. Effects of dietary fibre and its components on metabolic health Nutrients 2010 2 1266 1289 10.3390/nu2121266 22254008 39. Rabadan A. Alvarez-Orti M. Pardo J.E. A comparison of the effect of genotype and weather conditions on the nutritional composition of most important commercial nuts Sci. Hortic. 2019 244 218 224 10.1016/j.scienta.2018.09.064 40. Ruggeri S. Cappelloni M. Gambelli L. Nicoli S. Carnovale E. Chemical composition and nutritive value of nuts grown in Italy Ital. J. Food Sci. 1998 3 243 251 41. Romero A. Vargas F.J. Tous J. Ninot A. Miarnau X. New almond varieties from IRTA's breeding programme:(1) chemical composition Proceedings of the V International Symposium on Pistachios and Almonds Sanliurfa, Turkey 6-10 October 2009 Volume 912 477 484 42. Mandalari G. Tomaino A. Arcoraci T. Martorana M. Turco V.L. Cacciola F. Rich G. Bisignano C. Saija A. Dugo P. Characterization of polyphenols, lipids and dietary fibre from almond skins (Amygdalus communis L.) J. Food Compos. Anal. 2010 23 166 174 10.1016/j.jfca.2009.08.015 43. Barreira J.C. Casal S. Ferreira I.C. Peres A.M. Pereira J.A. Oliveira M.B.P. Supervised chemical pattern recognition in almond (Prunus dulcis) portuguese PDO cultivars: PCA-and LDA-based triennial study J. Agric. Food Chem. 2012 60 9697 9704 10.1021/jf301402t 22954238 44. Barreira J.C. Casal S. Ferreira I.C. Peres A.M. Pereira J.A. Oliveira M.B.P. Chemical characterization of chestnut cultivars from three consecutive years: Chemometrics and contribution for authentication Food Chem. Toxicol. 2012 50 2311 2317 10.1016/j.fct.2012.04.008 22525865 45. Yada S. Huang G. Lapsley K. Natural variability in the nutrient composition of California-grown almonds J. Food Comp. Anal. 2013 30 80 85 10.1016/j.jfca.2013.01.008 46. Rico R. Bullo M. Salas-Salvado J. Nutritional composition of raw fresh cashew (Anacardium occidentale L.) kernels from different origin Food Sci. Nutr. 2016 4 329 338 10.1002/fsn3.294 27004123 47. Goncalves B. Borges O. Costa H.S. Benett R. Santos M. Silva A.P. Metabolite composition of chestnut (Castanea sativa Mill.) upon cooking: Proximate analysis, fibre, organic acids and phenolics Food Chem. 2010 122 154 160 10.1016/j.foodchem.2010.02.032 48. Borges O. Goncalves B. Carvalho J.L.S. Correia P. Silva A.P. Nutritional quality of chestnut (Castanea sativa Mill.) cultivars from Portugal Food Chem. 2008 106 976 984 10.1016/j.foodchem.2007.07.011 49. Barreira J.C. Casal S. Ferreira I.C. Oliveira M.B.P. Pereira J.A. Nutritional, fatty acid and triacylglycerol profiles of Castanea sativa Mill. cultivars: A compositional and chemometric approach J. Agric. Food Chem. 2009 57 2836 2842 10.1021/jf803754u 19334758 50. Pena-Mendez E.M. Hernandez-Suarez M. Diaz-Romero C. Rodriguez-Rodriguez E. Characterization of various chestnut cultivars by means of chemometrics approach Food Chem. 2008 107 537 544 10.1016/j.foodchem.2007.08.024 51. Pereira-Lorenzo S. Ramos-Cabrer A.M. Diaz-Hernandez M.B. Ciordia-Ara M. Rios-Mesa D. Chemical composition of chestnut cultivars from Spain Sci. Hortic. 2006 107 306 314 10.1016/j.scienta.2005.08.008 52. De Vasconcelos M.C. Bennett R.N. Rosa E.A.S. Ferreira-Cardoso J.V. Metabolites composition of fresh kernels from three cultivars of Portuguese chestnut Proceedings of the II Iberian Congress on Chestnut Vila Real, Portugal 20-22 June 2007 Volume 784 43 48 53. Dimitri G. Mastrocola D. Sacchetti G. Neri L. Chemical composition and functional properties of three sweet chestnut (Castanea sativa Mill.) ecotypes from Italy Proceedings of the IV International Chestnut Symposium Beijing, China 25-28 September 2008 Volume 844 41 46 54. Alasalvar C. Shahidi F. Liyanapathirana C.M. Ohshima T. Turkish tombul hazelnut (Corylus avellana L.). 1. Compositional characteristics J. Agric. Food Chem. 2003 51 3790 3796 10.1021/jf0212385 12797745 55. Savage G.P. McNeil D.L. Chemical composition of hazelnuts (Corylus avellana L.) grown in New Zealand Int. J. Food Sci. Nutr. 1998 49 199 203 10.3109/09637489809086412 10616661 56. Solar A. Stampar F. Characterisation of selected hazelnut cultivars: Phenology, growing and yielding capacity, market quality and nutraceutical value J. Sci. Food Agric. 2011 91 1205 1212 10.1002/jsfa.4300 21328360 57. Kamangar T. Farsam H. Composition of pistachio kernels of various Iranian origins J. Food Sci. 1977 42 1135 1136 10.1111/j.1365-2621.1977.tb12689.x 58. Maskan M. Karatas S. Storage stability of whole-split pistachio nuts (Pistachia vera L.) at various conditions Food Chem. 1999 66 227 233 10.1016/S0308-8146(99)00055-2 59. Dreher M.L. Pistachio nuts: Composition and potential health benefits Nutr. Rev. 2012 70 234 240 10.1111/j.1753-4887.2011.00467.x 22458696 60. Bullo M. Juanola-Falgarona M. Hernandez-Alonso P. Salas-Salvado J. Nutrition attributes and health effects of pistachio nuts Br. J. Nutr. 2015 113 S79 S93 10.1017/S0007114514003250 26148925 61. Terzo S. Baldassano S. Caldara G.F. Ferrantelli V. Lo Dico G. Mule F. Amato A. Health benefits of pistachios consumption Nat. Prod. Res. 2019 33 715 726 10.1080/14786419.2017.1408093 29241364 62. Ozkan G. Koyuncu M.A. Physical and chemical composition of some walnut (Juglans regia L.) genotypes grown in Turkey Grasas y Aceites 2005 56 141 146 10.3989/gya.2005.v56.i2.122 63. Savage G.P. Chemical composition of walnuts (Juglans regia L.) grown in New Zealand Plant Foods Hum. Nutr. 2001 56 75 82 10.1023/A:1008175606698 11213171 64. Gharibzahedi S.M.T. Mousavi S.M. Ghahderijani M. Dadashpour A. Assessment of physical and chemical aspects of new Persian walnut cultivars to optimize process conditions Int. J. Food Eng. 2012 8 1 10.1515/1556-3758.2168 65. Gharibzahedi S.M.T. Mousavi S.M. Hamedi M. Khodaiyan F. Determination and characterization of kernel biochemical composition and functional compounds of Persian walnut oil J. Food Sci. Technol. 2014 51 34 42 10.1007/s13197-011-0481-2 24426045 66. Ozcan M.M. Some nutritional characteristics of fruit and oil of walnut (Juglans regia L.) growing in Turkey Iran. J. Chem. Chem. Eng. 2009 28 57 62 10.30492/IJCCE.2009.6915 67. Ozcan M.M. Iman C. Arslan D. Physicochemical properties, fatty acid and mineral content of some walnuts (Juglans regia L.) types Agric. Sci. 2010 1 62 10.4236/as.2010.12009 68. Amarowicz R. Gong Y. Pegg R.B. Recent advances in our knowledge of the biological properties of nuts Wild Plants, Mushrooms and Nuts: Functional Food Properties and Applications Ferreira I.C.F.R. Morales P. Barros L. John Wiley & Sons, Ltd. London, UK 2017 377 409 10.1002/9781118944653.ch11 9781118944653 69. Summo C. Palasciano M. De Angelis D. Paradiso V.M. Caponio F. Pasqualone A. Evaluation of the chemical and nutritional characteristics of almonds (Prunus dulcis (Mill). DA Webb) as influenced by harvest time and cultivar J. Sci. Food Agric. 2018 98 5647 5655 10.1002/jsfa.9110 29708600 70. Barreca D. Nabavi S.M. Sureda A. Rasekhian M. Raciti R. Silva A.S. Annunziata G. Arnone A. Tenore G.C. Suntar I. Almonds (Prunus dulcis Mill. D. A. Webb): A source of nutrients and health-promoting compounds Nutrients 2020 12 672 10.3390/nu12030672 32121549 71. Oliveira I. Meyer A.S. Afonso S. Aires A. Goufo P. Trindade H. Goncalves B. Phenolic and fatty acid profiles, a-tocopherol and sucrose contents, and antioxidant capacities of understudied Portuguese almond cultivars J. Food Biochem. 2019 43 e12887 10.1111/jfbc.12887 31353720 72. Oliveira I. Meyer A.S. Afonso S. Sequeira A. Vilela A. Goufo P. Trindade H. Goncalves B. Effects of different processing treatments on almond (Prunus dulcis) bioactive compounds, antioxidant activities, fatty acids, and sensorial characteristics Plants 2020 9 1627 10.3390/plants9111627 33238547 73. Alasalvar C. Bolling B.W. Review of nut phytochemicals, fat-soluble bioactives, antioxidant components and health effects Br. J. Nutr. 2015 113 S68 S78 10.1017/S0007114514003729 26148924 74. Kaur N. Chugh V. Gupta A. Essential fatty acids as functional components of foods--A review J. Food Sci. Technol. 2014 51 2289 2303 10.1007/s13197-012-0677-0 25328170 75. Barreira J.C.M. Ferreira I.C.F.R. Oliveira M.B.P.P. Bioactive compounds of Chestnut (Castanea sativa Mill.) BT Bioactive Compounds in Underutilized Fruits and Nuts Murthy H.N. Bapat V.A. Springer International Publishing Cham, Switzerland 2020 303 313 978-3-030-30182-8 76. Borges O.P. Carvalho J.S. Correia P.R. Silva A.P. Lipid and fatty acid profiles of Castanea sativa Mill. Chestnuts of 17 native Portuguese cultivars J. Food Comp. Anal. 2007 20 80 89 10.1016/j.jfca.2006.07.008 77. Benatti P. Peluso G. Nicolai R. Calvani M. Polyunsaturated fatty acids: Biochemical, nutritional and epigenetic properties J. Am. Coll. Nutr. 2004 23 281 302 10.1080/07315724.2004.10719371 15310732 78. Espana M.S.A. Galdon B.R. Romero C.D. Rodriguez E.R. Fatty acid profile in varieties of chestnut fruits from Tenerife (Spain) CyTA-J. Food 2011 9 77 81 10.1080/19476331003686858 79. Lucchetti S. Ambra R. Pastore G. Effects of peeling and/or toasting on the presence of tocopherols and phenolic compounds in four Italian hazelnut cultivars Eur. Food Res. Technol. 2018 244 1057 1064 10.1007/s00217-017-3028-6 80. Pannico A. Cirillo C. Giaccone M. Scognamiglio P. Romano R. Caporaso N. Sacchi R. Basile B. Fruit position within the canopy affects kernel lipid composition of hazelnuts J. Sci. Food Agric. 2017 97 4790 4799 10.1002/jsfa.8348 28374458 81. Oliveira I. Sousa A. Morais J.S. Ferreira I.C. Bento A. Estevinho L. Pereira J.A. Chemical composition, and antioxidant and antimicrobial activities of three hazelnut (Corylus avellana L.) cultivars Food Chem. Toxicol. 2008 46 1801 1807 10.1016/j.fct.2008.01.026 18316150 82. Kalogeropoulos N. Chiou A. Ioannou M.S. Karathanos V.T. Nutritional evaluation and health promoting activities of nuts and seeds cultivated in Greece Int. J. Food Sci. Nutr. 2013 64 757 767 10.3109/09637486.2013.793298 23641668 83. Freitas J.B. Naves M.M.V. Chemical composition of nuts and edible seeds and their relation to nutrition and health Rev. Nutr. 2010 23 269 279 10.1590/S1415-52732010000200010 84. Koksal A.I. Artik N. Simsek A. Gunes N. Nutrient composition of hazelnut (Corylus avellana L.) varieties cultivated in Turkey Food Chem. 2006 99 509 515 10.1016/j.foodchem.2005.08.013 85. Turan A. Effect of drying methods on fatty acid profile and oil oxidation of hazelnut oil during storage Eur. Food Res. Technol. 2018 244 2181 2190 10.1007/s00217-018-3128-y 86. Rodriguez-Bencomo J.J. Kelebek H. Sonmezdag A.S. Rodriguez-Alcala L.M. Fontecha J. Selli S. Characterization of the aroma-active, phenolic, and lipid profiles of the pistachio (Pistacia vera l.) nut as affected by the single and double roasting process J. Agric. Food Chem. 2015 63 7830 7839 10.1021/acs.jafc.5b02576 26301818 87. Roozban M.R. Mohamadi N. Vahdati K. Fat content and fatty acid composition of four Iranian Pistachio (Pistacia vera L.) varieties grown in Iran Proceedings of the IV International Symposium on Pistachios and Almonds Tehran, Iran 22-25 May 2005 Volume 726 573 578 88. Harmankaya M. Ozcan M.M. Juhaimi F.A. Mineral contents and proximate composition of Pistacia vera kernels Environ. Monit. Assess. 2014 186 4217 4221 10.1007/s10661-014-3692-y 24676991 89. Rabadan A. Alvarez-Orti M. Gomez R. de Miguel C. Pardo J.E. Influence of genotype and crop year in the chemometrics of almond and pistachio oils J. Sci. Food Agric. 2018 98 2402 2410 10.1002/jsfa.8732 29023801 90. Zribi F. Ghrab M. Mnafki N. Ayadi M. Ben Mimoun M. Biochemical characterization of pistachio germplasm grown in Tunisia Proceedings of the V International Symposium on Pistachios and Almonds Sanliurfa, Turkey 6-10 October 2009 Volume 912 469 476 91. Catalan L. Alvarez-Orti M. Pardo-Gimenez A. Gomez R. Rabadan A. Pardo J.E. Pistachio oil: A review on its chemical composition, extraction systems, and uses Eur. J. Lipid Sci. 2017 119 1600126 10.1002/ejlt.201600126 92. Muradoglu F. Oguz H.I. Yildiz K. Some chemical composition of walnut (Juglans regia L.) selections from Eastern Turkey Afr. J. Agric. Res. 2010 5 2379 2385 93. Beyhan O. Ozcan A. Ozcan H. Kafkas E. Kafkas S. Sutyemez M. Ercisli S. Fat, fatty acids and tocopherol content of several walnut genotypes Not. Bot. Horti Agrobot. Cluj Napoca 2017 45 437 441 10.15835/nbha45210932 94. Pereira J.A. Oliveira I. Sousa A. Ferreira I.C. Bento A. Estevinho L. Bioactive properties and chemical composition of six walnut (Juglans regia L.) cultivars Food Chem. Toxicol. 2008 46 2103 2111 10.1016/j.fct.2008.02.002 18334279 95. Amaral J.S. Casal S. Pereira J.A. Seabra R.M. Oliveira B.P. Determination of sterol and fatty acid compositions, oxidative stability, and nutritional value of six walnut (Juglans regia L.) cultivars grown in Portugal J. Agric. Food Chem. 2003 51 7698 7702 10.1021/jf030451d 14664531 96. Erturk U. Sisman T. Yerlikaya C. Erturk O. Karadeniz T. Chemical Composition and Nutritive Value of Selected Walnuts (Juglans regia L.) from Turkey Proceedings of the VII International Walnut Symposium Taiyuan, China 20-23 July 2013 Volume 1050 231 234 97. Zhai M.Z. Wang D. Tao X.D. Wang Z.Y. Fatty acid compositions and tocopherol concentrations in the oils of 11 varieties of walnut (Juglans regia L.) grown at Xinjiang, China J. Hortic. Sci. Biotechnol. 2015 90 715 718 10.1080/14620316.2015.11668736 98. Hayes D. Angove M.J. Tucci J. Dennis C. Walnuts (Juglans regia) chemical composition and research in human health Crit Rev. Food Sci. Nutr. 2016 56 1231 1241 10.1080/10408398.2012.760516 25747270 99. Vinson J.A. Cai Y. Nuts, especially walnuts, have both antioxidant quantity and efficacy and exhibit significant potential health benefits Food Funct. 2012 3 134 140 10.1039/C2FO10152A 22187094 100. Pribis P. Bailey R.N. Russell A.A. Kilsby M.A. Hernandez M. Craig W.J. Grajales T. Shavlik D.J. Sabate J. Effects of walnut consumption on cognitive performance in young adults Br. J. Nutr. 2012 107 1393 1401 10.1017/S0007114511004302 21923981 101. Poulose S.M. Miller M.G. Shukitt-Hale B. Role of walnuts in maintaining brain health with age J. Nutr. 2014 144 561S 566S 10.3945/jn.113.184838 24500933 102. Lamuel-Raventos R.M. St. Onge M.-P. Prebiotic nut compounds and human microbiota Crit. Rev. Food Sci. Nutr. 2017 57 3154 3163 10.1080/10408398.2015.1096763 27224877 103. Gervasi T. Barreca D. Lagana G. Mandalari G. Health benefits related to tree nut consumption and their bioactive compounds Int. J. Mol. Sci. 2021 22 5960 10.3390/ijms22115960 34073096 104. Sabate J. Wien M. Nuts, blood lipids and cardiovascular disease Asia Pac. J. Clin. Nutr. 2010 19 131 136 20199998 105. Del Gobbo L.C. Falk M.C. Feldman R. Lewis K. Mozaffarian D. Effects of tree nuts on blood lipids, apolipoproteins, and blood pressure: Systematic review, meta-analysis, and dose-response of 61 controlled intervention trials Am. J. Clin. Nutr. 2015 102 1347 1356 10.3945/ajcn.115.110965 26561616 106. Cardona F. Andres-Lacueva C. Tulipani S. Tinahones F.J. Queipo-Ortuno M.I. Benefits of polyphenols on gut microbiota and implications on human health J. Nutr. Biochem. 2013 24 1415 1422 10.1016/j.jnutbio.2013.05.001 23849454 107. Liu F. Chang X. Hu X. Brennan C.S. Guo X. Effect of thermal processing on phenolic profiles and antioxidant activities in Castanea mollissima Int. J. Food Sci. Technol. 2017 52 439 447 10.1111/ijfs.13299 108. Smeriglio A. Mandalari G. Bisigano C. Filocarmo A. Barreca D. Belloco E. Trombeta D. Polyphenolic content and biological properties of Avola almond (Prunus dulcis Mill. D.A. Webb) skin and its industrial by products Ind. Crops Prod. 2015 83 283 293 10.1016/j.indcrop.2015.11.089 109. Tas G.N. Gokmen V. Phenolic compounds in natural and roasted nuts and their skins: A brief review Curr. Opin. Food Sci. 2017 14 103 109 10.1016/j.cofs.2017.03.001 110. Wojdylo A. Turkiewicz I.P. Tkacz K. Nowicka P. Bobak L. Nuts as functional foods: Variation of nutritional and phytochemical profile and their in vitro bioactive properties Food Chem. X 2022 15 100418 10.1016/j.fochx.2022.100418 36211787 111. Chandrasekara N. Shahidi F. Effect of roasting on phenolic content and antioxidant activities of whole cashew nuts, kernels, and testa J. Agric. Food Chem. 2011 59 5006 5014 10.1021/jf2000772 21438525 112. Dolatabadi K.S.M. Dehghan G. Hosseini S. Jahanban Esfahlan A. Effect of five years storage on total phenolic content and antioxidant capacity of almond (Amygdalus communis L.) hull and shell from different genotypes Avicenna J. Phytomed. 2015 5 26 33 25767754 113. Pycia K. Kapusta I. Jaworska G. Changes in antioxidant activity, profile, and content of polyphenols and tocopherols in common hazel Seed (Corylus avellana L.) depending on variety and harvest date Molecules 2019 25 43 10.3390/molecules25010043 31877675 114. Granata M.U. Bracco F. Catoni R. Cavalloro V. Martino E. Secondary metabolites profile and physiological leaf traits in wild and cultivated Corylus avellana under different nutritional status Nat. Prod. Res. 2019 30 3100 3107 10.1080/14786419.2019.1682577 115. Caltagirone C. Peano C. Sottile F. Post-harvest industrial processes of almond (Prunus dulcis L. Mill) in Sicily influence the nutraceutical properties of by-products at harvest and during storage Front. Nutr. 2021 8 659378 10.3389/fnut.2021.659378 34150827 116. Bhagwat S. Haytowitz D.B. Holden J.M. USDA Database for the Flavonoid Content of Selected Foods, Release 3.1 Available online: (accessed on 18 October 2018) 117. Gultekin-Ozgunven M. Davarci F. Pasli A.A. Demir N. Ozcelic B. Determination of phenolic compounds by ultra-high liquid chromatography-tandem mass spectrometry: Applications in nuts LWT-Food Sci. Technol. 2015 64 42 49 10.1016/j.lwt.2015.05.014 118. Abe L.T. Lajolo F.M. Genovese M.I. Comparison of phenol content and antioxidant capacity of nuts Food Sci. Technol. 2010 1 (Suppl. 30) 254 259 10.1590/S0101-20612010000500038 119. Otles S. Seleck I. Phenolics compounds and antioxidant activities of chestnut (Castanea sativa Mill.) fruits Qual. Assur. Saf. Crops Foods 2012 4 199 205 10.1111/j.1757-837X.2012.00180.x 120. Ozdemir K.S. Yilmaz C. Durmaz G. Gokmen V. Hazelnut skin powder: A new brown colored functional ingredient Food Res. Int. Part B 2014 65 291 297 10.1016/j.foodres.2014.01.060 121. Bertolino M. Belviso S. Bello B.D. Ghirardello D. Giordano M. Rolle L. Gerbi V. Zeppa G. Influence of the addition of different hazelnut skins on the physicochemical, antioxidant, polyphenol and sensory properties of yogurt LWT-Food Sci. Technol. 2015 63 1145 1154 10.1016/j.lwt.2015.03.113 122. Yuan B. Lu M. Eskridge K.M. Isom L.D. Hanna M.A. Extraction, identification, and quantification of antioxidant phenolics from hazelnut (Corylus avellana L.) shells Food Chem. 2018 244 7 15 10.1016/j.foodchem.2017.09.116 29120806 123. Cabo S. Aires A. Carvalho R. Vilela A. Pascual-Seva N. Silva A.P. Goncalves B. Kaolin, Ascophyllum nodosum and salicylic acid mitigate effects of summer stress improving hazelnut quality J. Sci. Food Agric. 2021 101 459 475 10.1002/jsfa.10655 32648605 124. Sales J.M. Resurreccion A.V.A. Phenolic profile, antioxidants, and sensory acceptance of bioactive-enhanced peanuts using ultrasound and UV Food Chem. 2010 122 795 803 10.1016/j.foodchem.2010.03.058 125. De Camargo A.C. Regitano-d'Arce M.A.B. Gallo C.R. Shahidi F. Gamma-irradiation induced changes in microbiological status, phenolic profile and antioxidant activity of peanut skin J. Funct. Foods 2015 12 129 143 10.1016/j.jff.2014.10.034 126. Attree R. Du B. Xu B. Distribution of phenolic compounds in seed coat and cotyledon, and their contribution to antioxidant capacities of red and black seed coat peanuts (Arachis hypogaea L.) Ind. Crops Prod. 2017 67 448 456 10.1016/j.indcrop.2015.01.080 127. Grace M.H. Esposito D. Timmers M.A. Xiong J. Yousef G. Komarnytskya S. Lila M.A. In vitro lipolytic, antioxidant and anti-inflammatory activities of roasted pistachio kernel and skin constituents Food Funct. 2016 7 4285 4298 10.1039/C6FO00867D 27604963 128. Jakopic J. Verberic R. Stampar F. Extraction of phenolic compounds from green walnut fruits in different solvents Acta Agric. Slov. 2009 93 11 15 10.2478/v10014-009-0002-4 129. Bujdso G. Vegvari G. Hajnal V. Ficzek G. Toth M. Phenolic profile of the kernel of selected Persian walnut (Juglans regia L.) cultivars Not. Bot. Horti Agrobot. 2014 42 24 29 10.15835/nbha4219400 130. Brown R.C. Gray A.R. Tey S.L. Chisholm A. Burley V. Greenwood D.C. Cade J. Associations between nut consumption and health vary between omnivores, vegetarians, and vegans Nutrients 2017 9 1219 10.3390/nu9111219 29113145 131. Musarra-Pizzo M. Ginestra G. Smeriglio A. Pennisi R. Sciortino M.T. Mandalari G. The antimicrobial and antiviral activity of polyphenols from almond (Prunus dulcis L.) skin Nutrients 2019 11 2355 10.3390/nu11102355 31623329 132. Gorji N. Moeini R. Memariani Z. Almond, hazelnut and walnut, three nuts for neuroprotection in Alzheimer's disease: A neuropharmacological review of their bioactive constituents Pharmacol. Res. 2018 129 115 127 10.1016/j.phrs.2017.12.003 29208493 133. Mexis S. Badeka A. Kontominas M. Quality evaluation of raw ground almond kernels (Prunus dulcis): Effect of active and modified atmosphere packaging, container oxygen barrier and storage conditions Innov. Food Sci. Emerg. Technol. 2009 10 580 589 10.1016/j.ifset.2009.05.002 134. Lee J. Xiao L. Zhang G. Ebeler S. Mitchell A. Influence of storage on volatile profiles in roasted almonds (Prunus dulcis) J. Agric. Food. Chem. 2014 62 11236 11245 10.1021/jf503817g 25347382 135. Valdes A. Beltran A. Karabagias I. Badeka A. Kontominas M. Garrigos M. Monitoring the oxidative stability and volatiles in blanched, roasted and fried almonds under normal and accelerated storage conditions by DSC, thermogravimetric analysis and ATR-FTIR Eur. J. Lipid Sci. Technol. 2015 117 1199 1213 10.1002/ejlt.201400384 136. Erten E. Cadwallader K. Identification of predominant aroma components of raw, dry roasted and oil roasted almonds Food Chem. 2017 217 244 253 10.1016/j.foodchem.2016.08.091 27664632 137. Oliveira I. Malheiro R. Meyer A.S. Pereira J.A. Goncalves B. Application of chemometric tools for the comparison of volatile profile from raw and roasted regional and foreign almond cultivars (Prunus dulcis) J. Food Sci. Technol. 2019 56 3764 3776 10.1007/s13197-019-03847-x 31413403 138. Garcia A.V. Romero R.S. Polo A.J. Moya S.P. Perez S.E.M. Sanahuja A.B. Volatile profile of nuts, key odorants and analytical methods for quantification Foods 2021 10 1611 10.3390/foods10071611 34359483 139. Elmore J.S. Nisyrios I. Mottram D.S. Analysis of the headspace aroma compounds of walnuts (Juglans regia L.) Flavour Fragr. J. 2005 20 501 506 10.1002/ffj.1477 140. Kiefl J. Schieberle P. Evaluation of process parameters governing the aroma generation in three hazelnut cultivars (Corylus avellana L.) by correlating quantitative key odorant profiling with sensory evaluation J. Agric. Food Chem. 2013 61 5236 5244 10.1021/jf4008086 23663154 141. Mexis S.F. Kontominas M.G. Effect of g-irradiation on the physicochemical and sensory properties of hazelnuts (Corylus avellana L.) Radiat. Phys. Chem. 2009 78 407 413 10.1016/j.radphyschem.2009.03.008 142. Wilson-Kakashita G. Gerdes D.L. Hall W.R. The effect of gamma irradiation on the quality of English walnuts (Juglans regia) LWT-Food Sci. Technol. 1995 28 17 20 10.1016/S0023-6438(95)80006-9 143. Duduzile Buthelezi N.M. Samukelo Magwaza L. Zeray Tesfay S. Postharvest pre-storage processing improves antioxidants, nutritional and sensory quality of macadamia nuts Sci. Hortic. 2019 251 197 208 10.1016/j.scienta.2019.03.026 144. Chang S.K. Alasalvar C. Bolling B.W. Shahidi F. Nuts and their co-products: The impact of processing (roasting) on phenolics, bioavailability, and health benefits--A comprehensive review J. Funct. Foods 2016 26 88 122 10.1016/j.jff.2016.06.029 145. Hojjati M. Calin-Sanchez A. Razavi S.H. Carbonell-Barrachina A.A. Effect of roasting on colour and volatile composition of pistachios (Pistacia vera L.) Int. J. Food Sci. 2013 48 437 443 10.1111/j.1365-2621.2012.03206.x 146. Saklar S. Katnas S. Ungan S. Determination of optimum hazelnut roasting conditions Int. J. Food Sci. 2001 36 271 281 10.1046/j.1365-2621.2001.00457.x 147. Tas N.G. Gokmen V. Maillard reaction and caramelization during hazelnut roasting: A multiresponse kinetic study Food Chem. 2017 221 1911 1922 27979180 148. Lin J.T. Liu S.C. Hu C.C. Shyu C.Y. Hsu Y.D.J. Effects of roasting temperature and duration on fatty acid composition, phenolic composition, Maillard reaction degree and antioxidant attribute of almond (Prunus dulcis) kernel Food Chem. 2016 190 520 528 10.1016/j.foodchem.2015.06.004 26213005 149. Schmitzer V. Slatnar A. Veberic R. Stampar F. Solar A. Roasting affects phenolic composition and antioxidative activity of hazelnuts (Corylus avellana L.) J. Food Sci. 2011 76 S14 S19 10.1111/j.1750-3841.2010.01898.x 21535710 150. Alamprese C. Ratti S. Rossi M. Effects of roasting conditions on hazelnut characteristics in a two-step process J. Food Eng. 2009 95 272 279 10.1016/j.jfoodeng.2009.05.001 151. Cuadrado C. Sanchiz A. Linacero R. Nut Allergenicity: Effect of food processing Allergies 2021 1 150 162 10.3390/allergies1030014 152. Maleki S.J. Chung S.Y. Champagne E.T. Raufman J.P. The effects of roasting on the allergenic properties of peanut proteins J. Allergy Clin. Immunol. 2000 106 763 768 10.1067/mai.2000.109620 11031348 153. Venkatachalam M. Teuber S.S. Roux K.H. Sathe S.K. Effects of roasting, blanching, autoclaving, and microwave heating on antigenicity of almond (Prunus dulcis L.) proteins J. Agric. Food Chem. 2002 50 3544 3548 10.1021/jf020012z 12033826 154. Schlormann W. Birringer M. Bohm V. Lober K. Jahreis G. Lorkowski S. Muller A.K. Schone F. Glei M. Influence of roasting conditions on health-related compounds in different nuts Food Chem. 2015 180 77 85 10.1016/j.foodchem.2015.02.017 25766804 155. Arinola S.O. Adesina K. Effect of thermal processing on the nutritional, antinutritional, and antioxidant properties of Tetracarpidium conophorum (African walnut) J. Food Proc. 2014 2014 418380 10.1155/2014/418380 156. Kita A. Figiel A. Effect of roasting on properties of walnuts Pol. J. Food Nutr. Sci. 2007 57 89 94 157. Tian Y. Rao H. Zhang K. Tao S. Xue W.T. Effects of different thermal processing methods on the structure and allergenicity of peanut allergen Ara h 1 Food Sci. Nutr. 2018 6 1706 1714 10.1002/fsn3.742 30258615 158. Verhoeckx K.C.M. Vissers Y.M. Baumert J.L. Faludi R. Feys M. Flanagan S. Herouet-Guicheney C. Holzhauser T. Shimojo R. van der Bolt N. Food processing and allergenicity Food Chem. Toxicol. 2015 80 223 240 10.1016/j.fct.2015.03.005 25778347 159. Cuadrado C. Cheng H. Sanchiz A. Ballesteros I. Easson M. Grimm C.C. Dieguez M.C. Linacero R. Burbano C. Maleki S.J. Influence of enzymatic hydrolysis on the allergenic reactivity of processed cashew and pistachio Food Chem. 2018 241 372 379 10.1016/j.foodchem.2017.08.120 28958542 160. Samoggia A. Nicolodi S. Consumer's perception of fruit innovation J. Int. Food Agribus. Mark. 2017 9 92 108 10.1080/08974438.2016.1266567 161. Patil B.S. Uckoo R.M. Jayaprakasha G.K. Palma M.A. Consumers' changing perceptions of quality: Revisiting the science of fruit and vegetable cultivation for improved health benefits Acta Hortic. 2016 1120 459 468 10.17660/ActaHortic.2016.1120.71 162. Van Duyn M.A.S. Year 2000 Dietary Guidelines: The Case for Fruits and Vegetables First Produce for Better Health Foundation Wilmington, DE, USA 1999 163. Kader A. Importance of fruits, nuts and vegetables in human nutrition and health Perish. Handl. Q. 2001 106 6 164. Strohle A. Vegetables and fruits in prevention: The German Nutrition Society (DGE) opinion confirms: High consumption of vegetables and fruits reduces risk of contracting diseases Dtsch. Apoth. Ztg. 2012 152 75 77 10.1007/s00394-012-0380-y 165. Scheerens J.C. Phytochemicals and the consumer: Factors affecting fruit and vegetable consumption and the potential for increasing small fruit in the diet Hort Technol. 2001 11 547 556 10.21273/HORTTECH.11.4.547 166. Kris-Etherton P.M. Hecker K.D. Bonanome A. Coval S.M. Binkoski A.E. Hilpert K.F. Etherton T.D. Bioactive compounds in foods: Their role in the prevention of cardiovascular disease and cancer Am. J. Med. 2002 113 71 88 10.1016/S0002-9343(01)00995-0 167. Jenab M. Sabate J. Slimani N. Ferrari P. Mazuir M. Casagrande C. Deharveng G. Tjonneland A. Olsen A. Overvad K. Consumption and portion sizes of tree nuts, peanuts and seeds in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohorts from 10 European countries Br. J. Nutr. 2006 96 S12 S23 10.1017/BJN20061859 168. O'Neil C.E. Keast D.R. Nicklas T.A. Fulgoni V.L. Out-of-hand nut consumption is associated with improved nutrient intake and health risk markers in US children and adults: National Health and Nutrition Examination Survey 1999-2004 Nutr. Res. 2012 32 185 194 10.1016/j.nutres.2012.01.005 22464805 169. O'Neil C.E. Nicklas T.A. Fulgoni V.L. Tree nut consumption is associated with better nutrient adequacy and diet quality in adults: National Health and Nutrition Examination Survey 2005-2010 Nutrients 2015 7 595 607 10.3390/nu7010595 25599274 170. Flores-Mateo G. Rojas-Rueda D. Basora J. Ros E. Salas-Salvado J. Nut intake and adiposity: Meta-analysis of clinical trials Am. J. Clin. Nutr. 2013 97 346 1355 10.3945/ajcn.111.031484 171. Alasalvar C. Shahidi F. Tree Nuts: Composition, Phytochemicals, and Health Effects CRC Press Boca Raton, FL, USA 2008 340p 9780849337352 172. Wang J. Lee Bravatti M.A. Johnson E.J. Raman G. Daily almond consumption in cardiovascular disease prevention via LDL-C change in the U.S. population: A cost-effectiveness analysis BMC Public Health 2020 20 558 10.1186/s12889-020-08642-4 32334551 173. Pawlak R. Colby S. Herring J. Beliefs, benefits, barriers, attitude, intake and knowledge about peanuts and tree nuts among WIC participants in eastern North Carolina Nutr. Res. Pract. 2009 3 220 225 10.4162/nrp.2009.3.3.220 20090888 174. Pawlak R. London H.A. Colby S. Wall-Bassett E. Sira N. Perception of nut intake among individuals with or at risk for heart disease and/or diabetes J. Behav. Health 2012 1 185 188 10.5455/jbh.20120501060143 175. Souza R.G. Gomes A.C. Naves M.M. Mota J.F. Nuts and legume seeds for cardiovascular risk reduction: Scientific evidence and mechanisms of action Nutr. Rev. 2015 73 335 347 10.1093/nutrit/nuu008 26011909 176. Yong L.C. Gray A.R. Chisholm A. Leong S.L. Tey S.L. Brown R.C. Barriers to and facilitators and perceptions of nut consumption among the general population in New Zealand Public Health Nutr. 2017 20 3166 3182 10.1017/S1368980017002464 28965535 177. Li H. Li X. Yuan S. Jin Y. Lu J. Nut consumption and risk of metabolic syndrome and overweight/obesity: A meta-analysis of prospective cohort studies and randomized trials Nutr. Metab. 2018 15 46 10.1186/s12986-018-0282-y 178. Viale P.H. The benefits of nuts for cancer prevention J. Adv. Pract. Oncol. 2019 10 102 103 10.6004/jadpro.2019.10.2.1 31538022 179. Shirmohammadi M. Chandrasekaran I. Singh C. Effect of post-harvest processes and storage conditions on aging and quality of fruit nuts Proceedings of the CSBE/SCGAB 2018 Annual Conference School of Engineering, University of Guelph, Guelph, ON, Canada 22-25 July 2018 Available online: (accessed on 17 October 2022) 180. Kader A.A. Impact of nut postharvest handling, deshelling, drying and storage on quality Improving the Safety and Quality of Nuts Harris L.J. Woodhead Publishing Ltd. Cambridge, UK 2013 Volume 1 22 34 181. Mexis S.F. Badeka A.V. Riganakos K.A. Karakostas K.X. Kontominas M.G. Effect of packaging and storage conditions on quality of shelled walnuts Food Control 2009 20 743 751 10.1016/j.foodcont.2008.09.022 182. Fernandes L. Pereira E.L. Fidalgo M.C. Gomes A. Ramalhosa E. Effect of modified atmosphere, vacuum and polyethylene packaging on physicochemical and microbial quality of chestnuts (Castanea sativa) during storage Int. J. Fruit Sci. 2020 20 S785 S801 10.1080/15538362.2020.1768619 183. Koster E.P. Diversity in the determinants of food choice: A psychological perspective Food Qual. Prefer. 2009 20 70 82 10.1016/j.foodqual.2007.11.002 184. Grunert K.G. Larsen H.H. Madsen T.K. Baadsgaard A. Market Orientation in Food and Agriculture Kluwer Academic Dordrecht, The Netherlands 1996 185. Brunso K. Birch D. Memery J. Temesi A. Lakner Z. Lang M. Dean D. Grunert K.G. Core dimensions of food-related lifestyle: A new instrument for measuring food involvement, innovativeness and responsibility Food Qual. Prefer. 2021 91 104192 10.1016/j.foodqual.2021.104192 |
PMC10000570 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050649 healthcare-11-00649 Article Awareness of Human Papillomavirus among Male and Female University Students in Saudi Arabia Aldawood Esraa Conceptualization Validation Formal analysis Investigation Data curation Writing - original draft Writing - review & editing Supervision 1* Alzamil Lama Writing - original draft Writing - review & editing 1 Faqih Layla Writing - original draft 1 Dabbagh Deemah Conceptualization Formal analysis Writing - original draft Visualization 1 Alharbi Sarah Data curation Writing - review & editing 1 Hafiz Taghreed A. Writing - review & editing 1 Alshurafa Hassan H. Methodology Investigation Data curation 2 Altukhais Wajd F. 1 Dabbagh Rufaidah Formal analysis Writing - review & editing 3 1 Department of Clinical Laboratories Sciences, College of Applied Medical Sciences, King Saud University, Riyadh 12372, Saudi Arabia 2 College of Medicine, King Saud University, Riyadh 11461, Saudi Arabia 3 Department of Family and Community Medicine, College of Medicine, King Saud University, Riyadh 11461, Saudi Arabia * Correspondence: [email protected] 23 2 2023 3 2023 11 5 64911 1 2023 18 2 2023 21 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Background: Human papillomavirus (HPV) is a common sexually transmitted infection globally. Investigating HPV awareness can reduce the burden of HPV-related cancers. Aims: (1) Assessing HPV awareness and knowledge among health college students at King Saud University, (2) comparing these outcomes across sociodemographic characteristics. Methods: A cross-sectional survey study was conducted from November to December 2022 and included 403 health college students. Associations of HPV awareness and knowledge with sociodemographic characteristics were assessed using logistic regression analysis and linear regression analysis, respectively. Results: Only 60% of students were aware of HPV, with awareness higher among females, although their knowledge scores were comparable to males. The odds of awareness of HPV were greater among medical students compared to other colleges and among students belonging to older age groups compared to the younger age group (18-20). The odds of HPV awareness among hepatitis B vaccinated students were 2.10 times that among unvaccinated students (AOR = 2.10; 95% CI = 1.21, 3.64). Conclusions: The low level of HPV awareness among college students warrants the need for HPV educational campaigns to improve HPV awareness and to promote HPV vaccination in the community. HPV awareness knowledge health colleges Deputyship for Research & Innovation, Ministry of Education, Saudi ArabiaIFKSURG-2-366 This research was financially supported by The Deputyship for Research & Innovation, Ministry of Education, Saudi Arabia, by funding this research work through project no. (IFKSURG-2-366). pmc1. Introduction Human papillomavirus (HPV) infection is the most prevalent sexually transmitted infection (STI) among sexually active men and women worldwide. HPV infects the squamous cell lining the inner surface of the cervix, oropharynx, anus, penis, vagina, and vulva . It is responsible for a variety of cutaneous and mucosal epithelial lesions, as well as 15 different cancers, including cervical, penile, anal, and oropharyngeal cancers . Cervical cancer is considered the fourth most common cancer among females . A pap smear is usually the gold standard for cervical cancer screening worldwide; however, due to sociocultural factors, the screening test coverage in Saudi Arabia is lower than in most countries . Although data on the prevalence of HPV infection and subsequently associated cancers are limited in Saudi Arabia, reported cancer statistics for the year 2020 showed an estimated number of 358 annual new cervical cancer cases in Saudi Arabia, with a crude mortality rate of 1.22 per 100,000 women per year . In recent years, the number of cervical cancer cases in developed countries has significantly dropped due to increased awareness of HPV infection and its vaccination . Studies found that HPV awareness and knowledge are important predictors for HPV vaccine acceptability . Moreover, it has been reported that physicians can positively influence parents' decisions regarding their children's HPV vaccine uptake . Therefore, educating future healthcare providers about HPV infection is crucial for HPV prevention. Globally, awareness of HPV is higher among healthcare professionals compared to the general population . In addition, studies indicate that factors such as increasing age, being female, and higher education level are positively correlated with HPV awareness . In Saudi Arabia, several studies targeting various populations in different regions reported unsatisfactory knowledge about HPV, pap smears, and the HPV vaccine . While most local studies targeted the female population and examined HPV awareness mainly in the context of cervical cancer , one study targeting male medical students alone revealed low HPV knowledge . Another study targeting both males and females from different regions in Saudi Arabia reported poor HPV knowledge among both genders, but more so among males . With the exception of the aforementioned studies, little is known about gender differences in HPV awareness and knowledge among Saudi university students, let alone students from health colleges. The aim of this study was to determine HPV awareness and knowledge among male and female students in health colleges at King Saud University (KSU), Riyadh, and to compare awareness and knowledge across sociodemographic characteristics. 2. Materials and Methods 2.1. Study Design and Setting We performed a cross-sectional survey-based study at the five health colleges: College of Medicine, College of Dentistry, College of Pharmacy, College of Applied Medical Sciences, and College of Nursing at KSU, Riyadh, Saudi Arabia. The sample size for this study was calculated to be 377 participants, using Raosoft, Inc. (Seattle, WA, USA) ) (accessed on 26 July 2022), with 95% confidence and an error rate of 5%. Nevertheless, we included more responses until we reached 403 participants to make the study more meaningful. 2.2. Data Collection Data were collected over two months, from November to December 2022, using a self-administered 28-item questionnaire, adapted from a study that surveyed male medical students in Jeddah, Saudi Arabia , who share similar cultural and medical backgrounds as our target population. A pilot study with 23 participants was conducted to test the questions' clarity and the time needed to answer them. However, the results of the pilot study were not included in the final analysis. The first section comprised questions on the following sociodemographic characteristics: age, gender, nationality, marital status, cumulative grade point average (GPA), smoking status, hepatitis B (HBV) vaccination, and history of STIs. To investigate the awareness of HPV, participants were asked whether or not they had heard of HPV. If their answer was yes, they were transferred to the second section which comprised 17 "True/False/I Don't Know" questions about HPV knowledge, formulated and validated by Waller et al. . Each correct response was given a score of 1, while each incorrect or "I Don't Know" response was given a 0 score; therefore, the total maximum possible score was 17. Based on the percentage of scores achieved by the participants, the knowledge scores were grouped into the following: good (>75%), fair (50-75%), and poor (<50%) . 2.3. Procedure for Data Collection After we received the ethical approval, the survey link, which was prepared using Google forms, was distributed to the participants. Eligibility criteria were being an undergraduate student and studying at a health college at KSU. Postgraduate students were excluded. Data collectors from all of these health colleges were recruited to distribute the link to their classmates and through social media. 2.4. Ethical Consideration Participants were aware of the study's purpose and were asked voluntarily to fill out the survey after informed consent. Responses were completely anonymous and confidential. The study was ethically approved by the Institutional Review Board, KSU (ref. No. 22/0843/IRB). All data were kept confidential and were available only to the research team. 2.5. Statistical Analysis We calculated percentages for categorical variables, while mean and standard deviation (SD) were calculated for the total knowledge scores. To assess the association of awareness with sociodemographic characteristics, we conducted a logistic regression model using binary awareness (yes vs. no) as the dependent variable, while age group, gender, college, history of HBV vaccination, history of STIs, and smoking status were independent variables, for which we report adjusted odds ratios (AORs) and 95% confidence intervals (CIs). To assess the association of HPV knowledge with the covariates, we conducted linear regression analysis using the knowledge score as the dependent variable, and the other covariates as independent variables, for which we report adjusted beta estimates and 95% CIs. The independent variables were added to the models because of their predictive properties based on the previous literature . The data were analyzed using the Statistical Package for Social Sciences (SPSS) version 26 (Armonk, NY, USA) for IBM. 3. Results 3.1. Sociodemographic Characteristics of the Respondents The survey was distributed to 450 students, out of which 403 (89.6%) responded with a completion rate of 100%, of whom 52.6% were female and 47.4% were male. Most of the respondents were Saudi nationals (98.5%) and belonged to the age group of 21-23 years (54.5%). Out of all the participating health colleges, the highest responses came from students at the College of Applied Medical Sciences (42.2%). The majority of respondents (98%) were single. Overall, 50.9% of the students had previously received the HBV vaccine and 2.5% had a history of STI (Table 1). 3.2. HPV Awareness and Knowledge 3.2.1. Awareness of HPV Out of the 403 respondents, 161 (40%) responded that they had never heard of HPV. Awareness of HPV was significantly lower among males (57.1%) compared to female students (62.7%) (p-value = 0.003) . When comparing awareness across colleges, students from the College of Medicine had a significantly higher proportion of awareness (91.8%) compared to the rest of the colleges (X2 = 101.4; df = 4; p-value < 0.001). They were followed by the College of Dentistry (89.2%), Pharmacy (77.2%), Nursing (50.0%), and Applied Medical Sciences (35.3%). Because the total number of students reporting awareness of HPV was 242 (60%), only 242 responded to the section regarding knowledge about HPV. 3.2.2. Knowledge about HPV Out of the 242 students responding to knowledge questions, the frequency of female students exhibiting fair or good knowledge about HPV was greater than that among males, yet not statistically significant (p-value = 0.502) . The overall mean knowledge score among these students was 8.8 (SD = 3.6). There was no significant difference between the mean knowledge scores for males (8.7, SD = 3.9) and females (8.9, SD = 3.5) (p-value = 0.675). With respect to individual knowledge questions, the highest proportion of correct responses was recorded for both genders in the item "HPV can be passed on during sexual intercourse", with 77.1% of males and 83.5% of females answering it correctly. On the other hand, the lowest proportion of correct responses for both genders was recorded in the item "HPV usually doesn't need any treatment", with only 11.9% of males and 13.5% of females responding correctly (Table 2). It is worth mentioning that around 50% of the students from both genders believed that HPV infection can be treated with antibiotics, and around 60% thought that HPV can cause HIV. 3.2.3. Association of Awareness and Knowledge with Sociodemographic Characteristics Results from binary analyses suggest that the respondents' level of knowledge regarding HPV was significantly associated with their age (kh2 = 40.293, p < 0.001) and college type (kh2 = 34.582, p < 0.001). Interestingly, there was a higher level of knowledge about HPV among students who had previously taken the HBV vaccine (kh2 = 11.926, p < 0.001) (Table 3). As shown in Table 3, the majority of students were Saudi and single. Thus, these two variables were not included in regression analyses. Compared to students who were between 18 and 20 years, the odds for being aware of HPV were around 3-fold among students aged 21 to 23 years (AOR = 2.99; 95% CI = 1.71, 5.22), and were around 4-fold for students between the ages of 24 and 26 years (AOR = 4.13; 95% CI = 1.03, 16.54), suggesting that awareness about HPV may be associated with increasing age. However, this trend was not observed with the age group of 27 years or older (Table 4). Although the odds for awareness of HPV were lower among male students compared to female students, this point-estimate was not statistically significant. Students enrolled in medical college seemed to have greater odds of awareness compared to those from other colleges. This was particularly significant for students from the applied medical sciences, in which their odds of awareness were 89% lower than those for students in medical college (AOR = 0.11; 95% CI = 0.04, 0.29), and from the College of Nursing who had 90% lower odds (AOR = 0.10; 95% CI = 0.04, 0.29). Additionally, having an HBV vaccine seemed to significantly increase the odds of awareness (AOR = 2.10; 95% CI = 1.21, 3.64). Surprisingly, smoking was associated with around 5-fold odds of having good knowledge about HPV (AOR = 4.95; 95% CI = 1.84, 13.31). Finally, and although not statistically significant, having a history of STI was associated with lower odds of awareness about HPV. Compared to students who were aged 18 to 20 years, on average, students aged 21 to 23 years had higher knowledge scores by 1.59 points (b = 1.59; 95% CI = 0.44, 2.74). On the other hand, and although not statistically significant, older students had lower knowledge scores (Table 5). As with awareness, the mean knowledge scores for students from the colleges of pharmacy, applied medical sciences, nursing, and dentistry were at least 2 points lower than that for students from the College of Medicine. However, this estimate was lower and not statistically significant for students from the College of Dentistry (b = -1.37; 95% CI = -2.79, 0.04). As with what was reported from bivariate analysis, on average, the knowledge score for males was not significantly lower than that for females when controlling for other covariates. Moreover, having the HBV vaccination and having an STI were not significantly associated with a change in knowledge score. 4. Discussion Because HPV infection is a preventable disease, it is crucial to fill the gap in knowledge for future healthcare providers before their enrollment in the workforce. To do so, we evaluated the HPV knowledge in health students from the five health colleges at KSU in the central province of Saudi Arabia. The response rate was 89.6%, and the students who did not respond did not do so mainly because they were not students at any health colleges at KSU or did not agree to participate. Those who responded answered all the survey questions completely. Findings from the current study can be summarized in the following points. First, 40% of the students in our study reported not hearing about HPV. Additionally, the proportion of awareness was significantly greater among females compared to males. This trend was corroborated by the results from multivariate analyses, albeit not statistically significant. Second, the average HPV knowledge scores were only 8.9 and 8.7 for females and males, respectively. Third, we could not find significant differences in knowledge correctness or knowledge scores between male and female students. Fourth, compared to the youngest student group, students between the ages of 21 and 26 had greater odds of awareness of HPV, controlling for other student characteristics. Fifth, when controlling for other covariates, the odds of awareness were greater for medical students compared to students from all other health specialties, as were the mean knowledge scores. Finally, having the HBV vaccine seemed to increase the odds of awareness of HPV. In this study, about 60% of the students were aware of HPV. Previous studies in Saudi Arabia reported awareness among college students ranging from 33.7% to 49.9% . Even lower estimates were reported among university students in Morocco (10%) . Although this estimate is greater than what has been reported in the Middle East and regionally, awareness among our study population still falls short compared to Western college students, in which awareness is reported as high as 95.3% . Thus, HPV infection and vaccination awareness should be raised in local universities for both women and men. Our study suggests that the overall awareness is greater among females. This comes in agreement with several international studies . Possible explanations for this discrepancy could be that the complications of HPV infection are more commonly manifested in women, in addition to the recent regional introduction of the HPV vaccination into the vaccination schedule of adolescent girls, but not boys. Although awareness significantly differed between males and females on bivariate analysis, this significant association was not maintained through multivariate analysis. Additionally, we could not find significant differences between the genders in terms of knowledge scores. This is contrary to the worldwide belief that women consistently have better knowledge about HPV compared to men . To the best of our knowledge, our study is the only local study to have compared HPV awareness and knowledge among university students of both genders within the central region of Saudi Arabia. A previous study conducted in the central province by AlShaikh et al. reported a lack of knowledge and misinformation regarding HPV among Saudi female university students alone, which corresponds with what we report . Similar studies have also been published from the western , north-western , and south-western regions of Saudi Arabia, reporting a lower prevalence of HPV awareness. For example, the study that was conducted among male medical students in Jeddah found that only 70% had ever heard of HPV . Another study reported that 73% of medical students had heard of HPV . Students in the College of Medicine had the highest proportion of awareness about HPV (91.8%) and scored the highest on HPV knowledge. Moreover, they had greater odds of awareness and knowledge about HPV compared to students from other health colleges. This has been repeatedly observed in other regional studies . Perhaps the rigorous exposure of medical students to infectious diseases in their curriculum might explain this observation. Raising awareness of HPV is needed in all health colleges. Awareness of HPV was more likely among students of greater age groups compared to those who were between 18 and 20. This has also been reported in a study from Turkey . However, this trend was not observed when comparing knowledge scores across the age groups. This contradicts the notion that knowledge about HPV increases with age among university students . On the other hand, one study that sampled the general population reported a slight decrease in knowledge with each one-year increase in age . In both studies, age was measured as a continuous variable, unlike our categorical recording of age in the questionnaire, which may have contributed to this discrepancy. We observed that having received the HBV vaccine increased the odds of HPV awareness and knowledge. This observation corresponds with previous reports that used the same methodology . Not only does HBV vaccination increase ones' awareness and knowledge of HPV, but it has been associated with a reduction in HPV vaccination hesitancy . Moreover, HBV vaccination may be an indicator for preventative care, which increases the chance of awareness. Targeting both viruses with vaccination programs may help increase community acceptance for vaccine reception, aiding in the simultaneous prevention and control of both HBV and HPV infections. Because the subject of STIs is greatly affected by society, in a culturally and religiously conservative country such as Saudi Arabia, sexual education is not well-provided during school years. Thus, we find a need to fill the gap in knowledge in the community, starting by improving the overall awareness and knowledge of future healthcare providers. This is crucial for preparing health students for their future role as public health educators and potential immunizers. The current study has highlighted the urgent need to increase HPV awareness and knowledge among health college students in Saudi Arabia. This could be achieved by implementing explicit HPV educational modules in their undergraduate programs. Our study is not free of limitations. For example, we only sampled from one university in the central province of Saudi Arabia. Therefore, our findings cannot be generalized to the entire Saudi university population, and larger nationwide studies may be needed. Secondly, only a small proportion of students reported having a history of STIs (10), which could reflect underreporting for this estimate due to social desirability. Because participation in this study was voluntary, students who showed interest in the study may be more health aware in general. This could have potentially introduced selection bias that may have overestimated our outcomes. Furthermore, our model only controlled for the measured variables, so we cannot exclude confounding by unmeasured variables that are important in assessing the association between gender and HPV awareness and knowledge. Thus, the reader should interpret the results with these points of caution. Despite these limitations, our study sheds light on the understudied level of awareness and knowledge about HPV among health college students in the heavily conservative Saudi society. Health college students play a crucial role in spreading awareness about HPV infection and other STIs in their communities. Improving their knowledge about such diseases and empowering their community health-educational skills is fundamental to STI control and prevention. 5. Conclusions The current study suggests a 40% gap in awareness of HPV among university students and more so among males. Additionally, students in health disciplines exhibited lower levels of awareness and knowledge compared to medical students. The data presented in this study provide a point of reference around the level of HPV knowledge among KSU health college students, which can be used to frame an effective awareness program, especially among male students. We recommend further incorporation of HPV infection, its complication, and vaccine benefits into the curricula of all health college university students. Applying university-wide interventions and educational campaigns can boost HPV knowledge and awareness, which can ultimately promote community adoption of HPV vaccination. Acknowledgments The authors extend their appreciation to the Deputyship for Research & Innovation, Ministry of Education in Saudi Arabia for funding this research work through the project no. (IFKSURG-). Author Contributions Conceptualization, E.A. and D.D.; methodology, E.A. and H.H.A.; validation, E.A. and W.F.A.; formal analysis, R.D., E.A. and D.D.; investigation, E.A., H.H.A., W.F.A. and S.A.; resources, S.A.; data curation, H.H.A., W.F.A. and E.A.; writing--original draft preparation, L.A., E.A., L.F. and D.D.; writing--review and editing, T.A.H., S.A., E.A. and R.D.; visualization, D.D. and L.A.; supervision, E.A.; project administration, E.A.; funding acquisition, E.A. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of King Saud University (ref. No. 22/0843/IRB). Informed Consent Statement Participation in the online questionnaire-based survey implied consent for the study. Data Availability Statement The data that support the findings of this study are available upon request from the corresponding author E.A. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Awareness and knowledge of HPV among male and female students. Chi-square descriptive comparative test was applied. (A) Participants were asked whether they had previously heard of HPV. The number of participants responding with "Yes" or "No" responses are depicted as percentages (%). Males' awareness of HPV was less than females' (p-value = 0.003). (B) Participants answering yes to the question in (A) (N = 242) were asked a series of factual questions on HPV infection to assess the level of their knowledge about HPV. Knowledge scores were grouped into the following: good (>75%), fair (50-75%), and poor (<50%). The knowledge categories were not significantly different between male and female students (p-value = 0.502). healthcare-11-00649-t001_Table 1 Table 1 Sociodemographic characteristics of the respondents. Item Total (N = 403) Male (N = 191) Female (N = 212) N % N % N % Age 18-20 years 174 41.4 79 44.8 95 43.2 21-23 years 204 54.5 104 47.2 100 50.6 24-26 years 22 3.7 7 7.1 15 5.5 27 or more 3 0.5 1 0.9 2 0.7 In which college are you studying? Medicine 85 21.1 41 21.5 44 20.8 Pharmacy 57 14.1 50 26.2 7 3.3 College of Applied Medical Sciences 170 42.2 83 43.5 87 41.0 Nursing 54 13.4 6 3.1 48 22.6 Dentistry 37 9.2 11 5.8 26 12.3 Nationality Saudi 397 98.5 188 98.4 209 98.6 Non-Saudi 6 1.5 3 1.6 3 1.4 Marital Status Single 395 98.0 188 98.4 207 97.6 Married 6 1.5 3 1.6 3 1.4 Divorced 2 0.9 2 .9 Cumulative grade point average (GPA) 4 or more 290 64.4 118 61.8 172 81.1 Less than 4 114 25.1 73 38.2 40 18.9 Do you smoke? Yes 55 13.6 42 22.0 13 6.1 No 348 86.4 149 78.0 199 93.9 Did you take Hepatitis B Vaccine? Yes 205 50.9 84 44.0 121 57.1 No 198 49.1 107 56.0 91 42.9 Do you have a history of any sexually transmitted infections? Yes 10 2.5 4 2.1 6 2.8 No 393 97.5 187 97.9 206 97.2 Have you heard of the human papillomavirus (HPV)? Yes 242 60.0 109 57.1 133 62.7 No 161 40.0 82 42.9 79 37.3 Note: the number of responses is denoted as N with the corresponding percentages (%). healthcare-11-00649-t002_Table 2 Table 2 Distribution of the respondents regarding knowledge. Item Male (N = 109) Female (N = 133) p-Value Incorrect/ Did Not Know Correct Incorrect/ Did Not Know Correct N % N % N % N % Human papillomavirus (HPV) is very rare 46 42.2 63 57.8 43 32.3 90 67.7 0.140 HPV always has visible signs or symptoms 57 52.3 52 47.7 63 47.4 70 52.6 0.518 HPV can cause cervical cancer in females 26 23.6 83 76.1 23 17.3 110 82.7 0.260 HPV can cause oropharyngeal cancers in males 62 56.9 47 43.1 88 66.2 45 33.8 0.146 HPV can be passed on by genital skin-to-skin contact 46 42.2 63 57.8 49 36.8 84 63.2 0.429 There are many types of HPV 73 67.0 36 33.0 89 66.9 44 30.1 0.435 HPV can cause HIV/AIDS 67 61.5 42 38.5 89 66.9 44 33.1 0.419 HPV can be passed on during sexual intercourse 25 22.9 84 77.1 22 16.5 111 83.5 0.253 HPV can cause genital warts 36 33.0 73 67.0 42 31.6 91 68.4 0.890 Men cannot get HPV 28 25.7 81 74.3 38 28.6 95 71.4 0.665 Using condoms reduces the risk of getting HPV 38 34.9 71 65.1 41 30.8 92 69.2 0.582 HPV can be cured with antibiotics 55 50.5 54 49.5 60 45.1 73 54.9 0.439 Having many sexual partners increases the risk of getting HPV 25 22.9 84 77.1 19 14.3 114 85.7 0.095 HPV usually doesn't need any treatment 96 88.1 13 11.9 115 86.5 18 13.5 0.847 Most sexually active people will get HPV at some point in their lives 77 70.6 32 29.4 94 70.7 39 29.3 0.553 A person could have HPV for many years without knowing it 42 38.5 67 61.5 41 30.8 92 69.2 0.223 Having sexual intercourse at an early age increases the risk of getting HPV 52 47.7 57 52.3 69 51.9 64 48.1 0.605 Note: the number of responses is denoted as N with the corresponding percentages (%). healthcare-11-00649-t003_Table 3 Table 3 Relationship between the sociodemographic characteristics and the HPV knowledge levels. Sociodemographic Characteristics Knowledge about HPV Poor (<50%) Fair (50-75%) Good (>75%) p-Value N % N % N % Age 18-20 years 31 53.4 20 20.0 8 9.5 <0.001 21-23 years 23 39.7 73 73.0 67 79.8 24-26 years 4 6.9 6 6.0 9 10.7 27 or more 0 0.0 1 1.0 0 0.0 In which college are you studying? Medicine 8 13.8 29 29.0 41 48.8 <0.001 Pharmacy 14 24.1 16 16.0 14 16.7 College of Applied Medical Sciences 22 37.9 22 22.0 16 19.0 Nursing 11 19.0 13 13.0 3 3.6 Dentistry 3 5.2 20 20.0 10 11.9 Gender Male 30 51.7 43 43.0 36 42.9 0.502 Female 28 48.3 57 57.0 48 57.1 Nationality Saudi 47 98.3 98 98.0 83 98.8 0.911 Non-Saudi 1 1.7 2 2.0 1 1.2 Marital Status Single 57 98.3 99 99.0 83 98.8 Married 1 1.7 1 1.0 1 1.2 0.923 Divorced 0 0.0 0 0.0 0 0.0 Cumulative grade point average (GPA) 4 or more 143 60.3 121 98.0 26 86.7 <0.001 Less than 4 94 39.7 15 11.0 4 13.3 Do you smoke? Yes 6 10.3 12 12.0 16 19.0 0.254 No 52 89.7 88 88.0 68 81.0 Did you take hepatitis B Vaccine? Yes 28 48.3 70 70.0 63 75.0 <0.001 No 30 51.7 30 30.0 21 25.0 Do you have a history of any sexually transmitted infections? Yes 0 0.0 2 2.0 2 2.4 0.516 No 58 100.0 98 98.0 82 97.6 Note: the number of responses denoted as N, with the corresponding percentages (%). healthcare-11-00649-t004_Table 4 Table 4 Prediction of HPV awareness by sociodemographic characteristics. Sociodemographic Characteristics Awareness of HPV AOR 95% CI Age 18-20 years Ref 21-23 years 2.99 1.71, 5.22 24-26 years 4.13 1.03, 16.54 27 or more 0.47 0.03, 7.31 In which college are you studying? Medicine Ref Pharmacy 0.42 0.14, 1.24 College of Applied Medical Sciences 0.11 0.04, 0.29 Nursing 0.10 0.04, 0.29 Dentistry 0.48 0.12, 1.91 Gender Male 0.59 0.32, 1.08 Female Ref Cumulative grade point average (GPA) 4 or more 1.54 0.86, 2.77 Less than 4 Ref Do you smoke? Yes 1.19 0.55, 2.58 No Ref Did you take hepatitis B Vaccine? Yes 2.10 1.21, 3.64 No Ref Do you have a history of any sexually transmitted infections? Yes 0.41 0.08, 2.01 No Ref Notes: AOR = adjusted odds ratio. CI = confidence interval. healthcare-11-00649-t005_Table 5 Table 5 Prediction of HPV knowledge score by sociodemographic characteristics. Sociodemographic Characteristics HPV Knowledge Score Beta Estimate 95% CI Age 18-20 years Ref 21-23 years 1.59 0.44, 2.74 24-26 years -0.13 -2.02, 1.76 27 or more -1.02 -7.94, 5.90 In which college are you studying? Medicine Ref Pharmacy -2.18 -3.52, -0.85 College of Applied Medical Sciences -2.22 -3.62, -0.82 Nursing -2.84 -4.43, -1.25 Dentistry -1.37 -2.79, 0.04 Gender Male -0.91 -1.96, 0.14 Female Ref Cumulative grade point average (GPA) 4 or more 0.34 -0.97, 1.65 Less than 4 Ref Do you smoke? Yes 1.72 0.32, 3.13 No Ref Did you take hepatitis B Vaccine? Yes 0.51 -0.51, 1.54 No Ref Do you have a history of any sexually transmitted infections? Yes 1.66 -1.84, 5.16 No Ref Notes: CI = confidence interval. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Burd E.M. Human papillomavirus and cervical cancer Clin. Microbiol. Rev. 2003 16 1 17 10.1128/CMR.16.1.1-17.2003 12525422 2. Kovachev S.M. Slavov V.D. Correlation between abnormal cytological findings and human papillomavirus infection of the uterine cervix in Bulgarian women Biotechnol. Biotechnol. Equip. 2016 30 1168 1172 10.1080/13102818.2016.1222885 3. Rositch A.F. Silver M.I. Burke A. Viscidi R. Chang K. Duke C.M. Shen W. Gravitt P.E. The correlation between human papillomavirus positivity and abnormal cervical cytology result differs by age among perimenopausal women J. Low. Genit. Tract Dis. 2013 17 38 47 10.1097/LGT.0b013e3182503402 22885643 4. Alhamlan F.S. Khayat H. Ramisetty-Mikler S. Al-Muammar T. Tulbah A. Al-Badawi I. Kurdi W. Tulbah M. Alkhenizan A. Hussain A. Sociodemographic characteristics and sexual behavior as risk factors for human papillomavirus infection in Saudi Arabia Int. J. Infect. Dis. 2016 46 94 99 10.1016/j.ijid.2016.04.004 27062984 5. Ferlay J. Colombet M. Soerjomataram I. Parkin D.M. Pineros M. Znaor A. Bray F. Cancer statistics for the year 2020: An overview Int. J. Cancer 2021 149 778 789 10.1002/ijc.33588 6. Kombe Kombe A.J. Li B. Zahid A. Mengist H.M. Bounda G.-A. Zhou Y. Jin T. Epidemiology and Burden of Human Papillomavirus and Related Diseases, Molecular Pathogenesis, and Vaccine Evaluation Front. Public Health 2020 8 552028 10.3389/fpubh.2020.552028 33553082 7. Gerend M.A. Lee S.C. Shepherd J.E. Predictors of human papillomavirus vaccination acceptability among underserved women Sex. Transm. Dis. 2007 34 468 471 10.1097/01.olq.0000245915.38315.bd 17139233 8. Naing C. Pereira J. Abe T. Wei D.E.Z. Bajera I.B.A.R. Perera U.H.K. Predictors associated with the willingness to take human papilloma virus vaccination J. Community Health 2012 37 288 293 10.1007/s10900-011-9473-4 21928103 9. Loke A.Y. Kwan M.L. Wong Y.-T. Wong A.K.Y. The Uptake of Human Papillomavirus Vaccination and Its Associated Factors Among Adolescents: A Systematic Review J. Prim. Care Community Health 2017 8 349 362 10.1177/2150131917742299 29161946 10. Sherman S.M. Bartholomew K. Denison H.J. Patel H. Moss E.L. Douwes J. Bromhead C. Knowledge, attitudes and awareness of the human papillomavirus among health professionals in New Zealand PLoS ONE 2018 13 e0197648 10.1371/journal.pone.0197648 30596646 11. Trucchi C. Restivo V. Amicizia D. Fortunato F. Manca A. Martinelli D. Montecucco A. Piazza M.F. Prato R. Tisa V. Italian Health Care Workers' Knowledge, Attitudes, and Practices Regarding Human Papillomavirus Infection and Prevention Int. J. Environ. Res. Public Health 2020 17 5278 10.3390/ijerph17155278 32707867 12. McBride K.R. Singh S. Predictors of Adults' Knowledge and Awareness of HPV, HPV-Associated Cancers, and the HPV Vaccine: Implications for Health Education Health Educ. Behav. 2018 45 68 76 10.1177/1090198117709318 28595454 13. Alhamlan F. Obeid D. Khayat H. Asma T. Al-Badawi I.A. Almutairi A. Almatrrouk S. Fageeh M. Bakhrbh M. Nassar M. Prognostic impact of human papillomavirus infection on cervical dysplasia, cancer, and patient survival in Saudi Arabia: A 10-year retrospective analysis Ann. Saudi Med. 2021 41 350 360 10.5144/0256-4947.2021.350 34873934 14. Almehmadi M.M. Salih M.M. Al-Hazmi A.S. Awareness of human papillomavirus infection complications, cervical cancer, and vaccine among the Saudi population. A cross-sectional survey Saudi Med. J. 2019 40 555 559 10.15537/smj.2019.6.24208 31219489 15. Farsi N.J. Baharoon A.H. Jiffri A.E. Marzouki H.Z. Merdad M.A. Merdad L.A. Human papillomavirus knowledge and vaccine acceptability among male medical students in Saudi Arabia Hum. Vaccines Immunother. 2021 17 1968 1974 10.1080/21645515.2020.1856597 16. Mousa M. Al-Amri S.S. Degnah A.A. Tolah A.M. Abduljabbar H.H. Oraif A.M. Abduljabbar H.S. Mirza A.A. Azhar E. Hashem A.M. Prevalence of human papillomavirus in Jeddah, Saudi Arabia Ann. Saudi Med. 2019 39 403 409 10.5144/0256-4947.2019.403 31804132 17. Rodriguez-Alvarez M.I. Gomez-Urquiza J.L. Ahmed H.H.-E. Albendin-Garcia L. Gomez-Salgado J. la Fuente G.A.C.-D. Prevalence and Risk Factors of Human Papillomavirus in Male Patients: A Systematic Review and Meta-Analysis Int. J. Environ. Res. Public Health 2018 15 2210 10.3390/ijerph15102210 30309014 18. Akkour K. Alghuson L. Benabdelkamel H. Alhalal H. Alayed N. AlQarni A. Arafah M. Cervical Cancer and Human Papillomavirus Awareness among Women in Saudi Arabia Medicina 2021 57 1373 10.3390/medicina57121373 34946318 19. Altamimi T. Human papillomavirus and its vaccination: Knowledge and attitudes among female university students in Saudi Arabia J. Fam. Med. Prim. Care 2020 9 1849 1855 10.4103/jfmpc.jfmpc_1205_19 20. Waller J. Ostini R. Marlow L.A. McCaffery K. Zimet G. Validation of a measure of knowledge about human papillomavirus (HPV) using item response theory and classical test theory Prev. Med. 2013 56 35 40 10.1016/j.ypmed.2012.10.028 23142106 21. Hafiz T.A. D'Sa J.L. Zamzam S. Dionaldo M.L.V. Mubaraki M.A. Tumala R.B. Helicobacter pylori Infection: Comparison of Knowledge between Health Science and Non-Health Science University Students Int. J. Environ. Res. Public Health 2021 18 8173 10.3390/ijerph18158173 34360466 22. Newman P.A. Logie C.H. Doukas N. Asakura K. HPV vaccine acceptability among men: A systematic review and meta-analysis Sex. Transm. Infect. 2013 89 568 574 10.1136/sextrans-2012-050980 23828943 23. Faneye A.O. Adeiga A.A. Awoderu O.B. Fayemiwo A.S. Human Papilloma Virus Vaccine Awareness and Vaccination History in Patients Attending STI Clinics in Lagos and Ibadan, Nigeria Arch. Basic Appl. Med. 2018 6 95 98 30294662 24. Chen Z. Schiffman M. Herrero R. DeSalle R. Anastos K. Segondy M. Sahasrabuddhe V.V. Gravitt P.E. Hsing A.W. Chan P. Classification and evolution of human papillomavirus genome variants: Alpha-5 (HPV26, 51, 69, 82), Alpha-6 (HPV30, 53, 56, 66), Alpha-11 (HPV34, 73), Alpha-13 (HPV54) and Alpha-3 (HPV61) Virology 2018 516 86 101 10.1016/j.virol.2018.01.002 29331867 25. Alshammari F. Khan K.U. Knowledge, attitudes and perceptions regarding human papillomavirus among university students in Hail, Saudi Arabia PeerJ 2022 10 e13140 10.7717/peerj.13140 35345591 26. Aga S.S. Yasmeen N. Khan M.A. Hakami A.Y. Awadh A. Malli I.A. Knowledge, attitudes, and perception toward human papillomavirus among health profession students: A cross-sectional study in King Saud Bin Abdulaziz University for Health Sciences J. Educ. Health Promot. 2022 11 141 10.4103/jehp.jehp_640_21 35677262 27. El Mansouri N. Ferrera L. Kharbach A. Achbani A. Kassidi F. Rogua H. Wahmane S.A. Belmouden A. Chouham S. Nejmeddine M. Awareness and knowledge associated to Human papillomavirus infection among university students in Morocco: A cross-sectional study PLoS ONE 2022 17 e0271222 10.1371/journal.pone.0271222 35802731 28. Kasymova S. Harrison S.E. Pascal C. Knowledge and Awareness of Human Papillomavirus Among College Students in South Carolina Infect. Dis. Res. Treat. 2019 12 1178633718825077 10.1177/1178633718825077 29. Bencherit D. Kidar R. Otmani S. Sallam M. Samara K. Barqawi H.J. Lounis M. Knowledge and Awareness of Algerian Students about Cervical Cancer, HPV and HPV Vaccines: A Cross-Sectional Study Vaccines 2022 10 1420 10.3390/vaccines10091420 36146498 30. Rashid S. Labani S. Das B.C. Knowledge, Awareness and Attitude on HPV, HPV Vaccine and Cervical Cancer among the College Students in India PLoS ONE 2016 11 e0166713 10.1371/journal.pone.0166713 27861611 31. Yin G. Zhang Y. Chen C. Ren H. Guo B. Zhang M. Have you ever heard of Human Papillomavirus (HPV) vaccine? The awareness of HPV vaccine for college students in China based on meta-analysis Hum. Vaccines Immunother. 2021 17 2736 2747 10.1080/21645515.2021.1899731 33787459 32. Preston S.M. Darrow W.W. Are Men Being Left Behind (Or Catching Up)? Differences in HPV Awareness, Knowledge, and Attitudes Between Diverse College Men and Women Am. J. Men's Health 2019 13 1557988319883776 10.1177/1557988319883776 31787066 33. Al-Shaikh G.K. Almussaed E.M. Fayed A.A. Khan F.H. Syed S.B. Al-Tamimi T.N. Elmorshedy H.N. Knowledge of Saudi female university students regarding cervical cancer and acceptance of the human papilloma virus vaccine Saudi Med. J. 2014 35 1223 1230 25316467 34. Darraj A.I. Arishy A.M. Alshamakhi A.H. Osaysi N.A. Jaafari S.M. Sumayli S.A. Mushari R.Y. Alhazmi A.H. Human Papillomavirus Knowledge and Vaccine Acceptability in Jazan Province, Saudi Arabia Vaccines 2022 10 1337 10.3390/vaccines10081337 36016225 35. Cinar I.O. Ozkan S. Aslan G.K. Alatas E. Knowledge and Behavior of University Students toward Human Papillomavirus and Vaccination Asia-Pac. J. Oncol. Nurs. 2019 6 300 307 10.4103/apjon.apjon_10_19 31259227 36. Akalin A. Knowledge and attitude towards human papillomavirus and its vaccination and affecting factors among nursing and medical students: A questionnaire study J. Obstet. Gynaecol. 2022 42 3315 3321 10.1080/01443615.2022.2124851 36129449 37. Osazuwa-Peters N. Boakye E.A. Chen B.Y. Clancy J. Vallot P.L. Su J.L. Beck G.E. Varvares M.A. Sociodemographic Factors Associated with Knowledge and Risk Perception of Human Papillomavirus and Human Papillomavirus-Associated Oropharyngeal Squamous Cell Carcinoma Among a Predominantly Black Population JAMA Otolaryngol. Head Neck Surg. 2017 143 117 124 10.1001/jamaoto.2016.2784 27711922 |
PMC10000571 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050678 healthcare-11-00678 Article Utilization of Maternal Healthcare Services among Adolescent Mothers in Indonesia Gayatri Ratih Virta Conceptualization Methodology Formal analysis Investigation Data curation Writing - original draft 12 Hsu Yu-Yun 3* Damato Elizabeth G. Validation Writing - review & editing 4 Rizzo Giuseppe Academic Editor 1 International Doctoral Program in Nursing, Department of Nursing, College of Medicine, National Cheng Kung University, Tainan 70101, Taiwan 2 National Polytechnic of Health Bandung Ministry of Health, Republic of Indonesia, Bandung 40171, Indonesia 3 Department of Nursing, College of Medicine, National Cheng Kung University, Tainan 70101, Taiwan 4 School of Nursing, Case Western Reserve University, Cleveland, OH 44106, USA * Correspondence: [email protected]; Tel.: +886-6-2353535 (ext. 5036) 25 2 2023 3 2023 11 5 67814 12 2022 19 2 2023 23 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Providing maternal healthcare services is one of the strategies to decrease maternal mortality. Despite the availability of healthcare services, research investigating the utilization of healthcare services for adolescent mothers in Indonesia is still limited. This study aimed to examine the utilization of maternal healthcare services and its determinants among adolescent mothers in Indonesia. Secondary data analysis was performed using the Indonesia Demographic and Health Survey 2017. Four hundred and sixteen adolescent mothers aged 15-19 years were included in the data analysis of frequency of antenatal care (ANC) visits and place of delivery (home/traditional birth vs. hospital/birth center) represented the utilization of maternal healthcare services. Approximately 7% of the participants were 16 years of age or younger, and over half lived in rural areas. The majority (93%) were having their first baby, one-fourth of the adolescent mothers had fewer than four ANC visits and 33.5% chose a traditional place for childbirth. Pregnancy fatigue was a significant determinant of both antenatal care and the place of delivery. Older age (OR 2.43; 95% CI 1.12-5.29), low income (OR 2.01; 95% CI 1.00-3.74), pregnancy complications of fever (OR 2.10; 95% CI 1.31-3.36), fetal malposition (OR 2.01; 95% CI1.19-3.38), and fatigue (OR 3.63; 95% CI 1.27-10.38) were significantly related to four or more ANC visits. Maternal education (OR 2.14; 95% CI 1.35-3.38), paternal education (OR 1.62; 95% CI 1.02-2.57), income level (OR 2.06; 95% CI 1.12-3.79), insurance coverage (OR 1.68; 95% CI 1.11-2.53), and presence of pregnancy complications such as fever (OR 2.03; 95% CI 1.33-3.10), convulsion (OR 7.74; 95% CI 1.81-32.98), swollen limbs (OR 11.37; 95% CI 1.51-85.45), and fatigue (OR 3.65; 95% CI 1.50-8.85) were significantly related to the place of delivery. Utilization of maternal healthcare services among adolescent mothers was determined by not only socioeconomic factors but also pregnancy complications. These factors should be considered to improve the accessibility, availability, and affordability of healthcare utilization among pregnant adolescents. adolescent mothers maternal healthcare pregnancy complications Indonesia This research received no external funding. pmc1. Introduction Adolescent pregnancy is an important issue around the world and around 12 million of them aged 15-19 years old experience childbirth every year especially in developing countries such as Indonesia. Based on the Indonesia Demographic and Health Survey 2017, the live birth rate for adolescent pregnancy is 36 per 1000 women . Adolescent mothers are a vulnerable population because they tend to incur high-risk pregnancies that contribute to the higher rates of maternal mortality seen in developing countries. The maternal mortality rate in Indonesia is 305/100,000 live births . Although the maternal mortality rate during the 2010-2017 period in Indonesia reflects a decline, this figure is still far from the 2030 Sustainable Development Goals target of 70/100,000 live births . Adolescent females may be reproductively disadvantaged because their physiology is not fully developed, and their reproductive organs may not be able to fully carry out their function to sustain a prolonged labor . Complications of childbirth such as anemia, hypertension, preeclampsia and eclampsia, spontaneous abortion, assisted delivery, stillbirth and gestational diabetes can occur in this vulnerable age range . Furthermore, early childbearing in adolescent women is also associated with low birth-weight infants, preterm delivery, and severe neonatal complications . In middle-income countries, complications from pregnancy and childbirth are a leading cause of death among girls aged 15-19 years old . One strategy to decrease maternal mortality is utilization of antenatal care services that monitor maternal health status during pregnancy . Over half a million maternal deaths during pregnancy are due to the unavailability and poor utilization of maternal health care services around the world . The use of antenatal care services provides an opportunity to promote the safety and well-being of mothers and their babies . Accessibility, availability, and affordability of antenatal care services are still the determining factors for improving pregnancy outcomes . Both the use of antenatal care services and place of delivery are two important indicators of maternal health care services. Economic and demographic factors underlie women's considerations for accessing maternal healthcare services in both urban and rural areas . Geographic distance limits the use of healthcare services, and transportation expenses to access maternal healthcare services are a burden to pregnant adolescents . Giving birth at home with the help of traditional birth attendants is common in developing countries . A qualitative study conducted in Indonesia found that pregnant women in rural areas often prefer to use traditional birth attendants and give birth at home . This suggests that financial difficulties and geographical distance may be obstacles preventing pregnant adolescents from using modern healthcare services (e.g., maternity hospitals and/or birthing centers) for antenatal and intrapartum care. Reduction of maternal mortality rate is a main component of the 2030 Sustainable Development Goals (SDGs). As mention before, antenatal care is related to maternal mortality rate. Although several studies have explored the use of antenatal care services by adolescents in Indonesia, past studies only examined antenatal care and did not explore the place of delivery . Evidence of specific issues facing adolescents regarding antenatal care utilization, health history, pregnancy complications, and place of delivery is scarce, particularly in developing countries such as Indonesia. Therefore, this study aimed to investigate the utilization of maternal healthcare services and its determinants among adolescent mothers in Indonesia in which the utilization of maternal healthcare services refers to the frequency of antenatal care and place of delivery. The findings of this study are anticipated to support strategic policies and initiatives that prioritize the use of maternal healthcare services among adolescent mothers. 2. Materials and Methods 2.1. Study Design This study utilized secondary data analysis of the Indonesian Demographic Data Survey 2017 (2017 IDHS). IDHS is part of the International Demographic and Health Survey (DHS) program conducted by the Inner-City Fund (ICF). IDHS is a nationally representative cross-sectional survey using a multistage sampling design. The 2017 IDHS was implemented under the supervision of the Health Research and Development Agency (Balitbangkes), which is one of the main units of the Indonesian Ministry of Health. The Balitbangkes performed an independent ethics review of the 2017 IDHS protocol by obtaining informed consent from all participants. Data collection interviews were conducted after informed consent was obtained. The 2017 IDHS sample included 1970 census blocks covering urban and rural areas. The sample was nationally representative and covered the entire population residing in non-institutional dwellings in Indonesia. The census blocks obtained a household sample of 49,627 respondents who were married and aged 15-49 years (98% response rate). Permission to use the data in this study was approved by the international ICF, which is a part of the DHS program (DHS 2015). 2.2. Participants Eligibility criteria for the secondary data analysis were women aged 15-19 years giving birth to their first child from January 2012 until the IDHS survey was conducted in 2017. Of the 49,627 respondents, 416 adolescent mothers met the study criteria 2.3. Variables Demographic variables included maternal age, residence (rural or urban), maternal education level (primary, secondary, and above), father's education level (primary, secondary, and above), employment status (not working or working), knowledge of pregnancy danger signs (know, do not know), baby's birth order (first, second, and more), and insurance coverage (no, yes). The economic status quintile was regrouped into three levels: low income (<USD 200), middle income (USD 200), and high income (>USD 200). Medical history of pregnancy complications (no, yes) assessed included fever, convulsions, fetal malposition, swollen limbs, and fatigue. Two dependent variables were used as the measure of maternal healthcare services: number of antenatal care (ANC) and place of delivery. At least four ANC visits during pregnancy were expected because the health institute in Indonesia provided four free ANC visits during the IDHS 2017 survey period. For this study, the number of ANC visits was classified into two categories: 'less than four visits' and 'four visits or more'. Place of delivery was categorized as either delivery at home with a traditional birth attendant or delivery at a hospital, birthing center, public health center, or other health clinic 2.4. Statistical Analysis This study used SPSS version 24.0 (IBM, Inc., Armonk, NY, USA) to analyze the data. Simple logistic regressions were performed to examine the relationship of each socio-demographic characteristic and complications of pregnancy with antenatal care visits as well as place of delivery. Multiple logistic regressions were then performed to build the models containing explanatory factors predicting the utilization of maternal healthcare services. 3. Results 3.1. Characteristics of Subjects Of the 49,627 women in the 2017 IDHS data set, 416 met the inclusion criteria and their data were analyzed for the study (place of delivery data was missing for 4 women). Mean maternal age was 18.1 years (SD = 0.98; range 15 to 19). Approximately 7% of the participants were 16 years of age or younger, over half lived in rural areas, three-fourths were unemployed, two-thirds were low income and 43% lacked insurance coverage. The majority (93%) were having their first baby, and nearly half had no awareness of the danger signs of pregnancy complications. Around one-fourth of the adolescent mothers had fewer than four ANC visits and one-third of the adolescent mothers chose a traditional place for childbirth. Detailed descriptive statistics are provided in Table 1. 3.2. Bivariate Analysis Associated with ANC Visits Table 2 details variables associated with frequency of antenatal care visits. Older age (OR 2.43; 95% CI 1.12-5.29) and low income (OR 2.01; 95% CI 1.00-4.04) were significantly associated with increased use of antenatal care visits. Adolescent mothers older than 16 years were over twice as likely to report 4 or more ANC visits versus adolescents younger than 16 years. Adolescent mothers with low income were approximately twice as likely to report 4 or more ANC visits than adolescent mothers with middle-income levels. Pregnancy complications significantly related to frequency of antenatal care visits were fever (OR 2.10; 95% CI 1.31-3.36), fetal malposition (OR 2.01; 95% CI 1.19-3.38), and fatigue (OR 3.62; 95% CI 1.26-10.38). Adolescent mothers with fever or fetal malposition were approximately twice as likely to have 4 or more ANC visits than adolescent mothers without fever or fetal malposition. Adolescent mothers reporting pregnancy fatigue were 3.5 times more likely to have 4 or more ANC visits than adolescent mothers without pregnancy fatigue. Variables not associated with frequency of antenatal care visits were place of residence, maternal education level, paternal education, insurance coverage, parity, knowledge of pregnancy danger signs, and convulsions or swollen limbs during pregnancy. 3.3. Bivariate Analysis Associated with Place of Delivery Five variables were significantly associated with delivery location including place of residence (OR 0.22; 95% CI 0.13-0.36), maternal education (OR 2.14; 95% CI 1.35-3.38), paternal education (OR 1.62; 95% CI 1.02-2.57), income (OR 2.06; 95% CI 1.12-3.79) and insurance coverage (OR 1.68; 95% CI 1.11-2.54). Adolescent mothers who lived in rural areas were less likely to choose a hospital for delivery compared to those who lived in urban areas. Adolescent mothers with at least a secondary education level were twice as likely to choose a hospital for delivery than those with a primary education level. Similarly, adolescents at a high income level were more likely to choose hospital for delivery than those adolescent mothers at a middle-income level. Adolescent mothers with insurance coverage were also more likely to choose a hospital for delivery than those without insurance coverage. Pregnancy complications significantly related to delivery in a hospital included fever (OR 2.03, 95% CI 1.33-3.10), convulsion (OR 7.74, 95% CI 1.81-32.98), swollen limbs (OR 11.37, 95% CI 1.51-85.45), and fatigue (OR 3.65, 95% CI 1.50-8.85). Maternal age, employment status, parity, knowledge of pregnancy danger signs and fetal malposition were not associated with place of delivery. Detailed statistics are presented in Table 3. 3.4. Hierarchical Logistic Regression Models of Maternal Healthcare Services Utilization Two sequential logistic regressions analyses with three steps were performed to examine the predictors of antenatal care use and choice of delivery location. First, the four demographic predictors (maternal age, place of residence, maternal education, paternal education) were added, followed by the socioeconomic predictors of income level, and insurance coverage. Pregnancy complications of fever, convulsion, fetal malposition, swollen limbs, and fatigue were added in Step 3. In Step 1, for antenatal care use, maternal age (B = 0.89, p < 0.05) was the only significant demographic predictor, accounting for 2% of the variance. In Step 2, after controlling for the demographic variables, the only significant socioeconomic variable related to antenatal care use was low income level (B = 0.36, p < 0.05). In Step 3, after controlling for the demographic and socioeconomic variables, none of the pregnant complications significantly contributed to the variance in antenatal care. Table 4 illustrates the increase in explained variance from 5% at Step 2 to 8% at Step 3. The overall model explained only a small amount of the variance of antenatal care services (R2 = 0.08). Place of residence (B = -1.41, p < 0.001) was a significant predictor of place of delivery in Step 1, accounting for 15% of the variance (Nagelkerke R2). In step 2, high income level (B = 0.80, p < 0.05) was a significant predictor, contributing an additional 3% of the variance. In Step 3, after controlling for the demographic and socioeconomic variables, only the pregnancy complication of fatigue (B = 1.33, p <= 0.01) was significant, accounting for 7% of the variance. Table 5 illustrates the overall model which explained 25% of the variance for place of delivery. 4. Discussion The current study applied secondary data analysis to examine relationships among demographic characteristics, pregnancy complications, and utilization of maternal healthcare services among Indonesian adolescent mothers. Approximately 25% of the adolescent mothers had less than four ANC visits and one-third of the adolescent mothers chose to deliver their infants outside hospital or clinic settings. In addition, the present study reveals that adolescent mothers' age, residence, maternal education, paternal education, income level, insurance coverage, and pregnancy complications such as fever, fetal malposition, and fatigue are related to the utilization of maternal healthcare services. The findings of the current study show that adolescent mothers of a young age (less than 16 years old) are less likely to have 'four or more' ANC visits than adolescent mothers who were older than 16 years old. These findings are in concordance with one Indonesian study that adolescent mothers were less likely to have 4 or more ANC visits than young women aged 20-24 years . One possible explanation for fewer ANC visits in young adolescent mothers is that younger adolescent mothers may have less awareness of the importance of antenatal care. Another possible explanation may be that younger adolescent mothers may be more dependent on an adult or others to access antenatal care . Findings from the current study indicate that one-third of adolescent mother chose a traditional childbirth in the home. Similar to findings from previous studies, our study indicates place of delivery is influenced by where the mother lives. Indonesian adolescent mothers who live in rural areas tend to deliver a baby in the home . In rural areas, culture and traditional belief are deeply rooted from generation to generation, supporting use of traditional birth attendants in the home . In Indonesia, traditional birth attendants are trusted to provide complete services ranging from massaging during pregnancy, providing pregnancy advice, assisting with childbirth, and caring for the placenta as well as performing rituals during pregnancy and 40 days after giving birth. Traditional birth attendees even give spiritual service to laboring mothers by whispering a prayer or chanting incantations . It should also be acknowledged that women who live in urban areas are more likely to have access to modern health facilities such as hospitals for childbirth . We found that adolescent mothers who have higher education are more likely to have four or more antenatal care visits and choose a modern childbirth facility compared to those who have a lower educational background, similar to what has been previously reported . Educational background, is well known to influence health seeking behavior . Higher paternal education level was also associated with the adolescent mother's use of four or more ANC visits and in-hospital delivery. This finding supports other studies that found paternal education influences the awareness of health needs among pregnant women and decision making related to the place of delivery . The patriarchal system is prevalent in Indonesia; men are often expected to take on main family responsibilities and to be the breadwinner ; thus, they are often the dominant decision maker. Future research is needed to specifically explore the role of the husband's decision-making regarding the utilization of maternal healthcare among pregnant adolescents. It is not surprising that a significant relationship exists between economic status and the utilization of maternal healthcare services. This study found that adolescent women with a low income level are more than two times as likely to have four or more ANC visits than adolescent mothers with a medium income level. Those with high income are over two times more likely to use a modern place of delivery than those women with a low income level. Low income status is a deterrent to accessing modern childbirth facilities , whereas high income status is positively associated with delivery in a modern childbirth facility . In Indonesia the average income per capita is IDR 56,000,000/year (USD 1679-3918/year) and around 10.19% of the total population is living under the poverty line . The findings of this study show that those who have insurance as much as 1.68 times as likely to prefer to deliver in a modern healthcare facility than those who do not have insurance. Similar to previous reports , our study found that women with health insurance coverage were more likely to give birth in a modern healthcare facility than pregnant women without insurance coverage. The design of maternal healthcare services for Indonesian adolescents must include attention to low-income adolescent mothers and those living in rural areas . We found that pregnancy complications such as fever, convulsion, swollen limbs, and fatigue significantly affect choice of delivery for adolescent mothers. Previous evidence supports a correlation of ANC use and choice of delivery location with pregnancy complications of adolescent mothers . Fatigue is a pregnancy complication usually associated with sleep disturbances and can have serious consequences for a pregnant woman and her infant . In Indonesia, fatigue is attributed to cultural factors and family customs that bind pregnant women to all the rules that must be obeyed both during pregnancy and after giving birth. This condition can be in the form of limited space for pregnant women to carry out activities outside the home, restriction of foods considered to have an adverse effect on pregnancy, and the strong belief in things beyond reason that can interfere with pregnant women and affect their pregnancy . The results of this study indicate that pregnancy complications as well as demographic and socioeconomic factors determine the utilization of maternal health care services among adolescent mothers in Indonesia. Future studies are needed to examine the impact of maternal healthcare service utilization, both number of ANC visits and choice of delivery location, on maternal and infant outcomes for adolescent mothers. The final multivariate model only accounted for 8% of the variance for antenatal care use and 25% for place of delivery choice. These findings are similar to a previous study from sub-Saharan Africa, which showed that demographic and socioeconomic variables explained 11% of the variance in antenatal care use and 30% of the variance in delivery location . It is possible that other factors may contribute to the utilization of maternal healthcare services, such as awareness of pregnancy complications, health literacy, and social support. Further research is needed to investigate whether these variables are related to the utilization of maternal care services among adolescent mothers in Indonesia. However, in our hierarchy logistics regression, the Hosmer and Lemeshow tests in the last block shows r-value is 0.49 for antenatal care use and 0.69 for place of delivery, respectively. The findings indicate that both models fit the data. 4.1. Strength The findings of this study highlight that young adolescent mothers (less than 16 years) are less likely to utilize more than four antenatal care visits and are less likely to choose to deliver their infant in a hospital or modern childbirth center. Furthermore, paternal education plays a key determinant in the utilization of maternal care services. 4.2. Limitations Some limitations should be considered. First, the Indonesia Demographic Health Survey (IDHS) is carried out every five years and all existing data are obtained from self-reported questionnaires. Recall bias cannot be excluded from the survey. Second, data from the IDHS provide limited information on certain variables, particularly sexual and reproductive information. Finally, the study acknowledges four or more ANC visits as a standard, whereas the World Health Organization (WHO) has updated a new ANC guideline since 2016, in which at least eight ANC visits during pregnancy are suggested. A limitation in generalization of the study findings to those adolescent pregnant females who have at least eight ANC visits cannot be excluded. 5. Conclusions This study has identified that adolescent mothers' age, socioeconomic status, and insurance coverage are the significant determinants of utilization of maternal healthcare services in Indonesia. Young adolescent mothers limit their use of ANC visits and hospitals for delivery. Pregnancy complications such as fever, convulsion, fetal malposition, swollen limbs, and fatigue are key determinants of the utilization of maternal healthcare services among Indonesian adolescent mothers. A tailored intervention addressing economics, health resources, insurance coverage, and education is required for pregnant adolescent mothers in Indonesia. Acknowledgments We would like to acknowledge the Indonesian Demographic Health Survey (IDHS) for providing data collected in 2017. We would also like to acknowledge the Seventh Pan-Pacific Nursing Conference which took place in May 2021 in Hong Kong wherein this work was accepted for oral presentation. Author Contributions R.V.G. contributed to conceptualization, data curation, formal analysis, methodology, writing--original draft, and investigation. Y.-Y.H. contributed informal analysis, supervision, validation, and writing--review and editing. E.G.D. contributed to validation and writing--review and editing. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Permission to use the data in this study was approved by the international ICF, which is a part of the DHS program (DHS 2015), Institutional Review Board Findings Form ICF IRB FWA00000845. Further information about ethical review is available on the website: Accessed on 10 January 2020. Informed Consent Statement Written informed consent for each individual was obtained by DHS. Data Availability Statement This study used data sets available from USAID's Demographic and Health Survey (DHS) program. After registration on the website, data sets can be downloaded and used via the DHS program website: Accessed on 10 January 2020. Conflicts of Interest The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. healthcare-11-00678-t001_Table 1 Table 1 Demographic characteristics of adolescent mothers (n = 416). Characteristic n (%) Age (Mean = 18.1, SD = 0.98) Maternal age (years) <=16 29 (7.0) >=17 387 (93.0) Residence Rural 249 (59.9) Urban 167 (40.1) Maternal education Primary 106 (25.5) Secondary and above 310 (74.5) Paternal education Primary 155 (37.3) Secondary and above 261 (62.7) Income level Low Income (<USD 200) 275 (66.1) Middle (USD 200) 69 (16.6) High income (>USD 200) 72 (17.3) Maternal employment status Not working 327 (78.6) Working 89 (21.4) Parity 1 390 (93.8) >=2 26 (6.2) Insurance coverage No 178 (42.8) Yes 238 (57.2) Knowledge of pregnancy danger signs No 180 (43.3) Yes 236 (56.7) ANC visits Less than four visits 99 (23.8) Four visits or more 317 (76.2) Place of delivery Home 142 (34.1) Hospital or birthing center 274 (65.9) Pregnancy complications Fever Yes 200 (48.1) No 216 (51.9) Convulsions Yes 30 (7.2) No 386 (92.8) Fetal Malposition Yes 143 (34.4) No 273 (65.6) Swollen Limbs Yes 22 (5.3) No 394 (94.7) Fatigue Yes 46 (11.1) No 370 (88.9) healthcare-11-00678-t002_Table 2 Table 2 Independent variables with relation to use of antenatal care. Variables OR (95% CI) p-Value Demographic variables Maternal age (years) <=16 (Ref) 1 >=17 2.43 (1.12-5.29) 0.02 Residence Urban (Ref) 1 Rural 0.72 (0.45-1.16) 0.17 Maternal education Primary (Ref) 1 Secondary and above 1.13 (0.68-1.88) 0.63 Paternal education Primary (Ref) 1 Secondary and above 0.72 (0.41-1.25) 0.24 Socioeconomic variables Income level Low Income (<USD 200) 2.01 (1.00-4.04) 0.04 Middle (Ref) (USD 200) 1 High income (>USD 200) 1.91 (0.97-3.75) 0.06 Maternal employment status Not working (Ref) 1 Working 0.60 (0.36-1.01) 0.06 Insurance coverage No (Ref) 1 Yes 1.22 (0.77-1.91) 0.39 Health History Parity 1 (Ref) 1 >=2 0.68 (0.29-1.63) 0.39 Knowledge of pregnancy danger signs No (Ref) 1 Yes 0.74 (0.34-1.61) 0.73 Pregnancy complications Fever No (Ref) 1 Yes 2.10 (1.31-3.36) <0.01 Convulsions No (Ref) 1 Yes 2.98 (0.88-8.67) 0.07 Fetal Malposition No (Ref) 1 Yes 2.01 (1.19-3.38) <0.01 Swollen Limbs No (Ref) 1 Yes 6.95 (0.92-52.36) 0.06 Fatigue No (Ref) 1 Yes 3.63 (1.27-10.38) 0.01 healthcare-11-00678-t003_Table 3 Table 3 Independent variables with relation to place of delivery. Variables OR (95% CI) p-Value Demographic variables Maternal age (years) <=16 (Ref) 1 >=17 0.88 (0.39-2.00) 0.77 Residence Urban (Ref) 1 Rural 0.22 (0.13-0.36) <0.001 Maternal education Primary (Ref) 1 Secondary and above 2.14 (1.36-3.38) 0.001 Paternal education Primary (Ref) 1 Secondary and above 1.62 (1.02-2.57) 0.04 Socioeconomic variables Income level Low Income ( < USD 200) 1.47 (0.82-2.61) 0.18 Middle (Ref) (USD 200) 1 High income (>USD 200) 2.06 (1.12-3.79) 0.02 Employment status Not working (Ref) 1 Working 0.75 (0.46-1.22) 0.25 Insurance coverage No (Ref) 1 Yes 1.68 (1.11-2.53) 0.01 Health History Parity 1 (Ref) 1 >=2 0.79 (0.35-1.79) 0.58 Knowledge of danger signs No (Ref) 1 Yes 0.82 (0.44-1.54) 0.54 Pregnancy Complications Fever No (Ref) 1 Yes 2.03 (1.33-3.10) 0.001 Convulsions No (Ref) 1 Yes 7.74 (1.81-32.98) <0.01 Fetal Malposition No (Ref) 1 Yes 1.50 (0.96-2.34) 0.07 Swollen Limbs No (Ref) 1 Yes 11.37 (1.51-85.45) 0.02 Fatigue No (Ref) 1 Yes 3.65 (1.50-8.85) <0.01 healthcare-11-00678-t004_Table 4 Table 4 Hierarchical logistic regression models for antenatal care use. Model 1 Model 2 Model 3 B OR CI 95% B OR CI 95% B OR CI 95% Maternal Age (years) (Ref: <= 16) 0.89 2.43 * (1.12-5.29) 0.99 2.70 ** (1.22-5.97) 0.96 2.61 * (1.16-5.87) Low Income (Ref: Middle income) 0.36 2.11 * (1.04-4.26) 0.67 1.96 (0.96-4.01) Fever (Ref: No) 0.28 1.32 (0.61-2.86) Fetal Malposition (Ref: No) 0.27 1.32 (0.58-3.0) Fatigue (Ref: No) 1.01 2.75 (0.89-8.47) Hosmer and Lemeshow test 0.10 0.21 5.43 Sig 0.01 0.89 0.49 Nagelkerke R2 0.02 0.05 0.08 Significance and 95% confidence levels based on odds ratios; * r < 0.05; ** r < 0.01. healthcare-11-00678-t005_Table 5 Table 5 Hierarchical logistic regression models for place of delivery. Model 1 Model 2 Model 3 B OR CI 95% B OR CI 95% B OR CI 95% Residence (Ref: Urban) -1.41 0.24 *** (0.14-0.40) -1.48 0.22 *** (0.13-0.38) -1.54 0.23 *** (0.12-0.37) Maternal education (Ref: Primary) 0.48 1.62 (0.93-2.81) 0.37 1.45 (0.83-2.55) 0.27 1.31 (0.73-2.33) Paternal education (Ref:Primary) 0.24 1.28 (0.75-2.18) 0.18 1.20 (0.69-2.08) 0.15 1.16 (0.66-2.05) High income (Ref: Middle income) 0.80 2.24 * (1.13-4.43) 0.78 2.19 * (1.07-4.46) Insurance coverage (Ref: No) 0.44 1.56 (0.97-2.54) 0.34 1.40 (0.86-2.29) Fever (Ref: No) -0.18 0.82 (0.48-1.41) Convulsions (Ref: No) 1.43 4.18 (0.89-19.5) Swollen Limbs (Ref: No) 1.95 7.06 (0.83-59) Fatigue (Ref: No) 1.33 3.78 ** (1.29-11) Hosmer and Lemeshow test 1.76 16.26 5.55 Sig 0.88 0.03 0.69 Nagelkerke R2 0.15 0.18 0.25 Significance and 95% confidence levels based on odds ratios; * r < 0.05; ** r < 0.01; *** r < 0.001. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Indonesia Demographic and Health Survey 2017 Available online: (accessed on 14 September 2019) 2. World Health Organization Trends in Maternal Mortality: 1990-2015 Estimates from WHO, UNICEF, UNFPA, World Bank Group and the United Nations Population Division World Health Organization Geneva, Switzerland 2015 3. The World Bank Maternal Morality Ratio (Modeled Estimate, Per 100,000 Live Births)-Indonesia 2019 Available online: (accessed on 30 September 2020) 4. Chopra M. Daviaud E. Pattinson R. Fonn S. Lawn J.E. Saving the lives of South Africa's mothers, babies, and children: Can the health system deliver? Lancet 2009 374 835 846 10.1016/S0140-6736(09)61123-5 19709729 5. Azevedo W.F.d. Diniz M.B. Fonseca E.S.V.B.d. Azevedo L.M.R.d. Evangelista CB. Complications in adolescent pregnancy: Systematic review of the literature J. Einstein. 2015 13 618 626 10.1590/S1679-45082015RW3127 26061075 6. Wong S.P.W. Twynstra J. Gilliland J.A. Cook J.L. Seabrook J.A. Risk Factors and Birth Outcomes Associated with Teenage Pregnancy: A Canadian Sample J. Pediatr. Adolesc. Gynecol. 2020 33 153 159 10.1016/j.jpag.2019.10.006 31634579 7. World Health Organization Adolescent Pregnancy World Health Organization Geneva, Switzerland 2014 8. Milligan R. Wingrove B.K. Richards L. Rodan M. Monroe-Lord L. Jackson V. Hatcher B. Harris C. Henderson C. Johnson A.A. Perceptions about prenatal care: Views of urban vulnerable groups BMC Public Health 2002 2 25 10.1186/1471-2458-2-25 12421466 9. Nuamah G.B. Agyei-Baffour P. Mensah K.A. Boateng D. Quansah D.Y. Dobin D. Addai-Donkor K. Access and utilization of maternal healthcare in a rural district in the forest belt of Ghana BMC Pregnancy Childbirth 2019 19 6 10.1186/s12884-018-2159-5 30612557 10. Amongin D. Benova L. Nakimuli A. Nakafeero M. Kaharuza F. Atuyambe L. Hanson C. Trends and determinants of adolescent childbirth in Uganda-analysis of rural and urban women using six demographic and health surveys, 1988-2016 J. Reprod. Health 2020 17 74 10.1186/s12978-020-00925-8 32456705 11. World Health Organization WHO Recommendations on Antenatal Care for a Positive Pregnancy Experience World Health Organization Geneva, Switzerland 2016 12. Tolera H. Gebre-Egziabher T. Kloos H. Using Andersen's behavioral model of health care utilization in a decentralized program to examine the use of antenatal care in rural western Ethiopia PLoS ONE 2020 15 e0228282 10.1371/journal.pone.0228282 31986187 13. Houweling T.A. Ronsmans C. Campbell O.M. Kunst A.E. Huge poor-rich inequalities in maternity care: An international comparative study of maternity and child care in developing countries Bull. World Health Organ. 2007 85 745 754 10.2471/BLT.06.038588 18038055 14. Shahabuddin A. Delvaux T. Abouchadi S. Sarker M. De Brouwere V. Utilization of maternal health services among adolescent women in Bangladesh: A scoping review of the literature J. Trop. Med. Int. Health 2015 20 822 829 10.1111/tmi.12503 15. Gebregziabher N.K. Zeray A.Y. Abtew Y.T. Kinfe T.D. Abrha D.T. Factors determining choice of place of delivery: Analytical cross-sectional study of mothers in Akordet town, Eritrea BMC Public Health 2019 19 924 10.1186/s12889-019-7253-8 31291919 16. Tancred T. Marchant T. Hanson C. Schellenberg J. Manzi F. Birth preparedness and place of birth in Tandahimba district, Tanzania: What women prepare for birth, where they go to deliver, and why BMC Pregnancy Childbirth 2016 16 165 10.1186/s12884-016-0945-5 27422526 17. Kifle M.M. Kesete H.F. Gaim H.T. Angosom G.S. Araya M.B. Health facility or home delivery? Factors influencing the choice of delivery place among mothers living in rural communities of Eritrea J. Health Popul. Nutr. 2018 37 22 10.1186/s41043-018-0153-1 30348219 18. Titaley C.R. Hunter C.L. Dibley M.J. Heywood P. Why do some women still prefer traditional birth attendants and home delivery?: A qualitative study on delivery care services in West Java Province, Indonesia BMC Pregnancy Childbirth 2010 10 43 10.1186/1471-2393-10-43 20701762 19. Efendi F. Chen C.-M. Kurniati A. Berliana S.M. Determinants of utilization of antenatal care services among adolescent girls and young women in Indonesia J. Women Health 2017 57 614 629 10.1080/03630242.2016.1181136 27230132 20. Laksono A.D. Rukmini R. Wulandari R.D. Regional disparities in antenatal care utilization in Indonesia PLoS ONE 2020 15 e0224006 10.1371/journal.pone.0224006 32053621 21. Wulandari R.D. Laksono A.D. Nantabah Z.K. Effect of Marital Status on Completeness of Antenatal Care Visits among Childbearing Age Women in Rural Indonesia J. Med.-Legal Update 2020 20 453 458 22. Tripathi V. Singh R. Regional differences in usage of antenatal care and safe delivery services in Indonesia: Findings from a nationally representative survey BMJ Open 2017 7 e013408 10.1136/bmjopen-2016-013408 28159851 23. Erfina E. Widyawati W. McKenna L. Reisenhofer S. Ismail D. Exploring Indonesian adolescent women's healthcare needs as they transition to motherhood: A qualitative study Women Birth 2019 32 e544 e551 10.1016/j.wombi.2019.02.007 30928175 24. Alex-Ojei C.A. Odimegwu C.O. Akinyemi J.O. Patterns of delivery assistance among adolescent mothers in Nigeria J. Midwifery 2020 82 102619 10.1016/j.midw.2019.102619 25. Agus Y. Horiuchi S. Iida M. Women's choice of maternal healthcare in Parung, West Java, Indonesia: Midwife versus traditional birth attendant J. Women Birth 2018 31 513 519 10.1016/j.wombi.2018.01.007 26. Niehof A. Traditional birth attendants and the problem of maternal mortality in Indonesia J. Pac. Aff. 2014 87 693 713 10.5509/2014874693 27. Bobo F.T. Yesuf E.A. MJIjfeih W. Inequities in utilization of reproductive and maternal health services in Ethiopia Int. J. Equity Health 2017 16 105 10.1186/s12939-017-0602-2 28629358 28. Iacoella F. Tirivayi N. Determinants of maternal healthcare utilization among married adolescents: Evidence from 13 Sub-Saharan African countries J. Public Health 2019 177 1 9 10.1016/j.puhe.2019.07.002 29. Wulandari R.D. Putri N.K. Laksono A.D. Socioeconomic Disparities in Antenatal Care Utilisation in Urban Indonesia J. Innov. Creat. Change 2020 14 498 514 30. Andersen Ronald M. National Health Surveys and the Behavioral Model of Health Services Use Med. Care 2008 46 647 653 10.1097/MLR.0b013e31817a835d 18580382 31. Shahabuddin A.S. Delvaux T. Nostlinger C. Sarker M. Bardaji A. Sharkey A. Adhikari R. Koirala S. Rahman M.A. Mridha T. Maternal health care-seeking behaviour of married adolescent girls: A prospective qualitative study in Banke District, Nepal PLoS ONE 2019 14 e0217968 10.1371/journal.pone.0217968 31237907 32. Melani A.A. Yani A. Damayanti R. The Qualitative Study on Family Decision Making Related to the Perinatal Complication In Cianjur, West Java J. Kedokt. Brawijaya 2013 24 1 6 33. Johar M. Soewondo P. Pujisubekti R. Satrio H.K. Adji A. Inequality in access to health care, health insurance and the role of supply factors J. Soc. Sci. Med. 2018 213 134 145 10.1016/j.socscimed.2018.07.044 30077959 34. Rai R.K. Singh P.K. Singh L. Utilization of maternal health care services among married adolescent women: Insights from the Nigeria Demographic and Health Survey, 2008 J. Women's Health Issues 2012 22 e407 e414 10.1016/j.whi.2012.05.001 35. Mekonnen T. Dune T. Perz J. Ogbo F.A. Trends and determinants of antenatal care service use in Ethiopia between 2000 and 2016 Int. J. Environ. Res. Public Health 2019 16 748 10.3390/ijerph16050748 30832268 36. Malqvist M. Lincetto O. Du N.H. Burgess C. Hoa D.T. Maternal health care utilization in Viet Nam: Increasing ethnic inequity Bull. World Health Organ. 2013 91 254 261 10.2471/BLT.12.112425 23599548 37. Statistic Yearbook Indonesia 2020 Available online: (accessed on 19 May 2020) 38. Aizawa T. The impact of health insurance on out-of-pocket expenditure on delivery in Indonesia J. Health Care Women Int. 2019 40 1374 1395 10.1080/07399332.2019.1578778 30985260 39. Wulandari R.D. Laksono A.D. Nantabah Z.K. Rohmah N. Zuardin Z. Hospital utilization in Indonesia in 2018: Do urban-rural disparities exist? BMC Health Serv. Res. 2022 22 491 10.1186/s12913-022-07896-5 35413914 40. Santos G.H.N.d. Martins M.d.G. Sousa M.d.S. Teenage pregnancy and factors associated with low birth weight J. Rev. Bras. Ginecologia Obs. 2008 30 224 231 41. Lauche R. Hall H. Adams J. Steel A. Broom A. Sibbritt D. Health-care utilisation amongst pregnant women who experience sleeping problems and/or tiredness or fatigue: Secondary analysis of a cross-sectional survey of 1835 pregnant women J. Epidemiol. 2016 20 355 362 10.1007/s11325-015-1250-7 42. Setyowati S. Rosnani R. Women's efforts to solve postpartum problems based on the culture of South Sumatera J. Enferm. Clin. 2019 29 653 658 10.1016/j.enfcli.2019.04.100 |
PMC10000572 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051060 foods-12-01060 Article Incidence of Potentially Toxic Elements and Perfluoroalkyl Substances Present in Canned Anchovies and Their Impact on Food Safety Nobile Maria Conceptualization Methodology Formal analysis Investigation Data curation Writing - original draft Writing - review & editing 1 Mosconi Giacomo Validation Formal analysis 1 Chiesa Luca Maria Conceptualization Investigation Supervision Project administration 1 Panseri Sara Conceptualization Investigation Resources Supervision Project administration 1* Danesi Luigi Formal analysis 1 Falletta Ermelinda Formal analysis Writing - original draft 2 Arioli Francesco Conceptualization Investigation Writing - original draft Writing - review & editing 1 Pichardo Silvia Academic Editor 1 Department of Veterinary Medicine and Animal Science, University of Milan, Via dell'Universita' 6, 26900 Lodi, Italy 2 Department of Chemistry, University of Milan, Via Golgi 19, 20133 Milan, Italy * Correspondence: [email protected]; Tel.: +39-0250334611 02 3 2023 3 2023 12 5 106025 1 2023 21 2 2023 01 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Fish plays a key role in a healthy and balanced Italian diet, but it is also subject to the bioaccumulation of different contaminants depending on the geographical or anthropogenic context from which it is derived. In recent years, the European Food Safety Authority (EFSA) has been focusing its attention on consumer toxicological risk, considering emerging contaminants such as perfluoroalkyl substances (PFASs) and potentially toxic elements (PTEs). Regarding fish, anchovies are among the five small pelagic main commercial species in the European Union and the top five fresh species consumed by households in Italy. Considering the lack of data on PFASs and PTEs in this species, our aim was to investigate the mentioned contaminants in salted and canned anchovies collected over 10 months from different fishing areas, even those far apart, to verify possible variations in bioaccumulation and to consider the risk for the consumer. According to our results, the assessed risk was very reassuring also for large consumers. The only concern, related to Ni acute toxicity, also dependent on the different consumers' sensitivity, was related to only one sample. fish anchovy perfluoroalkyl substances toxic elements risk characterization food safety This research received no external funding. pmc1. Introduction Fish constitutes a fundamental component of the Italian diet, representing a relevant source of protein, polyunsaturated fatty acids and micronutrients; however, humans, through the consumption of fish products, are exposed to various contaminants in relation to the quality of the environment from which they are derived . The pollution of marine waters is mainly due to the development of anthropogenic activities that result in the direct or indirect introduction of substances capable of having harmful effects on living organisms--and, consequently, on human health--into the aquatic environment. In particular, it depends on contaminants transported to the sea by rivers and inland watersheds, along which numerous industrial, agro-livestock activities and/or intense urbanization phenomena are present, while a significant contribution is due to the direct sources, in coastal waters, of urban landfills and industrial discharges . The toxicological risk assessment of humans as consumers of fish is part of the broader topic of food safety, which has long been the goal of the European Food Safety Authority (EFSA). In addition to persistent and emerging contaminants, such as perfluoroalkyl substances (PFASs), among others , potentially toxic elements (PTEs) may also be of concern to the consumer because of their ability to bioaccumulate along the trophic chain. Perfluoroalkyl substances (PFASs), usually, are differentiated as short-chain compounds, but this is reductive as they can be sub-grouped in different ways considering their terminal groups and structures. They have unique chemical-physical properties, such as stability under intense conditions of heat, light and pH and persistence in the environment . They also have surfactant functions, e.g., as water and fat repellents, which is the reason for their wide use in different industry sectors, including food-contact materials, construction and household products, food processing, medical articles, fire-fighting, textiles, electronics, aerospace, automotive, aviation, etc. . Their release and circulation through water and air causes groundwater and drinking water contamination, with subsequent accumulation in animals and humans . Some PFASs are categorized as disruptors of the endocrine hormonal system, toxic for reproduction and the development of a fetus, and are suspected carcinogens. The primary source of PFAS exposure is food, especially fish. PFAS contamination in fish depends on the type of fish, age, geographical area, trophic level, etc. . Here, our attention is focused on four PFASs from 2020, to carry out the assessment of the sum of perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), perfluorohexanesulfonic acid (PFHxS) and perfluorooctanesulfonic acid (PFOS), with respect to a tolerable weekly intake (TWI) of 4.4 ng kg-1 body weight (bw) per week . On the other hand, the presence of PTEs such as Hg, Cd, As, Pb and others in fishery products represents one of the most serious chemical risks to food safety in the seafood supply chain. Metals are important from two points of view: for their toxicity and for their essentiality. Cu, Fe and Zn, for example, are essential metals that can cause toxic effects at high-concentration intakes. Other metals can be classified as potentially toxic, as well as Pb, Cd and Hg, when ingested for long periods, even at low concentrations . The progressive increase in sea pollution, the globalization of raw material supply markets and location of processing plants and the growing consumer awareness of food safety issues make this risk a critical factor for the development and competitiveness of the sector. The origin, characteristics, mechanisms and toxic effects of these elements on humans are well known. It should be remembered that in fish products (muscles of fish and crustaceans), the maximum levels (MLs), i.e., maximum allowable residue concentrations of heavy metals, are set by Reg. 1881/2006. In canned and processed products, however, MLs are not set for PTEs. Regarding fish, anchovies are among the five main small pelagic commercial species in the EU. In 2018, the landings of anchovy in the EU reached a 10-year peak of 135.460 tons, where the trend was led by Spain, followed by Italy, Croatia and Greece. In particular, anchovies are among the top five fresh species (in volume and nominal value) consumed by households in Italy and fall under the quality schemes in the EU seafood sector as Traditional Specialties Guaranteed (TSG), registered up to August 2020 in the EUMOFA report . If we consider that small pelagic scombroids (anchovies, herring, mackerel, sardines, etc.) are also used in aquaculture for the production of fishmeal and fish oil, we could consider this process a type of biomagnification phenomenon occurring in aquatic ecosystems, resulting in the contamination of other fish or other animals, and therefore, at the end, of the consumer . In the literature, there are few studies about the detection of PFASs or PTEs in anchovies and, if present, only few samples are considered among different types of fish (Table 1) and never a comprehensive study on both these classes of contaminants, referring also to a single provenance. Liquid chromatography coupled to high-resolution mass spectrometry (HRMS) is a powerful instrumental technique of choice for PFAS investigations, combining high selectivity in the identification through the exact mass of the target and untargeted molecules and high sensitivity in the order of pg g-1 added to the high scanning speed with different acquisition mode possibilities. On the basis of the considerations mentioned above, the aim of this study was to detect PFASs and PTEs through a survey on salted and canned anchovies of different types, collected in 10 months from different fishing geographical areas, to verify possible variations in bioaccumulation and to assess the risk for the consumer. In addition, we also considered PFASs, which are a very topical subject, considering the entry into force of the new regulations. 2. Materials and Methods 2.1. Chemicals and Reagents Solvents and reagents were obtained from Merck (Darmstadt, Germany). All studied PFASs (perfluorobutanoic acid (PFBA), perfluoropentanoic acid (PFPeA), perfluorohexanoic acid (PFHxA), perfluorobutane sulphonic acid (PFBS), perfluoroheptanoic acid (PFHpA), PFOA, perfluorohexane sulphonate (PFHxS), perfluorononanoic acid (PFNA), perfluorodecanoic acid (PFDA), PFOS, perfluorododecanoic acid (PFDoA), perfluoroundecanoic acid (PFUnDA), perfluorotridecanoic acid (PFTrDA), perfluorotetradecanoic acid (PFTeDA), perfluorohexadecanoic acid (PFHxDA), and perfluorooctadecanoic acid (PFODA)) and the two 13C-labeled internal standards (ISs) perfluoro-(1,2,3,4,5-13C5)nonanoic acid (MPFNA) and perfluoro-(1,2,3,4-13C4)octanesulfonic acid (MPFOS) were purchased from Chemical Research 2000 Srl (Rome, Italy). The purification columns Strata PFAS (WAX/GCB, 200 mg/50 mg/6 mL) were by Phenomenex (Torrance, CA, USA). Hg, Ni, Cd, Cr, As, Pb, Al, Sn and the internal standard Yttrium (Y) of 1000 mg L-1 in concentration were from Fisher Scientific (USA), as was HNO3 (67-69% superpure) and H2O2 (30 wt%). 2.2. Sample Collection A total of 258 sample pools of salted and canned anchovies, each one consisting of at least 10 elemental samples--as described by the COMMISSION REGULATION (EC) No. 333/2007 of 28 March 2007, outlining the methods of sampling and analysis for the official control of the levels of lead, cadmium, mercury, inorganic tin, 3-MCPD and benzo (a) pyrene in foodstuffs comprising PET, aluminum and glass product packages--collected during the period from January 2020 to October 2020 to assess an entire annual production cycle, were distributed according to their different origins: 42 from Tunisia, 71 from the Cantabrian Sea and 143 from the Mediterranean Sea (Croatia). 2.3. Analytical Protocol of PFASs The protocol has been thoroughly described in our previous work . Briefly, 5 g of homogenized sample was spiked with internal standards at 5 ng g-1 in a matrix, followed by the addition of 10 mL of acetonitrile for protein precipitation and extraction. After 1 min of vortexing and 15 min of sonication and centrifugation (2500 g, 4 degC, 10 min), the supernatant was dried, resuspended in 5 mL of water and purified by STRATA PFAS cartridges. In particular, 4 mL of 0.3% ammonium hydroxide (NH4OH) in MeOH, 4 of mL MeOH and 4 mL of ultrapure water were used during preconditioning, followed by the sample loading. Then, 2 washes with 2 x 4 mL of water were carried out, and 2 x 4 mL of 0.3% NH4OH in MeOH was used to elute compounds. After being dried, the extract was resuspended in 200 mL of 20 mM MeOH:ammonium formiate (20:80 v/v) and eventually centrifuged in Eppendorf for 2 min if precipitate was present. The analysis was performed by an UPLC-HRMS system composed of a Vanquish device (Thermo Fisher Scientific, Waltham, MA, USA) coupled to a Thermo OrbitrapTM Exploris 120 (Thermo Fisher Scientific, Waltham, MA, USA), equipped with a heated electrospray ionization (HESI) source. A Raptor ARC-18 5 mm, 150 x 2.1 mm column (Restek, Bellefonte, PA, USA) was used for the separation. Moreover, a small Megabond WR C18 column, 5 cm, 4.6 mm, i.d. 10 mm, was introduced before the injector to delay eventual PFASs present in the system. Mobile phases A (20 mM aqueous ammonium formate) and B (MeOH) were mixed during the gradient, which started with 20% B, increasing to 95% in 7 min and remaining until the 10th min. After 1 min, the initial conditions were reestablished until the 15th. The flow was set at 0.3 mL min-1. With regard to the detector, the capillary and vaporizer temperatures were set at 330 and 280 degC, respectively, the sheath and auxiliary gas at 35 and 15 arbitrary units (AU) and the electrospray voltage at 3.50 kV in negative mode. The full-scan (FS) acquisition was combined with parallel reaction monitoring (PRM) mode for the confirmatory response based on an inclusion list. The FS worked with a resolution of 60,000 FWHM, a scan range of 150-950 m/z, a standard automatic gain control (AGC), an RF lens % of 70 and an automatic maximum injection time. The PRM acquisition operated at 15,000 FWHM, with a standard AGC target, an automatic maximum injection time and scan range mode and an isolation window of 1 m/z. Fragmentation of the precursors was optimized with a two-step normalized collision energy (10 and 70 eV). XcaliburTM 4.5 (Thermo Fisher Scientific, Waltham, MA, USA) was the software used. 2.4. Analytical Protocol of Potentially Toxic and Essential Elements These analyses were conducted by inductively coupled plasma optical emission spectroscopy (ICP-OES, Optima 8000, Perkin Elmer, Waltham, MA, USA) detection, which allows quantification at ppb levels for all compounds examined, consistent with the legal limits set for some metals. The method parameters were related to digestion using a microwave digestion system (Mars One, CEM Corporation, Matthews, NC, USA) equipped with TFM closed vessels, followed by metal analyses. For the digestion procedure, 1.2 g of sample was accurately weighed into a TFM vessel. The sample was properly mineralized with 6 mL HNO3 (67-69% superpure), 2 mL H2O2 (30 wt%) and 2 mL ultrapure H2O. After acid digestion, the sample was cooled and diluted by ultrapure water (Milli-QTM system, Millipore, MA, USA) to the final volume of 30 mL. At the end, the sample was subjected to instrumental analysis by the ICP-OES technique. 2.5. Validation of PFAS Protocol Validation of the method was carried out following the SANTE Guidance 11312/2021 . Specificity/selectivity was evaluated by analyzing 20 blank samples, verifying the absence of interferences by the lack of peaks with a signal-to-noise ratio S/N >3 close to the retention times of selected analytes. The matrix-matched calibration curves were constructed by spiking 5 g of blank anchovy sample with the standards for five calibration points (0.05, 1, 3, 5 and 10 ng g-1) in duplicate. The limit of quantification (LOQ) of the methods was the lowest spiked level meeting the requirements of recovery within the range of 70-120% and an RSD <= 20%, assessed on 6 replicates. Recoveries were calculated at LOQ for all compounds, and thus on 6 replicates, by comparing the peak areas of PFASs spiked before extraction to those spiked after extraction. The matrix effect was also calculated by comparing the peak areas of PFASs spiked after extraction of a blank anchovy sample to the peak areas of a standard solution mix, expressed as a percentage. The intra-day repeatability was evaluated through 6 replicates and expressed as repeatability and within-laboratory reproducibility (%RSD), and the inter-day precision by 6 replicates in 3 different days, using one-way analysis of variance (ANOVA). 2.6. Validation of PTE Protocol For the evaluation of the concentrations of PTEs in the samples, the instrument was calibrated with standard solutions of concentrations between 14 and 200 mg kg-1, which were properly prepared from available stock solutions. Yttrium was used as an internal standard and high-purity argon was used as an inert gas. Method validation was carried out by evaluating each metal's limit of detection (LOD), limit of quantification (LOQ) and precision (% RSD) through six replicates. Determination of metal concentrations was carried out in triplicate per sample. The metals' recovery from the matrix was evaluated by the use of a certified reference material (ERM-CE278k mussel tissue, Joint Research Centre Institute for Reference Materials and Measurements, EU) as reported in Table 2, and the metal concentration was corrected for the results. 2.7. Statistical Analysis Statistical analyses were carried out using Graphpad Instat 3 software (Graphpad Instat Software, San Diego, CA, USA). Due to the extremely varying numbers of fishing areas, a statistical comparison between the various provenances could not be carried out with a reliable result. Therefore, the Kolmogorov-Smirnov test was only performed on the entire number of samples to test whether the distribution of the concentration values of the different elements was normal or not (p < 0.05). None of the compounds and elements analyzed were found to be normally distributed. Instead, the mean, median and maximum concentration values of the different elements were calculated. 3. Results and Discussion 3.1. Validation Performance of PFAS Method The PFAS protocol validation parameters are reported in Table 3, verifying all the requirements set by SANTE 11312/2021 . Briefly, the method had high selectivity with an S/N >3, where analytes were present, starting from the LOQ level, and high specificity, with the absence of interference close to the retention time of the detected PFASs. The recoveries ranged between 70 and 120%, revealing the good efficiency of the analytical protocol. Precision with RSD lower than 20% was in accordance with the tolerance range indicated. The LOQs in the range from 0.050 to 0.10 ng g-1 demonstrated good method sensitivity on this complex matrix, and they perfectly complied with the indicative levels recommended in fish by the new European Recommendations . Matrix calibration curves revealed a good fit over the five calibration points, with R2 > 0.99 for all PFASs. The matrix effect showed a lower influence (<20%) with a percentage variation from 89 to 106%. 3.2. Validation Performance of PTE Method The PTE protocol validation parameters are reported in Table 4, containing the limit of detection (LOD) and the limit of quantification (LOQ), along with the precision, expressed as % RSD. 3.3. Risk Characterization The average Italian consumption of processed anchovies, both in oil and in salt, was first assessed. We estimated the Italian consumers based on the Italian National Institute of Statistics data on the population updated for June 2021, subtracting roughly 10%, a value corresponding to the subgroup of the population represented by children aged between 0 and 10 years. The reason for their exclusion lies in the presumable difficulty of consumption presented by such individuals for a food with a particular taste, but also in the need to decrease the consumer population to make the subsequent data more precautionary. Following this, risk characterization was carried out considering a population of 53,961,715 inhabitants, and the "EUMOFA case study (European market observatory for fisheries and aquaculture products): anchovy transformed in Italy, February 2018" was taken into consideration. In this case study, which reports the data of the Italian market in 2015, the anchovies consumed in Italy were calculated at 43,830 tons of live weight per year, of which 38% are processed, corresponding to 16,484 tons of live weight per year. The case study reports that starting from 2.25 kg live weights of anchovies, 0.57 kg of clean fillet is produced, i.e., only 25.3% of the initial weight is transformed into the final product. If the final yield is 25.3% , the apparent consumption of transformed anchovies is 4170 tons per year, with a per capita consumption figure of 0.08 kg per person per year, corresponding to 0.22 g per person per day. Subsequently, the Estimated Daily Intake (EDI) of the PTEs and PFASs was calculated as follows: EDI = C x DC/BW,(1) where C is the mean concentration (considering that the mean was always equal to or higher than the median and therefore precautionarily considered), DC is the daily fish consumption per capita in Italy and BW is the consumer bodyweight, considered equal to 70 kg. We also considered the 95th percentile estimated seafood consumption, which was calculated as 2.9 times that of the median or mean consumers . We also calculated the Target Hazard Quotient (THQ), which is the ratio between the exposure and the health-based guidance value (HBGV) indicated by EFSA (2020) for each element/compound and each end-point, eventually recalculated on a daily basis: THQ = EDI/HBGV.(2) For a very conservative approach, the highest concentrations were accounted for. The Hazard Index (HI) was therefore calculated as (3) HI=i=8nTHQ Finally, the HIs for the EDI via preserved anchovies of the 95th percentile fish consumers were calculated. The difference in the number of samples from the different fishing areas did not allow a statistical comparison with a reliable result. Moreover, the distribution of the elements was not Gaussian. We therefore calculated the values of the average, median and maximal concentrations of the analyzed elements, as reported in Table 5. With an analogous argument, the data on the detected PPFAS are reported in Table 6. When the analyte was not quantifiable, the value corresponding to half the LOQ was used to calculate the average concentration. The toxicity and the Health-Based Guidance Values (HBGVs), relating to the elements and PFASs analyzed in anchovies, are shown in Table 7. Table 8 shows the data relating to the estimated daily intakes of elements and PFASs. Note that arsenic is considered only for a tenth of its value found in the analysis. This is because only inorganic arsenic (iAs), 10% of total arsenic, is considered toxic. It must be highlighted that only PFOS and PFOA were detected in the samples. Therefore, the group TWI stated by EFSA for four PFASs refers to the sum of the relieved two of the four PFASs. Only Hg and Sn exceeded 1% of the safe exposure and only considered the maximum values detected. It should be highlighted that the value of the four most toxic organic components of tin was considered as the toxicity value, and, as a further precautionary approach, the tin detected in the analyzed anchovies was represented only by these four compounds. Furthermore, if only the average and median values measured are taken into consideration, an even more favorable situation is observed, represented by approximately 0.5% of the safe exposure for tin and mercury, and even less than 0.5% of the safe exposure for all other elements. The sum of PFOA and PFOS, the two found PFASs of the four considered by EFSA for the evaluation of the TWI, represent 0.06 and 0.5 percent of the TWI for the mean and the highest concentrations detected. It must moreover be noticed that the two highest concentrations of PFOS and PFOA did not belong to the same sample, thus making the calculation safer. Additionally, calculating the THQs as above described and considering the highest concentrations detected, the HIs were always lower than 0.02, thus indicating the absence of concern deriving from the consumption of the preserved anchovies (Table 9). As the cancer risk regards only As, only the non-carcinogenic effects were considered. Following the considerations, the outcome of the oral exposure risk characterization of the researched elements and PFASs shows that, from a chronic toxicity point of view, the consumption of preserved anchovies does not constitute a cause for concern in relation to the sampling data from January to October 2020. The EDI for large consumers also shows that this population subclass is not a matter of concern as regards PTEs and PFASs in preserved anchovies (Table 10). Finally, the acute toxicity of Ni, consisting of contact dermatitis, must be considered. However, due to the peculiarity of the preserved anchovies, it is not easy to define consumption in terms of fish meal. Unlike fresh fish, preserved anchovies are usually an ingredient--for example, for pizza or some pasta sauces. We accounted for individuals of 70, 55 and 40 kg of body weight as representative of a man of average build and height, a woman of medium build and height and a woman of slender and not particularly tall height. A meal of 15 g is the usual serving size of anchovies: the aforementioned individuals could eat 18.4 g, 14.5 g and 10.5 g, respectively, of the sample with the highest concentration (0.418 mg kg-1) without incurring allergic dermatitis. Therefore, due to the acute reference value of 1.1 mg kg-1 bw Ni with an Moe value of 10 that accounts for the different individual susceptibilities, the portion should be reduced, but only for hypothetical consumers of 55 kg and 40 kg in weight. Considering the second highest value (0.201 mg kg-1), the consumption could double to 38.3, 30.1 and 21.9, considerably decreasing the risk of systemic contact dermatitis. Based on our results, only the sample with the highest content of Ni could cause an allergic reaction in Ni-sensitive, low-weight individuals, thus representing a moderate concern. 4. Conclusions Considering the lack of data on PFASs and PTEs in anchovies, the aim of the present work was to investigate the mentioned contaminants in salted and canned anchovies collected from different geographical areas in different months of the year to verify possible variations in bioaccumulation. For this purpose, two analytical methods were developed, one for PFASs and one for PTEs. The PFAS method showed high selectivity/specificity, good recovery from 70 and 120%, good sensitivity with LOQs in the range of 0.050 to 0.10 ng g-1, high precision (RSD < 20%) and a lower matrix effect (<20%). Additionally, the PTE method showed good sensitivity (LOD from 0.400 to 3.60 ng g-1 and LOQ from 1.20 to 12.1 ng g-1) and precision (RSD < 25%). A total of 258 sample pools of salted and canned anchovies were analyzed. Regarding PFASs, only PFBA, PFOS and PFOA were detected (from <LOQ to 8.40 ng g-1); instead, Hg, Cd, Pb, Cr, iAs, Sn, Al and Ni were the elements detected in the anchovies, with concentrations from 90.00 to 4940 ng g-1. According to our results, a risk characterization of the exposure to elements and PFASs was carried out for the Italian consumer, thus permitting us to comprehend the average European consumer. The very low hazard indexes for cumulative toxicities for consumers of large amounts were very reassuring. The only concern derived from Ni acute toxicity, which, also accounting for different individual sensitivities, related to only one sample. Acknowledgments The authors thank Irene Rizzoli and Stefano Valenti of Delicius Rizzoli S.P.A., Parma (Italy) for their valuable collaboration in the study design and sample procurement. Delicius Rizzoli S.P.A., Parma (Italy) funded this study. Author Contributions Conceptualization, F.A., L.M.C., M.N. and S.P.; methodology, M.N.; validation, M.N. and G.M.; formal analysis, M.N., G.M., L.D. and E.F.; investigation, M.N., F.A., S.P. and L.M.C.; resources, L.M.C. and S.P.; data curation, M.N. and L.D.; writing--original draft preparation, M.N., F.A. and E.F.; writing--review and editing, M.N. and F.A.; supervision, S.P. and L.M.C.; project administration, L.M.C. and S.P. All authors have read and agreed to the published version of the manuscript. Data Availability Statement The data are available from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. foods-12-01060-t001_Table 1 Table 1 State of the art on detection of PFASs and potentially toxic elements in fresh and canned anchovies. PFASs Reference Analytes Matrix Extraction Technique Instrumental Analysis LOD-LOQ ng g-1 Application Range Conc. ng g-1 5 PFCAs; 3 PFSAs Different fish including 5 anchovies Alkaline digestion, SPE LC-MS/MS 0.0030-0.050 <LOQ-0.80 13 PFCAs; 4 PFSAs Different foods including anchovies and sardines Basic methanol extraction, acidification, SPE WAX LC-MS/MS 0.00090-0.46 0.0090-9.3 PFOS Eggs, 5 sardines, 5 anchovies Acetonitrile extraction, incubation, purification by activated carbon and glacial acetic acid LC-MS/MS 0.023 0.54-1.5 9 PFCAs; 3 PFSAs; PFOSA Different fish including anchovies Basic methanol extraction, SPE WAX LC-MS/MS 0.48-10 0.51-15 7 PFCAs; 3 PFSAs; 3 PFOSA, 2 diPAP, 3 PFPiA, 2 FTCA, 1 FTUCA Water, sediments and small fish including 15 anchovies Alkaline digestion, methanol extraction, Pesti-Carb cartridges clean up LC-MS/MS 0. 00020-0.056 0.011-0.47 5 PFCAs; 3 PFSAs; PFOSA Different fish including 8 sardines Methanol extraction, concentration, treatment with aqueous KOH, SPE WAX LC-MS/MS 1.0 0.010-3.6 Heavy metals Reference Analytes Matrix Extraction Technique Instrumental Analysis LOD-LOQ ng g-1 Application Range Conc. ng g-1 Fe, Zn, Cu, Cd, Sn, Hg and Pb Canned anchovies and canned rainbow trout Digestion with concentrated (65%) nitric acid (HNO3) 30% hydrogen peroxide (H2O2), microwave and washing ICP-MS / 1.0-5.1 x 104 Pb, As, Cd, Zn, Cu Canned seafood products Homogenization, drying, digestion with (HCl:HNO3 = 1:1), evaporation Atomic absorption spectrometer / 27 x 10-7.1 x 104 Li, Na, Mg, P, Ca, V, Mn, Fe, Co, Ni, Cu, Zn, Ga, As, Se, Rb, Sr, Mo, Pd, Cd, Cs, Ba, Hg, Tl, Pb, U Sardine and anchovy from 6 Greek coastal areas Freeze-drying, homogenization, microwave-assisted acid digestion ICP-MS 2.0-14 x104 40-18 x 104 As, Cd, Co, Cr, Cu, Mn, Mo, Ni, P, Pb, V, Zn, Ca, K, Na, Mg, S and Sr Indian anchovy Digestion with 65% nitric acid ICP-OES 1.0-4.9 x 105 40-75 x 105 Al, Zn, Mn, Co, Cr, Cu, Fe, Ni, Cd, Pb, Se, As and Hg Anchovy of Black Sea Homogenization, drying, digestion with nitric acid and hydrochloric acid, dilution, filtration ICP-MS 0.10-29 3.0-14 x 102 Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Cd, Pb Fresh and salt-dried anchovy from Kuala Terengganu Drying, digestion with deionized water-nitric acid (49:1, v/v) ICP-MS 4.0 4.0-65 x 104 Hg, Cd and Pb Salted anchovies Digestion with HNO3-HClO4 (8:3) for Cd and Pb and with H2SO4-HNO3 (1:1) for Hg Atomic absorption spectroscopy 5.0-10 40-50 x 10 /: not reported in the article; PFCAs: perfluoroalkyl carboxylic acids; PFSAs: perfluoroalkyl sulfonic acids; PFOSA: perfluorooctane sulfonamide; diPAP: fluorotelomer phosphate diester; PFPiA: perfluorophosphinate; FTCA: perfluoroakyl saturated carboxylates; FTUCA: perfluoroakyl unsaturated carboxylates; LC-MS/MS: Liquid Chromatography tandem mass spectrometry; ICP-MS: Inductively coupled plasma mass spectrometry; ICP-OES: Inductively coupled plasma - optical emission spectrometry. foods-12-01060-t002_Table 2 Table 2 Metal concentration reported in the certificate of the reference material, metal concentration in the certified reference materials by ICP-OES and percentage of metal recovery from the reference material. Metal Certificate of Analysis of the Reference Material (ng g-1) From ICP-OES Analysis (ng g-1) Recovery (%) As 6700.00 6740.00 101 Cd 340.000 320.000 94 Pb 2180.00 1990.00 91 Ni 690.000 640.000 93 Cr 730.000 770.000 105 Al Not reported 136,900 - Hg 71.0000 160.000 125 Sn Not reported 270.000 - -: Not calculated. foods-12-01060-t003_Table 3 Table 3 PFAS list with their formula, parent exact mass and validation performance. Compound Formula Parent Exact Mass [m/z] Observed Parent Mass [m/z] Main Fragment ion [m/z] LOQ (ng g-1) Recovery % Intra-Day % RSD Inter-Day % RSD PFBA C4HF7O2 212.97920 212.97929 168.98955 0.050 115 7 18 PFPeA C5HF9O2 262.97601 262.97609 218.98612 0.050 114 11 11 PFBS C4F9HO3S 298.94299 298.94290 98.95434 0.050 104 11 12 PFHxA C6HF11O2 312.97281 312.97279 268.98352 0.050 117 7 10 PFHpA C7HF13O2 362.96962 362.96959 318.97977 0.050 113 6 10 PFHxS C6F13HO3S 398.9366 398.93650 79.95743 0.050 90 6 9 PFOA C8HF15O2 412.96643 412.96638 368.97698 0.050 113 6 9 PFNA C9HF17O2 462.96323 462.96318 418.97384 0.050 92 10 14 PFOS C8F17HO3S 498.93022 498.93024 79.95741 0.050 94 12 15 PFDA C10HF19O2 512.96004 512.96008 468.96990 0.050 86 11 16 PFUnDA C11HF21O2 562.95684 562.95679 518.96770 0.050 85 8 13 PFDS C10F21HO3S 598.92383 598.92377 79.95743 0.050 94 9 13 PFDoA C12HF23O2 612.95365 612.95360 568.96405 0.050 74 9 13 PFTrDA C13HF25O2 662.95046 662.95041 618.96097 0.10 80 8 15 PFTeDA C14HF27O2 712.94726 712.94719 668.95808 0.10 88 15 20 PFHxDA C16HF31O2 812.94088 812.94078 768.94913 0.10 88 16 20 PFODA C18HF35O2 912.93449 912.93440 868.94544 0.10 87 17 20 All compounds were detected in the deprotonated form. foods-12-01060-t004_Table 4 Table 4 Limit of detection (LOD), limit of quantification (LOQ) and precision (% RSD) of the investigated PTEs. Element LOD (ng g-1) LOQ (ng g-1) RSD % Hg 0.70 2.3 9 Cd 0.60 1.9 3 Pb 2.6 8.6 22 Cr 0.60 2.1 3 As 3.6 12 22 Sn 2.0 6.7 25 Al 0.40 1.2 5 Ni 1.3 4.2 8 foods-12-01060-t005_Table 5 Table 5 Mean, median and maximum concentrations of elements detected in the anchovies. Concentration (ng g-1) Hg Cd Pb Cr iAs Sn Al Ni Positives % 71 100 99 100 100 100 100 100 Mean 290.0 60.00 170.0 90.00 230.0 200.0 1880 50.00 Median 280.0 50.00 170.0 80.00 210.0 200.0 1720 50.00 Maximum 650.0 90.00 750.0 710.0 510.0 670.0 4940 420.0 foods-12-01060-t006_Table 6 Table 6 Mean, median and maximum concentrations of PFASs detected in the anchovies. Concentration (ng g-1). PFBA PFOA PFOS Positives % 100 1 83 Mean 2.08 <LOQ 0.0860 Median 1.76 0 <LOQ Maximum 8.40 <LOQ 1.15 foods-12-01060-t007_Table 7 Table 7 Potential toxicity and Health-Based Guidance Values of searched elements and PFAS. Element/ PFAS Potential Toxicity Health-Based Guidance Value Mercury Nervous system dysfunction such as tremors, irritability, memory problems, impaired vision and hearing. Exposure of mothers could lead to the birth of babies with permanent dysfunction of the nervous system. TWI a = 1.3 mg kg-1 bw per week of methylmercury, expressed as mercury for neurodevelopmental outcomes after prenatal exposure Cadmium Kidney and respiratory diseases TWI a = 2.5 mg kg-1 bw per week for tubular damage Lead Severe brain and kidney damage, possible miscarriage BMDL10 b Neurodevelopmental toxicity: 0.50 mg kg-1 bw per day Blood pressure: 1.5 mg kg-1 bw per day kidney: 0.63 mg kg-1 bw per day Chromium Respiratory problems, cough, asthma and allergic reactions. Chronic exposure could cause liver and kidney cancer, in particular, linked to Cr (VI), which, however, is rare to find in food due to its reduction to Cr (III). TDI c = 300 mg kg-1 bw per day of Cr (III)) on reproductive and developmental toxicity Arsenic In its organic form, it has negligible toxicity due to the fast excretion kinetics. In the form of inorganic arsenic, less than 10% of the total arsenic in fish is linked to skin, lung and bladder cancer. BMDL01 d for skin lesions and lung, bladder and skin cancers, ranges 0.30 and 8.0 mg kg-1 bw per day Tin Inorganic tin interferes with the metabolism of zinc, copper and iron, with the synthesis and catabolism of the heme group. Metallic tin and inorganic tin compounds are relatively nontoxic . TDI c of 0.10 mg kg-1 bw per day for immunotoxic effects of tributyltin, dibutyltin, triphenyltin and di-n-octyltin. This very precautionary value for organic tin (the last two used as additives in PVC and in materials in contact with food) is considered . Aluminium For professional exposure only, the target organs are the lungs and bones. The toxicity on the central nervous system (CNS) includes dementia in dialyzed patients (due to aluminum entering the circulation with dialysis) or oral exposure to Al hydroxide administered to patients and Parkinson's disease; however, it should be emphasized that in the two syndromes, the serum or cerebral levels of aluminum could be an effect of the syndrome and not the cause. TWI a = 1.0 mg kg-1 bw per week for effects on the developing nervous system Nickel Long-term toxicity: cancerogenic, immunotoxic, hepatotoxic, neurotoxic and nephrotoxic only through inhalation. Long-term exposure could cause reproductive diseases. Acute toxicity: allergic and eczematous reactions in sensitive individuals. TDI c = 2.8 mg kg-1 bw per day reproductive and developmental toxicity BMDL10 b = 1.1 mg kg bw with a margin of exposure (MOE) equal to or greater than 10, accounting for the variability of the response in sensitized individuals PFAS Toxicity on immune system, on reproduction and development TWI = 4.4 ng kg-1 bw per week for the sum of PFOA, PFOS, PFHxS and PFNA for the decrease in immune response to vaccination individuals exposed even during the mother's pregnancy. a TWI = tolerable weekly intake; b BMDL10 = benchmark dose--lower bound 10%; c TDI = tolerable daily intake; d BMDL01 = benchmark dose--lower bound 0.1%. foods-12-01060-t008_Table 8 Table 8 Data relating to the characterization of the risk from oral exposure to the studied elements through the consumption of preserved anchovies (HBGVs and EDIs of elements expressed as mg kg-1 bw per day; HBGV and EDI of PFAS as ng kg-1 bw per day). Hg Cd Pb Cr iAs Sn Al Ni PFOA + PFOS HBGV a 0.180 0.380 0.500 300 0.300 0.100 143 2.80 0.630 EDImean 0.000910 0.000190 0.000530 0.000280 0.000720 0.000630 0.00590 0.000160 0.000350 EDIMaximum 0.00200 0.000280 0.00240 0.00220 0.00160 0.00210 0.0160 0.00130 0.00370 a The HBGVs are reported on a daily basis even when they are stated by EFSA on a weekly basis. foods-12-01060-t009_Table 9 Table 9 Hypothetical target hazard quotients and hazard indexes for the estimated daily exposure (EDI) through preserved anchovies at the highest concentrations detected. All the Health-Based Guidance Values (HBGVs) are by EFSA. TWIs are recalculated and expressed on a daily basis. As is reported as inorganic arsenic and only Cr (III) is considered. EDI and HBGV are expressed as mg kg-1 day-1 for elements and as ng kg-1 day-1 for PFASs. HBGV Element EDI Target Hazard Quotient (HBGV Expressed on a Daily Basis) Neurodevelopment Kidney Blood Pressure Reproduction/Development Skin Immune System TWI Hg 0.00200 0.0100 (0.180) TWI Cd 0.000290 0.00080 (0.38) BMDL10 Pb 0.00230 0.00500 (0.500) 0.00400 (0.630) 0.00200 (1.50) TDI Cr 0.00230 0.0000800 (300) BMDL10 iAs 0.00160 0.0050 (0.3) TDI Sn 0.00210 0.020 (0.1) TWI Al 0.0160 0.000100 (142) TDI Ni 0.00130 0.000500 (2.80) TWI PFOS + PFOA 0.000600 (0.630) Hazard Index 0.0150 0.00480 0.00200 0.000580 0.00500 0.0210 foods-12-01060-t010_Table 10 Table 10 Hazard indexes for the estimated daily exposure to the studied elements via preserved anchovies of the 95th percentile fish consumers, accounting for the highest concentrations detected. Hazard Index 95th Percentile Consumers Neurodevelopment Hg + Pb + Al Kidney Cd + Pb Blood Pressure Pb Reproduction/Development Cr + Ni Skin As Immune System Sn + PFOS + PFOA 0.060 0.014 0.0060 0.0020 0.015 0.060 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Mantovani A. Ferrari D. Frazzoli C. Sustainability, security and safety in the feed-to-fish chain: Focus on toxic contamination Int. J. Nutr. Food Sci. 2015 4 6 24 10.11648/j.ijnfs.s.2015040202.12 2. Chiesa L.M. Nobile M. Ceriani F. Malandra R. Arioli F. Panseri S. Risk characterisation from the presence of environmental contaminants and antibiotic residues in wild and farmed salmon from different FAO zones Food Addit. Contam. Part A 2019 36 152 162 10.1080/19440049.2018.1563723 3. Chiesa L.M. Labella G.F. Panseri S. Pavlovic R. Bonacci S. Arioli F. Distribution of persistent organic pollutants (POPs) in wild Bluefin tuna (Thunnus thynnus) from different FAO capture zones Chemosphere 2016 153 162 169 10.1016/j.chemosphere.2016.03.010 27016811 4. Christensen K.Y. Raymond M. Blackowicz M. Liu Y. Thompson B.A. Anderson H.A. Turyk M. Perfluoroalkyl substances and fish consumption Environ. Res. 2017 154 145 151 10.1016/j.envres.2016.12.032 28073048 5. European Environment The Use of PFAS and Fluorine-Free Alternatives in Textiles, Upholstery, Carpets, Leather and Apparel 2020 Available online: (accessed on 25 January 2023) 6. EFSA Perfluorooctane sulfonate (PFOS), perfluorooctanoic acid (PFOA) and their salts. Scientific Opinion of the Panel on Contaminants in the Food chain EFSA J. 2008 653 1 131 7. EPA State of the Great Lakes 2009. Prepared by Environment Canada and United States Environmental Protection Agency 2009 Available online: (accessed on 25 January 2023) 8. Chiesa L.M. Pavlovic R. Arioli F. Nobile M. Di Cesare F. Mosconi G. Falletta E. Malandra R. Panseri S. Presence of perfluoroalkyl substances in Mediterranean Sea and North Italian Lake fish addressed to Italian consumer Int. J. Food Sci. Technol. 2022 57 1303 1316 10.1111/ijfs.15532 9. EFSA Scientific Opinion on the risk to human health related to the presence of perfluoroalkyl substances in food EFSA J. 2020 18 6223 6613 10.2903/j.efsa.2020.6223 10. Mol S. Determination of trace metals in canned anchovies and canned rainbow trouts Food Chem. Toxicol. 2011 49 348 351 10.1016/j.fct.2010.11.005 21074591 11. Mohamadi S. Mahmudiono T. Zienali T. Sadighara P. Omidi B. Limam I. Fakhri Y. Probabilistic health risk assessment of heavy metals (Cd, Pb, and As) in Cocoa powder (Theobroma cacao) in Tehran, Iran market Int. J. Environ. Res. Public Health 2022 1 16 10.1080/09603123.2022.2146070 36612323 12. Kruszewski B. Obiedzinski M.W. Multivariate analysis of essential elements in raw cocoa and processed chocolate mass materials from three different manufacturers LWT 2018 98 113 123 10.1016/j.lwt.2018.08.030 13. EUMOFA European Market Observatory For Fisheries And Aquaculture Products. The EU Fish Market 2020 Edition Available online: (accessed on 25 January 2023) 14. Llorca M. Farre M. Pico Y. Barcelo D. Development and validation of a pressurized liquid extraction liquid chromatography-tandem mass spectrometry method for perfluorinated compounds determination in fish J. Chromatogr. A 2009 1216 7195 7204 10.1016/j.chroma.2009.06.062 19616213 15. Domingo J.L. Jogsten I.E. Eriksson U. Martorell I. Perello G. Nadal M. van Bavel B. Human dietary exposure to perfluoroalkyl substances in Catalonia, Spain. Temporal trend Food Chem. 2012 135 1575 1582 10.1016/j.foodchem.2012.06.054 22953896 16. Bertolero A. Vicente J. Meyer J. Lacorte S. Accumulation and maternal transfer of perfluorooctane sulphonic acid in yellow-legged (Larus michahellis) and Audouin's gull (Larus audouinii) from the Ebro Delta Natural Park Environ. Res. 2015 137 208 214 10.1016/j.envres.2014.12.018 25575371 17. Sedlak M.D. Benskin J.P. Wong A. Grace R. Greig D.J. Per-and polyfluoroalkyl substances (PFASs) in San Francisco Bay wildlife: Temporal trends, exposure pathways, and notable presence of precursor compounds Chemosphere 2017 185 1217 1226 10.1016/j.chemosphere.2017.04.096 28477851 18. Chen M. Wang Q. Shan G. Zhu L. Yang L. Liu M. Occurrence, partitioning and bioaccumulation of emerging and legacy per-and polyfluoroalkyl substances in Taihu Lake, China Sci. Total Environ. 2018 634 251 259 10.1016/j.scitotenv.2018.03.301 29627548 19. Fernandes A.R. Mortimer D. Holmes M. Rose M. Zhihua L. Huang X. Smith F. Panton S. Marshall L. Occurrence and spatial distribution of chemical contaminants in edible fish species collected from UK and proximate marine waters Environ. Int. 2018 114 219 230 10.1016/j.envint.2018.02.047 29522986 20. Ismail S. Azhari S. Ngah C.W.Z.C.W. Determination of selected toxic metal (As, Cd, Pb) and essential (Zn, Cu) elements in local canned seafood products AIP Conf. Proc. 2018 1972 030020 10.1063/1.5041241 21. Sofoulaki K. Kalantzi I. Machias A. Pergantis S.A. Tsapakis M. Metals in sardine and anchovy from Greek coastal areas: Public health risk and nutritional benefits assessment Food Chem. Toxicol. 2019 123 113 124 10.1016/j.fct.2018.10.053 30366072 22. Alizada N. Malik S. Muzaffar S.B. Bioaccumulation of heavy metals in tissues of Indian anchovy (Stolephorus indicus) from the UAE coast, Arabian Gulf Mar. Pollut. Bull. 2020 154 111033 10.1016/j.marpolbul.2020.111033 32319889 23. Karsli B. Determination of metal content in anchovy (Engraulis encrasicolus) from Turkey, Georgia and Abkhazia coasts of the Black Sea: Evaluation of potential risks associated with human consumption Mar. Pollut. Bull. 2021 165 112108 10.1016/j.marpolbul.2021.112108 33581567 24. Nurulnadia M.Y. Nik-Nurasyikin N.M.A. Ling K.H. Zahid B.M. Adiana G. Nurlemsha B.I. Metal concentrations in fresh and salt-dried anchovy, Encrasicholina devisi, and estimation of target hazard quotient for consumers in Kuala Terengganu Reg. Stud. Mar. Sci. 2021 41 101595 10.1016/j.rsma.2020.101595 25. Storelli M.M. Giachi L. Giungato D. Storelli A. Occurrence of heavy metals (Hg, Cd, and Pb) and polychlorinated biphenyls in salted anchovies J. Food Prot. 2011 74 796 800 10.4315/0362-028X.JFP-10-453 21549051 26. European Commission Analytical Quality Control and Method Validation Procedures for Pesticide Residues Analysis in Food and Feed (SANTE 11312/2021) 2021 Available online: (accessed on 25 January 2023) 27. European Union Commission Recommendation (EU) 2022/1431 of 24 August 2022 on the monitoring of perfluoroalkyl substances in food OJEU 2022 L221 105 109 28. ISTAT Available online: (accessed on 25 January 2023) 29. EUMOFA European Market Observatory for Fisheries and Aquaculture Product: Anchovy Transformed in Italy. Case Study (February) 2018 Available online: (accessed on 25 January 2023) 30. Mosconi G. Di Cesare F. Arioli F. Nobile M. Tedesco D.E. Chiesa L.M. Panseri S. Organohalogenated Substances and Polycyclic Aromatic Hydrocarbons in Fish from Mediterranean Sea and North Italian Lakes: Related Risk for the Italian Consumers Foods 2022 11 2241 10.3390/foods11152241 35954008 31. Chiesa L.M. Ceriani F. Caligara M. Di Candia D. Malandra R. Panseri S. Arioli F. Mussels and clams from the italian fish market. is there a human exposition risk to metals and arsenic? Chemosphere 2018 194 644 649 10.1016/j.chemosphere.2017.12.041 29241139 32. Soeborg T. Frederiksen H. Andersson A.M. Cumulative risk assessment of phthalate exposure of Danish children and adolescents using the hazard index approach Int. J. Androl. 2012 35 245 252 10.1111/j.1365-2605.2011.01240.x 22320716 33. EFSA CONTAM Panel EFSA Panel on Contaminants in the Food Chain (CONTAM); Scientific opinion on the risk for public health related to the presence of mercury and methylmercury in food EFSA J. 2012 10 2985 3225 10.2903/j.efsa.2012.2985 34. EFSA CONTAM Panel EFSA Panel on Contaminants in the Food Chain (CONTAM); Cadmium in food-Scientific opinion of the panel on contaminants in the food chain EFSA J. 2009 7 980 10.2903/j.efsa.2009.980 35. EFSA CONTAM Panel EFSA Panel on Contaminants in the Food Chain (CONTAM); Scientific opinion on lead in food EFSA J. 2010 8 1570 1720 10.2903/j.efsa.2010.1570 36. EFSA CONTAM Panel EFSA Panel on Contaminants in the Food Chain (CONTAM); Scientific opinion on the risks to public health related to the presence of chromium in food and drinking water EFSA J. 2014 9 11 25 10.2903/j.efsa.2014.3595 37. EFSA CONTAM Panel EFSA Panel on Contaminants in the Food Chain (CONTAM); Scientific opinion on arsenic in food EFSA J. 2009 7 199 1351 10.2903/j.efsa.2009.1351 38. EFSA Chronic dietary exposure to inorganic arsenic EFSA J. 2021 19 6380 6429 10.2903/j.efsa.2021.6380 39. Tokar E.J. Boyd W.A. Freedman J.H. Waalkes M.P. Toxic effects of metals Casarett and Doull's Toxicology: The Basic Science of Poisons 8th ed. Klaassen C.D. McGraw Hill London, UK 2013 Available online: (accessed on 25 January 2023) 40. EFSA Opinion of the Scientific Panel on contaminants in the food chain (CONTAM) to assess the health risks to consumers associated with exposure to organotins in foodstuffs EFSA J. 2004 102 1 119 10.2903/j.efsa.2004.102 41. EFSA Safety of aluminium from dietary intake--Scientific Opinion of the Panel on Food Additives, Flavourings, Processing Aids and Food Contact Materials (AFC) EFSA J. 2008 754 1 34 10.2903/j.efsa.2008.754 42. EFSA Scientific opinion on the risks to public health related to the presence of nickel in food and drinking water EFSA J. 2015 13 4002 4203 10.2903/j.efsa.2015.4002 |
PMC10000573 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050647 healthcare-11-00647 Article Application of Forensic DNA Phenotyping for Prediction of Eye, Hair and Skin Colour in Highly Decomposed Bodies Fabbri Matteo 1+ Alfieri Letizia 1*+ Mazdai Leila 2 Frisoni Paolo 3 Gaudio Rosa Maria 4 Neri Margherita 1 Feola Alessandro Academic Editor Arcangeli Mauro Academic Editor 1 Department of Medical Sciences, University of Ferrara, 44121 Ferrara, Italy 2 Department of Neuroscience and Rehabilitation, University of Ferrara, 44121 Ferrara, Italy 3 Unit of Legal Medicine, Azienda USL di Ferrara, 44121 Ferrara, Italy 4 Department of Translational Medicine, University of Ferrara, 44121 Ferrara, Italy * Correspondence: [email protected] + These authors contributed equally to this work and share first authorship. 23 2 2023 3 2023 11 5 64731 1 2023 18 2 2023 21 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). In the last few years, predicting externally visible characteristics (EVCs) by adopting informative DNA molecular markers has become a method in forensic genetics that has increased its value, giving rise to an interesting field called "Forensic DNA Phenotyping" (FDP). The most meaningful forensic applications of EVCs prediction are those in which, having only a DNA sample isolated from highly decomposed remains, it is essential to reconstruct the physical appearance of a person. Through this approach, we set out to evaluate 20 skeletal remains of Italian provenance in order to associate them with as many cases of missing persons as possible. To achieve the intended goal, in this work we applied the HIrisPlex-S multiplex system through the conventional short tandem repeats (STR) method to confirm the expected identity of subjects by evaluating phenotypic features. To investigate the reliability and accuracy of the DNA-based EVCs prediction, pictures of the cases were compared as they were available to researchers. Results showed an overall prediction accuracy greater than 90% for all three phenotypic features--iris, hair, and skin colour--at a probability threshold of 0.7. The experimental analysis showed inconclusive results in only two cases; this is probably due to the characteristics of subjects who had an intermediate eye and hair colour, for which the DNA-based system needs to improve the prediction accuracy. forensic DNA phenotyping predictive DNA analysis HIrisPlex-S eye colour skin colour hair colour identification This research received no external funding. pmc1. Introduction Many forensic techniques are used today to identify a human corpse, mainly relating to the conditions of the remains. According to the Interpol Protocol , the four post-mortem recognised methods for corpse identification are finding of physical indications , matching of fingerprints , when pre-mortem inked prints are available, dental examination , when pre-mortem dental radiographs are available, and DNA analysis . Poor conditions of preservation of the remains, due to an advanced state of decomposition, intervention of physical/chemical elements or micro/macro fauna, often result in high fragmentation, degradation, and mixing of biological samples . In these cases, it is customary to resort to the evaluation of bone remains with methods of forensic anthropology. Forensic anthropology can be defined as the study of the morphometric characteristics of bone tissue. This approach is inescapably useful in the first phase of the identification process, such as the collection of personal identification characteristics (such as gender, age, height, ethnic group, body conformation, etc.), especially for bodies in an advanced state of decomposition. However, for what concerns the identification processes, the data that can be obtained through anthropological investigations are often insufficient . It seems therefore fundamental that these data have to be completed by adding other individual characteristics such as the chromatic somatic characters (skin pigmentation, eyes, and hair), iatrogenic or pathological outcomes (prostheses, metal plates, metal stitches, etc.), and other signs (tattoos, piercings, etc.) . Especially in those corpses found in a scant state of preservation, however, these characters are not detectable, and the anthropological investigation does not look capable of providing a sufficient probabilistic value, such as to ascertain the identity of the subject, as it can only allow an evaluation in terms of concordance between the data of ante and post-mortem. For this reason, DNA profiling is the most widespread and reliable method of personal identification in forensics when this must take place on highly degraded human remains and no dental records are available or the concordance of the fingerprints cannot be determined . For genetic identification of remains, the Interpol DNA-specific recommendations for Disaster Victim Identification (DVI) are also proposed by the ISFG membership for single-case identification in order to provide an internationally standardised method. The timing and proper storage of post-mortem samples affect the success rates of DNA typing. During the autopsy or external examination, the forensic geneticist, or a pathologist with a background in forensic genetics, should ideally be available for consultation during DNA sample collection. However, genetic identification linked to DNA profiling still poses some challenges today. One method to redesign STR primers that generate shorter amplicons is to improve the success rate of degraded DNA. During collaborative European exercises, the employment of several STR systems with size-reduced fragments gave positive results when applied to artificially degraded DNA . In cases where the starting material was of a reduced quantity, as in contact traces derived from the victims' personal items, the available DNA extracted reflects the low level of material for genetic analysis. In other cases, relatives' biological specimens may be unavailable for genetic analysis and comparison as well . These are the major limitations that confront forensic DNA (fDNA), as well as in an anthropological context where samples on which to apply molecular investigations are often both inadequate and unrepresentative . In these cases--especially on anthropological remains--it is possible to refer to mitochondrial DNA (mtDNA), which, by an intrinsic property of the organelle, allows for greater protection of the nucleic acid and consequently better preservation . In addition, within a eukaryotic cell, mtDNA is present in significantly more copies than nuclear DNA (nuDNA), which is present in single units . However, mtDNA does not allow a true comparison investigation; this is because it is inherited exclusively matrilineally and, for this reason, is identical in individuals descended from the same matrilineal line . Therefore, it is intuitive to understand why a mitochondrial DNA match cannot lead to absolute identification because, for example, comparison references must come from maternal relatives since mtDNA is inherited exclusively from the maternal parent, and unfortunately, this is not always practicable . The direct consequence of degenerative processes occurring in the post-mortem period is cellular breakdown, which is certainly promoted by extreme environmental conditions (e.g., high temperatures and oxygen deficiency) . All of this results in a completely random degradation of the nuDNA molecule, promoted and facilitated through the action of intracellular enzymes (e.g., nuclease, lipase, and protease) that are released due to membrane failure . Comparatively few environmental conditions provide better preservation and protection of biological matrices from degradation. The more stabilised DNA will be more likely to be found in skeletal tissue than in the soft tissues of the organism. Indeed, the gold standard in these situations--anthropological and forensic--is represented by the bone matrix . As per forensic practice, depending on corpse preservation, different tissue types should be collected, as suggested by Prinz, Carracedo et al. . In particular, for very decomposed or segmented remains, it would be preferable to collect fragments of any available bone or healthy teeth to be used as a substrate for DNA extraction. Cancellous bone can be rich in DNA, but its preservation is not reliable, so in these cases, dense cortical bone should always be the first choice, preferably taken from long bones of the lower limbs . Given this premise, in those cases where standard forensic DNA analysis based on short tandem repeats (STRs) or other types of polymorphic DNA markers such as single nucleotide polymorphisms (SNPs) fails to identify the donor of a human biological trace found at a crime scene or on an unidentified corpse via comparative DNA profiling, alternative approaches are needed . Forensic DNA phenotyping (FDP) analysis, as a matter of fact, could be crucial in the forensic field in the event that standard DNA analysis is not possible in the absence of a comparison sample . The aim of phenotype studies, through the analysis of DNA sequences, is to predict external somatic traits by processing biological material from unknown corpses or from identified biological traces found during crime scene recognition. Throughout the years, the evolution of this technology has allowed the development of a promising technique due to the evidence it can produce from biological tests . FDP can lead to the prediction of the visible external characteristics of a subject, useful for its identification, starting either from traces of DNA detected at the scene of a crime or from DNA extracted from human remains, even those that are highly decomposed. This can be very useful in identifying intent in cases of missing persons or in the need to use alternative methods for the identification of victims of a mass disaster, especially where ante-mortem reference specimens and/or known relatives are unobtainable . Although external visible characteristics (EVC) are considered complex traits, in which multiple genes contribute to the phenotype along with environmental factors, human pigmentation traits are found to be the least complicated of the EVCs, with only a few genes providing the vast amount of phenotypic information. For this reason, currently, the understanding of the genetic basis of chromatic traits is more studied and known than for other EVCs. The problem with extremely complex genetic characteristics, as extensively studied for many common diseases, is that any single gene contributes only a small portion of the phenotypic variance, and only the combination of a large number of genes accounts for the inherited trait . Precisely in this study, we applied a DNA-based method for the prediction of eye, hair, and skin colour on 20 identification cases dealing with skeletal remains previously identified by the laboratory using conventional STRs analysis. Bone samples, taken from 20 different missing subjects, were chosen for the experimental application because DNA showed a high degree of degradation, resulting in partial STR profiles. The cases have been chosen because pictures of the victims were available as a reference document; photographs were used in order to test and verify the fidelity and accuracy of the DNA multiplex reaction. Among the currently available statistical models, we choose to apply the HIrisPlex-S system (HIrisPlex-S typing software, open source software available at accessed on 17 February 2023, Copyright of the Department of Genetic Identification of Erasmus MC, Rotterdam, The Netherlands), which consists of 41 DNA variants: 24 included in the HIrisPlex assay and 17 variants investigated with another analysis. Both multiplex assays can be carried out even with a minimum amount of DNA of 63 pg . The analysis stipulates that subsequently the obtained genotype data are entered into a tool published on the site accessed on 17 February 2023, which is useful for evaluating prediction probabilities for 3 iris colours, 4 hair colours, and 5 skin colour categories . Moreover, this system has been preferred because it has been shown to be experimentally applied to naturally degraded genetic material. In fact, the HIrisPlex system was applied by J. Draus-Barini et al. to DNA samples extracted from old and ancient tissues, demonstrating its suitability in deteriorated DNA analysis. Their research gave promising results on very ancient and therefore very fragmented genetic material, revealing that out of the 26 DNA extracts from bones and teeth between 1 and about 800 years of post-mortem age, 23 yielded a complete 24 SNP profile . Furthermore, in 2014, King, T. et al. applied the HIrisPlex system to even older human tissues, in particular skeletal remains dating back to 1400 AD, leading researchers to the prediction of somatic characteristics (eye and hair colour) by King Richard III of England (1452-1485 AD). The predicted features appeared consistent with the known portraits of the sovereign . Furthermore, the HIrisPlex-S system is forensically endorsed, as forensic validation tests have been carried out and published for both multiplexes, in accordance with the guidelines of the Scientific Working Group on DNA Analysis Methods (SWGDAM) . 2. Materials and Methods 2.1. Sample Selection The preliminary step that preceded the effective analytical steps involved the evaluation of bone matrix samples from the 20 subjects under study. A total of 14 femurs and 6 tibias were selected, respectively, in case there were bodies for whom visual identification was impossible, due to the very poor conservation state of the corpse. Indeed, cases for the present study have been selected from those subjected to advanced decomposition or massive damage that occurred due to the particular circumstances of death or due to attempts to conceal or destroy the body. Specifically, corpses prone to thermal or mechanical damage (e.g., combustion, explosion, dismemberment) were included in the study. The post-mortem interval (PMI) of the cadavers included in the study ranged from a few days, for those cases in which massive body alteration had occurred secondary to the peculiar causes of death (explosion, charring), up to about 30 years in cases of exhumation of skeletal remains. The cases were selected from those received, due to the need for multidisciplinary forensic analysis, at the Laboratory of Genetics of the Unit of Legal Medicine of the University of Ferrara. For each case, a complete external inspection, autopsy (if possible due to the presence of residual soft tissues), anthropological evaluation, and identification process were carried out. 2.2. DNA Extraction and Phenotyping In all cases, a standard forensic DNA analysis based on short tandem repeats (STR) was performed. For each case, it has been possible to reach a correct personal identification by comparing the data obtained from the cadaver analysis, with data stored in the register of missing persons or with investigative files. As a result of the personal identification, photographic reproductions of the face of the missing person were available in order to trace the externally visible characteristics. The photographs were extracted from the files provided by the prosecutor's office, from the obituary pages of the local press, or from websites specialised in the search for missing persons. More details about the samples tested are summarised in Table 1. After appropriate and careful sampling was performed, the analytical laboratory steps were proceeded with, which are outlined in Figure 1. For every corpse, one sample of bone powder was collected (femur or tibia as reported in Table 1), except for samples named I7 and I10, where two bone powder samples were collected (femur sample I7; tibia sample I10) due to the high degree of degradation shown by these specimens. For the phenotyping investigation, we decided to apply the HIrisPlex-S system following the recommendations available in the literature . We applied the prediction system through the combined approach of multiplex PCR with SNaPshot single base extension (SBE) reactions to analyse multiple SNP in samples with different levels of gDNA degradation . For nucleic acid extraction, all samples of the bone matrix were treated appropriately. First, each bone sample was chemically treated, using diluted bleach, and then irradiated with UV light for 30 min, before being powdered as much as possible. Then, after the appropriate powdering of the bone matrix has been achieved, genomic DNA (gDNA) extraction can proceed. The procedure is performed using an input of 0.5 g of pulverised bone material as a starting point. The gDNA was extracted from the samples under study using the QIAamp DNA Investigator Kit (Qiagen(r) purchased at QIAGEN S.r.l.--Via Filippo Sassetti n. 16, 20124 Milano, Italy) according to the manufacturer's guidelines. The extracted and purified nucleic acid obtained was quantified by following the instructions of the Investigator Quantiplex Hyres Kit (Qiagen(r) purchased at QIAGEN S.r.l.--Via Filippo Sassetti n. 16, 20124 Milano, Italy). Initially, short adjacent regions of marker SNPs are amplified in a multiplex PCR reaction with a total volume of 10 mL. For the design of multiplex reactions and SBE-PCR primers, guidelines from Chaitanya L. published applying the HIrixPlex-S approach were followed . Specifically, 10 mL of extracted and purified gDNA was used in the multiplex PCR reaction, for which there were distinct concentrations for each sample under investigation as a result of degradation. A Verity 96-Well Fast thermal cycler (Thermo Scientific(r) purchased at Thermo Scientific Inc.--Via G. B. Tiepolo, 18, 20900 Monza, Italy) was used for all amplification reactions. The amplified PCR products were purified in order to remove the excess of unincorporated primers in the reaction by enzymatic method using Exonuclease-I (Thermo Scientific(r) purchased at Thermo Scientific Inc.--Via G. B. Tiepolo, 18, 20900 Monza, Italy). Following purification, the SBE reaction was set up: after the assembly of the primers located immediately beside the target position, the extension allows a fluorescent ddNTP to be linked to the variable site . Next, the extended products were purified by the enzymatic method again, but this time with alkaline shrimp phosphatase (SAP) to remove excess reaction oligonucleotides and prevent them from producing aberrant fluorescence signals during the detection step. In the end, the products of the second purification step were examined with an ABI 310 HID gene capillary sequencer with POP-4 polymer on a 47 cm-long capillary. For allele calling and analysis of the results, the Gene Mapper ID-X v1.1 software program was used. The genotype data obtained by the two multiplex PCR assays were entered in the tool provided by the internet site (last accessed on 17 February 2023) to generate individual prediction probabilities for eye, hair, and skin colour categories . As an additional analysis control, the victim's iris, hair, and skin colour were subjectively and objectively determined for all 20 skeletal remains analysed by four independent laboratory operators. The HIrisPlex-S system allows free evaluation of analytical results in the web interface of (last accessed on 17 February 2023). This approach provides a colour prediction of the three phenotypic traits based on genotypic outcomes, associating a probabilistic threshold of accuracy. Initially, the EVC prediction models started to estimate only iris colouration using the IrisPlex approach , which achieved a prediction accuracy expressed in the area under the receiver operating curve (AUC) of 0.94 for "blue eye colour", 0.74 for "intermediate eye colour", and 0.95 for "brown eye colour". Following that, with the HIrisPlex model, it was possible to evaluate both eye colour and hair colour at the same time . Using the latter model--of predicting eye and hair colour--it obtains an AUC performance of 0.93 for "red colour", 0.81 for "blond", 0.74 for "brown", and 0.86 for "black". Finally, the HIrisPlex-S system is developed, which adds skin colour assessment to the power of prediction, realising an AUC performance of 0.83 for "very light", 0.76 for "pale", 0.78 for "intermediate", 0.98 for "dark", and 0.99 for "dark to black" . For the definition of the categories to which the colour shades of the eyes, hair, and skin colour of individual cases belonged, four examiners were involved in order to classify the colour subcategory (e.g., "pale" or "very light") from the photographic material available. The results were consistent for about 89% of photographic evaluations. As expected, the evaluation of the pictures showed the highest inconstancy in identifying the intermediate eye and skin colour categories, for which all four examiners provided opposing descriptions. This limitation in the prediction of intermediate colour categories is attributed to a database that can still be refined and extended to the greatest number of pigmentation combinations of all three phenotypic traits under evaluation. By enriching the database in this way, there will certainly be an increase in the prediction power and accuracy of the method. An actual improvement in the phenotypic evaluation of forensic samples, in this particular case, will certainly be given by an increase in SNPs describing each trait in terms of colourations. 3. Results Results of the assay showed an overall prediction accuracy of 91.6%, 90.4%, and 91.2%, respectively, for iris, hair, and skin colour, at the 0.7 prediction probability threshold. Although DNA degradation was observed in the genetic profiles achieved by the STR analysis, the overall PCR amplification of the selected SNPs was successful for both the HIrisPlex and HIrisPlex-S assays. Figure 2 shows an example of an SNP profile achieved after the analysis of the sample identified as I1. Among the tested samples, only two turned out to have inconclusive results as compared to the HIrisPlex-S database. These results matched the somatic features shown by the pictures of these specific subjects, but, as documented, these samples showed an intermediate eye and hair colour. Figure 3 shows the prediction values for samples I7 and I10. Moreover, these specimens were particularly interesting because they were collected from two corpses showing the worst conservation status and the longest post-mortem interval (PMI), respectively, of 7 months and 30 years prior to DNA extraction. These samples showed the highest level of DNA degradation among those tested in the present work. The evaluation of the STR profiles achieved from the analysis of these samples allowed to identify the loss of larger fragments used in commercial STR kits (greater than 150-170 bp), together with allelic dropout. 4. Discussion Here we present an experimental application in forensic cases of the DNA-based method HIrisPlex-S for the simultaneous prediction of eye, hair, and skin colour. The applied multiplex genotyping system was previously designed to deal with degraded DNA and was based on a genotyping technology that relies on equipment widely used by the forensic community. In order to test the HIrisPlex-S assay in forensic cases, the multiplex was performed on a total of 20 different samples, chosen for the assessment, thorough cases in which the biological material was particularly affected. To avoid DNA contamination, bone samples were cleansed chemically, using diluted bleach, and irradiated with UV light for 30 min prior to DNA analysis. Results showed an overall prediction accuracy of 91.6%, 90.4%, and 91.2%, respectively, for iris, hair, and skin colour, at the 0.7 threshold, with only 2 inconclusive results (samples I7 and I10). The application of SNaPshot-SBE technology has the advantage of preserving the DNA template and optimising data availability . However, the requirement to increase the robustness and accuracy of the phenotyping method and thus the development of an increased gene panel will result in an increase in the number of loci amplified in multiplex reactions. In contrast, by doing so, it increases the chances of nonspecific bindings, the occurrence of allelic dropouts, and inevitably the preferential amplification of smaller products . In addition, designing primers for similar target fragment lengths involves similar annealing temperatures, so this step could be extremely limiting . For all these reflections, it might therefore be a future prospect to evaluate a Next Generation Sequencing (NGS) approach . Past works and our experience showed that the assay performs successfully in possibly degraded DNA from bodies remains of various age and storage conditions . However, also under such a design, the possibility of allelic dropouts and drop-ins, which have been described as typical phenomena associated with the analysis of low template DNA samples, cannot be eliminated completely . This is to be taken into account, although the cases here analysed were in a much better state of preservation, with PMI being much more recent than the findings in the literature. The incorrect calls that occur by applying the method used in this article may derive from the problem described above. However, the difficulties in the phenotypic determination of those chromatic characters that can vary over time must also be taken into consideration (e.g., a subject with blond hair that turns brown during growth). The SNaPshot test, as applied here, shows several limitations, among which the most important is related to the limited number of DNA variants that can be evaluated for each test. This implies carrying out several tests with the relative consumption of a larger quantity of DNA. This limitation can be overcome by using massively parallel sequencing (MPS), which has already been applied in other studies with good results . While many purport the usefulness of FDP in this regard, its probabilistic nature as well as its ability to disclose information about an individual that may be considered private raise a range of ethical and social concerns . Currently, there are only three European countries that have adopted specific laws in favour of the FDP (the Netherlands, Germany, and Slovakia), although all have limitations in the application of the method. However, in most European countries, even in the absence of specific rules, FDP can be practiced according to the direction of general laws. In the United States, there is no federal law that regulates phenotyping, as there is extreme regulatory variability between states . One dilemma about the FDP is the judicial applicability of the assay and its application in criminal and civil matters. Indeed, FDP estimates of individual appearance are probabilistic and not deterministic: FDP can create a most likely appearance but cannot generate a perfect likeness of the person. The penal legislative system is based on probabilities close to certainty, unlike civil law, which is based on a probabilistic preponderance. This opens up doubts about the possibility of using the FDP in any context. Actually, it is not necessary to admit FDP in court because once the investigative phase (which can certainly benefit from the FDP) is over, it is quite simple to conduct a standard DNA genotyping and report a match with stronger probabilities. In this way, FDP must only be used for investigative purposes and should never be used as evidence at trial. Genetic diseases are clearly excluded from the phenotyping technique starting with DNA, as it is believed that the use of this data in a forensic context would violate privacy in an unacceptable way . 5. Conclusions The results demonstrate the reliability of the HIrisPlex-S genotyping system also in highly degraded DNA samples isolated from bone samples. These findings may encourage the application of the multiplex for the prediction of eye, hair, and skin traits from DNA in forensic cases, particularly in cases where no other identification method can be applied or where chromatic information can support an anthropological investigation. In fact, the FDP shows a very broad spectrum of applications in the forensic field, both in terms of identification based on traces and cadaveric identification in multiple fields (e.g., missing persons, human trafficking, and mass disasters). In the future, the implementation of DNA markers related to skin chromatism and their addition to the multiplex system would prove fundamental to improving the prediction accuracies of the "pale" and "intermediate" categories, which with the described model are detected with a much lower accuracy compared to the "dark" and "dark black" skin colour categories . In addition, the future application of this method to a wider case study, even forensic, will allow for the identification of new SNPs for these traits and other visible traits, as well as the development of new prediction models, as already suggested . Definitely, DNA prediction of EVC will become more widely used in genetic studies of human remains in evolutionary, anthropological, and forensic investigations. This preliminary study aims to widen the case studies by combining the cases already selected and further cases of cadavers with different PMI and storage conditions in order to increase the experience in the field. A stronger awareness of the possibilities of this method could lead to its systematic use in forensic cases dealing with unidentifiable corpses, either through the classic genetic method or as a support for classical anthropological analysis. Author Contributions Conceptualization, L.A. and M.F.; investigation, L.A., M.F. and L.M.; methodology, L.M.; project administration, M.N.; supervision, R.M.G.; validation, P.F. and M.N.; visualization, R.M.G.; writing--original draft, L.A., M.F. and L.M.; writing--review and editing, P.F. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement All the data used for the article are in the availability of the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Illustration of the workflow for multiplex PCR phenotyping. Outlining of the main work steps: (a) sampling, (b) gDNA extraction, (c) multiplex PCR, purification of amplicons with exonuclease, SBE reactions, enzymatic purification with SAP, capillary electrophoresis, (d) integration of genetic data in the on-line table to obtain phenotypic results, and finally, comparison of these with the photograms of the individuals. Figure 2 HIrisPlex-S SNP electropherogram obtained from sample I1. The electropherogram shows the amplified SNPs in the same order as those reported in Figure 3. Peaks show the different DNA bases, pointed using the following colour code: green (adenine), black (cytosine), blue (guanine), and red (timine). Figure 3 Phenotype and trait prediction shown by samples I7, shown on the left, and I10, shown on the right. Charts taken from entering data in the tool available on the website (last accessed on 17 February 2023). In the figure, "0" indicates if the input allele is not present; "1" indicates the presence of heterozygosity; and "2" shows the presence of the homozygous input allele; "NA" shows missing SNP. healthcare-11-00647-t001_Table 1 Table 1 Bone samples tested in the present work. Identified Corpses Bone Sample Corpse Preservation I1 Femur Skeletonized and burnt I2 Femur Highly decomposed I3 Femur Burnt I4 Tibia Burnt I5 Tibia Burnt I6 Femur Burnt I7 Femur Skeletonized I8 Tibia Highly decomposed I9 Tibia Highly decomposed I10 Tibia Skeletonized I11 Femur Highly decomposed I12 Femur Highly decomposed I13 Femur Skeletonized I14 Femur Highly decomposed I15 Femur Highly decomposed I16 Femur Highly decomposed I17 Tibia Highly decomposed I18 Femur Highly decomposed I19 Femur Skeletonized I20 Femur Skeletonized Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Disaster Victim Identification (DVI) Available online: (accessed on 29 June 2021) 2. Cattaneo C. De Angelis D. Porta D. Grandi M. Personal Identification of Cadavers and Human Remains Forensic Anthropology and Medicine Schmitt A. Cunha E. Pinheiro J. Humana Press Totowa, NJ, USA 2006 359 379 10.1007/978-1-59745-099-7_15 3. Kahana T. Grande A. Tancredi D.M. Penalver J. Hiss J. Fingerprinting the deceased: Traditional and new techniques J. Forensic. Sci. 2001 46 908 912 10.1520/JFS15067J 11451076 4. Jahagirdar B.P. Anand M. Vidhii S. Role of forensic odontologist in post mortem person identification Dent. Res. J. 2012 9 522 530 10.4103/1735-3327.104868 5. Butler J.M. Forensic DNA Typing: Biology, Technology and Genetics of STR Markers 2nd ed. Elsevier Academic Press Burlington, MA, USA 2005 541 560 6. Eglinton G. Logan G.A. Molecular preservation Philos. Trans. R. Soc. Lond B Biol. Sci. 1991 333 315 328 315-327; discussion 327-328 10.1098/rstb.1991.0081 1684047 7. Archer M.S. Ranson D.L. 3. Potential Contamination of Forensic Entomology Samples Collected in the Mortuary Med. Sci. Law 2005 45 89 91 10.1258/rsmmsl.45.1.89 15745280 8. Dirkmaat D.C. Cabo L.L. Forensic Archaeology and Forensic Taphonomy: Basic Considerations on how to Properly Process and Interpret the Outdoor Forensic Scene Acad. Forensic. Pathol. 2016 6 439 454 10.23907/2016.045 31239919 9. Rizzi E. Lari M. Gigli E. De Bellis G. Caramelli D. Ancient DNA studies: New perspectives on old samples Genet Sel Evol. 2012 44 21 10.1186/1297-9686-44-21 22697611 10. Marquez-Grant N. An overview of age estimation in forensic anthropology: Perspectives and practical considerations Ann. Hum. Biol. 2015 42 308 322 10.3109/03014460.2015.1048288 26366990 11. Adserias-Garriga J. Thomas C. Ubelaker D.H. Zapico S.C. When forensic odontology met biochemistry: Multidisciplinary approach in forensic human identification Arch. Oral. Biol. 2018 87 7 14 10.1016/j.archoralbio.2017.12.001 29241027 12. De Angelis D. Cattaneo C. Implant bone integration importance in forensic identification J. Forensic. Sci. 2015 60 505 508 10.1111/1556-4029.12640 25387697 13. Kate Brookes G. Thompson T. The impact of personal perception on the identification of tattoo pattern in human identification J. Forensic. Leg. Med. 2019 64 34 41 10.1016/j.jflm.2019.03.002 30959313 14. Montelius K. Lindblom B. DNA analysis in disaster victim identification Forensic. Sci. Med. Pathol. 2012 8 140 147 10.1007/s12024-011-9276-z 22009165 15. Turingan R.S. Brown J. Kaplun L. Smith J. Watson J. Boyd D.A. Steadman D.W. Selden R.F. Identification of human remains using Rapid DNA analysis Int. J. Legal. Med. 2020 134 863 872 10.1007/s00414-019-02186-y 31781850 16. Alvarez-Cubero M.J. Saiz M. Martinez-Gonzalez L.J. Alvarez J.C. Eisenberg A.J. Budowle B. Lorente J.A. Genetic identification of missing persons: DNA analysis of human remains and compromised samples Pathobiology 2012 79 228 238 10.1159/000334982 22722562 17. Alonso A. Andelinovic S. Martin P. Sutlovic D. Erceg I. Huffine E. de Simon L.F. Albarran C. Definis-Gojanovic M. Fernandez-Rodriguez A. DNA typing from skeletal remains: Evaluation of multiplex and megaplex STR systems on DNA isolated from bone and teeth samples Croat Med. J. 2001 42 260 266 11387635 18. Piccinini A. Betti F. Capra M. Cattaneo C. Comino A. L'identificazione delle vittime del disastro aereo di Linate (8 ottobre 2001) Lo stato dell'arte in Genetica forense a cura di De Leo, Turrina, Orrico Giuffre Editore Milano, Italy 2003 289 302 19. Alonso A. Martin P. Albarran C. Garcia P. Fernandez de Simon L. Jesus Iturralde M. Capilla J. Garcia-Hirschfeld J. Martinez P. Vallejo G. Challenges of DNA Profiling in Mass Disaster Investigations Croat. Med. J. 2005 46 540 548 16100756 20. Hofreiter M. Sneberger J. Pospisek M. Vanek D. Progress in forensic bone DNA analysis: Lessons learned from ancient DNA Forensic. Sci. Int. Genet. 2021 54 102538 10.1016/j.fsigen.2021.102538 34265517 21. Butler J.M. Recent advances in forensic biology and forensic DNA typing: INTERPOL review 2019-2022 Forensic. Sci. Int. Synerg. 2022 6 100311 10.1016/j.fsisyn.2022.100311 36618991 22. Syndercombe Court D. Mitochondrial DNA in forensic use Emerg. Top Life Sci. 2021 5 415 426 10.1042/ETLS20210204 34374411 23. Andreasson H. Nilsson M. Budowle B. Lundberg H. Allen M. Nuclear and mitochondrial DNA quantification of various forensic materials Forensic. Sci. Int. 2006 164 56 64 10.1016/j.forsciint.2005.11.024 16427750 24. Amorim A. Fernandes T. Taveira N. Mitochondrial DNA in human identification: A review PeerJ 2019 7 e7314 10.7717/peerj.7314 31428537 25. Holland M.M. Fisher D.L. Mitchell L.G. Rodriquez W.C. Canik J.J. Merril C.R. Weedn V.W. Mitochondrial DNA sequence analysis of human skeletal remains: Identification of remains from the Vietnam War J. Forensic. Sci. 1993 38 542 553 10.1520/JFS13439J 8515208 26. Christensen A.M. Passalacqua N.V. Bartelink E.J. Chapter 5--Forensic Taphonomy, Forensic Anthropology Academic Press Cambridge, MA, USA 2014 119 147 9780124186712 27. Alaeddini R. Walsh S.J. Abbas A. Forensic implications of genetic analyses from degraded DNA--a review Forensic. Sci. Int. Genet. 2010 4 148 157 10.1016/j.fsigen.2009.09.007 20215026 28. Grela M. Jakubczak A. Kowalczyk M. Listos P. Gryzinska M. Effectiveness of various methods of DNA isolation from bones and teeth of animals exposed to high temperature J. Forensic. Leg. Med. 2021 78 102131 10.1016/j.jflm.2021.102131 33561692 29. Prinz M. Carracedo A. Mayr W.R. Morling N. Parsons T.J. Sajantila A. Scheithauer R. Schmitter H. Schneider P.M. International Society for Forensic Genetics. DNA Commission of the International Society for Forensic Genetics (ISFG): Recommendations regarding the role of forensic genetics for disaster victim identification (DVI) Forensic. Sci. Int. Genet. 2007 1 3 12 10.1016/j.fsigen.2006.10.003 19083722 30. Yagasaki K. Mabuchi A. Higashino T. Hao Wong J. Nishida N. Fujimoto A. Tokunaga K. Practical forensic use of kinship determination using high-density SNP profiling based on a microarray platform, focusing on low-quantity DNA Forensic. Sci. Int. Genet. 2022 61 102752 10.1016/j.fsigen.2022.102752 35987117 31. Urtiaga G.O. Domingues W.B. Komninou E.R. Martins A.W.S. Blodorn E.B. Dellagostin E.N. Woloski R.D.S. Pinto L.S. Brum C.B. Tovo-Rodrigues L. DNA microarray for forensic intelligence purposes: High-density SNP profiles obtained directly from casework-like samples with and without a DNA purification step Forensic. Sci. Int. 2022 332 111181 10.1016/j.forsciint.2022.111181 35042181 32. de Vries J.H. Kling D. Vidaki A. Arp P. Kalamara V. Verbiest M.M.P.J. Piniewska-Rog D. Parsons T.J. Uitterlinden A.G. Kayser M. Impact of SNP microarray analysis of compromised DNA on kinship classification success in the context of investigative genetic genealogy Forensic. Sci. Int. Genet. 2022 56 102625 10.1016/j.fsigen.2021.102625 34753062 33. Kayser M. Forensic DNA Phenotyping: Predicting human appearance from crime scene material for investigative purposes For. Sci. Int. Genet. 2015 18 33 48 10.1016/j.fsigen.2015.02.003 34. Wray N.R. Yang J. Hayes B.J. Price A.L. Goddard M.E. Visscher P.M. Pitfalls of predicting complex traits from SNPs Nat. Rev. Genet. 2013 14 507 515 10.1038/nrg3457 23774735 35. Walsh S. Liu F. Wollstein A. Kovatsi L. Ralf A. Kosiniak-Kamysz A. Branicki W. Kayser M. The HIrisPlex system for simultaneous prediction of hair and eye colour from DNA Forensic. Sci. Int. Genet. 2013 7 98 115 10.1016/j.fsigen.2012.07.005 22917817 36. Chaitanya L. Breslin K. Zuniga S. Wirken L. Pospiech E. Kukla-Bartoszek M. Sijen T. Knijff P. Liu F. Branicki W. The HIrisPlex-S system for eye, hair and skin colour prediction from DNA: Introduction and forensic developmental validation Forensic. Sci. Int. Genet. 2018 35 123 135 10.1016/j.fsigen.2018.04.004 29753263 37. Draus-Barini J. Walsh S. Pospiech E. Kupiec T. Glab H. Branicki W. Kayser M. Bona fide colour: DNA prediction of human eye and hair colour from ancient and contemporary skeletal remains Investig. Genet. 2013 4 3 10.1186/2041-2223-4-3 38. King T.E. Fortes G.G. Balaresque P. Thomas M.G. Balding D. Maisano Delser P. Neumann R. Parson W. Knapp M. Walsh S. Identification of the remains of King Richard III Nat. Commun. 2014 5 5631 10.1038/ncomms6631 25463651 39. Walsh S. Liu F. Ballantyne K.N. van Oven M. Lao O. Kayser M. IrisPlex: A sensitive DNA tool for accurate prediction of blue and brown eye colour in the absence of ancestry information Forensic. Sci. Int. Genet. 2011 5 170 180 10.1016/j.fsigen.2010.02.004 20457092 40. Nimbkar P.H. Bhatt V.D. A review on touch DNA collection, extraction, amplification, analysis and determination of phenotype Forensic. Sci. Int. 2022 336 111352 10.1016/j.forsciint.2022.111352 35660243 41. Sanchez J.J. Endicott P. Developing multiplexed SNP assays with special reference to degraded DNA templates Nat. Protoc. 2006 1 1370 1378 10.1038/nprot.2006.247 17406424 42. Daniel R. Santos C. Phillips C. Fondevila M. van Oorschot R.A. Carracedo A. Lareu M.V. McNevin D. A SNaPshot of next generation sequencing for forensic SNP analysis Forensic. Sci. Int. Genet. 2015 14 50 60 10.1016/j.fsigen.2014.08.013 25282603 43. Walsh S. Lindenbergh A. Zuniga S.B. Sijen T. de Knijff P. Kayser M. Ballantyne K.N. Developmental validation of the IrisPlex system: Determination of blue and brown iris colour for forensic intelligence Forensic. Sci. Int. Genet. 2011 5 464 471 10.1016/j.fsigen.2010.09.008 20947461 44. Pastinen T. Kurg A. Metspalu A. Peltonen L. Syvanen A.C. Minisequencing: A specific tool for DNA analysis and diagnostics on oligonucleotide arrays Genome Res. 1997 7 606 614 10.1101/gr.7.6.606 9199933 45. Walsh S. Chaitanya L. Clarisse L. Wirken L. Draus Barini J. Kovatsi L. Maeda H. Ishikawa T. Sijen T. de Knijff P. Developmental validation of the HIrisPlex system: DNA-based eye and hair colour prediction for forensic and anthropological usage Forensic. Sci. Int. Genet 2014 9 150 161 10.1016/j.fsigen.2013.12.006 24528593 46. Walsh S. Chaitanya L. Breslin K. Muralidharan C. Bronikowska A. Pospiech E. Koller J. Kovatsi L. Wollstein A. Branicki W. Global skin colour prediction from DNA Hum. Genet. 2017 136 847 863 10.1007/s00439-017-1808-5 28500464 47. Fortes G.G. Speller C.F. Hofreiter M. King T.E. Phenotypes from ancient DNA: Approaches, insights and prospects Bioessays 2013 35 690 695 10.1002/bies.201300036 23703035 48. Pruvost M. Reissmann M. Benecke N. Ludwig A. From genes to phenotypes--Evaluation of two methods for the SNP analysis in archaeological remains: Pyrosequencing and competitive allele specific PCR (KASPar) Ann. Anat. 2012 194 74 81 10.1016/j.aanat.2011.10.007 22154270 49. Foley M.M. Oldoni F. A global snapshot of current opinions of next-generation sequencing technologies usage in forensics Forensic. Sci. Int. Genet. 2022 63 102819 10.1016/j.fsigen.2022.102819 36509023 50. Kukla-Bartoszek M. Szargut M. Pospiech E. Diepenbroek M. Zielinska G. Jarosz A. Piniewska-Rog D. Arciszewska J. Cytacka S. Spolnicka M. The challenge of predicting human pigmentation traits in degraded bone samples with the MPS-based HIrisPlex-S system Forensic. Sci. Int. Genet. 2020 47 102301 10.1016/j.fsigen.2020.102301 32387914 51. Ragazzo M. Puleri G. Errichiello V. Manzo L. Luzzi L. Potenza S. Strafella C. Peconi C. Nicastro F. Caputo V. Evaluation of OpenArrayTM as a Genotyping Method for Forensic DNA Phenotyping and Human Identification Genes 2021 12 221 10.3390/genes12020221 33546406 52. Zupanic Pajnic I. Zupanc T. Leskovar T. Cresnar M. Fattorini P. Eye and Hair Color Prediction of Ancient and Second World War Skeletal Remains Using a Forensic PCR-MPS Approach Genes 2022 13 1432 10.3390/genes13081432 36011343 53. Haned H. Egeland T. Pontier D. Pene L. Gill P. Estimating drop-out probabilities in forensic DNA samples: A simulation approach to evaluate different models Forensic. Sci. Int. Genet. 2011 5 525 531 10.1016/j.fsigen.2010.12.002 21216685 54. Melchionda F. Silvestrini B. Robino C. Bini C. Fattorini P. Martinez-Labarga C. De Angelis F. Tagliabracci A. Turchi C. Development and Validation of MPS-Based System for Human Appearance Prediction in Challenging Forensic Samples Genes 2022 13 1688 10.3390/genes13101688 36292573 55. Breslin K. Wills B. Ralf A. Ventayol Garcia M. Kukla Bartoszek M. Pospiech E. Freire-Aradas A. Xavier C. Ingold S. de La Puente M. HIrisPlex-S system for eye, hair, and skin color prediction from DNA: Massively parallel sequencing solutions for two common forensically used platforms Forensic. Sci. Int. Genet 2019 43 102152 10.1016/j.fsigen.2019.102152 31518964 56. Samuel G. Prainsack B. Civil society stakeholder views on forensic DNA phenotyping: Balancing risks and benefits Forensic. Sci. Int. Genet. 2019 43 102157 10.1016/j.fsigen.2019.102157 31518963 57. Samuel G. Prainsack B. The Regulatory Landscape of Forensic DNA Phenotyping in Europe 2018 Available online: (accessed on 17 February 2023) 58. Zieger M. Forensic DNA phenotyping in Europe: How far may it go? J. Law Biosci. 2022 9 lsac024 10.1093/jlb/lsac024 36120387 59. Westermark H. Aronovitz A. Curran J. Fausch I. Fournier J. Hohenecker L. Kleczewski A. Pretelli I. Polanco Lazo R. Topaz Druckman K. The Regulation of the Use of DNA in Law Enforcement, Current to: 28.08.2020, E-AVIS ISDC 2020-02, Swiss Institute of Comparative Law 1, 9 2020 Available online: (accessed on 17 February 2023) 60. Schneider P.M. Prainsack B. Kayser M. The Use of Forensic DNA Phenotyping in Predicting Appearance and Biogeographic Ancestry Dtsch Arztebl. Int. 2019 51-52 873 880 10.3238/arztebl.2019.0873 31941575 |
PMC10000574 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050980 diagnostics-13-00980 Review Recent Advances in Ultrasound Breast Imaging: From Industry to Clinical Practice Catalano Orlando 1 Fusco Roberta 2* De Muzio Federica 3 Simonetti Igino 4 Palumbo Pierpaolo 56 Bruno Federico 56 Borgheresi Alessandra 78 Agostini Andrea 78 Gabelloni Michela 9 Varelli Carlo 1 Barile Antonio 10 Giovagnoni Andrea 78 Gandolfo Nicoletta 11 Miele Vittorio 612 Granata Vincenza 4 Ekpo Ernest Usang Academic Editor 1 Department of Radiology, Istituto Diagnostico Varelli, 80126 Naples, Italy 2 Medical Oncology Division, Igea SpA, 80013 Naples, Italy 3 Department of Medicine and Health Sciences "V. Tiberio", University of Molise, 86100 Campobasso, Italy 4 Division of Radiology, "Istituto Nazionale Tumori IRCCS Fondazione Pascale-IRCCS di Napoli", 80131 Naples, Italy 5 Department of Diagnostic Imaging, Area of Cardiovascular and Interventional Imaging, Abruzzo Health Unit 1, 67100 L'Aquila, Italy 6 Italian Society of Medical and Interventional Radiology (SIRM), SIRM Foundation, 20122 Milan, Italy 7 Department of Clinical, Special and Dental Sciences, University Politecnica delle Marche, 60126 Ancona, Italy 8 Department of Radiology, University Hospital "Azienda Ospedaliera Universitaria delle Marche", 60126 Ancona, Italy 9 Department of Translational Research, Diagnostic and Interventional Radiology, University of Pisa, 56126 Pisa, Italy 10 Department of Applied Clinical Sciences and Biotechnology, University of L'Aquila, 67100 L'Aquila, Italy 11 Diagnostic Imaging Department, Villa Scassi Hospital-ASL 3, Corso Scassi 1, 16149 Genoa, Italy 12 Department of Emergency Radiology, Careggi University Hospital, 50134 Florence, Italy * Correspondence: [email protected] 04 3 2023 3 2023 13 5 98029 12 2022 02 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Breast ultrasound (US) has undergone dramatic technological improvement through recent decades, moving from a low spatial resolution, grayscale-limited technique to a highly performing, multiparametric modality. In this review, we first focus on the spectrum of technical tools that have become commercially available, including new microvasculature imaging modalities, high-frequency transducers, extended field-of-view scanning, elastography, contrast-enhanced US, MicroPure, 3D US, automated US, S-Detect, nomograms, images fusion, and virtual navigation. In the subsequent section, we discuss the broadened current application of US in breast clinical scenarios, distinguishing among primary US, complementary US, and second-look US. Finally, we mention the still ongoing limitations and the challenging aspects of breast US. ultrasound breast imaging clinical practice This research received no external funding. pmc1. Introduction--Where Do We Come from? In 1977, Dr. Dodd wrote: "The spatial resolution presently obtainable in ultrasonograms is inadequate for the detection of subclinical [breast] cancer" . Later still in 1983, Sickles and co-workers claimed: "These data indicate that sonography is not an acceptable substitute for mammography in the detection and diagnosis of breast cancer" . For many years, the main use of breast ultrasound (US) was for differentiating cystic lesions from solid lesions . Consequently, US has become an integral component of the diagnostic work-up of breast abnormalities only in the last two decades . The seminal article on using US for differential diagnosis was only published in 1995 , and the first BI-RADS atlas addressing the issue of US, in addition to mammography, was only published in 2003 . The dramatic increase in spatial resolution is the main explanation for the advancing role of US. The application of tissue harmonics, spatial compound imaging, and speckle reduction techniques further refined the images . Appropriate scanner settings and optimized scanning methodologies are also mandatory for an up-to-date US examination of the breast. Technological advances now allow a comprehensive US diagnosis, management, and treatment of breast abnormalities. In this review, we discuss the technical revolution that happened over the years and point out the consequent clinical revolution that has occurred in daily practice. 2. Technological Developments--What the Industry Has Made Available to Us 2.1. Conventional Doppler Techniques and New Microvasculature Imaging Techniques Cancer growth is based on neoangiogenesis, i.e., the tumor-induced development of a vascular network. Detecting these flow signals and assessing their characteristics in terms of number, distribution, and appearance are consequently of paramount importance in tumor characterization and monitorization . Even if appropriately set to identify small, low-flow vessels, conventional Doppler techniques, including color Doppler and power Doppler, have a limited sensitivity. In recent years, almost all companies have developed filtered techniques capable of working at a higher frame rate and consequently detecting tiny peri-tumoral flow signals. With these new software facilities, the background and tissue motion artefacts are suppressed and US sensitivity, spatial resolution, and temporal resolution are significantly improved . Some companies have also developed systems capable of quantifying the number of colored pixels within the box, thus quantifying the flow intensity . If a new microvascular tool is available on the scanner, we encourage users to refrain from still using conventional Doppler modalities for breast imaging and to employ more advanced techniques . 2.2. High-Frequency Transducers The transmission frequencies employed to scan breasts have increased through the years, moving from 7.5 MHz initially to 10-12 MHz and currently to 13-18 MHz. High-frequency transducers provide increased axial and soft tissue resolution, permitting improved differentiation of subtle shades of gray, margin resolution, and lesion conspicuity in the background of normal breast parenchyma . Current transducers have a broad band of frequencies, allowing the operator to choose the most appropriate one in relation to the size of the given breast and based on the depth of the area of interest within it. However, most US companies now sell probes reaching frequencies up to 22-24 MHz. These transducers have been developed to study the skin, but there are a number of circumstances where they can be adopted to investigate the breast. These include skin abnormalities of the breast and axilla, nipple-areolar complex abnormalities, very small-size breasts, superficial areas in any size breasts, prepuberal breasts, male breasts, breast parenchyma abnormalities in subjects with implants, post-mastectomy chest wall, and intraoperative breast sonography . 2.3. Extended Field-of-View Scanning US has a limited field of view (FOV), not allowing to display, in a single scan, a breast area larger than the probe footprint itself . Extended FOV systems allow to partially encompass this limitation. Starting from a real-time translational movement of the probe over the skin, the software can continuously compare the position and create a large 2D image, without any loss in terms of spatial resolution. Extended FOV scans allow to display and measure large breast lesions. Additionally, it becomes possible to show the spatial relationship and the distance between a lesion and some anatomic landmarks, such as the nipple, or between multiple lesions, as for multifocal/multicentric cancer . Images of the whole implant can be obtained, being particularly useful for plastic surgeons. 2.4. Elastography US elastography measures small tissue motions due to pressure forces, i.e., the viscoelastic properties of breast abnormalities . Current techniques include strain elastography, based on the manual pressure from the operator; acoustic radiation force impulse imaging (ARFI); point shear-wave elastography; and 2D/3D shear-wave elastography, based on the changes provoked from the focused ultrasounds themselves. Based on the different techniques, qualitative (subjective scoring), semi-quantitative (strain-to-fat ratio), and/or quantitative data can be obtained . The use of elastography has been mostly focused on breast nodule characterization, but aspects such as tumor detection, tumor extent assessment, axillary lymph node status, percutaneous procedure guidance, and tumor response to treatment assessment must also be considered . As a general rule, stiffer nodules (i.e., score 1 or 2, low strain ratio, low shear-wave speed), opposing a significant resistance to the changes, are thought to be malignant, while elastic ones (i.e., score 4 or 5, high strain ratio, high shear-wave speed) are usually categorized as benign . Breast nodules may, however, show an intermediate behavior (score 3), with overlapping findings and possible false positives (fibrotic and calcified fibroadenomas, sclerosing adenosis, radial scar, steatonecrosis, etc.) and false negatives (e.g., lymphomas, in situ ductal carcinomas, and small or low-grade or tubular invasive carcinomas) . Elastography may increase US specificity, although this may come at the cost of decreased sensitivity. Consequently, images' interpretation should never rely on elastography alone and should always be correlated with the grayscale appearance . 2.5. Contrast-Enhanced Ultrasound As for Doppler vascular signal, the intensity of contrast medium enhancement represents an indirect indicator of tumor microvascular density, correlating with a malignant nature and with the cancer grade. First introduced to boost the low-sensitivity Doppler systems of the 1990s, microbubble contrast media are now employed during real-time, low mechanical-index, grayscale US. However, despite a high number of publications, contrast-enhanced US (CEUS) has not significantly impacted breast imaging. This is mostly due to the kind of contrast media available until now, which work better at the low frequencies employed to scan the abdomen than at the higher frequencies needed to image superficial structures such as the breast . Malignant breast nodules show a quick and strong but transient enhancement, with variable degrees of heterogeneity . Irregular, tortuous, radially distributed vessels are seen around and inside tumors . A time-intensity curve can be obtained to quantify the microbubbles' behavior over time. Three-dimensional images may better render the tumor angioarchitecture . CEUS may measure the tumor size better than conventional B-mode US. Additionally, it may assess the breast tumor response to neoadjuvant treatment, showing 87% aggregated sensitivity and 84% aggregated specificity in a recent meta-analysis . Tissue harmonic imaging (THI) and contrast harmonic imaging (CHI) also have value in the detection and characterization of breast tumors. 2.6. Three-Dimensional Ultrasound Both hand-held and automated linear transducers are available for use in high-resolution 3D breast imaging. Three-dimensional US, also called volumetric US, allows for obtaining a surface rendering of normal and abnormal breast structures. With a single pass of the ultrasound beam, a 3D reconstructed image is formed in the coronal, sagittal, and transverse planes, allowing a more accurate assessment of anatomical structures and tumor margins . Coronal images, being parallel to the skin, represent a unique opportunity allowed from 3D US. A vivid representation of breast tumors can be obtained, with a retraction, star-like profile strongly supporting the diagnosis of malignancy . Additionally, lesions can be automatically or semi-automatically delimited on the three planes, and their volumes can be quantified. Nodules' growth and cancer regression during treatment can be objectively evaluated through serial scans. Three-dimensional US can be combined with harmonic imaging, Doppler techniques, and CEUS . Additionally, percutaneous procedures can be advantaged by viewing multiplanar and three-dimensional images. 2.7. MicroPure Grouped microcalcifications represent a very important finding in breast imaging and a significant US limitation, being hard or impossible to detect, particularly if not associated with a nodule . MicroPureTM is a software from Canon (Tokyo, Japan) that allows to highlight microcalcifications . MicroPure combines non-linear imaging with speckle suppression, extracting the calcification from the heterogeneous background. Filtered microcalcifications are shown as bright dots inside a dark blue background overimposed to a grayscale US. 2.8. Automated Breast Ultrasound So-called "manual" or "hand-held" US is limited by operator dependence, non-reproducibility, and the inability to image extensive breast areas and store volumes . To overcome these limitations, an automated breast US (ABUS) system, or automated breast volume US, has been developed. ABUS allows a technologist to simultaneously acquire large volumes of tissue, from the skin to the chest wall, by using a large, specific transducer (15 cm) . These volumes, consisting of a high number of thin scans, are stored to be visualized and reformatted by a doctor at another moment or another place. Coronal views are of special value . By now, ABUS has been considered an adjunct to mammography to screen patients with dense breasts. ABUS has a shown good agreement with manual US in terms of detection rate and BI-RADS categorization . Architectural distortions and peritumoral infiltrations may be better displayed than with manual US, while large masses can be easily measured . A good correlation with MRI in the assessment of tumor response to treatment has been demonstrated . ABUS is less time-consuming than manual US and causes less fatigue to the operator . It should be considered, however, that any probe-mediated palpation finding is missed with ABUS and that the direct interaction between the patient and the physician is completely lost. 2.9. Computer-Assisted Diagnosis--S-Detect Owing to deep learning algorithms, artificial intelligence systems are able to automatically detect and quantify a number of features from US images . This may allow for the simple and reproducible detection and characterization of breast lesions as well as the prediction of the response to treatment in patients with locally advanced breast cancer. Computer-assisted diagnosis (CAD) can be employed as a second reader to improve the accuracy of the operator in US or CEUS imaging of the breast . The software analyzes the targets identified by the operator, showing their shape and their risk of malignancy based on the BI-RADS lexicon or other descriptors. A CAD system works through four successive phases: pre-processing, segmentation, feature extraction and selection, and classification . The analysis can be approved or rejected by the operator . CAD systems can be offline, located in a personal computer, or inserted directly on the scanner. An example of the latter is represented by S-DetectTM, a semi-automatic tool from Samsung (Busan, Republic of Korea) . The operator manually places a marker inside a lesion, and then, the software traces the border (adjustable) and analyzes and classifies the lesion according to the US descriptors from the BI-RADS . In a study on the differential diagnosis of breast lesions, the sensitivity of five operators of different experience levels was >90% and specificity was 50-75%, while S-Detect had 90% sensitivity and 71% specificity . Advanced systems are fully automatic, are based on convolutional neural networks, and have a high capability of recognizing the images. 2.10. Ultrasound Nomograms Radiomics can extract many quantitative features from US images through a computer algorithm . To create a nomogram, first, well-established US features are selected and extracted by the training experts; then a model is built; and finally, the model is tested, possibly using both an internal and an external validation cohort of patients. The quantitative features extracted through computerized algorithms can be employed for the differential diagnosis of breast abnormalities or to predict the prognosis (risk of lymph-node metastasis, risk of high nodal burden, etc.) or establish in advance the response to treatment of breast tumors . 2.11. Images Fusion and Virtual Navigation Fusion imaging consists of merging digital images from two different modalities to improve the overall performance, with special reference to lesions' localization at second-look US and to the percutaneous approach to those abnormalities poorly visible with US. Real-time images of the US system are combined (directly overimpressed or shown synchronized side by side) with those previously uploaded from a previous mammography, CT, MRI, or PET exam. This allows real-time, virtual navigation through the volume. The structures invisible under US but visible with other modalities can be operated using US-guided biopsy navigated by the other modality . The indirect systems employ artificial skin markers placed before the exam, while the direct systems use some anatomical landmarks as spatial references . Algorithms have been developed for assessing organ motion induced by breathing and movement . BreastNav(r) is an interesting system from Esaote (Genoa, Italy) allowing fusion between a prone MR scan and a supine US scan through 2D remodeling applied to the breast from an anatomical landmark established during the MR scans. An electromagnetic sensor is applied to the transducer. This allows a targeted second-look US assessment of findings not determined at MRI. The US fusion volume navigation technique can be used to scan the breast nodules requiring follow-up . Needle-tracking software facilities, based on specific sensors, allow to simulate the needle path during US-guided percutaneous procedures. This tool may increase effectiveness and safety. 3. Changing Clinical Scenarios--The Current Impact of US in Breast Practice 3.1. Primary Ultrasound There are a good number of clinical settings where whole-breast, bilateral, axilla-including US must be regarded as the first imaging modality to be employed, after physical examination. In many circumstances, US will be conclusive to solve the issue, while in others, US results will prompt further investigation. A palpable breast mass or swelling always requires imaging assessment, and US will effectively differentiate cysts (BI-RADS 2), non-suspected nodules (BI-RADS 3), and suspected nodules (BI-RADS 4 and 5) . Dense breasts in young women cannot be imaged efficaciously with mammography. However, more and more women nowadays are asking for an imaging assessment to feel themselves followed and to reduce their anxiety. These subjects can be asymptomatic or may present with symptoms such as breast pain or tenderness. US has become the supplemental screening tool of choice for cancer detection in women with dense breast tissue. MRI is the most effective tool in measuring the primary breast tumor extent, and it allows the detection of additional foci of mammographically and/or sonographically occult disease in women with newly diagnosed cancer. However, US is also useful for preoperative breast cancer staging, with special reference to axillary lymph nodes' status, allowing to perform examination of the whole axilla with level 1-3 lymph nodes. Nipple discharge, a common circumstance of special relevance if bloody or dark, can now be imaged with US. Galactography is no longer employed, and MRI is reserved to selected cases. US works well for newborns, children, and adolescents with breast abnormalities (grading of premature thelarche, etc.). Breast disorders in pregnancy should be studied using US, at least because of radioprotection needs . Lactating breasts are also difficult to image with mammography because of the glandular density, and US is the primary imaging modality in women with abnormal symptoms during lactation . Male breast abnormalities should be studied with US, which can easily differentiate gynecomastia from cancer. Breast assessment in women scheduled for augmentation surgery or for hormonal therapy because of infertility is also based on US, particularly for younger women. Periodic assessment after mastoplasty is carried out with US, while MRI is employed when an implant rupture is suspected . Owing mainly to the real-time characteristic, US is the method of choice to guide diagnostic and therapeutic percutaneous procedures whenever the target is adequately visible with US . US can also be useful during treatment planning for breast radiation therapy . Intraoperative US can guide the surgeon towards a quicker and more effective approach at the operating table, decreasing the incidence of positive margins and the consequent need for re-excision . Though focused on the mammary parenchyma, breast US has the ability to assess all the tissue layers, from superficial to the gland (dermis and hypodermis) and deeper to the gland (retro-glandular fat, chest wall, and pleuro-pulmonary area) . Patients presenting to a US exam for abnormalities arising from these areas, as well as non-glandular incidental findings during breast US, can be readily evaluated. Second-level imaging modalities are employed in selected cases . 3.2. Complementary Ultrasound US is both an adjunct and a complement to mammography . Screening of asymptomatic women is based on mammography. However, younger women are now asking to be imaged for an early diagnosis of breast cancer. In this setting, mammography alone is not enough, and combination with US, or with ABUS, is mandatory . Surveillance of women at high risk, bearing a hereditary/familial risk, is currently performed with contrast-enhanced MRI. However, US can represent a useful adjunct, at least to increase the time interval between each MRI exam. US complements initial mammography and, if needed, MRI in the assessment of local and regional tumor extent. MRI is the gold standard in the assessment of response to treatment both in patients with neoadjuvant therapy for a locally advanced breast cancer and in patients with a metastatic breast cancer. However, being an easily repeatable exam, US can be performed serially and allows multiple measurements of the size changes. US complements annual mammography in the loco-regional follow-up of patients with a history of breast cancer . 3.3. Second-Look Ultrasound US is frequently used as a targeted, second-look option for patients imaged with other breast imaging modalities. Since the beginning, US has been employed to better define any abnormality found with mammography. US is employed to differentiate cystic and solid opacities, to assess the level of suspicion of any nodule, and to assess distortion or any other changes. US is performed immediately after a mammography requiring further work-up or during patient recall. Finally, US is employed in patients with a palpable abnormality and negative mammograms. Contrast-enhanced MRI may require a targeted US scan in the case of enhancing or apparently enhancing focal changes or when an intense background parenchymal enhancement may mimic or masquerade a breast nodule . Lesions detected on MRI are often mammographically occult, but many of them can be detected with targeted US . Lesions found with MRI can be located by US and, consequently, can be histologically clarified by US-guided biopsy . Breast uptakes with a whole-body PET scan can also be further investigated with US to avoid false positive diagnoses. This applies to patients undergoing a PET exam because of breast cancer but also to subjects with non-mammary primary tumors during their staging or follow-up with molecular imaging. Breast nodules are frequently detected during contrast-enhanced or unenhanced CT exams, chest scans, abdominal scans, or whole-body scans . US works well as a quick and simple tool to confirm or rule out a nodule and to establish the need for further investigation or for patient follow-up. 4. Conclusions--Not Everything That Glitters Is Gold Despite the tremendous advances illustrated in this review, it is important to highlight a number of ongoing limitations and pitfalls. A large, fatty breast is still a problem for US scan, and US should never be employed as the only modality to evaluate such a case . Microcalcification must be accurately detected and characterized, and this aspect is still a duty of mammography as US has limited sensitivity, especially when microcalcifications are not located inside a nodule. Moreover, a careful, multiparametric, US exam of the whole breasts and axillary cavities is time-consuming. Both in the case of sonographer-performed US and physician-performed US, this examination requires an adequate amount of time. Breast US should be carried out with top-level scanners, equipped with all the software capabilities allowing to perform a multiparametric investigation. Due to continuous direct contact with the patient and the high emotional involvement of women with any breast-related trouble, mammary US requires special abilities from the operator to concentrate and, at the same time, show good empathy . Additionally, despite many attempts at methodological and lexical standardization and the introduction of ABUS, US is still a subjective exam that depends on the skills of the single operator, with limitations in terms of objectivity and intra-observer and inter-observer reproducibility. Finally, it must be considered that competition from other imaging modalities is strong. As with US, mammography and MRI have also shown significant improvements through the years, and other techniques are being proposed as well . Digital tomosynthesis and contrast-enhanced spectral mammography are quite important options, now employed routinely . Quantitative perfusion MRI, MR lymphography, blood oxygenation level-dependent MRI, and diffusion-weighted MRI have all increased the impact of this modality . However, it should always be kept in mind that all imaging modalities have their points of strength and weakness. Consequently, all modalities must be employed in the most appropriate way to cover the diagnostic needs of each single patient. Acknowledgments The authors are grateful to Alessandra Trocino, librarian at the National Cancer Institute of Naples, Italy. Author Contributions O.C., R.F., F.D.M., I.S., P.P., F.B., A.B. (Alessandra Borgheresi), A.A., M.G., C.V., A.B. (Antonio Barile), A.G., N.G., V.M. and V.G. written and revised the manuscript. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement All data are reported in the manuscript. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Breast invasive ductal carcinoma. Vascularization as assessed with power Doppler (left) and with MV-Flow (right). Recently developed techniques such as MV-Flow are more sensitive to slow flows and offer a better display of tumor vessels. Figure 2 Breast intraductal papilloma. Detailed morpho-structural assessment working with a high-frequency transducer (22 MHz). Power Doppler detection of tumor vessels. Figure 3 Invasive ductal carcinoma of the breast. (A) US scan showing a 4-mm nodule (calipers). (B) Extended FOV scan displaying the nodule (arrow) within the whole breast. Figure 4 Breast fibroadenoma. Strain ratio allows a semi-quantitative assessment of the lesion-to-fat stiffness. Figure 5 Invasive ductal carcinoma of the breast. Coronal 3D display of infiltrating tumor margins. Figure 6 Fibroadenoma in a patient with history of augmentation mammoplasty. MicroPure allows to easily detect a bright dot (arrow) due to a small calcification. Figure 7 S-Detect automatic measurement and categorization of a breast nodule (fibroadenoma) using BI-RADS descriptors. Figure 8 Sub-muscular breast implant as displayed using an extended FOV acquisition. Figure 9 Ingrown areolar hair causing abscess formation in a young male. (A) Clinical photograph. (B) Power Doppler US imaging. The 6-mm linear echoic hair and the surrounding strong hyperemia are readily recognizable, allowing a confident differential diagnosis with Montgomery glands inflammation. Figure 10 Palpable mass in a woman with history of augmentation mammoplasty. (A) Mammography oblique digital view was negative in this case, despite use of tomosynthesis. (B) US could detect a nodule immediately above the implant. The tumor proved to be a triple-negative invasive ductal carcinoma at surgery. Figure 11 Pure tubular carcinoma detected incidentally during whole-body CT in a female patient with rectal cancer. (A) Contrast-enhanced, venous-phase CT scan detecting a nodule within the right breast (arrow). (B) Targeted US scan displaying the malignant lesion. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Dodd G.D. Present status of thermography, ultrasound and mammography in breast cancer detection Cancer 1977 39 2796 2805 10.1002/1097-0142(197706)39:6<2796::AID-CNCR2820390667>3.0.CO;2-0 872067 2. Sickles E.A. Filly R.A. Callen P.W. Breast cancer detection with sonography and mammography: Comparison using state-of-the-art equipment Am. J. Roentgenol. 1983 140 843 845 10.2214/ajr.140.5.843 6601422 3. Dempsey P.J. The History of Breast Ultrasound J. Ultrasound Med. 2004 23 887 894 10.7863/jum.2004.23.7.887 15292555 4. Agarwal M. van der Pol C.B. Patlas M.N. Udare A. Chung A.D. Rubino J. Optimizing the radiologist work environment: Actionable tips to improve workplace satisfaction, efficiency, and minimize burnout Radiol. Med. 2021 126 1255 1257 10.1007/s11547-021-01397-x 34269963 5. Albano D. Stecco A. Micci G. Sconfienza L.M. Colagrande S. Reginelli A. Grassi R. Carriero A. Midiri M. Lagalla R. Whole-body magnetic resonance imaging (WB-MRI) in oncology: An Italian survey Radiol. Med. 2020 126 299 305 10.1007/s11547-020-01242-7 32572763 6. Ahmed S.A. Samy M. Ali A.M. Hassan R.A. Architectural distortion outcome: Digital breast tomosynthesis-detected versus digital mammography-detected Radiol. Med. 2021 127 30 38 10.1007/s11547-021-01419-8 34665431 7. Argalia G. Ventura C. Tosi N. Campioni D. Tagliati C. Tufillaro M. Cucco M. Svegliati Baroni G. Giovagnoni A. Comparison of point shear wave elastography and transient elastography in the evaluation of patients with NAFLD Radiol. Med. 2022 127 571 576 10.1007/s11547-022-01475-8 35290569 8. Argalia G. Tarantino G. Ventura C. Campioni D. Tagliati C. Guardati P. Kostandini A. Marzioni M. Giuseppetti G.M. Giovagnoni A. Shear wave elastography and transient elastography in HCV patients after direct-acting antivirals Radiol. Med. 2021 126 894 899 10.1007/s11547-020-01326-4 33492651 9. Grani G. Lamartina L. Biffoni M. Giacomelli L. Maranghi M. Falcone R. Ramundo V. Cantisani V. Filetti S. Durante C. Sonographically Estimated Risks of Malignancy for Thyroid Nodules Computed with Five Standard Classification Systems: Changes over Time and Their Relation to Malignancy Thyroid 2018 28 1190 1197 10.1089/thy.2018.0178 30105949 10. American College of Radiology ACR-BI-RADS-Ultrasound ACR Breast Imaging Reporting and Data System, Breast Imaging Atlas American College of Radiology Reston, VA, USA 2003 11. Hooley R.J. Scoutt L.M. Philpotts L.E. Breast ultrasonography: State of the art Radiology 2013 268 642 659 10.1148/radiol.13121606 23970509 12. Brizi M.G. Perillo F. Cannone F. Tuzza L. Manfredi R. The role of imaging in acute pancreatitis Radiol. Med. 2021 126 1017 1029 10.1007/s11547-021-01359-3 33982269 13. Catalano O. Wortsman X. Dermatology ultrasound. imaging technique, tips and tricks, high-resolution anatomy Ultrasound Q. 2020 36 321 327 10.1097/RUQ.0000000000000520 33298770 14. Catalano O. Varelli C. Sbordone C. Corvino A. De Rosa D. Vallone G. Wortsman X. A bump: What to do next? ultrasound imaging of superficial soft-tissue palpable lesions J. Ultrasound 2020 23 287 300 10.1007/s40477-019-00415-z 31786796 15. Catalano O. Mattace Raso M. D'Aiuto M. Illiano L.A. Saturnino P.P. Siani A. Additional role of colour Doppler ultrasound imaging in intracystic breast tumours Radiol. Med. 2009 114 253 266 10.1007/s11547-008-0346-6 19082789 16. Chang R.F. Huang S.F. Moon W.K. Lee Y.H. Chen D.R. Solid Breast Masses: Neural Network Analysis of Vascular Features at Three-dimensional Power Doppler US for Benign or Malignant Classification Radiology 2007 243 56 62 10.1148/radiol.2431060041 17312276 17. Chiti G. Grazzini G. Flammia F. Matteuzzi B. Tortoli P. Bettarini S. Pasqualini E. Granata V. Busoni S. Messserini L. Gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs): A radiomic model to predict tumor grade Radiol. Med. 2022 127 928 938 10.1007/s11547-022-01529-x 35917099 18. Di Serafino M. Vallone G. The role of point of care ultrasound in radiology department: Update and prospective. A statement of Italian college ultrasound Radiol. Med. 2021 126 636 641 10.1007/s11547-020-01301-z 33146845 19. Horvath E. Cuitino M.J. Pinochet M.A. Sanhueza P. Color Doppler in the study of the breast: How do we perform it? Rev. Chil. Radiol. 2011 17 19 27 20. Granata V. Simonetti I. Fusco R. Setola S.V. Izzo F. Scarpato L. Vanella V. Festino L. Simeone E. Ascierto P.A. Management of cutaneous melanoma: Radiologists challenging and risk assessment Radiol. Med. 2022 127 899 911 10.1007/s11547-022-01522-4 35834109 21. Granata V. Faggioni L. Grassi R. Fusco R. Reginelli A. Rega D. Maggialetti N. Buccicardi D. Frittoli B. Rengo M. Structured reporting of computed tomography in the staging of colon cancer: A Delphi consensus proposal Radiol. Med. 2022 127 21 29 10.1007/s11547-021-01418-9 34741722 22. Granata V. Grassi R. Fusco R. Setola S.V. Belli A. Ottaiano A. Nasti G. La Porta M. Danti G. Cappabianca S. Intrahepatic cholangiocarcinoma and its differential diagnosis at MRI: How radiologist should assess MR features Radiol. Med. 2021 126 1584 1600 10.1007/s11547-021-01428-7 34843029 23. Cappabianca S. Granata V. Di Grezia G. Mandato Y. Reginelli A. Di Mizio V. Grassi R. Rotondo A. The role of nasoenteric intubation in the MR study of patients with Crohn's disease: Our experience and literature review Radiol. Med. 2010 116 389 406 English, Italian 10.1007/s11547-010-0605-1 20981501 24. Granata V. Fusco R. De Muzio F. Cutolo C. Setola S.V. Grassi R. Grassi F. Ottaiano A. Nasti G. Tatangelo F. Radiomics textural features by MR imaging to assess clinical outcomes following liver resection in colorectal liver metastases Radiol. Med. 2022 127 461 470 10.1007/s11547-022-01477-6 35347583 25. Granata V. Fusco R. De Muzio F. Cutolo C. Setola S.V. Dell'Aversana F. Grassi F. Belli A. Silvestro L. Ottaiano A. Radiomics and machine learning analysis based on magnetic resonance imaging in the assessment of liver mucinous colorectal metastases Radiol. Med. 2022 127 763 772 10.1007/s11547-022-01501-9 35653011 26. Sansone M. Marrone S. Di Salvio G. Belfiore M.P. Gatta G. Fusco R. Vanore L. Zuiani C. Grassi F. Vietri M.T. Comparison between two packages for pectoral muscle removal on mammographic images Radiol. Med. 2022 127 848 856 10.1007/s11547-022-01521-5 35816260 27. Gabelloni M. Faggioni L. Cioni D. Mendola V. Falaschi Z. Coppola S. Corradi F. Isirdi A. Brandi N. Coppola F. Extracorporeal membrane oxygenation (ECMO) in COVID-19 patients: A pocket guide for radiologists Radiol. Med. 2022 127 369 382 10.1007/s11547-022-01473-w 35279765 28. Ossola C. Curti M. Calvi M. Tack S. Mazzoni S. Genesio L. Venturini M. Genovese E.A. Role of ultrasound and magnetic resonance imaging in the prognosis and classification of muscle injuries in professional football players: Correlation between imaging and return to sport time Radiol. Med. 2021 126 1460 1467 10.1007/s11547-021-01396-y 34309765 29. Bakdik S. Arslan S. Oncu F. Durmaz M.S. Altunkeser A. Eryilmaz M.A. Unlu Y. Effectiveness of Superb Microvascular Imaging for the differentiation of intraductal breast lesions Med Ultrason. 2018 20 306 312 10.11152/mu-1433 30167583 30. Cai S.-M. Wang H.-V. Zhang X.-Y. Zhang L. Zhu Q.-L. Li J.-C. Sun Q. Jiang Y.-X. The Vascular Index of Superb Microvascular Imaging Can Improve the Diagnostic Accuracy for Breast Imaging Reporting and Data System Category 4 Breast Lesions Cancer Manag. Res. 2020 12 1819 1826 10.2147/CMAR.S242101 32210624 31. Corvino A. Varelli C. Cocco G. Corvino F. Catalano O. Seeing the unseen with superb microvascular imaging: Ultrasound depiction of normal dermis vessels J. Clin. Ultrasound 2022 50 121 127 10.1002/jcu.23068 34761407 32. Corvino A. Varelli C. Catalano F. Cocco G. Delli Pizzi A. Boccatonda A. Corvino F. Basile L. Catalano O. Use of high-frequency transducers in breast sonography J. Pers. Med. 2022 12 1960 10.3390/jpm12121960 36556182 33. Corvino A. Catalano O. Varelli C. Cocco G. Delli Pizzi A. Corvino F. Tafuri D. Caruso M. Non-glandular findings on breast ultrasound. Part II: A pictorial review of chest wall lesions J. Ultrasound 2023 1 10 10.1007/s40477-022-00773-1 36595199 34. Kim S. Lee H.J. Ko K.H. Park A.Y. Koh J. Jung H.K. New Doppler imaging technique for assessing angiogenesis in breast tumors: Correlation with immunohistochemically analyzed microvessels density Acta Radiol. 2018 59 1414 1421 10.1177/0284185118769690 29667882 35. Park A.Y. Seo B.K. Up-to-date Doppler techniques for breast tumor vascularity: Superb microvascular imaging and contrast-enhanced ultrasound Ultrasonography 2018 37 98 106 10.14366/usg.17043 29025210 36. Park A.Y. Kwon M. Woo O.H. Cho K.R. Park E.K. Cha S.H. Song S.E. Lee J.-H. Cha J. Son G.S. A Prospective Study on the Value of Ultrasound Microflow Assessment to Distinguish Malignant from Benign Solid Breast Masses: Association between Ultrasound Parameters and Histologic Microvessel Densities Korean J. Radiol. 2019 20 759 772 10.3348/kjr.2018.0515 30993927 37. Zhu Y.-C. Zhang Y. Deng S.-H. Jiang Q. Diagnostic Performance of Superb Microvascular Imaging (SMI) Combined with Shear-Wave Elastography in Evaluating Breast Lesions Experiment 2018 24 5935 5942 10.12659/MSM.910399 38. BBarile A. Correction to: Some thoughts and greetings from the new Editor-in-Chief Radiol. Med. 2021 126 1377 10.1007/s11547-021-01346-8 34559382 39. Catalano O. Corvino A. Basile L. Catalano F. Varelli C. Use of new microcirculation software allows the demonstration of dermis vascularization J. Ultrasound 2022 1 6 10.1007/s40477-022-00710-2 40. Sivakumaran L. Alturkistani H. Lerouge S. Bertrand-Grenier A. Zehtabi F. Therasse E. Roy-Cardinal M.-H. Bhatnagar S. Cloutier G. Soulez G. Strain Ultrasound Elastography of Aneurysm Sac Content after Randomized Endoleak Embolization with Sclerosing vs. Non-sclerosing Chitosan-based Hydrogels in a Canine Model J. Vasc. Interv. Radiol. 2022 33 495 504 10.1016/j.jvir.2022.02.003 35150836 41. Bartolotta T.V. Orlando A.A.M. Dimarco M. Zarcaro C. Ferraro F. Cirino A. Matranga D. Vieni S. Cabibi D. Diagnostic performance of 2D-shear wave elastography in the diagnosis of breast cancer: A clinical appraisal of cutoff values Radiol. Med. 2022 127 1209 1220 10.1007/s11547-022-01546-w 36114930 42. Ding S.-S. Liu C. Zhang Y.-F. Sun L.-P. Xiang L.-H. Liu H. Fang Y. Ren W.-W. Zhao H. Sun X.-M. Contrast-enhanced ultrasound in the assessment of Crohn's disease activity: Comparison with computed tomography enterography Radiol. Med. 2022 127 1068 1078 10.1007/s11547-022-01535-z 35943658 43. Elia D. Fresilli D. Pacini P. Cardaccio S. Polti G. Guiban O. Celletti I. Kutrolli E. De Felice C. Occhiato R. Can strain US-elastography with strain ratio (SRE) improve the diagnostic accuracy in the assessment of breast lesions? Preliminary results J. Ultrasound 2020 24 157 163 10.1007/s40477-020-00505-3 32651770 44. Hyodo R. Takehara Y. Naganawa S. 4D Flow MRI in the portal venous system: Imaging and analysis methods, and clinical applications Radiol. Med. 2022 127 1181 1198 10.1007/s11547-022-01553-x 36123520 45. Ruan S.-M. Huang H. Cheng M.-Q. Lin M.-X. Hu H.-T. Huang Y. Li M.-D. Wang W. Shear-wave elastography combined with contrast-enhanced ultrasound algorithm for noninvasive characterization of focal liver lesions Radiol. Med. 2022 1 10 10.1007/s11547-022-01575-5 34665430 46. Ruscitti P. Esposito M. Gianneramo C. Di Cola I. De Berardinis A. Martinese A. Tochap G.N. Conforti A. Masciocchi C. Cipriani P. Nail and enthesis assessment in patients with psoriatic disease by high frequency ultrasonography: Findings from a single-centre cross-sectional study Radiol. Med. 2022 127 1400 1406 10.1007/s11547-022-01568-4 36260243 47. Salaffi F. Carotti M. Di Matteo A. Ceccarelli L. Farah S. Villota-Eraso C. Di Carlo M. Giovagnoni A. Ultrasound and magnetic resonance imaging as diagnostic tools for sarcopenia in immune-mediated rheumatic diseases (IMRDs) Radiol. Med. 2022 127 1277 1291 10.1007/s11547-022-01560-y 36125609 48. Ventura C. Baldassarre S. Cerimele F. Pepi L. Marconi E. Ercolani P. Floridi C. Argalia G. Goteri G. Giovagnoni A. 2D shear wave elastography in evaluation of prognostic factors in breast cancer Radiol. Med. 2022 127 1221 1227 10.1007/s11547-022-01559-5 36123521 49. Youk J.H. Gweon H.M. Son E.J. Shear-wave elastography in breast ultrasonography: The state of the art Ultrasonography 2017 36 300 309 10.14366/usg.17024 28513127 50. Barr R.G. Nakashima K. Amy D. Cosgrove D. Farrokh A. Schafer F. Bamber J.C. Castera L. Choi B.I. Chou Y.-H. WFUMB guidelines and recommendations for clinical use of ultrasound elastography: Part 2, breast Ultrasound Med. Biol. 2015 41 1148 1160 10.1016/j.ultrasmedbio.2015.03.008 25795620 51. Barr R.G. Future of breast elastography Ultrasonography 2019 38 93 105 10.14366/usg.18053 30884636 52. Luo S. Yao G. Hong Z. Zhang S. Wang W. Zhang J. Zhang Y. Wu J. Zhang L. Cheng H. Qualitative Classification of Shear Wave Elastography for Differential Diagnosis Between Benign and Metastatic Axillary Lymph Nodes in Breast Cancer Front. Oncol. 2019 9 533 10.3389/fonc.2019.00533 31312611 53. Lee S.H. Chang J.M. Cho N. Koo H.R. Yi A. Kim S.J. Youk J.H. Son E.J. Choi S.H. Kook S.H. Practice guideline for the performance of breast ultrasound elastography Ultrasonography 2013 33 3 10 10.14366/usg.13012 24936489 54. Sezgin G. Coskun M. Apaydin M. Sari A.A. The role of rare breast cancers in the false negative strain elastography results Radiol. Med. 2021 126 349 355 10.1007/s11547-020-01270-3 32894448 55. Trombadori C.M.L. D'Angelo A. Ferrara F. Santoro A. Belli P. Manfredi R. Radial Scar: A management dilemma Radiol. Med. 2021 126 774 785 10.1007/s11547-021-01344-w 33743143 56. Kapetas P. Woitek R. Clauser P. Bernathova M. Pinker K. Helbich T.H. Baltzer P.A. A Simple Ultrasound Based Classification Algorithm Allows Differentiation of Benign from Malignant Breast Lesions by Using Only Quantitative Parameters Mol. Imaging Biol. 2018 20 1053 1060 10.1007/s11307-018-1187-x 29633108 57. De Muzio F. Grassi F. Dell'Aversana F. Fusco R. Danti G. Flammia F. Chiti G. Valeri T. Agostini A. Palumbo P. A Narrative Review on LI-RADS Algorithm in Liver Tumors: Prospects and Pitfalls Diagnostics 2022 12 1655 10.3390/diagnostics12071655 35885561 58. Catalano O. Sandomenico F. Vallone P. Setola S.V. Granata V. Fusco R. Lastoria S. Mansi L. Petrillo A. Contrast-Enhanced Ultrasound in the Assessment of Patients with Indeterminate Abdominal Findings at Positron Emission Tomography Imaging Ultrasound Med. Biol. 2016 42 2717 2723 10.1016/j.ultrasmedbio.2016.06.023 27495384 59. Granata V. Grassi R. Fusco R. Setola S. Belli A. Piccirillo M. Pradella S. Giordano M. Cappabianca S. Brunese L. Abbreviated MRI Protocol for the Assessment of Ablated Area in HCC Patients Int. J. Environ. Res. Public Heal. 2021 18 3598 10.3390/ijerph18073598 33808466 60. Barretta M.L. Catalano O. Setola S.V. Granata V. Marone U. Gallipoli A.D. Gallbladder metastasis: Spectrum of imaging findings Abdom. Imaging 2011 36 729 734 10.1007/s00261-011-9696-y 21318378 61. Granata V. Fusco R. Catalano O. Avallone A. Palaia R. Botti G. Tatangelo F. Granata F. Cascella M. Izzo F. Diagnostic accuracy of magnetic resonance, computed tomography and contrast enhanced ultrasound in radiological multimodality assessment of peribiliary liver metastases PLoS ONE 2017 12 e0179951 10.1371/journal.pone.0179951 28632786 62. Granata V. Castelguidone E.D.L.D. Fusco R. Catalano O. Piccirillo M. Palaia R. Izzo F. Gallipoli A.D. Petrillo A. Irreversible electroporation of hepatocellular carcinoma: Preliminary report on the diagnostic accuracy of magnetic resonance, computer tomography, and contrast-enhanced ultrasound in evaluation of the ablated area Radiol. Med. 2015 121 122 131 10.1007/s11547-015-0582-5 26345332 63. Chou C.-P. Huang J.-S. Wang J.-S. Pan H.-B. Contrast-enhanced ultrasound features of breast capillary hemangioma: A case report and review of literature J. Ultrasound 2022 25 103 106 10.1007/s40477-020-00550-y 33409863 64. Guo R. Lu G. Qin B. Fei B. Ultrasound Imaging Technologies for Breast Cancer Detection and Management: A Review Ultrasound Med. Biol. 2018 44 37 70 10.1016/j.ultrasmedbio.2017.09.012 29107353 65. Faccioli N. Santi E. Foti G. D'Onofrio M. Cost-effectiveness analysis of including contrast-enhanced ultrasound in management of pancreatic cystic neoplasms Radiol. Med. 2022 127 349 359 10.1007/s11547-022-01459-8 35230618 66. Fresilli D. Di Leo N. Martinelli O. Di Marzo L. Pacini P. Dolcetti V. Del Gaudio G. Canni F. Ricci L.I. De Vito C. 3D-Arterial analysis software and CEUS in the assessment of severity and vulnerability of carotid atherosclerotic plaque: A comparison with CTA and histopathology Radiol. Med. 2022 127 1254 1269 10.1007/s11547-022-01551-z 36114929 67. Sofia C. Solazzo A. Cattafi A. Chimenz R. Cicero G. Marino M.A. D'Angelo T. Manti L. Condorelli E. Ceravolo G. Contrast-enhanced voiding urosonography in the assessment of vesical-ureteral reflux: The time has come Radiol. Med. 2021 126 901 909 10.1007/s11547-021-01360-w 33954899 68. Varghese B.A. Lee S. Cen S. Talebi A. Mohd P. Stahl D. Perkins M. Desai B. Duddalwar V.A. Larsen L.H. Characterizing breast masses using an integrative framework of machine learning and CEUS-based radiomics J. Ultrasound 2022 25 699 708 10.1007/s40477-021-00651-2 35040103 69. Jia K. Li L. Wu X.J. Hao M.J. Xue H.Y. Contrast-enhanced ultrasound for evaluating the pathologic response of breast cancer to neoadjuvant chemotherapy Medicine 2019 98 e14258 10.1097/MD.0000000000014258 30681622 70. Bin L. Huihui Y. Weiping Y. Changyuan W. Qinghong Q. Weiyu M. Value of three-dimensional ultrasound in differentiating malignant from benign breast tumors. A systematic review and meta-analysis Ultrasound Q. 2019 35 68 73 10.1097/RUQ.0000000000000433 30807546 71. Satake H. Ishigaki S. Ito R. Naganawa S. Radiomics in breast MRI: Current progress toward clinical application in the era of artificial intelligence Radiol. Med. 2021 127 39 56 10.1007/s11547-021-01423-y 34704213 72. Rossi F. Lambertini M. Brunetti N. De Giorgis S. Razeti M.G. Calabrese M. Tagliafico A.S. Muscle mass loss in breast cancer patients of reproductive age (<=45 years) undergoing neoadjuvant chemotherapy Radiol. Med. 2022 1 9 10.1007/s11547-022-01574-6 34665430 73. Fiaschetti V. Ubaldi N. De Fazio S. Ricci A. Maspes F. Cossu E. Digital tomosynthesis spot view in architectural distortions: Outcomes in management and radiation dose Radiol. Med. 2022 1 14 10.1007/s11547-022-01570-w 34665430 74. Clauser P. Londero V. Como G. Girometti R. Bazzocchi M. Zuiani C. Comparison between different imaging techniques in the evaluation of malignant breast lesions: Can 3D ultrasound be useful? Radiol. Med. 2014 119 240 248 10.1007/s11547-013-0338-z 24297584 75. Park A.Y. Seo B.K. Cho K.R. Woo O.H. The utility of MicroPureTM ultrasound technique in assessing grouped microcalcifications without a mass on mammography J. Breast Cancer 2016 19 83 86 10.4048/jbc.2016.19.1.83 27066098 76. van Zelst J.C. Mann R.M. Automated three-dimensional breast US for screening: Technique, artifacts, and lesion characterization Radio Graph. 2018 38 663 683 10.1148/rg.2018170162 77. Zanotel M. Bednarova I. Londero V. Linda A. Lorenzon M. Girometti R. Zuiani C. Automated breast ultrasound: Basic principles and emerging clinical applications Radiol. Med. 2018 123 1 12 10.1007/s11547-017-0805-z 28849324 78. Schiaffino S. Gristina L. Tosto S. Massone E. De Giorgis S. Garlaschi A. Tagliafico A. Calabrese M. The value of coronal view as a stand-alone assessment in women undergoing automated breast ultrasound Radiol. Med. 2020 126 206 213 10.1007/s11547-020-01250-7 32676876 79. Barr R.G. DeVita R. Destounis S. Manzoni F. De Silvestri A. Tinelli C. Agreement between an automated volume breast scanner and hand held ultrasound for diagnostic breast examinations J. Ultrasound. Med. 2017 36 2087 2092 10.1002/jum.14248 28569407 80. Brem R.F. Tabar L. Duffy S.W. Inciardi M.F. Guingrich J.A. Hashimoto B.E. Lander M.R. Lapidus R.L. Peterson M.K. Rapelyea J.A. Assessing Improvement in Detection of Breast Cancer with Three-dimensional Automated Breast US in Women with Dense Breast Tissue: The SomoInsight Study Radiology 2015 274 663 673 10.1148/radiol.14132832 25329763 81. D'Angelo A. Orlandi A. Bufi E. Mercogliano S. Belli P. Manfredi R. Automated breast volume scanner (ABVS) compared to handheld ultrasound (HHUS) and contrast-enhanced magnetic resonance imaging (CE-MRI) in the early assessment of breast cancer during neoadjuvant chemotherapy: An emerging role to monitoring tumor response? Radiol. Med. 2021 126 517 526 10.1007/s11547-020-01319-3 33385300 82. Girometti R. Zanotel M. Londero V. Bazzocchi M. Zuiani C. Comparison between automated breast volume scanner (ABVS) versus hand-held ultrasound as a second look procedure after magnetic resonance imaging Eur. Radiol. 2017 27 3767 3775 10.1007/s00330-017-4749-4 28120030 83. Coppola F. Faggioni L. Regge D. Giovagnoni A. Golfieri R. Bibbolino C. Miele V. Neri E. Grassi R. Artificial intelligence: Radiologists' expectations and opinions gleaned from a nationwide online survey Radiol. Med. 2021 126 63 71 10.1007/s11547-020-01205-y 32350797 84. Nardone V. Reginelli A. Grassi R. Boldrini L. Vacca G. D'Ippolito E. Annunziata S. Farchione A. Belfiore M.P. Desideri I. Delta radiomics: A systematic review Radiol. Med. 2021 126 1571 1583 10.1007/s11547-021-01436-7 34865190 85. Tadayyon H. Gangeh M. Sannachi L. Trudeau M. Pritchard K. Ghandi S. Eisen A. Look-Hong N. Holloway C. Wright F. A priori prediction of breast tumour response to chemotherapy using quantitative ultrasound imaging and artificial neural networks Oncotarget 2019 10 3910 3923 10.18632/oncotarget.26996 31231468 86. Vicini S. Bortolotto C. Rengo M. Ballerini D. Bellini D. Carbone I. Preda L. Laghi A. Coppola F. Faggioni L. A narrative review on current imaging applications of artificial intelligence and radiomics in oncology: Focus on the three most common cancers Radiol. Med. 2022 127 819 836 10.1007/s11547-022-01512-6 35771379 87. Simonetti I. Bruno F. Fusco R. Cutolo C. Setola S.V. Patrone R. Masciocchi C. Palumbo P. Arrigoni F. Picone C. Multimodality Imaging Assessment of Desmoid Tumors: The Great Mime in the Era of Multidisciplinary Teams J. Pers. Med. 2022 12 1153 10.3390/jpm12071153 35887650 88. Granata V. Fusco R. Avallone A. De Stefano A. Ottaiano A. Sbordone C. Brunese L. Izzo F. Petrillo A. Radiomics-Derived Data by Contrast Enhanced Magnetic Resonance in RAS Mutations Detection in Colorectal Liver Metastases Cancers 2021 13 453 10.3390/cancers13030453 33504085 89. Granata V. Fusco R. Risi C. Ottaiano A. Avallone A. De Stefano A. Grimm R. Grassi R. Brunese L. Izzo F. Diffusion-Weighted MRI and Diffusion Kurtosis Imaging to Detect RAS Mutation in Colorectal Liver Metastasis Cancers 2020 12 2420 10.3390/cancers12092420 32858990 90. Wilding R. Sheraton V.M. Soto L. Chotai N. Tan E.Y. Deep learning applied to breast imaging classification and segmentation with human expert intervention J. Ultrasound 2022 25 659 666 10.1007/s40477-021-00642-3 35000127 91. Ciritsis A. Rossi C. Eberhard M. Marcon M. Becker A.S. Boss A. Automatic classification of ultrasound breast lesions using a deep convolutional neural network mimicking human decision-making Eur. Radiol. 2019 29 5458 5468 10.1007/s00330-019-06118-7 30927100 92. Zeng D. Xu M. Liang J.-Y. Cheng M.-Q. Huang H. Pan J.-M. Huang Y. Tong W.-J. Xie X.-Y. Lu M.-D. Using new criteria to improve the differentiation between HCC and non-HCC malignancies: Clinical practice and discussion in CEUS LI-RADS 2017 Radiol. Med. 2021 127 1 10 10.1007/s11547-021-01417-w 34665430 93. Jiang M. Li C.-L. Luo X.-M. Chuan Z.-R. Lv W.-Z. Li X. Cui X.-W. Dietrich C.F. Ultrasound-based deep learning radiomics in the assessment of pathological complete response to neoadjuvant chemotherapy in locally advanced breast cancer Eur. J. Cancer 2021 147 95 105 10.1016/j.ejca.2021.01.028 33639324 94. Ilesanmi A.E. Chaumrattanakul U. Makhanov S.S. Methods for the segmentation and classification of breast ultrasound images: A review J. Ultrasound 2021 24 367 382 10.1007/s40477-020-00557-5 33428123 95. Scapicchio C. Gabelloni M. Barucci A. Cioni D. Saba L. Neri E. A deep look into radiomics Radiol. Med. 2021 126 1296 1311 10.1007/s11547-021-01389-x 34213702 96. Wu G.-G. Zhou L.-Q. Xu J.-W. Wang J.-Y. Wei Q. Deng Y.-B. Cui X.-W. Dietrich C.F. Artificial intelligence in breast ultrasound World J. Radiol. 2019 11 19 26 10.4329/wjr.v11.i2.19 30858931 97. Bartolotta T.V. Orlando A. Cantisani V. Matranga D. Ienzi R. Cirino A. Amato F. Di Vittorio M.L. Midiri M. Lagalla R. Focal breast lesion characterization according to the BI-RADS US lexicon: Role of a computer-aided decision-making support Radiol. Med. 2018 123 498 506 10.1007/s11547-018-0874-7 29569216 98. Bartolotta T.V. Orlando A.A.M. Spatafora L. Dimarco M. Gagliardo C. Taibbi A. S-Detect characterization of focal breast lesions according to the US BI RADS lexicon: A pictorial essay J. Ultrasound 2020 23 207 215 10.1007/s40477-020-00447-w 32185702 99. Bartolotta T.V. Orlando A.A.M. Di Vittorio M.L. Amato F. Dimarco M. Matranga D. Ienzi R. S-Detect characterization of focal solid breast lesions: A prospective analysis of inter-reader agreement for US BI-RADS descriptors J. Ultrasound 2020 24 143 150 10.1007/s40477-020-00476-5 32447631 100. Zhao C. Xiao M. Jiang Y. Liu H. Wang M. Wang H. Sun Q. Zhu Q. Feasibility of computer-assisted diagnosis for breast ultrasound: The results of the diagnostic performance of S-detect from a single center in China Cancer Manag. Res. 2019 11 921 930 10.2147/CMAR.S190966 30774422 101. Di Segni M. de Soccio V. Cantisani V. Bonito G. Rubini A. Di Segni G. Lamorte S. Magri V. De Vito C. Migliara G. Automated classification of focal breast lesions according to S-detect: Validation and role as a clinical and teaching tool J. Ultrasound 2018 21 105 118 10.1007/s40477-018-0297-2 29681007 102. Chiao J.-Y. Chen K.-Y. Liao K.Y.-K. Hsieh P.-H. Zhang G. Huang T.-C. Detection and classification the breast tumors using mask R-CNN on sonograms Medicine 2019 98 e15200 10.1097/MD.0000000000015200 31083152 103. Gao Y. Luo Y. Zhao C. Xiao M. Ma L. Li W. Qin J. Zhu Q. Jiang Y. Nomogram based on radiomics analysis of primary breast cancer ultrasound images: Prediction of axillary lymph node tumor burden in patients Eur. Radiol. 2021 31 928 937 10.1007/s00330-020-07181-1 32845388 104. Fausto A. Rizzatto G. Preziosa A. Gaburro L. Washburn M.J. Rubello D. Volterrani L. A new method to combine contrast-enhanced magnetic resonance imaging during live ultrasound of the breast using volume navigation technique: A study for evaluating feasibility, accuracy and reproducibility in healthy volunteers Eur. J. Radiol. 2012 81 e332 e337 10.1016/j.ejrad.2011.11.001 22104088 105. Kucukkaya F. Aribal E. Tureli D. Altas H. Kaya H. Use of a Volume Navigation Technique for Combining Real-Time Ultrasound and Contrast-Enhanced MRI: Accuracy and Feasibility of a Novel Technique for Locating Breast Lesions Am. J. Roentgenol. 2016 206 217 225 10.2214/AJR.14.14101 26700355 106. Papalexis N. Parmeggiani A. Facchini G. Miceli M. Carbone G. Cavallo M. Spinnato P. Current concepts in the diagnosis and treatment of adhesive capsulitis: Role of diagnostic imaging and ultrasound-guided interventional procedures Radiol. Med. 2022 127 1390 1399 10.1007/s11547-022-01566-6 36376543 107. Crisan D. Wortsman X. Alfageme F. Catalano O. Badea A. Scharffetter-Kochanek K. Sindrilaru A. Crisan M. Ultrasonography in dermatologic surgery: Revealing the unseen for improved surgical planning JDDG J. der Dtsch. Dermatol. Ges. 2022 20 913 926 10.1111/ddg.14781 35616215 108. Moschetta M. Sardaro A. Nitti A. Telegrafo M. Maggialetti N. Scardapane A. Brunese M.C. Lavelli V. Ferrari C. Ultrasound evaluation of ductal carcinoma in situ of the breast J. Ultrasound 2021 25 41 45 10.1007/s40477-020-00551-x 33409862 109. Stavros A.T. Thickman D. Rapp C.L. Dennis M.A. Parker S.H. Sisney G.A. Solid breast nodules: Use of sonography to distinguish between benign and malignant lesions Radiology 1995 196 123 134 10.1148/radiology.196.1.7784555 7784555 110. Deandrea S. Cavazzana L. Principi N. Luconi E. Campoleoni M. Bastiampillai A.J. Bracchi L. Bucchi L. Pedilarco S. Piscitelli A. Screening of women with aesthetic prostheses in dedicated sessions of a population-based breast cancer screening programme Radiol. Med. 2021 126 946 955 10.1007/s11547-021-01357-5 33954896 111. Fusco R. Petrillo A. Catalano O. Sansone M. Granata V. Filice S. D'Aiuto M. Pankhurst Q. Douek M. Procedures for location of non-palpable breast lesions: A systematic review for the radiologist Breast Cancer 2014 21 522 531 10.1007/s12282-012-0427-1 23115016 112. Nori J. Bicchierai G. Amato F. De Benedetto D. Boeri C. Vanzi E. Di Naro F. Bianchi S. Miele V. A new technique for the histological diagnosis of Paget's disease of the breast using a semiautomated core needle biopsy with a 14-gauge needle Radiol. Med. 2021 126 936 945 10.1007/s11547-021-01358-4 33900526 113. Panzironi G. Moffa G. Galati F. Pediconi F. Ultrasound-guided 8-Gauge vacuum-assisted excision for selected B3 breast lesions: A preliminary experience Radiol. Med. 2022 127 57 64 10.1007/s11547-021-01429-6 34800217 114. Petrillo A. Di Giacomo R. Esposito E. Vallone P. Setola S.V. Raso M.M. Granata V. Barretta M.L. Siani C. Rinaldo C. Preoperative localisation of nonpalpable breast lesions using magnetic markers in a tertiary cancer centre Eur. Radiol. Exp. 2022 6 28 10.1186/s41747-022-00280-2 35790602 115. Bellardita L. Colciago R.R. Frasca S. De Santis M.C. Gay S. Palorini F. La Rocca E. Valdagni R. Rancati T. Lozza L. Breast cancer patient perspective on opportunities and challenges of a genetic test aimed to predict radio-induced side effects before treatment: Analysis of the Italian branch of the REQUITE project Radiol. Med. 2021 126 1366 1373 10.1007/s11547-021-01395-z 34268681 116. Carbonaro L.A. Rizzo S.S. Schiaffino S. Mainini A.P. Berger N. Trimboli R.M. Sardanelli F. Biennial screening mammography: How many women ask for more? Estimate of the interval mammogram rate in an organised population-based screening programme Radiol. Med. 2021 126 200 205 10.1007/s11547-020-01238-3 32577906 117. Caruso M. Catalano O. Bard R. Varelli C. Corvino F. Caiazzo C. Corvino A. Non-glandular findings on breast ultrasound. Part I: A pictorial review of superficial lesions J. Ultrasound 2022 25 783 797 10.1007/s40477-021-00619-2 35438461 118. Maio F. Tari D.U. Granata V. Fusco R. Grassi R. Petrillo A. Pinto F. Breast Cancer Screening during COVID-19 Emergency: Patients and Department Management in a Local Experience J. Pers. Med. 2021 11 380 10.3390/jpm11050380 34066425 119. Neri E. Granata V. Montemezzi S. Belli P. Bernardi D. Brancato B. Caumo F. Calabrese M. Coppola F. Cossu E. Structured reporting of x-ray mammography in the first diagnosis of breast cancer: A Delphi consensus proposal Radiol. Med. 2022 127 471 483 10.1007/s11547-022-01478-5 35303247 120. Deandrea S. Sardanelli F. Calabrese M. Ferre F. Vainieri M. Sestini E. Caumo F. Saguatti G. Bucchi L. Cataliotti L. Provision of follow-up care for women with a history of breast cancer following the 2016 position paper by the Italian Group for Mammographic Screening and the Italian College of Breast Radiologists by SIRM: A survey of Senonetwork Italian breast centres Radiol. Med. 2022 127 484 489 10.1007/s11547-022-01485-6 35347582 121. Granata V. Bicchierai G. Fusco R. Cozzi D. Grazzini G. Danti G. De Muzio F. Maggialetti N. Smorchkova O. D'Elia M. Diagnostic protocols in oncology: Workup and treatment planning. Part 2: Abbreviated MR protocol Eur. Rev. Med. Pharmacol. Sci. 2021 25 6499 6528 10.26355/eurrev_202111_27094 34787854 122. Leung J.W. Second-look ultrasound: Only for biopsy or more? Eur. J. Radiol. 2012 81 s87 s89 10.1016/S0720-048X(12)70035-9 23083617 123. Sansone M. Grassi R. Belfiore M.P. Gatta G. Grassi F. Pinto F. La Casella G.V. Fusco R. Cappabianca S. Granata V. Radiomic features of breast parenchyma: Assessing differences between FOR PROCESSING and FOR PRESENTATION digital mammography Insights Into Imaging 2021 12 147 10.1186/s13244-021-01093-4 34674061 124. Fusco R. Granata V. Pariante P. Cerciello V. Siani C. Di Bonito M. Valentino M. Sansone M. Botti G. Petrillo A. Blood oxygenation level dependent magnetic resonance imaging and diffusion weighted MRI imaging for benign and malignant breast cancer discrimination Magn. Reson. Imaging 2020 75 51 59 10.1016/j.mri.2020.10.008 33080334 125. Fusco R. Granata V. Raso M.M. Vallone P. De Rosa A. Siani C. Di Bonito M. Petrillo A. Sansone M. Blood Oxygenation Level Dependent Magnetic Resonance Imaging (MRI), Dynamic Contrast Enhanced MRI, and Diffusion Weighted MRI for Benign and Malignant Breast Cancer Discrimination: A Preliminary Experience Cancers 2021 13 2421 10.3390/cancers13102421 34067721 126. Girometti R. Linda A. Conte P. Lorenzon M. De Serio I. Jerman K. Londero V. Zuiani C. Multireader comparison of contrast-enhanced mammography versus the combination of digital mammography and digital breast tomosynthesis in the preoperative assessment of breast cancer Radiol. Med. 2021 126 1407 1414 10.1007/s11547-021-01400-5 34302599 127. Catalano O. Raso M.M. Petrillo A. D'Errico A.G. Extended field of view in breast sonography Ultraschall Med 2011 32 198 202 10.1055/s-0029-1245836 21465424 128. Petralia G. Zugni F. Summers P.E. Colombo A. Pricolo P. Grazioli L. Colagrande S. Giovagnoni A. Padhani A.R. On behalf of the Italian Working Group on Magnetic Resonance Whole-body magnetic resonance imaging (WB-MRI) for cancer screening: Recommendations for use Radiol. Med. 2021 126 1434 1450 10.1007/s11547-021-01392-2 34338948 129. Petrillo A. Catalano O. Fusco R. Filice S. Vallone P. Setola S. Granata V. Raiano C. Avino F. Di Bonito M. Optical imaging of the breast: Evaluation of deoxyhemoglobin concentration alteration in 166 patients with suspicious breast lesions Eur. Radiol. Exp. 2018 2 8 10.1186/s41747-018-0038-5 29708210 130. Lohitvisate W. Pummee N. Kwankua A. Mammographic and ultrasonographic features of triple-negative breast cancer compared with non-triple-negative breast cancer J. Ultrasound 2022 1 8 10.1007/s40477-022-00709-9 131. Cellina M. Gibelli D. Martinenghi C. Giardini D. Soresina M. Menozzi A. Oliva G. Carrafiello G. Non-contrast magnetic resonance lymphography (NCMRL) in cancer-related secondary lymphedema: Acquisition technique and imaging findings Radiol. Med. 2021 126 1477 1486 10.1007/s11547-021-01410-3 34379302 132. Fusco R. Sansone M. Granata V. Di Bonito M. Avino F. Catalano O. Botti G. Petrillo A. Use of Quantitative Morphological and Functional Features for Assessment of Axillary Lymph Node in Breast Dynamic Contrast-Enhanced Magnetic Resonance Imaging BioMed Res. Int. 2018 2018 2610801 10.1155/2018/2610801 30003092 133. Nicosia L. Bozzini A.C. Palma S. Montesano M. Signorelli G. Pesapane F. Latronico A. Bagnardi V. Frassoni S. Sangalli C. Contrast-Enhanced Spectral Mammography and tumor size assessment: A valuable tool for appropriate surgical management of breast lesions Radiol. Med. 2022 127 1228 1234 10.1007/s11547-022-01561-x 36149581 134. Romanucci G. Mercogliano S. Carucci E. Cina A. Zantedeschi E. Caneva A. Benassuti C. Fornasa F. Diagnostic accuracy of resection margin in specimen radiography: Digital breast tomosynthesis versus full-field digital mammography Radiol. Med. 2021 126 768 773 10.1007/s11547-021-01337-9 33625658 |
PMC10000576 | Background: Metaplastic breast cancer (MpBC) is an aggressive histologic type of breast cancer. Although MpBC has a poor prognosis and is responsible for a large proportion of breast cancer mortalities, the clinical features of MpBC compared with invasive ductal carcinoma (IDC) are not well known, and the optimal treatment has not been identified. Methods: We retrospectively reviewed medical records of 155 MpBC patients and 16,251 IDC cases who underwent breast cancer surgery in a single institution between January 1994 and December 2019. The two groups were matched 1:4 by age, tumor size, nodal status, hormonal receptor status, and HER2 status using propensity-score matching (PSM). Finally, 120 MpBC patients were matched with 478 IDC patients. Disease-free survival and overall survival of MpBC and IDC patients both before and after PSM were analyzed by Kaplan-Meier survival, and multivariable Cox regression analysis was performed to identify variables affecting long-term prognosis. Results: The most common subtype of MpBC was triple-negative breast cancer, and nuclear and histologic grades were higher than those of IDC. Pathologic nodal staging of the metaplastic group was significantly lower than that of the ductal group, and more frequent adjuvant chemotherapy was performed in the metaplastic group. Multivariable Cox regression analysis indicated that MpBC was an independent prognostic factor for disease-free survival (HR = 2.240; 95% CI, 1.476-3.399, p = 0.0002) and overall survival (HR = 1.969; 95% CI, 1.147-3.382, p = 0.0140). However, survival analysis revealed no significant difference between MpBC and IDC patients in disease-free survival (HR = 1.465; 95% CI, 0.882-2.432, p = 0.1398) or overall survival (hazard ratio (HR) = 1.542; 95% confidential interval (CI), 0.875-2.718, p = 0.1340) after PSM. Conclusion: Although the MpBC histologic type had poor prognostic factors compared with IDC, it can be treated according to the same principles as aggressive IDC. metaplastic breast cancer invasive ductal carcinoma triple negative breast neoplasm propensity score survival This research received no external funding. pmc1. Introduction Female breast cancer was the most common cancer in 2020, surpassing lung cancer . Nearly half of breast cancer cases are diagnosed in Asia . Metaplastic breast cancer (MpBC) is a rare histologic subtype that accounts for only 0.2%-2% of breast cancers, and it has been found that there is a high incidence of MpBC diagnosis among the black race in some studies . This cancer is comprised of poorly differentiated tumors and heterogeneous histology involving both mesenchymal and epithelial components of spindle cells, squamous differentiation, and chondroids . Variants of MpBC include carcinosarcoma, matrix-producing carcinoma, sarcomatoid carcinoma, pseudosarcoma, and mixed tumors of the breast . MpBC is a clinically aggressive breast cancer with poor prognosis . Such tumors tend to be larger and have more node negativity than other types. In addition, it carries a greater risk of metastasis than other histologic subtypes, patients present with more advanced disease stage, and the cancer is less responsive to conventional adjuvant and neoadjuvant chemotherapy . MpBC is characterized by chemoresistance and hematogenous spread, in contrast to the lymphatic dissemination typically seen with invasive ductal carcinoma (IDC) . MpBC is more likely to be treated with mastectomy than breast-conserving surgery, and such patients have higher mastectomy rates than patients with IDC. Despite this higher rate, this type of surgery was not associated with a survival difference . The molecular subtypes of MpBC are basal-like breast cancer with estrogen receptor negativity, progesterone receptor negativity, and human epidermal growth factor receptor 2 (HER2) negativity; and MpBC is more aggressive than triple negative breast cancer (TNBC), having a worse prognosis . In addition, the TNBC subtype of MpBC is worse than that of IDC . In addition, although many studies have analyzed MpBC, only one simple comparison with IDC was performed and did not adjust for prognostic factors of MpBC. There are no treatment guidelines specific to the management of MpBC, which is frequently misdiagnosed or unrecognized on pathologic review because of its rarity and heterogeneity . MpBC tends to be over-treated compared to typical IDC, even at a lower pathologic stage because of its aggressiveness . However, there are few studies comparing clinical features and prognosis between MpBC and IDC because of the rarity of MpBC. In the present study, we investigated the characteristics of MpBC in the real world and compared prognosis between MpBC and IDC using propensity-score matching to adjust confounding factors. 2. Patients and Methods 2.1. Study Population We retrospectively reviewed the clinical data of 27,675 patients with breast cancer who were treated with surgery from 1994 to 2019 in the Samsung Medical Center. We investigated extensive pathologic findings about histopathology, multiplicity, lymphatic invasion, extensive intraductal component, nuclear grade, and histologic grade. We reviewed molecular pathologic markers of estrogen receptor status, progesterone receptor status, and HER2 status. We also investigated patients who underwent chemotherapy, radiation therapy, or hormone therapy instead of surgery. 2.2. Propensity-Score Matching Propensity-score matching (PSM) is a method for filtering experimental and control cases of similar characteristics, which are called the matching variables, for comparison in a retrospective analysis . PSM can be used to adjust for baseline characteristics and reduce the effect of selection bias. The variables for PSM in this study were age (years), histopathology, multiplicity, lymphatic invasive, extensive intraductal component, nuclear and histologic grade, tumor size, nodal status, estrogen receptor, progesterone receptor, HER2, hormonal therapy, chemotherapy, and radiation therapy. We used a ratio of 1:4 for nearest neighbor matching within 0.2 standard deviations of the logit of the propensity score. The propensity score was analyzed using logistic regression. We assessed the balance of covariates in the two groups after PSM, the results of which are shown in Supplementary Table S1. 2.3. Statistical Analyses Patient characteristics were compared using the independent t-test for continuous variables and the chi-square or Fisher's exact test for categorical variables. Univariable and multivariable analyses were conducted using Cox regression analysis models. After excluding patients due to missing values, Cox regression analysis models were tested on 12,905 out of 16,406 total patients. Statistical significance was established at p < 0.05. All statistical analyses were performed using the Statistical Analysis System (SAS) version 9.4 (SAS Institute Inc., Cary, NC, USA). 3. Results We investigated 27,675 breast cancer patients treated with surgery from 1994 to 2019 at the Samsung Medical Center. We excluded patients with stage IV cancer or carcinoma in the breast at the time of diagnosis and patients receiving neoadjuvant chemotherapy. We also excluded patients confirmed with a mixed tumor subtype greater than grade two on final pathology to remove confounding variables and analyze by pure subtype. Among 18,370 cases of invasive carcinoma, the ductal type comprised 88.5% (16,251 patients), and the metaplastic type was 0.8% (155 patients); other types, such as lobular, mucinous, and papillary, accounted for 10.8% of the population (1964 patients). After excluding patients with missing values, the completed set was analyzed during propensity-score matching. Finally, after 1:4 propensity-score matching, 478 patients of ductal type and 120 patients of metaplastic type were compared . Demographics and clinical characteristics before PSM are shown in Table 1. There was no difference in age between the two groups, but there were differences in all other variables. Multiplicity, lymphatic invasion, and extensive intraductal component (EIC) were significantly higher in the ductal group than the metaplastic group. Nuclear and histologic grade and tumor size were significantly higher in the metaplastic group than the ductal group. In the metaplastic group, the proportion of patients with estrogen receptor positivity was about 10%, and more than 95% of cases were negative for progesterone receptor and HER2. About 60% of the ductal group received adjuvant chemotherapy, compared with more than 90% of the metaplastic group. After 1:4 PSM, there was no difference in any variables except progesterone receptor status (Supplementary Table S2). The median follow-up duration of the overall population was 68 months (range 3-277 months). Disease-free survival (DFS) of the ductal group was 89.98%, and that of the metaplastic group was 78.06%, with a hazard ratio of 2.893 (95% confidence interval (CI): 1.929-4.340) before PSM (p = 0.0002). Overall survival (OS) of the ductal group was 94.04%, and that of the metaplastic group was 85.16%, with a hazard ratio of 3.839 (95% CI: 2.289-6.439) before PSM (p = 0.0140) . After 1:4 PSM, the DFS of the ductal group was 87.45%, and that of the metaplastic group was 82.50%, with an HR of 1.465 (95% CI: 0.882-2.432) (p = 0.1398). The OS of the ductal group was 91.84%, and that of the metaplastic group was 88.33%, with an HR of 1.542 (95% CI: 0.875-2.718) (p = 0.1340) . We analyzed the clinically significant prognostic factors using a Cox regression model for univariable and multivariable analyses. This analysis showed lymphatic invasion, histologic grade, pT and pN stage, ER, radiotherapy, and chemotherapy as significant factors for DFS and OS (Table 2, Table 3, Table 4 and Table 5). Inclusion in the metaplastic group was a significant, poor prognostic factor for DFS (HR = 2.240; 95% CI: 1.476-3.399, p = 0.0002) and OS (HR = 1.969; 95% CI: 1.147-3.382, p = 0.0140). 4. Discussion As shown in previous studies, metaplastic breast cancer has more aggressive features than other types . Patients with MpBC present with a higher histologic grade, with less nodal involvement, and at an advanced stage. Due to the advanced stage, their prognosis is poorer than that of IDC. However, after performing PSM to correct for variables that affect a tumor, there were no differences in disease-free and overall survival between MpBC and IDC groups. MpBC is a rare disease that is clinically characterized by a poor prognosis and has shorter disease-specific survival than IDC. Owing to this, when MpBC is diagnosed, there is a tendency toward aggressive treatment . However, the intrinsic nature of MpBC might not significantly differ from that of IDC based on our propensity-score-matched data. It is possible that we did not have the correct general knowledge of the nature of MpBC because of a lack of comparative analysis of MpBC and IDC using PSM. Although MpBC is a rare disease, the histological subtypes and genomic profiles of MpBC are diverse. Histologically, MpBC tumors are heterogeneous, poorly differentiated tumors with epithelial and mesenchymal components, including ductal carcinoma . The epithelial types of MpBC are largely divided into squamous cell carcinoma, adenocarcinoma with spindle cell differentiation, and adenosquamous carcinoma. The mesenchymal types of MpBC consist of carcinoma with chondroid metaplasia, carcinoma with osseous metaplasia, and carcinosarcoma. Some molecular and genetic features of MpBC are known to be involved in its pathogenesis, including the epithelial-mesenchymal transition pathway; the mitogen-activated protein (MAP) kinase signaling pathway; the phosphoinositide 3-kinase (PI3K) pathway; and the epithelial growth factor pathways, including those of protein kinase B (Akt), mammalian target of rapamycin (mTOR), and epithelial growth factor receptor (EGFR). MpBC shows aggressive clinical characteristics, including a larger tumor size than other subtypes and histologically high nuclear and tumor grades. Oberman reported that a tumor size less than 4 cm was associated with a good prognosis, and that tumor size at the time of diagnosis was correlated with prognosis . MpBC mainly demonstrates hematogenous spread and is prone to metastasize to other organs, thereby having a poor prognosis. Additionally, patients with MpBC are generally over 60 years old, and those who have distant metastases and do not receive radiotherapy have worse prognoses in terms of disease-free survival and overall survival than patients with IDC. Thus, when MpBC is diagnosed, chemotherapy and radiotherapy generally are performed, along with surgical treatment. In this study, in the MpBC group, chemotherapy was performed in 90% of cases, in contrast to 60% in the ductal group. MpBC is thought to be chemoresistant, even in a metastatic setting . However, the diverse histological subtypes and genomic profiles of MpBC result in diverse response to chemotherapy. Yam C et al. reported a 23% pathologically complete response (pCR) rate in MpBC patients treated with neoadjuvant chemotherapy (NAC) and suggested that patients with MpBC receiving NAC should have careful monitoring of cancer progression because its pCR rate is lower than that of ductal TNBC . In patients with clinically node-negative MpBC, NAC was linked to worse outcomes compared to adjuvant chemotherapy, but no such association was observed for those with clinically node-positive MpBC. Thus, it is inaccurate to describe MpBC as chemoresistant because the responsiveness to such therapy should be evaluated considering various factors, such as tumor size, nodal status, nuclear and histologic grade, and Ki-67. Although MpBC can be resistant to chemotherapy, responsiveness and survival may differ by subtype. Schroeder et al. reported that survival of HER2-positive MpBC was improved and similar to that of HER2-positive IDC, and suggested that HER2-positive MpBC might be responsive to additional HER2-directed therapy . When considering chemoresistance, the regimen should be assessed . Through the development of immuno-oncology (IO), targeted therapy is in the spotlight for breast cancer. MpBC frequently involves tumor infiltrating lymphocytes (TILs) in the tumor microenvironment, sharing many features with the basal-like subtype . Recent studies have demonstrated frequent overexpression of programmed death-ligand 1 (PD-L1) in MpBC, prompting interest in combining immune checkpoint inhibitors with conventional chemotherapy . Use of IO for MpBC can be improved by identifying immune checkpoint inhibitor markers that increase expression in TIL. Although systemic therapy is important for the treatment of MpBC, radical surgical treatment with sufficient margin should not be overlooked for locoregional control. Surgical treatment of MpBC is as successful as surgery for IDC . In MpBC patients, mastectomy is performed at a higher rate than in IDC patients because of the larger tumor size . In our study, the mastectomy rate was not high because there were many patients with pathologic T2 staging and few with pathologic T1 and T3 stages. However, no difference in disease-free survival or overall survival has been shown between mastectomy and lumpectomy . The survival rate for MpBC compared to other types of breast cancer is not affected by the extent of surgical intervention. However, LI XIA et al. reported that overall survival was significantly higher in patients who received lumpectomy with radiation therapy compared to those who had a mastectomy . MpBC generally demonstrates hematogenous rather than lymphatic metastases. Several studies have reported that MpBC patients had a lower rate of nodal involvement in the axillary area than IDC patients . Tseng and Martinez reported 22% of MpBC patients had a metastatic axillary lymph node, and Pezzi et al. also reported 22% of MpBC patients had axillary lymph node involvement, in contrast with 34% of IDC patients . Even if MpBC metastasizes less frequently to lymph nodes than does IDC, axillary lymph node surgery is important for local control and clear staging for adjuvant therapy. Although the rate of nodal involvement is low in MpBC, axillary nodal surgery is important because lymph-node metastases are present in about 20% of cases. The main limitation of this study is that it did not present a difference in prognosis according to the subtype of MpBC because histologic grade, hormonal receptor status, and HER2 depend on subtype. Therefore, it is important to establish a molecular subtype and to target therapy or immunotherapy rather than basic treatment. Another limitation is that the study was conducted at a single institution with a retrospective design and was not a randomized controlled study (RCT). Despite these limitations, the large number of patients might allow a representative conclusion because MpBC is rare and RCT is difficult to perform in this type of study. 5. Conclusions MpBC is an aggressive histologic subtype and has a poor prognosis compared with IDC. Based on PSM, there were no differences in disease-free and overall survival between MpBC and IDC. However, since there was a difference in tendency pf around 1.5 times in the hazard ratio, it is difficult to say that there is no complete difference. This suggests that treatment of MpBC should be in line with the treatment principles of the TNBC subtype of aggressive IDC. In the future, additional studies on the characteristics of MpBC, and new treatment modalities, such as IO, are expected to be applied. Supplementary Materials The following supporting information can be downloaded at: Table S1: Summary of balance table of propensity-score matching. Table S2: Patient demographic and clinical characteristics after propensity-score matching. Click here for additional data file. Author Contributions Conceptualization, J.Y. and J.-H.L.; methodology, B.J.C. and J.E.L.; formal analysis, J.Y. and J.-H.L.; investigation, J.M.R. and S.K.L.; data curation, B.J.C. and S.W.K.; writing-original draft preparation, J.-H.L.; writing--review and editing, J.Y., J.M.R. and S.K.L.; supervision, S.W.K. and S.J.N. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The present study was approved by the appropriate review committees (no. 2021-08-131) and was conducted according to the principles outlined in the Declaration of Helsinki (approval date: 26 August 2021). Informed Consent Statement The need for informed consent from patients was waived by the institutional review board due to the retrospective study design. Data Availability Statement The datasets used and/or analyzed during the study are available from the corresponding author, on reasonable request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Consort diagram of this study. Figure 2 Disease-free survival and overall survival before matching. (A) Disease-free survival before matching; (B) Overall survival before matching. Figure 3 Disease-free survival and overall survival after matching. (A) Disease-free survival after matching; (B) Overall survival after matching. cancers-15-01556-t001_Table 1 Table 1 Patient demographics and clinical characteristics before propensity-score matching. Metaplastic N (%) Ductal N (%) p-Value Age (years) 0.8205 Mean 50.16 49.87 Std 9.94 11.34 Multiplicity <0.0001 No 148 (95.48) 12,483 (76.81) Yes 7 (4.52) 3725 (22.92) Unknown 0 (0.00) 43 (0.27) Lymphatic invasion 0.0001 No 126 (81.29) 10,868 (66.88) Yes 23 (14.84) 4787 (29.46) Unknown 6 (3.87) 43 (3.66) EIC <0.0001 No 130 (83.87) 10,506 (64.65) Yes 14 (9.03) 4866 (29.94) Unknown 11 (7.1) 879 (5.61) Nuclear grade <0.0001 Low 1 (0.64) 1894 (11.65) Intermediate 21 (13.55) 8761 (53.91) High 129 (83.23) 5402 (33.24) Unknown 4 (2.58) 194 (1.20) Histologic grade <0.0001 Good 4 (2.58) 3456 (21.27) Moderate 22 (14.19) 7443 (45.80) Poor 124 (80.00) 4957 (30.50) Unknown 5 (3.23) 395 (2.43) pT stage <0.0001 T1 42 (27.10) 9882 (60.81) T2 102 (65.81) 5837 (35.92) T3 10 (6.45) 501 (3.08) T4 1 (0.65) 26 (0.16) Unknown 0 (0.00) 5 (0.03) pN stage 0.0002 N0 124 (80.00) 10,257 (63.12) N1 25 (16.13) 4275 (26.31) N2 5 (3.23) 1070 (6.58) N3 1 (0.65) 595 (3.66) Unknown 0 (0.00) 54 (0.33) Estrogen receptor <0.0001 Negative 135 (87.10) 3751 (23.08) Positive 19 (12.26) 12,361 (76.06) Unknown 1 (0.64) 139 (0.86) Progesterone receptor <0.0001 Negative 146 (94.19) 4972 (30.60) Positive 8 (5.16) 11,133 (68.51) Unknown 1 (0.65) 146 (0.89) HER2 <0.0001 Negative 142 (91.61) 12,079 (74.33) Positive 9 (5.81) 3475 (21.38) Unknown 4 (2.58) 697 (4.29) Hormone therapy <0.0001 No 131 (84.52) 3736 (22.99) Yes 21 (13.55) 12,246 (75.36) Unknown 3 (1.93) 269 (1.65) Radiation therapy 0.0094 No 26 (16.77) 4267 (26.26) Yes 124 (80.00) 11,678 (71.86) Unknown 5 (3.23) 306 (1.88) Chemotherapy <0.0001 No 11 (7.10) 5558 (34.20) Yes 140 (90.32) 10,435 (64.21) Unknown 4 (2.58) 258 (1.59) Abbreviations: EIC = extensive intraductal component; pT = pathologic T stage; pN = pathologic N stage; N = number of patients. cancers-15-01556-t002_Table 2 Table 2 Univariable analysis of disease-free survival (DFS). DFS Univariable Analysis (12,905 Patients) HR (95% CI) p-Value Age (years) 0.982 (0.975-0.989) <0.0001 Histology Metaplastic vs. Ductal 2.893 (1.929-4.340) <0.0001 Lymphatic invasion Negative vs. Positive 0.498 (0.348-0.567) <0.0001 Histologic grade Moderate vs. Good 2.419 (1.866-3.136) <0.0001 Poor vs. Good 3.923 (3.039-5.064) <0.0001 pT stage pT2 vs. pT1 1.960 (1.665-2.309) <0.0001 pT3 vs. pT1 3.215 (2.350-4.400) <0.0001 pT4 vs. pT1 4.022 (1.208-13.388) 0.0168 pN stage pN1 vs. pN0 1.365 (1.136-1.642) 0.0002 pN2 vs. pN0 2.086 (1.609-2.704) <0.0001 pN3 vs. pN0 4.778 (3.704-6.164) <0.0001 Estrogen receptor Negative vs. Positive 1.923 (1.683-2.197) <0.0001 Radiotherapy Negative vs. Positive 1.556 (1.354-1.789) <0.0001 Chemotherapy Negative vs. Positive 0.642 (0.547-0.755) <0.0001 Abbreviation: pT = pathologic T stage; pN = pathologic N stage. cancers-15-01556-t003_Table 3 Table 3 Multivariable analysis of disease-free survival (DFS). DFS Multivariable Analysis (12,905 Patients) HR (95% CI) p-Value Age (years) 0.981 (0.975-0.988) <0.0001 Histology Metaplastic vs. Ductal 2.240 (1.476-3.399) 0.0002 Lymphatic invasion Negative vs. Positive 0.718 (0.615-0.839) <0.0001 Histologic grade Moderate vs. Good 2.006 (1.531-2.627) <0.0001 Poor vs. Good 2.785 (2.071-3.745) <0.0001 pT stage pT2 vs. pT1 1.377 (1.145-1.655) 0.0001 pT3 vs. pT1 1.751 (1.234-2.485) 0.0004 pT4 vs. pT1 2.699 (0.800-9.108) 0.1524 pN stage pN1 vs. pN0 1.218 (0.988-1.501) 0.0716 pN2 vs. pN0 1.897 (1.405-2.553) <0.0001 pN3 vs. pN0 3.975 (2.898-5.451) <0.0001 Estrogen receptor Negative vs. Positive 1.553 (1.325-1.820) <0.0001 Radiotherapy Negative vs. Positive 1.815 (1.565-2.106) <0.0001 Chemotherapy Negative vs. Positive 1.797 (1.468-2.200) <0.0001 Abbreviation: pT = pathologic T stage; pN = pathologic N stage. cancers-15-01556-t004_Table 4 Table 4 Univariable analysis of overall survival (OS). OS Univariable Analysis (12,905 Patients) HR (95% CI) p-Value Age (years) 1.035 (1.025-1.045) <0.0001 Histology Metaplastic vs. Ductal 3.839 (2.289-6.439) <0.0001 Lymphatic invasion Negative vs. Positive 0.425 (0.350-0.514) <0.0001 EIC Negative vs. Positive 1.374 (1.106-1.708) 0.0042 Histologic grade Moderate vs. Good 2.358 (1.545-3.600) <0.0001 Poor vs. Good 5.275 (3.518-7.910) <0.0001 pT stage pT2 vs. pT1 2.996 (2.321-3.869) <0.0001 pT3 vs. pT1 5.416 (3.552-8.256) <0.0001 pT4 vs. pT1 5.678 (1.032-31.225) 0.0443 pN stage pN1 vs. pN0 2.053 (1.551-2.716) <0.0001 pN2 vs. pN0 3.216 (2.232-4.633) <0.0001 pN3 vs. pN0 7.272 (5.110-10.350) <0.0001 Estrogen receptor Negative vs. Positive 2.972 (2.455-3.598) <0.0001 Radiotherapy Negative vs. Positive 1.440 (1.167-1.777) 0.0007 Chemotherapy Negative vs. Positive 0.706 (0.556-0.898) 0.0045 Abbreviation: EIC = extensive intraductal component; pT = pathologic T stage; pN = pathologic N stage. cancers-15-01556-t005_Table 5 Table 5 Multivariable analysis of overall survival (OS). OS Multivariable Analysis (12,905 Patients) HR (95% CI) p-Value Age (years) 1.024 (1.014-1.034) <0.0001 Histology Metaplastic vs. Ductal 1.969 (1.147-3.382) 0.0140 Lymphatic invasion Negative vs. Positive 0.662 (0.526-0.834) 0.0005 EIC Negative vs. Positive 1.160 (0.923-1.457) 0.2028 Histologic grade Moderate vs. Good 1.731 (1.114-2.689) 0.0105 Poor vs. Good 2.710 (1.699-4.322) <0.0001 pT stage pT2 vs. pT1 1.823 (1.365-2.436) <0.0001 pT3 vs. pT1 2.947 (1.835-4.733) <0.0001 pT4 vs. pT1 1.535 (0.269-8.749) 1.0000 pN stage pN1 vs. pN0 1.851 (1.355-2.531) <0.0001 pN2 vs. pN0 2.654 (1.744-4.039) <0.0001 pN3 vs. pN0 5.012 (3.223-7.793) <0.0001 Estrogen receptor Negative vs. Positive 2.437 (1.936-3.067) <0.0001 Radiotherapy Negative vs. Positive 1.636 (1.299-2.062) <0.0001 Chemotherapy Negative vs. Positive 2.576 (1.900-3.492) <0.0001 Abbreviation: EIC = extensive intraductal component; pT = pathologic T stage; pN = pathologic N stage. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000577 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050750 healthcare-11-00750 Article Key Factors for Enhancing Home Care Workers' Intention to Stay by Multiple-Criteria Decision Analysis Hsu Wei Conceptualization Methodology Validation Formal analysis Investigation Writing - original draft Visualization Supervision Project administration Funding acquisition 1* Shih Fang-Ping 2 Nakano Hideki Academic Editor Ramirez Gilbert Academic Editor 1 Department of Health Care Management, National Taipei University of Nursing and Health Sciences, Taipei 112303, Taiwan 2 Department of Human Resources--Personnel Section, Koo Foundation Sun Yat-Sen Cancer Center, Taipei 11259, Taiwan * Correspondence: [email protected] 03 3 2023 3 2023 11 5 75016 1 2023 25 2 2023 28 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). The ageing population is increasing rapidly in Taiwan, where the ageing rate exceeds even that of Japan, the United States and France. The increase in the disabled population and the impact of the COVID-19 pandemic have resulted in an increase in the demand for long-term professional care, and the shortage of home care workers is one of the most important issues in the development of such care. This study explores the key factors that promote the retention of home care workers through multiple-criteria decision making (MCDM) to help managers of long-term care institutions retain home care talent. A hybrid model of multiple-criteria decision analysis (MCDA) combining Decision-Making Trial and Evaluation Laboratory (DEMATEL) and the analytic network process (ANP) was employed for relative analysis. Through literature discussion and interviews with experts, all factors that promote the retention and desire of home care workers were collected, and a hierarchical MCDM structure was constructed. Then, the hybrid MCDM model of DEMATEL and the ANP was used to analyze the questionnaire data of seven experts to evaluate the factor weights. According to the study results, the key direct factors are improving job satisfaction, supervisor leadership ability and respect, while salary and benefits are the indirect factor. This study uses the MCDA research method and establishes a framework by analyzing the facets and criteria of different factors to promote the retention of home care workers. The results will enable institutions to formulate relevant approaches to the key factors that promote the retention of domestic service personnel and to strengthen the intention of Taiwan's home care workers to stay in the long-term care industry. home care workers intention to stay multiple-criteria decision analysis (MCDA) decision-making trial and evaluation laboratory (DEMATEL) analytic network process (ANP) National Taipei University of Nursing & Health Sciences107EH12-506 111D019 This research was funded by National Taipei University of Nursing & Health Sciences, under grant number 107EH12-506 and 111D019 to Wei Hsu. The APC was funded by National Taipei University of Nursing & Health Sciences. pmc1. Introduction It is estimated that the proportion of the elderly population in the European Union member states will continue to grow . In the Italian social welfare system, hospital facilities for providing social health care to the elderly are limited. In conventional forms of old-age care, although systems and governance vary from region to region, care for the elderly can also be handled at the regional level by nonprofit institutions such as social cooperatives, which receive funding from municipalities and government departments to provide a range of regional services for the elderly . With the increase in the ageing population and the general miniaturization of the family structure in Taiwan, the care needs of the disabled elderly have risen sharply. The proportion of the disabled population will also increase significantly, resulting in an increase in the demand and burden of long-term care that will aggravate the shortage of home care attendants. The services of home care attendants range from companionship to life assistance, personal hygiene and cleaning, rehabilitation activities and housework. Since the content of home care services is quite diverse, there are life assistants and care attendants, both of which are important and indispensable roles for home care workers on the front line. Therefore, when implementing follow-up care plans for individual cases, the presence and retention of staff are important in regard to whether home care can be offered. Simply training and cultivating diverse and professional care service staff takes a large amount of time. In addition, the relationship and emotional challenges faced by home care attendants underscore the need for training . The number of trainees and the turnover rate are key to the rapid growth of the home care service industry. In terms of the ageing rate in Taiwan, the number of people requiring long-term care services is rising, and with the limited government resources, long-term home care-related industries are facing changes in the social structure. The shortage of health care staff is a serious and chronic problem for long-term care in the 21st century, and research has shown that the reasons include disrespect, cultural influences, too much paperwork and changes in work style . Due to the high turnover rate of resident staff and the decline in retention, the lack of staff will burden the management operations if it is not easy to find new resident staff and cultivate service staff talent. Long-term care institutions and the government must face this problem squarely; otherwise, it will lead to high labour costs, and the quality of care in overloaded long-term care institutions will be reduced. Thus, the issue of home care workers' retention should be highlighted. Home care services were provided by volunteers in the past, but in recent years, they have mainly been provided by hiring regular staff due to the long-term care policy. Taiwan is under the influence of socioeconomic changes, and people who participate in home care services have different personal backgrounds and service perceptions. Because the human capital of home care services is important for the long-term care of elderly and disabled individuals and home care attendants' willingness to stay in their jobs can directly affect the quality of long-term care services, the retention of home care attendants must be evaluated and analyzed to understand their needs. Numerous factors influence home care workers' intention to stay. There have been many past studies on the factors influencing home care attendants' willingness to stay, and most studies have used expert analysis methods, regression analysis methods and structural equations. Each method has advantages and disadvantages. For example, the expert research method is fast but overly subjective. The regression analysis method, under the assumption of a normal matrix, needs a large sample for research and discussion and cannot directly deal with the problem of category variable data. Multiple-criteria decision making (MCDM) is well suited to solve such problems. MCDM was designed to solve the uncertainty of planning (such as technology and risk) with multiple criteria and uses the structured method to stratify complex problems. The analytic hierarchy process (AHP) in MCDM has been used in different fields to discuss the brain drain and human capital care of enterprises, but the factors in the architecture may have interdependence or feedback relationships. If only the AHP method is considered, the causal relationships among the criteria cannot be considered, and the final overall priority weight is calculated using the average method. The weight of the final evaluation result will be higher or lower than the real situation, and there will be inaccuracies . The analytic network process (ANP) can integrate the relationship among factors and AHP concepts. This study categorized the factors identified in the literature review to establish a preliminary research framework, considering the interaction among the criteria for promoting the retention of home care workers, which cannot be fully explained by the structure of the AHP. Therefore, it adopted the characteristics of solving the problem of internal dependence and external feedback of the cluster ANP. To determine the relationships among the criteria of home care workers' intention to stay, this study employed the powerful Decision-Making Trial and Evaluation Laboratory (DEMATEL) method, which uses expert questionnaires and graphical causal theory to understand the dependence and feedback relationship among the criteria and can be used to explore the core objectives and key criteria and to transform the degree of interaction between complex facets or clusters into characteristics of causality. DEMATEL can introduce the interaction impact weight among clusters into the ANP step and establish the relationship structure for the ANP method. To clarify the causal relationship and influence among the criteria, the MCDA model in this study combined DEMATEL and ANP to identify the key factors that promote home care attendants' intention to stay. Therefore, this study adopted the MCDA model, a combination of the DEMATEL and ANP methods, to establish a framework to determine the key factors that could significantly increase home care attendants' willingness to stay. There are many work difficulties for home care workers in Taiwan, and some negative elements cause high turnover rates of home care employees, so workers in home care service institutions should be vaccinated. The findings of this study can provide managers of long-term care institutions with an understanding of the reasons for promoting the retention of front-line staff and a reference for formulating corresponding strategies or human resource management methods to reduce large organizational labour costs and loads. This study can improve the staffing problem of home care attendants and serve as a priority reference for long-term care institutions to improve the retention of home care attendants in the future. In addition, the findings may be helpful to provide concepts for policy makers to overcome the current global shortage of long-term care staff. This paper also enhances the human resources literature by demonstrating the key factors that promote home care workers' intention to stay. 2. Literature Review Due to the annually increasing needs of home care services and the lack of home care workers, many scholars have begun to pay attention to home care attendants' intention to stay . As home care providers face high turnover and problems in recruiting new employees, they should understand the factors that lead to the loss of home care staff in order to effectively provide better-quality home care . In particular, employees need to feel valued for their contribution to the organization, and how they measure their willingness to stay is closely related to their motivation to stay. Employees' retention can be measured in terms of burnout, organization and tendency to leave, and when employees subjectively believe that their needs are not met, they will inevitably not contribute to the organization's goals or tasks and will eventually choose to leave . Employees' different perceptions of supervisors' leadership styles can also lead to differences in their work attitudes . For example, the more respectful managers are, the more engaged their subordinates are, and their willingness to stay then increases. The external retention motivation created by the organization effectively encourages caregivers to continue in the profession. Retention factors are particularly important for young or novice caregivers with precarious employment status. Many caregivers believe that internal retention motivations, such as commitment, conscience, interest in caregiving and responsibility, are the most important factors in convincing them to stay in the profession . To identify the factors that strengthen and promote the retention intention of home care workers, the relevant literature was searched and collected, and the factors that scholars have found to promote employees retaining their intention to stay were comprehensively sorted. Then, appropriate evaluation criteria and sub-criteria were determined, and three aspects of the assessment criteria are summarized as "working environment", "organization" and "future development". The selected assessment components and criteria variables are explained in detail by means of a literature review. 2.1. Working Environment Factors related to the desired working environment include relationship issues, i.e., problems with employers, supervisors, coworkers, doctors or subordinates. Regarding the aspect of the working environment, there may be bullying and harassment in the workplace. Emotional difficulties are composed of a lack of psychological support and excessive emotional needs. Problems associated with time pressure and quality of care include too many responsibilities, too many tasks, fear of making mistakes, making mistakes, not having enough time to care for patients, reduced conditions for caring for patients and time pressure leading to suboptimal work quality. In addition, employee independence was listed as a possible factor in the decision to leave. Work scheduling difficulties include unsatisfactory work plans, chaotic shifts, too many night shifts and too much overtime. Dissatisfaction with pay includes personal income needs, workload and low wage trends . A number of scholars have proposed that the relevant factors of the working environment that promote employee retention include salary and benefits, flexible scheduling, low work pressure and workload, "good relationships with clients" and "avoiding workplace harassment". 2.1.1. Salary and Benefits Salary and benefits refer to any form of financial compensation and welfare received by employees in a way that motivates their morale and improves their work efficiency so that employees with the required knowledge and ability can be retained . Employees' attitude towards benefits is usually an important predictor of their turnover rate, and salary satisfaction is quite important for employees. Some scholars have used a combination of customary domain and AHP to explore factors associated with the tendency to leave. The expert interviews revealed that the top five criteria are salary, health, bonus benefits, promotion opportunities and family care, with salary and benefits being among the top three criteria to which employees pay the most attention . According to equity theory, when the staff perceive a mismatch between the pays/benefits ratio and their comparison to others' ratios, they experience inequity. Associations were found between staff equity perceptions of benefits and intention to leave for staff working in services for individuals . 2.1.2. Flexible Scheduling Flexible scheduling refers to shift work, which usually refers to variable working hours, that is, working hours that are not fixed hours and schedules. In the face of changes in output demand, the shift system can adjust the amount of staffing in a timely manner and reduce labour costs, and if a flexible and coordinated shift and scheduling system is achieved, it can effectively promote employees' intention to stay . Home care attendants have a higher chance of working shifts or fixed night shifts, which not only increases work pressure but also reduces their willingness to stay. Some home care cases require 24-h care, so caregivers may be required to work shifts, and employment contracts between caregivers and institutions may affect and limit the shifts or scheduling patterns of caregivers. 2.1.3. Low Work Pressure and Workload Reducing work stress and load refers to improving employees' overall health, increasing their performance and reducing burnout. For example, staff work independently or with a small number of colleagues, which increases their work pressure . A study in Sweden using a cross-sectional survey design to explore the effects of work stress on dementia care specialists and other staff in home care services using t-tests and multiple linear regression analysis showed that due to the expected increase in the population of elderly people living at home, the need for extensive and complex health care is increasing, and there is a greater need for staff to work alone . This strengthens their work pressure and generates more stress . 2.1.4. Good Relationships with Clients During the work of the home care attendant, there may be difficulties in interaction between the attendant and the client, and the family may also interfere with the provision of services or disagree with the home care attendant due to different views on the care, which may affect the progress of the care work. For home care attendants, communication problems with clients and families are one of the possible work difficulties. Salary and benefits, training, supervision and institutional support and relationship with the client are important factors in the retention of home care attendants . If the home care attendant does not establish a good relationship with the client, the emotional problems reflected by the client may cause the attendant to reduce his or her intention to stay. Therefore, in the process of work, if attendants feel the respect of the client and the client's family, like their relationship with the client, enjoy the feeling of being needed and expect to help make others' lives better, their willingness to stay will be higher . 2.1.5. Avoiding Workplace Harassment Avoiding workplace harassment means that when a resident staff member is required to enter the home to provide services, when there are concerns about their personal safety in relation to a person of the opposite sex, such as sexual harassment, measures should be put in place to avoid problems . The workplace of the home attendant is the area of the client's life, and some activities or behaviors in the household's private life may not be appropriate when the care attendant is present. One study of 1214 care workers found that 12.8% had experienced sexual assault and 27.6% had experienced sexual harassment and felt that their personal safety was compromised when dealing with issues such as sexual harassment . 2.2. Organization An organization is a social entity that has a specific structure and a system of coordinated activities and exists to achieve a specific goal. Organizations are formed by relationships between people, and organizations already exist when people interact with each other to achieve goals . In the connection between the organization and employees, employees' identification with the organization plays an important role. This study synthesizes the past literature and proposes the relevant factors of organizations that promote employee retention. Therefore, the organizational assessment includes professional image, respect, sense of accomplishment, organizational commitment, job satisfaction and supervisor leadership style. 2.2.1. Professional Image Professional image refers to the beginning of the employee's professional experience, which is gradually internalized in the formation of professional behavior. After the profession interacts with its environment, the public forms an image of the profession in its best state . An Iranian study showed that the general public's attitude towards the nursing profession is not positive. The lack of understanding of the professional image of patients and their families is another factor affecting caregivers, and the turnover rate of home care workers should be improved by improving working conditions and raising the social status and image of the profession . At present, the public's perception of care attendants is part of the impression of care workers, making it difficult for them to pay attention to their profession. On the one hand, the dignity of the workplace cannot be improved. On the other hand, because of the lack of career promotion channels, coupled with low salary levels and lack of labour condition protection, even if there are long-term care students who graduate and obtain a resident service license, their willingness to enter the long-term care industry is low, resulting in a shortage of front-line personnel. The serious shortage not only leads to the loss of long-term care staff at the grassroots level but also wastes investment in educational resources. Therefore, government agencies should strengthen the professional image of resident service workers, increase professional recognition and enable those who wish to participate in the long-term care industry to clearly understand the progression from home care workers to instructors and supervisors so that they can develop professional competence and qualifications. 2.2.2. Respect Past research has mentioned that caregiver self-improvement, effective communication and support from managers to caregivers, integration of organizational structures, improved working conditions and creation of caregiver cultural competencies can promote a respectful working environment and reduce the likelihood of staff departure. Current changes in social value systems and implicit social norms are considered determinants of respectful workplaces, while caregiver respect refers to showing respect for others, perceived respect for the environment and self-respect. Sociocultural factors are challenges that affect respect for the client and his or her family and other members of the care team . 2.2.3. Sense of Accomplishment Sense of accomplishment refers to a deep-seated motivation for individuals to engage in challenging tasks and improve their performance to achieve their goals . In addition, increasing a personal sense of accomplishment can also reduce burnout and increase intention to stay . High burnout and severe declines in job satisfaction and personal sense of accomplishment are obvious problems for caregivers in China, and improving the working environment may be an effective strategy for health care organizations to improve care outcomes . 2.2.4. Organizational Commitment Organizational commitment refers to the degree of individual recognition of and commitment to a particular organization, creating a work team with high performance and loyalty to the organization. This enables members to realize their potential and voluntarily accept the goals and values of the organization so that they will have a strong willingness to stay in order to effectively take advantage of the organization's abundant human resources . Managing and monitoring employee turnover is an important retention factor for organizations, as are determining appropriate criteria and minimizing negative consequences with effective solutions. 2.2.5. Job Satisfaction Job satisfaction refers to the employee's feelings about the enterprise or aspects of it and is a subjective response to the work situation, including the individual's physical and psychological satisfaction with the working environment and the work itself and the degree to which he or she likes (is satisfied with) or does not like (is dissatisfied with) his or her work. Expressing job satisfaction represents the extent to which an individual experiences happiness in an organizational context . According to a cross-sectional Australian study, job satisfaction is the only important predictor of willingness to leave, so understanding and addressing job satisfaction is important. Understanding the experience of caregivers contributes to workforce development, supporting skilled caregivers in working within their area of practice, retaining experienced caregivers and supporting the recruitment of new caregivers . Facility managers and human resources practitioners need to develop programs to increase caregivers' job satisfaction, thereby fostering caregivers' sense of responsibility and increasing their retention. In addition, dissatisfaction with the opportunity to use competence and autonomy is caused by more responsibilities, high career challenges, not enough freedom of decision making, difficulty of work, too much routine work, etc. These problems will also affect retention factors, and the lack of improvement can decrease the chances of employee retention . 2.2.6. Supervisor Leadership Style Home care supervisor leadership refers to the two-way interaction between supervisors and staff, which is important for interpersonal relationships in the workplace. If supervisors do not identify their own emotional problems or recognize employees' feelings, which can easily lead to the breakdown of interactive relationships, they cannot lead employees to reduce negative emotions to improve work results and prevent burnout and separation . Good leadership is considered a superior model because it generates greater trust and motivation among subordinates. Caregivers experiencing work-related bullying and burnout are more likely to want to leave their jobs and careers, highlighting the importance of leadership management in preventing negative outcomes for employees and organizations . Additionally, establishing strong, authentic leadership and management can help reduce caregiver burnout and turnover propensities . The education and support function of the supervision system can reduce the impact of risk factors in the working environment and have a positive effect on the continuation of staff members' careers; when the education, support and coordination functions of the supervisor are better executed, the willingness of the staff to stay will be higher . 2.3. Future Development Each decision to leave is driven by identifiable factors, such as a desire to change careers, start a business, seek opportunities or switch companies. Due to its negative impact, employee turnover is a challenge for many companies. Increased employee turnover often reduces a company's performance, especially for smaller companies with limited resources. When an employee resigns, the company must look for a replacement (high cost), which includes recruitment and training costs . Therefore, the assessment of future employee development, including training programs, promotion opportunities and career development, is regarded as an important retention factor in the working environment of employees. 2.3.1. Training Programs Training programs refers to one of the ways an organization can persuade employees that the organization values them and that their development is to prepare for future jobs in addition to effectively performing current work. Training and development is also a radical change in employees' knowledge, attitudes and skills to improve the quality and efficiency of their work, representing a significant investment in human capital . Employee training programs affect employee turnover, and as management training increases, employees become less willing to leave . To improve employee retention, companies should adopt appropriate career development policies, such as arranging training courses and seminars or providing incentives to learn new technologies to help employees overcome their fear of obsolescence and motivate them to contribute to the organization . A Canadian study used a generalized estimation equation model to assess individual and organizational predictors of job satisfaction among allied health care providers and found that providing adequate education improved job satisfaction, reduced staff turnover and led to the provision of better-quality residential care . 2.3.2. Promotion Opportunities Opportunities for advancement mean higher job grades, higher salaries or greater responsibilities. As employees pursue a sense of professional competence, promotion reduces the link between under improvement and the intention to improve professional competence . In particular, research has found that the chances of promotion within a company were significantly positively correlated with job satisfaction and retention, and promotion opportunities had a positive predictive effect on job satisfaction . Promotion means an employee moves to a higher position relative to someone who does not. Employees can derive satisfaction not only from higher earnings than their peers but also from higher rankings. For managers, retention measures for resident service staff can achieve appropriate results by establishing a good promotion channel and training system. After the basic professional knowledge and skills of the staff are upgraded, individuals can be promoted or developed to have more specific professional abilities and work independently . Additionally, career promotion can increase the expectations of the staff in the organization and their willingness to stay . 2.3.3. Career Development Career development refers to the continuous course of an individual's work or career throughout his or her life and positively affects whether he or she can find other jobs in the future. A lack of management skills means that training opportunities and departmental objectives are not clearly communicated to staff; a lack of career coaching, poor retention strategies and inadequate promotion opportunities all contribute to declining career prospects . The results of a cross-sectional questionnaire survey showed that career development and higher salaries ranked highest as factors to motivate retention, and incentive programs aligned with employees' work-related and personal needs were developed to improve employee satisfaction and reduce turnover, thereby improving the quality of health care . By combining the above factors that promote the retention of home care attendants, managers can first better understand the retention factors and then improve and increase the quality and continuity of home-based care services. 3. Research Methodology In this study, the MCDM method was used to construct the key factors that promote the voluntary retention of home care attendants. Additionally, it was used to identify the causal relationship between key factors and their mutual causation by combining the DEMATEL and ANP methods to improve the reliability of identifying the factors that promote the voluntary retention of home care attendants. The data collection of this study was divided into two phases. The first stage was the DEMATEL stage, which mainly explored the interaction among the important factors that promote home care workers' intention to stay. The second stage was the ANP stage, which evaluated the weight of each factor. 3.1. DEMATEL DEMATEL was used to clarify whether the components and criteria of key factors that promote the retention of home care attendants are related or feedback; additionally, this method can be used to construct a structural composition of the analytical network program. DEMATEL in this MCDA model has the following four steps:STEP 1: Define the factors and measurement scale The 3 facets and 14 criteria were selected through the literature review, and the first stage of the DEMATEL questionnaire was defined and designed. The content of the first part of the DEMATEL questionnaire contained basic information, including service unit, service title and service seniority. The second part contained a brief definition of the 14 criteria that are important factors in promoting the retention of home care workers; for example, the future environment component comprises the definitions of three programs, namely training programs, promotion opportunities and career development. The third part explored the interrelationship between key factors that promote the retention of home care attendants, so the experts were given five levels, 0~4, divided into the following responses: 0 is no impact, 1 is low impact, 2 is medium impact, 3 is high impact and 4 is very high impact. A total of 188 questions were included in the DEMATEL questionnaire. In terms of data collection, this study began by e-mailing invitations to act as experts to chief executives who had managed a long-term care institution in Taiwan for more than three years. Four managers of long-term care institutions replied to the e-mail and voluntarily participated in March 2022. To improve the reliability of the home care attendant discretionary evaluation system and the construction of the MCDM model, this study extended the invitation to participate in the data collection to one long-term care academic professor (with more than 20 years of experience), one senior home care service supervisor (with more than 3 years of service) and one senior home care worker (with more than 5 years of service). Then, the seven experts were asked to complete the DEMATEL questionnaires. STEP 2: Calculate the direct/indirect relationship matrix After the experts completed the questionnaire and judged the influence of the two criteria, the degree of influence between factors was determined, and the correlation between factors was expressed by a matrix (Equation (1)). (1) X=0x12x1nx21x2nxn1xn20 Then, a normalized direct relationship matrix was calculated. There are two ways to calculate the normalized direct relationship matrix: one is to take the column vector and the largest as the normalization base, and the other is to take the vector and the largest of the column or columns as the normalization base. In this study, the column vector and the largest were used as the normalization base. The calculation formula is shown in Equation (2). (2) l=1Max1<=i<=nj=1nxij The normalized direct correlation matrix N can be calculated by Equations (1) and (2), with the direct correlation matrix X multiplied by the lambda value (Equation (3)). After the normalized relationship matrix was obtained, the direct/indirect relationship matrix T was established using Equation (4). (3) N=lX (4) T=limk-N+N2++Nk=NI-N-1 Equations (5) and (6) were used to calculate the total intensity Di of the affected criteria and the total intensity of the affected Rj. (5) Di=j=1ntij i=1,2,,n (6) Rj=i=1ntij j=1,2,,n STEP 3: Calculate the prominence and relation (Dk + Rk) is the prominence, and k = i = j =1, 2, ..., n, which represents the degree to which this factor is affected, according to which the core degree of factor k in all problems can be revealed. (Dk - Rk) is defined as the degree of causation, which indicates the degree of difference between the influence of this factor. According to this value, the degree of causality to which factor k belongs in all problems can be displayed, and if it is positive, the factor is biased towards the cause class. If it is a negative table, the factor is biased towards the result class. STEP 4: Establish thresholds and causal diagrams To present a more significant causal relationship, this study calculated the values of the direct and indirect relationship matrices of each facet, determined whether there were values with lower correlation that needed to be deleted by setting a threshold value, and selected values greater than or equal to the threshold value. Thresholds were determined in the following ways. First, they were set on the basis of the recommendations of decision makers or jointly with multiple experts without specifying the specific process of how to determine them. Second, the statistical mean or median of the elements in the total relationship matrix was the mainstay, but the average value was easily affected by the extreme value and lacked a theoretical basis. Although the median was not affected by extreme values, it still lacked sensitivity to the representativeness of skewed data and was too subjective for the causal relationship of systemic factors. Third, based on the statistical distribution, the threshold could be calculated using m + s or m + 1.5s , where m and s are the mean and standard deviation (SD) of the elements of the full relationship matrix, respectively. Essentially, the total relational matrix data do not necessarily follow a normal distribution and therefore may not be consistent with reality . Furthermore, if the threshold is set too high, the relationship between the facets may not be distinguishable, and if it is too low, the relationship between the facets may be complicated so that the important inter-facet relationship cannot be revealed. After discussion with the experts, the threshold value of this study was set in the third quartile (Q3), which was suggested by the previous research . With Q3 as the threshold, some factors may not have entered the ANP questionnaire when assessing the relative importance of the interdependence between various components and criteria. Causality diagrams simplify the complex relationships between factors, show the influence of each factor on other factors, facilitate understanding of the topic to be studied and provide direction for subsequent discussions. When plotting the causal diagram, prominence (Dk + Rk) was taken as the horizontal axis and the relation (Dk - Rk) as the vertical axis. The coordinate points of the characteristics of the above factors formed a coordinate graphic. When (Dk - Rk) is positive, attribute k is classified as a cause class or influence group, and when (Dk - Rk) is negative, attribute k is classified as a result class or affected group. A higher (Dk - Rk) indicates that the attribute affects and is affected by other attributes to a greater extent. Depending on the coordinate position of the relation and prominence, attributes can be divided into the following four categories: (a) Core region: The relation (Dk - Rk) is positive, the prominence (Dk + Rk) value is high, and the degree of influence of other characteristics plus the degree of influence of other characteristics and the sum of influence are strongly affected. However, the degree of influence on other characteristics is greater and is biased towards the cause class. The representative attribute is the cause class, which is the driver of the problem. (b) Drive region: The relation (Dk - Rk) is positive, the prominence (Dk + Rk) is low, and the degree of influence of other characteristics is high, but the degree of total influence is low, which belongs to the quality characteristic factors of the cause of the low sum influence. This indicates that attributes are independent and can affect only a few other attributes. (c) Affected region: The relation (Dk - Rk) is negative, the prominence (Dk + Rk) value is higher, the degree of influence of other quality characteristics is higher, and the degree of total influence is low. However, the sum of the two effects is highly influential, biased towards the quality characteristics of the outcome category. Representative properties are the core problem that needs to be solved, but because they are properties of the result class, they cannot be directly improved. (d) Independent region: The relation (Dk - Rk) is negative, and the prominence (Dk + Rk) value is low, which is highly affected by other characteristics. However, the degree of sum influence is low, which belongs to the result quality characteristic factors affected by the low sum. This indicates that the attribute is independent, has a low correlation with other factors and is affected by only a few other attributes. 3.2. The Procedure of the Analytic Network Process (ANP) In the second stage, the ANP is used to understand the relative importance of the evaluation system criteria and for the network hierarchy of factors that can improve the construction to promote the retention of home care attendants. Therefore, the network analysis procedure method is used to obtain the relative weight of the evaluation system criteria and the criteria. The ANP is derived from the AHP decision-making model, and its decision-making process is closer to people's decision-making process. The AHP, on the other hand, is limited by the fact that the criteria are independent of each other and there is no interaction between them, which may oversimplify the problem; thus, its assessment may be biased. In the ANP, feedback within clusters (internal dependence) and between clusters (external dependence) is allowed, the complex impact of interaction in human society can be more accurately expressed through feedback, and a more complete decision-making structure is the result. Elements in the cluster can be linked to other cluster elements according to the user's need to investigate the process. The evaluation of decision problems using ANP includes the following three steps :STEP 1: Establish the structure of the guidelines and their relationships Through the literature discussion and summarizing and formulating the important factors and criteria for promoting the retention of home care attendants, we established the structure of the guidelines and their relationships. We started the literature review using relevant key words, searching ("home care workers", "home health aides", "nursing aides", "willingness to stay" and "intention to leave"). The literature records were identified through database searching (Web of Science and EBSCO). After removing duplicates and screening abstracts, we only kept the records with full text for eligibility. Then, we excluded the records which only discussed home care workers' personal reasons (such as personal background, socioeconomic status, commuting convenience and stress from family conflict) because these factors could not be affected directly by the home care institutions and were not appropriate for the study aims. There were in total 19 records included, and 11 records were related to home care workers' working environment, 10 were related to the organization factor and 7 were related to home care workers' future development (some records mentioned more than two factors of home care workers' intention to stay). There were three records mentioning "Salary and benefit (A1)", one record mentioning "Flexible scheduling (A2)", four records mentioning "Low working pressure and load (A3)", four records mentioning "Good relationships with clients (A4)", two records mentioning "Avoiding workplace harassment (A5)", one record mentioning "Professional image (B1)", two records mentioning "Respect (B2)", two records mentioning "Sense of accomplishment (B3)", one record mentioning "Organizational commitment (B4)", two records mentioning "Job satisfaction (B5)", three records mentioning "Supervisor's leadership style (B6)", two records mentioning "Training program (C1)", three records mentioning "Promotion opportunities (C2)" and two records mentioning "Career development (C3)". After integrating the included literature records, we built the structural chart of the factors which influence home care workers' intention to stay. In total, 3 facets and 14 criteria of factors which influence home care workers' intention to stay were included. The structural chart of the ANP criteria is shown in Figure 1. The MCDM research model was constructed through the DEMATEL and ANP method questionnaires to obtain appropriate factors to promote home care attendants' intention to stay, so this structural chart was employed in both the DEMATEL and ANP stages. STEP 2: Make pairwise comparisons This was the second stage of the questionnaire; the seven experts were asked to compare pairs between criteria with a total of 134 questions in the ANP questionnaire in May 2022. It included pairwise comparisons between dimensions and comparisons of criteria within dimension groups. There were nine levels (1-9) of standards to fill in: 1 is equal importance, 3 is slightly important, 5 is quite important, 7 is extremely important, 9 represents absolutely important, and 2, 4, 6, and 8 are intermediate values of adjacent scales. According to the seven experts' scores of the relative importance of the criteria, the eigenvectors of each pairwise comparison matrix were finally calculated in the same way as in the analytical hierarchy program. Using Equation (7) to generate group priority matrix A, the factors of dependence were compared in pairs, and the priority weights were generated. (7) A=1A12A1n1/A12A2n1/A1n1/A2n1 The ANP questionnaire was distributed to the same seven experts as the DEMATEL questionnaire, the same situation and background were discussed in depth and the research questions were analyzed more intensively. A total of seven questionnaires were completed. To comply with the consistency requirement, the CR value should be less than or equal to 0.1 when comparing the aspects and criteria of the evaluation system in compliance with conformity verification. STEP 3: Verify Conformity and Obtain Weights After the ANP questionnaire was collected, the consistency verification procedure was carried out, and the consistency verification principle had to meet the principle of superiority and inferiority and the reproducibility of the strength relationship (transitivity) to ensure the validity of the questionnaire and the judgement of the experts who completed the questionnaire. Consistency verification is based on the consistency ratio of a pairwise comparison matrix; C.R. = C.I./R.I. where C.I. is consistency index and R.I. is random index. C.R. <= 0.1 was suggested to indicate that the degree of bias of decision makers in the judgement of the weights of each element when establishing the pairwise comparison matrix was still within an acceptable range, that is, there was consistency . Then, a supermatrix structure was built, with all priority vectors placed in the supermatrix in the appropriate positions to produce an unweighted supermatrix. The criterion weights of the unweighted supermatrix were multiplied by the weights of the relevant facets so that the sum of the values of the rows was 1. The limit supermatrix was calculated from the weighted supermatrix. The weighted supermatrix was multiplied multiple times until the values of the columns were consistent and the final relative priority of each criterion could be obtained. The normalized limit supermatrix identified the relative weights of the group (criterion) and the element (alternative). The criterion with the higher weight indicated the criterion with the highest priority . 4. Study Results This study mainly summarized the preliminary research framework through a literature review and categorized the aspects and criteria for promoting home care workers' intention to stay. Seven experts and scholars were invited to complete the two questionnaires. 4.1. DEMATEL Results and Network Relationship Establishment From the data collected by the DEMATEL questionnaire, the total influence matrix T of the three dimensions was obtained and is shown in Table 1. After filtering by threshold (Q3), the prominence (D + R) and the relation (D - R) calculated from the sum of rows and columns of the total influence matrix are shown in Table 1. Ranking by prominence (D + R), "Organization (B)" is the highest (21.424), followed by "Working environment (A)" and "Future development (C)." The results show that "Organization (B)" is the most visual component dimension to enhance home care workers' intention to stay. Staff working in health organizations not only expect their basic physiological needs to be met but also gradually pay attention to the pursuit of self-image, sense of accomplishment and identity, and seek recognition of and respect for their professional skills. Meanwhile, regarding the values of the relation (D - R) from the table, "Work environment (A)" (1.347) is the "cause group" since the value is positive, while "Organization (B)" (-0.4 77) and "Future development (C)" (-0.871) are the "affected group". The information shown in Table 1 was plotted as a diagram of the influence relationship between the dimensions in Figure 2. The X-axis is the prominence (D + R), the Y-axis is the relation (D - R), the dimension of the "cause group" is at the top and the two dimensions of the "affected group" are at the bottom. The direction of the significant influences filtered through the threshold value of Q3 (3.658) is represented as the arrow. When the impact value was greater than or equal to the threshold value, the influence relationship was chosen and displayed by an arrow in the impact digraph map. In addition, the impact digraph map of the dimensions shows that "Working environment (A)" is the "cause group" and "Organization (B)" and "Future development (C)" are the "affected group", where the solid arrows point in the three one-way influence directions as follows: "Work environment (A)" affects "Organization (B)" and "Future development (C)", and "Organization (B)" affects "Future development (C)." According to the prominence (D + R) and the relation (D - R), "Working environment (A)" was in the "core region" as the key influence factor and should be listed as the priority dimension. "Organization (B)" and "Future development (C)" belong to the "Affected region," which means that their linkage could be improved when "Working environment (A)" is improved. Similar to the dimension analysis process, the values of the same columns were summed to determine the average value, and the matrix was then normalized. Each element of the matrix was divided by the maximum column sum to obtain the normalized matrix and brought into the formula, and the threshold (Q3) was obtained by the total influence matrix T of the criteria in Table 2. The prominence (D + R) and the relation (D - R) were calculated from the sum of the rows and columns of the total influence matrix, as shown in Table 2. Table 2 shows the central degree of the prominence (D + R) of the criteria. According to the DEMATEL results, the top three criteria with the highest overall impact on the retention of home care workers were "Job satisfaction (B5)", "Career development (C3)" and "Organizational commitment (B4)". Regarding the values of the relation (D - R) of the criteria, "Salary and benefits (A1)", "Flexible scheduling (A2)", "Training Program (C1)", "Professional image (B1)", "Avoiding workplace harassment (A5)" and "Supervisor Leadership (B6)" were all in the "cause group" since these criteria obtained positive values. "Building a good relationship with the client (A4)", "Promotion opportunity (C2)", "Respect (B2)", "Career development (C3)", "Organizational commitment (B4)", "Job satisfaction (B5)", "Sense of accomplishment (B3)" and "Reduce work stress and workload" (A3)" with negative values of the relation (D - R) were the "affected group". The values of the prominence (D + R) and the relation (D - R) obtained in Table 2 were used to plot the impact digraph map to visualize the causal effects among the criteria, as shown in Figure 3. The X-axis is the prominence (D + R), and the Y-axis is the relation (D - R). The significant influences were filtered from the total-related matrix T of the criteria by setting the threshold (Q3 = 0.309). All the significant influences in which the influence value was greater than or equal to the threshold value were drawn as arrows to indicate the influence relationships among criteria to affect home care workers' intention to stay. In the impact digraph map of criteria , "Salary and benefits (A1)", "Flexible scheduling (A2)", "Avoiding workplace harassment (A5)", "Professional image (B1)", "Supervisor's leadership style (B6)" and "Training program (C1)" were the "cause group", and "Low work pressure and workload (A3)", "Build a good relationship with the client (A4)", "Respect (B2)", "Sense of accomplishment (B3)", "Organizational commitment (B4)", "Job satisfaction (B5)", "Promotion opportunities (C2)" and "Career Development (C3)" were the "affected group". Among the criteria, the solid arrows indicate the one-way influence direction. The dotted arrows in Figure 3 show two-way influences. The six criteria "Salary and benefits (A1)", "Flexible scheduling (A2)", "Avoiding workplace harassment (A5)", "Professional image (B1)", "Supervisor's leadership style (B6)" and "Training program (C1)" were in the "core region" that addresses the key impact factors of home care workers' intention to stay and should be prioritized. The seven criteria of "Reduction of work pressure and workload (A3)", "Good relationships with clients (A4)", "Respect (B2)", "Sense of accomplishment (B3)", "Organizational commitment (B4)", "Job satisfaction (B5)", "Promotion opportunities (C2)" and "Career development (C3)" were in the "affected region", and their linkages would improve when those criteria in the "core region" were improved. 4.2. Analytic Network Process (ANP) Results After the DEMATEL analysis to understand the complex causal relationships between the dimensions and the criteria, this section continued to employ the ANP method to calculate the quantitative weight of each criterion. Table 3 shows the results of the criteria weights. When the ANP weights were ranked, "Job satisfaction (B5)" was the highest (31.90%), followed by "Supervisor's leadership style (B6)", "Respect (B2)", "Career development (C3)", "Low work pressure and workload (A3)", "Organizational commitment (B4)", "Sense of accomplishment (B3)", "Good relationships with clients (A4)", "Promotion opportunities (C2)" and "Professional image (B1)". The ANP weights of the remaining four criteria, "Salary and benefits (A1)", "Flexible scheduling (A2)", "Avoiding workplace harassment (A5)" and "Training program (C1)", all converged to 0.0%. In addition, the weights of the criteria by the AHP method are displayed in Table 3, and the top three AHP weights were "Salary and benefits (A1)", "Low work pressure and workload (A3)" and "Job satisfaction (B5)". 4.3. Discussion The prominence (D + R) values analyzed by DEMATEL indicated that the criterion "Job satisfaction (B5)" from the dimension "Organization (B)" had the highest overall impact; regarding the relation (D - R) values, the criterion "Salary and benefits package (A1)" in the dimension "Working environment (A)" had the highest net impact on home care workers' intention to stay in the profession. The ANP method was used to determine the priority of importance of the factors that promote the retention of home care staff by considering the interdependent relationships conducted by DEMATEL simultaneously. The ANP results indicated that the top three key criteria for enhancing the retention of home care attendants were "Job satisfaction (B5)", "Supervisor's leadership style (B6)" and "Respect (B2)" in the dimension of "Organization (B)". Overall, "Job satisfaction (B5)" had the highest weight and was the most important criterion. According to past studies, employees who believe that managers genuinely care about their welfare will have higher job satisfaction, which in turn will reduce their willingness to leave . Therefore, "Job satisfaction (B5)" was the main reason for willingness to stay and an important predictor of employees' intention to stay. When a job does not meet employees' needs, they might have the idea of leaving to find a better job. Being able to take care of and help clients so that they can have a sufficient sense of accomplishment and satisfaction and feel that they have contributed to society may strengthen their willingness to stay. The satisfaction of autonomy among home care workers has also often been mentioned in previous studies, and home workers often hope that their need for work autonomy can be satisfied. Previous research has noted that there is a clear positive correlation between job satisfaction and career retention, and job satisfaction was the main determinant of employee retention even during the COVID-19 pandemic . The more satisfied employees are with their feelings about human interaction and their need for responsibility, the higher the likelihood of retention. If the organization fails to improve these retention factors to a satisfactory level, it may lead to departures, and these results are consistent with the literature. However, when home care workers are fully engaged in care services, they can not only significantly increase their job satisfaction and reduce their willingness to leave but also promote organizational citizenship behaviors , such as taking on more responsibilities and dedicating personal time to work. In addition, the retention of resident staff was positively correlated with an improvement in the quality of life of the residents under their care . Therefore, if the job satisfaction of resident staff can be improved, it is actually indirectly helpful for residents and the organization as a whole. Earlier research has shown that managers can treat others in a way that is related to their ability to achieve organizational goals, suggesting that the influence of managers or supervisors on employees comes mainly from their ability to create positive intrinsic motivation for staff, which sometimes affects employees more than money . Therefore, the way supervisors treat staff is an important factor in promoting the voluntary behavior of resident staff. Different supervisor leadership styles produce different employee behaviors that can improve management performance, promote employee productivity and sustain team development. Therefore, adequate communication between home-based caregivers and their supervisor can reduce burnout and improve retention. Home care workers themselves are becoming increasingly aware that care is a professional job and understanding the value and significance of this service field. However, some members of the public may not have a good understanding of home care work, and there is insufficient respect and recognition in their attitude towards home care workers. This will cause great frustration to home attendants, as when they receive unfriendly treatment from society, they will feel uncomfortable. This may even lead to passive, withdrawn behavior or self-questioning among home care staff. Furthermore, although "Salary and benefits (A1)" is the highest net factor in promoting employee retention in the DEMATEL method, it is not the reason with the greatest weight in the ANP. For example, although managers can consider raising salaries or increasing annual bonuses, other institutions can increase their financial temptation in the same way. Therefore, it will not be mainly money considerations that will reduce their intention to stay. Since most home care workers are female, staff may differ in the weightings they give specific inputs or rewards, and pay was found to be less of an issue for female direct-care leavers . "Salary and benefits (A1)" in the "core region" might not be a direct factor but influence home care workers' intention to stay through the impact on other factors. "Job satisfaction (B5)", "Supervisor treatment (B6)" and "Respect (B2)" can lead to satisfactory support and communication; therefore, these are the key factors in becoming a home care worker with a high level of caring. This also means that while the government is striving to increase the salary and benefits of home care attendants in order to retain human resources for home care, it may be more important to consider how to improve the respect for home care workers in society, meet the needs of the working environment and work itself, strengthen and improve the support system for supervision by agency heads, etc., to effectively retain the home care staff who already provide services in the practical field and attract young and new workers who are willing to devote themselves to the work of home care. If a manager or supervisor has strong communication with staff and can properly allocate time and labour, this has a positive impact on retention. This finding is also consistent with the literature, and leadership style is one of the key factors in promoting the retention of staff members. To meet the long-term care staffing needs in Taiwan, the quality of home care services should be improved; the mutual circulation of resident staff, health care staff and the job market should be promoted; employment opportunities should be increased; and training courses should be integrated. The government organizes training courses every year, and at present, home care workers are required to attend the care worker training course to obtain a certificate of completion and to obtain a vocational technician certificate. Staff should regularly participate in training courses every year to maintain their expertise, which can indirectly enable them to feel confidence in their work and to show caring, patience and a professional image. In summary, from the perspective of ANP weighting, "Job satisfaction (B5)", "Supervisor's leadership style (B6)" and "Respect (B2)" are the key factors in promoting the retention of home care workers. "Salary and benefits (A1)", "Flexible scheduling (A2)", "Avoiding workplace harassment (A5)", Professional image (B1)" and "Training program (C1)" were not weighted in the ANP results, but in the DEMATEL analysis, those factors were in the "cause group". This means that these factors might not directly affect home care workers' intention to stay, but they might indirectly influence home care workers' intention to stay through linkage effects. The current situation of home care workers in Taiwan is more flexible than that of hospital attendants in the past, and home care workers can arrange their own time and cases. Additionally, the organization allows more space for home care workers to choose their own cases to avoid sexual harassment in the workplace. If we use the traditional AHP to understand the key factors that promote intention to stay, the findings reveal that the three most important criteria are "Salary and benefits (A1)", "Low work pressure and workload (A3)" and "Job satisfaction (B5)". However, through the MCDA model combined with DEMATEL and ANP, the three most criteria were "Job satisfaction (B5)", "Supervisor's leadership style (B6)" and "Respect (B2)". This result indicates that if only the traditional AHP is considered, the results would be different. The reason is that the AHP assumed independency among the criteria and ignored the impact relationships among them. Although the hierarchy of the traditional AHP is linear, it does not consider that there may be feedback and dependence among the criteria, which can easily cause decision makers to make mistakes. In the MCDA model proposed in this study, the criteria are used in feedback and dependence simultaneously, so the results are more valuable and reduce the chance of misjudgment. 5. Conclusions and Recommendations for Future Research With the development of the home care service industry, under the influence of the ageing society and the pandemic, society faces huge home care problems. Home care workers are the front line of home care staffing. They contact service recipients most frequently and are the most trusted by service recipients. In addition, clients may also face the dilemma of readapting and building a relationship of trust due to changes in caregivers if a hired home care attendant is a novice and is not familiar with the home care services. This might lead to a decline in the quality of care if it is impossible to recruit suitable home care workers quickly enough. This will cause work pressure and an excessive load on existing home care helpers. Although most caregivers faced new challenges in internal motivation to continue working during the COVID-19 pandemic, organizational rewards, financial incentives and the hope of changing employment status were all motivating factors for them to continue working . This paper aimed to further explore the promotion of the retention of home care workers in the context of understanding the characteristics of home care work. Through various research and analysis methods and comprehensive results, this study identified the key factors that promote the retention of home care staff. The results of the study show that organizations should focus on improving job satisfaction, supervisor leadership style and respect because job satisfaction is a true reflection of employees who can express themselves directly. For staff members, overall job satisfaction and the degree of feeling that affects their mood are related to promoting retention. Therefore, as an important retention factor, managers and human resources practitioners need to develop relevant programs to improve the job satisfaction of caregivers to cultivate their sense of responsibility . They should pay more attention to the work of home attendants and have empathy. When home attendants encounter work setbacks, they can help calm their emotions and provide encouragement and support, provide positive affirmation with an attitude of appreciation and encouragement, publicly praise excellent home care attendants and publish information on websites or in institutional publications that can effectively enhance home care attendants' work motivation and sense of belonging to the organization. In addition, if resident staff can share their care experience or care skills at regular meetings to enhance their sense of honor and achievement, when they engage in work, the clients they care for will also receive good service, thus achieving the win-win care goal of supply and demand. Home care supervisors or managers should improve their leadership communication and training skills so that both the clients and the staff can feel fairly treated. Additionally, they should teach staff to take care of the practice to help make their work smoother and regularly take the initiative to investigate the work needs of staff to establish efficient communication channels and feedback channels, support their work based on the service staff's response and create a friendly workplace (including supervision, a promotion system and festive event parties). This will encourage home care attendants to stay in their jobs for a long time. When dispatching client services, home care institutions should consider the service route of the home attendants, assist them in applying for accident insurance, emphasize the service content to the clients and advocate the professionalism of the home attendants so that the public can improve their understanding of the roles and responsibilities of care attendants. This will shape the professional image of the care staff and enhance respect for them. This is a suggestion to increase their intention to stay, as it is necessary to improve this aspect to effectively achieve the retention of staff members. Although "Salary and benefits (A1)" is the factor with the highest net impact on promoting the retention of staff in the DEMATEL, it is not the most weighted factor in the ANP results. Therefore, although "Salary and benefits (A1)" is the basic element of employee satisfaction, it is not the only factor. Additionally, the amount and quality of salary and benefits are no longer the most direct and main factors generally. Employers or managers think that employees mainly desire money and leave, but this finding reverses the stereotype that salary or pay is the only thing that workers care about when considering retention. This indicates that managers in the home care industry should pay attention to the differences from factors that have promoted the retention of employees in the past if they want to provide improved salary and benefits. The opinions of the resident staff about salary and benefits should be investigated regularly, the salary and benefits of the resident staff of different home care institutions should be compared so that the appropriate salary and benefits treatment can be given according to the situation of the institution. This result also coincides with the concept of equity theory. When providing reasonable salary and benefits, attention should be paid to the leadership and management of the supervisor and the due respect of the home care staff. The hybrid MCDA model of DEMATEL and the ANP, compared with the AHP, is currently the most suitable method not only for determining the mutual influence relationship but also for effectively solving the problem of incompatibility and feedback between AHP aspects and criteria. If the combination of the DEMATEL and ANP methods is not considered, researchers may be unable to determine the truly important factors. The purpose of this study is to understand the key factors of the success of home care attendants that are truly important to help managers of long-term care institutions retain home care talent. If home care institution managers can understand the important factors that promote the retention of home care staff, on the one hand, agency managers will be able to understand that home care staff want to stay in the institution, and on the other hand, they will be able to face problems or difficulties in the future and introduce practical and policy-related suggestions for future improvement. In this way, they can achieve the goal of effectively promoting the retention of staff. This study mainly uses the MCDM research method to explore the factors that promote the retention of home care attendants. Although the causal influence and key factors can be understood, in the future, big data analysis might be used to consider home care workers as the research object and to verify the relationships among the factors that promote the retention of home care attendants. This would resolve the limitations of the MCDA model, such as the few expert observations and the inability to verify the mediation or moderation effect analysis. In addition, it is suggested that in the future, the scope of research could be extended to other long-term care industries (e.g., community or institutional) to compare whether the key factors of care workers' intention to stay are similar. Finally, this study ignores the ambiguity of decision-makers' judgements, so future research can apply the concept of fuzzy theory for in-depth discussion so that researchers can obtain objective information, opinions and insights through the independent judgement of multiple experts in the process of information collection. In this way, agency managers could gain a deeper understanding of the different views of staff retention in their organizations, and the depth and breadth of research could be expanded to achieve the goal of promoting the retention of employees. In the past, there has been no study using the DEMATEL and ANP methods to discuss the factors that promote the retention of home attendants. This research method solves the problem of stratification in a structured way to quantitatively determine the problem of decomposition in different aspects and hierarchies. It provides an overall final evaluation score and judgement of the criteria to construct an MCDM structure of the factors that promote the retention of home care workers. In practice, this study provides managers of home care institutions with a better understanding of ways to promote the retention of home care workers and allows them to consider coping methods based on the factors generated by the results of this study. Managers can implement activities and plans to increase the retention of home care staff in their institutions. Therefore, this study provides practical ideas for promoting the retention of home care workers and benefiting people with home care needs, the entire home care industry and government agencies. The future ageing problem is a major challenge that the country must face, and how to cultivate and retain exceptional home care talents is an important issue. Author Contributions Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Validation, Visualization, Writing--original, draft, W.H.; Data curation, Resources, Software, Writing--review & editing, F.-P.S. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 ANP-based model for enhancing home care workers' intention of to stay. Figure 2 The impact digraph map of dimensions. Figure 3 The impact digraph map of criteria. healthcare-11-00750-t001_Table 1 Table 1 The total-related matrix T of dimensions (threshold value >=3.658). Dimension A B C D R D + R D - R Ranking Result D + R D - R A 3.272 4.003 3.873 11.149 9.801 20.950 1.347 2 1 B 3.382 3.433 3.658 10.474 10.951 21.424 -0.477 1 2 C 3.147 3.514 3.118 9.779 10.650 20.429 -0.871 3 3 Note: A = Working environment; B = Organization; C = Future development. D + R is the prominence and D - R is the relation. healthcare-11-00750-t002_Table 2 Table 2 The total-related matrix T of the criteria (threshold value >=0.309). Criterion A1 A2 A3 A4 A5 B1 B2 B3 B4 B5 B6 C1 C2 C3 D R D + R D - R Ranking Result D + R D - R A1 0.202 0.205 0.310 0.287 0.169 0.300 0.317 0.333 0.355 0.376 0.303 0.274 0.282 0.367 4.080 3.149 7.229 0.930 9 1 A2 0.193 0.111 0.252 0.223 0.134 0.189 0.226 0.208 0.265 0.295 0.243 0.188 0.195 0.263 2.986 2.327 5.314 0.659 13 2 A3 0.217 0.174 0.201 0.253 0.155 0.226 0.249 0.248 0.298 0.329 0.270 0.220 0.235 0.308 3.383 3.775 7.158 -0.393 10 14 A4 0.220 0.176 0.286 0.214 0.180 0.276 0.308 0.296 0.300 0.337 0.282 0.221 0.254 0.325 3.675 3.719 7.394 -0.043 8 7 A5 0.140 0.108 0.188 0.197 0.088 0.181 0.207 0.179 0.212 0.236 0.191 0.157 0.159 0.219 2.462 2.271 4.734 0.191 14 5 B1 0.253 0.165 0.263 0.291 0.190 0.223 0.328 0.323 0.316 0.343 0.296 0.262 0.272 0.338 3.861 3.660 7.522 0.201 7 4 B2 0.222 0.162 0.269 0.294 0.197 0.298 0.248 0.321 0.331 0.351 0.312 0.244 0.258 0.340 3.845 4.088 7.933 -0.243 4 9 B3 0.231 0.156 0.266 0.273 0.152 0.268 0.303 0.235 0.314 0.336 0.285 0.246 0.269 0.335 3.669 4.044 7.713 -0.375 6 13 B4 0.248 0.193 0.295 0.285 0.169 0.289 0.312 0.317 0.265 0.349 0.311 0.270 0.290 0.350 3.945 4.305 8.250 -0.360 3 11 B5 0.265 0.194 0.326 0.325 0.182 0.287 0.341 0.352 0.365 0.304 0.336 0.278 0.300 0.381 4.235 4.599 8.834 -0.364 1 12 B6 0.229 0.192 0.311 0.285 0.186 0.284 0.333 0.308 0.340 0.351 0.240 0.251 0.268 0.345 3.924 3.911 7.835 0.013 5 6 C1 0.239 0.152 0.259 0.266 0.169 0.272 0.285 0.287 0.292 0.312 0.267 0.187 0.259 0.317 3.562 3.340 6.902 0.222 12 3 C2 0.222 0.144 0.250 0.238 0.141 0.257 0.289 0.292 0.300 0.308 0.262 0.250 0.195 0.318 3.466 3.550 7.016 -0.085 11 8 C3 0.269 0.193 0.300 0.287 0.159 0.312 0.342 0.345 0.352 0.373 0.314 0.292 0.315 0.292 4.144 4.498 8.642 -0.353 2 10 Note: Working environment (A): A1 = Salary and benefits; A2 = Flexible shift scheduling; A3 = Low working pressure and load; A4 = Good relationships with clients; A5 = Avoiding workplace harassment. Organization (B): B1 = Professional Image; B2 = Respect; B3 = Sense of accomplishment; B4 = Organizational Commitment; B5 = Job Satisfaction; B6 = Supervisor's leadership Style. Future Development (C): C1 = Training Program; C2 = Promotion Opportunities; C3 = Career Development. D + R is the prominence and D - R is the relation. healthcare-11-00750-t003_Table 3 Table 3 Comparison of critical success factors weights in AHP and ANP. Dimension Criterion AHP Weights (%) ANP Weights (%) AHP (Ranking) ANP (Ranking) Working environment (A) A1 Salary and benefits 33.58 0.00 1 11 A2 Flexible shift scheduling 7.80 0.00 4 12 A3 Low working pressure and load 10.72 9.99 2 5 A4 Good relationships with clients 5.76 3.40 6 8 A5 Avoiding workplace harassment 6.27 0.00 5 14 Organization (B) B1 Professional image 1.53 0.76 14 10 B2 Respect 3.10 11.61 9 3 B3 Sense of accomplishment 2.95 6.77 11 7 B4 Organizational commitment 2.30 8.08 13 6 B5 Job satisfaction 9.68 31.90 3 1 B6 Supervisor's leadership style 5.58 15.40 7 2 Future development (C) C1 Training program 5.14 0.00 8 13 C2 Promotion opportunities 3.10 1.85 10 9 C3 Career development 2.49 10.24 12 4 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Gazzaroli D. D'Angelo C. Corvino C. Home-Care Workers' Representations of Their Role and Competences: A Diaphanous Profession Front. Psychol. 2020 11 581399 10.3389/fpsyg.2020.581399 33362645 2. Scrinzi F. Beyond 'women's work': Gender, ethnicity, and the management of paid care work in nonprofit domiciliary services in Italy J. Immigr. Refugee Stud. 2019 17 441 456 10.1080/15562948.2018.1538472 3. Barney S.M. The nursing shortage: Why is it happening? J. Healthc. Manag. 2002 47 153 155 10.1097/00115514-200205000-00004 12055897 4. Ou Yang Y.P. Shieh H.M. Leu J.D. Tzeng G.H. A novel hybrid MCDM model combined with DEMATEL and ANP with applications Int. J. Oper. Res. 2008 5 160 168 10.1016/j.matdes.2014.03.071 5. Steinmetz S. de Vries D.H. Tijdens K.G. Should I stay or should I go? The impact of working time and wages on retention in the health workforce Hum. Resour. Health 2014 12 23 10.1186/1478-4491-12-23 24758705 6. Carpenter M.L. Blaskewicz Boron J. Beadle J. Potter J.F. Understanding Influential Factors in Turnover within the Home Care Workforce Home Health Care Manag. Pract. 2021 33 147 153 10.1177/1084822320981012 7. Mak B.L. Sockel H. A confirmatory factor analysis of IS employee motivation and retention Inf. Manag. 2001 38 265 276 10.1016/S0378-7206(00)00055-0 8. Testa M.R. Sipe L. Service-Leadership Competencies for Hospitality and Tourism Management Int. J. Hosp. Manag. 2012 31 648 658 10.1016/j.ijhm.2011.08.009 9. Jeong S.A. Kim J. Factors influencing nurses' intention to care for patients with COVID-19: Focusing on positive psychological capital and nursing professionalism PLoS ONE 2022 17 e0262786 10.1371/journal.pone.0262786 35045117 10. Estryn-Behar M. Van Der Heijden B.I. Fry C. Hasselhorn H.M. Longitudinal analysis of personal and work-related factors associated with turnover among nurses Nurs. Res. 2010 59 166 177 10.1097/NNR.0b013e3181dbb29f 20421841 11. Carraher S.M. Turnover prediction using attitudes towards benefits, pay, and pay satisfaction among employees and entrepreneurs in Estonia, Latvia, and Lithuania Balt. J. Manag. 2011 6 25 52 10.1108/17465261111100905 12. Disley P. Hatton C. Dagnan D. Applying equity theory to staff working with individuals with intellectual disabilities J. Intellect. Dev. Disabil. 2009 34 55 66 10.1080/13668250802684701 19234979 13. Teclaw R. Osatuke K. Nurse perceptions of workplace environment: Differences across shifts J. Nurs. Manag. 2015 23 1137 1146 10.1111/jonm.12270 25491124 14. Li X. Zhang Y. Yan D. Wen F. Zhang Y. Nurses' intention to stay: The impact of perceived organizational support, job control and Job satisfaction J. Adv. Nurs. 2020 76 1141 1150 10.1111/jan.14305 31957044 15. Mirzaei A. Rezakhani Moghaddam H. Habibi Soola A. Identifying the predictors of turnover intention based on psychosocial factors of nurses during the COVID-19 outbreak Nurs. Open 2021 8 3469 3476 10.1002/nop2.896 33960721 16. Bratt C. Gautun H. Should I stay or should I go? Nurses' intensions to leave nursing homes and home nursing J. Nurs. Manag. 2018 26 1074 1082 10.1111/jonm.12639 29707851 17. Sandberg L. Borell L. Edvardsson D. Rosenberg L. Bostrom A.M. Job strain: A cross-sectional survey of dementia care specialists and other staff in Swedish home care services J. Multidiscip. Healthc. 2018 11 255 266 10.2147/JMDH.S155467 29861636 18. Faul A.C. Schapmire T.J. D'Ambrosio J. Feaster D. Oak C.S. Farley A. Promoting Sustainability in Frontline Home Care Aides: Understanding Factors Affecting Job Retention in the Home Care Workforce Home Health Care Manag. Pract. 2010 22 408 416 10.1177/1084822309348896 19. Tourangeau A. Patterson E. Rowe A. Saari M. Thomson H. MacDonald G. Cranley L. Squires M. Factors influencing home care nurse intention to remain employed J. Nurs. Manag. 2014 22 1015 1026 10.1111/jonm.12104 23905629 20. Hanson G.C. Perrin N.A. Moss H. Laharnar N. Glass N. Workplace Violence against Homecare Workers and Its Relationship with Workers Health Outcomes: ACross-Sectional Study BioMed Cent. Public Health 2015 15 11 10.1186/s12889-014-1340-7 25595487 21. Daft R.L. Organization Theory and Design Cengage Learning Boston, MA. USA 2015 22. Ngwenya V.C. Job performance: Working conditions of urban teachers in Zimbabwean schools SA J. Hum. Resour. Manag. 2021 19 13 10.4102/sajhrm.v19i0.1454 23. Alilu L. Zamanzadeh V. Valizadeh L. Habibzadeh H. Gillespie M. A Grounded theory study of the intention of nurses to leave the profession1 Rev. Lat.-Am. Enferm. 2017 25 1 9 10.1590/1518-8345.1638.2894 24. Nouri A. Sanagoo A. Jouybari L. Taleghani F. Contextual barriers of Respectful workplace in nursing: A focused ethnography Iran. J. Nurs. Midwifery Res. 2021 26 349 354 10.4103/ijnmr.IJNMR_274_20 34422616 25. Fodor E.M. Schultheiss O.C. Brunstein J.C. Power Motivation Implicit Motives Oxford, England 2010 3 29 26. Na S.Y. Park H. The effect of nurse's emotional labor on turnover intention: Mediation effect of burnout and moderated mediation effect of authentic leadership J. Korean Acad. Nurs. 2019 49 286 297 10.4040/jkan.2019.49.3.286 31266925 27. Zhang L.F. You L.M. Liu K. Zheng J. Fang J.B. Lu M.M. Bu X.Q. The association of Chinese hospital work environment with nurse burnout, Job satisfaction, and intention to leave Nurs. Outlook 2014 62 128 137 10.1016/j.outlook.2013.10.010 24345617 28. Thompson J.E. An alternative control model for event studies J. Bus. Financ. Account. 1989 16 507 513 10.1111/j.1468-5957.1989.tb00033.x 29. Luz C.M.D.R. de Paula S.L. de Oliveira L.M.B. Organizational commitment, Job satisfaction and their possible influences on intent to turnover Revista Gestao 2018 25 84 101 10.1108/REGE-12-2017-008 30. Halcomb E. Bird S. Job satisfaction and Career Intention of Australian General Practice Nurses: A Cross: Sectional Survey J. Nurs. Scholarsh. 2020 52 270 280 10.1111/jnu.12548 32187812 31. Gebregziabher D. Berhanie E. Berihu H. Belstie A. Teklay G. The relationship between Job satisfaction and turnover intention among nurses in Axum comprehensive and specialized hospital Tigray, Ethiopia BioMed Cent. Nurs. 2020 19 79 10.1186/s12912-020-00468-0 32. Laschinger H.K.S. Fida R. A time-lagged analysis of the effect of authentic leadership on workplace bullying, burnout, and occupational turnover intentions Eur. J. Work. Organ. Psychol. 2014 23 739 753 10.1080/1359432X.2013.804646 33. McCaughey D. Kim J. McGhan G. Jablonski R. Brannon D. Who Needs Caring? We Do! Workplace Injury and Its Effect on Home Health Aides Acad. Manag. Annu. Meet. Proc. 2010 8 1 6 10.5465/ambpp.2010.54499996 34. Memon M.A. Sallaeh R. Rosli Baharom M.N. Nordin S.M. Ting H. The relationship between training satisfaction, organisational citizenship behaviour, and turnover intention: A PLS-SEM approach J. Organ. Eff. People Perform. 2017 4 267 290 10.1108/JOEPP-03-2017-0025 35. Malek K. Kline S.F. DiPietro R. The impact of manager training on employee turnover intentions J. Hosp. Tour. Insights 2018 1 203 219 10.1108/JHTI-02-2018-0010 36. Ogbonnaya C. Tillman C.J. Gonzalez K. Perceived organizational support in health care: The importance of teamwork and training for employee well-being and patient satisfaction Group Organ. Manag. 2018 43 475 503 10.1177/1059601118767244 37. Aloisio L.D. Gifford W.A. McGilton K.S. Lalonde M. Estabrooks C.A. Squires J.E. Individual and organizational predictors of allied healthcare providers' Job satisfaction in residential long-term care BioMed Cent. Health Serv. Res. 2018 18 491 10.1186/s12913-018-3307-3 38. Rajamohan S. Porock D. Chang Y.P. Understanding the relationship between staff and Job satisfaction, stress, turnover, and staff outcomes in the person-centered care nursing home arena J. Nurs. Scholarsh. 2019 51 560 568 10.1111/jnu.12488 31245922 39. Castle N.G. Degenholtz H. Rosen J. Determinants of staff Job satisfaction of caregivers in two nursing homes in Pennsylvania BMC Health Serv. Res. 2006 6 60 10.1186/1472-6963-6-60 16723022 40. Pierson M.A. Liggett C. Moore K.S. Twenty years of experience with a clinical ladder: A tool for professional growth, evidence-based practice, recruitment, and retention J. Contin. Educ. Nurs. 2010 41 33 40 10.3928/00220124-20091222-06 20102141 41. Kosteas V.D. Job satisfaction and promotions Ind. Relat. A J. Econ. Soc. 2011 50 174 194 10.1111/j.1468-232X.2010.00630.x 42. Erasmus B.J. Perceptions of administrative staff on career advancement realities at a South African university Manag.-J. Contemp. Manag. Issues 2020 25 69 88 10.30924/mjcmi.25.1.4 43. Al-Qathmi A. Zedan H. The effect of incentive management system on turnover rate, Job satisfaction and motivation of medical laboratory technologists Health Serv. Res. Manag. Epidemiol. 2021 8 2333392820988404 10.1177/2333392820988404 33614829 44. Hsu C.C. Evaluation criteria for blog design and analysis of causal relationships using factor analysis and DEMATEL Expert Syst. Appl. 2012 39 187 193 10.1016/j.eswa.2011.07.006 45. Wang T.C. The interactive trade decision-making research: An application case of novel hybrid MCDM model Econ. Model. 2012 29 926 935 10.1016/j.econmod.2012.02.001 46. Cebi S. Determining importance degrees of website design parameters based on interactions and types of websites Decis. Support Syst. 2013 54 1030 1043 10.1016/j.dss.2012.10.036 47. Horng J.S. Liu C.H. Chou S.F. Tsai C.Y. Creativity as a critical criterion for future restaurant space design: Developing a novel model with DEMATEL application Int. J. Hosp. Manag. 2013 33 96 105 10.1016/j.ijhm.2012.06.007 48. Bai C. Sarkis J. A grey-based DEMATEL model for evaluating business process management critical success factors Int. J. Prod. Econ. 2013 146 281 292 10.1016/j.ijpe.2013.07.011 49. Rajesh R. Ravi V. Modeling enablers of supply chain risk mitigation in electronic supply chains: A Grey-DEMATEL approach Comput. Ind. Eng. 2015 87 126 139 10.1016/j.cie.2015.04.028 50. Tzeng G.H. Huang C.Y. Combined DEMATEL technique with hybrid MCDM methods for creating the aspired intelligent global manufacturing & logistics systems Ann. Oper. Res. 2012 197 159 190 10.1007/s10479-010-0829-4 51. Chen J.K. Multi-layer hierarchical DEMATEL method: Analysis of soft factors in TQM practice J. Qual. 2022 29 85 10.6220/joq.202202_29.0005 52. Bolanos R. Fontela E. Nenclares A. Pastor P. Using interpretive structural modelling in strategic decision: Making groups Manag. Decis. 2005 43 877 895 10.1108/00251740510603619 53. Meade L.M. Sarkis J.J.I.J. Analyzing organizational project alternatives for agile manufacturing processes: An analytical network approach Int. J. Prod. Res. 1999 37 241 261 10.1080/002075499191751 54. Saaty T.L. Decision Making with Dependence and Feedback: The Analytic Network Process RWS Publications Pittsburgh, PA, USA 1996 Volume 4922 55. Yang C.H. Hsu W. Wu Y.L. A hybrid multiple-criteria decision portfolio with the resource constraints model of a smart healthcare management system for public medical centers Socio-Econ. Plan. Sci. 2022 80 101073 10.1016/j.seps.2021.101073 56. Babakus E. Yavas U. Karatepe O.M. Avci T. The effect of management commitment to service quality on employees' affective and performance outcomes J. Acad. Mark. Sci. 2003 31 272 286 10.1177/0092070303031003005 57. McCreary D.D.J. Home health nursing Job satisfaction and retention: Meeting the growing need for home health nurses Nurs. Clin. 2020 55 121 132 10.1016/j.cnur.2019.11.002 58. Nemteanu M.S. Dinu V. Dabija D.C. Job insecurity, job instability and job satisfaction in the context of COVID-19 pandemic J. Compet. 2021 13 65 82 10.7441/joc.2021.02.04 59. Nemteanu M.S. Dinu V. Pop R.A. Dabija D.C. Predicting job satisfaction and work engagement behavior in the COVID-19 pandemic: A conservation of resources theory approach E&M Econ. Manag. 2022 25 23 40 10.15240/tul/001/2022-2-002 60. Hsu W. Yang F.C. Factors associated with home health aides' turnover intention and organizational citizenship behavior in long-term care services Healthcare 2022 10 1743 10.3390/healthcare10091743 36141356 61. Zebral L.P. The influence of leadership and payment for performance on individual performance J. Appl. Leadersh. Manag. 2017 5 76 89 62. Cohen S. Nica E. COVID-19 pandemic-related emotional anxiety, perceived risk of infection, and acute depression among primary care providers Psychosociological Issues Hum. Resour. Manag. 2021 9 7 20 10.22381/pihrm9220211 |
PMC10000578 | Epigenetic mechanisms have emerged as an important contributor to tumor development through the modulation of gene expression. Our objective was to identify the methylation profile of the imprinted C19MC and MIR371-3 clusters in patients with non-small cell lung cancer (NSCLC) and to find their potential target genes, as well as to study their prognostic role. DNA methylation status was analyzed in a NSCLC patient cohort (n = 47) and compared with a control cohort including COPD patients and non-COPD subjects (n = 23) using the Illumina Infinium Human Methylation 450 BeadChip. Hypomethylation of miRNAs located on chromosome 19q13.42 was found to be specific for tumor tissue. We then identified the target mRNA-miRNA regulatory network for the components of the C19MC and MIR371-3 clusters using the miRTargetLink 2.0 Human tool. The correlations of miRNA-target mRNA expression from primary lung tumors were analyzed using the CancerMIRNome tool. From those negative correlations identified, we found that a lower expression of 5 of the target genes (FOXF2, KLF13, MICA, TCEAL1 and TGFBR2) was significantly associated with poor overall survival. Taken together, this study demonstrates that the imprinted C19MC and MIR371-3 miRNA clusters undergo polycistronic epigenetic regulation leading to deregulation of important and common target genes with potential prognostic value in lung cancer. DNA methylation C19MC and MIR371-3 clusters miRNA-target mRNA expression prognosis non-small cell lung cancer (NSCLC) The Ministry of Health and Social Welfare of the Junta de AndaluciaRC-0004-2020 PECART-0091-2020 Andalusian Research, Development and Innovation PlanPY20_00992 Instituto de Salud Carlos IIIPI20/01109 The Ministry of Health and Social Welfare of the Junta de AndaluciaRH-0051-2020 Instituto de Salud Carlos IIIF21/00226 Instituto de Salud Carlos IIICM20/00178 The Andalusian Research, Development and Innovation PlanPY20_00951 European Regional Development Fund (ERDF)Asociacion Espanola Contra el CancerTRNSC18004PAZ Instituto de Salud Carlos IIIPI20/00870 AC20/0070 Centro de Investigacion Biomedica en RedCD16/12/00442 FEDER from Regional Development European Funds (European Union)S.M.-P. was funded by the Ministry of Health and Social Welfare of the Junta de Andalucia (Nicolas Monardes Program RC-0004-2020, and PECART-0091-2020), Andalusian Research, Development and Innovation Plan (PY20_00992), and ISCIII (PI20/01109) and co-funded by FEDER from Regional Development European Funds (European Union). L.B. was funded by the Ministry of Health and Social Welfare of Junta de Andalucia (RH-0051-2020). J.F.N.-U. was funded by a PFIS predoctoral fellowship from ISCIII (F21/00226). J.C.B. was funded by ISCIII (CM20/00178) and co-funded by FEDER from Regional Development European Funds (European Union). RBC was funded by the Andalusian Research, Development and Innovation Plan (PY20_00951) and co-financed by the European Regional Development Fund (ERDF). L.P.-A. was funded by AECC (TRNSC18004PAZ), ISCIII (PI20/00870; AC20/0070) and CIBERONC (CD16/12/00442), and co-funded by FEDER from Regional Development European Funds (European Union). pmc1. Introduction Lung cancer remains the most common cause of mortality worldwide, and tobacco exposure increases its risk of developing . Furthermore, the incidence of lung cancer is significantly higher in patients diagnosed with chronic obstructive lung disease (COPD), reflecting the impact of smoking habits in both pathologies . Approximately 85% of all lung cancers are non-small cell lung cancer (NSCLC). Histologically, NSCLC is classified as adenocarcinoma (ADC), squamous cell carcinoma and large cell carcinoma (SCC) . Molecular analyses have led to advances in our understanding of NSCLC genetics and even in the identification of biomarkers that can predict its occurrence . This includes the role of microRNAs (miRNAs) in the disease, involved in the complexity of gene expression regulation. Therefore, a single miRNA can exert its regulatory function on several target mRNAs, and a particular target can be regulated by multiple miRNAs . There are numerous miRNAs involved in cancer-relevant processes, and many of them are clustered on the genome and act in coordinated regulatory networks. Furthermore, some evidence has even been provided suggesting that several miRNAs are able to identify patients with an increased risk of developing lung cancer, as well as COPD . For a more extensive review of oncomiRs in lung cancer, see . In this context, epigenetics appears to play an important role in the regulation of miRNA expression levels . Aberrations in methylation profiles can promote silencing of tumor suppressor microRNAs or overexpression of oncogenic miRNAs (oncomiRs) . For example, a significant upregulation of the miR-17-92 cluster has been reported in lung cancer . In addition, some oncomiRs can be located in imprinted genomic regions. Such is the case of the imprinted delta-like homolog 1 gene and the type III iodothyronine deiodinase gene (DLK1-DIO3) cluster, which includes two large miRNA clusters between other coding and non-coding transcripts, and has been reported to contribute to tumorigenesis in the lung, leukemia, breast, and hepatoblastoma, among others . In addition, alterations in other clusters located in imprinted regions are also attracting interest due to their downstream targets and their involvement in oncogenic and drug resistance mechanisms, such as the chromosome 19 microRNA (C19MC) and MIR371-3 clusters . The C19MC and MIR371-3 clusters are located on chromosome 19q13.42. The first cluster includes forty-six miRNA genes and the second one contains four miRNAs (miR-371, miR-372, miR-373 and miR-373*). Both clusters are only expressed from their paternal allele, so they are functionally haploid and, furthermore, they are expressed mainly in embryonic tissue, particularly in the placenta . However, aberrations that involve some miRNAs from both these clusters have been linked to tumoral processes, such as immunomodulation, angiogenesis, invasion, and cell reprograming . In fact, some miRNAs through exosomes have been proposed as specific cell-to-cell communication mediators . Regarding lung cancer, the regulation of imprinted C19MC and MIR371-3 has not been extensively and systematically reviewed in patients with NSCLC. For this reason, we have analyzed the methylation profile of the C19MC and MIR371-3 clusters in lung tumors compared to non-tumoral lung tissue in NSCLC patients. We have also assessed the methylation patterns of both clusters in COPD patients to study their association in a population at high risk of developing lung cancer. In addition, we were able to experimentally identify and validate deregulated targets, as well as their prognostic role in the disease. 2. Materials and Methods 2.1. Patients and Clinical Specimens The present study was carried out on 70 subjects from the Virgen del Rocio University Hospital (Seville, Spain). Samples were divided into 2 cohorts according to the underlying pathology. The first cohort consisted of 47 NSCLC patients who had undergone surgical resection at an early clinical stage. During surgical resection, adjacent normal and tumor tissue samples were collected from all patients and immediately frozen at -80 degC until further use. The clinical characteristics of patients with NSCLC (n = 47) are summarized in Supplementary Table S1. The second cohort was used as control without lung cancer (n = 23). This control cohort consisted of COPD patients and non-COPD subjects (Supplementary Table S1) who had undergone bullectomy or bronchoscopic biopsy with a negative diagnosis of lung cancer. Both cohorts were used for the analysis of the methylation profile. The protocol of the study and the use of human samples were approved by the Ethics Committee of our hospital (1381-N-21). Written informed consent was obtained from all patients included in the study. 2.2. DNA Sample Genomic DNA was extracted from 15 mg adjacent normal and tumor tissue samples using the QIAamp DNA mini kit (QIAGEN, Hilden, Germany). DNA was quantified using the QuantiFluor dsDNA system (Promega, Madison, WI, USA) according to the manufacturer's instructions. 2.3. Illumina 450 K Methylation Assay DNA methylation status at the CpG sites within the C19MC and MIR371-3 clusters was identified using the Illumina Infinium Human Methylation 450 BeadChip (Illumina Inc., San Diego, CA, USA). 500 ng of DNA were treated with sodium bisulfate using the EZ DNA MethylationTM Kit and cleaned with the ZR-96 DNA Clean-up KitTM (Zymo Research, Irvine, CA, USA). Subsequently, the following steps were performed: amplification, hybridization and imaging. Intensity data was analyzed with Illumina's GenomeStudio, from which, b-scores (i.e., the proportion of total fluorescence signal from the methylation-specific probe or color channel) were obtained. Infinium HD-based assays included sample-dependent and sample-independent controls for the highest quality data. 2.4. Methylome Data Processing The methylome data was processed using the R/Bioconductor package RnBeads . After a quality check, intensity normalization was performed by SWAN method and converted to b values. The probes were tested for differential methylation with the limma linear model followed by empirical Bayes methods for the comparisons of interest . Statistical significance was established using the Benjamini-Hochberg false discovery rate (FDR) with a value lower than 0.05. The DNA methylation status and CpG chromosomal location were displayed using the Circos software . Furthermore, the methylation data was visualized by the Wash U Epigenome Browser . 2.5. Integrated Analysis of the Target mRNA-miRNA Regulatory Network Strong validated miRNA-mRNA interactions were identified for the C19MC and MIR371-3 miRNA clusters with the miRTargetLink 2.0 Human tool. miRTargetLink collects information from various databases, including the sources miRBase (v.22.1) and miRTarBase (v.8) . Gene expression analysis on tumor and normal lung tissue were analyzed from the Cancer Genome Atlas datasets (TCGA) using GEPIA 2.0 accessed on 6 September 2021). GEPIA is an interactive web server that compiles data from TCGA and GTEx projects, using a standard processing pipeline. This tool allows for a customizable analysis of the collected data . The molecular function and proposed biological process of the experimentally verified mRNAs were determined using the PANTHER program accessed on 29 September 2021). The PANTHER classification system contains a comprehensive, annotated "library" of phylogenetic trees of gene families designed to classify proteins (and their genes) to facilitate high-throughput analysis . Besides, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database was used to know the molecular interaction network accessed on 3 October 2021) . 2.6. Correlation Analysis in Lung Primary Tumors from TCGA Datasets Transcriptome profiling data from TCGA datasets were downloaded using the CancerMIRNome tool . MicroRNA and mRNA expression data from primary tumors were retrieved from the LUAD (Lung Adenocarcinoma) and LUSC (Lung Squamous Carcinoma) datasets included in TCGA. Only samples labeled as "tumor" and expression level (Log2 Counts per Million (CPM)) > -3.322 were used for the analyses. To test the association between paired miRNA-mRNA profiles, the Pearson correlation coefficients and p-values were computed. p-values lower than 0.05 were considered statistically significant. 2.7. Survival Analysis to Assess the Prognostic Value of Validated Target Genes of the C19MC and MIR371-3 miRNA Clusters in NSCLC Patients To analyze the prognosis associated with the target genes of the C19MC and MIR371-3 miRNA clusters, the Kaplan-Meier survival plots to overall survival time were obtained using the Kaplan-Meier (KM) plotter website , where unprocessed. CEL files from the Gene Expression Omnibus (GEO), the European Genome-phenome Archive (EGA) and the Cancer Genome Atlas (TCGA) repositories were normalized in the R environment. The datasets included in the Kaplan-Meier plotter website are GSE3141, GSE4573, GSE8894, GSE14814, GSE19188, GSE29013, GSE31210, GSE37745, EGA and TCGA (n = 1715 patients). The best-performing threshold from computed lower and upper quartiles was used as cut-off point for the definition of high and low expression of the analyzed genes. Overall survival (OS) was determined from the date of diagnosis to the date of death. p-values lower than 0.05 were considered statistically significant. 3. Results 3.1. DNA Methylation Pattern of the C19MC and MIR371-3 miRNA Clusters in Lung Cancer To evaluate the potential role of the C19MC and MIR371-3 miRNA clusters in lung cancer, we analyzed the methylation status of both of them in human lung tissues from a NSCLC patient cohort (n = 47) and a control cohort (n = 23) of the Virgen del Rocio University Hospital (Seville, Spain). The methylation profile of these clustered miRNAs, which are located on chromosome 19q13.42, were evaluated in human tumor samples compared to paired non-tumoral tissue by using the Illumina Infinium Human Methylation 450 BeadChip. The methylation levels in lung cancer versus paired non-tumoral tissues are represented in Figure 1A and Supplementary Table S2. Patients with lung cancer at our hospital showed DNA hypomethylation at 50 miRNAs included in the C19MC and MIR371-3 clusters after standardisation with non-tumoral control samples. Statistically significant differences (adjusted p-value < 0.05) were detected in the large C19MC cluster (46 miRNAs) and the closely distal MIR371-3 cluster (miR-371a, miR-371b, miR-372 and miR-373). Among all miRNAs, miR-520b, miR-520c, miR-520f, miR-526a1, miR-1283-1 and miR-1283-2 showed greater changes in the DNA hypomethylation pattern in NSCLC patients . We next analyzed the DNA methylation pattern of both clusters in patients at high risk of developing lung cancer, such as patients with COPD. In these patients, we found that the C19MC cluster methylation profile showed no statistical differences compared to the control group . We even found that miRNA-520e, miR-524 and miR-516b2 were hypermethylated in COPD patients versus the non-tumoral control group . However, these differences did not reach statistical significance. In the case of the MIR371-3 cluster, changes in the DNA methylation levels were negligible between both patients groups . 3.2. Transcriptional Mapping of the C19MC and MIR371-3 miRNA Clusters To analyze genomic features associated with different mechanisms of the transcriptional regulation of the C19MC and MIR371-3 miRNA clusters, we used the Wash U Epigenome Browser to display the epigenomic mapping of both clusters (19q13.42) on the reference human genome (hg19; chr19: 54,030,000-54,430,000 genomic coordinates) . Thus, it is shown which transcriptional mechanisms (direct regulators and structural determinants) act at each locus throughout both clusters . We observed that enhancers are distributed outside the regions where the C19MC and MIR371-3 miRNA clusters are located (yellow histograms). In these same regions, we found low activity of polycomb protein-mediated epigenetic regulator (grey histograms). In addition, we identified two heterochromatin-rich domains located in the C19MC cluster region (purple histograms). One of these highly condensed regions is observed at 5' of the C19MC miRNA cluster. Interestingly, the 5' and 3' flanking regions of both clusters showed transcription initiation activity, marked by small active transcriptional start sites (TSS) shown by red histograms. No other region of the C19MC and MIR371-3 miRNA clusters showed active TSS. Behind these active TSS, we found regions with strong transcription activity (green histograms) . We identified a CpG island in the 5' region at the beginning of the C19MC miRNA cluster (~54,150,000 bp; CpG count: 86; and citosine base count plus guanine base count: 762) and another in the MIR371-3 miRNA cluster (~54,270,000 bp; CpG count: 23; and citosine count plus guanine count: 157) (green lines) . However, the CpG sites were distributed throughout both clusters (white lines) . The CG-content was similar from 54,030,000 to 54,430,000 bp on chromosome 19 . 3.3. Methylation Profile of the C19MC and MIR371-3 miRNA Clusters by Histological Subtypes Due to lung cancer-specific hypomethylation, in order to evaluate the potential role of the C19MC and MIR371-3 miRNA clusters as biomarkers in different histological subtypes of lung cancer, we analyzed the methylation pattern in SCC versus adenocarcinoma, pre-normalizing each patient with the methylation status of the matched non-tumoral tissue. The DNA-methylation levels of the C19MC and MIR371-3 clusters were consistently hypomethylated in both histological subtypes in comparison with non-tumoral tissue (Table S2). Furthermore, these differences were statistically significant for both clusters (C19MC, p < 0.001; MIR371-2, p = 0.030) . 3.4. Experimentally Validated miRNA-Target Interactions for the C19MC and MIR371-2 miRNA Clusters We have graphically represented a network with experimentally validated miRNA-mRNA interactions for each component of the C19MC and MIR371-3 clusters in order to study their functional relevance . According to the data included in miRTargetLink software and considering only strong evidence targets, a total of 115 genes were found targeted by at least one of the aforementioned miRNAs. In the network of the C19MC miRNA cluster, the nodes with several connections were those corresponding to miRNA-512-5p, miR-518a-5p, miR-519b-3p, miR-519a-3p, miR-519d-3p, miR-520a-3p, miR-520c-3p, miR-520g-3p, miR-520h, miR-524-5p and miR-525-3p . Among the MIR371-3 cluster compounds, miR-371a-3p, miR-372-3p and miR-373-3p showed a high number of connections. The common target genes possessed a high interaction grade with different miRNAs simultaneously . Of these targets, we identified six genes (CD44, CDKN1A, MTOR, SIRT1, TGFBR2 and VEGFA) that were targeted by miRNAs from both clusters. 3.5. Aberrant Expression of Validated Target Genes for Both miRNA Clusters in Lung Cancer Patients We explored the transcriptional levels of the previously identified validated target genes in cancer and normal lung tissues in the TCGA/GTEx data available in GEPIA2 . Of the 115 target genes validated by miRTargetLink, 31 were significantly underexpressed in at least one of the histological subtypes of NSCLC compared to non-tumor tissue . Seventeen of them showed significant differences in both histological subtypes of lung cancer (adenocarcinoma and SCC). It is worth highlighting those genes that we found targeted by several miRNAs from both clusters, such as the tumor suppressor genes (TSG) CDKN1A (regulated by miR-372-3p, miR-512-5p, miR-515-3p, miR-519a-3p, miR-519b-3p, miR-519e-3p and miR-520a-3p) and TGFBR2 (targeted by miRNA-372-3p, miRN-373-3p and miR-520a-3p). As displayed in Figure 6, we classified the validated targets by gene ontology (GO) molecular function and biological processes using the PANTHER software . The main GO molecular functions were binding (GO:0005488) (39.0%), catalytic activity (GO:0003824) (26.8%) and transcription regulator activity (GO:0001071) (14.6%). The primary binding types were protein binding (GO:0005515) and organic cyclic compound binding (GO:0097159). In the case of catalytic activity, we primarily found hydrolase (GO:0016787), transferase (GO:0016740) and catalytic (GO:0140096) activities. Validated targets displayed three main biological processes: response to stimulus (GO:0050896) (15.0%), biological regulation (GO:0065007) (17.7%) and cellular process (GO:0009987) (19.5%). In the latter, we found represented cell communication (GO:0030234), signal transduction (GO:0007165) and cellular metabolic process (GO:0044237). Finally, we found that 31% of validated targets with aberrant expression in lung cancer were classified in cancer by the KEGG pathways (adjusted p = 7.0 x 10-3). 3.6. Correlation of the Validated Target mRNA-miRNA Expression in Lung Primary Tumors Since we had compelling data on the functional relationship between miRNA-target for the C19MC and MIR371-3 miRNA clusters, as well as differential expression of the validated target genes in lung cancer, we tested whether miRNA-target correlations extended to primary tumors. We studied these correlations using external expression data obtained from the TCGA datasets through the CancerMIRNome tool . We found 11 significant negative correlations for lung adenocarcinoma , all of them for miRNAs included only in the C19MC cluster. On the other hand, 8 were found for SCC , this time including miRNAs from both clusters. It should be noted that CDKN1A was the target gene with the highest number of significant negative correlations (p < 0.05) in both adenocarcinoma (miRNA-512-5p, miRNA-512-3p, miRNA-520a-3p and miRNA-520h) and SCC (miRNA-515-3p, miRNA-519b-3p, miRNA-520a-3p and mir-372-3p). Furthermore, DAPK2 was also negatively correlated with miRNA-520g-3p and miRNA-520h in both histological subtypes of lung cancer (p < 0.01). Other significant correlations for adenocarcinoma were PTK2B-miRNA-517c-3p (p < 0.001), FOXF2-miRNA-519a-3p (p = 0.001), TGFBR2-miRNA-520a-3p (p = 0.009), MICA-miRNA-520c-3p (p = 0.007) and TECEAL1-miRNA-520g-3p (p = 0.037); while for SCC, they were JAG1-miRNA-524-5p (p < 0.001) and KLF13-miRNA-372-3p (p = 0.003). 3.7. Prognostic Role of the Target Gene Network of the C19MC and MIR371-3 Clusters in Lung Cancer To evaluate whether the genes targeted by the C19MC and MIR371-3 clusters with significant negative correlations were associated with clinical outcomes in patients with lung cancer, we analyzed their gene expression levels according to OS data using the KM Plotter website . We found that five target genes were also significantly associated with worsening OS . These genes were FOXF2 (HR = 0.66, 95% CI = 0.58-0.75, p < 0.001), KLF13 (HR = 0.61, 95% CI = 0.52-0.72, p < 0.001), MICA (HR = 0.75, 95% CI = 0.66-0.86, p < 0.001), TCEAL1 (HR = 0.72, 95% CI = 0.63-0.82, p < 0.001) and TGFBR2 (HR = 0.66, 95% CI = 0.58-0.75, p < 0.001). The expression levels of the rest of the genes did not show significant differences regarding the OS in patients with lung cancer. 4. Discussion In the present study, we identified the methylation pattern of the imprinted C19MC and MIR371-3 clusters in patients with NSCLC. Specifically, we identified that all compounds of both clusters are hypomethylated in tumor lung tissue compared to paired normal lung tissue. Importantly, this imprinted cluster-specific methylation signature is restricted to lung cancer, as it is absent in patients with COPD, who have an increased risk of developing lung cancer. Furthermore, these epigenetic changes observed in patients with NSCLC are negatively associated with the expression of relevant target genes in the disease; even five of them were significantly associated with the prognosis of this disease. This methylation profile is consistent with the epigenetic features found at the 19q13.42 locus, in which, the C19MC and MIR371-3 miRNA clusters are located. Epigenetic modifications are characterized by not altering the nucleotide sequence, but by regulating genetic structure and expression reversibly. Early events in tumorigenesis and its progression are related to DNA methylation, histone modifications, nucleosome remodeling and miRNA, which are key epigenetic players . In the particular case of C19MC and MIR371-3 miRNA clusters, the mechanisms involved in regulation were the presence of enhancers, CpG islands, active TSS, heterochromatin-rich domains, as well as low activity of polycomb complexes. In addition, these differences were evident for both histological subtypes of NSCLC. Thus, we found that although all subtypes are hypomethylated, this is more accentuated in SCC than in lung adenocarcinoma. At the same locus, the Protein Tyrosine Phosphatase Receptor type H (PTPRH) has also been confirmed to be hypomethylated and this is correlated with increased gene expression and leads to a poor prognosis in NSCLC . Therefore, the epigenetic features of this region have a significant effect on the disease. Members of the C19MC and MIR371-3 clusters are expressed almost exclusively in human embryonic stem cells (hESC) and rapidly down-regulated during the differentiation process . However, they are over-expressed in cancer , suggesting a tumorigenic role and a possible maintenance function of tumor-associated progenitor cells for these clusters when reactivated . Higher expression of members of the C19MC and MIR371-3 clusters has also been reported in thyroid adenomas and parathyroid carcinomas , germ cell tumors , retinoblastoma , breast cancer , gastric adenocarcinoma and esophageal cancer , among others. And this increase affects tumor growth, differentiation, progression and aggressiveness and, ultimately, patient survival. This cluster overexpression phenomenon also occurs in other sets of miRNAs in NSCLC, such as the miR-23a/27a/24-2 cluster, which has predictive value in early stages and stimulates postoperative progression by inducing tumor suppressor gene silencing . Activation of all miRNAs from the DLK1-DIO3 locus has also been described for human lung adenocarcinoma samples, which is associated with cell stemness and its targets are involved in embryogenesis . Moreover, it is hypomethylated in current and former smokers with NSCLC, suggesting a relevant role in the pathogenesis of lung cancer . The 14q32 miRNA cluster is another example of up-regulation due to DNA hypomethylation in metastatic lung adenocarcinoma patients. Overexpression of this cluster induces cell migration and invasion and has prognostic value . Furthermore, overexpression of the miR-17-92 cluster, which is a highly conserved oncogene cluster, has been frequently reported in lung cancer, especially in the small cell lung cancer subtype, promoting cell growth . Therefore, the gene regulation mechanism mediated by miRNAs is very interesting due to its ability to associate with mRNAs of multiple targets. On the other hand, we identified a total of 115 strongly validated targets for the C19MC or MIR371-3 clusters, and 31 of them presented a significant lower transcriptional level in NSCLC tissue compared to normal. In other words, we identified only those miRNA-mRNA interactions that had been experimentally validated previously. Of these, only six genes were targeted by members of both clusters: CD44, CDKN1A, MTOR, SIRT1, TGFBR2 and VEGFA. In addition, these miRNA-target interactions were verified with TCGA data for primary NSCLC tumors, corroborating significant miRNA-target negative correlations in both clusters and histological subtypes of the disease. Interestingly, redundancy is observed in the miRNAs belonging to C19MC and MIR371-3, since many of them share targets, as it can be inferred from our results. An explanation for this is the high degree of homology that has been described for some C19MC miRNAs through a seed region (5'-AAGUGC-3'), which can be found in several members of the cluster at different positions . This suggests that, in addition to a polycistronic regulation, the targets of these miRNAs are common or share functions. Bioinformatic predictions for this seed region relate these miRNAs to cellular proliferation and apoptosis . Something similar occurs with the MIR371-3 cluster, which has identical seed sequences that are similar to its murine miR290-295 homolog . In this study, we even found that this redundancy occurs between the two clusters C19MC and MIR371-3. The CDKN1A and DAPK2 genes are notable for being negatively correlated with and targeted by multiple miRNAs from these clusters in the two most common subtypes of NSCLC (lung adenocarcinoma and SCC). CDKN1A, also known as p21, plays a critical role in the cellular response to DNA damage, and its overexpression results in p53-mediated cell cycle arrest . It has been reported that CDKN1A/p21 can be blocked in NSCLC by oncomiRs, such as miR-212/132 or miR-93. The latter can act directly or indirectly, thereby inhibiting liver kinase B1 (LKB1) . This promotes proliferation and metastases through the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT) pathway in NSCLC. Other authors have also confirmed the relationship between CDKN1A/p21 and members of the MIR371-3 cluster in hESCs , specifically miR-372. On the other hand, DAPK2 is a serine/threonine kinase that promotes cell apoptosis and autophagy by activating the oncogenic nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) signaling pathway , which sensitizes resistant cells to tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL)-mediated death . DAPK2 expression has been reported to be significantly associated with the poor prognosis in NSCLC . As in this study, DAPK2 is also downregulated by miR-520h in breast cancer and miR-520g in epithelial ovarian cancer and it contributes to chemoresistance. Finally, we found that five of the genes targeted by members of the C19MC and MIR371-3 clusters correlated with worse OS in lung cancer. These genes were FOXF2, KLF13, MICA, TCEAL1 and TGFBR2. They are known transcription regulators and oncogenes, some involved in immune evasion and resistance to immune checkpoint inhibitors , which places them in the focus of current cancer research. For example, according to our results, decreased FOXF2 expression has been reported as an independent predictor of poor prognosis for patients with early-stage NSCLC . FOXF2 deregulation ought to an aberrant DNA methylation status has also recently been identified for gastric cancer . In the case of MICA, Okita et al. reported that PD-L1low/MICA/Bhigh is associated with a better clinical outcome in patients with stage I-IIIA NSCLC. On the other hand, both oncogenic and tumor suppressor roles have been attributed to the transforming growth factor beta (TGF-b) pathway, depending on both the type of tumor and its stage . In this matter, TGFBR2 has recently been proposed as a tumor suppressor with prognostic value in early-stage NSCLC; however, there is no prior evidence of an association between KLF13 and lung cancer prognosis, so further research is required. Our study has the limitation that the mechanism underlying the alterated methylation status of the MIR371-3-imprinted clusters was not evaluated. Cancer-related gene hypomethylation is common in solid tumors, which could contribute to increased expression of oncogenes. In this case, miRNAs may act as oncomiRs by inhibiting the expression of tumor suppressor genes. DNA methylation is a dynamic process regulated by the action of DNA demethylases and DNA methyltransferases (DNMT) . Alterations in the expression or activity of these enzymes can lead to changes in DNA methylation patterns that can trigger alterations in key genes involved in cancer development. For example, Zhang et al. have reported that DNMT1, DNMT3A and DNMT3B show frequency alterations in approximately 3% to 5% of lung cancer patients from the cBioPortal datasets . However, further research is needed to validate the mechanisms in detail, e.g., in organoid in vitro and/or in preclinical in vivo models, to fully understand the mechanisms upstream involved in the aberrant methylation of both clusters and to assess the potential therapeutic applications of targeting them. In addition, we identify relevant miRNA-mRNA interactions in NSCLC patients. In other words, the regulation of the expression of these genes can be partly explained by the activity of these miRNAs; nevertheless, it should be noted that additional mechanisms may also be involved in the regulation of these genes, such as deletion, amplification, mutation, fusion and multiple alterations, and even other miRNAs not included in these clusters, as their regulation can be mediated by the action of many miRNAs. Despite these limitations, this study provides important insights into the consequences of these epigenetic alterations in NSCLC patients and highlights potential targets for future research and therapy. On the other hand, another limitation would be the sample size of patients with COPD included in the study. We have analyzed the DNA methylation pattern of the C19MC and MIR371-3 clusters in a group of COPD patients because they have a six-fold increased risk of developing lung cancer, even if they quit smoking . In our study, we found no statistical differences between the control group without lung cancer and COPD. However, sample size is an important consideration because it may affect the accuracy and reliability of the results. A larger size should be considered to conclude the effect of methylation status of both clusters in this disease. 5. Conclusions In conclusion, this study demonstrates that the imprinted C19MC and MIR371-3 clusters undergo polycistronic epigenetic regulation that leads to differential tumor expression in NSCLC patients. These differences, in turn, deregulate the expression of important and common target genes, many of them with a clear oncogenic and regulatory role in this disease, highlighting five genes (FOXF2, KLF13, MICA, TCEAL1, TGFBR2) that have also a potential prognostic value in NSCLC patients. All these characteristics make the different components of both clusters an interesting target in oncology that needs to be further investigated; it would even be interesting to evaluate the use of methylation agents as an alternative approach to lung cancer therapy. Acknowledgments We would like to acknowledge patients and their families for donating tumor tissue. We would also like to thank Carolina Castillo, as well as the Biobank of the Hospital Virgen del Rocio for kindly providing the samples used in this study. Supplementary Materials The following supporting information can be downloaded at: Table S1: Characteristics of the study and control cohorts; Table S2: Methylation levels of the C19MC and MIR371-3 clusters in lung cancer versus non-tumoral tissue. Click here for additional data file. Author Contributions Conception and design: S.M.-P.; Methodology: L.B., J.F.N.-U., A.C.-P., A.S., M.A., A.S.-G., J.C.B., R.B.-C., L.P.-A. and S.M.-P.; provision of study materials or patients: M.A., A.S.-G., R.B.-C. and L.P.-A.; collection and assembly of data: L.B., A.S., M.A., A.S.-G. and J.C.B.; data analysis and interpretation: L.B. and S.M.-P.; Validation: L.B. and S.M.-P.; supervision: S.M.-P.; Writing--original draft preparation: L.B., J.F.N.-U., A.C.-P., A.S., A.S.-G., M.A., J.C.B., R.B.-C., L.P.-A. and S.M.-P.; Writing--review and editing: L.B. and S.M.-P. Project administration: S.M.-P.; Funding acquisition: S.M.-P. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of Virgen del Rocio University Hospital, Seville, Spain (1381-N-21; 01/07/2020). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are available in this article and supplementary material. Conflicts of Interest L.P.-A. has received honoraria for scientific advice and speaker fees from Lilly, Merck Sharp & Dohme, Bristol-Myers Squibb, Roche, PharmaMar, Merck, AstraZeneca, Novartis, Boehringer Ingelheim, Celgene, Servier, Sysmex, Amgen, Incyte, Pfizer, Ipsen, Adacap, Sanofi, Bayer and Blueprint, and participates as an external member on the board of Genomica. He is founder and board member of Altum sequencing and has received institutional support for contracted research from Merck Sharp & Dohme, Bristol-Myers Squibb, AstraZeneca and Pfizer. The remaining authors declare no potential conflicts of interest. Figure 1 Methylation profiles of the C19MC and MIR371-3 miRNA clusters in lung cancer. (A) Circos plot showing the methylation levels on chromosome 19. From inside to outside: methylation levels, ideogram and gene labels. Hypermethylation (red dots and green background) and hypomethylation events (green dots and red background) in patients with lung cancer versus paired non-tumoral samples. NA: Not available. (B) Observed methylation changes (log2 ratio) in the C19MC cluster. Relative levels of methylation in patients with lung cancer relative to the control group are represented in blue bars, whereas methylation levels of COPD patients without lung cancer compared to the non-tumoral control group are represented by red bars. A grey background represents statistically significant differences (adjusted p-value < 0.05) of methylation levels relative to the control group. (C) Observed methylation changes (log2 ratio) in the MIR371-3 cluster. Relative levels of methylation in patients with lung cancer relative to the control group are represented in blue bars, whereas methylation levels of COPD patients without lung cancer compared to the non-tumoral control group are represented by red bars. A grey background represents statistically significant differences (adjusted p-value < 0.05) of methylation levels in comparison to the non-tumoral control group. Figure 2 Transcriptional mapping of the C19MC and MIR371-3 clusters on the reference human genome hg19. (A) Chromosome 19 ideogram. The C19MC and MIR371-3 clusters are located on 19q13.42. The exact position of the cluster in the region is marked with a blue square. (B) Chromosome position. Base pairs of the C19MC and MIR371-3 clusters on chromosome 19 (highlighted in yellow). Transcriptional mechanisms underlying the expression of both clusters, such as transcription start sites (TSS), enhancer regions (Enh), zinc finger (ZNF), packed form of DNA and polycomb group proteins. (C) CpG islands in the C19MC and MIR371-3 clusters (green lines). (D) Genomic distribution of CpG sites in the C19MC and MIR371-3 clusters (white lines). (E) CG percentage in the C19MC and MIR371-3 clusters (blue). Finally, the reference sequences of the members of the C19MC and MIR371-3 cluster are represented at the bottom. Figure 3 Methylation levels of the C19MC and MIR371-3 clusters components according to the histological subtype of lung cancer. (A) C19MC methylation levels in patients with squamous cell carcinoma (SCC) relative to the control group is represented in purple bars, whereas methylation levels in patients with adenocarcinoma (ADC) compared to the control group are represented in orange bars. (B) Relative MIR371-3 methylation levels in patients with SCC (purple) and ADC (orange) relative to the control group. Relative methylation changes in b values (log2) are represented on the y-axis. Figure 4 miRNAs-target genes network. Strong interactions are displayed for experimentally validated target genes and the components of the C19MC (A) and MIR371-3. (B) clusters generated by miRTargetLink 2.0. Network layout: Random. Nodes are represented as circles: miRNAs (blue) and target genes (green). Figure 5 Transcripts per million (TPM) in normal (green) and tumor (red) lung tissues of those genes identified as strongly validated miRNA targets. (A) C19MC cluster and (B) MIR371-3 cluster. T: Tumor Tissue; N: Normal Tissue; ADC: Adenocarcinoma Lung; SCC: Squamous Cell Carcinoma. * ANOVA test with p-value < 0.01. Figure 6 Classification of validated targets by gene ontology (GO) molecular function and biological processes. Summary of molecular functions and biological processes for validated target genes of miRNAs included in both clusters. Figure 7 miRNA-target correlations in lung primary tumors. Significant negative correlations between miRNA-target for the C19MC and MIR371-3 miRNA clusters in (A) ADC and (B) SCC lung primary tumors. r: Pearson's correlation coefficient; n: sample size, ADC: Lung Adenocarcinoma; SCC: Lung Squamous. p-values lower than 0.05 were considered statistically significant. Figure 8 Clinical outcomes significantly associated with target gene expression in lung cancer. The Kaplan-Meier survival plots were obtained using the Kaplan-Meier plotter website. HR: hazard ratio. cancers-15-01466-t001_Table 1 Table 1 Statistically differences of the methylation levels of C19MC and MIR371-3 clusters between human lung tumor samples and normal lung. Genes Relative Methylation Changes (log2) Adjusted p-Value C19MC miR-512-1 -0.307 6.3946 x 10-14 miR-512-2 -0.341 7.2597 x 10-9 miR-1323 -0.245 1.0347 x 10-6 miR-498 -0.241 2.7704 x 10-5 miR-520e -0.214 5.7827 x 10-5 miR-515-1 -0.346 2.0661 x 10-12 miR-519e -0.340 3.1486 x 10-11 miR-520f -0.402 4.644 x 10-13 miR-515-2 -0.258 2.686 x 10-9 miR-519c -0.264 2.9066 x 10-10 miR-1283-1 -0.601 4.8535 x 10-17 miR-520a -0.279 4.1063 x 10-11 miR-526b -0.329 5.2657 x 10-14 miR-519b -0.348 1.5969 x 10-15 miR-525 -0.360 1.2282 x 10-10 miR-523 -0.267 1.1234 x 10-11 miR-518f -0.368 7.9986 x 10-15 miR-520b -0.441 4.5322 x 10-16 miR-518b -0.290 1.7394 x 10-9 miR-526a1 -0.474 3.4512 x 10-14 miR-520c -0.590 4.7234 x 10-18 miR-518c -0.271 6.5541 x 10-9 miR-524 -0.261 1.5518 x 10-12 miR-517a -0.129 1.6843 x 10-5 miR-519d -0.140 3.7082 x 10-5 miR-521-2 -0.343 1.1838 x 10-14 miR-520d -0.378 2.9354 x 10-13 miR-517b -0.327 2.6624 x 10-12 miR-520g -0.293 9.6092 x 10-11 miR-516b2 -0.215 1.1738 x 10-8 miR-526a2 -0.303 9.5413 x 10-10 miR-518e -0.168 6.8329 x 10-7 miR-518a1 -0.139 7.0465 x 10-5 miR-518d -0.412 2.144 x 10-12 miR-516b1 -0.269 9.8287 x 10-10 miR-518a2 -0.364 2.9253 x 10-14 miR-517c -0.199 1.476 x 10-10 miR-520h -0.211 2.5861 x 10-10 miR-521-1 -0.183 8.9968 x 10-5 miR-522 -0.235 5.605 x 10-10 miR-519a1 -0.217 9.8489 x 10-11 miR-527 -0.130 7.4475 x 10-9 miR-516a1 -0.400 2.5393 x 10-9 miR-1283-2 -0.453 2.9539 x 10-13 miR-516a2 -0.340 3.3131 x 10-16 miR-519a2 -0.344 1.2518 x 10-14 MIR371-3 miR-371b -0.209 1.4784 x 10-8 miR-371a -0.106 7.6971 x 10-7 miR-372 -0.106 7.6971 x 10-7 miR-373 -0.137 1.183 x 10-7 cancers-15-01466-t002_Table 2 Table 2 Common miRNA-target interactions for both C19MC and MIR371-2 miRNA clusters. Gene Gene Description miRNA BTG1 B-Cell Translocation Gene 1 miRNA-372-3p, miRNA-373-3p CD44 Cluster of Differentiation 44 miRNA-520a-3p, miRNA-520c-3p, miRNA-373-3p CDK2 Cyclin Dependent Kinase 2 miRNA-524-5p, miRNA-372-3p CDKN1A Cyclin Dependent Kinase Inhibitor 1A miRNA-512-5p, miRNA-515-3p, miRNA-519a-3p, miRNA-519b-3p, miRNA-519d-3p, miRNA-519e-3p, miRNA-520a-3p, miRNA-520h, miRNA-373-3p DAPK2 Death Associated Protein Kinase 2 miRNA-520h, miRNA-520g-3p DKK1 Dickkopf WNT Signaling Pathway Inhibitor 1 miRNA-371a-3p, miRNA-372-3p, miRNA-373-3p ELAVL1 ELAV (Embryonic Lethal, Abnormal Vision, Drosophila)-Like RNA Binding Protein 1 miRNA-519a-3p, miRNA-519b-3p LATS2 Large Tumor Suppressor Kinase 2 miRNA-372-3p, miRNA-373-3p LEFTY1 Left-Right Determination Factor 1 miRNA-372-3p, miRNA-373-3p MCL1 Myeloid Cell Leukemia Sequence 1 miRNA-512-5p, miRNA-518a-5p MMP2 Matrix Metallopeptidase 2 miRNA-519d-3p, miRNA-520g-3p, miRNA-524-5p MTOR Mechanistic Target of Rapamycin Kinase miRNA-520c-3p, miRNA-373-3p NFIB Nuclear Factor I B miRNA-372-3p, miRNA-373-3p PTEN Phosphatase and Tensin Homolog miRNA-518c-3p, miRNA-519a-3p, miRNA-519d-3p PTK2B Protein Tyrosine Kinase 2 Beta miRNA-517a-3p, miRNA-517c-3p SIRT1 Sirtuin 1 miRNA-520c-3p, miRNA-373-3p SAMD7 Sterile Alpha Motif Domain Containing 7 miRNA-519d-3p, miRNA-520g-3p STAT3 Signal Transducer and Activator of Transcription 3 miRNA-519a-3p, miRNA-519g-3p, miRNA-520c-3p TGFBR2 Transforming Growth Factor Beta Receptor 2 miRNA-520a-3p, miRNA-372-3p, miRNA-373-3p TNFAIP1 Tumor Necrosis Factor Alpha-Induced Protein 1 miRNA-372-3p, miRNA-373-3p VEGFA Vascular Endothelial Growth Factor A miRNA-520g-3p, miRNA-520h, miRNA-372-3p, miRNA-373-3p cancers-15-01466-t003_Table 3 Table 3 Target genes with significant expression differences in NSCLC tissue compared to normal lung tissue. Symbol Description ARRB1 Arrestin Beta 1 AKT3 AKT Serine/Threonine Kinase 3 CCL2 C-C Motif Chemokine Ligand 2 CDKN1A Cyclin Dependent Kinase Inhibitor 1A DAPK2 Death Associated Protein Kinase 2 DICER1 Double-Stranded RNA-Specific Endoribonuclease FOXF2 Forkhead Box F2 GPC3 Glypican 3 JAG1 Jagged Canonical Notch Ligand 1 JAK1 Janus Kinase 1 KLF13 Kruppel Like Factor 13 LEFTY2 Left-Right Determination Factor 2 LATS2 Large Tumor Suppressor Kinase 2 MBNL2 Muscleblind-Like Protein 2 MCL1 Myeloid Cell Leukemia Sequence 1 MICA Major Histocompatibility Complex Class I Chain-Related Protein A MMP2 Matrix Metallopeptidase 2 NFIB Mechanistic Target of Rapamycin Kinase (MTOR), Nuclear Factor I B NR4A2 Nuclear Receptor Subfamily 4 Group A Member 2 PIK3C2A Phosphatidylinositol-4-Phosphate 3-Kinase Catalytic Subunit Type 2 Alpha PLCB4 Phospholipase C Beta 4 PTK2B Protein Tyrosine Kinase 2 Beta TCEAL1 Transcription Elongation Factor A Like 1 TEAD4 TEA Domain Transcription Factor 4 TGFBR2 Transforming Growth Factor Beta Receptor 2 TIMP2 Tissue Inhibitor of Metalloproteinases 2 TXNIP Thioredoxin Interacting Protein RASSF1 Ras Association Domain Family Member 1 RECK Reversion Inducing Cysteine Rich Protein with Kazal Motifs SMAD7 SMAD (small Mothers Against Decapentaplegic) family member 7 STX12 Syntaxin 12 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000579 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050719 cells-12-00719 Review Roles of Astrocytic Endothelin ETB Receptor in Traumatic Brain Injury Michinaga Shotaro 1 Hishinuma Shigeru 1 Koyama Yutaka 2* Brenner Michael Academic Editor Parpura Vladimir Academic Editor 1 Department of Pharmacodynamics, Meiji Pharmaceutical University, 2-522-1 Noshio, Tokyo 204-8588, Japan 2 Laboratory of Pharmacology, Kobe Pharmaceutical University, 4-19-1 Motoyama-Kita Higashinada, Kobe 668-8558, Japan * Correspondence: [email protected]; Tel.: +81-78-441-7572 24 2 2023 3 2023 12 5 71925 11 2022 08 2 2023 22 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Traumatic brain injury (TBI) is an intracranial injury caused by accidents, falls, or sports. The production of endothelins (ETs) is increased in the injured brain. ET receptors are classified into distinct types, including ETA receptor (ETA-R) and ETB receptor (ETB-R). ETB-R is highly expressed in reactive astrocytes and upregulated by TBI. Activation of astrocytic ETB-R promotes conversion to reactive astrocytes and the production of astrocyte-derived bioactive factors, including vascular permeability regulators and cytokines, which cause blood-brain barrier (BBB) disruption, brain edema, and neuroinflammation in the acute phase of TBI. ETB-R antagonists alleviate BBB disruption and brain edema in animal models of TBI. The activation of astrocytic ETB receptors also enhances the production of various neurotrophic factors. These astrocyte-derived neurotrophic factors promote the repair of the damaged nervous system in the recovery phase of patients with TBI. Thus, astrocytic ETB-R is expected to be a promising drug target for TBI in both the acute and recovery phases. This article reviews recent observations on the role of astrocytic ETB receptors in TBI. astrocyte traumatic brain injury endothelin ETB receptor blood-brain barrier neuroinflammation Japan Society for the Promotion of Science20K16016 21K06609 This research was funded by the Japan Society for the Promotion of Science (grant numbers: 20K16016 and 21K06609). pmc1. Introduction Traumatic brain injury (TBI) is critical damage to the brain caused by a sudden insult, such as traffic accidents, falls, collisions, and sporting activities. TBI is a major cause of death and disability worldwide. Even in surviving patients, TBI causes severe sequelae in motor, sensory, mental, and cognitive functions, resulting in decreased quality of life (QOL). Therefore, much effort has been devoted to realizing effective therapies for TBI, that is, treatments to protect the brain from damage in the acute phase and promote the recovery of neurological function in TBI patients with sequelae. Brain damage caused by TBI is classified as primary or secondary . Primary damage includes direct physical injury to the brain parenchyma, such as skull fractures, intracranial hemorrhage, compression/deformation of nerve tissue, diffuse axonal injury, and crushing of blood vessels. Biochemical, cellular, and physiological alterations induced by primary damage propagate from the impact core to the peripheral area and aggravate brain injury (secondary damage) . Pathological events that induce secondary damage in TBI around the impact core include excitotoxicity, cerebral hypoperfusion, brain edema, and neuroinflammation. While primary damage is irreversible and difficult to reduce, secondary damage is partly reversible and remediable. Therefore, therapies for TBI in the acute phase focus on reducing secondary damage. Current treatments for the acute phase of TBI include decompressive craniotomy, hyperosmolar treatment, barbiturates, sedation, and hypothermia therapy. However, these treatments are insufficient and may have adverse effects in some cases. Several candidate drugs have shown beneficial effects in preclinical studies using experimental TBI animal models, but clinical trials have failed to show significant effects in patients with TBI . In addition, for TBI patients with sequelae, treatments to promote the recovery of neurological functions impaired by TBI are required, which are currently performed by physical therapy. Although many studies have shown that physical therapy promotes synaptic regeneration in damaged nervous systems , no medication is clinically used to enhance its efficiency. Therefore, research and development of medicines applied in the acute and recovery phases of TBI have been extensively conducted. Many studies have clarified the roles of astrocytes in nerve damage and recovery processes in several brain disorders, including TBI . Based on these studies, the regulation of astrocytic functions has been proposed as a novel therapeutic strategy for TBI. Endothelin (ET) is one of the factors regulating the pathophysiological functions of astrocytes in damaged nerve tissues . ET receptor signaling-mediated pathophysiological reactions include ischemia, neuropathic pain, and disruption of the blood-brain barrier (BBB) . We previously observed that ETB receptor (ETB-R) but not ETA receptor (ETA-R) was highly distributed in astrocytes and that an ETB-R antagonist but not an ETA-R antagonist alleviated pathological conditions, including the proliferation of reactive astrocytes, BBB disruption, and brain edema in TBI model mice . In this review, the roles of the ETB receptor (ETB-R) in astrocytic functions in TBI are reviewed. In addition, the possibility of astrocytic ETB-R as a novel drug target for TBI is discussed. 2. Pathophysiological Responses of Astrocytes to TBI In response to brain disorders, astrocytes change their phenotype to that of reactive astrocytes, which are characterized by increased glial fibrillary acidic protein (GFAP) expression and hypertrophy. Reactive astrocytes are involved in the progression of many brain pathologies and the regeneration of the injured nervous system. In patients with TBI, phenotypic conversion to reactive astrocytes is predominantly observed in damaged areas . Similarly, reactive astrocytes were also increased in experimental TBI model animals . Brain edema occurs during the acute phase of TBI. Increased intracranial pressure accompanied by brain edema causes impairment of the nervous system and often results in the death of patients with TBI. In addition, neuroinflammation in the acute phase exacerbates neuronal damage caused by TBI and causes various neurological dysfunctions in patients affecting motor, sensory, and cognitive activities. Disruption of the BBB underlies the development of brain edema and neuroinflammation caused by TBI. That is, the hyperpermeability of brain microvascular endothelial cells, which constitute the BBB, can allow the infiltration of inflammatory cells and serum proteins into the cerebral parenchyma damaged by TBI. Astrocytes support the integrity of the BBB, and their end feet surround a large part of the basolateral side of brain microvessels. The permeability of brain microvascular endothelial cells responsible for the BBB is regulated by their interaction with astrocytes. Functional alterations in astrocytes in TBI lead to excessive hyperpermeability of the BBB, which allows the entry of inflammatory cells and serum proteins . Increased production of various cytokines and chemokines in the acute phase of TBI has been reported . The production of astrocytic cytokines and chemokines is stimulated by several signaling molecules released from damaged cells . In TBI, astrocytic IL-33 is increased in the human and mouse brain and promotes the accumulation of microglia/macrophages at the site of injury . Xue et al. showed that astrocytes produce C-C Motif Chemokine Ligand 7 (CCL7), which promotes microglia-mediated inflammation in a TBI rat model. Other astrocytic vascular permeability regulators, such as matrix metalloproteinase 9 (MMP9) and vascular endothelial growth factor-A (VEGF-A), are also upregulated by TBI and cause disruption of the BBB . These results indicate that astrocytes have a detrimental effect on BBB function during the acute phase of TBI. However, some reports suggest that astrocytes have supporting roles in BBB function, by which brain edema and neuroinflammation in TBI are reduced. Hu et al. showed that the ablation of astrocytes exacerbated the infiltration of monocytes into the cerebral parenchyma and neuronal loss in mice with brain stab injuries. Gao et al. found that programmed cell death 1 (PD-L1) signaling in reactive astrocytes prevented excessive neuroimmune and neuronal damage in a controlled cortical impact-induced TBI mouse model. Astrocyte-derived exosomes also protect hippocampal neurons by suppressing mitochondrial oxidative stress and apoptosis in rats with TBI . We also found that astrocyte-derived vascular protective factors, such as angiopoietin-1 (ANG-1) and sonic hedgehog (SHH), were increased in TBI model mice . These findings suggest that astrocytes suppress TBI-induced neuroinflammation and BBB disruption and exert protective actions against neuronal damage in TBI. Additionally, expression levels of MMP-9 and VEGF-A increased at 6 h to 5 days after TBI, whereas expression levels of ANG-1 and SHH increased at 3 to 10 days after TBI . This finding shows a possibility that conversion to reactive astrocytes enhances both detrimental and supportive actions on BBB function depending on the phase of TBI. During the recovery phase of TBI, new synapses are formed in damaged areas, which are supported by neurogenesis from neural progenitors and axonal elongation . This remodeling of the damaged nervous system is the mechanism that underlies the recovery of brain functions impaired by TBI. Some astrocyte-derived factors have been shown to promote the remodeling of the nervous system damaged by TBI. Astrocyte-produced apolipoprotein E and S100b promote neurogenesis and recovery of cognitive function impairments in TBI. Thrombospondins (TSPs) are astrocyte-secreted proteins that promote synaptogenesis . Cheng et al. reported that TSP-1 was increased in TBI and that TSP-1 knockout mice exhibited significantly worse neurological deficits in motor and cognitive functions . Production of neurotrophin family neurotrophic factors is upregulated by TBI and promotes neuroprotection in the acute phase, as well as regeneration in the recovery phase . Astrocytes are the major source of nerve growth factor (NGF), which is also upregulated by TBI . Administration of NGF to the rat brain reversed the decrease in cholinergic nerves induced by TBI and enhanced cognitive function . Treatment to increase the production of brain-derived neurotrophic factor (BDNF) in astrocytes restored neuronal function impaired by TBI . Hao et al. showed that exogenous neurotrophin-3 (NT-3) administration to TBI model rats promoted neural stem cell proliferation and synaptogenesis . These findings indicate that the ability of astrocytes to produce neurotrophic factors is beneficial in promoting nerve regeneration during the recovery phase of TBI. 3. Endothelin in TBI 3.1. Endothelin Receptor Signal Pathways and Pathophysiological Reactions The ET ligand family, ET-1, ET-2, and ET-3, was initially discovered as vasoconstrictor peptides produced by vascular endothelial cells. ET-1 is present in the brain and is one of the factors involved in pathophysiological responses of damaged nerve tissues . ET-2 is largely limited to the gastrointestinal tract, sex organs, and pituitary gland, and ET-3 is abundantly expressed in the intestine, pituitary gland, and brain . Receptors for ETs, which are seven-transmembrane G-protein-coupled receptors, are classified into two distinct types: ETA-R and ETB-R, which are encoded by the EDNRA and EDNRB genes, respectively . ETA-R shows higher affinities for ET-1 and ET-2 than for ET-3, whereas ETB-R has an equal affinity for all three ET ligands. Both ETA-R and ETB-R are linked to the Gq protein and increase intracellular Ca2+ by activating phospholipase C (PLC). However, ETA-R and ETB-R have different regulatory mechanisms for adenylate cyclase-mediated signals. ETA-R is Gs-linked to increase cAMP, whereas ETB-R is linked to Gi and suppresses the signal . Both ETA-R and ETB-R are also linked to the G12/13 protein . The G12/13 protein-mediated signal activates the Rho protein, a low molecular weight G-protein, and stimulates Rho-associated protein kinase (ROCK), which regulates cellular proliferation, Ca2+, and cytoskeletal actin reorganization . Several studies imply that ET-R-mediated calcium-dependent signaling contributes to neuroinflammation. ET-1 increases intracellular Ca2+ by the influx of calcium and release of intracellular calcium stores. ETA-R-mediated calcium-dependent responses include activation and degranulation of neutrophils by calcium influx . ETB-R-mediated calcium-dependent responses include the chemotactic migration of neutrophils by the release of intracellular calcium . In addition, astrocytic dysregulation of Ca2+ homeostasis promotes the release of inflammatory factors, which cause neuroinflammation in Alzheimer's disease . Because ETB-R is highly expressed in astrocytes and ETB-R signaling promotes the production of astrocytic bioactive factors , ETB-R signaling may be involved in neuroinflammation by astrocytic dysregulation of Ca2+ homeostasis. These observations suggest that activation of ET-Rs and neuroinflammation are correlated to calcium homeostasis dysregulation. The pathophysiological roles of ETs in the cardiovascular system, such as arterial hypertension, myocardial infarction, preeclampsia, and coronary atherosclerosis, have been well investigated. In the clinical state, several ET-R antagonists, including bosentan (non-selective), ambrisentan (ETA-R selective), and macitentan (non-selective), have been applied as therapeutic drugs for pulmonary arterial hypertension. ET antagonists are also expected to be therapeutic drugs for cardiovascular and other disorders . In patients with several neurodegenerative disorders, cerebral ischemia, and subarachnoid hemorrhage, expression levels of ETs are increased . Some studies showed that in addition to vascular endothelial cells, astrocytes are also a primary source of ET-1. The expression level of astrocytic ET-1 is increased by several factors, including cytokines, hypoxia, and ET-1 itself . 3.2. Relationships of Endothelin and TBI Brain ET-1 production was found to be increased by TBI, whereas productions of ET-2 and ET-3 have not been investigated. Studies in patients with TBI and experimental animal models indicate that increases in brain ET-1 are closely related to the pathological conditions of TBI. Maier et al. showed that ET-1 is increased in the plasma and cerebrospinal fluid of patients with TBI. Chen et al. also found that ET-1 levels were significantly higher in the severe TBI group than in the mild/moderate TBI and control groups. Additionally, ET-1 is also increased in the cerebrospinal fluid and is associated with unfavorable outcomes in children after severe TBI . In experimental TBI models, ET-1 levels are increased in the brain . Histological observations have shown that reactive astrocytes produce ET-1 in TBI . An examination using cultured cells showed that a direct physical effect on astrocytes is partially involved in TBI-induced astrocytic ET-1 production . TBI causes cerebral circulation dysfunction, such as vasospasm and hyperpermeability of brain microvessels, which aggravates secondary damage . Experimental animal models of cerebral ischemia and subarachnoid hemorrhage have shown that increases in brain ET-1 cause vasospasm through ETA-R . ETA-R mediates vasospasm through protein kinase C, whereas ETB-R mediates vasodilation through nitric oxide synthesis . These experimental results suggest that antagonism of ETA-R may also alleviate vasospasm in the brain damaged by TBI. ETB-Rs are expressed in astrocytes and brain vascular endothelial cells , whereas ETA-Rs are highly expressed in brain blood vessels but not in astrocytes . Expression of astrocytic ETB-Rs is also upregulated in response to brain disorders, as well as astrocytic ET-1 . We found that ETB-R expression in reactive astrocytes was significantly increased in the injured cerebrum of TBI mice . The concomitant increases in both astrocytic ET ligands and receptors suggest an autocrine/paracrine mechanism mediated by ET-1/ETB-R signaling, which becomes more prominent in the regulation of astrocytic pathophysiological actions in TBI . Therefore, the role of ETB-R-mediated signaling in astrocytic pathophysiological responses has been investigated. 4. Regulation of Astrocytic Functions by ETB-R 4.1. ETB-R-Mediated Conversion to Reactive Astrocytes Conversion of resting astrocytes to reactive phenotypes in response to brain injury is characterized by increased GFAP expression and hypertrophy. Hyperplasia of reactive astrocytes causes glial scar formation in injured areas, which prevents axonal elongation during the recovery of damaged nerve tissues. ET-1 is one of the factors that induce conversion to reactive astrocytes in injured nerve tissues. Intracerebroventricular administration of Ala1,3,11,15-ET-1, a selective ETB-R agonist, promotes the reactive conversion of astrocytes in the rat brain . In experimental brain injury models, inhibition of ETB-R decreases the number of reactive astrocytes . We also found that BQ788, a selective ETB-R antagonist, reduced the induction of reactive astrocytes in TBI model mice . 4.2. ETB-R-Mediated Proliferation Signal Pathways in Astrocytes The proliferation of reactive astrocytes is regulated by multiple intracellular signaling pathways. Several studies have suggested that signal transducer and activator of transcription 3 (STAT3) is a key transcription factor for astrocytic proliferation in brain injury . Selective nuclear localization of phosphorylated STAT3 in reactive astrocytes was observed in the cerebrum of TBI model rats . We also found that phosphorylated STAT3 increased in the cerebrum of mice with TBI . In cultured astrocytes, ET-1 promotes STAT3 activation through ETB-R, and inhibition of STAT3 reduces cell proliferation , suggesting that STAT3 is involved in ET-induced astrocytic proliferation. Transcriptional target molecules of STAT3 include cyclin D1 and S-phase kinase-associated protein 2 (Skp-2), cell cycle regulators that stimulate the G1/S transition . Histological observations in brain injury models have shown that cyclin D1 is expressed in reactive astrocytes . The increase in cyclin D1 expression mediates astrocytic proliferation triggered by ETB-R/STAT3 signaling . Other transcription factors than STAT3 are also involved in astrocytic cyclin D1 expression. We found that ET-1-activated specificity protein-1 (SP-1) promotes the expression of cyclin D1 and the proliferation of cultured astrocytes through ETB-R . ET-1-induced SP-1 activation and cyclin D1 expression were suppressed by inhibitors of mitogen-activated protein kinases (MAPKs) . Gadea et al. also showed that ETB-R-mediated astrocyte proliferation and activation occur through the activation of JNK-dependent pathways. ET-induced astrocytic proliferation is coordinately regulated by both cell adhesion-dependent and independent pathways under ETB-R activation . The expression of cyclin D1 through the ERK and JNK pathways is in a cell adhesion-independent pathway, whereas astrocytic cyclin D3, another G1/S cyclin, is increased by the stimulation of ETB-R in a cell adhesion-dependent manner. In the adhesion-dependent mechanism, a small G-protein (Rho) and focal adhesion kinase (FAK) mediate the cell proliferation signal from ETB-R to cyclin D3 . This diversity of proliferation pathways characterizes the ETB-R signal in hyperplasia of reactive astrocytes, leading to glial scar formation . 4.3. ETB- Production of Bioactive Factors in Astrocytes In the damaged brain, reactive astrocytes produce multiple bioactive factors that increase the permeability of microvascular endothelial cells. The production of these factors disrupts the BBB, which leads to brain edema and infiltration of inflammatory cells. Activation of ETB-R in cultured astrocytes and rat brains stimulated the production of MMP-9, MMP-2, VEGF-A, and stromelysin-1 , which increases vascular permeability. At brain injury sites, activated astrocytes produce monocyte chemoattractant protein-1 (MCP-1)/C-C motif chemokine 2 (CCL2) and cytokine-induced neutrophil chemoattractant-1 (CINC-1)/C-X-C motif chemokine ligand 1 (CXCL1) . In addition to increasing the vascular permeability of brain microvessels, these chemokines serve as chemoattractants that lead to blood inflammatory cells in the brain. The production of astrocytic MCP-1/CCL2 and CINC-1/CXCL1 is increased by the stimulation of ETB-R . In contrast, stimulation of ETB-R decreased the expression of astrocytic ANG-1 and SHH , which are factors that enhance the integrity of brain microvessels and protect the BBB from brain injury . These actions of ET-1 on vascular permeability-regulating factors suggest that the activation of astrocytic ETB-R signaling impairs BBB function, resulting in brain edema and neuroinflammation. As described above, astrocytes also play a role in supporting neuroprotection and repair of damaged nerve systems. Stimulation of ETB-R increases the production of factors promoting neuroprotection/neuronal regeneration, including glial cell line-derived neurotrophic factor (GDNF), BDNF, NT-3, NGF, and basic fibroblast growth factor (bFGF) . Ephrin is a membrane-associated protein family that regulates adhesion-dependent cellular functions by interacting with Eph receptors . Several ephrin subtypes are expressed in astrocytes. Increased expression of astrocytic ephrins in brain disorders inhibited axon elongation and synaptogenesis during the regeneration of damaged nerve tissues . The expression of ephrin-A2, -A4, -B2, and -B3 in cultured astrocytes was decreased by stimulation of ETB-R . Both the increase in neurotrophic factors and the decrease in ephrins suggest that stimulation of astrocytic ETB receptors plays a role in promoting neuroprotection/neuronal regeneration of damaged nerve tissues. 5. Roles of Astrocytic ETB-R in the Acute Phase of TBI Under normal conditions, the function of the BBB is maintained by a balance between astrocyte-derived factors that increase vascular permeability and enhance the integrity of brain microvascular endothelial cells. In the acute phase of TBI, the production of astrocytic factors that enhance vascular permeability causes BBB disruption in the injured areas, which results in brain edema and neuroinflammation. ET-1 is upregulated in the acute phase of TBI and regulates the production of these astrocytic factors through ETB-R . Therefore, we investigated the effects of ETB-R antagonists on BBB disruption in fluid percussion injury (FPI)-induced TBI model mice . In TBI model mice, BBB permeability was remarkably increased one to five days after FPI. Furthermore, an increased expression level of ETB-R and proliferation of reactive astrocytes were also observed . Thus, proliferation signal pathways of astrocytic ETB-R are activated and promote BBB permeability in TBI. Intracerebroventricular administration of the ETB-R antagonist BQ788 from 2 to 5 days after FPI significantly reduced TBI-induced BBB disruption and brain edema in the acute phase (three to five days after FPI), accompanied by a reduction in the number of GFAP-positive reactive astrocytes . Thus, an ETB-R antagonist inhibits excessive proliferation signal pathways of astrocytic ETB-R, resulting in the reduction in BBB disruption and brain edema. On the other hand, administration of the ETA-R antagonist FR139317 from 2 to 5 days after FPI had no significant effects on BBB disruption and increased reactive astrocytes . Guo et al. found that BQ788 reduced neutrophils and monocytes and decreased the expression levels of cytokines and inflammatory mediators in the spinal cord after contusion injury. Although no selective ETB receptor antagonists, including BQ788, have been clinically applied, some ET antagonists have been used as therapeutics for pulmonary arterial hypertension . We also found that the intravenous administration of bosentan, a non-selective ET antagonist, reduced FPI-induced BBB disruption and brain edema . These results indicate that antagonism of ETB-R during the acute phase of TBI is effective in protecting the BBB. BBB protection by ETB-R antagonists is mediated by the normalization of the expression of astrocyte-derived vascular permeability regulators. TBI increased the expression levels of MMP-9 and VEGF-A in reactive astrocytes . Administration of BQ788 and bosentan significantly decreased the expression of MMP-9 and VEGF-A in FPI-induced TBI model mice . On the other hand, expression levels of astrocytic ANG-1 and SHH, which were increased by TBI, were further increased by the administration of BQ788 and bosentan . Supporting the results in a TBI animal model, we confirmed that the altered expression of these permeability regulators by ET-1 was inhibited by BQ788 in cultured astrocytes . These results imply that antagonism of ETB-R in the acute phase of TBI can alleviate BBB disruption by controlling the expression of astrocyte-derived bioactive factors . 6. Roles of Astrocytic ETB-R in the Recovery Phase of TBI Neuroinflammation and cerebral ischemia in the acute phase of TBI induce neuronal degradation in injured areas, leading to impairment of motor, cognitive, and emotional functions in patients with TBI . To improve these nerve dysfunctions, it is important to reduce damage in the acute phase of TBI and promote neurogenesis and synaptic regeneration in the recovery phase, which is conventionally performed by physical therapy. The direct application of several neurotrophic factors was initially trialed to recover nerve dysfunction in patients with brain insults and neurodegenerative diseases. However, owing to the difficulty of intracerebral delivery and the occurrence of side effects, none of them has been clinically used . Recently, a clinical trial of genetically modified bone marrow-derived mesenchymal stem cells (SB623) in patients with TBI in the recovery phase was conducted to examine the effectiveness of recovery of impaired nerve functions . However, no low molecular weight drugs used for peripheral administration have been clinically applied, and a new therapeutic strategy is desired. Reactive astrocytes produce many factors that regulate the regeneration of the nervous system, including neurotrophic factors and cell adhesion molecules . Regulating the production of astrocyte-derived factors is proposed as an effective approach to promote the regeneration of the damaged nervous system . Some astrocyte-derived factors have been shown to play roles in nerve regeneration after TBI , suggesting that impaired nerve function can be recovered by regulating astrocytic function. ETB-R signaling stimulates the production of neurotrophic factors, including NGF, bFGF, BDNF, NT-3, and VEGF, . In addition, ETB-R signaling decreased the expression of ephrins, which are negative regulators of neurite outgrowth and neurogenesis in the repair of damaged nerve systems . The effects of ETB-R signaling on the expression of trophic factors and ephrins suggest that stimulation of astrocytic ETB-R promotes the regeneration of the nervous system. In TBI model mice, we found that increased expression of ETB-R and reactive astrocytes were also observed not only in the acute phase but also in the recovery phase (seven days after FPI) . Thus, the proliferation signal pathway of astrocytic ETB-R is also activated in the recovery phase and enhances the production of astrocytic neurotrophic factors. In addition, ET-1 may affect neural progenitor cells to promote the regeneration of the nervous system through ETB-R. During the development of the enteric nervous system, the proliferation and migration of neural progenitor cells are regulated by the ET-3/ETB-R signal . Nishikawa et al. observed in an ex vivo culture system of mouse cortical tissue that an ETB-R agonist, IRL-1620 (sovateltide), enhanced interkinetic nuclear migration of cerebral neural progenitor cells, whereas BQ788, an ETB-R antagonist, reduced it . This report also showed that BQ788 decreased the number of newborn neurons. Although the role of ETB-R in nerve regeneration after TBI has not been investigated, some studies in animal models have suggested that stimulation of ETB-R promotes regeneration of the damaged nervous system. Leonard et al. showed that intravenous treatment with IRL-1620 enhanced angiogenesis and neurogenesis and reduced neurological damage in a rat model of permanent cerebral ischemia. Briyal et al. suggested that IRL-1620 prevents beta-amyloid-induced oxidative stress and cognitive impairment in rats. These findings raise the possibility that stimulation of ETB-R during the recovery phase of TBI also has a beneficial effect on the improvement of impaired nerve function . 7. Significance of Astrocytic ETB-R as a Therapeutic Target of TBI In this article, we have discussed the importance of ETB-R signaling in regulating the pathophysiological functions of astrocytes, which show both detrimental and beneficial effects on the nervous system in response to brain disorders. Both astrocytic actions were promoted by stimulation with ETB-R. This indicates that the selective use of ETB-R antagonists/agonists according to the pathological course of TBI is necessary to obtain therapeutic benefits. In the acute phase of TBI, the increased production of ET-1 induces BBB disruption and brain edema through astrocytic ETB-R. Therefore, ETB-R antagonists may be effective in reducing secondary damage. Some drugs with ETB-R antagonistic activity, such as bosentan and macitentan, have been clinically applied for pulmonary hypertension . These ET antagonists are expected to expand their clinical application to protect against nerve damage during the acute phase of TBI. In contrast, stimulation of astrocytic ETB-R increases the production of several neurotrophic factors and decreases the expression of ephrin subtypes. Because these ETB-R-mediated functional alterations of astrocytes can promote regeneration of the impaired nervous system, the application of ETB-R agonists may show beneficial effects in patients with TBI during the recovery phase. In clinical trials of selective ETB agonists, IRL-1620 has been evaluated in patients with acute cerebral ischemic stroke . Despite societal demands, no effective drug treatments are currently in clinical use for patients with TBI. In addition, targeting drugs for astrocytes have not been developed although multiple studies suggest that astrocytes play key roles in the pathogenesis of TBI. Because astrocytic ETB-R signaling is closely involved in the pathogenesis of TBI and regulates astrocyte functions including activation, proliferation, and production of bioactive factors, targeting drugs for astrocytic ETB-R may resolve two problems at the same time. Although further preclinical verification is required regarding its efficacy, astrocytic ETB-R is expected to be a novel target for drugs that aim to protect/improve brain functions impaired by TBI. Author Contributions Writing, review, and editing, Y.K., S.H. and S.M. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Involvement of astrocytes in BBB disruption in the acute phase of traumatic brain injury (TBI). In the acute phase of TBI, activation of astrocytes occurs around the damaged brain area. The astrocytic activation is accompanied by increased expressions of astrocyte-derived vascular permeability factors, such as vascular endothelial growth factor-A (VEGF-A), matrix metalloproteinase 9 (MMP9), and chemokines. This impairs the barrier function of the blood-brain barrier (BBB) and allows entry of inflammatory cells and serum proteins into the brain parenchyma, which leads to neuroinflammation and brain edema. Figure 2 Classification of endothelin receptors. Endothelins (ETs) exert their physiological actions through two G-protein-coupled receptor subtypes, ETA (ETA-R) and ETB receptors (ETB-R). ETA receptor signal pathways mediate Gq, Gs, and G12/13 resulting in activation of Ca2+/phospholipase C (PLC), cAMP, and G12/13/Rho/Rock signal pathways. ETB-R signal pathways mediate Gq, Gi, and G12/13 resulting in activation of Ca2+/PLC and Rho/Rock signal pathways and suppression of cAMP signaling pathways . Figure 3 Promotion of astrocytic activation through ET-1/ETB-R signaling. In TBI, ET-1 induces reactive astrocytes through the following mechanisms: (1) TBI promotes the conversion of resting astrocytes to reactive ones. (2) ET-1 is produced and released from reactive astrocytes. The physical impact of TBI directly stimulates ET-1 production, although the mechanism is unclear . (3) As well as ET-1, expression of astrocytic ETB-R is upregulated, accompanied by astrocytic activation. (4) Astrocyte-derived ET-1 stimulates ETB-R through an autocrine/paracrine mechanism. (5) ETB-R stimulation further promotes conversion to reactive astrocytes and enhances their functions. Figure 4 Proliferation signal pathways triggered by astrocytic ETB-R. ETB-R-mediated intracellular signals for astrocytic proliferation include cell adhesion-dependent and -independent pathways. In the cell adhesion-dependent pathway, stimulation of ETB-R causes cytoskeletal actin reorganization through a small G-protein, rho. The formation of focal adhesions accompanied by cytoskeletal reorganization activates FAK and increases cyclinD3 expression. In the cell adhesion-independent signal pathway, stimulation of ETB-R increases the activities of STAT3 and SP-1 through ERK and JNK. These transcription factors increase the expressions of cyclinD1 and S-phase kinase-associated protein 2 (Skp-2). The upregulation of these cell cycle regulatory proteins by ET-1 induces the proliferation of reactive astrocytes in brain disorders, including TBI. Figure 5 Astrocytic ETB-R as a therapeutic target of TBI. Astrocytic ETB-R may have different significance as a therapeutic target for patients with TBI in the acute and recovery phases. (A) In acute TBI, increased ET-1 increases the expression of MMPs and VEGF-A through the astrocytic ETB-R. It reduces angiopoietin-1 (ANG-1) and sonic hedgehog (SHH) production. Altered production of these vascular permeability regulators causes BBB disruption that induces brain edema. Stimulation of ETB-Rs also attracts inflammatory cells to the site of injury through increased production of chemokines. Administration of ETB-R antagonists in the acute phase of TBI reduces these astrocytic functions and may protect the nervous system. (B) In the recovery phase of TBI, the administration of ETB-R agonists is expected to be effective in recovering neuronal function impaired by TBI. Stimulation of astrocytic ETB-Rs promotes the production of neurotrophic factors, which promotes synaptogenesis and neurogenesis. ETB-R agonists also reduce the expression of ephrins. Then, regeneration of the damaged nervous system is promoted. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Jarrahi A. Braun M. Ahluwalia M. Gupta R.V. Wilson M. Munie S. Ahluwalia P. Vender J.R. Vale F.L. Dhandapani K.M. Revisiting traumatic brain injury: From molecular mechanisms to therapeutic interventions Biomedicines 2020 8 389 10.3390/biomedicines8100389 33003373 2. Thapa K. Khan H. Singh T.G. Kaur A. Traumatic brain injury: Mechanistic insight on pathophysiology and potential therapeutic targets J. Mol. Neurosci. 2021 9 1725 1742 10.1007/s12031-021-01841-7 3. Maas A.I. Roozenbeek B. Manley G.T. Clinical trials in traumatic brain injury: Past experience and current developments Neurotherapeutics 2010 7 115 126 10.1016/j.nurt.2009.10.022 20129503 4. Jha R.M. Kochanek P.M. Simard J.M. Pathophysiology and treatment of cerebral edema in traumatic brain injury Neuropharmacology 2019 145 230 246 10.1016/j.neuropharm.2018.08.004 30086289 5. Lerouet D. Marchand-Leroux C. Besson V.C. Neuropharmacology in traumatic brain injury: From preclinical to clinical neuroprotection? Fundam. Clin. Pharmacol. 2021 35 524 538 10.1111/fcp.12656 33527472 6. Maugeri G. D'Agata V. Trovato B. Roggio F. Castorina A. Vecchio M. Di Rosa M. Musumeci G. The role of exercise on peripheral nerve regeneration: From animal model to clinical application Heliyon 2021 7 e08281 10.1016/j.heliyon.2021.e08281 34765794 7. Theisen C.C. Sachdeva R. Austin S. Kulich D. Kranz V. Houle J.D. Exercise and peripheral nerve grafts as a strategy to promote regeneration after acute or chronic spinal cord injury J. Neurotrauma 2017 34 1909 1914 10.1089/neu.2016.4640 28437223 8. Sachdeva R. Theisen C.C. Ninan V. Twiss J.L. Houle J.D. Exercise dependent increase in axon regeneration into peripheral nerve grafts by propriospinal but not sensory neurons after spinal cord injury is associated with modulation of regeneration-associated genes Exp. Neurol. 2016 276 72 82 10.1016/j.expneurol.2015.09.004 26366525 9. Lee H.G. Wheeler M.A. Quintana F.J. Function and therapeutic value of astrocytes in neurological diseases Nat. Rev. Drug Discov. 2022 21 339 358 10.1038/s41573-022-00390-x 35173313 10. Gorshkov K. Aguisanda F. Thorne N. Zheng W. Astrocytes as targets for drug discovery Drug Discov. Today 2018 23 673 680 10.1016/j.drudis.2018.01.011 29317338 11. Zhou Y. Shao A. Yao Y. Tu S. Deng Y. Zhang J. Dual roles of astrocytes in plasticity and reconstruction after traumatic brain injury Cell Commun. Signal. 2020 18 62 10.1186/s12964-020-00549-2 32293472 12. Koyama Y. Endothelin ETB receptor-mediated astrocytic activation: Pathological roles in brain disorders Int. J. Mol. Sci. 2021 22 4333 10.3390/ijms22094333 33919338 13. Michinaga S. Kimura A. Hatanaka S. Minami S. Asano A. Ikushima Y. Matsui S. Toriyama Y. Fujii M. Koyama Y. Delayed administration of BQ788, an ETB antagonist, after experimental traumatic brain injury promotes recovery of blood-brain barrier function and a reduction of cerebral edema in mice J. Neurotrauma 2018 35 1481 1494 10.1089/neu.2017.5421 29316834 14. Castejon O.J. Morphological astrocytic changes in complicated human brain trauma. A light and electron microscopic study Brain Inj. 1998 12 409 427 10.1080/026990598122539 9591144 15. Dunn C. Sturdivant N. Venier S. Ali S. Wolchok J. Balachandran K. Blood-brain barrier breakdown and astrocyte reactivity evident in the absence of behavioral changes after repeated traumatic brain injury Neurotrauma Rep. 2021 2 399 410 10.1089/neur.2021.0017 34901939 16. Zhai Y. Ye S.Y. Wang Q.S. Xiong R.P. Fu S.Y. Du H. Xu Y.W. Peng Y. Huang Z.Z. Yang N. Overexpressed ski efficiently promotes neurorestoration, increases neuronal regeneration, and reduces astrogliosis after traumatic brain injury Gene Ther. 2022 10.1038/s41434-022-00320-x 35132206 17. Prabhakar N.K. Khan H. Grewal A.K. Singh T.G. Intervention of neuroinflammation in the traumatic brain injury trajectory: In vivo and clinical approaches Int. Immunopharmacol. 2022 108 108902 10.1016/j.intimp.2022.108902 35729835 18. Goodman J.C. Van M. Gopinath S.P. Robertson C.S. Pro-inflammatory and proapoptotic elements of the neuroinflammatory response are activated in traumatic brain injury Acta Neurochir. Suppl. 2008 102 437 439 10.1007/978-3-211-85578-2_85 19388362 19. Pedrazzi M. Patrone M. Passalacqua M. Ranzato E. Colamassaro D. Sparatore B. Pontremoli S. Melloni E. Selective proinflammatory activation of astrocytes by high-mobility group box 1 protein signaling J. Immunol. 2007 179 8525 8532 10.4049/jimmunol.179.12.8525 18056400 20. Gorina R. Font-Nieves M. Marquez-Kisinousky L. Santalucia T. Planas A.M. Astrocyte TLR4 activation induces a proinflammatory environment through the interplay between MyD88-dependent NFkB signaling, MAPK, and Jak1/Stat1 pathways Glia 2011 59 242 255 10.1002/glia.21094 21125645 21. Koyama Y. Kotani M. Sawamura T. Kuribayashi M. Konishi R. Michinaga S. Different actions of endothelin-1 on chemokine production in rat cultured astrocytes: Reduction of CX3CL1/fractalkine and an increase in CCL2/MCP-1 and CXCL1/CINC-1 J. Neuroinflamm. 2013 10 51 10.1186/1742-2094-10-51 23627909 22. Wicher G. Wallenquist U. Lei Y. Enoksson M. Li X. Fuchs B. Abu Hamdeh S. Marklund N. Hillered L. Nilsson G. Interleukin-33 promotes recruitment of microglia/macrophages in response to traumatic brain injury J. Neurotrauma 2017 34 3173 3182 10.1089/neu.2016.4900 28490277 23. Xue J. Zhang Y. Zhang J. Zhu Z. Lv Q. Su J. Astrocyte-derived CCL7 promotes microglia-mediated inflammation following traumatic brain injury Int. Immunopharmacol. 2021 99 107975 10.1016/j.intimp.2021.107975 34293712 24. Ding J.Y. Kreipke C.W. Schafer P. Schafer S. Speirs S.L. Rafols J.A. Synapse loss regulated by matrix metalloproteinases in traumatic brain injury is associated with hypoxia inducible factor-1alpha expression Brain Res. 2009 1268 125 1234 10.1016/j.brainres.2009.02.060 19285046 25. Suehiro E. Fujisawa H. Akimura T. Ishihara H. Kajiwara K. Kato S. Fujii M. Yamashita S. Maekawa T. Suzuki M. Increased matrix metalloproteinase-9 in blood in association with activation of interleukin-6 after traumatic brain injury: Influence of hypothermic therapy J. Neurotrauma 2004 21 1706 1711 10.1089/neu.2004.21.1706 15684762 26. Gao W. Zhao Z. Yu G. Zhou Z. Zhou Y. Hu T. Jiang R. Zhang J. VEGI attenuates the inflammatory injury and disruption of blood-brain barrier partly by suppressing the TLR4/NF-kB signaling pathway in experimental traumatic brain injury Brain Res. 2015 1622 230 239 10.1016/j.brainres.2015.04.035 26080076 27. Michinaga S. Onishi K. Shimizu K. Mizuguchi H. Hishinuma S. Pharmacological inhibition of transient receptor potential vanilloid 4 reduces vasogenic edema after traumatic brain injury in mice Biol. Pharm. Bull. 2021 44 1759 1766 10.1248/bpb.b21-00512 34719652 28. Hu X. Li S. Shi Z. Lin W.J. Yang Y. Li Y. Li H. Xu Y. Zhou M. Tang Y. Partial ablation of astrocytes exacerbates cerebral infiltration of monocytes and neuronal loss after brain stab injury in mice Cell Mol. Neurobio. 2022 10.1007/s10571-022-01224-5 29. Gao X. Li W. Syed F. Yuan F. Li P. Yu Q. PD-L1 signaling in reactive astrocytes counteracts neuroinflammation and ameliorates neuronal damage after traumatic brain injury J. Neuroinflamm. 2022 19 43 10.1186/s12974-022-02398-x 30. Zhang W. Hong J. Zhang H. Zheng W. Yang Y. Astrocyte-derived exosomes protect hippocampal neurons after traumatic brain injury by suppressing mitochondrial oxidative stress and apoptosis Aging (Albany NY) 2021 13 21642 21658 10.18632/aging.203508 34516406 31. Michinaga S. Tanabe A. Nakaya R. Fukutome C. Inoue A. Iwane A. Minato Y. Tujiuchi Y. Miyake D. Mizuguchi H. Angiopoietin-1/Tie-2 signal after focal traumatic brain injury is potentiated by BQ788, an ETB receptor antagonist, in the mouse cerebrum: Involvement in recovery of blood-brain barrier function J. Neurochem. 2020 154 330 348 10.1111/jnc.14957 31957020 32. Michinaga S. Inoue A. Sonoda K. Mizuguchi H. Koyama Y. Down-regulation of astrocytic sonic hedgehog by activation of endothelin ETB receptors: Involvement in traumatic brain injury-induced disruption of blood brain barrier in a mouse model Neurochem. Int. 2021 146 105042 10.1016/j.neuint.2021.105042 33838160 33. Jamjoom A.A.B. Rhodes J. Andrews P.J.D. Grant S.G.N. The synapse in traumatic brain injury Brain 2021 144 18 31 10.1093/brain/awaa321 33186462 34. Yu T.S. Tensaouti Y. Stephanz E.P. Chintamen S. Rafikian E.E. Yang M. Kernie S.G. Astrocytic ApoE underlies maturation of hippocampal neurons and cognitive recovery after traumatic brain injury in mice Commun. Biol. 2021 4 1303 10.1038/s42003-021-02841-4 34795427 35. Baecker J. Wartchow K. Sehm T. Ghoochani A. Buchfelder M. Kleindienst A. Treatment with the neurotrophic protein S100B increases synaptogenesis after traumatic brain injury J. Neurotrauma 2020 37 1097 1107 10.1089/neu.2019.6475 31597524 36. Christopherson K.S. Ullian E.M. Stokes C.C. Mullowney C.E. Hell J.W. Agah A. Lawler J. Mosher D.F. Bornstein P. Barres B.A. Thrombospondins are astrocyte-secreted proteins that promote CNS synaptogenesis Cell 2005 120 421 433 10.1016/j.cell.2004.12.020 15707899 37. Lu Z. Kipnis J. Thrombospondin 1--A key astrocyte-derived neurogenic factor FASEB J. 2010 24 1925 1934 10.1096/fj.09-150573 20124433 38. Cheng C. Yu Z. Zhao S. Liao Z. Xing C. Jiang Y. Yang Y.G. Whalen M.J. Lo E.H. Sun X. Thrombospondin-1 gene deficiency worsens the neurological outcomes of traumatic brain injury in mice Int. J. Med. Sci. 2017 14 927 936 10.7150/ijms.18812 28924363 39. Lin P.H. Kuo L.T. Luh H.T. The roles of neurotrophins in traumatic brain injury Life 2021 12 26 10.3390/life12010026 35054419 40. Goss J.R. O'Malley M.E. Zou L. Styren S.D. Kochanek P.M. DeKosky S.T. Astrocytes are the major source of nerve growth factor upregulation following traumatic brain injury in the rat Exp. Neurol. 1998 149 301 309 10.1006/exnr.1997.6712 9500953 41. Dixon C.E. Flinn P. Bao J. Venya R. Hayes R.L. Nerve growth factor attenuates cholinergic deficits following traumatic brain injury in rats Exp. Neurol. 1997 146 479 490 10.1006/exnr.1997.6557 9270059 42. Furtado A.B.V. Goncalves D.F. Hartmann D.D. Courtes A.A. Cassol G. Nunez-Figueredo Y. Argolo D.S. do Nascimento R.P. Costa S.L. da Silva V.D.A. JM-20 treatment after mild traumatic brain injury reduces glial cell pro-inflammatory signaling and behavioral and cognitive deficits by increasing neurotrophin expression Mol. Neurobiol. 2021 58 4615 4627 10.1007/s12035-021-02436-4 34148214 43. Zhao J. Qu D. Xi Z. Huan Y. Zhang K. Yu C. Yang D. Kang J. Lin W. Wu S. Mitochondria transplantation protects traumatic brain injury via promoting neuronal survival and astrocytic BDNF Transl. Res. 2021 235 102 114 10.1016/j.trsl.2021.03.017 33798765 44. Hao P. Duan H. Hao F. Chen L. Sun M. Fan K.S. Sun Y.E. Williams D. Yang Z. Li X. Neural repair by NT3-chitosan via enhancement of endogenous neurogenesis after adult focal aspiration brain injury Biomaterials 2017 140 88 102 10.1016/j.biomaterials.2017.04.014 28641124 45. Schinelli S. Pharmacology and physiopathology of the brain endothelin system: An overview Curr. Med. Chem. 2006 13 627 638 10.2174/092986706776055652 16529555 46. Dashwood M.R. Loesch A. Endothelin-1 as a neuropeptide: Neurotransmitter or neurovascular effects? J. Cell Commun. Signal. 2010 4 51 62 10.1007/s12079-009-0073-3 19847673 47. Tykocki N.R. Watts S.W. The interdependence of endothelin-1 and calcium: A review Clin. Sci. 2010 119 361 372 10.1042/CS20100145 48. Di Benedetto G. Burgaletto C. Bellanca C.M. Munafo A. Bernardini R. Cantarella G. Role of Microglia and astrocytes in Alzheimer's disease: From neuroinflammation to Ca2+ homeostasis dysregulation Cells 2022 11 728 10.3390/cells11172728 35203377 49. Enevoldsen F.C. Sahana J. Wehland M. Grimm D. Infanger M. Kruger M. Endothelin receptor antagonists: Status quo and future perspectives for targeted therapy J. Clin. Med. 2020 9 824 10.3390/jcm9030824 32197449 50. Sapira V. Cojocaru I.M. Lilios G. Grigorian M. Cojocaru M. Study of endothelin-1 in acute ischemic stroke Rom. J. Intern. Med. 2010 48 329 332 21528761 51. Palmer J.C. Barker R. Kehoe P.G. Love S. Endothelin-1 is elevated in Alzheimer's disease and upregulated by amyloid-b J. Alzheimers 2012 29 853 861 10.3233/JAD-2012-111760 22330820 52. Yan Y. Fu J. Plasma ApoA-1 and endothelin-1 levels changes in early Parkinson disease and its relationship with cognitive function and cerebral white matter structure change Pak. J. Pharm. Sci. 2021 34 487 492 34275798 53. Bellapart J. Nasrallah F. Winearls J. Lassig-Smith M. Stuart J. Boots R. Winter C. Flaws D. Bulmer A. Jones L. Diagnostic sensitivity of plasma endothelin-1 for cerebral vasospasm after subarachnoid hemorrhage, a multicentre double-blind Sstudy J. Intensive Care Med. 2022 37 769 775 10.1177/08850666211056568 34898303 54. Liu X. Deng F. Yu Z. Xie Y. Hu C. Chen L. Inhibition of endothelin A receptor protects brain microvascular endothelial cells against hypoxia-induced injury Int. J. Mol. Med. 2014 34 313 320 10.3892/ijmm.2014.1744 24736835 55. Hostenbach S. D'Haeseleer M. Kooijman R. De Keyser J. Modulation of cytokine-induced astrocytic endothelin-1 production as a possible new approach to the treatment of multiple sclerosis Front. Pharmacol. 2020 10 1491 10.3389/fphar.2019.01491 31969819 56. Michinaga S. Inoue A. Yamamoto H. Ryu R. Inoue A. Mizuguchi H. Koyama Y. Endothelin receptor antagonists alleviate blood-brain barrier disruption and cerebral edema in a mouse model of traumatic brain injury: A comparison between bosentan and ambrisentan Neuropharmacology 2020 175 108182 10.1016/j.neuropharm.2020.108182 32561219 57. Maier B. Lehnert M. Laurer H.L. Marzi I. Biphasic elevation in cerebrospinal fluid and plasma concentrations of endothelin 1 after traumatic brain injury in human patients Shock 2007 27 610 614 10.1097/shk.0b013e31802f9eaf 17505299 58. Chen L.X. Zhang W.F. Wang M. Jia P.F. Relationship of calcitonin gene-related peptide with disease progression and prognosis of patients with severe traumatic brain injury Neural Regen. Res. 2018 13 1782 1786 10.4103/1673-5374.238619 30136693 59. Salonia R. Empey P.E. Poloyac S.M. Wisniewski S.R. Klamerus M. Ozawa H. Wagner A.K. Ruppel R. Bell M.J. Feldman K. Endothelin-1 is increased in cerebrospinal fluid and associated with unfavorable outcomes in children after severe traumatic brain injury J. Neurotrauma 2010 27 1819 1825 10.1089/neu.2010.1402 20684675 60. Curvello V. Pastor P. Hekierski H. Armstead W.M. Inhaled nitric oxide protects cerebral autoregulation and reduces hippocampal necrosis after traumatic brain injury through inhibition of ET-1, ERK MAPK and IL-6 upregulation in pigs Neurocrit. Care 2019 30 467 477 10.1007/s12028-018-0638-1 30386963 61. Liu T. Liao X.Z. Zhou M.T. Ulinastatin alleviates traumatic brain injury by reducing endothelin-1 Transl. Neurosci. 2021 12 1 8 10.1515/tnsci-2021-0001 33505713 62. Jullienne A. Obenaus A. Ichkova A. Savona-Baron C. Pearce W.J. Badaut J. Chronic cerebrovascular dysfunction after traumatic brain injury J. Neurosci. Res. 2016 94 609 622 10.1002/jnr.23732 27117494 63. Clozel M. Watanabe H. BQ-123, a peptidic endothelin ETA receptor antagonist, prevents the early cerebral vasospasm following subarachnoid hemorrhage after intracisternal but not intravenous injection Life Sci. 1993 52 825 834 10.1016/0024-3205(93)90081-D 8437512 64. Yeung P.K. Shen J. Chung S.S. Chung S.K. Targeted over-expression of endothelin-1 in astrocytes leads to more severe brain damage and vasospasm after subarachnoid hemorrhage BMC Neurosci. 2013 14 131 10.1186/1471-2202-14-131 24156724 65. Ishikawa N. Takemura M. Koyama Y. Shigenaga Y. Okada T. Baba A. Endothelins promote the activation of astrocytes in rat neostriatum through ETB receptors Eur. J. Neurosci. 1997 9 895 901 10.1111/j.1460-9568.1997.tb01440.x 9182942 66. Rogers S.D. Peters C.M. Pomonis J.D. Hagiwara H. Ghilardi J.R. Mantyh P.W. Endothelin B receptors are expressed by astrocytes and regulate astrocyte hypertrophy in the normal and injured CNS Glia 2003 41 180 190 10.1002/glia.10173 12509808 67. Wilhelmsson U. Li L. Pekna M. Berthold C.H. Blom S. Eliasson C. Renner O. Bushong E. Ellisman M. Morgan T.E. of glial fibrillary acidic protein and vimentin prevents hypertrophy of astrocytic processes and improves post-traumatic regeneration J. Neurosci. 2004 24 5016 5021 10.1523/JNEUROSCI.0820-04.2004 15163694 68. Peters C.M. Rogers S.D. Pomonis J.D. Egnaczyk G.F. Keyser C.P. Schmidt J.A. Ghilardi J.R. Maggio J.E. Mantyh P.W. Endothelin receptor expression in the normal and injured spinal cord: Potential involvement in injury-induced ischemia and gliosis Exp. Neurol. 2003 180 1 13 10.1016/S0014-4886(02)00023-7 12668144 69. LeComte M.D. Shimada I.S. Sherwin C. Spees J.L. Notch1-STAT3-ETBR signaling axis controls reactive astrocyte proliferation after brain injury Proc. Natl. Acad. Sci. USA 2015 112 8726 8731 10.1073/pnas.1501029112 26124113 70. Koyama Y. Takemura M. Fujiki K. Ishikawa N. Shigenaga Y. Baba A. BQ788, an endothelin ETB receptor antagonist, attenuates stab wound injury-induced reactive astrocytes in rat brain Glia 1999 26 268 271 10.1002/(SICI)1098-1136(199905)26:3<268::AID-GLIA8>3.0.CO;2-G 10340767 71. Gadea A. Schinelli S. Gallo V. Endothelin-1 regulates astrocyte proliferation and reactive gliosis via a JNK/c-Jun signaling pathway J. Neurosci. 2008 28 2394 2408 10.1523/JNEUROSCI.5652-07.2008 18322086 72. Yamasaki R. Fujii T. Wang B. Masaki K. Kido M.A. Yoshida M. Matsushita T. Kira J.I. Allergic inflammation leads to neuropathic pain via glial cell activation J. Neurosci. 2016 36 11929 11945 10.1523/JNEUROSCI.1981-16.2016 27881779 73. Herrmann J.E. Imura T. Song B. Qi J. Ao Y. Nguyen T.K. Korsak R.A. Takeda K. Akira S. Sofroniew M.V. STAT3 is a critical regulator of astrogliosis and scar formation after spinal cord injury J. Neurosci. 2008 28 7231 7243 10.1523/JNEUROSCI.1709-08.2008 18614693 74. Tsuda M. Kohro Y. Yano T. Tsujikawa T. Kitano J. Tozaki-Saitoh H. Koyanagi S. Ohdo S. Ji R.R. Salter M.W. JAK-STAT3 pathway regulates spinal astrocyte proliferation and neuropathic pain maintenance in rats Brain 2011 134 1127 1139 10.1093/brain/awr025 21371995 75. Wanner I.B. Anderson M.A. Song B. Levine J. Fernandez A. Gray-Thompson Z. Ao Y. Sofroniew M.V. Glial scar borders are formed by newly proliferated, elongated astrocytes that interact to corral inflammatory and fibrotic cells via STAT3-dependent mechanisms after spinal cord injury J. Neurosci. 2013 33 2870 12886 10.1523/JNEUROSCI.2121-13.2013 76. Oliva A.A. Jr. Kang Y. Sanchez-Molano J. Furones C. Atkins C.M. STAT3 signaling after traumatic brain injury J. Neurochem. 2012 120 710 720 10.1111/j.1471-4159.2011.07610.x 22145815 77. Koyama Y. Sumie S. Nakano Y. Nagao T. Tokumaru S. Michinaga S. Endothelin-1 stimulates expression of cyclin D1 and S-phase kinase-associated protein 2 by activating the transcription factor STAT3 in cultured rat astrocytes J. Biol. Chem. 2019 294 3920 3933 10.1074/jbc.RA118.005614 30670587 78. Kitazawa S. Kitazawa R. Maeda S. Transcriptional regulation of rat cyclin D1 gene by CpG methylation status in promoter region J. Biol. Chem. 1999 274 28787 28793 10.1074/jbc.274.40.28787 10497251 79. Di Giovanni S. Movsesyan V. Ahmed F. Cernak I. Schinelli S. Stoica B. Faden A.I. Cell cycle inhibition provides neuroprotection and reduces glial proliferation and scar formation after traumatic brain injury Proc. Natl. Acad. Sci. USA 2005 102 8333 8338 10.1073/pnas.0500989102 15923260 80. Michinaga S. Ishida A. Takeuchi R. Koyama Y. Endothelin-1 stimulates cyclin D1 expression in rat cultured astrocytes via activation of Sp1 Neurochem. Int. 2013 63 25 34 10.1016/j.neuint.2013.04.004 23619396 81. Cazaubon S. Chaverot N. Romero I.A. Girault J.A. Adamson P. Strosberg A.D. Couraud P.O. Growth factor activity of endothelin-1 in primary astrocytes mediated by adhesion-dependent and -independent pathways J. Neurosci. 1997 17 6203 6212 10.1523/JNEUROSCI.17-16-06203.1997 9236231 82. Koyama Y. Yoshioka Y. Shinde M. Matsuda T. Baba A. Focal adhesion kinase mediates endothelin-induced cyclin D3 expression in rat cultured astrocytes J. Neurochem. 2004 90 904 912 10.1111/j.1471-4159.2004.02546.x 15287896 83. Koyama Y. Tanaka K. Endothelins stimulate the production of stromelysin-1 in cultured rat astrocytes Biochem. Biophys. Res. Commun. 2008 371 659 663 10.1016/j.bbrc.2008.04.064 18439420 84. Koyama Y. Tanaka K. Intracerebroventricular administration of an endothelin ET(B)-receptor agonist increases expression of matrix metalloproteinase-2 and -9 in rat brain J. Pharmacol. Sci. 2010 114 433 443 10.1254/jphs.10195FP 21127388 85. Koyama Y. Maebara Y. Hayashi M. Nagae R. Tokuyama S. Michinaga S. Endothelins reciprocally regulate VEGF-A and angiopoietin-1 production in cultured rat astrocytes: Implications on astrocytic proliferation Glia 2012 60 1954 1963 10.1002/glia.22411 22927341 86. Koyama Y. Hayashi M. Nagae R. Tokuyama S. Konishi T. Endothelin-1 increases the expression of VEGF-R1/Flt-1 receptors in rat cultured astrocytes through ETB receptors J. Neurochem. 2014 130 759 769 10.1111/jnc.12770 24862165 87. Kim J.Y. Ko A.R. Hyun H.W. Kang T.C. ETB receptor-mediated MMP-9 activation induces vasogenic edema via ZO-1 protein degradation following status epilepticus Neuroscience 2015 304 355 367 10.1016/j.neuroscience.2015.07.065 26232046 88. Gourmala N.G. Buttini M. Limonta S. Sauter A. Boddeke H.W. Differential and time-dependent expression of monocyte chemoattractant protein-1 mRNA by astrocytes and macrophages in rat brain: Effects of ischemia and peripheral lipopolysaccharide administration J. Neuroimmunol. 1997 74 35 44 10.1016/S0165-5728(96)00203-2 9119977 89. Ishizuka K. Kimura T. Igata-yi R. Katsuragi S. Takamatsu J. Miyakawa T. Identification of monocyte chemoattractant protein-1 in senile plaques and reactive microglia of Alzheimer's disease Psychiatry Clin. Neurosci. 1997 51 135 138 10.1111/j.1440-1819.1997.tb02375.x 9225377 90. Koyama Y. Baba A. Matsuda T. Production of monocyte chemoattractant protein-1 and cytokine-induced neutrophil chemoattractant-1 in rat brain is stimulated by intracerebroventricular administration of an endothelin ETB receptor agonist Neuroreport 2007 18 1275 1279 10.1097/WNR.0b013e32825a67f1 17632282 91. Koyama Y. Tsujikawa K. Matsuda T. Baba A. Endothelin-1 stimulates glial cell line-derived neurotrophic factor expression in cultured rat astrocytes Biochem. Biophys. Res. Commun. 2003 303 1101 1105 10.1016/S0006-291X(03)00491-1 12684049 92. Koyama Y. Tsujikawa K. Matsuda T. Baba A. Intracerebroventricular administration of an endothelin ETB receptor agonist increases expressions of GDNF and BDNF in rat brain Eur. J. Neurosci. 2003 18 887 894 10.1046/j.1460-9568.2003.02797.x 12925014 93. Koyama Y. Tsujikawa K. Matsuda T. Baba A. Endothelin increases expression of exon exon IV-containing brain-derived neurotrophic factor transcripts in cultured astrocytes and rat brain J. Neurosci. Res. 2005 80 809 816 10.1002/jnr.20512 15898104 94. Goldshmit Y. McLenachan S. Turnley A. Roles of Eph receptors and ephrins in the normal and damaged adult CNS Brain Res. Rev. 2006 52 327 345 10.1016/j.brainresrev.2006.04.006 16774788 95. Jiao J.W. Feldheim D.A. Chen D.F. Ephrins as negative regulators of adult neurogenesis in diverse regions of the central nervous system Proc. Natl. Acad. Sci. USA 2008 105 8778 8783 10.1073/pnas.0708861105 18562299 96. Wu L.Y. Yu X.L. Feng L.Y. Connexin 43 stabilizes astrocytes in a stroke-like milieu to facilitate neuronal recovery Acta Pharmacol. Sin. 2015 36 928 938 10.1038/aps.2015.39 26095039 97. Li Y. Chen Y. Tan L. Pan J.Y. Lin W.W. Wu J. Hu W. Chen X. Wang X.D. RNAimediatedephrin-B2 silencing attenuates astroglial-fibrotic scar formation and improves spinal cord axon growth CNS Neurosci. Ther. 2017 23 779 789 10.1111/cns.12723 28834283 98. Ren Z. Chen X. Yang J. Kress B.T. Tong J. Liu H. Takano T. Zhao Y. Nedergaard M. Improved axonal regeneration after spinal cord injury in mice with conditional deletion of ephrin B2 under the GFAP promoter Neuroscience 2013 241 89 99 10.1016/j.neuroscience.2013.03.010 23518227 99. Dixon K.J. Mier J. Gajavelli S. Turbic A. Bullock R. Turnley A.M. Liebl D.J. EphrinB3 restricts endogenous neural stem cell migration after traumatic brain injury Stem Cell Res. 2016 17 504 513 10.1016/j.scr.2016.09.029 27771498 100. Duffy P. Wang X. Siegel C.S. Tu N. Henkemeyer M. Cafferty W.B. Strittmatter S.M. Myelin-derived ephrinB3 restricts axonal regeneration and recovery after adult CNS injury Proc. Natl. Acad. Sci. USA 2012 109 5063 5068 10.1073/pnas.1113953109 22411787 101. Koyama Y. Tsuboi S. Mizogui F. Endothelin-1 decreases the expression of Ephrin-A and B subtypes in cultured rat astrocytes through ETB receptors Neurosci. Lett. 2021 741 135393 10.1016/j.neulet.2020.135393 33279571 102. Guo J. Li Y. He Z. Zhang B. Li Y. Hu J. Han M. Xu Y. Li Y. Gu J. Targeting endothelin receptors A and B attenuates the inflammatory response and improves locomotor function following spinal cord injury in mice Int. J. Mol. Med. 2014 34 74 82 10.3892/ijmm.2014.1751 24756152 103. Mira R.G. Lira M. Cerpa W. Traumatic brain injury: Mechanisms of glial response Front. Physiol. 2021 12 740939 10.3389/fphys.2021.740939 34744783 104. Kawabori M. Weintraub A.H. Imai H. Zinkevych L. McAllister P. Steinberg G.K. Frishberg B.M. Yasuhara T. Chen J.W. Cramer S.C. Cell therapy for chronic TBI: Interim analysis of the randomized controlled STEMTRA Trial Neurology 2021 96 e1202 e1214 10.1212/WNL.0000000000011450 33397772 105. Chiareli R.A. Carvalho G.A. Marques B.L. Mota L.S. Oliveira-Lima O.C. Gomes R.M. Birbrair A. Gomez R.S. Simao F. Klempin F. The Role of astrocytes in the neurorepair process Front. Cell Dev. Biol. 2021 9 665795 10.3389/fcell.2021.665795 34113618 106. Tan C.X. Eroglu C. Cell adhesion molecules regulating astrocyte-neuron interactions Curr. Opin. Neurobiol. 2021 69 170 177 10.1016/j.conb.2021.03.015 33957433 107. Michinaga S. Koyama Y. Pathophysiological responses and roles of astrocytes in traumatic brain injury Int. J. Mol. Sci. 2021 22 6418 10.3390/ijms22126418 34203960 108. Wu J. Li H. He J. Tian X. Luo S. Li J. Li W. Zhong J. Zhang H. Huang Z. Downregulation of microRNA-9-5p promotes synaptic remodeling in the chronic phase after traumatic brain injury Cell Death Dis. 2021 12 9 10.1038/s41419-020-03329-5 33414448 109. Nikolakopoulou A.M. Koeppen J. Garcia M. Leish J. Obenaus A. Ethell I.M. Astrocytic ephrin-B1 regulates synapse remodeling following traumatic brain injury ASN Neuro 2016 8 1 18 10.1177/1759091416630220 26928051 110. Barlow A. De Graaff E. Pachnis V. Enteric nervous system progenitors are coordinately controlled by the G protein-coupled receptor EDNRB and the receptor tyrosine kinase RET Neuron 2003 40 905 916 10.1016/S0896-6273(03)00730-X 14659090 111. Pawolski V. Schmidt M.H.H. Neuron-glia interaction in the developing and adult enteric nervous system Cells 2020 10 47 10.3390/cells10010047 33396231 112. Nishikawa K. Ayukawa K. Hara Y. Wada K. Aoki S. Endothelin/endothelin-B receptor signals regulate ventricle-directed interkinetic nuclear migration of cerebral cortical neural progenitors Neurochem. Int. 2011 58 261 272 10.1016/j.neuint.2010.11.013 21130129 113. Leonard M.G. Briyal S. Gulati A. Endothelin B receptor agonist, IRL-1620, reduces neurological damage following permanent middle cerebral artery occlusion in rats Brain Res. 2011 1420 48 58 10.1016/j.brainres.2011.08.075 21959172 114. Leonard M.G. Gulati A. Endothelin B receptor agonist, IRL-1620, enhances angiogenesis and neurogenesis following cerebral ischemia in rats Brain Res. 2013 1528 28 41 10.1016/j.brainres.2013.07.002 23850649 115. Briyal S. Shepard C. Gulati A. Endothelin Receptor Type B Agonist, IRL-1620, Prevents Beta Amyloid (Ab) induced oxidative stress and cognitive impairment in normal and diabetic rats Pharmacol. Biochem. Behav. 2014 120 65 72 10.1016/j.pbb.2014.02.008 24561065 116. Gulati A. Agrawal N. Vibha D. Misra U.K. Paul B. Jain D. Pandian J. Borgohain R. Safety and efficacy of sovateltide (IRL-1620) in a multicenter randomized controlled clinical trial in patients with acute cerebral ischemic stroke CNS Drugs 2021 35 5 104 10.1007/s40263-020-00783-9 33428177 |
PMC10000580 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050676 cells-12-00676 Article HLA-B*57:01/Carbamazepine-10,11-Epoxide Association Triggers Upregulation of the NFkB and JAK/STAT Pathways Haukamp Funmilola Josephine Conceptualization Methodology Software Validation Formal analysis Investigation Writing - original draft 1* Hartmann Zoe Maria Methodology Validation 1 Pich Andreas Software Formal analysis 23 Kuhn Joachim Formal analysis Data curation 4 Blasczyk Rainer Resources 1 Stieglitz Florian Software Formal analysis Investigation Writing - original draft Visualization 1+ Bade-Doding Christina Conceptualization Formal analysis Writing - original draft Supervision 1+ Munz Christian Academic Editor 1 Institute of Transfusion Medicine and Transplant Engineering, Hannover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany 2 Institute of Toxicology, Hannover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany 3 Core Facility Proteomics, Hannover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany 4 Institute for Laboratory and Transfusion Medicine, Heart and Diabetes Center North Rhine-Westphalia, Ruhr University Bochum, Georgstrasse 11, 32545 Bad Oeynhausen, Germany * Correspondence: [email protected]; Tel.: +49-511-532-9774; Fax: +49-511-532-2079 + These authors contributed equally to this work. 21 2 2023 3 2023 12 5 67621 12 2022 17 2 2023 18 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Measure of drug-mediated immune reactions that are dependent on the patient's genotype determine individual medication protocols. Despite extensive clinical trials prior to the license of a specific drug, certain patient-specific immune reactions cannot be reliably predicted. The need for acknowledgement of the actual proteomic state for selected individuals under drug administration becomes obvious. The well-established association between certain HLA molecules and drugs or their metabolites has been analyzed in recent years, yet the polymorphic nature of HLA makes a broad prediction unfeasible. Dependent on the patient's genotype, carbamazepine (CBZ) hypersensitivities can cause diverse disease symptoms as maculopapular exanthema, drug reaction with eosinophilia and systemic symptoms or the more severe diseases Stevens-Johnson-Syndrome or toxic epidermal necrolysis. Not only the association between HLA-B*15:02 or HLA-A*31:01 but also between HLA-B*57:01 and CBZ administration could be demonstrated. This study aimed to illuminate the mechanism of HLA-B*57:01-mediated CBZ hypersensitivity by full proteome analysis. The main CBZ metabolite EPX introduced drastic proteomic alterations as the induction of inflammatory processes through the upstream kinase ERBB2 and the upregulation of NFkB and JAK/STAT pathway implying a pro-apoptotic, pro-necrotic shift in the cellular response. Anti-inflammatory pathways and associated effector proteins were downregulated. This disequilibrium of anti-inflammatory processes clearly explain fatal immune reactions following CBZ administration. adverse drug reaction HLA carbamazepine hypersensitivity carbamazepine carbamazepine-10,11-epoxide proteome HLA-B*57:01 This research received no external funding. pmc1. Introduction The approval of a medical product requires extensive and distinct clinical trials. Yet, the preselected group of volunteers who attend those clinical trials is limited. Every single person has a unique genetic profile affecting the functionality of any cell type of the immune system. It becomes obvious that drug-hypersensitivity reactions in most cases disorganize the adaptive immune system, resulting in severe cellular autoimmune reactions. In the past, these reactions resulted in the mandatory determination of distinct genetic profiles and at worst in the exclusion of patients from the desired medication. It is clear that these scenarios of hypersensitivity reactions following drug treatment represent an unpredictable challenge for the health care system. Adverse events occur when harmful symptoms arise after administration of a certain drug. If the harm is caused by application of the respective drug, the immunological reaction is termed an adverse drug event; if the drug was applied correctly at normal dosage, the reaction is termed an adverse drug reaction (ADR) . ADRs usually occur in a dose-dependent and predictable manner and can be explained by the pharmacological toxicity of the drug . Nevertheless, in 20% of all ADRs, their occurrence seems idiosyncratic; those reactions are termed type B ADRs . Yet, type B ADRs are often related to the immune system . Since 2002, more and more type B ADRs have been described to be associated with the highly polymorphic human leukocyte antigen (HLA) molecules . HLA molecules are cell surface proteins with a central function in immune surveillance. They present peptides to immune receptors of T and NK cells and, based on the origin of the presented peptide (i.e., self-peptide or pathogen-derived peptides), effector cell responses are prevented or induced . The presentation of a diversity of peptides derived from different origins is unique in the ligand/receptor biology, since every peptide bound to an HLA molecule results in structural and biophysical alteration of the peptide-HLA complex. Therefore, it becomes clear that every subtle variation in the HLA molecule might facilitate binding and presentation of peptides that have not undergone selection by the thymus; the biological consequences are autoimmune-like reactions . The anticonvulsant carbamazepine (CBZ) is widely used to treat various neurological diseases such as epilepsy, bipolar disorders or schizophrenia. However, CBZ administration can cause cutaneous type B ADRs in certain patients. These reactions have been described to be associated with the human leukocyte antigen (HLA) class I genotypes HLA-B*15:02 and HLA-A*31:01 . Depending on the patient's genotype, CBZ-induced ADRs are characterized by differential disease phenotypes. The symptoms range from mild skin rash such as maculopapular exanthema (MPE) and drug reaction with eosinophilia and systemic symptoms (DRESS) to more severe and potentially fatal Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) . It has been shown that the more severe SJS and TEN occur mainly in HLA-B*15:02+ patients, whereas MPE and DRESS following CBZ-treatment more likely arise in HLA-A*31:01+ patients . Positive and negative predictive values indicate that the clinical picture of HLA-associated ADRs cannot be explained exclusively by the presence of a certain HLA allele , hence other factors have to be taken into account . We could demonstrate that CBZ treatment in soluble HLA-A*31:01-expressing cells and EPX treatment in soluble HLA-B*15:02-expressing cells result in different alterations in the cellular proteome that might contribute to the explanation of distinct clinical pictures of the diseases . Recently, a further association of CBZ-induced ADRs has been described. The allele HLA-B*57:01 is unambiguously associated with SJS/TEN following treatment with CBZ in Europeans : The analysis included 28 European patients with CBZ-induced SJS, SJS/TEN-overlap or TEN, 11 of them were carrying HLA-B*57:01 (39.29%), whereas the frequency of this allele was 6.69% in European general population controls. The onset of SJS/TEN following drug application should be close-meshed monitored, an algorithm of drug causality (ALDEN) has been adjusted to provide safe diagnosis . The allele HLA-B*57:01 is originally known to be strongly associated with hypersensitivity to the antiretroviral drug abacavir (ABC) . ABC-induced ADRs for HLA-B*57:01+patients vary from fever, fatigue, gastrointestinal symptoms to severe multiorgan failure. For this disease pattern, autologous cytotoxic T cells that attack in a manner like an autoimmune reaction, the body itself could be verified to be responsible . Illing et al. could impressively show that ABC occupies the peptide binding region (PBR) of HLB-B*57:01 resulting in a conformational change of the HLA molecule and therefore, in CD8+-mediated foreign recognition of the self-HLA-B*57:01 molecules bound to a foreign peptide. Since then, this finding provides the gold standard for understanding HLA-mediated ADRs . Patients with susceptible HLA variants have not been permitted to take certain drugs. However, more and more clinical studies recently have demonstrated that drug-tolerant patients exist ; namely, patients with a certain HLA type who still could receive the questionable drug even though no immunological reactions occurred. This seems difficult to believe since drug binding and, subsequently, loading of a different peptide repertoire into the peptide binding region of the respective HLA molecule should still occur. However, in some cases a slight alteration in the amino acid sequence of bound peptides is not sufficient to trigger T cell responses. This would lead to maintained T cell tolerance in certain patients . These drug-tolerant patients could receive the respective drug regardless of their HLA type. Appreciation of this phenomenon can certainly take place by a real-time view on the proteomic content of cell with the susceptible HLA type and the respective drug. Modern proteomics provide deep insight into the health status, biological and functional opportunities of a cell, and would therefore provide a stage to monitor pharmacovigilance. The observation of an association between HLA-B*57:01 and CBZ-mediated ADRs is in this respect remarkable, since it further emphasizes that CBZ hypersensitivity seems to be associated with several HLA alleles that differ structurally. CBZ hypersensitivity was formerly an immunological reaction that targeted patients with HLA-A*31:01 or B*15:02 following drug administration. We were recently able to demonstrate why CBZ hypersensitivity features completely different clinical pictures depending on the HLA type. Although HLA-A*31:01 would bind CBZ, B*15:02 would preferably bind EPX . Both small-molecule (drug)/protein (HLA) complexes would alter the HLA-specific peptidome by the occupation of the PBR, yet the T cell response manifested by the HLA-specific clinical picture would differ significantly. Clarification for the relation between HLA molecule and drug could in this case be delivered by complete proteome analysis . The aim of this work is to give a first insight into the complex molecular basics of HLA-B*57:01-associated CBZ-mediated ADRs. This knowledge will contribute to a comprehensive understanding of the mechanisms of CBZ hypersensitivities that seem to represent disparate diseases. To achieve sufficiency in genetically based CBZ immune effects, we performed full proteome analysis of HLA-B*57:01 expressing cells in response to CBZ or EPX treatment. Understanding the pharmacological and biological basis of distinct genetic profiles and drug interplay will guide towards personalized and safe medication. 2. Materials and Methods 2.1. Detection of CBZ and EPX Bound to sHLA-B*57:01 Molecules The human B-lymphoblastoid cell line LCL721.221 (LGG promochem, Wesel, Germany) has been transduced with a lentiviral construct encoding for HLA-B*57:01 Exon 1-4, as previously described . LCL721.221 cells expressing sHLA-B*57:01 molecules were cultured in RPMI 1640 (Lonza, Basel, Switzerland) supplemented with 10% fetal calf serum (FCS, Lonza), 2 mM L-glutamine (c. c. pro, Oberdorla, Germany), 100 U/mL penicillin, and 100 mg/mL streptomycin (c. c. pro) at 37 degC and 5% CO2 in the presence of 25 mg/mL CBZ or EPX and cell culture supernatants were collected twice a week. Affinity purification of sHLA-B*57:01 molecules post drug treatment was performed and protein concentration was calculated by Bicinchoninic Acid Assay (BCA) Protein Quantitation Kit (Interchim, San Diego, CA, USA). 150 ng purified drug-treated sHLA-B*57:01 molecules were applied to mass spectrometric drug quantification in solution as previously described . 2.2. Detection of EPX-Induced Modifications of the LCL721.221/HLA-B*57:01 Proteome Proteome analysis was performed with 1 x 106 untreated, EPX-treated LCL721.221 and LCL721.221/sHLA-B*57:01 cells. Parental and sHLA-B*57:01-expressing LCL721.221 cells are not able to metabolize CBZ to EPX; this enables the analysis of CBZ and EPX treatment orthogonally. Cells were cultured in addition of 25 mg/mL CBZ or EPX for 48 h. After 24 h, drug addition was repeated. Cell harvest in RIPA lysis was performed as previously described and calculation of protein concentration was performed by Bicinchoninic Acid Assay (BCA) Protein Quantitation Kit (Interchim, San Diego, CA, USA). Sample preparation, protein digestion and MS analysis was performed as previously described . 2.3. Data Analysis The MaxQuant software (version 1.6.3.3; ) was used to search the obtained spectra against the Swiss-Prot reviewed UniProtKB database (version 01/2021, 20,395 entries; ). Propionamid of cysteine was set as fixed modification and oxidation of methionine, N-terminal acetylation, deamidation of glutamine, and asparagine were set as variable modifications The data were processed with the Perseus software (version 1.6.2.3; ). In brief, proteins that resemble a possible contamination, only identified by sight or were reversed were excluded from further analysis as well as proteins that were not measured in all replicates. To exclude potential effects on protein abundance caused by transduction with sHLA-B*57:01, the proteome of untreated LCL721.221/sHLA-B*57:01 and parental LCL721.221 cells were subtracted from the corresponding EPX-treated cells. Visualization was performed with R . In particular, the R packages complex heat map and ggplot2 were used. The heat map was generated by including the proteins that were positively tested in a Benjamini Hochberg FDR-based ANOVA. The Ingenuity Pathway Analysis software was used to perform an upstream analysis of significantly regulated proteins (IPA, QIAGEN Inc., (accessed on 24 November 2022)). Gene ontology analysis was performed with the GSEA software . The mass spectrometry proteomics data were deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD037502. 3. Results 3.1. CBZ and EPX Bind to sHLA-B*57:01 To verify binding of CBZ or EPX to sHLA-B*57:01 molecules, sHLA-B*57:01 expressing cells were cultured in the presence of 25 mg/mL CBZ or EPX, and sHLA-B*57:01 containing cell culture supernatant was collected twice a week. Functional sHLA-B*57:01 molecules were affinity purified by an NHS column coupled to the mAb W6/32 and protein concentration was determined as previously described . 150 ng EPX-treated sHLA-B*57:01 molecules were applied to UPLC-MS/MS analysis for detection of CBZ or EPX in solution . CBZ as well as EPX could be verified to bind to sHLA-B*57:01 molecules. In the solution with 150 ng CBZ/sHLA-B*57:01 molecules 0.033 ng/mL CBZ could be detected and in the EPX-containing sHLA-B*57:01 solution 0.020 ng/mL EPX could be detected . 3.2. Quantitative Proteomic Analysis after CBZ and EPX Treatment The cellular proteomes of parental LCL721.221 cells and LCL721.221/sHLA-B*57:01 cells were analyzed in an LFQ-based approach. For comparison of CBZ or EPX treatment of HLA-B*57:01 expression and parental LCL721.221 cells, the proteomic content of untreated LCL721.221 and LCL721.221/sHLA-B*57:01 cells was subtracted from the drug-treated proteome abundances. In total, 4519 proteins could be identified. To exclude proteins that were induced through transduction with the HLA-B*57:01 allele, only proteins were included in the analysis that were measured in all replicates without imputation. After filtering, 2713 proteins were feasible for further research. By subtracting the untreated LCL721.221/sHLA-B*57:01 and parental LCL721.221 cells from the corresponding EPX-treated cells, possible effects on the proteome introduced by the transduction were excluded. The data were analyzed for their examinability through dimensionality reduction with t-SNE, and clustering of the different treatments confirmed that the data were feasible for further analysis. Distinct clustering also occurred in the heat map analysis . 3.3. EPX Treatment Induced a Strong Reaction in the Proteome of LCL721.221/sHLA-B*57:01 Cells CBZ treatment induced a significant change of abundance (p < 0.05) of 335 proteins with only 35 changes more than 2-fold in LCL721.221/sHLA-B*57:01 cells when compared to parental cells . However, EPX treatment led to 776 significantly changed proteins and 134 proteins with a difference greater than 2-fold . Furthermore, we found ten proteins showing an overlapped regulation between EPX-treated regulation with one protein being co-upregulated and nine proteins being co-downregulated . An upstream analysis with the IPA software was performed to find central regulators responsible for the change in abundance. For CBZ treatment, the serine/threonine kinase IKBKE was suggested as the only upstream kinase that is activated (p-value 0.0457; Z-Score 2.0). At the same time, treatment with EPX led to the regulation of 14 kinases, with the receptor tyrosine kinase ERBB2 as the most activated, and the insulin receptor INSR predicted to be the most inactivated kinase. According to IPA upstream analysis, ERBB2 is responsible for the upregulation of IKBKB, MCM5, POLD2 and MCM7. In contrast, INSR leads to the downregulation of SLC39A7, MTCH2, ECI1, TOMM40 and LSS. Other activated indicated upstream regulators were part of the MAPK protein family or involved in the MAPK signaling cascade. In comparison, downregulated upstream regulators were predicted to be G Protein alpha, Rb, PRKAA, ERN1 and CDKN1A . Further analysis of function classes of significantly regulated proteins via the IPA software showed that EPX treatment induced expression of proteins involved in "DNA Replication, Recombination, and Repair". Furthermore, cell cycle-related proteins were found to be upregulated . Downregulated proteins were involved in "organismal death" and "glycogenesis". CBZ treatment led to the downregulation of proteins involved in "organismal death", "necroptosis", and "cell death of epithelial cells" whereas proteins involved in "cell proliferation of Tumor cell lines" were upregulated. Global GSEA enrichment analysis showed enrichment (Enrichment score: 0.58) in protein expression involved in an inflammatory pathway ("HP_CHRONIC_OTITIS_MEDIA") in LCL721.221/sHLA-B*57:01 cells that were treated with EPX compared to parental LCL721.221 cells treated with EPX. ELF4H, NCE1, STAT3, DNAAF5, RAZIB, NFKB1 were upregulated following EPX treatment in LCL721.221/sHLA-B*57:01 cells and were downregulated in parental cells after EPX treatment . EPX treatment induced the regulation of 14 pathways in the 25 most significant pathways predicted by the IPA software whereas CBZ treatment induced the regulation of 10 pathways. The most activated pathway after EPX treatment was predicted to be the "Necroptosis Signaling Pathway", and "2-ketoglutarate Dehydrogenase Complex" was predicted to be most downregulated. CBZ treatment induced the most robust activation of the "G2/M DNA Damage Checkpoint Regulation" and inhibited "ELF2 Signaling" . 4. Discussion Recent studies have demonstrated that besides HLA-A*31:01 and HLA-B*15:02, HLA-B*57:01 is also strongly linked to CBZ-induced ADRs. Although CBZ administration in HLA-A*31:01+ patients causes diseases such as MPE and DRESS, CBZ administration in HLA-B*57:01+ Europeans resulted in SJS/TEN disease phenotypes as observed for HLA-B*15:02+ patients . SJS and TEN manifest severe life-threatening cutaneous and mucosal necrosis and have to be treated by specialized burns units . When SJS/TEN emerge as HLA-mediated ADRs that involve T cell recognition of foreign peptide/HLA-complexes, the withdrawal of the drug should assure recovery of the clinical state. SJS/TEN is such an intense impairment of the affected skin that recovery is rarely possible. Therefore, prevention of such an adverse condition is mandatory in pharmacovigilance management strategies. Although the prophylaxis of HLA-mediated ADRs is not feasible and individual patient cases are often underreported due to the polymorphic character of HLA molecules, the need for conscientious analysis of HLA-mediated ADRs immediately following their establishment should be obvious. HLA molecules exhibit unique properties in the immune system. Host HLA molecules bind foreign antigens. This exceptional co-recognition requires exquisite specificity and genetical restriction for the host T cells . HLA diversity and corresponding T cell diversity restrict a comprehensive screening of patient cohorts in phase I, II and III studies where pharmacokinetics and pharmacodynamics prior to admission of a drug are tested. In the era of fast and sophisticated methods to view into the cellular content, proteomics are the instrument for understanding and long-term prevention of HLA-mediated ADRs. HLA-restricted peptidomics and T cell analysis deliver indisputable answers to understand immune compatibility, but in some cases the host T cells fail to recognize the presented peptide/drug/HLA ligand of host origin. Understanding indistinct intracellular activities as metabolism, cytokine expression, and downregulation of certain proteins in drug-tolerant patients would certainly be beneficial for drug-sensitive patients with a susceptible HLA type. Utilizing proteomics as a mirror into cellular events should support this objective. In the present study, we aimed to illuminate the underlying mechanism of HLA-B*57:01-mediated hypersensitivity to CBZ by full proteome analysis of EPX-treated LCL721.221/sHLA-B*57:01 cells. We chose the lymphoblastoid LCL721.221 cells, because these cells are not able to metabolize CBZ to EPX. The metabolization of CBZ to EPX occurs exclusively in hepatocytes and is catalyzed by cytochrome P450 enzymes . Thus, the impact of CBZ and EPX treatment on the cellular proteome of LCL721.221 cells can be analyzed orthogonally. Prior to proteome analysis, the specificity of drug-HLA interaction was determined via UPLC-MS/MS analysis. The selection of CBZ or the metabolite EPX by the respective HLA molecule is decisive for the fate of the HLA-expressing cell as previously described . We could previously demonstrate that CBZ binds exclusively to HLA-A*31:01, leading to severe skin lesions and that the exclusive interaction between EPX-HLA-B*15:02 and not CBZ-HLA-B*15:02 led to life-threatening SJS/TEN diseases. The present study showed that HLA-mediated ADRs have to be meticulously analyzed to comprehensively understand their clinical outcome. In this paper, we can show that both drug conditions CBZ and EPX are able to engage with HLA-B*57:01. The main question occurs if both or one drug condition would, in cooperation with HLA-B*57:01, impact the cellular content of the antigen-presenting cells and possibly their microenvironment. Therefore, LCL721.221 cells have been transduced with sHLA-B*57:01 and exposed to the respective drug. LCL721.221 cells are not able to metabolize CBZ to EPX and are thus a meaningful instrument to answer the question. LCL721.221/sHLA-B*57:01 cells were treated with 25 mg/mL CBZ or EPX, respectively, and cell lysates were applied to full proteome analysis. By subtracting the proteomic changes that were introduced through transduction of the cells with the HLA-B*57:01 allele, we were able to observe the independent effects that occurred due to the interplay of CBZ or EPX with the HLA-B*57:01 molecule. We found that EPX treatment induced significant changes in the proteome of LCL721.221/sHLA-B*57:01 cells. In contrast, CBZ treatment resulted in minimal changes in the proteome of LCL721.221/sHLA-B*57:01 cells. CBZ treatment of LCL721.221/sHLA-B*57:01 cells led to only 35 significantly regulated proteins whereas EPX treatment of the cells resulted in 134 significantly regulated proteins. Only a slight overlap of ten significantly regulated proteins could be detected in both EPX-treated cells . Upstream regulator analysis via IPA revealed just one activated upstream regulator, the serine/threonine kinase IKBKE, responsible for the change in protein abundance of CBZ-treated cells. In contrast, 14 kinases were detected as regulated in EPX-treated cells . Although UPLC-MS/MS analysis revealed equal binding of CBZ and EPX to sHLA-B*57:01 molecules , the CBZ-induced changes of the cellular proteome of LCL721.221/sHLA-B*57:01 seem to be marginal when compared to EPX-induced changes. Following EPX treatment, the receptor tyrosine kinase ERBB2 could be described to be the most activated upstream regulator . ERBB2 is mainly involved in inflammatory and growth-associated processes . Consequently, proteins that were predicted to be influenced and significantly two-fold upregulated were IKBKB, MCM5, MCM7, and POLD2. IKBKB is described to activate NFkB that is involved in inflammatory, pro-apoptotic and necrotic processes . Additionally, NFkB has been found to be significantly enriched in the GSEA enrichment analysis in an overall inflammatory process . MCM5 and MCM7 are involved in DNA replication and are responsive to cytokine-induced gene transcription. MCM5 has been shown to be central for STAT1-mediated gene transcription . In line with this, JAK1 has also been predicted to be activated . The JAK/STAT pathway plays a central role in reaction to external inflammatory stimuli . Consistent with this finding, STAT5 upregulation has also been described for HLA-B*15:02 after EPX treatment . The comparison of proteomic profiling of cells with intracellular small molecule/protein engagement features clearly that EPX/HLA engagement triggers the upregulation of inflammatory pathways. The sudden upregulation of proteins that are described to be part of signal transduction pathways and thus triggers of autoimmune reactions through effector cell activation could not be predicted by conventional methods. We further describe the upregulation of POLD2, an enzyme that is involved in DNA repair processes and preserving DNA integrity . POLD2 could recently be uncovered as a tumor suppressor and prognostic biomarker in distinct cancers . In coherence with POLD2 upregulation was the finding that more than 50 proteins involved in DNA repair, replication and recombination were regulated in response to EPX treatment and DNA regulatory processes were predicted to be activated . Moreover, FLT1 and EGFR were also indicated as activated. Both are known for their potential to induce apoptosis through either NFkB (FLT1) or STAT3 (EGFR) activation . FLT1 could be described as a therapeutic target in inflammatory events whereas EGFR is known as a key regulator in cell division and cancer development . In conclusion, the engagement of EPX/HLA-B*57:01 induces inflammatory, pro-apoptotic and necrotic processes in LCL721.221 cells when compared to parental LCL721.221 cells. These findings seem to be consistent with and might be a coherent explanation of the disease phenotype of SJS/TEN in HLA-B*57:01+ patients that is associated with keratinocyte death, cutaneous blistering and epidermal detachment . In coherence with the upregulation of proinflammatory proteins, INSR could be observed to be inactivated after EPX treatment . INSR has been described as inhibiting inflammatory and cytokine-mediated processes when overexpressed . These data illustrate the dignified intracellular cooperation between signal transduction proteins and the unpredictable interference of drug/HLA complexes. In addition, CDKN1A is predicted to be inhibited . Although CDKN1A is an inhibitor for cell proliferation in B cells, CDKN1A acts as an activator of proliferation and is closely regulated through Caspase-3 mediated degradation . CDKN1A downregulation might suggest Caspase-3 activation and cleavage of CDKN1A. The "Necroptosis Signaling Pathway" was detected to be the most activated pathway following EPX treatment of LCL721.221/sHLA-B*57:01 cells. Cell death through necroptosis is a form of programmed necrosis that is mediated by pattern recognition receptors (PRR) and diverse cytokines. Necroptosis of cells results in the secretion of damage-related pattern molecules (DAMPs) and, subsequently, an inflammatory immune response . Our observations indicate an unbalance of anti-inflammatory processes through rather downregulation of certain proteins that might lead to an excessive immune reaction in the affected patients with the susceptible HLA allele that is mainly caused by EPX. Taken together, although EPX/HLA-B*57:01 cooperation introduced the described drastic changes in the proteome, alterations through CBZ/HLA-B*57:01 cooperation were only marginal. It seems obvious that, similar to CBZ-induced hypersensitivity in HLA-B*15:02+ patients, EPX is the main driver for the SJS/TEN phenotype in HLA-B*57:01 patients. The engagement of EPX and the HLA molecule seems to perturb the intracellular processing of healthy cells and produces a stress response resulting in DNA damage and consequently, extensive inflammation. The possibility to study the effect of EPX/HLA and CBZ/HLA orthogonally in the present study offers the potential to appreciate the different clinical outcomes of HLA-mediated CBZ hypersensitivity. The metabolization of CBZ to EPX occurs in the cytochrom P450 system in the liver. Although CBZ is metabolized to EPX, the balance between both drugs shifts towards an excess of EPX, the inflammatory life-threatening condition of concerned patients therefore becomes more severe and might shift from the initiation of SJS to TEN. TEN is a serious and fatal condition for which the outcome is in >50% of affected patients lethal or at least leads to incurable long-term harm. To embed these findings into the biological context of systemic inflammation, the key processes that might drive the hypersensitivity reaction are pathways that were found to be upregulated, indicating a beginning of cell death in HLA-B*57:01 transduced cells, for example the strong activation of necroptosis pathway after EPX treatment . A recent study of the hypersensitivity reaction in HLA-B*15:02+ patients revealed that the presence of CD4+CD25+CD127loCD39+ Treg that can reduce the presence of extracellular ATP by degrading it via CD39 and CD73 to adenosine determines the conversion from a non-responder to a responder . By taking this study into account, we hypothesize that releasing intracellular ATP into the extracellular matrix facilitated by inflammatory processes induced by EPX is the initial step towards a systemic inflammation when not enough CD4+CD25+CD127loCD39+ Treg are present to reduce the effect of extracellular ATP and subsequent IFNg production. 5. Conclusions The future of pharmacological appreciation of drug and medical safety relies on the comprehension of the functional consequences of individual immunogenetics. Acknowledgments The excellent technical assistance and scientific contribution of Ulrike Schrameck, Karsten Heidrich and Wiebke Hiemisch is gratefully acknowledged. Supplementary Materials The following supporting information can be downloaded at: Figure S1: Ingenuity pathway analysis of proteomic changes induced by (A) EPX or (B) CBZ treatment in LCL721.221/mHLA-B*57:01 cells vs. parental cells. Figure S2: Analysis of proteomic changes and predicted upstream regulators after treatment of LCL721.221/sHLA-B*57:01 and parental LCL721.221 cells with CBZ or EPX. Volcano plot of protein expression in LCL721.221/sHLA-B*57:01 vs. parental LCL721.221 cells after CBZ (A) or EPX treatment (B). Labeled proteins overlap between both treatment conditions. Click here for additional data file. Author Contributions Conceptualization, F.J.H., C.B.-D. and R.B.; methodology, F.J.H. and Z.M.H.; software, F.J.H., J.K., A.P. and F.S.; validation, F.J.H. and Z.M.H.; formal analysis, F.J.H., C.B.-D., J.K., A.P. and F.S.; investigation, F.J.H., F.S. and C.B.-D.; resources, R.B.; writing--original draft preparation, F.J.H., F.S. and C.B.-D.; writing--review and editing, F.J.H., F.S. and C.B.-D.; visualization, F.S.; supervision, C.B.-D. and R.B. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD037502. Conflicts of Interest The authors declare no conflict of interest. Figure 1 CBZ and EPX concentration on sHLA-B*57:01 molecules post CBZ or EPX treatment. 150 ng affinity purified sHLA-B*57:01 molecules post drug treatment were measured in four replicates (n = 4) by UPLC-MS/MS. Figure 2 Proteome analysis LCL721.221/sHLA-B*57:01 cells or parental LCL721.221 cells treated with either CBZ or EPX. Heat map of proteins significantly changed (p < 0.05) after FDR-based ANOVA testing. Figure 3 Analysis of proteomic changes and predicted upstream regulators after treatment of LCL721.221/sHLA-B*57:01 and parental LCL721.221 cells with CBZ or EPX. (A) Volcano plot of protein expression in LCL721.221/sHLA-B*57:01 vs. parental LCL721.221 cells after CBZ treatment. (B) Volcano plot of protein expression in LCL721.221/sHLA-B*57:01 vs. parental LCL721.221 cells after EPX treatment. Highlighted proteins were predicted to be responsive to the predicted upstream regulator ERBB2 or ISNR after EPX treatment and are altered significantly with at least a two-fold change in expression. Horizontal dotted line indicates two-fold difference, vertical line p-value threshold of p < 0.05, blue dots: significant two-fold downregulation, red dots: Two-fold significant upregulation. downregulated proteins that overlap between EPX-treated cells are depicted in green. (C) Predicted upstream regulators by IPATM. Figure 4 IPA disease and function analysis of proteomic changes induced by EPX or CBZ treatment in LCL721.221/mHLA-B*57:01 cells vs. parental LCL721.221 cells. Only functions are depicted that were predicted to be downregulated (Z-score > 2 or <-2) by the IPA software. Figure 5 GSEA global enrichment analysis of proteomic changes induced by EPX treatment in LCL721.221/mHLA-B*57:01 cells vs. parental cells. (A) Heatmap of regulated proteins in HLA-B*57:01 expressing and parental LCL721.221 cells following EPX treatment. (B) GSEA enrichment plot HP_CHRONIC_OTTIS_MEDIA. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Simper G. Celik A.A. Kunze-Schumacher R.B.H. Blasczyk R. Bade-Doding C. Physiology and Pathology of Drug Hypersensitivity: Role of Human Leukocyte Antigens IntechOpen London, UK 2017 2. Ho G.-G.T. Hiemisch W. Pich A. Matern M. Graser L.S. Blasczyk R. Bade-Doeding C. Simper G.S. Small Molecule/HLA Complexes Alter the Cellular Proteomic Content IntechOpen London, UK 2021 3. Nebeker J.R. Barach P. Samore M. Clarifying adverse drug events: A clinician's guide to terminology, documentation, and reporting Ann. Intern. Med. 2004 140 795 801 10.7326/0003-4819-140-10-200405180-00009 15148066 4. Gurwitz J.H. Field T.S. Avorn J. McCormick D. Jain S. Eckler M. Benser M. Edmondson A.C. Bates D.W. Incidence and preventability of adverse drug events in nursing homes Am. J. Med. 2000 109 87 94 10.1016/S0002-9343(00)00451-4 10967148 5. Bates D.W. Cullen D.J. Laird N. Petersen L.A. Small S.D. Servi D. Laffel G. Sweitzer B. Shea B.F. Hallisey R. Incidence of adverse drug events and potential adverse drug events. Implications for prevention. ADE Prevention Study Group JAMA 1995 274 29 34 10.1001/jama.1995.03530010043033 7791255 6. Edwards I.R. Aronson J.K. Adverse drug reactions: Definitions, diagnosis, and management Lancet 2000 356 1255 1259 10.1016/S0140-6736(00)02799-9 11072960 7. Rawlins M.D. Clinical pharmacology. Adverse reactions to drugs Br. Med. J. (Clin. Res. Ed.) 1981 282 974 976 10.1136/bmj.282.6268.974 6781682 8. Pichler W.J. Hausmann O. Classification of Drug Hypersensitivity into Allergic, p-i, and Pseudo-Allergic Forms Int. Arch. Allergy Immunol. 2016 171 166 179 10.1159/000453265 27960170 9. Hetherington S. Hughes A.R. Mosteller M. Shortino D. Baker K.L. Spreen W. Lai E. Davies K. Handley A. Dow D.J. Genetic variations in HLA-B region and hypersensitivity reactions to abacavir Lancet 2002 359 1121 1122 10.1016/S0140-6736(02)08158-8 11943262 10. Mallal S. Nolan D. Witt C. Masel G. Martin A. Moore C. Sayer D. Castley A. Mamotte C. Maxwell D. Association between presence of HLA-B*5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse-transcriptase inhibitor abacavir Lancet 2002 359 727 732 10.1016/S0140-6736(02)07873-X 11888582 11. Chung W.H. Hung S.-I. Hong H.-S. Hsih M.-S. Yang L.-C. Ho H.-C. Wu J.-Y. Chen Y.-T. Medical genetics: A marker for Stevens-Johnson syndrome Nature 2004 428 486 10.1038/428486a 12. Deshpande P. Hertzman R.J. Palubinsky A.M. Giles J.B. Karnes J.H. Gibson A. Phillips E.J. Immunopharmacogenomics: Mechanisms of HLA-Associated Drug Reactions Clin. Pharmacol. Ther. 2021 110 607 615 10.1002/cpt.2343 34143437 13. Zinkernagel R.M. Doherty P.C. Restriction of in vitro T cell-mediated cytotoxicity in lymphocytic choriomeningitis within a syngeneic or semiallogeneic system Nature 1974 248 701 702 10.1038/248701a0 4133807 14. Neefjes J. Jongsma M.L.M. Paul P. Bakke O. Towards a systems understanding of MHC class I and MHC class II antigen presentation Nat. Rev. Immunol. 2011 11 823 836 10.1038/nri3084 22076556 15. Falk K. Rotzschke O. Stevanovic S. Jung G. Rammensee H.-G. Allele-specific motifs revealed by sequencing of self-peptides eluted from MHC molecules. 1991 J. Immunol. 2006 177 2741 2747 16920906 16. Wieczorek M. Abualrous E.T. Sticht J. Alvaro-Benito M. Stolzenberg S. Noe F. Freund C. Major Histocompatibility Complex (MHC) Class I and MHC Class II Proteins: Conformational Plasticity in Antigen Presentation Front. Immunol. 2017 8 292 10.3389/fimmu.2017.00292 28367149 17. McCormack M. Alfirevic A. Bourgeois S. Farrell J.J. Kasperaviciute D. Carrington M. Sills G.J. Marson T. Jia X. De Bakker P.I. HLA-A*3101 and Carbamazepine-Induced Hypersensitivity Reactions in Europeans N. Engl. J. Med. 2011 364 1134 1143 10.1056/NEJMoa1013297 21428769 18. Hari Y. Frutig-Schnyder K. Hurni M. Yawalkar N. Zanni M.P. Schnyder B. Kappeler A. Von Greyerz S. Braathen L.R. Pichler W.J. T cell involvement in cutaneous drug eruptions Clin. Exp. Allergy 2001 31 1398 1408 10.1046/j.1365-2222.2001.01164.x 11591190 19. Nassif A. Bensussan A. Boumsell L. Deniaud A. Moslehi H. Wolkenstein P. Bagot M. Roujeau J.-C. Toxic epidermal necrolysis: Effector cells are drug-specific cytotoxic T cells J. Allergy Clin. Immunol. 2004 114 1209 1215 10.1016/j.jaci.2004.07.047 15536433 20. Yip V.L.M. Pirmohamed M. The HLA-A*31:01 allele: Influence on carbamazepine treatment Pharm. Pers. Med. 2017 10 29 38 10.2147/PGPM.S108598 21. Li Y. Deshpande P. Hertzman R.J. Palubinsky A.M. Gibson A. Phillips E.J. Genomic Risk Factors Driving Immune-Mediated Delayed Drug Hypersensitivity Reactions Front. Genet. 2021 12 641905 10.3389/fgene.2021.641905 33936169 22. Hertzman R.J. Deshpande P. Gibson A. Phillips E.J. Role of pharmacogenomics in T-cell hypersensitivity drug reactions Curr. Opin. Allergy Clin. Immunol. 2021 21 327 334 10.1097/ACI.0000000000000754 34039850 23. Simper G.S. Graser L.S. Celik A.A. Kuhn J. Kunze-Schumacher H. Ho G.-G.T. Blasczyk R. Pich A. Bade-Doeding C. The Mechanistic Differences in HLA-Associated Carbamazepine Hypersensitivity Pharmaceutics 2019 11 536 10.3390/pharmaceutics11100536 31618895 24. Mockenhaupt M. Wang C. Hung S. Sekula P. Schmidt A.H. Pan R. Chen C. Dunant A. Le Gouvello S. Schumacher M. HLA-B*57:01 confers genetic susceptibility to carbamazepine-induced SJS/TEN in Europeans Allergy 2019 74 2227 2230 10.1111/all.13821 30972788 25. Sassolas B. Haddad C. Mockenhaupt M. Dunant A. Liss Y. Bork K. Haustein U.F. Vieluf D. Roujeau J.C. Le Louet H. ALDEN, an Algorithm for Assessment of Drug Causality in Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis: Comparison With Case-Control Analysis Clin. Pharmacol. Ther. 2010 88 60 68 10.1038/clpt.2009.252 20375998 26. Hetherington S. McGuirk S. Powell G. Cutrell A. Naderer O. Spreen B. Lafon S. Pearce G. Steel H. Hypersensitivity reactions during therapy with the nucleoside reverse transcriptase inhibitor abacavir Clin. Ther. 2001 23 1603 1614 10.1016/S0149-2918(01)80132-6 11726000 27. Mallal S. Phillips E. Carosi G. Molina J.-M. Workman C. Tomazic J. Jagel-Guedes E. Rugina S. Kozyrev O. Cid J.F. HLA-B*5701 Screening for Hypersensitivity to Abacavir N. Engl. J. Med. 2008 358 568 579 10.1056/NEJMoa0706135 18256392 28. Chessman D. Kostenko L. Lethborg T. Purcell A.W. Williamson N.A. Chen Z. Kjer-Nielsen L. Mifsud N.A. Tait B.D. Holdsworth R. Human Leukocyte Antigen Class I-Restricted Activation of CD8+ T Cells Provides the Immunogenetic Basis of a Systemic Drug Hypersensitivity Immunity 2008 28 822 832 10.1016/j.immuni.2008.04.020 18549801 29. Illing P.T. Vivian J.P. Dudek N.L. Kostenko L. Chen Z. Bharadwaj M. Miles J.J. Kjer-Nielsen L. Gras S. Williamson N.A. Immune self-reactivity triggered by drug-modified HLA-peptide repertoire Nature 2012 486 554 558 10.1038/nature11147 22722860 30. Saag M. Balu R. Phillips E. Brachman P. Martorell C. Burman W. Stancil B. Mosteller M. Brothers C. Wannamaker P. High Sensitivity of Human Leukocyte Antigen-B*5701 as a Marker for Immunologically Confirmed Abacavir Hypersensitivity in White and Black Patients Clin. Infect. Dis. 2008 46 1111 1118 10.1086/529382 18444831 31. Genin E. Chen D.-P. Hung S.-I. Sekula P. Schumacher M.A. Chang P.-Y. Tsai S.-H. Wu T.-L. Bellon T. Tamouza R. HLA-A*31:01 and different types of carbamazepine-induced severe cutaneous adverse reactions: An international study and meta-analysis Pharm. J. 2013 14 281 288 10.1038/tpj.2013.40 32. Simper G.S. Ho G.-G. Celik A. Huyton T. Kuhn J. Kunze-Schumacher H. Blasczyk R. Bade-Doding C. Carbamazepine-Mediated Adverse Drug Reactions: CBZ-10,11-epoxide but Not Carbamazepine Induces the Alteration of Peptides Presented by HLA-B *15:02 J. Immunol. Res. 2018 2018 5086503 10.1155/2018/5086503 30302345 33. Haukamp F.J. Gall E. Ho G.-G.T. Hiemisch W. Stieglitz F. Kuhn J. Blasczyk R. Pich A. Bade-Doding C. Unravelling the Proteomics of HLA-B*57:01+ Antigen Presenting Cells during Abacavir Medication J. Pers. Med. 2022 12 40 10.3390/jpm12010040 35055355 34. Ho G.-G.T. Heinen F.J. Huyton T. Blasczyk R. Bade-Doding C. HLA-F*01:01 presents peptides with N-terminal flexibility and a preferred length of 16 residues Immunogenetics 2019 71 353 360 10.1007/s00251-019-01112-1 30941482 35. Gall E. Stieglitz F. Pich A. Behrens G.M.N. Kuhn J. Blasczyk R. Haukamp F.J. Bade-Doding C. Proteomic Profiling and T Cell Receptor Usage of Abacavir Susceptible Subjects Biomedicines 2022 10 693 10.3390/biomedicines10030693 35327495 36. Cox J. Mann M. MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification Nat. Biotechnol. 2008 26 1367 1372 10.1038/nbt.1511 19029910 37. UniProt C. UniProt: The universal protein knowledgebase in 2021 Nucleic Acids Res. 2021 49 D480 D489 33237286 38. Tyanova S. Temu T. Sinitcyn P. Carlson A. Hein M.Y. Geiger T. Mann M. Cox J. The Perseus computational platform for comprehensive analysis of (prote)omics data Nat. Methods 2016 13 731 740 10.1038/nmeth.3901 27348712 39. R Core Team R: A Language and Environment for Statistical Computing R Foundation for Statistical Computing Vienna, Austria 2013 Available online: (accessed on 11 November 2022) 40. Gu Z. Eils R. Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data Bioinformatics 2016 32 2847 2849 10.1093/bioinformatics/btw313 27207943 41. Wickham H. ggplot2: Elegant Graphics for Data Analysis Springer New York, NY, USA 2016 42. Subramanian A. Tamayo P. Mootha V.K. Mukherjee S. Ebert B.L. Gillette M.A. Paulovich A. Pomeroy S.L. Golub T.R. Lander E.S. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles Proc. Natl. Acad. Sci. USA 2005 102 15545 15550 10.1073/pnas.0506580102 16199517 43. Mootha V.K. Lindgren C.M. Eriksson K.-F. Subramanian A. Sihag S. Lehar J. Puigserver P. Carlsson E. Ridderstrale M. Laurila E. PGC-1a-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes Nat. Genet. 2003 34 267 273 10.1038/ng1180 12808457 44. Griss J. Perez-Riverol Y. Lewis S. Tabb D.L. Dianes J.A. Del-Toro N. Rurik M. Walzer M. Kohlbacher O. Hermjakob H. Recognizing millions of consistently unidentified spectra across hundreds of shotgun proteomics datasets Nat. Methods 2016 13 651 656 10.1038/nmeth.3902 27493588 45. Kunze-Schumacher H. Blasczyk R. Bade-Doeding C. Soluble HLA Technology as a Strategy to Evaluate the Impact of HLA Mismatches J. Immunol. Res. 2014 2014 1 8 10.1155/2014/246171 25254222 46. Charlton O.A. Harris V.R. Phan K. Mewton M.E. Jackson C.J. Cooper A. Toxic Epidermal Necrolysis and Steven-Johnson Syndrome: A Comprehensive Review Adv. Wound Care 2020 9 426 439 10.1089/wound.2019.0977 32520664 47. Phase 0 trials: A platform for drug development? Lancet 2009 374 176 10.1016/S0140-6736(09)61309-X 19616703 48. Le Tourneau C. Lee J. Siu L. Dose escalation methods in phase I cancer clinical trials J. Natl. Cancer Inst. 2009 101 708 720 10.1093/jnci/djp079 19436029 49. Kerr B.M. Thummel K.E. Wurden C.J. Klein S.M. Kroetz D.L. Gonzalez F.J. Levy R. Human liver carbamazepine metabolism: Role of CYP3A4 and CYP2C8 in 10,11-epoxide formation Biochem. Pharmacol. 1994 47 1969 1979 10.1016/0006-2952(94)90071-X 8010982 50. Madson J.G. Lynch D.T. Tinkum K.L. Putta S.K. Hansen L.A. Erbb2 Regulates Inflammation and Proliferation in the Skin after Ultraviolet Irradiation Am. J. Pathol. 2006 169 1402 1414 10.2353/ajpath.2006.060082 17003495 51. Cardinez C. Miraghazadeh B. Tanita K. Da Silva E. Hoshino A. Okada S. Chand R. Asano T. Tsumura M. Yoshida K. Gain-of-function IKBKB mutation causes human combined immune deficiency J. Exp. Med. 2018 215 2715 2724 10.1084/jem.20180639 30337470 52. Snyder M. He W. Zhang J.J. The DNA replication factor MCM5 is essential for Stat1-mediated transcriptional activation Proc. Natl. Acad. Sci. USA 2005 102 14539 14544 10.1073/pnas.0507479102 16199513 53. Banerjee S. Biehl A. Gadina M. Hasni S. Schwartz D.M. JAK-STAT Signaling as a Target for Inflammatory and Autoimmune Diseases: Current and Future Prospects Drugs 2017 77 521 546 10.1007/s40265-017-0701-9 28255960 54. Fuchs J. Cheblal A. Gasser S. Underappreciated Roles of DNA Polymerase delta in Replication Stress Survival Trends Genet. 2021 37 476 487 10.1016/j.tig.2020.12.003 33608117 55. Zhang Z. POLD2 is activated by E2F1 to promote triple-negative breast cancer proliferation Front. Oncol. 2022 12 981329 10.3389/fonc.2022.981329 36119494 56. Cong F. Long J. Liu J. Deng Z. Yan B. Liang C. Huang X. Liu J. Tang W. An integrative analysis revealing POLD2 as a tumor suppressive immune protein and prognostic biomarker in pan-cancer Front. Genet. 2022 13 877468 10.3389/fgene.2022.877468 36081989 57. Orlandi A. Ferlosio A. Arcuri G. Scioli M.G. De Falco S. Spagnoli L.G. Flt-1 expression influences apoptotic susceptibility of vascular smooth muscle cells through the NF-kappaB/IAP-1 pathway Cardiovasc. Res. 2010 85 214 223 10.1093/cvr/cvp288 19720604 58. Jackson N.M. Ceresa B.P. EGFR-mediated apoptosis via STAT3 Exp. Cell Res. 2017 356 93 103 10.1016/j.yexcr.2017.04.016 28433699 59. Hall I.P. FLT1: A potential therapeutic target in sepsis-associated ARDS? Lancet Respir. Med. 2020 8 219 220 10.1016/S2213-2600(19)30404-7 31982042 60. Sabbah D.A. Hajjo R. Sweidan K. Review on Epidermal Growth Factor Receptor (EGFR) Structure, Signaling Pathways, Interactions, and Recent Updates of EGFR Inhibitors Curr. Top. Med. Chem. 2020 20 815 834 10.2174/1568026620666200303123102 32124699 61. Paul C. Wolkenstein P. Adle H. Wechsler J. Garchon H. Revuz J. Roujeau J. Apoptosis as a mechanism of keratinocyte death in toxic epidermal necrolysis Br. J. Dermatol. 1996 134 710 714 10.1111/j.1365-2133.1996.tb06976.x 8733377 62. Liu W. Chen Y. Zeng G. Yang T. Song W. INSR mediated by transcription factor KLF4 and DNA methylation ameliorates osteoarthritis progression via inactivation of JAK2/STAT3 signaling pathway Am. J. Transl. Res. 2020 12 7953 7967 33437372 63. Woo M. Hakem R. Furlonger C. Hakem A. Duncan G.S. Sasaki T. Bouchard D. Lu L. Wu G.E. Paige C.J. Caspase-3 regulates cell cycle in B cells: A consequence of substrate specificity Nat. Immunol. 2003 4 1016 1022 10.1038/ni976 12970760 64. Seo J. Nam Y.W. Kim S. Oh D.-B. Song J. Necroptosis molecular mechanisms: Recent findings regarding novel necroptosis regulators Exp. Mol. Med. 2021 53 1007 1017 10.1038/s12276-021-00634-7 34075202 65. Bertheloot D. Latz E. Franklin B.S. Necroptosis, pyroptosis and apoptosis: An intricate game of cell death Cell. Mol. Immunol. 2021 18 1106 1121 10.1038/s41423-020-00630-3 33785842 66. Shen M. Lim J.M.E. Chia C. Ren E.C. CD39+ regulatory T cells modulate the immune response to carbamazepine in HLA-B*15:02 carriers Immunobiology 2019 225 151868 10.1016/j.imbio.2019.11.003 31784043 |
PMC10000581 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051070 foods-12-01070 Article An Authentication Survey on Retail Seafood Products Sold on the Bulgarian Market Underlines the Need for Upgrading the Traceability System Tinacci Lara Conceptualization Methodology Formal analysis Investigation Writing - original draft 1* Stratev Deyan Conceptualization Methodology Project administration Funding acquisition 2 Strateva Mariyana Investigation Data curation 3 Zhelyazkov Georgi Investigation Data curation 4 Kyuchukova Ralica Investigation Data curation 2 Armani Andrea Conceptualization Methodology Writing - review & editing Supervision 1 Beltran Gracia Jose Antonio Academic Editor 1 Department of Veterinary Sciences, University of Pisa, Via delle Piagge 2, 56124 Pisa, Italy 2 Department of Food Quality and Safety and Veterinary Legislation, Faculty of Veterinary Medicine, Trakia University, 6000 Stara Zagora, Bulgaria 3 Department of Veterinary Anatomy, Histology and Embryology, Faculty of Veterinary Medicine, Trakia University, 6000 Stara Zagora, Bulgaria 4 Department of Animal Husbandry--Non-Ruminants and Other Animals, Faculty of Agriculture, Trakia University, 6000, Stara Zagora, Bulgaria * Correspondence: [email protected] 02 3 2023 3 2023 12 5 107026 1 2023 13 2 2023 27 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Economically motivated or accidental species substitutions lead to economic and potential health damage to consumers with a loss of confidence in the fishery supply chain. In the present study, a three-year survey on 199 retail seafood products sold on the Bulgarian market was addressed to assess: (1) product authenticity by molecular identification; (2) trade name compliance to the list of official trade names accepted in the territory; (3) adherence of the list in force to the market supply. DNA barcoding on mitochondrial and nuclear genes was applied for the identification of whitefish (WF), crustaceans (C) and mollusks (cephalopods--MC; gastropods--MG; bivalves--MB) except for Mytilus sp. products for which the analysis was conducted with a previously validated RFLP PCR protocol. Identification at the species level was obtained for 94.5% of the products. Failures in species allocation were reconducted due to low resolution and reliability or the absence of reference sequences. The study highlighted an overall mislabeling rate of 11%. WF showed the highest mislabeling rate (14%), followed by MB (12.5%), MC (10%) and C (7.9%). This evidence emphasized the use of DNA-based methods as tools for seafood authentication. The presence of non-compliant trade names and the ineffectiveness of the list to describe the market species varieties attested to the need to improve seafood labeling and traceability at the national level. seafood labeling authenticity mislabeling DNA barcoding PCR-RFLP analysis seafood marking Bulgarian Ministry of Education and Science577/17.08.2018 This study was conducted within the National Scientific Program, "Healthy Foods for a Strong Bio-Economy and Quality of Life," approved by DCM No. 577/17.08.2018 and funded by the Bulgarian Ministry of Education and Science. pmc1. Introduction Global fish consumption has grown steadily in the last five decades, driven by a combination of population growth, rising incomes and changes in food habits, as well as strong expansion in fish production . Aquatic foods are increasingly recognized for their key role in food security and global nutrition. In addition, they represent an invaluable source of income and employment with a direct impact on the livelihoods of a substantial percentage of the population, mainly in Asia and Africa. According to FAO, the amount of seafood destined for human consumption in 2020 was estimated to be 20.2 kg per capita, doubling the average of 9.9 kg per capita noted in 1960. While a plateau has been reached for the volumes of aquatic products obtained from fishing activities, this trend is set to grow further, mainly due to the expansion and modernization of aquaculture systems. The increase in per capita consumption is also favorably influenced by the speeding-up of the supply chain and the improvement of transport logistics, which have led to the exponential expansion of the global product offer . The European Union (EU) holds a considerable position in this scenario, with consumption estimates of around 25 kg per capita per year . However, a significant variation between Member States exists. Seafood prices, purchasing ability of the consumer and net income, and culturally-driven dietary preference are the factors most frequently affecting the variability of consumption at EU and international levels . In Bulgaria, apparent domestic seafood consumption appears to be limited, ranging from 5.3 kg to 7.5 kg per capita per year, and an average value of 6 kg per capita, according to EUMOFA and the Bulgarian Ministry of Agriculture for the years 2019-2020 . In addition, according to data published by the EU Commission on consumer habits regarding fishery and aquaculture products for the year 2021, the consumption rate generally does not exceed a monthly frequency . Evidence of this limited consumption is also attested by a consumer survey conducted in 2018 by Stancheva , from which emerged that seafood products have an average consumption rate not exceeding 1-2 times per month by about the 46% of the respondents, while only 26% declared weekly consumption. The low buying appetite towards seafood is generally due to (1) the consumer's lack of perception of the benefits of a regular consumption and (2) the medium/high price of seafood . These aspects could therefore constitute major limiting factors to the expansion of seafood market demand in Bulgaria. However, despite the unchanged consumption of fresh water and marine seafood from both inland waters and the coast of the Black Sea, an increase in fresh, chilled, frozen and variously processed marine seafood imports (fish, crustaceans and mollusks) of EU and extra-EU origin on the Bulgarian market has been recently reported . This evidence was also previously highlighted in a survey conducted in 2019, in which the presence of imported seafood, especially at the large retail level, was described as mostly pre-packaged fresh and frozen . It is well known that globalization and the complexity of the supply chain, together with the distribution system, can potentially expose the seafood market to an increased opportunity for the perpetration of deceptive behaviors . Fraudulent incidents are frequently described, leading to misdescriptions, mislabeling and economically motivated species substitutions, resulting in consumer economic damage and loss of confidence in the sector chain. These incidents could potentially include health risks whenever substitution events include the illicit presence of toxic species or the omission of allergens . To ensure safety, transparency, and fair trading, which are founding principles of EU food law , specific provisions have been issued for the labeling and presentation of the sale of fishery products by EU Regulation No.1379/2013 . Pursuant to this regulation, the Member States are delegated to publish and update a list reporting the official seafood trade names corresponding to the species' scientific names accepted in the national territories for product sales. According to the current EU legislation, the responsibility for the correctness of the information of the product to the consumer lies directly with the food business operator who prepares the product for sale; the organization of official controls on labeling is entrusted to a central competent authority identified by each member state . Specifically, in Bulgaria, the responsibility for official controls in this area is delegated to the Bulgaria Food Safety Agency (BFSA; accessed on 30 October 2022). Nonetheless, fraudulent incidents in the seafood supply chain are still well documented both at the international and EU level , although data as regards labeling compliance and mislabeling seafood rates in Central-Eastern and Eastern European countries are rarely reported . In this respect, the regulation also promotes control measures aimed at verifying the identity of products through the use of available technology, including DNA testing, to deter fraudulent substitution practices. Several DNA-testing techniques have been successfully applied in market surveys to verify the labeling compliance of commercial seafood collected at retail. Particularly, sequencing-based methods like FINS and DNA barcoding based on the analysis of mitochondrial (COI, cytb; 16S rRNA) targets, are recognized as valuable tools for seafood species identification. Nuclear targets (rhodopsin, Phosphoenolpyruvate Carboxykinase (PEPCK), and Polyphenolic Adhesive Protein (PAP)) have been additionally selected in the presence of introgression or hybridization phenomena . Additionally and alternatively, PCR-restriction fragment length polymorphism (RFLP), amplified fragment length polymorphism (AFLP), or single-stranded conformational polymorphism (SSCP) on both mitochondrial and nuclear targets have been applied as targeted methods to design genus or species-specific assays . In this regard, the validity of an RFLP protocol on the PAP gene for species-specific discrimination of bivalve mollusks--mussels--belonging to the genus Mytilus sp., has been confirmed . The exposure of the Bulgarian market to fraudulent phenomena was first documented in a pilot study conducted in 2017 . In that survey, a DNA-barcoding approach applied to assess the product's authenticity highlighted an overall substitution rate of 17.7%. The species substitutions were equally distributed between unprocessed and processed products (smoked and marinated products). Particularly, five plausibly deceptive incidents for economic gain were finally pointed out, two of which consisted in the replacement of Gadus chalcogramma (Alaska pollock) with Pangasianodon hypophthalmus (striped catfish) and the remaining three cases in the replacement of Illex argentinus (Argentine squid) with Dosidicus gigas (Humboldt squid). The remaining mislabeling incidents, consisting of species substitution between species of similar commercial value, were linked the improper training of the operators in species morphological identification . Furthermore, the study highlighted the urgency of the revision of the Bulgarian official list of seafood trade names. This evidence was reaffirmed in a further study on the validity and accuracy of the new official Bulgarian list of seafood trade and the list evolution and adherence to the Bulgarian market's trend . Considering the evidence acquired in the pilot study on products belonging to two taxonomic classes (fish and gastropod mollusks) purchased at retail in a single city (Stara Zagora) , a three-year survey was conducted for the assessment of the authenticity of seafood products belonging to five distinct taxonomic classes, purchased in retail settings in four different cities across the country. The products' molecular identification to verify the compliance of the labeled species designations was conducted by means of DNA barcoding and PCR-RFLP based on mitochondrial and nuclear markers. Furthermore, the labeled designations were checked against the Bulgarian official list of seafood trade names to verify the use of authorized commercial designations. Finally, the comparison of the labeled designation with the accepted designation in force was performed to verify the adherence of the official list to the products basket available at retail. 2. Materials and Methods 2.1. Samples Collection: Sampling Strategy and Products Variety In the present study, the sampling plan for selecting the seafood categories was structured by considering the characteristics of the Bulgarian market in terms of production, imports and exports and the most frequent seafood fraud incidents reported in a previous study and at the EU level. Within the fish taxonomical class, the sampling was targeted on white fish (WF) since Tinacci et al. observed that deliberate substitutions mainly involved Alaskan pollock (Gadus chalcogrammus) products. This choice was also performed in response to clear evidence of deliberate substitution phenomena involving valuable WF, particularly those belonging to the Gadidae and Merluccidae families, both at the international and EU levels . In addition, in a nationwide survey conducted on the Bulgarian fish retail market, Gadidae and Merluccidae were reported to be among the products (frozen and filleted) most frequently found on sale, especially in large retail settings . The survey also revealed the association of a few generic commercial designations for a considerable number of species belonging to the two families characterized by heterogeneous commercial values. This aspect was considered by the authors as favoring consumers' confusion on fish value and market exposure to deceitful incidents for economic gain. . Within cephalopods class, the number of products was mainly selected in relation to the results of the pilot study and the sampling was primarily targeted at squid products, for which three substitutions were shown in the limited sampling of products under analysis (N = 11) . Moreover, the sampling was also extended to crustaceans, mollusks bivalves, and gastropods classes by virtue of the increased interest in the Bulgarian market, showing crustaceans to be second in terms of nationwide production and import, followed by mollusk categories . For gastropod mollusks (MG), the sampling was deliberately reduced as, according to bibliographic data, their consumption is mostly limited to the Rapana venosa species . The products were purchased from large distribution retailers in four different towns (Stara Zagora, Shumen, Varna and Dobrich) from March 2019 to July 2021, and the sampling resulted in the collection of 199 seafood products, including whitefish (WF; N = 100), cephalopod mollusks (MC; N = 40), crustaceans (C; N = 38), mollusks bivalves (MB; N = 16), and mollusks gastropods (MG; N = 5). Therefore, the different proportion of products collected for the five product categories (WF = 100, 50.3%; MC = 40, 20.1%; C = 38, 19.1%; MB = 16, 8.0%; MG; N = 5, 2.5%) more or less reflects the weight of the various categories on the national market. The products were transferred to the Department of Food Quality and Safety (Faculty of Veterinary Medicine, Trakia University). Each product was classified according to Regulation (EC) No. 852/2004 as unprocessed (fresh, frozen, beheaded, cleaned and filleted) or processed (marinated and precooked), then provided with a progressive numerical code, which was recorded in a single file together with labeling information (Table S1). Finally, 1-5 g of muscle tissue were collected, dehydrated by means of 95% ethanol and sent to FishLab (Department of Veterinary Science, University of Pisa) for molecular identification. 2.2. DNA Extraction, Molecular Target Selection, Amplification, and Sequencing Total DNA extraction from each dehydrated tissue sample was performed according to the salting-out procedure proposed by Armani et al. , starting from 50 to 100 mg of tissue. Final DNA concentrations and quality were checked with a Nanodrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) according to the manufacturer guidelines and absorbance ratios A260/A280 > 2.0 and A260/A230 >1.8 were set as minimum values of nucleic acid purity. The selection of the molecular target followed the decisional procedure (decision tree) proposed by Tinacci et al. . A fragment of 655-658 bp of the Cytochrome C oxidase subunit I (COI) gene was chosen as an elective target for species identification for all the product categories. Additionally, two mitochondrial gene coding for the enzyme cytochrome b and 16S ribosomal RNA (cytb and 16SrRNA) and one nuclear gene coding for the enzyme Phosphoenolpyruvate Carboxykinase (PEPCK) were applied in the study to enhance the discriminatory molecular ability when the elective target alone failed to achieve species identification according to the seafood product's taxonomical class under analysis. Products labeled as Mytilus sp. (Table S1) were only tested by the PCR-RFLP technique for the analysis of a non-repetitive region of the nuclear Polyphenolic Adhesive Foot Protein (PAP) gene as described in Giusti et al. . Details on elective and additional targets, primer pairs and amplification protocol applied in the study are summarized in Table 1. After amplification, 5 mL of each PCR product was checked on a 1.8% agarose gel previously stained with GelRedTM Nucleic Acid Gel Stain (Biotium, Hayward, CA, USA). The presence of the expected amplicon and the final concentration was verified by comparison with the standard molecular marker SharpMassTM50-DNA (Euroclone Spa, Pero; Milano, Italy). A concentration equal to or greater than 5 mg/mL of PCR product was set as a threshold value for subsequent sequencing reaction according to the sequencing laboratory operative procedures and sent to Eurofins Genomics laboratories (Eurofins Genomics, Ebersberg, Germany). 2.3. Sequences Analysis and Molecular Identification The obtained sequences were aligned and edited with Clustal W in BioEdit version 7.0.9 , and the final sequences were queried against the reference sequences available in GenBank accessed on 30 July 2022 ) and, in case of COI, the BOLD accessed on 30 July 2022) databases. The species was finally allocated by setting target-specific identity score cut-offs as specified below. For the COI and cytb targets, a top match identity score >98% was applied for species allocation ; for the 16SrRNA and PEPCK targets, an identity score of 100% was defined as the cut-off threshold for species identification . Furthermore, for species allocation with the additional targets, the Neighbor-joining method and Kimura 2-parameter model with 1000 bootstrap re-samplings were applied to infer distance-based dendrograms in MEGA-11 . 2.4. Assessment of Products Labeling Compliance and Mislabeling Rate The information reported on the labels was verified against the mandatory requirements of Regulation (EU) No. 1379/2013 . Specifically, for each product included in the scope of Article 35, commercial designation of the species and its scientific name, an indication of production method, catch or farming area for aquaculture products, fishing method and thawing before the sale were considered. The scientific names were verified by consulting the official databases listed in Article 37 of the Regulation (FAO Sealife base; FAO Fishbase, World Register of Marine Species (WorMS) and ASFIS databases). Thus, a direct comparison between the molecular results and the labeled scientific names was performed. The products were declared non-compliant if the molecularly identified species did not match the scientific names labeled on the product. Furthermore, the adherence of the labeled designations against the list of accepted scientific names in force during the sampling period, included in Ordinance no. 4 of 13.01.2006 , was verified. The labeled designations were finally compared with the new list, published with Ordinance No. 13 of 30.11.2021 , published after the completion of the sampling phase. The comparison with both lists was made to assess the evolution of the list of official designations in terms of adherence to the Bulgarian market supply. 3. Results and Discussion 3.1. Distribution of Product Variety Resulting from Sampling The definition of the sampling strategy should be based on a preliminary study of the different product categories available on the market to better define the representative sample size for each taxonomical class . Thus, the sample products collected and analyzed in this study reflect the weight of the different seafood taxonomical classes and products as assessed in the preliminary sampling strategy (see Section 2.1). The study highlighted the availability at large retail levels of a fairly limited variety of species in agreement with several findings from an extensive survey conducted in 2019 on the Bulgarian seafood market, both at the local and large retail levels, to assess the diversity of available retail fish products . When limiting the comparison to retail data, the survey had already revealed increased attention towards marine imported products of Atlantic and Pacific origins. This finding showed a clear evolution of the consumption trend compared to a study conducted by Stancheva , from which the consumers' preferences appeared to be mainly oriented towards local freshwater products or locally caught marine fish products in accordance with national culinary traditions. Prominent species were immediately identifiable for each taxonomical class (WF, MC, MG, C). In detail, in WF, G. chalcogrammus accounted for 46% of products collected in the study, followed by Argentine hake (Merluccius hubbsi) (23%). The C and MC categories were led respectively by Penaeus vannameii and Dosidicus gigas representing 58% and 60% of the products, respectively. A greater equilibrium of species is observed within the MB in which the species M. chilensis is predominant but in balance with M. galloprovincialis, while the third species Perna canaliculus, was only occasionally available for sampling. With regard to MG, consumer interest, as previously introduced, is focused exclusively on R. venosa, the only species sampled for the taxonomical class and, according to Keskin et al. , one of the most exploited species along Black Sea coasts. Overall, 76.4% (N = 152) of the samples were made of unprocessed products (75.4% frozen and 1% chilled) and the remaining 23.6 % (N = 47) of distinct types of processed products (14% precooked frozen, 5.0% marinated and 4.5% canned) with different frequencies among the categories included in the study . A different presentation of the products at purchase for the different taxonomic classes was observed . In particular, the greatest variability was observed in WF where, however, the prevalence of variously cut products (59% in fillets and 9% in slices) compared to whole products can be observed. In 100% of the MC products, a standard level of processing was observed with the absence of whole products at retail. A similar assessment applies to MB, where 100% of the products presented for sale in whole form without shells. In MG, a distribution between whole products without shells and sliced products was observed where the aquatic organism was no longer immediately recognizable as it had been reduced to a pulp. For class C finally, a clear predominance of peeled products (68.6%, 26/38) over whole specimens (31.6%, 12/38) was observed. Thus, according to the results, the retailers imported from EU and non-EU producers seem to almost exclusively target frozen unprocessed and precooked products and less frequently canned products, except for MB. This evidence agrees with the data collected in Todorov , confirming the Bulgarian large retail fishery market trades are mainly oriented towards imports of prepared (frozen cleaned, filleted or sliced) and processed (canned, marinated) products of EU and non-EU origin. This is in agreement with the description also given by Vindigni and colleagues on the type of products most imported at the EU level based on data provided by the EU Commission in 2020. It is also interesting to note the degree of preparation found not only in processed products (ready-to-eat or breaded precooked) but also in unprocessed products. This observation is consistent with that reported by Stancheva regarding consumer demand for easy-to-use, fresh, frozen or canned products. 3.2. Molecular Identification and Its Limitations Total DNA sample was successfully extracted from all 199 products included in the study. All the 199 DNA samples were successfully amplified producing 208 PCR products (COI N = 186; cytb N = 11; 16S rRNA N = 8; PEPCK N = 3) for further sequencing and 13 PCR products (belonging to products labeled as Mytilus sp.) to be analyzed by RFLP (Table S2). All the 208 PCR products intended for post-sequencing analysis returned readable sequences. The length of the final sequences and the results of the post-sequencing analysis are shown in Table S2. Overall, a final species allocation was reached in 188 out of 199 products included in the study (94.5%) using: (1) a COI barcoding approach in 159 products (F = 81; MC = 39; MB = 3, C = 35); (2) or by the analysis of an additional target in 16 products (F = 11; MG = 5); (3) by the application of a PCR-RFLP protocol specifically for Mytilus sp. species allowed the final species allocation in 13 MB products as detailed in Figure 3. Specifically, 10 MB were finally assigned to the species M. chilensis, and the remaining 3 MB products (MB-3, MB-8, and MB-16) were assigned to the species M. galloprovincialis. In particular, 11 WF products finally identified as M. productus (N = 7) and M. gayi (N = 4) by the analysis of COI and cytb targets (Section 3.2.1) and five MG products were finally assigned at the species level by the analysis of COI and 16S rRNA targets (Section 3.2.2). In the remaining 11 products, the analysis exclusively led to a genus-level allocation with the assignment of 8 F products to Alepocephalus sp. and 3 C products to Metapenaeus sp. (N = 2) and Heterocarpus sp. (N = 1). Nevertheless, all the data obtained were sufficient for the subsequent verification of labeling compliance. Major limitations to the successful species identification could be due to (1) low target barcode resolution, (2) the low reliability of reference sequences or (3) the absence of reference sequences for the comparative analysis, as already highlighted by Fernandes et al. . 3.2.1. Low Target Barcode Resolution Low target barcode resolution is typically described in recently diverged and/or closely related species that are geographically isolated or are in the presence of species complexes and hybrids . In the present study, a low resolution of COI barcodes was highlighted during the analysis of products finally identified as belonging to species from Merluccius sp. and products finally not assigned at species level belonging to Metapenaeus sp. and Heterocarpus sp. Limitations in species allocation have been described in the analysis of hake products (Merluccius sp.) for which the evaluation of different mitochondrial targets such as cytb or Control Region barcodes have been proposed in numerous studies . In the present study, the use of cytb, in additional to COI, was conclusive for the product allocation to two species, M. productus and M. gayi. In fact, the similarity analysis on the COI target computed with BLASTn and K2P analysis computed with BOLD IDs had shown overlapping identity values within the threshold set for species-specific identification. The distance-based dendrogram produced on cytb target by including reference sequences belonging to 11 Merluccius sp. species and 11 sequences obtained in this study from products shows distinct and discrete species clustering and the allocation of the products sequences within the cluster M. products (F15, F16, F17, F46, F63, F66, F95) and M. gayi (F61, F79, F81, F82) clusters . Thus, in accordance with the multitarget approach proposed by Tinacci et al. , the introduction of an additional target offered a decisive improvement of the species disambiguation and the final allocation of the products at the species level. With respect to Metapenaeus sp., although the DNA barcoding technique offers an efficient tool for species identification to ensure traceability and identity verification of crustacean products in the present study, objective limitations were highlighted to the final species allocation within the genus even despite the use of the additional target 16S rRNA or PEPCK. Similar considerations can be applied to the failure of species identification for a product (C10) identified as belonging to the genus Heterocarpus sp. Limits in molecular identification of penaeid shrimps were also highlighted by Rajkumar . In this respect, several authors concurred in identifying the gene targets selected in the present study as valid for the identification of Penaeid shrimp species, nonetheless emphasizing the importance of selecting the bioinformatics method to be applied for an accurate definition of the relationships and clustering of the various species within genera clades . 3.2.2. Reference Sequences' Unreliability and Absence Reference sequences' unreliability and absence have been extensively debated and generally attributed to incorrect species-sequence associations or the submission of sequences belonging to not taxonomically validated specimens . The presence of reference sequences of uncertain and debatable reliability was highlighted in the study during the post-sequencing analysis of COI target barcodes obtained from an MC product (MC32) and from five MG products (MG1, MG2, MG3, MG4 and MG5). In the first case, MC32 was finally assigned to the species Notodarus sloanii despite an initial overlapping top identity match of the MC32 COI sequence with the reference sequences of Nototodarus sloanii (100-99.84%), and 4 reference sequences (DQ373957-60) deposited by Carlini et al. for Illex argentinus (100-99.84%). The four sequences were excluded by assuming the hypothesis of a potential error occurring during the morphological identification of the specimens. The assumption was strengthened by the results of a further cross-BLAST analysis of the sequences deposited by Carlini et al. against the GenBank repository of all I. argentinus COI reference sequences from which the identity values lower than 85% were highlighted for each of the 4 investigated sequences. A similar assumption was pursued for the final allocation at the species level of the five MG products included in the study. In this respect, the post-sequencing analysis of the COI barcodes obtained from five MG products highlighted overlapping similarity scores within the ID score threshold (98%) with R. venosa, R. bezoar and three reference sequences deposited for Turbo cornutus (HM180929, HM180930 and HM180931) deposited by Kim at al. . The three sequences were excluded following a cross-BLAST analysis against the GenBank repository of all Turbo cornutus COI sequences from which identity values lower than 78% were highlighted for each of them. Thus, the final allocation of the products to the species Rapana venosa was achieved through distance analysis conducted on the target 16S rRNA , which had been previously successfully applied in a comprehensive phylogeny study on Rapaninae, Muricidae taxonomical group in association with COI, 18S rRNA, 12S rRNA . In the case of Alepocephalus sp., given the low resolution of the elective target for species-specific identification revealed during BLAST analysis, the analysis was not conclusive due to the lack of deposited reference sequences for the additional target gene provided by the operative protocol (cytb). This evidence underlined a major limitation of the DNA barcoding technique and the need to continuously update reference databases. The need for continuous updating of the comprehensive DNA barcode reference libraries has been extensively emphasized in a study conducted by Weigand and colleagues on the Barcode of Life Data Systems (BOLD) and NCBI GenBank databases. From this study a clear lack of homogeneity in the deposit of sequence records among taxonomic groups emerged among geographic regions. The authors also highlighted that the presence of a large proportion of species (up to 50%) in several taxonomic groups is only represented by private data with obvious implications on the actual possibility of their use. The authors, therefore, emphasize the need for a coordinated and systematic action of database improvement to close the information gap and to maximize phylogenetic representativeness, thereby yielding to the collection of reference barcodes of representative species from missing orders, families and genera. 3.3. Assessment of Products Labeling Compliance and Mislabeling Rate 3.3.1. Labeling Compliance to Regulation (EU) No. 1379/2013 It is first appropriate to emphasize that according to Regulation (EU) No. 1379/2013 , the labeling obligations stated in Article 35 do not apply to processed products which, except for marinated and smoked products, do not fall within the scope of the provision. For other products, the application of the regulatory requirements is exclusively subject to the FBO's will, although strongly advocated by the EU Parliament to promote informed consumer choice at the time of purchase . Voluntary extensions of the regulation's requirements for the labeling of out-of-scope products have been documented in numerous studies of species identification in variously processed seafood . This considered the labels' analysis confirmed for all the products within the scope of Regulation (EU) No. 1379/2013 (N = 160), the proper application of Article 35, indicating commercial designation and species scientific name, an indication of production method, catching area, and thawing process. The only exception was represented by the fishing method, which was not declared in 6.8% of the products (11/160) (Table S1). This confirmed the substantial labeling compliance of products sold at large retail, as already highlighted by Tinacci et al. . Interestingly, the analysis of the data also revealed the voluntary application of the Regulation requirements in the remaining products (N = 39), consisting of breaded precooked (N = 26) and canned (N = 13) products. In all these products the commercial designation and scientific name, origin and catching area was reported while the fishing method was only reported in 38.46% (15/39) of the products. Notwithstanding, since any information included on the label, whether mandatory or voluntarily introduced, is subject to transparency and authenticity obligations according to Regulation (EU) No. 1169/2011 (article 7) , all products under study were included in the calculation of the mislabeling rate overall and for the individual product categories (F, MC, MB, MG and C) discussed below. 3.3.2. Mislabeling Rate and Products Origin In the comparison of the molecular results with the labeled scientific names, the presence of 22 substitutions out of 199 products, corresponding to an overall mislabeling rate of 11%, was found All substitution incidents are presented in Table 2. The overall mislabeling rate appears to be lower than in the pilot study , which stood at around 17%. The highest percentage was found in WF (14%), followed by MB (12.5%) and MC (10%). The overall percentage and percentages per taxonomical class fall within the substitution range (4-14%) identified by Luque and Donlan in a meta-analysis study conducted on scientific data and publications produced up to 2017. In particular, the WF mislabeling rate appears next to the percentage (12.9%) highlighted by Minoudi et al. in a market survey conducted on the Greek market and mostly directed toward whitefish species. In this regard, we have to highlight that in 2015, the frequency and impact of fraud perpetrated on this category prompted the EU Commission to organize a coordinated control program across all member states to assess the extent of mislabeling in the fishery sector with a specific focus on the whitefish market . Although not internationally harmonized, the definition of species substitution is broadly described as the intentional deception of a food product for economic gain or to conceal other illegal actions such as illegal, unreported and unregulated fishing achieved through the misrepresentation of food products or alteration of the associated documentation . Conversely, in addition to fraudulent incidents, the existence of involuntary substitutions due to the lack of adequate training in morphological species identification of fishermen and operators at the first sale level should also be emphasized . Both incidents result in damage to both consumers and food business operators and underline the fishery sector's vulnerability and the need to acquire in-depth knowledge of seafood chain traceability systems and individual business practices . Within WF substitution, incidents between Alaska pollock (G. chalcogrammus) and species belonging to Merluccius sp. of medium commercial value were frequently observed, followed by substitutions of Atlantic cod (G. morhua) with Saithe (Pollachius virens) and substitutions of hake species (Merluccius productus and Merluccius hubbsi) of medium commercial value replaced with Gadidae species of lower commercial value (Micromesistius australis). The substitutions observed in the study are all widely described and attributable to fraudulent incidents for economic gain . Similar evidence within the Gadidae and Merluccidae families was found by Minoudi et al. , wherein they identified misidentification during fishing and plausible fraudulent actions perpetrated at the product distribution level as possible causes of substitution. The study of the natural geographical distribution of substitute and replaced species could provide useful elements to potentially determine the origin of the fraud, as most potentially perpetrated at the first sale or at an intermediate level during the products processing . In the present study, the analysis of the labeled origins revealed a mostly non-Mediterranean supply mainly oriented to products of Atlantic or Pacific origin. Within WF, the fishing areas most frequently highlighted at purchase were Northeast/Northwest Pacific (FAO 61/67), corresponding to the distribution area of G. chalcogrammus. Less frequent but nonetheless relevant were Northeast/Northwest Atlantic (FAO 21/27) and Southeast Atlantic (FAO 41), corresponding to the distribution areas of Atlantic cod (G. morhua), Saithe (P. virens), Baird's smooth-head (Alepocephalus bairdii) and a few hake species (M. hubbsi, Macruronus sp.) all appreciably represented in large-scale distribution. Product of Mediterranean origin and, in particular, Black Sea origin (FAO 37.4) was exclusively found in products labeled as Merlangius merlangus (whiting), a fish species belonging to the Gadidae family of local interest and of medium commercial value. Thus, from the observation of the results collected in Table 3, fraudulent substitution phenomena within WF possibly occurred both at fishing/first sale and during processing or packaging. In detail, fraudulent substitutions at the first-sale level are plausibly conceivable for replaced and substitute species that share the geographic area, as in the substitution of G. chalcogrammus with M. productus or G. morhua with P. virens. Conversely, fraudulent substitution phenomena occurring at a more advanced stage of the chain (processing, packaging and distribution) are speculated in cases where a geographically distant substitute species were highlighted, such as in the substitution of G. chalcogrammus with M. hubbsi and M. productus with M. hubbsi or M. australis. In terms of environmental sustainability, it is pertinent to emphasize that the perpetration of substitution between geographically distant products could also conceal the attempt to reallocate products belonging to illegal fishing . Within MC, Humboldt squid (D. gigas) was verified as the dominant species on the market, bringing the Southwest and Southeast Pacific (FAO 81/87) into a prominent position among exporting areas. The species was also the most frequently substituted species in the pilot study and in the studies concerning mislabeling of cephalopod products on the Chinese and EU retail market . In fact, D. gigas has been thoroughly described as one of the elective substitute species, especially due to its high availability and low commercial value, which render the species an appealing candidate in the perpetration of deceptive frauds for economic gain . In MB products, the analysis of products origin highlighted a market orientation towards imported products belonging to North Atlantic (FAO 27) or South Pacific (FAO 81/87), specifically represented by the Mediterranean mussel (M. galloprovincialis), Chilean mussel (M. chilensis) and New Zealand green-lipped mussel (Perna canaliculus). The only two substitutions encountered consisted of the replacement of M. galloprovincialis with M. chilensis. Similar substitutions are described in products imported from Chile and have been attributed to unintentional accidents related to the coexistence of the two species in fishing and aquaculture areas along Chilean coasts . However, this observation does not apply to one of the two mislabeled products in the study (MB13), a canned marinated product, for which the clear North Atlantic (FAO 27) origin of the product was declared on the label and a fraudulent action is clearly hypothesized and collocable at the product processing stage. Lastly, in terms of C-class from the data collected at purchase, the market supply appeared directed to aquaculture products (P. vannamei) of Asian origin (Vietnam, Bangladesh, India) except for the sporadic presence of P. borealis of the Northwestern Atlantic origin and Metapenaeus sp./Penaeus sp. products from the Indian Ocean. Thus, with respect to the substitutions highlighted, given the considerations outlined above concerning the difficulties encountered in the morphological species identification and given the high degree of overlap of the geographical distribution areas of the substituted and substituted species (Sealifebase.org), it is plausible to affirm the occurrence of involuntary substitution phenomena for these products. Nevertheless, this aspect further underlines the need to promote and improve the operator's awareness and skills for species recognition, also implementing molecular identification systems that can be used at processing plants . Improved traceability tools represent, in fact, an essential support for FBOs who, in any case, hold the responsibility of verifying the identity of their products for consumer protection in accordance with EU legislation . In this regard, rapid identification methods have been promoted and developed in recent years for the most traded species, potentially applicable to self-monitoring by various operators, especially at the distribution level . 3.3.3. Adherence of Labeled Designation to Official Designations Accepted in the National Territory Table 3 presents the results of the comparison between the scientific names found in association with the products and the official list of accepted trade names in force during the sampling period , and the updated list . The comparison of the labeled scientific names and those listed in the ministerial ordinance, including the accepted designations, clearly highlighted the ineffectiveness of the list in force at the time of sampling in describing the basket of species present on the market. This aspect is extensively investigated and described by Tinacci et al. in a study aimed at assessing the validity and accuracy of the new official Bulgarian list of seafood trade names in compliance with EU requirements. The authors, in this regard, stated that there was a clear contradiction between the official records listed in the ordinance and the market needs. The authors particularly emphasized the ineffectiveness of the list in describing imported products that are widely available on the Bulgarian market and especially at the large retail level, which is even more evident in the newly promulgated list, in which some previously included scientific names of relevance, such as G. morhua or Pollachius virens, have disappeared. Therefore, an urgent need for further revision and expansion of the list of official denominations is advisable and pivotal to meet the obligation of periodic updating imposed by Regulation (EU) No. 1379/2013 (Article 37) and to provide FBOs with an effective tool to guarantee consumer rights on informed choice. In the revision of the list, as pointed out by Tinacci et al. , Tinacci et al. and in the present study, all taxonomical classes should be considered since, albeit to a lesser extent, besides fish, both mollusks and crustaceans are variously represented on the national market. An example of exponential evolution and expansion of the list of official denominations in relation to market needs is presented by Tinacci et al. for the Italian context. In a comparative retrospective analysis of the lists promulgated for the marketing of seafood products on the national territory from 2002 to 2017, the authors highlighted the market inputs at the origin of the evolution of the list, which were driven both by the demand of the average Italian consumer and by the ethnic groups and migrant populations permanently present on the national territory. Finally, regarding taxonomical validity, three invalid names, referring to an obsolete classification were highlighted (Table 3). It is, admittedly, true that updating scientific designations is an extremely challenging issue, given the continuous advancement in fish and seafood phylogeny research . In this regard, therefore, in conjunction with the revision of the list in which, as highlighted, these names are not yet included, the promotion of a specific FBO training would be appropriate to monitor and encourage the replacement of obsolete names and update the labeling of new product batches. 4. Conclusions The overall mislabeling rate of 11% and the analysis of the substitution incidents that emerged in the study highlighted the need to promote the implementation of DNA-based monitoring systems oriented towards supplier selection to be applied amongst FBOs at various levels of the production chain (processing and retail). This could be reduce involuntary substitutions and prevent deceptive practices to protect both seafood supply chain and consumers' rights. In this light, an integrated approach for species-specific polymorphisms analysis, through the association of different DNA analytical methods, may represent a useful and effective strategy for univocal seafood identification. In addition, this study confirms how the Bulgarian fish market, with reference to large-scale retail sales, still appears to be targeting a limited, albeit expanding, number of species. This aspect is stressed by the apparent inadequacy of the list of official names currently in force in the territory to describe the variety of products on sale. Therefore, as pointed out in a previously published study by the authors , a further update and expansion of the carnet of official commercial designations authorized in the territory are required. The data obtained from this survey could constitute inputs for implementing a monitoring plan promoted by governmental agencies in agreement with seafood stakeholders (wholesalers and sellers) for seafood authentication, contributing to the transparency of the seafood market at the national level. Supplementary Materials The following supporting information can be downloaded at: Figure S1: Distance-based dendrogram inferred with Neighbor-joining method on Kimura 2 parameter model computed in MEGA 11 involving 67 nucleotide sequences which include a selection of 56 reference sequences belonging to species of the genus Merluccius sp and 11 sequences produced in the study (F15, F16, F17, F46, F61, F63, F66, F79, F81, F82 and F95) highlighted with the symbol (*). All positions containing gaps and missing data were eliminated. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown above the branches, only bootstrap values higher than 70% are shown in the figure; Figure S2: Distance-based dendrogram inferred with Neighbor-joining method on Kimura 2 parameter model computed in MEGA 11 involving 37 nucleotide sequences which include 32 reference sequences belonging to R. bezoar, R. rapiformis, R. venosa, T. cornutus and five sequences produced in the study (MG 1, MG2, MG3, MG4 and MG5) highlighted with the symbol (*). The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown above the branches, only bootstrap values higher than 70% are shown in the figure; Table S1: Labeling information of the products collected in the study; Table S2: Results of molecular analysis of the products. Mislabeled products are highlighted in grey; Table S3: Estimates of interspecies divergence rate over sequence pairs between groups conducted with K2P model on a selection of 32 reference sequences deposited for R. venosa (N = 10); R. bezoar (N = 10); R. rapiformis (N = 7); T. cornutus (N = 5). Click here for additional data file. Author Contributions Conceptualization, L.T., D.S. and A.A.; data curation, M.S., G.Z. and R.K.; formal analysis, L.T.; funding acquisition, D.S.; investigation, L.T., M.S., G.Z. and R.K.; methodology, L.T., D.S. and A.A.; project administration, D.S.; supervision, A.A.; writing--original draft, L.T.; writing--review & editing, A.A. All authors have read and agreed to the published version of the manuscript. Data Availability Statement Data is included in the manuscript and available in supplementary tables attached to the text; further data is also available at the authors on request. Conflicts of Interest The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. Figure 1 The number of seafood products collected during the sampling period at retail outlets per species declared on the label within each seafood taxonomical class Distribution of products types classified according to Regulation (EC) No. 852/2004 . (B) distribution of product types classified according to the type of presentation for sale (without previous preparation, post evisceration, post cutting). Figure 3 Results of PCR-RFLP analysis on DNA samples extracted from the 13 MB Mytilus sp. products. Electrophoretic run shows post Aci-I digestion outcomes on PCR products of PAP target according to Giusti et al. protocol. The length (bp) of the fragments is reported above the bands. L--ladder. foods-12-01070-t001_Table 1 Table 1 Molecular targets selected in this study, primers pairs and amplification protocol applied. WF--whitefish; MB--bivalve mollusks; MC = cephalopod mollusks; MG--gastropod mollusks; C--crustaceans. Target Length Taxonomical Class Reference Primer Sequences Amplification Program COI (Elective) 655-658 bp WF Handy et al. FISH_COILBC TCAACYAATCAYAAAGATATYGGCAC FISH_COIHBC ACTTCYGGGTGRCCRAARAATCA 40 cycles: Denaturation 95 degC/30 s Annealing 55 degC/30 s Extension 72 degC/30 s MB Mikkelsen et al. COIF-ALT ACAAATCAYAARGAYATYGG COIR-ALT TTCAGGRTGNCCRAARAAYCA 40 cycles: Denaturation 95 degC/30 s Annealing 47 degC/35 s Extension 72 degC/20 s MC MG C Folmer et al. LCO1490 GGTCAACAAATCATAAAGATATTGG HC02198 TAAACTTCAGGGTGACCAAAAAATCA 40 cycles: Denaturation 95 degC/30 s Annealing 51 degC/30 s Extension 72 degC/30 s cytb (additional) 1150 bp WF Sevilla et al. GLUFISH-F AACCACCGTTGTTATTCAACTACAA THR-Fish-R ACCTCCGATCTTCGGATTACAAGACC 40 cycles: Denaturation 95 degC/30 s Annealing 57 degC/30 s Extension 72 degC/45 s 16S rRNA (additional) ~550 bp WF MB MC MG C Palumbi 16Sar_L CGCCTGTTTATCAAAAACAT 16Sbr_H CCGGTCTGAACTCAGATCACGT 40 cycles: Denaturation 95 degC/30 s Annealing 57 degC/15 s Extension 72 degC/15 s PEPCK (additional) 595 bp C Tsang et al. PEPCK FOR2 GCAAGACCAACCTGGCCATGATGAC PEPCK REV3 CGGGYCTCCATGCTSAGCCARTG 40 cycles: Denaturation 95 degC/30 s Annealing 59 degC/30 s Extension 72 degC/35 s PAP RFLP protocol MB (Mytilus sp. only) Giusti et al. Me15m-F GTATACAAACCTGTGAAGACAAGT Me16m-R TGTTGTCTTAATAGGTTTGTAAGATG Aci-I enzymatic digestion 40 cycles: Denaturation 95 degC/30 s Annealing 58 degC/30 s Extension 72 degC/30 s Incubation 37 degC/30 min Inactivation 65 degC/20 min foods-12-01070-t002_Table 2 Table 2 List of mislabeled products. * Geographical distribution area verified from FAO databases (Sealife.org and FishBase.org). Taxonomical Class Substitution Rate Percentage Test Target Code Product Type, Description Declared Scientific Name (Real Distribution Area) * Molecular Identification (Real Distribution Area) * Whitefish (WF) N = 100 Substitution rate: 14% Barcoding COI WF3 Unprocessed frozen, fillets Gadus morhua (FAO 21-27) Pollachius virens (FAO 21-27) Barcoding COI WF6 Unprocessed frozen, fillets Gadus morhua (FAO 21-27) Pollachius virens Barcoding COI, cytb WF15 Unprocessed frozen, w/o head and eviscerated Theragra chalcogramma (FAO 61-67) Merluccius productus (FAO 67) Barcoding COI, cytb WF16 Unprocessed frozen, w/o head and eviscerated Theragra chalcogramma (FAO 61-67) Merluccius productus (FAO 67) Barcoding COI, cytb WF17 Unprocessed frozen, w/o head and eviscerated Theragra chalcogramma (FAO 61-67) Merluccius productus (FAO 67) Barcoding COI WF19 Unprocessed, frozen, fillets Theragra chalcogramma (FAO 61-67) Merluccius hubbsi (FAO 41) Barcoding COI WF21 Unprocessed, frozen, fillets Theragra chalcogramma (FAO 61-67) Merluccius hubbsi (FAO 41) Barcoding COI WF25 Unprocessed frozen, w/o head and eviscerated Macruronus magellanicus (FAO 41-87) Merluccius hubbsi (FAO 41) Barcoding COI WF36 Unprocessed, frozen, fillets Gadus morhua (FAO 21-27) Pollachius virens (FAO 21-27) Barcoding COI WF39 Unprocessed frozen, w/o head and eviscerated Merluccius hubbsi (FAO 41) Micromesistius australis (FAO 81) Barcoding COI WF45 Unprocessed frozen, w/o head and eviscerated Merluccius productus (FAO 67) Micromesistius australis (FAO 81) Barcoding COI WF47 Unprocessed, frozen, fillets Theragra chalcogramma (FAO 61-67) Merluccius hubbsi (FAO 41) Barcoding COI, cytb WF79 Processed frozen, breaded fillets Theragra chalcogramma (FAO 61-67) Merluccius gayi (FAO 87) Cephalopod Mollusks (MC) N = 40 Substitution rate: 10% Barcoding COI MC6 Unprocessed frozen, whole peeled and eviscerated Ommastrephes bartramii (Cosmopolitan) Dosidicus gigas (FAO 77-87) Barcoding COI MC21 Unprocessed frozen, sliced Todarodes pacificus (FAO 61) Dosidicus gigas (FAO 77-87) Barcoding COI MC22 Unprocessed frozen, whole peeled and eviscerated Ommastrephes bartramii (Cosmopolitan) Dosidicus gigas (FAO 77-87) Barcoding COI MC40 Unprocessed frozen, sliced Nototodarus sloanii (FAO 81) Todaropsis eblanae (cosmopolitan, no FAO 77-87) Bivalve Mollusks (MB) N = 16 Substitution rate: 12.5% RFLP PAP MB13 Processed canned Mytilus galloprovincialis (FAO 27-37-41-87) Mytilus chilensis (FAO 41-87) RFLP PAP MB15 Processed canned Mytilus galloprovincialis (FAO 27-37-41-87) Mytilus chilensis (FAO 41-87) Crustaceans (C) N = 38 Substitution rate: 7.9% Barcoding COI, PEPCK C10 Unprocessed frozen, w/o head peeled Solenocera Melantho (FAO 61-71-81) Heterocarpus sp. (FAO51-57-61-71) Barcoding COI C31 Processed frozen, w/o head peeled precooked Metapenaeus sp (FAO 51-57) Fenneropenaeus indicus (FAO 51-57) Barcoding COI C36 Unprocessed frozen, whole Solenocera Melantho (FAO 61-71-81) Plesionika quasigrandis (FAO 51-57) foods-12-01070-t003_Table 3 Table 3 Comparison between the scientific names found on the labels, the scientific names included in Ministerial, the official list of designations in force during the sampling period and the updated list currently in force. Species Labeled Ordinance No. 4 of 13.01.2006 Ordinance No. 13 30.11.2021 Theragra chalcogramma no no Gadus morhua yes no Merlangius merlangus euxinus yes yes Pollachius virens yes no Merluccius sp. no no Merluccius hubbsi no no Merluccius productus no no Merluccius australis no no Micromesistius australis no no Macruronus magellanicus no no Macruronus novaezelandiae no no Macruronus australis no no Alepocephalus bairdii no no Litopenaeus vannamei no no Penaeus monodon no no Metapenaeus affinis no no Metapenaeus sp. no no Solenocera melantho no no Pandalus borealis yes no Nephrops norvegicus yes no Dosidicus gigas no no Ommastrephes bartramii no no Illex argentinus no no Todarodes pacificus no no Nototodarus sloanii no no Loligo gahi/patagonica no no Octopus vulgaris no no Sepia sp. no (Sepia officinalis) no Mytilus chilensis no no Mytilus galloprovincialis yes yes Mytilus sp. no no Perna canaliculus no no Rapana venosa no yes Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. FAO The State of World Fisheries and Aquaculture 2022 FAO Rome, Italy 2022 978-92-5-136364-5 2. European Commission. Directorate General for Maritime Affairs and Fisheries The EU Fish Market 2022nd ed. EU Publications Office Luxembourg City, Luxembourg 2022 3. Issifu I. Deffor E.W. Deyshappriya N.P.R. Dahmouni I. Sumaila U.R. Drivers of Seafood Consumption at Different Geographical Scales J. Sustain. Res. 2022 4 e220012 10.20900/jsr20220012 4. MZHG Agraren Doklad 2020 Available online: (accessed on 19 January 2023) 5. Eurobarometer 2271/SP515 EU Consumer Habits Regarding Fishery and Aquaculture Products Available online: (accessed on 19 January 2023) 6. Stancheva M. A Survey on Fish Consumption in Bulgaria Proceedings of the Poljoprivreda--Zbornitsi COBISS.SR-ID 270329868 Belgrade-Zemun, Serbia 22 November 2018 Faculty of Agriculture Belgrade-Zemun, Serbia 7. Cantillo J. Martin J.C. Roman C. Determinants of Fishery and Aquaculture Products Consumption at Home in the EU28 Food Qual. Prefer. 2021 88 104085 10.1016/j.foodqual.2020.104085 8. Todorov A. Fish and Seafood Market Brief--Bulgaria US Department of Agriculture (USDA) Washington, DC, USA 2020 9. Tinacci L. Stratev D. Zhelyazkov G. Kyuchukova R. Strateva M. Nucera D. Armani A. Nationwide Survey of the Bulgarian Market Highlights the Need to Update the Official Seafood List Based on Trade Inputs Food Control 2020 112 107131 10.1016/j.foodcont.2020.107131 10. Everstine K. Supply Chain Complexity and Economically Motivated Adulteration Food Protection and Security Elsevier Amsterdam, The Netherlands 2017 1 14 978-1-78242-251-8 11. Lawrence S. Elliott C. Huisman W. Dean M. van Ruth S. The 11 Sins of Seafood: Assessing a Decade of Food Fraud Reports in the Global Supply Chain Compr. Rev. Food Sci. Food Saf. 2022 21 3746 3769 10.1111/1541-4337.12998 35808861 12. Giusti A. Malloggi C. Tinacci L. Nucera D. Armani A. Mislabeling in Seafood Products Sold on the Italian Market: A Systematic Review and Meta-Analysis Food Control 2023 145 109395 10.1016/j.foodcont.2022.109395 13. Luque G.M. Donlan C.J. The Characterization of Seafood Mislabeling: A Global Meta-Analysis Biol. Conserv. 2019 236 556 570 10.1016/j.biocon.2019.04.006 14. Regulation EC No. 178/2002 Regulation (EC) No 178/2002 of the European Parliament and of the Council of 28 January 2002 Laying down the General Principles and Requirements of Food Law, Establishing the European Food Safety Authority and Laying down Procedures in Matters of Food Safety Off. J. Eur. Communities 2002 L31 1 24 15. Regulation EU No. 1169/2011 Regulation (EU) No 1169/2011 of the European Parliament and of the Council of 25 October 2011 on the Provision of Food Information to Consumers, Amending Regulations (EC) No 1924/2006 and (EC) No 1925/2006 of the European Parliament and of the Council, and Repealing Commission Directive 87/250/EEC, Council Directive 90/496/EEC, Commission Directive 1999/10/EC, Directive 2000/13/EC of the European Parliament and of the Council, Commission Directives 2002/67/EC and 2008/5/EC and Commission Regulation (EC) No 608/2004 Text with EEA Relevance Off. J. Eur. Union 2011 L304 18 63 16. Regulation EU No. 1379/2013 Regulation (EU) No 1379/2013 of the European Parliament and of the Council of 11 December 2013 on the Common Organisation of the Markets in Fishery and Aquaculture Products, Amending Council Regulations (EC) No 1184/2006 and (EC) No 1224/2009 and Repealing Council Regulation (EC) No 104/2000 Off. J. Eur. Union 2013 L354 1 21 17. Regulation (EU) 2017/625 Regulation (EU) 2017/625 of the European Parliament and of the Council of 15 March 2017 on Official Controls and Other Official Activities Performed to Ensure the Application of Food and Feed Law, Rules on Animal Health and Welfare, Plant Health and Plant Protection Products Off. J. Eur. Union 2017 L 95 1 142 18. Fernandes T.J.R. Amaral J.S. Mafra I. DNA Barcode Markers Applied to Seafood Authentication: An Updated Review Crit. Rev. Food Sci. Nutr. 2021 61 3904 3935 10.1080/10408398.2020.1811200 32838560 19. Silva A.J. Hellberg R.S. Hanner R.H. Seafood Fraud Food Fraud Elsevier Amsterdam, The Netherlands 2021 109 137 978-0-12-817242-1 20. Silva A.J. Hellberg R.S. DNA-Based Techniques for Seafood Species Authentication Advances in Food and Nutrition Research Elsevier Amsterdam, The Netherlands 2021 Volume 95 207 255 978-0-12-821520-3 21. Giusti A. Malloggi C. Tosi F. Boldini P. Larrain Barth M.A. Araneda C. Arcangeli G. Armani A. Mislabeling Assessment and Species Identification by PCR-RFLP of Mussel-Based Products (Mytilus Spp.) Sold on the Italian Market Food Control 2022 134 108692 10.1016/j.foodcont.2021.108692 22. Tinacci L. Stratev D. Vashin I. Chiavaccini I. Susini F. Guidi A. Armani A. Seafood Labelling Compliance with European Legislation and Species Identification by DNA Barcoding: A First Survey on the Bulgarian Market Food Control 2018 90 180 188 10.1016/j.foodcont.2018.03.007 23. Tinacci L. Stratev D. Strateva M. Zhelyazkov G. Kyuchukova R. Armani A. New Official Bulgarian List of Seafood Trade Names: Coping with EU Labelling Requirements and Market Trends to Enhance Consumers' Informed Choice J. Consum. Prot. Food Saf. 2022 17 395 406 10.1007/s00003-022-01397-7 24. Brechon A.L. Hanner R. Mariani S. A Systematic Analysis across North Atlantic Countries Unveils Subtleties in Cod Product Labelling Mar. Policy 2016 69 124 133 10.1016/j.marpol.2016.04.014 25. Helyar S.J. Lloyd H.D. de Bruyn M. Leake J. Bennett N. Carvalho G.R. Fish Product Mislabelling: Failings of Traceability in the Production Chain and Implications for Illegal, Unreported and Unregulated (IUU) Fishing PLoS ONE 2014 9 e98691 10.1371/journal.pone.0098691 24921655 26. Miller D. Jessel A. Mariani S. Seafood Mislabelling: Comparisons of Two Western European Case Studies Assist in Defining Influencing Factors, Mechanisms and Motives: Cod Fraud: Mechanisms and Motives Fish Fish. 2012 13 345 358 10.1111/j.1467-2979.2011.00426.x 27. Shehata H.R. Bourque D. Steinke D. Chen S. Hanner R. Survey of Mislabelling across Finfish Supply Chain Reveals Mislabelling Both Outside and within Canada Food Res. Int. 2019 121 723 729 10.1016/j.foodres.2018.12.047 31108801 28. Janssen R. Knudsen S. Todorova V. Hosgor A.G. Managing Rapana in the Black Sea: Stakeholder Workshops on Both Sides Ocean Coast. Manag. 2014 87 75 87 10.1016/j.ocecoaman.2013.10.015 29. Regulation EC 852/2004 Regulation (EC) No 852/2004 of the European Parliament and of the Council of 29 April 2004 on the Hygiene of Foodstuffs Off. J. Eur. Union 2004 L139 1 54 30. Armani A. Tinacci L. Xiong X. Titarenko E. Guidi A. Castigliego L. Development of a Simple and Cost-Effective Bead-Milling Method for DNA Extraction from Fish Muscles Food Anal. Methods 2014 7 946 955 10.1007/s12161-014-9792-z 31. Tinacci L. Guidi A. Toto A. Guardone L. Giusti A. D'Amico P. Armani A. DNA Barcoding for the Verification of Supplier's Compliance in the Seafood Chain: How the Lab Can Support Companies in Ensuring Traceability Ital. J. Food Saf. 2018 7 6894 10.4081/ijfs.2018.6894 30046552 32. Handy S.M. Deeds J.R. Ivanova N.V. Hebert P.D.N. Hanner R.H. Ormos A. Weigt L.A. Moore M.M. Yancy H.F. A Single-Laboratory Validated Method for the Generation of DNA Barcodes for the Identification of Fish for Regulatory Compliance J. AOAC Int. 2011 94 201 210 10.1093/jaoac/94.1.201 21391497 33. Mikkelsen P.M. Bieler R. Kappner I. Rawlings T.A. Phylogeny of Veneroidea (Mollusca: Bivalvia) Based on Morphology and Molecules Zool. J. Linn. Soc. 2006 148 439 521 10.1111/j.1096-3642.2006.00262.x 34. Folmer O. Black M. Hoeh W. Lutz R. Vrijenhoek R. DNA Primers for Amplification of Mitochondrial Cytochrome c Oxidase Subunit I from Diverse Metazoan Invertebrates Mol. Mar. Biol. Biotechnol. 1994 3 294 299 7881515 35. Sevilla R.G. Diez A. Noren M. Mouchel O. Jerome M. Verrez-Bagnis V. Van Pelt H. Favre-Krey L. Krey G. Consortium T.F. Primers and Polymerase Chain Reaction Conditions for DNA Barcoding Teleost Fish Based on the Mitochondrial Cytochrome b and Nuclear Rhodopsin Genes Mol. Ecol. Notes 2007 7 730 734 10.1111/j.1471-8286.2007.01863.x 36. Palumbi S. Nucleic Acids II: The Polymerase Chain Reaction Molecular Systematics Sinauer Associates, Inc. Sunderland, MA, USA 1996 205 247 978-0-87893-282-5 37. Tsang L.M. Ma K.Y. Ahyong S.T. Chan T.-Y. Chu K.H. Phylogeny of Decapoda Using Two Nuclear Protein-Coding Genes: Origin and Evolution of the Reptantia Mol. Phylogenet. Evol. 2008 48 359 368 10.1016/j.ympev.2008.04.009 18501643 38. Hall T. BioEdit: A User-Friendly Biological Sequence Alignment Editor and Analysis Program for Windows 95/98/NT Nucleic Acids Symp. Ser. 1999 41 95 98 39. Pardo M.A. Jimenez E. Vidarsson J.R. Olafsson K. Olafsdottir G. Danielsdottir A.K. Perez-Villareal B. DNA Barcoding Revealing Mislabeling of Seafood in European Mass Caterings Food Control 2018 92 7 16 10.1016/j.foodcont.2018.04.044 40. Armani A. Tinacci L. Xiong X. Castigliego L. Gianfaldoni D. Guidi A. Fish Species Identification in Canned Pet Food by BLAST and Forensically Informative Nucleotide Sequencing (FINS) Analysis of Short Fragments of the Mitochondrial 16s Ribosomal RNA Gene (16S RRNA) Food Control 2015 50 821 830 10.1016/j.foodcont.2014.10.018 41. Armani A. Tinacci L. Lorenzetti R. Benvenuti A. Susini F. Gasperetti L. Ricci E. Guarducci M. Guidi A. Is Raw Better? A Multiple DNA Barcoding Approach (Full and Mini) Based on Mitochondrial and Nuclear Markers Reveals Low Rates of Misdescription in Sushi Products Sold on the Italian Market Food Control 2017 79 126 133 10.1016/j.foodcont.2017.03.030 42. Saitou N. Nei M. The Neighbor-Joining Method: A New Method for Reconstructing Phylogenetic Trees Mol. Biol. Evol. 1987 4 406 425 10.1093/oxfordjournals.molbev.a040454 3447015 43. Kimura M. A Simple Method for Estimating Evolutionary Rates of Base Substitutions through Comparative Studies of Nucleotide Sequences J. Mol. Evol. 1980 16 111 120 10.1007/BF01731581 7463489 44. Tamura K. Stecher G. Kumar S. MEGA11: Molecular Evolutionary Genetics Analysis Version 11 Mol. Biol. Evol. 2021 38 3022 3027 10.1093/molbev/msab120 33892491 45. Ordinance No. 4 of 13.01.2006 Regulation no. 4 of January 13, 2006 on the conditions and procedures for the first sale of fish and other aquatic organisms Off. SG 2006 73 80 46. Ordinance No. 13 of 30.11.2021 Ordinance No 13 of 30 November 2021 of the Ministry of Agriculture and Forestry on the terms and conditions for the first sale of fishery products Off. SG 2021 22 Available online: (accessed on 20 October 2022) 47. Keskin C. Ulman A. Zylich K. Raykov V. Daskalov G.M. Pauly D. Zeller D. The Marine Fisheries in Bulgaria's Exclusive Economic Zone, 1950-2013 Front. Mar. Sci. 2017 4 53 10.3389/fmars.2017.00053 48. Vindigni G. Pulvirenti A. Alaimo S. Monaco C. Spina D. Peri I. Bioinformatics Approach to Mitigate Mislabeling in EU Seafood Market and Protect Consumer Health Int. J. Environ. Res. Public. Health 2021 18 7497 10.3390/ijerph18147497 34299949 49. Cawthorn D.-M. Duncan J. Kastern C. Francis J. Hoffman L.C. Fish Species Substitution and Misnaming in South Africa: An Economic, Safety and Sustainability Conundrum Revisited Food Chem. 2015 185 165 181 10.1016/j.foodchem.2015.03.113 25952855 50. Haldar C. Nath S. DNA Barcoding for Fish Species Identification: Current Status and Future Prospective Int. J. Fauna Biol. Stud. 2020 7 72 75 51. Blanco-Fernandez C. Garcia-Vazquez E. Machado-Schiaffino G. Seventeen Years Analysing Mislabelling from DNA Barcodes: Towards Hake Sustainability Food Control 2021 123 107723 10.1016/j.foodcont.2020.107723 52. Ramirez J.L. Simbine L. Marques C.G. Zelada-Mazmela E. Reyes-Flores L.E. Lopez A.S. Gusmao J. Tavares C. Galetti P.M. Freitas P.D. DNA Barcoding of Penaeidae (Decapoda; Crustacea): Non-Distance-Based Species Delimitation of the Most Economically Important Shrimp Family Diversity 2021 13 460 10.3390/d13100460 53. Shen Y. Kang J. Chen W. He S. DNA Barcoding for the Identification of Common Economic Aquatic Products in Central China and Its Application for the Supervision of the Market Trade Food Control 2016 61 79 91 10.1016/j.foodcont.2015.08.038 54. Rajkumar G. Saravana Bhavan P. Udayasuriyan R. Vadivalagan C. Molecular Identification of Shrimp Species, Penaeus Semisulcatus, Metapenaeus Dobsoni, Metapenaeus Brevicornis, Fenneropenaeus Indicus, Parapenaeopsis Stylifera and Solenocera Crassicornis Inhabiting in the Coromandel Coast (Tamil Nadu, India) Using MT-COI Gene Int. J. Fish. Aquat. Stud. 2015 2 96 106 55. Hurzaid A. Chan T. Mohd Nor S.A. Muchlisin Z.A. Chen W. Molecular Phylogeny and Diversity of Penaeid Shrimps (Crustacea: Decapoda) from South-East Asian Waters Zool. Scr. 2020 zsc.12428. 10.1111/zsc.12428 56. Armani A. Guardone L. Castigliego L. D'Amico P. Messina A. Malandra R. Gianfaldoni D. Guidi A. DNA and Mini-DNA Barcoding for the Identification of Porgies Species (Family Sparidae) of Commercial Interest on the International Market Food Control 2015 50 589 596 10.1016/j.foodcont.2014.09.025 57. Leray M. Knowlton N. Ho S.-L. Nguyen B.N. Machida R.J. GenBank Is a Reliable Resource for 21st Century Biodiversity Research Proc. Natl. Acad. Sci. USA 2019 116 22651 22656 10.1073/pnas.1911714116 31636175 58. Meiklejohn K.A. Damaso N. Robertson J.M. Assessment of BOLD and GenBank--Their Accuracy and Reliability for the Identification of Biological Materials PLoS ONE 2019 14 e0217084 10.1371/journal.pone.0217084 31216285 59. Carlini D.B. Kunkle L.K. Vecchione M. A Molecular Systematic Evaluation of the Squid Genus Illex (Cephalopoda: Ommastrephidae) in the North Atlantic Ocean and Mediterranean Sea Mol. Phylogenet. Evol. 2006 41 496 502 10.1016/j.ympev.2006.05.011 16782358 60. Kim D.-W. Yoo W.G. Park H.C. Yoo H.S. Kang D.W. Jin S.D. Min H.K. Paek W.K. Lim J. DNA Barcoding of Fish, Insects, and Shellfish in Korea Genomics Inform. 2012 10 206 10.5808/GI.2012.10.3.206 23166532 61. Claremont M. Vermeij G.J. Williams S.T. Reid D.G. Global Phylogeny and New Classification of the Rapaninae (Gastropoda: Muricidae), Dominant Molluscan Predators on Tropical Rocky Seashores Mol. Phylogenet. Evol. 2013 66 91 102 10.1016/j.ympev.2012.09.014 23026810 62. Weigand H. Beermann A.J. Ciampor F. Costa F.O. Csabai Z. Duarte S. Geiger M.F. Grabowski M. Rimet F. Rulik B. DNA Barcode Reference Libraries for the Monitoring of Aquatic Biota in Europe: Gap-Analysis and Recommendations for Future Work Sci. Total Environ. 2019 678 499 524 10.1016/j.scitotenv.2019.04.247 31077928 63. European Parliament Resolution No. 2016/2532 European Parliament Resolution of 12 May 2016 on Traceability of Fishery and Aquaculture Products in Restaurants and Retail (2016/2532(RSP)) Off. J. Eur. Union 2018 C76 1 3 64. Fernandes T.J.R. Costa J. Oliveira M.B.P.P. Mafra I. DNA Barcoding Coupled to HRM Analysis as a New and Simple Tool for the Authentication of Gadidae Fish Species Food Chem. 2017 230 49 57 10.1016/j.foodchem.2017.03.015 28407939 65. Tinacci L. Guardone L. Castro-Palomino Rubio J. Riina M.V. Stratev D. Guidi A. Armani A. Labelling Compliance and Species Identification of Herring Products Sold at Large Scale Retail Level within the Italian Market Food Control 2019 106 106707 10.1016/j.foodcont.2019.106707 66. Giusti A. Tinacci L. Sotelo C.G. Acutis P.L. Ielasi N. Armani A. Authentication of Ready-to-Eat Anchovy Products Sold on the Italian Market by BLAST Analysis of a Highly Informative Cytochrome b Gene Fragment Food Control 2019 97 50 57 10.1016/j.foodcont.2018.10.018 67. Mottola A. Piredda R. Catanese G. Lorusso L. Ciccarese G. Di Pinto A. Species Authentication of Canned Mackerel: Challenges in Molecular Identification and Potential Drivers of Mislabelling Food Control 2022 137 108880 10.1016/j.foodcont.2022.108880 68. Minoudi S. Karaiskou N. Avgeris M. Gkagkavouzis K. Tarantili P. Triantafyllidou D. Palilis L. Avramopoulou V. Tsikliras A. Barmperis K. Seafood Mislabeling in Greek Market Using DNA Barcoding Food Control 2020 113 107213 10.1016/j.foodcont.2020.107213 69. European Commission Fish Substitution 2015 Available online: (accessed on 19 January 2023) 70. Manning L. Soon J.M. Food Fraud Vulnerability Assessment: Reliable Data Sources and Effective Assessment Approaches Trends Food Sci. Technol. 2019 91 159 168 10.1016/j.tifs.2019.07.007 71. Robson K. Dean M. Haughey S. Elliott C. A Comprehensive Review of Food Fraud Terminologies and Food Fraud Mitigation Guides Food Control 2021 120 107516 10.1016/j.foodcont.2020.107516 72. Barendse J. Roel A. Longo C. Andriessen L. Webster L.M.I. Ogden R. Neat F. DNA Barcoding Validates Species Labelling of Certified Seafood Curr. Biol. 2019 29 R198 R199 10.1016/j.cub.2019.02.014 30889387 73. Christiansen H. Fournier N. Hellemans B. Volckaert F.A.M. Seafood Substitution and Mislabeling in Brussels' Restaurants and Canteens Food Control 2018 85 66 75 10.1016/j.foodcont.2017.09.005 74. Feldmann F. Ardura A. Blanco-Fernandez C. Garcia-Vazquez E. DNA Analysis Detects Different Mislabeling Trend by Country in European Cod Fillets Foods 2021 10 1515 10.3390/foods10071515 34209308 75. Fox M. Mitchell M. Dean M. Elliott C. Campbell K. The Seafood Supply Chain from a Fraudulent Perspective Food Secur. 2018 10 939 963 10.1007/s12571-018-0826-z 76. Fernandez S. Rodriguez-Muniz L.J. Molina J. Muniz-Rodriguez L. Jimenez J. Garcia-Vazquez E. Borrell Y.J. Lab Experience with Seafood Control at the Undergraduate Level: Cephalopods as a Case Study Biochem. Mol. Biol. Educ. 2020 48 236 246 10.1002/bmb.21332 31944524 77. Wen J. Tinacci L. Acutis P.L. Riina M.V. Xu Y. Zeng L. Ying X. Chen Z. Guardone L. Chen D. An Insight into the Chinese Traditional Seafood Market: Species Characterization of Cephalopod Products by DNA Barcoding and Phylogenetic Analysis Using COI and 16SrRNA Genes Food Control 2017 82 333 342 10.1016/j.foodcont.2017.07.011 78. Espineira M. Vieites J.M. Rapid Method for Controlling the Correct Labeling of Products Containing Common Octopus (Octopus vulgaris) and Main Substitute Species (Eledone cirrhosa and Dosidicus gigas) by Fast Real-Time PCR Food Chem. 2012 135 2439 2444 10.1016/j.foodchem.2012.07.056 22980826 79. Shehata H.R. Naaum A.M. Garduno R.A. Hanner R. DNA Barcoding as a Regulatory Tool for Seafood Authentication in Canada Food Control 2018 92 147 153 10.1016/j.foodcont.2018.04.045 80. Fernandes T.J.R. Silva C.R. Costa J. Oliveira M.B.P.P. Mafra I. High Resolution Melting Analysis of a COI Mini-Barcode as a New Approach for Penaeidae Shrimp Species Discrimination Food Control 2017 82 8 17 10.1016/j.foodcont.2017.06.016 81. Sharma L. Watts E. Singh P. High Resolution Real-Time PCR Melting Curve Assay for Identification of Top Five Penaeidae Shrimp Species LWT 2020 133 109983 10.1016/j.lwt.2020.109983 82. Wilwet L. JeyaShakila R. Sivaraman B. Nayak B.B. Kumar H.S. Jaiswar A.K. Ratrey V.P. Jeyasekaran G. In-House and on-Field Validation of the Multiplex PCR Assay Developed for Authentication of Three Commercially Important Shrimp Species LWT 2021 148 111701 10.1016/j.lwt.2021.111701 83. Tinacci L. Giusti A. Guardone L. Luisi E. Armani A. The New Italian Official List of Seafood Trade Names (Annex I of Ministerial Decree n. 19105 of September the 22nd, 2017): Strengths and Weaknesses in the Framework of the Current Complex Seafood Scenario Food Control 2019 96 68 75 10.1016/j.foodcont.2018.09.002 |
PMC10000582 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050740 cells-12-00740 Article The Effect of Acidic and Alkaline Seawater on the F-Actin-Dependent Ca2+ Signals Following Insemination of Immature Starfish Oocytes and Mature Eggs Limatola Nunzia Conceptualization Methodology Formal analysis Investigation Data curation Writing - original draft Writing - review & editing Visualization 1* Chun Jong Tai Conceptualization Formal analysis Investigation Resources Data curation Writing - original draft Writing - review & editing Visualization 2 Schneider Suzanne C. Resources 3 Schmitt Jean-Louis Resources 3 Lehn Jean-Marie Resources 3 Santella Luigia Conceptualization Methodology Formal analysis Investigation Resources Data curation Writing - original draft Writing - review & editing Visualization Supervision Project administration Funding acquisition 1* Llano Elena Academic Editor 1 Department of Research Infrastructures for Marine Biological Resources, Stazione Zoologica Anton Dohrn, 80121 Napoli, Italy 2 Department of Biology and Evolution of Marine Organisms, Stazione Zoologica Anton Dohrn, 80121 Napoli, Italy 3 Laboratory of Supramolecular Chemistry, Institut de Science et d'Ingenierie Supramoleculaires ISIS, Universite de Strasbourg, 8 Allee Gaspard Monge, 67000 Strasbourg, France * Correspondence: [email protected] (N.L.); [email protected] (L.S.); Tel.: +39-081-583-3229 (N.L.); +39-081-583-3289 (L.S.) 25 2 2023 3 2023 12 5 74013 1 2023 07 2 2023 23 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). In starfish, the addition of the hormone 1-methyladenine (1-MA) to immature oocytes (germinal vesicle, GV-stage) arrested at the prophase of the first meiotic division induces meiosis resumption (maturation), which makes the mature eggs able to respond to the sperm with a normal fertilization response. The optimal fertilizability achieved during the maturation process results from the exquisite structural reorganization of the actin cytoskeleton in the cortex and cytoplasm induced by the maturing hormone. In this report, we have investigated the influence of acidic and alkaline seawater on the structure of the cortical F-actin network of immature oocytes of the starfish (Astropecten aranciacus) and its dynamic changes upon insemination. The results have shown that the altered seawater pH strongly affected the sperm-induced Ca2+ response and the polyspermy rate. When immature starfish oocytes were stimulated with 1-MA in acidic or alkaline seawater, the maturation process displayed a strong dependency on pH in terms of the dynamic structural changes of the cortical F-actin. The resulting alteration of the actin cytoskeleton, in turn, affected the pattern of Ca2+ signals at fertilization and sperm penetration. maturation starfish pH acrosome reaction Ca2+ signaling jelly coat fertilization polyspermy Stazione Zoologica Anton Dohrn (SZN)Institutional fundsThis work was financially supported by institutional funds granted by the Stazione Zoologica Anton Dohrn (SZN). N.L. was supported by Institutional funds. pmc1. Introduction The landmark discovery of 1-methyladenine (1-MA) as the hormone inducing the resumption of the meiotic cycle (maturation) of immature starfish (GV-stage) oocytes arrested at the first prophase of the meiotic division has allowed profound in vitro studies on the structural and biochemical changes that immature oocytes must undergo to reach the state of optimum fertilizability following the maturation process . It has been reported that 1-MA activates a yet unknown receptor that releases the bg subunits of its coupled heterotrimeric G-protein from the a subunit upon hormonal stimulation. The interaction of bg subunits with effectors leads to an intracellular pH increase and activation of cyclin B-Cdk1, which is known as MPF (Maturation Promoting Factor) inducing GVBD (Germinal Vesicle Breakdown). The latter events are required for the spindle assembly at metaphase I . The 1-MA-induced morpho-functional changes in the oocyte are necessary to make the mature eggs competent to respond to the activating sperm with a normal fertilization response and subsequent embryonic development. The possibility of inducing meiotic maturation in vitro provides an advantage for starfish oocytes as an experimental model. Furthermore, at variance with sea urchin eggs that are spawned and fertilized in seawater after the second maturation division, starfish oocytes can be inseminated before, during, and after the course of the maturation process to study the nature of the altered physiological conditions leading to polyspermic fertilization . Previous studies have highlighted the importance of the structural dynamics of the cortical F-actin network following hormonal stimulation of immature oocytes in eliciting Ca2+ signals in the cytoplasm and nucleus and at different stages of the maturation process . Specifically, the two modes of Ca2+ increases taking place a few minutes after hormonal stimulation and during GVBD have underscored the contributions made by the state of the cortical actin cytoskeleton and by intermixing the nuclear and cytoplasmic components in producing mature eggs apt for normal fertilization . Indeed, immature oocytes whose surface topography and structural organization of the oocyte cortex are strikingly different from mature eggs experience multiple Ca2+ waves at fertilization due to polyspermy . In fact, numerous sperm are incorporated into the GV-stage oocytes, but they still remain localized in the cortical region without undergoing structural changes such as DNA decondensation and the formation of aster and pronucleus due to their cytoplasmic "immaturity" . It has been shown that the 1-MA-dependent maturation of the intracellular Ca2+ stores enables a higher Ca2+ response to inositol 1,4,5-trisphosphate (InsP3) in mature starfish eggs and the fertilizing sperm. The sensitization of the Ca2+ stores to InsP3 might be attributed to the structural changes in the endoplasmic reticulum during the maturation process of the oocytes ; however, it was then suggested that it was linked to the remodeling of the MPF-dependent actin cytoskeleton in the cortex and cytoplasm induced by the 1-MA . Notably, the F-actin-dependent morphological modifications of the cortex of maturing oocytes also include the shortening of microvilli (filled with actin filaments) on the plasma membrane of mature eggs . This microvillar morphology restructuring leads to changes in plasma membrane resting potential, which constitute an essential physiological condition for a normal electrophysiological and Ca2+ response upon insemination of mature eggs , in line with the idea of a cytoskeletal modulation of Ca2+ signals in the cells mediated by microvilli . Following insemination, a mature starfish egg experiences electrical and Ca2+ responses due to its interaction with the tip of the acrosomal process (approximately 20 mm long) formed on the sperm head when it comes into contact with the jelly coat (JC) surrounding the egg . The first detectable Ca2+ signal occurs in the form of a fast (1-to 3 sec) subitaneous Ca2+ increase in the periphery of the egg cortex (the cortical flash, CF), which is followed by a Ca2+ wave starting at the site of sperm-egg interaction and propagating to the opposite pole . The findings from the starfish mature eggs on the morpho-functional importance of the cortical actin filaments in inducing a normal Ca2+ response and monospermic penetration have been extended to Paracentrotus lividus sea urchin eggs . Recent studies have highlighted the essential roles played by the integrity of the vitelline layer tightly bound to microvilli and the F-actin-associated cortical granules and vesicles by showing that microvillar morphology changes and disruption, relocation, or fusion of the cortical vesicles all lead to abnormal fertilization responses . Confirmation of a crucial role of the structural morphology of microvilli and egg cortex in achieving a normal Ca2+ signal and monospermic fertilization in sea urchin eggs has been provided by showing that the alteration of the egg cortex induced by exposure to unfertilized P. lividus eggs to seawater containing low sodium or an acidic pH (pH 6.8) compromised the sperm-induced Ca2+ signals and embryo development . A more recent study on the effect of the incubation and fertilization of P. lividus eggs in acidic (pH 5.5) and alkaline seawater (pH 9) has highlighted the strict interdependence between seawater pH and the fertilization response due to the changes in the morpho-functional alterations of the F-actin of the cortex of unfertilized eggs. Specifically, the acidic and alkaline pH of the seawater evoked depolymerizing and polymerizing effects on cortical F-actin, respectively, and this, in turn, influenced the Ca2+ response at fertilization and sperm entry . Finally, the elongation of microvilli and F-actin spikes formation in the perivitelline space as well as translocation of actin fibers from the zygote surface , sperm incorporation, and cleavage have sanctioned the indispensable role of the egg cortical F-actin structure and its remodeling in the regulation of all the aspects of the fertilization process. The present contribution has analyzed the effect of the acidic and alkaline seawater pH on the response of immature oocytes of starfish to insemination. It is well known that the GV-stage (immature) oocytes respond to sperm addition by eliciting multiple Ca2+ signals due to polyspermic interactions . Polyspermy probably arises from the different organization of the oocyte's surface and the actin cytoskeleton of their cortex compared to mature eggs . Thus, our experimental design focused on how the acidic or alkaline conditions would affect the structural F-actin dynamics and polyspermy in immature oocytes as well as the process of meiotic maturation and the pattern of the fertilization response. 2. Materials and Methods 2.1. Gametes Collection, Maturation, and Fertilization In Vitro Astropecten aranciacus (starfish) were collected from the end of January to May in the Gulf of Naples and Gaeta and maintained at 16 degC in circulating seawater. GV-stage oocytes were isolated from the ovary by making a small hole in the dorsal region of the female animal. The oocytes were sieved with gauze and collected in natural seawater (NSW, pH 8.1) filtered with a Millipore membrane of 0.2 mm pore size (Nalgene vacuum filtration system, Rochester, NY, USA). The collected oocytes kept in NSW were used for maturation and fertilization experiments within the next 3 hr. The dry sperm collected from the testis were diluted in NSW and used to inseminate oocytes and eggs at a final concentration of 1 x 106 sperm/mL. In vitro maturation was performed by adding the hormone 1-methyladenine (1-MA) (Acros Organics, Fisher Scientific, Milan, Italy) to the GV-stage oocytes suspended in NSW at different pH at a final concentration of 10 mL/mL. The lowering of pH was obtained immediately before the experiment by adding HCl to NSW (pH 8.1) until reaching the needed value (pH 6.8) . The pH of seawater was raised to 9.0 by adding sufficient NH4OH to NSW (pH 8.1), about 1 mmol of NH4OH/l, which was previously used in studying membrane potential and cortical F-actin structural changes upon the exposure of sea urchin eggs to NH4OH seawater . 2.2. Light Microscopy and Transmission Electron Microscopy (TEM) Light microscopy was performed with the Leica DMI6000 B system to monitor the surface and cortical changes following the insemination of immature oocytes and mature eggs in acidic or alkaline NSW at different time points. For TEM analyses, immature oocytes and oocytes matured and fertilized in acidic or alkaline seawater were fixed in NSW containing 0.5% glutaraldehyde (pH 8.1) for 1 h at room temperature. After extensive washing in NSW, the samples were post-fixed with 1% osmium tetroxide and 0.8% K3Fe(CN)6 for 1 h at 4 degC. The samples were washed in NSW and rinsed with distilled water (3 times, 10 min each), and finally treated with 0.15% tannic acid for 1 min at room temperature. The specimens were then dehydrated in ethanol with increasing concentrations. Residual ethanol was removed with propylene oxide before embedding in EPON 812. Ultrathin sections were made with the ultramicrotome Leica EM UC7 (Leica Microsystems, Wetzlar, Germany) and observed under a Zeiss LEO 912 AB (Carl Zeiss Microscopy Deutschland GmbH, Oberkochen, Germany) without staining. 2.3. Microinjection, Ca2+ Imaging, Fluorescent Labeling of F-Actin and Extracellular Matrix Intact immature and maturing oocytes were microinjected using an air-pressure transjector (Eppendorf Femto-Jet, Hamburg, Germany). To monitor intracellular Ca2+ level changes in the GV-stageoocytes and mature eggs (treated with 1-MA for 70 min) were microinjected 10 min before hormonal stimulation or insemination with 500 mM Calcium Green 488 conjugated with 10 kDa dextran mixed with 35 mM Rhodamine Red (Molecular Probes, Eugene, OR, USA) in the injection buffer (10 mM Hepes, 0.1 M potassium aspartate, pH 7.0). The fluorescence images of cytosolic Ca2+ were captured with a cooled CCD camera (Micro-Max, Princeton Instruments Inc., Trenton, NJ, USA) mounted on a Zeiss Axiovert 200 with a Plan-Neofluar 40/0.75 objective at about 2 s intervals, and the data were analyzed with MetaMorph (Universal Imaging Corporation, Molecular Devices, LLC, San Jose, CA, USA). Following the formula, Frel = [F-F0]/F0, where F represents the average fluorescence level of the entire egg and F0 the baseline fluorescence, the overall Ca2+ signals were quantified for each moment, and Frel was expressed as RFU (relative fluorescence unit) for plotting the Ca2+ trajectories. Applying the formula Finst = [Ft-F(t-1)]/F(t-1), the instantaneous increment of the Ca2+ level was analyzed to locate the specific area of transient Ca2+ increase. The values of Ca2+ signals were obtained from three independent experiments (N) using three different females, and the number of eggs (n) being analyzed for each condition is specified in the Results. To visualize the actin cytoskeleton, 50 mM (pipette concentration in methanol) of the fluorescent F-actin probe AlexaFluor 568-phalloidin (Molecular Probes, Thermo Fisher Scientific, Oregon, USA) was microinjected into GV-stage oocytes and the unfertilized mature eggs in two independent experiments (N = 2), using two different females. To study the effect of the altered pH of the seawater on the morpho-functionality of the jelly coat triggering the sperm acrosome reaction, living immature oocytes and eggs matured with 1-MA in seawater at various pH conditions were incubated with 25 mM of the synthetic fluorescent polyamine BPA-C8-Cy5 to label the extracellular matrix (File S1). The distinct excitation and emission spectra of BPA-C8-Cy5 and Alexa568-phallodin enabled us to label the extracellular matrix and F-actin simultaneously. The fluorescent polyamine BPA-C8-Cy5 also visualized the VL, and the sperm acrosomal processes during insemination in acidic or alkaline seawater of immature oocytes and mature eggs. Alexa568-phalloidin, on the other hand, allowed us to monitor the sequential changes of the cortical F-actin in the same oocyte/egg. The acrosomal processes in A. aranciacus sperm were visualized by diluting sperm in seawater at different pH containing the fluorescent polyamine before insemination in given conditions. The signals of the fluorescent probes from the samples were detected with a Leica TCS SP8X confocal laser scanning microscope equipped with a white light laser and hybrid detectors using the Lightning deconvolution mode (Leica Microsystem, Wetzlar, Germany). The number of eggs examined for each condition (n) is specified in the Results. 2.4. Visualization of Sperm Inside the Fertilized Eggs Diluted sperm were stained with 5 mM Hoechst-33342 (Sigma-Aldrich, Saint Louis, MO, USA) for 30 s before insemination. The labeled sperm nuclei incorporated into A. aranciacus oocytes and eggs were counted in the cytoplasm of fertilized eggs 10-15 min after insemination using a cooled CCD (Charge-Coupled Device) camera (MicroMax, Princeton Instruments Inc., Trenton, NJ, USA) mounted on a Zeiss Axiovert 200 microscope with a Plan-Neofluar 40X/0.75 objective with a UV fluorescent lamp. The number of fertilized eggs examined (n) for each condition in three or four independent experiments (N) using three different females is shown in the Results and Table 1. 2.5. Statistical Analysis The numerical MetaMorph data were compiled and analyzed with Excel (Microsoft Office 2010) and reported as mean +- standard deviation (SD) in all cases in this manuscript. Oneway ANOVA was performed through Prism 5.00 (GraphPad Software), and p < 0.05 was considered statistically significant. For ANOVA results showing p < 0.05, the statistical significance of the difference between the two groups was assessed by Tukey's post hoc tests. The two groups of data showing significant differences from each other were marked with symbols indicating the p values in the figure legends. For each experimental condition, 5 or 6 oocytes were used, and they were from the same animal. Now, of course, the same experiment was repeated 2-4 times (N) utilizing the oocytes from as many (N) animals. Thus, N refers to the number of replicates. On the other hand, the numerical data for Ca2+ signal and other values were pooled together for statistical Analysis by using all the oocytes treated in the same way. This number of oocytes was referred to with 'n'. 3. Results 3.1. Structural Changes of Cortical F-Actin in Immature Starfish Oocytes Incubated at pH 6.8 and pH 9 In the starfish Astropecten aranciacus, immature oocytes isolated from the ovary in natural seawater (NSW, pH 8.1) are arrested at the prophase of meiosis I. At this stage of the maturation process, they are penetrated by multiple sperm upon insemination and fail to undergo the cortical changes, such as exocytosis of the cortical granules and separation of the vitelline layer (VL) from the egg plasma membrane that are experienced by the mature eggs. Indeed, when the immature oocytes are stimulated in vitro with 1-MA and inseminated, the optimum period to achieve a successful monospermic fertilization response is between the germinal vesicle breakdown (GVBD) and the formation of the first polar body. A fully grown A. aranciacus immature oocyte isolated from the ovary contains the large nucleus (GV) at the animal pole. The oocyte is surrounded by the jelly coat (JC) to which follicle cells (FC) adhere . The microinjection of the fluorescent F-actin filament probe AlexaFluor 568-phalloidin in living immature oocytes (n = 18) allows the visualization of the network of actin filaments characteristic of the cytoplasm of the GV-stage oocytes. The VL and the JC of immature oocytes were visualized with the fluorescent polyamine BPA-C8-Cy5 . The addition of sperm double-labeled with the polyamine probe and the DNA dye Hoechst-33342 to the immature oocytes in NSW (pH 8.1) induced the formation of numerous blebs on the oocyte surface due to polyspermic insemination . The binding and fusion between the two gametes promote the formation of the fertilization cones due to the protrusion of the ectoplasmic region of the oocyte as the result of the reorganization of the cortical F-actin in preparation for sperm incorporation . Figure 1D also shows no separation of the VL from the oocyte surface after insemination due to the F-actin-dependent structural organization of the oocyte cortex, which is distinct from that of mature eggs concerning the microvilli length and the orientation of cortical granules . Although the incubation of A. aranciacus immature oocytes in NSW at pH 6.8 for 20 min apparently did not remove the JC surrounding the oocytes , it reduced blebs formation and sperm penetration following insemination of the oocytes in acidic seawater . The effect of the acidic seawater on the structure of the cortical actin cytoskeleton was evident in these oocytes (n = 12), which showed a clear difference in the F-actin fibers in the specific cortical regions that are formed to incorporate the sperm after insemination in acidic seawater as compared to those in the control oocytes shown in Figure 1F arrowheads. Furthermore, the transmission electron microscopy (TEM) micrograph in Figure 2E showed the presence of cortical granules (CG) in the cytoplasm of oocytes incubated in acidic seawater that were dislodged from the plasma membrane as a result of the structural rearrangement of the cortical F-actin. Upon insemination, occasionally, it was possible to observe sporadic exocytotic events on the cortex of the oocyte that did not expand to the entire oocyte surface . At variance with the low rate of sperm penetration in immature oocytes approached by sperm in acidic seawater, GV-stage oocytes incubated for 20 min and inseminated in NSW (pH 9) were penetrated by a more significant number of sperm as a result of the increased receptivity to the sperm, which is ascribable to the pH-dependent structural changes of the oocyte cortex . That the incubation of immature oocytes in alkaline seawater promoted the alteration of the cortical F-actin structure is also indicated by the elevation of the VL following exocytosis of the CG upon fusion and the subsequent release of their contents into the perivitelline space (n = 10). The confocal image showed numerous F-actin structures beneath the fertilization cones formed in the cortical cytoplasm to incorporate multiple sperm (arrowheads). However, it is worth noting that the elevation of the VL upon insemination of immature GV-stage oocytes in alkaline seawater, which was also visualized at TEM , did not resemble that occurring in mature eggs as it collapsed within a few minutes on the oocyte surface, unlike in the mature eggs. Similarly, the elevation of the VL was also inhibited when immature oocytes were microinjected with a fluorescent phalloidin and inseminated in seawater containing the fluorescent polyamine to visualize the formation of the spikes in the perivitelline space due to the F-actin extension with the confocal microscope. 3.2. Effect of the Acidic and Alkaline Seawater on the F-Actin-Dependent Maturation Process Induced in Vitro by the Hormone 1-MA At variance with immature oocytes that permit entry of multiple sperm, mature eggs obtained by incubation of the same oocytes in the presence of 1-MA exhibit a normal fertilization response with monospermy due to the restructuration of the oocyte cortex occurring a few minutes after hormonal stimulation and around the time of GVBD . Figure 3A shows an A. aranciacus egg matured with 1-MA for 70 min which is the time for this species to achieve the optimal physiological conditions necessary to respond correctly to the fertilizing sperm. The hormonal stimulation induces the disassembly of the nuclear envelope of the GV and the intermixing of the nuclear components with the cytoplasm . Insemination in NSW (pH 8.1) activates the eggs that respond with the elevation of the fertilization envelope (FE) as a result of CG exocytosis . Figure 3C,D shows a confocal image of the dramatic rearrangement of the F-actin on the outer region of the egg cytoplasm at fertilization. Before fertilization, the subplasmalemmal actin filaments were regularly oriented perpendicularly to the egg surface , but appeared sparse after fertilization (n = 13). Figure 3D shows the network of actin filaments characteristic of the outer region of the cytoplasm of living mature eggs, surrounded by the VL and the JC (visualized in green color with BPA-C8-Cy5) in the same living egg. The image also shows the translocation of the actin filaments from the cortex of the fertilized egg to the center (orange color) and the elevation of the fertilization envelope (FE) and F-actin spike formation in the perivitelline space (PS), labeled with BPA-C8-Cy5 upon incubation of unfertilized eggs. The Hoechst-33342-stained sperm inside the cortical cytoplasm, visualized 6 min after insemination, is visible in blue beneath the fertilization cone (arrow). Transmission electron microscopy (TEM) observations revealed the ultrastructure of the surface of an unfertilized control egg matured with 1-MA for 70 min . Microvilli (MV) distributed uniformly and regularly on the egg surface are embedded in the VL, covering the plasma membrane. Upon insemination, the VL, which will form the FE, separates from the egg plasma membrane following the release of the content of the CG in the PS . Elongation of microvilli (MV) in the PS is also visible in the eggs fixed 5 min after insemination. When GV-stage oocytes (A. aranciacus) are stimulated with 1-MA to induce maturation in acidic seawater (pH 6.8), the transmitted light microscopic images of GVBD and the elevation of the fertilization envelope resembled those of the control eggs matured and inseminated in NSW (pH 8.1) shown in Figure 3A,B. However, subtle structural differences were observed in the cortex of mature eggs before and after fertilization due to the effect of the acidic seawater on the cortical F-actin. The confocal image of Figure 4A shows 1-MA-induced reorganization of the F-actin in the egg cortex (n = 14), highlighted by the conspicuous absence of F-actin filaments (orange color) that are typically oriented perpendicularly to the plasma membrane in control eggs, as exemplified in Figure 3C. Following fertilization and elevation of FE , the lack of the centripetal translocation of the actin fibers from the egg surface is evident in the confocal image of Figure 4B (orange color), with an F-actin remodeling events occurring in sea urchin eggs, too . The effect of the acidic seawater on the microvillar morphology of the unfertilized mature egg is highlighted in the TEM micrograph , showing MV of different lengths embedded in the VL and CG detached from the egg surface. Upon insemination, CG are no longer visible in the cortical cytoplasm due to their exocytosis, which is also evidenced by the released material (arrow) in the PS in the TEM image in D. In immature oocytes stimulated with 1-MA for 70 min in alkaline seawater (pH 9), GVBD did not occur, as shown in the transmitted light image . The confocal microscopic image of the cortical and cytoplasmic F-actin in the eggs matured with 1-MA in alkaline seawater (n = 13) shows, at variance with the control, the inhibition of the dramatic reorganization of the F-actin in the cytoplasm and cortex of the oocytes following GVBD and lack of the disappearance of the network of the actin filaments in the cytoplasm . The outer region of the cytoplasm of the oocytes treated with 1-MA in alkaline seawater was morphologically altered, as shown by the ultrastructural analysis by TEM . The image highlights an increase and not a decrease in microvillar length, the latter of which (shortening of microvilli) is a morphological change that usually occurs at the time of the GVBD and is thought to be essential for the establishment of a normal fertilization potential and Ca2+ response at fertilization . Furthermore, the alkaline seawater induced an increased polyspermic incorporation rate, as judged by the higher number of F-actin structures formed to incorporate sperm into the cytoplasm . The TEM micrograph of an egg matured in alkaline seawater, fertilized, and fixed 5 min upon insemination shows the elevation of the FE as a result of the exocytosis of the CG into the PS, which was not detectable under a transmitted light (B) and a confocal microscope (D) as it slightly elevated from the egg surface. Since GVBD did not take place in the oocytes stimulated with 1-MA in seawater at pH 9, the GV-stage oocytes were first treated for 50 min with 1-MA in NSW (pH 8.1) to allow the breakdown of the nuclear envelope and then transferred to seawater at pH 9 for 20 min . A maturation time of 70 min is required to achieve optimal fertilizability conditions for the eggs of this starfish species. In these experimental conditions, even if the intermixing of the nucleoplasm with the cytoplasm normally occurred, 20 min of exposure to alkaline seawater (pH 9) compromised the fertilization response, as shown in Figure 6. The visualization of sperm entry in these eggs showed that, even if sperm incorporation was not affected (see below), the time required for the eggs to incorporate the sperm was much longer than in the control eggs fertilized in NSW . Indeed, 18 min after insemination, the sperm is still trapped in the fertilization cone, as shown in Figure 6B (arrowhead). This is probably due to the alteration (stabilization) of the F-actin core of microvilli (MV), which was more prominent in these experimental conditions than in control . The alteration of the morpho-functionality of the F-actin of the egg cortex caused by the alkaline seawater-induced maturation process was also evident following fertilization. Figure 6B shows an accumulation of longer actin fibers (orange color) on one side of the egg surface 18 min after insemination that failed to translocate toward the center of the fertilized egg (n = 11), which the control eggs in NSW (pH 8.1) achieved in 6 min . Nonetheless, the stabilization of the actin filaments in the egg cortex did not interfere with the elevation of the FE due to the exocytosis of CG in the PS (arrow), as shown in the confocal and TEM images in B and D. 3.3. Effect of the Acidic or Alkaline Seawater on the Polyspermy Rate of Immature Oocytes and the Sperm Entry Following Oocytes Maturation in Altered Seawater pH To test whether the lower or higher pH of seawater could affect the polyspermy, characteristic of immature starfish oocytes, GV-stage oocytes were incubated for 20 min in NSW (pH 8.1, control), 6.8 (acidic), or 9.0 (alkaline) and subsequently inseminated with Hoechst 33342-stained sperm (Table 1). The number of oocyte-incorporated sperm was counted 10-15 min after insemination. Figure 7A and the histogram (green color) show the fluorescently labeled DNA of multiple sperm inside the oocytes inseminated in NSW at pH 8.1 (4.13 +- 0.90, n = 80). At variance with that, immature oocytes incubated and inseminated in seawater at pH 6.8 were penetrated by a lower number of sperm than the control (1.22 +- 0.75, n = 80, p < 0.01). Since the organization of the F-actin structures in the oocyte cortex that are formed to incorporate sperm is quite different in the oocytes inseminated in acidic seawater , as compared to those in control (pH 8.1) , the reduced rate of polyspermy observed in the oocytes inseminated in acidic seawater may be attributed to the alteration of the structural organization of the cortical F-actin. By contrast, GV-stage oocytes incubated and inseminated in alkaline seawater (pH 9) were penetrated by much more sperm than the control (15.35 +- 2.65, n = 80, p < 0.01), as shown in the fluorescent image of Figure 7A (third fluorescent image of the panel) and the histogram (purple color). In a slightly different protocol, we examined the effect of acidic or alkaline seawater on oocyte maturation and its consequence at fertilization . As expected, A. aranciacus oocytes incubated with 1-MA for 70 min in NSW (pH 8.1) led to monospermic fertilization in most cases (48 out of 60 eggs). In the remaining mature eggs, the number of sperm counted was two or three per egg to make the average 1.23 +- 0.25 (n = 60). When oocytes were induced to undergo maturation with 1-MA in acidic seawater (pH 6.8) , only 6 out of 60 mature eggs showed a single sperm in the cytoplasm. When GV-stage oocytes were induced to undergo maturation with 1-MA in NSW (pH 8.1) for 50 min and then transferred to seawater at pH 9 for further (20 min) incubation, the eggs experiencing the later stage of maturation in alkaline seawater exhibited a higher rate of polyspermy than the control eggs. Of 60 mature eggs being analyzed under a fluorescence microscope, 34 eggs were monospermic, and the remaining eggs had 2 or 3 sperm inside to make the average 1.82 +- 0.46 sperm per egg (n = 60), which was modestly higher than the value in the control (non-significant). By contrast, when maturation was promoted in alkaline seawater (pH 9), a higher number of sperm entered the egg under these experimental conditions (20.42 +- 2.55 n = 60, p < 0.01 in comparison with the control NSW pH 8.1), which is also indicated by the visualization of one fertilized egg with seven sperm inside . 3.4. Insemination of the GV-Stage Oocytes in Seawater with Altered pH Significantly Affects the Sperm-Induced Ca2+ Response Previous studies from our laboratory showed that the sperm-induced Ca2+ response in the GV-stage starfish oocytes (not treated with the maturing hormone) is dependent on the structural organization of the oocyte cortex. At this stage of the maturation process (prophase I), the immature oocytes have longer microvilli and more irregular actin meshwork underneath the plasma membrane in comparison with the mature eggs . The F-actin morphological features of the cortex of immature oocytes mirror their characteristic Ca2+ response at fertilization, which is quite different from that in mature and overmatured eggs . Figure 8A shows that the addition of sperm to a GV-stage oocyte suspended in NSW (pH 8.1) promotes two small releases of Ca2+ due to the interaction of two sperm with the oocyte plasma membrane (arrowheads). These Ca2+ releases are followed by a subitaneous increase in Ca2+ at the periphery of the oocyte (cortical flash, CF) that lasts only a few seconds and is followed by two Ca2+ waves that converge and propagate to the opposite pole in about two minutes, as indicated by the pseudocolor image of the relative fluorescence analysis (n = 13). The graph in Figure 8D and the histogram in E (green colors) show that when the immature oocytes were inseminated in NSW, the time lag until the first Ca2+ release was 22.5 +- 7.7 s. The Ca2+ waves reached an amplitude of 0.37 +- 0.04 RFU in 358.7 +- 36.3 s. At variance with the control, oocytes inseminated in acidic seawater (pH 6.8) showed a Ca2+ response only in 4 out of 13 oocytes with a much longer delay (215.3 +- 87.5 s, n = 4, p < 0.01). The peak amplitude of the Ca2+ wave was also significantly lower than in the control . Moreover, the occurrence of the CF, which followed the Ca2+ waves elicited by two sperm visualized in the third pseudocolor image was substantially delayed as it appeared 1 min and 32 s after the initiation of the first Ca2+ signal, as shown in the fourth relative fluorescence image of Figure 8B. Differences were also observed in the time required to reach the highest release of Ca2+ (time to reach the Ca2+ peak, 0.19 +- 0.06 RFU in 494.7 +- 49.9 s, n = 4) as compared to the control (358.7 +- 36.3 s n = 13, p < 0.05) . As for the Ca2+ response induced by sperm addition to immature oocytes incubated in seawater at pH 9 (n = 14), even if the sequence of occurrence of the initial Ca2+ spots and CF was similar to that of the control, striking differences were observed in the pattern of Ca2+ release and the decline as shown in the pseudocolor images . The results show that the time to reach the Ca2+ peak amplitude was faster than the control (167.6 +- 25.3 versus 358.7 +- 36.3 s, p < 0.01) and that the higher Ca2+ increase (0.45 +- 0.04 RFU, p < 0.05) induced by the fusion of multiple Ca2+ waves propagated faster than the control (44 s versus 120 s). The results also show that insemination of the oocytes kept in alkaline seawater dramatically alters the pattern of Ca2+ release by the heavy stimulation of the oocyte surface caused by the interaction and subsequent penetration of many more sperm (15.35 +- 2.65, n = 80) than in the control. 3.5. A. aranciacus Oocytes Matured in Seawater at pH 6.8 and pH 9 Show an Altered Ca2+ Response at Fertilization as Compared to the Control In normal conditions, upon insemination, eggs matured in NSW (pH 8.1) with 1-MA at a final concentration of 10 mM experience the first Ca2+ response (the CF), which takes place simultaneously at the periphery of the egg cortex as a result of the activation of the L-type Ca2+ channels promoting Ca2+ influx . The pseudocolor images show that the CF is followed by a Ca2+ wave propagating to the opposite pole . The graph and histogram in D and E (green colors) show the fertilization Ca2+ responses representative of 13 control eggs matured and fertilized in NSW. The Ca2+ wave in control eggs reached a peak amplitude of approximately 0.49 +- 0.03 RFU and took about 2 min to get to the opposite pole (traverse time, 131.1 +- 16.1 s) . In addition to a normal Ca2+ response, optimal physiological conditions of the eggs are also required for a typical separation of the vitelline layer, which will form the FE seen in Figure 3B. The GV-stage oocytes stimulated with the hormone while suspended in acidic seawater (pH 6.8) can undergo GVBD and, if fertilized, display apparently normal elevation of the FE similar to that of the control. However, it is interesting to note that, at variance with immature oocytes in which incubation and insemination in seawater at pH 6.8 heavily reduced the Ca2+ response and the number of sperm entries, all the eggs matured in acidic seawater (n = 13) responded to sperm stimulation by elevating their intracellular Ca2+ with a CF, which was followed by a Ca2+ wave . The Analysis of the Ca2+ changes following sperm addition showed that the only parameter of the sperm-induced Ca2+ response that was altered was the propagation time of the Ca2+ signal in the egg , which was significant for a longer duration (160.7 +- 14.5 s). On the other hand, when GV-stage oocytes were induced to mature in seawater at pH 9, they all failed to undergo GVBD. As aforementioned, 70 min after the addition of 1-MA, a GV with an abnormally elliptical shape was still visible . Upon insemination, these "mature" eggs (n = 14) elicited a Ca2+ response that was heavily altered as compared to that of the control eggs matured and fertilized in NSW . The relative fluorescence images of Figure 9C show that at variance with the control, the Ca2+ response could initiate with a Ca2+ spot (arrowhead) which was then followed by a CF and a series of Ca2+ waves as a result of the fusion of multiple sperm transducing several Ca2+ signals in 8 eggs out of 14. The Ca2+ waves then converged and propagated faster to the opposite pole (101.4 +- 11.6 s, n = 14) than the control (131.1 +- 16.1 s, n = 13, p < 0.05). The analysis of the relative fluorescence changes has also indicated that 3 out of 14 eggs did not elicit any CF. Furthermore, the Ca2+ peak amplitude was also affected when eggs were matured in alkaline seawater (0.35 +- 0.04 RFU, p < 0.01), as shown in the graph of Figure 9D (brown colors) and the histogram in Figure 9E. A lesser effect on the Ca2+ response was highlighted in eggs inseminated after being matured for 50 min in NSW (pH 8.1) and 20 min in seawater at pH 9 (n = 12). While the Ca2+ wave always followed the CF as in control in A, its amplitude was significantly lower than that of the control (0.37 +- 0.02 RFU, p < 0.01), as shown in the graph in D (light brown color) and histogram in Figure 9E. 4. Discussion Eggs of marine invertebrates have been broadly used in laboratories to study fertilization because the eggs are spawned freely into the sea, where they are immediately mixed with the spermatozoa. Thus, the fertilization process and the following developmental changes of the embryo can be easily observed in the petri dish. It has been known that the structural organization of the cortical F-actin in the oocytes and eggs profoundly impacts how they respond to the fertilizing sperm with Ca2+ increase and sperm incorporation . Interestingly, our recent studies with sea urchin eggs in acidic and alkaline seawater conditions suggested that the pH of the seawater has profound effects on many aspects of fertilization that are linked to the cortical F-actin structural dynamics, such as patterns of Ca2+ signaling and actin cytoskeletal remodeling that take place in the eggs after fertilization . In this communication, by using starfish as the experimental model, we have surveyed the effect of acidic and alkaline seawater not only on fertilization, but also on meiotic maturation. To begin with, the structural organization of the GV-stage oocytes and mature eggs are strikingly different in terms of the actin cytoskeleton (including microvilli) and the distribution of cortical granules and vesicles . Because of that, immature oocyte insemination results in polyspermic entries. However, various manipulation of the pH of the media (seawater) either intensified the tendency of polyspermy or alleviated it (Table 1). In line with that, the changes of the pH of the incubation media in various conditions resulted in remarkable alterations in terms of organization of the F-actin cortical network, meiotic progression (GVBD), Ca2+ signaling pattern upon insemination, sperm incorporation, and cortical granule exocytosis in some cases. The results are summarized in Table 2. Another advantage of using starfish is that the formation of the acrosomal process at the head of sperm is much more prominent in starfish than in sea urchins. Furthermore, the location of the occurrence of the sperm acrosomal process is well known. Indeed, both in vitro and in vivo, it has been shown that the acrosomal process (AP) formation occurs when the fertilizing sperm contacts the outer layer of the jelly coat . The tip of the long (20 mm) and thin acrosomal process made by actin filaments enters openings in the vitelline layer and interacts with the egg plasma membrane to trigger electrical and Ca2+ changes upon their fusion . Immature oocytes isolated from the gonad of A. aranciacus in natural seawater (pH 8.1) are prone to polyspermy due to their structural organization of the oocyte surface and cortex . Following insemination, the interaction of the tips of the sperm acrosomal filament with the oocyte plasma membrane elicits numerous Ca2+ responses with a pattern of Ca2+ release , which is different from the single one produced by a mature egg undergoing regular cortical F-actin restructuring during maturation . The different organization of the cortical F-actin in the female gametes at the two different maturation stages is also indicated by the lack of the separation of the vitelline layer from the oocyte plasma membrane at fertilization . Finally, the incorporation of numerous sperm by the specialized F-actin structures polymerized in the oocyte cytoplasm beneath the fertilization cones further shows differences in the F-actin-dependent mode of sperm incorporation in comparison with mature eggs . In starfish oocytes, following hormonal stimulation, the cortical cytoskeletal reorganization by G-proteins' activation modulates the Ca2+ signals during the early and late phases of the maturation process in starfish . Acidic and alkaline seawater incubation also altered the structure of the F-actin of the immature oocyte cortex, which, in turn, impaired the sperm-oocyte binding and the pattern of the sperm-induced Ca2+ signals. The role of intracellular pH in regulating actin polymerization and Ca2+ signaling in eggs has been demonstrated in sea urchins and Xenopus . In this regard, besides the faster and higher Ca2+ release upon insemination in alkaline seawater, immature oocytes displayed a faster decline in the intracellular Ca2+ level down to the baseline level , as previously shown by sea urchin eggs , as well as an increase in the rate of F-actin dependent sperm incorporation . An additional indication of the striking effect of the higher pH on the structure of the cortical F-actin is given by the observation under a light microscope that the vitelline layer could separate from the plasma membrane of immature oocytes incubated and inseminated in alkaline seawater . Since such an event never happens in control, these results add weight to the suggested role played by the state of polymerization of the F-actin of the cortex of immature oocytes and mature eggs in the regulation of the exocytosis of cortical granules and independently of the Ca2+ increase . When maturation was induced in alkaline seawater, the inhibition of the structural changes of the cortical F-actin was also striking . The cortical F-actin alteration also inhibits the breakdown of the nuclear envelope of the large nucleus GV anchored to the oocyte surface by actin filaments , the disassembly of which is dependent on the F-actin dynamic changes at the nucleoplasmic face that are essential to rupture the nuclear envelope . As a result, the absence of the intermixing of the nucleoplasm with the cytoplasm by preventing the rearrangement of the cortical F-actin at the time of GVBD impairs the acquisition of cytoplasmic maturity, a normal fertilization potential and Ca2+ response upon insemination of mature eggs . Interestingly, the given condition of acidity (pH 6.8) did not make much difference to the morphology of the JC of immature oocytes, as judged by the fluorescent labeling by BPA-C8-Cy5 , and did not affect sperm motility as they arrived in the vicinity of the surface of the oocyte with no time lag. However, the pH appears to affect the sperm acrosomal process formation. As shown in Figure 2D, the acrosomal processes traversing the JC are more easily labeled in the oocytes inseminated in the alkaline seawater (pH 9). On the other hand, insemination in acidic conditions (pH 6.8) made it more challenging to observe the acrosomal processes over the oocytes compared with the oocytes inseminated in NSW pH 8.1 . This observation is comparable with the idea that the formation of the acrosomal process per se is facilitated in higher seawater pH. This is possibly due to the effect of the seawater pH on the sugar moiety of the sulfated glycoprotein, i.e., the inducer of the acrosome reaction . Interestingly, at fertilization of oocytes matured in acidic seawater, only the initiation of the sperm-induced Ca2+ response was not affected, indicating a standard functionality of the JC but the propagation of the Ca2+ wave and sperm entry were compromised due to the alteration of the cortical F-actin dynamics . Drastic increases in the Ca2+ response and in the rate of sperm penetration were observed in GV-stage oocytes challenged with sperm and after maturation of oocytes in alkaline seawater. In light of the fact that the acrosomal processes are more easily visible in alkaline seawater and much less so in acidic seawater , this enhanced sperm entry into the oocytes and eggs is probably due to the facilitated induction of the sperm acrosome reaction in starfish, which is known to be stimulated by an increased pH . Furthermore, the results showed that GVBD inhibition was due to the effect of the alkaline seawater on the F-actin structures of microvilli and cortex, which failed to undergo morphological changes underlying the maturation process . Thus, alkaline seawater may not represent the optimal pH for cdk1 activity and for the F-actin rearrangement to fragment nuclear membranes leading to GVBD . Finally, even if a less drastic effect of the alkaline seawater on polyspermic was observed when immature oocytes were stimulated with 1-MA for 50 min in NSW and 20 min in alkaline seawater, the time delay in sperm incorporation and alteration of the F-actin translocation following fertilization also indicated the strict interdependence of the effect of the alkaline pH on the structural modification of the cortical F-actin, leading to an altered Ca2+ response at fertilization . The advantage of designing controlled experimental conditions in the laboratory to study the effect of the altered seawater pH on the physiology of the fertilization process of marine eggs is of great importance. Indeed understanding how changes in the pH values of seawater affect the structural organization and dynamics of the actin cytoskeleton of the egg cortex may shed light on the molecular mechanisms regulating the F-actin-dependent Ca2+ signaling in other cell types as well. In line with this, acidic and alkaline seawater alteration of the cortical actin cytoskeleton of starfish oocytes, which is differently organized in the two oocyte maturation stages (this contribution), and sea urchin eggs affect the sperm-induced Ca2+ signaling and the subsequent F-actin reorganization necessary for cleavage . These results may also light up a similar scenario occurring in natural environmental conditions, i.e., reducing the surface ocean pH due to the absorption of the increasing atmospheric carbon dioxide. Even if a seawater pH 6.8 would correspond to an extreme condition, seawater acidification will inevitably limit the animal population by significantly impacting the initiation of the fertilization process. 5. Conclusions In starfish, it has been well known that the breakdown of the immature oocytes' large nucleus (germinal vesicle, GV) induced by hormonal stimulation must occur to make the mature egg ready to be successfully fertilized with normal Ca2+ response and penetration by only one sperm. The oocyte maturation process following the addition of 1-MA brought about by the intermixing of the nucleoplasm with the cytoplasm ensures the F-actin-dependent reorganization of the cortex of the mature egg for a normal fertilization response. In line with this, immature oocytes are prone to polyspermic fertilization. The results of this contribution have shown that the acidic or alkaline seawater pH reduced or increased polyspermy rates, respectively, and the altered Ca2+ responses at fertilization. Altered pH also affects the acrosome reaction of the sperm and the structure and dynamics of the actin filaments in the oocyte cortex. Morphological modifications of the cortical F-actin network of eggs matured in seawater at different pHs influencing the sperm-induced fertilization response were also observed. The results have highlighted the strict interdependence between the seawater pH and the sperm-induced Ca2+ signals in immature oocytes and mature eggs, which reflect the different organization of the cortical F-actin in their two maturation stages. Acknowledgments The authors are grateful to Davide Caramiello for maintaining the starfish and to the technicians of the Advanced Microscopy Center at the SZN for processing fixed samples for SEM and TEM analyses and assisting in the observations with the electron microscopes. The authors want to thank Giovanni Gragnaniello for his help in preparing the Figures. Supplementary Figure S1 was created using BioRender.com. Supplementary Materials The following supporting information can be downloaded at: Figure S1: Summary of the experimental design. (A-C) GV-stage oocytes were incubated in seawater at different pH for 20 min and then inseminated in the same medium. (D-F) GV-stage oocytes were first matured with 1-methyladenine (1-MA) in seawater at different pH and subsequently inseminated in the same media. (G) GV-stage oocytes were first stimulated with 1-MA in natural seawater (pH 8.1) for 50 min but exposed to seawater (pH 9.0) for 20 min before the insemination in the alkaline seawater. File S1: Synthesis of the fluorescent molecular probe BPA-C8-Cy5. Click here for additional data file. Author Contributions Conceptualization, L.S., J.T.C. and N.L.; methodology, N.L. and L.S.; resources, J.T.C., S.C.S., J.-L.S., J.-M.L. and L.S.; formal analysis, N.L., L.S., J.T.C.; investigation, N.L, L.S, J.T.C.; data curation, N.L., L.S. and J.T.C.; visualization N.L., L.S. and J.T.C.; writing--original draft preparation, L.S., J.T.C. and N.L.; writing--review and editing, L.S.; J.T.C. and N.L.; supervision, L.S.; project administration, L.S.; funding acquisition, L.S. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Starfish used for the present study were collected according to the Italian legislation (DPR 1639/68, 19 September 1980 and confirmed on 1 October 2000). All the experimental procedures were carried out following the guidelines of the European Union (Directive 609/86). Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Structural changes of the actin cytoskeleton following insemination of immature oocytes in NSW (pH 8.1). (A) A. aranciacus oocytes isolated from the ovary in NSW (pH 8.1) are blocked at the prophase I of the meiotic division, showing a large nucleus (germinal vesicle, GV). These oocytes also surrounded by the follicle cells (FC) adhering to the extracellular jelly coat (JC). (B) In confocal microscopy, actin filaments were visualized with AlexaFluor 568-phalloidin (orange color), and the vitelline layer (VL) and JC were disclosed with fluorescent polyamine BPA-C8-Cy5 (green color). (C) Sperm heads were labeled with Hoechst-33342 (blue color) and the tails with BPA-C8-Cy5. (D) Upon sperm addition, several fertilization cones (i.e., polyspermic) are formed on the oocyte surface as a result of the protrusion of the cytoplasm (blebs, arrow). (E) JC triggers the formation of the acrosomal process, which is now visible as if there is a tether between the sperm head and the fertilization cone (arrow). (F) Formation of thick bundles of F-actin where multiple sperm are incorporated. Figure 2 Penetration of sperm into GV-stage oocytes incubated and inseminated in acidic (pH 6.8, left panels) and alkaline (pH 9, right panels) seawater. (A) In the acidic seawater, the morphology of the fertilization cones (blebs, arrows) was similar to that of the oocytes fertilized in NSW at pH 8.1 , and there was no elevation of the fertilization envelope (FE). (B) In alkaline seawater, FE elevates but the oocytes are predominantly polyspermic. (C,D) The shapes of the fertilizing cones labeled with AlexaFluor 568-phalloidin (orange color, arrowheads) are strikingly different in oocytes fertilized in acidic (round) and alkaline (sickle-formed) seawater. In alkaline seawater, a strikingly higher number of sperm entered the eggs, and the acrosomal processes were more easily visualized by BPA-C8-Cy5 (arrow) in the perivitelline space. (E) The TEM images show that cortical granules (CG) are dislodged from the surface of the oocytes in acidic seawater, which hampered their exocytosis and resulted in a lack of FE elevation. (F) In alkaline seawater, the VL readily elevates upon fertilization, creating a vast perivitelline space (PS). However, an even higher number of sperm entered (see the text). Note that microvilli (MV) in the oocytes of the acidic seawater (E) are not remodeled, unlike the ones in the alkaline seawater (F). Figure 3 Cortical reaction and F-actin changes the eggs matured and fertilized in NSW pH8.1. GV-stage oocytes of A. aranciacus were stimulated with the maturing hormone 1-methyladenine (1-MA) in NSW (pH 8.1) for 70 min prior to fertilization. The left panels represent the eggs before fertilization, and the right ones 5 min after fertilization. (A) Transmission view of the mature egg. GV breakdown is evident. (B) Fertilized eggs show the elevation of the fertilization envelope (FE) and the expansion of the perivitelline space (PS). Arrow indicates the single site of sperm entry. (C,D) Confocal microscopy images showing actin filaments (AlexaFluor568-phalloidin, orange color), extracellular matrix jelly coat (JC and VL; BPA-C8-Cy5, green color), and sperm head (Hoechst-33342, blue color). The subplasmalemmal actin meshwork in an unfertilized egg (C) disappears as many microvilli are extended in the PS (D). (E,F) TEM images of the egg surface before (E) and after fertilization (F). By 5 min after fertilization (F), the content of the CG (arrow) is deposited in the perivitelline space (PS) containing elongated microvilli (MV). Figure 4 Structural changes of the actin cytoskeleton in the cortex of A. aranciacus eggs matured and fertilized in acidic seawater (pH 6.8). Left panels before fertilization, right ones after. (A,B) Confocal images. (C,D) TEM images. The elevation of the fertilization envelope (FE) following insemination still occurs in this acidic condition (B,D). Note that the jelly coat (JC) is still visible in the eggs matured in acidic seawater (panel (A); JC visualized by BPA-C8-Cy5, green color). The arrow in (B) shows the DNA-labeled sperm in the cytoplasm of the fertilized egg. Following oocytes maturation in acidic seawater, the impairment of the cortical F-actin remodeling is also evident in the TEM image in (C), showing microvilli (containing actin filaments) with a different length embedded in the vitelline layer (VL). The cortical granules (CG) are not positioned near the plasma membrane, unlike the eggs matured in NSW with pH 8.1. Microvilli (MV) elongation in the perivitelline space (PS) 5 min after insemination is visible in the TEM image in (D). Figure 5 Starfish oocytes stimulated with maturation hormone in alkaline seawater fail to undergo GVBD. The GV-stage oocytes of A. aranciacus were exposed to the alkaline seawater (pH 9) for 70 min before insemination in the same media. The left panels represent images before insemination, and the right ones 5 min after insemination. (A,B) Light microscopy images. (C,D) Confocal micrographs. (E,F) TEM images. Note that incubation and maturation of GV-stage oocytes in alkaline seawater (pH 9) blocks the disassembly of the nuclear envelope of the GV (GVBD). However, the GV in these oocytes appeared flattened after the incubation in the alkaline seawater. Note also the formation of many fertilization cones containing F-actin structures apparently encompassing the head of incorporated sperm ((D), orange and blue colors, arrowhead). Numerous acrosomal filaments (arrow) impressively traversing the JC ((D), arrow; stained with BPA-C8-Cy5, green color), indicating high tendency of polyspermy. The elevation of the fertilization envelope (FE) 5 min after sperm addition is modest, which was visible only at the TEM level (F), and not at light microscopy (B). Microvilli (MV) elongate in the perivitelline space (PS) in (F). Figure 6 Cortical F-actin dynamics in A. aranciacus oocytes that were initially matured in NSW (pH 8.1) for 50 min and then in alkaline seawater (pH 9) for 20 min. The confocal (A,B) and TEM (C,D) were captured before (A,C) and after insemination (B,D). In the confocal images, BPA-C8-Cy5 (green color) delineated the vitelline layer (VL) and jelly coat (JC), and AlexaFluor 568-phalloidin (orange color) visualized F-actin. The confocal image in (B) shows that even if the elevation of the fertilization envelope (FE) is not affected, the translocation of the subplasmalemmal actin filaments (orange color) is much delayed. Note also the fertilizing sperm trapped beneath the fertilization cone in panel (B) (arrowhead) even 18 min after insemination. (C) After the second exposure to the alkaline seawater, the eggs exhibited prominent actin filaments within the core of some microvillar (MV, arrow), which were never observed in control eggs . The TEM micrograph in panel (D) shows the cortical granules (arrow) content in the perivitelline space (PS) following their exocytosis. Figure 7 Pretreatment of GV-stage oocytes in acidic or alkaline seawater alters the rate of polyspermy at fertilization and sperm entry following oocyte maturation in altered seawater pH. (A) A. aranciacus immature oocytes were preincubated for 20 min in NSW at pH 8.1 (control), pH 6.8 (acidic), or pH 9 (alkaline) and inseminated in the same media with Hoechst 33822-prestained sperm. The number of oocyte-incorporated sperm 10-15 min after insemination was counted using epifluorescence microscopy for each case and presented in histograms. (B) Sperm incorporation in A. aranciacus eggs matured with 1-MA in a variety of pH conditions prior to fertilization: (i) NSW (pH 8.1) for 70 min; (ii) acidic seawater (pH 6.8) for 70 min; (iii) 50 min in NSW (pH 8.1) and then transferred to alkaline seawater (pH 9) for 20 min; (iv) alkaline seawater for 70 min. The number of sperm inside the eggs was counted in the same method and presented in the histograms. # p < 0.05, * p < 0.01. Figure 8 Insemination of GV-stage oocytes in acidic and alkaline seawater alters the sperm-induced Ca2+ response. The pseudocolored images in (A) represent the instantaneous increases in Ca2+ levels extracted from a time-lapse acquisition following the insemination of an immature A. aranciacus oocyte kept in normal seawater (NSW, pH 8.1). The first Ca2+ signal detected upon sperm addition is the initiation of multiple Ca2+ waves (arrowheads) as a result of a polyspermic stimulation that is followed by a cortical Ca2+ release (the cortical flash, CF), which occurs simultaneously at the periphery of the oocyte cortex. The Ca2+ waves run together to propagate to the opposite pole. Immature oocytes loaded with the calcium dye were fertilized in acidic (pH 6.8, (B)) and alkaline seawater (pH 9, (C)). The traces of Ca2+ signals were quantified in (D,E). Color codes: green, pH 8.1; blue, pH 6.8; purple, pH 9. Tukey's post hoc test * p < 0.01, # p < 0.05. RFU = Relative Fluorescence Unit. The time in the pseudocolor fluorescent images (A-C) indicates minutes and seconds from the standard time format (mm: ss). Figure 9 The Ca2+ response at fertilization mature eggs of A. aranciacus was altered when immature oocytes were induced to undergo maturation in acidic and alkaline seawater. (A) Ca2+ response in the control eggs matured and fertilized in NSW (pH 8.1). (B) Ca2+ response in the eggs matured and fertilized in the acidic seawater (pH 6.8). (C) Ca2+ response in the eggs matured and fertilized in the alkaline seawater (pH 9). The pseudo-color images of the Ca2+ response in the eggs first matured in NSW (pH 8.1) then fertilized in alkaline seawater sere similar to that the control eggs, and thus not shown here. (D,E) Quantification of the Ca2+ response in each condition. Note that in the eggs matured and fertilized in alkaline seawater (C), the cortical flash CF) precedes the Ca2+ wave, unlike the eggs in all other conditions, and that the generation of multiple Ca2+ signals in different areas indicate polyspermic gamete interaction. The higher pH of the seawater significantly reduced the peak amplitude of the Ca2+ wave (D,E), and shortened the propagation time required for the Ca2+ waves to reach the opposite pole. Tukey's post hoc test * p < 0.01, # p < 0.05. RFU = Relative Fluorescence Unit. cells-12-00740-t001_Table 1 Table 1 Summary of the effect of the acidic or alkaline seawater on sperm entry in GV-stage oocytes and mature eggs. The letters (A to G) in the column of the experimental conditions marked with the asterisk refer to the experimental scheme specified in Figure S1. Experimental Conditions * n. of Inseminated Oocytes/Eggs n. of Oocytes/Eggs Successfully Penetrated by Sperm Percentage of Monospermy Percentage of Polyspermy A 80 80 0% 100% B 80 38 15.8% 84.2% C 80 80 0% 100% D 60 60 80% 20% E 60 6 100% 0% F 60 60 0% 100% G 60 60 56.7% 43.3% cells-12-00740-t002_Table 2 Table 2 Summary of the effect of the acidic or alkaline seawater on GV-stage oocytes and eggs matured in different seawater pH at fertilization. The letters indicating the experimental conditions refer to those provided in Figure S1. Experimental Conditions Morphology (Light and TEM Observations) F-Actin Distribution Before and After Insemination Ca2+ Changes Sperm Incorporation FE Elevation A Long MV CG dislodged from PM Network of F-actin in the oocyte cytoplasm. Formation of fertilization cones after insemination. CF after or together with multiple CW Polyspermy NO B Long MV CG dislodged from PM Altered distribution of the cortical F-actin before insemination. Reduced formation of the fertilization cones. Failure of Ca2+ release. Delay and reduced amplitude of CF and CW Reduced polyspermy NO C Long MV CG exocytosis at insemination Altered F-actin redistribution. Increased formation of the fertilization cones. Higher CW amplitude. Multiple CW Faster Ca2+ reuptake Increased polyspermy YES, but collapsed D Shortened MV CG beneath PM F-actin fibers perpendicularly oriented in the unfertilized egg cortex. Centripetal translocation of F-actin fibers following fertilization. One fertilization cone. CF before CW Single CW Monospermy YES E CG detached from egg surface Different length of MV Lack of F-actin distribution in the unfertilized egg cortex following maturation. Inhibition of actin fibers translocation following fertilization. Slower CW propagation Inhibition of sperm entry YES F Longer MV GVBD inhibition Altered F-actin organization in the cortex and cytoplasm following maturation and fertilization. Increased number of fertilization cones. CF after multiple CW Reduced CW amplitude Faster CW propagation Polyspermy NO G F-actin core of MV more evident Alteration of the cortical F-actin organization following maturation and translocation upon fertilization. Reduced CW amplitude Monospermy YES Abbreviations: Microvilli (MV), Plasma Membrane (PM), Cortical Ca2+ Flash (CF), Ca2+ Wave (CW), Cortical Granules (CG), Fertilization Envelope (FE), Germinal Vesicle Breakdown (GVBD), Transmission Electron Microscopy (TEM). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Kanatani H. Shirai H. Nakanishi K. Kurokawa T. Isolation and Identification of Meiosis Inducing Substance in Starfish Asterias amurensis Nature 1969 221 273 274 10.1038/221273a0 5812580 2. Meijer L. Guerrier P. Maturation and Fertilization in Starfish Oocytes Int. Rev. Cytol. 1984 86 129 196 10.1016/s0074-7696(08)60179-5 6423562 3. Kishimoto T. MPF-based meiotic cell cycle control: Half a century of lessons from starfish oocytes Proc. Jpn. Acad. Ser. B 2018 94 180 203 10.2183/pjab.94.013 29643273 4. Chiba K. Oocyte Maturation in Starfish Cells 2020 9 476 10.3390/cells9020476 32092921 5. Just E.E. The Biology of the Cell Surface P. Blakiston's Son & Co., Inc. Philadelphia, PA, USA 1939 6. Fujimori T. Hirai S. Differences in starfish oocyte susceptibility to polyspermy during the course of maturation Biol. Bull. 1979 157 249 257 10.2307/1541052 7. Santella L. Limatola N. Chun J.T. Actin Cytoskeleton and Fertilization in Starfish Eggs Sexual Reproduction in Animals and Plants--Part II: Gametogenesis, Gamete Recognition, Activation, and Evolution Sawada H. Inuoe N. Iwano M. Springer Open Tokyo, Japan 2014 141 155 10.1007/978-4-431-54589-7_13 8. Limatola N. Vasilev F. Chun J.T. Santella L. Altered actin cytoskeleton in ageing eggs of starfish affects fertilization process Exp. Cell Res. 2019 381 179 190 10.1016/j.yexcr.2019.05.007 31082375 9. Schroeder T.E. Stricker S.A. Morphological changes during maturation of starfish oocytes: Surface ultrastructure and cortical actin Dev. Biol. 1983 98 373 384 10.1016/0012-1606(83)90366-4 6683686 10. Kyozuka K. Chun J.T. Puppo A. Gragnaniello G. Garante E. Santella L. Actin cytoskeleton modulates calcium signaling during maturation of starfish oocytes Dev. Biol. 2008 320 426 435 10.1016/j.ydbio.2008.05.549 18599031 11. Limatola N. Chun J.T. Kyozuka K. Santella L. Novel Ca2+ increases in the maturing oocytes of starfish during the germinal vesicle breakdown Cell Calcium 2015 58 500 510 10.1016/j.ceca.2015.08.002 26318405 12. Lim D. Kyozuka K. Gragnaniello G. Carafoli E. Santella L. NAADP+ initiates the Ca2+ response during fertilization of starfish oocytes FASEB J. 2001 15 2257 2267 10.1096/fj.01-0157com 11641253 13. Hirai S. Nagahama Y. Kishimoto T. Kanatani H. Cytoplasmic Maturity Revealed by the Structural Changes in Incorporated Spermatozoon during the Course of Starfish Oocyte Maturation Dev. Growth Differ. 1981 23 465 478 10.1111/j.1440-169X.1981.00465.x 14. Chiba K. Kado R.T. Jaffe L.A. Development of calcium release mechanisms during starfish oocyte maturation Dev. Biol. 1990 140 300 306 10.1016/0012-1606(90)90080-3 2373255 15. Jaffe L.A. Terasaki M. Structural changes in the endoplasmic reticulum of starfish oocytes during meiotic maturation and fertilization Dev. Biol. 1994 164 579 587 10.1006/dbio.1994.1225 8045353 16. Lim D. Lange K. Santella L. Activation of oocytes by latrunculin A FASEB J. 2002 16 1050 1056 10.1096/fj.02-0021com 12087066 17. Lim D. Ercolano E. Kyozuka K. Nusco G.A. Moccia F. Lange K. Santella L. The M-phase-promoting Factor Modulates the Sensitivity of the Ca2+ Stores to Inositol 1,4,5-Trisphosphate via the Actin Cytoskeleton J. Biol. Chem. 2003 278 42505 42514 10.1074/jbc.M301851200 12867432 18. Vasilev F. Limatola N. Park D.-R. Kim U.-H. Santella L. Chun J.T. Disassembly of Subplasmalemmal Actin Filaments Induces Cytosolic Ca2+ Increases in Astropecten aranciacus Eggs Cell. Physiol. Biochem. 2018 48 2011 2034 10.1159/000492523 30099438 19. Longo F.J. Woerner M. Chiba K. Hoshi M. Cortical changes in starfish (Asterina pectinifera) oocytes during 1-methyladenine-induced maturation and fertilisation/activation Zygote 1995 3 225 239 10.1017/S0967199400002628 8903792 20. Santella L. Limatola N. Vasilev F. Chun J.T. Maturation and fertilization of echinoderm eggs: Role of actin cytoskeleton dynamics Biochem. Biophys. Res. Commun. 2018 506 361 371 10.1016/j.bbrc.2018.09.084 30297103 21. Santella L. Limatola N. Chun J.T. Cellular and molecular aspects of oocyte maturation and fertilization: A perspective from the actin cytoskeleton Zool. Lett. 2020 6 5 21 10.1186/s40851-020-00157-5 22. Miyazaki S.-I. Hirai S. Fast polyspermy block and activation potential: Correlated changes during oocyte maturation of a starfish Dev. Biol. 1979 70 327 340 10.1016/0012-1606(79)90031-9 478165 23. Santella L. Limatola N. Chun J.T. Calcium and actin in the saga of awakening oocytes Biochem. Biophys. Res. Commun. 2015 460 104 113 10.1016/j.bbrc.2015.03.028 25998739 24. Lange K. Microvillar Ca++ signaling: A new view of an old problem J. Cell. Physiol. 1999 180 19 34 10.1002/(SICI)1097-4652(199907)180:1<19::AID-JCP3>3.0.CO;2-K 10362014 25. Lange K. Microvillar Ion Channels: Cytoskeletal Modulation of Ion Fluxes J. Theor. Biol. 2000 206 561 584 10.1006/jtbi.2000.2146 11013115 26. Dan J.C. Studies on the acrosome: VI. Fine structure of the starfish acrosome Exp. Cell Res. 1960 19 13 28 10.1016/0014-4827(60)90033-1 13813919 27. Dale B. Dan-Sohkawa M. De Santis A. Hoshi M. Fertilization of the starfish Astropecten aurantiacus Exp. Cell Res. 1981 132 505 510 10.1016/0014-4827(81)90132-4 7194189 28. Puppo A. Chun J.T. Gragnaniello G. Garante E. Santella L. Alteration of the Cortical Actin Cytoskeleton Deregulates Ca2+ Signaling, Monospermic Fertilization, and Sperm Entry PLoS ONE 2008 3 e3588 10.1371/journal.pone.0003588 18974786 29. Hoshi M. Moriyama H. Matsumoto M. Structure of acrosome reaction-inducing substance in the jelly coat of starfish eggs: A mini review Biochem. Biophys. Res. Commun. 2012 425 595 598 10.1016/j.bbrc.2012.08.033 22925680 30. Santella L. Lim D. Moccia F. Calcium and fertilization: The beginning of life Trends Biochem. Sci. 2004 29 400 408 10.1016/j.tibs.2004.06.009 15362223 31. Santella L. Puppo A. Chun J.T. The role of the actin cytoskeleton in calcium signaling in starfish oocytes Int. J. Dev. Biol. 2008 52 571 584 10.1387/ijdb.072560ls 18649271 32. Santella L. Vasilev F. Chun J.T. Fertilization in echinoderms Biochem. Biophys. Res. Commun. 2012 425 588 594 10.1016/j.bbrc.2012.07.159 22925679 33. Santella L. Chun J.T. Structural actin dynamics during oocyte maturation and fertilization Biochem. Biophys. Res. Commun. 2022 633 13 16 10.1016/j.bbrc.2022.09.001 36344149 34. Chun J.T. Limatola N. Vasilev F. Santella L. Early events of fertilization in sea urchin eggs are sensitive to actin-binding organic molecules Biochem. Biophys. Res. Commun. 2014 450 1166 1174 10.1016/j.bbrc.2014.06.057 24960199 35. Vasilev F. Limatola N. Chun J.T. Santella L. Contributions of suboolemmal acidic vesicles and microvilli to the intracellular Ca2+ increase in the sea urchin eggs at fertilization Int. J. Biol. Sci. 2019 15 757 775 10.7150/ijbs.28461 30906208 36. Limatola N. Chun J.T. Cherraben S. Schmitt J.-L. Lehn J.-M. Santella L. Effects of Dithiothreitol on Fertilization and Early Development in Sea Urchin Cells 2021 10 3573 10.3390/cells10123573 34944081 37. Limatola N. Chun J.T. Santella L. Species-Specific Gamete Interaction during Sea Urchin Fertilization: Roles of the Egg Jelly and Vitelline Layer Cells 2022 11 2984 10.3390/cells11192984 36230946 38. Limatola N. Vasilev F. Chun J.T. Santella L. Sodium-mediated fast electrical depolarization does not prevent polyspermic fertilization in Paracentrotus lividus eggs Zygote 2019 27 241 249 10.1017/S0967199419000364 31397235 39. Limatola N. Chun J.T. Santella L. Effects of Salinity and pH of Seawater on the Reproduction of the Sea Urchin Paracentrotus lividus Biol. Bull. 2020 239 13 23 10.1086/710126 32812816 40. Limatola N. Chun J.T. Santella L. Regulation of the Actin Cytoskeleton-Linked Ca2+ Signaling by Intracellular pH in Fertilized Eggs of Sea Urchin Cells 2022 11 1496 10.3390/cells11091496 35563801 41. Chun J.T. Puppo A. Vasilev F. Gragnaniello G. Garante E. Santella L. The Biphasic Increase of PIP2 in the Fertilized Eggs of Starfish: New Roles in Actin Polymerization and Ca2+ Signaling PLoS ONE 2010 5 e14100 10.1371/journal.pone.0014100 21124897 42. Limatola N. Vasilev F. Santella L. Chun J.T. Nicotine Induces Polyspermy in Sea Urchin Eggs through a Non-Cholinergic Pathway Modulating Actin Dynamics Cells 2020 9 63 10.3390/cells9010063 43. Chun J.T. Vasilev F. Limatola N. Santella L. Fertilization in Starfish and Sea Urchin: Roles of Actin Cell Differ. 2018 65 33 47 10.1007/978-3-319-92486-1_3 44. Steinhardt R.A. Mazia D. Development of K+-Conductance and Membrane Potentials in Unfertilized Sea Urchin Eggs after Exposure to NH4OH Nature 1973 241 400 401 10.1038/241400a0 4735605 45. Terasaki M. Actin filament translocations in sea urchin eggs Cell Motil. Cytoskelet. 1996 34 48 56 10.1002/(SICI)1097-0169(1996)34:1<48::AID-CM5>3.0.CO;2-E 46. Limatola N. Chun J.T. Santella L. Fertilization and development of Arbacia lixula eggs are affected by osmolality conditions Biosystems 2021 206 104448 10.1016/j.biosystems.2021.104448 34058296 47. Kyozuka K. Chun J.T. Puppo A. Gragnaniello G. Garante E. Santella L. Guanine Nucleotides in the Meiotic Maturation of Starfish Oocytes: Regulation of the Actin Cytoskeleton and of Ca2+ Signaling PLoS ONE 2009 4 e6296 10.1371/journal.pone.0006296 19617909 48. Chun J.T. Santella L. The actin cytoskeleton in meiotic maturation and fertilization of starfish eggs Biochem. Biophys. Res. Commun. 2009 384 141 143 10.1016/j.bbrc.2009.04.087 19393631 49. A Begg D.A. I Rebhun L. pH regulates the polymerization of actin in the sea urchin egg cortex J. Cell Biol. 1979 83 241 248 10.1083/jcb.83.1.241 41844 50. Grandin N. Charbonneau M. Changes in intracellular free calcium activity in Xenopus eggs following imposed intracellular pH changes using weak acids and weak bases Biochim. Biophys. Acta (BBA)-Mol. Cell Res. 1991 1091 242 250 10.1016/0167-4889(91)90068-9 51. Mori M. Somogyi K. Kondo H. Monnier N. Falk H.J. Machado P. Bathe M. Nedelec F. Lenart P. An Arp2/3 Nucleated F-Actin Shell Fragments Nuclear Membranes at Nuclear Envelope Breakdown in Starfish Oocytes Curr. Biol. 2014 24 1421 1428 10.1016/j.cub.2014.05.019 24909322 52. Wesolowska N. Avilov I. Machado P. Geiss C. Kondo H. Mori M. Lenart P. Actin assembly ruptures the nuclear envelope by prying the lamina away from nuclear pores and nuclear membranes in starfish oocytes elife 2020 9 e49774 10.7554/eLife.49774 31989921 53. Sugizaki A. Sato K. Chiba K. Saito K. Kawagishi M. Tomabechi Y. Mehta S.B. Ishii H. Sakai N. Shirouzu M. POLArIS, a versatile probe for molecular orientation, revealed actin filaments associated with microtubule asters in early embryos Proc. Natl. Acad. Sci. 2021 11 8e2019071118 10.1073/pnas.2019071118 33674463 54. Dale B. de Santis A. Hoshi M. Membrane response to 1-methyladenine requires the presence of the nucleus Nature 1979 282 89 90 10.1038/282089a0 503193 55. Mohri T. Kyozuka K. Starfish oocytes of A. pectinifera reveal marked differences in sperm-induced electrical and intracellular calcium changes during oocyte maturation and at fertilization Mol. Reprod. Dev. 2022 89 3 22 10.1002/mrd.23544 34729824 56. Hoshi M. Nishigaki T. Ushiyama A. Okinaga T. Chiba K. Matsumoto M. Egg-jelly signal molecules for triggering the acrosome reaction in starfish spermatozoa Int. J. Dev. Biol. 1994 38 167 174 7981026 |
PMC10000583 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050684 cells-12-00684 Review COVID-19 and Multiple Sclerosis: A Complex Relationship Possibly Aggravated by Low Vitamin D Levels Fernandes de Souza William Danilo 1* da Fonseca Denise Morais 2+ Sartori Alexandrina 1+ Brandenburg Lars Ove Academic Editor 1 Department of Chemical and Biological Sciences, Institute of Biosciences, Sao Paulo State University (UNESP), Botucatu 18618-689, Brazil 2 Laboratory of Mucosal Immunology, Department of Immunology, Institute of Biomedical Sciences, University of Sao Paulo, Sao Paulo 05508-000, Brazil * Correspondence: [email protected] + These authors contributed equally to this work. 21 2 2023 3 2023 12 5 68431 12 2022 21 1 2023 08 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is an exceptionally transmissible and pathogenic coronavirus that appeared at the end of 2019 and triggered a pandemic of acute respiratory disease, known as coronavirus disease 2019 (COVID-19). COVID-19 can evolve into a severe disease associated with immediate and delayed sequelae in different organs, including the central nervous system (CNS). A topic that deserves attention in this context is the complex relationship between SARS-CoV-2 infection and multiple sclerosis (MS). Here, we initially described the clinical and immunopathogenic characteristics of these two illnesses, accentuating the fact that COVID-19 can, in defined patients, reach the CNS, the target tissue of the MS autoimmune process. The well-known contribution of viral agents such as the Epstein-Barr virus and the postulated participation of SARS-CoV-2 as a risk factor for the triggering or worsening of MS are then described. We emphasize the contribution of vitamin D in this scenario, considering its relevance in the susceptibility, severity and control of both pathologies. Finally, we discuss the experimental animal models that could be explored to better understand the complex interplay of these two diseases, including the possible use of vitamin D as an adjunct immunomodulator to treat them. SARS-CoV-2 COVID-19 multiple sclerosis immunopathogenesis vitamin D FAPESP2021/06881-5 CNPq scholarship313429/2020-0 307269/2017-5 JBS S.A.Improvement of Higher Education Personnel (CAPES)88882.495054/2020-01 D.M.F. was supported by FAPESP 2021/06881-5 and CNPq scholarship 313429/2020-0; A.S. was supported by JBS S.A. and CNPq scholarship 307269/2017-5; W.D.F.S. received scholarships from the Coordination for the Improvement of Higher Education Personnel (CAPES) 88882.495054/2020-01. pmc1. Introduction COVID-19 is a disease caused by the SARS-CoV-2 virus infection whose severity varies depending primarily on host conditions and specific mutations found in the several virus variants. Even though the lungs are primarily affected, other organs and tissues such as kidneys, heart, and the nervous system can be injured. Most of its pathophysiological findings have been attributed to a hyperinflammatory syndrome resulting mainly from a dysregulated innate immune response. This review focuses on the interrelationship between COVID-19 and multiple sclerosis (MS), an autoimmune pathology that damages the central nervous system (CNS). Initially, we will discuss the main COVID-19 clinical manifestations, its etiologic agent and immunopathogenesis, and the nervous system involvement which takes place in some patients. MS clinical manifestations, the most relevant stages of its immunopathogenesis and the recognized relevance of viral infections to its development will then be described. The possible role of low vitamin D levels for the development and severity of both diseases and the recommendation of the patient's supplementation to improve its anti-inflammatory potential will also be addressed. The final section of this review will be dedicated to briefly describe the current animal disease models that could be employed to investigate, simultaneously, both diseases or some aspects related to their immunopathogenesis. 2. COVID-19: Clinical Manifestations, Etiology, and Immunopathogenesis A novel coronavirus named 2019-nCoV or SARS-CoV-2 was originally detected in Wuhan, China, in 2019, but in a few months, it spread out to most countries, initiating a pandemic. This viral agent causes COVID-19 which predominantly affects the respiratory system, causing flu-like symptoms such as fever, cough, sore throat, dyspnea and fatigue . Variable disease severity is observed among patients; around 80% of the affected patients display mild symptoms or can even be asymptomatic, while about 15% may develop more severe symptoms. The remaining 5% of patients can evolve to severe pathological conditions characterized by acute respiratory distress syndrome (ARDS), septic shock, and multiorgan failures associated with an elevated risk of death . Evolution to more severe conditions has been especially linked to advanced age, existence of comorbidities such as hypertension, diabetes and heart diseases, and genetic and epigenetic factors . Although this infection is mostly characterized by a significant respiratory impairment, it can also trigger several extrapulmonary manifestations including thrombotic complications, myocardial dysfunction and arrhythmia, acute coronary syndromes, kidney injury, gastrointestinal symptoms, hepatocellular lesions, hyperglycemia and ketosis, neurologic alterations, visual disturbances and dermatologic complications . The pulmonary and extrapulmonary manifestations of COVID-19 have been mainly attributed to a direct virus damage, given that ACE2 (angiotensin-converting enzyme), which is the entry receptor for the SARS-CoV-2, is expressed in the lungs and in these other extrapulmonary tissues. Analogously to other coronaviruses, SARS-CoV-2 consists of four structural proteins: spike (S), membrane (M), envelope (E) and nucleocapsid (N). The spike protein comprises two functional subunits: S1, which binds to the target cell, and S2, which triggers the fusion between the viral and the target cell membrane. SARS-CoV-2 uses two host proteins to enter the target cell; the ACE2 that is used for the attachment to S1 and the transmembrane serine protease 2 (TMPRSS2) that activates the protease activity of S2. For detailed information about the molecular mechanisms involved in this initial interaction between the virus and the target cell, see . An overview of the viral structure and the initial process of interaction with pneumocytes are illustrated in Figure 1A,B, respectively. The first line of response against pathogens, including SARS-CoV-2, is the innate immunity. In the case of SARS-CoV-2, its recognition by tissue-resident immune cells within the lung provides a local immune response resulting in the recruitment of further cells from the blood. Innate immune cells, including monocytes, macrophages, polymorphonuclear cells (PMNs) and innate lymphoid cells (ILCs) express pattern recognition receptors (PRRs) which identify pathogen-associated molecular patterns (PAMPs) and danger-associated molecular patterns (DAMPs). SARS-CoV-2 is able to initiate the activation of innate immunity by interacting with various PRRs, especially toll-like receptors (TLRs), retinoic acid-inducible gene 1 (RIG)-like receptors (RLRs), nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs) and inflammasomes. PRR signaling triggered by SARS-CoV-2 induces the concurrent release of IFNs, mainly I and III types, and other pro-inflammatory cytokines such as TNF-a, interleukin-1 (IL-1), IL-6 and IL-18 . Together these cytokines will induce antiviral programs in target cells and potentiate the specific immune response, which will eventually control the infection . Contrasting with this well balanced and effective innate immunity, the evolution of SARS-CoV-2 infection to a severe condition has been associated with a reduced or delayed type I IFN response together with high levels of other pro-inflammatory cytokines and high viral titers . This defective IFN response has been attributed to inborn errors of type I IFN immunity , and to the presence of autoantibodies against this cytokine . Interestingly, a sustained increase in the levels of type I IFN in a later phase of the infection can also promote a poor clinical outcome . Indeed, the signaling mechanisms involved in early (beneficial) or delayed (deleterious) type I IFN production are distinct. A rapid detection of viral RNA by TLR3, 6, and 7 and RLRs triggers a protective response, whereas a later activation of the cGAS-STING by DNA leads to cell death and a damaging production of type I IFN. Much of COVID-19 severity has been attributed to immune dysregulation manifested by a low production of interferons, remarkable inflammatory response and delayed adaptive immune response. This subject has been intensively investigated and reviewed and will be mentioned here only briefly to reinforce possible connection routes between COVID-19 and MS. The hallmark of most severe cases of COVID-19 is a strong inflammatory process that may ultimately lead to organ failure and patient death . The cytokine storm, also known as cytokine release syndrome is characterized by the extensive activation of macrophages, dendritic cells (DCs), NK, B and T cells and the subsequent production of high levels of TNF-a, IL-1b, IL-6, IL-12, IL-18, IL-33, IFN-I, IFN-g, CCL2, CCL3, CCL5, CXCL8, CXCL9 and CXCL10 . Several components are probably involved in the cytokine storm associated with SARS-CoV-2, including the interaction of viral RNA and proteins with PRRs, the binding of the virus to ACE2 and inflammasome activation . It has been demonstrated that in human and mouse epithelial cells the SARS-CoV-2 spike protein binds to TLR2 and induces inflammation via the activation of the NF-kB pathway. The interaction of ACE2 with SARS-CoV-2 is followed by its reduced expression on the surface because it is internalized together with the virus. As the biological function of ACE2 is to inactivate angiotensin II, there is an increased serum level of this molecule. Increased angiotensin II contributes to COVID-19 severity by inducing specific autoantibodies whose presence correlates with blood pressure dysregulation and increased cytokine production . This strong inflammatory process is largely mediated by the NLRP3 system that promotes inflammation through the cleavage and activation of specialized molecules, including active caspase-1 (Casp1p20), IL-1b, and IL-18. The analysis of samples from moderate and severe COVID-19 cases indicated active NLRP3 inflammasome in PBMCs and tissues of postmortem patients . These authors also observed the correlation of serum inflammasome-derived products, such as Casp1p20 and IL-18 with a poor clinical evolution. According to these authors, inflammation prompted by NLRP3 is initiated by IL-1b which induces the secretion of TNF-a, IL-6 and IL-8 by monocytes. These cytokines determine the influx of PMNs into lung tissue, gasdermin D activation and the subsequent formation of neutrophil extracellular traps (NETs), which can recruit platelets and promote hypercoagulability. This crucial inflammasome role in COVID-19 pathogenesis has been investigated as a potential target for therapy. To that end, a plethora of inflammasome inhibitors, including natural products as well as already authorized drugs, should be tested in pre-clinical and clinical studies . Even though future studies are still required, clinical findings obtained in a randomized and double-blind placebo-controlled trial in which mefenamic acid was administered to ambulatory patients , showed that it significantly reduced their symptomatology in comparison to the placebo group. This efficacy was attributed to both the anti-viral and the anti-inflammatory properties of mefenamic acid. Concerning the NETs mentioned above, their formation is even more accentuated in most severe COVID-19 cases, what has been attributed to increased immature PMNs and the presence of anti-NET antibodies. In addition, these antibodies can impair NET clearance and possibly enhance virus-mediated thrombo-inflammation . IL-1b and IL-6 can also directly contribute to coagulation in the lung vasculature by decreasing adherens junctions in endothelial cells. Tissue factor positive extracellular vesicles (EVs) released by pyroptotic monocytes can also directly activate the clotting cascade and promote coagulation in COVID-19. In addition to the direct virus damage and the deleterious immune response, pieces of evidences reinforce the view that the gut-lung axis will affect both the susceptibility and efficacy of the immune response against the virus. It is well known that the virus affects mainly the respiratory system; however, the gastrointestinal system is also a critical target. Gastrointestinal manifestations such as nausea, vomiting and diarrhea are present in a high percentage of COVID-19 patients. These symptoms have been attributed to the infection of gut epithelial cells by SARS-CoV-2 and the local dysbiosis characterized by alterations in microbiota bacterial composition and diversity . The respiratory tract has its own microbiota and it was already demonstrated that infections by other respiratory viruses induce local inflammation which contributes to gut dysbiosis . A similar effect could be expected from a lung SARS-CoV-2 infection. Many patients have reported the persistence of symptoms as fatigue, exercise intolerance, dyspnea, muscle pain, insomnia, chest pain, anosmia, cough, and brain fog after the acute disease stage . This condition has been denominated Post-COVID-19 syndrome or Long-COVID-19. Interestingly, in addition to the degree of infection severity, antibiotic usage has been considered one of the main risk factors for Long-COVID-19 development . Antibiotic prescription, which is expected to be more common in severe COVID-19 patients, would alter the gut microbiota composition. This hypothesis is strongly sustained by evidence showing that antibiotics are major disruptors of gut microbiota . In addition, gut dysbiosis triggered by excessive antibiotic administration, together with poorly controlled glycaemia and not well-regulated steroid administration were also identified as risk factors for COVID-19-associated mucormycosis , a rare and lethal fungal infection. Despite the complex gut dysbiosis scenario induced by the virus itself, as demonstrated in both a hamster experimental model and human patients , which is aggravated by antibiotic use, there is already a variety of promising microbiota-oriented strategies being suggested as prophylactic or therapeutic interventions such as probiotics, prebiotics, microbiota-derived metabolites and even fecal transplantation . Notably, whether lung dysbiosis associated with SARS-CoV-2 infection or antibiotic usage impacts the poor outcomes of COVID-19 is still an open question. Besides the microbiota-mediated gut-lung communication axis, another important systemic axis of immune communication impacts COVID-19 and MS: the gut-brain axis, as addressed afterward in this review. 3. Neurological Involvement Associated with COVID-19 The neuropathology associated with COVID-19 is a complex condition related to the local presence of the virus, the induced local and peripheral immune responses and to alterations in the microbiota, the vascular and the coagulation systems. The following neurological manifestations have been reported in COVID-19 patients: headache, myalgia, dizziness and fatigue, described as mild; hyposmia, hypogeusia, visual disturbances, encephalopathy, epilepsy, paralysis and consciousness disorder, identified in moderate to severe cases; and cerebrovascular events, acute necrotizing encephalopathy, meningitis, encephalitis and Guillain-Barre syndrome, considered as severe conditions . The prevalence of these manifestations seems to be particularly increased in hospitalized patients . The pathogenesis of CNS infection by SARS-CoV-2 and the neurological complications are still poorly understood. Most of these symptoms have been attributed to the ingress of the virus into the nervous system. Neuro-invasion by SARS-CoV-2 has been confirmed by the virus detection in the cerebrospinal fluid of a patient suffering from Guillain-Barre syndrome , in infected brain organoids, in mice expressing human ACE2 and autopsies from deceased patients . Two major routes have been associated with this neuro-invasion: through peripheral neurons and by hematogenous dissemination . The peripheral nerve endings are believed to be the most common route used by SARS-CoV-2 to reach the CNS. The olfactory nerve is considered the major candidate because it is located very close to the olfactory epithelium which, by expressing ACE2 and TMPRSS2, allows initial virus replication . Indeed, in a non-human primate model, it was demonstrated that SARS-CoV-2 can invade the CNS primarily via the olfactory bulb . This process has been described as transcribrial route because it occurs across the cribriform plate of the ethmoid bone, followed by retrograde viral spread via transsynaptic transfer using an endocytosis or exocytosis mechanism and a rapid axonal transport. The virus could also gain access to the CNS via the vagal afferents from the upper airways and the enteric nervous system . Another possible route is through the virus's presence in the bloodstream from where it can reach the nervous system by a direct interaction with brain capillary cells or by means of an infected leukocyte. The detection of viral particles in capillary endothelial cells in the front lobe tissue obtained in a post mortem sample gives support to this interaction with endothelial cells . Infected leukocytes could also pass through the blood-brain barrier (BBB), acting as a Trojan horse. More recently, this hypothesis that infected cells could cross the BBB as a Trojan horse has been extended to include exosomes and high-density lipoproteins associated with SARS-CoV-2 . The most recurrent neuropathological findings in COVID-19 patients include microglial activation, lymphoid inflammation with a clear predominance of TCD8+ cells, astrogliosis, myelin loss, hypoxia-ischemic changes, brain infarcts and hemorrhage and mi-crothrombi . Part of these findings is due to the SARS-CoV-2 infection of microglia and neurons which express different receptors for spike as ACE2, ephrin (Eph) ligands and the Eph receptors, neuropilin 1 (NRP-1), P2X7 and CD147 . Similar to peripheral infection, CNS infection also triggers a cytokine and chemokine avalanche causing neurotoxicity, disruption of the neuroglia homeostasis and neuronal death . As previously addressed in item 2 of this review, the dissemination of the virus to the gastrointestinal system is an aggravating condition that can also affect the nervous system due to an altered microbiota gut-brain axis. It is of note that the gut-microbiota signatures shared by COVID-19 patients and neurological and psychiatric disorders have been described . Such signatures are characterized by a reduction in microbial diversity and richness, an expansion of opportunistic proinflammatory pathogens and a reduction in anti-inflammatory-promoting bacteria. One of the consequences of this disrupted axis is a decreased secretion of short-chain fatty acid (SCFA), whose anti-inflammatory ability is well recognized. Therefore, the potential benefit of direct SCFA supplementation or reliance on probiotics prescription is being suggested for COVID-19 patients . The disturbed synthesis of other gut-brain axis mediators, such as cytokines, 5-hydroxytryptamine and cholecystokinin, can additionally contribute to neurological manifestations during COVID-19 . Notably, the gut-brain axis is also dysfunctional in MS , disclosing another link in this already puzzling interplay. 4. Multiple Sclerosis: Clinical Manifestations and Immunopathogenesis Multiple sclerosis (MS) is classically described as an inflammatory and demyelinating disease originating from an autoimmune disturbance. It is characterized by multifocal and scattered lesions through the grey and white matter from the brain and spinal cord. A damaged myelin sheath commonly results in vision and coordination loss, muscle weakness, stiffness and spasms, pain, and changes in bladder and bowel function. An MS diagnosis is usually based on clinical presentation, supported by neuroimaging, and in some cases, by cerebrospinal fluid analysis to search for inflammatory markers and oligoclonal antibody bands . Even though MS is considered a single disease, it can manifest under different phenotypes. According to , this characteristic is due to its multifactorial etiology that includes a genetic predisposition together with environmental elements such as infectious agents, mainly viruses, and vitamin D (vitD) levels, as will be commented on later. Understanding how these factors affect this disease is fundamental because this is a handicapping pathology whose incidence and prevalence are increasing worldwide. Four basic disease courses, which can also be referred to as phenotypes or types, are recognized: clinically isolated syndrome, relapsing-remitting MS (RRMS), secondary progressive MS (SPMS) and primary progressive MS (PPMS). RRMS is considered the most common phenotype, affecting around 85% of patients. As indicated by its designation, it is characterized by alternating relapses and remissions, which are periods of neurological dysfunction or absence of neurological symptoms, respectively . A noteworthy finding is the strong association of relapses with infections, postpartum period, genetic risk factors, stress, and vitD levels . In addition, an increased relative risk for relapses has been associated with infections located in the upper respiratory system or affecting the urinary and gastrointestinal tracts . MS immunopathogenesis has been disclosed by using both patient samples and information from animal models, especially experimental autoimmune encephalomyelitis (EAE). This disease is induced in mice and rats by immunization with myelin-derived proteins and peptides in the presence of complete Freund's adjuvant and pertussis toxin . The sequence of events leading to the onset of MS is outlined in Figure 3 and is briefly described below. 4.1. Peripheral Activation of Myelin-Specific Lymphocytes It is accepted that the activation of myelin-specific T cells would occur in peripheral lymphoid organs by different mechanisms such as the recognition of microbial epitopes sharing homology with self-antigens (molecular mimicry), the release of myelin from the CNS by a local insult, or by bystander activation . The perspective that molecules derived from a dysbiotic gut microbiota could induce neuroinflammation and symptoms of MS by mimicking autoantigens has been interestingly proposed . Other mechanisms, including the induction of co-stimulation, polyclonal activation, altered processing and expression of cryptic antigens could also be induced by infectious agents and contribute to the onset of autoimmunity . 4.2. Presumed T Cell Licensing A fairly new concept has emerged from EAE studies in the last few years and is called "licensing" or more clearly "licensing for pathogenicity". This stage, which seems to occur in the lungs, spleen or maybe both, allows the T cells to become pathogenic, reach the CNS and orchestrate a local inflammatory process . This licensing process is accomplished by a change in the gene expression pattern marked by the downregulation of proliferation/activation-related genes and upregulation of migration-promoting genes . Th17 cells play a critical encephalitogenic role by opening the BBB and promoting neurodegeneration . 4.3. Expansion of Th17 in the Intestine The immune system associated with the intestinal mucosa has recently been recognized as pivotal to MS and EAE development. This crucial role in the pathogenesis occurs mainly by promoting the activation and acquisition of the Th17 phenotype . It is assumed that the activation of effector Th17 cells occurs mostly in the murine small intestine, regardless of their ensuing function . Interestingly, the development of steady-state or pathogenic Th17 cells is critically determined by microbiota composition. Segmented filamentous bacteria, for example, induce brain autoimmunity in mice by selectively privileging Th17 differentiation . 4.4. Breakdown of the Blood Barriers and Cell Migration to the CNS Two barriers protect the CNS integrity and functionality: the BBB and the blood-cerebrospinal fluid (B-CSF) barrier. They are located in distinct CNS compartments and their dysfunction can allow leukocyte access and the ensuing neurodegeneration in MS and EAE . The integrity of these barriers can be disturbed by peripheral inflammation, infections and proinflammatory cytokines. There are a number of mechanisms by which an exaggerated inflammatory process can induce the disruption of these barriers, including changes in tight junctions, damage to endothelial cells, the activation of astrocytes and microglia, and the effects of peripheral immune cells . The invasion of the CNS by neurotropic viruses, by hematogenous or non-hematogenous routes, can be associated with structural and functional BBB alterations that also lead to its breakdown . Specific cytokines have been more often linked with alterations in these barriers as shown by , which found that periodontal inflammation-induced IL-6 is associated with neuroinflammation and BBB disruption in the hippocampus. 4.5. Local Inflammation and Neurodegeneration A variety of infiltrating cells, mainly gdT, Th1 and Th17 cells, are locally expanded and release cytokines that will activate microglia and oligodendrocytes. Together with the ones mobilized from the periphery, these activated cells will release inflammatory mediators such as IL-8, IL-17, GM-CSF, CCL2 and enzymes that will trigger neurodegeneration . A great deal of contribution to this damaging process has been imputed to B cells, mitochondrial dysfunction, oxidative stress and inflammasome activation . This process will eventually be controlled by regulatory mechanisms involving different cell subsets such as Tregs CD25+Foxp3+, Tr1, Qa-1 restricted CD8, regulatory B cells, NK and CNS-derived myeloid-derived suppressor cells . However, defined viruses, namely EBV and HHV6, could trigger relapses through a peripheral mechanism rather than a direct effect through intrathecal multiplication . 5. Connection between MS and Virus (EBV and SARS-CoV-2) The Epstein-Barr virus (EBV) establishes lifelong infection, usually subclinical, in more than 90% of the adult population worldwide. However, it is also the causal agent of infectious mononucleosis, some types of cancer and severe lymphoproliferative diseases. Epidemiological, serological and virological pieces of evidences support its role also in MS development . Primary EBV infection occurs in the squamous epithelial cells where it replicates and from where it reaches and infects the B lymphocytes from Waldeyer s tonsillar ring. Consequently, naive B cells undergo a germinal center-like activation and differentiation program affecting more than 11,000 genes. This process culminates in proliferating immortalized B cells resembling plasmablasts and early plasma cells . Therefore, in the context of MS, EBV is understood as a disease trigger. This possibility was tested in a large cohort of more than 10 million US army personnel . A 32-fold increase in MS diagnosis in individuals who became seropositive, compared with those that remained seronegative, was observed. This role of EBV has been mainly attributed to cross-reactivity between self and EBV antigens, involving both cellular and humoral immunity . Additionally, as B cells and plasma cells have been identified in the brain of deceased MS patients, but not in controls , it is believed that induction of B cell trafficking to the CNS is also involved in this mechanism . The recognized efficacy of monoclonal antibodies such as rituximab, which targets the B cell surface marker CD20, adds more credibility to the contribution of EBV to MS development . Considering that certain deregulated immunological pathways found during severe COVID-19 coincide with immune alterations present in MS, it has been postulated that SARS-CoV-2 could be a risk factor for the triggering or worsening of MS in prone individuals. Through a system biology study, found the expressive interaction of SARS-CoV-2 with genes associated with autoimmunity, especially MS. In this scenario, these authors highlighted three intersecting pathways: type-1 IFN response, Th17 axis, and inflammasome pathway, which were considered critical in this COVID-19 vs MS interplay and that will be briefly commented on. The hypothesis that MS is a type of IFN I deficiency syndrome was initially proposed in and later expanded to encompass other autoimmune diseases as well . Type I IFN appears to play a pivotal role in the CNS, avoiding both severe infections, especially viral ones through its antiviral effect, and local inflammation through its immunomodulatory potential . In 1993, type I IFN, mainly the b-1a type, was adopted as the first disease-modifying therapy for MS . Coincidentally, dysregulated and/or delayed type I IFN responses are also associated with severe COVID-19 prognosis. Dysfunctional IFN I production can be caused by inborn errors, by the presence of anti-IFN I autoantibodies and by inhibition of type I IFN production by several SARS-CoV-2 proteins . In this context, one can speculate that MS patients with IFN I deficiencies would be more susceptible to SARS-CoV-2. To the best of our knowledge, this possibility has not been investigated yet. On the other hand, MS patients under type I IFN therapy would be more protected against severe forms of COVID-19. This possibility was reinforced in , which showed a lower, even though not significant risk of infection by this virus in these MS patients. Moreover, only minor COVID-19 symptoms were described in an MS patient under IFN I therapy . Given this scenario, we could presume that COVID-19 is potentially able, considering its ability to reach the CNS and to infect an IFN I deficient MS patient, to trigger an MS relapse or worsen disease symptoms. The Th1/Th17 axis, which is responsible for IFN-y and IL-17 production, is also shared by both pathologies. The contribution of these Th subsets to MS immunopathogenesis is strongly supported by the literature . Interestingly, the Th17 subset develops in the intestine and its higher frequency correlates with microbiota alterations . Coincidentally, severe COVID-19 is associated with high levels of IFN-y, Th17 polarization and gut microbiota dysbiosis . This scenario suggests that a Th1/Th17 active axis in COVID-19 or MS patients could aggravate MS or COVID-19 symptoms, respectively. Another critical link at the crossroad between COVID-19 and MS is the inflammasome, a complex molecular platform comprising a sensor protein, inflammatory caspases and, in some cases, an adapter protein that bridges the two other components. Its activation by DAMPs and PAMPs promotes IL-1b and IL-18 production and pyroptosis . Dysregulated inflammasome activation can be associated with infections and inflammatory pathologies. A strong contribution has been attributed to inflammasomes, especially the NLRP3, to the development of MS and its experimental animal model (EAE). NLRP3 activation is involved in various MS stages such as initial inflammation, T cell polarization, CNS barrier breakdown and neurodegeneration . The ability of SARS-CoV-2 to activate this platform has also been clearly demonstrated during COVID-19 and is accentuated in the more aggressive disease . This association makes sense, considering that a significant cause of COVID-19 pathogenesis and subsequent severity is the cytokine storm associated with NLRP3 overactivation . Despite the expectation of a deleterious effect of COVID-19 on MS manifestations, a retrospective study concluded that, regardless of its severity, COVID-19 was not associated with an increased risk of MS relapse shortly after infection . The authors attributed these non-expected results to post-COVID-19 lymphopenia and to the use of immunomodulatory drugs to control MS. However, the authors also do not rule out the possibility that COVID-19's deleterious effects will be observed later as sequels associated with post-COVID condition. The stages of MS immunopathogenesis that could most likely be deleteriously affected by SARS are schematically suggested in Figure 4. 6. Connection between Multiple Sclerosis and Vitamin D Numerous questions have been raised concerning the relationship between MS and vitD levels. Some of the most relevant ones, considering practical purposes, are: Is there a vitD deficiency in MS patients and is this a risk factor to develop this disease and to present a more severe pathology? Is vitD supplementation indicated for MS patients? How would vitD control MS pathogenesis? The literature data concerning the relationship between vitD and MS support the concepts that there is a vitD deficiency in these patients, that this is a risk factor for disease development and that this deficit probably contributes to a more severe pathology. The hypothesis that adequate vitD levels were relevant to preventing MS development emerged from the realization that this disease was more prevalent in geographical regions with a lower solar incidence where the production of this vitamin by the skin, stimulated by UV light, is low. The high prevalence of vitD deficiency in MS patients was formally described in 1994 . This finding was subsequently validated by other authors who also showed a significant correlation between this deficiency, MRI load and disease severity . Several studies have been conducted to determine if vitD supplementation could be considered as an add-on therapy for MS. A study performed in relapsing-remitting patients investigated its supplementation in patients under IFN-b-1b . Even though no difference has been observed in the annual relapse rate, the vitD supplemented group presented a reduced disease activity indicated by MRI. According to , the supplementation with vitD of MS patients under natalizumab treatment corrected hypovitaminosis and decreased annualized relapse rate. As vitD efficacy could rely on high doses and considering that this could trigger fatigue, muscle weakness, renal failure and cardiac arrhythmia, some clinical trials were designed to test possible deleterious outcomes. It was found, however, that even vitD supplements that elevated twice the top of physiological range did not elicit hypercalcemia, hypercalciuria or any other detrimental effect . Nonetheless, a clinical trial concerning the efficacy of a high vitD dose was inconclusive, neither supporting nor discrediting its potential benefit . So far, there is no substantial evidence to approve this vitamin as an add-on therapy for MS . Despite the literature discrepancies, these authors recommend implementing standardized study designs with well-defined variables concerning the kind of vitD supplement, its concentration, cohort composition, and the clinical and laboratory parameters to be evaluated. They also recommend the collection and storage of samples to assemble a bio-bank for further evaluations. Detailed reasoning and suggestions regarding their proposed study design are available in their publication. Our research team has been testing vitD efficacy in EAE. Our findings indicate a clear window of opportunity for therapy with vitD in an animal model which could also be relevant for the human disease. We found that vitD is still effective when administered during the preclinical disease phase. However, it is important to clarify that earlier supplementation determines a more pronounced therapeutic effect . It is well established that vitD signaling pathways are able to regulate both innate and adaptive immunity, keeping the associated inflammatory response within physiological limits. The immunomodulatory ability of vitD is clearly pleiotropic and reaches the majority of the immune cells in different phases of the immune response. This is explained by its interaction with the vitamin D receptor (VDR) expressed in immune cells including PMNs, macrophages, DCs, and B and T lymphocytes . Understanding this essential VitD biological activity has sparked much interest in two aspects: vitD level in patients with inflammatory diseases and the possibility of its supplementation as a therapeutic measure. The interaction of the active form of vitD with VDR and the main effects over the immune system are illustrated in Figure 5. An overview of the main outcomes from vitD interaction with immune cells is presented in Table 1. Its potential as an adjunct or alternative therapy in inflammatory diseases is exemplified in Table 1. The presumed protective effect of vitD on MS may occur in both the periphery and the CNS. As previously mentioned in this review, it has been suggested that the immune response against neural antigens is initiated in the peripheral lymphoid organs. VitD could already be effective at this initial stage by modulating the differentiation and function of APCs. It has been widely demonstrated that vitD affects the differentiation, maturation and function of DCs, directing them to a tolerogenic profile . These APCs will then modulate naive TCD4+ lymphocytes toward a functional hypo-reactive state. VitD-induced tolerogenic DCs are also capable of driving the differentiation of Tregs. This effect was demonstrated in EAE by the adoptive transference of vitD-induced IDO+ DCs, which are immature and tolerogenic cells. This procedure increased the percentage of CD4+CD25+Foxp3+ Tregs in the lymph nodes of rats with EAE . The ability of VitD to decrease proliferation and differentiation of effector proinflammatory T cells in EAE was also demonstrated . This effect was associated with the downmodulation of various metabolic and signaling routes which are essential for Th1/Th17 polarization. This inhibition was concurrent with reduced DNA methylation and the upregulation of many non-coding RNA classes. One of the critical stages of MS immunopathogenesis is the BBB breakdown which can be detected in patients via gadolinium-enhanced MRI in the CNS . An effect at this level is expected, considering that BBB endothelial cells express VDR. Our research team showed that calcitriol administration to a murine model of MS decreases neuroinflammation and reduces BBB disruption . Other experimental pieces of evidence have disclosed the molecular mechanisms underlying this protective effect. By using an in vitro system comprising a human brain microvascular endothelial cell line stimulated with TNF-a or sera from MS patients, the authors in showed that vitD determined the upregulation of tight junction proteins and downregulation of cell adhesion molecules. VitD can also downmodulate neuroinflammation by targeting additional CNS cells. It decreases, for example, the production of TNF-a, IL-1b and the expression of IL-4 by astrocytes stimulated with LPS in vitro. This likely outcome was reproduced in neonatal rats injected with LPS . The addition of vitD to stimulated microglial cells reduces the expression of Iba-1, MHC-II, CD86 and TLR-4 in vitro, and in EAE . By using a model of demyelination in rats, the authors in described that VitD also operates at the level of remyelination. VitD promotes the proliferation and differentiation of neural stem cells and their migration to the lesion site, where they subsequently differentiate into oligodendrocyte lineage cells and produce myelin basic protein. Lastly, the possibility that vitD is also acting through inflammasome inhibition must be considered. This system operates in the periphery during the initial induction/expansion of Th1/Th17 cells and also during the inflammatory process in the CNS. The demonstration in that vitD negatively regulates the NLRP3 inflammasome via VDR signaling, effectively inhibiting IL-1b secretion, gives some support to this additional level of vitD protection in MS. 7. Connection between COVID-19 and Vitamin D Analogous to the approach used to analyze the relationship between MS and vitD, we also limited the complex connection between COVID-19 and vitD to answer the same three questions: Is there vitD deficiency in COVID-19 patients and is this a risk factor for getting infected and to develop a more severe pathology? Is vitD supplementation indicated for COVID-19 patients? How would vitD control COVID-19 pathogenesis? An analysis between vitD levels and the number of COVID-19 cases in 20 European countries showed a significant negative correlation, suggesting that higher levels of this hormone could afford some protection against SARS-CoV-2 infection . These findings were reinforced by the observation that many hospitalized COVID-19 patients presented vitD serum levels considered below the normal expected ones . The majority of the findings also support an inverse correlation between vitD deficiency and a poor prognosis for COVID-19. According to , low vitD levels at hospital admission were associated with increased IL-6 production and predicted the severity of respiratory distress and mortality during the course of hospitalization. This association between vitD deficiency and severe COVID-19 has been confirmed by several other researchers . From a theoretical point of view, based mainly on vitD's immunomodulatory properties, its adoption as an adjunct therapy for COVID-19 seems consistent. For instance, the authors in have investigated the effect of oral vitD supplementation in mild to moderate COVID-19 patients with low levels of this vitamin. They observed that 5000 IU of vitD reduced the recovery time related to cough and loss of taste and smell. The adequate levels of vitD in the host have been associated with the reduced release of proinflammatory cytokines, thus lowering the risk of a cytokine storm; increased levels of anti-inflammatory cytokines; and enhanced secretion of natural antimicrobial peptides. It may also be involved in the enhancement of the Th2 immune response and activation of defensive cells such as macrophages, as illustrated in Figure 5. Contrary to these findings, several studies have concluded that there is no direct association between vitD concentrations and a poor prognosis of the disease . By employing a meta-analysis and GRADE assessment of cohort studies and RCTs, the authors of inferred that low vitD levels do not play a role in disease severity and that supplementation does not improve outcomes in hospitalized patients. The explanation for these conflicting results could be partially related to inter-cohort variability. Other parameters including supplementation protocols such as doses, period of supplementation, patient's age, presence of comorbidities and even the risk of bias, could contribute to this variability. This ambiguous scenario has prevented an official recommendation concerning the prophylactic or therapeutic use of vitD for COVID-19 control . The presumptive ability of vitD to control SARS-CoV-2 infectivity and COVID-19 severity would be mediated by different mechanisms. Some of them are related to the capacity of this hormone to increase the production of antimicrobial peptides, in particular cathelicidin antimicrobial peptide, also known as LL-37 . LL-37 is produced by immune cells and epithelial cells from the skin and respiratory tract. Experimental data strongly suggests that this peptide can inhibit SARS-CoV-2 infection and other alterations that contribute to disease severity. Human cathelicidin can inhibit virus infection by directly interacting with the SARS-CoV-2 RDB and also by masking the ACE2 . According to , a plethora of other biological activities have been ascribed to LL-37 and could contribute to its eventual preventive and therapeutic adoption against COVID-19. In this sense, this peptide is endowed with an immunomodulatory ability, could facilitate efficient NET clearance by macrophages and speed endothelial repair. These authors also addressed the fact that further investigations about the VitD/LL-37 axis in the context of COVID-19 are highly recommended considering that vitD could be a widely accessible strategy. One of the hallmarks of severely affected COVID-19 patients is the presence of a cytokine storm, mainly triggered by the activation of cells from innate immunity. The well-established ability of vitD to directly control cytokine and chemokine production could provide another mechanism for vitD usefulness as an adjunct therapy for COVID-19. This effect derives mostly from the downmodulatory ability of vitD over Th1 and Th17 differentiation and cytokine production . This mechanism is already suggested by clinical findings in COVID-19 patients. The authors in described that vitD supplementation in geriatric intensive care patients suffering from COVID-19 reduced many inflammatory parameters, including IL-16, C-reactive protein, procalcitonin, D-Dimer, ferritin and lactate dehydrogenase. Its activity against endothelial dysfunction , and vascular thrombosis could also contribute to the ability to control COVID-19 immunopathogenesis. 8. Experimental Animal Models to Decipher the Complex COVID-19 and MS Interplay Many questions concerning the relationship between COVID-19 and MS have already been raised and partially answered by experts in the field. We believe, however, that experimental animal models could add more knowledge to the remaining gaps of the complex interplay between these two pathologies. Insightful and updated reviews have been published regarding the most useful animal models to investigate, separately, these two pathologies . For the sake of objectivity, only the models that seem to be immediately or more easily available to investigate aspects concomitantly related to these two pathologies will be briefly described, as summarized in Figure 6. Syrian hamsters (Mesocricetus auratus) are widely used in the research of respiratory viruses. In addition, their ACE2 receptor binds tightly to SARS-CoV-2 which makes then naturally susceptible to infection by this virus. The experimental trans-nasal inoculation of SARS-CoV-2 in 4-8-week-old hamsters triggers a reproducible infection characterized by a short-term, self-limiting, epitheliotropic infection of the lungs and intestine with almost complete elimination of the virus before 14 days post infection. Details of these lesions, which are similar to the ones found in humans infected with SARS-CoV-2, were described in . This experimental disease can progress with different degrees of severity, depending upon the hamster strain . It was recently described in that hamsters develop a condition that clearly resembles the post-acute sequels of COVID-19. After virus clearance, these animals presented a clear inflammatory process in both the olfactory bulb and the olfactory epithelium. This process included the activation of myeloid and T cells, and the production of proinflammatory cytokines, including IFN-y. We believe that this is an interesting model to be explored in the context of these two diseases. The use of hamsters to model MS are scarce. However, older publications by the authors of showed that Syrian hamsters immunized with guinea pig spinal cord derived antigen, in the presence of adjuvants, developed a chronic paralysis after 50-100 days, which was often relapsing. Seventy percent of these animals presented mononuclear cell infiltration and focal demyelination in the neuraxis also demonstrated that the susceptibility of these rodents to develop EAE was highly dependent on the specific inbred strain, with some being able to develop acute paralysis around 10-21 days after immunization. Interesting, this model was already used to test the possible interference of a virus on EAE development showed that the persistence of the measles virus in the CNS exacerbated EAE manifestations. For many reasons, including the availability of reagents to perform immunological characterizations, mice constitute the first choice for these investigations. However, wild-type murine ACE2 does not bind adequately to the viral spike protein, rendering them resistant to the infection . Different strategies have been engendered to overcome this obstacle. We believe that transgenic mice expressing human ACE2 would be worthwhile to be tested, considering that COVID-19 severity could be controlled by the level of ACE2 expression which would allow a more precise investigation about the potential role of this infection on EAE aggravation. These transgenic mice need to have a C57BL/6 and SJL background to be able to develop the classic EAE pathology. To the best of our knowledge, the suitability of ACE2 transgenic C57BL/6 and SJL mice strains to develop EAE was not tested yet. This investigation is mandatory considering that transgenesis could alter the evolution of EAE in these animals. The most employed animal model for MS studies is experimental autoimmune encephalomyelitis (EAE). Murine EAE is usually induced by active immunization with myelin-derived peptides emulsified with complete Freund's adjuvant in the presence of pertussis toxin. C57BL/6 and SJL/J mice strains immunized with specific immunodominant peptides develop a chronic and a relapsing-remitting form of the disease, respectively . EAE in mice is characterized by an ascending paralysis that starts by the tail, followed by limb and forelimb paralysis, and its clinical severity can be easily classified by a clinical score based on a five-point scale , together with weight loss. The immunization of SJL/J mice with PLP139-151 can result in an initial paralytic attack, followed by multiple remissions and relapses, whereas immunization of C57BL6/J mice with MOG35-55 usually causes a chronic disease course in which an initial attack does not resolve. Caenorhabditis elegans (C. elegans) is a nematode species which has been increasingly employed as a model to investigate human diseases. This has been possible because humans and C. elegans share some identity concerning the digestive, the nervous and the reproductive systems. Indeed, many important signaling pathways are highly conserved between this worm and humans. Even though this worm lacks the classic adaptive immunity system, which is typical of vertebrates, it is endowed with a variety of innate mechanisms that have been studied to understand microbe-host interactions, originally during bacterial infections . Later on, it was discovered that C. elegans could be also employed to investigate anti-viral defense mechanisms . This was demonstrated by using both natural viruses such as Orsay, Santeuil and Le Blanc and non-natural ones such as Flock House and stomatitis virus . Similar to the murine models, this nematode could be adapted to SARS-CoV-2 research by expressing the human ACE2 receptor and TMPRSS2 co-factor. In the context of this review, this transgenic nematode model could be especially useful to investigate alterations in innate immunity and the nervous system associated with SARS-CoV-2 infection. As far as we know, C. elegans has not being directly used to study aspects related to MS; however, it is a well-established model to investigate neurodegenerative diseases in a general way, offering many advantages over other model systems to decipher the involved mechanisms. Of particular importance for the study of neurodegenerative processes are the nematode's small genome, the anatomical simplicity and the availability of a complete 3D map of the 302-cell nervous system . It has been suggested that the control of more severe COVID-19 cases will require a poly-therapeutic approach including both anti-viral and anti-inflammatory medicines. According to , C. elegans was recently included as an additional system to establish a combination therapy platform to treat COVID-19. Interestingly, C. elegans express DAF-12 that is a nuclear hormone receptor which is homologous to the VDR expressed in human cells. The authors in demonstrated that the uptake of vitD by C. elegans via their traditional E. coli food source results in a significantly extended lifespan. The capacity of this worm to respond to vitD could be additionally useful in studies involving neurodegeneration considering that this hormone has a well-defined neuroprotective role . The characteristics of this nematode model system which includes relative simplicity, ease of use, exquisite genetics, and an available genomic sequence, provides an extremely useful model system in many areas of study. Indeed, many important signaling pathways are highly conserved between C. elegans and humans; this worm has more than 7500 genes with human homologs . Having in mind that most of COVID-19's pathogenesis is due to a hyperinflammatory reaction and that this process can affect MS, a few models of inflammation induction by virus antigens are also succinctly described. Ref. observed that the spike protein is able to induce inflammatory cytokines and chemokines, including IL-6, IL-1b, TNF-a, CXCL1, CXCL2 and CCL2, but not IFNs, in human cells, in mouse macrophages or lung epithelial cells. The potential of the spike protein to induce inflammation in vivo was shown by . However, the most relevant data from their work was the characterization of a lung inflammation model induced by coadministration of aerosolized S protein and LPS to the lungs. This procedure triggered a strong pulmonary inflammation and a cytokine profile similar to that observed in more severe COVID-19. According to the authors, this model mimics better the more stringent lung involvement in patients with comorbidities such as diabetes, obesity and chronic obstructive pulmonary disease. These patients frequently present abnormal gut permeability allowing the translocation of LPS through the gut epithelia and, therefore, its availability to interact with the virus spike. The ability of the spike to strongly bind to LPS and boost the proinflammatory activity was previously demonstrated in . The whole inactivated virus has also been employed to elicit inflammation. The intratracheal instillation of human ACE2-transgenic mice with formaldehyde-inactivated SARS-CoV-2 caused weight loss and pulmonary pathologic alterations such as consolidation, hemorrhage, necrotic debris and hyaline membrane formation. IL-1b, TNF-a, IL-6 and the infiltration of activated neutrophils, inflammatory monocytes, macrophages and T cells were also detected in the lungs . We recently established a model of lung inflammation via the intranasal instillation of UV-inactivated SARS-CoV-2. This procedure triggered an exuberant inflammatory process composed of various cell types and mediators similar to lung inflammation associated with COVID-19 . This inflammatory process was significantly downmodulated by intranasal vitD administration, suggesting that this hormone has the potential to be an adjunct therapy for COVID-19. In addition, considering our previous data that support a strong protective effect of vitD on EAE development , we believe that IN vitD administration could downmodulate inflammatory reactions occurring simultaneously in the lungs and the CNS. 9. Conclusions COVID-19 and MS are associated with several immunological disturbances that could, theoretically, interfere with each other's disease onset or outcome. SARS-CoV-2 displays, for example, molecular mimicry with CNS epitopes and causes microbiota and BBB disruption which are crucial for MS development. This virus can also reach and inflame the CNS itself which is the target of the autoimmune inflammatory reaction that characterizes MS. These two pathologies also share a possible type I IFN deficient production and hyperactivation of both the Th1/Th17 axis and the NLRP3 inflammasome platform which could mutually cause disease aggravation. The role of vitD levels in susceptibility, severity and possible adjunctive therapy in both diseases have been investigated and highly discussed but not well-established yet. This complex interplay between COVID-19 and MS urgently needs further and in-depth investigations. A plethora of experimental animal models, usually employed to study each of these pathologies individually, as is the case of C. elegans, hamster strains and transgenic mice, could be explored to investigate aspects related to both diseases simultaneously. These disease models could not only complement the current knowledge but also possible future questions, bearing in mind that more severe neurological changes associated with long-term COVID are possible. Author Contributions All authors equally contributed for the conceptualization, construction and revision of the manuscript. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Schematic representation of SARS-CoV-2 structure and host cell invasion. (A) This is an enveloped, positive-sense RNA virus containing the following main structural proteins: spike (S) and membrane (M) glycoproteins, and envelope (E) and nucleocapsid (N) proteins. (B) Virus exposition occurs primarily through the upper airways, with tracheal and lung cells being the primary targets of infection. The interaction with these cells involves the spike protein expressed on the surface of the viral particle and comprising the S1 and S2 domains which interact with the host membrane proteins ACE2 and TRPMSS2, respectively, resulting the virus/cell fusion. Once the processes of fusion and the passage of the virus's genetic material into the cell are completed, replication starts. Source: Created with Biorender.com. Figure 2 Development of systemic inflammation in the severe COVID-19. (A) After infecting lung epithelial cells, innate receptors distributed in the several lung immune cells recognize viral components and promote the secretion of several cytokines and chemokines. Such immune mediators act locally by promoting local inflammation and recruitment of inflammatory cells, such as neutrophils and monocytes, and by activating adaptive immunity. (B) Finally, in the severe forms of the disease, the overproduction of inflammatory mediators and lack of regulatory mechanisms favor the dissemination of the inflammatory process, which becomes systemic. (C) This process is called a cytokine storm, which drives the development of severe lung damage (D) and multiple organ dysfunction (E). Source: Created with Biorender.com. Figure 3 Immunopathogenesis of multiple sclerosis/experimental autoimmune encephalomyelitis. (A) Activation of self-reactive T cells specific for myelin antigens in secondary lymphoid organs, (B) licensing of self-reactive cells in the lungs and spleen (C), differentiation of Th17 cells in the intestine, (D) disruption of the blood-brain barrier and cell migration to the central nervous system, (E) local reactivation and expansion of Th cells, (F) local inflammatory process that leads to demyelination and neurodegeneration, (G) cells and molecules that mediate the control of disease via regulatory mechanisms. Source: Created with Biorender.com. Figure 4 Stages of MS immunopathogenesis which could be affected by SARS-CoV-2. Source: Created with Biorender.com. Figure 5 Immunomodulatory effects of vitamin D on innate and adaptive immunity. (A) Calcifediol (25(OH)D3) becomes biologically active forming calcitriol (1,25(OH)2D3) through two consecutive hydroxylations, with the last one being performed by 1a-hydroxylase (CYP27B1) which is present in numerous cells of innate immunity. (B) Calcitriol's biological actions are mediated through binding to the VDR. This binding induces a conformational change that facilitates interaction with RXR and the coregulatory complexes required for the transcription of target genes. (C) Downmodulatory effects of vitD on adaptive immunity. Source: Crated with Biorender.com. Figure 6 Infection models and immunological/histopathological studies. Source: Created with Biorender.com. cells-12-00684-t001_Table 1 Table 1 Impact of vitamin D on immune response cells. The arrows | and | indicate reduction or increase in expression, respectively. Cell Type/Source Experimental Model Treatment Main Outcome Dendritic cells PENNA and ADORINI 2000 BERER et al., 2000 Cell culture from peripheral blood monocytes 1,25 (OH)2D3 added to the culture medium Inhibition of differentiation and maturation Apoptosis induction | MHC II, CD40, CD80, CD86, IL-12 and IL-23 | IL-10 expression Macrophages VERWAY et al., 2013 LIU et al., 2006 GOMBARD; BORREGAARD; KOEFLER, 2005 Co-culture of macrophages and human lung epithelial cells Culture of bone marrow cells from humans and mice 1,25 (OH)2D3 added to the culture medium | IL-b, IL-8, TNF-a, CCL3, CCL4 and CCL8 | TLR2 and TLR4 Induces cathelicidin synthesis Peripheral mononuclear cells KHOO et al., 2011 Human peripheral blood cell culture 1,25 (OH)2D3 added to the culture medium | Dose-dependent IL-6, TNF-a and IFN-y | cathelicidin Neutrophils YANG et al., 2015 CHEN; EAPEN; ZOSKI 2015 ARAZ-CIBRIAN; GIRALDO; URCUQUI-ICHIMA 2019 Human peripheral blood cell culture 1,25 (OH)2D3 added to the culture medium | Apoptosis in chronic obstructive pulmonary disease | IL-8 levels | NETs formation Eosinophils MATHEU et al., 2003 Knockout mice Vitamin D subcutaneous injection Eosinophilic narrowing of the upper airways | IL-5 synthesis Mast cells BIGGS et al., 2010 Knockout Vitamin D subcutaneous injection Eosinophilic narrowing of the upper airways | IL-5 synthesis Th1 cells SKROBOT; DEMKOW; WACHOWSKA 2018 RAUSCH-FAN et al., 2002 Human peripheral blood cell culture 1,25 (OH)2D3 added to the culture medium | Synthesis IL-2, IFN-y, TNF-a Inhibition of IL-12 synthesis Th2 cell SKROBOT; DEMKOW; WACHOWSKA 2018 BOONSTRA et al., 2001 Knockout mice | Synthesis of IL-4, IL-5, IL-10 | Transcription of GATA3 Th17 cells IKEDA et al. 2010 JOSHI et al., 2011 Human peripheral blood cell culture Knockout mice IP treatment with 1,25 (OH)2D3 1,25 (OH)2D3 added to the culture medium | Synthesis of IL-17, IL-21 and IL-22 Promotion of regulatory T cell differentiation URRY et al., 2012 KANG et al., 2012 Human peripheral blood cell culture Knockout mice (tissue culture) 1,25 (OH)2D3 added to the culture medium | Synthesis of IL-10 and of FoxP3 transcription factor B cells CHEN et al., 2007 Human peripheral blood cell culture 1,25 (OH)2D3 added to the culture medium | B cell maturation into plasmocytes and memory cells | Isotype switch Multiple sclerosis SLOKA et al., 2011 CHAG et al., 2010 COSTA et al., 2016 Human peripheral blood cell culture Knockout mice (tissue culture) 1,25 (OH)2D3 administered IP in mouse 1,25 (OH)2D3 added to the culture medium |Th2 | Th1, Th17, IFN-y and IL-17 Rheumatoid arthritis ZHOU et al., 2019 Knockout mice IP treatment with 1,25 (OH)2D3 1,25 (OH)2 D3 administered together with the chow Stopped disease progression | IL-17 and | Tregs Systemic Lupus Erythematosus ABOU-RAYA; ABOU-RAYA; HELMII 2013 PIANTONI et al., 2015 Measurement of serum calciferol levels in humans Human peripheral blood cell culture Oral supplementation with cholecalciferol | IL-8, IL-1, IL-6 and TNF-a | Tregs Inflammatory bowel disease DANIEL et al., 2008 BARTELS et al., 2007 CANTORNA et al., 2000 Human peripheral blood cell culture BALB/c mice 1,25 (OH)2D3 added to the culture medium IP treatment with 1,25 (OH)2D3 | IL-10, IL-4, TGF-b | Th1, IFN-y and TNF-a Airway Diseases PFEFFER and HAWRYLOWICZ 2018 BREHM et al., 2010 GUPTA et al., 2014 URRY et al., 2012 SUBRAMANINA; BERGMAN; NORMAK 2017 Serum vitD dosage Asthmatic children Knockout mice 1,25 (OH)2D3 added to peripheral blood culture and to co-culture of neutrophils and pneumococcus Low vitamin D levels associated with severe asthma | Tregs and IL-10 | IgE Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Huang C. Wang Y. Li X. Ren L. Zhao J. Hu Y. Zhang L. Fan G. Xu J. Gu X. Clinical Features of Patients Infected with 2019 Novel Coronavirus in Wuhan, China Lancet 2020 395 497 506 10.1016/S0140-6736(20)30183-5 31986264 2. Da Rosa Mesquita R. Francelino Silva Junior L.C. Santos Santana F.M. Farias de Oliveira T. Campos Alcantara R. Monteiro Arnozo G. Rodrigues da Silva Filho E. Galdino dos Santos A.G. Oliveira da Cunha E.J. Salgueiro de Aquino S.H. Clinical Manifestations of COVID-19 in the General Population: Systematic Review Wien. Klin. Wochenschr. 2021 133 377 382 10.1007/s00508-020-01760-4 33242148 3. Pietrobon A.J. Teixeira F.M.E. Sato M.N. Imunosenescence and Inflammaging: Risk Factors of Severe COVID-19 in Older People Front. Immunol. 2020 11 579220 10.3389/fimmu.2020.579220 33193377 4. Shahid Z. Kalayanamitra R. McClafferty B. Kepko D. Ramgobin D. Patel R. Aggarwal C.S. Vunnam R. Sahu N. Bhatt D. COVID-19 and Older Adults: What We Know J. Am. Geriatr. Soc. 2020 68 926 929 10.1111/jgs.16472 32255507 5. Yildirim Z. Sahin O.S. Yazar S. Bozok Cetintas V. Genetic and Epigenetic Factors Associated with Increased Severity of COVID-19 Cell Biol. Int. 2021 45 1158 1174 10.1002/cbin.11572 33590936 6. Johnson K.D. Harris C. Cain J.K. Hummer C. Goyal H. Perisetti A. Pulmonary and Extra-Pulmonary Clinical Manifestations of COVID-19 Front. Med. 2020 7 526 10.3389/fmed.2020.00526 7. Nelwan E.J. Tunjungputri R.N. Tetrasiwi E.N. Lauditta R.K. Nainggolan L. Extrapulmonary Manifestations COVID-19 Acta Med. Indones. 2022 54 314 315 35818654 8. Jackson C.B. Farzan M. Chen B. Choe H. Mechanisms of SARS-CoV-2 Entry into Cells Nat. Rev. Mol. Cell Biol. 2022 23 3 20 10.1038/s41580-021-00418-x 34611326 9. Diamond M.S. Kanneganti T.-D. Innate Immunity: The First Line of Defense against SARS-CoV-2 Nat. Immunol. 2022 23 165 176 10.1038/s41590-021-01091-0 35105981 10. Vabret N. Britton G.J. Gruber C. Hegde S. Kim J. Kuksin M. Levantovsky R. Malle L. Moreira A. Park M.D. Immunology of COVID-19: Current State of the Science Immunity 2020 52 910 941 10.1016/j.immuni.2020.05.002 32505227 11. Hadjadj J. Yatim N. Barnabei L. Corneau A. Boussier J. Smith N. Pere H. Charbit B. Bondet V. Chenevier-Gobeaux C. Impaired Type I Interferon Activity and Inflammatory Responses in Severe COVID-19 Patients Science 2020 369 718 724 10.1126/science.abc6027 32661059 12. Blanco-Melo D. Nilsson-Payant B.E. Liu W.-C. Uhl S. Hoagland D. Moller R. Jordan T.X. Oishi K. Panis M. Sachs D. Imbalanced Host Response to SARS-CoV-2 Drives Development of COVID-19 Cell 2020 181 1036 1045.e9 10.1016/j.cell.2020.04.026 32416070 13. Zhang Q. Bastard P. Liu Z. Le Pen J. Moncada-Velez M. Chen J. Ogishi M. Sabli I.K.D. Hodeib S. Korol C. Inborn Errors of Type I IFN Immunity in Patients with Life-Threatening COVID-19 Science 2020 370 eabd4570 10.1126/science.abd4570 32972995 14. Bastard P. Rosen L.B. Zhang Q. Michailidis E. Hoffmann H.-H. Zhang Y. Dorgham K. Philippot Q. Rosain J. Beziat V. Autoantibodies against Type I IFNs in Patients with Life-Threatening COVID-19 Science 2020 370 eabd4585 10.1126/science.abd4585 32972996 15. Domizio J.D. Gulen M.F. Saidoune F. Thacker V.V. Yatim A. Sharma K. Nass T. Guenova E. Schaller M. Conrad C. The CGAS-STING Pathway Drives Type I IFN Immunopathology in COVID-19 Nature 2022 603 145 151 10.1038/s41586-022-04421-w 35045565 16. Wong L.-Y.R. Perlman S. Immune Dysregulation and Immunopathology Induced by SARS-CoV-2 and Related Coronaviruses --Are We Our Own Worst Enemy? Nat. Rev. Immunol. 2022 22 47 56 10.1038/s41577-021-00656-2 34837062 17. Ramatillah D.L. Gan S.H. Pratiwy I. Syed Sulaiman S.A. Jaber A.A.S. Jusnita N. Lukas S. Abu Bakar U. Impact of Cytokine Storm on Severity of COVID-19 Disease in a Private Hospital in West Jakarta Prior to Vaccination PLoS ONE 2022 17 e0262438 10.1371/journal.pone.0262438 35077495 18. Khan S. Shafiei M.S. Longoria C. Schoggins J.W. Savani R.C. Zaki H. SARS-CoV-2 Spike Protein Induces Inflammation via TLR2-Dependent Activation of the NF-KB Pathway eLife 2021 10 e68563 10.7554/eLife.68563 34866574 19. Briquez P.S. Rouhani S.J. Yu J. Pyzer A.R. Trujillo J. Dugan H.L. Stamper C.T. Changrob S. Sperling A.I. Wilson P.C. Severe COVID-19 Induces Autoantibodies against Angiotensin II That Correlate with Blood Pressure Dysregulation and Disease Severity Sci. Adv. 2022 8 eabn3777 10.1126/sciadv.abn3777 36206332 20. Almutlaq M. Mansour F.A. Alghamdi J. Alhendi Y. Alamro A.A. Alghamdi A.A. Alamri H.S. Alroqi F. Barhoumi T. Angiotensin II Exaggerates SARS-CoV-2 Specific T-Cell Response in Convalescent Individuals Following COVID-19 IJMS Int. J. Mol. Sci. 2022 23 8669 10.3390/ijms23158669 35955801 21. Vora S.M. Lieberman J. Wu H. Inflammasome Activation at the Crux of Severe COVID-19 Nat. Rev. Immunol. 2021 21 694 703 10.1038/s41577-021-00588-x 34373622 22. Shah A. Novel Coronavirus-Induced NLRP3 Inflammasome Activation: A Potential Drug Target in the Treatment of COVID-19 Front. Immunol. 2020 11 1021 10.3389/fimmu.2020.01021 32574259 23. Guzman-Esquivel J. Galvan-Salazar H. Guzman-Solorzano H. Cuevas-Velazquez A. Guzman-Solorzano J. Mokay-Ramirez K. Paz-Michel B. Murillo-Zamora E. Delgado-Enciso J. Melnikov V. Efficacy of the Use of Mefenamic Acid Combined with Standard Medical Care vs. Standard Medical Care Alone for the Treatment of COVID-19: A Randomized Double-blind Placebo-controlled Trial Int. J. Mol. Med. 2022 49 29 10.3892/ijmm.2022.5084 35029292 24. Zuo Y. Yalavarthi S. Navaz S. Hoy C. Harbaugh A. Gockman K. Zuo M. Madison J.A. Shi H. Kanthi Y. Autoantibodies Stabilize Neutrophil Extracellular Traps in COVID-19; Infectious Diseases (except HIV/AIDS) arXiv 2021 10.1101/2021.03.31.21254692 25. De Oliveira G.L.V. Oliveira C.N.S. Pinzan C.F. de Salis L.V.V. Cardoso C.R.d.B. Microbiota Modulation of the Gut-Lung Axis in COVID-19 Front. Immunol. 2021 12 635471 10.3389/fimmu.2021.635471 33717181 26. Groves H.T. Cuthbertson L. James P. Moffatt M.F. Cox M.J. Tregoning J.S. Respiratory Disease Following Viral Lung Infection Alters the Murine Gut Microbiota Front. Immunol. 2018 9 182 10.3389/fimmu.2018.00182 29483910 27. Asadi-Pooya A.A. Akbari A. Emami A. Lotfi M. Rostamihosseinkhani M. Nemati H. Barzegar Z. Kabiri M. Zeraatpisheh Z. Farjoud-Kouhanjani M. Long COVID Syndrome-associated Brain Fog J. Med. Virol. 2022 94 979 984 10.1002/jmv.27404 34672377 28. Guzman-Esquivel J. Mendoza-Hernandez M.A. Guzman-Solorzano H.P. Sarmiento-Hernandez K.A. Rodriguez-Sanchez I.P. Martinez-Fierro M.L. Paz-Michel B.A. Murillo-Zamora E. Rojas-Larios F. Lugo-Trampe A. Clinical Characteristics in the Acute Phase of COVID-19 That Predict Long COVID: Tachycardia, Myalgias, Severity, and Use of Antibiotics as Main Risk Factors, While Education and Blood Group B Are Protective Healthcare 2023 11 197 10.3390/healthcare11020197 36673565 29. Ramirez J. Guarner F. Bustos Fernandez L. Maruy A. Sdepanian V.L. Cohen H. Antibiotics as Major Disruptors of Gut Microbiota Front. Cell. Infect. Microbiol. 2020 10 572912 10.3389/fcimb.2020.572912 33330122 30. Bilgic A. Kodjikian L. Sudhalkar A. Dwivedi S. Vasavada V. Shah A. Dziadzko M. Mathis T. Risk Factors for COVID-19 Associated Mucormycosis: The Ophthalmologist's Perspective J. Fungi 2022 8 271 10.3390/jof8030271 35330273 31. Sencio V. Machelart A. Robil C. Benech N. Hoffmann E. Galbert C. Deryuter L. Heumel S. Hantute-Ghesquier A. Flourens A. Alteration of the Gut Microbiota Following SARS-CoV-2 Infection Correlates with Disease Severity in Hamsters Gut Microbes 2022 14 2018900 10.1080/19490976.2021.2018900 34965194 32. Yeoh Y.K. Zuo T. Lui G.C.-Y. Zhang F. Liu Q. Li A.Y. Chung A.C. Cheung C.P. Tso E.Y. Fung K.S. Gut Microbiota Composition Reflects Disease Severity and Dysfunctional Immune Responses in Patients with COVID-19 Gut 2021 70 698 706 10.1136/gutjnl-2020-323020 33431578 33. Wang B. Zhang L. Wang Y. Dai T. Qin Z. Zhou F. Zhang L. Alterations in Microbiota of Patients with COVID-19: Potential Mechanisms and Therapeutic Interventions Signal Transduct. Target. Ther. 2022 7 143 10.1038/s41392-022-00986-0 35487886 34. Abboud H. Abboud F.Z. Kharbouch H. Arkha Y. El Abbadi N. El Ouahabi A. COVID-19 and SARS-CoV-2 Infection: Pathophysiology and Clinical Effects on the Nervous System World Neurosurg. 2020 140 49 53 10.1016/j.wneu.2020.05.193 32474093 35. Chou S.H.-Y. Beghi E. Helbok R. Moro E. Sampson J. Altamirano V. Mainali S. Bassetti C. Suarez J.I. McNett M. Global Incidence of Neurological Manifestations Among Patients Hospitalized With COVID-19--A Report for the GCS-NeuroCOVID Consortium and the ENERGY Consortium JAMA Netw. Open 2021 4 e2112131 10.1001/jamanetworkopen.2021.12131 33974053 36. Araujo N.M. Ferreira L.C. Dantas D.P. Silva D.S. dos Santos C.A. Cipolotti R. Martins-Filho P.R. First Report of SARS-CoV-2 Detection in Cerebrospinal Fluid in a Child With Guillain-Barre Syndrome Pediatr. Infect. Dis. J. 2021 40 e274 e276 10.1097/INF.0000000000003146 33990525 37. Song E. Zhang C. Israelow B. Lu-Culligan A. Prado A.V. Skriabine S. Lu P. Weizman O.-E. Liu F. Dai Y. Neuroinvasion of SARS-CoV-2 in Human and Mouse Brain J. Exp. Med. 2021 218 e20202135 10.1084/jem.20202135 33433624 38. Reza-Zaldivar E.E. Hernandez-Sapiens M.A. Minjarez B. Gomez-Pinedo U. Marquez-Aguirre A.L. Mateos-Diaz J.C. Matias-Guiu J. Canales-Aguirre A.A. Infection Mechanism of SARS-CoV-2 and Its Implication on the Nervous System Front. Immunol. 2021 11 621735 10.3389/fimmu.2020.621735 33584720 39. Gasmi A. Tippairote T. Mujawdiya P.K. Gasmi Benahmed A. Menzel A. Dadar M. Bjorklund G. Neurological Involvements of SARS-CoV-2 Infection Mol. Neurobiol. 2021 58 944 949 10.1007/s12035-020-02070-6 33064267 40. Jiao L. Yang Y. Yu W. Zhao Y. Long H. Gao J. Ding K. Ma C. Li J. Zhao S. The Olfactory Route Is a Potential Way for SARS-CoV-2 to Invade the Central Nervous System of Rhesus Monkeys Signal Transduct. Target. Ther. 2021 6 169 10.1038/s41392-021-00591-7 33895780 41. Liu J. Tan B. Wu S. Gui Y. Suo J. Li Y. Evidence of Central Nervous System Infection and Neuroinvasive Routes, as Well as Neurological Involvement, in the Lethality of SARS-CoV-2 Infection J. Med. Virol. 2021 93 1304 1313 10.1002/jmv.26570 33002209 42. Esposito G. Pesce M. Seguella L. Sanseverino W. Lu J. Sarnelli G. Can the Enteric Nervous System Be an Alternative Entrance Door in SARS-CoV-2 Neuroinvasion? Brain Behav. Immun. 2020 87 93 94 10.1016/j.bbi.2020.04.060 32335192 43. Paniz-Mondolfi A. Bryce C. Grimes Z. Gordon R.E. Reidy J. Lednicky J. Sordillo E.M. Fowkes M. Central Nervous System Involvement by Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) J. Med. Virol. 2020 92 699 702 10.1002/jmv.25915 32314810 44. Lam S.M. Huang X. Shui G. Neurological Aspects of SARS-CoV-2 Infection: Lipoproteins and Exosomes as Trojan Horses Trends Endocrinol. Metab. 2022 33 554 568 10.1016/j.tem.2022.04.011 35613979 45. Priori A. Neurology of COVID-19 Milano University Press Milan, Italy 2021 9791280325358 46. Zalpoor H. Akbari A. Samei A. Forghaniesfidvajani R. Kamali M. Afzalnia A. Manshouri S. Heidari F. Pornour M. Khoshmirsafa M. The Roles of Eph Receptors, Neuropilin-1, P2X7, and CD147 in COVID-19-Associated Neurodegenerative Diseases: Inflammasome and JaK Inhibitors as Potential Promising Therapies Cell. Mol. Biol. Lett. 2022 27 10 10.1186/s11658-022-00311-1 35109786 47. Savelieff M.G. Feldman E.L. Stino A.M. Neurological Sequela and Disruption of Neuron-Glia Homeostasis in SARS-CoV-2 Infection Neurobiol. Dis. 2022 168 105715 10.1016/j.nbd.2022.105715 35364273 48. Sajdel-Sulkowska E.M. Neuropsychiatric Ramifications of COVID-19: Short-Chain Fatty Acid Deficiency and Disturbance of Microbiota-Gut-Brain Axis Signaling BioMed Res. Int. 2021 2021 7880448 10.1155/2021/7880448 34651049 49. Wlodarczyk J. Czerwinski B. Fichna J. Short-Chain Fatty Acids-Microbiota Crosstalk in the Coronavirus Disease (COVID-19) Pharmacol. Rep. 2022 74 1198 1207 10.1007/s43440-022-00415-7 36166147 50. Vakili K. Fathi M. Yaghoobpoor S. Sayehmiri F. Nazerian Y. Nazerian A. Mohamadkhani A. Khodabakhsh P. Reus G.Z. Hajibeygi R. The Contribution of Gut-Brain Axis to Development of Neurological Symptoms in COVID-19 Recovered Patients: A Hypothesis and Review of Literature Front. Cell. Infect. Microbiol. 2022 12 983089 10.3389/fcimb.2022.983089 36619768 51. Parodi B. Kerlero de Rosbo N. The Gut-Brain Axis in Multiple Sclerosis. Is Its Dysfunction a Pathological Trigger or a Consequence of the Disease? Front. Immunol. 2021 12 718220 10.3389/fimmu.2021.718220 34621267 52. Garg N. Smith T.W. An Update on Immunopathogenesis, Diagnosis, and Treatment of Multiple Sclerosis Brain Behav. 2015 5 e00362 10.1002/brb3.362 26445701 53. Confavreux C. Vukusic S. Natural History of Multiple Sclerosis: A Unifying Concept Brain 2006 129 606 616 10.1093/brain/awl007 16415308 54. Klineova S. Lublin F.D. Clinical Course of Multiple Sclerosis Cold Spring Harb. Perspect. Med. 2018 8 a028928 10.1101/cshperspect.a028928 29358317 55. Xie Y. Tian Z. Han F. Liang S. Gao Y. Wu D. Factors Associated with Relapses in Relapsing-Remitting Multiple Sclerosis: A Systematic Review and Meta-Analysis Medicine 2020 99 e20885 10.1097/MD.0000000000020885 32629678 56. Vollmer T. The Natural History of Relapses in Multiple Sclerosis J. Neurol. Sci. 2007 256 S5 S13 10.1016/j.jns.2007.01.065 17346747 57. Robinson A.P. Harp C.T. Noronha A. Miller S.D. The Experimental Autoimmune Encephalomyelitis (EAE) Model of MS Handbook of Clinical Neurology Elsevier Amsterdam, The Netherlands 2014 Volume 122 173 189 978-0-444-52001-2 58. Fujinami R.S. von Herrath M.G. Christen U. Whitton J.L. Molecular Mimicry, Bystander Activation, or Viral Persistence: Infections and Autoimmune Disease Clin. Microbiol. Rev. 2006 19 80 94 10.1128/CMR.19.1.80-94.2006 16418524 59. Elsayed N.S. Aston P. Bayanagari V.R. Shukla S.K. The Gut Microbiome Molecular Mimicry Piece in the Multiple Sclerosis Puzzle Front. Immunol. 2022 13 972160 10.3389/fimmu.2022.972160 36045671 60. Khan U. Ghazanfar H. T Lymphocytes and Autoimmunity International Review of Cell and Molecular Biology Elsevier Amsterdam, The Netherlands 2018 Volume 341 125 168 978-0-12-815385-7 61. Odoardi F. Sie C. Streyl K. Ulaganathan V.K. Schlager C. Lodygin D. Heckelsmiller K. Nietfeld W. Ellwart J. Klinkert W.E.F. T Cells Become Licensed in the Lung to Enter the Central Nervous System Nature 2012 488 675 679 10.1038/nature11337 22914092 62. Tan C. Wandu W.S. Lee R.S. Hinshaw S.H. Klinman D.M. Wawrousek E. Gery I. Shedding New Light on the Process of "Licensing" for Pathogenicity by Th Lymphocytes J. Immunol. 2017 198 681 690 10.4049/jimmunol.1502108 27986906 63. Ransohoff R.M. Licensed in the Lungs Nature 2012 488 595 596 10.1038/488595a 22932379 64. Kebir H. Kreymborg K. Ifergan I. Dodelet-Devillers A. Cayrol R. Bernard M. Giuliani F. Arbour N. Becher B. Prat A. Human TH17 Lymphocytes Promote Blood-Brain Barrier Disruption and Central Nervous System Inflammation Nat. Med. 2007 13 1173 1175 10.1038/nm1651 17828272 65. Moser T. Akgun K. Proschmann U. Sellner J. Ziemssen T. The Role of TH17 Cells in Multiple Sclerosis: Therapeutic Implications Autoimmun. Rev. 2020 19 102647 10.1016/j.autrev.2020.102647 32801039 66. Ivanov I.I. Atarashi K. Manel N. Brodie E.L. Shima T. Karaoz U. Wei D. Goldfarb K.C. Santee C.A. Lynch S.V. Induction of Intestinal Th17 Cells by Segmented Filamentous Bacteria Cell 2009 139 485 498 10.1016/j.cell.2009.09.033 19836068 67. Cosorich I. Dalla-Costa G. Sorini C. Ferrarese R. Messina M.J. Dolpady J. Radice E. Mariani A. Testoni P.A. Canducci F. High Frequency of Intestinal TH 17 Cells Correlates with Microbiota Alterations and Disease Activity in Multiple Sclerosis Sci. Adv. 2017 3 e1700492 10.1126/sciadv.1700492 28706993 68. Gaboriau-Routhiau V. Rakotobe S. Lecuyer E. Mulder I. Lan A. Bridonneau C. Rochet V. Pisi A. De Paepe M. Brandi G. The Key Role of Segmented Filamentous Bacteria in the Coordinated Maturation of Gut Helper T Cell Responses Immunity 2009 31 677 689 10.1016/j.immuni.2009.08.020 19833089 69. Lee Y.K. Menezes J.S. Umesaki Y. Mazmanian S.K. Proinflammatory T-Cell Responses to Gut Microbiota Promote Experimental Autoimmune Encephalomyelitis Proc. Natl. Acad. Sci. USA 2011 108 4615 4622 10.1073/pnas.1000082107 20660719 70. Ortiz G.G. Pacheco-Moises F.P. Macias-Islas M.A. Flores-Alvarado L.J. Mireles-Ramirez M.A. Gonzalez-Renovato E.D. Hernandez-Navarro V.E. Sanchez-Lopez A.L. Alatorre-Jimenez M.A. Role of the Blood-Brain Barrier in Multiple Sclerosis Arch. Med. Res. 2014 45 687 697 10.1016/j.arcmed.2014.11.013 25431839 71. Huang X. Hussain B. Chang J. Peripheral Inflammation and Blood-Brain Barrier Disruption: Effects and Mechanisms CNS Neurosci. Ther. 2021 27 36 47 10.1111/cns.13569 33381913 72. Chen Z. Li G. Immune Response and Blood-Brain Barrier Dysfunction during Viral Neuroinvasion Innate Immun. 2021 27 109 117 10.1177/1753425920954281 32903111 73. Furutama D. Matsuda S. Yamawaki Y. Hatano S. Okanobu A. Memida T. Oue H. Fujita T. Ouhara K. Kajiya M. IL-6 Induced by Periodontal Inflammation Causes Neuroinflammation and Disrupts the Blood-Brain Barrier Brain Sci. 2020 10 679 10.3390/brainsci10100679 32992470 74. Karpus W.J. Cytokines and Chemokines in the Pathogenesis of Experimental Autoimmune Encephalomyelitis J. Immunol. 2020 204 316 326 10.4049/jimmunol.1900914 31907274 75. McGinley A.M. Edwards S.C. Raverdeau M. Mills K.H.G. Th17 Cells, Gd T Cells and Their Interplay in EAE and Multiple Sclerosis J. Autoimmun. 2018 87 97 108 10.1016/j.jaut.2018.01.001 76. Gharibi T. Babaloo Z. Hosseini A. Marofi F. Ebrahimi-kalan A. Jahandideh S. Baradaran B. The Role of B Cells in the Immunopathogenesis of Multiple Sclerosis Immunology 2020 160 325 335 10.1111/imm.13198 32249925 77. Lang Y. Chu F. Shen D. Zhang W. Zheng C. Zhu J. Cui L. Role of Inflammasomes in Neuroimmune and Neurodegenerative Diseases: A Systematic Review Mediat. Inflamm. 2018 2018 1549549 10.1155/2018/1549549 78. Ruiz F. Vigne S. Pot C. Resolution of Inflammation during Multiple Sclerosis Semin. Immunopathol. 2019 41 711 726 10.1007/s00281-019-00765-0 31732775 79. Pandit L. Malli C. D'Cunha A. Sudhir A. Role of Viral Infections in Multiple Sclerosis Pathogenesis among Indian Population Neurol. India 2021 69 681 10.4103/0028-3886.319209 34169868 80. Bjornevik K. Cortese M. Healy B.C. Kuhle J. Mina M.J. Leng Y. Elledge S.J. Niebuhr D.W. Scher A.I. Munger K.L. Longitudinal Analysis Reveals High Prevalence of Epstein-Barr Virus Associated with Multiple Sclerosis Science 2022 375 296 301 10.1126/science.abj8222 35025605 81. Mrozek-Gorska P. Buschle A. Pich D. Schwarzmayr T. Fechtner R. Scialdone A. Hammerschmidt W. Epstein-Barr Virus Reprograms Human B Lymphocytes Immediately in the Prelatent Phase of Infection Proc. Natl. Acad. Sci. USA 2019 116 16046 16055 10.1073/pnas.1901314116 31341086 82. Soldan S.S. Lieberman P.M. Epstein-Barr Virus and Multiple Sclerosis Nat. Rev. Microbiol. 2023 21 51 64 10.1038/s41579-022-00770-5 35931816 83. Serafini B. Severa M. Columba-Cabezas S. Rosicarelli B. Veroni C. Chiappetta G. Magliozzi R. Reynolds R. Coccia E.M. Aloisi F. Epstein-Barr Virus Latent Infection and BAFF Expression in B Cells in the Multiple Sclerosis Brain: Implications for Viral Persistence and Intrathecal B-Cell Activation J. Neuropathol. Exp. Neurol. 2010 69 677 693 10.1097/NEN.0b013e3181e332ec 20535037 84. Robinson W.H. Steinman L. Epstein-Barr Virus and Multiple Sclerosis Science 2022 375 264 265 10.1126/science.abm7930 35025606 85. Milo R. Therapies for Multiple Sclerosis Targeting B Cells Croat. Med. J. 2019 60 87 98 10.3325/cmj.2019.60.87 31044580 86. Bellucci G. Rinaldi V. Buscarinu M.C. Renie R. Bigi R. Pellicciari G. Morena E. Romano C. Marrone A. Mechelli R. Multiple Sclerosis and SARS-CoV-2: Has the Interplay Started? Front. Immunol. 2021 12 755333 10.3389/fimmu.2021.755333 34646278 87. Brod S.A. Hypothesis: Multiple Sclerosis Is a Type I Interferon Deficiency Syndrome Exp. Biol. Med. 1998 218 278 283 10.3181/00379727-218-44295 9714071 88. Paul S. Ricour C. Sommereyns C. Sorgeloos F. Michiels T. Type I Interferon Response in the Central Nervous System Biochimie 2007 89 770 778 10.1016/j.biochi.2007.02.009 17408841 89. Cohan S.L. Hendin B.A. Reder A.T. Smoot K. Avila R. Mendoza J.P. Weinstock-Guttman B. Interferons and Multiple Sclerosis: Lessons from 25 Years of Clinical and Real-World Experience with Intramuscular Interferon Beta-1a (Avonex) CNS Drugs 2021 35 743 767 10.1007/s40263-021-00822-z 34228301 90. Sui L. Zhao Y. Wang W. Wu P. Wang Z. Yu Y. Hou Z. Tan G. Liu Q. SARS-CoV-2 Membrane Protein Inhibits Type I Interferon Production Through Ubiquitin-Mediated Degradation of TBK1 Front. Immunol. 2021 12 662989 10.3389/fimmu.2021.662989 34084167 91. Palermo E. Di Carlo D. Sgarbanti M. Hiscott J. Type I Interferons in COVID-19 Pathogenesis Biology 2021 10 829 10.3390/biology10090829 34571706 92. Sormani M.P. Salvetti M. Labauge P. Schiavetti I. Zephir H. Carmisciano L. Bensa C. De Rossi N. Pelletier J. Cordioli C. DMTs and COVID-19 Severity in MS: A Pooled Analysis from Italy and France Ann. Clin. Transl. Neurol. 2021 8 1738 1744 10.1002/acn3.51408 34240579 93. Gemcioglu E. Davutoglu M. Ozdemir E.E. Erden A. Are Type 1 Interferons Treatment in Multiple Sclerosis as a Potential Therapy against COVID-19? Mult. Scler. Relat. Discord. 2020 42 102196 10.1016/j.msard.2020.102196 94. Kaskow B.J. Baecher-Allan C. Effector T Cells in Multiple Sclerosis Cold Spring Harb. Perspect. Med. 2018 8 a029025 10.1101/cshperspect.a029025 29358315 95. Franchi L. Eigenbrod T. Munoz-Planillo R. Nunez G. The Inflammasome: A Caspase-1-Activation Platform That Regulates Immune Responses and Disease Pathogenesis Nat. Immunol. 2009 10 241 247 10.1038/ni.1703 19221555 96. Barclay W. Shinohara M.L. Inflammasome Activation in Multiple Sclerosis and Experimental Autoimmune Encephalomyelitis (EAE): Inflammasomes in CNS Autoimmunity Brain Pathol. 2017 27 213 219 10.1111/bpa.12477 27997058 97. Rodrigues T.S. de Sa K.S.G. Ishimoto A.Y. Becerra A. Oliveira S. Almeida L. Goncalves A.V. Perucello D.B. Andrade W.A. Castro R. Inflammasomes Are Activated in Response to SARS-CoV-2 Infection and Are Associated with COVID-19 Severity in Patients J. Exp. Med. 2021 218 e20201707 10.1084/jem.20201707 33231615 98. Dutta D. Liu J. Xiong H. NLRP3 Inflammasome Activation and SARS-CoV-2-Mediated Hyperinflammation, Cytokine Storm and Neurological Syndromes Int. J. Physiol. Pathophysiol. Pharmacol. 2022 14 138 160 35891930 99. Etemadifar M. Sedaghat N. Aghababaee A. Kargaran P.K. Maracy M.R. Ganjalikhani-Hakemi M. Rayani M. Abhari A.P. Khorvash R. Salari M. COVID-19 and the Risk of Relapse in Multiple Sclerosis Patients: A Fight with No Bystander Effect? Mult. Scler. Relat. Disord. 2021 51 102915 10.1016/j.msard.2021.102915 33799284 100. Nieves J. Cosman F. Herbert J. Shen V. Lindsay R. High Prevalence of Vitamin D Deficiency and Reduced Bone Mass in Multiple Sclerosis Neurology 1994 44 1687 10.1212/WNL.44.9.1687 7936297 101. Fahmi R.M. Lotfy S.M. Mohamed W.S. Elsaid A.F. Murad M.H. Abdulmoneem G. Vitamin D Levels in Patients with Multiple Sclerosis Egypt. J. Neurol. Psychiatry Neurosurg. 2014 51 145 152 102. Soilu-Hanninen M. Aivo J. Lindstrom B.-M. Elovaara I. Sumelahti M.-L. Farkkila M. Tienari P. Atula S. Sarasoja T. Herrala L. A Randomised, Double Blind, Placebo Controlled Trial with Vitamin D3 as an Add on Treatment to Interferon b-1b in Patients with Multiple Sclerosis J. Neurol. Neurosurg. Psychiatry 2012 83 565 571 10.1136/jnnp-2011-301876 22362918 103. Laursen J.H. Sondergaard H.B. Sorensen P.S. Sellebjerg F. Oturai A.B. Vitamin D Supplementation Reduces Relapse Rate in Relapsing-Remitting Multiple Sclerosis Patients Treated with Natalizumab Mult. Scler. Relat. Disord. 2016 10 169 173 10.1016/j.msard.2016.10.005 27919484 104. Kimball S.M. Ursell M.R. O'Connor P. Vieth R. Safety of Vitamin D3 in Adults with Multiple Sclerosis Am. J. Clin. Nutr. 2007 86 645 651 10.1093/ajcn/86.3.645 17823429 105. Dorr J. Backer-Koduah P. Wernecke K.-D. Becker E. Hoffmann F. Faiss J. Brockmeier B. Hoffmann O. Anvari K. Wuerfel J. High-Dose Vitamin D Supplementation in Multiple Sclerosis--Results from the Randomized EVIDIMS (Efficacy of Vitamin D Supplementation in Multiple Sclerosis) Trial Mult. Scler. J.-Exp. Transl. Clin. 2020 6 205521732090347 10.1177/2055217320903474 106. Boltjes R. Knippenberg S. Gerlach O. Hupperts R. Damoiseaux J. Vitamin D Supplementation in Multiple Sclerosis: An Expert Opinion Based on the Review of Current Evidence Expert Rev. Neurother. 2021 21 715 725 10.1080/14737175.2021.1935878 34058936 107. Mimura L.A.N. Fraga-Silva T.F.d.C. Oliveira L.R.C.d. Ishikawa L.L.W. Borim P.A. Machado C.d.M. Junior J.d.A.d.C.e.H. Fonseca D.M.d. Sartori A. Preclinical Therapy with Vitamin D3 in Experimental Encephalomyelitis: Efficacy and Comparison with Paricalcitol IJMS Int. J. Mol. Sci. 2021 22 1914 10.3390/ijms22041914 33671896 108. Goncalves Zorzella-Pezavento S. Nishiyama Mimura L. Denadai M. Fernandes de Souza W. de Campos Fraga-Silva T. Sartori A. Is There a Window of Opportunity for the Therapeutic Use of Vitamin D in Multiple Sclerosis? Neural. Regen. Res. 2022 17 1945 10.4103/1673-5374.335139 35142671 109. Wang Y. Zhu J. DeLuca H.F. Where Is the Vitamin D Receptor? Arch. Biochem. Biophys. 2012 523 123 133 10.1016/j.abb.2012.04.001 22503810 110. Penna G. Adorini L. 1a,25-Dihydroxyvitamin D3 Inhibits Differentiation, Maturation, Activation, and Survival of Dendritic Cells Leading to Impaired Alloreactive T Cell Activation J. Immunol. 2000 164 2405 2411 10.4049/jimmunol.164.5.2405 10679076 111. Berer A. Stockl J. Majdic O. Wagner T. Kollars M. Lechner K. Geissler K. Oehler L. 1,25-Dihydroxyvitamin D3 Inhibits Dendritic Cell Differentiation and Maturation in Vitro Exp. Hematol. 2000 28 575 583 10.1016/S0301-472X(00)00143-0 10812248 112. Verway M. Bouttier M. Wang T.-T. Carrier M. Calderon M. An B.-S. Devemy E. McIntosh F. Divangahi M. Behr M.A. Vitamin D Induces Interleukin-1b Expression: Paracrine Macrophage Epithelial Signaling Controls M. Tuberculosis Infection PLoS Pathog. 2013 9 e1003407 10.1371/journal.ppat.1003407 23762029 113. Liu P.T. Stenger S. Li H. Wenzel L. Tan B.H. Krutzik S.R. Ochoa M.T. Schauber J. Wu K. Meinken C. Toll-Like Receptor Triggering of a Vitamin D-Mediated Human Antimicrobial Response Science 2006 311 1770 1773 10.1126/science.1123933 16497887 114. Gombart A.F. Borregaard N. Koeffler H.P. Human Cathelicidin Antimicrobial Peptide (CAMP) Gene Is a Direct Target of the Vitamin D Receptor and Is Strongly Up-regulated in Myeloid Cells by 1,25-dihydroxyvitamin D3 FASEB J. 2005 19 1067 1077 10.1096/fj.04-3284com 15985530 115. Khoo A.-L. Chai L.Y.A. Koenen H.J.P.M. Oosting M. Steinmeyer A. Zuegel U. Joosten I. Netea M.G. van der Ven A.J.A.M. Vitamin D3 Down-Regulates Proinflammatory Cytokine Response to Mycobacterium Tuberculosis through Pattern Recognition Receptors While Inducing Protective Cathelicidin Production Cytokine 2011 55 294 300 10.1016/j.cyto.2011.04.016 21592820 116. Yang H. Long F. Zhang Y. Yu R. Zhang P. Li W. Li S. Jin X. Xia J. Dong L. 1a,25-Dihydroxyvitamin D3 Induces Neutrophil Apoptosis through the P38 MAPK Signaling Pathway in Chronic Obstructive Pulmonary Disease Patients PLoS ONE 2015 10 e0120515 10.1371/journal.pone.0120515 25905635 117. Chen L. Eapen M.S. Zosky G.R. Vitamin D Both Facilitates and Attenuates the Cellular Response to Lipopolysaccharide Sci. Rep. 2017 7 45172 10.1038/srep45172 28345644 118. Agraz-Cibrian J.M. Giraldo D.M. Urcuqui-Inchima S. 1,25-Dihydroxyvitamin D3 Induces Formation of Neutrophil Extracellular Trap-like Structures and Modulates the Transcription of Genes Whose Products Are Neutrophil Extracellular Trap-Associated Proteins: A Pilot Study Steroids 2019 141 14 22 10.1016/j.steroids.2018.11.001 30414422 119. Matheu V. Dual Effects of Vitamin D-Induced Alteration of TH1/TH2 Cytokine Expression Enhancing IgE Production and Decreasing Airway Eosinophilia in Murine Allergic Airway Disease J. Allergy Clin. Immunol. 2003 112 585 592 10.1016/S0091-6749(03)01855-4 13679819 120. Biggs L. Yu C. Fedoric B. Lopez A.F. Galli S.J. Grimbaldeston M.A. Evidence That Vitamin D3 Promotes Mast Cell-Dependent Reduction of Chronic UVB-Induced Skin Pathology in Mice J. Exp. Med. 2010 207 455 463 10.1084/jem.20091725 20194632 121. Skrobot A. Demkow U. Wachowska M. Immunomodulatory Role of Vitamin D: A Review Current Trends in Immunity and Respiratory Infections Pokorski M. Advances in Experimental Medicine and Biology Springer International Publishing Cham, Switzerland 2018 Volume 1108 13 23 978-3-030-01634-0 122. Rausch-Fan X. Leutmezer F. Willheim M. Spittler A. Bohle B. Ebner C. Jensen-Jarolim E. Boltz-Nitulescu G. Regulation of Cytokine Production in Human Peripheral Blood Mononuclear Cells and Allergen-Specific Th Cell Clones by 1a,25-Dihydroxyvitamin D3 Int. Arch. Allergy Immunol. 2002 128 33 41 10.1159/000058001 12037399 123. Boonstra A. Barrat F.J. Crain C. Heath V.L. Savelkoul H.F.J. O'Garra A. 1a,25-Dihydroxyvitamin D3 Has a Direct Effect on Naive CD4 + T Cells to Enhance the Development of Th2 Cells J. Immunol. 2001 167 4974 4980 10.4049/jimmunol.167.9.4974 11673504 124. Ikeda U. Wakita D. Ohkuri T. Chamoto K. Kitamura H. Iwakura Y. Nishimura T. 1a,25-Dihydroxyvitamin D3 and All-Trans Retinoic Acid Synergistically Inhibit the Differentiation and Expansion of Th17 Cells Immunol. Lett. 2010 134 7 16 10.1016/j.imlet.2010.07.002 20655952 125. Joshi S. Pantalena L.-C. Liu X.K. Gaffen S.L. Liu H. Rohowsky-Kochan C. Ichiyama K. Yoshimura A. Steinman L. Christakos S. 1,25-Dihydroxyvitamin D3 Ameliorates Th17 Autoimmunity via Transcriptional Modulation of Interleukin-17A Mol. Cell. Biol. 2011 31 3653 3669 10.1128/MCB.05020-11 21746882 126. Urry Z. Dimeloe S. Hawrylowicz C.M. Vitamin D and Regulatory T Cells Vitamin D and the Lung Litonjua A.A. Humana Press Totowa, NJ, USA 2012 85 101 978-1-61779-887-0 127. Kang S.W. Kim S.H. Lee N. Lee W.-W. Hwang K.-A. Shin M.S. Lee S.-H. Kim W.-U. Kang I. 1,25-Dihyroxyvitamin D3 Promotes FOXP3 Expression via Binding to Vitamin D Response Elements in Its Conserved Noncoding Sequence Region J. Immunol. 2012 188 5276 5282 10.4049/jimmunol.1101211 22529297 128. Chen S. Sims G.P. Chen X.X. Gu Y.Y. Chen S. Lipsky P.E. Modulatory Effects of 1,25-Dihydroxyvitamin D3 on Human B Cell Differentiation J. Immunol. 2007 179 1634 1647 10.4049/jimmunol.179.3.1634 17641030 129. Sloka S. Silva C. Wang J. Yong V.W. Predominance of Th2 Polarization by Vitamin D through a STAT6-Dependent Mechanism J. Neuroinflamm. 2011 8 56 10.1186/1742-2094-8-56 130. Chang J.-H. Cha H.-R. Lee D.-S. Seo K.Y. Kweon M.-N. 1,25-Dihydroxyvitamin D3 Inhibits the Differentiation and Migration of TH17 Cells to Protect against Experimental Autoimmune Encephalomyelitis PLoS ONE 2010 5 e12925 10.1371/annotation/6da7b65e-dda2-467d-bcb3-82d5669f6bc6 20886077 131. Da Costa D.S.M.M. Hygino J. Ferreira T.B. Kasahara T.M. Barros P.O. Monteiro C. Oliveira A. Tavares F. Vasconcelos C.C. Alvarenga R. Vitamin D Modulates Different IL-17-Secreting T Cell Subsets in Multiple Sclerosis Patients J. Neuroimmunol. 2016 299 8 18 10.1016/j.jneuroim.2016.08.005 27725127 132. Zhou L. Wang J. Li J. Li T. Chen Y. June R.R. Zheng S.G. 1,25-Dihydroxyvitamin D3 Ameliorates Collagen-Induced Arthritis via Suppression of Th17 Cells Through MiR-124 Mediated Inhibition of IL-6 Signaling Front. Immunol. 2019 10 178 10.3389/fimmu.2019.00178 30792721 133. Abou-Raya A. Abou-Raya S. Helmii M. The Effect of Vitamin D Supplementation on Inflammatory and Hemostatic Markers and Disease Activity in Patients with Systemic Lupus Erythematosus: A Randomized Placebo-Controlled Trial J. Rheumatol. 2013 40 265 272 10.3899/jrheum.111594 23204220 134. Piantoni S. Andreoli L. Scarsi M. Zanola A. Dall'Ara F. Pizzorni C. Cutolo M. Airo P. Tincani A. Phenotype Modifications of T-Cells and Their Shift toward a Th2 Response in Patients with Systemic Lupus Erythematosus Supplemented with Different Monthly Regimens of Vitamin D Lupus 2015 24 490 498 10.1177/0961203314559090 25801892 135. Daniel C. Sartory N.A. Zahn N. Radeke H.H. Stein J.M. Immune Modulatory Treatment of Trinitrobenzene Sulfonic Acid Colitis with Calcitriol Is Associated with a Change of a T Helper (Th) 1/Th17 to a Th2 and Regulatory T Cell Profile J. Pharmacol. Exp. Ther. 2008 324 23 33 10.1124/jpet.107.127209 17911375 136. Bartels L.E. Jorgensen S.P. Agnholt J. Kelsen J. Hvas C.L. Dahlerup J.F. 1,25-Dihydroxyvitamin D3 and Dexamethasone Increase Interleukin-10 Production in CD4+ T Cells from Patients with Crohn's Disease Int. Immunopharmacol. 2007 7 1755 1764 10.1016/j.intimp.2007.09.016 17996686 137. Cantorna M.T. Munsick C. Bemiss C. Mahon B.D. 1,25-Dihydroxycholecalciferol Prevents and Ameliorates Symptoms of Experimental Murine Inflammatory Bowel Disease J. Nutr. 2000 130 2648 2652 10.1093/jn/130.11.2648 11053501 138. Pfeffer P.E. Hawrylowicz C.M. Vitamin D in Asthma Chest 2018 153 1229 1239 10.1016/j.chest.2017.09.005 28923762 139. Brehm J.M. Schuemann B. Fuhlbrigge A.L. Hollis B.W. Strunk R.C. Zeiger R.S. Weiss S.T. Litonjua A.A. Serum Vitamin D Levels and Severe Asthma Exacerbations in the Childhood Asthma Management Program Study J. Allergy Clin. Immunol. 2010 126 52 58.e5 10.1016/j.jaci.2010.03.043 20538327 140. Gupta A. Dimeloe S. Richards D.F. Chambers E.S. Black C. Urry Z. Ryanna K. Xystrakis E. Bush A. Saglani S. Defective IL-10 Expression and in Vitro Steroid-Induced IL-17A in Paediatric Severe Therapy-Resistant Asthma Thorax 2014 69 508 515 10.1136/thoraxjnl-2013-203421 24347461 141. Subramanian K. Bergman P. Henriques-Normark B. Vitamin D Promotes Pneumococcal Killing and Modulates Inflammatory Responses in Primary Human Neutrophils J. Innate Immun. 2017 9 375 386 10.1159/000455969 28241127 142. Navarro-Barriuso J. Mansilla M.J. Quirant-Sanchez B. Ardiaca-Martinez A. Teniente-Serra A. Presas-Rodriguez S. ten Brinke A. Ramo-Tello C. Martinez-Caceres E.M. MAP7 and MUCL1 Are Biomarkers of Vitamin D3-Induced Tolerogenic Dendritic Cells in Multiple Sclerosis Patients Front. Immunol. 2019 10 1251 10.3389/fimmu.2019.01251 31293564 143. Farias A.S. Spagnol G.S. Bordeaux-Rego P. Oliveira C.O.F. Fontana A.G.M. de Paula R.F.O. Santos M.P.A. Pradella F. Moraes A.S. Oliveira E.C. Vitamin D3 Induces IDO+ Tolerogenic DCs and Enhances Treg, Reducing the Severity of EAE CNS Neurosci. Ther. 2013 19 269 277 10.1111/cns.12071 23521914 144. Zeitelhofer M. Adzemovic M.Z. Gomez-Cabrero D. Bergman P. Hochmeister S. N'diaye M. Paulson A. Ruhrmann S. Almgren M. Tegner J.N. Functional Genomics Analysis of Vitamin D Effects on CD4+ T Cells in Vivo in Experimental Autoimmune Encephalomyelitis Proc. Natl. Acad. Sci. USA 2017 114 E1678 E1687 10.1073/pnas.1615783114 28196884 145. Martin-Aguilar L. Presas-Rodriguez S. Rovira A. Capellades J. Massuet-Vilamajo A. Ramio-Torrenta L. Tintore M. Brieva-Ruiz L. Moral E. Cano-Orgaz A. Gadolinium-Enhanced Brain Lesions in Multiple Sclerosis Relapse Neurologia 2022 37 557 563 10.1016/j.nrl.2021.10.005 36064284 146. De Oliveira L.R.C. Mimura L.A.N. Fraga-Silva T.F.d.C. Ishikawa L.L.W. Fernandes A.A.H. Zorzella-Pezavento S.F.G. Sartori A. Calcitriol Prevents Neuroinflammation and Reduces Blood-Brain Barrier Disruption and Local Macrophage/Microglia Activation Front. Pharmacol. 2020 11 161 10.3389/fphar.2020.00161 32226379 147. Takahashi S. Maeda T. Sano Y. Nishihara H. Takeshita Y. Shimizu F. Kanda T. Active Form of Vitamin D Directly Protects the Blood-Brain Barrier in Multiple Sclerosis Clin. Exp. Neuroimmunol. 2017 8 244 254 10.1111/cen3.12398 148. Jiao K.-P. Li S.-M. Lv W.-Y. Jv M.-L. He H.-Y. Vitamin D3 Repressed Astrocyte Activation Following Lipopolysaccharide Stimulation in Vitro and in Neonatal Rats NeuroReport 2017 28 492 497 10.1097/WNR.0000000000000782 28430709 149. Galoppin M. Kari S. Soldati S. Pal A. Rival M. Engelhardt B. Astier A. Thouvenot E. Full Spectrum of Vitamin D Immunomodulation in Multiple Sclerosis: Mechanisms and Therapeutic Implications Brain Commun. 2022 4 fcac171 10.1093/braincomms/fcac171 35813882 150. Gomez-Pinedo U. Cuevas J.A. Benito-Martin M.S. Moreno-Jimenez L. Esteban-Garcia N. Torre-Fuentes L. Matias-Guiu J.A. Pytel V. Montero P. Matias-Guiu J. Vitamin D Increases Remyelination by Promoting Oligodendrocyte Lineage Differentiation Brain Behav. 2020 10 e01498 10.1002/brb3.1498 31837111 151. Rao Z. Chen X. Wu J. Xiao M. Zhang J. Wang B. Fang L. Zhang H. Wang X. Yang S. Vitamin D Receptor Inhibits NLRP3 Activation by Impeding Its BRCC3-Mediated Deubiquitination Front. Immunol. 2019 10 2783 10.3389/fimmu.2019.02783 31866999 152. Ali N. Role of Vitamin D in Preventing of COVID-19 Infection, Progression and Severity J. Infect. Public Health 2020 13 1373 1380 10.1016/j.jiph.2020.06.021 32605780 153. Demir M. Demir F. Aygun H. Vitamin D Deficiency Is Associated with COVID-19 Positivity and Severity of the Disease J. Med. Virol. 2021 93 2992 2999 10.1002/jmv.26832 33512007 154. Hernandez J.L. Nan D. Fernandez-Ayala M. Garcia-Unzueta M. Hernandez-Hernandez M.A. Lopez-Hoyos M. Munoz-Cacho P. Olmos J.M. Gutierrez-Cuadra M. Ruiz-Cubillan J.J. Vitamin D Status in Hospitalized Patients with SARS-CoV-2 Infection J. Clin. Endocrinol. Metab. 2021 106 e1343 e1353 10.1210/clinem/dgaa733 155. Campi I. Gennari L. Merlotti D. Mingiano C. Frosali A. Giovanelli L. Torlasco C. Pengo M.F. Heilbron F. Soranna D. Vitamin D and COVID-19 Severity and Related Mortality: A Prospective Study in Italy BMC Infect. Dis. 2021 21 566 10.1186/s12879-021-06281-7 34126960 156. Carpagnano G.E. Di Lecce V. Quaranta V.N. Zito A. Buonamico E. Capozza E. Palumbo A. Di Gioia G. Valerio V.N. Resta O. Vitamin D Deficiency as a Predictor of Poor Prognosis in Patients with Acute Respiratory Failure Due to COVID-19 J. Endocrinol. Investig. 2021 44 765 771 10.1007/s40618-020-01370-x 32772324 157. Karonova T.L. Andreeva A.T. Golovatuk K.A. Bykova E.S. Simanenkova A.V. Vashukova M.A. Grant W.B. Shlyakhto E.V. Low 25(OH)D Level Is Associated with Severe Course and Poor Prognosis in COVID-19 Nutrients 2021 13 3021 10.3390/nu13093021 34578898 158. Infante M. Buoso A. Pieri M. Lupisella S. Nuccetelli M. Bernardini S. Fabbri A. Iannetta M. Andreoni M. Colizzi V. Low Vitamin D Status at Admission as a Risk Factor for Poor Survival in Hospitalized Patients With COVID-19: An Italian Retrospective Study J. Am. Nutr. Assoc. 2022 41 250 265 10.1080/07315724.2021.1877580 33600292 159. Vanegas-Cedillo P.E. Bello-Chavolla O.Y. Ramirez-Pedraza N. Rodriguez Encinas B. Perez Carrion C.I. Jasso-Avila M.I. Valladares-Garcia J.C. Hernandez-Juarez D. Vargas-Vazquez A. Antonio-Villa N.E. Serum Vitamin D Levels Are Associated With Increased COVID-19 Severity and Mortality Independent of Whole-Body and Visceral Adiposity Front. Nutr. 2022 9 813485 10.3389/fnut.2022.813485 35155539 160. Sabico S. Enani M.A. Sheshah E. Aljohani N.J. Aldisi D.A. Alotaibi N.H. Alshingetti N. Alomar S.Y. Alnaami A.M. Amer O.E. Effects of a 2-Week 5000 IU versus 1000 IU Vitamin D3 Supplementation on Recovery of Symptoms in Patients with Mild to Moderate COVID-19: A Randomized Clinical Trial Nutrients 2021 13 2170 10.3390/nu13072170 34202578 161. Dhawan M. Priyanka Choudhary O.P. Immunomodulatory and Therapeutic Implications of Vitamin D in the Management of COVID-19 Hum. Vaccines Immunother. 2022 18 2025734 10.1080/21645515.2022.2025734 162. Chen J. Mei K. Xie L. Yuan P. Ma J. Yu P. Zhu W. Zheng C. Liu X. Low Vitamin D Levels Do Not Aggravate COVID-19 Risk or Death, and Vitamin D Supplementation Does Not Improve Outcomes in Hospitalized Patients with COVID-19: A Meta-Analysis and GRADE Assessment of Cohort Studies and RCTs Nutr. J. 2021 20 89 10.1186/s12937-021-00744-y 34719404 163. Farid N. Rola N. Koch E.A.T. Nakhoul N. Active Vitamin D Supplementation and COVID-19 Infections: Review Ir. J. Med. Sci. 2021 190 1271 1274 10.1007/s11845-020-02452-8 33409846 164. White J.H. Vitamin D as an Inducer of Cathelicidin Antimicrobial Peptide Expression: Past, Present and Future J. Steroid Biochem. Mol. Biol. 2010 121 234 238 10.1016/j.jsbmb.2010.03.034 20302931 165. Wang C. Wang S. Li D. Chen P. Han S. Zhao G. Chen Y. Zhao J. Xiong J. Qiu J. Human Cathelicidin Inhibits SARS-CoV-2 Infection: Killing Two Birds with One Stone ACS Infect. Dis. 2021 7 1545 1554 10.1021/acsinfecdis.1c00096 33849267 166. Aloul K.M. Nielsen J.E. Defensor E.B. Lin J.S. Fortkort J.A. Shamloo M. Cirillo J.D. Gombart A.F. Barron A.E. Upregulating Human Cathelicidin Antimicrobial Peptide LL-37 Expression May Prevent Severe COVID-19 Inflammatory Responses and Reduce Microthrombosis Front. Immunol. 2022 13 880961 10.3389/fimmu.2022.880961 35634307 167. Chauss D. Freiwald T. McGregor R. Yan B. Wang L. Nova-Lamperti E. Kumar D. Zhang Z. Teague H. West E.E. Autocrine Vitamin D Signaling Switches off Pro-Inflammatory Programs of TH1 Cells Nat. Immunol. 2022 23 62 74 10.1038/s41590-021-01080-3 34764490 168. Zhang H. Shih D.Q. Zhang X. Mechanisms Underlying Effects of 1,25-Dihydroxyvitamin D3 on the Th17 Cells Eur. J. Microbiol. Immunol. 2013 3 237 240 10.1556/EuJMI.3.2013.4.1 24294492 169. BiLiR Y. BiLiR A. Saracoglu A. CiYiLtepe F. Bombaci E. Saracoglu K.T. DemiRhan R. Does Vitamin D Supplementation Reduce Cytokine Storm and Mortality in Geriatric Intensive Care Patients Diagnosed with COVID-19 J. Contemp. Med. 2022 12 233 237 10.16899/jcm.990057 170. Mohd S. Sharma S. Mishra A. Zahid Ashraf M. Vitamin D and Its Relationship with the Pathways Related to Thrombosis and Various Diseases Vitamin D Ozdemir O. IntechOpen London, UK 2021 978-1-83969-349-6 171. Constantinescu C.S. Farooqi N. O'Brien K. Gran B. Experimental Autoimmune Encephalomyelitis (EAE) as a Model for Multiple Sclerosis (MS): EAE as Model for MS Br. J. Pharmacol. 2011 164 1079 1106 10.1111/j.1476-5381.2011.01302.x 21371012 172. Fan W. Sun S. Zhang N. Zhang Y. Jiao P. Wang J. Gao G.F. Liu W. Bi Y. Yang L. Nasal Delivery of Thermostable and Broadly Neutralizing Antibodies Protects Mice against SARS-CoV-2 Infection Signal Transduct. Target. Ther. 2022 7 55 10.1038/s41392-022-00911-5 35190525 173. Munoz-Fontela C. Dowling W.E. Funnell S.G.P. Gsell P.-S. Riveros-Balta A.X. Albrecht R.A. Andersen H. Baric R.S. Carroll M.W. Cavaleri M. Animal Models for COVID-19 Nature 2020 586 509 515 10.1038/s41586-020-2787-6 32967005 174. Shou S. Liu M. Yang Y. Kang N. Song Y. Tan D. Liu N. Wang F. Liu J. Xie Y. Animal Models for COVID-19: Hamsters, Mouse, Ferret, Mink, Tree Shrew, and Non-Human Primates Front. Microbiol. 2021 12 626553 10.3389/fmicb.2021.626553 34531831 175. Gruber A.D. Firsching T.C. Trimpert J. Dietert K. Hamster Models of COVID-19 Pneumonia Reviewed: How Human Can They Be? Vet. Pathol. 2022 59 528 545 10.1177/03009858211057197 34856819 176. Frere J.J. Serafini R.A. Pryce K.D. Zazhytska M. Oishi K. Golynker I. Panis M. Zimering J. Horiuchi S. Hoagland D.A. SARS-CoV-2 Infection in Hamsters and Humans Results in Lasting and Unique Systemic Perturbations after Recovery Sci. Transl. Med. 2022 14 eabq3059 10.1126/scitranslmed.abq3059 35857629 177. Massanari R.M. Acceleration of Experimental Allergic Encephalomyelitis in Hamsters with Antecedent Virus Infection Clin. Immunol. Immunopathol. 1981 19 457 462 10.1016/0090-1229(81)90089-1 7249420 178. Massanari R.M. Wilson M.L. Differing Degrees of Susceptibility to Experimental Allergic Encephalomyelitis among Inbred Strains of Hamsters Clin. Immunol. Immunopathol. 1983 26 370 379 10.1016/0090-1229(83)90121-6 6191900 179. Massanari R.M. Paterson P.Y. Lipton H.L. Potentiation of Experimental Allergic Encephalomyelitis in Hamsters with Persistent Encephalitis Due to Measles Virus J. Infect. Dis. 1979 139 297 303 10.1093/infdis/139.3.297 448185 180. Wan Y. Shang J. Graham R. Baric R.S. Li F. Receptor Recognition by the Novel Coronavirus from Wuhan: An Analysis Based on Decade-Long Structural Studies of SARS Coronavirus J. Virol. 2020 94 e00127-20 10.1128/JVI.00127-20 31996437 181. Procaccini C. De Rosa V. Pucino V. Formisano L. Matarese G. Animal Models of Multiple Sclerosis Eur. J. Pharmacol. 2015 759 182 191 10.1016/j.ejphar.2015.03.042 25823807 182. Rangachari M. Kuchroo V.K. Using EAE to Better Understand Principles of Immune Function and Autoimmune Pathology J. Autoimmun. 2013 45 31 39 10.1016/j.jaut.2013.06.008 23849779 183. Irazoqui J.E. Ausubel F.M. 99th Dahlem Conference on Infection, Inflammation and Chronic Inflammatory Disorders: Caenorhabditis elegans as a Model to Study Tissues Involved in Host Immunity and Microbial Pathogenesis Clin. Exp. Immunol. 2010 160 48 57 10.1111/j.1365-2249.2010.04122.x 20415851 184. Gammon D.B. Caenorhabditis elegans as an Emerging Model for Virus-Host Interactions J. Virol. 2017 91 e00509-17 10.1128/JVI.00509-17 28931683 185. Felix M.-A. Ashe A. Piffaretti J. Wu G. Nuez I. Belicard T. Jiang Y. Zhao G. Franz C.J. Goldstein L.D. Natural and Experimental Infection of Caenorhabditis nematodes by Novel Viruses Related to Nodaviruses PLoS Biol. 2011 9 e1000586 10.1371/journal.pbio.1000586 21283608 186. Franz C.J. Renshaw H. Frezal L. Jiang Y. Felix M.-A. Wang D. Orsay, Santeuil and Le Blanc Viruses Primarily Infect Intestinal Cells in Caenorhabditis nematodes Virology 2014 448 255 264 10.1016/j.virol.2013.09.024 24314656 187. Hobert O. Specification of the Nervous System WormBook 2005 10.1895/wormbook.1.12.1 18050401 188. Hopkins C. Onweni C. Zambito V. Fairweather D. McCormick K. Ebihara H. Caulfield T. Zhang Y.S. Freeman W.D. Platforms for Personalized Polytherapeutics Discovery in COVID-19 J. Mol. Biol. 2021 433 166945 10.1016/j.jmb.2021.166945 33753053 189. Messing J. Heuberger R. Schisa J. Effect of Vitamin D3 on Lifespan in Caenorhabditis elegans CAS Curr. Aging Sci. 2013 6 220 224 10.2174/18746098113066660038 24304198 190. Wrzosek M. Lukaszkiewicz J. Wrzosek M. Jakubczyk A. Matsumoto H. Piatkiewicz P. Radziwon-Zaleska M. Wojnar M. Nowicka G. Vitamin D and the Central Nervous System Pharmacol. Rep. 2013 65 271 278 10.1016/S1734-1140(13)71003-X 23744412 191. Hendricks G. Mylonakis E. Expanding the Nematode Model System: The Molecular Basis of Inflammation and Infection Recovery in C. elegans Virulence 2017 8 244 245 10.1080/21505594.2016.1239011 27668316 192. Puthia M. Tanner L. Petruk G. Schmidtchen A. Experimental Model of Pulmonary Inflammation Induced by SARS-CoV-2 Spike Protein and Endotoxin ACS Pharmacol. Transl. Sci. 2022 5 141 148 10.1021/acsptsci.1c00219 35774232 193. Petruk G. Puthia M. Petrlova J. Samsudin F. Stromdahl A.-C. Cerps S. Uller L. Kjellstrom S. Bond P.J. Schmidtchen A. SARS-CoV-2 Spike Protein Binds to Bacterial Lipopolysaccharide and Boosts Proinflammatory Activity J. Mol. Cell Biol. 2021 12 916 932 10.1093/jmcb/mjaa067 194. Bi Z. Hong W. Que H. He C. Ren W. Yang J. Lu T. Chen L. Lu S. Peng X. Inactivated SARS-CoV-2 Induces Acute Respiratory Distress Syndrome in Human ACE2-Transgenic Mice Signal Transduct. Target. Ther. 2021 6 439 10.1038/s41392-021-00851-6 34952899 195. De Souza W.D.F. Zorzella-Pezavento S.F.G. Ayupe M.C. Salgado C.L. Oliveira B.D.C. Moreira F. da Silva G.W. Muraro S.P. de Souza G.F. Proenca-Modena J.L. Experimental Lung Inflammation Induced by Inactivated SARS-CoV-2 is Controlled by Intranasal Instillation of Vitamin D. Biology 2023 |
PMC10000584 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050769 cells-12-00769 Review Implications of Hypothalamic Neural Stem Cells on Aging and Obesity-Associated Cardiovascular Diseases Plakkot Bhuvana Di Agostino Ashley Subramanian Madhan * Grilli Mariagrazia Academic Editor Department of Physiological Sciences, Oklahoma State University, Stillwater, OK 74078, USA * Correspondence: [email protected] 28 2 2023 3 2023 12 5 76925 1 2023 14 2 2023 25 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). The hypothalamus, one of the major regulatory centers in the brain, controls various homeostatic processes, and hypothalamic neural stem cells (htNSCs) have been observed to interfere with hypothalamic mechanisms regulating aging. NSCs play a pivotal role in the repair and regeneration of brain cells during neurodegenerative diseases and rejuvenate the brain tissue microenvironment. The hypothalamus was recently observed to be involved in neuroinflammation mediated by cellular senescence. Cellular senescence, or systemic aging, is characterized by a progressive irreversible state of cell cycle arrest that causes physiological dysregulation in the body and it is evident in many neuroinflammatory conditions, including obesity. Upregulation of neuroinflammation and oxidative stress due to senescence has the potential to alter the functioning of NSCs. Various studies have substantiated the chances of obesity inducing accelerated aging. Therefore, it is essential to explore the potential effects of htNSC dysregulation in obesity and underlying pathways to develop strategies to address obesity-induced comorbidities associated with brain aging. This review will summarize hypothalamic neurogenesis associated with obesity and prospective NSC-based regenerative therapy for the treatment of obesity-induced cardiovascular conditions. aging cardiovascular conditions hypothalamus neural stem cells neuroinflammation obesity AHA-AIREA959725 NIH-NHLBIR15 HL14884401 OCASCR ResearchCollege of Veterinary Medicine Oklahoma State UniversityPLAKKOT-FY23 SUBRAMANIAN-FY23 This research was funded by AHA-AIREA Grant #959725 (M.S.), NIH-NHLBI R15 HL14884401 (M.S.), OCASCR Research Grant (M.S.), Research Advisory Committee Grant College of Veterinary Medicine Oklahoma State University PLAKKOT-FY23 (B.P.) and SUBRAMANIAN-FY23 (M.S.). pmc1. Introduction Neural stem cells (NSCs) in an adult brain are responsible for neurogenesis and regeneration of brain functions. The two primary NSC reservoirs (neurogenic niches) in an adult mammalian brain are the sub-ventricular zone (SVZ) of the lateral ventricles and the hippocampal dentate gyrus (DG) . In recent times, a third NSC pool, hypothalamic neural stem cells (htNSCs), were discovered . The htNSC population is sensitive to variations in nutrient intake and signaling. An increase in neurogenesis in the hypothalamus was observed upon acutely feeding a high fat diet (HFD) , whereas a reduced neurogenesis in the hypothalamus was noticed as a result of chronic HFD feeding , and 'inflammation' was suggested as a major factor in causing such pronounced changes in neurogenesis. Upon htNSCs culturing, we observed a significant increase in htNSCs after eight months in HFD-fed C57BL/6J male adult mice compared to the chow-fed controls (unpublished). Cellular senescence is an irreversible growth arrest in proliferating cells, which has been implicated in several neurodegenerative diseases . During the process of senescence, the NSCs lose their ability to proliferate and generate neurons . Supplementing mono-unsaturated fatty acids, such as oleic acid, in the diet caused lipid droplets to develop in ependymal cells and contributed to a decrease in neurogenesis in SVZ in the Alzheimer's disease mouse model, 3xTg-AD . Likewise in obesity, SVZ showed an increase in senescent glial cells carrying excessive fat deposits, and genetically ablating these senescent glial cells restored neurogenesis . Thus, modifying the lipid content in the diet can replenish the old neurogenic pool. In this review, we will summarize hypothalamic neurogenesis associated with obesity and aging and explore the possibilities of NSC-based regenerative therapy to treat obesity-induced cardiovascular conditions. 2. HtNSCs and Obesity NSCs are multipotent and they generate neurons, oligodendrocytes, and glia in the nervous system . Varied levels of neural inflammation are observed in many neurological disorders or neurodegenerative diseases in human beings . Their progression involves mediators of inflammation that are synthesized and secreted by various CNS cells, such as astrocytes, microglia, and oligodendrocytes . Both beneficial and detrimental effects are observed in inflammatory conditions, which makes it unclear to specify the exact role of inflammation on NSCs. Certain pathways, after long term activation, cause energy imbalance, abnormal nutrient metabolism, restricted neurogenesis, proliferation, and differentiation of neural stem cells leading to metabolic and cognitive abnormalities. In the hypothalamus, the medio-basal hypothalamus (MBH) and the 3rd ventricle wall are observed to be the NSC niches . Some studies state that mainly adult NSCs are observed in the MBH . The MBH is a predominant region for physiological homeostasis of the entire body. Many neural progenitors or specialized ependymal cells that line the 3rd ventricle are observed to be glia-like tanycytes. They send processes to the arcuate nucleus and ventro-medial nucleus of the hypothalamus. Functionally these tanycytes are observed to be glucosensitive, reacting to metabolic stimulation and signal variations caused by feeding and energy balance . Properties of tanycytes include ATP release, purinergic P2Y1 receptors, ectonucleoside triphosphate diphosphohydrolase 2 (NTPDase2) expression , and reacting to the activation of these receptors by the means of intense Ca2+ waves . This is similar to the signaling mechanisms in stem cells. Expression of doublecortin-like proteins, nestin and vimentin , linked to neural precursor cells are observed in humans and rodent tanycytes. The expression of Sox2 , a nuclear transcription factor and NSC marker, is found in a few of the tanycytes, especially in the subventricular zone and dentate gyrus. In adult mice, it is mainly expressed in a group of cells in the MBH, particularly within the hypothalamic third-ventricle wall . However, a few studies have shown rare occurrences of proliferating neurogenic progenitors in the human dentate gyrus . One of the studies also observed human paralaminar nuclei of the amygdala showing persistence of immature excitatory neurons for decades . Thus, the possibility of observing immature non-proliferative hypothalamic neurons cannot be denied and future studies focusing on confirming their ability to proliferate and differentiate could possibly reveal their normal functionality. The MBH regulates body weight, feeding, and glucose balance via melanocortin signals based in the arcuate nucleus (ARC), mainly via orexigenic agouti-related peptide (AGRP) neurons and anorexigenic proopiomelanocortin (POMC) neurons . Leptin and insulin, which vary with different fat mass conditions and feeding patterns, affect these two neurons and the process is crucial for body weight homeostasis . The studies also showed decrease in responsiveness to leptin and insulin by these neurons upon chronic feeding of a high-fat diet (HFD), resulting in type-2 diabetes (T2D) and HFD-induced obesity. A 10% loss in POMC neurons was observed in the hypothalamus upon long term HFD feeding . Neural precursors giving rise to different neurons were observed to have POMC gene expression . Considering these data and mechanisms, there is evidence of dysregulation of neurogenesis in the hypothalamus of obese subjects. Based on many recent studies, neurogenesis has been observed in adult rodents and htNSCs in adult MBH contribute to the regulation of metabolic physiology . Hence, future studies could be focused on developing htNSCs as a treatment regimen for obesity and its related disorders, such as diabetes. 3. HtNSCs and Inflammation Microglia are brain-resident macrophages that contribute to reduced neurogenesis in aging and play a predominant role in the inflammatory response . Through microglia sorting studies, we observed a significant elevation of activated microglia in the hypothalamus of four-month HFD-fed young adult male mice compared to the chow-fed controls (unpublished). Activated microglia have the potential to release proinflammatory cytokines that can be harmful to NSCs, neurons and other glial cells. Among the complex neural immune reactions in adult NSCs, inflammatory cytokines are observed to majorly affect differentiation, proliferation, migration, and survival . Inhibition of neurogenesis is achieved by pro-inflammatory cytokines whereas an increase in neurogenesis is observed by anti-inflammatory cytokines . The gene expression studies in our lab revealed a significant increase in proinflammatory markers, such as IL1b, MCP1, and TNFa, in the whole hypothalamus of middle-aged, eight-month HFD-fed male mice compared to controls (unpublished). However, an anti-inflammatory cytokine, such as the transforming growth factor-beta (TGFb), can enhance endothelial cells of adult NSC during aging . In addition to these, a chemokine, CCL11, was observed to be increased in aged mice, both in blood and cerebrospinal fluid (CSF), which further caused a decline in neurogenesis leading to cognitive function impairment . Exercise and restriction of calories can cause variations in systemic factors, and hence, act as adult NSC function modulators . Upon over-nutrition, the IkB kinase-b/nuclear transcription factor NF-kB (IKKb/NF-kB) pathway, that plays a crucial role in many physiological processes, gets activated; this can cause SOCS3, a suppressor of cytokine signaling-3 gene upregulation in the hypothalamus, to inhibit insulin and leptin signaling, leading to resistance . Studies have confirmed that, in the neurons of the hypothalamus in mice, SOCS3 knockout leads to an improvement in central leptin signaling and reduced obesity . Similar effects were observed in central IKKb knockout mice and, in the MBH, SOCS3 overexpression decreased the neural IKKb inhibition effect on obesity reduction . Like SOCS3, protein tyrosine phosphatase 1B (PTP1B) causes inhibition of leptin and insulin signaling and was observed to have a role in the IKKb/NF-kB inflammatory pathway. PTP1B expression in the hypothalamus can be increased by TNF-a by activating the IKKb/NF-kB pathway, mainly by being a transcriptional target . Inhibition of PTP1B in neurons resolved leptin resistance, glucose disorders, and obesity induced by over-nutrition . It is assumed that neural PTP1B may form a link with metabolic disease pathways and neurodegenerative diseases as it had an effect on genetic mouse models of Alzheimer's disease . In the forebrain, degeneration of GABAergic interneurons was mediated by an overproduction of the cytokine interleukin-6 in diabetes and obesity, which leads to NF-kB activation and release of neurotoxic inflammatory products . Therefore, alleviating chronic diet-induced neuroinflammation by exploring the pathways associated with the metabolic control function of htNSC and identifying their therapeutic potential is essential. 4. Nrf2, an Important Transcription Factor Affecting NSC Populations in Obesity Various factors affect NSC populations in obesity, including hormonal factors, transcription factors, inflammatory factors such as cytokines and chemokines, epigenetic changes and chromatin stability, oxidative stress, DNA damage, hyperlipidemia/hyperglycemia, etc. Nuclear factor E2-related factor 2 (Nrf2) is a major transcription factor that regulates basal and induced expression of antioxidant response element genes in response to oxidative stress. Functions of Nrf2 also include stem cell survival, apoptosis, autophagy, mitochondrial biogenesis, and many more, in addition to aging processes . Studies in our lab observed an elevated expression of Nrf2 in the hypothalamus of adult obese male mice, along with a significant increase in htNSCs (unpublished). In a previous study, increased oxidation, or reactive oxygen species in adult mouse NSCs, promoted their ability to generate neurons and proliferate . Self-renewal of stem cells was observed to be regulated by Nrf2, along with differentiation initiation with the support of epigenetic factors and transcription regulators . Nrf2 expression and transcriptional activity steadily increased during the induced oluripotent stem cells (iPSC) differentiation process that peaked in later stages . Restoration of age-related loss of hippocampal function was evidenced by transplanting Nrf2-overexpressing young NSCs , indicating the critical role of Nrf2 in mediating NSC/neural progenitor cell (NPC)- dependent neurogenesis in aging. Redox homeostasis by Nrf2 critically mediates the differentiation ability of different stem cell types to survive oxidative stress, which could gradually reduce during aging . Thus, obtaining insight into one of the main transcription factors, Nrf2, that can resist oxidative stress, could provide fundamental knowledge about changes in htNSCs during neuroinflammation and lead to development of an associated therapeutic strategy. 5. HtNSCs and Aging A continuous decline in physiological integrity is observed during aging. Characteristic intertwining factors that contribute to the complex aging process include deregulated nutrient sensing, cellular senescence, epigenetic alterations, genomic instability, loss of proteostasis, mitochondrial dysfunction, telomere attrition, change in intercellular communication, and exhaustion of stem cells . It has been observed in various research that the hypothalamus is particularly important in aging but the underlying cellular mechanism is not known in depth. The IkB kinase-b (IKKb) pro-inflammatory axis in the hypothalamus and its downstream nuclear transcription factor, NF-kB, (IKKb/NF-kB signaling) is over-stimulated in over-nutrition or aging . Systemic aging is directed by the hypothalamic IKKb/NF-kB pathway via inflammatory crosstalk between neurons and microglia by inhibiting gonadotropin-releasing hormone (GnRH) production, and so counteracting inflammation or GnRH therapy could partly regress degenerative signs of aging . Maternal inflammation has been observed to cause reduced ventricular cell proliferation in developing fetal mouse brain . In young mice, a high number of cells co-expressing Sox2 and the polycomb complex protein, Bmi-1, a nuclear protein that is vital for self-renewal of NSCs and hematopoietic stem cells , were observed in the third-ventricle wall, whereas the ones in the MBH were found to be sparse. However, a gradual decrease in these cells was observed as age increased, which was initiated in the ventral region of 3rd ventricle wall within the MBH in 11-16-month-old mice and was totally lost in 22-months-and-older ones. Thus, various studies that aim to evaluate the exact time required to intervene in an inflammatory condition/pathway in the brain will provide more understanding upon which to formulate therapeutic clinical strategies for different neural stem cell niches. Senescent glial cell accumulation is observed in proximity to the lateral ventricles along with excessive fat deposition within them. Upon removal of senescent cells from HFD or obese mice deficient in leptin receptors, neurogenesis being restored and a decline in anxiety-related behavior was observed . Hence from subsequent studies, they concluded that the topmost contributors to obesity-induced anxiety are senescent cells. Therefore, senolytic drugs have opened a novel therapeutic pathway to treat neuropsychiatric disorders. Alterations in mitochondrial structure and function may cause deleterious effects in adult NSC, which could drive the aging process . Abnormal toxic by-product accumulation, including of reactive oxygen species (ROS), accompanies this event . SOD2, an antioxidant enzyme that is regulated by FoxO3, a transcription factor associated with longevity , protects adult NSCs in mice . An increased level of ROS and a decrease in the potential for self-renewal of adult NSCs was observed in mice that were deficient in FoxO1, 3, and 4 . Other dysfunctions of mitochondria that contribute to the aging of NSCs include mitochondrial protein oxidation, variations in mitochondrial membrane composition, and abnormal mitophagy . During mammalian NSC division, protein segregation is affected by age, mainly by means of diffusion barrier alteration. The stem cells are kept free of damage by the diffusion barrier that facilitates asymmetric segregation of damaged proteins among daughter cells . Like yeast, efficient compartmentalization of cellular damage is achieved in young rodent NSCs and that can protect these proliferative cells. As age advances, this efficiency is reduced, causing aged NSCs to be exposed to excessive cell damage . A mitochondrial function regulator, hypoxia-inducible factor-1a (HIF-1a), is essential for the maintenance of adult NSCs in their hypoxic niches. HIF-1a plays a major part in cell adaptation under hypoxia by inducing transcriptional responses. Thus, for proper adult NSC proliferation and subsequent differentiation, oxygen availability is critically important . An abnormal oxygen-sensing pathway may be responsible for the neurogenic decline in aging . Thus, the use of anti-inflammatory agents along with senolytic and associated htNSC therapy have the potential to strategically counteract diet-induced chronic neuroinflammation and aging. This could possibly pave way to new therapeutic regimens in obesity-induced cardiovascular conditions. 6. Molecular Pathways Associated with NSC Inflammation and Aging Certain nutrient-sensing mechanisms that can be associated with aging have been considered modifiers of adult NSCs. Adult NSC proliferation and differentiation can be stimulated by insulin-like growth factor 1 (IGF-1) , and a reduced IGF-1 level has been associated with cognitive aging . However, lifelong IGF-1 exposure may be the reason for an age-related reduction in adult neurogenesis . An important metabolic regulation coordinator is the mammalian target of rapamycin (mTOR), which has two types, viz., mTORC1 and mTORC2 . Regulation of body weight and feeding behavior is primarily controlled by mTOR1 using ghrelin and leptin signaling, in addition to control of gluconeogenesis and adipogenesis peripherally in many tissues . Size, morphology, and neuronal cell numbers are controlled by mTORC2, along with energy balance regulation in the hypothalamus. In POMC neurons in aged mice, an elevation in mTOR activity was observed , which can indirectly lead to POMC neuronal soma enlargement and a decline in the projection of neurites to the paraventricular nucleus (PVN), which causes age-dependent obesity . It has been observed that, upon intracerebral injection, rapamycin causes mTOR inhibition which further leads to neurite projection and neuronal excitability in POMC, establishing a decline in body weight and food consumption; hence, age acceleration is achieved by the mTOR pathway. Therefore, to delay aging and improve the lifespan, this pathway can be considered a potential target for therapeutic intervention. As previously discussed, during aging, a decrease in htNSCs was observed . In addition, mice models with gene silencing mediating Bmi1+ depletion in stem cells showed a significant reduction in cognition, sociality, muscle endurance, coordination, and spatial memory. In other mice models, a decline in lifespan was observed in Sox2+ stem cell-depleted animals. Hence, replenishing new htNSC from a newborn mouse into the MBH of a middle-aged mouse could enhance the lifespan and delay age-associated physiological decline . Exogenous implantation of stem cells into the hypothalamus caused secretion of microRNA-containing exosomes, which delayed physiological deficits in aging. Suppression of NF-kB activation was achieved in neurons due to these microRNAs, and GnRH secretion was also restored . As a result, during aging, htNSC loss might cause systemic physiological changes due to underlying inflammation. Through Wingless-related integration site (Wnt)-mediated signaling by astrocytes, adult NSC expansion is induced in a paracrine manner . As age increases, Wnt3 expression reduces in astrocytes, which causes further neurogenic decline . Expression of survivin is decreased in adult NSCs due to an age-associated decline in Wnt-mediated signaling in the astrocytes that leads to a quiescent phase in adult NSCs . Release of the Wnt inhibitor, DKK1, from astrocytes is increased in NSC niches during aging, which decreases neurogenesis . Extracellular matrix composition, mechanical properties, and arrangement have a role in adult NSC function, which varies with injury, disease, and aging . A high fat mass expression and obesity associated gene (FTO) was observed in adult NSCs . A smaller number of BrdU+ and Ki67+ cells was also observed during FTO loss, showing a decline in adult NSC and reduced proliferating capacity, along with a decline in glial and neuronal differentiation, making adult NSC less multipotent. In addition, in adult mice, FTO loss was observed to decrease adult NSC proliferation and caused inhibition of neuronal differentiation in both SGZ and SVZ regions. Thus, adult NSC activity modulation is achieved by FTO through m6A modification regulation of selective transcripts that can indirectly affect the gene expression . Alteration of important signaling for neurodevelopment is observed when a change in nutrient or neurotrophic environment is observed. Brain abnormalities and decreased brain weight, along with altered glial and neuronal protein expression is observed in mice that have a paucity of leptin signaling and varied expression of neuronal and glial proteins . Elevated proliferation and decline in neural stem/progenitor cells (NPC) are observed in adult rats with type II diabetes and in rats showing hyperglycemia. NPCs do not respond to growth factors and form neurospheres (NS) that are smaller in size. In addition, a decline in neurogenesis is observed in type I diabetic mice or rats treated with streptozotocin . Along with anorexigenic response signaling, during fetal life, insulin and leptin help in neuronal development and their neurotrophic effects are mediated by the MAPK (ERK/MAPK) pathway that resulted in phosphorylation of ERK1/2 . Significant neuronal differentiation was induced by leptin in differentiation conditions along with elevated early and late neuronal marker expression ; whereas the late neuronal marker neuronal nuclei (NeuN) showed no significant increase and a normal elevation in early neuronal markers, such as doublecortin (DCX) and neuron-specific class III b tubulin (Tuj1), were observed upon insulin exposure . According to these studies, it was concluded that maternal diabetes and differential exposure of the fetus to insulin and leptin could result in reduced growth or macrosomia that could have a significant effect on the development of a fetal brain. 7. The Hypothalamus and the Sympathoexcitatory Effect In various studies of the effects of leptin on the hypothalamus, it has been observed that a-melanocyte-stimulating hormone (a-MSH) or melanotan II (agonist of MC3/4R (MTII)), upon intracerebroventricular (ICV) administration, enhanced sympathetic nerve activity (SNA), however agouti-related protein or MC3/4R broad brain inhibition with ICV SHU9119 blocked leptin's sympathoexcitatory effect . This is based on the understanding that a-MSH and glutamate are two major excitatory signals to the PVN, a cardiogenic center in the hypothalamus , that can mediate leptin's sympathoexcitatory effects. POMC neurons synthesize and release a-MSH. These neurons are in arcuate nucleus (ArcN), which projects to various sites in the hypothalamus, including the PVN , and regulates autonomic activity; however, the role of PVN MC3/4 is ambiguous. Glutamatergic signals are received by the PVN from various regions that include the dorsal medial hypothalamus, ventral medial hypothalamus, lateral hypothalamus, and ArcN, wherein elevated SNA is observed due to the action of leptin . A small group of POMC neurons in the ArcN also expresses the glutamate vesicular transporter (VGLUT-2) . PVN glutamate receptors blockade decreases the ArcN's non-specific chemical stimulation-mediated sympathoexcitatory effects . Along with this excitatory signaling, inhibitory neurons, such as neuropeptide Y (NPY) neurons of the ArcN, are projected into the PVN . NPY neurons are inhibited by leptin in the ArcN and PVN neuron firing is inhibited by NPY, which gets stimulated by a-MSH or plasma leptin elevation. By various studies it has been identified that elevated SNA is observed in obesity, especially in the kidney and hindlimb, for which a leptin increase and hypothalamic melanocortin activity elevation are predominant activities . In mice and rats, expression of NPY in the ArcN/PVN is reduced by diet-induced obesity or, in the ArcN, by NPY mRNA levels . In the PVN region, obesity-prone rats that were inbred showed a reduction in agouti-related protein/NPY processes . Tonic NPY inhibition decline is essential for leptin-induced sympathetic outflow driven via PVN MC3/4R. It is inferred from this that obesity plays a role in SNA inhibition and it is due to tonic activity of NPY, which further reveals an elevated a-MSH excitation . htNSCs are predominantly found adjacent to the PVN of the hypothalamus lining the 3rd ventricle . Based on these studies, there is a need for detailed investigation into the link between the variation in NSC levels associated with different conditions, such as age, diet etc., and sympathoexcitatory activity. 8. Time-Restricted Feeding and Its Effect on NSCs Reduced energy consumption without any effect on nutritional value is characteristic of dietary restriction (DR). It can be alternatively described as caloric restriction (CR) and, in a broader way, termed as periodic fasting, short-term starvation, intermittent fasting (IF), and fasting-mimetic diets . In maintaining proper health and physiology, a crucial role is played by the type and amount of diet . Adult stem cells are important for tissue regeneration and homeostasis and these stem cells can differentiate and self-renew into specialized cell types. Dietary changes, environment, and nutrient variation influence the stem cells via function alteration. In various studies, a positive effect was observed in stem cells when calories were restricted, especially an increase in the function of intestine and skeletal muscle stem cells, in addition to an elevated quiescence of hematopoietic stem cells (HSCs). In addition, time-restricted feeding has been shown to protect neuronal stem cells, intestinal stem cells, and HSCs from injury, especially stroke and neurodegenerative diseases in the brain . HFD impairs neurogenesis and hematopoiesis, and it can create opportunities for tumorigenesis. Characteristic changes in metabolic pathways in the brain are achieved by IF, mainly by ketogenic amino acid and fatty acid breakdown and an elevation in stress resistance . A neuroprotective effect can be achieved via IF by activation of many signaling pathways . IF in rodents has shown an increase in long-term potentiation (LTP) at synapses in the hippocampus and an increase in hippocampal neurogenesis in comparison with animals with a sedentary lifestyle that are fed ad libitum (AL) diet. BrdU-labeled cell number in the dentate gyrus was elevated in the mice that were intermittently fasted . They also used Ki67 as a marker to evaluate cell proliferation by identifying an increase in dentate gyrus Ki67-labeled cells in mice that were fed with an IF diet. Mice subjected to IF for three months showed an elevated level of hippocampal nestin and NeuN (protein markers for precursor/neuronal stem cells), and also PSD95 (a scaffolding protein that is a potent regulator of synaptic strength) compared to AL mice , which demonstrated an increase in hippocampal neurogenesis and a strengthening of synaptic connections after IF. The researchers also showed that a pathway essential for neural stem cell maintenance in the mammalian brain , the Notch 1 signaling pathway, was shown to become activated mainly by upregulation of full-length Notch 1, Notch intracellular domain (NICD1), and transcription factor HES5 (involved in the formation of neurospheres) after IF. The stress resistance ability of brain cells is activated by IF by causing various changes in brain metabolic pathways . The changes in metabolic pathways during IF can be injurious to the brain and through activation of the brain-derived neurotrophic factor (BDNF) signaling pathway, a neuroprotective state is achieved. Downstream transcription factor activation that helps in energy balance and neurogenesis is made possible by BDNF, and one such transcription factor is cAMP response element-binding protein (CREB). To differentiate stem cells into matured neurons, collaboration between the Notch signaling pathway and the CREB and BDNF signaling pathways is essential . An increase in BDNF and p-CREB expression has been seen in IF compared to AL animals . Without leading to malnutrition, CR is a 20-40% reduction in intake of calories. It is known to cause life-span increase, prolonged onset of diseases that are age related, and decrease in the incidence of cancer in different tissues and organisms . The link between CR and longevity is under the influence of the downregulation of major nutrient sensing pathways, including those of insulin or IGF-1, and signaling by mTOR . Very few studies have been documented on the positive and negative effects of CR on NSCs. Two-days-a-week fasting or alternate-days fasting (IF) in animals have been shown to decrease clinical symptoms caused by age-related neuronal diseases such as Alzheimer's disease, and the animals that were fasted also perform better after stroke, which is an acute injury . After three weeks of a three-month period of IF, an elevated NSC proliferation in the rats and mice dentate gyrus was observed . An elevated BDNF was associated with these positive effects. However, various studies showed that neuronal survival ability was altered by fasting rather than induction of NSC proliferation. In the dentate gyrus of mice, an increase in neuron and glia numbers was observed within 72 h of feeding a fasting-mimicking diet (FMD), along with a reduced IGF-1/PKA signaling . In addition, an increase in mesenchymal stem and progenitor cell number and proliferation were observed on FMD repeated feeding in aged animals, and in aged mice; rebalanced output from HSCs and progenitors were also observed . Therefore, time restricting feeding can be a neuroprotective strategy for replenishing lost NSCs in chronic neuroinflammatory conditions. 9. Exosomes from the HtNSCs HtNSCs have a distinct endocrine function, to release excessive amounts of microRNAs (miRNAs)-containing exosomes . In addtion, they have certain long non-coding RNAs (lncRNAs) that possess the ability to maintain pluripotency and embryonic stem cell neurogenesis , self-renewal of cancer stem cells , and reprogramming of pluripotent stem cells . LncRNAs may play a unique role in determining the fate of these stem cells in cellular senescence regulation . An abundant lncRNA, Hnscr in the htNSCs of young mice, drastically reduces as the mice age . Hnscr regulates htNSC senescence and mouse aging by binding to YB-1, a multi-functional protein that controls protein translation , and also regulates DNA repair , protecting it from protein degradation and ubiquitination. YB-1 acts as a repressor of transcription, inhibiting p16INK4A expression in htNSCs , and hence could be targeted to modulate senescence in htNSC. According to , TF2A treatment, isomeric theaflavin monomer, and a black tea derivative , improved YB-1 stability, diminished htNSC senescence, and decreased the level of aging related physiological downturn in mice. By various studies it has been observed that htNSC loss causes systemic aging within a short time and the exosomal miRNAs secreted by these cells mediate anti-aging properties . Aging can be correlated with modulation of some gene expressions by certain non-coding RNAs. In aged adult NSC, heterochronic micro-RNA let-7b upregulation is observed. Repression of Hmga2, a high mobility group transcriptional regulator, is observed upon let-7b overexpression, which indirectly potentiates p16lnk4a (an inhibitor of cyclin-dependent kinase and activator of Rb) and p19Arf expression, improving the stability of p53 protein . Therefore, it slows down the progression of cell cycle and induces senescence , leading to reduced adult NSC functioning and neurogenesis. However, deficiency of p16INK4a in aged mice diminished this effect . Let-7b initiates differentiation and inhibits proliferation of neural stem cells by targeting Tlx and cyclin D1 in adult NSC and embryonic brains . A higher-to --lower/quiescent shift in NSC proliferative state from fetus to adult is contributed to by Imp1, a different let-7b target, even though it is not expressed in adult NSC . As a result, changes in let-7b may initiate aging in adult NSC. The gene regulation mediated by micro RNAs impacts healthy aging as well as aging associated with neurodegenerative diseases . Administering exosomes derived from NSCs (exo-NSCs) could restore BDNF signaling and memory in HFD mice , providing suggestive evidence of the potential therapeutic effect of exo-NSCs on HFD-induced NSC dysregulation in obesity. Hence, further studies on differential expression of certain exosomal non-coding RNAs must be performed to form an understandable association with pathological and healthy aging. 10. Challenges Associated with NSCs for Regenerative Medicine and Future Perspectives Immunological rejection is one of the major difficulties in stem cell therapy. This could be addressed by isolating NSCs from the same subjects that require the therapy to prevent immunological reaction to the newly transplanted stem cells. Administering immunosuppressive drugs could be an additional or alternative option even though it has a lot of side effects. Another challenge is to make certain that the transplanted cells grow enough without causing tumor development and karyotypic instability. According to , there is a critical challenge in isolating multipotent NSCs from cell culture for transplantation as the majority of neurospheres in vitro are heterogenic with varying developmental stages and gene expression mainly due to ex vivo culture conditions. Overcoming these challenges and establishing NSC-based therapy for obesity-induced comorbidities, especially cardiovascular conditions, to improve functional outcomes through associated multimodal mechanisms is tremendously foresighted. Based on the difficulty in accessing the brain to collect tissues for processing from live animals, using induced pluripotent stem cell (iPSC) technology is a solution that could produce in vitro NSCs or neurons for transplantation. As iPSCs can be non-invasively obtained from live subjects, and to reduce the risk of immune rejection, reprogramming these cells to NSCs or neurons could provide autologous engraftments . 11. Conclusions HtNSCs could be a potential therapy in obesity-induced cardiovascular diseases. Exosomes derived from htNSCs could be an alternative to or a conjunction with NSC therapy, being a minimally invasive technique to reverse aging and degenerative changes in the CNS. The relationship between htNSC dysregulation and sympathetic nerve response in obesity has never been studied. As brain microglia activation is a predominant indicator of neuroinflammation in hypertension, restoring a normal population of glia and neurons within the cardiogenic centers of the brain cannot be ruled out. Thus, identifying associated htNSC mechanisms and pathways could bring novel insight to therapeutic strategies in obesity-associated hypertension or sympathetic nerve overactivity. Acknowledgments We would like to acknowledge Rajasingh Johnson, Department of Bioscience Research, UTHSC for the suggestions made while forming this review. Author Contributions B.P. has made substantial contributions to the design and concept of the article including figures, data acquisition and analysis of the quoted unpublished data and interpretation of relevant literature. A.D.A. produced figures and critically revised the literature. M.S. conceptualized and critically revised the article for intellectual content. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement No new data was created as this is a review, and some data is unavailable due to privacy or ethical restrictions. Data will be published with ethical considerations in future manuscript. Conflicts of Interest The authors declare no conflict of interest. Figure 1 A simplified pathway showing excitatory circulatory signaling from SFO and OVLT, which are circumventricular organs lacking blood brain barrier, passing to the adjacent PVN in the hypothalamus and RVLM in the brainstem, causing an efferent sympathetic nerve response from the intermediolateral column of the spinal cord to the heart and blood vessels, resulting in increased heart rate and increased blood pressure associated with vasoconstriction. Figure 2 A schematic illustration of the response to feeding chow (left) and high fat diet (HFD) (right) on hypothalamic paraventricular nucleus (PVN) and hypothalamic neural stem cells (htNSCs) lining the 3rd ventricle in mice. HFD could potentially cause an increase in senescent glial cells within the PVN that can release senescence-associated secretory phenotype (SASP) factors causing a proinflammatory response, further leading to htNSC dysregulation. This could, again, cause a reduction in functional glia and neurons in the PVN. The circulatory inflammatory mediators could also potentially cross the blood brain barrier and cause a direct effect on the htNSCs. Chow-fed control showed normal cell populations. Figure 3 A simplified protocol for the potential therapeutic effect of exo-NSCs in obesity-induced NSC dysregulation. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Bond A.M. Ming G.L. Song H. Adult Mammalian Neural Stem Cells and Neurogenesis: Five Decades Later Cell Stem Cell 2015 17 385 395 10.1016/j.stem.2015.09.003 26431181 2. Stolp H.B. Molnar Z. Neurogenic niches in the brain: Help and hindrance of the barrier systems Front. Neurosci. 2015 9 20 10.3389/fnins.2015.00020 25691856 3. Silva-Vargas V. Crouch E.E. Doetsch F. Adult neural stem cells and their niche: A dynamic duo during homeostasis, regeneration, and aging Curr. Opin. Neurobiol. 2013 23 935 942 10.1016/j.conb.2013.09.004 24090877 4. Bolborea M. Dale N. Hypothalamic tanycytes: Potential roles in the control of feeding and energy balance Trends Neurosci. 2013 36 91 100 10.1016/j.tins.2012.12.008 23332797 5. Pellegrino G. Trubert C. Terrien J. Pifferi F. Leroy D. Loyens A. Migaud M. Baroncini M. Maurage C.-A. Fontaine C. A comparative study of the neural stem cell niche in the adult hypothalamus of human, mouse, rat and gray mouse lemur (Microcebus murinus) J. Comp. Neurol. 2018 526 1419 1443 10.1002/cne.24376 29230807 6. Xu Y. Tamamaki N. Noda T. Kimura K. Itokazu Y. Matsumoto N. Dezawa M. Ide C. Neurogenesis in the ependymal layer of the adult rat 3rd ventricle Exp. Neurol. 2005 192 251 264 10.1016/j.expneurol.2004.12.021 15755543 7. Lee D.A. Bedont J.L. Pak T. Wang H. Song J. Miranda-Angulo A. Takiar V. Charubhumi V. Balordi F. Takebayashi H. Tanycytes of the hypothalamic median eminence form a diet-responsive neurogenic niche Nat. Neurosci. 2012 15 700 702 10.1038/nn.3079 22446882 8. Li J. Tang Y. Cai D. IKKb/NF-kB disrupts adult hypothalamic neural stem cells to mediate a neurodegenerative mechanism of dietary obesity and pre-diabetes Nat. Cell Biol. 2012 14 999 1012 10.1038/ncb2562 22940906 9. Campisi J. Cancer, aging and cellular senescence In Vivo 2000 14 183 188 10757076 10. Balasubramanian P. Branen L. Sivasubramanian M.K. Monteiro R. Subramanian M. Aging is associated with glial senescence in the brainstem-implications for age-related sympathetic overactivity Aging 2021 13 13460 10.18632/aging.203111 34038388 11. Nada M.B. Slomianka L. Vyssotski A.L. Lipp H.P. Early age-related changes in adult hippocampal neurogenesis in C57 mice Neurobiol. Aging 2010 31 151 161 18455269 12. Chinta S.J. Woods G. Rane A. DeMaria M. Campisi J. Andersen J.K. Cellular senescence and the aging brain Exp. Gerontol. 2015 68 3 7 10.1016/j.exger.2014.09.018 25281806 13. Kuhn H.G. Dickinson-Anson H. Gage F.H. Neurogenesis in the dentate gyrus of the adult rat: Age-related decrease of neuronal progenitor proliferation J. Neurosci. 1996 16 2027 2033 10.1523/JNEUROSCI.16-06-02027.1996 8604047 14. Maslov A.Y. Barone T.A. Plunkett R.J. Pruitt S.C. Neural stem cell detection, characterization, and age-related changes in the subventricular zone of mice J. Neurosci. 2004 24 1726 1733 10.1523/JNEUROSCI.4608-03.2004 14973255 15. Hamilton L.K. Dufresne M. Joppe S.E. Petryszyn S. Aumont A. Calon F. Barnabe-Heider F. Furtos A. Parent M. Chaurand P. Aberrant lipid metabolism in the forebrain niche suppresses adult neural stem cell proliferation in an animal model of Alzheimer's disease Cell Stem Cell 2015 17 397 411 10.1016/j.stem.2015.08.001 26321199 16. Ogrodnik M. Zhu Y.I. Langhi L.G. Tchkonia T. Kruger P. Fielder E. Victorelli S. Ruswhandi R.A. Giorgadze N. Pirtskhalava T. Obesity-induced cellular senescence drives anxiety and impairs neurogenesis Cell Metab. 2019 29 1061 1077.e8 10.1016/j.cmet.2018.12.008 30612898 17. Clarke D.L. Johansson C.B. Wilbertz J. Veress B. Nilsson E. Karlstrom H. Lendahl U. Frisen J. Generalized potential of adult neural stem cells Science 2000 288 1660 1663 10.1126/science.288.5471.1660 10834848 18. Das S. Basu A. Inflammation: A new candidate in modulating adult neurogenesis J. Neurosci. Res. 2008 86 1199 1208 10.1002/jnr.21585 18058947 19. Lucin K.M. Wyss-Coray T. Immune activation in brain aging and neurodegeneration: Too much or too little? Neuron 2009 64 110 122 10.1016/j.neuron.2009.08.039 19840553 20. Block M.L. Hong J.-S. Microglia and inflammation-mediated neurodegeneration: Multiple triggers with a common mechanism Prog. Neurobiol. 2005 76 77 98 10.1016/j.pneurobio.2005.06.004 16081203 21. McNay D. Briancon N. Kokoeva M.V. Maratos-Flier E. Flier J.S. Remodeling of the arcuate nucleus energy-balance circuit is inhibited in obese mice J. Clin. Investig. 2012 122 142 152 10.1172/JCI43134 22201680 22. Dietrich M.O. Horvath T.L. Fat incites tanycytes to neurogenesis Nat. Neurosci. 2012 15 651 653 10.1038/nn.3091 22534576 23. Braun N. Sevigny J. Mishra S.K. Robson S.C. Barth S.W. Gerstberger R. Hammer K. Zimmermann H. Expression of the ecto-ATPase NTPDase2 in the germinal zones of the developing and adult rat brain Eur. J. Neurosci. 2003 17 1355 1364 10.1046/j.1460-9568.2003.02567.x 12713638 24. Frayling C. Britton R. Dale N. ATP-mediated glucosensing by hypothalamic tanycytes J. Physiol. 2011 589 2275 2286 10.1113/jphysiol.2010.202051 21486800 25. Saaltink D.-J. Havik B. Verissimo C.S. Lucassen P. Vreugdenhil E. Doublecortin and doublecortin-like are expressed in overlapping and non-overlapping neuronal cell population: Implications for neurogenesis J. Comp. Neurol. 2012 520 2805 2823 10.1002/cne.23144 22605579 26. Baroncini M. Allet C. Leroy D. Beauvillain J.-C. Francke J.-P. Prevot V. Morphological evidence for direct interaction between gonadotrophin-releasing hormone neurones and astroglial cells in the human hypothalamus J. Neuroendocrinol. 2007 19 691 702 10.1111/j.1365-2826.2007.01576.x 17680884 27. Barrett P. Ivanova E. Graham E.S. Ross A.W. Wilson D. Ple H. Mercer J.G. Ebling F.J. Schuhler S. DuPre S.M. Photoperiodic regulation of cellular retinoic acid-binding protein 1, GPR50 and nestin in tanycytes of the third ventricle ependymal layer of the Siberian hamster J. Endocrinol. 2006 191 687 698 10.1677/joe.1.06929 17170225 28. Wei L.-C. Shi M. Chen L.-W. Cao R. Zhang P. Chan Y. Nestin-containing cells express glial fibrillary acidic protein in the proliferative regions of central nervous system of postnatal developing and adult mice Dev. Brain Res. 2002 139 9 17 10.1016/S0165-3806(02)00509-6 12414089 29. Bolborea M. Laran-Chich M.-P. Rasri K. Hildebrandt H. Govitrapong P. Simonneaux V. Pevet P. Steinlechner S. Klosen P. Melatonin controls photoperiodic changes in tanycyte vimentin and neural cell adhesion molecule expression in the Djungarian hamster (Phodopus sungorus) Endocrinology 2011 152 3871 3883 10.1210/en.2011-1039 21846800 30. Chauvet N. Prieto M. Alonso G. Tanycytes present in the adult rat mediobasal hypothalamus support the regeneration of monoaminergic axons Exp. Neurol. 1998 151 1 13 10.1006/exnr.1998.6784 9582250 31. Sidibe A. Mullier A. Chen P. Baroncini M. Boutin J.A. Delagrange P. Prevot V. Jockers R. Expression of the orphan GPR50 protein in rodent and human dorsomedial hypothalamus, tanycytes and median eminence J. Pineal Res. 2010 48 263 269 10.1111/j.1600-079X.2010.00750.x 20210849 32. Alvarez-Buylla A. Cebrian-Silla A. Sorrells S.F. Nascimento M.A. Paredes M.F. Garcia-Verdugo J.M. Yang Z. Huang E.J. Comment on "Impact of neurodegenerative diseases on human adult hippocampal neurogenesis" Science 2022 376 eabn8861 10.1126/science.abn8861 35420933 33. Terreros-Roncal J. Moreno-Jimenez E.P. Flor-Garcia M. Rodriguez-Moreno C.B. Trinchero M.F. Cafini F. Rabano A. Llorens-Martin M. Rose M.C. Styr B. Impact of neurodegenerative diseases on human adult hippocampal neurogenesis Science 2021 374 1106 1113 10.1126/science.abl5163 34672693 34. Sorrells S.F. Paredes M.F. Velmeshev D. Herranz-Perez V. Sandoval K. Mayer S. Chang E.F. Insausti R. Kriegstein A.R. Rubenstein J.L. Immature excitatory neurons develop during adolescence in the human amygdala Nat. Commun. 2019 10 2748 10.1038/s41467-019-10765-1 31227709 35. Dietrich M.O. Horvath T.L. Synaptic plasticity of feeding circuits: Hormones and hysteresis Cell 2011 146 863 865 10.1016/j.cell.2011.08.031 21925311 36. Dietrich M.O. Horvath T.L. Hypothalamic control of energy balance: Insights into the role of synaptic plasticity Trends Neurosci. 2013 36 65 73 10.1016/j.tins.2012.12.005 23318157 37. Flier J.S. Regulating energy balance: The substrate strikes back Science 2006 312 861 864 10.1126/science.1127971 16690851 38. Niswender K.D. Baskin D. Schwartz M. Insulin and its evolving partnership with leptin in the hypothalamic control of energy homeostasis Trends Endocrinol. Metab. 2004 15 362 369 10.1016/j.tem.2004.07.009 15380807 39. Coll A.P. Farooqi I.S. O'Rahilly S. The hormonal control of food intake Cell 2007 129 251 262 10.1016/j.cell.2007.04.001 17448988 40. Munzberg H. Myers M.G. Molecular and anatomical determinants of central leptin resistance Nat. Neurosci. 2005 8 566 570 10.1038/nn1454 15856064 41. Olofsson L.E. Unger E.K. Cheung C.C. Xu A.W. Modulation of AgRP-neuronal function by SOCS3 as an initiating event in diet-induced hypothalamic leptin resistance Proc. Natl. Acad. Sci. USA 2013 110 E697 E706 10.1073/pnas.1218284110 23386726 42. Thaler J.P. Yi C.-X. Schur E.A. Guyenet S.J. Hwang B.H. Dietrich M.O. Zhao X. Sarruf D.A. Izgur V. Maravilla K.R. Obesity is associated with hypothalamic injury in rodents and humans J. Clin. Investig. 2012 122 153 162 10.1172/JCI59660 22201683 43. Padilla S.L. Carmody J.S. Zeltser L.M. Pomc-expressing progenitors give rise to antagonistic neuronal populations in hypothalamic feeding circuits Nat. Med. 2010 16 403 405 10.1038/nm.2126 20348924 44. Kokoeva M.V. Yin H. Flier J.S. Evidence for constitutive neural cell proliferation in the adult murine hypothalamus J. Comp. Neurol. 2007 505 209 220 10.1002/cne.21492 17853440 45. Pierce A.A. Xu A.W. De novo neurogenesis in adult hypothalamus as a compensatory mechanism to regulate energy balance J. Neurosci. 2010 30 723 730 10.1523/JNEUROSCI.2479-09.2010 20071537 46. Fonseca R.S. Mahesula S. Apple D.M. Raghunathan R. Dugan A. Cardona A. O'Connor J. Kokovay E. Neurogenic niche microglia undergo positional remodeling and progressive activation contributing to age-associated reductions in neurogenesis Stem Cells Dev. 2016 25 542 555 10.1089/scd.2015.0319 26857912 47. Kokaia Z. Martino G. Schwartz M. Lindvall O. Cross-talk between neural stem cells and immune cells: The key to better brain repair? Nat. Neurosci. 2012 15 1078 1087 10.1038/nn.3163 22837038 48. Breton J. Mao-Draayer Y. Impact of cytokines on neural stem/progenitor cell fate J. Neurol. Neurophysiol. 2011 S4 1 12 10.4172/2155-9562.S4-001 49. Kandasamy M. Lehner B. Kraus S. Sander P.R. Marschallinger J. Rivera F.J. Trumbach D. Ueberham U. Reitsamer H.A. Strauss O. TGF-beta signalling in the adult neurogenic niche promotes stem cell quiescence as well as generation of new neurons J. Cell. Mol. Med. 2014 18 1444 1459 10.1111/jcmm.12298 24779367 50. Wachs F.-P. Winner B. Couillard-Despres S. Schiller T. Aigner R. Winkler J. Bogdahn U. Aigner L. Transforming growth factor-b1 is a negative modulator of adult neurogenesis J. Neuropathol. Exp. Neurol. 2006 65 358 370 16691117 51. Katsimpardi L. Litterman N.K. Schein P.A. Miller C.M. Loffredo F.S. Wojtkiewicz G.R. Chen J.W. Lee R.T. Wagers A.J. Rubin L.L. Vascular and neurogenic rejuvenation of the aging mouse brain by young systemic factors Science 2014 344 630 634 10.1126/science.1251141 24797482 52. Bouchard J. Villeda S.A. Aging and brain rejuvenation as systemic events J. Neurochem. 2015 132 5 19 10.1111/jnc.12969 25327899 53. Zhang X. Zhang G. Zhang H. Karin M. Bai H. Cai D. Hypothalamic IKKb/NF-kB and ER stress link overnutrition to energy imbalance and obesity Cell 2008 135 61 73 10.1016/j.cell.2008.07.043 18854155 54. Kievit P. Howard J.K. Badman M.K. Balthasar N. Coppari R. Mori H. Lee C.E. Elmquist J.K. Yoshimura A. Flier J.S. Enhanced leptin sensitivity and improved glucose homeostasis in mice lacking suppressor of cytokine signaling-3 in POMC-expressing cells Cell Metab. 2006 4 123 132 10.1016/j.cmet.2006.06.010 16890540 55. Mori H. Hanada R. Hanada T. Aki D. Mashima R. Nishinakamura H. Torisu T. Chien K.R. Yasukawa H. Yoshimura A. Socs3 deficiency in the brain elevates leptin sensitivity and confers resistance to diet-induced obesity Nat. Med. 2004 10 739 743 10.1038/nm1071 15208705 56. Reed A.S. Unger E.K. Olofsson L.E. Piper M.L. Jr. Myers M.G. Xu A.W. Functional role of suppressor of cytokine signaling 3 upregulation in hypothalamic leptin resistance and long-term energy homeostasis Diabetes 2010 59 894 906 10.2337/db09-1024 20068134 57. Zabolotny J.M. Kim Y.-B. Welsh L.A. Kershaw E.E. Neel B.G. Kahn B.B. Protein-tyrosine phosphatase 1B expression is induced by inflammation in vivo J. Biol. Chem. 2008 283 14230 14241 10.1074/jbc.M800061200 18281274 58. Banno R. Zimmer D. De Jonghe B.C. Atienza M. Rak K. Yang W. Bence K.K. PTP1B and SHP2 in POMC neurons reciprocally regulate energy balance in mice J. Clin. Investig. 2010 120 720 734 10.1172/JCI39620 20160350 59. Bence K. Delibegovic M. Xue B. Gorgun C.Z. Hotamisligil G.S. Neel B.G. Kahn B.B. Neuronal PTP1B regulates body weight, adiposity and leptin action Nat. Med. 2006 12 917 924 10.1038/nm1435 16845389 60. Picardi P.K. Calegari V.C. Prada P.D.O. Moraes J.C. Araujo E. Marcondes M.C.C.G. Ueno M. Carvalheira J.B.C. Velloso L.A. Saad M.J.A. Reduction of hypothalamic protein tyrosine phosphatase improves insulin and leptin resistance in diet-induced obese rats Endocrinology 2008 149 3870 3880 10.1210/en.2007-1506 18467448 61. Liao G. Cheung S. Galeano J. Ji A.X. Qin Q. Bi X. Allopregnanolone treatment delays cholesterol accumulation and reduces autophagic/lysosomal dysfunction and inflammation in Npc1-/- mouse brain Brain Res. 2009 1270 140 151 10.1016/j.brainres.2009.03.027 19328188 62. Dugan L.L. Ali S.S. Shekhtman G. Roberts A.J. Lucero J. Quick K.L. Behrens M.M. IL-6 mediated degeneration of forebrain GABAergic interneurons and cognitive impairment in aged mice through activation of neuronal NADPH oxidase PLoS ONE 2009 4 e5518 10.1371/journal.pone.0005518 19436757 63. Niture S.K. Jaiswal A.K. Nrf2 protein up-regulates antiapoptotic protein Bcl-2 and prevents cellular apoptosis J. Biol. Chem. 2012 287 9873 9886 10.1074/jbc.M111.312694 22275372 64. Paek J. Lo J.Y. Narasimhan S.D. Nguyen T.N. Glover-Cutter K. Robida-Stubbs S. Suzuki T. Yamamoto M. Blackwell T.K. Curran S.P. Mitochondrial SKN-1/Nrf mediates a conserved starvation response Cell Metab. 2012 16 526 537 10.1016/j.cmet.2012.09.007 23040073 65. Murakami S. Motohashi H. Roles of Nrf2 in cell proliferation and differentiation Free Radic. Biol. Med. 2015 88 168 178 10.1016/j.freeradbiomed.2015.06.030 26119783 66. Sykiotis G.P. Bohmann D. Keap1/Nrf2 signaling regulates oxidative stress tolerance and lifespan in Drosophila Dev. Cell 2008 14 76 85 10.1016/j.devcel.2007.12.002 18194654 67. Balasubramanian P. Asirvatham-Jeyaraj N. Monteiro R. Sivasubramanian M.K. Hall D. Subramanian M. Obesity-induced sympathoexcitation is associated with Nrf2 dysfunction in the rostral ventrolateral medulla Am. J. Physiol.-Regul. Integr. Comp. Physiol. 2020 318 R435 R444 10.1152/ajpregu.00206.2019 31823672 68. Le Belle J.E. Orozco N.M. Paucar A.A. Saxe J.P. Mottahedeh J. Pyle A.D. Wu H. Kornblum H.I. Proliferative neural stem cells have high endogenous ROS levels that regulate self-renewal and neurogenesis in a PI3K/Akt-dependant manner Cell Stem Cell 2011 8 59 71 10.1016/j.stem.2010.11.028 21211782 69. Dodson M. Anandhan A. Zhang D.D. Madhavan L. An NRF2 perspective on stem cells and ageing Front. Aging 2021 2 690686 10.3389/fragi.2021.690686 36213179 70. Zhou G. Meng S. Li Y. Ghebre Y.T. Cooke J.P. Optimal ROS signaling is critical for nuclear reprogramming Cell Rep. 2016 15 919 925 10.1016/j.celrep.2016.03.084 27117405 71. Ray S. Corenblum M.J. Anandhan A. Reed A. Ortiz F.O. Zhang D.D. Barnes C.A. Madhavan L. A role for Nrf2 expression in defining the aging of hippocampal neural stem cells Cell Transplant. 2018 27 589 606 10.1177/0963689718774030 29871525 72. Lopez-Otin C. Blasco M.A. Partridge L. Serrano M. Kroemer G. The hallmarks of aging Cell 2013 153 1194 1217 10.1016/j.cell.2013.05.039 23746838 73. Nunan E. Wright C.L. Semola O.A. Subramanian M. Balasubramanian P. Lovern P.C. Fancher I.S. Butcher J.T. Obesity as a premature aging phenotype--Implications for sarcopenic obesity GeroScience 2022 44 1393 1405 10.1007/s11357-022-00567-7 35471692 74. Dacks P.A. Moreno C.L. Kim E.S. Marcellino B.K. Mobbs C.V. Role of the hypothalamus in mediating protective effects of dietary restriction during aging Front. Neuroendocrinol. 2013 34 95 106 10.1016/j.yfrne.2012.12.001 23262258 75. Sadagurski M. Landeryou T. Cady G. Bartke A. Bernal-Mizrachi E. Miller R.A. Transient early food restriction leads to hypothalamic changes in the long-lived crowded litter female mice Physiol. Rep. 2015 3 e12379 10.14814/phy2.12379 25907790 76. Satoh A. Brace C.S. Rensing N. Cliften P. Wozniak D.F. Herzog E.D. Yamada K.A. Imai S.-I. Sirt1 extends life span and delays aging in mice through the regulation of Nk2 homeobox 1 in the DMH and LH Cell Metab. 2013 18 416 430 10.1016/j.cmet.2013.07.013 24011076 77. Zhang G. Li J. Purkayastha S. Tang Y. Zhang H. Yin Y. Li B. Liu G. Cai D. Hypothalamic programming of systemic ageing involving IKK-b, NF-kB and GnRH Nature 2013 497 211 216 10.1038/nature12143 23636330 78. Cai D. Neuroinflammation and neurodegeneration in overnutrition-induced diseases Trends Endocrinol. Metab. 2013 24 40 47 10.1016/j.tem.2012.11.003 23265946 79. Posey K.A. Clegg D.J. Printz R.L. Byun J. Morton G.J. Vivekanandan-Giri A. Pennathur S. Baskin D.G. Heinecke J.W. Woods S.C. Hypothalamic proinflammatory lipid accumulation, inflammation, and insulin resistance in rats fed a high-fat diet Am. J. Physiol. -Endocrinol. Metab. 2009 296 E1003 E1012 10.1152/ajpendo.90377.2008 19116375 80. Stolp H.B. Turnquist C. Dziegielewska K.M. Saunders N. Anthony D. Molnar Z. Reduced ventricular proliferation in the foetal cortex following maternal inflammation in the mouse Brain 2011 134 3236 3248 10.1093/brain/awr237 21964917 81. Zhang Y. Kim M.S. Jia B. Yan J. Zuniga-Hertz J.P. Han C. Cai D. Hypothalamic stem cells control ageing speed partly through exosomal miRNAs Nature 2017 548 52 57 10.1038/nature23282 28746310 82. Molofsky A.V. Pardal R. Iwashita T. Park I.-K. Clarke M.F. Morrison S.J. Bmi-1 dependence distinguishes neural stem cell self-renewal from progenitor proliferation Nature 2003 425 962 967 10.1038/nature02060 14574365 83. Stoll E.A. Cheung W. Mikheev A.M. Sweet I.R. Bielas J.H. Zhang J. Rostomily R.C. Horner P.J. Aging neural progenitor cells have decreased mitochondrial content and lower oxidative metabolism J. Biol. Chem. 2011 286 38592 38601 10.1074/jbc.M111.252171 21900249 84. Finkel T. Deng C.X. Mostoslavsky R. Recent progress in the biology and physiology of sirtuins Nature 2009 460 587 591 10.1038/nature08197 19641587 85. Hagenbuchner J. Ausserlechner M.J. Mitochondria and FOXO3: Breath or die Front. Physiol. 2013 4 147 10.3389/fphys.2013.00147 23801966 86. Sakata H. Narasimhan P. Niizuma K. Maier C.M. Wakai T. Chan P.H. Interleukin 6-preconditioned neural stem cells reduce ischaemic injury in stroke mice Brain 2012 135 3298 3310 10.1093/brain/aws259 23169920 87. Paik J.-H. Ding Z. Narurkar R. Ramkissoon S. Muller F. Kamoun W.S. Chae S.-S. Zheng H. Ying H. Mahoney J. FoxOs cooperatively regulate diverse pathways governing neural stem cell homeostasis Cell Stem Cell 2009 5 540 553 10.1016/j.stem.2009.09.013 19896444 88. Gemma C. Vila J. Bachstetter A. Bickford P.C. Oxidative stress and the aging brain: From theory to prevention Brain Aging 2007 353 374 89. Lionaki E. Markaki M. Palikaras K. Tavernarakis N. Mitochondria, autophagy and age-associated neurodegenerative diseases: New insights into a complex interplay Biochim. Biophys. Acta (BBA)-Bioenerg. 2015 1847 1412 1423 10.1016/j.bbabio.2015.04.010 90. Moore D.L. Pilz G.A. Arauzo-Bravo M.J. Barral Y. Jessberger S. A mechanism for the segregation of age in mammalian neural stem cells Science 2015 349 1334 1338 10.1126/science.aac9868 26383951 91. Mazumdar J. O'Brien W.T. Johnson R. LaManna J. Chavez J.C. Klein P.S. Simon M.C. O2 regulates stem cells through Wnt/b-catenin signalling Nat. Cell Biol. 2010 12 1007 1013 10.1038/ncb2102 20852629 92. Rabie T. Kunze R. Marti H.H. Impaired hypoxic response in senescent mouse brain Int. J. Dev. Neurosci. 2011 29 655 661 10.1016/j.ijdevneu.2011.06.003 21704147 93. Aberg M.A. Aberg N.D. Hedbacker H. Oscarsson J. Eriksson P.S. Peripheral infusion of IGF-I selectively induces neurogenesis in the adult rat hippocampus J. Neurosci. 2000 20 2896 2903 10.1523/JNEUROSCI.20-08-02896.2000 10751442 94. Sonntag W.E. Ramsey M. Carter C.S. Growth hormone and insulin-like growth factor-1 (IGF-1) and their influence on cognitive aging Ageing Res. Rev. 2005 4 195 212 10.1016/j.arr.2005.02.001 16024298 95. Chaker Z. Aid S. Berry H. Holzenberger M. Suppression of IGF-I signals in neural stem cells enhances neurogenesis and olfactory function during aging Aging Cell 2015 14 847 856 10.1111/acel.12365 26219530 96. Cota D. Proulx K. Blake Smith K.A. Kozma S.C. Thomas G. Woods S.C. Seeley R.J. Hypothalamic mTOR signaling regulates food intake Science 2006 312 927 930 10.1126/science.1124147 16690869 97. de Morentin P.M. Martinez-Sanchez N. Roa J. Ferno J. Nogueiras R. Tena-Sempere M. Dieguez C. Lopez M. Hypothalamic mTOR: The rookie energy sensor Curr. Mol. Med. 2014 14 3 21 10.2174/1566524013666131118103706 24236459 98. Yang S.-B. Tien A.-C. Boddupalli G. Xu A.W. Jan Y.N. Jan L.Y. Rapamycin ameliorates age-dependent obesity associated with increased mTOR signaling in hypothalamic POMC neurons Neuron 2012 75 425 436 10.1016/j.neuron.2012.03.043 22884327 99. Mori H. Inoki K. Munzberg H. Opland D. Faouzi M. Villanueva E.C. Ikenoue T. Kwiatkowski D. MacDougald O.A. Myers M.G. Critical role for hypothalamic mTOR activity in energy balance Cell Metab. 2009 9 362 374 10.1016/j.cmet.2009.03.005 19356717 100. Song H. Stevens C.F. Gage F.H. Astroglia induce neurogenesis from adult neural stem cells Nature 2002 417 39 44 10.1038/417039a 11986659 101. Okamoto M. Inoue K. Iwamura H. Terashima K. Soya H. Asashima M. Kuwabara T. Reduction in paracrine Wnt3 factors during aging causes impaired adult neurogenesis FASEB J. 2011 25 3570 3582 10.1096/fj.11-184697 21746862 102. Miranda C.J. Braun L. Jiang Y. Hester M.E. Zhang L. Riolo M. Wang H. Rao M. Altura R.A. Kaspar B.K. Aging brain microenvironment decreases hippocampal neurogenesis through Wnt-mediated survivin signaling Aging Cell 2012 11 542 552 10.1111/j.1474-9726.2012.00816.x 22404871 103. Seib D.R. Corsini N.S. Ellwanger K. Plaas C. Mateos A. Pitzer C. Niehrs C. Celikel T. Martin-Villalba A. Loss of Dickkopf-1 restores neurogenesis in old age and counteracts cognitive decline Cell Stem Cell 2013 12 204 214 10.1016/j.stem.2012.11.010 23395445 104. Gattazzo F. Urciuolo A. Bonaldo P. Extracellular matrix: A dynamic microenvironment for stem cell niche Biochim. Biophys. Acta (BBA)-Gen. Subj. 2014 1840 2506 2519 10.1016/j.bbagen.2014.01.010 24418517 105. Keung A.J. de Juan-Pardo E.M. Schaffer D.V. Kumar S. Rho GTPases mediate the mechanosensitive lineage commitment of neural stem cells Stem Cells 2011 29 1886 1897 10.1002/stem.746 21956892 106. Li L. Zang L. Zhang F. Chen J. Shen H. Shu L. Liang F. Feng C. Chen D. Tao H. Fat mass and obesity-associated (FTO) protein regulates adult neurogenesis Hum. Mol. Genet. 2017 26 2398 2411 10.1093/hmg/ddx128 28398475 107. Haussmann I.U. Bodi Z. Sanchez-Moran E. Mongan N.P. Archer N. Fray R.G. Soller M. m6A potentiates Sxl alternative pre-mRNA splicing for robust Drosophila sex determination Nature 2016 540 301 304 10.1038/nature20577 27919081 108. Nainar S. Marshall P.R. Tyler C.R. Spitale R.C. Bredy T.W. Evolving insights into RNA modifications and their functional diversity in the brain Nat. Neurosci. 2016 19 1292 1298 10.1038/nn.4378 27669990 109. Yue Y. Liu J. He C. RNA N6-methyladenosine methylation in post-transcriptional gene expression regulation Genes Dev. 2015 29 1343 1355 10.1101/gad.262766.115 26159994 110. Bouret S.G. Simerly R.B. Minireview: Leptin and development of hypothalamic feeding circuits Endocrinology 2004 145 2621 2626 10.1210/en.2004-0231 15044371 111. Lang B.T. Yan Y. Dempsey R.J. Vemuganti R. Impaired neurogenesis in adult type-2 diabetic rats Brain Res. 2009 1258 25 33 10.1016/j.brainres.2008.12.026 19138677 112. Beauquis J. Saravia F. Coulaud J. Roig P. Dardenne M. Homo-Delarche F. De Nicola A. Prominently decreased hippocampal neurogenesis in a spontaneous model of type 1 diabetes, the nonobese diabetic mouse Exp. Neurol. 2008 210 359 367 10.1016/j.expneurol.2007.11.009 18190910 113. Cui H. Cai F. Belsham D.D. Leptin signaling in neurotensin neurons involves STAT, MAP kinases ERK1/2, and p38 through c-Fos and ATF1 FASEB J. 2006 20 2654 2656 10.1096/fj.06-5989fje 17077290 114. Desai M. Li T. Ross M.G. Fetal hypothalamic neuroprogenitor cell culture: Preferential differentiation paths induced by leptin and insulin Endocrinology 2011 152 3192 3201 10.1210/en.2010-1217 21652728 115. Dunbar J.C. Lu H. Leptin-induced increase in sympathetic nervous and cardiovascular tone is mediated by proopiomelanocortin (POMC) products Brain Res. Bull. 1999 50 215 221 10.1016/S0361-9230(99)00197-5 10566984 116. Haynes W.G. Morgan N.A. Djalali A. Sivitz W.I. Mark A.L. Interactions between the melanocortin system and leptin in control of sympathetic nerve traffic Hypertension 1999 33 542 547 10.1161/01.HYP.33.1.542 9931162 117. Cone R.D. Anatomy and regulation of the central melanocortin system Nat. Neurosci. 2005 8 571 578 10.1038/nn1455 15856065 118. King C. Sarvetnick N. The incidence of type-1 diabetes in NOD mice is modulated by restricted flora not germ-free conditions PLoS ONE 2011 6 e17049 10.1371/journal.pone.0017049 21364875 119. Myers M.G. Munzberg H. Leinninger G.M. Leshan R.L. The geometry of leptin action in the brain: More complicated than a simple ARC Cell Metab. 2009 9 117 123 10.1016/j.cmet.2008.12.001 19187770 120. Ulrich-Lai Y.M. Jones K.R. Ziegler D.R. Cullinan W.E. Herman J. Forebrain origins of glutamatergic innervation to the rat paraventricular nucleus of the hypothalamus: Differential inputs to the anterior versus posterior subregions J. Comp. Neurol. 2011 519 1301 1319 10.1002/cne.22571 21452198 121. Mercer A.J. Hentges S.T. Meshul C.K. Low M.J. Unraveling the central proopiomelanocortin neural circuits Front. Neurosci. 2013 7 19 10.3389/fnins.2013.00019 23440036 122. Kawabe T. Kawabe K. Sapru H.N. Cardiovascular responses to chemical stimulation of the hypothalamic arcuate nucleus in the rat: Role of the hypothalamic paraventricular nucleus PLoS ONE 2012 7 e45180 10.1371/journal.pone.0045180 23028831 123. Kawabe T. Kawabe K. Sapru H.N. Effect of barodenervation on cardiovascular responses elicited from the hypothalamic arcuate nucleus of the rat PLoS ONE 2012 7 e53111 10.1371/journal.pone.0053111 23300873 124. Cassaglia P.A. Shi Z. Li B. Reis W.L. Clute-Reinig N.M. Stern J.E. Brooks V.L. Neuropeptide Y acts in the paraventricular nucleus to suppress sympathetic nerve activity and its baroreflex regulation J. Physiol. 2014 592 1655 1675 10.1113/jphysiol.2013.268763 24535439 125. Jobst E.E. Enriori P.J. Cowley M.A. The electrophysiology of feeding circuits Trends Endocrinol. Metab. 2004 15 488 499 10.1016/j.tem.2004.10.007 15541648 126. Varela L. Horvath T.L. Leptin and insulin pathways in POMC and AgRP neurons that modulate energy balance and glucose homeostasis EMBO Rep. 2012 13 1079 1086 10.1038/embor.2012.174 23146889 127. Hall J.E. da Silva A.A. do Carmo J.M. Dubinion J. Hamza S. Munusamy S. Smith G. Stec D.E. Obesity-induced hypertension: Role of sympathetic nervous system, leptin, and melanocortins J. Biol. Chem. 2010 285 17271 17276 10.1074/jbc.R110.113175 20348094 128. Kohsaka A. Laposky A.D. Ramsey K.M. Estrada C. Joshu C. Kobayashi Y. Turek F.W. Bass J. High-fat diet disrupts behavioral and molecular circadian rhythms in mice Cell Metab. 2007 6 414 421 10.1016/j.cmet.2007.09.006 17983587 129. La Fleur S.E. Van Rozen A.J. Luijendijk M.C.M. Groeneweg F. Adan R.A.H. A free-choice high-fat high-sugar diet induces changes in arcuate neuropeptide expression that support hyperphagia Int. J. Obes. 2010 34 537 546 10.1038/ijo.2009.257 130. Lin S. Storlien L.H. Huang X.F. Leptin receptor, NPY, POMC mRNA expression in the diet-induced obese mouse brain Brain Res. 2000 875 89 95 10.1016/S0006-8993(00)02580-4 10967302 131. Bouret S.G. Gorski J.N. Patterson C.M. Chen S. Levin B.E. Simerly R.B. Hypothalamic neural projections are permanently disrupted in diet-induced obese rats Cell Metab. 2008 7 179 185 10.1016/j.cmet.2007.12.001 18249177 132. Shi Z. Li B. Brooks V.L. Role of the paraventricular nucleus of the hypothalamus in the sympathoexcitatory effects of leptin Hypertension 2015 66 1034 1041 10.1161/HYPERTENSIONAHA.115.06017 26370892 133. Plakkot B. Subramanian M. Evaluation of Hypothalamic Neural Stem Cell Niche-Implications on Obesity-Induced Sympathoexcitation FASEB J. 2022 36 10.1096/fasebj.2022.36.S1.R6154 134. Mattson M.P. Arumugam T.V. Hallmarks of brain aging: Adaptive and pathological modification by metabolic states Cell Metab. 2018 27 1176 1199 10.1016/j.cmet.2018.05.011 29874566 135. Mana M.D. Kuo E.Y.S. Yilmaz O.H. Dietary regulation of adult stem cells Curr. Stem Cell Rep. 2017 3 1 8 10.1007/s40778-017-0072-x 28966904 136. Arumugam T. Phillips T.M. Cheng A. Morrell C.H. Mattson M.P. Wan R. Age and energy intake interact to modify cell stress pathways and stroke outcome Ann. Neurol. 2010 67 41 52 10.1002/ana.21798 20186857 137. Halagappa V.K.M. Guo Z. Pearson M. Matsuoka Y. Cutler R.G. LaFerla F.M. Mattson M.P. Intermittent fasting and caloric restriction ameliorate age-related behavioral deficits in the triple-transgenic mouse model of Alzheimer's disease Neurobiol. Dis. 2007 26 212 220 10.1016/j.nbd.2006.12.019 17306982 138. Manzanero S. Erion J.R. Santro T. Steyn F.J. Chen C. Arumugam T.V. Stranahan A.M. Intermittent fasting attenuates increases in neurogenesis after ischemia and reperfusion and improves recovery J. Cereb. Blood Flow Metab. 2014 34 897 905 10.1038/jcbfm.2014.36 24549184 139. Bruce-Keller A.J. Umberger G. McFall R. Mattson M.P. Food restriction reduces brain damage and improves behavioral outcome following excitotoxic and metabolic insults Ann. Neurol. Off. J. Am. Neurol. Assoc. Child Neurol. Soc. 1999 45 8 15 10.1002/1531-8249(199901)45:1<8::AID-ART4>3.0.CO;2-V 140. Kim J. Kang S.-W. Mallilankaraman K. Baik S.-H. Lim J.C. Balaganapathy P. She D.T. Lok K.-Z. Fann D.Y. Thambiayah U. Transcriptome analysis reveals intermittent fasting-induced genetic changes in ischemic stroke Hum. Mol. Genet. 2018 27 1497 1513 10.1093/hmg/ddy057 29447348 141. Baik S. Rajeev V. Fann D.Y. Jo D. Arumugam T.V. Intermittent fasting increases adult hippocampal neurogenesis Brain Behav. 2020 10 e01444 10.1002/brb3.1444 31804775 142. Hitoshi S. Alexson T. Tropepe V. Donoviel D. Elia A.J. Nye J.S. Conlon R.A. Mak T.W. Bernstein A. van der Kooy D. Notch pathway molecules are essential for the maintenance, but not the generation, of mammalian neural stem cells Genes Dev. 2002 16 846 858 10.1101/gad.975202 11937492 143. Brai E. Marathe S. Astori S. Ben Fredj N. Perry E. Lamy C. Scotti A. Alberi L. Notch1 regulates hippocampal plasticity through interaction with the reelin pathway, glutamatergic transmission and CREB signaling Front. Cell. Neurosci. 2015 9 447 10.3389/fncel.2015.00447 26635527 144. He W. Tian X. Yuan B. Chu B. Gao F. Wang H. Rosuvastatin improves neurite extension in cortical neurons through the Notch 1/BDNF pathway Neurol. Res. 2019 41 658 664 10.1080/01616412.2019.1610226 31023175 145. Scholzke M.N. Schwaninger M. Transcriptional regulation of neurogenesis: Potential mechanisms in cerebral ischemia J. Mol. Med. 2007 85 577 588 10.1007/s00109-007-0196-z 17429598 146. Colman R.J. Beasley T.M. Kemnitz J.W. Johnson S.C. Weindruch R. Anderson R.M. Caloric restriction reduces age-related and all-cause mortality in rhesus monkeys Nat. Commun. 2014 5 3557 10.1038/ncomms4557 24691430 147. Fontana L. Partridge L. Longo V.D. Extending healthy life span--From yeast to humans Science 2010 328 321 326 10.1126/science.1172539 20395504 148. Longo V.D. Fontana L. Calorie restriction and cancer prevention: Metabolic and molecular mechanisms Trends Pharmacol. Sci. 2010 31 89 98 10.1016/j.tips.2009.11.004 20097433 149. Mercken E.M. Carboneau B.A. Krzysik-Walker S.M. de Cabo R. Of mice and men: The benefits of caloric restriction, exercise, and mimetics Ageing Res. Rev. 2012 11 390 398 10.1016/j.arr.2011.11.005 22210414 150. Berryman D. Christiansen J.S. Johannsson G. Thorner M.O. Kopchick J.J. Role of the GH/IGF-1 axis in lifespan and healthspan: Lessons from animal models Growth Horm. IGF Res. 2008 18 455 471 10.1016/j.ghir.2008.05.005 18710818 151. Longo V.D. Mattson M.P. Fasting: Molecular mechanisms and clinical applications Cell Metab. 2014 19 181 192 10.1016/j.cmet.2013.12.008 24440038 152. Lee J. Duan W. Long J.M. Ingram D.K. Mattson M.P. Dietary restriction increases the number of newly generated neural cells, and induces BDNF expression, in the dentate gyrus of rats J. Mol. Neurosci. 2000 15 99 108 10.1385/JMN:15:2:99 11220789 153. Lee J. Seroogy K.B. Mattson M.P. Dietary restriction enhances neurotrophin expression and neurogenesis in the hippocampus of adult mice J. Neurochem. 2002 80 539 547 10.1046/j.0022-3042.2001.00747.x 11905999 154. Brandhorst S. Choi I.Y. Wei M. Cheng C.W. Sedrakyan S. Navarrete G. Dubeau L. Yap L.P. Park R. Vinciguerra M. A periodic diet that mimics fasting promotes multi-system regeneration, enhanced cognitive performance, and healthspan Cell Metab. 2015 22 86 99 10.1016/j.cmet.2015.05.012 26094889 155. Cheng C.W. Adams G.B. Perin L. Wei M. Zhou X. Lam B.S. Da Sacco S. Mirisola M. Quinn D.I. Dorff T.B. Prolonged fasting reduces IGF-1/PKA to promote hematopoietic-stem-cell-based regeneration and reverse immunosuppression Cell Stem Cell 2014 14 810 823 10.1016/j.stem.2014.04.014 24905167 156. Ng S.Y. Johnson R. Stanton L.W. Human long non-coding RNAs promote pluripotency and neuronal differentiation by association with chromatin modifiers and transcription factors EMBO J. 2012 31 522 533 10.1038/emboj.2011.459 22193719 157. Wang Y. He L. Du Y. Zhu P. Huang G. Luo J. Yan X. Ye B. Li C. Xia P. The long noncoding RNA lncTCF7 promotes self-renewal of human liver cancer stem cells through activation of Wnt signaling Cell Stem Cell 2015 16 413 425 10.1016/j.stem.2015.03.003 25842979 158. Loewer S. Cabili M.N. Guttman M. Loh Y.H. Thomas K. Park I.H. Garber M. Curran M. Onder T. Agarwal S. Large intergenic non-coding RNA-RoR modulates reprogramming of human induced pluripotent stem cells Nat. Genet. 2010 42 1113 1117 10.1038/ng.710 21057500 159. Xiao Y.Z. Yang M. Xiao Y. Guo Q. Huang Y. Li C.J. Cai D. Luo X.H. Reducing hypothalamic stem cell senescence protects against aging-associated physiological decline Cell Metab. 2020 31 534 548.e5 10.1016/j.cmet.2020.01.002 32004475 160. Bernstein H.-G. Lindquist J.A. Keilhoff G. Dobrowolny H. Brandt S. Steiner J. Bogerts B. Mertens P.R. Differential distribution of Y-box-binding protein 1 and cold shock domain protein A in developing and adult human brain Brain Struct. Funct. 2015 220 2235 2245 10.1007/s00429-014-0786-9 24817634 161. En-Nia A. Yilmaz E. Klinge U. Lovett D.H. Stefanidis I. Mertens P.R. Transcription factor YB-1 mediates DNA polymerase a gene expression J. Biol. Chem. 2005 280 7702 7711 10.1074/jbc.M413353200 15615704 162. Izumi H. Imamura T. Nagatani G. Ise T. Murakami T. Uramoto H. Torigoe T. Ishiguchi H. Yoshida Y. Nomoto M. Y box-binding protein-1 binds preferentially to single-stranded nucleic acids and exhibits 3'- 5' exonuclease activity Nucleic Acids Res. 2001 29 1200 1207 10.1093/nar/29.5.1200 11222770 163. Kotake Y. Ozawa Y. Harada M. Kitagawa K. Niida H. Morita Y. Tanaka K. Suda T. Kitagawa M. YB 1 binds to and represses the p16 tumor suppressor gene Genes Cells 2013 18 999 1006 10.1111/gtc.12093 24165022 164. Anandhan A. Tamilselvam K. Radhiga T. Rao S. Essa M.M. Manivasagam T. Theaflavin, a black tea polyphenol, protects nigral dopaminergic neurons against chronic MPTP/probenecid induced Parkinson's disease Brain Res. 2012 1433 104 113 10.1016/j.brainres.2011.11.021 22138428 165. Nishino J. Kim I. Chada K. Morrison S.J. Hmga2 promotes neural stem cell self-renewal in young but not old mice by reducing p16Ink4a and p19Arf Expression Cell 2008 135 227 239 10.1016/j.cell.2008.09.017 18957199 166. Lowe S.W. Sherr C.J. Tumor suppression by Ink4a-Arf: Progress and puzzles Curr. Opin. Genet. Dev. 2003 13 77 83 10.1016/S0959-437X(02)00013-8 12573439 167. Molofsky A.V. Slutsky S.G. Joseph N.M. He S. Pardal R. Krishnamurthy J. Sharpless N.E. Morrison S.J. Increasing p16INK4a expression decreases forebrain progenitors and neurogenesis during ageing Nature 2006 443 448 452 10.1038/nature05091 16957738 168. Zhao C. Sun G. Li S. Lang M.-F. Yang S. Li W. Shi Y. MicroRNA let-7b regulates neural stem cell proliferation and differentiation by targeting nuclear receptor TLX signaling Proc. Natl. Acad. Sci. USA 2010 107 1876 1881 10.1073/pnas.0908750107 20133835 169. Nishino J. Kim S. Zhu Y. Zhu H. Morrison S.J. A network of heterochronic genes including Imp1 regulates temporal changes in stem cell properties Elife 2013 2 e00924 10.7554/eLife.00924 24192035 170. Szafranski K. Abraham K.J. Mekhail K. Non-coding RNA in neural function, disease, and aging Front. Genet. 2015 6 87 10.3389/fgene.2015.00087 25806046 171. Spinelli M. Natale F. Rinaudo M. Leone L. Mezzogori D. Fusco S. Grassi C. Neural Stem Cell-Derived Exosomes Revert HFD-Dependent Memory Impairment via CREB-BDNF Signalling Int. J. Mol. Sci. 2020 21 8994 10.3390/ijms21238994 33256199 172. Reekmans K. Praet J. Daans J. Reumers V. Pauwels P. Van Der Linden A. Berneman Z.N. Ponsaerts P. Current challenges for the advancement of neural stem cell biology and transplantation research Stem Cell Rev. Rep. 2012 8 262 278 10.1007/s12015-011-9266-2 21537994 173. Bellin M. Marchetto M.C. Gage F.H. Mummery C.L. Induced pluripotent stem cells: The new patient? Nat. Rev. Mol. Cell Biol. 2012 13 713 726 10.1038/nrm3448 23034453 |
PMC10000585 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051002 foods-12-01002 Article Impact of Climate Changes on the Natural Prevalence of Fusarium Mycotoxins in Maize Harvested in Serbia and Croatia Janic Hajnal Elizabet Conceptualization Investigation Resources Data curation Writing - original draft Writing - review & editing Visualization Funding acquisition 1* Kos Jovana Validation Investigation Resources Data curation Writing - original draft Writing - review & editing Supervision Project administration Funding acquisition 1 Radic Bojana Conceptualization Methodology Validation Formal analysis Writing - original draft Writing - review & editing 1 Anic Mislav Resources Writing - review & editing 2 Radovic Radmila Methodology Validation Formal analysis 1 Kudumija Nina Methodology Validation Formal analysis 3 Vulic Ana Methodology Formal analysis 3 Dekic Sanja Validation Resources Funding acquisition 4 Pleadin Jelka Conceptualization Investigation Resources Data curation Writing - original draft Writing - review & editing Project administration Funding acquisition 3 Khaneghah Amin Mousavi Academic Editor Abrunhosa Luis Academic Editor 1 Institute of Food Technology, University of Novi Sad, Bulevar cara Lazara 1, 21000 Novi Sad, Serbia 2 Croatian Meteorological and Hydrological Service, Ravnice 48, 10000 Zagreb, Croatia 3 Croatian Veterinary Institute, Laboratory for Analytical Chemistry, Savska Cesta 143, 10000 Zagreb, Croatia 4 Faculty of Chemistry, Department of Analytical Chemistry, University of Belgrade, Sudentski trg 12-16, 11158 Belgrade, Serbia * Correspondence: [email protected] 27 2 2023 3 2023 12 5 100228 12 2022 18 2 2023 23 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Ongoing climate change may affect the susceptibility of plants to attacks by pathogenic, mostly mycotoxigenic fungi with a consequent increase in the presence of mycotoxins. Fusarium fungi represent one of the most important producers of mycotoxins, and are also important pathogens of agricultural crops. Therefore, the main aim of the study was to estimate the impact of weather parameters on the natural occurrence of Fusarium mycotoxins, such as deoxynivalenol (DON), fumonisins B1 and B2 (FUMs), zearalenone (ZEN), T-2, and HT-2 toxins (T-2/HT-2) in maize samples harvested from two neighboring countries, Serbia and Croatia, during a four-year production period (2018-2021). The frequency and contamination level of examined Fusarium mycotoxins varied by maize year of production and could be linked to weather conditions per investigated country. Among them, FUMs were found to be the most common contaminants (84-100%) of maize in both Serbia and Croatia. Additionally, a critical assessment of Fusarium mycotoxins occurrence in the last 10 years (2012-2021), for both Serbia and Croatia, was done. Results pointed out the highest contamination of maize from 2014, especially with DON and ZEN, in connection to extreme levels of precipitation observed in both Serbia and Croatia, whereas FUMs occurred with high prevalence from each of the ten investigated years. deoxynivalenol fumonisins zearalenone T-2 HT-2 maize Serbia Croatia weather conditions The Ministry of Education, Science and Technological Development of the Republic of Serbia451-03-68/2022-14/20022 Croatian Veterinary Institute in ZagrebThis paper is a result of the research funded by The Ministry of Education, Science and Technological Development of the Republic of Serbia (451-03-68/2022-14/20022), and Croatian Veterinary Institute in Zagreb. pmc1. Introduction One of the major global food and feed safety issues is the presence of mycotoxigenic fungi and their secondary metabolites, mycotoxins. They reduce the quality of agricultural crop commodities and impact negatively on human and animal health . The fungal genus Fusarium is one of the most important mycotoxigenic fungal genera in food and feed. Several species of the genus Fusarium can infect cereals, while the predominant species vary depending on the type of crop, geographical region, and environmental conditions . Fusarium mycotoxins are produced by toxigenic fungi that naturally contaminate cereals, representing a source of serious concern in cereals and cereal-based products, which leads to harmful contamination of food and feed . The main and most frequently occurring mycotoxins produced by Fusarium fungi in cereals are fumonisins, trichothecenes, and zearalenone (ZEN), followed by "emerging" mycotoxins (such as moniliformin, fusaproliferin, enniatins, and beauvericin) and many other metabolites (such as bikaverin, culmorin, aurofusarin, fusaric acid, fusapyron, and butenolid) . Fumonisins are mainly produced by Fusarium (F.) verticillioides and F. proliferatum and represent one of the most important mycotoxins that occur in maize (Zea mays), particularly when maize is grown in warmer regions . Recent studies on laboratory animals have clearly shown that prenatal exposure to fumonisins is highly detrimental to the offspring, which can be manifested by altered chemical coding of enteric neurons of the gut. Furthermore, the neurotoxic effect of fumonisins is also linked to the enteric nervous system of the fetus . Fumonisin analogues are classified into four main groups: fumonisin A, B (FB), C, and P. However, among all analogs of fumonisins, FB1, FB2, and FB3 generally occur in nature most frequently, while FB1 usually occurs with the highest concentrations . F. langsethiae, F. poae, F. equiseti, and F. sporotrichoides mainly produce trichothecenes type A (T-2 toxin, HT-2 toxins, neosolaniol, monoacetoxyscirpenol, and diacetoxyscirpenol), while F. graminearum and F. culmorum typically produce trichothecenes type B (mainly deoxynivalenol (DON), 3-acetyldeoxynivalenol, 15-acetyldeoxynivalenol, and nivalenol). DON is the most frequently detected mycotoxin among trichothecenes in cereals and cereal-based products . Furthermore, fungal species of F. culmorum, F. graminearum, F. equiseti, F. cerealis, and F. semitectum are the most important producers of ZEN which very often infect cereals worldwide, mostly in temperate climates . In general, at any stage of the food production process (pre-harvest, during harvesting, and storage), fungal production of mycotoxins can occur and expose consumers to the risk of contamination, either through food consumption directly or through feed indirectly . The health risks, which are recognized worldwide, are associated with the consumption of contaminated cereal products with Fusarium mycotoxins . These toxins have been associated with human esophageal cancer and birth defects, while animals manifest carcinogenic and genotoxic effects, and lead to reproductive disorders . The International Agency for Research on Cancer (IARC) classifies FB1 into Group 2B as possibly carcinogenic to humans . Furthermore, the frequent co-occurrence of Fusarium mycotoxins in maize raises the question of synergistic or additive actions in the manifestation of toxicity . In order to protect human and animal exposure to mycotoxins, and also to reduce financial losses and improve international trade, numerous measures are needed at different levels. As a result, regulatory limits on significant levels of mycotoxins in food and feed are established by various authorities worldwide. One of them is the European Commission (EC) No. 1881/2006 , which has set maximum levels (MLs) for Fusarium mycotoxins such as FB1 and FB2, ZEN, and DON in different food and feed products, and Commission Recommendation 2013/165/EU on the presence of T-2 and HT-2 toxin in cereals and cereal products . In addition to the direct health risk, the economic losses resulting from mycotoxicosis are enormous. Annual agricultural and industrial losses due to crops contaminated with mycotoxins are estimated in billions of dollars . Among cereals, maize is one of the most susceptible to contamination by mycotoxigenic fungi and mycotoxins. Furthermore, maize is the leading cereal by production in the world and one of the most important agricultural products . Its popularity as a crop is largely due to its versatile functionality as a food source for both humans and animals. In addition to contributing to human nutrition and animal feed, maize has great economic importance for Serbia and Croatia. With an average annual production of 7.4 million tons and an average annual export estimated at 2.7 million tons of maize per year (average data for the period 2018-2020), Serbia is one of the leading maize producers and exporters in Europe . Furthermore, Serbia consumes about 4.7 million tons of maize annually, with the vast majority used for animal feed (4.3 million tons), and 0.4 million tons for food, seed, and industrial uses . In Croatia, maize is the largest single crop with an average annual production of about 2.3 million tons and with average annual exports of about 0.9 million tons (average data for the period 2018-2020) . The presence of toxigenic fungi and mycotoxins in maize is primarily influenced by climatic conditions, genetic factors, fungal activity, agronomic practices, and storage conditions . Extreme weather conditions, as well as reduced annual environmental variability, can cause biotic and abiotic stress in plants . It is thought that this can have a significant effect on the extent of Fusarium infection in plants . Therefore, it is very important to take into account climatic conditions, as well as climate changes, when studying and analyzing mycotoxins, such as Fusarium mycotoxins. Given that Fusarium mycotoxins represent a potential danger for primary agricultural production in Europe and the world, as well as a high risk to public health, it is necessary to continuously monitor them. Furthermore, increasingly significant changes in mycotoxin occurrence patterns in cereals due to extreme weather conditions, which are part of ongoing climate change, are of concern. Thus, the aim of this study was to examine and compare the influence of weather conditions on the occurrence and concentration of Fusarium mycotoxins, including DON, ZEN, FUMs (FB1 and FB2), T-2, and HT-2 in maize samples during a four-year period (2018-2021), and to discuss the results of these mycotoxin studies performed in the last 10 years (2012-2021) in two neighboring countries, Serbia and Croatia. 2. Materials and Methods 2.1. Samples Maize samples (n = 400) were randomly collected from the main maize producing areas in Northern Serbia (Autonomous Province of Vojvodina, latitude 45deg18'' (N), longitude 20deg09'' (E), and altitude 111 m) during a period from 2018 to 2021 . A total of 100 maize samples from each year were taken evenly from the entire region of Northern Serbia. In Croatia 268 maize samples were analysed on DON, 191 samples on FUMs, 382 samples on ZEN, and 262 samples on T-2/HT-2 toxin concentrations. Samples were collected within the period 2018-2021 from different fields situated in four Croatian regions (Eastern, Central, Northern and Western Croatia), with latitude 45deg10'' (N), longitude 15deg30'' (E), and altitude 142 m . Since the largest production of maize is in Eastern Croatia, about 70% of the samples originate from that part of the country. Sampling in both countries was performed according to Commission Regulation (EC) No. 401/2006 , while in Serbia the Serbian Regulation was also taken into account. With the aim to examine the influence of weather conditions on the contamination of maize samples with Fusarium mycotoxins, and to avoid the possibility of secondary contamination, maize samples from both countries were collected during September and October (2018-2021), immediately after harvest or from dryers, before further storage in silo and distribution. In Serbia after harvest in each year, at the Institute of Food technology in Novi Sad, representative maize samples were prepared from approximately 10 kg of aggregate samples. Preparation of representative samples included: homogenization (Nauta mixer, model 19387, Nauta patenten, The Netherlands), quartering, milling (KnifetecTM 1095 mill, Foss, Hoganas, Sweden), packing in zip lock bags (150-200 g), and storage at -18 degC. The laboratory samples were removed from freezing at the beginning of 2022, and again homogenized (Rotary laboratory mixer RRM Mini-II, Ludwigshafen, Germany) before analysis on high-performance liquid chromatography-tandem mass spectrometry (LC-MS/MS). Before mycotoxins analysis in maize samples from Croatia by ELISA methods, each year, the prepared test portions were ground into a fine powder having a particle size of 1.0 mm using an analytical mill (Cylotec 1093, Tecator, Sweden), and then stored at 4 degC until analyses of particular mycotoxins within 48 h. 2.2. Fusarium Mycotoxins Analysis In Serbia LC-MS/MS was used for analysis of the collected maize samples. The following chemicals were used: acetonitrile (ACN) (Fisher Scientific, Geel, Belgium) of HPLC grade and water (Fisher Scientific, Geel, Belgium), methanol (MeOH) (Carlo Erba, Val de Reuil, France), and formic acid (Fluka Analytical, Sigma Aldrich, Steinheim, Germany) of LC-MS/MS grade. Ultra-pure water was produced by Adrona Crystal EX HPLC Water Purification system (Riga, Latvia). Mycotoxin standards DON (100 mg/mL), FUMs (FB1 and FB2, each 50 mg/mL), HT-2 toxin (100 mg/mL), and ZEN (100 mg/mL) were purchased from Fluka Analytical (Steinheim, Germany), except for T-2 toxin (100 mg/mL) which was purchased from LGC Standards (Wesel, Germany). A mixed standard stock solution of 6 different mycotoxin standards was dissolved in MeOH/water (50:50, v/v). Correspondingly, matrix-matched standards were prepared by diluting an appropriate volume of the appropriate stock standard solution in the blank sample extract with MeOH/water (50:50, v/v), yielding concentration levels from 5 to 150 ng/mL for DON, from 0.5 to 200 ng/mL for FB1 and FB2, from 0.05 to 30 ng/mL for HT-2 toxin and T-2 toxin, and from 1 to 200 ng/mL for ZEN. Both mycotoxin standards and stock standard solutions were stored in the freezer at -18 degC until analysis. Sample preparation and analysis were performed according to Application Note from Thermo Fisher Scientific . Briefly, five grams of milled and homogenized sample was weighed into a 50 mL centrifuge tube. Twenty milliliters of ACN/water (80:20, v/v) extraction solvent was added and the tube was shaken on a horizontal shaker (6 Hz, 60 min). This was followed with centrifugation at room temperature (67 Hz, 5 min). Supernatant (400 mL) was diluted with 600 mL MeOH/water (50:50, v/v), and shaken on a vortex shaker. The supernatant was filtered through a 0.2 mm PTFE filter into HPLC vials and injected into the LC-MS/MS system. LC-MS/MS analysis was performed using an HPLC Vanquish Core system (Thermo Fisher Scientific, Waltham, MA, USA) equipped with a TSQ Quantis Triple Quadrupole mass spectrometer equipped with a heated electrospray ionization (HESI) source (Thermo Fisher Scientific, Waltham, MA, USA). A ZORBAX Eclipse Plus C18 column (2.1 x 100 mm, 1.8 mm) (Agilent, Santa Clara, CA, USA) was selected for chromatographic separation. A mobile phase consisting of water with 0.1% formic acid (mobile phase A) and 0.1% formic acid in methanol (mobile phase B) was used. The gradient elution program was as follows: 0 min 95% mobile phase A, 0.5 min 95% mobile phase A, 7 min 30% mobile phase A, 9 min 0% mobile phase A, 12 min 0% mobile phase A, 12.1 min 95% mobile phase A, and 15 min 95% mobile phase A. The mobile phase flow rate was 0.3 mL/min. The temperature of the column was set at 40 degC and the autosampler tray temperature was set at 20 degC. The sample injection volume was 10 mL. The mass spectrometer analyses were carried out using selected reaction monitoring (SRM) mode. The HESI source was operated under positive (3.5 kV) mode, except for ZEN in negative (2.5 kV) mode. The applied parameters were as follows: ion transfer tube, 325 degC; vaporizer temperature, 350 degC; cycle time, 0.5 s; collision induced dissociation gas pressure, 1.5 mTorr; auxiliary gas, 6 Arb; sheath gas, 30 Arb; and sweep gas, 1 Arb. Argon served as collision gas, while nitrogen served as auxiliary, sheath, and sweep gas. Data were acquired and analyzed using Thermo Scientific TraceFinder software TSQ Quantis 3.2 Tune (Thermo Fisher Scientific, Waltham, MA, USA). The LC-MS/MS method was validated in accordance with the performance criteria outlined in Commission Decision and Technical Report CEN/TR 16059:2010 . The method was validated in terms of linearity, limit of quantification (LOQ), trueness, recovery, repeatability, and reproducibility. The validation study was conducted by the analysis of quality control materials (for DON, FB1, FB2, ZEN) and spiked uncontaminated maize samples (for T-2 and HT-2 toxin). The quality control materials, maize flours, for the examined mycotoxins were provided by: Trilogy analytical laboratory for FUMs (product code TR-F100; contained 2400 mg/kg of FUMs) and ZEN (product code TR-Z100; contained 59.4 mg/kg of ZEN), and Food Analysis Performance Assessment Scheme (FAPAS) for DON (Food Chemistry Proficiency Test 04378; contained 1200 mg/kg of DON). Matrix effects were accounted by using matrix matched calibration (MMC). The matrix effects were expressed as the signal suppression/enhancement (SSE) and calculated from the slope ratio for MMC and solvent calibration. For each examined Fusarium mycotoxin, the primary product ion which corresponds to the most abundant product ion was used for quantification, while the other two were used for confirmation: DON (249.0, 231.0, and 175.0), FB1 (334.1, 352.1, and 704.4), FB2 (336.1, 354.2, and 688.5), ZEN (175.0, 131.0, and 273.0), T-2 (245.1, 327.0, and 387.1), and HT-2 toxin (345.1, 255.1, and 285.1). The retention times for the examined Fusarium mycotoxins were the following: DON, 1.05 min; FB1, 10.85 min; FB2, 11.50 min; ZEN, 11.66 min; T-2, 11.30 min; and HT-2, 10.86 min. All obtained validation parameters, for the applied LC-MS/MS method in this study, fulfil the criteria given under the Regulation 2006/401/EC and the Technical Report . All mycotoxins were quantified using MMC curves, because for all investigated mycotoxins SSEs were higher than +-20%. For all curves the squared correlation coefficients (R2) were above 0.998. LOQs for DON, FB1, FB2, ZEN, T-2, and HT-2 toxin were 50.0, 5.0, 5.0, 10.0, 0.50, and 0.50 mg/kg, respectively. The analysis of quality control materials yielded values for trueness: DON (104%), FB1 (83%), FB2 (81%), ZEN (110%), while the analysis of spiked uncontaminated maize samples yielded recovery: T-2 (96%), HT-2 toxin (88%). Repeatability and reproducibility were calculated as relative standard deviations and none of them exceeded 20%. Therefore, all obtained values for trueness, recovery, repeatability, and reproducibility were in accordance to the criteria specified in used regulations . In Croatia, concentrations of mycotoxins were determined using competitive ELISA Ridascreen(r) test kits for DON (Art. No. R5906), ZEN (Art. No. R1401), FUMs (Art. No. R3401), and T-2/HT-2 toxin (Art. No. R3805). Analyses were performed completely as instructed by the kits manufacturer (R-Biopharm, Darmstadt, Germany). Each kit contains: a microtiter plate with 96 wells coated with antibodies; standard solutions containing different concentrations of mycotoxin standards: DON (0, 3.7, 11.1, 33.3, 100 mg/L); ZEN (0, 50, 150, 450, 1350, 4050 ng/L); FUM (0, 0.025, 0.074, 0.222, 0.666, 2 mg/L), and T2-HT2 (0, 1, 3, 6, 12, 36 mg/L); an enzyme conjugate; an anti-antibody; the substrate and chromogen solution (urea peroxide/tetramethylbenzidine); washing and dilution buffers and stop solution. All other chemicals used in the analysis were of analytical grade. ELISA tests were evaluated using a ChemWell auto-analyzer (Awareness Technology Inc. 2910, Palm City, FL, USA) with the absorbance being measured at 450 nm. In order to determine mycotoxin concentrations in the sample, a standard curve was plotted based on the sample extract dilution factors. Implemented ELISA methods were validated in earlier studies and final concentrations were calculated per each mycotoxin based on the average recoveries. Further, applied ELISA methods were validated as accredited in agreement with the ISO/IEC 17025 Standard . LOQ values for DON, ZEN, FUMs, and T-2/HT-2 equaled to 22, 3, 24, and 5 mg/kg, respectively. For all ELISA tests the validation was performed by determination of recovery and intermediate precision. The recovery rates were determined at three different levels (50, 100, and 200 mg/kg) by spiking the control maize samples with the standard in-house mycotoxin working solution (300 mg/L) corresponding to the assessed content levels. Regarding the determination of intermediate precision, the same steps were repeated on two other occasions by two different analysts and under the same analytical conditions. The mean recovery rates for DON, ZEN, FUMs, and T-2/HT-2 toxins were 98.2%, 89.2%, 75.2%, and 97.6%, while for intermediate precision the mean rates equaled to 95.9%, 86.8%, 72.1%, and 92.9%, respectively. Quality control was performed by analysis of different reference materials (RMs) in parallel with each batch of the studied samples, so as to check whether the obtained concentration falls within the assigned range. The producer (FAPAS, Sand Hutton, York, UK) assigned ranges of RMs were: T04354QC, maize--1104 mg/kg (756-1452 mg/kg) for DON and 90.7 mg/kg (50.8-130.5 mg/kg) for ZEN; TYG079RM, maize flour--854 mg/kg (791-917 mg/kg) for FUM; and TYG087RM, cereal-based animal feed--523 mg/kg (496-550 mg/kg) for T-2/HT-2. Mycotoxin concentrations obtained with RMs were compared to the assigned values given by the manufacturer and were found to be within the defined ranges. 2.3. Weather Analysis With the aim to investigate the influence of weather conditions on the natural occurrence of DON, FUMs, ZEN, T-2, and HT-2 toxin, a detailed analysis of weather condition parameters for maize growing seasons (April-September) in a period of four years (2018-2021) for investigated regions in Serbia and Croatia, was conducted. Weather condition parameters related to the monthly average air temperature and sum of precipitation, as well as drought indicators (Standardized Precipitation Index (SPI-2), Palmer Z Drought Index, and Palmer Drought Severity Index (PDSI)) were acquired from the Republic Hydrometeorological Service of Serbia and from Croatian Meteorological and Hydrological Service . Deviations of mentioned data were determined in comparison to the data recorded in the long-term period, from 1981 to 2010. 2.4. Statistical Analysis For statistical analysis of data obtained in the validation study, as well as for the analysis of mycotoxin occurrence, and weather condition parameters, Microsoft Excel 2010 was used. The following functions have been used: percentage, minimum, maximum, mean, median, standard deviation, and sum. The statistical analysis of data was performed only on positive samples in which the determined concentrations were higher than the LOQs of methods applied. 3. Results and Discussion In this study, the natural occurrence of DON, FUMs, ZEN, T-2, and HT-2 toxins was examined in Serbian and Croatian maize samples collected over a four-year period (2018-2021), and the obtained results are interpreted in relation to the recorded weather data. 3.1. Serbia 3.1.1. Fusarium Mycotoxins Occurrence in Serbia in the Period 2018-2021 Among 400 analyzed maize samples (100 samples each year), the overall average incidence of Fusarium mycotoxins in the investigated four-year period was as follows: 10% of maize samples contained DON, 97% FB1, 97% FB2, 1% ZEN, and 29% T-2 toxin, while 8% of samples contained HT-2 toxin. The frequency and the level of contamination of examined Fusarium mycotoxins varied by production year, as can be seen in Table 1. Among examined mycotoxins in this study, FUMs (FB1 and FB2) were the most prevalent in each year of investigation. FB1 and FB2 contaminated 100% of the analyzed maize samples from years 2018, 2020, and 2021, while 89% and 88% of analyzed maize samples from year 2019 were contaminated with FB1 and FB2, respectively. The highest mean and median concentration of FB1 and FB2 were detected in maize samples from the year 2021 (Table 1). The second most common mycotoxin which occurred in Serbian maize samples was T-2 toxin, with the highest incidence (66%) in the year 2019, but with a very low average and median concentration (Table 1). In regard to DON and HT-2 toxin, it could be observed that those mycotoxins occurred in maize samples collected from the investigated period in the range from 5% to 17% and from 6% to 12%, respectively. Further, ZEN was the least frequently detected among examined Fusarium mycotoxins in this study. In the maize growing season 2019, none of the analyzed samples was contaminated with this mycotoxin, while in the other three years, the presence of ZEN was determined only between 1% to 2% of analyzed maize samples. From Table 1 it could be noted that the obtained concentrations of DON, ZEN, T-2, and HT-2 toxins in examined maize samples from Serbia did not exceed the maximum levels (MLs) intended for human or animal nutrition as regulated by both the European Union and Serbian Regulative . Conversely, in examined maize samples from the four-year period, the sum of FB1 and FB2 concentrations exceeded ML of 4000 mg/kg in 16%, making them unsuitable for human nutrition, while none of the analyzed maize samples contained FUMs above the ML of 60,000 mg/kg intendent for animal nutrition. It is well known that the weather conditions (especially air temperature and amount of precipitation) during the maize growing season represent factors with the strongest influence on the occurrence of mycotoxins in maize in general . Therefore, for a better interpretation of the obtained results in this study, an analysis of weather witnessed for the period 2018-2021 was undertaken. 3.1.2. Weather in Serbia in 2018-2021 The weather conditions (average air temperatures and sum of precipitation) compared to the long-term average values (1981-2010) during the maize growing season (April-September 2018-2021) in Northern Serbia are summarized and shown in Figure 2a,b. Additionally, drought indicators were analyzed for the period of the generative phase of maize from May to September (2018-2021), and are presented in Table 2. It can be noticed that the monthly average air temperatures in each study year were mostly above the long-term average values (1981-2010) during the maize growing season. Only during April 2021 and May 2019, 2020, and 2021 were lower air temperatures recorded in comparison to the long-term average temperatures (1981-2010). On the other hand, considerable differences are observed in the sum of precipitation in the same period compared to long-term average values . Namely, an extremely high amount of precipitation was recorded in May 2019 and June 2018, while in June 2021 a significantly lower sum of precipitation occurred. Further, the amount of precipitation during July was higher than the long-term average values in the 2018 and 2021 maize growing seasons, while in the other two years of investigation it was around (2020) or lower (2019) than the long-term average sum values (1981-2010). In all investigated years, air temperature considerably deviated in August in comparison to the long-term average, while the sum of average precipitation was around (2018, 2020) or lower (2019, 2021) than the long-term average. The air temperature was considerably higher for all years in September followed by a considerably lower sum of precipitation compared to the long-term average. To determine the dry/wet conditions during investigated years, values of three different drought indicators were considered (Table 2). According to the SPI-2, which represents deviations of the observed total precipitation over a 2-month accumulation period, 2018 has been characterized mainly as normal except for May, which was characterized as moderately humid. Further, May and June in 2019 were extremely humid, while the other three months of the observed period were normal. In 2020 all investigated months were characterized as normal except May, which was characterized as severe drought. The year 2021 was characterized with normal (May, July, August) to moderate drought condition (June-September). Palmer Z Drought Index as a measure of short-term drought on a monthly scale indicated weather conditions from normal to extreme drought during a period of May-September in 2021, and from extremely humid to moderate drought in 2019 in the same period. In 2018 and 2019 (May-September), normal to severe drought and extremely humid to moderate drought, respectively, were recorded. Furthermore, the Palmer Drought Severity Index is a standardized index based on a simplified soil water balance; estimated relative soil moisture conditions indicate normal conditions during period May-September in 2018 and 2019, and dry weather conditions during 2020 and 2021, particularly in 2020. Briefly, the weather conditions in the period from April to September during the four-year (2018-2021) investigation were as follows: 2018 warmer and slightly wetter than average conditions; 2019 and 2020 warmer with normal humidity; and 2021 warmer and drier than average condition . However, in 2021 climatic extremes were recorded, such as late frosts during the spring, drought during the summer months, and there were hailstorms and storms, which also had a negative impact on agricultural crop production . Weather conditions in the investigated period (2018-2021), as can be seen from Table 2, did not prove to be suitable for the synthesis of most of the analyzed Fusarium mycotoxins, with the exception of FB1 and FB2. As already stated above, the highest mean concentration of FB1 (4684 +- 3517 mg/kg) and FB2 (1300 +- 994 mg/kg) and median concentration of FB1 (3921 mg/kg) and FB2 (1073 mg/kg) in 2021, were determined. Furthermore, in 2021 the highest quantified concentrations of FB1 and FB2, 21,239 mg/kg and 5825 mg/kg, respectively, were observed. These data could be related to the amount of precipitation and the average daily temperature during the flowering period of maize. Namely, in July 2021, a considerably higher average amount of precipitation was recorded (93.6 mm), which represents 49% more precipitation compared to the long-term average (63 mm). The precipitation in July was often local and came in the form of rainstorms with a huge precipitation fall in a short amount of time. In the same month, deviation of the average daily temperature from the long-term average of 3.1 degC was recorded, while according to the drought indicators (SPI-2, Z, and PDSI index), normal weather conditions were registered . When we consider the values of the SPI-1, which represent deviations of the observed total precipitation over a 1-month accumulation period, July in 2021 was characterized with moderate humid conditions according the SPI-1 Index (1.0). Considering all mentioned facts, high average daily temperatures, high sum of precipitation, and moderate humid weather conditions during July in 2021 were favorable for infection and growth of Fusarium species; as reported worldwide, FUMs are a group of mycotoxins mainly produced by F. verticillioides and F. proliferatum under favorable high-temperature and humid climates . Furthermore, the lack of precipitation and higher average mean daily temperatures during August and September in 2021 led to FUMs synthesis in high concentrations, as a response of Fusarium species to the unfavorable weather conditions. This conclusion is based on the results of a study conducted by Marin et al. . Namely, in their investigation , the effects of ecophysiological factors, temperature, and solute potential on both growth and the regulation of the FUMs biosynthetic FUM1 gene were studied and compared in one isolate of each of the two closely related FUMs-producing maize pathogens F. verticillioides and F. proliferatum. Both of the investigated Fusarium species have shown similar profiles of growth (17-35 degC), but optimal growth condition for F. verticillioides was maintained at higher temperatures and lower solute potential values. The results of this study indicated that environmental conditions leading to water stress (drought) might result in increased risk of FUMs contamination of maize caused by F. verticillioides. On the other hand, F. proliferatum showed a stable expression pattern of the FUM1 gene regardless of water potential conditions . It could be assumed that both F. verticillioides and F. proliferatum occurred in Serbian maize, since FUMs were present in very high frequency in the maize samples from all four investigated growing seasons (2018-2021). 3.2. Croatia 3.2.1. Fusarium Mycotoxins Occurrence in Croatia in the Period 2018-2021 The frequency and the contamination level of maize with Fusarium mycotoxins determined in Croatia during the period 2018-2021 are shown in Table 3. The overall average incidence of Fusarium mycotoxins in maize for the investigated four-year period showed that 78% of maize samples are contaminated with DON, 89% with FUMs, 45% with ZEN, and 50% with T-2/HT-2 toxin. In Croatia during the period 2018-2021, the highest occurrence for DON (95%) was observed in 2021 and its highest mean concentration was observed in 2020 (922 +- 1511 mg/kg). Occurrence for ZEN (63%) and FUMs (95%) was the highest in 2020, but contrary to DON, the highest mean concentrations for these mycotoxins were determined in 2021 (ZEN 182 +- 242 mg/kg and FUMs 1314 +- 1740 mg/kg). Values higher than MLs defined for maize as food were determined in 4% of the samples analyzed for DON, 3% for FUMs, and 0.5% for ZEN, whereas for T-2/HT-2 toxin 1% of the samples was not in accordance with recommended values defined for maize as food. However, according to the obtained concentrations for all four analyzed mycotoxins, all samples were compliant to be used as feed. In order to connect the results of mycotoxin occurrence and concentrations obtained in this study, with the weather witnessed for the period 2018-2021 in Croatia, those data were further presented and interpreted. 3.2.2. Weather in Croatia in the Period 2018-2021 The average air temperatures and sum of precipitation in the period April-September 2018-2021 compared to the long-term average values in the period 1981-2010 during the maize growing season in Croatia are shown in Figure 3a,b. Additionally, drought indicators for the period from June to August (2018-2021), are presented in Table 4. Analysis of meteorological data from stations in the lowland part of Croatia indicates lower air temperatures during May (2019, 2020, and 2021) when booting of maize usually occurs. In addition to lower air temperatures, an extremely high amount of precipitation was recorded during May of 2019. The amount of precipitation also increased during May 2021, while in May of 2018 and 2020 precipitation did not deviate considerably from the long-term average of 1981-2010. The amount of precipitation during July, when the flowering of maize mainly takes place, was higher than the long-term average in all analyzed years. Air temperature was considerably higher in July of 2021, while during July of 2018, 2019, and 2020, air temperature did not deviate considerably from the long-term average. August of 2018, 2019, and 2020 was warmer in comparison to the long-term average, while the total amount of precipitation recorded during August of 2018, 2019, and 2021 was lower than the long-term average. Values of three different drought indicators were examined to determine the dry/wet conditions during the analyzed years (Table 4). The SPI-2 index shows deviations of the observed total precipitation over a 2-month accumulation period. The Palmer Z index was calculated on a monthly basis, while the PDSI carries information about long-term drought. According to the SPI-2 and Z drought indicators, wet conditions occurred during June 2019, while the Z drought indicator indicates slightly wet conditions in June 2021 as well. The PDSI indicator indicates dry conditions during 2018, 2019, and 2020. It is known that humid and cool conditions, in addition to moderate temperatures (between 20 and 30 degC) during the period of maize booting, together with high relative humidity (90%), frequent rainfall during and after flowering, extended periods of high moisture, and the occurrence of air currents promote Fusarium spp. development. Data also suggests that precipitation levels influence fungi and mycotoxin synthesis to a greater extent compared to the influence of temperature . Taking into account drought indicators registered in Croatia for the period May-September (Table 4), results of this study determined for DON, ZEN, and FUM in general can be put into connection with the higher amount of precipitation in May 2019 and 2021 in comparison to the long-term average for this month. Lower contamination in 2018 can be put into connection with data which shows that May of 2018 did not deviate considerably from the long-term average. Additionally, data show that August 2020 was slightly rainier than average. For other monthly average indicators of air temperatures and sum of precipitation for months of the period 2018-2021 in Croatia, there is no clear link with mycotoxin contamination. The highest prevalence of T-2/HT-2 toxin was determined in 2019 (84%), which can be explained with a very to extremely wet June 2019, which followed an extremely wet May 2019. However, due to the interaction of many various factors that may affect the biosynthesis of Fusarium mycotoxins during maize cultivation, such as genetic factors, mechanical damage of kernels, pest infestation, mineral plant nutrition, poor harvest and storage practices, and/or chemical treatment, as observed in some earlier studies , the contamination observed in this study also cannot be put into relation solely to the weather. 3.3. Comparison of Fusarium Mycotoxins Occurrence for the Period 2012-2021 Taking into account the period of ten years (2012-2021), it can be observed that investigated Fusarium mycotoxins occurred very frequently in maize cultured in Serbia and Croatia . Figure 4 shows an overview of the occurrence data for DON, ZEN, FUMs, and T-2/HT-2 toxins, represented as a percentage of contaminated samples. In Figure 4, published data for maize growing seasons 2012-2015 and 2016-2017 for Serbia were summarized. For Croatia, data were partially published , while the other data were not published. Earlier unpublished data and data obtained in this study are also included in Figure 4. The same sites and methodology of sampling were applied as described for the period 2018-2021 for both countries. Further, in Croatia, all mycotoxin analyses, in the period 2012-2017, were conducted by the ELISA methods as described earlier in this study. Serbian maize samples for the period 2012-2015 and 2016-2017 were analyzed by the validated LC-MS/MS methods described in detail by Malachova et al. and Sulyok et al. , respectively. Merged data for Serbia, presented in Figure 4, obtained in our previous studies investigated the contamination of maize with Fusarium mycotoxins in the period 2012-2017. The drought indicators for the period 2012-2017 are given in the Table 5, which describes the summer months of the investigated years in our previous studies. In these studies, the examined period of six years included maize-growing seasons with extreme drought (2012), hot and dry conditions (2013, 2015, and 2017), extreme rainfall (2014), and weather without significant deviations (2016) compared to the long-term period (Table 5). The results showed that FB1 and FB2, regardless of considerable differences in recorded weather conditions, occur in maize samples with a very high frequency of contamination, most often between 96-100%. However, the highest concentrations of FB1 (27,103 mg/kg) and FB2 (4651 mg/kg) were determined in maize samples collected in 2014, in which the maximum recorded value of precipitation registered occurred, since the beginning of meteorological observations in Serbia. In terms of DON and ZEN, the obtained results indicate significant differences in their occurrence in maize samples collected over the period of six years. Both mycotoxins occurred with a prevalence of 100% in maize samples collected from 2014, which was characterized as an extremely rainy year. DON was detected in the concentration range from 428 to 16,350 mg/kg, and even in 84% of samples its concentrations were higher than the ML of 1750 mg/kg defined for maize in food. Further, ZEN was also detected in the highest concentrations (15-2596 mg/kg) in maize from 2014, while in 51% of samples its concentration exceeded ML (350 mg/kg) for maize in food. In less than, or around 10% of samples, detected concentrations of both DON and ZEN were higher than the ML stipulated for maize intended to be used as a feed material. Among six investigated years, in 2014 DON and ZEN were detected in maize with the highest concentrations as well as contamination frequency. In the other five examined years, DON (6-76%) and ZEN (9-62%) occurred with different contamination frequency, while their determined concentrations were lower than MLs for maize intended for food and feed. T-2 and HT-2 contaminated between 6-66% and 12-16%, respectively, of maize samples examined in the period 2012-2017, while no considerable differences were found between their concentrations determined in maize samples originated from different years. Results of earlier studies concluded that maize is one of the most frequent crops in Croatia that is often contaminated with Fusarium mycotoxins . As in Serbia, studies performed in the last period in Croatia showed huge variations of their concentrations, with high incidence influenced by environmental factors such as temperature, humidity, drought, and rainfall during pre-harvest and harvest periods. High concentrations of these mycotoxins could be linked to frequent rainfalls, low temperatures, and significantly lower than average temperatures in the period of maize growth, which increased the contamination of maize with Fusarium moulds and production of their secondary metabolites. The drought indicators for the period 2012-2017 in Croatia are given in the Table 6. Moderate occurrence of all investigated Fusarium mycotoxins in Croatia during 2012 and 2013 (unpublished data) can be connected with normal to mild or extreme drought indicators registered during the period of maize growth to harvest. Higher mycotoxin concentrations determined in Croatia during 2014 can be linked with moderate humid drought indicators. These indicators in 2015 were found to be normal to mild drought or mild humid during the period of June to September. Furthermore, generally lower maize contamination with Fusarium mycotoxins determined in 2016 and 2017 (unpublished data), especially lower in comparison to the most contamination in 2014, can be put into connection with normal drought indicators during the same observed period. Concentrations of DON determined in Croatia in 2014 were higher than MLs in 27% of maize samples, while in 2015 in 44% of maize samples . Levels higher than the guidance values given for feedstuffs (2006/576/EC) were observed in 3% (in 2014) and 7% (in 2015) of maize samples. The highest mean DON concentration of 1998 +- 2517 mg/kg established in 2014 and 3711 +- 2710 mg/kg established in 2015 were determined in maize. Maximal DON levels observed in 2014 and 2015 were 9270 mg/kg and 9560 mg/kg, respectively. The mean ZEN concentration determined in 2014 was 810 +- 858 mg/kg, in 2015 the concentration was 1519 +- 1754 mg/kg, the maximal ZEN concentration in 2014 was 3217 mg/kg, and in 2015 it was 7874 mg/kg. Kis et al. analysed T-2/HT-2 toxin in maize from different fields located in three Croatian regions during 2017-2018. The highest mean sum of concentration in maize was 54 +- 85 mg/kg. In two maize samples, sum concentrations of T-2/HT-2 toxin were higher (332 mg/kg and 253 mg/kg, respectively) than the indicative level stipulated for maize (200 mg/kg). Authors linked obtained results to substantial temperature variations and high precipitation seen during the maize growth and harvesting periods. In regard to both countries, as can be seen from Figure 4, the greatest prevalence of DON and ZEN among the investigated period of 10 years (2012-2021) was noticed in maize samples collected in 2014. The authors of the previously reported studies indicated that extremely high amounts of precipitation in 2014 influenced 96% and 100% of maize samples from Serbia, and 98% and 91% of samples contaminated with DON and ZEN, respectively. As can be seen from Figure 4, the percentage of contaminated maize samples with DON in the period from 2017 to 2021 was low (6-13%) in Serbia, while Croatian maize samples were contaminated with higher frequency in the same period (12-95%). Further, in terms of ZEN, it was also detected with lower frequency in Serbia (0-9%) in comparison to Croatia (12-63%) in the period from 2017 to 2021. Regarding prevalence of the sum of T-2 and HT-2 toxins, it can be noted that in Croatian maize they occurred more frequently (9-84%) than in Serbian maize (0-66%) in the investigated ten year-period. The highest percentage of contaminated maize with T-2 and HT-2 toxins in both countries was recorded in 2019, which was characterized with a higher amount of sum of precipitation in the period May-August. In the investigated ten year-period, FUMs were the most prevalent Fusarium mycotoxins in Serbian maize, since in 7 of 10 investigated years, 100% of the examined samples were contaminated with FB1 and FB2. As in Serbia, FUMs in maize from Croatia occurred with the highest prevalence during the investigated ten year-period (82-100%). Based on the obtained results in this study, as well as the results of our previous study, it can be noticed that in the period of ten years (2012-2021), investigated Fusarium mycotoxins occurred very frequently in maize from Serbia and Croatia. Changes in weather conditions, especially increased amounts of precipitation, influenced the increased prevalence of certain Fusarium mycotoxins, particularly DON and ZEN. Contrary to this, FUMs occurred with high frequency in maize from each of the ten investigated years. It could also be observed that maize samples from each year in the period 2012-2021 represent a mixture of Fusarium mycotoxins. This fact should be taken into consideration, since a co-occurrence of different mycotoxins may generate synergistic or additive effects in humans and animals . The climate of Serbia and Croatia is described as a moderate continental climate. However, in recent years, from 2012 to the present, changes in weather conditions followed by dry and hot conditions with occasional extreme precipitation have been recorded. It is obvious that the trend of increasing frequency of extreme weather events and a changing climate is becoming more and more noticeable. All of the mentioned above has already had an impact on the occurrence of Fusarium mycotoxins in maize from Serbia and Croatia. In both Serbia and Croatia, in the last 10 years, FUMs became the most prevalent Fusarium mycotoxins in cultivated maize. On the other side, changes in weather events in the period 2012-2021, especially dry and hot conditions in 2012, 2013, 2015, 2017, and 2021 in both countries, influenced aflatoxin contamination of cultivated maize . It is in line with recent quantitative estimations which have shown that in certain regions of Europe, increased DON (in wheat) and aflatoxin B1 (in maize) contaminations are expected, as a result of global warming and changeable weather events. 4. Conclusions The frequency and contamination level of Fusarium mycotoxins examined in this study during the period 2018-2022 are shown to vary by year of maize production, and could be linked to climatic conditions per investigated country. Among them, FUMs were found to be the most common contaminants of maize in both Serbia and Croatia. The results indicate that dry and hot weather witnessed in the year 2021 resulted in the highest mean content of FUMs (5984 +- 4500 mg/kg) in maize samples in Serbia, while in Croatia the highest mean content of FUMs (1371 +- 2235 mg/kg) was observed in 2019. Considering the period of 2012-2022, the highest concentrations of Fusarium mycotoxins, especially DON and ZEN, were determined in maize samples cultivated in 2014, which can be liked to an extreme level of precipitation observed for that year in both countries, whereas FUMs occurred with high prevalence in maize from each of the ten investigated years. However, due to the interaction of many various factors that may affect the biosynthesis of Fusarium mycotoxins during maize cultivation, the contamination observed in this study cannot be put into relation solely to the climate conditions. Therefore, further research is needed on the influence of numerous meteorological and agrotechnical factors on the occurrence of mycotoxins in maize, as well as in other cereals. Acknowledgments The authors would like to thank the company Analysis DOO from Belgrade, the Republic of Serbia, which co-financed the publication of this study. Author Contributions Conceptualization, E.J.H., B.R. and J.P.; methodology, B.R., R.R., A.V. and N.K.; validation, B.R., R.R., S.D., J.K. and N.K.; formal analysis, B.R., R.R., N.K. and A.V.; investigation J.K., E.J.H. and J.P.; resources, M.A., J.P., E.J.H. and J.K.; data curation, E.J.H., J.K. and J.P.; writing--original draft preparation, E.J.H., B.R., J.P. and J.K.; writing--review and editing, J.P., E.J.H., J.K., B.R. and M.A.; visualization, E.J.H.; supervision, J.K.; project administration, J.P. and J.K.; funding acquisition, J.K., J.P. and E.J.H. All authors have read and agreed to the published version of the manuscript. Data Availability Statement Data is contained within the article. Conflicts of Interest The authors declare that this study received funding from the Ministry of Education, Science and Technological Development of the Republic of Serbia (451-03-68/2022-14/20022), Croatian Veterinary Institute in Zagreb, and Analysis DOO from Belgrade, the Republic of Serbia. The funders were not involved in the study design, collection, analysis, interpretation of data, the writing of this article, or the decision to submit it for publication. Figure 1 Regions of maize sampling in Serbia and Croatia. Figure 2 Monthly average (a) air temperatures and (b) sum of precipitation in Northern Serbia in the period 2018-2021 (April-September) in comparison with the long-term average values for the period 1981-2010. Figure 3 Monthly average (a) air temperatures and (b) sum of precipitation in Croatia in the period 2018-2021 (April-September) in comparison with the long-term average values for the period 1981-2010. Figure 4 An overview of Fusarium mycotoxins occurrence in maize from Serbia and Croatia in the period 2012-2021. foods-12-01002-t001_Table 1 Table 1 Fusarium mycotoxins occurrence in maize from Serbia in the period 2018-2021. Mycotoxin Year N (%) 1 Min-Max (mg/kg) 2 Mean +- Sdv (mg/kg) 3 Median (mg/kg) 4 DON 2018 5 (5) 50-148 85 +- 43 60 2019 3 (3) 50-277 160 +- 114 152 2020 17 (17) 53-392 144 +- 82 123 2021 13 (13) 56-752 162 +- 181 139 FB1 2018 100 (100) 19-4403 978 +- 732 830 2019 89 (89) 9-3793 614 +- 663 382 2020 100 (100) 45-2435 739 +- 583 533 2021 100 (100) 401-21,239 4684 +- 3517 3921 FB2 2018 100 (100) 17-1297 268 +- 206 221 2019 88 (88) 5-1143 171 +- 192 100 2020 100 (100) 12-638 186 +- 155 136 2021 100 (100) 120-5825 1300 +- 994 1073 ZEN 2018 2 (2) 17-43 30 +- 19 30 2019 nd 5 nd nd nd 2020 2 (2) 17-57 37 +- 28 37 2021 1 (1) 27 nd nd T-2 toxin 2018 14 (14) 0.5-24 5 +- 7 2 2019 66 (66) 0.7-17 3 +- 3 3 2020 12 (12) 0.6-6 2 +- 2 1 2021 25 (25) 0.5-62 7 +- 12 3 HT-2 toxin 2018 6 (6) 1.5-32 13 +- 12 7 2019 9 (9) 0.6-24 6 +- 8 1 2020 6 (6) 0.6-13 6 +- 4 6 2021 12 (12) 0.7-36 8 +- 11 2 1 N (%): number (percentage) of contaminated samples; 2 Min-Max: minimum and maximum concentrations (mg/kg); 3 Mean +- Std: mean concentration (mg/kg) +- standard deviation (mg/kg); 4 Median: median concentration (mg/kg); 5 nd: not detected, i.e., below the limit of quantification (LOQ). foods-12-01002-t002_Table 2 Table 2 Drought indicators registered in Serbia in May-September in the period 2018-2021. Month Year 2018 2019 2020 2021 Standardized Precipitation Index for 60 days (SPI-2) May -0.3 N 1 1.9 EH 2 -1.4 SD 3 -0.1 N June -0.3 N 1.7 EH 0.3 N -1.0 MoD 4 July 1.2 MoH 5 -0.2 N 0.6 N -0.1 N August 0.4 N -0.3 N 0.8 N 0.6 N September -0.6 N -0.3 N 0.0 N -1.2 MoD Palmer Z Drought Index (Z) May -1.2 N 3.6 EH -1.6 MoD 0.2 N June -0.4 N 1.1 MoH 1.3 MoH -3.9 ED 6 July 0.7 N -1.7 MoD -0.3 N 0.5 N August -0.8 MoD -1.6 MoD 0.4 N -1.1 MoD September -2.3 SD -1.0 N -2.8 ED -2.8 ED Palmer Drought Severity Index (PDSI) May -0.7 N -0.9 N -3.9 SD -0.4 N June -0.6 N 0.4 N -3.0 SD 3 -2.2 MoD July 0.7 N -0.7 N -2.6 MoD -1.6 N August 0.1 N -1.4 N -2.0 MoD -2.0 MoD September -1.6 N -1.9 N -2.9 MoD -2.8 MoD 1 N: normal; 2 EH: extremely humid; 3 SD: severe drought; 4 MoD: moderate drought; 5 MoH: moderate humid; 6 ED: extreme drought. foods-12-01002-t003_Table 3 Table 3 Fusarium mycotoxins occurrence in Croatia in the period 2018-2021. Mycotoxin Year N 1 N (%) 2 Min-Max (mg/kg) 3 Mean +- Sdv (mg/kg) 4 Median (mg/kg) 5 DON 2018 78 47 (60) 22-1232 235 +- 324 104 2019 72 51 (71) 27-4688 473 +- 862 168 2020 56 48 (86) 93-9923 922 +- 1511 493 2021 62 59 (95) 43-5134 875 +- 924 572 FUMs 2018 61 51 (84) 28-13800 1006 +- 2170 336 2019 61 57 (93) 30-11530 1371 +- 2235 504 2020 37 35 (95) 24-5920 962 +- 1453 476 2021 32 27 (84) 61-6330 1314 +- 1740 398 ZEN 2018 99 15 (15) 3.6-479 76 +- 125 23 2019 114 53 (46) 3.1-658 69 +- 146 9 2020 84 53 (63) 3.1-1241 120 +- 214 26 2021 85 47 (55) 8.2-1170 182 +- 242 95 T-2/HT-2 2018 93 34 (37) 11-332 58 +- 73 36 2019 94 79 (84) 10-283 32 +- 40 18 2020 39 9 (23) 11-42 18 +- 10 13 2021 36 20 (56) 11-407 56 +- 93 25 1 N: number of total samples; 2 N (%): number (percentage) of contaminated samples; 3 Min-Max: minimum and maximum concentrations (mg/kg); 4 Mean +- Std: mean concentration (mg/kg) +- standard deviation (mg/kg); 5 Median: median concentration (mg/kg). foods-12-01002-t004_Table 4 Table 4 Drought indicators registered in Croatia for May-September in the period 2018-2021. Month Year 2018 2019 2020 2021 Standardized Precipitation Index for 60 days (SPI-2) May -0.4 N 1 1.9 EW 2 -0.7 N 0.7 N June 0.5 N 1.5 VW 3 0.0 N -0.8 N July 0.7 N 0.3 N 0.2 N -1.3 SD 4 August -0.1 N -0.1 N 0.7 N 0.1 N September -0.6 N 0.0 N 0.3 N -0.9 N Palmer Z Drought Index (Z) May -1.5 MoD 5 -0.1 N -3.2 ED 6 0.1 N June -0.4 N 4.1 EW -0.1 N 1.5 SW 7 July 0.8 N -0.6 N -1.0 MoD -3.5 ED August 0.6 N 0.5 N 0.5 N -0.1 N September -1.0 MoD -1.3 MoD 0.2 N -1.1 MoD Palmer Drought Severity Index (PDSI) May 0.4 N -1.0 N -2.2 MoD -0.6 N June 0.4 N -1.1 MiD 8 -2.3 MoD -1.8 MiD July 0.5 N -1.0 N -1.8 MiD -1.7 MiD August -0.5 N -1.3 MiD -1.6 MiD -1.9 MiD September -0.8 N -1.1 MiD -1.7 MiD -2.2 MoD 1 N: normal; 2 EW: extremely wet; 3 VW: very wet; 4 SD: severe drought; 5 MoD: moderate drought; 6 ED: extreme drought; 7 SW: slightly wet, 8 MiD: mild drought. foods-12-01002-t005_Table 5 Table 5 Drought indicators registered in Serbia in June-September in the period 2012-2017. Month Year 2012 2013 2014 2015 2016 2017 Standardized Precipitation Index for 60 days (SPI-2) June -0.8 N 1 0.6 N 1.0 MoH 2 -0.3 N 1.0 MoH -0.6 N July -1.1 MoD 3 -0.9 N 0.7 N -2.7 ExD 4 0.9 N -1.3 SD 5 August -0.8 N -0.6 N 1.2 MoH -0.4 N 0.5 N -0.8 N September -2.1 ED 0.4 N 1.3 MoH 0.6 N 0.4 N 0.0 N Palmer Z Drought Index (Z) June -3.9 ED 6 -0.5 N -1.5 SD -2.8 ED 1.8 MoH -3.1 ED July -1.4 MoD -2.4 SD 2.9 VH 7 -3.9 ED 0.0 N -3.1 ED August -3.9 ED -1.4 MoD 1.2 MoH 0.2 N 0.5 N -2.3 SD September -0.7 MoD 0.8 N 3.9 EH 8 -0.1 N 0.0 N 0.1 N Palmer Drought Severity Index (PDSI) June -3.9 SD 1.1 N -0.9 N 0.4 N 1.4 N -2.9 MoD July -4.1 ED -0.1 N 1.6 MoH -2.8 MoD 1.3 N -3.6 SD August -4.9 ED -2.0 SD 1.9 MoH -2.4 MoD 1.3 N -4.0 ED September -4.0 ED -0.9 N 3.4 VH -2.0 MoD 1.0 N -3.6 SD 1 N: normal; 2 MoH: moderate humid; 3 MoD: moderate drought; 4 ExD: exceptional drought; 5 SD: severe drought; 6 ED: extreme drought; 7 VH: very humid; 8 EH: extremely humid. foods-12-01002-t006_Table 6 Table 6 Drought indicators registered in Croatia in June-September in the period 2012-2017. Month Year 2012 2013 2014 2015 2016 2017 Standardized Precipitation Index for 60 days (SPI-2) June 0.4 N 1 -0.3 N 1.1 MoH 2 0.7 N 0.5 N -0.8 N July -0.7 N -1.1 MoD 3 0.8 N -1.5 SD 4 0.7 N -0.8 N August -2.2 ED 5 -0.3 N 1.4 MoH -0.4 N 0.4 N -0.7 N September -1.1 MoD 0.4 N 1.8 MoH 0.1 N -0.5 N 0.5 N Palmer Z Drought Index (Z) June -0.2 N 0.8 N 3.1 SH 6 2.9 SH 0.7 N -0.6 N July -1.6 MiD 7 -1.4 MiD -0.1 N -1.8 MiD 0.5 N -2.1 MoD August -3.0 SD -1.1 MiD 2.4 MoH -1.5 MiD 1.3 MiH 8 -1.4 MiD September -4.2 ED -0.8 N 2.3 MoH -0.6 N -0.1 N -2.5 MoD Palmer Drought Severity Index (PDSI) June -5.1 ED -0.3 N 1.4 MiH 2.1 MoH 1.9 MiH -0.8 N July -5.4 ED -0.6 N 1.9 MiH 1.7 MiH 1.4 MiH -1.0 N August -6.0 ED -0.7 N 2.3 MoH 1.5 MiH 1.0 N -1.4 MiD September -5.5 ED -0.6 N 3.3 SH 1.4 MiH 0.6 N 0.0 N 1 N: normal; 2 MH: moderate humid; 3 MoD: moderate drought; 4 SD: severe drought; 5 ED: extreme drought; 6 SH: severe humid; 7 MiD: mild drought, 8 MiH: mild humid. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Moretti A. Pascale M. Logrieco A.F. Mycotoxin risks under a climate change scenario in Europe Trends Food. Sci. Technol. 2019 84 38 40 10.1016/j.tifs.2018.03.008 2. Medina A. Gonzalez-Jartin J.M. Sainz M.J. Impact of global warming on mycotoxins Curr. Opin. Food Sci. 2017 18 76 81 10.1016/j.cofs.2017.11.009 3. Kabak B. Dobson A.D. Var I.I.L. Strategies to prevent mycotoxin contamination of food and animal feed: A review Crit. Rev. Food Sci. Nutr. 2006 46 593 619 10.1080/10408390500436185 17092826 4. Ekwomadu T.I. Akinola S.A. Mwanza M. Fusarium mycotoxins, their metabolites (free, emerging, and masked), food safety concerns, and health impacts Int. J. Environ. Res. Public Health 2021 18 11741 10.3390/ijerph182211741 34831498 5. Summerell B.A. Resolving Fusarium: Current status of the genus Annu. Rev. Phytopathol. 2019 57 323 339 10.1146/annurev-phyto-082718-100204 31226019 6. Ferrigo D. Raiola A. Causin R. Fusarium toxins in cereals: Occurrence, legislation, factors promoting the appearance and their management Molecules 2016 21 627 10.3390/molecules21050627 27187340 7. Marin S. Ramos A.J. Cano-Sancho G. Sanchis V. Mycotoxins: Occurrence, toxicology, and exposure assessment Food Chem. Toxicol. 2013 60 218 237 10.1016/j.fct.2013.07.047 23907020 8. Gruber-Dorninger C. Novak B. Nagl V. Berthiller F. Emerging mycotoxins: Beyond traditionally determined food contaminants J. Agric. Food Chem. 2017 65 7052 7070 10.1021/acs.jafc.6b03413 27599910 9. Zinedine A. Soriano J.M. Molto J.C. Manes J. Review on the toxicity, occurrence, metabolism, detoxification, regulations and intake of zearalenone: An oestrogenic mycotoxin Food Chem. Toxicol. 2007 45 1 18 10.1016/j.fct.2006.07.030 17045381 10. Agriopoulou S. Stamatelopoulou E. Varzakas T. Advances in occurrence, importance, and mycotoxin control strategies: Prevention and detoxification in foods Foods 2020 9 137 10.3390/foods9020137 32012820 11. Janic Hajnal E. Kos J. Malachova A. Steiner D. Stranska M. Krska R. Sulyok M. Mycotoxins in maize harvested in Serbia in the period 2012-2015. Part 2: Non-regulated mycotoxins and other fungal metabolites Food Chem. 2020 317 126409 10.1016/j.foodchem.2020.126409 32087516 12. Scott P.M. Recent research on fumonisins: A review Food Addit. Contam. A 2012 29 242 248 10.1080/19440049.2010.546000 21337235 13. Kras K. Rudyk H. Muszynski S. Tomaszewska E. Dobrowolski P. Kushnir V. Muzuka V. Brezvyn O. Arciszewski M.B. Kotsyumbas I. Morphology and Chemical Coding of Rat Duodenal Enteric Neurons following Prenatal Exposure to Fumonisins Animals 2021 12 1055 10.3390/ani12091055 35565482 14. Lee H.J. Ryu D. Worldwide occurrence of mycotoxins in cereals and cereal-derived food products: Public health perspectives of their co-occurrence J. Agric. Food Chem. 2017 33 7034 7051 10.1021/acs.jafc.6b04847 15. Bertero A. Moretti A. Spicer L.J. Caloni F. Fusarium molds and mycotoxins: Potential species-specific effects Toxins 2018 10 244 10.3390/toxins10060244 29914090 16. Alshannaq A. Yu J.H. Occurrence, toxicity, and analysis of major mycotoxins in food Int. J. Environ. Res. Public Health 2017 14 632 10.3390/ijerph14060632 28608841 17. Rai A. Das M. Tripathi A. Occurrence and toxicity of a Fusarium mycotoxin, zearalenone Crit. Rev. Food Sci. Nutr. 2020 60 2710 2729 10.1080/10408398.2019.1655388 31446772 18. Neme K. Mohammed A. Mycotoxin occurrence in grains and the role of postharvest management as a mitigation strategies. A review Food Control 2017 78 412 425 10.1016/j.foodcont.2017.03.012 19. Pereira V.L. Fernandes J.O. Cunha S.C. Mycotoxins in cereals and related foodstuffs: A review on occurrence and recent methods of analysis Trends Food Sci. Technol. 2014 36 96 136 10.1016/j.tifs.2014.01.005 20. Huang D. Cui L. Sajid A. Zainab F. Wu Q. Wang X. Yuan Z. The epigenetic mechanisms in Fusarium mycotoxins induced toxicities Food Chem. Toxicol. 2019 123 595 601 10.1016/j.fct.2018.10.059 30599843 21. International Agency for Research on Cancer Chemical Agents and Related Occupations, a Review of Human Carcinogens in IARC Monograph on the Evaluation of Carcinogenic Risk to Humans Voluem 100F World Health Organization, IARC Lyon, France 2012 1 628 22. Speijers G.J.A. Speijers M.H.M. Combined toxic effects of mycotoxins Toxicol. Lett. 2004 153 91 98 10.1016/j.toxlet.2004.04.046 15342085 23. Commission of the European Communities Commission Regulation EC (1881/2006) Off. J. Eur. Union 2006 49 5 24 24. EC-European Commission Commission recommendation of 27 March 2013 on the presence of T-2 and HT-2 toxin in cereals and cereal products (2013/165/EU) Off. J. Eur. Comm. L 2013 91 12 15 25. FAO Food and Agriculture Organization of the United Nations, 2018-2020 Available online: (accessed on 14 November 2022) 26. Maslac T. Grain and Feed Annual--Serbia United States Department of Agriculture (USDA) Washington, DC, USA 2022 1 17 27. Zadravec M. Markov K. Lesic T. Frece J. Petrovic D. Pleadin J. Biocontrol Methods in Avoidance and Downsizing of Mycotoxin Contamination of Food Crops Processes 2022 10 655 10.3390/pr10040655 28. Bryla M. Pierzgalski A. Zapasnik A. Uwineza P.A. Ksieniewicz-Wozniak E. Modrzewska M. Waskiewicz A. Recent research on Fusarium mycotoxins in maize--A review Foods 2022 11 3465 10.3390/foods11213465 36360078 29. European Regulation No 2006/401/EC of 23 February 2006 laying down the methods of sampling and analysis for the official control of the levels of mycotoxins in foodstuffs Off. J. Eur. Union 2006 70 1 23 30. Serbian Regulation Methods for physical and chemical analysis for quality control of cereals, mills and bakery products, pasta and fast frozen dough Off. Bull. SFRJ 1988 74 1 93 31. Hofmann S. Scheibner O. Quantification of 48 Phytoxins in Cereal Using Liquid Chromotography-Triple Quadrupole Mass Spectrometry. ThermoFisher Scientific, Application Note 65969 2021 1 9 Available online: (accessed on 14 January 2022) 32. European Commission Commission Regulation 2002/657/EC of 14 August 2002 implementing council directive 96/23/EC concerning the performance of analytical methods and the interpretation of results Off. J. Eur. Communities 2002 221 8 36 33. CEN/TR 16059:2010 Food Analysis-Performance Criteria for Single Laboratory Validated Methods of Analysis for the Determination of Mycotoxins European Committee for Standardization, Menagmente Centre Brussels, Belgium 2012 34. Pleadin J. Sokolovic M. Persi N. Zadravec M. Jaki V. Vulic A. Contamination of maize with deoxynivalenol and zearalenone in Croatia Food Control 2012 28 94 98 10.1016/j.foodcont.2012.04.047 35. Pleadin J. Persi N. Mitak M. Zadravec M. Sokolovic M. Vulic A. Jaki V. Brstilo M. The natural occurance of T-2 toxin and fumonisins in maize samples in Croatia B Environ. Contam. Tox. 2012 88 863 866 10.1007/s00128-012-0559-1 36. Pleadin J. Frece J. Lesic T. Zadravec M. Vahcic N. Malenica Staver M. Markov K. Deoxynivalenol and zearalenone in unprocessed cereals and soybean from different cultivation regions in Croatia Food Addit. Contam. B 2017 10 268 274 10.1080/19393210.2017.1345991 28635371 37. Pleadin J. Vasilj V. Kudumija N. Petrovic D. Vilusic M. Skrivanko M. Survey of T-2/HT-2 toxins in unprocessed cereals, food and feed coming from Croatia and Bosnia & Herzegovina Food Chem. 2017 224 153 159 28159250 38. ISO/IEC 17025 General Requirements for the Competence of Testing and Calibration Laboratories ISO Geneva, Switzerland 2017 39. Republic Hydrometeorological Service of Serbia, 2012-2021 Available online: (accessed on 1 November 2022) 40. Croatian Meteorological and Hydrological Service Available online: (accessed on 1 November 2022) 41. European Commission Commission Recommendation 2006/576/EC of 17 August 2006 on the presence of deoxynivalenol, zearalenone, ochratoxin A, T-2 and HT-2 and fumonisins in products intended for animal feeding Off. J. Eur. Union 2006 229 7 9 42. Serbian Regulation Maximum allowed contents of contaminants in food and feed Off. Bull. Repub. Serb. 2011 28 1 16 43. Serbian Regulation Quality of animal feed Off. Bull. Repub. Serb. 2017 54 1 10 44. Agrometeorological Conditions in Production 2020/2021 Years on the Territory of the Republic of Serbia Available online: (accessed on 16 November 2022) 45. Cendoya E. Chiotta M.L. Zachetti V. Chulze S.N. Ramirez M.L. Fumonisins and fumonisin-producing Fusarium occurrence in wheat and wheat by products: A review J. Cereal. Sci. 2018 80 158 166 10.1016/j.jcs.2018.02.010 46. Marin P. Magan N. Vazquez C. Gonzalez-Jaen M.T. Differential effect of environmental conditions on the growth and regulation of the fumonisin biosynthetic gene FUM1 in the maize pathogens and fumonisin producers Fusarium verticillioides and Fusarium proliferatum FEMS Microbiol. Ecol. 2010 73 303 311 10.1111/j.1574-6941.2010.00894.x 20491926 47. Hjelkrem A.G.R. Aamot H.U. Brodal G. Strand E.C. Torp T. Edwards S.G. Dill-Macky R. Hofgaard I.S. HT-2 and T-2 toxins in Norwegian oat grains related to weather conditions at different growth stages Eur. J. Plant Pathol. 2018 151 501 514 10.1007/s10658-017-1394-3 48. Stanciu O. Juan C. Berrada H. Miere D. Loghin F. Manes J. Study on Trichothecene and Zearalenone Presence in Romanian Wheat Relative to Weather Conditions Toxins 2019 11 163 10.3390/toxins11030163 30875933 49. Kis M. Vulic A. Kudumija N. Sarkanj B. Jaki Tkalec V. Aladic K. Skrivanko M. Furmeg S. Pleadin J. A two-year occurrence of Fusarium T-2 and HT-2 toxin in Croatian cereals relative of the regional weather Toxins 2021 13 39 10.3390/toxins13010039 33430378 50. Kos J. Hajnal E.J. Malachova A. Steiner D. Stranska M. Krska R. Poschmaier B. Sulyok M. Mycotoxins in maize harvested in Republic of Serbia in the period 2012-2015. Part 1: Regulated mycotoxins and its derivatives Food Chem. 2020 312 126034 10.1016/j.foodchem.2019.126034 31875527 51. Radic B. Kos J. Janic Hajnal E. Malachova A. Krska R. Sulyok M. Fusarium metabolites in maize from regions of Northern Serbia in 2016-2017 Food Addit. Contam. B 2021 14 295 305 10.1080/19393210.2021.1961877 52. Malachova A. Sulyok M. Beltran E. Berthiller F. Krska R. Optimization and validation of a quantitative liquid chromatography-tandem mass spectrometric method covering 295 bacterial and fungal metabolites including all regulated mycotoxins in four model food matrices J. Chromatog. A 2014 1362 145 156 10.1016/j.chroma.2014.08.037 25175039 53. Sulyok M. Stadler D. Steiner D. Krska R. Validation of an LC-MS/MS-based dilute-and-shoot approach for the quantification of > 500 mycotoxins and other secondary metabolites in food crops: Challenges and solutions Anal. Bioanal. Chem. 2020 412 2607 2620 10.1007/s00216-020-02489-9 32078002 54. Njobeh P.B. Dutton M.F. Koch S.H. Chuturgoon A.A. Stoev S.D. Mosonik J.S. Simultaneous occurrence of mycotoxins in human food commodities from Cameroon Mycotoxin Res. 2010 26 47 57 10.1007/s12550-009-0039-6 23605240 55. Pleadin J. Kos J. Radic B. Vulic A. Kudumija N. Radovic R. Janic Hajnal E. Mandic A. Anic M. Aflatoxins in Maize from Serbia and Croatia: Implications of Climate Change Foods 2023 12 548 10.3390/foods12030548 36766076 56. Van der Fels-Klerx H.J. Liu C. Battilani P. Modelling climate change impacts on mycotoxin contamination World Mycotoxin J. 2016 9 717 726 10.3920/WMJ2016.2066 57. Battilani P. Toscano P. der Fels-Klerx V. Moretti A. Camardo Leggieri M. Brera C. Rortais A. Goumperis T. Robinson T. Aflatoxin B1 contamination in maize in Europe increases due to climate change Sci. Rep. 2016 6 24328 10.1038/srep24328 27066906 |
PMC10000587 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050952 diagnostics-13-00952 Article Blood Transfusions and Adverse Events after Colorectal Surgery: A Propensity-Score-Matched Analysis of a Hen-Egg Issue Catarci Marco 1* Guadagni Stefano 2 Masedu Francesco 3 Montemurro Leonardo Antonio 1 Ciano Paolo 1 Benedetti Michele 1 Delrio Paolo 4 Garulli Gianluca 5 Pirozzi Felice 6 Scatizzi Marco on behalf of the Italian ColoRectal Anastomotic Leakage (iCral) Study Group 7+ Cruciani Mario Academic Editor Pati Ilaria Academic Editor Pupella Simonetta Academic Editor 1 General Surgery Unit, Sandro Pertini Hospital, ASL Roma 2, 00157 Rome, Italy 2 General Surgery Unit, University of L'Aquila, 67100 L'Aquila, Italy 3 Department of Applied Clinical Sciences and Biotechnology, University of L'Aquila, 67100 L'Aquila, Italy 4 Colorectal Surgical Oncology, Istituto Nazionale per lo Studio e la Cura dei Tumori, "Fondazione Giovanni Pascale IRCCS-Italia", 80131 Napoli, Italy 5 General Surgery Unit, Infermi Hospital, 47900 Rimini, Italy 6 General Surgery Unit, ASL Napoli 2 Nord, 80078 Pozzuoli (NA), Italy 7 General Surgery Unit, Santa Maria Annunziata & Serristori Hospital, 50012 Firenze, Italy * Correspondence: [email protected] + See the Appendix A for the complete list of iCral3 investigators. 02 3 2023 3 2023 13 5 95210 2 2023 27 2 2023 01 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Blood transfusions are considered a risk factor for adverse outcomes after colorectal surgery. However, it is still unclear if they are the cause (the hen) or the consequence (the egg) of adverse events. A prospective database of 4529 colorectal resections gathered over a 12-month period in 76 Italian surgical units (the iCral3 study), reporting patient-, disease-, and procedure-related variables, together with 60-day adverse events, was retrospectively analyzed identifying a subgroup of 304 cases (6.7%) that received /or postoperative blood transfusions (IPBTs). The endpoints considered were overall and major morbidity (OM and MM, respectively), anastomotic leakage (AL), and mortality (M) rates. After the exclusion of 336 patients who underwent neo-adjuvant treatments, 4193 (92.6%) cases were analyzed through a 1:1 propensity score matching model including 22 covariates. Two well-balanced groups of 275 patients each were obtained: group A, presence of IPBT, and group B, absence of IPBT. Group A vs. group B showed a significantly higher risk of overall morbidity (154 (56%) vs. 84 (31%) events; OR 3.07; 95%CI 2.13-4.43; p = 0.001), major morbidity (59 (21%) vs. 13 (4.7%) events; OR 6.06; 95%CI 3.17-11.6; p = 0.001), and anastomotic leakage (31 (11.3%) vs. 8 (2.9%) events; OR 4.72; 95%CI 2.09-10.66; p = 0.0002). No significant difference was recorded between the two groups concerning the risk of mortality. The original subpopulation of 304 patients that received IPBT was further analyzed considering three variables: appropriateness of BT according to liberal transfusion thresholds, BT following any hemorrhagic and/or major adverse event, and major adverse event following BT without any previous hemorrhagic adverse event. Inappropriate BT was administered in more than a quarter of cases, without any significant influence on any endpoint. The majority of BT was administered after a hemorrhagic or a major adverse event, with significantly higher rates of MM and AL. Finally, a major adverse event followed BT in a minority (4.3%) of cases, with significantly higher MM, AL, and M rates. In conclusion, although the majority of IPBT was administered with the consequence of hemorrhage and/or major adverse events (the egg), after adjustment accounting for 22 covariates, IPBT still resulted in a definite source of a higher risk of major morbidity and anastomotic leakage rates after colorectal surgery (the hen), calling urgent attention to the implementation of patient blood management programs. blood transfusion colorectal surgery transfusion hazards anastomotic leakage morbidity mortality This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. pmc1. Introduction Preoperative anemia is a very common finding, affecting more than 30% of patient candidates for major digestive surgery . Consequently, it is the strongest predictor of blood transfusions (five-fold) in the postoperative period . Postoperative anemia affects up to 90% of patients after major surgery . The immediate and most widely used treatment for postoperative anemia is blood transfusion, entailing the risk of several complications, culminating in a higher incidence of morbidity and mortality . A recent meta-analysis identified blood transfusions (BTs) as a risk factor for poorer early postoperative outcomes, and previous multicenter prospective studies by the Italian The ColoRectal Anastomotic Leakage (iCral) study group showed /or postoperative BT (IPBT) was independently associated with higher morbidity, anastomotic leakage, and mortality rates after colorectal surgery. However, the results of these studies do not allow one to solve the hen-egg issue in which it is still unclear whether blood transfusions are a definite risk factor for poorer outcomes rather than a marker of bad performers: on the one hand, perioperative blood transfusions may induce immunomodulation (transfusion-related immunomodulation, TRIM) because of the infusion of cytokines, lipids, and allogenic leukocytes, leading to immune activation and resulting in transfusion-related acute lung injury (TRALI) or immune suppression, increasing susceptibility to infectious complications; on the other hand, blood transfusions are generally more frequently administered in patients with major comorbidities, more extensive and longer procedures, more advanced cancer stages, and higher intraoperative blood loss. The iCral study group therefore decided to reappraise the results of its last prospective study (iCral3), trying to solve this hen-egg issue. 2. Materials and Methods This is a retrospective analysis of the iCral3 study, designed to assess the influence of adherence to an enhanced recovery pathway (ERP) on patient-reported outcome measures and return to intended oncologic therapy after colorectal surgery. Seventy-six Italian surgical centers voluntarily participated in a prospective enrolment carried out from November 2020 to October 2021, upon explicit inclusion and exclusion criteria . Adherence to twenty-six items of the ERP was measured for each enrolled case upon criteria adapted from the 2018 ERAS SocietyTM and 2019 national guidelines. For the purposes of this study, the population of 4529 enrolled cases was divided in two groups according to the presence (No. = 304; 6.7%) or absence (No. = 4225; 93.3%) of IPBT. Continuous variables were categorized according to their median value. The Mini Nutritional Assessment--Short Form (MNA-SF ) was categorized for values < 12, indicating potential malnutrition. Surgical procedures were categorized as standard (anterior resection, right colectomy, and left colectomy) versus non-standard (splenic flexure resection, transverse colectomy, Hartmann's reversal, subtotal and total colectomy, and other) resections . Biometric data, patient-, disease-, treatment-, and center-related variables (Table 1) were compared among the two groups using cross tabulation and chi-square or Fisher's exact test where indicated. All analyses were conducted using StatsDirectTM statistical software (StatsDirect Ltd., Wirral, UK); the significance level was set at p < 0.05. 2.1. Outcomes The study endpoints were overall morbidity (OM, any adverse event), major morbidity (MM, any adverse event grade > II according to Clavien-Dindo and the Japanese Clinical Oncology Group (JCOG) extended criteria ), anastomotic leakage (AL), defined according to international consensus , and mortality (M, any death) rates at 60 days post-surgery. 2.2. Propensity-Score-Matched Analysis Neo-adjuvant therapy is a treatment variable exclusively impacting a subgroup of patients; therefore, to avoid bias in the study design, 336 patients who received a neo-adjuvant treatment were excluded and a cohort of 4193 cases was divided into two groups according to the presence (Group A; No. = 280; 6.7%) or absence (Group B; No. = 3913; 93.3%) of /or postoperative blood transfusions (IPBTs). The propensity score matching analysis (PSMA) model was based on (a) IPBT as the treatment (exposure) variable; (b) group A as the true population of interest; (c) group B as the control population; and (d) the following 22 covariates (confounding variables): sex, age, American Society of Anesthesiologists (ASA) class, body mass index (BMI), diabetes, chronic renal failure, dialysis, chronic liver disease, MNA-SF < 12, surgery for malignancy, urgent admission, preoperative steroids, open approach, standard procedure, associated procedures, operation length, ERP adherence rates, preoperative anemia screening, preoperative BT, hospital type, surgical unit type, and center volume. Adjusted logistic regression was used to estimate the propensity scores in the treatment and control groups. Based on the conditioning categorical variables selected, each patient was assigned a propensity score estimated by the standardized mean difference (a standardized mean difference less than 0.1 typically indicates a negligible difference between the means of the groups). No outcome variable was included . As balance is the main goal of PSMA, the analysis was performed using the software "R(c)" (The R Foundation(c) for Statistical Computing, Vienna, Austria) with the following specifications: (a) seed 100 for the reproducibility of the analysis; (b) method for distance metric = nearest, distance = logit, caliper = 0.1, replace = false (without sampling replacement), ratio = 1; (c) adjusted logistic regression to estimate the association between the exposure/treatment variable and the outcomes. The following R(c) libraries/programs were used: "matchit", "glm", "publish", "Tablone", "Plot", and "cobalt" . Balance in the matched groups was assessed by calculating the standardized mean difference (SMD) and general variance ratio (a variance ratio close to 1 means that variances are equal in the two groups). For outcome modeling, an adjusted logistic regression, based on IPBT as the treatment variable and on the same 22 covariates selected for the PSMA, was performed, presenting odds ratios (ORs) and their 95% confidence intervals (95%CI). The eventual effect of any unobserved confounder was tested via sensitivity analysis , using the R(c) software library "SensitivityR5" and presenting the G values (each 0.1 increment of G values representing a 10% odds of differential assignment to treatment due to any unobserved variable). 2.3. Subgroup Analysis in the IPBT Population Considering the population of 304 patients who received one or more IPBT (No. = 304), BT was considered appropriated when administered for Hb levels below liberal transfusion thresholds (<=80 g/L for ASA class I-II, absence of hemodynamic instability, and absence of myocardial ischemia; <=100 g/L for ASA class III, presence of hemodynamic instability, and/or myocardial ischemia). Furthermore, BT was considered (the egg) secondary to bleeding and/or any major adverse event (B/MAE-BT) if it was administered during the operation and/or within 24 h from it, and/or if there was evidence of any previous hemorrhagic (i.e., abdominal bleeding, trocar/wound site bleeding, or anastomotic bleeding) or major adverse event (MAE). Conversely, any MAE was considered (the hen) secondary to BT when it occurred after any BT without any previous hemorrhagic adverse event (BT-MAE). Again, these three BT categories were further tested for the endpoints, individually and combined in several scenarios, using cross tabulation and the chi-square or Fisher's exact test where indicated. All analyses were conducted using StatsDirectTM statistical software (StatsDirect Ltd., Wirral, UK); the significance level was set at p < 0.05. 3. Results The outcomes recorded in the whole population are shown in Table 2. 3.1. Propensity-Score-Matched Analysis After propensity score matching, 3643 cases were excluded (5 with IPBT and 3638 without IPBT, and two groups of 275 patients each were generated: group A (IPBT, true population of interest) and group B (no IPBT, control population)). A good balance between the two groups was achieved , with a model variance ratio of 1.005. After adjusted logistic regression, group A vs. group B (Table 4) showed a significantly higher risk of OM (154 (56.0%) vs. 84 (30.5%) events; OR 3.07; 95%CI 2.13-4.43; p = 0.001), MM (59 (21.4%) vs. 13 (4.7%) events; OR 6.06; 95%CI 3.17-11.6; p = 0.001), and AL (31 (11.3%) vs. 8 (2.9%) events; OR 4.72; 95%CI 2.09-10.66; p = 0.0002). No difference was recorded between the two groups (8 (2.9%) vs. 5 (1.8%) events; OR 1.57; 95%CI 0.42-5.79; p = 0.50) concerning the risk of mortality. Compared to local/regional hospitals, metropolitan/academic hospitals showed a significantly lower risk of OM (125/326 (38.3%) vs. 113/224 (50.4%) events; OR 0.61; 95%CI 0.41-0.92; p = 0.0166) and mortality (3/326 (0.9%) vs. 10/224 (4.5%) events; OR 0.17; 95%CI 0.03-0.90; p = 0.0366). Male vs. female sex was associated with a significantly higher risk of OM (144/309 (46.6%) vs. 94/241 (39.0%) events; OR 1.47; 95%CI 1.0-2.15; p = 0.0487) and MM (50/309 (16.2%) vs. 22/241 (9.1%) events; OR 2.26; 95%CI 1.26-4.08; p = 0.0066). At the same time, operation length > vs. <= 180 min was associated with a significantly higher risk of OM (107/214 (50.0%) vs. 131/336 (39.0%) events; OR 1.60; 95%CI 1.08-2.38; p = 0.0183), enrolment > vs. <= 44 cases to MM (63/438 (14.4%) vs. 9/112 (8.0%) events; OR 2.36; 95%CI 1.04-5.34; p = 0.0397), and presence vs. absence of chronic renal failure to mortality (5/62 (8.1%) vs. 8/488 (1.6%) events; OR 5.11; 95%CI 1.06-24.54; p = 0.0416). 3.2. Subgroup Analysis in the IPBT Population Outcome rates according to individual evaluation of the three BT categories (appropriateness; B/MAE-BT; BT-MAE) are shown in Table 5. Inappropriate BT was administered in more than a quarter of cases, without any significant influence on any endpoint. On the other hand, the majority of BTs were administered after a hemorrhagic or a major adverse event, with significantly higher rates of MM and AL, but not OM or M. Finally, a BT-MAE was recorded in a minority (4.3%) of cases, showing significantly higher MM, AL, and M rates. Six different scenarios were recorded after matching the three BT categories (Table 6). All of the scenarios related to BT determined a significant variation in MM, AL, and M rates, with the worst scenario represented by a major adverse event following an appropriate BT. 4. Discussion The comparison of raw data in the subgroups of the whole population (Table 1) fully agrees with the previous findings of the iCral 1 and 2 studies ; the IPBT subgroup is a reservoir of bad performers (with most of the considered variables showing a significant unfavorable pattern in this subgroup of patients), with significant higher rates of unfavorable outcomes (Table 2). In this setting, it seems that the egg was born before the hen (IPBT may represent the consequence, rather than the cause, of poorer outcomes). Once a nearly perfect balance of the 22 confounding variables was achieved through propensity score matching , the paradigm appeared to be totally reversed; the adjusted logistic regression analysis clearly showed (Table 4) that group A (IPBT), compared to group B (no IPBT), is linked to an independent and significant higher risk of OM, MM, and AL (with the lack of statistical significance of the difference concerning the risk of mortality being possibly due to the small number of recorded events). According to these results, it seems that the hen was born before the egg (IPBT may be the cause, rather than the consequence, of poorer outcomes). Assuming that the probabilities of random assignment to the two treatment groups could be different, the sensitivity analysis (Table 4) showed that the relative impact of unknown and/or unmeasured confounding variables should double (G = 2.3) for OM and triple (G = 3.3) for MM to alter the results and/or their statistical significance. Therefore, the repercussions of this finding on everyday clinical practice are quite relevant: the absolute risk reductions linked to no IPBT recorded in the present study led to small number needed to treat; this could be sufficient to avoid IPBT in 4, 6, and 12 patients to avoid one adverse event, one major adverse event, and one anastomotic leakage, respectively. Another consequence of these findings is that, although the described relationship between blood transfusion and poorer outcomes is not new, a clear understanding of the mechanism by which IPBT may worsen the early outcomes after colorectal surgery is still lacking. Apart from the long-standing and updated concept of TRIM and transient immunosuppression , a recent retrospective propensity-score-matched study on colorectal cancer surgery patients suggested that the worst early outcomes after surgery for colorectal cancer may be mediated by an exaggerated perioperative systemic inflammatory response in patients receiving perioperative blood transfusions. Moreover, recent experimental evidence suggests a direct link between the gut flora composition (microbiota) and the development of antibody-mediated TRALI in mice. The recent introduction of metabolomics and proteomics to transfusion medicine will possibly clarify how the microbiome and gut microbiota can affect the immune system shaping the antigenicity and contributing to TRIM and the potential transmission of infection by blood donors. As the vast majority of colorectal resections are commonly performed for cancer, representing a particularly vulnerable population and showing significant immunosuppression and altered microbiota , further clinical investigation on this issue is warranted. Most of the other significant findings of the logistic regression analysis, such as higher risk of adverse outcomes in male vs. female sex, metropolitan/academic vs. local/regional hospitals, operation length > vs. <= 180 min, and presence vs. absence of chronic renal failure, were expected, having been already recorded in previous studies . On the other hand, the finding of a higher risk of major morbidity recorded in high vs. low volume centers seems to confirm that the surgeon's volume may be more relevant than center volume . Although perioperative BT rates have been declining in the last decade, no change in the risk of mortality after surgery was recorded , and there is still a wide variability in perioperative transfusion practices in colorectal surgery . We decided, therefore, to consider liberal (Hb <= 80-100 g/L) rather than recommended restrictive (Hb <= 70-80 g/L) transfusion thresholds in the analysis of the original subpopulation of patients that received IPBT. Even considering liberal thresholds, inappropriate IPBT was still administered in more than a quarter of cases (Table 5), although this did not determine any significant difference in the outcomes. Anyway, the majority of IPBTs were administered after hemorrhagic and/or major adverse events with a small subgroup of patients (4.3%), in which the BT preceded the major adverse event without any previous hemorrhagic event (Table 5), showing the highest rates of adverse outcomes (Table 6). Applying the long-time-honored 20-80 rule, also known as the "Pareto Principle" , it could be argued that improving transfusion appropriateness and eliminating this small subgroup of patients may allow for a significant improvement in the outcomes. This is the main aim of the recent call toward the urgent need for patient blood management (PBM) program implementation by the World Health Organization and the Italian Surgical Association . Actually, a recent . post-PBM implementation study regarding colorectal cancer surgery from Korea showed a significant decrease in the total transfusion rate, Hb threshold before transfusion (Hb trigger), anastomotic leakage rate, and postoperative length of stay. For these reasons, the iCral study group is currently enrolling patients in its fourth observational multicenter prospective study , designed to test the effect of adherence to a combined ERP-PBM pathway on blood transfusion rates and outcomes. The main strength of this study is its methodology: a large database gathered during a prospective multicenter study was analyzed using a PSMA perfectly responding to the EQUATOR (Enhancing the Quality and Transparency of Health Research) network reporting guidelines . Although observational studies cannot be regarded as a replacement for randomized studies, data generated from large observational cohorts have been increasingly used to evaluate important clinical questions where data from randomized trials are limited or do not exist , mainly because of lower barriers and cost regarding subject recruitment. PSMA offers an alternative approach for estimating treatment effects with observational data when randomized trials are not feasible or unethical, or when researchers need to assess treatment effects based on real life data, collected through the observation of systems as they operate in normal practice without any intervention implemented via randomized assignment rules, responding to the frequent need to draw conditioned casual inferences from quasi-experimental studies. To account for the conditional probability of treatment selection, thus reducing confounding bias, PSMA presents analytical and interpretation challenges that need to be addressed to maintain the reproducibility of its results, which in recent years has been recognized as a crucial element of high-quality research . The relevant quality of the PSMA used in the present study is based on (1) rigorous patient selection from the parent population, performed adhering to explicit criteria; (2) the inclusion of 22 conditioning variables (covariates), such as hospital type, unit type, and accrual volume, to account for the potential heterogeneity of multicenter, clustered data and adherence to the ERP to account for the potential heterogeneity of medical, anesthesiologic, and surgical perioperative management; (3) a clear, sheer, and restrictive balance algorithm , particularly regarding caliper = 0.1, matching ratio = 1:1, and complete balance assessment; (4) complete description of the software package and of its related analytic details; (5) evaluation of the treatment effect through an adjusted multiple regression model including the same 22 covariates used for matching; and (6) a sensitivity analysis to account for unmeasured confounders. The other strength of this study is the large number of enrolled patients in a well-defined time-lapse in a large number of centers, representing a very wide sample of surgical units performing colorectal resections in Italy. While the multicenter nature of the parent database may be a definite source of clustering bias, it is undoubtedly representative of real-life data. However, this study is subject to several limitations, and its results should be interpreted with caution. Several potential confounders were not measured or recorded in the parent study: the number and age of transfused packed red blood cells , postoperative Hb levels, iron and Hb status before and after BT , the management of preoperative and postoperative anemia through high-dose i.v. iron preparations , and, as reported above, the composition of blood donors and recipients' microbiome . Finally, although data quality control was performed and repeated at various levels, we could not rule out potential measurement errors caused by the participating investigators. 5. Conclusions This retrospective PSMA of a large prospective multicenter database confirmed that IPBTs are a definite risk factor for morbidity and anastomotic leakage after colorectal resections even after a well-balanced matching of 22 potential confounders. Although most IPBTs are administered in response to intraoperative blood loss and early postoperative hemorrhagic adverse events, in a minority of cases a major adverse event is triggered by IPBT. In this setting, the avoidance of inappropriate (or unnecessary) BT through the implementation of PBM programs in colorectal surgery may significantly influence the incidence of perioperative adverse outcomes. Author Contributions M.C., iCral study group coordinator, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: M.C., S.G., L.A.M., P.C., M.B., P.D., G.G., F.P. and M.S. Acquisition, analysis, or interpretation of data: M.C., S.G., F.M., L.A.M., P.C., M.B., P.D., G.G., F.P. and M.S. Drafting of the manuscript: M.C., S.G. and F.M. Critical revision of the manuscript for important intellectual content: M.C., S.G., F.M., L.A.M., P.C., M.B., P.D., G.G., F.P. and M.S. Statistical analysis: F.M., S.G. and M.C. All other co-authors participated in data acquisition and quality control, and read and approved the final version of this manuscript. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and the Guideline for Good Clinical Practice E6 (R2) principles. The study protocol was approved by the coordinating center's ethics committee (Comitato Etico Regionale delle Marche--C.E.R.M. #2020/192, approved on 30 July 2020) and was registered at ClinicalTrials.gov (NCT04397627). Thereafter, all participating centers obtained authorization from the local institutional review board. Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement Individual participant-level anonymized datasets are available upon reasonable request by contacting the study coordinator. Conflicts of Interest Catarci reports personal fees from Baxter Spa outside the submitted work. Guadagni, Masedu, Montemurro, Ciano, Benedetti, Delrio, Garulli, Pirozzi, Scatizzi, and all other co-authors have no competing interests to declare. Appendix A + iCral3 study group investigators with shared authorship are as follows: Maria Sole Mattei, MD, Elena Belloni, MD, Matteo Di Carlo, MD, Daniela Apa, MD, General Surgery Unit; Sandro Pertini Hospital, ASL Roma 2; Marco Clementi, MD, General Surgery Unit, University of L'Aquila; Ugo Pace, MD, Andrea Fares Bucci, MD, Colorectal Surgical Oncology, Istituto Nazionale per lo Studio e la Cura dei Tumori, "Fondazione Giovanni Pascale IRCCS-Italia", Napoli; Francesco Monari, MD, General Surgery Unit, Infermi Hospital, Rimini; Antonio Sciuto, MD, General Surgery Unit, ASL Napoli 2 Nord, Pozzuoli (NA); Lorenzo Pandolfini, MD, Alessandro Falsetto, MD, General Surgery Unit, Santa Maria Annunziata and Serristori Hospital, Firenze; Giacomo Ruffo, MD, Elisa Bertocchi, MD, Gaia Masini, MD, General Surgery Unit, IRCCS Sacro Cuore Don Calabria Hospital, Negrar di Valpolicella (VR); Massimo Giuseppe Viola, MD, Amedeo Altamura, MD, Francesco Rubichi, MD, General Surgery Unit, Cardinale G. Panico Hospital, Tricase (LE); Ferdinando Ficari, MD, Francesco Giudici, MD, Fabio Cianchi, MD, General Surgery and IBD Unit, Careggi University Hospital, Firenze; Felice Borghi, MD, Oncologic Surgery Unit, Candiolo Cancer Institute, FPO-IRCCS, Candiolo (TO); Desiree Cianflocca, MD, Marco Migliore, MD, General and Oncologic Surgery Unit, Department of Surgery, Santa Croce e Carle Hospital, Cuneo; Raffaele De Luca, MD, Department of Surgical Oncology, IRCCS Istituto Tumori "Giovanni Paolo II", Bari; Alessandro Rizzo, MD, Department of Medical Oncology, IRCCS Istituto Tumori "Giovanni Paolo II", Bari; Anna Albano, MS, Trial Office (AA), IRCCS Istituto Tumori "Giovanni Paolo II", Bari; Alberto Patriti, MD, Marcella Lodovica Ricci, MD, Department of Surgery, Marche Nord Hospital, Pesaro e Fano (PU); Walter Siquini, MD, Alessandro Cardinali, MD, General Surgery Unit, S. Lucia Hospital, Macerata; Stefano D'Ugo, MD, PhD, FEBS, FACS, Marcello Spampinato, MD, PhD, FEBS (HPB), General Surgery Unit, "V. Fazzi" Hospital, Lecce; Stefano Scabini, MD, Alessandra Aprile, MD, Domenico Soriero, MD, General and Oncologic Surgery Unit, IRCCS "San Martino" National Cancer Center, Genova; Marco Caricato, MD, FACS, Gabriella Teresa Capolupo, MD, FACS, Colorectal Surgery Unit, Policlinico Campus BioMedico, Roma; Giusto Pignata, MD, Jacopo Andreuccetti, MD, Ilaria Canfora, MD, Second General Surgery Unit 2, Spedali Civili di Brescia; Andrea Liverani, MD, Andrea Scarinci, MD, General Surgery Unit, Regina Apostolorum Hospital, Albano Laziale (RM); Roberto Campagnacci, MD, Angela Maurizi, MD, General Surgery Unit, "C. Urbani" Hospital, Jesi (AN); Pierluigi Marini, MD, Grazia Maria Attina, MD, General and Emergency Surgery Unit, San Camillo-Forlanini Hospital, Roma; Ugo Elmore, MD, Giulia Maggi, MD, Department of Gastrointestinal Surgery Unit, San Raffaele Research Hospital and "Vita-Salute" San Raffaele University, Milano; Francesco Corcione, MD, Umberto Bracale, MD, Roberto Peltrini, MD, Maria Michela Di Nuzzo, MD, Minimally Invasive General and Oncologic and Surgery Unit, "Federico II" University, Napoli; Roberto Santoro, MD, Pietro Amodio, MD, General Oncologic Surgery Unit, Belcolle Hospital, Viterbo; Massimo Carlini, MD, FACS, Domenico Spoletini, MD, PhD, FACS, Rosa Marcellinaro, MD, Giorgio Lisi, MD, General Surgery Unit, S. Eugenio Hospital, ASL Roma 2; Antonio Giuliani, MD, Giovanni Del Vecchio, MD, General Surgery Unit, S. Carlo Hospital, Potenza; Mario Sorrentino, MD, Massimo Stefanoni, MD, General Surgery Unit, Latisana-Palmanova Hospital, Friuli Centrale University (UD); Giovanni Ferrari, MD, Carmelo Magistro, MD, General Oncologic and Mininvasive Surgery Unit, Great Metropolitan Niguarda Hospital, Milano; Gianandrea Baldazzi, MD, Diletta Cassini, MD, General Surgery Unit, ASST Ovest Milanese, Nuovo Ospedale di Legnano, Legnano (MI); Alberto Di Leo, MD, Lorenzo Crepaz, MD, General and Minimally Invasive Surgery Unit, San Camillo Hospital, Trento; Augusto Verzelli, MD, Andrea Budassi, MD, General Surgery Unit, Profili Hospital, Fabriano (AN); Giuseppe Sica, MD, Bruno Sensi, MD, Minimally Invasive Surgery Unit, Policlinico Tor Vergata University Hospital, Roma; Stefano Rausei, MD, Silvia Tenconi, MD, General Surgery Unit, Gallarate Hospital (VA); Davide Cavaliere, MD, Leonardo Solaini, MD, Giorgio Ercolani, MD, General and Oncologic Surgery Unit, AUSL Romagna, Forli (FC); Gian Luca Baiocchi, MD, FACS, Sarah Molfino, MD, General Surgery Unit 3, Department of Clinical and Experimental Sciences, University of Brescia; Marco Milone, MD, Giovanni Domenico De Palma, MD, General and Endoscopic Surgery Unit, "Federico II" University, Napoli; Giovanni Ciaccio, MD, Paolo Locurto, MD, General Surgery Unit, S. Elia Hospital, Caltanissetta; Giovanni Domenico Tebala, MD, FACS, FRCS, Antonio Di Cintio, MD, General Surgery Unit, S. Maria Hospital, Terni; Luigi Boni, MD, FACS, Elisa Cassinotti, MD, General Surgery Unit, Fondazione IRCCS Ca' Granda, Policlinico Maggiore Hospital, Milano; Stefano Mancini, MD, Andrea Sagnotta, MD, PhD, General and Oncologic Surgery Unit, San Filippo Neri Hospital, ASL Roma 1; Mario Guerrieri, MD, Monica Ortenzi, MD, Surgical Clinic, Torrette Hospital, University of Ancona; Roberto Persiani, MD, Alberto Biondi, MD, General Surgery Unit, Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Roma; Andrea Lucchi, MD, FACS, Alban Cacurri, MD, General Surgery Unit, "Ceccarini" Hospital, Riccione (RN); Dario Parini, MD, Maurizio De Luca, MD, General Surgery Unit, S. Maria della Misericordia Hospital, Rovigo; Antonino Spinelli, MD, Francesco Carrano, MD, Department of Biomedical Sciences, Humanitas University, Pieve Emanuele (MI) and IRCCS Humanitas Research Hospital, Rozzano (MI); Michele Genna, MD, Francesca Fior, MD, General Surgery Unit, University Hospital, Verona; Vincenzo Bottino, MD, Antonio Ferronetti, MD, General and Oncologic Surgery Unit, Evangelico Betania Hospital, Napoli; Andrea Coratti, MD, Giuseppe Giuliani, MD, Roberto Benigni, MD, General and Emergency Surgery Unit, Misericordia Hospital, Grosseto; Dario Scala, MD, Graziella Marino, MD, Battistino Puppio, MD, Abdominal Oncologic Surgery Unit, IRCCS CROB Basilicata Referral Cancer Center, Rionero in Vulture (PZ); Andrea Muratore, MD, Patrizia Marsanic, MD, Nicoletta Sveva Pipitone Federico, MD, General Surgery Unit, "E. Agnelli" Hospital, Pinerolo (TO); Maurizio Pavanello, MD, Carlo Di Marco, MD, General Surgery Unit, AULSS2 Marca Trevigiana, Conegliano Veneto (TV); Umberto Rivolta, MD, Camillo Leonardo Bertoglio, MD, PhD, General Surgery Unit, Fornaroli Hospital, ASST Ovest Milanese, Magenta (MI); Micaela Piccoli, MD, FACS, Francesca Pecchini, MD, General Surgery Unit, Civil Hospital, Baggiovara (MO); Carlo Talarico, MD, Vincenzo Greco, MD, General Surgery Unit, Villa dei Gerani Hospital, Vibo Valentia (VV); Alessandro Carrara, MD, Michele Motter, MD, Giuseppe Tirone, MD, First General Surgery Unit, S. Chiara Hospital, Trento; Mauro Totis, MD, Nicolo Tamini, MD, Colorectal Surgery Unit, San Gerardo Hospital, ASST Monza; Franco Roviello, MD, Riccardo Piagnerelli, MD, General and Oncologic Surgery Unit, AOU Senese, Siena; Alessandro Anastasi, MD, Giuseppe Canonico, MD, General Surgery Unit, San Giovanni di Dio Hospital, Firenze; Gianluca Guercioni, MD, Simone Cicconi, MD, General Surgery Unit, "C. e G. Mazzoni" Hospital, Ascoli Piceno; Giuseppe Maria Ettorre, MD, Marco Colasanti, MD, General and Transplant Surgery Unit, San Camillo-Forlanini Hospital, Roma; Mauro Montuori, MD, Enrico Pinotti, MD, General and Mininvasive Surgery Unit, S. Pietro Hospital, Ponte San Pietro (BG); Pierpaolo Mariani, MD, Roberta Carminati, MD, General Surgery Unit, Pesenti Fenaroli Hospital, Alzano Lombardo (BG); Nicolo de Manzini, MD, Edoardo Osenda, MD, Surgical Clinic, University of Trieste; Annibale Donini, MD, Luigina Graziosi, MD, General and Emergency Surgery Unit, University of Perugia; Mariano Fortunato Armellino, MD, Ciro De Martino, MD, Giovanna Ioia, MD, General and Emergency Surgery Unit, S. Giovanni di Dio e Ruggi d'Aragona Hospital, Salerno; Lucio Taglietti, MD, Arianna Birindelli, MD, General Surgery Unit, ASST Valcamonica, Esine (BS); Gabriele Anania, MD, Matteo Chiozza, MD, General and Laparoscopic Surgery Unit, University Hospital, Ferrara; Mariantonietta Di Cosmo, MD, Daniele Zigiotto, MD, General and Upper GI Surgery Unit, University Hospital, Verona; Carlo Vittorio Feo, MD, Fioralba Pindozzi, MD, General Surgery Unit, Delta Hospital, Lagosanto (FE); Paolo Millo, MD, Manuela Grivon, MD, General Surgery Unit, "U. Parini" Regional Hospital, Aosta; Corrado Pedrazzani, MD, Cristian Conti, MD, General and HPB Surgery Unit, University Hospital, Verona; Silvio Guerriero, MD, Lorenzo Organetti, MD, General Surgery Unit, "A. Murri" Hospital, Fermo; Andrea Costanzi, MD, Michela Monteleone, MD, General Surgery Unit, S. Leopoldo Hospital, Merate (LC); Nereo Vettoretto, MD, Emanuele Botteri, MD, General Surgery Unit, Spedali Civili of Brescia, Montichiari (BS); Federico Marchesi, MD, Giorgio Dalmonte, MD, Surgical Clinic, University of Parma; Massimo Basti, MD, Diletta Frazzini, MD, General Surgery Unit, Spirito Santo Hospital, Pescara; Graziano Longo, MD, Simone Santoni, MD, General Surgery Unit, Policlinico Casilino, Roma; Moreno Cicetti, MD, Gabriele La Gioia, MD, General Surgery Unit, S. Maria della Misericordia Hospital, Urbino (PU); Italy. Figure 1 Study flowchart according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement guidelines and to the Reporting and Guidelines in Propensity Score Analysis : iCral: Italian ColoRectal Anastomotic Leakage study group; ASA: American Society of Anesthesiologists; IPBT: /or postoperative blood transfusion(s); and ERP: enhanced recovery pathway. Figure 2 (a) Jitter plot distribution of propensity scores in treatment and control groups; (b) Love plot of covariates' standardized mean differences between treatment and control groups before and after matching; the vertical lines represent the interval of +-0.1 and within which balance is considered to be acceptable. diagnostics-13-00952-t001_Table 1 Table 1 Comparative analysis of study variables in the two groups. No a IPBT (No. = 4225) a IPBT (No. = 304) Variable No. (%) No. (%) p Male sex 2218 (52.5) 171 (56.2) 0.205 Age > 69 years 2014 (47.7) 207 (68.1) <0.0001 b ASA class III 1373 (32.5) 164 (53.9) <0.0001 c BMI <= 25.0 kg/m2 2030 (48.0) 160 (52.6) <0.0001 Diabetes 573 (13.6) 57 (18.7) 0.014 Chronic renal failure 161 (3.8) 34 (11.2) <0.0001 Dialysis 7 (0.2) 3 (1.0) 0.025 Chronic liver disease 42 (1.0) 10 (3.3) 0.002 d MNA-SF < 12 1417 (33.5) 164 (53.9) <0.0001 Disease Cancer 3021 (71.5) 262 (86.2) <0.0001 Diverticular 491 (11.6) 13 (4.3) Endometriosis 165 (3.9) 0 (---) e IBD 143 (3.4) 12 (3.9) Lymphoma 3 (0.1) 1 (0.3) Polyp(s) 223 (5.3) 7 (2.3) Other 179 (4.2) 9 (3.0) Elective admission 3970 (94.0) 266 (87.5) <0.0001 Neo-adjuvant therapy 312 (7.4) 24 (7.9) 0.743 Preoperative steroids 71 (1.7) 8 (2.6) 0.221 Open approach 554 (13.1) 76 (25.0) <0.0001 Procedure Right colectomy 1492 (35.5) 172 (56.6) <0.0001 Transverse colectomy 87 (2.1) 4 (1.3) Splenic flexure colectomy 123 (2.9) 10 (3.3) Left colectomy 1111 (26.3) 41 (13.5) Anterior resection 964 (22.8) 37 (12.2) f TaTME 44 (1.0) 7 (2.3) Hartmann reversal 102 (2.4) 9 (3.0) (Sub) total colectomy 81 (1.9) 9 (3.0) Other 221 (5.2) 15 (4.9) g Standard procedures 3567 (84.4) 250 (82.2) 0.311 Associated procedures 753 (17.8) 78 (25.7) 0.0007 Operation length > 180' 2023 (47.9) 131 (43.1) 0.106 h ERP adherence > median (69.3%) 1902 (45.0) 107 (35.2) 0.001 Preoperative anemia screening 464 (11.0) 88 (29.0) <0.0001 Preoperative blood transfusion(s) 212 (5.0) 61 (20.1) <0.0001 Hospital type Local/Regional 1959 (46.4) 127 (41.8) 0.217 Metropolitan/Academic 2266 (53.6) 177 (58.2) Unit type General 3564 (84.4) 262 (86.2) 0.006 Colorectal/Oncologic 661 (15.6) 42 (13.8) Center volume >=44 enrolled cases 3261 (77.2) 245 (80.6) 0.008 a /or postoperative blood transfusion(s); b American Society of Anesthesiologists; c body mass index; d Mini Nutritional Assessment--Short Form; e inflammatory bowel disease; f transanal total mesorectal excision; g the sum of right colectomies, left colectomies, and anterior resections; and h enhanced recovery pathway. diagnostics-13-00952-t002_Table 2 Table 2 Comparative analysis of outcomes in the two groups. No a IPBT (No. = 4225) a IPBT (No. = 304) Variable No. (%) No. (%) p Overall morbidity 1039 (24.6) 175 (57.6) <0.0001 Major morbidity 272 (6.4) 69 (22.7) <0.0001 Anastomotic leakage 167 (3.9) 38 (12.5) <0.0001 Mortality 52 (1.2) 10 (3.3) 0.008 a /or postoperative blood transfusion(s). diagnostics-13-00952-t003_Table 3 Table 3 Variable distribution in treatment and control groups before and after propensity score matching. Before Propensity Score Matching After Propensity Score Matching a IPBT No a IPBT a IPBT No a IPBT Variable Pattern No. = 280 (6.7%) No. = 3913 (93.3%) b p c SMD No. = 275 (50.0%) No. = 275 (50.0%) b p c SMD Sex Male 154 2017 0.001 1.18 151 158 0.69 0.03 Female 126 1896 0.001 1.13 124 117 0.66 -0.03 Age (years) <69 84 2021 0.001 1.26 84 72 0.34 -0.06 >=69 196 1892 0.001 1.06 191 203 0.49 0.05 d ASA class I-II 125 2630 0.001 1.65 125 114 0.46 -0.05 III 155 1283 0.001 0.76 150 161 0.50 0.04 e BMI (kg/m2) <=25 149 1873 0.001 1.10 145 144 1.00 -0.004 >25 131 2040 0.001 1.22 130 131 1.00 0.004 Diabetes Yes 51 532 0.001 0.46 50 56 0.61 0.04 No 229 3381 0.001 2.33 225 219 0.76 -0.02 f CRF Yes 31 152 0.001 0.20 31 31 1.00 0.00 No 249 3761 0.001 3.08 244 244 1.00 0.00 Dialysis Yes 2 5 0.45 0.02 2 1 1.00 -0.03 No 278 3908 0.001 3.46 273 274 1.00 0.004 g CLD Yes 9 38 0.001 0.09 7 4 0.54 -0.05 No 271 3875 0.001 3.36 268 271 0.90 0.01 h MNA-SF <=12 128 2609 0.001 1.63 128 131 0.89 0.01 >12 152 1304 0.001 0.78 147 144 0.89 -0.01 Cancer Yes 239 2713 0.001 1.57 234 237 0.90 0.01 No 41 1200 0.001 0.85 41 38 0.82 -0.02 Admission Elective 243 3661 0.001 2.83 239 239 1.00 0.00 Urgent 37 252 0.001 0.28 36 36 1.00 0.00 Preoperative steroids Yes 7 67 0.001 0.15 7 10 0.62 0.04 No 273 3846 0.001 3.26 268 265 0.90 -0.01 j MI surgery Yes 209 3396 0.001 2.40 207 211 0.85 0.01 No 71 517 0.001 0.43 68 64 0.78 -0.02 Standard procedure Yes 233 3298 0.001 2.20 229 232 0.90 0.01 No 47 615 0.001 0.52 46 43 0.83 -0.02 Associated procedures Yes 71 705 0.001 0.54 69 49 0.06 -0.12 No 209 3208 0.001 2.12 206 226 0.24 0.07 Operation length (min.) <=180 167 2128 0.001 1.23 164 172 0.65 0.03 >180 113 1785 0.001 1.08 111 103 0.59 -0.04 k ERP adherence (%) <=69.3 181 2142 0.001 1.23 177 188 0.52 0.04 >69.3 99 1771 0.001 1.09 98 87 0.42 -0.05 l Preop AS Yes 83 445 0.001 0.36 79 82 0.86 0.02 No 197 3468 0.001 2.55 196 193 0.90 -0.01 m Preop BT Yes 59 206 0.001 0.20 55 55 1.00 0.00 No 221 3707 0.001 3.01 220 220 1.00 0.00 Hospital type n LR 117 1846 0.001 1.11 114 110 0.82 -0.02 o MA 163 2067 0.001 1.20 161 165 0.84 0.02 Unit type p GS 245 3331 0.001 2.23 240 234 0.76 -0.02 q CO 35 582 0.001 0.52 35 41 0.55 0.04 Enrolment (no. of cases) <=44 54 909 0.001 0.68 54 58 0.76 0.02 >44 226 3004 0.001 1.86 221 217 0.85 -0.01 r OM Yes 158 944 0.001 0.58 154 84 0.01 -0.31 No 122 2969 0.001 1.98 121 191 0.01 0.29 s MM Yes 60 236 0.001 0.23 59 13 0.001 -0.34 No 220 3677 0.001 2.94 216 262 0.006 0.17 t AL Yes 31 135 0.001 0.18 31 8 0.001 -0.23 No 249 3778 0.001 3.13 244 267 0.18 0.08 Mortality Yes 9 48 0.001 0.11 8 5 0.58 -0.05 No 271 3865 0.001 3.33 267 270 0.90 0.01 a /or postoperative blood transfusion(s); b Student's test for proportions; c standardized mean difference d American Society of Anesthesiologists; e body mass index; f chronic renal failure; g chronic liver disease; h Mini Nutritional Assessment--Short Form; j Mininvasive; k enhanced recovery pathway; l preoperative anemia screening; m preoperative blood transfusion(s); n local/regional; o metropolitan/academic; p general surgery; q colorectal/oncologic surgery; r overall morbidity; s major morbidity; and t AL: anastomotic leakage. diagnostics-13-00952-t004_Table 4 Table 4 Adjusted multiple regression analysis for endpoints. Overall Morbidity Major Morbidity Overall a AL Mortality Variable Pattern b OR (95%CI) p b OR (95%CI) p b OR (95%CI) p b OR (95%CI) p c IPBT Yes 3.07 (2.13-4.43) 0.001 6.06 (3.17-11.6) 0.001 4.72 (2.09-10.66) 0.0002 1.57 (0.42-5.79) 0.50 No Reference Reference Reference Reference Sex Male 1.47 (1.00-2.15) 0.05 2.26 (1.26-4.08) 0.007 1.24 (0.60-2.56) 0.56 3.56 (0.69-18.42) 0.13 Female Reference Reference Reference Reference Age (years) <69 Reference Reference Reference Reference >=69 0.96 (0.62-1.51) 0.87 0.76 (0.40-1.46) 0.41 0.87 (0.37-2.04) 0.74 4.81 (0.42-55.12) 0.21 d ASA class I-II Reference Reference Reference Reference III 1.02 (0.67-1.54) 0.94 1.10 (0.59-2.06) 0.76 1.88 (0.82-4.31) 0.13 1.63 (0.30-8.72) 0.57 Body mass index (kg/m2) <=25 Reference Reference Reference Reference >25 0.92 (0.62-1.36) 0.68 0.97 (0.54-1.72) 0.91 0.97 (0.47-2.02) 0.94 0.43 (0.10-1.84) 0.26 Diabetes Yes 0.84 (0.52-1.35) 0.47 0.46 (0.20-1.04) 0.06 0.68 (0.25-1.83) 0.45 0.19 (0.02-2.10) 0.18 No Reference Reference Reference Reference Chronic renal failure Yes 0.81 (0.43-1.51) 0.51 1.49 (0.63-3.53) 0.36 1.41 (0.50-3.98) 0.52 5.11 (1.06-24.54) 0.04 No Reference Reference Reference Reference Dialysis Yes 1.97 (0.16-24.8) 0.60 1.52 (0.10-23.6) 0.77 Not Estimable 7.15 (0.15-344.47) 0.32 No Reference Reference Reference Chronic liver disease Yes 0.61 (0.16-2.36) 0.47 0.58 (0.06-5.29) 0.63 Not Estimable Not Estimable No Reference Reference e MNA-SF <=12 Reference Reference Reference Reference >12 1.46 (0.99-2.14) 0.053 0.80 (0.46-1.41) 0.45 1.57 (0.75-3.27) 0.23 2.21 (0.47-10.40) 0.32 Surgery for malignancy Yes 1.22 (0.69-2.17) 0.50 1.42 (0.60-3.37) 0.42 0.77 (0.28-2.13) 0.62 0.51 (0.09-3.01) 0.46 No Reference Reference Reference Reference Admission Elective Reference Reference Reference Reference Urgent 0.88 (0.49-1.56) 0.66 0.81 (0.33-2.00) 0.65 0.77 (0.24-2.47) 0.66 0.70 (0.11-4.69) 0.72 Preoperative steroids Yes 1.13 (0.39-3.33) 0.82 0.31 (0.03-2.82) 0.30 0.68 (0.07-6.11) 0.73 3.75 (0.26-53.80) 0.33 No Reference Reference Reference Reference f MI surgery Yes 0.86 (0.54-1.37) 0.52 0.73 (0.36-1.50) 0.39 1.10 (0.44-2.73) 0.84 0.50 (0.09-2.77) 0.43 No Reference Reference Reference Reference Standard procedures Yes 1.25 (0.75-2.10) 0.39 0.51 (0.26-1.02) 0.06 0.78 (0.31-1.96) 0.60 1.38 (0.23-8.23) 0.73 No Reference Reference Reference Reference Associated procedures Yes 0.87 (0.55-1.38) 0.56 0.56 (0.27-1.15) 0.11 0.44 (0.16-1.17) 0.10 0.68 (0.14-3.32) 0.63 No Reference Reference Reference Reference Operation length (min.) <=180 Reference Reference Reference Reference >180 1.60 (1.08-2.38) 0.02 1.07 (0.60-1.90) 0.83 1.92 (0.92-3.98) 0.08 2.25 (0.56-9.08) 0.26 g ERP adherence (%) <=69.3 Reference Reference Reference Reference >69.3 1.30 (0.84-2.01) 0.24 1.10 (0.58-2.08) 0.76 0.59 (0.25-1.41) 0.24 0.60 (0.13-2.87) 0.52 h Preop AS Yes 1.36 (0.78-2.39) 0.28 1.40 (0.63-3.09) 0.41 1.53 (0.54-4.34) 0.42 0.74 (0.09-6.05) 0.78 No Reference Reference Reference Reference i Preop BT Yes 0.92 (0.49-1.69) 0.78 0.77 (0.32-1.84) 0.55 0.54 (0.16-1.80) 0.31 3.28 (0.47-22.68) 0.23 No Reference Reference Reference Reference Hospital type l LR Reference Reference Reference Reference m MA 0.61 (0.41-0.92) 0.02 0.81 (0.45-1.48) 0.50 1.06 (0.48-2.33) 0.88 0.17 (0.03-0.90) 0.04 Unit type n GS Reference Reference Reference Reference o CO 1.04 (0.58-1.85) 0.90 1.00 (0.42-2.37) 0.99 0.86 (0.28-2.65) 0.79 0.90 (0.07-12.12) 0.94 Enrolment (no. of cases) <=44 Reference Reference Reference Reference >44 0.84 (0.52-1.37) 0.49 2.36 (1.04-5.34) 0.04 1.90 (0.69-5.20) 0.21 1.99 (0.30-13.30) 0.48 Sensitivity analysis G G G G 2.3 0.051 3.3 0.05 2.3 0.06 1 0.29 a AL: anastomotic leakage; b OR (95%CI): odds ratio and 95% confidence intervals; c /or postoperative blood transfusions; d ASA: American Society of Anesthesiologists; e Mini Nutritional Assessment--Short Form; f Mininvasive; g ERP: enhanced recovery pathway; h preoperative anemia screening; i preoperative blood transfusion(s); l local/regional; m metropolitan/academic; n general surgery; and o colorectal/oncologic surgery. diagnostics-13-00952-t005_Table 5 Table 5 Outcome rates according to individual BT categories. Endpoint Overall Morbidity Major Morbidity Anastomotic Leakage Mortality a BT Category Pattern No. (%) No. (%) p No. (%) p No. (%) p No. (%) p Appropriateness Yes 225 (74.0) 131 (58.2) 0.845 51 (22.7) 0.983 28 (12.4) 0.391 10 (4.4) 0.07 No 79 (26.0) 45 (57.0) 18 (22.8) 7 (8.9) 0 (- -.-) b B/MAE-BT Yes 155 (51.0) 89 (57.4) 0.864 52 (33.5) <0.001 24 (15.5) 0.027 6 (3.9) 0.751 No 149 (49.0) 87 (58.4) 17 (11.4) 11 (7.4) 4 (2.7) c BT-MAE Yes 13 (4.3) 10 (76.9) 0.250 12 (92.3) <0.0001 5 (38.5) 0.001 2 (15.4) 0.063 No 291 (95.7) 166 (57.0) 57 (19.6) 26 (10.3) 8 (2.7) a Blood transfusion(s); b /or major-adverse-event-related blood transfusion(s); and c blood-transfusion(s)-related major adverse event. diagnostics-13-00952-t006_Table 6 Table 6 Matching scenarios of BT categories. Scenario Overall Morbidity Major Morbidity Anastomotic Leakage Mortality No. (%) No. (%) a p No. (%) a p No. (%) a p No. (%) a p Appropriate a BT No b B/MAE-BT No c BT-MAE 99 (32.6) 56 (56.6) 0.746 4 (4.0) <0.0001 3 (3.0) 0.0003 2 (2.0) 0.0026 Inappropriate a BT No b B/MAE-BT No c BT-MAE 37 (12.2) 21 (56.7) 2 (5.4) 3 (8.1) 0 (-.-) Appropriate a BT b B/MAE-BT No c BT-MAE 120 (39.5) 70 (58.3) 40 (33.3) 22 (18.3) 6 (5.0) Inappropriate a BT b B/MAE-BT No c BT-MAE 35 (11.5) 19 (66.7) 12 (34.3) 2 (5.7) 0 (-.-) Inappropriate a BT No b B/MAE-BT c BT-MAE 6 (2.0) 4 (66.7) 4 (66.7) 2 (16.7) 0 (-.-) Appropriate a BT No b B/MAE-BT c BT-MAE 7 (2.3) 6 (85.7) 7 (100.0) 3 (42.8) 2 (28.6) a Two by six chi-square test with five degrees of freedom; a blood transfusion(s); b /or major-adverse-event-related blood transfusion(s); and c blood-transfusion(s)-related major adverse event. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Beattie W.S. Karkouti K. Wijeysundera D.N. Tait G. Risk associated with preoperative anemia in noncardiac surgery: A single-center cohort study Anesthesiology 2009 110 574 581 10.1097/ALN.0b013e31819878d3 19212255 2. Fowler A.J. Ahmad T. Phull M.K. Allard S. Gillies M.A. Pearse R.M. Meta-analysis of the association between preoperative anaemia and mortality after surgery Br. J. Surg. 2015 102 1314 1324 10.1002/bjs.9861 26349842 3. Munoz M. Acheson A.G. Auerbach M. Besser M. Habler O. Kehlet H. Liumbruno G.M. Lasocki S. Meybohm P. Rao Baikady R. International consensus statement on the peri-operative management of anaemia and iron deficiency Anaesthesia 2017 72 233 247 10.1111/anae.13773 27996086 4. Ponnusamy K.E. Kim T.J. Khanuja H.S. Perioperative blood transfusions in orthopaedic surgery J. Bone Joint Surg. Am. 2014 96 1836 1844 10.2106/JBJS.N.00128 25378512 5. Kumar A. Perioperative management of anemia: Limits of blood transfusion and alternatives to it Clevel. Clin. J. Med. 2009 76 (Suppl. 4) S112 S118 10.3949/ccjm.76.s4.18 6. Saleh A. Small T. Chandran Pillai A.L. Schiltz N.K. Klika A.K. Barsoum W.K. Allogenic blood transfusion following total hip arthroplasty: Results from the nationwide inpatient sample, 2000 to 2009 J. Bone Joint Surg. Am. 2014 96 e155 10.2106/JBJS.M.00825 25232085 7. Pang Q.Y. An R. Liu H.L. Perioperative transfusion and the prognosis of colorectal cancer surgery: A systematic review and meta-analysis World J. Surg. Oncol. 2019 17 7 17 10.1186/s12957-018-1551-y 30611274 8. Italian ColoRectal Anastomotic Leakage (iCral) Study Group Risk factors for adverse events after elective colorectal surgery: Beware of blood transfusions Updates Surg. 2020 72 811 819 10.1007/s13304-020-00753-2 32399595 9. Catarci M. Ruffo G. Viola M.G. Pirozzi F. Delrio P. Borghi F. Garulli G. Baldazzi G. Marini P. Sica G. ERAS program adherence-institutionalization, major morbidity and anastomotic leakage after elective colorectal surgery: The iCral2 multicenter prospective study Surg. Endosc. 2022 36 3965 3984 10.1007/s00464-021-08717-2 34519893 10. The Italian Colorectal Anastomotic Leakage (iCral) Study Group Patient-reported outcomes and return to intended oncologic therapy after colorectal enhanced recovery pathway: The iCral3 prospective study Ann. Surg. Open 2023 in press 11. Gustafsson U.O. Scott M.J. Hubner M. Nygren J. Demartines N. Francis N. Rockall N.A. Young-Fadok T.M. Hill A.G. Soop M. Guidelines for perioperative care in elective colorectal surgery: Enhanced Recovery After Surgery (ERAS(r)) Society recommendations: 2018 World J. Surg. 2019 43 659 695 10.1007/s00268-018-4844-y 30426190 12. Ficari F. Borghi F. Catarci M. Scatizzi M. Alagna V. Bachini I. Baldazzi G. Bardi U. Benedetti M. Beretta L. Enhanced recovery pathways in colorectal surgery: A consensus paper by the Associazione Chirurghi Ospedalieri Italiani (ACOI) and the PeriOperative Italian Society (POIS) Il Giornale di Chirurgia 2019 40 (Suppl. 4) 1 40 32003714 13. Kaiser M.J. Bauer J.M. Ramsch C. Uter W. Guigoz Y. Cederholm T. Thomas D.R. Anthony P. Charlton K.E. Maggio M. Validation of the Mini Nutritional Assessment short-form (MNA-SF): A practical tool for identification of nutritional status J. Nutr. Health Aging 2009 13 782 788 10.1007/s12603-009-0214-7 19812868 14. Clavien P.A. Barkun J. de Oliveira M.L. Vauthey J.N. Dindo D. Schulick R.D. De Santibanes E. Pekolj J. Slankamenac K. Bassi C. The Clavien-Dindo classification of surgical complications: Five-year experience Ann. Surg. 2009 250 187 196 10.1097/SLA.0b013e3181b13ca2 19638912 15. Katayama H. Kurokawa Y. Nakamura K. Ito H. Kanemitsu Y. Masuda N. Tsubosa Y. Satoh T. Yokomizo A. Fukuda H. Extended Clavien-Dindo classification of surgical complications: Japan Clinical Oncology Group postoperative complications criteria Surg. Today 2016 46 668 685 10.1007/s00595-015-1236-x 26289837 16. Rahbari N.N. Weitz J. Hohenberger W. Heald R.J. Moran B. Ulrich A. Holm T. Wong W.D. Tiret E. Moriya Y. Definition and grading of anastomotic leakage following anterior resection of the rectum: A proposal by the International Study Group of Rectal Cancer Surgery 2010 147 339 351 10.1016/j.surg.2009.10.012 20004450 17. Von Elm E. Altman D.G. Egger M. Pocock S.J. Gotzsche P.C. Vandenbroucke J.P. STROBE Initiative The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies J. Clin. Epidemiol. 2008 61 344 349 10.1016/j.jclinepi.2007.11.008 18313558 18. Yao X.I. Wang X. Speicher P.J. Hwang E.S. Cheng P. Harpole D.H. Berry M.F. Schrag D. Pang H.H. Reporting and Guidelines in Propensity Score Analysis: A Systematic Review of Cancer and Cancer Surgical Studies J. Natl. Cancer Inst. 2017 109 djw323 10.1093/jnci/djw323 28376195 19. Austin P.C. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies Multivar. Behav. Res. 2011 46 399 424 10.1080/00273171.2011.568786 21818162 20. Rosenbaum P.R. Rubin D.B. The Central Role of the Propensity Score in Observational Studies for Causal Effects Biometrika 1983 70 41 55 10.1093/biomet/70.1.41 21. Brookhart M.A. Schneeweiss S. Rothman K.J. Glynn R.J. Avorn J. Sturmer T. Variable selection for propensity score models Am. J. Epidemiol. 2006 163 1149 1156 10.1093/aje/kwj149 16624967 22. Austin P.C. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies Pharm. Stat. 2011 10 150 161 10.1002/pst.433 20925139 23. Rosenbaum P.R. The power of a sensitivity analysis and its limit Design of Observational Studies 2nd ed. Springer Series in Statistics Springer Nature Switzerland AG Cham, Switzerland 2020 317 336 24. Carson J.L. Stanworth S.J. Dennis J.A. Trivella M. Roubinian N. Fergusson D.A. Triulzi D. Doree C. Hebert P.C. Transfusion thresholds for guiding red blood cell transfusion Cochrane Database Syst. Rev. 2021 12 CD002042 34932836 25. Opelz G. Sengar D.P. Mickey M.R. Terasaki P.I. Effect of blood transfusions on subsequent kidney transplants Transplant. Proc. 1973 5 253 259 4572098 26. Vamvakas E.C. Blajchman M.A. Transfusion-related immunomodulation (TRIM): An update Blood Rev. 2007 21 327 348 10.1016/j.blre.2007.07.003 17804128 27. McSorley S.T. Tham A. Dolan R.D. Steele C.W. Ramsingh J. Roxburgh C. Horgan P.G. McMillan D.C. Perioperative Blood transfusion is associated with postoperative systemic inflammatory response and poorer outcomes following surgery for colorectal cancer Ann. Surg. Oncol. 2020 27 833 843 10.1245/s10434-019-07984-7 31664621 28. Kapur R. Kim M. Rebetz J. Hallstrom B. Bjorkman J.T. Takabe-French A. Kim N. Liu J. Shanmugabhavananthan S. Milosevic S. Gastrointestinal microbiota contributes to the development of murine transfusion-related acute lung injury Blood Adv. 2018 2 1651 1663 10.1182/bloodadvances.2018018903 29991496 29. D'Alessandro A. Zolla L. Proteomic analysis of red blood cells and the potential for the clinic: What have we learned so far? Expert Rev. Proteom. 2017 14 243 252 10.1080/14789450.2017.1291347 30. Goubran H. Seghatchian J. Radosevic J. Ragab G. Burnouf T. The microbiome and transfusion in cancer patients Transfus. Apher. Sci. 2017 56 330 335 10.1016/j.transci.2017.05.023 28633955 31. Garcia-Granero E. Navarro F. Cerdan Santacruz C. Frasson M. Garcia-Granero A. Marinello F. Flor-Lorente B. Espi A. Individual surgeon is an independent risk factor for leak after double-stapled colorectal anastomosis: An institutional analysis of 800 patients Surgery 2017 162 1006 1016 10.1016/j.surg.2017.05.023 28739093 32. Nordestgaard A.T. Rasmussen L.S. Sillesen M. Steinmetz J. Eid A.I. Meier K. Kaafarani H.M.A. Velmahos G.C. Red blood cell transfusion in surgery: An observational study of the trends in the USA from 2011 to 2016 Anaesthesia 2020 75 455 463 10.1111/anae.14900 31667830 33. Aquina C.T. Blumberg N. Probst C.P. Becerra A.Z. Hensley B.J. Noyes K. Monson J.R. Fleming F.J. Large Variation in Blood Transfusion Use After Colorectal Resection: A Call to Action Dis. Colon Rectum 2016 59 411 418 10.1097/DCR.0000000000000588 27050603 34. Pareto V. Cours d'Economie Politique Professe a l'Universite de Lausanne F. Rouge Lausanne, Switzerland 1896 Volume I 35. Juran J.J. Quality Control Handbook McGraw-Hill New York, NY, USA 1951 36. Hofmann A. Farmer S. Towler S.C. Strategies to preempt and reduce the use of blood products: An Australian perspective Curr. Opin. Anaesthesiol. 2012 25 66 73 10.1097/ACO.0b013e32834eb726 22157195 37. Isbister J.P. The three-pillar matrix of patient blood management ISBT Sci. Ser. 2015 10 (Suppl. 1) 286 294 10.1111/voxs.12135 38. World Health Organization The Urgent Need to Implement Patient Blood Management: Policy Brief Electronic Version World Health Organization Geneva, Switzerland 2021 978-92-4-003574-4 39. Catarci M. Borghi F. Ficari F. Scatizzi M. Perioperative anemia and its implications II Giornale di Chirurgia--JISA 2022 42 e01 10.1097/IA9.0000000000000001 40. Shin S.H. Piozzi G.N. Kwak J.M. Baek S.J. Kim J. Kim S.H. Effect of a Patient Blood Management system on perioperative transfusion practice and short-term outcomes of colorectal cancer surgery Blood Transfus. 2022 20 475 482 35848631 41. Italian ColoRectal Anastomotic Leakage (iCral) Study Group Enhanced Recovery and Patient Blood Management in Colorectal Surgery: The Italian ColoRectal Anastomotic Leakage Study Group (iCral 4). NCT05227014; 2022; ClinicalTrials.gov--NIH--US National Library of Medicine Available online: (accessed on 28 December 2022) 42. Silverman S.L. From randomized controlled trials to observational studies Am. J. Med. 2009 122 114 120 10.1016/j.amjmed.2008.09.030 19185083 43. Ioannidis J.P.A. Greenland S. Hlatky M.A. Khoury M.J. Macleod M.R. Moher D. Schulz K.F. Tibshirani R. Increasing value and reducing waste in research design, conduct, and analysis Lancet 2014 383 166 175 10.1016/S0140-6736(13)62227-8 24411645 44. Kim Y. Amini N. Gani F. Wagner D. Johnson D.J. Scott A. Ejaz A. Margonis G.A. Xu L. Buettner S. Age of transfused blood impacts perioperative outcomes among patients who undergo major gastrointestinal surgery Ann. Surg. 2017 265 103 110 10.1097/SLA.0000000000001647 28009733 45. Halabi W.J. Jafari M.D. Nguyen V.Q. Carmichael J.C. Mills S. Pigazzi A. Stamos M.J. Blood transfusions in colorectal cancer surgery: Incidence, outcomes, and predictive factors: An American College of Surgeons National Surgical Quality Improvement Program analysis Am. J. Surg. 2013 206 1024 1032 10.1016/j.amjsurg.2013.10.001 24296103 46. Quinn E.M. Meland E. McGinn S. Anderson J.H. Correction of iron-deficiency anaemia in colorectal surgery reduces perioperative transfusion rates: A before and after study Int. J. Surg. 2017 38 1 8 10.1016/j.ijsu.2016.12.029 28011177 47. Khalafallah A.A. Yan C. Al-Badri R. Robinson E. Kirkby B.E. Ingram E. Gray Z. Khelgi V. Robertson I.K. Kirkby B.P. Intravenous ferric carboxymaltose versus standard care in the management of postoperative anaemia: A prospective, open-label, randomised controlled trial Lancet Haematol. 2016 3 e415 e425 10.1016/S2352-3026(16)30078-3 27570088 48. Kleinman S. Busch M.P. Murphy E.L. Shan H. Ness P. Glynn S.A. National Heart, Lung, and Blood Institute Recipient Epidemiology and Donor Evaluation Study (REDS-III) The National Heart, Lung, and Blood Institute Recipient Epidemiology and Donor Evaluation Study (REDS-III): A research program striving to improve blood donor and transfusion recipient outcomes Transfusion 2014 54 942 955 10.1111/trf.12468 24188564 |
PMC10000588 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050935 diagnostics-13-00935 Article Bronchoalveolar Lavage Cell Count and Lymphocytosis Are the Important Discriminators between Fibrotic Hypersensitivity Pneumonitis and Idiopathic Pulmonary Fibrosis Sobiecka Malgorzata 1* Szturmowicz Monika 1 Lewandowska Katarzyna B. 1 Baranska Inga 2 Zimna Katarzyna 1 Lyzwa Ewa 1 Dybowska Malgorzata 1 Langfort Renata 3 Radwan-Rohrenschef Piotr 1 Rozy Adriana 4 Tomkowski Witold Z. 1 Grenier Philippe A. Academic Editor 1 1st Department of Lung Diseases, National Tuberculosis and Lung Diseases Research Institute, Plocka 26, 01-138 Warsaw, Poland 2 Department of Radiology, National Tuberculosis and Lung Diseases Research Institute, Plocka 26, 01-138 Warsaw, Poland 3 Department of Pathology, National Tuberculosis and Lung Diseases Research Institute, Plocka 26, 01-138 Warsaw, Poland 4 Department of Genetics and Clinical Immunology, National Tuberculosis and Lung Diseases Research Institute, Plocka 26, 01-138 Warsaw, Poland * Correspondence: [email protected] 01 3 2023 3 2023 13 5 93530 1 2023 21 2 2023 27 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Background: Fibrotic hypersensitivity pneumonitis (fHP) shares many features with other fibrotic interstitial lung diseases (ILD), and as a result it can be misdiagnosed as idiopathic pulmonary fibrosis (IPF). We aimed to determine the value of bronchoalveolar lavage (BAL) total cell count (TCC) and lymphocytosis in distinguishing fHP and IPF and to evaluate the best cut-off points discriminating these two fibrotic ILD. Methods: A retrospective cohort study of fHP and IPF patients diagnosed between 2005 and 2018 was conducted. Logistic regression was used to evaluate the diagnostic utility of clinical parameters in differentiating between fHP and IPF. Based on the ROC analysis, BAL parameters were evaluated for their diagnostic performance, and optimal diagnostic cut-offs were established. Results: A total of 136 patients (65 fHP and 71 IPF) were included (mean age 54.97 +- 10.87 vs. 64.00 +- 7.18 years, respectively). BAL TCC and the percentage of lymphocytes were significantly higher in fHP compared to IPF (p < 0.001). BAL lymphocytosis >30% was found in 60% of fHP patients and none of the patients with IPF. The logistic regression revealed that younger age, never smoker status, identified exposure, lower FEV1, higher BAL TCC and higher BAL lymphocytosis increased the probability of fibrotic HP diagnosis. The lymphocytosis >20% increased by 25 times the odds of fibrotic HP diagnosis. The optimal cut-off values to differentiate fibrotic HP from IPF were 15 x 106 for TCC and 21% for BAL lymphocytosis with AUC 0.69 and 0.84, respectively. Conclusions: Increased cellularity and lymphocytosis in BAL persist despite lung fibrosis in HP patients and may be used as important discriminators between IPF and fHP. bronchoalveolar lavage lymphocytosis bronchoalveolar lavage cell count fibrotic hypersensitivity pneumonitis idiopathic pulmonary fibrosis diagnosis This research received no external funding. pmc1. Introduction Recently, the American Thoracic Society (ATS), Japanese Respiratory Society (JRS) and Association Latinoamericana del Torax (ALAT), as well as the American College of Chest Physicians (ACCP) have independently published practical clinical guidelines for the diagnosis of hypersensitivity pneumonitis (HP) . Both documents provide algorithms to establish the diagnosis in a patient with interstitial lung disease (ILD) suspected to have HP based on a combination of specific features grouped into three domains: 1. exposure identification, 2. high-resolution computed tomography (HRCT) pattern, and 3. bronchoalveolar lavage (BAL) lymphocytosis. Based on the ACCP guidelines, a confident diagnosis of HP can be made in a patient who has an inciting antigen identified and typical pattern on a HRCT scan, without the need for invasive procedures . According to the ATS/JRS/ALAT guidelines, a highly confident diagnosis of HP additionally requires the presence of lymphocytosis in the bronchoalveolar lavage fluid (BALF) equal to 30% or greater . HP is a complex, immune-mediated interstitial lung disease caused by the repeated inhalation of organic dust in susceptible individuals . The disorder shares the features of other acute/inflammatory or chronic/fibrotic pulmonary diseases. As a result, fibrotic hypersensitivity pneumonitis (fHP) can be misdiagnosed as idiopathic pulmonary fibrosis (IPF), especially when the inciting antigen has not been identified, despite a thorough history . Until recently, HP patients have been classified according to symptom chronicity in acute and chronic forms , but both of the last published guidelines propose that HP be simply classified as fibrotic or nonfibrotic based on the presence or absence of fibrosis on the HRCT of the chest and/or histopathological analysis . Based on new data, stratification according to the presence of fibrosis seems more in line with the prognosis and may have important diagnostic and therapeutic implications . The most frequent and challenging diagnostic dilemma encountered by ILD-experienced clinicians is the differentiation of fibrotic HP from IPF due to overlapping symptoms, HRCT pattern, and histological findings. The discrimination between fibrotic HP and IPF represents a frequent diagnostic challenge even in the best tertiary lung disease centres. An integrated approach to the assessment of clinical features, radiological patterns, and bronchoalveolar lavage fluid and/or histopathological findings (as appropriate) should be applied according to the recent ATS/European Respiratory Society (ERS)/JRS/ALAT guidelines for IPF and ATS/JRS/ALAT or ACCP guidelines for fHP . A thorough medical history of exposure and evaluation of HRCT images by an ILD-experienced radiologist are crucial in the differential diagnosis of fHP and IPF. However, in many cases of fHP, an exposure is not identified (up to approximately 50% of cases). In addition, lung fibrosis on HRCT, not accompanied by the radiologic features of active HP, such as the centrilobular nodules or the areas of grand-glass opacities, may be difficult to differentiate from IPF. On such occasions, the coexistence of fibrosis and air trapping and the distribution of the lesions (with no basal predominance of the lesion) may determine the suspicion of fHP . Thus, only fair agreement on a diagnosis of HP across multidisciplinary discussion groups has been demonstrated, while for other ILDs, the agreement was good . BALF analysis is frequently used in patients with newly identified ILD, as a low-risk procedure that can narrow the differential diagnosis, and in some cases, may eliminate the need for a lung biopsy . Increased cellularity with lymphocytosis is associated with HP and the threshold of 30% for lymphocytes in BAL has been proposed to be reasonable in distinguishing HP from non-HP ILD . Re-evaluation of patients with different fibrotic ILD driven by BAL results as a complementary tool can lead to a change in diagnosis in some cases . However, the role of BALF cell count and lymphocytosis and their discriminative performance in the distinction between fHP and IPF needs to be clarified. We aimed to determine the frequency and magnitude of BALF lymphocytosis in patients with fHP and IPF, and to evaluate the best cut-off points for total cell count and percentage of lymphocytes discriminating these two fibrotic ILDs. We hypothesized that BALF lymphocytosis and total cell count may have a valuable complementary role when differentiating between fHP and IPF. 2. Materials and Methods 2.1. Ethical Approval The study was approved by the Bioethical Committee at the National Tuberculosis and Lung Diseases Research Institute (approval No. KB-14/2019 and KB-5/2022) and conducted in accordance with the Declaration of Helsinki. Patients' consent was waived by the Bioethical Committee because of the study's retrospective nature. All personal data were anonymized. The publication does not include any data or features enabling the identification of any individual patient in the analysis. 2.2. Study Population Patients with fibrotic hypersensitivity pneumonitis and idiopathic pulmonary fibrosis were identified retrospectively from a cohort of consecutive patients diagnosed with different ILDs in our tertiary referral centre between 2 January 2005, and 31 December 2018, and their initial diagnoses were re-evaluated by our multidisciplinary team. The diagnosis of fHP was based on the current ATS/JRS/ALAT guidelines through multidisciplinary discussion (MDD), following the integration of clinical data, exposure history, HRCT pattern, bronchoalveolar lavage, and histology, when available. IPF was diagnosed according to the 2018 ATS/ERS/JRS/ALAT statement and its update from 2022 . Only patients who underwent bronchoalveolar lavage in the diagnostic process were eligible for the study. 2.3. Data Collection and Procedures Data regarding age, gender, smoker status, history of exposure, HP precipitin serology, pulmonary function test, 6 min walk test distance and desaturation, and bronchoalveolar lavage fluid results were extracted from the hospital database. A detailed description of the diagnostic procedures used by our group has been published previously . Spirometry and the whole body plethysmography were performed as routine measures in all patients with an integrated measuring device Master Screen Body/Diffusion by Jaeger (Germany 2002), following the ERS/ATS recommendations and reported as percentages of predictive values according to the ERS reference equations . The transfer factor of the lungs for carbon monoxide (TL,co) was measured with a single breath method, with helium gas as a marker. The results were presented as a percentage of the predicted values with a correction to haemoglobin concentration . The six-minute walk test was performed on a corridor in accordance with ATS guidelines and the distance and desaturation at the sixth minute were noted. Bronchoalveolar lavage was performed for diagnostic purposes according to the ATS recommendations . The bronchoscope was placed in the wedge position in a subsegmental bronchus of the middle lobe. Then, warmed up to 37 Celsius degrees 0.9% saline was instilled through a working channel in 20 mL aliquots, up to a maximum of 200 mL. The recovered BAL fluid (at least 50% of infused volume) was filtered through sterile gauze, and centrifuged (4 degC, 400x g, 15 min). Cell viability was assessed by trypan blue exclusion. The total cell count was counted using a Burker chamber and differential cell count was evaluated in light microscopy on May-Grunwald-Giemsa stained slides by counting a minimum of 600 cells . None of the patients had received glucocorticosteroids or immunosuppressants prior to the BAL being performed. 2.4. Statistical Analysis Statistical analyses were performed in Stata 15.1. Two-sided a = 0.05 was used to determine statistical significance. The one-way ANOVA and chi-squared test were used to assess differences in descriptive statistics in continuous and categorical variables, respectively. Logistic regression was used to evaluate the potential diagnostic utility of clinical findings in differentiating fHP and IPF, adjusting for age, sex, and smoking status (never-smoker versus ever-smoker). BALF parameters with evidence of discriminatory performance were evaluated for their diagnostic performance based on the receiver operating characteristic curve (ROC), and positive and negative predictive values. Optimal diagnostic cut-offs were established based on maximising the product of sensitivity and specificity and they were rounded to the nearest integer . As a sensitivity analysis, multivariable logistic regression models were further adjusted for estimated pack-years of smoking (n participants = 127). p <= 0.05 was defined as a significant difference. 3. Results 3.1. Baseline Characteristics of the Study Group A total of 136 patients (65 with fibrotic hypersensitivity pneumonitis and 71 with idiopathic pulmonary fibrosis) were included in the study. The baseline clinical characteristics and pulmonary function test findings of the study population are summarized in Table 1 and Table 2, respectively. The patients in the fHP group were significantly younger, more often female and never-smokers than the IPF patients. An exposure history to various antigens was recognized in 72.3% of fHP patients and it was dominated by avian antigens. There were no significant differences in the results of six-minute walk distance, desaturation and pulmonary function tests between fHP and IPF patients, except for FEV1 and FVC, which were lower in the patients with fHP. 3.2. BALF Cell Analysis BALF cell analysis results and distribution of lymphocytosis by diagnosis are presented in Table 3. The total cell count and the percentage of lymphocytes were significantly higher in the fHP group compared to the IPF group (p < 0.001) . Among patients with fHP BALF lymphocytosis equal to or greater than 20% was observed in 75% of cases, while only in 10% of patients with IPF. In addition, BALF lymphocytosis above 30% was found in 60% of fHP patients, and none of the IPF patients. 3.3. Predictors of Fibrotic HP Diagnosis A logistic regression analysis adjusted for age, gender, and smoking status was performed to determine predictors of fHP diagnosis. The analysis revealed that younger age, never smoker status, identified exposure, lower FEV1 (in % predicted value), higher BALF cell count and lymphocytosis significantly increased the odds of fHP diagnosis rather than IPF. The BALF lymphocytosis exceeding 20% increased the odds of being classified as fHP rather than IPF by 25-fold (Table 4). The sensitivity analysis using the multivariable logistic regression models further adjusted for estimated pack-years, provided similar findings (Table S1). The optimal cut-off values to differentiate fibrotic HP from IPF were 15 x 106 for BALF cell count and 21% for lymphocytosis. With the use of calculated cut-off values, the sensitivity and specificity of BALF lymphocytosis for the recognition of fHP were 75% and 93%, respectively, and the corresponding values for BALF cell count, were 80% and 59%, respectively (Table 5). The receiver operating characteristic (ROC) curves illustrating the diagnostic utility of BALF total cell count and percentage of lymphocytes in differentiating fibrotic HP from IPF are shown on Figure 2 and Figure 3, respectively. 4. Discussion There are no randomised controlled trials or controlled observational studies evaluating the BALF lymphocyte percentage as a diagnostic test for fHP. On the other hand, many authors present the difficulties in the differential diagnosis between fHP and IPF in clinical practise. This is caused by a relatively large population of fHP patients, in whom the HRCT pattern is not typical of HP, or even presents the usual interstitial pneumonia (UIP)-like features . On such occasions, the MDD experts decide to perform invasive diagnostic procedures to ascertain the diagnosis . In our retrospective study, we found that not only lymphocytosis, but also total cell count in BALF is important in distinguishing fHP from IPF, with an AUC at 0.90 for lymphocytosis and at 0.71 for cell count. We also showed that the BALF lymphocytosis exceeding the 20% increased by twenty-five times the probability of being classified as fHP rather than IPF. In addition, in our study, it was demonstrated that a lymphocytosis equal to or higher than 21% and a total cell count of 15 x 106 in BALF would be appropriate cut-offs to consider fHP as a diagnosis, with a sensitivity of 75% and 80%, and with a specificity of 93% and 59%, respectively, for discriminating fHP from IPF. In the diagnostic evaluation of patients with fibrotic ILD, the distinction between fHP and IPF can be challenging even for experienced clinicians, in particular when the exposure to the antigen has been hidden or forgotten . Nevertheless, this differentiation is crucial for disease management and prognostication, especially in the era of antifibrotic treatment. The avoidance of further exposure to the identified antigen and consideration of immunosuppressive therapy remain essential in the management of patients with fHP, while the prompt initiation of antifibrotic treatment is essential in IPF . The primary goal in the diagnosis of ILD is to make a confident diagnosis using the least invasive approach, considering that these patients are often elderly and have some comorbidities. Hence, the usefulness of BALF cellular analysis, including lymphocytosis in the differentiation of fHP from IPF and other fibrotic ILDs, as a minimally invasive method, is a subject of debate in the literature. Two recent systemic reviews and meta-analyses have pooled the data on the value of BALF lymphocytosis in the diagnosis of HP and found higher lymphocyte percentages in chronic/fibrotic HP compared to IPF . In a meta-analysis of 42 studies, Adderley et al. demonstrated that the pooled estimate for the BALF lymphocyte percentage in chronic HP was 43% compared to 10% of lymphocytes in IPF . Importantly, a similar pooled estimate of 44% for BALF lymphocytes was provided when the sensitivity analysis using 26 studies that defined chronic HP based on signs of fibrosis on HRCT and/or lung biopsy. Additionally, the authors chose to use individual patient data from eight studies, pooled and analysed as a single cohort, to calculate the performance characteristics of BALF at different lymphocyte percentage thresholds to discriminate chronic HP from IPF/non-IPF idiopathic interstitial pneumonia (IIP). The thresholds that maximised sensitivity and specificity were 20% and 50%, respectively. If the cut-off level of BALF lymphocyte percentage was set at 20%, the positive predictive value was only 57% for suggesting chronic HP vs. IPF/non-IPF IIP. On the other hand, if the cut-off level was set high at 50%, fewer subjects with chronic HP would be captured but with minimising false-positives (PPV increased to 78%). The BALF lymphocytosis value that concurrently optimised sensitivity (70.7%) and specificity (67.6%) was 21% . Similarly, in a meta-analysis of 36 studies used to inform the ATS/JRS/ALAT guidelines specifically, Patolia et al. reported comparable sensitivity (69%) and specificity (61%) for a lymphocytosis threshold of 20% distinguishing fHP from IPF . However, despite a relatively high mean difference of 21% in the BALF lymphocyte percentage (95% confidence interval, 14-27%) of fHP versus IPF patients, the area under the receiver-operating-characteristic curve (AUC) was only 0.54. After checking the results the suggested primary explanation was the high standard deviation (SD) observed for the BALF lymphocyte percentage in both fHP and IPF populations within many studies. A threshold of 40% greatly increased specificity to 93%, but markedly reduced sensitivity to 41%, whereas a threshold of 30% gave a specificity of 80% and the sensitivity of 55% in distinguishing fHP from IPF . In our study, the mean difference in the BALF lymphocyte percentage between fHP and IPF was slightly higher (25%), and the AUC value was excellent (0.90) for lymphocytosis and fair (0.71) for total cell count. The ATS/JRS/ALAT guideline committee did not identify a threshold proportion of BALF lymphocytes that distinguishes HP from non-HP ILD, given the poor area under the curve for the comparisons. Based on the committee's collective clinical experience, the 30% threshold for lymphocytosis in BALF has been considered to be reasonable . In contrast, in a recent Delphi online survey involving 45 ILD experts from 14 countries, the vast majority of experts rated BALF lymphocytosis more than 40% as "important" for the diagnosis of chronic HP, while levels between 30 and 39% were considered a grey zone (did not meet consensus), and values between 20 and 29% as unhelpful . In turn, the absence of BALF lymphocytosis supports the diagnosis of IPF . Ohshimo et al. were the first to demonstrate that the cut off level of <30% for lymphocytes in BALF had a favourable discriminative power for the diagnosis of IPF . Tzilas et al. re-evaluated the initial diagnoses of patients with fibrotic ILD and indeterminate for UIP HRCT pattern during multidisciplinary team discussion. The authors reported that a BALF lymphocytosis of >=20% was of added diagnostic value in their retrospective cohort of undiagnosed fibrotic ILD, changing the diagnosis from IPF to HP in 15% of the cases . They also stated that even a mild BALF lymphocytosis (>20-25%) should increase vigilance for the search of an underlying inciting antigen. In our group, the percentage of lymphocytes in the BALF did not exceed the threshold of 30% in any of the patients with IPF, and only in 10% of the patients with IPF was the lymphocytosis in the BALF higher than 20%. On the other hand, among patients with fHP, the threshold of 30%, considered reasonable by the ATS/JRS/ALAT guidelines committee, was exceeded by 60% of patients, while the threshold of 40% proposed by the DELPHI survey was exceeded by only 42% of patients . Very little data exist on the total cell count in BAL fluid. Domagala-Kulawik et al. reported that elevated total cell count in the BALF is common in patients with ILD . According to the authors, the risk of HP diagnosis increased by 37% per one million of BALF cells compared to a control group of healthy volunteers. Bergantini et al. showed that the total cell count was significantly higher in the fibrotic HP group than in the connective tissue disease (CTD)-ILD and cryptogenic organizing pneumonia (COP) groups . However, we have shown that both the total cell count and the percentage of lymphocytes were significantly higher in the fHP group compared to the IPF group. The probability of fHP diagnosis increased in our study by 4% per one million of BALF cells. Including in the diagnostic process of fibrotic HP not only the percentage of lymphocytes, but also the total number of cells could be an additional factor supporting the diagnosis. The above-mentioned systematic reviews and meta-analyses included only observational studies reporting on the BALF lymphocyte percentage in patients with HP, as no randomised controlled trials or controlled observational studies exist on this topic . What is important is that prior studies assessing the diagnostic utility of BALF lymphocytosis in HP may be limited by varied diagnostic criteria. Additional confounding may also involve the unclear differentiation of fibrotic vs. nonfibrotic HP with the use of prior classification schemes (acute, subacute, and chronic), where BALF lymphocytosis findings may intrinsically vary. Our results are derived from a cohort of well-characterized fHP and IPF patients, in which the diagnosis was re-evaluated according to the latest international guidelines . BALF total and differential cell count may be influenced by several confounding factors such as age, smoking status, systemic corticosteroid therapy, the presence and extent of fibrosis and duration of symptoms . According to the Domagala-Kulawik et al., cigarette smoking caused approximately two-fold increase in the BALF total cell count (TCC) when healthy smokers and non-smokers have been compared . Additionally, the authors observed the higher mean TCC values in ILD patients than in healthy smokers. Adderley et al., using individual patient data from eight studies, showed that older age and ever smoking have been associated with lower BALF lymphocyte percentages in patients with chronic HP . On the other hand, an increase in the TCC with a significant elevation in the BALF lymphocyte percentage (>30%) is considered to play a key role in distinguishing fHP from IPF . Hence, some authors have suggested using the threshold of 20% for BALF lymphocytosis in smoking HP patients . In our study, the patients with fHP were significantly younger, more often female and never-smokers in comparison to IPF patients. Given that age and smoking were major potential confounding factors, we used them as covariates in our multivariable analyses, thus eliminating or minimising any bias that they may have had on our results. Additionally, the sensitivity analysis using the multivariable logistic regression models further adjusted for estimated pack-years of smoking, provided similar findings (Table S1 in Supplementary Material). Our study has some limitations. First, an obvious limitation is the retrospective, single-centre nature of the study. On the other hand, BAL is a procedure performed in our centre by experienced staff for several decades, always according to international guidelines. Hence, the impact of the BAL methodology on the results, as may be seen in a multicentre study, was limited. Second, a relatively small number of patients were included in the study because we focused on IPF and fHP patients who underwent a BAL procedure. Finally, an essential limitation of the study was incorporation bias, since the presence of BALF lymphocytosis was considered one of the criteria to diagnose fHP. However, BALF lymphocytosis was not the only test to confirm or rule out the HP diagnosis. A significant proportion of our patients with fHP had TBLB performed simultaneously with BAL and subsequently, in doubtful cases, cryobiopsy or surgical lung biopsy, based on which the diagnosis could be established without taking the BAL results into account. 5. Conclusions The percentage of lymphocytes and the total cell count in the BALF are important in separating fHP from IPF. Increased cellularity with lymphocytosis in BALF persists despite lung fibrosis in HP and may be an additional value in improving the diagnostic likelihood of fHP and allowing us to avoid more invasive tests such as lung biopsy. A BALF lymphocytosis exceeding the threshold of 20% increases the probability of fHP diagnosis by twenty-five times compared to IPF. Therefore, even a mild BALF lymphocytosis should increase diagnostic vigilance and call attention to a more thorough search for a causative antigen or history of exposure. Acknowledgments The authors are grateful for the help of Jakub Sobiecki, who assisted with the statistical analyses. Supplementary Materials The following supporting information can be downloaded at: Table S1: Clinical predictors of fibrotic hypersensitivity pneumonitis vs. idiopathic pulmonary fibrosis diagnosis (multivariable logistic regression analysis# further adjusted for estimated pack-years). Click here for additional data file. Author Contributions M.S. (Malgorzata Sobiecka), conceptualization, visualization, formal analysis, data curation, writing--original draft preparation, M.S. (Monika Szturmowicz), conceptualization, writing--review and editing, supervision, K.B.L., writing--review and editing, I.B., radiological scans assessment, writing--review and editing, K.Z., investigation, writing--review and editing, E.L., investigation, writing--review and editing, M.D., writing--review and editing, R.L., writing--review and editing, P.R.-R., BAL performing, writing--review and editing, A.R., BALF analysis, writing--review and editing, W.Z.T., writing--review and editing, supervision. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Ethics Committee National Tuberculosis and Lung Diseases research Institute (protocol code KB-14/2019 and date of approval 27 February 2019, and KB-5/2022 and date of approval 13 April 2022). Informed Consent Statement Patient consent was waived by the Ethics Committee because of the study's retrospective nature. All personal data were anonymized. The publication does not include any data or features enabling the identification of any individual patient in the analysis. Data Availability Statement The data are available from the corresponding author upon request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Bronchoalveolar lavage fluid results in relation to diagnosis fibrotic hypersensitivity pneumonitis (fHP) and idiopathic pulmonary fibrosis (IPF). (a) The percentage of BALF lymphocytes and (b) the BALF total cell count in both groups. Figure 2 Receiver operating characteristic (ROC) curve for diagnostic utility of total cell count in differentiating between fHP (n = 65) and IPF (n = 71). Dots indicate individual data points used to construct the graph and calculate the value of the area under the ROC curve. Figure 3 Receiver operating characteristic (ROC) curve for diagnostic utility of bronchoalveolar lavage fluid lymphocytosis in differentiating between fHP (n = 65) and IPF (n = 71). Dots indicate individual data points used to construct the graph and calculate the value of the area under the ROC curve. diagnostics-13-00935-t001_Table 1 Table 1 Baseline characteristics of the patients with fibrotic hypersensitivity pneumonitis and idiopathic pulmonary fibrosis. Parameter fHP IPF p-Value Subjects 65 71 Age at diagnosis, years 54.97 (10.87) 64.00 (7.18) <0.001 Sex 0.012 Male 32 (49.2) 50 (70.4) Female 33 (50.8) 21 (29.6) Smoking status <0.001 Smokers and ex-smokers 26 (40.0) 55 (77.5) Never-smokers 39 (60.0) 16 (22.5) Pack-years 8.75 (16.68) 35.07 (22.08) <0.001 Identified exposure to: 47 (72.3) 8 (11.3) <0.001 Poultry 14 2 Pigeons 16 1 Parrots 4 4 Hay/feed 15 1 Other 10 0 Data are presented as n, mean +- SD or n (%), unless otherwise stated, p-values were calculated using one-way ANOVA and chi-squared test; fHP: fibrotic hypersensitivity pneumonitis; IPF: idiopathic pulmonary fibrosis. diagnostics-13-00935-t002_Table 2 Table 2 Baseline pulmonary function test and echo results of the study population. Parameter fHP n = 65 IPF n = 71 p-Value FVC, % predicted 77.85 (19.29) 87.14 (17.69) 0.004 FEV1, %predicted 73.92 (19.12) 89.55 (17.02) <0.001 FEV1/FVC 80.66 (8.36) 80.80 (8.85) 0.924 TLC, %predicted 79.67 (15.81) 84.56 (16.85) 0.088 TL,co, %predicted 47.20 (16.25) 52.29 (16.00) 0.071 6MWD, m 496.00 (92.67) 465.94 (111.12) 0.092 Desaturation during 6MWT, % 7.83 (5.91) 6.38 (5.89) 0.157 TVPG, mmHg 31.02 (11.22) 31.45 (7.72) 0.815 Data are presented as mean +- SD, unless otherwise stated, p-values were calculated using one-way ANOVA; fHP: fibrotic hypersensitivity pneumonitis; IPF: idiopathic pulmonary fibrosis; FVC: forced vital capacity, FEV1: forced expiratory volume in one second, FEV1/FVC: forced expiratory volume in one second to forced vital capacity ratio; TLC: total lung capacity, TL,co: transfer factor of the lungs for carbon monoxide, TVPG: tricuspid valve transvalvular pressure gradient; 6MWD: 6-min Walk Distance; 6MWT: 6-min Walk Test. diagnostics-13-00935-t003_Table 3 Table 3 Bronchoalveolar lavage fluid percent differential and lymphocytosis distribution in the study group. Parameter fHP n = 65 IPF n = 71 p-Value Total cell count (x106/mL) 26.23 (15.25) 16.20 (10.38) <0.001 Lymphocytes (%) 35.24 (17.50) 10.17 (6.80) <0.001 Eosinophils (%) 3.11 (3.68) 3.81 (4.30) 0.323 Neutrophils (%) 7.73 (6.93) 6.74 (6.11) 0.379 Macrophages (%) 54.51 (17.47) 79.28 (10.01) <0.001 Lymphocytosis >= 20%, N (%) 49 (75.4) 7 (9.9) <0.001 BALF lymphocytosis, N (%) <10% 8 (12) 44 (62) - 10-20% 8 (12) 21 (30) - 21-30% 10 (15) 6 (8) - 31-40% 12 (18) 0 - 41-50% 15 (23) 0 - 51-60% 7 (11) 0 - >60% 5 (8) 0 - Data are presented as mean +- SD or N (%), unless otherwise stated, p-values were calculated using one-way ANOVA and chi-squared test; BALF: bronchoalveolar lavage fluid; fHP: fibrotic hypersensitivity pneumonitis; IPF: idiopathic pulmonary fibrosis. diagnostics-13-00935-t004_Table 4 Table 4 Clinical predictors of fibrotic hypersensitivity pneumonitis vs. idiopathic pulmonary fibrosis diagnosis (logistic regression analysis #). Characteristics OR 95% CI p-Value Male 0.90 0.36-2.22 0.814 Age at diagnosis 0.90 0.86-0.94 <0.001 Ever smoker 0.20 0.08-0.49 <0.001 Identified exposure 17.21 5.93-50.00 <0.001 FVC, % predicted 0.99 0.97-1.01 0.424 FEV1, %predicted 0.97 0.95-0.99 0.018 FEV1/FVC 0.97 0.93-1.02 0.239 TLC, %predicted 0.99 0.96-1.02 0.423 TL,co, %predicted 0.97 0.95-1.00 0.068 6MWD, m 1.00 1.00-1.00 0.885 Desaturation during 6MWT, % 1.04 0.97-1.12 0.237 Total cell count in BALF 1.04 1.01-1.08 0.020 Neutrophils in BALF 1.03 0.97-1.10 0.340 Eosinophils in BALF 0.98 0.89-1.08 0.692 Lymphocytes in BALF 1.16 1.09-1.23 <0.001 Lymphocytosis in BALF > 20% 25.06 7.43-84.50 <0.001 # Adjusted for a priori age, sex and smoking history. BALF: bronchoalveolar lavage fluid; FVC: forced vital capacity, FEV1: forced expiratory volume in one second, FEV1/FVC: forced expiratory volume in one second to forced vital capacity ratio; TLC: total lung capacity, TL,co: transfer factor of the lungs for carbon monoxide; 6MWD: 6-min Walk Distance; 6MWT: 6-min Walk Test. diagnostics-13-00935-t005_Table 5 Table 5 Optimal cut-off values of lymphocytosis and total cell count in BALF for differentiation of fibrotic hypersensitivity pneumonitis from idiopathic pulmonary fibrosis. Parameter Cut-off Specificity Sensitivity PPV NPV AUC Lymphocytosis in BALF (%) 21 0.93 0.75 0.91 0.80 0.84 Total cell count in BALF (x106) 15 0.59 0.80 0.62 0.77 0.69 BALF: bronchoalveolar lavage fluid, PPV: positive predictive value; NPV: negative predictive value; AUC: area under the receiver operating characteristic (ROC) curve. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Fernandez Perez E.R. Travis W.D. Lynch D.A. Brown K.K. Johannson K.A. Selman M. Ryu J.H. Wells A.U. Tony Huang Y.-C. Pereira C.A.C. Diagnosis and Evaluation of Hypersensitivity Pneumonitis: CHEST Guideline and Expert Panel Report Chest 2021 160 e97 e156 10.1016/j.chest.2021.03.066 33861992 2. Raghu G. Remy-Jardin M. Ryerson C.J. Myers J.L. Kreuter M. Vasakova M. Bargagli E. Chung J.H. Collins B.F. Bendstrup E. Diagnosis of Hypersensitivity Pneumonitis in Adults. An Official ATS/JRS/ALAT Clinical Practice Guideline Am. J. Respir. Crit. Care Med. 2020 202 e36 e69 10.1164/rccm.202005-2032ST 32706311 3. Lacasse Y. Selman M. Costabel U. Dalphin J.-C. Ando M. Morell F. Erkinjuntti-Pekkanen R. Muller N. Colby T.V. Schuyler M. Clinical diagnosis of hypersensitivity pneumonitis Am. J. Respir. Crit. Care Med. 2003 168 952 958 10.1164/rccm.200301-137OC 12842854 4. Morell F. Villar A. Montero M.-A. Munoz X. Colby T.V. Pipvath S. Cruz M.-J. Raghu G. Chronic hypersensitivity pneumonitis in patients diagnosed with idiopathic pulmonary fibrosis: A prospective case-cohort study Lancet Respir. Med. 2013 1 685 694 10.1016/S2213-2600(13)70191-7 24429272 5. Vasakova M. Morell F. Walsh S. Leslie K. Raghu G. Hypersensitivity Pneumonitis: Perspectives in Diagnosis and Management Am. J. Respir. Crit. Care Med. 2017 196 680 689 10.1164/rccm.201611-2201PP 28598197 6. Raghu G. Remy-Jardin M. Myers J.L. Richeldi L. Ryerson C.J. Lederer D.J. Behr J. Cottin V. Danoff S.K. Morell F. Diagnosis of Idiopathic Pulmonary Fibrosis. An Official ATS/ERS/JRS/ALAT Clinical Practice Guideline Am. J. Respir. Crit. Care Med. 2018 198 e44 e68 10.1164/rccm.201807-1255ST 30168753 7. Walsh S.L.F. Wells A.U. Desai S.R. Poletti V. Piciucchi S. Dubini A. Nunes H. Valeyre D. Brillet P.Y. Kambouchner M. Multicentre evaluation of multidisciplinary team meeting agreement on diagnosis in diffuse parenchymal lung disease: A case-cohort study Lancet Respir. Med. 2016 4 557 565 10.1016/S2213-2600(16)30033-9 27180021 8. Meyer K.C. Raghu G. Baughman R.P. Brown K.K. Costabel U. du Bois R.M. Drent M. Haslam P.L. Kim D.S. Nagai S. An official American Thoracic Society clinical practice guideline: The clinical utility of bronchoalveolar lavage cellular analysis in interstitial lung disease Am. J. Respir. Crit. Care Med. 2012 185 1004 1014 10.1164/rccm.201202-0320ST 22550210 9. Reynolds H.Y. Bronchoalveolar lavage and other methods to define the human respiratory tract milieu in health and disease Lung 2011 189 87 99 10.1007/s00408-011-9284-5 21350888 10. Caillaud D.M. Vergnon J.M. Madroszyk A. Melloni B.M. Murris M. Dalphin J.C. French Group of Environmental Immunoallergic Bronchopulmonary Diseases Bronchoalveolar lavage in hypersensitivity pneumonitis: A series of 139 patients Inflamm. Allergy-Drug Targets 2012 11 15 19 10.2174/187152812798889330 22309080 11. Tzilas V. Tzouvelekis A. Bouros E. Karampitsakos T. Ntasiou M. Katsaras M. Costabel U. Wells A. Bouros D. Diagnostic value of BAL lymphocytosis in patients with indeterminate for usual interstitial pneumonia imaging pattern Eur. Respir. J. 2019 54 1901144 10.1183/13993003.01144-2019 31320457 12. Raghu G. Remy-Jardin M. Richeldi L. Thomson C.C. Inoue Y. Johkoh T. Kreuter M. Lynch D.A. Maher T.M. Martinez F.J. Idiopathic Pulmonary Fibrosis (an Update) and Progressive Pulmonary Fibrosis in Adults: An Official ATS/ERS/JRS/ALAT Clinical Practice Guideline Am. J. Respir. Crit. Care Med. 2022 205 e18 e47 10.1164/rccm.202202-0399ST 35486072 13. Szturmowicz M. Baranska I. Jedrych M.E. Bartoszuk I. Radwan-Roehrenschef P. Rozy A. Bestry I. Chorostowska-Wynimko J. Langfort R. Kus J. Hypersensitivity pneumonitis recognised in a single pulmonary unit, between 2005 and 2015--Comparison with recently proposed diagnostic criteria Adv. Respir. Med. 2019 87 83 89 10.5603/ARM.2019.0014 31038718 14. Szturmowicz M. Baranska I. Skoczylas A. Jedrych M.E. Demkow U. Correlation of bronchoalveolar lavage lymphocyte count with the extent of lung fibrosis and with plethysmographic lung volumes in patients with newly recognized hypersensitivity pneumonitis Cent.-Eur. J. Immunol. 2020 45 276 282 10.5114/ceji.2020.101246 33437179 15. Miller M.R. Hankinson J. Brusasco V. Burgos F. Casaburi R. Coates A. Crapo R. Enright P. van der Grinten C.P.M. Gustafsson P. Standardisation of spirometry Eur. Respir. J. 2005 26 319 338 10.1183/09031936.05.00034805 16055882 16. Wanger J. Clausen J.L. Coates A. Pedersen O.F. Brusasco V. Burgos F. Casaburi R. Crapo R. Enright P. van der Grinten C.P.M. Standardisation of the measurement of lung volumes Eur. Respir. J. 2005 26 511 522 10.1183/09031936.05.00035005 16135736 17. Quanjer P.H. Tammeling G.J. Cotes J.E. Pedersen O.F. Peslin R. Yernault J.C. Lung volumes and forced ventilatory flows Eur. Respir. J. 1993 6 (Suppl. 16) 5 40 10.1183/09041950.005s1693 18. Graham B.L. Brusasco V. Burgos F. Cooper B.G. Jensen R. Kendrick A. MacIntyre N.R. Thompson B.R. Wanger J. 2017 ERS/ATS standards for single-breath carbon monoxide uptake in the lung Eur. Respir. J. 2017 49 1600016 10.1183/13993003.00016-2016 28049168 19. ATS Committee on Proficiency Standards for Clinical Pulmonary Function Laboratories ATS statement: Guidelines for the six-minute walk test Am. J. Respir. Crit. Care Med. 2002 166 111 117 10.1164/ajrccm.166.1.at1102 12091180 20. Liu X. Classification accuracy and cut point selection Stat. Med. 2012 31 2676 2686 10.1002/sim.4509 22307964 21. Chiba S. Tsuchiya K. Akashi T. Ishizuka M. Okamoto T. Furusawa H. Tateishi T. Kishino M. Miyazaki Y. Tateishi U. Chronic Hypersensitivity Pneumonitis With a Usual Interstitial Pneumonia-Like Pattern: Correlation Between Histopathologic and Clinical Findings Chest 2016 149 1473 1481 10.1016/j.chest.2015.12.030 26836921 22. Morisset J. Johannson K.A. Vittinghoff E. Aravena C. Elicker B.M. Jones K.D. Fell C.D. Manganas H. Dube B.-P. Wolters P.J. Use of Mycophenolate Mofetil or Azathioprine for the Management of Chronic Hypersensitivity Pneumonitis Chest 2017 151 619 625 10.1016/j.chest.2016.10.029 27816444 23. Lewandowska K.B. Baranska I. Sobiecka M. Radwan-Rohrenschef P. Dybowska M. Franczuk M. Rozy A. Skoczylas A. Bestry I. Kus J. Factors Predictive for Immunomodulatory Therapy Response and Survival in Patients with Hypersensitivity Pneumonitis-Retrospective Cohort Analysis Diagnostics 2022 12 2767 10.3390/diagnostics12112767 36428827 24. Piotrowski W.J. Bestry I. Bialas A.J. Boros P.W. Grzanka P. Jassem E. Jastrzebski D. Klimczak D. Langfort R. Lewandowska K. Guidelines of the Polish Respiratory Society for diagnosis and treatment of idiopathic pulmonary fibrosis Adv. Respir. Med. 2020 88 41 93 10.5603/ARM.2020.0081 32153010 25. Adderley N. Humphreys C.J. Barnes H. Ley B. Premji Z.A. Johannson K.A. Bronchoalveolar lavage fluid lymphocytosis in chronic hypersensitivity pneumonitis: A systematic review and meta-analysis Eur. Respir. J. 2020 56 2000206 10.1183/13993003.00206-2020 32265306 26. Patolia S. Tamae Kakazu M. Chami H.A. Chua A. Diaz-Mendoza J. Duggal A. Jenkins A.R. Knight S.L. Raghu G. Wilson K.C. Bronchoalveolar Lavage Lymphocytes in the Diagnosis of Hypersensitivity Pneumonitis among Patients with Interstitial Lung Disease Ann. Am. Thorac. Soc. 2020 17 1455 1467 10.1513/AnnalsATS.202005-420OC 32757946 27. Morisset J. Johannson K.A. Jones K.D. Wolters P.J. Collard H.R. Walsh S.L.F. Ley B. HP Delphi Collaborators Identification of Diagnostic Criteria for Chronic Hypersensitivity Pneumonitis: An International Modified Delphi Survey Am. J. Respir. Crit. Care Med. 2018 197 1036 1044 10.1164/rccm.201710-1986OC 29172641 28. Nagai S. Handa T. Ito Y. Takeuchi M. Izumi T. Bronchoalveolar lavage in idiopathic interstitial lung diseases Semin. Respir. Crit. Care Med. 2007 28 496 503 10.1055/s-2007-991522 17975777 29. Ryu Y.J. Chung M.P. Han J. Kim T.S. Lee K.S. Chun E.-M. Kyung S.Y. Jeong S.H. Colby T.V. Kim H. Bronchoalveolar lavage in fibrotic idiopathic interstitial pneumonias Respir. Med. 2007 101 655 660 10.1016/j.rmed.2006.06.003 16919929 30. Ohshimo S. Bonella F. Cui A. Beume M. Kohno N. Guzman J. Costabel U. Significance of bronchoalveolar lavage for the diagnosis of idiopathic pulmonary fibrosis Am. J. Respir. Crit. Care Med. 2009 179 1043 1047 10.1164/rccm.200808-1313OC 19246718 31. Domagala-Kulawik J. Skirecki T. Maskey-Warzechowska M. Grubek-Jaworska H. Chazan R. Bronchoalveolar lavage total cell count in interstitial lung diseases--Does it matter? Inflammation 2012 35 803 809 10.1007/s10753-011-9378-5 21882075 32. Bergantini L. d'Alessandro M. Cameli P. Perrone A. Cekorja B. Boncompagni B. Mazzei M.A. Sestini P. Bargagli E. Integrated approach to bronchoalveolar lavage cytology to distinguish interstitial lung diseases Eur. J. Intern. Med. 2021 89 76 80 10.1016/j.ejim.2021.04.005 33931268 33. Salisbury M.L. Myers J.L. Belloli E.A. Kazerooni E.A. Martinez F.J. Flaherty K.R. Diagnosis and Treatment of Fibrotic Hypersensitivity Pneumonia. Where We Stand and Where We Need to Go Am. J. Respir. Crit. Care Med. 2017 196 690 699 10.1164/rccm.201608-1675PP 28002680 |
PMC10000589 | The mitochondrial voltage-dependent anion channel 1 (VDAC1) protein is involved in several essential cancer hallmarks, including energy and metabolism reprogramming and apoptotic cell death evasion. In this study, we demonstrated the ability of hydroethanolic extracts from three different plants, Vernonanthura nudiflora (Vern), Baccharis trimera (Bac), and Plantago major (Pla), to induce cell death. We focused on the most active Vern extract. We demonstrated that it activates multiple pathways that lead to impaired cell energy and metabolism homeostasis, elevated ROS production, increased intracellular Ca2+, and mitochondria-mediated apoptosis. The massive cell death generated by this plant extract's active compounds involves the induction of VDAC1 overexpression and oligomerization and, thereby, apoptosis. Gas chromatography of the hydroethanolic plant extract identified dozens of compounds, including phytol and ethyl linoleate, with the former producing similar effects as the Vern hydroethanolic extract but at 10-fold higher concentrations than those found in the extract. In a xenograft glioblastoma mouse model, both the Vern extract and phytol strongly inhibited tumor growth and cell proliferation and induced massive tumor cell death, including of cancer stem cells, inhibiting angiogenesis and modulating the tumor microenvironment. Taken together, the multiple effects of Vern extract make it a promising potential cancer therapeutic. apoptosis cancer metabolism mitochondria plant extract VDAC1 IGP LLCThis research was sponsored by IGP LLC. pmc1. Introduction Numerous natural products with anti-cancer activity used clinically, such as paclitaxel docetaxel and taxol, are derived from plants . Moreover, some compounds such as resveratrol that are produced in plant species considered to have health benefits are also shown to have pro-apoptotic effects, inducing cell death, and, as such, they act as anti-cancer agents . Similarly, curcumin expresses a variety of therapeutic properties, including antioxidant, anti-inflammatory, and antiseptic activities, as well as anticancer effects in a variety of biological pathways involved in mutagenesis, apoptosis, tumorigenesis, cell cycle regulation, and metastasis . Quercetin, a polyphenol derived from plants, has also been shown to have a wide range of biological actions, including anti-carcinogenic, anti-inflammatory, and antiviral activities, as well as attenuating lipid peroxidation, platelet aggregation, and capillary permeability . In addition to the many plant species that are already used to treat or prevent the development of cancer, several species of plants have demonstrated anti-cancer properties and are used as herbal medicines in developing countries . Here, we focus on the activity of extracts derived from three different plants. The Vernonanthura nudiflora species of perennial plant in the family Asteraceae includes more than 23,500 species spread over about 1600 genera , with distribution in Argentina, Brazil, and Uruguay . The Vernonanthura (Vernonia) genera includes more than 1000 species . The anti-proliferative and antioxidant activities of an organic extract of Vernonanthura nudiflora and some of its constituents have already been reported . In addition, some metabolites isolated from the flowers of Vernonanthura nudiflora showed antimicrobial activities The second plant tested for cytotoxicity is the Baccharis trimera used in folk medicine for the treatment of gastrointestinal disorders and hepatic diseases ; other biological activities reported for B. trimera include antihepatotoxic, antidiabetic, schistosomicidal, antioxidant, antinociceptive, and anti-inflammatory effects that are attributable to flavonoids, diterpenes, triterpenes, saponins, essential oils, and caffeoylquinic acids . The third plant is Plantago major from the Plantaginaceae family, commonly known as great plantain, and used as a medicinal plant . Plantago major contains several active compounds such as flavonoids, polysaccharides, terpenoids, lipids, iridoid glycosides, and caffeic acid derivatives . It is used to treat various diseases such as constipation, cough, wounds, infection, fever, bleeding, and inflammation. In addition, water and ethanol extracts of Plantago major leaves show anti-inflammatory activity on oral epithelial cells. Here, we present the pro-apoptotic activity of the hydroethanolic extracts from the indicated three plants while deciphering their mode of pro-apoptotic, anti-cancer activity involving mitochondria-mediated apoptosis and the mitochondrial gatekeeper protein, the voltage-dependent channel 1 (VDAC1). Mitochondria are central to essential life functions for the generation of cellular energy and critical components of the biosynthetic pathways, and function as points for cellular decisions leading to apoptosis. One of the proteins in control of these cellular life and death decisions is the mitochondrial protein VDAC1. Proper cell activity requires an efficient exchange of molecules between the mitochondria and cytoplasm. Lying in the outer mitochondrial membrane (OMM), VDAC1 assumes a crucial position in the cell, forming the main interface between the mitochondrial and cellular metabolisms . VDAC1 is a key protein in regulating metabolism, controlling the passage of adenine nucleotides, other metabolites, and Ca2+ in and out of mitochondria. VDAC1 is also an essential protein in regulating mitochondria-mediated apoptotic cell death and controls other biological and cellular functions . VDAC1 is overexpressed in various cancer cell lines and different tumors, pointing to its importance for their survival . The crucial role it plays in regulating the metabolic and energetic functions of mitochondria in cancer cells is demonstrated by findings that downregulating VDAC1 expression using siRNA decreases energy production and cell growth and inhibits tumor growth . VDAC1 is overexpressed in many diseases other than cancer, and its overexpression is associated with cell death induction . Mitochondria play a central role in apoptosis. During the transduction of an apoptotic signal into the cell, an alteration in the mitochondrial membrane permeability occurs . This allows the release of apoptogenic proteins such as cytochrome c (Cyto c), apoptosis-inducing factor (AIF), and second mitochondria-derived activator of caspase/direct inhibitor of apoptosis-binding protein with low pI (Smac/DIABLO) . When released from the mitochondria, all participate in the complex processes resulting in the activation of proteases and nucleases, leading to DNA and protein degradation and ultimately to apoptotic cell death . Several mechanisms for releasing the pro-apoptotic proteins have been proposed . These include a large channel formed by Bax and/or Bak oligomers and a channel formed by hetero-oligomers of VDAC1 and Bax or VDAC1 oligomers . Defects in the regulation of apoptosis are often associated with drug resistance and diseases such as cancer , with apoptosis evasion being a cancer hallmark . All the apoptotic proteins known to translocate to the cytoplasm following an apoptotic stimulus reside in the mitochondrial intermembrane space (IMS). Thus, only the permeability of the OMM needs to be modified for their release . Hence, VDAC1 as an OMM channel could mediate Cyto c release. Recently, we demonstrated that VDAC1 oligomers form a large channel that mediates the release of Cyto c and other pro-apoptotic proteins . Moreover, we found that cisplatin, selenite, H2O2, UV light, and more lead to apoptosis by inducing VDAC1 overexpression, thereby shifting the equilibrium towards oligomers, which leads to the release of pro-apoptotic proteins and, subsequently, apoptosis . In the present study, we demonstrate that hydroethanolic extracts from three different plants, Vernonanthura nudiflora (Vern), Baccharis trimera (Bac), and Plantago major (Pla), can induce apoptotic cell death by increasing VDAC1 expression levels, its oligomerization, and subsequent apoptotic cell death. In addition, the plant extracts increased intracellular Ca2+ and ROS production and reduced cell survival. Vern extract and one of its compounds, phytol, inhibited tumor growth and reversed tumor oncogenic properties. The plant extracts tested here showed pro-apoptotic anti-cancer activity and thus represent a promising therapeutic candidate for cancer treatment. 2. Materials and Methods 2.1. Materials 4',6-diamidino-2-phenylindole (DAPI), dimethyl sulfoxide (DMSO), propidium iodide (PI), Tris, carbonyl cyanide p-trifluoro-methoxyphenyl hydrazone (FCCP), tetramethylrhodamine, methyl ester (TMRM), and trypan blue were purchased from Sigma (St. Louis, MO, USA). Dithiothretol (DTT) was purchased from Thermo Fisher Scientific (Waltham, MA, USA). Annexin-V (FITC) was obtained from Alexis Biochemicals (Lausen, Switzerland). Dulbecco's modified Eagle's medium (DMEM) and phosphate-buffered saline (PBS) were purchased from Gibco (Grand Island, NY, USA). Fluo-4-AM and MitoSOX-Red were acquired from Invitrogen (Waltham, MA, USA). TUNEL was obtained from Promega (Madison, WI, USA). Primary and secondary antibodies used in immunoblotting and immunofluorescence (IF), as well as their dilutions, are listed in Table S1, and XTT cell viability assay kits were obtained from Biological Industries (Beit Haemek, Israel). TLC silica gel 60 F254 plates were obtained from Merck (Darmstadt, Germany). 2.2. Cell Lines and Culture U-87MG (human glioblastoma), SH-SY5Y (human neuroblastoma), HeLa (human cervix adenocarcinoma), MEFs (mouse embryonic fibroblasts), and PC-3 (human prostate cancer cells) were maintained at 37 degC and 5% CO2 in DMEM medium supplemented with 10% FBS, 1 mM L-glutamine, 100 U/mL penicillin, and 100 mg/mL streptomycin. Mycoplasma contamination was routinely evaluated on cell lines. 2.3. Plant Extracts Vernonanthura nudiflora, Baccharis trimera, and the Plantago major plants were washed with distilled water, and the aerial parts were naturally dried up to a 50% reduction in total mass. The material was introduced into reactors for maceration, with 0.20 g botanic material per 1 mL and 70% of a hydroalcoholic solution of ethanol/water 70/30, and agitated for a period of 21 days. The extract was later filtered through a filter with a pore size of 15 mm and kept at 4 degC. Before use, the hydroalcoholic was centrifuged for 5 min at 15,000x g. The mixture was composed of a hydroethanolic filtered extract of Vernonanthura nudiflora (40%), Baccharis trimera (40%), and Plantago major (20%). 2.4. Cell Treatment with Plant Extracts and Cell Death Analysis Cells (6 x 105/mL at 70-80% confluence) were incubated with ethanol extract from Vern, Bac, or Pla plant extracts or their mixture at the indicated dilution in 2000 mL culture medium for 24 h or the indicated time at 37 degC and 5% CO2. The cells were then trypsinized, centrifuged (1500x g, 5 min), washed with PBS, and analyzed for the desired activity. For cell death analysis, propidium iodide (PI) staining was performed by adding PI (6.25 mg/mL) to the cells, followed by immediate analysis by flow cytometry with the iCyt sEC800 -flow cytometry analyzer (Sony Biotechnology Inc., San Jose, CA, USA) and analysis with EC800 software. For PI and annexin V-FITC staining, cells (2 x 105), untreated or treated with the plant extracts, were collected (1500x g for 5 min), washed, and resuspended in 200 mL binding buffer (10 mM HEPES-NaOH, pH 7.4, 140 mM NaCl, and 2.5 mM CaCl2). Annexin V-FITC staining was performed according to the recommended protocol. Cells were then washed once with binding buffer and resuspended in 200 mL binding buffer, to which PI was added immediately before flow cytometric analysis by flow cytometry with the iCyt sEC800 -flow cytometry analyzer (Sony Biotechnology Inc., San Jose, CA, USA) and analysis with EC800 software. At least 10,000 events were collected and recorded on a dot plot. 2.5. Cell Viability Assay The effect of the plant extracts on SH-SY5Y cell survival was assayed using an XTT-based kit (Biological Industries, Beit Haemek, Israel) according to the manufacturer's protocol. Cells were seeded in a 96-well plate and incubated at 37 degC with 5% CO2, and 24 h later were treated with different concentrations of the extracts for the time indicated in the figure legends. XTT reagent was added to each well, and the absorbance was measured at 450 nm and 630 nm (Tecan, Infinite M1000, Mannedorf, Switzerland). The absorbance obtained at 630 nm was subtracted from the absorbance at 450 nm to obtain the specific reduced XTT reaction product. 2.6. Determination of Reactive Oxygen Species (ROS), Mitochondria Membrane Potential, and Intracellular Ca2+ Levels Mitochondrial ROS Determination: For measuring mitochondrial accumulated ROS, SH-SY5Y cells were seeded in a 6-well plate (1 x 105/well). Cells were treated for 24 h with the indicated plant extract, and then were incubated with MitoSOX-Red, a mitochondrial superoxide indicator for live-cell imaging, for 10 min at 37 degC. Fluorescence was measured using flow cytometry (iCyt, Sony Biotechnology, San Jose, CA, USA). At least 10,000 events were recorded on the FL2 detector, represented as a histogram, and analyzed with EC800 software (Sony Biotechnology, San Jose, CA, USA). Positive cells showed a shift to an enhanced level of green fluorescence (FL2). Mitochondrial Membrane Potential Determination: Mitochondrial membrane potential was determined using TMRM, a potentially sensitive dye, and a plate reader. SH-SY5Y cells were treated for 24 h with the indicated Vern plant extract and subsequently incubated with TMRM (400 nM, 20 min). The cells were then washed twice with PBS and examined withby flow cytometry with the iCyt sy3200 Benchtop Cell Sorter/Analyzer (Sony Biotechnology Inc., San Jose, CA, USA) and analysis with EC800 software. FCCP-mediated dissipation served as the control. Cytosolic Ca2+ levels [Ca2+]i measurements: [Ca2+]i was analyzed using Fluo-4-AM. Cells were harvested after the appropriate treatment, collected (1500x g for 10 min), washed with HBSS buffer (5.33 mM KCl, 0.44 mM KH2PO4, 138 mM NaCl, 4 mM NaHCO3, 0.3 mM Na2HPO4, 5.6 mM glucose), supplemented with 1.8 mM CaCl2 (HBSS+), and incubated with 2 mM Fluo-4 in 200 mL of HBSS(+) buffer in the dark for 20 min at 37 degC. After washing the remaining dye, the cells were incubated with 200 mL HBSS(+) buffer, and changes in [Ca2+]i were measured immediately by FACS and analyzed by flow cytometry with the iCyt sy3200 Benchtop Cell Sorter/Analyzer (Sony Biotechnology Inc., San Jose, CA, USA) and analysis with EC800 software. Positive cells showed a shift to an enhanced level of green fluorescence (FL1). Changes in cellular Ca2+ were monitored in live cells using the high content Operetta screening system (Perkin-Elmer, Hamburg, Germany). In each well, ten fields were imaged using a 20x wide field objective with an excitation filter of 520-550 nm and emission filter of 560-630 nm. 2.7. Cross-Linking Experiments Cells were treated with the plant extract for the indicated time and concentration, harvested, washed with PBS, pH 8.3, and incubated for 15 min with the cross-linking reagent EGS at a ratio of 1 mg protein/mL/100 mM EGS. Aliquots (30 mg protein) were subjected to SDS-PAGE and immunoblotting using anti-VDAC1 antibodies. Quantitative analysis of VDAC1 dimers was performed using FUSION-FX (Vilber Lourmat, Marne-la-Vallee, France). 2.8. Gel Electrophoresis and Immunoblotting Cells or tumor tissues were lysed using lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1.5 mM MgCl2, 10% glycerol, 1% Triton X-100, supplemented with a protease inhibitor cocktail (Calbiochem, Welwyn Garden City, UK)). The lysates were then centrifuged at 12,000x g (10 min at 4 degC), and protein concentration was determined. Aliquots (10-20 mg of protein) were subjected to SDS-PAGE and immunoblotting using various primary antibodies (sources and dilutions are provided in Supplementary Information Table S1), followed by incubation with appropriate HRP-conjugated secondary antibodies (i.e., anti-mouse, anti-rabbit). Blots were developed using enhanced chemiluminescence (Biological Industries, Beit Haemek, Israel). Band intensities were analyzed by densitometry using FUSION-FX (Vilber Lourmat, Marne La Vallee, France) software, and values were normalized to the intensities of the appropriate b-actin signal that served as a loading control. 2.9. Gas Chromatography-Mass Spectroscopy (GC-MS) Analysis CG-MS analysis was carried out using a 7890B Mass-Detector; 5977A, Agilent Technologies; Column 5MS UI. The compounds were identified using Library Name W 10N 14L (NIST MS Search 2.2). The various names representing each compound, quality of identification (maximum is 100%), and peak area (Ab*s) are given in Table S2. 2.10. TLC Separation TLC silica gel 60 F254 plates (Merck, 20 x 20 cm) were used to separate the ethanol plant extracts and phytol and ethyl linoleate, using a mobile phase mixture of petroleum ether:diethyl ether:acetic acid (85:15:1 V/V/V). The plates were air dried and visualized by exposure to iodine vapor. 2.11. Xenograft Mouse Model U-87MG cells (1.8 x 106 cells/mouse) were inoculated subcutaneously (s.c.) into the hind leg flanks of athymic eight-week-old male nude mice (Envigo). Tumor size was measured using a digital caliper, and volume was calculated. When it reached 50 mm3, mice were randomly divided into several groups (5 mice/group). One group was intratumorally injected with PBS containing 5% ethanol (final concentration in the tumor was 0.14%), and other groups were treated with Vern plant extract to a final dilution of 1:100, 1:250, 1:300, or 1:500 or with phytol to a final concentration of 75 mM. The xenografts were injected three times a week. The mice were sacrificed 34 days post-cell inoculation, and tumors were excised. Tumors were fixed in 4% buffered formaldehyde, paraffin-embedded, and processed for immunofluorescence (IF). These experimental protocols were approved by the Institutional Animal Care and Use Committee of Ben-Gurion University. 2.12. Immunofluorescence (IF) of Tumor Tissue Sections Formalin-fixed, paraffin-embedded sections (5 mm thick) of U-87MG cell-derived tumors from control and Vern plant phytol-treated tumors were deparaffinized by placing the slides at 60 degC for 1 h and using xylene, followed by rehydration with a graded ethanol series (100-50%). Antigen retrieval was performed in 0.01 M citrate buffer (pH 6.0) at 95 degC-98 degC for 20 min. After washing sections in PBS, pH 7.4, sections were incubated in 10% normal goat serum, 1% BSA in PBS containing 0.1% Triton X100 for 2 h, followed by overnight incubation at 4 degC with primary antibodies (Table S1). Sections were washed thoroughly with PBS, pH 7.4, containing 0.1% Triton-X100 (PBST), incubated with the fluorescently labeled secondary antibodies (Table S1) for 2 h, washed five times with PBST, and cover-slipped with fluoroshield mounting medium (Immunobioscience, Mukilteo, WA, USA). Fluorescent images were viewed with an Olympus IX81 confocal microscope. Quantitation of protein levels, as reflected in the staining intensity, was analyzed in the whole area of the sections using Image J software. 2.13. TUNEL Assay Paraffin-embedded fixed tumor sections (5 mm thick) were processed for a Terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick-end labeling TUNEL assay using the Dead End Fluorometric TUNEL system according to the manufacturer's instructions. Sections were deparaffinized, equilibrated in PBS, permeabilized with proteinase K (20 mg/mL in PBS), post-fixed in 4% paraformaldehyde, and incubated in TdT reaction mix for 1 h at 37 degC in the dark. Slides were then washed in saline-sodium citrate buffer, counter-stained with PI (1 mg/mL), and cover slipped with fluoroshield mounting medium. Fluorescent images of apoptotic cells (green) and cell nuclei (red) were captured using a confocal microscope (Olympus 1 x 81). Quantification analysis of stained slides was performed using an Image J program. 2.14. Statistics and Data Analysis Means +- SE of results obtained from three independent experiments are presented. Statistical significance is reported at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), or p < 0.0001 (****). 3. Results 3.1. Apoptosis Induction by the Hydroethanolic Plant Extracts The cell death activity of the three plant extracts (Vern, Bac, Pla) alone and combined was analyzed by propidium iodide (PI) staining and flow cytometry analysis . The results show that Vern extract was the most potent, triggering massive cell death, followed by the Pla extract and then the Bac extract. Vern extract (1:500) induced more cell death than all three extracts combined at a 1:100 dilution and Bac extract alone (1:166), suggesting that it is threefold more active, respectively . To determine whether the observed cell death induced by Vern was apoptosis, we analyzed apoptosis by Annexin V/PI and via FACS . The results show that Vern extract induced apoptotic cell death. Next, we analyzed the effects of the plant extract on cell survival using the XTT assay . The Vern extract reduced cell survival following 24, 48, and 72 h incubation. In contrast, Bac extract showed some decreased survival following incubation for 48 h and 72 h, and Pla extract showed no decrease in cell survival. Considering that the XTT assay is based on reduced levels of NADH produced in the mitochondria, these results suggest that Vern plant extract, but not Bac and Pla, induced mitochondrial dysfunction. In addition, the results suggest that the cell death caused by the three plant extracts involves different active compounds and modes of action. Vern plant extract similar to SH-SY5Y cells induced cell death in other cancer cell lines such as Hela and PC-3 . In contrast, non-cancerous cell lines, such as MEFs, were less sensitive to the Vern plant extract . The following experiments were conducted to reveal the extract active compounds' possible modes of action. The ethanolic plant extracts induced VDAC1 overexpression and oligomerization. We have shown that many apoptosis triggers, such as chemotherapy drugs, stress, and radiation, induce VDAC1 overexpression and oligomerization. We suggest that this is a general mechanism common to numerous apoptosis stimuli, although they act via different initiating cascades . Thus, we tested the effects of the plant extracts on VDAC1 expression levels and its oligomerization . The most active Vern plant extract induced VDAC1 overexpression in both cell lines tested: the neuroblastoma-derived cell line SH-SY5Y and glioblastoma-derived U-87MG cell line . In both cell lines, Vern extract highly increased the expressed VDAC1 level by fourfold and induced similar pro-apoptotic activity (IC50 = 1:800) . The three extracts, in a concentration-dependent manner, increased VDAC1 oligomeric forms as stabilized by chemical cross-linking using EGS and monitored by immunoblotting . However, the highest level of VDAC1 oligomerization was induced by the Vern extract , which aligns with the superior potency of its cell death-inducing activity. Interestingly, we observed the presence of VDAC1 oligomers even without chemical cross-linking, even after exposing the cells to a high concentration (1%) of the detergent SDS and heating at 70 degC for 5 min . The level of oligomeric VDAC1 was highest with Vern extract, as found for VDAC1 overexpression, oligomerization, and apoptosis induction. This suggests that the VDAC1 oligomers induced by the plant extract are very stable. The results support the suggestion that the active compounds in the Vern plant extract, via enhancing VDAC1 expression levels, lead to VDAC1 oligomerization and apoptosis. The active compounds in Vern extract are resistant to high temperatures, as heating the extract for 10 min at 40, 60, or 80 degC had no effect on the extract cell death activity (Table S3). 3.2. Vern Extract Increased Intracellular Ca2+ and ROS Production Reactive oxygen species (ROS) were shown to induce apoptosis . Thus, we measured mitochondrial ROS and found that cell treatment with Vern extract induced their production . This increase in mitochondrial ROS suggests that Vern extract at high concentrations induces dysfunction of the mitochondria, as also reflected in the decrease in XTT reduction. Several studies have shown that an increase in [Ca2+] is involved in apoptosis induction and that Ca2+ is required for apoptosis-stimuli-induced VDAC1 overexpression and VDAC1 oligomerization . Vern extract's effect on cellular [Ca2+] levels was analyzed using Fluo-4 and FACS or by Operetta . Both assays demonstrated that at high levels, this extract increased cellular [Ca2+] levels. Cell treatment with Vern plant extract reduced the mitochondrial membrane potential (Dps) only when high (over 80%) cell death was obtained . The findings that the increase in ROS production, [Ca2+] levels, and dissipation of (Dps) were not correlated with cell death and were observed only at high concentrations of the Vern plant extract and over 80% cell death suggest that these effects are due to cell destruction, including of the mitochondria. 3.3. GS-MS Analysis of Extracts from Plants Vern, Bac, and Pla To identify some of the chemical compounds present in the hydroethanolic extracts from the three different plants, Vern, Bac, and Pla, the extracts were subjected to gas chromatography-mass spectroscopy (GC-MS) analysis. Considering only those compounds with a score of over 90% certainty, 12 and 13 compounds were identified in Vern and Pla extracts, respectively, and 20 in the Bac extract. Some of these compounds are common to the three plant extracts, while others are only two or unique to just one (Table S2). As expected for ethanol extract, all identified compounds are hydrophobic, containing fatty acid derivatives such as palmitic acid ethyl ester (hexadecanoic acid ethyl ester, linolenic acid ethyl ester, (9,12,15-octadecatrienoic acid ethyl ester), and stearic acid ethyl ester (octadecanoic acid, ethyl ester). We tested the cell death induction activity of two compounds found in the extracts, namely phytol and ethyl linoleate (Table S2), both commercially available. The relative amounts of phytol and ethyl linoleate in the plant extracts were determined using known quantities of the two compounds and TLC . The amounts of phytol were about 1800, 800, and 1200 nmol/mL (mM) in Vern, Bac, and Pla extracts, respectively. Next, the activity of the phytol and ethyl linoleate in inducing cell death was analyzed by incubating SH-SY5Y cells with different concentrations (50-200 mM) for 24 or 48 h . Phytol induced cell death with half-maximal cell death (IC50) of about 80% at 70 mM. Ethyl linoleate showed weak cell death activity, increasing from 15% in non-treated cells to about 40% at 200 mM of ethyl linoleate . Phytol also highly reduced cell survival, as analyzed using the XTT assay , suggesting that its effect involves mitochondria dysfunction. Next, we tested whether ethyl linoleate-induced cell death was associated with increased VDAC1 expression levels and oligomerization. Both compounds, at the high concentrations used, induced VDAC1 overexpression and VDAC1 oligomerization in a concentration-dependent manner. In correlation with the higher activity of phytol in cell death induction, phytol increased both VDAC1 overexpression and oligomerization. Since Vern extract induced cell death at a dilution of 500-1000, with phytol concentrations in these dilutions of 1.8 to 3.6 mM, cell death was observed at over 50 mM of phytol, suggesting that other compounds in the plant extracts are involved in cell death induction. 3.4. Anti-Tumor Activity of Vern Extract and Phytol Next, we tested the effect of Vern extract at two dilutions and phytol on tumor growth . U-87MG cells were inoculated subcutaneously (s.c.) into the hind leg flanks of 7-week-old male athymic nude mice. When the tumor volume was around 50 mm3, the mice were divided into four groups with a similar average volume and treated with Vern extract or phytol . Control tumors were injected with PBS containing 5% ethanol (final concentration in the tumor 0.14%). Groups 2 and 3 were treated with Vern extract at a final dilution in the tumor of 1:250 or 1:500, and Group 4 was treated with phytol (75 mM). Treatment was given three times a week, and tumor growth was monitored . All mice were sacrificed 34 days post-cell inoculation; tumors were excised, weighed , and fixed; and sections were immunofluorescent-stained for selected proteins. The results show that the tumors in the control grew exponentially with time and in a similar way when injected with Vern extract to a final dilution of 1:250. On the other hand, tumors treated with a higher dilution of Vern extract, 1:500, showed about a 70% decrease in tumor volume and weight . The results indicate that Vern extract at a higher concentration (1:250) is less effective than at the dilution of 1:500. Similar results with no effect on tumor growth were obtained using a 1:100 dilution of Vern extract . The decreased anti-cancer effect with increased Vern extract concentration may result from the protective activity of other compounds in the extract. The results also show that phytol at the concentration used (75 mM) significantly inhibited tumor growth, yet less than Vern extract at 1:500 dilution . Phytol has been shown to induce apoptosis in cells in culture , and, as also shown here, it also decreased cell survival . For the first time, it induced VDAC1 overexpression and VDAC1 oligomerization and inhibited tumor growth . Tumor-fixed paraffin-embedded sections were stained for Ki-67, a proliferation marker, showing that both Vern extract and phytol inhibited cell proliferation by about 80% . Finally, we analyzed apoptosis by TUNEL staining . While no TUNEL-positive cells were apparent in control tumors, most of the cells were TUNEL-positive in the Vern extract-treated tumors and to a lesser extent in phytol-treated tumors, with staining co-localizing with PI nuclear staining . Thus, Vern extract was more effective than phytol in cell death induction. The results indicate that the treatments induced apoptotic cell death and suggest that the marked decrease in tumor size in the Vern phytol-treated xenografts can be attributed to both inhibition of cell proliferation (decreased Ki-67 staining) and cell death induction. Next, we analyzed the effect of tumor treatment with Vern extract or phytol on the expression levels of proteins associated with metabolism, the microenvironment, and cancer stem cells . 3.5. Vern Extract and Phytol Reduced the Expression of the Metabolic Enzyme in a Xenograft Mouse Model The metabolic alterations that occur during malignant transformation involve a spectrum of functional aberrations and mutations that contribute to elevated glycolysis and increased expression levels of glucose transporters (Glut-1) and glycolytic enzymes as hexokinase (HK-I) . IF of the control tumors derived from U-87MG cells showed high expression levels of Glut-1 and glyceraldehyde three-phosphate dehydrogenase (GAPDH) that were decreased in tumors treated with Vern extract or phytol . Similarly, the expression of VDAC1 and HK-I was reduced in the Vern phytol-treated tumors . The decreased expression levels of metabolism-related enzymes in the Vern phytol-treated tumors suggest decreased energy production in these treated tumors. 3.6. Vern Extract and Phytol Modulate the Tumor Microenvironment A tumor contains cancer cells and non-cancerous cells, creating the tumor microenvironment (TME). Besides cancer cells, a tumor has fibroblasts , immune system cells , blood vessels , and extracellular matrix (ECM) components . The TME plays a vital role in cancer growth and spread. Angiogenesis is an underlying promoter of tumor growth, invasion, and metastases, with glioblastoma (GBM) being highly angiogenic . Immunostaining of endothelial cell marker CD-31 showed that in the Vern phytol-treated tumors, there was a significant decrease in the number of blood vessels, with quantitation revealing decreases of about 70% and 60% in Vern extract and phytol-treated tumors, respectively, relative to control tumors . The effects of Vern extract and phytol on the TME were analyzed by IF staining for the fibroblast marker alpha-smooth muscle actin (a-SMA) . Tumors treated with either Vern extract or phytol showed decreased a-SMA expression. Accumulated recent evidence supports the cancer stem cell (CSC) hypothesis, which suggests that a sub-population of malignant cells exhibit the stem cell properties of self-renewal and differentiation. CSCs are resistant to conventional cytotoxic/anti-proliferative therapies. In GBM, the proteins Sox2, CD133, SSEA1, CD49f, Musashi-1, and Nestin are considered to be glioma stem cell CSC markers. The IF staining for Sox2 and Nestin of the tumor specimens demonstrated that in tumors treated with plant Vern extract or phytol, the expression levels of these CSC markers was highly reduced, by about 70% . These results indicate that Vern plant extract and phytol treatment of U-87MG-derived tumors eliminated CSCs associated with tumor recurrence. 4. Discussion Recently, significant attention has been placed on using nutraceuticals as therapeutic agents inducing cell death and suppressing cancer growth as an alternative treatment for cancer or in combination with chemotherapy . Screening for plant-derived compounds with anti-neoplastic activity has contributed to identifying resveratrol, quercetin, curcumin, and others. In addition to their anti-cancer activity, these compounds also showed antioxidant, anti-inflammatory, anti-viral, and neuroprotective properties, lowered blood pressure, and improved cardio-metabolic markers and anti-aging effects . Here, we present the activity of hydroethanolic extracts from three different plants--Vernonanthura nudiflora, Baccharis trimera, Plantago major, and their mixture--in cell death induction, VDAC1 overexpression, and oligomerization. The effects of Vern plant extract and one of its constituents, phytol, on tumor growth and oncological properties were tested. 4.1. Plant Extracts Inducing Cell Death Involve VDAC1 Overexpression and Oligomerization Recently, VDAC1 has been recognized as a regulator of mitochondria-mediated apoptosis . We demonstrated that many apoptosis inducers lead to VDAC1 overexpression and its oligomerization, resulting in the formation of a large channel that enables the release of pro-apoptotic protein from the IMS to the cytosol, thereby activating apoptosis . This study showed that the three plant extracts induced massive cell death at relatively low doses (1:1000 of the original ethanolic extract). The plant extracts' active compounds inducing apoptosis involved increased VDAC1 expression levels and oligomerization and, thereby, apoptosis. Thus, we propose that phytol and the plant extracts' activation of apoptosis involve overexpression of VDAC1, shifting the equilibrium towards VDAC1 oligomers, allowing Cyto c release, and thereby, apoptosis . We have proposed a model for plant extract and phytol inducing VDAC1 overexpression leading to VDAC1 oligomerization, forming a large channel, mediating the release of apoptogenic proteins from the intermembrane space (IMS) to the cytosol, and activating the apoptosis cascade. The active compound(s) in the plant extracts inducing VDAC1 overexpression may involve several possible mechanisms, such as the increase in ROS and intracellular Ca2+ levels, as both were shown to regulate gene expression. Ca2+-dependent gene transcription has been demonstrated in neurons and other cells . ROS was shown to upregulate gene expression for death receptors such as TRAIL (TNF-related apoptosis-inducing ligand) that appear to be mediated by transcription factors such as CHOP (C/EBP homologous protein) and p53 . Here, we demonstrated that Vern extract highly elevated intracellular Ca2+ levels, as monitored using Fluo-4 and FACS analysis or by cell imaging using the Operetta imaging system . Similarly, Vern extract induced ROS production . Thus, these findings suggest the involvement of Ca2+ and ROS in triggering transcription factors controlling VDAC1 expression, with the increased VDAC1 level leading to its oligomerization, and apoptotic cell death . The active molecule(s) in the extracts responsible for VDAC1 overexpression are yet to be identified. 4.2. GC-MS Analysis of the Ethanolic Plant Extracts A preliminary GC-MS study showed that the ethanolic plant extracts contain many chemical entities at different levels (Table S2). The major compounds are fatty acids in their ethyl ester forms, such as palmitic acid ethyl ester (hexadecanoic acid ethyl ester), linolenic acid ethyl ester (9,12,15-octadecatrienoic acid ethyl ester), and stearic acid ethyl ester (octadecanoic acid ethyl ester). Interestingly, analysis of the constituents of ethanol root extract of the plant Rauwolfia vomitoria using GC-MS also showed the presence of fatty acids. Still, these were different from those found in the plant extracts tested here, such as ethyl oleate (10.59%), 9,12-octadecadienoic acid ethyl ester (8.26%), and palmitic acid, also known as hexadecanoic acid ethyl ester (8.11%) . Here, we showed that one of the identified compounds in the ethanolic plant extracts, phytol, was found to induce cell death and VDAC1 overexpression and oligomerization. Phytol has been reported to induce both apoptosis and protective autophagy . Quantitative analysis of phytol amounts in the three plant extracts indicates that its concentration was highest in the Vern extract (5.9 mM) and about 7 mM and 2 mM in extracts B and C, respectively. Phytol-induced cell death was obtained at high concentrations (50 to 200 mM). Based on the estimated phytol concentration in the plant extracts and that Vern extract induced cell death at a 1000-fold dilution, the phytol concentration is about 13.5 mM below its effective concentrations in cell death induction . Thus, it is most likely that other compounds in the extracts contribute to its biological activity; however, their identification is the topic of another study. Interestingly, our findings that phytol induced VDAC1 overexpression and oligomerization agree with the results that phytol was shown to modulate transcription in cells via a transcription factor--the peroxisome proliferator-activated receptor alpha (PPAR)--involved in regulating lipid metabolism in various tissues . Phytol directly activates PPAR-alpha and regulates gene expression involved in lipid metabolism in PPAR-alpha-expressing HepG2 hepatocytes. It also modulates the retinoid X receptors (RXRs), which are nuclear receptors activated by various endogenous and natural ligands such as 9-cis retinoic acid, n-3 polyunsaturated fatty acids, and phytanic acid . Thus, phytol may enhance VDAC1 expression by modulating transcription factor(s). 4.3. Vern Extract and Phytol Inhibit Tumor Growth and Alter Tumor Oncogenic Properties The effects of Vern extract and phytol on tumor growth and tumor oncogenic properties were tested using a U-87MG-cell-derived tumor-based GBM mouse model. GBM is an aggressive brain cancer with high rates of relapse and mortality, mutational diversity, and poor treatment options. The Vern extract decreased tumor size by about 70% when used at the high dilution of 1:500 but was less effective at higher amounts, such as at 1:200 or 1:100 dilutions . This finding can be explained when considering the composition of the extract compounds revealed by both GS-MS and LC-MS/MS analyses. In addition to the cell death-induced compounds, Vern plant extract contains compounds that support cell growth and are protective against cell death (Table S2). We suggest that the pro-survival compounds are active at high concentrations and overcome the activity of the pro-cell death compounds. At high dilution, the levels of pro-survival compounds are below their active concentrations; thereby, the effects of the pro-apoptotic compounds, which act at low concentrations, are evident. This observation suggests that it is possible to control the desired activity, supportive/ pro-cell death, according to the extract concentration. The results also show that phytol at the concentration used (75 mM) inhibited tumor growth. Phytol in cells in culture has been shown to induce apoptosis and protective autophagy . Here, for the first time, it was demonstrated to inhibit tumor growth, and as discussed below, similar to the Vern extract, it induced multiple effects on the tumor. The impact of Vern extract on the tumors cannot be due to phytol, as at a dilution of 1:500, it contains about 30 mM phytol, which induced only 10% cell death . The inhibition of tumor growth involves both inhibition of cell proliferation, reflected in an 80% decrease in the levels of the cell proliferation marker, Ki-67 , and cell death induction, showing that the Vern extract is about fivefold more active than phytol . Tumors require changes in the cellular metabolism and bioenergy of cancer cells , and their metabolic adaptation provides the tumor with the precursor needed for the biosynthesis of nucleic acids, fatty acids, cholesterol, and porphyrins . Mitochondrial metabolism plays a vital role in the survival and development of cancer cells . Here, we demonstrated that both Vern extract and phytol treatment of GBM in a mouse model significantly decreased the expression of metabolism-related enzymes involved in glycolysis and the TCA cycle , leading to reduced cell function and survival. The decrease in the expression of metabolism-related enzymes in the tumors treated with Vern plant extract or phytol for 27 days may result from the massive cell death leading to cell distraction, including in the mitochondria and in the degradation of many cell proteins . Similar results were obtained with the VDAC1-based peptide . Vern extract and phytol tumor treatment altered the tumor microenvironment, disrupting tumor-host interactions. The treated tumors showed a reduced expression of angiogenesis markers such as CD-31, decreasing blood supply. Chronic inflammation caused by cancer cells stimulates surrounding cells, including fibroblasts and activated fibroblasts, with a-SMA expression producing an extracellular matrix including collagen. The fact that Vern extract and phytol alter the tumor microenvironment is reflected in the decreased a-SMA and Sirius red staining . a-SMA produced by cancer-associated fibroblasts (CAFs) contributes to remodeling and reconstitution to promote invasion and metastasis via the extracellular matrix, growth factors, and protease production , as well as to metastasis, and poorer prognosis was highly decreased; thus, the treatment reduced these tumor properties. As tumorigenesis is considered an interplay between tumor cells and the surrounding stroma host cells , alteration in the tumor microenvironment by the treatments suggests that this affects cancer progression, invasiveness, and treatment response. Finally, CSCs, with their ability to self-regenerate, are considered to be responsible for initiating tumor growth and recurrence after therapeutic interventions and are associated with tumor resistance to anti-cancer therapies . Here, we showed that tumor treatment with Vern extract or phytol resulted in the elimination of CSCs, as indicated by the decreased expression of the specific markers Nestin and Sox2 . 5. Conclusions We found that Vern extract at a high dilution and one of its compounds, phytol, have various effects on tumors, with their anti-cancer effects involving (i) apoptosis induction, (ii) inhibition of cell proliferation, (iii) re-modulation of the tumor microenvironment, (iv) impairment of cancer cell metabolism, and (v) eliminating CSCs, all leading to the observed inhibition of tumor growth. These findings, and considering that the side effects of these plant extracts are minor relative to those of conventional chemotherapy, suggest that the plant extract or a combination of its active compounds (yet to be identified) are a promising therapeutic approach for GBM and various other cancers. Supplementary Materials The following supporting information can be downloaded at: Figure S1: Apoptosis analysis using Annexin V and PI and FACS analysis; Figure S2: Vern plant extract effect on different cell lines; Figure S3: Vern plant extract effect on mitochondrial membrane potential; Figure S4: Inhibition of tumor development by Vern plant extract in glioblastoma xenograft mouse model; Table S1: Antibodies used in this study; Table S2: The compounds identified in plant extracts Vern, Bac, and Pla using GC-MS analysis, Table S3: Activity of Vern plant extract at various temperatures. Click here for additional data file. Author Contributions Conceptualization, V.S.-B.; methodology, A.N., A.S.-K. and S.K.P.; resources, J.O., D.K. and V.S.-B.; writing--review and editing, V.S.-B.; supervision, V.S.-B. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement All experiments were approved by the Animal Care and Use Committee of Ben-Gurion University of the Negev, as required by Israeli legislation, and all efforts were taken to minimize animal suffering. Authorization Number: IL-36-07-2018(E). Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest Juan Ortas and Daniel Kerekes from Sigma Inc. No sponsor is received from this company. The current affiliation of Swaroop Kumar Pandey is Department of Biotechnology, GLA University, Mathura 281406, India; ([email protected]). All authors declare no conflicts of interest. Figure 1 Plant extracts derived from Vernonanthura nudiflora (Vern), Baccharis trimera (Bac), and Plantago major (Pla) inducing apoptotic cell death. Cancer cells, SH-SY5Y, were incubated (24 h) with the indicated dilutions of the ethanolic plant extracts: Vern (Vernonanthura nudiflora), Bac (Baccharis trimera), Pla (Plantago major), or their mixture (Vern+Bac+Pla). (A) A representative FACS analysis of propidium iodine (PI)-stained cells showing live and dead cells in the control and cells incubated for 24 h with Vern plant extract (1:500). (B) Apoptosis was analyzed in cells incubated with the indicated different dilutions of extract from plant Vern, Bac, or Pla extracts, and death cells were analyzed as in (A). The results are the mean +- SEM of three independent experiments. (C) SH-SY5Y cells were incubated (24 h) with the indicated dilution of Vern plant extract, and then analyzed for apoptosis using Annexin V/PI staining and FACS. (*) and (^) represent results from two different experiments. (D-F) SH-SY5Y cell survival as revealed by the XTT assay of cells incubated with different dilutions of the indicated plant extract for 24, 48, or 72 h. Figure 2 Vern plant extract induced VDAC1 overexpression, oligomerization, and cell death. (A,B) SH-SY5Y or U-87MG cells were incubated for 24 h with the indicated dilution of Vern plant extract, and then analyzed for VDAC1 expression levels (A) by immunoblotting using anti-VDAC1-specific antibodies. Immunoblotting with b-actin as a loading control is also shown. The levels of VDAC1 are given below the blot as relative units (RUs). Samples were also analyzed for cell death (B). (C,D) SH-SY5Y cells (1 mg/mL) were incubated for 24 h with the indicated dilutions of extracts Vern, Bac, Pla, or their mixture and then analyzed for VDAC1 oligomerization by incubation with the cross-linking reagent EGS (100 mM), followed by immunoblotting using anti-VDAC1 antibodies (C). The positions of the VDAC1 monomers, dimers, trimers, tetramers, and higher oligomers are indicated. The level of VDAC1 dimers was analyzed using FUSION-FX software (D) and presented relative to its levels in control cells subjected to EGS. (E,F) SH-SY5Y cells were incubated for 24 h with the indicated dilutions of plant extracts Vern, Bac, Pla, or their mixture without EGS treatment, and subjected to immunoblotting. The position of the VDAC1 monomers, dimers, trimers, and tetramers is indicated. The level of VDAC1 dimers was analyzed using FUSION-FX software and presented relative to its levels in control cells (F). Figure 3 Vern plant extract induced increased intracellular Ca2+ and ROS production. SH-SY5Y cells were incubated for 24 h with the indicated dilutions of Vern plant extract and analyzed for ROS production using MitoSox Red reagent and FACS analysis (A), and for intracellular [Ca2+]i, using the calcium indicator Fluo-4 reagent and FACS analysis (B), or for Vern plant extract treatment at 1:333 dilution is visualized using Operetta imaging (C), and quantified (D) **** p < 0.0001. Figure 4 GC-MS analysis of the plant extracts and phytol as one of the cell death active compounds. (A) Ethanolic extracts from Vern, Bac, and Pla plant extracts were subjected to TLC separation, along with known amounts of phytol and ethyl linoleate using the solvent mixture of petroleum ether:diethyl ether:acetic acid (85:15:1, V:V:V) and developed by exposure to iodine vapor. (B) Quantification of phytol and ethyl linoleate in plant extracts using the compounds' calibration curves and Image J software (n = 3). (C) SH-SY5Y cells were incubated for 24 h or 48 h (C) with the indicated concentration of phytol or ethyl linoleate, and cell death was evaluated using PI staining and FACS analysis. Results are the mean +- SEM (n = 3). (D) Cell survival was assayed following 24, 48, or 72 h incubation with the indicated concentrations of phytol, using an XTT assay. Figure 5 Phytol and ethyl linoleate VDAC1 overexpression and oligomerization. (A,B) SH-SY5Y cells were incubated for 24 h with the indicated concentrations of phytol or ethyl linoleate, and then analyzed for VDAC1 expression levels by immunoblotting (A). The level of VDAC1 is given as relative units (RUs) quantified using Image J software (B). (C,D) Control and ethyl linoleate-treated cells (1 mg/mL) were also analyzed for VDAC1 oligomerization by incubation with the cross-linking reagent EGS (100 mM), followed by immunoblotting using anti-VDAC1 antibodies (C) and quantified for dimer level (D). The position of the VDAC1 monomers, dimers, trimers, and tetramers is indicated. Results are the mean +- SEM (n = 3). Figure 6 Inhibition of tumor growth by Vern plant extract and phytol in a GBM xenograft mouse model. U-87MG cells (1.8 x 106 cells/mouse) were s.c. inoculated into nude mice. Tumor volume was monitored (using a digital caliper), and on day 14, when the tumor volume was about 50 mm3, the mice were divided into four groups with a similar average volume calculated per group (5 or 6 mice per group) (A). The four mice groups were subjected to the following treatments: control (ethanol to a final concentration 0.14%) or Vern plant extract to a final dilution of 1:250 or 1:500, and phytol to a final 75 mM, calculated according to the tumor volume. (B) The calculated average tumor volumes as a function of time are presented as means +- SEM (n = 5 or 6 mice). (C) Tumor-calculated volume before scarifying the mice (day 34), presented as % of the control. (D) The calculated average tumor weights are presented as means +- SEM. * p < 0.05. (E,F) Confocal images of representative paraffin-embedded sections from U-87MG-derived control, Vern plant extract (1:500)-, or phytol (75 mM)-treated tumors, immunofluorescent-stained with antibodies against the proliferation marker, Ki-67 (E), and quantification of staining intensity (F). Results are the mean +- SEM (n = 3), **** p < 0.0001. (G,H) TUNEL staining on paraffin-embedded sections cut from control, Vern plant extract-, or phytol-treated tumors. TUNNEL staining of tumor sections was carried out as described in the Section 2. Representative confocal images with red staining indicates PI nuclear staining, and green-stained cells indicate TUNEL staining (G). TUNEL-positive cells were quantified and presented as TUNEL-positive cells (H). Results are the mean +- SEM (n = 3), * p < 0.05, *** p < 0.001. Figure 7 Tumor treatment with Vern plant extract or phytol resulted in reduced cell metabolism. (A-D) Confocal images of sections from U-87MG-derived tumors, control or treated with Vern plant extract (1:500) or phytol (75 mM), immunofluorescent-stained for Glut-1 and GAPDH (A,B) or for VDAC1 and HK-I (C,D) using specific antibodies. Staining intensity was quantitative using Image J software (B,D). Results reflect the mean +- SEM (n = 3), **** p <= 0.0001. Figure 8 Vern plant extract and phytol treatment markedly reduced angiogenesis and the tumor microenvironment in U-87MG cell-derived tumors. Confocal images of sections from U-87MG-derived tumors, control or Vern plant extract (1:500) or phytol (75 mM)-treated tumors were immunofluorescent-stained for CD-31 (A,B) or a-SMA (C,D), and their quantification is presented (B,D). Results are means +- SEM (n = 3 mice), **** p <= 0.0001. Figure 9 Vern plant extract and phytol treatment markedly reduced cancer stem cells markers in U-87MG cell-derived tumors. Representative IF staining of tumor sections from U-87MG-derived tumors, control or treated with Vern plant extract (1:500) or phytol (75 mM), co-immunofluorescent-stained with specific antibodies against the CSCs markers, Sox2 and Nestin (A), and their quantitative analysis (B). Results are means +- SEM (n = 3 tumors), **** p <= 0.0001. Figure 10 Proposed model for active molecules of plant extracts inducing VDAC1 overexpression and VDAC1 oligomerization leading to apoptosis. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000590 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050727 cells-12-00727 Review The Two Faces of Immune-Related lncRNAs in Head and Neck Squamous Cell Carcinoma Bueno-Urquiza Lesly J. 1 Martinez-Barajas Marcela G. 1 Villegas-Mercado Carlos E. 2 Garcia-Bernal Jonathan R. 1 Pereira-Suarez Ana L. 3 Aguilar-Medina Maribel 4 Bermudez Mercedes 2* Alessandro Poggi Academic Editor Dudas Jozsef Academic Editor 1 Department of Physiology, University Center for Health Sciences, University of Guadalajara, Guadalajara 44340, Mexico 2 Faculty of Dentistry, Autonomous University of Chihuahua, Chihuahua 31000, Mexico 3 Department of Microbiology and Pathology, University Center for Health Sciences, University of Guadalajara, Guadalajara 44340, Mexico 4 Faculty of Biological and Chemical Sciences, Autonomous University of Sinaloa, Culiacan, Sinaloa 80030, Mexico * Correspondence: [email protected]; Tel.: +52-(614)-439-1834 24 2 2023 3 2023 12 5 72721 12 2022 15 1 2023 21 1 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Head and neck squamous cell carcinoma (HNSCC) is a group of cancers originating from the mucosal epithelium in the oral cavity, larynx, oropharynx, nasopharynx, and hypopharynx. Molecular factors can be key in the diagnosis, prognosis, and treatment of HNSCC patients. Long non-coding RNAs (lncRNAs) are molecular regulators composed of 200 to 100,000 nucleotides that act on the modulation of genes that activate signaling pathways associated with oncogenic processes such as proliferation, migration, invasion, and metastasis in tumor cells. However, up until now, few studies have discussed the participation of lncRNAs in modeling the tumor microenvironment (TME) to generate a protumor or antitumor environment. Nevertheless, some immune-related lncRNAs have clinical relevance, since AL139158.2, AL031985.3, AC104794.2, AC099343.3, AL357519.1, SBDSP1, AS1AC108010.1, and TM4SF19-AS1 have been associated with overall survival (OS). MANCR is also related to poor OS and disease-specific survival. MiR31HG, TM4SF19-AS1, and LINC01123 are associated with poor prognosis. Meanwhile, LINC02195 and TRG-AS1 overexpression is associated with favorable prognosis. Moreover, ANRIL lncRNA induces resistance to cisplatin by inhibiting apoptosis. A superior understanding of the molecular mechanisms of lncRNAs that modify the characteristics of TME could contribute to increasing the efficacy of immunotherapy. HNSCC tumor microenvironment LncRNAs cancer-associated fibroblasts This research received no external funding. pmc1. Introduction Cancer, a group of multifactorial diseases, is considered one of the main public health problems, being the second cause of death worldwide . According to GLOBOCAN, HNSCC incidence and mortality are about 800,000 and 400,000 cases, respectively, positioning it as the sixth most common cause of cancer death around the world . HNSCC develops from squamous cells in the mucosal epithelium lining the oral cavity, larynx, oropharynx, nasopharynx, and hypopharynx . This type of cancer is more common in men, with a 3:1 ratio compared with women , and occurs mainly after the age of 55 . The main factors related to the development of this type of cancer are the consumption of alcohol and tobacco , whose effect is proportional to the intensity of exposure . Additionally, it has been described that infection with high-risk human papillomavirus (HPV), mainly genotypes 16, 18, 31, 33, and 35, acts synergistically in carcinogenesis. In this regard, HNSCC can be classified as HPV-negative and HPV-positive . HPV infection is responsible for up to 60% of HNSCC cases, as it participates in the development of oropharyngeal tumors, being the 90% of HPV-positive tumors related to HPV 16 infection. Interestingly, HPV infection, in addition to being an etiological factor, is related to the prognosis of patients. It has been observed that HPV-positive cases show a favorable prognosis, unlike those that are not . Tumor Microenvironment The TME is a very complex construct composed of extracellular matrices (ECM) and cellular components such as tumor cells, immune cells, and cancer-associated fibroblasts (CAFs), among others . For this, tumors can be classified according to the cellular infiltrate as inflamed tumors and immune-excluded and immune-desert tumors. Inflamed tumors are characterized by abundant intratumoral and stromal immunological infiltrate. Immune-excluded tumors have immunological infiltrate restricted to the stroma. Immune-deserts lack infiltrate both in the tumor and in the stroma . Even though in inflamed tumors there is an infiltrate of immune cells, in an immunosuppressive environment the tumor can evade the host response and progress ; this also depends on the infiltrate and its relationship with a positive or negative prognosis. The most common model to explain the tumor behavior is "cancer immunoediting", which refers to a dual action that the immune system can take, one of which is the protection towards the host by eliminating tumor cells, the other is the programming of, those cells of the immune system that are associated with the tumor and help tumor progression . This process can be divided into three phases that are called the "three E" (elimination, equilibrium, and escape). First, elimination refers to immunosurveillance mediated by the immune cells; second, equilibrium is where the immune system promotes the generation of tumor cells that survive the attack; finally, once immunological anergy and tolerance are achieved, escape leads to cancer cells that can form tumors . The immune infiltrate in TME includes cells from the adaptive immunity such as cytotoxic T lymphocytes (CTL) that recognize and kill tumor cells through the release of granzymes and perforins, CD4+ T cells that are essential for the proliferation and differentiation of CD8+ T cells that infiltrate the tumor, innate immune cells such as natural killer (NK) cells that have cytotoxic and cytokine-producing activity, tumor-associated macrophages (TAMs) classified into two subpopulations (M1 with antitumor activity and M2 with protumor activity and an immunosuppressive profile), mast cells that release preformed inflammatory mediators in their granules, and finally stromal cells such as CAFs that are fibroblasts functionally different from the normal population and participate in the remodeling of the extracellular matrix and the production of protumoral cytokines . The antitumor immune response is characterized by an infiltrate of CTL, B lymphocytes, CD4+ Th1 lymphocytes, regulatory T cells (Treg), M1 macrophages, and NK cells, while CD4+ Th2 lymphocytes, M2 macrophages, neutrophils, and CAFs, among others, participate in the protumoral immune response . These cell populations have intercellular communication through cytokines, chemokines, and non-coding RNAs (ncRNAs) , which will modulate the characteristics of TME . ncRNAs represent a large percentage of the genome with relevant functions in biological processes since they control the expression of genes. ncRNAs can be classified according to their length in microRNAs (miRNA), which have a length of approximately 22 nucleotides, and the lncRNA, which are longer than 200 nucleotides . 2. Long Non-Coding RNAs LncRNAs are non-coding chains of 200 to 100,000 nucleotides transcribed by RNA polymerase II . Generally, they have a poly-A tail and can be subjected to splicing processes . Their mechanisms of action are diverse both in the cytoplasm and in the nucleus. In the cytosol, they are related to the regulation of mRNA decay as well as its stability, functioning as sponges for miRNAs. Meanwhile, in the nucleus, they are associated with promoter sites, participating in transcriptional repression, epigenetic regulation, and nuclear architecture . LncRNAs play an important role in both innate and adaptive immune responses; it has been shown that they affect essential processes such as differentiation or the immune function . It has been observed that some of the biological processes they regulate are cell activation, proliferation, metabolism, and death . 3. LncRNAs in HNSCC Tumor Microenvironment 3.1. Tumor Cells In recent years, the participation of lncRNAs in the tumorigenesis of cancer cells involving the tumor microenvironment has gained relevance given that some of the lncRNAs are associated with poor prognosis (Table 1). In this regard, the lncRNA MIR31HG is associated with poor prognosis since its expression is significantly correlated with advanced stages in laryngeal squamous cell carcinoma (LSCC) samples and in vitro and in vivo analysis found that it promotes cancer cell growth . In addition, USP2-AS1 promotes progression through proliferation, tumor growth, invasion, and the transition from G0/G1 to the S phase of the cell cycle in both in vitro and in vivo models . On the other hand, the lncRNA TM4SF19-AS1 acts as a sponge for miR-153-3p since it binds to LAMC1 (laminin gamma 1 subunit), which has been reported to be upregulated in patients with HNSCC ; thus, TM4SF19-AS1 enhances proliferation, migration, invasion, and epithelial-mesenchymal transition (EMT) through the expression of mesenchymal markers (vimentin, N-cadherin) . Furthermore, LINC00460 is associated with the regulation of proliferation, migration, invasion, and mesenchymal marker expression in vitro . In the case of HCG18, lncRNA is overexpressed in cell lines and patients with HNSCC regulating migration, invasion, and modulating progression through the expression of cyclin D, which is a key protein in the WNT signaling pathway and is directly associated with a poor prognosis of patients . The lncRNAs not only act in the progression of cancer but also in tumor suppression, being associated with a good prognosis. For instance, HNSCAT1 is downregulated in samples of advanced HNSCC, meanwhile its overexpression is associated with the formation of minor tumors in vivo . 3.2. CAFs Due to the heterogeneity of CAFs, several pathways participate in their activation. Recently, the role of some lncRNAs that participate in the modulation of their activation has been described as finding the stimulus with the factor PDGF-BB (platelet-derived growth factor-BB), associated with differentiation towards CAFs , also increases the expression of the lncRNA LURAP1L-AS1 (leucine-rich adaptor protein 1-like antisense RNA 1) as well as the classical markers of CAFs (a-SMA (a-smooth muscle actin), FSP-1 (fibroblast-specific protein 1), and FAP (fibroblast activation protein)).When LURAP1L-AS1 silencing is performed, the expression of the markers decreases; it also participates in the regulation of NF-kB through the LURAP1L-AS1/LURAP1L/IKKa/IkBa/NF-kB axis . Another lncRNA overexpressed is FLJ22447 or lncRNA-CAF that, in conjunction with IL-33, participates in NF activation toward CAFs. LncRNA-CAF silencing has an impact on the decreased expression of classical CAF markers and lncRNA-CAF functions as a lncRNA scaffold to maintain IL-33 protein stability and inhibit its degradation . Recently, the lncRNA LOC100506114 was found to be overexpressed in the tumor stroma, indicating that expression is driven by mesenchymal cells. Subsequently, increased expression of LOC100506114 was found in CAFs isolated from patients in comparison with NF . Furthermore, functional analysis on tumor cells co-cultured with CAF-conditioned medium determined the increase in migration, proliferation, and expression of mesenchymal markers. Briefly, the studies showed that growth differentiation factor 10 (GDF10) promotes the functional transformation of an NF to a CAF via LOC100506114 that binds to the transcription factor RUNX2, which, in turn, participates in tumor growth, invasion, and metastasis . Some lncRNAs participate in the regulation of glucose metabolism of oral CAFs, as reported by Yang et al., where lncRNA H19 was identified to modulate glucose metabolism . When its expression is suppressed, it decreases glucose uptake and lactate secretion. It also regulates fundamental processes such as proliferation and migration. It has been reported that H19 exerts sponge or precursor functions of various miRNAs. In this case, it was reported to be a precursor of Hsa-miR-675 that interacts with the PFKFB3 gene in the glycolysis pathway in oral CAFs . Important processes such as angiogenesis and metastasis are regulated by lncRNAs in CAFs. However, they can result in a better or worse prognosis for patients, depending on the regulation at the gene level. For instance, patients who overexpress FOXF1 adjacent noncoding developmental regulatory RNA (FENDRR) have a better prognosis, because, when it is overexpressed, there is less migration in vitro and it also can regulate proangiogenic activity through the PI3K/AKT pathway . Conversely, a lncRNA associated with a poor prognosis is the new one called TIRY, which indirectly regulates cancer cells due to the effect of CAF-conditioned medium on tumor cells, where it was reported that TIRY is upregulated and facilitates increased invasion, migration, and metastasis in addition to acting as a miRNA sponge of miR-14 and inducing activation of the WNT/b-catenin pathway resulting in increased EMT . There are molecules secreted by CAFs that regulate the expression of lncRNAs in tumor cells as reported by Zhang et al., reporting that the Midkine molecule (MK) secreted by the tumor stroma regulates the expression of the lncRNA ANRIL and participates directly in the resistance to cisplatin, showing that CAF-conditioned medium in stimulated cancer cells induces cisplatin resistance, thus suggesting that the MK secreted by CAFs in a paracrine manner towards tumor cells regulates the resistance to cisplatin by inhibiting apoptosis . The use of lncRNAs as therapeutic targets has gained relevance in recent years since they could act in response to chemotherapy. For instance, it has been reported that, when lncRNA IL7R is silenced and a TLR3 inhibitor is used, tumor cells are more sensitive to treatment and apoptosis increases in epithelial cells cocultured with CAFs, in addition to increasing the immune infiltrate with immune cells associated with a better prognosis such as dendritic cells and CD8+ lymphocytes . 3.3. Immune-Related lncRNAs In TME, tumor cells interact with other cell populations such as CAFs, endothelial cells, and cells of the immune system through complex communication networks, enhancing tumor modulation of the microenvironment. Thus, TME plays an essential role in the initiation, tumor growth, invasion, and metastasis . In addition, the HNSCC TME is highly infiltrated by immune cells, which, depending on tumor biology, may mediate immune surveillance or evasion through various mechanisms . Recently, increasing evidence has revealed that lncRNAs regulate the immune response in TME by controlling the type of cellular infiltration, differentiation, and functions of immune cells , which can suppress or favor the progression of cancer. Hence, the study of the involvement of immune-related lncRNAs on the evolution of HNSCC has gained importance. Recent studies have identified immune-related lncRNAs in HNSCC impacting the prognosis of patients. Using bioinformatic tools, Chen et al. selected seven immune-related lncRNAs associated with OS: AL139158.2, AL031985.3, AC104794.2, AC099343.3, AL357519.1, SBDSP1, and AC108010.1. With these lncRNAs, they built a prognostic signature and classified HNSCC patients as high-risk. Furthermore, they identified that low-risk cases have a more significant infiltration of immune cells and enrichment of pathways associated with the immune response. In contrast, high-risk cases are related to the enrichment of metabolic pathways . This result is consistent with previous reports that identified nine immune-related lncRNAs in nasopharyngeal carcinoma, where low-risk patients have active pathways associated with the immune response and a greater intratumoral infiltrate of CD8+ T cells and B cells. In contrast, in high-risk patients, there is an association with pathways involved in metabolism . In the case of OSCC, previous research divided samples according to the expression of eight ferroptosis-related lncRNAs with implications in the prognosis. In the low-risk group, a significant decrease in AL512274.1, MIAT, and AC079921.2 was found, related to a more intense immune response compared with the high-risk group, where the expression of FIRRE, AC099850.3, and AC090246.1 increased . Multiple reports relate the cases of HNSCC that present a better prognosis with an active immune response, which can be associated with an abundant infiltrate of immune cells . However, there are tumor characteristics that can modify the expression of specific immune-related lncRNAs and, with this, induce an immunosuppressive TME. Mutations in the tumor suppressor genes TP53 and CDKN2A are frequent in HNSCC, and tumor cells that present these modifications can alter their pattern of expression of lncRNAs. At the same time, conditions such as hypoxia induce the expression of specific lncRNAs in immune cells. This crosstalk between tumor cells and immune cells induces the formation of an immunosuppressive TME . On the other hand, it has been observed that there are lncRNAs expressed in tumor cells that promote immune activation. PRINS, an overexpressed lncRNA in some cases of HPV-positive HNSCC, is related to the activation of genes involved in the immune response. Among HPV-positive tumors, those with higher PRINS levels are associated with a better prognosis . Due to this, research is focused on studying how lncRNAs regulate the differentiation and function of specific populations of immune cells in TME, which ultimately impacts tumor progression. Bioinformatic analysis of HNSCC samples has allowed the identification of various lncRNAs associated with genomic instability or with the immune response and related to the prognosis of patients for grouping the cases in low and high risk. Furthermore, since the inflammatory infiltrates in TME can exert a dual protumor function, the types of immune cell populations that infiltrate both groups of tumors have been studied. Different reports agree that in the low-risk group, there is a greater infiltrate of activated CD8+ T cells, activated CD4+ T cells, T follicular helper (T fh) cells , Treg cells , NK cells, B cells , and resting mast cells , as well as decreased numbers of M0 macrophages, activated mast cells , and CAFs . In contrast, the high-risk group is characterized by increased eosinophil infiltration, naive CD4+ T cells, resting NK cells, M0 macrophages , M2 macrophages , activated mast cells , and CAFs , accompanied by the decrease in the expression of human leukocyte antigen (HLA) molecules necessary to sustain the activation of the immune response. These findings show that the cell populations in low-risk cases are associated with an anti-tumor immune response, whereas an immunosuppressive microenvironment predominates in high-risk tumors. For a better understanding of the factors involved in the modulation of TME characteristics, the role played by some lncRNAs in the differentiation or polarization of immune cells has been studied. In LSCC, a considerable infiltrate of M2 macrophages plays a protumor role. HOTAIR is a lncRNA expressed in LSCC tumor cells and can be released into exosomes, which is related to M2 macrophage polarization via the downregulation of PTEN and the upregulation of PI3K and AKT expression. In in vitro analysis, the co-culture of M2 macrophages polarized with exosomes and LSCC cell lines increased the proliferation and migration of tumor cells. Interestingly, the in vivo injection of exosome-treated macrophages promoted an increase in tumor size, downregulation of the epithelial marker E-cadherin, and increased expression levels of the mesenchymal marker N-cadherin related to the EMT . Regarding M1 macrophages, with bioinformatic analysis of databases, it has been identified that in OSCC, LINC00460 is positively correlated with this cell phenotype and CASC9 is negatively correlated; both have a strong correlation with the prognosis of patients . Within the group of innate immune cells present in TME are neutrophils, the study of which has gained importance in recent years due to the impact that the functions of these cells have on tumor progression . In an evaluation of the lncRNAs associated with NETosis (formation of neutrophil extracellular traps (NETs)), in HNSCC, it was found that low-risk patients present enrichment of pathways associated with the immune response. At the same time, high-risk cases correspond to cold tumors associated with NETosis activation. Among the lncRNAs identified, LINC00426 is a protective factor. When nasopharyngeal carcinoma cell lines are transfected with this lncRNA, its overexpression significantly increased the expression levels of p-STING, p-TBK1, and p-IRF3. In addition, activation of the STING signaling pathway promotes the secretion of cytokines necessary for the recruitment of T cells and B cells, such as CXCL10, CCL5, ISG15, and ISG56 . MANCR is highly expressed in HNSCC tissue and cell lines, related to poor OS and disease-specific survival. In vitro, MANCR silencing inhibits the proliferation, migration, and invasion of HNSCC cell lines. However, in addition to acting as an oncogene, bioinformatic analyses have revealed that its expression is positively correlated with the infiltrate of neutrophils and gd T cells but negatively with the presence of CD8+ T cells and B cells . The type of inflammatory infiltrate is modulated by the cytokines secreted in the TME; this cytokine secretion can, in turn, be affected by the expression of lncRNAs. For example, BARX1-DT, KLHL7-DT, and LINC02154 are expressed in LSCC. These immune-related lncRNAs can promote an immunosuppressive TME by decreasing the expression of CCR3, CXCL9, and CXCL10, decreasing the recruitment of CD8+ T cells . There are immune-related lncRNAs that, in addition to being involved with the secretion of cytokines, are also related to the expression of other molecules necessary to mount the antitumor immune response. For example, TRG-AS1 is expressed in warm tumors with a high infiltration of cytotoxic cells, related to a better prognosis. It has been shown in vitro that the silencing of TRG-AS1 in an OSCC cell line suppresses the expression of HLA-A, HLA-B, and HLA-C molecules necessary for antigen presentation, as well as CXCL9, CXCL10, and CXCL11 . On the other hand, LINC02195 has high expression in the nucleus and cytoplasm of HNSCC cells, which is associated with a good prognosis as it has a positive correlation with the infiltrate of CD4+ T cells and CD8+ T cells, in addition to being involved in the expression of the MHC-I, antigen processing, and presentation . The set of molecules that regulate the immune response includes costimulatory and coinhibitory molecules that are also targets for modification by lncRNAs. IFITM4P progressively increases its expression from premalignant lesions such as OL to OSCC, thus acting as an oncogene. In a murine model of carcinogenesis in the tongue, lipopolysaccharides (LPS) bind to its receptor TLR4, which induces an increase in the expression of IFITM4P, acceleration of the carcinogenesis process, and immune escape through overexpression of the PD-L1 immunoregulatory ligand. IFITM4P induces PD-L1 expression in two different ways. In the cytoplasm it acts as a scaffold for the recruitment of SASH1, which binds and phosphorylates TAK1; this increases NF-kB phosphorylation, which ultimately induces PD-L1 expression. In the nucleus, IFITM4P reduces the transcription of PTEN by increasing the binding of KDM5A to its promoter and, with this, it upregulates PD-L1. In contrast, the overexpression of IFITM4P increases the sensitivity to treatment with PD-1 mAb . LncRNAs are associated with tumor immune evasion. LINC01123 is overexpressed in HNSCC tissue and cell lines, mainly in the cytoplasm of the cells, which, together with the overexpression of the immune checkpoint B7-H3, is associated with a poor prognosis by promoting tumor immune evasion. Furthermore, LINC01123 is competitively bound to miR-214-3p, and miR-214-3p, specifically targeting B7-H3; this inhibits CD8+ T cell activation and favors tumor progression. By silencing LINC01123 in HNSCC cell lines, the cytotoxic activity of CD8+ T cells increased, thereby decreasing tumorigenicity and increasing the secretion of factors associated with immune activation in vivo . LncRNAs also participate in sculpting the TME and include the activation or inhibition of specific pathways. For example, LINC01355 is overexpressed in OSCC and is associated with antitumor evasion by inhibiting the activity of CD8+ T cells through activation of the Notch pathway. Conversely, by deleting LINC01355 in OSCC cells, apoptosis of CD8+ T cell is retrained, proliferation and cytolysis activity is enhanced, and tumor cell proliferation, migration, and invasion are decreased . The expression of lncRNAs in TME immune system cells has been less studied. However, the reported evidence shows that they have an impact on tumor progression due to the bidirectional communication that exists between tumor cells and stromal cells. DCST1-AS1 is overexpressed in OSCC tumor cells and M2 macrophages. Silencing this lncRNA has been shown in vitro and in vivo to block NF-kB signaling, therefore repressing tumor cell emergence, migration, and invasion, as well as protumor M2 polarization of macrophages . In the case of CRNDE, it is expressed in OSCC, mainly in advanced stages in tumor cells and tumor-infiltrating T lymphocytes (TILs). Its expression in cancer cells exerts a protumor function by sponging miR-545-5p, which leads to increased expression of the immune checkpoint TIM-3 and suppresses the cytotoxicity of CD8+ T cells by contributing to their depletion . In a mouse model, injecting CD8+ T cells with CRNDE-knockdown decreases tumor size, increases the number of IFN-g and TNF-a-producing CD8+ T cells, decreases TIM-3 expression, and increases miR expression -545-5p, activating the antitumor immune response of CD8+ T lymphocytes . Finally, HOTTIP is a lncRNA expressed by HNSCC tumor cells and present in the exosomes of M1 macrophages. Although it has been associated with a protumor function, one study reported that exosomes from M1 macrophages, primarily through HOTTIP, inhibit HNSCC progression by activating the TLR5/NF-kB signaling pathway by competitively sponging miR-19a-3p and miR-19b-3p. In addition, they polarize circulating monocytes and TAMs toward an antitumor M1 phenotype, inducing positive feedback . 3.4. LncRNAs: Therapeutic Targets and Clinical Relevance in HNSCC Despite significant advances in the treatment of HNSCC, the mortality rate remains around 50% . It is essential to explore new therapeutic strategies to improve patients' time and quality of life. Lately, immunotherapy has received rising attention in cancer treatment for the OS advantages it offers; however, the overall response rate to immunotherapy in patients with HNSCC is less than 20% . The understanding of the molecular mechanisms that modify the characteristics of the TME can contribute to the detection of lncRNAs as novel biomarkers to provide new ideas for clinical diagnosis, immune-targeted therapy, and drug discovery . For example, identifying that HOTTIP polarizes circulating monocytes towards an antitumor M1 phenotype and suppresses HNSCC progression through the upregulation of the TLR5/NF-kB signaling pathway may provide novel insight into HNSCC immunotherapy . LncRNAs may function as potential therapeutic targets , as it has been reported that, when their lnc-IL7R function is suppressed, there is better sensitivity to chemotherapy in oral cancer cell lines . The mechanism by which IFITM4P induces PD-L1 expression is known, so this lncRNA may serve as a new therapeutic target in the blockage of oral carcinogenesis . However, for most of the lncRNAs that show alteration in expression in HNSCC, the exact mechanism by which TME conditions are modified remains to be unknown; for this reason, more studies are required to clarify this information to develop new therapeutic strategies . Hereby, we present data that show that some immune-related lncRNAs have clinical relevance, since AL139158.2, AL031985.3, AC104794.2, AC099343.3, AL357519.1, SBDSP1 , and AC108010.1 TM4SF19-AS1 have been associated with overall survival (OS). MANCR is also related to poor OS and disease-specific survival. MiR31HG , TM4SF19-AS1 , and LINC01123 are associated with poor prognosis. Meanwhile, LINC02195 and TRG-AS1 overexpression is associated with favorable prognosis. Moreover, ANRIL lncRNA induces resistance to cisplatin by inhibiting apoptosis. A superior understanding of the molecular mechanisms of lncRNAs that modify the characteristics of TME could contribute to increasing the efficacy of immunotherapy. 4. Conclusions LncRNAs involved in TME are clinically relevant, being indicators of survival, acting in important processes such as chemoresistance and being indicators of prognosis. The study of lncRNAs in cancer can contribute to a better understanding of the molecular mechanisms that modify the characteristics of TME, allowing the detection of possible therapeutic targets and biomarkers that contribute to the best selection of patients who are candidates for immunotherapy, resulting in the increase in efficacy of this type of treatment in HNSCC. Author Contributions L.J.B.-U., M.G.M.-B., and M.B. conceived the manuscript; L.J.B.-U., M.G.M.-B., C.E.V.-M., J.R.G.-B., A.L.P.-S., M.A.-M., and M.B. wrote the manuscript and analyzed data. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Interactions in the TME are mediated by crosstalk between cells of the immune system, stromal cells, and tumor cells, primarily regulating the behavior of the immune system to induce tumor escape. The image shows examples of crosstalk between some cellular components of the TME through communication with lncRNAs and the effect that is observed. Figure 2 Potential lncRNAs that could be targeted. In the green background the lncRNAs associated to inhibition of tumorigenesis are listed; the therapeutic strategies could be directed to induce them within the TME. In the pink background there is a list of lncRNAs related to promotion of tumorigenesis in HNSCC; the therapeutic strategy could be directed to inhibit their expression within the TME. cells-12-00727-t001_Table 1 Table 1 Overview of lncRNAs in the HNSCC tumor microenvironment. lncRNA Status of Expression Model Participation in HNSCC MiR31HG Upregulated LSCC cancer tissue Plays an oncogenic role and its overexpression can serve as a poor prognosis marker . USP2-AS1 In vitro model (HNSCC cell lines) Inhibits cellular senescence, acts as an oncogenic molecule, and promotes progression through proliferation, tumor growth, and invasion . TM4SF19-AS1 In vitro model (HNSCC cell lines), RNA sequencing dataset Acts like sponge del miR-153-3p , associated with OS and prognosis . cLINC00460 HNSCC tissues and in vitro model (HNSCC cell lines) Regulates cancer progression and mesenchymal marker expression in CAFs . HCG18 Tissue samples of HNSCC, HNSCC cell lines, and xenograft model/in vitro model (laryngeal and hypopharyngeal squamous cell carcinoma cell lines) Promotes cell proliferation and metastasis and modulates progression through the WNT signaling pathway . HNSCAT1 In vitro model of primary keratinocytes Overexpression of HNSCAT1 significantly inhibited tumor progression through HNSCAT1 interaction with miR-1254 . LURAP1L-AS1 In vitro model of oral fibroblasts Activation of the canonical NF-kB pathway, inducing the transformation of NFs (normal fibroblasts) into CAFs . FLJ22447/lncRNA-CAF In vitro model of oral squamous cell carcinoma (OSCC) (primary culture of CAF) and OSCC cell line Regulate IL-33 levels and prevented p62-dependent autophagy-lysosome degradation of IL-33 . LOC100506114 In vitro model of OSCC (primary culture of CAFs) Regulates fibroblast activation and promotes OSCC cell proliferation and migration through activation of TGFbR1/X2 and migration through activation of the TGFbR1/Smad3/ERK pathway of OSCC cells . H19 In vitro model of OSCC (primary culture of CAFs) Regulates the expression of enzymes, regulatory molecules, and oncogenes and/or oncogenes that indirectly modulate pathways involved in glucometabolic processes . TIRY In vitro model of OSCC (primary culture of CAFs) It acts as a miRNA sponge and downregulates miR-14 expression, promoting invasion and metastasis through WNT-b-catenin activation in oral cancer cells . ANRIL In vitro model (OSCC cell lines) Encodes 3 tumor-suppressor proteins, p15INK4b, p14ARF, and p16INK4a; its transcription is a key requirement for replicative or oncogene-induced senescence and constitutes an important barrier for tumor growth . LncRNA-IL17R In vivo model of OSCC Regulate response to chemotherapy, and cancer progression . PRINS HNSCC RNA sequencing datasets High expression in HPV-positive patients is associated with better OS. Is involved in the immune mechanisms, in mounting an antiviral response by affecting some pattern recognition receptors (PRRs) . HOTAIR Tissue samples; in vitro and in vivo models of LSCC Highly expressed in the advanced clinical stages of LSCC . Exosomal HOTAIR induces macrophages to M2 polarization by PI3K/p-AKT/AKT signaling pathway and these M2 macrophages facilitate the migration, proliferation, and EMT of LSCC in vitro and in vivo . MANCR Tissue samples and in vitro model of HNSCC Is a high-risk factor in patients with HNSCC. Is associated with peripheral nerves and the extracellular matrix for highly expressed genes and hence may play a crucial role in the occurrence of HNSCC . BARX1-DT KLHL7-DT LINC02154 RNA sequencing datasets and in vitro model of LSCC Patients with LSCC and high expression of BARX1-DT , KLHL7-DT, and LINC02154 have worse OS. These lncRNAs may boost the development of an immunosuppressive TME by downregulating the expression of key immunomodulators such as CCR3, CXCL10, and CXCL9 and subsequently decreasing the recruitment of effector CD8+ T cells . TRG-AS1 RNA sequencing datasets and in vitro model of HNSCC The high expression indicates a favorable prognosis in HNSCC. Is an essential lncRNA involving TME formation. Knockdown of TRG-AS1 inhibited the expression of HLA-A, HLA-B, HLA-C, CXCL9, CXCL10, and CXCL11 in vitro . LINC02195 RNA sequencing datasets, tissue samples, and in vitro model of HNSCC There is a correlation between high LINC02195 expression and favorable prognosis in HNSCC. Is associated with genes encoding MHC-I molecules, antigen processing, and presentation and is related to an increased number of CD8+ and CD4+ T cells . IFITM4P Oral leukoplakia (OL) and OSCC tissue samples, in vitro and in vivo models of OL and HNSCC Acts as a scaffold to facilitate the recruitment of SASH1 to bind and phosphorylate TAK1 and further increase the phosphorylation of NF-kB to induce PD-L1 transcription, hence promoting immune evasion . LINC01123 Tissue samples, in vitro and in vivo models of HNSCC High expression is associated with poor prognosis in patients with HNSCC. Acts as a miR-214-3p sponge to inhibit the activation of CD8+ T cells and promote tumor immune escape by upregulating B7-H3 . LINC01355 In vitro and in vivo models of OSCC Could induce the development of OSCC via modulating the Notch signal pathway that represses CD8+ T cell activity . DCST1-AS1 In vitro and in vivo models of OSCC Contributes to cancer progression by enhancing the NF-kB signaling pathway to promote OSCC development and M2 macrophage polarization . CRNDE Tissue samples, in vitro and in vivo models of OSCC The expression is higher in stage IV of OSCC than early stages. Can exhibit a crucial role in activating CD8+ T cell exhaustion by sponge miR-545-5p to induce TIM-3 expression . HOTTIP RNA sequencing datasets, in vitro/in vivo models of HNSCC Is highly expressed in stages III-IV of HNSCC . Overexpression of HOTTIP inhibits HNSCC progression and induces the polarization of M1 macrophages because it activates the TLR5/ NF-kB signaling pathway by competitively sponging miR-19a-3p and miR-19b-3p . FENDRR Downregulated In vitro model of OSCC (primary culture of CAF) Downregulation of FENDRR can activate the PI3K/AKT pathway in NFs and increases matrix metalloproteinase 9 (MMP9) expression . LINC00426 Downregulated in nasopharyngeal carcinoma cell lines CNE1, HNE1, and TW03 HNSCC RNA sequencing datasets and in vitro model with nasopharyngeal carcinoma cell lines Contributes to the innate immune cGAS-STING signaling pathway, related to the secretion of cytokines to recruit B cells and T cells, and promoting immune cell infiltration . Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Siegel R.L. Miller K.D. Jemal A. Cancer statistics, 2022 CA Cancer J. Clin. 2022 72 7 33 10.3322/caac.21708 35020204 2. Sung H. Ferlay J. Siegel R.L. Laversanne M. Soerjomataram I. Jemal A. Bray F. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries CA Cancer J. Clin. 2021 71 209 249 10.3322/caac.21660 33538338 3. Johnson D.E. Burtness B. Leemans C.R. Lui VW Y. Bauman J.E. Grandis J.R. Head and neck squamous cell carcinoma Nat. Rev. Dis. Primers 2020 6 92 10.1038/s41572-020-00224-3 33243986 4. Pena-Flores J.A. Bermudez M. Ramos-Payan R. Villegas-Mercado C.E. Soto-Barreras U. Muela-Campos D. Alvarez-Ramirez A. Perez-Aguirre B. Larrinua-Pacheco A.D. Lopez-Camarillo C. Emerging role of lncRNAs in drug resistance mechanisms in head and neck squamous cell carcinoma Front. Oncol. 2022 12 3750 10.3389/fonc.2022.965628 5. Contreras W. Venegas B. Virus papiloma humano en cancer oral y orofaringeo: Revision de la literatura Int. J. Odontostomatol. 2015 9 427 435 10.4067/S0718-381X2015000300012 6. Chow L.Q.M. Head and Neck Cancer N. Engl. J. Med. 2020 382 60 72 10.1056/NEJMra1715715 31893516 7. Di Credico G. Polesel J. Maso L.D. Pauli F. Torelli N. Luce D. Radoi L. Matsuo K. Serraino D. Brennan P. Alcohol drinking and head and neck cancer risk: The joint effect of intensity and duration Br. J. Cancer 2020 123 1456 1463 10.1038/s41416-020-01031-z 32830199 8. Dhull A.K. Atri R. Dhankhar R. Chauhan A.K. Kaushal V. Major Risk Factors in Head and Neck Cancer: A Retrospective Analysis of 12-Year Experiences World J. Oncol. 2018 9 80 84 10.14740/wjon1104w 29988794 9. Granados-Garcia M. Oropharyngeal cancer: An emergent disease? Salud Publica Mexico 2016 58 285 290 10.21149/spm.v58i2.7798 27557387 10. Dok R. Nuyts S. HPV Positive Head and Neck Cancers: Molecular Pathogenesis and Evolving Treatment Strategies Cancers 2016 8 41 10.3390/cancers8040041 27043631 11. He X. Xiao Y. Liu S. Deng R. Li Z. Zhu X. Predicting the Immune Microenvironment and Prognosis with a NETosis-Related lncRNA Signature in Head and Neck Squamous Cell Carcinoma BioMed Res. Int. 2022 2022 3191474 10.1155/2022/3191474 36147630 12. Leemans C.R. Braakhuis B.J.M. Brakenhoff R.H. The molecular biology of head and neck cancer Nat. Rev. Cancer 2011 11 9 22 10.1038/nrc2982 21160525 13. Hanahan D. Coussens L.M. Accessories to the crime: Functions of cells recruited to the tumor microenvironment Cancer Cell 2012 21 309 322 10.1016/j.ccr.2012.02.022 22439926 14. Hanahan D. Weinberg R.A. Hallmarks of cancer: The next generation Cell 2011 144 646 674 10.1016/j.cell.2011.02.013 21376230 15. Gavrielatou N. Doumas S. Economopoulou P. Foukas P.G. Psyrri A. Biomarkers for immunotherapy response in head and neck cancer Cancer Treat. Rev. 2020 84 101977 10.1016/j.ctrv.2020.101977 32018128 16. Allen C.T. Judd N.P. Bui J.D. Uppaluri R. The clinical implications of antitumor immunity in head and neck cancer Laryngoscope 2012 122 144 157 10.1002/lary.21913 22183634 17. Dunn G.P. Bruce A.T. Ikeda H. Old L.J. Schreiber R.D. Cancer immunoediting: From immunosurveillance to tumor escape Nat. Immunol. 2002 3 991 998 10.1038/ni1102-991 12407406 18. Dunn G.P. Old L.J. Schreiber R.D. The Immunobiology of Cancer Immunosurveillance and Immunoediting Immunity 2004 21 137 148 10.1016/j.immuni.2004.07.017 15308095 19. Chen X. Yan B. Lou H. Shen Z. Tong F. Zhai A. Wei L. Zhang F. Immunological network analysis in HPV associated head and neck squamous cancer and implications for disease prognosis Mol. Immunol. 2018 96 28 36 10.1016/j.molimm.2018.02.005 29477933 20. Sun Q. Li Y. Yang X. Wu X. Liu Z. Mou Y. Song X. Identification and Validation of 17-lncRNA Related to Regulatory T Cell Heterogeneity as a Prognostic Signature for Head and Neck Squamous Cell Carcinoma Front. Immunol. 2021 12 782216 10.3389/fimmu.2021.782216 34880875 21. Tampa M. Mitran M.I. Mitran C.I. Sarbu M.I. Matei C. Nicolae I. Caruntu A. Tocut S.M. Popa M.I. Caruntu C. Mediators of Inflammation--A Potential Source of Biomarkers in Oral Squamous Cell Carcinoma J. Immunol. Res. 2018 2018 1061780 10.1155/2018/1061780 30539028 22. Wang G. Zhang M. Cheng M. Wang X. Li K. Chen J. Chen Z. Chen S. Chen J. Xiong G. Tumor microenvironment in head and neck squamous cell carcinoma: Functions and regulatory mechanisms Cancer Lett. 2021 507 55 69 10.1016/j.canlet.2021.03.009 33741424 23. Bozgeyik E. Bozgeyik I. Non-coding RNA variations in oral cancers: A comprehensive review Gene 2023 851 147012 10.1016/j.gene.2022.147012 36349577 24. Du Y. Ji Z. Liao J. Liu H. Peng H. Identification of Molecular Subtypes in Head and Neck Squamous Cell Carcinoma Based on Dysregulated Immune LncRNAs J. Oncol. 2022 2022 9702789 10.1155/2022/9702789 35126519 25. Taft R.J. Pang K.C. Mercer T.R. Dinger M. Mattick J.S. Non-coding RNAs: Regulators of disease J. Pathol. 2010 220 126 139 10.1002/path.2638 19882673 26. Whiteside T.L. The tumor microenvironment and its role in promoting tumor growth Oncogene 2008 27 5904 5912 10.1038/onc.2008.271 18836471 27. Bermudez M. Aguilar-Medina M. Lizarraga-Verdugo E. Avendano-Felix M. Silva-Benitez E. Lopez-Camarillo C. Ramos-Payan R. LncRNAs as Regulators of Autophagy and Drug Resistance in Colorectal Cancer Front. Oncol. 2019 9 1008 10.3389/fonc.2019.01008 31632922 28. Huang Q. You Q. Zhu N. Wu Z. Xiang Z. Wu K. Ren J. Gui Y. Prognostic prediction of head and neck squamous cell carcinoma: Construction of cuproptosis-related long non-coding RNA signature J. Clin. Lab. Anal. 2022 36 e24723 10.1002/jcla.24723 36189780 29. Guglas K. Bogaczynska M. Kolenda T. Rys M. Teresiak A. Blizniak R. Lasinska I. Mackiewicz J. Lamperska K. lncRNA in HNSCC: Challenges and potential Contemp. Oncol. 2017 21 259 266 10.5114/wo.2017.72382 30. Hombach S. Kretz M. Non-coding RNAs: Classification, Biology and Functioning Adv. Exp. Med. Biol. 2016 937 3 17 27573892 31. Wapinski O. Chang H.Y. Long noncoding RNAs and human disease Trends Cell Biol. 2011 21 354 361 10.1016/j.tcb.2011.04.001 21550244 32. Cao R. Cui L. Zhang J. Ren X. Bin Cheng B. Xia J. Immune-related lncRNA classification of head and neck squamous cell carcinoma Cancer Cell Int. 2022 22 25 10.1186/s12935-022-02450-z 35033066 33. Wang L. Yang G. Liu G. Pan Y. Identification of lncRNA Signature of Tumor-Infiltrating T Lymphocytes With Potential Implications for Prognosis and Chemotherapy of Head and Neck Squamous Cell Carcinoma Front. Pharmacol. 2021 12 795205 10.3389/fphar.2021.795205 35242027 34. Wilusz J.E. Sunwoo H. Spector D.L. Long noncoding RNAs: Functional surprises from the RNA world Genes Dev. 2009 23 1494 1504 10.1101/gad.1800909 19571179 35. Wang R. Ma Z. Feng L. Yang Y. Tan C. Shi Q. Lian M. He S. Ma H. Fang J. LncRNA MIR31HG targets HIF1A and P21 to facilitate head and neck cancer cell proliferation and tumorigenesis by promoting cell-cycle progression Mol. Cancer 2018 17 162 10.1186/s12943-018-0916-8 30458787 36. Tang J. Wu Z. Wang X. Hou Y. Bai Y. Tian Y. Hypoxia-Regulated lncRNA USP2-AS1 Drives Head and Neck Squamous Cell Carcinoma Progression Cells 2022 11 3407 10.3390/cells11213407 36359803 37. Jiang P. He S. Li Y. Xu Z. Identification of Therapeutic and Prognostic Biomarkers of Lamin C (LAMC) Family Members in Head and Neck Squamous Cell Carcinoma Experiment 2020 26 e925735 10.12659/MSM.925735 38. Yu Z. Wang X. Niu K. Sun L. Li D. LncRNA TM4SF19-AS1 exacerbates cell proliferation, migration, invasion, and EMT in head and neck squamous cell carcinoma via enhancing LAMC1 expression Cancer Biol. Ther. 2022 23 1 9 10.1080/15384047.2022.2116923 36411963 39. Zhang C. Cao W. Wang J. Liu J. Liu J. Wu H. Li S. Zhang C. A prognostic long non-coding RNA-associated competing endogenous RNA network in head and neck squamous cell carcinoma PeerJ 2020 8 e9701 10.7717/peerj.9701 32983633 40. Jiang Y. Cao W. Wu K. Qin X. Wang X. Li Y. Yu B. Zhang Z. Wang X. Yan M. LncRNA LINC00460 promotes EMT in head and neck squamous cell carcinoma by facilitating peroxiredoxin-1 into the nucleus J. Exp. Clin. Cancer Res. 2019 38 365 10.1186/s13046-019-1364-z 31429766 41. Mao B. Wang F. Zhang J. Li Q. Ying K. Long non-coding RNA human leucocyte antigen complex group-18 HCG18 (HCG18) promoted cell proliferation and migration in head and neck squamous cell carcinoma through cyclin D1-WNT pathway Bioengineered 2022 13 9425 9434 10.1080/21655979.2022.2060452 35389764 42. Peng H. Ge P. Long non-coding RNA HCG18 facilitates the progression of laryngeal and hypopharyngeal squamous cell carcinoma by upregulating FGFR1 via miR-133b Mol. Med. Rep. 2022 25 46 10.3892/mmr.2021.12562 34878161 43. Zhao Y. Huang X. Zhang Z. Li H. Zan T. The Long Noncoding Transcript HNSCAT1 Activates KRT80 and Triggers Therapeutic Efficacy in Head and Neck Squamous Cell Carcinoma Oxidative Med. Cell. Longev. 2022 2022 4156966 10.1155/2022/4156966 35965679 44. Ren X. Li L. Wu J. Lin K. He Y. Bian L. PDGF-BB regulates the transformation of fibroblasts into cancer-associated fibroblasts via the lncRNA LURAP1L-AS1/LURAP1L/IKK/IkB/NF-kB signaling pathway Oncol. Lett. 2021 22 537 10.3892/ol.2021.12798 34079593 45. Ding L. Ren J. Zhang D. Li Y. Huang X. Hu Q. Wang H. Song Y. Ni Y. Hou Y. A novel stromal lncRNA signature reprograms fibroblasts to promote the growth of oral squamous cell carcinoma via LncRNA-CAF/interleukin-33 Carcinogenesis 2018 39 397 406 10.1093/carcin/bgy006 29346528 46. Zhang D. Song Y. Li D. Liu X. Pan Y. Ding L. Shi G. Wang Y. Ni Y. Hou Y. Cancer-associated fibroblasts promote tumor progression by lncRNA-mediated RUNX2/GDF10 signaling in oral squamous cell carcinoma Mol. Oncol. 2022 16 780 794 10.1002/1878-0261.12935 33657265 47. Fan C. Tang Y. Wang J. Xiong F. Guo C. Wang Y. Zhang S. Gong Z. Wei X. Zhaoyang Z. Role of long non-coding RNAs in glucose metabolism in cancer Mol. Cancer 2017 16 130 10.1186/s12943-017-0699-3 28738810 48. Yang J. Shi X. Yang M. Luo J. Gao Q. Wang X. Wu Y. Tian Y. Wu F. Zhou H. Glycolysis reprogramming in cancer-associated fibroblasts promotes the growth of oral cancer through the lncRNA H19/miR-675-5p/PFKFB3 signaling pathway Int. J. Oral Sci. 2021 13 12 10.1038/s41368-021-00115-7 33762576 49. Jin N. Jin N. Bu W. Li X. Liu L. Wang Z. Tong J. Li D. Long non-coding RNA TIRY promotes tumor metastasis by enhancing epithelial-to-mesenchymal transition in oral cancer Exp. Biol. Med. 2020 245 585 596 10.1177/1535370220903673 50. Aguilo F. Zhou M.M. Walsh M.J. Long noncoding RNA, polycomb, and the ghosts haunting INK4b-ARF-INK4a expression Cancer Res. 2011 71 5365 5369 10.1158/0008-5472.CAN-10-4379 21828241 51. Ding L. Ren J. Zhang D. Li Y. Huang X. Ji J. Hu Q. Wang H. Ni Y. Hou Y. The TLR3 Agonist Inhibit Drug Efflux and Sequentially Consolidates Low-Dose Cisplatin-Based Chemoimmunotherapy while Reducing Side Effects Mol. Cancer Ther. 2017 16 1068 1079 10.1158/1535-7163.MCT-16-0454 28138030 52. Kopczynska M. Kolenda T. Guglas K. Sobocinska J. Teresiak A. Blizniak R. Mackiewicz A. Mackiewicz J. Lamperska K. PRINS lncRNA Is a New Biomarker Candidate for HPV Infection and Prognosis of Head and Neck Squamous Cell Carcinomas Diagnostics 2020 10 762 10.3390/diagnostics10100762 32998396 53. Wang J. Zhou Y. Lu J. Sun Y. Xiao H. Liu M. Tian L. Combined detection of serum exosomal miR-21 and HOTAIR as diagnostic and prognostic biomarkers for laryngeal squamous cell carcinoma Med. Oncol. 2014 31 148 10.1007/s12032-014-0148-8 25099764 54. Wang J. Wang N. Zheng Z. Che Y. Suzuki M. Kano S. Lu J. Wang P. Sun Y. Homma A. Exosomal lncRNA HOTAIR induce macrophages to M2 polarization via PI3K/ p-AKT /AKT pathway and promote EMT and metastasis in laryngeal squamous cell carcinoma BMC Cancer 2022 22 1208 10.1186/s12885-022-10210-5 36424539 55. Tang J. Bao M. Chen J. Bin X. Xu X. Fang X. Tang Z. Long-Noncoding RNA MANCR is Associated With Head and Neck Squamous Cell Carcinoma Malignant Development and Immune Infiltration Front. Genet. 2022 13 911733 10.3389/fgene.2022.911733 35873456 56. Wang X. Pan Y. Ou Y. Duan T. Zou Y. Zhou X. Construction and validation of immune-related LncRNAs classifier to predict prognosis and immunotherapy response in laryngeal squamous cell carcinoma World J. Surg. Oncol. 2022 20 164 10.1186/s12957-022-02608-z 35606755 57. Gong S. Xu M. Zhang Y. Shan Y. Zhang H. The Prognostic Signature and Potential Target Genes of Six Long Non-coding RNA in Laryngeal Squamous Cell Carcinoma Front. Genet. 2020 11 413 10.3389/fgene.2020.00413 32411183 58. Li H. Xiong H.-G. Xiao Y. Yang Q.-C. Yang S.-C. Tang H.-C. Zhang W.-F. Sun Z.-J. Long Non-coding RNA LINC02195 as a Regulator of MHC I Molecules and Favorable Prognostic Marker for Head and Neck Squamous Cell Carcinoma Front. Oncol. 2020 10 615 10.3389/fonc.2020.00615 32435615 59. Shi L. Yang Y. Li M. Li C. Zhou Z. Tang G. Wu L. Yao Y. Shen X. Hou Z. LncRNA IFITM4P promotes immune escape by up-regulating PD-L1 via dual mechanism in oral carcinogenesis Mol. Ther. 2022 30 1564 1577 10.1016/j.ymthe.2022.01.003 35051616 60. Li H. Yang Z. Yang X. Zhang F. Wang J. Wu Z. Wanyan C. Meng Q. Gao W. Yang X. LINC01123 promotes immune escape by sponging miR-214-3p to regulate B7-H3 in head and neck squamous-cell carcinoma Cell Death Dis. 2022 13 109 10.1038/s41419-022-04542-0 35115487 61. Zou C. Wu S. Wei H. Luo H. Tang Z. Li X. Lv X. Ai Y. LINC01355 Contributes to Malignant Phenotype of Oral Squamous Cell Carcinoma and Cytotoxic T Cell Infiltration via Activating Notch Signaling Pathway J. Immunol. Res. 2021 2021 1830790 10.1155/2021/1830790 34355042 62. Ai Y. Liu S. Luo H. Wu S. Wei H. Tang Z. Li X. Zou C. lncRNA DCST1-AS1 Facilitates Oral Squamous Cell Carcinoma by Promoting M2 Macrophage Polarization through Activating NF-kB Signaling J. Immunol. Res. 2021 2021 5524231 10.1155/2021/5524231 34414241 63. Ai Y. Wu S. Gao H. Wei H. Tang Z. Li X. Zou C. Repression of CRNDE enhances the anti-tumour activity of CD8 + T cells against oral squamous cell carcinoma through regulating miR-545-5p and TIM-3 J. Cell Mol. Med. 2021 25 10857 10868 10.1111/jcmm.16909 34729919 64. Yin X. Yang W. Xie J. Wei Z. Tang C. Song C. Wang Y. Cai Y. Xu W. Han W. HOTTIP Functions as a Key Candidate Biomarker in Head and Neck Squamous Cell Carcinoma by Integrated Bioinformatic Analysis BioMed Res. Int. 2019 2019 5450617 10.1155/2019/5450617 31032351 65. Jiang H. Zhou L. Shen N. Ning X. Wu D. Jiang K. Huang X. M1 macrophage-derived exosomes and their key molecule lncRNA HOTTIP suppress head and neck squamous cell carcinoma progression by upregulating the TLR5/NF-kB pathway Cell Death Dis. 2022 13 183 10.1038/s41419-022-04640-z 35210436 66. Xu Y. Jiang E. Shao Z. Shang Z. LncRNA FENDRR in Carcinoma-Associated Fibroblasts Regulates the Angiogenesis of Oral Squamous Cell Carcinoma Through the PI3K/AKT Pathway Front. Oncol. 2021 11 616576 10.3389/fonc.2021.616576 34327132 67. Zhang D. Wang Y. Shi Z. Liu J. Sun P. Hou X. Zhang J. Zhao S. Zhou B.P. Mi J. Metabolic reprogramming of cancer-associated fibroblasts by IDH3a downregulation Cell Rep. 2015 10 1335 1348 10.1016/j.celrep.2015.02.006 25732824 68. Zhang D. Ding L. Li Y. Ren J. Shi G. Wang Y. Zhao S. Ni Y. Hou Y. Midkine derived from cancer-associated fibroblasts promotes cisplatin-resistance via up-regulation of the expression of lncRNA ANRIL in tumour cells Sci. Rep. 2017 7 16231 10.1038/s41598-017-13431-y 29176691 69. Mowel W. Kotzin J.J. McCright S. Neal V.D. Henao-Mejia J. Control of Immune Cell Homeostasis and Function by lncRNAs Trends Immunol. 2018 39 55 69 10.1016/j.it.2017.08.009 28919048 70. Chen Y. Luo T.Q. Xu S.S. Chen C.Y. Sun Y. Lin L. Mao Y.P. An immune-related seven-lncRNA signature for head and neck squamous cell carcinoma Cancer Med. 2021 10 2268 2285 10.1002/cam4.3756 33660378 71. Liang Y.-L. Zhang Y. Tan X.-R. Qiao H. Liu S.-R. Tang L.-L. Mao Y.-P. Chen L. Li W.-F. Zhou G.-Q. A lncRNA signature associated with tumor immune heterogeneity predicts distant metastasis in locoregionally advanced nasopharyngeal carcinoma Nat. Commun. 2022 13 2996 10.1038/s41467-022-30709-6 35637194 72. Qiu L. Tao A. Liu F. Ge X. Li C. Potential prognostic value of a eight ferroptosis-related lncRNAs model and the correlative immune activity in oral squamous cell carcinoma BMC Genet. Data 2022 23 80 10.1186/s12863-022-01097-z 36384476 73. Ma B. Jiang H. Luo Y. Liao T. Xu W. Wang X. Dong C. Ji Q. Wang Y. Tumor-Infiltrating Immune-Related Long Non-Coding RNAs Indicate Prognoses and Response to PD-1 Blockade in Head and Neck Squamous Cell Carcinoma Front. Immunol. 2021 12 692079 10.3389/fimmu.2021.692079 34737735 74. Jing L. Du Y. Fu D. Characterization of tumor immune microenvironment and cancer therapy for head and neck squamous cell carcinoma through identification of a genomic instability-related lncRNA prognostic signature Front. Genet. 2022 13 979575 10.3389/fgene.2022.979575 36105083 75. Liu Z. Liu Q. Wang X. Liu L. Shi L. Li H. Multiomics Immune-Related lncRNA Analysis of Oral Squamous Cell Carcinoma and Its Correlation with Prognosis Dis. Markers 2022 2022 6106503 10.1155/2022/6106503 36118668 76. Wang X. Cao K. Guo E. Mao X. Guo L. Zhang C. Guo J. Wang G. Yang X. Sun J. Identification of Immune-Related LncRNA Pairs for Predicting Prognosis and Immunotherapeutic Response in Head and Neck Squamous Cell Carcinoma Front. Immunol. 2021 12 658631 10.3389/fimmu.2021.658631 33995377 77. Qian L. Ni T. Fei B. Sun H. Ni H. An immune-related lncRNA pairs signature to identify the prognosis and predict the immune landscape of laryngeal squamous cell carcinoma BMC Cancer 2022 22 545 10.1186/s12885-022-09524-1 35568824 78. Zhou C. Shen Y. Jin Y. Shen Z. Ye D. Shen Y. Deng H. A novel Pyroptosis-related long non-coding RNA signature for predicting the prognosis and immune landscape of head and neck squamous cell carcinoma Cancer Med. 2022 11 5097 5112 10.1002/cam4.4819 35567376 79. Li C. Zhao L. Wang Q. Ma S. Sun J. Ma C. Liu J. Jing X. Ai D. Nan Z. Neutrophils infiltration and its correlation with human papillomavirus status in the oral squamous cell carcinoma Cancer Manag. Res. 2019 11 5171 5185 10.2147/CMAR.S202465 31239772 80. Hanna G.J. Lizotte P. Cavanaugh M. Kuo F.C. Shivdasani P. Frieden A. Chau N.G. Schoenfeld J.D. Lorch J.H. Uppaluri R. Frameshift events predict anti-PD-1/L1 response in head and neck cancer JCI Insight 2018 3 e98811 10.1172/jci.insight.98811 29467336 |
PMC10000591 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050970 diagnostics-13-00970 Systematic Review Subepidermal Low-Echogenic Band--Its Utility in Clinical Practice: A Systematic Review Nicolescu Alin Codrut 1 Ionescu Sinziana 23 Ancuta Ioan 24 Popa Valentin-Tudor 5 Lupu Mihai 6 Soare Cristina 27 Cozma Elena-Codruta 78* Voiculescu Vlad-Mihai 27 Bini Fabiano Academic Editor 1 "Agrippa Ionescu" Emergency Clinical Hospital, 011773 Bucharest, Romania 2 Surgery Department, "Carol Davila" University of Medicine and Pharmacy, 050474 Bucharest, Romania 3 Surgery Department, "Prof. Dr. Al. Trestioreanu" Oncology Institute Bucharest, 022328 Bucharest, Romania 4 Rheumatology Department, "Dr. I. Cantacuzino" Clinical Hospital, 020475 Bucharest, Romania 5 Dermatology Department, Center for Morphologic Study of the Skin MORPHODERM, "Victor Babes" University of Medicine and Pharmacy, 300041 Timisoara, Romania 6 Department of Dermatology, MEDAS Medical Center, 030447 Bucharest, Romania 7 Dermatology Department, Elias University Emergency Hospital, 011461 Bucharest, Romania 8 Pathophysiology Department, University of Medicine and Pharmacy of Craiova, 200349 Craiova, Romania * Correspondence: [email protected] 03 3 2023 3 2023 13 5 97028 1 2023 17 2 2023 28 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). High-frequency ultrasonography (HF-USG) is a relatively new imaging method that allows the evaluation in a non-invasive manner of the skin layers and skin appendages. It is a diagnostic tool with increasing usefulness in numerous dermatological pathologies. High reproducibility, non-invasiveness and short diagnostic time make this method an increasingly used tool in dermatological practice. The subepidermal low-echogenic band is a relatively newly described parameter that seems to be a marker not only of intrinsic and extrinsic skin aging, but also of inflammatory processes taking place at the skin level. This systematic review aims to evaluate the role that SLEB has in the diagnosis and monitoring of the treatment of some inflammatory and non-inflammatory dermatological conditions, as well as its utility as a disease marker. high-frequency ultrasonography sub-epidermal low-echogenic band echogenicity non-invasive diagnosis atopic dermatitis This research received no external funding. pmc1. Introduction High-frequency ultrasonography (HF-USG) is a non-invasive imaging technique used since the second half of the 20th century (1979) to evaluate skin layers and structures. This is accomplished by using devices with high frequencies (over 15Mhz) which allow the examiner to increase the resolution of ultrasonographic images, sacrificing at the same time the depth of penetration . Although this is a method known for over a century, only in the last decade scientific attention has been directed towards it and the possibility of its use in numerous inflammatory or tumoral dermatological diseases . The possibility of assessing the thickness of epidermis and dermis, their echogenicity and density, as well as the visualization of a new parameter, namely, the subepidermal low-echogenic band (SLEB), opened the possibility of using this method not only for the descriptive purpose and evaluation of tumor margins, but also as the interface of the processes of inflammation and aging (intrinsic or photoinduced) that occur at the skin level . SLEB is described as a hypoechoic, well-defined band, located immediately under the epidermis layer. The two layers of the dermis (papillary and reticular), although different in composition and organization of the collagen fibers (in the case of the papillary dermis being more loosely packed), are very difficult to differentiate ultrasonographically. However, SLEB, by its hypoechoic appearance, seems to correspond to lax papillary dermis . Variations in its thickness and echogenicity are closely related to photo-aging processes, through the degradation of collagen fibers at this level, but also to the invasion of this area by inflammatory cells and the accumulation of water molecules (edema). Thus, SLEB has become an increasingly frequent parameter in cutaneous HF-USG studies, regardless of the studied disease . This work represents the first systematic review that highlights the role and usefulness of SLEB in the diagnosis, staging, monitoring of disease activity, as well as in determining the response to topical therapies in various dermatological diseases. 2. Materials and Methods This systematic review is based on the protocol elaborated by Moher et al. regarding "The Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA)" . The investigators performed an advanced search in several international databases (PubMed/Medline, Web of Science, SCOPUS) for articles published during 2012-2022, using the word combination" subepidermal low echogenic band". Several inclusion criteria had to be met for articles to be included in this systematic review, namely: 1. the papers were original articles, research letters, or case series that evaluated the role of SLEB in the diagnosis, prognosis, description of the stage of a disease or its role in monitoring different invasive dermatological procedures in physiological processes (such as skin aging) or pathological processes (inflammatory, autoimmune and neoplastic diseases); 2. the selected papers were presented in a language spoken by the authors (English, French or Romanian); 3. the selected papers were published after 1 January 2012; 4. the full text of the articles/case series/research letters could be obtained. Exclusion criteria included the following: 1. the papers represented reviews, single case reports, letters to the editors, or abstracts presented at scientific meetings; 2. the papers were written in a language not known by any of the investigators; 3. the studies were performed on cell cultures or animals; 4. the studies were published before 1 January 2012; 5. the investigators could not obtain the full-text version of the articles; 6. the papers did not offer sufficient information regarding the role of SLEB in the diagnosis, prognosis, description of diseases or in the monitoring of invasive dermatological procedures. The selection process of the articles to be included consisted in a preliminary scan of the titles of the articles resulting from the search based on the combination of words mentioned above; then, the next step involved the reading of the abstracts and the full-text articles of the works resulting from the first scan. A database in Microsoft Excel was also created which included the following key information on the selected studies: first author name, year of publication, country where the study was conducted, study design, number of patients included, dermatological conditions studied by ultrasound, frequency of the ultrasound probe used, reporting of the level at which the SLEB measurement was made, pathological changes of the SLEB. In addition, the investigators ensured the scientific quality of the selected studies and assessed the risk of bias by using two essential tools, namely, the Methodological Index for Non-Randomized Observational Studies (MINORS), and the Mixed Methods Appraisal Tool (MMAT) . 3. Results Figure 1 represents the diagram of the selection process of the 24 articles included in this systematic review. From the initial group of 39 articles resulting from the use of the search key, duplicates and, subsequently, articles whose titles and/or abstracts did not match our research question were eliminated, resulting in 34 articles that were read in full text. Of these, 24 papers met the inclusion criteria and are discussed below (Table 1). 3.1. SLEB and Atopic Dermatitis (AD) A total of three studies evaluated the role of SLEB in establishing the severity of the examined disease and in monitoring therapy efficacy as well as the best moment for switching from a reactive to a proactive treatment in patients with atopic dermatitis (AD) . In an observational study, Polanska et al. observed a group of 39 patients (21 women, mean age 26.3 +- 12.8 years old) with AD who underwent reactive (one application/day, between 1 and 6 weeks) and proactive treatment (2-3 applications/weeks) with tacrolimus 0.1%, followed by the evaluation of the group clinically, evaporimetrically and by HF-USG. The patients were evaluated at the initial presentation and then monthly, up to 6 months of treatment, examining both the affected area (right antecubital fold) and an area of apparently normal skin, 10 cm away from the affected area. Regarding the HF-USG measurements, significant differences were found in the size of SLEB, both in the affected skin and in apparently intact skin. Thus, SLEB decreased from 0.227 mm to 0.03 mm after 24 weeks in the affected skin (p < 0.001) and from 0.03 mm to 0.00 mm (p < 0.001) after 24 weeks, respectively, in the apparently healthy skin. In addition, Polanska et al. observed a decrease in the number of patients in which SLEB was visualized: from 100% to 18% (p < 0.001) in the affected skin and from 20.5% to 5.1% (p < 0.001) in the healthy skin after 24 weeks of treatment . Another study conducted by Sorokina et al. on a group of 22 children with AD (14 females, mean age 3.7 +- 2.4 years) and a control group of 18 healthy children (11 females, mean age 3.9 +- 2.2 years) evaluated several parameters of skin ultrasonography, both in the lesional skin as well as in the normal skin of the control group and in the healthy perilesional skin of children with AD. The areas analyzed were the cheek, the ulnar region and the popliteal fossa. Regarding the evaluation of SLEB, the presence of SLEB was found in 22% of healthy children, with sizes between 53 microns and 92 microns. In the group of children with AD, SLEB was identified in 100% of the examined lesions and in 77% of the apparently healthy perilesional skin, with thicknesses between 68 microns and 106 microns. The lowest values found in both groups were at the ulnar level . Sabau et al. evaluated, in a group of 10 patients with AD (8 women, mean age 26 years), ultrasound aspects of lesional and nonlesional skin (SLEB, skin thickness and skin intensity), obtaining an average SLEB on the lesional skin of 164 microns and on healthy skin of 13 microns; the average thickness was 1409 mm and 0.8755 mm, respectively, and the average skin intensity was 37.95 and 60.4, respectively. Moreover, they observed that the Disease Life Quality Index (DLQI) correlated with skin thickness (p = 0.039, c = 0.657) and skin intensity (p = 0.032, c = 0.675), but not with SLEB or disease severity assessed by the SCORAD score (Scoring Atopic Dermatitis) . 3.2. SLEB and Skin Aging A number of seven studies evaluated by HF-USG different skin aspects (including SLEB) related to skin aging caused both by the cumulative effect of ultraviolet (UV) radiation and by aging . Crisan et al. evaluated in a study conducted on a group of 160 patients (80 women, mean age = 40.4 +- 21.1), divided into four age categories (under 20 years, 21-40 years, 41-60 years, over 80 years), the ultrasonographic aspects of the skin in three photoexposed areas (dorsal forearm, medial arm, zygomatic arc) and one non-photoexposed area (medial arm). They visualized SLEB in all patients over 20 years old, especially in photoexposed areas, compared to the population under 20 years of age. They also found a significant increase in the thickness of the dermis (p = 0.035) in patients in the age range of 21-40 years . Rayner et al. evaluated in a pilot study the test-retest reliability of measuring SLEB, skin thickness and skin intensity in a group of 31 patients (22 women, mean age 83.3 +- 4.23) by means of the intraclass correlation coefficient (ICC) and the Lin's concordance correlation coefficient (CCC). Rayner et al. assessed the degree of skin aging and the risk of skin tearing. The ultrasonographic characteristics were measured at the level of the mid dorsal forearm and the upper quarter of the lateral lower leg, bilaterally (places characterized by skin aging and tear) and at the half distance between the umbilicus and the left iliac crest (control area). They observed an almost perfect test-retest reliability for SLEB (0.95-0.99) and for skin thickness (0.95-0.99) . The above-mentioned study was continued a few years later by the same group that evaluated 173 patients (123 women, mean age = 87.8 +- 6.7 years) in order to establish the risk of developing purpura at the forearm level, through two separated measurements, at a distance of 6 months. The researchers found that SLEB and skin thickness were characteristics of aged skin that were statistically significantly associated with purpura in multivariable analysis (p = 0.014), but not in univariate one (p = 0.92). They also observed a positive correlation between SLEB and forearm skin thickness (p < 0.01, r = 0.512) . Tedeschi et al. evaluated the SLEB changes after performing two hyaluronic acid injection protocols at the level of the dorsal face of the hand, bilaterally. The study was carried out on 22 women (mean age 50.5 years) who were administered 1 mL of hyaluronic acid weekly for one month and then monthly, for 4 months the first group, and for 9 months the second group. SLEB was measured before the injection and one week after it at the level of the second metacarpal web, and an increase in echogenicity of SLEB was observed in 24% of the participants (p < 0.01) after 4 weeks, and in 18% of them after 4 and 10 months (p < 0.05) . Wakade et al. evaluated, in a group of 40 women, the impact on SLEB of two aesthetic procedures performed weekly for 4 weeks on the two halves of the face (radiofrequency and chemical peeling with glycolic acid). SLEB was measured bilaterally, both at the level of the external canthus of the eye and at the level of the nasolabial fold. They observed an insignificant decrease in SLEB at the eye level with both types of procedures and a significant decrease at the nasolabial fold level (a decrease in SLEB from 0.32 mm to 0.25 mm on the half of the face subjected to radiofrequency and a decrease from 0.3 to 0.2 mm on the half of the face subjected to chemical peeling) . Mondon et al. evaluated in an ex vivo and in vivo study several histopathological, confocal microscopy and skin ultrasonography aspects of skin aging. They assessed the effects of palmitoyl oligo and tetrapeptides applied on the face and unilateral forearm, compared to those of a placebo. The peptides were applied daily for 2 months to a group of 28 women over 50 years old (mean age = 59 +- 5.4 years). A significant improvement was observed in SLEB thickness (which decreased by 11-14.4% at the forearm level, p < 0.01) and SLEB density (which increased by 15% at the forearm level, p < 0.01) . Arisi et al. investigated, in a group of eight patients with field cancerization, skin ultrasonography aspects related to the presence of actinic keratosis and photoaged perilesional skin before and after 3 months of cold atmospheric plasma therapy. They highlighted an increase in the density of the dermis and perilesional SLEB (p = 0.02) and a decrease in the thickness of the perilesional SLEB (from 223 to 146.5 microns, p = 0.04) after the therapy. No statistically significant changes were observed regarding the thickness of the epidermis or dermis before and after the treatment . 3.3. SLEB and Skin Lymphomas A number of five studies evaluated SLEB changes and its role in evaluating disease progression and response to treatment in patients with cutaneous lymphomas (mycosis fungoides, folliculotropic mycosis fungoides and Sesary syndrome) . Polanska et al. evaluated, in 2017, in a group of 18 patients (5 women, mean age = 52.2 years) with mycosis fungoides stages I-IIA, the changes in SLEB and skin thickness measured by HF-USG before and after treatment with UVA1 and PUVA (7-10 weeks). They observed the presence of SLEB in all the examined lesions, but not in normal skin, regardless of the stage of the disease. Moreover, after the treatment, the disappearance of SLEB was observed in 66% of the cases, as well as a significant decrease in its thickness, from 0.256 mm to 0.064 mm (p < 0.001). In addition, after the treatment, a change in echogenicity of the skin was observed, both in the lesions (where it increased from 7.49% to 7.94%) and at the level of apparently intact skin (from 7.8% to 8.46%) (p < 0.01) . Later, the same team evaluated, in a group of three patients with mycosis fungoides stage IB (mean age = 48.3 years), the evolution of SLEB after treatment during a period of follow-up of approximately 5 years. Thus, before the initiation of the treatment, a mean SLEB = 0.44 mm was observed, which subsequently decreased to 0.13 mm (p = 0.01). A correlation of SLEB with clinical response was also observed, with SLEB disappearing completely in patients with complete response, unlike in those with partial response (SLEB = 0.16mm) . Another study by Polanska et al. in a group of 10 patients (1 woman) with mycosis fungoides highlighted the correlations between the presence of SLEB and that of a histological lymphocytic infiltrate. They observed an average SLEB thickness of 0.488 mm, with higher values in the plaque stage (0.544 mm) compared to the patch stage (0.265mm). Statistically significant positive correlations between the size of SLEB and the size of the lymphocytic infiltrate observed at the histopathological examination (r = 0.994, p < 0.01), with higher values of the former, were also detected . Niu et al. evaluated ultrasonographic skin differences (epidermal morphology and thickness, level of inflammatory infiltrate, SLEB thickness, margins and echogenicity) in patients with early mycosis fungoides (19 patients) and psoriasis vulgaris/eczema (48 patients). They observed a lower thickness of the epidermis (p < 0.01) and SLEB (p = 0.006) in patients with early mycosis fungoides compared to those with psoriasis/eczema. Moreover, the cut-off values for these patients were 0.2375 mm for the thickness of the epidermis (sensitivity 88.9% and specificity 73%) and 0.2655 mm for the thickness of SLEB (sensitivity 55.6% and specificity 90.9%) . The latest study led by Yukun Wang et al. evaluated the margins and echogenicity of SLEB in 26 patients with cutaneous lymphomas (23 with classic mycosis fungoides, 2 with folliculotropic mycosis fungoides and 1 with Sezary syndrome). Both in the patch and in the plaque stages, the presence of SLEB with clearly defined edges and homogeneous echogenicity was observed, ultrasound differences being observed at the level of the epidermal layer, with a homogeneous epidermis in the patch stage and a wavy epidermis in the plaque stage. At the same time, in patients with both folliculotropic and Sezary Syndrome forms, the same aspect of SLEB described above was found, but associated with the presence of hypoechoic perifollicular areas . 3.4. SLEB and Psoriasis We identified two studies that evaluated the features of SLEB in patients with psoriasis . Oweczarczyk-Saczonek et al. used HF-USG to monitor the response to treatment with dichiroinositol 1% and 0.25% compared to a placebo in 40 patients with mild psoriasis. They found no statistically significant differences regarding the thickness of the epidermis or SLEB in the two groups compared to the placebo group. However, at the end of the treatment, SLEB was identified in only 25% of the patients, whereas it was visualized in 70% of them at the initiation of the therapy . Regarding the ultrasonographic aspects of psoriasis, Odrzywolek et al. evaluated skin thickness, skin density and SLEB at the level of lesions in 71 patients, observing significantly lower skin densities in patients with psoriasis (p = 0.0003), as well as a thickened epidermis (p = 0.257). These two parameters were proportionally inversely correlated (p = 0.001). Higher SLEB thickness values were also observed in these patients, with the greatest thickness found at the knee level (0.389 mm) . 3.5. Other Applications of HF-USG in Skin Diseases Several studies evaluated the usefulness of HF-USG in general and of SLEB in particular in assessing different skin structures in special areas (genital) or in the presence of certain dermatological pathologies (leg ulcer, diabetic ulcer, lichen planus, facial granuloma, psoriasis vulgaris, talar calluses, limb oedema) . Migda et al. evaluated, in a study carried out in 50 women (age between 20-80 years), the ultrasonographic characteristics of the skin and adjacent structures at the level of the mons pubis and the labia majora and minora. They found the presence of SLEB in 80% of the examined patients, all cases identified having previously undergone cosmetic procedures which might have influenced the results . Krause et al. observed ultrasonographic changes in a group of eight patients (four women, mean age = 70.2 +- 11.6 years) with shin ulcer without signs of healing in the last 2 months, for which they decided to use bio-stimulation by laser therapy (two sessions/week). They observed an increase in the granulation tissue measured by HF-USG (from 2.66 +- 1.44 mm to 2.97 +- 0.7mm) and a decrease in SLEB measured in the perilesional tissue (from 0.78 +- 0.3 mm to 0.57 +- 0.37 mm) . Regarding foot ulcers, this time in patients with diabetes, Chao et al. evaluated a group of 19 patients with diabetic ulcers, 35 patients with diabetes but without neuropathy and 33 healthy patients. They observed an increase in the thickness of the plantar epidermis in both groups with diabetes compared to the healthy population (p < 0.05). In addition, in both groups SLEB was higher than in the healthy population (p < 0.001), with a more pronounced increase in patients with ulcers (64.7%) compared to those without ulcers (11.8%). Moreover, they observed a negative correlation between SLEB thickness and the thickness of the epidermis at the level of the hallux (p = 0.002, r = -0.366) . Yazdanparast et al. evaluated changes in healthy, perilesional and lesional skin in 21 patients with lichen planus (13 women, mean age = 47.62 +- 15.36 years), observing the presence of SLEB in 76.2% of the lichen lesions. Moreover, a statistically significant increase in SLEB was observed at the level of the lesions (from 22.64 +- 38.54 microns to 346.33 +- 281.55 microns, p < 0.01) compared to healthy skin, but not at the level of the perilesional skin . Another inflammatory lesion evaluated ultrasonographical is the facial granuloma, whose HF-USG aspects were examined by Morgado-Carrasco et al. in five patients. They highlighted the presence of the lesion as a hypoechoic, heterogeneous, poorly defined mass that invaded the dermis and hypodermis, most frequently associated with the presence of SLEB in all the patients . Suehiro et al. also evaluated SLEB thickness and several parameters related to subcutaneous echogenicity in 30 patients (mean age 67 years) with unilateral lymphedema in the context of breast cancer. They found statistically significantly higher values of all measured parameters (skin thickness, SLEB, subcutaneous tissue thickness, subcutaneous echogenicity, subcutaneous echo-free space) (p < 0.05) in all measured areas (medial upper arm, lateral upper arm, medial forearm, lateral forearm, dorsum of the hand) compared to the contralateral arm without lymphedema . Talar callosities represent another pathology for which HF-USG is useful. Luna-Bastante et al. evaluated ultrasonographic aspects in a group of four children with talar callosities that can cause diagnostic difficulties. They observed the disappearance of SLEB associated with dermo-hypodermic thickening, allowing a differential diagnosis with respect to other inflammatory lesions associated with a thickening of SLEB . diagnostics-13-00970-t001_Table 1 Table 1 Changes in SLEB thickness and echogenicity in different pathologies. Year Study Type No. of Patients Studied Disease Ultrasound Frequency Is the Level of SLEB Measuring Reported? Is SLEB Present in Non Lesional Skin? SLEB Changes Reference 2015 Interventional study 39 Atopic dermatitis 15-25 MHZ Yes Yes DT 2019 Observational study 10 Atopic dermatitis 15-25 MHZ No Yes IT, DE 2020 Observational study 18 Atopic dermatitis >25 MHZ No Yes IT 2012 Case series 160 Skin aging 15-25 MHZ Yes Yes IT 2015 Interventional study 28 Skin aging >25 MHZ Yes Not applied IT, IE 2015 Interventional study 22 Skin aging (fillers with hyaluronic acid) 15-25 MHZ Yes Not applied IE 2016 Interventional study 40 Skin aging (chemical peels and radiofrequency) >25 MHZ Yes Not applied DT 2017 Case series 31 Skin aging 15-25 MHZ No Yes IT 2019 Observational study 173 Skin aging 15-25 MHZ No Yes IT, DE 2021 Interventional study 12 Aktinic keratosis and photoaging >25 MHZ No Not applied IT, DE 2017 Interventional study 18 Cutaneous lymphoma (Mycosis fungoides) 15-25 MHZ No No DT, IE 2018 Observational study 3 Cutaneous Lymphoma (Mycosis fungoides) 15-25 MHZ No No IT, DE 2019 Observational study 10 Cutaneous lymphoma 15-25 MHZ No No IT, DE 2020 Observational study 26 Cutaneous Lymphoma (Mycosis fungoides) >25 MHZ No No IT, DE 2021 Case series 67 Cutaneous lymphoma (Mycosis fungoides), psoriasis/eczema >25 MHZ No Not applied IT, DE 2012 Observational study 87 Diabetic foot >25 MHZ Yes Yes IT 2016 Observational study 30 Arm lymphedema 1-14 MHZ Yes Not applied IT 2019 Interventional study 50 Vaginal rejuvenation 15-25 MHZ Yes No IT, DE 2019 Observational study 21 Lichen planus >25 MHZ No Yes IT 2020 Interventional study 9 Venous leg ulcer (cold atmospheric plasma treatment) 15-25 MHZ Yes Yes IT 2021 Observational study 5 Granuloma faciale 15-25 MHZ No No IT, DE 2021 Case series 4 Talar callosity 1-14 MHZ Yes Yes DT, IE 2021 Interventional study 46 Psoriasis (D-chiro-inositol treatment) 15-25 MHZ No Yes IT, DE 2022 Observational study 71 Psoriasis >25 MHZ Yes Yes IT, DE IT, increased thickness (measured in mm, perpendicular to skin surface), DT, decreased thickness (measured in mm, perpendicular to skin surface), IE, increased echogenicity, DE, decreased echogenicity. 4. Discussion HF-USG represents a relatively new, non-invasive imaging method that allows the in vivo evaluation of physiological and pathological aspects of the skin, as well as the influence of endogenous and exogenous factors . This non-invasive imaging technique finds its usefulness more and more often in the evaluation of inflammatory skin diseases, being possible in the future perhaps even to use it as a marker of treatment efficiency. This is supported by studies that show the presence of SLEB in AD in all cases before the start of treatment, followed by a significant decrease of it after approximately 6 months of treatment (p < 0.001) . Moreover, in the same dermatological condition, the presence and thickness of SLEB can orient us with a fairly high accuracy in order to establish the optimal moment in which to switch from a reactive treatment to a proactive treatment, in order to reduce relapses, SLEB values being influenced by edema and inflammatory infiltrate at the skin level . At the same time, the thickness of SLEB correlates with the degree of parakeratosis, hyperkeratosis, spongiosis and inflammatory infiltrate, but not with the DLQI and the SCORAD score that assesses the severity of the disease perceived by patient and doctor . Another inflammatory condition associated with ultrasonographic changes is lichen planus, which is associated with an increase in the thickness of the dermis and a decrease in its density on HF-USG, but also with an increase in SLEB thickness, most likely due to edema and inflammatory infiltrate. Thus, HF-USG can be used in patients with this pathology to increase the diagnostic accuracy, but also to monitor the response to treatment . To increase the accuracy of a non-invasive diagnosis, we can also use this method in the diagnosis of facial granulomas, HF-USG allowing their differentiation from other facial lesions, such as cutaneous lupus erythematosus or cutaneous lymphomas . However, in this sense, the characteristic hypoechoic-heterogeneous appearance, with poorly defined edges, is more useful than the presence of SLEB (which probably corresponds to the inflammatory infiltrate present histopathologically), this being a non-specific parameter also found in the inflammatory diseases described above and especially as a result of photoaging to which the skin of the face is subjected . Moreover, versatility, reproducibility, relatively high accuracy, as well as minimal discomfort make HF-USG an effective technique in the evaluation of lesions in the pediatric population, where the percentage of atopic dermatitis is increasing. Thus, although SLEB, is not pathognomonic, it allows the correlation in pediatric patients with edema and inflammatory infiltrate . Moreover, through the obtained images we can appreciate in real time aspects related to skin morphology and pathophysiological mechanisms, which allow us to observe the content of collagen, the orientation and distribution of collagen fibers at the skin level. Thus, the echogenicity of the dermis becomes one of the most important parameters that allow us to examine the ultrasonographic echogenicity, which is influenced by the orientation of the collagen fibers and the water content at this level . Regarding aspects related to skin aging, HF-USG allows the assessment of skin aging characteristics . Thus, SLEB can be perceived as a skin-aging marker, correlating with the thickness of the skin at the forearm level (p < 0.01, r = 0.512), an increased SLEB indicating the destruction of collagen fibers at the level of the papillary dermis and their replacing with elastin deposits and glycosaminoglycans. However, no correlation of SLEB with intrinsic aging was observed, suggesting a greater impact in the skin aging process of UV radiation than of aging itself . Thus, it was observed that the echogenicity of SLEB is inversely proportional to skin aging . Moreover, the same parameter, SLEB, can be used in the elderly population to assess the risk of developing a purpura at the level of the forearms, SLEB thickness measured at this level being positively correlated with this risk . The thickness and density of SLEB can also be used to assess the response to treatment of skin with signs of photoaging and field cancerization. Arisi et al. observed an improvement in this parameter after therapy (increase in density and decrease in thickness), as well as an increase in the density of the dermis, both, probably, through collagen synthesis and through the activation of immune system cells . Regarding the usefulness of SLEB in aesthetic medicine, it can be used as an objective marker to assess the collagen remodeling process that occurs after minimally invasive aesthetic procedures, allowing a non-invasive, long-term follow-up of the results . Moreover, Mondon et al. highlighted the fact that HF-USG can be used complementary to confocal reflectance microscopy to assess non-invasively aspects related to the improvement of skin quality (changes in SLEB's increased echogenicity, decreased thickness and papillary dermis' increased echogenicity) after topical application of anti-aging dermatocosmetics . Thus, HF-USG can become a useful tool to appreciate the real effectiveness of these products and highlight how they exert their anti-aging effects. HF-USG is also useful in the management of cutaneous lymphomas, especially type T (mycosis fungoides). SLEB is an objective parameter that allows the appreciation of the degree of lymphocytic inflammatory infiltrate that correlates statistically significantly with histopathological aspects (p < 0.01, r = 0.99) and with the stage of the disease . This imaging method is also useful in increasing the diagnostic accuracy, compared to main differential diagnoses (psoriasis, non-specific eczema), some studies showing statistically significantly lower values in patients with early mycosis fungoides compared to those with psoriasis/eczema in the thickness of both epidermis and SLEB, which reflects the pathophysiological processes underlying these conditions . However, HF-USG is not a diagnostic method, as the ultrasonographic aspects are not specific to this pathology, and histopathological and immunohistochemical confirmation is still necessary. Ultrasonography, by visualizing SLEB and its modifications along with the evolution of the disease, therefore allows the administration of topical therapy and the assessment of the partial or complete response to treatment, as well as of the residual disease, being thus an indicator for the continuation or interruption of therapy . Cutaneous ultrasonography can also be used in the assessment of the skin and adnexal structures at the genital level in the female population, the ultrasound data correlating with the histopathological ones, as HF-USG is able to identify the vulvo-vaginal and cervical anatomical structures, the thickness of the skin and its layers and the accessory glands. Moreover, the identification of SLEB at this level allows the identification of inflammatory pathologies, such as dermatitis, eczema, inverted psoriasis or irritations caused by various local procedures . In the case of leg ulcers, HF-USG can be used to assess the granulation tissue developed before and after the use of topical or interventional therapies so to assess the vascularity of the ulcer base or the perilesional tissue, as well as to visualize de novo epithelization. SLEB measured in these patients, in addition to the phenomena of skin aging and degradation of collagen fibers, may also reflect water retention at the level of the papillary dermis. Thus, a subsequent decrease in SLEB along with the improvement of the appearance of the lesion can be explained by a decrease in edema at this level . This theory is also be supported by the study conducted by Suehiro et al., which highlighted a thicker SLEB in all examined patients with lymphedema . The same theory is also supported by Chao et al., who explained the thicker SLEB in patients with diabetic ulcers and diabetes without neuropathy by the presence of edema in the papillary dermis (clinically manifested also by edema of the lower limbs in these patients) which consequently determined atrophy of the underlying skin, more pronounced in patients with ulcers than in those without ulcers. Thus, the negative correlation between the thickness of the SLEB and that of the epidermis entails numerous pathophysiologic implications for these patients and may in the future be a parameter for assessing the patient risk of developing skin tears . 5. Conclusions Important progress has been made in the last decade regarding the histopathological correspondence of SLEB, as well as the translation of its presence into pathophysiological processes. However, SLEB is not a parameter specific for a certain dermatological pathology, being rather associated with collagen degradation processes, inflammation and edema at the level of the papillary dermis. Although it can currently be used to follow the activity of a disease, guide the duration of treatment or assess the response to treatment in the presence of numerous inflammatory diseases, we are still far from relying on HF-USG as the only diagnostic tool for dermatological diseases, the existing studies up to the current time, although with statistically significant results, not having analyzed large, representative groups of patients. Thus, the need for randomized controlled studies in large groups of patients, the presence of clear histopathological correlations, as well as the increase in resolution in order to differentiate the skin layers more accurately are still essential elements necessary for the large-scale applicability of this parameter in clinical practice. Author Contributions Conceptualization: V.-M.V., V.-T.P. and A.C.N.; methodology, E.-C.C., V.-T.P., M.L., V.-M.V., writing--original draft preparation, S.I., I.A., C.S., E.-C.C.; writing--review and editing: A.C.N., E.-C.C., C.S., I.A., S.I.; supervision, A.C.N., V.-M.V. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest The authors declare no conflict of interest. Figure 1 PRISMA 2009 flow diagram depicting the selection process of the articles included in the systematic review . Figure 2 HF-USG aspects of normal skin (A) and skin with atopic dermatitis (B) and psoriasis (C). Visualization of epidermis (red arrows), SLEB (yellow arrow) and dermis (yellow star). The images were captured using a device with a 20 Mhz transducer (personal library of Dr. Alin Nicolescu). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Shung K. Cannata J. Zhou M.Q. Lee J. High frequency ultrasound: A new frontier for ultrasound Proceedings of the 2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society Minneapolis, MN, USA 3-6 September 2009 Volume 2009 1953 1955 10.1109/IEMBS.2009.5333463 2. Levy J. Barrett D.L. Harris N. Jeong J.J. Yang X. Chen S.C. High-frequency ultrasound in clinical dermatology: A review Ultrasound J. 2021 13 24 10.1186/s13089-021-00222-w 33877462 3. Mandava A. Ravuri P.R. Konathan R. High-resolution ultrasound imaging of cutaneous lesions Indian J. Radiol. Imaging 2013 23 269 277 10.4103/0971-3026.120272 24347861 4. Corvino A. Varelli C. Cocco G. Corvino F. Catalano O. Seeing the unseen with superb microvascular imaging: Ultrasound depiction of normal dermis vessels J. Clin. Ultrasound 2022 50 121 127 10.1002/jcu.23068 34761407 5. Moher D. Liberati A. Tetzlaff J. Altman D.G. Group P. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement BMJ 2009 339 b2535 10.1136/bmj.b2535 19622551 6. Slim K. Nini E. Forestier D. Kwiatkowski F. Panis Y. Chipponi J. Methodological index for non-randomized studies (minors): Development and validation of a new instrument ANZ J. Surg. 2003 73 712 716 10.1046/j.1445-2197.2003.02748.x 12956787 7. Hong Q.N. Gonzalez-Reyes A. Pluye P. Improving the usefulness of a tool for appraising the quality of qualitative, quantitative and mixed methods studies, the Mixed Methods Appraisal Tool (MMAT) J. Eval. Clin. Pract. 2018 24 459 467 10.1111/jep.12884 29464873 8. Polanska A. Silny W. Jenerowicz D. Kniola K. Molinska-Glura M. Danczak-Pazdrowska A. Monitoring of therapy in atopic dermatitis--Observations with the use of high-frequency ultrasonography Ski. Res. Technol. 2015 21 35 40 10.1111/srt.12153 9. Sorokina E. Mikailova D. Krakhaleva J. Krinitsyna J. Yakubovich A. Sergeeva I. Ultrasonography patterns of atopic dermatitis in children Ski. Res. Technol. 2020 26 482 488 10.1111/srt.12839 10. Sabau M. Boca A.N. Ilies R.F. Tataru A. Potential of high-frequency ultrasonography in the management of atopic dermatitis Exp. Ther. Med. 2019 17 1073 1077 10.3892/etm.2018.6984 30679976 11. Crisan D. Lupsor M. Boca A. Crisan M. Badea R. Ultrasonographic assessment of skin structure according to age Indian J. Dermatol. Venereol. Leprol. 2012 78 519 10.4103/0378-6323.98096 12. Rayner R. Carville K. Leslie G. Dhaliwal S.S. Measurement of morphological and physiological skin properties in aged care residents: A test-retest reliability pilot study Int. Wound J. 2017 14 420 429 10.1111/iwj.12621 27218422 13. Rayner R.L. Carville K.J. Leslie G.D. Dhaliwal S.S. Clinical purpura and elastosis and their correlation with skin tears in an aged population Arch. Dermatol. Res. 2019 311 231 247 10.1007/s00403-019-01899-9 30783769 14. Tedeschi A. Lacarrubba F. Micali G. Mesotherapy with an Intradermal Hyaluronic Acid Formulation for Skin Rejuvenation: An Intrapatient, Placebo-Controlled, Long-Term Trial Using High-Frequency Ultrasound Aesthet. Plast. Surg. 2015 39 129 133 10.1007/s00266-014-0432-1 15. Wakade D.V. Nayak C.S. Bhatt K.D. A Study Comparing the Efficacy of Monopolar Radiofrequency and Glycolic Acid Peels in Facial Rejuvenation of Aging Skin Using Histopathology and Ultrabiomicroscopic Sonography (UBM)--An Evidence Based Study Acta Med. 2016 59 14 17 10.14712/18059694.2016.49 16. Mondon P. Hillion M. Peschard O. Andre N. Marchand T. Doridot E. Feuilloley M.G.J. Pionneau C. Chardonnet S. Evaluation of dermal extracellular matrix and epidermal-dermal junction modifications using matrix-assisted laser desorption/ionization mass spectrometric imaging, in vivo reflectance confocal microscopy, echography, and histology: Effect of age and peptide applications J. Cosmet. Dermatol. 2015 14 152 160 10.1111/jocd.12135 25817264 17. Arisi M. Soglia S. Pisani E.G. Venturuzzo A. Gelmetti A. Tomasi C. Zane C. Rossi M. Lorenzi L. Calzavara-Pinton P. Cold Atmospheric Plasma (CAP) for the Treatment of Actinic Keratosis and Skin Field Cancerization: Clinical and High-Frequency Ultrasound Evaluation Dermatol. Ther. 2021 11 855 866 10.1007/s13555-021-00514-y 33738749 18. Polanska A. Osmola-Mankowska A. Olek-Hrab K. Molinska-Glura M. Adamski Z. Zaba R. Danczak-Pazdrowska A. High-frequency ultrasonography in objective evaluation of the efficacy of PUVA and UVA 1 phototherapy in mycosis fungoides Arch. Dermatol. Res. 2017 309 645 651 10.1007/s00403-017-1767-7 28799043 19. Polanska A. Danczak-Pazdrowska A. Olek-Hrab K. Osmola-Mankowska A. Bowszyc-Dmochowska M. Zaba R. Adamski Z. High-frequency ultrasonography--NEW non-invasive method in assessment of skin lymphomas Ski. Res. Technol. 2018 24 517 521 10.1111/srt.12450 20. Polanska A. Bowszyc-Dmochowska M. Olek-Hrab K. Adamski Z. Zaba R. Danczak-Pazdrowska A. High-frequency ultrasonography a new quantitative method in evaluation of skin lymphomas--First comparative study in relation to histopathology Ski. Res. Technol. 2019 25 720 724 10.1111/srt.12708 21. Niu Z. Wang Y. Zhu Q. Liu J. Liu Y. Jin H. The value of high-frequency ultrasonography in the differential diagnosis of early mycosis fungoides and inflammatory skin diseases: A case-control study Ski. Res. Technol. 2021 27 453 460 10.1111/srt.12945 22. Yukun W. Zihan N. Jie L. Qingli Z. Zhaorui L. Yuehua L. Hongzhong J. Value of High-Frequency Ultrasound in Accurate Staging of Mycosis Fungoides/Sezary Syndrome J. Ultrasound Med. 2020 39 1927 1937 10.1002/jum.15297 32319698 23. Owczarczyk-Saczonek A. Czerwinska J. Wygonowska E. Kasprowicz-Furmanczyk M. Placek W. D-chiro-inositol as a treatment in plaque psoriasis: A randomized placebo-controlled clinical trial Dermatol. Ther. 2021 34 e14538 10.1111/dth.14538 33188584 24. Odrzywolek W. Deda A. Zdrada J. Wcislo-Dziadecka D. Lipka-Trawinska A. Blonska-Fajfrowska B. Wilczynski S. Assessment of Psoriatic Skin Features Using Non-Invasive Imaging Technique Processes 2022 10 985 10.3390/pr10050985 25. Migda M.S. Migda M. Slapa R. Mlosek R.K. Migda B. The use of high-frequency ultrasonography in the assessment of selected female reproductive structures: The vulva, vagina and cervix J. Ultrason. 2019 19 261 268 10.15557/JoU.2019.0039 32021707 26. Krauze A. Wozniak W. Mlosek R.K. Usefulness of high-frequency ultrasound to assess the healing progress of shin ulcers J. Ultrason. 2020 20 E254 E260 10.15557/JoU.2020.0045 33500792 27. Chao C.Y.L. Zheng Y.-P. Cheing G.L.Y. The Association between Skin Blood Flow and Edema on Epidermal Thickness in the Diabetic Foot Diabetes Technol. Ther. 2012 14 602 609 10.1089/dia.2011.0301 22512286 28. Yazdanparast T. Yazdani K. Humbert P. Khatami A. Nasrollahi S.A. Zartab H. Firouzabadi L.I. Firooz A. Biophysical and ultrasonographic changes in lichen planus compared with uninvolved skin Int. J. Womens Dermatol. 2019 5 100 104 10.1016/j.ijwd.2018.10.001 30997382 29. Morgado-Carrasco D. Giavedoni P. Mascaro J.M. Iranzo P. Ultrasound Features of Facial and Extrafacial Granuloma Facial. A Caso Series Actas Dermo-Sifiliogr. 2021 112 559 563 10.1016/j.ad.2019.11.015 30. Luna-Bastante L. Negrete E.L. Naranjo L.A. Roustan G. Alfageme F. Sonography of a case series of talar callosities J. Ultrasound 2022 25 983 987 10.1007/s40477-021-00566-y 35129778 31. Suehiro K. Morikage N. Yamashita O. Harada T. Samura M. Takeuchi Y. Mizoguchi T. Nakamura K. Hamano K. Skin and Subcutaneous Tissue Ultrasonography Features in Breast Cancer-Related Lymphedema Ann. Vasc. Dis. 2016 9 312 316 10.3400/avd.oa.16-00086 28018504 |
PMC10000592 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050858 diagnostics-13-00858 Article Fetal Health Classification from Cardiotocograph for Both Stages of Labor--A Soft-Computing-Based Approach Das Sahana 1 Mukherjee Himadri 2 Roy Kaushik 2* Saha Chanchal Kumar 3 Blaumeiser Bettina Academic Editor 1 School of Computer Science, Swami Vivekananda University, Kolkata 700121, India 2 Department of Computer Science, West Bengal State University, Kolkata 700126, India 3 Department of Obstetrics and Gynecology, Biraj Mohini Matrisadan and Hospital, Kolkata 700122, India * Correspondence: [email protected] 23 2 2023 3 2023 13 5 85809 11 2022 10 2 2023 16 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). To date, cardiotocography (CTG) is the only non-invasive and cost-effective tool available for continuous monitoring of the fetal health. In spite of a marked growth in the automation of the CTG analysis, it still remains a challenging signal processing task. Complex and dynamic patterns of fetal heart are poorly interpreted. Particularly, the precise interpretation of the suspected cases is fairly low by both visual and automated methods. Also, the first and second stage of labor produce very different fetal heart rate (FHR) dynamics. Thus, a robust classification model takes both stages into consideration separately. In this work, the authors propose a machine-learning-based model, which was applied separately to both the stages of labor, using standard classifiers such as SVM, random forest (RF), multi-layer perceptron (MLP), and bagging to classify the CTG. The outcome was validated using the model performance measure, combined performance measure, and the ROC-AUC. Though AUC-ROC was sufficiently high for all the classifiers, the other parameters established a better performance by SVM and RF. For suspicious cases the accuracies of SVM and RF were 97.4% and 98%, respectively, whereas sensitivity was 96.4% and specificity was 98% approximately. In the second stage of labor the accuracies were 90.6% and 89.3% for SVM and RF, respectively. Limits of agreement for 95% between the manual annotation and the outcome of SVM and RF were (-0.05 to 0.01) and (-0.03 to 0.02). Henceforth, the proposed classification model is efficient and can be integrated into the automated decision support system. cardiotocograph fetal heart rate random forest sensitivity specificity SVM This research received no external funding. pmc1. Introduction The perinatal mortality rate is defined as the total number of stillbirths and deaths that take place during the first seven days of life per 1000 live births. According to UNICEF, in middle-income countries this figure stood at 19 per 1000 births by 2018, whereas for high-income and upper-middle-income countries this value was 3 and 7 per 1000, respectively South Asia and Sub-Saharan Africa have the highest perinatal mortality at 26% and 28%, respectively as reported by UNICEF . Though in the beginning of 20th century the perinatal mortality in high-income countries was frighteningly high, it was drastically reduced with proper antenatal care, comprehensive indicators of C-section, and introduction of perinatal screening technologies such as cardiotocograph (CTG), ultrasound, fetal ECG, amnioscopy, and amniocentesis etc. . It was found that the most common reasons for birth-related complications and perinatal deaths are prematurity, birth asphyxia, maternal hypertension, and septicemia . Asphyxia arises due to prolonged deprivation of oxygen caused by the interruption in the placental blood flow that is caused by maternal pre-eclampsia, placental abruption, or umbilical cord prolapse. Inadequate delivery management causes the signs of asphyxia to be overlooked. Fetal heart rate during the intra-partum and ante-partum period gives indications of the signs of asphyxia that, if detected in time, can prevent irreversible organ damage or even death. The main objective of intrapartum fetal monitoring is to recognize fetuses with risks of mortality and morbidity in order to ensure timely intervention. Cardiotocograph (CTG) is the most commonly used external monitoring system that continuously records the fetal heart rate (FHR) and uterine contraction (UC) to produce a visual display either electronically or on thermal paper. CTG has been in clinical practice for the last six decades and has enabled clinical practitioners to detect signs of fetal compromise at an early stage. However, studies have shown that continuous CTG monitoring of low-risk pregnancies led to an increase in unnecessary interventions and C-sections. To avoid this international guidelines such as FIGO NICE, NICHD etc., have suggested the use of CTG only for high-risk pregnancies. 1.1. Physiology of Fetal Hypoxia The autonomic nervous system (ANS), baroreceptors, and chemoreceptors control the pace of the fetal cardiac activity. The ANS consists of the sympathetic nervous system, which accelerates the heart rate, and the parasympathetic nervous system, which tries to reduce the heart rate. The baseline of the FHR is determined by the balance between these two nervous systems. When blood pressure increases, baroreceptors help reduce the heart rate, which in turn reduces blood pressure. This reduction in blood pressure prompts the baroreceptors to reduce the tone of the parasympathetic nerve and stimulate an increase in FHR and blood pressure. Chemoreceptors can sense the depletion of oxygen levels in fetal blood, which causes an increase the FHR to intensify the rate of oxygen input from the placenta. However, during hypoxemia, chemoreceptors cause the FHR to drop, resulting in an increase in blood pressure. 1.2. Problems of CTG Interpretation Significant and methodical cardiotocography (CTG) interpretation is vital to minimize intrapartum fetal asphyxia and the related grim consequences. Nowadays, though there are widespread uses of CTG, it suffers from intra-observer variation which results in low specificity. To counter this problem Dawes et al. introduced the first computerized version of CTG for automated feature extraction. A study that involved 469 subjects established that the fetal mortality rate when monitored by computerized CTG was four times less than the population whose CTG was interpreted visually. Despite the worldwide acceptance of CTG, the medico-legal issues have risen due to reasons such as improper interpretation of CTG and subsequent lack of timely action Birth asphyxia-related litigations in the UK in the decade 2000-2010 totaled as high as GBP 3.1 billion. A total of 73.6% of US-based obstetricians faced litigations due to fetal neurological damage. These litigations not only affected the working lives of the clinicians but also had influenced them to opt for unnecessary C-sections as a safeguard. The lack of standard guidelines for interpretation and recognition of the FHR signals in the gray zone is another reason for misinterpretation. To address these issues, in recent times soft-computing-based methods have been explored to find better interpretation of FHR results, which yielded results that are comparable to the clinical interpretation. These systems are capable of distinguishing between normal and IUGR fetuses . Automated interpretation of CTGs have not produced the desired result in predicting the pathological conditions. Soft-computing-based techniques were thus sought to identify the high-risk features with better accuracy. In fetal healthcare, machine learning had been used successfully to measure fetal weight predict hypoxia, or estimate gestational age . In this work, the authors propose a machine learning (ML)-based technique to classify the CTG by taking into consideration both the stages of labor separately. Standard ML algorithms such as MLP, SVM, random forest, and bagging were used for classification, and the most suitable classifier was selected using various statistical analysis metrics. Arrangement of the paper is as follows: Literature study is elaborated in Section 2; Methodology is described in Section 3; Section 4 contains the Results and Discussion; and the Conclusion is given in Section 5. 2. Literature Study Since the classical method for the interpretation and classification of CTG failed to produce convincing results, in last decade the focus of research had shifted towards the application of soft-computing-based techniques for the classification of CTGs. Comert et al. used the CTU-UHB dataset with 552 raw CTG data points to predict the fetal state with 87.9% accuracy . In another work with the same dataset, they reported an accuracy of 91.8% and 93.4%, respectively, with an artificial neural network (ANN) and an extreme learning machine (ELM) Comert (2019) . These models have shown high accuracy for normal and pathological states; however, the accuracy dropped to 59% for suspicious CTGs. Sundar et al. generated a model using XGBoost with an accuracy of 96% for the pathological state but only 73% for suspicious state. Batra et al. performed feature selection on the UCI machine learning repository's CTG data. Impulse response function, FHR baseline, and variability were used as inputs to SVM by Warrick et al. to train the system to classify normal and pathological CTGs. A total of 50% of the pathological CTGs were correctly identified with FP = 7.5%. The training dataset contained 189 normal CTGs and 31 pathological ones. Dash et al. exploited the dynamic nature of the FHR signal. They segmented the CTG record into shorter segments and extracted the features from each segment separately. The sequence of features from the segments were used as inputs to a Bayesian classifier. The true positive rate (TPR) and true negative rate (TNR) were 0.817 and 0.609, respectively, and thus outperformed SVM on the same dataset. The hidden patterns from the FHR were extracted by Chamidah et al. (to classify the fetus as normal, irregular, or pathologic using k-means clustering with an accuracy of 90.64%. Brocklehurst et al. used numerical algorithms and ANN to effectively identify abnormal CTGs. Features such as signal quality, patient clinical data, and the major features of CTG were used as the input. Nuanes et al. used uterine contractions, acceleration, and deceleration as input features to identify the rates of acidosis. In both the works the targets were labelled manually by the clinicians. Deceleration capacity was also used to predict acidemia with pH < 7.05 . Another similar study was conducted with fractal analysis and the Hurst parameter as the input . In these two studies, the AUCs were found to be 0.665 and 0.87, respectively. Dash et al. used the standard parameters of CTG and applied generative models and Bayesian theory to predict acidemia; pH < 7.15 was used for providing the target labels. Petrozziello et al. used a multi-modal convolution neural network (MCNN) on 35000 CTG data points and had a false positive rate (FPR) of 15% and a TPR of 31%. Ogasawara et al. used a database containing 162 normal and 162 abnormal CTGs to show that a deep neural network (DNN) performed better in classifying CTGs than either SVM or k-means clustering. Rahamayanti et al. experimented with LSTM models that had 28, 14, and 42 nodes in their hidden layers. They used ReLU and Softsign activations with 32 batch size. The accuracies were 95-98% for different models. Besides class imbalance, the major challenge in developing a robust soft-computing-based CTG interpretation model is class labeling. Many class labeling methods use weak metrics such as pH and healthy against HIE (hypoxic ischemic encephalopathy) without taking into consideration other CTG abnormalities. In this work, the authors have addressed this problem by annotating the CTG by different clinicians. Also, the classification of CTG without considering the progress in labor is not ideal because definition of normal features varies from first stage to second stage. The authors have taken this into account and performed the classification based on the stage of labor. The studies discussed above do not take into consideration the second stage of labor, which reduces the usefulness of the decision support system. 3. Methodology Classification was performed separately on both the stages of labor. The duration of each stage of labor was roughly 30 min. Overview of the methodology is given in Figure 1. 3.1. Dataset Description The Czech Technical University and University Hospital of Brno (CTU-UHB) database consists of 552 intrapartum records collected using the OB TraceVue System between 27th April and 6th August 2012. The duration of each record was at most 90 min. In the CTG records, the second stage of labor was not more than 30 min. From the 552 cases, only 46 were caesarean deliveries, whereas the remaining were vaginal deliveries. The following selection criteria had been taken into account :Mother's age--plays a role in congenital disease. Women with age below 18 years were thus excluded from the study. Weeks of gestation--affects the shape and behavior of the FHR signal. Hence, only fetuses with a gestational age greater than 36 weeks were included in the study. Known fetal disease--a disease such as intrauterine growth retardation (IUGR) has an effect on FHR patterns. Fetuses with IUGR or other congenital defects were not part of the study. Singleton pregnancies without complication were included. Different labor outcome measures such as umbilical artery pH were also provided to facilitate objective classification. It is considered the most common outcome measure of fetal hypoxia. Besides pH, the following outcome measures were also provided in the database:Base excess (BE) is a sign of metabolic hypoxia. A base deficit in extracellular fluid (BDecf) is another measure of hypoxia. Complete neonatal records of each baby. There was no neonatal morbidity, hypoxic-ischemic encephalopathy (HIE), or seizures. Apgar score. The following criteria had been applied to the CTG records: Signal Length: Out of the 90 min of each CTG record, stage 1 was restricted to a maximum of 60 min. The minimum time length, Td, needed was:(1) {pH<=7.15;Td>=30minspH>7.15;Td>=40mins Time duration from the end of the first stage of labor to birth was less than 30 min. A maximum of 30 min of FHR was recorded for the second stage of labor to ensure the circumvention of adversarial outcomes that might occur at the end of this stage. Missing signal: The quantity of missing signal was a maximum of 50% in the first stage of labor and was kept to an optimum in the second stage. Noisy signal: Maternal heart rates were present in some of the recordings due to the inaccurate placement of the ultrasound probe. The main parameters of the CTG signal were:Length of stage 1 of the labor in minutes. Length of stage 2 of the labor in minutes. Time duration from the signal end to birth. Percentage of noisy data in Window 1. Percentage of noisy data in Window 1. Overall percentage in Window 1. Percentage of noisy data in Window 2. Percentage of noisy data in Window 2. Overall percentage in Window 2. 3.2. Extraction of the Features The main features of CTG had been extracted in some previous works by the author. The authors considered two frequency domain parameters, i.e., baseline variability and sinusoidal heart rate pattern (SHR) and three time domain parameters, i.e., FHR baseline, acceleration, and deceleration. A summary of the extraction of the features is given below. 3.2.1. Baseline of FHR FHR baseline (BL) was estimated iteratively based on the assessment of an initial virtual baseline within a 10 min window . The effect of this initial calculation of baseline is nullified due to the iterative calculation of the adjusted baseline, which ultimately approaches the original baseline of the FHR signal. Each stage of iteration removes accelerations and decelerations. This algorithm is applied in a 10 min window with 3 minutes of overlap. If within each window is an identifiable baseline of duration more than 2 min (not necessarily contiguous), then the baseline value is accepted; otherwise, it is rejected. Baseline was classified as Normal, Bradycardia, and Tachycardia based on the amplitude of FHR. 3.2.2. FHR Variability A given FHR signal was disintegrated to its component frequencies, where each frequency is represented by its amplitude expressed in bpm. FHR variability (FHRV) was calculated in each discreet 10 min slice by computing the deviation of the FHR amplitude from the baseline within a one minute segment. If the number of cycles in each segment is more than one, then that segment has an identifiable baseline variability value. FHR variability (FHRV) is the standard deviation of the FHR values at each point. Summation of the value of variability for all the segments where the number of cycles exceeds one, gives the variability of the FHR for a 10 min window, and the average FHRV of all the 10 min windows is the variability of the CTG trace . 3.2.3. Acceleration Three points are used to identify each acceleration (Ac): the beginning, where the fetal heart rate crosses the baseline; the peak of Ac; and the end, i.e., where FHR again crosses the baseline. The duration of each deviation and the peak are noted. If the duration is between 15 sec and 10 min and the difference between the peak and the baseline is greater than 15 bpm, then the deviation was identified as acceleration. Acceleration that lasts for more than 10 min is considered a baseline change. Acceleration was classified as either present or absent . 3.2.4. Deceleration Besides the FHR baseline, the next important feature is the type of deceleration, which is a major indicator of fetal health status. The classification of deceleration plays a major role in the classification of FHR patterns into the three-tier system, i.e., normal, suspicious, and pathological. The spatial and temporal relationship of FHR with the uterine contraction pressure (UCP) of the mother can help classify the deceleration as Early, Late, or Variable. The authors proposed a fuzzy-logic-based method to identify a deceleration, extract the event points of both FHR and UCP, and finally classify the decelerations. The algorithm proposed by the authors to estimate the deceleration assesses the width as well as the amplitude of any negative deviation from the baseline and identifies it as deceleration if both the amplitude and the width conforms to the definition provided by the different international obstetric bodies. Every deceleration, De, was identified using three points--the start, where the fetal heart rate falls below the baseline; the nadir of De; and the end, i.e., where FHR again goes above the baseline. The duration of each De was noted. According to NICHD guidelines, the length of deceleration should be at least 15 s but not more than 10 min, in which case it is considered a baseline change. The difference between the nadir and the baseline should be at least 15 bpm. However, in clinical scenarios may not conform to such a strict definition. Therefore, the authors used a fuzzy-logic-based approach to identify the length and width in order to define a period with negative deviation from the baseline as deceleration. Finally, the classification of deceleration was performed using machine-learning-based techniques. 3.3. Classification of FHR for First Stage of Labor The dataset is D={d1,d2,....,dn}, where each data sample di has a feature set consisting of 11 features X={x1,x2,....,x10}. The features are listed in Table 1. A total of 399 cases from the CTU-UHB dataset were used for the experiment. Six clinicians separately annotated the data. The final annotation was performed by taking a majority vote. The CTG trace is classified as Normal (1), Suspicious (2), and Pathological (3). The annotated data was used as input to the classifiers. The input vector is represented as:I = [v11..v1nCk....Ck.....vm1..vmnCk] where the feature value for each data sample is represented as vij and Ck, k=1,2,3 denotes the class of each sample. The four classifiers used for building different classification models are discussed below. 3.3.1. Random Forest (RF) RF is known to provide a higher level of accuracy. It uses the bagging method, which is a combination of learning models that provides a more accurate and stable outcome. It searches for the best feature among a random subset of features, leading to a wide variety that generally results in a better model. Hence, it is a suitable model for predictive modelling. The associated algorithm is given in Algorithm 1: Algorithm 1: FHR Classsification using Random Forest Input: Si{(d1,y1),....,(dn,yn)} Step1: Perform row and column sampling Step2: Decision tree DTi for each Si Step3: PredictionPi - output of eachSi Output: vote of P1,....,Pi 3.3.2. Multilayer Perceptron (MLP) MLP offers approximate solutions for complex problems. The input layer has 10 features along with weights, wi, and the annotated class of the sample as input units. The output layer has one output unit. There are three hidden layers with sigmoid as the activation function. The use of the sigmoid function for the hidden layers is appropriate for the problem since we need to consider a soft decision boundary. The associated algorithm is given in Algorithm 2: Algorithm 2: FHR Classsification using Multi Layer Perceptron Input: X=[x1,....,x10], class labels, initial weight vector w=[wi]T Step 1: Weighted sum of the input features g(x)-i=1n10wixi Step2: Pass the value to the sigmoid activation function f. Step3: Produce the output y-f(g(x)) Step4: Compute the error ei=hi-yi Step5: Adjustwto minimize ei. Output: y^=i=110wifii=110wi 3.3.3. Support Vector Machine Training samples are mapped to high-dimensional feature space. SVM can be easily extended into complex classification problems with more than two classes. In the current classification problem it is not always possible to find the hard boundary between the classes. SVM is capable of dealing with soft boundaries for multiclass classification problems by introducing a slack variable that allows the input di to be closer to the hyperplane. The associated algorithm is given in Algorithm 3: Algorithm 3: FHR Classsification using Support Vector Machine Input: D={(d1,y1),....,(dn,yn)}, class labels C=[c1|,c2,c3], initial weight vector w=[wi]T, hyper-parameters, training set T Step1: Define the hyperplane wT*d+b-0 Step3: Maximize the margin w*d+bw|w|-1|w| between the hyperplane and the plane median Step4: Minimize |w| to maximize the margin Output: wT+1 3.3.4. Bagging Bagging is a commonly used ensemble method that creates several training sets using boot-strapping. It is useful in medical data analysis because it constructs a classification tree by bootstrapping the training data and then aggregating the predictions to produce the outcome. During bootstrapping, multiple training sets are created choosing the samples randomly and repeatedly. The training of each learner results in the generation of multiple learning models. This improves the accuracy of the prediction. The associated algorithm is given in Algorithm 4: Algorithm 4: FHR Classsification using Bagging Input: D={(d1,y1),....,(dn,yn)}, class labels, base learning algorithm L, number of learning rounds j, training set T Step1: D Step3: ht-LDt Step4: j-j+1 Step5: Repeat steps 1-3 till j covers the entire training set T Output: y^(d)=argmaxj=1T(y=ht(d)) For the purpose of classification, the dataset was split into k-number of folds where each fold is used as a testing set at some point, whereas other folds are used for training. The value of k was set to 5. Rather than splitting the data into train-test sets, the cross validation was chosen because it gives a less biased estimate of the model skill. The algorithm for the classification with k-fold cross validation is given in Algorithm 5: Algorithm 5: Algorithm for the k-fold cross validation Input: D={d1,d2,....,dn}, class labels Step1: Initialize i-1,k-5 Step2: Split D into k-folds Step3: In iterationi, the ith fold is used for testing and the rest are used for training Step4: i-i+1 Step5: Repeat steps 3-4 while i<=k Step6: Evaluate model on test score 3.4. Data Augmentation A large dataset is important for the improved performance of a machine learning model. However, the existing dataset is imbalanced. Thus, to increase the diversity of the data, we have performed data augmentation using the synthetic minority oversampling technique (SMOTE). The normal and suspicious CTGs are the majority class C-, whereas pathological CTGs are the minority class C+. The total synthetic samples are:S = (C - C+) x C + size of the dataset after augmentation was 1968. The same set of classifiers were used with the augmented data and the accuracy was noted for both the training and testing data. SMOTE is applied only to the training dataset, whereas the test dataset remains unchanged to correctly represent the original data. 3.5. Classification of the Second Stage of Labor Another complication associated with the classification of CTG is the existence of two different stages of labor. In the first stage, cervical dilation occurs with regular contractions. The second stage is characterized by active pushing, during which uterine contraction is more frequent and so are the decelerations. Interpretation of the parameter values also change during this stage. While building a robust classification system most researchers consider either one of the labor stages or make no distinction between the stages. The first approach is methodologically correct but does not take into consideration the change in feature values that might lead to a better recognition of the fetal state. The second approach can lead to erroneous classification because it does not consider the radical changes in FHR parameter dynamics. In this work, the authors applied the classification and feature extraction algorithms to the second stage as well, while taking into consideration the changed dynamics in this stage. 4. Results and Discussion The analysis of the result can be categorized into four categories:Adequacy of the dataset. Outcome of the classification and the comparison of the performance of the classifiers. Agreement between the manual annotation and the classifier-based outcome. Comparison of the classifiers' performance for different stages of labor. 4.1. Kaiser-Meyer-Olkin (KMO) and Bartlett's Test Sample size is the most important factor to be considered to determine if the data is appropriate for the experiment. Sample size should be more than 100 and at least 300 for factor analysis . Also, the number of samples should be a least five times more than the number of features . KMO was conducted to measure the sample adequacy, which is used to compare the magnitudes of the observed correlation coefficients in relation to the partial correlation coefficients. KMO 1 is considered ideal; KMO < 0.5 is unacceptable. Bartlett's test of sphericity is used to establish that there is no redundancy between the features, i.e., an ideal value for Bartlett's sphericity means that the null hypothesis of uncorrelated features is rejected. Both the values are shown in Table 2. A scree plot is plotted to find out the number of features that can be retained by recognizing the point of inflexion. The number of features before the inflexion represents the important features to be extracted for the factor analysis. The plot is shown in Figure 2. According to the scree plot, all the features up to the ninth feature are important for the classification. For the classification, the first nine features are adequate. The 10th and 11th features have been added to ensure the robustness of the system. 4.2. Sensitivity Analysis of the 5-Fold Cross-Validation The authors have evaluated the choice of k value for the k-fold cross validation. Distribution of the outcome for each fold for each different classifiers are compared to check the robustness of the chosen configuration of k. Values of the performance metrics set:(2) M={Accuracy,TPR,FPR,Precision,Recall,F -Measure,ROC,kappa,RMSE} were compared and shown graphically in Figure 3. The correlation matrix is shown in Table 3, and graphical representation of the correlation is given in Figure 4. Metric measurements of the classifiers for each fold are almost equal except for MLP. Correlation of the performance outcome in Table 3 suggests high correlation among the classifier performance in fold 1, whereas in other folds, except for the highlighted, correlations are very high. The scatter plot of Figure 4 has very few outliers and performance measures for SVM and RF are almost identical. Thus, the choice of k = 5 is appropriate for the purpose. The correlation test not only established the appropriateness of the chosen test harness, but also established that the value of k performed equally well across the different classifiers. The scatter plot has revealed how closely the performance of the classifiers match when compared pairwise. 4.3. Model Performance Measure Metrics of the model performance measures for the classifiers are given in Table 4. Though accuracy is normally considered a measure for evaluating the performance of a classifier, it can be ambiguous if the dataset is not balanced. In such cases, more emphasis is on the majority class, which makes it difficult for the classifier to perform well on the minority class. The parameter values for all the classes are comparable, though the performance of the RF and SVM are better. 4.3.1. Combined Performance Measure The authors also establish that the proposed model balances between false positive rate (FPR) and false negative rate (FNR). Such measures cannot be provided by model performance metrics. Hence, the values of combined performance measures are given in Table 5. Since G-mean > 0.9 for all the classifiers, the performance of the classifiers in identifying positive classes is very high. However, the values of DP < 0.2 indicate that the classifiers have limited capacity in distinguishing between normal and abnormal (suspicious and pathological) classes. The average accuracy or the balanced accuracy for both the major and minor classes are fair. MCC values indicate that the correlation between classifier prediction and visual prediction are high for SVM and bagging and moderate for MLP and RF. Cohen's kappa > 0.8 is in almost perfect concordance with the classifier model prediction and the classified class. All the classification models are good at avoiding misclassification because Youden's index > 0.8. 4.3.2. Model Performance Comparison The number of data points where actual and predicted classifications differ along with the actual and predicted class are shown in Figure 5, along with the prediction probability. The prediction probability summarizes how well the predicted class matches the actual class. The maximum number of misclassifications is observed with MLP and bagging with 27 and 25 data points, respectively. Prediction probability is unvarying for SVM. 4.3.3. AUC-ROC Curve Since it is a multi-class classification problem, the AUC-ROC was used to visualize the performance of the classifiers. The ROC curve is plotted with TPR (sensitivity) against the FPR (1--specificity) in Figure 6. The AUC values are 0.966, 0.997, 0.966, and 0.989 for MLP, RF, SVM, and bagging, respectively. The AUC-ROC value greater than 0.9 indicates high discrimination capacity for all the classifiers for all the classes. All the classifiers exhibit very high discrimination capacity. 4.4. Performance Comparison of the Classifiers with Clinicians' Annotation Performance of the four classifiers are compared with the visual estimate using covariance correlation and a contingency table. A total of 399 observations were annotated as Normal (1), Suspicious (2), and Pathological (3). 4.4.1. Covariance Correlation The degree of association between the actual and predicted values are analyzed and the results for all four classifiers are given in Table 6; the scatter plots are shown in Figure 7. All the models show almost similar standard deviation between the actual and the predicted values. The scatter plot shows that many traces that had been annotated as Suspicious (2) or Pathological (3) were predicted as Normal (1) by all the classifiers. On the other hand, actual Pathological ones have been predicted by the classifiers as Normal or Suspicious. This anomaly in the prediction can cause serious damage to the fetus. The degree of disagreement is most evident with the bagging classifier. Also, this contradicts the result of Table 5 where Youden's Index indicates the avoidance of misclassification by all four classifiers. This leads us to hypothesis test. 4.4.2. Contingency Table Since we are dealing with categorical data, the contingency table is used for hypothesis testing along with estimation of model parameters. Contingency dals with:Hypothesis testing that verifies whether there is any association among the actual and predicted classifications or whether they are independent. Which model provides the better option for the overall classification of CTG? The null hypothesis H0: The rows and the columns of the table are independent. The alternate hypothesis Ha: There is a link between the rows and the columns of the table. The observed frequencies for the actual and predicted classifications are shown in Table 7. The result of the test for independence or the chi-squared test for degree of freedom (DF) = 4 and a = 0.05 are shown in Table 8. As the computed p-value is lower than the significance level alpha = 0.05, one should reject the null hypothesis H0 and accept the alternative hypothesis Ha, i.e., the annotated class and the predicted class are related; however, this test does not reveal the difference between the two methods of classification and the Bland-Altman test was performed. 4.4.3. Bland-Altman Analysis Bland-Altman (B-A) analysis was performed to find the difference between the actual and predicted classifications The difference between the two methods was measured by constructing limits of agreement (LoA). A graphical approach is used to verify the assumption of normality of differences. The lack of agreement is summarized by computing the bias, estimated by the mean difference and the standard deviation of the differences. Parameters of the B-A analysis are given in Table 9, and the plots for it shown in Figure 8. The number of outliers is insignificant, and the limits of agreement are also within acceptable limits. The bias is insignificant because the line of equality is within the CI of the mean difference. Thus, no method under-estimates compared with the actual annotation. Though the limits of agreement (LoA) for the difference is within an a priori limit, the LoA is narrowest at 0.5 and 0.6 for SVM and RF, respectively. The agreement of the upper and lower limits for SVM and RF are also smaller compared with the other classification methods. Thus, the correlation between the actual annotation and the predictions by SVM and RF are both high. 4.5. Result with Augmented Data Training models on limited data leads to overfitting, which is exhibited by high accuracy on the training dataset but a lack of generalization to the test data. This difference for both sets of data is shown in Figure 9. For both sets of data, there are differences in accuracy for the training and test data. This could mean overfitting, however, since the difference between the accuracy at each fold is below 4% the overfitting can be ignored. No improvement in accuracy is observed with the augmented data. 4.6. Class-Wise Accuracy Estimation for Stage 1 and Stage 2 Labor Feature values from the second stage of labor were fed into the same set of classifiers and the overall accuracies were 86.2%, 92.6%, 92.3%, and 88.2%, respectively, for MLP, RF, SVM, and bagging. The accuracy of each class for the two stages of labor are given in Table 10. 4.7. Comparison with Other Works Some of the recent works along with their outcomes are listed in Table 11. 5. Conclusions The primary objective of this work was to develop a robust classification model for CTG that can identify fetal distress at both stage 1 and stage 2 of labor. The accuracy for both stages for each class is presented in Table 10. The difference in accuracy was mainly due to the presence of noise and motion artefacts in the FHR signal in stage 2 of labor. A sample of adequate size had been used for the experiment as established by the KMO and Bartlett's Test. The experiment was repeated with augmented data, but the deviation in accuracy for each fold was less than 4%. The sensitivity and the specificity for different classifiers are in the ranges 92.7-96.4% and 92.8-98.4%, respectively. The AUC, which is greater than 0.9 for all classifiers, indicated their high discrimination capacity. The combined performance measure reveals that the discriminant power of SVM and RF were highest at 1.744 and the maximum number of misclassifications were with MLP and bagging. The authors used annotation by six clinicians, and the final annotation for each record was chosen using majority voting. The Bland-Altman plot shows the strong agreement between the clinicians' annotation and the classification obtained with SVM and RF. Performance measures for the four classifiers and the performance in comparison with clinicians' annotations are summarized in Table 12. SVM and RF had improved classification performance. The performance of the classifiers varies in different stages of labor. In the second stage of the labor, the accuracy of the classifiers is comparatively low; however, SVM and RF still performed better than the other two. The accuracy, sensitivity, specificity, and AUC, etc., are sufficiently high for both SVM and RF. These two classifiers also exhibited a high level of agreement between the system and the human annotators. The other feature-based approaches to the automation of CTG classification that were based on weak annotation did not achieve high AUC, sensitivity, and specificity. In fact, with weak annotations, the systems showed little improvement over the visual interpretation. Features used in stage 1 of labor are typically embedded in clinical exercise. However, features such as deceleration and average baseline are no longer applicable for stage 2. Since there is active maternal pushing, the decelerations are recurrent and large. The data is burstier and hence there is a dwindling of the accuracy level. In subsequent work, the authors intend to implement the classification of CTG in the second stage of labor after removing the noise and artefacts. Author Contributions S.D.: Developing the algorithms, Methodology, Validation, and Writing; H.M.: Data Pre-processing, Validation; K.R.: Conceptualization, Supervision; C.K.S.: Supervision, Validation. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study did not require ethical approval. Informed Consent Statement Not applicable. Data Availability Statement The authors used an open access dataset that is available from Conflicts of Interest The authors declare no conflict of interest. Figure 1 Overview of the proposed methodology. Figure 2 The scree plot begins to flatten from the 9th feature. Figure 3 Graphical representation of the performance comparison of the classifiers for each fold of the 5-fold cross validation in terms of the performance metrics. Figure 4 Scatter plot of the correlation analysis for the four classifiersestablishes how well the chosen value of k matches across the different algorithms. Figure 5 Data points where actual and predicted classifications differ for each of the classifiers. The prediction probability is shown as a yellow line. (a) MLP. (b) RF. (c) SVM. (d) Bagging. Figure 6 ROC curve for the four classifiers. Predicted has at least one tie between the positive actual state group and the negative actual state group. (a) MLP, AUC-ROC: 0.966. (b) RF, AUC-ROC: 0.997. (c) SVM, AUC-ROC: 0.966. (d) Bagging, AUC-ROC: 0.989. Figure 7 Scatter plots show the difference between the actual and predicted classifications. The bar charts show the number of instances in each class and the scatter plots show the number of outliers in each class for all four classifiers. (a) MLP. (b) RF. (c) SVM. (d) Bagging. Figure 8 Bland-Altman plot of the actual and predicted classifications for all the classifiers. (a) For MLP, the average difference is -0.01 and the limits of agreement: -0.82 and 0.81. (b) For RF the average difference is -0.03 and the limits of agreement: -0.7 and 0.65. (c) For SVM, the average difference is -0.01 and the limits of agreement: -0.59 and 0.58. (d) For bagging the average difference is -0.01 and the limits of agreement: -0.74 and 0.72. Figure 9 Comparison of the differences in accuracy for both the augmented and normal datasets for both the training data and test data. The result is shown for each fold of the 5-fold cross validation for (a) bagging, (b) random forest, (c) MLP, and (d) SVM. diagnostics-13-00858-t001_Table 1 Table 1 Description of the feature set for the classification of CTG. Feature Number Feature Set Description 1 Baseline (BL) Baseline of the FHR 2 Baseline type (B_type) Classification of BL as Normal, Bradycardia, or Tachycardia 3 Variability (V) Variability of FHR baseline 4 Variability type (V_type) Classification variability as Absent, Minimal, Moderate, or Marked 5 Acceleration (Ac) Acceleration of FHR 6 Number of accelerations Total number of accelerations present in a CTG trace. 7 Number of Early decelerations Number of Early decelerations present in a CTG trace 8 Number of Late decelerations Number of Late decelerations present in a CTG trace 9 Number of Variable decelerations Number of Variable decelerations present in a CTG trace 10 Sinusoidal heart rate (SHR) Sinusoidal heart rate pattern present or not. SHR in a CTG trace is denoted as Present, Absent, or Undetermined. 11 Stage of labor Stage of labor can be normal, stage 1, or stage 2. diagnostics-13-00858-t002_Table 2 Table 2 KMO and Bartlett's test. A KMO value greater than 0.8 indicates the presence of a strong partial correlation among the features and the number for the significance of Bartlett's test is less than 0.05. KMO and Bartlett's Test Kaiser-Meyer-Olkin Measure of Sampling Adequacy. 0.867 Bartlett's Test of Sphericity Approx. Chi-Square 5612.158 df 105 Sig. 0.000 diagnostics-13-00858-t003_Table 3 Table 3 Correlation matrix of the classifiers depicting the degree of correlation in terms of M. Fold 1 Variables Bagging MLP RF SVM Bagging 1 0.888 0.953 0.953 MLP 0.888 1 0.931 0.931 RF 0.953 0.931 1 1.000 SVM 0.953 0.931 1.000 1 Fold 2 Bagging 1 0.888 1.000 0.832 MLP 0.888 1 0.888 0.615 RF 1.000 0.888 1 0.832 SVM 0.832 0.615 0.832 1 Fold 3 Bagging 1 0.888 0.538 1.000 MLP 0.888 1 0.478 0.888 RF 0.538 0.478 1 0.538 SVM 1.000 0.888 0.538 1 Fold 4 Bagging 1 0.901 0.742 0.742 MLP 0.901 1 0.931 0.931 RF 0.742 0.931 1 1.000 SVM 0.742 0.931 1.000 1 Fold 5 Bagging 1 0.901 0.742 0.742 MLP 0.901 1 0.931 0.931 RF 0.742 0.931 1 1.000 SVM 0.742 0.931 1.000 1 diagnostics-13-00858-t004_Table 4 Table 4 Metric values of the model performance measure. Classifier Accuracy Sensitivity Specificity Precision MLP 0.927 0.927 0.962 0.928 RF 0.967 0.964 0.984 0.968 SVM 0.964 0.964 0.983 0.966 Bagging 0.936 0.936 0.968 0.936 diagnostics-13-00858-t005_Table 5 Table 5 Metrics values of the combined performance measure. Classifier G-Mean Discriminant Power (DP) Balanced Accuracy Matthew's Correlation Coefficient (MCC) Cohen's Kappa Youden's Index MLP 0.944 1.382 0.446 0.784 0.889 0.889 RF 0.974 1.774 0.474 0.696 0.950 0.948 SVM 0.973 1.774 0.474 0.970 0.946 0.947 Bagging 0.952 1.459 0.453 0.904 0.902 0.904 diagnostics-13-00858-t006_Table 6 Table 6 Correlation between the actual and the predicted values for the classifiers. Classifier Variable Mean Std. Deviation MLP Actual 1.927 0.781 Predicted 1.912 0.770 RF Actual 1.927 0.781 Predicted 1.947 0.786 SVM Actual 1.927 0.778 Predicted 1.932 0.785 Bagging Actual 1.937 0.782 Predicted 1.945 0.787 diagnostics-13-00858-t007_Table 7 Table 7 Observed frequencies for actual and predicted classification of CTG. Classifier Actual/Predicted 1 2 3 MLP 1 128 1 8 2 0 151 3 3 9 8 91 RF 1 131 0 6 2 0 151 3 3 3 1 104 SVM 1 132 0 4 2 1 152 3 3 4 0 103 Bagging 1 129 0 6 2 2 145 7 3 4 6 100 diagnostics-13-00858-t008_Table 8 Table 8 Contingency table showing the chi-squared values and the p-value for each classifier. Classifier Chi-Squared (Observed Value) Chi-Squared (Critical Value) p-Value MLP 624.194 9.488 <0.0001 RF 717.664 9.488 <0.0001 SVM 723.633 9.488 <0.0001 Bagging 650.510 9.488 <0.0001 diagnostics-13-00858-t009_Table 9 Table 9 Parameters of the Bland-Altman analysis. MLP Parameter Value 95%CI lower limit 95%CI upper limit Difference 0.02 -0.03 0.06 Upper Limit of Agreement 0.89 0.82 0.97 Lower Limit of Agreement -0.86 -0.94 -0.79 Intercept -0.02 -0.14 0.11 Slope 0.02 -0.04 0.08 RF Parameter Value 95%CI lower limit 95%CI upper limit Difference -0.02 -0.05 0.01 Upper Limit of Agreement 0.60 0.55 0.65 Lower Limit of Agreement -0.64 -0.69 -0.59 Intercept -0.01 -0.09 0.08 Slope -0.01 -0.05 0.03 SVM Parameter Value 95%CI lower limit 95%CI upper limit Difference -0.01 -0.03 0.02 Upper Limit of Agreement 0.58 0.53 0.64 Lower Limit of Agreement -0.59 -0.65 -0.54 Intercept 0.01 -0.07 0.09 Slope -0.01 -0.05 0.03 Bagging Parameter Value 95%CI lower limit 95%CI upper limit Difference -0.01 -0.04 0.03 Upper Limit of Agreement 0.72 0.66 0.78 Lower Limit of Agreement -0.74 -0.80 -0.67 Intercept 0.01 -0.09 0.11 Slope -0.01 -0.06 0.04 diagnostics-13-00858-t010_Table 10 Table 10 Accuracies for each class for each of the four classifiers. Classifier Stage of Labor Stage 1 Stage 2 Normal Suspicious Pathological Normal Suspicious Pathological MLP 93.4% 98% 84.3% 89.3% 88.7% 79.2% RF 95.6% 98% 96.3% 92.8% 89.3% 92.3% SVM 97% 97.4% 96.3% 91.2% 90.6% 92.7% Bagging 95.6% 94.2% 90.9% 89.8% 86.5% 86.7% diagnostics-13-00858-t011_Table 11 Table 11 A comparison with some of the recent works for the classification of CTG. Author Dataset Number of Samples Annotations Method Used Outcome Comert et al. (2017) UCI Machine Learning Repository 2126 with 21 features and 3 classifications. 3 obstetricians ANN with 21 inputs, 10 hidden layer, 3 output nodes. Tan sigmoid and softmax functions were used in the hidden layer. Sensitivity = 97.9% Specificity = 99.7% Signorini et al. (2019) Ob-Gyn Clinics at the Azienda Ospedaliera Universitaria Federico II, Napoli, Italy 60 normal and 60 IUGR samples Normal and IUGR Linear Logistic Regression, SVM, Naive Bayes, Random Forest, RIDGE Regression Sensitivity = 83.8-86.7% Specificity = 76.7-87.1% Rahmayanti et al.(2021) UCI Machine Learning Repository 2126 with 21 features and 3 classifications 3 obstetricians ANN and LSTM with ReLU activation and Adam optimizer. Accuracy = 9-37% Ogasawara et al (2021). ) CTG data from Keio University hospital 380 with 2 classifications--normal and abnormal Umbilical artery pH value and apgar score Deep Neural Network (DNN) with 3 convolution layers. Normal and abnormal are 2 outputs. AUC-ROC = 0.73 +- 0.04 Petroziello et al (2019) . Oxford archive 35,429 CTG data Umbilical artery pH value and apgar score Multimodal-CNN Only 1% improvement in FPR and 19% improvement in TPR compared to normal clinical practice. Zhao et al. (2019) CTU-UHB dataset for CTG 552 CTG data 9 obstetricians Deep CNN with 8 layers Sensitivity = 93.45%, Specificity = 91.22% Current Work CTU-UHB dataset for CTG 399 data with three classifications 6 obstetricians Best result with Random Forest Sensitivity = 96.4% Specificity = 98.4% diagnostics-13-00858-t012_Table 12 Table 12 Performance measure of the classifiers in terms of various statistical evaluation metrics and classification performance in comparison with the clinicians' interpretations. Evaluation Metrics MLP SVM RF Bagging Performance Measure of the Classifiers Combined Performance Measure x x Model Performance Measure x x AUC-ROC Sensitivity and Specificity x x Accuracy of Detecting Suspect Class x Performance Comparison with Clinicians' Annotation Covariance Correlation x Limits of Agreement of B-A Plot x x Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. UNICEF UNCEF, World Health Organization, the World Bank Group, and the United Nations 2019 Levels and trends in child mortality Report 2019. UNICEF. Obtido em 12 de 04 de 2022 Available online: (accessed on 10 February 2023) 2. Goldenberg R.L. Saleem S. Pasha O. Harrison M.S. Mcclure E.M. Reducing stillbirths in low-income countries Acta Obstet. Et Gynecol. Scand. 2016 95 135 143 10.1111/aogs.12817 3. Mahdizadeh J. Bouraghi H. Panahi S.S.G.S. Mohammadpour A. Shargh A.K. Mojarad M.R. Kahouei M. A theory map of the causes of perinatal death in a developing country Crescent J. Med. Biol. 2019 6 237 241 4. Alfirevic Z. Gyte G.M. Cuthbert A. Devane D. Continuous Cardiotocography (CTG) as a Form of Electronic Fetal Monitoring (EFM) for Fetal Assessment during Labour John Wiley & Sons, Ltd Hoboken, NJ, USA 2017 5. Lewis D. Downe S. FIGO consensus guidelines on intrapartum fetal monitoring: Intermittent auscultation Int. J. Gynecol. Obstet. 2015 131 9 12 10.1016/j.ijgo.2015.06.019 26433400 6. Macones G.A. The 2008 National Institute of Child Health and Human Development Workshop Report on Electronic Fetal Monitoring: Update on Definitions, Interpretation, and Research Guidelines Obstet. Gynecol. 2009 113 230 237 10.1097/AOG.0b013e318193be34 7. Dawes G.S. Redman C.W. Numerical analysis of the human fetal heart rate: The quality of ultrasound records Am. J. Obstet. Gynecol. 1981 141 43 52 10.1016/0002-9378(81)90673-6 7270621 8. Grivell R.M. Alfirevic Z. Gyte G.M. Devane D. Antenatal cardiotocography for fetal assessment Cochrane Database Syst. Rev 2015 2015 1 59 10.1002/14651858.CD007863.pub4 26363287 9. Georgieva A. Payne S.J. Moulden M. Redman C.W.G. Artificial neural networks applied to fetal monitoring in labor Neural Comput. Appl. 2013 22 85 93 10.1007/s00521-011-0743-y 10. Solt I. Caspi O. Beloosesky R. Weiner Z. Avdor E. Machine learning approach to fetal weight estimation Am. J. Obstet. Gynecol. 2019 220 S666 S667 10.1016/j.ajog.2018.11.1063 11. Ananthet C.V. Brandt J.S. Fetal growth and gestational age prediction by machine learning Lancet Digit. Health 2020 2 336 337 10.1016/S2589-7500(20)30143-6 12. Comert Z. Kocamaz A. Comparison of Machine Learning Techniques for Fetal Heart Rate Classification Acta Phys. Pol. 2017 132 451 454 10.12693/APhysPolA.132.451 13. Comert Z. Sengur A. Budak U. Kocamaz A.F. Prediction of intrapartum fetal hypoxia considering feature selection algorithms and machine learning models Health Inf. Sci. Syst. 2019 7 1 9 10.1007/s13755-019-0079-z 30588291 14. Sundar C. Chitradevi M. Geetharamani G. Classification of cardiotocogram data using neural network based machine learning technique Int. J. Comput. Appl. 2012 47 1 8 15. Batra A. Chandra A. Matoria V. Cardiotocography Analysis Using Conjunction of Machine Learning Algorithms 2017 International Conference on Machine Vision and Information Technology (CMVIT) IEEE Singapore 2017 1 6 16. Warrick P.A. Hamilton E.F. Precup D. Kearney R.E. Identification of the dynamic relationship between intra-partum uterine pressure and fetal heart rate for normal and hypoxic fetuses IEEE Trans. Biomed. Eng. 2009 56 1587 1597 10.1109/TBME.2009.2014878 19237336 17. Dash S. Quirk J.G. DjuriC P.M. Fetal Heart Rate Classification Using Generative Models IEEE Trans. Bimedical Eng. 2014 61 2796 2805 10.1109/TBME.2014.2330556 24951678 18. Chamidah N. Wasito I. Fetal state classification from cardiotocography based on feature extraction using hybrid K-Means and support vector machine Proceedings of the 2015 International Conference on Advanced Computer Science and Information Systems Piscataway, Indonesia 10-11 October 2015 19. Brocklehurst P. Field D. Greene K. Juszczak E. Keith R. Kenyon S. Linsell L. Mabey C. Newburn M. Plachcinski R. Computerised Interpretation of Fetal Heart Rate during Labour (INFANT): A Randomised Controlled Trial Lancet 2017 389 1719 1729 10.1016/S0140-6736(17)30568-8 28341515 20. Nunes I. Ayres-de-Campos D. Figueiredo C. Bernardes J. An overview of central fetal monitoring systems in labor J. Perinat. Med. 2013 41 93 99 10.1515/jpm-2012-0067 23093259 21. Doret M. Spilka J. Chudacek V. Goncalves P. Abry P. Fractal Analysis and Hurst Parameter for Intrapartum Fetal Heart Rate Variability Analysis: A Versatile Alternative to Frequency Bands and LF/HF Ratio PLoS ONE 2015 10 1 19 10.1371/journal.pone.0136661 22. Petrozziello A. Redman C.W. Papageorghiou A.T. Jordanov I. Georgieva A. Convolutional Neural Networks to Detect Fetal Compromise During Labor and Delivery IEEE Access 2019 7 112026 112036 10.1109/ACCESS.2019.2933368 23. Ogasawara J. Ikenoue S. Yamamoto H. Sato M. Kasuga Y. Mitsukura Y. Ikegaya Y. Yasui M. Tanaka M. Ochiai D. Deep neural network-based classification of cardiotocograms outperformed conventional algorithms Sci. Rep. 2021 11 1 9 10.1038/s41598-021-92805-9 33414495 24. Rahmayanti N. Pradani H. Pahlawan M. Vinarti R. Comparison of Machine Learning Algorithms to Classify Fetal Health Using Cardiotocogram Data Sixth Information Systems International Conference (ISICO 2021) Procedia Computer Science Surabaya, Indonesia 2021 162 171 25. Goldberger A.L. Amaral L.A. Glass L. Hausdorff J.M. Ivanov P.C. Mark R.G. Mietus J.E. Moody G.B. Peng C.K. Stanley H.E. PhysioBank, PhysioToolkit, and PhysioNet: Components of a New Research Resource for Complex Physiologic Signals Circulation 2000 101 e215 e220 10.1161/01.CIR.101.23.e215 10851218 26. Das S. Roy K. Saha C.K. Establishment of Automated Technique of FHR Baseline and Variability Detection Using CTG: Statistical Comparison with Expert's Analysis Int. J. Inf. Eng. Electron. Bus. 2019 10 1 9 10.5815/ijieeb.2019.01.04 27. Das S. Roy K. Saha C.K. A novel step towards machine diagnosis of fetal status in utero: Calculation of baseline variability Proceedings of the 2015 IEEE International Conference on Research in Computational Intelligence and Communication Networks (ICRCICN) Kolkata, India 20-22 November 2015 IEEE Press Kolkata, Indonesia 2015 230 234 28. Hair J.F. Black W.C. Babin B.J. RE Anderson Multivariate Data Analysis: A Global Perspective Pearson Prentice Hall New Jersey, NJ, USA 2020 29. Tabachnick B.G. Fidell L.S. Using Multivariate Statistics Pearson Education Inc. Boston, MA, USA 2007 30. Chan L.L. Idris N. Validity and Reliability of The Instrument Using Exploratory Factor Analysis and Cronbach's alpha Int. J. Acad. Res. Bus. Soc. Sci. 2017 7 400 410 31. Giavarina D. Understanding Bland Altman analysis Biochem. Med. 2015 25 141 151 10.11613/BM.2015.015 32. Signorini M.G. Pini N. Malovini A. Bellazzi R. Magenes G. Integrating machine learning techniques and physiology based heart rate features for antepartum fetal monitoring Comput. Methods Programs Biomed. 2019 185 1 29 10.1016/j.cmpb.2019.105015 31678794 33. Zhao Z. Deng Y. Zhang Y. Zhang Y. Zhang X. Shao L. DeepFHR: Intelligent prediction of fetal Acidemia using fetal heart rate signals based on convolutional neural network BMC Med Inf. Decis. Mak. 2019 19 1 15 10.1186/s12911-019-1007-5 |
PMC10000593 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050698 healthcare-11-00698 Article At Risk Safety Behaviors of the Perioperative Nursing Team: A Direct Observational Study Letvak Susan Conceptualization Methodology Formal analysis Investigation Writing - original draft 1* Apple Brandi Investigation Data curation Writing - review & editing 2 Jenkins Marjorie Methodology Data curation Writing - review & editing 3 Doss Carrie Investigation 1 McCoy Thomas P. 1 Chen Tin-Chih Toly Academic Editor Vetrugno Luigi Academic Editor 1 Adult Health Nursing Department, University of North Carolina, Greensboro, NC 27215, USA 2 Chapel Hill School of Nursing, University of North Carolina, Greensboro, NC 27215, USA 3 Cone Health System, Greensboro, NC 26170, USA * Correspondence: [email protected] 27 2 2023 3 2023 11 5 69823 12 2022 11 2 2023 22 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Background: The operating room setting has unique workforce hazards and extremely high ergonomic demands due to patient lifting/positioning requirements, long periods of standing, and the heavy equipment and supplies that are needed for surgical procedures. Despite worker safety policies, injuries among registered nurses are increasing. Most of the research on the ergonomic safety of nurses is conducted utilizing survey methodology, which may not provide accurate data. It is imperative to understand the at-risk safety behaviors that perioperative nurses face if we are to design interventions to prevent injury. Methods: Two perioperative nurses were directly observed during 60 different operating room surgical procedures (n = 120 different nurses). Data were collected utilizing the job safety behavioral observation process (JBSO), which is designed specifically for the operating room environment. Results: There were 82 total at-risk behaviors observed amongst the 120 perioperative nurses. More specifically, 13 (11%) of the surgical procedures had at least one perioperative nurse observed in a position of at-risk behavior, and a total of 15 (12.5%) individual perioperative nurses performed at least one at-risk behavior. Conclusion: More attention must be placed on the safety of the perioperative nurse if we are to retain a healthy, productive workforce to provide the highest quality patient care. perioperative setting at risk safety factors observational study CCI FoundationSpecial thanks to the CCI Foundation for funding this study. pmc1. Introduction The profession of nursing is known to face workplace hazards and has high ergonomic demands resulting in occupational injury. While only 61% of registered nurses (RNs) work in hospital settings, these nurses experience 74.1% of RN nonfatal workplace injuries and illnesses, with the majority of these injuries being work-related musculoskeletal injuries/disorders (MSD), which is significantly higher than the rate for all occupations . Despite having worker safety policies in all hospital settings, the US Bureau of Labor Statistics reports occupational injuries in registered nurses (RNs) resulting in time away from work, which increased by 42.2% from 2019 to 2020. Ergonomic-specific increases include increased contact with objects and equipment (1990 to 2830), falls/slips/trips (5320 to 5390), and over-exertion/bodily reactions (8610 to 10,510). Moreover, the state of North Carolina had a 335% increase in RN occupational injuries leading to time away from work increasing from 200 to 870 from 2019 to 2020. Workplace injuries in nurses impact not only the nurse but also have monetary and societal costs that necessitate the need to further understand the specific workplace hazards nurses face . Perioperative RNs make up less than six percent of the RN workforce and, thus, are often lost in the nursing workforce and ergonomic studies. However, RNs are extremely important members of the perioperative team in assuring the quality outcomes of patients undergoing surgical procedures. Shewchuk states it is impossible to understand why there is a need for RNs in the operating room without first understanding the nursing roles in the operating room. One essential role is that of the circulating nurse. The circulating nurse coordinates all aspects of the perioperative environment, especially communication, and assures compliance with policies and procedures. The circulating nurse assists the entire perioperative team in maintaining a safe environment and must also be ready to step into the scrub nurse role if needed. The role of the scrub nurse, who is part of the sterile field, is to have knowledge of all steps of surgical procedures to anticipate surgeon needs by passing the needed instruments, sponges, and supplies at the correct time. A newer role for RNs in the operating room is the registered nurse first assistant (RNFA). The RNFA is an advanced practice role for nurses who perform surgical interventions under the supervision of a surgeon . Specifically, the RNFA is allowed to make incisions, cauterize, suction, retract tissue, and suture surgical patients in the operating room. Finally, RNs may also have the role of certified registered nurse anesthetist (CRNA) in the operating room. The American Association of Nurse Anesthesiology defines CRNA as advanced practice nurses licensed as independent practitioners who plan and deliver anesthesia, pain management, and related care to patients of all health complexities across their lifespans. In the operating room, the CRNA is responsible for the delivery of anesthesia and pain relief for the patient undergoing surgery. The operating room setting is important for research because there are significant workforce hazards and extremely high ergonomic demands due to patient lifting/positioning requirements, long periods of standing, and the heavy equipment and supplies that are needed for surgical procedures. Moreover, incorporating ergonomic interventions into the operating room is more difficult than in other workforce settings because of demanding tasks and restrictive environments that prioritize patient safety, as well as sterility requirements . It is important to note that all RNs in the operating room share safety risks, including all RN roles that are required to lift and position patients. Because of the high ergonomic demands of the operating room, in 2011, the Association of Perioperative Registered Nurses (AORN) reviewed the research literature to identify seven ergonomic risk factors of perioperative nurses and developed ergonomic tool kits that were specific for each risk factor and updated in 2018. These tool kits provide research-supported specific safety measures for each ergonomic risk factor, such as the maximum weight to be lifted, pushed, or pulled and time limits for the use of hands as a tissue retractor and standing in one position. More recently, Yasak and Vural identified additional perioperative workforce hazards, which include awkward positioning, forcible exertion, lifting while bending or twisting, overreaching, repeated motions, static postures, floors that are wet or covered with debris, heavy operating room doors, and moving oxygen tanks. Moreover, there have been a few more recent research studies focusing on variables that influence safety hazards for perioperative nurses. Abdollahzade and colleagues measured the individual demographics, job type, and working posture of 147 operating room nurses in Iran. The researchers found that gender and the type of operating room (e.g., orthopedic versus cardiac) predicted poor posture, which could lead to musculoskeletal disorders and injury. Researchers in Italy analyzed the association between personal and job characteristics and the risk of upper limb work-related musculoskeletal disorders in 148 operating room nurses. Regression analyses revealed that being female and in the scrub nurse role was associated with a higher score on the disabilities of the arm, shoulder, and hand (DASH) questionnaire placing these nurses at great risk for upper limb disorders. Most of the research conducted on the ergonomic safety of nurses was conducted utilizing survey methodology and has now been outdated. Survey designs are known to have low response rates and may also have response distortion, which compromises the validity of self-report measures in high-stake situations , such as injury and worker safety reporting. Thus, the purpose of this study was to determine the at-risk safety factors perioperative nurses experience through direct observation in the operating room setting. It is only through understanding specific risks that interventions and needed safety education can be developed. This study was guided by applied behavioral analysis , which is derived from Skinner's behavioral science theory and offers a method to improve occupational safety by directly observing human behavior in the workplace setting to reduce worker injury. Observational data are useful when trying to understand situations, gather data on individual behaviors, and obtain knowledge about the physical setting . Additionally, observational data can be particularly useful in the operating room setting to provide insight into this uniquely complex environment. 2. Methods 2.1. Study Design and Human Subjects Protection We employed an observational design utilizing non-participant observation to observe the at-risk safety behaviors of perioperative nurses in the operating room. Creswell and Poth defined non-participant observations as the researcher being uninvolved with the group under study: only watching and taking notes from a distance. IRB approval was first received (#21-0070). An advantage of non-participant observation is that the researcher has a more objective view of what is being observed. As part of the IRB approval, prior to observations, the participating health system required that a consent to shadow and observe form be signed by the researchers. In addition to assuring complete confidentiality of all data collected, the observers agreed that only health system perioperative nurses would be observed, and no patients would be observed. 2.2. Study Setting and Sample This study was conducted in a large healthcare system located in the southeastern United States, which performs over 45,000 surgical procedures annually. The sample consisted of perioperative nurses working at the bedside in the operating room (RN circulating nurses, RN scrub nurses, CRNAs, and RNFAs). All participants had agreed to have an observer present during the surgery that was being performed. 2.3. Data Collection and Measurement Tool As stated above, data collection was required following the health system's policy of having an observer present in the operating room. The perioperative charge nurse assisted with determining suitable surgical cases which might have ergonomic risk factors to observe and obtained permission from the surgeon and perioperative team to have an observer present. Patient consent forms already included observers possibly being present. Operating room staff are used to observers being present; thus, there were no refusals or questions asked of the researcher who was present in the operating room. The researchers were careful to only observe the perioperative observations were made of the patient, other team members, or other observers that were present during the surgical procedures. To assure as much anonymity as possible, no demographic information (such as perceived gender, age, or race/ethnicity) of the perioperative nurses observed was recorded. Additionally, no written documentation was made of any conversations that may have been overheard by the researcher. McNaughton et al. recommend that the procedures to be observed took place over a number of sites and were best observed in a range of locations. Thus, direct observations were made at the major medical center hospital, three community hospital sites, and several free-standing surgical center sites. Sixty surgical procedures, as recommended in Yasak and Vural's observational study, were observed. The surgical procedures observed included a variety of specialties to ensure capturing the ergonomic demands of RNs working in different surgical specialty lines. Observational data were collected using the job safety behavioral observation process (JBSO) designed specifically for the operating room environment by Simon et al. and included all the AORN ergonomic safety risk factors. The form allowed the observer to document if the ergonomic risk factor being performed was safe (S) and followed ergonomic safety precautions or whether it was at-risk (AR) and not following ergonomic safety measures. An open-ended "comment" area was used for field notes. The co-PI, who is an experienced, certified perioperative nurse, collected observational data with a research assistant who was also highly experienced and certified in perioperative nursing. They conducted three observations together to assure inter-rater reliability with data collection. 3. Data Analysis All observation checklist data were entered into SPSS.v28 software . Descriptive statistics were then used to determine the specific risk factors that perioperative nurses were exposed to during surgical procedures. Findings Over the course of two months, two perioperative nurses were observed during 60 different operating room surgical procedures for a total of 120 perioperative nurses who were directly observed for at-risk safety behaviors. Of the 120 different perioperative nurses observed, 60 (50%) were in the circulating nurse role, 45 (38%) were certified nurse anesthetists (CRNAs), 11 (9%) were in the scrub nurse role, and 4 (3%) were RN first assistants (RNFAs). Of the 60 surgical procedures observed, there were 82 total at-risk behaviors observed amongst the 120 perioperative nurses. More specifically, 13 (11%) of the surgical procedures had at least one perioperative nurse who was observed conducting at-risk behavior, and a total of 15 (12.5%) of the individual perioperative nurses performed at least one at-risk behavior. Table 1 identifies specific at-risk behaviors with the number and type of perioperative nurses who were at risk. The top three at-risk behaviors observed included overreaching (16 observations), working in an awkward position (15 observations), and lifting while bending or twisting (13 observations). Circulating nurses had the largest number of at-risk behaviors but also made up 50% of those observed. CRNAs had 22% of the at-risk behaviors and made up 38% of those observed. Scrub nurses had 16% of the at-risk behaviors, followed by 6% for the RNFAs; however, they made up 9% and 3% of those observed, respectively. 4. Discussion Most hospitals prioritize patient safety; however, Fencl and colleagues argue that patient safety also requires a just culture and prioritizing nurse safety in the perioperative work setting. This study documents that despite well-publicized AORN safety tool kits, perioperative nurses have numerous personal at-risk behaviors, as well as environmental risks, while working in the operating room setting. First, environmental risks that place the perioperative nurse at risk of injury are heavy doors and operating room floors that have debris or are wet. Operating room doors are over-pressurized and hermetically sealed to prevent contaminated air from entering and increasing the risk of surgical infections . In this study, all the operating room doors were electronic; thus, there were no observations of having to open a heavy door, especially while pushing or carrying a heavy object. Of the 60 surgical procedures observed, only one operating room had a very wet floor which exposed the CRNA and circulating nurse to a risk of falling, and five circulating nurses and two scrub nurses were at risk for falling because of debris on the floors. The United States Department of Labor, Occupational Safety and Health Administration (OSHA) has a standard that walking-working surfaces must be clean and dry. When wet processes are used, dry-standing places (such as platforms or mats) must be provided, and all walking-work surfaces should be free of hazards. The majority of the safety risks observed in this study were at the personal level. Over-reaching was observed in each of the four perioperative nurse roles (CRNA, circulating nurse, scrub nurse, and RN first assistant). Overreaching was observed when performing tasks such as placing EKG leads over a very obese patient as well as while lifting and moving a 2000 mL irrigation bag over a patient. Reaching is considered an awkward body posture that can lead to injury . Surprisingly, despite the obvious strain over-reaching places on the spine, especially when reaching with weight in one's hands, no evidence was found in the literature that specified over-reaching as an ergonomic risk factor for nurses, and "reaching" was only discussed as an awkward body position. Fifteen of the 120 perioperative nurses were observed working in awkward positions, mostly in a twisted position while standing or hunched over. An additional 13 perioperative nurses were observed lifting patients or equipment while bending or twisting. The U.S. OSHA clearly lists ergonomic risk factors for MSD injury or disorders as working in awkward positions, including twisting while lifting. Additionally, seven different perioperative nurses were observed either lifting too-heavy objects, transferring a too-heavy patient (six), positioning a too-heavy patient (six), or pushing/pulling too heavy of weight (one). These behaviors are classified by OSHA as exerting excessive force, putting the worker at risk for MSD injuries and disorders. It is important to note that 11% of the 60 surgical procedures observed had at least one perioperative nurse performing an at-risk behavior. The circulating nurse role had the largest number of at-risk behaviors (56%); however, they also made up 50% of the perioperative nurses observed. Circulating nurses, who are non-sterile members of the operating room team, are in a multi-dimensional role and are "caretakers of the highly technical surgical environment and supervisor of surgical team's activities while concurrently directing patient care . These highly trained nurses are in the critical role of assuring the safety of patients, and more efforts must be directed toward their ergonomic safety. 5. Implications This observational study has implications for all who work in the perioperative setting. Each of the observed safety risk factors has the potential to put a perioperative nurse at risk for injury. Clearly, floors should be free of debris and water, and if certain surgical procedures produce excess debris and water, platforms should be provided to keep staff safe until the debris and water can be removed. Perioperative nurses must be continuously reminded of the dangers of lifting/pushing too-heavy weights and equipment over the recommended weight standards. Time pressure should not prevent perioperative nurses from seeking additional help for lifting and pushing heavy weights or obtaining the support needed for lifting equipment. Nurse managers must advocate for more time between operating room procedures to remove the time pressure placed on perioperative nurses. Perioperative nurses must be cognizant at all times of how long they are standing in one position or wearing too heavy of a weight. Specifically, they should not stand more than 30% of an 8 h workday or be in the same position for more than two hours, especially while wearing a lead apron. The AORN offers four small changes that can be made for ergonomic safety: (1) the removal of unnecessary risks: and even small fixes can prevent one injury. (2) The application of formal fixes should be made, as quick fixes may cause more safety risks. (3) Safety data should be reviewed and clear plans made. Injury data should be combined with audit data so that action plans can be developed. (4) Safety rules should be rewritten using a three-column chart format. The first column outlines the safety action, the second column describes the potential risk of injury, and the third column describes actions for mitigating the risk. There is a strong need for interventional research to reduce physical ergonomic demands on the perioperative team. Cha and colleagues conducted simulation-based research utilizing exoskeletons on the perioperative team, including nurses. Exoskeletons are external devices that are worn to support physical demands and tasks and are currently being used in agriculture and manufacturing settings. This study's findings demonstrate that perioperative team members were receptive to wearing an exoskeleton to decrease the risk of pain and injury; however, this technology is still in the development phase and is not ready for healthcare worker use. In addition to exoskeletons, the use of body sensors, or wearable technology, also offers promise for ergonomic safety for perioperative nurses. Stefana and colleagues conducted a systematic integrative review on the use of sensors for ergonomic safety. They identified 28 articles based on 24 studies, the majority of which propose wearing sensors to analyze unfavorable postures during working tasks, while others focused on sensors measuring physical loads. Most of the studies focused on construction workers, with only one study focused on healthcare workers. Unfortunately, as with exoskeletons, the use of sensors to improve ergonomic safety is still only being tested in lab studies and is not ready for the real-world context, and more research is needed. Overlooking safety risk factors has significant implications for both the perioperative team, as well as the healthcare system. To date, while technology is being investigated to assist with worker ergonomics, awareness and prevention of safety risk factors are at the individual and organizational levels. Ignoring safety risk factors can lead to perioperative nurse injury, causing both short and long-term musculoskeletal pain. Worker injury costs the healthcare system through absenteeism, compensation claims, as well as decreased productivity. 6. Limitations This study was conducted with perioperative nurses in only one large healthcare system. Only 120 different nurses in 60 different surgical procedures were observed; additional safety violations may have been missed. Additionally, while subjectivity may be a concern, the tool used to assess safety risks was developed from safety risks identified by the AORN and utilized in the operating room safety risk study by Yasak and Vural . 7. Conclusions The risk factors of working in a perioperative setting are often overlooked in our fast-paced working environments. This study documents the specific safety risk factors perioperative nurses face in the operating room setting. Interventional research is urgently needed to reduce the safety of perioperative nurses if we are to retain a healthy, productive workforce to provide the highest quality of patient care. Author Contributions Conceptualization: S.L., Methodology: S.L., B.A., M.J. and T.P.M.; Validation: S.L., B.A., M.J., C.D. and T.P.M.; Formal analysis: S.L. and B.A.; data curation: B.A. and C.D.; writing--original draft preparation: S.L.; Writing-review and editing: S.L., B.A., M.J., C.D. and T.P.M.; Funding acquisition: S.L. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Institutional Review Board Statement: This study was approved by the University of North Carolina at Greensboro's Institutional Review Board. IRB #21-0070. Informed Consent Statement Patient consent was waived due to observational design and not wanting subjects to change "normal" ergonomic behaviors due to being observed. The researchers signed a consent to shadow and observe form provided by the health system and agreed to not observe any patients or members of the operating room team other than the registered nurses' ergonomic behaviors. No overheard conversations were recorded. No individual names or any demographic or identifying information was recorded. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. healthcare-11-00698-t001_Table 1 Table 1 At risk safety issues by perioperative nurses. Ergonomic Safety Risk Circulating Nurse Scrub Nurse CRNA RNFA TOTAL Awkward position 9 4 1 1 15 Forcible exertion 1 0 1 1 3 Lifting while bending or twisting 7 2 3 1 13 Overreaching 9 1 5 1 16 Repeated motions 0 0 0 0 0 Static posture 0 0 0 0 0 Wrist deviation 1 0 2 1 4 Lifting too-heavy objects 5 1 0 0 6 Pushing/pulling too much weight 1 0 0 0 1 Wearing a lead vest for too long 0 1 0 0 1 Wet floor 1 0 1 0 2 Opening too-heavy of a door 0 0 0 0 0 Transferring too-heavy of a patient/not enough help 4 1 2 0 7 Positioning too-heavy of a patient/not enough help 3 0 3 0 6 Standing in one position for too long 0 1 0 0 1 Carrying/pushing heavy oxygen tank 0 0 0 0 0 Debris on floor 5 2 0 0 7 Total number of at-risk behaviors 46 13 18 5 82 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Dressner M. Kissinger S. Occupational injuries and illnesses among registered nurses Mon. Labor Rev. 2018 10.21916/mlr.2018.27 2. Bureau of Labor Statistics, U.S. Department of Labor Nonfatal Injuries and Illnesses Resulting in Days off Work among Nurses up 291 Percent in 2020 the Economics Daily Available online: (accessed on 15 December 2022) 3. Smiley R.A. Lauer P. Bienemy C. Berg J.G. Shireman E. Reneau K.A. Alexander M. The 2017 national nursing workforce survey J. Nurs. Regul. 2018 9 S1 S88 10.1016/S2155-8256(18)30131-5 4. Shewchuk M. Why a registered nurse (RN) in the OR? The perioperative registered nurse role--A well kept secret Can. Oper. Room Nurs. J. 2007 25 38 39 18193726 5. Pika R. O'Brien B. Murphy J. Markey K. O'Donnell C. The role of the registered nurse first assistant within the perioperative setting Br. J. Nurs. 2021 30 148 153 10.12968/bjon.2021.30.3.148 33565935 6. American Association of Nurse Anesthesiology Scope of Nurse Anesthesia Practice Available online: /professional-practice-manual/scope-of-nurse-anesthesia-practice.pdf?sfvrsn=250049b1_10 (accessed on 26 December 2022) 7. Cha J.S. Monfared S. Stefanidis D. Nussbaum M.A. Yu D. Supporting surgical teams: Identifying needs and barriers for exoskeleton implementation in the operating room Hum. Factors J. Hum. Factors Ergon. Soc. 2019 62 377 390 10.1177/0018720819879271 8. Yasak K. Vural F. Assessment of the Environmental and Physical Ergonomic Conditions of ORs in Turkey AORN J. 2019 110 517 523 10.1002/aorn.12841 31660590 9. Abdollahzade F. Mohammadi F. Dianat I. Asghari E. Asghari-Jafarabadi M. Sokhanvar Z. Working posture and its predictors in hospital operating room nurses Health Promot. Perspect. 2016 6 17 10.15171/hpp.2016.03 27123432 10. Clari M. Garzaro G. Di Maso M. Donato F. Godono A. Paleologo M. Dimonte V. Pira E. Upper limb work-related musculoskeletal disorders in operating room nurses: A multicenter cross-sectional study Int. J. Environ. Res. Public Health 2019 16 2844 10.3390/ijerph16162844 31395811 11. Apple B. Letvak S. Ergonomic challenges in the perioperative setting: A synthesis of the literature AORN J. 2021 113 339 348 10.1002/aorn.13345 33788237 12. Malham P.B. Saucier G. The conceptual link between social desirability and cultural normativity Int. J. Psychol. 2016 51 474 480 10.1002/ijop.12261 26914819 13. Williams J.H. Geller E. Behavior-based intervention for occupational safety: Critical impact of social comparison feedback J. Saf. Res. 2000 31 135 142 10.1016/S0022-4375(00)00030-X 14. Skinner B.F. Science and Human Behavior Simon and Schuster New York, NY, USA 1965 15. Centers for Disease Control Data Collection for Program Evaluation: Observation. Evaluation Brief. No. 16 Available online: (accessed on 23 December 2022) 16. Creswell J.W. Poth C.N. Qualitative Inquiry and Design: Choosing among Five Approaches 4th ed. Sage London, UK 2013 17. McNaughton Nichols C. Mills L. Kotecha M. Observation in qualitative research practice Qualitative Health Research in Health Care 2nd ed. Ritchie J. Lewis C. McNaughton Nicolls C. Ormston R. Sage London, UK 18. Simon R.W. DiTullio B.L. Foster L.A. Canacari E.G. Gugleilmi C.L. Eliminating workplace hazards in the OR Perioper. Care Oper. Management. 2016 2 18 27 10.1016/j.pcorm.2016.01.004 19. IBM Corp IBM SPSS Statistics for Windows, Version 28 Released Armonk, IBM Corp New York, NY, USA 2022 20. Fencl J.L. Willoughby C. Jackson K. Just culture: The foundation of staff safety in the perioperative environment AORN J. 2021 113 329 336 10.1002/aorn.13352 33788231 21. Sadrizadeh S. Pantelic J. Sherman M. Clark J. Abouali O. Airborne particle dispersion to an operating room environment during sliding and hinged door opening J. Infect. Public Health 2018 11 631 635 10.1016/j.jiph.2018.02.007 29526441 22. United States Department of Labor, Occupational Safety and Health Administration Occupational Safety and Health Standards Available online: (accessed on 2 February 2023) 23. Choi S.D. Brings K. Work-related musculoskeletal risks associated with nurses and nursing assistants handling overweight and obese patients: A literature review Work 2016 53 439 448 10.3233/WOR-152222 26835850 24. United States Department of Labor Ergonomics. Ergonomics--Overview|Occupational Safety and Health Administration Available online: (accessed on 30 January 2023) 25. Crocco I.M. Goodlitt L. Parkosewich J.A. The effect of an educational intervention on nurses' knowledge, perception, and use of nursing presence in the perioperative setting J. PeriAnesthesia Nurs. 2022 in press 10.1016/j.jopan.2022.06.005 36494296 26. Association of periOperative Nurses (AORN) 4 Small Changes to Improve Workplace Safety 2019 Available online: (accessed on 19 December 2022) 27. Stefana E. Marciano F. Rossi D. Cocca P. Tomasoni G. Wearable devices for ergonomics: A systematic literature review Sensors 2021 21 777 10.3390/s21030777 33498904 |
PMC10000594 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050770 healthcare-11-00770 Article Functional Disability and Symptomatic Slow-Acting Drugs for Osteoarthritis in Adults with Periodontitis Nik-Azis Nik-Madihah Conceptualization Methodology Formal analysis Investigation Writing - original draft Visualization Project administration 1* Mohd Nurulhuda Conceptualization Methodology Writing - review & editing Supervision Project administration Funding acquisition 1 Baharin Badiah Conceptualization Data curation Writing - review & editing Supervision Project administration Funding acquisition 1 Mohd Fadzilah Fazalina Conceptualization Methodology Formal analysis Resources Writing - review & editing Project administration 2 Mohamed Haflah Nor Hazla Methodology Investigation Data curation Writing - review & editing Project administration 3 Mohamed Said Mohd Shahrir Conceptualization Methodology Investigation Writing - review & editing Supervision Project administration 4 Kanno Takahiro Academic Editor Manicone Paolo Francesco Academic Editor De Angelis Paolo Academic Editor 1 Department of Restorative Dentistry, Faculty of Dentistry, Universiti Kebangsaaan Malaysia, Jalan Raja Muda Abdul Aziz, Kuala Lumpur 50300, Malaysia 2 Radiology Department, Sunway Medical Centre, Bandar Sunway, Selangor 47500, Malaysia 3 Department of Orthopaedics and Traumatology, Faculty of Medicine, Hospital Canselor Tuanku Mukhriz, Jalan Yaacob Latif, Bandar Tun Razak Cheras, Kuala Lumpur 56000, Malaysia 4 Faculty of Medicine, Hospital Canselor Tuanku Mukhriz, Jalan Yaacob Latif, Bandar Tun Razak Cheras, Kuala Lumpur 56000, Malaysia * Correspondence: [email protected]; Tel.: +60-3-92897745 06 3 2023 3 2023 11 5 77028 12 2022 28 2 2023 02 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Osteoarthritis (OA) patients have decreased functional ability and restricted access to healthcare facilities and are on a spectrum of medications. These can impact their oral health. This study aims to investigate the association between periodontal disease and OA disease parameters, specifically the functional disability and the medications taken. This was a cross-sectional study on OA participants recruited from the Hospital Canselor Tuanku Mukhriz. Periodontal health parameters were obtained from an oral examination of the participants. A Health Assessment Questionnaire (HAQ) was administered to ascertain the functional status of the participants. Out of the 130 participants recruited, 71 (54.6%) had periodontitis. There was a correlation between the teeth count with OA severity, where participants with a greater Kellgren-Lawrence score had less teeth (rs = 0.204, p = 0.025). Participants with a greater degree of functional limitation also had less teeth (rs = -0.181, p = 0.039) and a higher clinical attachment loss (rs = 0.239, p = 0.006). There were no associations found between the symptomatic slow-acting drugs in OA and periodontal health parameters. In conclusion, there was a high proportion of periodontitis in patients with OA. Functional disability was associated with measures of periodontal health. It is suggested that clinicians treating OA patients consider the need for a referral for dental care when managing this group of patients. periodontal disease health assessment questionnaire functional disability symptomatic slow-acting drugs in osteoarthritis Universiti Kebangsaan MalaysiaTAF316 This research was funded by Universiti Kebangsaan Malaysia (grant number TAF316). pmc1. Introduction Osteoarthritis (OA) is traditionally described as a 'wear and tear', mechanically driven disease, that is, a consequence of the aging process affecting mostly older adults . It is a disorder involving movable joints, characterized by cell stress and extracellular matrix degradation initiated by macro-injury. This process stimulates maladaptive repair responses, including pro-inflammatory pathways of the innate immunity . More contemporary views of OA argued that the inflammation contributes to OA synovitis and its pathology . OA is now viewed as a complex disease, with inflammatory mediators released by cartilage, bone and synovium and its progression governed by a set of multifactorial components . The role of inflammation in this new paradigm highlights inflammatory signals, including cytokines, surface-expressed pattern recognition receptors, complement factors, pathogen-associated molecular patterns and damage-associated molecular patterns that can lead to the degradation of the cartilage matrix in OA . Periodontal disease (PD) is a disease that affects the supporting structures of the teeth. It is characterized by microbially associated, host-mediated inflammation that results in loss of periodontal attachment . Periodontitis is an immune-inflammatory infection that can cause low-grade systemic inflammation, which may influence the development of systemic comorbidities. The periodontitis-associated systemic inflammation can cause hematogenous dissemination of periodontal bacteria and inflammatory mediators from periodontal tissues to the bloodstream . This can later spread to other parts of the body. Despite the similarities between the role of inflammation in the pathogenesis of OA and PD, limited studies have investigated the association between the two diseases and its related factors. Chung and co-workers studied the data of 7969 adults from the Korea National Health and Nutrition Examination Survey (KNHANES) during 2010-2014. They reported that severe OA and periodontitis were associated with each other in a subgroup analysis involving female patients . Kim et al. also reported that periodontitis was associated with the presence and severity of radiographic knee OA in their Korean nationwide study, also using data from the KNHANES . Other studies on periodontal disease and OA were case-control studies where OA participants were recruited as controls. These studies include those from United States veterans, reporting 26.4% of OA participants with periodontitis , and Vietnam, where 28% of OA participants had periodontitis . OA can involve the joints of the upper extremities, causing functional limitations of the hand. In a study on OA participants, radiographic changes due to OA were found to be associated with reduced grip and pinch strength . This can limit the ability of the participants to utilize oral hygiene aids and remove plaque efficiently . Apart from that, patients with OA can have mobility limitations that affect their ability to access the dental clinic for routine dental care . The accessibility of dental services and level of patient care, as well as transport availability, were suggested as factors that affect the use of dental services by patients with arthritis . Oral medications used to treat OA can broadly be categorized into analgesics, including paracetamol; non-steroidal anti-inflammatory drugs (NSAIDs); and symptomatic slow acting drugs for OA (SYSADOAs), including glucosamine, chondroitin sulphate, diacerein and avocado and soybean unsaponifiable (ASU). There is evidence that NSAIDs can cause reduced bleeding and a reduction in the rate of bone resorption in observational and intervention studies . Diacerein has been hypothesized to be a potential therapeutic drug for periodontitis due to its anti-inflammatory activities that selectively inhibits signal transduction affecting cytokine profiles and ameliorating disease breakdown . In an animal study investigating the effects of diacerein in the management of ligature-induced experimental periodontitis in rats, significant decreases in the IL-1ss level in the test group suggest that diacerein may play a therapeutic role as a potent anti-inflammatory drug in the management of periodontitis . ASU is routinely prescribed in OA and has been suggested as an adjunct for the treatment of periodontitis. ASU added to gingival fibroblasts in culture showed the potential to prevent the deleterious effects of IL-1ss in periodontitis . Animal studies in rats also showed that ASU prescribed as an adjunct to conventional mechanical debridement has subtle beneficial effects on periodontal repair . However, a clinical study found that ASU did not have a favorable effect in the treatment of chronic periodontitis . Current evidence shows that glucosamine exhibits anti-inflammatory effects by reducing the levels of pro-inflammatory factors, such as tumor necrosis factor-alpha, interleukin-1 and interleukin-6 . There is, however, limited data on the effects of glucosamine and chondroitin sulphate on periodontitis. A randomized controlled trial investigated the effects of Arthocare Forte containing chondroitin (400 mg) and glucosamine sulphates (500 mg) administered to patients with tooth mobility over a period of 5 years . The participants underwent non-surgical periodontal therapy, and half were treated with Arthocare Forte. They found that the medication speeds up the regenerative capacity and the stability of the periodontium compared to the control group. Generally, there is limited evidence on the effects of medications, such as the SYSADOAs and NSAIDs used to treat OA, on the oral health parameters and periodontal disease. Further studies are needed in this area to assist in understanding how these medications affect specific oral health parameters and whether there is potential for these medications to be used in the treatment of periodontitis. It was hypothesized that: (a) the prevalence and severity of PD is higher in participants with OA compared to the general population; (b) participants with functional limitations have a greater PD severity, and (c) there is a difference in the PD among participants taking the different SYSADOAs. Hence, this study aimed to investigate: (a) the prevalence and severity of PD among patients with OA compared to the general population; (b) the association between PD parameters and OA parameters, including functional limitations, and (c) the differences in PD among participants taking different SYSADOAs. 2. Materials and Methods The participants were recruited from the Osteoarthritis Clinic under the Orthopaedic Department in Hospital Canselor Tuanku Mukhriz, Malaysia, using convenience sampling. Consecutive patients attending the OA Clinic who met the study criteria were invited to participate in the study. The flowchart of the recruitment process is shown in Figure 1. The inclusion criteria were: (i) OA patients, as confirmed by the ACR Classification, (ii) above the age of 18 years old, (iii) dentate and (iv) able to give verbal and written consent. The exclusion criteria were: (1) patients who were completely unable to read, write or understand the Malay or English Language; (2) coexistence of other autoimmune diseases; (3) uncontrolled systemic disease or malignancy; (4) patients who were pregnant or planning to become pregnant; (5) any current or previous history of periodontal treatment, including root surface debridement/ periodontal surgery; and (6) any previous or current use of phenytoin or cyclosporin. This study was conducted as part of research assessing oral health in patients with joint diseases . Ethical approval for the study was obtained from the Ethical Board of the Universiti Kebangsaan Malaysia (UKM/PPI/111/8/JEP-2017-553), and the study conformed to the provisions of the Declaration of Helsinki (as revised in Brazil 2013). All participants of the study gave informed consent, and the anonymity of the participants has been preserved in the conduct and reporting of this study. 2.1. Oral Examination The assessment of the oral and periodontal health used parameters such as the number of remaining teeth, the Plaque Index , the Gingival Index , the probing pocket depth (periodontitis) and clinical attachment loss (CAL). For the Plaque Index, the original O'Leary protocol used a disclosing solution (such as Bismarck Brown Iodine Stain). However, the disclosing solution was not used in our study. A dichotomous scoring system of (present/absent) was used. Six sites per tooth, namely, the mesiobuccal, midbuccal, distobuccal, mesiopalatal, midpalatal and distopalatal sites, were assessed for both the Plaque Index and the Gingival Index. The periodontal charting was carried out using the University of North Carolina-15 (UNC15) probe. The clinical attachment loss (CAL) was then calculated as a sum of the pocket depth and recession. Diagnosis of periodontitis was made based on the criteria as outlined by Papapanou et al. . They are as follows: (1) interdental CAL is detectable at >=2 non-adjacent teeth, or (2) buccal or oral CAL >=3 mm with pocketing >=3 mm is detectable at >=2 teeth, but the observed CAL cannot be ascribed to non-periodontitis-related causes. The periodontal findings were also later recoded to the Community Periodontal Index (CPI) to allow for comparison with the national data available for Malaysia . Oral examination was conducted by a single examiner (NMNA). Prior to the initiation of the study, NMNA was calibrated against a gold-standard periodontist (NM) to ensure intra-examiner reliability. Patients who were diagnosed with any dental disease were given a follow-up appointment or advised to seek care at any dental clinic if they were not able to attend for the appointment given. 2.2. Measures of Functional Limitations The Malaysian version of the Health Assessment Questionnaire (HAQ) was administered to ascertain the functional status of the subject. The authors have permission to use this instrument from the copyright holders. The HAQ is a questionnaire with eight sections (Dressing and Grooming, Arising, Eating, Walking, Hygiene, Reach, Grip and Activities) that is validated to assess the functional limitations in participants with musculoskeletal diseases. The HAQ also included vertical, 100 mm, patient global and pain visual analogue scales. Administration of the questionnaire was done by a single researcher prior to collection of the oral health and OA parameters to minimize risk of bias. 2.3. Osteoarthritis Parameters The patient's medical notes were accessed to extract information regarding the disease duration, type of OA and prescribed medications. The prescription of any medications were noted, and the type of SYSADOAs and painkiller taken were specifically extracted from the hospital patient system. The participants' radiographs were read by a musculoskeletal radiologist (FMF) to determine the Kellgren-Lawrence score . Prior to reading the radiographs, FMF was calibrated to another independent musculoskeletal radiologist. The radiographs were each assigned a grade from 0 to 4. The grades correlated to the increasing severity of OA, with Grade 0 signifying no presence of OA and Grade 4 signifying severe OA. For participants suffering a combination of OA where more than one type of joint is affected, the Kellgren-Lawrence score was taken from the joint that was most severely affected. Examiner FMF was blinded to the clinical findings of the participants. The extracted information was checked to ensure that it was dated no longer than three months in duration from the clinical examination. 2.4. Statistical Analysis Statistical analysis of variables was performed with IBM SPSS version 19.0 (IBM Co., Armonk, NY, USA). Univariate comparisons were made using the Chi-squared test. The correlations between periodontal indices and OA disease characteristics were analyzed by Pearson or Spearman correlation coefficients to test for the association between the variables. The Mann-Whitney U test was applied for analysis of independent nonparametric variables. All p values are two-sided, and p values less than 0.05 were considered statistically significant. 3. Results 3.1. Demographic Data The flowchart of the steps of subject recruitment and data collection is reported in Figure 1. The demographic characteristics and periodontitis diagnosis of the study participants are shown in Table 1. The participants were mostly older adults with a mean age of 61.5 (+-9.3). Out of the 130 participants with OA, 98 (75.4%) had knee OA, 7 (5.4%) had hip OA and 6 (4.6%) had hand OA, while 19 (14.6%) had a combination type of OA. 3.2. Proportion of Osteoarthritis Participants with Periodontitis From the 130 participants recruited, 71 (54.6%) had periodontitis. Generally, the OA participants had a higher proportion of periodontitis compared to the Malaysian NOHSA findings of 48.5% . The severity of periodontitis suffered by the OA participants was also greater, with more participants having a CPI of 4 compared to the NOHSA findings. This is shown in Figure 2. 3.3. Osteoarthritis Parameters with Periodontal Health Parameters The mean +- standard deviation for the OA parameters measured are as follows: OA disease duration: 6.08 +- 6.40; Kellgren-Lawrence Score: 2.95 +- 1.01; HAQ Score: 0.40 +- 0.43; Pain Score: 5.04 +- 2.96; and Global Health Score 7.37 +- 1.80. The distribution of the PD parameters and their correlation with the OA parameters are shown in Table 2. There was a correlation using the Spearman Rho test between the Kellgren-Lawrence score of the participants with teeth count, where participants with a greater Kellgren-Lawrence score had less teeth (rs = 0.204, p = 0.025, two-tailed, N = 121). There was no other correlation between any parameters of PD (plaque index, gingival index, probing pocket depth and clinical attachment loss) with the Kellgren-Lawrence score. The overall HAQ score was correlated with the teeth count (rs = -0.181, p = 0.039), average PPD (rs = 0.209, p = 0.017) and CAL (rs = 0.239, p = 0.006). Participants with a greater degree of functional limitation, as shown by the HAQ score, had worst periodontal health, with less teeth, a higher PPD and a higher CAL. The global health score was inversely correlated with the plaque index (rs = -0.178, p = 0.043), where participants with a higher global health score had less plaque. The specific sections of the HAQ were further investigated against the oral health parameters to ascertain which type of functional limitations was correlated with which oral health parameter. Out of the eight HAQ sections, only 'Arising', 'Eating' and 'Walking' were correlated with specific oral health parameters. Participants reporting a limitation in 'Arising' and 'Eating' had deeper CAL (rs = 0.206, p = 0.019; rs = 0.211 p = 0.016) and PPD (rs = 0.245, p = 0.005; rs = 0.278 p = 0.001). Both CAL and PPD are parameters related to periodontal health, indicating that participants with a limitation in 'Arising' and 'Eating' had worst periodontal health. A limitation in 'Walking' was correlated with all 5 oral health parameters investigated, namely, teeth count (rs = -0.233, p = 0.008), plaque index (rs = 0.195, p = 0.026), gingival index (rs = 0.177, p = 0.045), average PPD (rs = 0.209, p = 0.017) and average CAL (rs = 0.265, p = 0.002). All five oral health parameters are measures of the overall oral health, indicating that participants with walking limitations have worst oral and periodontal health. 3.4. Medications and Periodontal Health Parameters Table 3 shows the type of SYSADOAs taken by the participants. Most of the participants were not taking any SYSADOAs (53.8%), while those who were taking SYSADOAs were mostly prescribed either Glucosamine (13.1%) or Piascledine (27.7%). There was no difference in the periodontal health parameters between participants taking different SYSADOA medications and when compared to those not on any medications. There was also no difference in the oral health parameters between the participants taking NSAIDs and those who were not. 4. Discussion Osteoarthritis is a complex condition affecting the whole joint, where an association with systemic inflammation could also be present . OA is now viewed as a complex biological response connecting biomechanics, inflammation and the immune system, rather than a purely mechanical disease due to the wear and tear of the cartilage in the joints . The prevalence of periodontitis among patients with OA in this study is higher at 54.6% when compared to the national average of 48.5% . Other reports on the prevalence of periodontitis in OA participants reported a relatively lower prevalence of 39.4% , 26% , 26.4% and 28% . This increased prevalence can be attributed to the limitations in function, as well as the increase in the overall systemic inflammation experienced by participants with OA. Other factors include the poorer general health of the subjects, as well as the older age of the participants. The results showed that participants who reported functional limitations in the HAQ had a more severe form of periodontitis with less teeth, a higher PPD and a greater CAL. Participants with a limitation in 'Arising' and 'Walking' had the worst periodontal health. This can be attributed to the mobility limitations that affect the OA participants' ability to access the dental clinic for routine dental care . Patients with arthritis can be limited in their access to dental services and may also have transportation restrictions, which are factors affecting the use of dental services by this group of patients . In view of the increased severity of periodontal disease in this group of patients, dental services should be made more accessible, such as by offering home visits and having the location of the dental clinics be more accessible, such as on the ground floor or with elevators or ramps to better facilitate the dental visits of these OA patients. Dental healthcare personnel can also visit musculoskeletal clinics to provide care at regular intervals. The correlation between a limitation in 'Eating' and the severity of periodontitis, as measured by CAL and PPD, suggests that OA patients with nutrition issues may be more predisposed to PD. Bone formation, healing of the periodontium and periodontal regeneration are affected by various vitamins, minerals and trace elements, such as vitamin C, vitamin D, magnesium and calcium . It is possible that the limitation in eating in this population contributes to the increased prevalence and severity of PD in patients with OA . The assessment of diet during dental visits can allow the clinician to better address the factors related to the periodontal disease among patients with OA. Referral to a dietitian can also be a strategy for the improvement of both the oral health and the general health of this group of patients. Restrictions in the ability to grip objects can lead to poorer oral hygiene, worsening of the periodontal health clinical parameters and poorer oral health. However, there was no correlation between limitations in 'Grip' with any oral health parameters in this study. This can be attributed to the relatively low number of subjects with hand OA, which was 14.6%. The tool used to assess functional limitations in this study, which was the HAQ, were also not designed to specifically assess hand function. This can explain the lack of correlation between 'Grip' limitations and the oral health parameters. This study found an association between the severity of OA and the number of teeth in the mouth, where participants with a more severe form of OA had less teeth. The association can be an indicator that the two diseases may be related. However, it is also possible that the association is due to the participant's age, where the mean age of the participants in this study is 61.5 years. With increasing age, participants will tend to have a greater severity of OA and will also have less teeth, as the estimate of tooth loss is 0.2 teeth per year . Tooth loss can lead to the reduced nutritional value of the food intake, as well as a reduction in social participation that can eventually lead to social isolation and depression. Preventive measures to limit tooth loss, as well as the provision of a prothesis to replace teeth, can be invaluable for these patients. Medications taken for OA termed SYSADOAs have been reported to have beneficial effects on the periodontium. Diacerein has anti-inflammatory activities that affects the cytokine profiles , while ASU cultured with gingival fibroblasts showed potential to prevent the deleterious effects of IL-1ss in periodontitis , and supplements containing chondroitin and glucosamine sulphate have been reported to speed up the regenerative capacity and stability of the periodontium . However, there was no significant difference of the periodontal health parameters between the participants taking the different medications for OA in this study. This can be due to the cross-sectional nature of this study, where the longitudinal effects of these medications cannot be accurately detected, as well as the relatively small numbers of participants taking the different types of SYSADOAs. In this group of participants, not all of the SYSADOAs are available in the clinics, and some of them had to be purchased privately. This may impact the type of medication taken for the OA, depending on the socio-economic background of the participants. The limitations of this study include the methods of measurement for the functional disability. This study utilized a self-reported assessment of function and disability as assessed by the HAQ. Other methods of assessment may be questionnaires that specifically assess the hand function, such as the Functional Index for Hand Arthropathies (FIHOA), or a performance-based test, such as the Arthritis Hand Function Test (AHFT). A more objective measurement of hand function will give a better picture of the ability of the subjects to perform regular oral hygiene care. The older age of the participants in this study may also be a confounding factor affecting the periodontal health, as well as the functional limitations, of the participants. Other limitations are that factors such as the duration, compliance and dose of the medications taken were not analyzed. These may play a role in the relationship between the medications and PD and should ideally be analyzed in the future. This study is one of very few studies reporting on the proportion of OA subjects with periodontitis. The findings of this study can be useful for dental practitioners and policymakers to have a better understanding of the factors and clinical parameters related to the oral and periodontal health of OA patients. It can also be used to design a more rigorous study investigating factors linking OA and periodontitis in the form of cohort studies or randomized controlled trials. 5. Conclusions There is a high proportion of periodontitis among patients with musculoskeletal disease, namely, OA. Functional disability among OA patients was associated with clinical measures of periodontal health. Within the limitations of this study, SYSADOAs taken by the participants were not associated with the diagnosis, stage, grade and periodontal health parameters in OA participants. It is suggested that clinicians treating OA patients consider the need for a referral for dental care when managing this group of patients. Author Contributions Conceptualization, N.-M.N.-A., N.M., B.B. and M.S.M.S.; methodology, N.-M.N.-A., N.H.M.H. and F.M.F.; software, N.-M.N.-A. and F.M.F.; validation, N.-M.N.-A., N.H.M.H., F.M.F. and N.M.; formal analysis, N.-M.N.-A.; resources, N.-M.N.-A., N.M., N.H.M.H. and M.S.M.S.; data curation, N.-M.N.-A., F.M.F. and M.S.M.S.; writing--original draft preparation, N.-M.N.-A.; writing--review and editing, B.B., M.S.M.S. and N.M.; visualization, N.-M.N.-A., F.M.F. and B.B.; supervision, N.M., B.B., N.H.M.H. and M.S.M.S.; project administration, N.M., B.B. and M.S.M.S.; funding acquisition, N.M. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Universiti Kebangsaan Malaysia (UKM/PPI/111/8/JEP-2017-553). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement Data available on request due to restrictions. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Flowchart of the recruitment and the data collection process, including the number of participants for every step. Figure 2 Community Periodontal Index (CPI) of osteoarthritis subjects compared to the data taken from the National Oral Health Survey, 2010. healthcare-11-00770-t001_Table 1 Table 1 Sociodemographic characteristics and the periodontal diagnosis of the study participants. Sociodemographic Characteristics (N = 130) Gender (n [%]) Female 106 (81.54) Male 24 (18.44) Race (n [%]) Malay 74 (56.92) Chinese 40 (30.77) Indian 14 (10.77) Other 2 (1.54) Age (mean +- SD) Age 61.5 (+- 9.3) Education (n [%]) Primary 48 (36.92) Secondary 53 (40.77) Tertiary 29 (22.31) BMI (mean +- SD) Kg/m2 28.4 (+- 5.1) Smoking (n [%]) No 121 (93.08) Past 4 (3.08) Current 5 (3.85) Diabetes (n [%]) Yes 44 (33.85) No 86 (66.15) Periodontal Diagnosis, Stage and Grade Diagnosis (n = 130) n (%) Healthy 34 (26.2) Gingivitis 25 (19.2) Periodontitis 71 (54.6) Stage of Periodontitis (n = 71) Stage 1 17 (13.1) Stage 2 19 (14.6) Stage 3 21 (16.2) Stage 4 14 (10.8) Grade of Periodontitis (n = 71) Grade A 6 (8.5) Grade B 58 (81.7) Grade C 7 (9.9) healthcare-11-00770-t002_Table 2 Table 2 PD parameters and their correlation with the OA parameters. Teeth Count Plaque Index Gingival Index Average PPD Average CAL mean +- SD 18.93 +- 7.63 45.45 +- 32.12 40.27 +- 31.90 12.56 +- 5.86 1.34 +- 1.39 Correlation between PD and OA parameters OA Disease Duration; rs (p-value) -0.173 (0.048 *) 0.119 (0.177) 0.124 (0.162) 0.060 (0.495) 0.039 (0.660) Kellgren-Lawrence Score; rs (p-value) -0.204 (0.025 *) 0.063 (0.492) 0.071 (0.438) 0.006 (0.951) 0.066 (0.471) HAQ; rs (p value) -0.181 (0.039 *) 0.143 (0.105) 0.153 (0.083) 0.209 (0.017 *) 0.239 (0.006 *) Pain Score; rs (p value) 0.041 (0.643) 0.047 (0.598) -0.074 (0.408) -0.032 (0.714) -0.047 (0.597) Global Health Score; rs (p value) 0.125 (0.158) -0.178 (0.043 *) 0.143 (0.105) -0.150 (0.088) -0.164 (0.062) * p-value < 0.05; PPD--periodontal probing depth; CAL--clinical attachment loss; HAQ--health assessment questionnaire. healthcare-11-00770-t003_Table 3 Table 3 The type of Symptomatic Slow Acting Drugs for Osteoarthritis (SYSADOAs) prescribed and the Periodontal Health parameters according to the SYSADOAs prescribed. Medication Type Glucosamine Diascerine Piascledine Combination No p-Value N (%) 17 (13.1) 3 (2.3) 36 (27.7) 4 (3.1) 70 (53.8) Periodontal Diagnosis Healthy; n (%) 5 (29.4) 1 (33.3) 12 (33.3) 1 (25.0) 34 (26.2) 0.667 Gingivitis; n (%) 4 (23.5) 0 5 (13.9) 2 (50.0) 25 (19.2) Periodontitis; n (%) 8 (47.1) 2 (66.7) 19 (52.8) 1 (25.0) 71 (54.6) Oral Health Parameters Teeth Count; (mean +- SD) 19.35 +- 6.89 19.33 +- 12.50 18.36 +- 8.03 26.50 +- 1.29 18.67 +- 7.54 0.373 Plaque Index; (mean +- SD) 40.88 +- 31.24 53.33 +- 45.17 44.17 +- 33.1 40.00 +- 14.14 47.19 +- 32.65 0.927 Gingival Index; (mean +- SD) 35.94 +- 33.28 51.67 +- 47.52 37.78 +- 34.44 36.25 +- 12.50 42.29 +- 30.89 0.875 Average PPD; (mean +- SD) 1.98 +- 1.25 2.84 +- 2.10 1.88 +- 1.09 1.50 +- 0.42 2.04 +- 1.29 0.660 Average CAL; (mean +- SD) 2.98 +- 2.33 4.51 +- 4.18 3.38 +- 2.13 3.38 +- 2.13 3.37 +- 2.38 0.871 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Yunus M.H.M. Nordin A. Kamal H. Pathophysiological Perspective of Osteoarthritis Medicina 2020 56 614 10.3390/medicina56110614 33207632 2. Kraus V.B. Blanco F.J. Englund M. Karsdal M.A. Lohmander L.S. Call for standardized definitions of osteoarthritis and risk stratification for clinical trials and clinical use Osteoarthr. Cartil. 2015 23 1233 1241 10.1016/j.joca.2015.03.036 25865392 3. Chow Y.Y. Chin K.-Y. The Role of Inflammation in the Pathogenesis of Osteoarthritis Mediat. Inflamm. 2020 2020 8293921 10.1155/2020/8293921 32189997 4. Martel-Pelletier J. Pathophysiology of osteoarthritis Osteoarthr. Cartil. 2004 12 31 33 10.1016/j.joca.2003.10.002 5. Woodell-May J.E. Sommerfeld S.D. Role of Inflammation and the Immune System in the Progression of Osteoarthritis J. Orthop. Res. 2020 38 253 257 10.1002/jor.24457 31469192 6. Tonetti M.S. Greenwell H. Kornman K.S. Staging and grading of periodontitis: Framework and proposal of a new classification and case definition J. Periodontol. 2018 89 159 172 10.1002/JPER.18-0006 7. Hajishengallis G. Chavakis T. Local and systemic mechanisms linking periodontal disease and inflammatory comorbidities Nat. Rev. Immunol. 2021 21 426 440 10.1038/s41577-020-00488-6 33510490 8. Di Spirito F. Iacono V.J. Alfredo I. Alessandra A. Sbordone L. Lanza A. Evidence based recommendations on periodontal practice and the management of periodontal patients during and after the COVID-19 era: Challenging infectious diseases spread by airborne transmission Open Dent. J. 2021 15 325 336 10.2174/1874210602115010325 9. Chung M.K. Koo N. Lee B.W. Lee J. Kwok S.-K. Park S.-H. Ju J. SAT0533 Association of osteoarthritis and periodontitis based on the Korea national health and nutrition examination survey Ann. Rheum. Dis. 2017 76 (Suppl. S2) 977 978 10. Kim J. Chung M.K. Lee J. Kwok S. Association of Periodontitis with Radiographic Knee Osteoarthritis J. Periodontol. 2019 91 369 376 10.1002/JPER.19-0068 31389022 11. Mikuls T.R. Payne J.B. Yu F. Thiele G.M. Reynolds R.J. Cannon G.W. Markt J. McGowan D. Kerr G.S. Redman R.S. Periodontitis and Porphyromonas gingivalis in Patients with Rheumatoid Arthritis Arthritis Rheumatol. 2014 66 1090 1100 10.1002/art.38348 24782175 12. Nguyen V.B. Nguyen T.T. Huynh N.C. Le T.A. Relationship between periodontitis and rheumatoid arthritis in Vietnamese patients Acta Odontol. Scand. 2020 78 522 528 10.1080/00016357.2020.1747635 32238080 13. Dominick K.L. Jordan J.M. Renner J.B. Kraus V.B. Relationship of Radiographic and Clinical Variables to Pinch and Grip Strength Among Individuals with Osteoarthritis Arthritis Rheum. 2005 52 1424 1430 10.1002/art.21035 15880347 14. Kelsey J.L. Lamster I.B. Influence of musculoskeletal conditions on oral health among older adults Am. J. Public Health 2008 98 1177 1183 10.2105/AJPH.2007.129429 18511715 15. Dolan T.A. Peek C.W. Stuck A.E. Beck J.C. Functional health and dental service use among older adults J. Gerontol. A Biol. Sci. Med. Sci. 1998 53 M413 M418 10.1093/gerona/53A.6.M413 9823744 16. Pokrajac-Zirojevic V. Slack-Smith L.M. Booth D. Arthritis and use of dental services: A population-based study Aust. Dent. J. 2002 47 208 213 10.1111/j.1834-7819.2002.tb00330.x 12405459 17. Hughes F.J. Bartold P.M. Periodontal complications of prescription and recreational drugs Periodontol. 2000 2018 78 47 58 10.1111/prd.12230 30198137 18. Salvi G.E. Lang N.P. The Effects of Non-Steroidal Anti-Inflammatory Drugs (Selective and Non-Selective) on the Treatment of Periodontal Diseases Curr. Pharm. Des. 2005 11 1757 1769 10.2174/1381612053764878 15892673 19. Huang R.Y. Lu S.H. Su K.W. Chen J.K. Fang W.H. Liao W.N. Chen S.Y. Shieh Y.S. Diacerein: A potential therapeutic drug for periodontal disease Med. Hypotheses 2012 79 165 167 10.1016/j.mehy.2012.04.024 22583561 20. Zaki B.M. Mahmoud E.A. Aly A.A. Diacerein: A potential therapeutic drug for the management of experimental periodontitis in rats Electron. Physician 2015 7 1290 1295 10.14661/1290 26435830 21. Kut-Lasserre C. Miller C.C. Ejeil A.L. Gogly B. Dridi M. Piccardi N. Guillou B. Pellat B. Godeau G. Effect of Avocado and Soybean Unsaponifiables on Gelatinase A (MMP-2), Stromelysin 1 (MMP-3), and Tissue Inhibitors of Matrix Metalloproteinase (TIMP-1 and TIMP-2) Secretion by Human Fibroblasts in Culture J. Periodontol. 2001 72 1685 1694 10.1902/jop.2001.72.12.1685 11811504 22. Oliveira G. Paula L. Souza J. Spin-Neto R. Stavropoulos A. Marcantonio R. Effects of avocado/soybean unsaponifiables (ASU) on the treatment of ligature-induced periodontitis in rats Braz. Oral Res. 2017 31 e28 10.1590/1807-3107bor-2017.vol31.0028 28403331 23. Paknejad M. Khorsand A. Yaghobee S. Nasiri M. The Effect of Piascledine in the Treatment of Chronic Periodontitis J. Dent. 2007 4 135 138 24. Al-Saadi H.M. Pang K.-L. Ima-Nirwana S. Chin K.-Y. Multifaceted Protective Role of Glucosamine against Osteoarthritis: Review of Its Molecular Mechanisms Sci. Pharm. 2019 87 34 10.3390/scipharm87040034 25. Anyanechi C. Chukwuneke F. Ngim N. Clinical Evaluation of the Efficacy of Arthocare Forte, A Chondro Protective and Anti Arthritic Drug in the Management of Bacterial Plaque Induced Chronic Periodontitis Ann. Med. Health Sci. Res. 2015 5 157 162 10.4103/2141-9248.157481 26097755 26. Nik-Azis N.-M. Mohd N. Mohd Fadzilah F. Mohamed Haflah N.H. Mohamed Said M.S. Baharin B. Rheumatoid arthritis serotype and synthetic disease-modifying anti-rheumatic drugs in patients with periodontitis: A case-control study PLoS ONE 2021 16 e0252859 10.1371/journal.pone.0252859 34153036 27. O'Leary J.O. Drake B. Naylor J.E. The Plaque Control Record J. Periodontol. 1972 43 38 10.1902/jop.1972.43.1.38 4500182 28. Ainamo J. Bay I. Problems and proposals for recording gingivitis and plaque Int. Dent. J. 1975 25 229 235 1058834 29. Papapanou P.N. Sanz M. Buduneli M. Dietrich T. Feres M. Periodontitis: Consensus report of workgroup 2 of the 2017 World Workshop on the Classification of Periodontal and Peri-Implant J. Periodontol. 2018 89 173 182 10.1002/JPER.17-0721 30. Kirch W. CPI (Community Periodontal Index) Encyclopedia of Public Health Springer Dordrecht, The Netherlands 2008 176 177 10.1007/978-1-4020-5614-7_604 31. National Oral Health Survey of Adults 2010 (NOHSA 2010) Oral Health Division, Ministry of Health Malaysia Putrajaya, Malaysia 2013 32. Hussein H. Mustafa R. Quek K.F. Hassanudin N.S. Shahid S. Cross-cultural adaptation and validation of a Malay Health Assessment Questionnaire for use in rheumatoid arthritis patients Int. J. Rheum. Dis. 2008 11 237 240 10.1111/j.1756-185X.2008.00384.x 33. Kellgren J.H. Lawrence J.S. Radiological Assessment of Osteo-arthrosis Ann. Rheum. Dis. 1957 16 494 502 10.1136/ard.16.4.494 13498604 34. Mohd-Dom T.N. Abdul-Muttalib K. Ayob R. Muttalib K.A. Asari A.S. Periodontal status and provision of periodontal services in Malaysia: Trends and way forward Malays. J. Public Health Med. 2013 13 38 47 10.1186/1472-6963-12-S1-P8 35. Glyn-Jones S. Palmer A.J.R. Agricola R. Price A.J. Vincent T.L. Weinans H. Carr A.J. Osteoarthritis Lancet 2015 386 376 387 10.1016/S0140-6736(14)60802-3 25748615 36. Dissick A. Redman R.S. Jones M. Rangan B.V. Reimold A. Griffiths G.R. Mikuls T.R. Amdur R.L. Richards J.S. Kerr G.S. Association of Periodontitis with Rheumatoid Arthritis: A Pilot Study J. Periodontol. 2010 81 223 230 10.1902/jop.2009.090309 20151800 37. Raindi D. Nutrition and Periodontal Disease Dent. Update 2016 43 66 72 10.12968/denu.2016.43.1.66 27024903 38. Najeeb S. Zafar M.S. Khurshid Z. Zohaib S. Almas K. The Role of Nutrition in Periodontal Health: An Update Nutrients 2016 8 530 10.3390/nu8090530 27589794 39. Nik-Azis N.M. Mohd-Fadzilah F. Mohd-Shahrir M.S. Baharin B. Mohd N. Impact of rheumatoid arthritis functional status on oral and periodontal health in a multi-ethnic population Sains Malays. 2019 48 645 652 10.17576/jsm-2019-4803-18 40. Needleman I. Garcia R. Gkranias N. Kirkwood K.L. Kocher T. Di Iorio A. Moreno F. Petrie A. Mean annual attachment, bone level, and tooth loss: A systematic review J. Periodont. 2018 89 120 139 10.1002/JPER.17-0062 29926956 |
PMC10000595 | Angiogenesis is the formation of new blood capillaries taking place from preexisting functional vessels, a process that allows cells to cope with shortage of nutrients and low oxygen availability. Angiogenesis may be activated in several pathological diseases, from tumor growth and metastases formation to ischemic and inflammatory diseases. New insights into the mechanisms that regulate angiogenesis have been discovered in the last years, leading to the discovery of new therapeutic opportunities. However, in the case of cancer, their success may be limited by the occurrence of drug resistance, meaning that the road to optimize such treatments is still long. Homeodomain-interacting protein kinase 2 (HIPK2), a multifaceted protein that regulates different molecular pathways, is involved in the negative regulation of cancer growth, and may be considered a "bona fide" oncosuppressor molecule. In this review, we will discuss the emerging link between HIPK2 and angiogenesis and how the control of angiogenesis by HIPK2 impinges in the pathogenesis of several diseases, including cancer. VEGF HIF-1 hypoxia micro-RNA circular HIPK2 p53 cancer diabetes diabetic retinopathy wound healing Italian Association for Cancer ResearchAIRC The research in G.D. lab was funded by the Italian Association for Cancer Research (AIRC, number IG11377), by the Italian Ministry of Research (Miur-PRIN, n. 2005059700_003) and by G. D'Annunzio University Grants (Fondi Ateneo 2011). pmc1. Introduction Angiogenesis is the formation of new blood capillaries taking place from preexisting functional vessels. In the adult, a physiologic vessels formation is transiently activated for tissue growth and regeneration during processes such as wound healing and the female reproductive cycle. However, angiogenesis may also have a pathologic role as it fuels inflammatory and malignant diseases . Deregulation of the normal vessels' growth is observed in many diseases including diabetic retinopathy, autoimmune diseases, rheumatoid arthritis, atherosclerosis, cerebral ischemia, cardiovascular diseases, psoriasis, and delayed wound healing . In many solid cancers, angiogenesis is constantly activated by the "angiogenic switch" that causes normally quiescent vasculature to continually sprouts new vessels. In this way, angiogenesis helps to sustain expanding neoplastic growth. For that reason, angiogenesis is considered a hallmark of cancer progression . The interaction between neoplastic cells by means of the angiogenic factors produced by them, and the newly formed vessels promotes the growth of solid tumors and the metastases formation, as well as the impairment of the efficacy of anticancer therapies . Interestingly, other than the tumor cells, there are also stromal cells such as tumor-associated macrophages (TAM) which can produce the angiogenic factors that promote angiogenesis and metastasis . Many reviews have extensively summarized the steps through which the vascular bed expands by sprouting and matures into a system of stable vessels in normal and pathological conditions; therefore, here the main molecules regulating angiogenesis will be only briefly described. Angiogenesis is regulated by the balance of many positive and negative factors released into the microenvironment . The positive regulators include vascular endothelial growth factors (VEGFs); A, B, and C fibroblasts growth factors (FGFs); 1 and 2 platelet-derived growth factor (PDGF); hepatocyte growth factor (HGF); and angiopoietins, while the negative regulators include angiostatin, endostatin, thrombospondin, and interferons . However, the most important mediator of angiogenesis is VEGFA, which acts on endothelial cells by binding two different receptors (R), namely VEGFR-1 and VEGFR-2 . The binding of VEGF to its receptor activates the PI3K/Akt, MEK, or FAK signaling pathways leading to the expression of genes whose proteins induce vascular permeability, cell proliferation, and motility, thus promoting angiogenesis . Since its discovery, VEGF has revolutionized the comprehension of the angiogenesis process in normal tissue development and in health conditions, as well as in the course of many diseases . The targeting of VEGF is therefore a therapeutic approach of high interest and, so far, hundreds of thousands of patients have been treated with blockers of VEGF even if the limited therapeutic efficacy due to the activation of resistance mechanisms remains an outstanding problem . A key driver of angiogenesis is the hypoxia-inducible factor-1 (HIF-1), a heterodimeric transcription factor that consists of two subunits: the oxygen-sensitive subunit HIF-1a (or its analogs HIF-2a and HIF-3a) that undergoes quick degradation under normoxic conditions and the constitutively expressed HIF-1b, also known as aryl hydrocarbon nuclear translocator (ARNT) . Under hypoxia, HIF-1a is stabilized via several post-translational modifications involving hydroxylation, acetylation, and phosphorylation. Following activation, HIF-1a translocates into the nucleus to bind HIF-1b and induce the transcription of several target genes involved in many aspects of cancer progression including angiogenesis (e.g., VEGF, PDGFB), metabolic adaptation (e.g., GLUT1, PDK1), apoptosis resistance (e.g., Bcl-2, MDR), invasion, and metastasis (e.g., CXCR4, MMP9) . When a tumor mass grows beyond 1-2 mm, it undergoes hypoxia because of the distance from the host microvasculature, which makes it difficult to efficiently supply the tumor with nutrients and oxygen. In order to survive to the hypoxia, tumors activate HIF-1 . Other than hypoxia, many genetic alterations inactivating tumor suppressors or activating oncoproteins have been reported to increase the basal levels of HIF-1a in cancers and contribute to tumor progression and angiogenesis . 2. HIPK2 and Tumor Angiogenesis Homeodomain-interacting protein kinase-2 (HIPK2) is a serine/threonine kinase which belongs to a family that includes four members (HIPK1, HIPK2, HIPK3, and HIPK4) of corepressors for homeodomain transcription factors whose structures and functions have been extensively summarized (for a review, see references ). HIPK2 modulates the activity of many transcriptional regulators and chromatin modifiers and, depending on the cell context, it can repress or promote the gene transcription . HIPK2 regulates the expression of several genes involved in cell development, cytokinesis, protein stability, apoptosis, and DNA damage response . One of the most important targets of HIPK2 is the oncosuppressor p53 that is phosphorylated by HIPK2 in Serine 46 to specifically activate the p53 apoptotic function essential for the success of the anticancer therapies . HIPK2 may also regulate p53-independent pathways and, for this reason, HIPK2 dysregulation is associated with neurological diseases and fibrosis other than with cancer progression . Recently, a role for HIPK2 in angiogenesis has been pointed out, not only in cancer, but also in other angiogenic diseases, and will be summarized below. 2.1. HIPK2 and HIF-1/VEGF in Tumor Angiogenesis Our previous studies showed that HIPK2 binds, along with histone deacetylase 1 (HDAC1), to the HIF-1a gene promoter repressing the HIF-1-mediated transcription of many target genes including VEGF, therefore restraining tumor growth . As a proof of principle, HIPK2 silencing with small interfering RNA upregulated HIF-1a in cancer cells, inducing a pseudohypoxic tumor phenotype in normoxic conditions, fueling tumor progression, and chemoresistance . To examine whether the effect of HIPK2 on the modulation of HIF-1a/VEGF pathway was associated with endothelial cell sprouting, the growth of human umbilical vein endothelial cells (HUVEC) was evaluated in vitro in the presence of conditioned media (CM) derived from colon cancer cells depleted or not of the HIPK2 function, and the results confirmed that HIPK2 silencing increases tumor vascularity in vitro . Interestingly, hypoxia-driven mechanisms lead to HIPK2 protein degradation . Therefore, a regulatory loop exists between HIPK2 and HIF-1a that affects the multiple downstream molecular pathways, including p53 and VEGF, regulated by both proteins , impinging on tumor growth and angiogenesis and/or on tumor regression . Thus, HIPK2 silencing increases the xenograft tumor growth and the physiologic relevance was assessed by analyzing the HIPK2 gene expression in human specimens collected from patients with the familial adenomatous polyposis (FAP) and with sporadic colorectal cancer (CRC). HIPK2 mRNA levels were lower in sporadic CRC tissues compared to FAP tissues and the HIPK2 expression in human CRC inversely correlate with the staging of the tumors , although the molecular mechanisms leading to HIPK2 mRNA downregulation were not unveiled. In the attempt to target hypoxia and restrain tumor angiogenesis, we have shown that zinc chloride induces HIF-1a protein degradation and inhibits the HIF-1-induced transcription of VEGF and angiogenesis, in vitro and in animal studies . In addition, zinc counteracts the hypoxia-induced HIPK2 deregulation restoring p53 apoptotic response to chemotherapy, underscoring the potential use of zinc supplementation in combination with chemotherapy to improve the efficacy of the anticancer treatments . The balance between HIPK2 and HIF-1 in angiogenesis was recently confirmed in a study on hepatocellular carcinomas (HCC) . In two independent patients' cohorts with, respectively, 90 and 52 paired HCC and adjacent normal tissues, the authors analyzed using immunohistochemistry (IHC) of the expression levels of HIPK2 protein and found that they were lower in the cancer tissues compared to the adjacent normal tissues . Studies performed in animal models showed that HIPK2 overexpression reduces tumor xenografts growth and metastasis formation. IHC analysis of the xenograft tumor tissues derived from Huh7 or BEL-7404 cancer cells, with or without HIPK2 overexpression, showed that HIPK2 downregulation significantly increases VEGFa level in the subcutaneous tumor and in the corresponding new blood vessels . Following this, in vitro studies evaluated the tube formation of HUVEC cultured with the conditioned medium (CM) of HCC cells with or without HIPK2 overexpression in hypoxic condition. The tube formation was reduced when HUVEC were cultured with the medium derived from cells with HIPK2 overexpression compared with the medium derived from the control cells , strengthening the finding that HIPK2 may inhibit hypoxia-induced angiogenesis in the HCC tumor, as observed in the above reported study on CRC . Mechanistically, the authors found that HIPK2 directly binds to and downregulates HIF-1a protein by inducing its proteasomal degradation as demonstrated by the use of the proteasome inhibitor MG132 . To further evaluate the antiangiogenic role of HIPK2 in HCC samples, the authors analyzed data retrieved from the Gene Expression Omnibus (GEO) database. They found lower HIPK2 expression in both metastasis tissues and the primary lesions with metastasis compared to the primary lesions without metastasis. After this, by knocking out with the CRISPR-Cas9 system the HIF-1a and HIPK2 genes individually or simultaneously, the authors determined the in vivo tumor growth capacities of Hepa1-6 cells in a xenograft mice model. They observed enhanced in vivo tumor growth of the HIPK2-/- cells while the tumor cells with HIF-1a knockout grew significantly slower compared to the control tumors . The double HIF-1a and HIPK2 knockout greatly counteracted the tumor growth caused by the HIPK2 knockout, suggesting that the effect of HIPK2 depletion on HCC progression was mediated by HIF-1a and by the HIF-1-induced angiogenesis , further strengthening the effect of the HIPK2/HIF-1a balance. Consistently, HIPK2 overexpression reduced the hypoxia-induced angiogenesis in vitro as well as the brain and bone metastasis of the highly metastatic HCC cell line CSQT-2 in a mouse model , thus connecting angiogenesis with metastasis . The above reported findings suggest that the lower expression of HIPK2 in cancer tissues, compared to the normal ones, could serve as a novel biomarker of HCC progression due to the HIF-1-induced angiogenesis, although the mechanisms leading to HIPK2 downregulation (e.g., hypoxia or microRNAs) in HCC have not been elucidated and might deserve further studies. The findings also confirm the key role of angiogenesis in the HCC progression and metastasis and highlight how its targeting might represent an efficacious strategy in the clinical treatment of HCC . Antiangiogenic drugs such as sorafenib and regorafenib or the multikinase inhibitors for VEGF receptors, PDGF receptors, and c-Kit, have been shown to be promising therapeutic agents against the HCC, although drug resistance may occur and contribute to the chemotherapeutic failure . Therefore, uncovering novel molecular mechanisms driving angiogenesis in HCC could provide novel potential therapeutical strategies. In this regard, it is tempting to hypothesize that combined therapies including zinc supplementation could, on one hand, inhibit the HIF-1-induced angiogenesis and, on the other hand, restore the HIPK2/p53 antitumor axis, as previously shown . 2.2. HIPK2 and microRNA in Tumor Angiogenesis Another mechanism that underscores the role of HIPK2 in tumor angiogenesis is the microRNA (miRNAs)-induced HIPK2 modulation . miRNAs are non-coding single strand RNAs of about 19-25 nucleotides which bind to the 3' untranslated region (3'UTR) of target mRNAs to inhibit the translation and induce degradation of the target mRNAs at the post-transcriptional level . miRNAs can be included into exosomes, a type of extracellular vesicles that are secreted by many cell types and that contain, other than miRNAs, all the main biomolecules including lipids, proteins, circulating tumor DNA (ctDNA), messenger RNAs, and oncoproteins . Tumor-derived exosomes (TEXs) perform intercellular transfer of components, locally and systemically, interacting with the surrounding cells in the tumor microenvironment. They are considered new players in tumor growth and invasion, tumor-associated angiogenesis, tissue inflammation, and immunologic remodeling . In this regard, it has been found that patients with colorectal cancer (CRC) show high levels of circulating exosomal (exo) miR-1229 which correlated with tumor size, lymphatic metastasis, angiogenesis, and poor survival . Mechanistically, exomiR-1229 targets the HIPK2 3'UTR. Thus, HIPK2 mRNA expression was found to be significantly downregulated in CRC tissues compared to the adjacent normal tissues , in agreement with the above-described study showing reduced HIPK2 mRNA expression in CRC tissues compared to the familial adenomatous polyposis (FAP) samples . Circulating exomiR-1229 reduced the HIPK2 protein levels in HUVECs leading to upregulation of the downstream VEGFA, VEGFR1, and p-Akt, thereby stimulating angiogenesis . Hu et al. showed that HIPK2 overexpression counteracts the exomiR-1229-induced upregulation of VEGFA, VEGFR1, and p-Akt, reducing both the extracellular release of VEGF and angiogenesis . The authors identified in the VEGF promoter a potential binding site for myocyte enhancer factor (MEF)-2C, a transcription factor that regulates sprouting angiogenesis directly downstream from VEGFA , and demonstrated that the MEF2-mediated activation of VEGF luciferase reporter may be suppressed by HIPK2 . In agreement, a previous study has shown that HIPK2 represses MEF2C-mediated transcriptional activation of VEGF and MMP10, regulating vascular morphogenesis . The authors concluded that the CRC-secreted exomiR-1229 can induce tumor angiogenesis by blocking the HIPK2-mediated suppression of VEGF expression. Hence, lower HIPK2 mRNA or protein levels in CRC tissues compared to the adjacent normal ones can be considered potential novel prognostic biomarkers of CRC progression. In addition, high levels of circulating exomiR-1229, associated with HIPK2 downregulation, could be also considered a potential prognostic biomarker in addition to being a potential therapeutic target for inhibiting tumor angiogenesis in CRC . Interestingly, exomiR-1229 was found to be upregulated in breast cancer and to trigger tumorigenesis by activating the Wnt/b-catenin pathway following targeting the key negative regulators of b-catenin such as glycogen synthase kinase (GSK)-3b, adenomatous polyposis coli (APC), and ICAT . The Wnt/b-catenin pathway is an evolutionarily conserved cellular signaling system involved in different biologic processes such as organogenesis, tissue homeostasis, as well as in the pathogenesis of many human diseases . The b-catenin transcription factor indeed induces the expression of several target genes involved in cell growth and angiogenesis including c-myc, cyclin D1, and VEGF . The b-catenin transcription factor is strongly involved in the early and stepwise events of the colon tumorigenesis and an aberrant activation of the Wnt/b-catenin signaling has been linked to the progression of many other cancer types . Interestingly, HIPK2 has been shown to phosphorylate and degrade b-catenin protein , therefore repressing the b-catenin-induced VEGF expression and tumorigenesis . Hence, it is tempting to speculate that high levels of exomiR-1229 might induce tumor angiogenesis not only by blocking the HIPK2-mediated suppression of VEGF expression but also by blocking the HIPK2-mediated inhibition of the b-catenin/VEGF pathway, although this latter hypothesis needs to be supported by further studies. Among the tumor-derived exosomes (TEXs), exomiR-1260b has been shown to target HIPK2 in HUVECs and promote angiogenesis, migration, invasion, and chemoresistance of non-small cell lung cancer (NSCLC) cells . Although the authors did not unveil the molecular mechanisms leading to the increased angiogenesis by exomiR-1260b-induced HIPK2 downregulation, they found a relationship between miR-1260b and HIPK2 and its clinical meaning. They found an inverse correlation between miR-1260b and HIPK2 by analyzing 124 paired NSCLC tissues and adjacent noncancerous lung tissues using quantitative Reverse Transcription (qRT)-PCR. The expression levels of HIPK2 transcripts were significantly lower in NSCLC tissues compared to the corresponding noncancerous lung tissues, while miR-1260b expression was higher in NSCLC tissues compared to the noncancerous lung tissues . Interestingly, HIPK2 downregulation and miR-1260b upregulation correlated with the presence of lymph node and distant metastasis, although the molecular mechanisms were not investigated. Further analyses showed that the level of exomiR-1260b was higher in the plasma of patients with NSCLC compared to that of healthy donors. In addition, Kaplan-Meier survival analysis showed that patients with high exomiR-1260b levels had worse overall survival rates than those with low exomiR-1260b levels . These findings suggest an inverse association between miR-1260b and HIPK2 and underline the new role of low HIPK2 levels as a prognostic indicator or predictor of metastasis in NSCLC. In addition, high levels of exomiR-1260b, associated with HIPK2 downregulation, could be considered a potential prognostic biomarker and a therapeutic target to inhibit NSCLC progression. Interestingly, it has been shown that exomiR-1260b promotes cell invasion through the Wnt/b-catenin signaling pathway in lung adenocarcinoma . Therefore, it can be hypothesized that exomiR-1260b-induced HIPK2 downregulation can consequently inhibit also the b-catenin signaling leading to angiogenesis and metastasis, as reported above for the exomiR-1229 . 2.3. HIPK2 and Circular RNA in Tumor Angiogenesis miRNAs can be regulated by circular RNAs (circRNAs), a large family of non-coding (nc) RNAs which are produced by "back splicing" of primary transcripts, and are more stable in vivo because they are protected from exonuclease degradation . Dysregulation of circRNAs is associated with the development of many diseases; hence, they are considered potentially useful biomarkers . In this regard, high expression of circHIPK2 was found in cisplatin (DDP)-resistant NSCLC cells and tissues . Bioinformatic analyses predicted that miR-1249-3p was the downstream target of circHIPK2, and the authors found that miR-1249-3p was indeed downregulated in NSCLC tissues and cells . The VEGFA expression positively correlated with circHIPK2 while negatively correlating with miR-1249-3p expression, as assessed by tumor xenograft studies. The authors showed that miR-1249-3p is a regulator of VEGFA expression and that VEGF was responsible of induction of angiogenesis and resistance to cisplatin. At the biological level, circHIPK2 silencing in lung cancer A549-DDP-resistant cells reduced their proliferation and inhibited the tube formation of HUVEC, leading to reduced tumor growth in vivo . They concluded that circHIPK2 has the malignant property to induce angiogenesis in NSCLC via miR-1249-3p/VEGF axis . High levels of circHIPK2 are starting to be found in a few other tumors and are being associated with increased tumor progression, although angiogenesis was not always analyzed in those studies . A high level of circHIPK2 has been found in CRC tissues compared to the adjacent normal tissue, and has been associated with lower overall and disease-free survival rate . CircHIPK2 has been found remarkably upregulated in nasopharyngeal carcinoma (NPC) tissues . In vitro and in vivo studies in animal models showed that circHIPK2 promotes proliferation of NPC cells while knockdown of circHIPK2 dampens the growth of NPC cells . Mechanistically, circHIPK2 downregulated HIPK2 at the protein levels and consequently increased the b-catenin protein expression. Hence, high levels of circHIPK2 have potential clinical significance in CRC and NPC progression; therefore, analysis of circHIPK2 may be worth of further studies also in other tumor types. The biological consequences of miRNAs-induced HIPK2 targeting and of the high circHIPK2 levels in tumors are summarized in Table 1. 3. HIPK2 and Other Angiogenic Diseases 3.1. HIPK2 and Angiogenesis in Gestational Complications and in Myocardial Infarction (MI) Gestational hypertension is the second leading cause of maternal death in developed countries . Angiogenesis plays a role in gestational hypertension through upregulation of angiopoietin-1 (ANG-1) or activation of the renin angiotensin system (RAS) that causes high circulating levels of angiotensin-II (ANG-II) . HIPK2 has been shown to play a role in angiogenesis of a model of gestational hypertension induced by hypoxia and reoxygenation (H/R). Human placental microvascular endothelial cells (HPMECs) undergoing H/R showed downregulation of miR-100-5p along with reduced concentrations of ANG-1 and ANG-2 and reduced VEGFA, TGF-b, and PLGF protein levels that correlated with reduced viability and angiogenesis of HPMECs . Rescue assays showed that miR-100-5p overexpression promoted HPMECs viability and angiogenesis restoring the levels of ANG-1, ANG-2, VEGFA, TGF-b, and PLGF inhibited by H/R . Interestingly, miR-100-5p overexpression significantly downregulated the expression levels of HIPK2 in HPMECs and, indeed, HIPK2 was found to targeted and negatively modulated by miR-110-5p . The authors showed that HIPK2 overexpression decreases the expression of VEGFA and TGF-b while increases the expression of anti-angiogenetic proteins (e.g., sFLT1 and sENG). Such overexpression reversed the effect induced by overexpression of miR-100-5p in terms of viability and angiogenesis in HPMECs exposed to the H/R. Mechanistically, miR-100-5p-induced HIPK2 downregulation led to the activation of the PI3K/AKT signaling pathway and such activation was reversed by HIPK2 overexpression . A link between HIPK2 and PI3K/AKT has been previously suggested. In that study, the authors found that the oncogene SPEN induces miR-4652-3p expression in nasopharyngeal carcinoma (NPC) by activation of the PI3K/AKT/c-JUN signaling and that miR-4652-3p targets and downregulates HIPK2 . The PI3K/AKT inhibitor LY294002 counteracted the increase in HPMEC viability and angiogenesis induced by miR-100-5p overexpression, an effect that was further strengthened by HIPK2 overexpression . Given the lack of effective therapies against pregnancy-induced hypertension, the discovery of new potential therapeutic targets such as the miR-100-5p could help to reduce the overall risk of cardiovascular, cerebrovascular, kidney diseases, and diabetes during pregnancy. Another complication that may occur during pregnancy is the increased risk of bronchopulmonary dysplasia (BPD) promoted by smoking . In a mouse model of gestational exposure to sidestream cigarette smoke (SS), the BPD-like condition correlated with impaired angiogenesis, suppression of VEGF, and increase in the alveolar cells' apoptosis . Mechanistically, gestational SS inhibited HIF-1a and increased pro-apoptotic factors including HIPK2 , in line with the role of HIF-1a in inducing the HIPK2 degradation, an interplay that is known to affect both apoptosis and angiogenesis . Among the miRNAs that regulate angiogenesis there is miR-126 that has been shown to directly repress the negative regulators of the VEGF pathway, including the Sprouty-related protein SPRED1 and the phosphoinositol-3 kinase regulatory subunit 2 (PIK3R2) . More recently, miR-126-5p was investigated in a model of myocardial infarction (MI), the most common cardiovascular disease in which hypoxia induces endothelial injury . The authors found that miR-126-5p was upregulated in hypoxia-treated HUVECs undergoing oxidative stress and apoptosis, an effect that was counteracted by inhibiting miR-126-5p via negative regulation of HIPK2 which was predicted as a target of miR-126-5p. However, the molecular mechanisms of miR-126-5p-induced HIPK2 regulation were not investigated in this study, neither was the ability of HUVEC to undergo angiogenesis . The biological consequences of miRNAs-induced HIPK2 targeting in angiogenic diseases are summarized below in Table 2. 3.2. HIPK2 in Diabetic Retinopathy (DR) and in Diabetic Wound Healing Diabetic retinopathy (DR) is the primary cause of blindness in the world. It is a complication of diabetes characterized by hyperglycemia that damages retina and has limited treatments options . A key role in the vascular complications of DR has been described for several classes of non-coding RNA including miRNAs . It has been shown that miR-4235-5p can increase proliferation, migration, and angiogenesis of retinal endothelial cells (RECs) cultured in high glucose (HG) condition . In agreement, elevated miR-423-5p levels were found to be present in the plasma of DR patients . RECs cultured in HG showed E2F1-dependent miR-423-5p upregulation that was responsible of HIPK2 downregulation and of HIF-1a and VEGF upregulation . Knockdown of E2F1 or miR-423-5p suppressed the HG-induced angiogenesis and restored the HIPK2 levels . In vivo studies in a mouse model of streptozotocin (STZ)-induced diabetes confirmed that VEGF was upregulated in the retina in correlation with the upregulation of E2F1 and miR-423-5p and the downregulation of HIPK2 . These data suggest that HIPK2 acts as a suppressor of angiogenesis in DR, likely through downregulation of the HIF-1a/VEGF axis, a role played also in angiogenesis during cancer development . The above reported study shed some light into the mechanisms driving DR progression and identified promising biomarkers, such as low HIPK2 levels, and potential targets, such as elevated miR-423-5p levels in the plasma, to predict the disease progression and to eventually design novel therapeutic strategies. The low expression of the HIPK2 levels in DR is also in agreement with our recent study showing that hyperglycemia triggers HIPK2 degradation via HIF-1, increasing tumor progression . Another frequent complication of hyperglycemia is the diabetic foot ulcer , a consequence of neurological disorders and peripheral vascular complications due to impaired angiogenesis that leads to reduced wound healing and increased risk of infections . It has been previously shown that endothelial progenitor cell-derived exosomes containing miR-221-3p alleviate diabetic ulcers improving wound healing . Subsequently, it has been shown that HG inhibited HUVEC migration and capillary formation, effects that could be reversed by treatment with miR-221-3p that promoted angiogenesis and improved the wound healing . The authors of this study also found an increased expression of HIPK2 in skin tissues of diabetic mice when compared to normal ones, as well as in HG-cultured HUVECs . This is in agreement with a finding showing that HG, by downregulating Siah1, increases HIPK2 expression in glomerular mesangial cells of a mouse model of diabetic nephropathy . HIPK2 inhibition with small interfering (si) RNA rescued HUVEC migration and tube formation under HG condition, while it did not affect HUVEC migration and tube formation in normal metabolic condition . HIPK2 was found to be targeted and negatively regulated by miR-221-3p. Subcutaneous injection of miR-221-3p agomir (which upregulates miRNA activity) into diabetic mice, suppressed HIPK2 expression in wound margin tissues and promoted wound healing . These findings indicate that HG condition reduces angiogenesis and impairs wound healing, effects that correlated with the increased expression of HIPK2. Although the authors did not elucidate the molecular mechanisms through which miR-221-3p/HIPK2 may affect angiogenesis in diabetic condition, they suggest that miR-221-3p analogs may be potentially useful for treating diabetic foot ulcers and for improving wound healing . The biological consequences of miRNAs-induced HIPK2 targeting in angiogenic diseases are summarized in Table 2. cancers-15-01566-t002_Table 2 Table 2 Summary of the miRNA/HIPK2 expression described in the reported references. miRNA Cell Type Disease Model Target Reference miR-100-5p HPMEC 1 Gestational hypertension | HIPK2 miR-126-5p HUVEC 2 Myocardial infarction (MI) | HIPK2 miR-423-5p REC 3 Diabetic retinopathy | HIPK2 miR-221-3p HUVEC 2 Diabetic foot ulcer | HIPK2 1 HPMEC: human placental microvascular endothelial cell; 2 HUVEC: human umbilical vein endothelial cell; 3 REC: retinal endothelial cell; 4 NPC: nasopharyngeal carcinoma; |: downregulation; |: upregulation. 4. Conclusions The studies performed in more than twenty years since its discovery have depicted HIPK2 as a central hub in a molecular network that controls several signaling pathways involved in cell death and proliferation and that restrain tumor growth. In this scenario, HIPK2 downregulation by hypoxia-driven mechanisms plays a key role in inducing tumor angiogenesis and solid tumor progression. The role of HIPK2 in restraining tumor angiogenesis has been strengthened by several studies also showing that miRNAs may induce HIPK2 downregulation. Based on these findings, mostly obtained in pre-clinical studies, we can hypothesize that low HIPK2 mRNA or protein levels in cancer tissues compared to the adjacent normal ones can be considered a potential novel prognostic biomarkers of cancer progression, especially if correlated with increased angiogenesis. Interestingly, HIPK2 downregulation by some miRNAs has been shown to be involved in diabetic retinopathy, diabetic wound ulcer, and gestational hypertension, by limiting HIF-1-induced VEGF, and/or b-catenin-induced-VEGF or by activating p53. Defining the molecular basis of angiogenic disorders in greater detail may provide new avenues to improve the prognosis of angiogenic diseases including cancer and to develop more tailored therapeutic strategies. In this regard, the lower expression of HIPK2 in angiogenic tissues compared to the normal ones could become a novel biomarker of angiogenic diseases that deserves to be supported by further studies. Acknowledgments The authors wish to thank people in the lab for helpful discussion. Author Contributions Conceptualization and writing, G.D.; review and editing, A.G., V.D., G.P., M.C. and G.D. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest The authors declare no conflict of interests. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. Figure 1 Schematic representation of the balance between hypoxia and HIPK2 in cancer. When hypoxia is activated ((left) panel) the hypoxia-inducible mechanisms inhibit HIPK2 and the effects of hypoxia (such as angiogenesis, chemoresistance, inhibition of apoptosis, and tumor growth) prevail. When HIPK2 is activated ((right) panel) the HIF-1-induced molecular mechanisms are inhibited and the antitumor effects (such as p53 activation, activation of apoptosis, and tumor regression) prevail. cancers-15-01566-t001_Table 1 Table 1 Summary of the miRNA/HIPK2 and circHIPK2 expression described in the reported references. miRNA/circHIPK2 Tumor Type Biological Effect Target Tissues Cell Lines Reference miR-1229 CRC 1 angiogenesis, metastasis | HIPK2 + + miR-1260b NSCLC 2 angiogenesis, metastasis | HIPK2 + + circHIPK2 DDP-resistant NSCLC 3 angiogenesis, drug resistance | miR-1249-3p, | VEGF + + circHIPK2 CRC 1 reduced overall survival ? + + circHIPK2 NPC 4 tumor progression ? + + 1 CRC: colorectal cancer; 2 NSCLC: non-small cell lung cancer; 3 DDP-resistant: cisplatin resistant; 4 NPC: nasopharyngeal carcinoma; |: downregulation; |: upregulation; +: present in analyzed tissues and cell lines; ?: not determined. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000596 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050878 diagnostics-13-00878 Article Clinicopathological and Imaging Features of Breast Papillary Lesions and Their Association with Pathologic Nipple Discharge Oh Jeongeum Software Formal analysis Data curation Writing - original draft Park Ji Yeon Conceptualization Methodology Validation Writing - review & editing Visualization Supervision * Ekpo Ernest Usang Academic Editor Department of Radiology, Inje University Ilsan Paik Hospital, Goyang 10380, Republic of Korea * Correspondence: [email protected] 24 2 2023 3 2023 13 5 87806 2 2023 22 2 2023 23 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). No studies have evaluated whether any clinicopathological or imaging characteristics of breast papillary lesions are associated with pathological nipple discharge (PND). We analyzed 301 surgically confirmed papillary breast lesions diagnosed between January 2012 and June 2022. We evaluated clinical (age of patient, size of lesion, pathologic nipple discharge, palpability, personal/family history of breast cancer or papillary lesion, location, multiplicity, and bilaterality) and imaging characteristics (Breast Imaging Reporting and Data System (BI-RADS), sonographic, and mammographic findings) and compared malignant versus non-malignant lesions and papillary lesions with versus without PND. The malignant group was significantly older than the non-malignant group (p < 0.001). Those in the malignant group were more palpable and larger (p < 0.001). Family history of cancer and peripheral location in the malignant group were more frequent than in the non-malignant group (p = 0.022 and p < 0.001). The malignant group showed higher BI-RADS, irregular shape, complex cystic and solid echo pattern, posterior enhancement on ultrasound (US), fatty breasts, visibility, and mass type on mammography (p < 0.001, 0.003, 0.009, <0.001, <0.001, <0.001, and 0.01, respectively). On multivariate logistic regression analysis, peripheral location, palpability, and age of >=50 years were factors significantly associated with malignancy (OR: 4.125, 3.556, and 3.390, respectively; p = 0.004, 0.034, and 0.011, respectively). Central location, intraductal nature, hyper/isoechoic pattern, and ductal change were more frequent in the PND group (p = 0.003, p < 0.001, p < 0.001, and p < 0.001, respectively). Ductal change was significantly associated with PND on multivariate analysis (OR, 5.083; p = 0.029). Our findings will help clinicians examine patients with PND and breast papillary lesions more effectively. breast papillary neoplasm nipple discharge mammography ultrasound This study received no external funding. pmc1. Introduction Nipple discharge is a frequently occurring symptom in women, accounting for 2-10% of the symptoms that women present with at breast clinics. Nipple discharge can be classified as a physiologic or a pathologic discharge and is generally considered to be pathologic if it is spontaneous, bloody, clear/serous, unilateral, or arising from a single duct . Common causes of pathologic nipple discharge (PND) are intraductal papilloma, duct ectasia, and malignancy; benign papilloma is the most common etiology (48 to 88%) of PNDs , followed by duct ectasia (up to 33%) , while malignancy accounts for only 5-23% of all PNDs . Papillary lesions of the breast are varied; these include benign papillomas, atypical papillomas, in situ papillary carcinoma, and invasive papillary carcinomas . It is often difficult to distinguish between benign and malignant papillary lesions, because these have been reported to have overlapping clinical symptoms and imaging features . Patient's age, lesion size, multiplicity, and peripheral location of the lesion have been reported to be significant clinical factors, whereas visibility and lesion density on mammography, as well as echo pattern, echogenic halo, orientation, posterior feature, and vascularity on ultrasound, have been reported to be significant radiologic factors, which can differentiate malignant papillary lesions from benign ones . However, previous studies have utilized small sample sizes of less than 200 papillary lesions or have focused on a small number of cases diagnosed as malignancies after being labelled as benign papillomas using core needle biopsies . To the best of our knowledge, no studies have evaluated whether clinicopathological or imaging characteristics of papillary lesions are associated with pathological nipple discharge (PND). Consequently, this study aimed to assess whether any clinical or imaging features can assist in the differentiation of breast papillary lesions and to evaluate whether the clinicopathological and imaging findings of breast papillary lesions are associated with PND. 2. Materials and Methods 2.1. Study Population Our institutional review board approved this retrospective study and waived the requirement of informed patient consent for the review of medical records and radiologic images. We searched the pathological database for patients diagnosed with papillary lesions of the breast via surgical excision between January 2012 and June 2022 at our hospital. During this period, 1993 patients underwent surgical excision of breast lesions. Among them, 257 patients with 325 papillary lesions of the breast were identified. We excluded cases that did not have ultrasound (US) and mammography images available, those that had lesions that were not detectable on US or mammography, and those that had incidental papillomas detected on excised specimens for breast cancer. A total of 301 papillary lesions in 238 women were finally included in this study. 2.2. Clinicopathologic Characteristics Medical records were reviewed for patient age, clinical symptoms (nipple discharge, palpability), location (central or peripheral), multiplicity, bilaterality, personal or family history (breast cancer, previous papillary lesion), and lesion size, as well as final pathological diagnosis of benign, atypical, or malignant papillary lesions. Pathological nipple discharge was defined as discharge from a single duct or a unilateral; spontaneous; bloody, serous, or clear nipple discharge. Lesions at a distance of <=2 cm from the nipple were categorized as central and those at >2 cm from the nipple were considered peripheral and the distance was measured on mammography or US. Multiple lesions were defined as >=2 lesions in the unilateral breast on imaging and pathology. The lesion size was defined as the largest diameter on US. 2.3. Imaging Analysis Mammography was performed using a Lorad Selenia (Hologic, Bedford, MA, USA). US examinations were performed using one of the following high-resolution US systems with an 11-18 MHz linear transducer: HDI 5000 (Philips Healthcare, Bothell, WA, USA), Aixplorer (SuperSonic Imagine, Aix-en-Provence, France), and Aplio I800 (Canon Medical Systems Corporation, Tokyo, Japan). Two radiologists with 2-12 years of experience in breast imaging interpretation retrospectively reviewed the radiologic images by consensus and were blinded to the final pathologic diagnosis. The images were assessed according to the Breast Imaging Reporting and Data System (BI-RADS) lexicon for mammography and US . Breast density, lesion visibility, and lesion type (mass, asymmetry, or calcification) were assessed using mammography. The shape (oval to round, irregular), margin (circumscribed, non-circumscribed), orientation (parallel, nonparallel), echo pattern (hyper-iso, hypo, complex cystic and solid), posterior feature (enhancement, shadowing, absent), calcification (absent, present), and vascularity (absent, present) of the lesion were evaluated on US. Associated ductal change, such as dilatation or continuation with adjacent ducts and the presence of an intraductal feature, were also recorded. 2.4. Statistical Analysis Differences in clinicopathological and imaging variables were compared between the benign/atypical (non-malignant) and malignant groups using Student's t-test, chi-square test, or Fisher's exact test. Differences between the PND and non-PND groups were also evaluated. Multiple linear regression analysis was used to evaluate significant factors associated with malignancy and PND. Statistical analyses were performed using SPSS 25.0 (IBM, Armonk, NY, USA) and, when the p-value was less than 0.05, it was considered to be statistically significant. 3. Results A total of 301 papillary lesions in 238 women (mean age: 49.31 +- 12.22 years, range: 12-88 years) were included in this study. Twenty-one women had bilateral papillary lesions, 13 had multiple lesions present simultaneously, and 22 had relapsed papillary lesions within the study period. There were 192 (63.8%) benign papillary lesions, 68 (22.6%) atypical lesions, and 41 (13.6%) malignant lesions. Furthermore, 55 (18.3%) papillary lesions were associated with PND and 246 (81.7%) were not. 3.1. Comparison of Clinical and Imaging Characteristics between Non-Malignant and Malignant Papillary Lesions Among the clinical factors, age, palpability, lesion size, family history of cancer, and location were significantly different between the non-malignant and malignant groups (Table 1). The mean ages of the benign, atypical, and malignant papillary lesion groups were 46.89 +- 11.18, 49.96 +- 11.49, and 59.56 +- 12.88 years, respectively; additionally, patients in the malignant group were significantly older than those in the non-malignant group (p < 0.001). In the benign, atypical, and malignant lesion groups, 32.3% (62/192), 35.3% (24/68), and 68.3% (29/41), respectively, were >=50 years of age (p < 0.001). Lesion size on US was available for 295 cases: the mean sizes of benign, atypical, and malignant papillary lesions were 0.96 +- 0.65 cm, 0.90 +- 0.43 cm, and 1.76 +- 1.36 cm, respectively. Moreover, lesions in the malignant group were significantly larger than those in the non-malignant group (p < 0.001). The proportion of lesions sized >=1 cm were 32.1% (61/190), 46.9% (30/64), and 65.9% (27/41) in the benign, atypical, and malignant groups, respectively, (p < 0.001). The proportion of lesions that were palpable were 9.4% (18/192), 10.3% (7/68), and 51.2% (21/42) in the benign, atypical, and malignant lesion groups, respectively. Malignant papillary lesions were more frequently palpable than non-malignant lesions (p < 0.001). Family history of cancer (12.2%, 5/41) and peripheral location (46.3%, 19/41) in the malignant group were more frequent than in the non-malignant group (3.8%, 21.5%, respectively) (p = 0.022 and p < 0.001, respectively) . Factors such as pathological nipple discharge, history of breast cancer, previous papillary lesions, multiplicity, and bilaterality were not statistically significant. Among the imaging characteristics, BI-RADS category, shape, echo pattern, posterior feature on US, breast density, visibility, and lesion type on mammography were significantly different between the non-malignant and malignant groups (Table 2). The malignant group showed a higher BI-RADS than the non-malignant group. Moreover, 51.2% (21/41) of malignancies were categorized as 4B, 4C, and 5 and 88.1% (229/260) of the lesions in the non-malignant group were scored as 3 and 4A (p < 0.001). Additionally, 53.7% (22/41) of malignant lesions had an irregular shape and 69.7% (177/254) of non-malignant lesions had an oval to round shape (p = 0.003). A complex cystic and solid echo pattern was found in 24.4% (10/41) of malignant lesions and hyper/iso echotexture was found in 19.2% (49/254) of non-malignant lesions; the proportion of hypoechoic patterns was similar between the two groups (p = 0.009). Posterior enhancement was more frequent in the malignant group (48.8%, 20/41) than in the non-malignant group (p < 0.001). Of the 227 cases with available mammographic data, fatty breasts were more frequent in cases of malignancies (56.8%, 21/37) and dense breasts were common in cases of non-malignancies (86.8%, 165/190) (p < 0.001). The malignant lesions were significantly more visible on mammography (86.5%, 32/37) (p < 0.001); 75.0% of malignancies were masses and 57.6% (38/66) of non-malignancies were presented as asymmetry or calcification only (p = 0.01). Factors such as the presence of intraductal features, margins, orientation, calcification, vascularity, and ductal change on US were not statistically significant. In the multivariate logistic regression analysis, peripheral location, palpability, and patient age of >=50 years were significant factors associated with malignant papillary lesions (OR: 4.125, 3.556, and 3.390, respectively; p = 0.004, 0.034, and 0.011, respectively) (Table 3). 3.2. Comparison of Clinicopathological and Imaging Characteristics of Papillary Lesions between PND and Non-PND Groups Table 4 and Table 5 show the comparison of papillary lesion with or without PND. The location, echo pattern, presence of intraductal features, and ductal changes were statistically significant between the two groups. Central location in the PND group was more frequent than in the non-PND group (90.9% vs. 71.5%, respectively, p = 0.003). In the PND group, 34.5% (19/55) of lesions were hyper/isoechoic, 60% (33/55) were hypoechoic, and 5.5% (3/55) were complex echoic. In the non-PND group, 14.2% (34/240) of lesions were hyper/isoechoic, 73.3% (176/240) were hypoechoic, and 12.5% (30/240) were complex echoic (p < 0.001). Ductal change and intraductal feature were more common in the PND group than in the non-PND group (p < 0.001) . Other factors, including the pathological results, showed no significant association with PND. Ductal change was the only significant factor associated with PND in the multivariate logistic regression analysis (OR: 5.083, p = 0.029) (Table 6). 4. Discussion Benign papilloma is the most common cause of PND and has been reported to account for up to 88% of PND cases . However, (i) the association between the pathology of papillary breast lesions and PND and (ii) the variable clinical or imaging factors affecting PND are unknown. There is a lack of consensus with respect to the management of papillary breast lesions because of differences in populations sampled and the methods employed by the different studies . To the best of our knowledge, our study is the largest series investigating papillary lesions along with their surgical results and the first study focusing on the association between papillary lesions and PND. Our study revealed that old age (>=50 years), palpability of the lesion, family history of breast cancer, large lesion size (>=1 cm), and peripheral location of the lesion were clinically associated with malignant papillary lesions. These results are consistent with those of previous studies that reported that peripheral location, lesion size, and old age are correlated with malignancy . Palpability was also reported to be a significant factor associated with malignant papillary lesions (50%, 5/10) . However, our results regarding family history of breast cancer were discordant with those of previous studies. A study reported no significant differences in family history between benign and atypical/malignant papillary lesions . One study showed a trend toward a higher upgrade rate to a malignancy in patients with a family history of breast cancer, without statistical significance (p = 0.09) . In another study evaluating benign papillomas diagnosed by core biopsy, family history was not associated with an upgraded diagnosis to malignancy (n = 14) . These studies had a limited number of cases of malignancies (up to 21 cases); therefore, a study with a larger number of papillary lesions is necessary in the future. In the present study, high BI-RADS, irregular shape, complex echogenicity, and posterior enhancement were the sonographic characteristics associated with malignancy. These results are supported by previous studies, and high BI-RADS scores were one of the significant factors in predicting malignant progression . A complex echo pattern was reported to be significantly frequent in malignant lesions . Several studies suggested that margins, echo pattern, and posterior features were significant sonographic features that assisted in differentiating papillary lesions . Our study showed that fatty breasts, visibility, and mass type were mammographic features significantly associated with malignancy. Old patients tend to have fatty breast density on mammography, and background fatty density and the large size of malignant papillary lesions may partly explain the greater visibility of the lesions on mammography. In a previous study, fatty breast density was reported to be a predictor of upgrade to malignancy . Visibility on mammography as a significant factor in predicting malignant papillary lesions was already proven by a previous study . Mass on mammography has been reported to be a risk factor for upgrade to malignancy in benign papillomas, diagnosed using core biopsy . The factors significantly associated with PND for papillary lesions in our study were central location of the lesion, hyperechoic/isoechoic pattern, intraductal lesion, and ductal changes on ultrasound. Because common causes of PND are intraductal papilloma or duct ectasia and solitary papillomas usually present as subareolar duct dilatation with internal solid echoes , these results are sufficiently predictable. Therefore, when a physician encounters a patient with PND, an intraductal echogenic lesion with central location or lesion with ductal changes on US may be predicted. This study showed that the surgical pathology of papillary lesions was not associated with PND. Among the 55 papillary lesions with PND, 85.4% of them were benign/atypical lesions and malignancy was only found in 14.5%. According to published studies, the predictive value of nipple discharge for malignant papillary lesions remains unclear. A study including 51 papillary lesions with surgical results reported that 32.4% of benign/atypical papillomas and 0% of malignant papillary lesions presented with a pathological nipple discharge . Wang et al. demonstrated that bloody nipple discharge is a significant indicator of papilloma with high-risk or malignant lesions (p = 0.009) . Ahn et al. reported that bloody nipple discharge was significantly associated with upgraded malignancy in benign papillomas diagnosed by core biopsy . Rizzo et al. reported that 16.2% of papillary lesions upgraded to a diagnosis of malignancy had bloody nipple discharge, but the difference was not statistically significant . However, these studies had only a small number of malignant cases or included only benign papillomas diagnosed by core biopsy. Our study has several limitations. First, this was a retrospective study conducted at a single institution, and the number of papillary breast lesions with pathological nipple discharge was small. Therefore, multicenter studies with larger sample sizes are necessary. Second, we considered multiple or recurrent lesions of each patient to be separate cases; this could have affected the results. Third, seven lesions of patients who had other concurrent malignancy in the ipsilateral breast were not excluded because the patients did not have PND. However, the strength of our study is that all lesions were surgically excised and histopathologically confirmed. Additionally, this is the first study to focus on the association between pathological nipple discharge and the clinicopathological or imaging features of breast papillary lesions. 5. Conclusions Among breast papillary lesions, old age, palpability of lesion, large lesion size, peripheral location of the lesion, and a family history of breast cancer were found to be clinically associated with malignancy. A high BI-RADS category, irregular shape, complex echo, posterior enhancement on US, fatty breasts, mammographic visibility, and type of mass on mammography were also imaging characteristics associated with malignancy. Hyper/iso echo pattern, central location, presence of intraductal feature, and ductal change were significantly associated with PND. Consequently, we propose that breast physicians and radiologists should pay particular attention to the clinical history, physical examination, and imaging findings when they encounter patients with PND and breast papillary lesions. Author Contributions Conceptualization, J.Y.P.; methodology, J.Y.P.; software, J.O.; validation, J.Y.P.; formal analysis, J.O.; data curation, J.O.; writing--original draft preparation, J.O.; writing--review and editing, J.Y.P.; visualization, J.Y.P.; supervision, J.Y.P. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the guidelines of the Declaration of Helsinki and was approved by the Institutional Review Board of Inje University Ilsan Paik Hospital (2022-11-009). Informed Consent Statement Patient consent was waived owing to the retrospective nature of the analysis. Data Availability Statement Data generated in this study are available upon request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 A 46-year-old woman presented with palpable masses in the breast accompanied by bloody nipple discharge. (a) Mammography shows a hyperdense mass with spiculated margins and an irregular shape in the upper central portion (arrowheads) and a circumscribed round hyperdense mass in the upper inner portion of right breast (arrows). (b) Ultrasound shows a 3.9 cm heterogeneous echoic mass (arrowheads) with irregular shape and internal vascularity in the right breast, in the 1 o'clock position, 5 cm from the nipple. Another 2.7 cm complex cystic and solid mass (arrows) is noted in the right breast in the 2 o'clock position, 6 cm from the nipple. Ductal carcinoma in situ with focal invasion in the background of papilloma was confirmed by surgery. The patient had a family history of breast cancer (mother). Figure 2 A 38-year-old woman presented with bloody nipple discharge. Ultrasound shows a 1.0 cm isoechoic intraductal mass with internal blood flow in the left breast subareolar area. Benign intraductal papilloma was confirmed on pathology after surgical excision. diagnostics-13-00878-t001_Table 1 Table 1 Clinical characteristics of papillary lesions according to pathological examination. Total Benign Group Atypical Group Malignant Group p-Value * Age (years) 49.31 +- 12.22 (12-88) 46.89 +- 11.18 (12-88) 49.96 +- 11.49 (28-74) 59.56 +- 12.88 (35-85) <0.001 Age group (n) Age < 50 years Age >= 50 years 301 187 (62.1%) 114 (37.9%) 192 130 (67.7%) 62 (32.3%) 68 44 (64.7%) 24 (35.3%) 41 13 (31.7%) 28 (68.3%) <0.001 Lesion size (cm) 1.06 +- 0.80 (0.3-6.8) 0.96 +- 0.65 (0.3-4.5) 0.90 +- 0.43 (0.3-2.0) 1.76 +- 1.36 (0.3-6.8) <0.001 Lesion size group (n) <1 cm >=1 cm 295 177 (58.80%) 118 (39.20%) 190 129 (67.9%) 61 (32.1%) 64 34 (53.1%) 30 (46.9%) 41 14 (34.1%) 27(65.9%) 0.001 Palpability (n) (-) (+) 301 255 (84.7%) 46 (15.30%) 192 174 (90.6%) 18 (9.4%) 68 61 (89.7%) 7 (10.3%) 41 20 (48.8%) 21 (51.2%) <0.001 PND (n) (-) (+) 301 246 (81.7%) 55 (18.3%) 192 152 (79.2%) 40 (20.8%) 68 61 (89.7%) 7 (10.3%) 41 33 (80.5%) 8 (19.5%) 0.825 Multiplicity (n) Single Multiple 301 272 (90.40%) 29 (9.60%) 192 174 (90.6%) 18 (9.4%) 68 63 (92.6%) 5 (7.4%) 41 35 (85.4%) 6 (14.6%) 0.243 Bilaterality (n) Unilateral Bilateral 301 258 (85.70%) 43 (14.30%) 192 164 (85.4%) 28 (14.6%) 68 59 (86.8%) 9 (13.2%) 41 35 (85.4%) 6 (14.6%) 0.945 Location (n) Central Peripheral 301 226 (75.1%) 75 (24.9%) 192 157 (81.8%) 35 (18.2%) 68 47 (69.1%) 21 (30.9%) 41 22 (53.7%) 19 (46.3%) <0.001 Previous history of cancer (n) (-) (+) 301 292 (97.0%) 9 (3.0%) 192 188 (97.9%) 4 (2.1%) 68 64 (94.1%) 4 (5.9%) 41 40 (97.6%) 1 (2.4%) >0.999 Family history of cancer (n) (-) (+) 301 286 (95.0%) 15 (5.0%) 192 183 (95.3%) 9 (4.7%) 68 67(98.5%) 1 (1.5%) 41 36(87.8%) 5 (12.2%) 0.022 Previous papillary lesion (n) (-) (+) 301 267 (88.7%) 34 (11.30%) 192 167 (87.0%) 25 (13.0%) 68 62 (91.2%) 6 (8.8%) 41 38 (92.7%) 3 (7.3%) 0.595 * Malignant group vs. benign and atypical groups. diagnostics-13-00878-t002_Table 2 Table 2 Imaging characteristics of papillary lesions according to final pathology. Total Benign Group Atypical Group Malignant Group p-Value * BI-RADS (n) Category 3 Category 4A Category 4B Category 4C Category 5 301 7 (2.30%) 242 (80.4%) 31 (10.3%) 12 (4.0%) 9 (3.0%) 192 4 (2.1%) 168 (87.5%) 18 (9.4%) 2 (1.0%) 0 (0%) 68 3 (4.4%) 54 (79.4%) 7 (10.3%) 2 (2.9%) 2 (2.9%) 41 0 (0%) 20 (48.8%) 6 (14.6%) 8 (19.5%) 7 (17.1%) <0.001 Shape on US (n) Oval to round Irregular 295 196 (65.10%) 99 (32.90%) 190 133 (70.0%) 57 (30.0%) 64 44 (68.8%) 20 (31.3%) 41 19 (46.3%) 22 (53.7%) 0.003 Margin on US (n) Circumscribed Non-circumscribed 295 99 (32.90%) 196 (65.10%) 190 69 (43.1%) 121 (63.7%) 64 21 (32.8%) 43 (67.2%) 41 9 (22.0%) 32 (78.0%) 0.09 Echo on US (n) Hyper/iso Hypo Complex 295 53 (17.60%) 209 (69.40%) 33 (11.0%) 190 39 (20.5%)$ 136 (71.6%) 15 (7.9%) 64 10 (15.6%) 46 (71.9%) 8 (12.5%) 41 4 (9.8%) 27 (65.9%) 10 (24.4%) 0.009 Orientation on US (n) Parallel Non-parallel 295 268 (89.00%) 27 (9.0%) 190 175 (92.1%) 15 (7.9%) 64 55 (85.9%) 9 (14.1%) 41 38 (92.7%) 3 (7.3%) >0.999 Posterior feature on US (n) Enhancement Shadowing No 295 71 (23.60%) 4 (1.30%) 220 (73.10%) 190 38 (20.0%) 2 (1.1%) 150 (78.9%) 64 13 (20.3%) 1 (1.6%) 50 (78.1%) 41 20 (48.8%) 1 (2.4%) 20 (48.8%) <0.001 Calcification on US (n) (-) (+) 295 281 (93.40%) 14 (4.70%) 190 183 (96.3%) 7 (3.7%) 64 59 (92.2%) 5 (7.8%) 41 39 (95.1%) 2 (4.9%) >0.999 Vascularity on US (n) (-) (+) 286 79 (26.20%) 207 (68.80%) 186 58 (31.2%) 128 (68.8%) 61 14 (23.0%) 47 (77.0%) 39 7 (17.9%) 32 (82.1%) 0.27 Intraductal feature on US (n) (-) (+) 295 186 (61.80%) 109 (36.20%) 190 111 ((58.4%) 79 (41.6%) 64 45 (70.3%) 19 (29.7%) 41 30 (73.2%) 11 (26.8%) 0.148 Ductal change on US (n) (-) (+) 295 180 (59.80%) 115 (38.20%) 190 107 (56.3%) 83 (43.7%) 64 45 (70.3%) 19 (29.7%) 41 28 (68.3%) 13 (31.7%) 0.303 Density on MG (n) Fatty Dense 227 46 (15.30%) 181 (60.10%) 136 16 (11.8%) 120 (88.2%) 54 9 (16.7%) 45 (83.3%) 37 21 (56.8%) 16 (43.2%) <0.001 Visibility on MG (n) (-) (+) 227 129 (56.8%) 98 (43.2%) 136 95 (69.9%) 41 (30.1%) 54 29 (53.7%) 25 (46.3%) 37 5 (13.5%) 32 (86.5%) <0.001 Type on MG Mass Asymmetrical Calcification only 98 52 (53.0%) 38 (38.8%) 8 (8.2%) 41 15 (36.6%) 23 (56.1%) 3 (7.3%) 25 13 (52.0%) 8 (32.0%) 4 (16.0%) 32 24 (75.0%) 7 (21.9%) 1 (3.1%) 0.01 * Malignant group vs. benign and atypical lesion groups. BI-RADS, Breast Imaging Reporting and Data System; MG, mammography; US, ultrasound. diagnostics-13-00878-t003_Table 3 Table 3 Multivariate logistic regression analysis to determine the risk factors associated with malignant papillary breast lesions. Variable Odds Ratio (95% CI) p-Value Location Peripheral 4.125 (1.582-10.753) 0.004 Palpability 3.556 (1.103-11.470) 0.034 Age group >=50 years 3.390 (1.327-8.661) 0.011 CI: confidence interval. diagnostics-13-00878-t004_Table 4 Table 4 Clinical characteristics of papillary lesions, classified according to pathologic nipple discharge. PND Group (n = 55) Non-PND Group (n = 246) p-Value Age (years) 47.42 +- 12.88 (range: 12-74 years) 49.73 +- 12.05 (range: 19-88 years) 0.205 Age group <50 years >=50 years 36 (65.5%) 19 (34.5%) 151 (61.4%) 95 (38.6%) 0.574 Lesion size 1.20 +- 0.88 (range: 0.4-5.0) 1.03 +- 0.78 (0.3-6.8) 0.163 Lesion size group (n) <1 cm >=1 cm 55 29 (52.7%) 26 (47.3%) 240 148 (61.7%) 92 (38.3%) 0.222 Pathology Benign Atypical Malignancy 40 (72.7%) 7 (12.7%) 8 (14.5%) 152 (61.8%) 61 (24.8%) 33 (13.4%) 0.151 Location Central Peripheral 50 (90.9%) 5 (9.1%) 176 (71.5%) 70 (28.5%) 0.003 Multiplicity Single Multiple 49 (89.1%) 6 (10.9%) 223 (90.7%) 23 (9.3%) 0.723 Bilaterality Unilateral Bilateral 48 (87.3%) 7 (12.7%) 36 (14.6%) 210 (85.4%) 0.715 History of previous cancer (-) (+) 54 (98.2%) 1 (1.8%) 238 (96.7%) 8 (3.3%) >0.999 Family history of cancer (-) (+) 50 (90.9%) 5 (9.1%) 236 (95.9%) 10 (4.1%) 0.121 Previous papillary lesion (-) (+) 53 (96.4%) 2 (3.6%) 214 (87.0%) 32 (13.0%) 0.057 PND, pathologic nipple discharge. diagnostics-13-00878-t005_Table 5 Table 5 Imaging characteristics of papillary lesions, classified according to pathologic nipple discharge. PND Group (n = 55) Non-PND Group (n = 246) p-Value BI-RADS Category 3 Category 4A Category 4B Category 4C Category 5 1 (1.8%) 41 (74.5%) 9 (16.4%) 1 (1.8%) 3 (5.5%) 6 (2.4%) 201 (81.7%) 22 (8.9%) 11 (4.5%) 6 (2.4%) 0.295 Shape on US (n) Oval to round Irregular 55 38 (69.1%) 17 (30.9%) 240 158 (65.8%) 82 (34.2%) 0.644 Margin on US (n) Circumscribed Non-circumscribed 55 23 (41.8%) 32 (58.2%) 240 76 (31.7%) 164 (68.3%) 0.150 Echo on US (n) Hyper/iso Hypo Complex 55 19 (34.5%) 33 (60.0%) 3 (5.5%) 240 34 (14.2%) 176 (73.3%) 30 (12.5%) 0.001 Orientation on US (n) Parallel Non-parallel 55 51 (92.7%) 4 (7.3%) 240 217 (90.4%) 23 (9.6%) 0.796 Posterior feature on US (n) Enhancement Shadowing No 55 11 (20.0%) 0 44 (80.0%) 240 60 (25.0%) 4 (1.7%) 176 (73.3%) 0.439 Calcification on US (n) (-) (+) 55 51 (92.7%) 4 (7.3%) 240 230 (95.8%) 10 (4.2%) 0.304 Vascularity on US (n) (-) (+) 54 10 (18.5%) 44 (81.5%) 232 69 (29.7%) 163 (70.3%) 0.209 Ductal change on US (n) (-) (+) 55 13 (23.6%) 42 (76.4%) 240 167 (69.6%) 73 (30.4%) <0.001 Intraductal feature on US (n) (-) (+) 55 15 (27.3%) 40 (72.7%) 240 171 (71.3%) 69 (28.8%) <0.001 Density on MG (n) Fatty Dense 47 11 (23.4%) 36 (76.6%) 180 35 (19.4%) 145 (80.6%) 0.548 Visibility on MG (n) (-) (+) 47 25 (53.2%) 22 (46.8%) 180 104 (57.8%) 76 (42.2%) 0.572 Type on MG (n) Mass Asymmetry Calcification only 22 9 (40.9%) 12 (54.5%) 1 (4.5%) 76 43 (56.6%) 26 (34.2%) 7 (9.2%) 0.216 PND, pathologic nipple discharge; BI-RADS, Breast Imaging Reporting and Data System; MG, mammography; US, ultrasound. diagnostics-13-00878-t006_Table 6 Table 6 Multivariate logistic regression analysis to determine the risk factors associated with pathologic nipple discharge in papillary breast lesions. Variable Odds Ratio (95% CI) p-Value Ductal change 5.083 (1.180-21.894) 0.029 CI: confidence interval. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Patel B.K. Falcon S. Drukteinis J. Management of nipple discharge and the associated imaging findings Am. J. Med. 2015 128 353 360 10.1016/j.amjmed.2014.09.031 25447625 2. Morrogh M. Park A. Elkin E.B. King T.A. Lessons learned from 416 cases of nipple discharge of the breast Am. J. Surg. 2010 200 73 80 10.1016/j.amjsurg.2009.06.021 20079481 3. Vargas H.I. Vargas M.P. Eldrageely K. Gonzalez K.D. Khalkhali I. Outcomes of clinical and surgical assessment of women with pathological nipple discharge Am. Surg. 2006 72 124 128 10.1177/000313480607200205 16536240 4. Nicholson B.T. Harvey J.A. Patrie J.T. Mugler J.P. 3D-MR ductography and contrast-enhanced MR mammography in patients with suspicious nipple discharge: A feasibility study Breast J. 2015 21 352 362 10.1111/tbj.12417 25882883 5. Cabioglu N. Hunt K.K. Singletary S.E. Stephens T.W. Marcy S. Meric F. Ross M.I. Babiera G.V. Ames F.C. Kuerer H.M. Surgical decision making and factors determining a diagnosis of breast carcinoma in women presenting with nipple discharge J. Am. Coll. Surg. 2003 196 354 364 10.1016/S1072-7515(02)01606-X 12648684 6. Park C.J. Kim E.K. Moon H.J. Yoon J.H. Kim M.J. Reliability of breast ultrasound BI-RADS final assessment in mammographically negative patients with nipple discharge and radiologic predictors of malignancy J. Breast Cancer 2016 19 308 315 10.4048/jbc.2016.19.3.308 27721881 7. Morrogh M. Morris E.A. Liberman L. Borgen P.I. King T.A. The predictive value of ductography and magnetic resonance imaging in the management of nipple discharge Ann. Surg. Oncol. 2007 14 3369 3377 10.1245/s10434-007-9530-5 17896158 8. Tavassoli F.A. Papillary Lesions Pathology of the Breast Tavassoli F.A. Appleton & Lange Norwalk, CN, USA 1992 193 227 9. Koerner F.C. Papilloma and Related-Benign Lesions Rosen's Breast Pathology 4th ed. Hoda S. Brogi E. Koerner F.C. Rosen P.P. Lippincott Williams & Wilkins Philadelphia, PA, USA 2014 95 152 10. Lam W.W.M. Chu W.C.W. Tang A.P.Y. Tse G. Ma T.K.F. Role of radiologic features in the management of papillary lesions of the breast AJR Am. J. Roentgenol. 2006 186 1322 1327 10.2214/AJR.04.1908 16632726 11. Jung S.Y. Kang H.S. Kwon Y. Min S.Y. Kim E.A. Ko K.L. Lee S. Kim S.W. Risk factors for malignancy in benign papillomas of the breast on core needle biopsy World J. Surg. 2010 34 261 265 10.1007/s00268-009-0313-y 19997916 12. Fadzli F. Rahmat K. Ramli M.T. Rozalli F.I. Hooi T.K. Fadzli A.N. Hoong S.M. Ramli N.M. Taib N.A.M. Spectrum of imaging findings of papillary breast disease. A radiopathological review in a tertiary center Medicine 2021 100 e25297 10.1097/MD.0000000000025297 33879660 13. Choi S.H. Jo S. Kim D.H. Park J.S. Choi Y. Kook S.H. Chung E.C. Lee S.Y. Clinical and imaging characteristics of papillary neoplasms of the breast associated with malignancy: A retrospective cohort study Ultrasound Med. Biol. 2014 40 2599 2608 10.1016/j.ultrasmedbio.2014.06.018 25220267 14. Kuzmiak C.M. Lewis M.Q. Zeng D. Liu X. Role of sonography in the differentiation of benign, high-risk, and malignant papillary lesions of the breast J. Ultrasound Med. 2014 33 1545 1552 10.7863/ultra.33.9.1545 25154934 15. Shin H.J. Kim H.H. Kim S.M. Yang H.R. Sohn J.H. Kwon G.Y. Gong G. Papillary lesions of the breast diagnosed at percutaneous sonographically guided biopsy: Comparison of sonographic features and biopsy methods AJR Am. J. Roentgenol. 2008 190 630 636 10.2214/AJR.07.2664 18287432 16. Kim T.H. Kang D.K. Kim S.Y. Lee E.J. Jung Y.S. Yim H. Sonographic differentiation of benign and malignant papillary lesions of the breast J. Ultrasound Med. 2008 27 75 82 10.7863/jum.2008.27.1.75 18096733 17. Youk J.H. Kim E.K. Kwak J.Y. Son E.J. Park B.W. Kim S.I. Benign papilloma without atypia diagnosed at US-guided 14-gauge core-needle biopsy: Clinical and US features predictive of upgrade to malignancy Radiology 2011 258 81 88 10.1148/radiol.10100728 20971773 18. MacColl C. Salehi A. Parpia S. Hodgson N. Ramonas M. Williams P. Benign breast papillary lesions diagnosed on core biopsy: Upgrade rate and risk factors associated with malignancy on surgical excision Virchows Arch. 2019 475 701 707 10.1007/s00428-019-02626-5 31372738 19. D'Orci C.J. Sickles E.A. Mendelson E.B. Morris E.A. ACR. BI-RADS Atlas, Breast Imaging Reporting And Data System American College of Radiology Washington, DC, USA 2013 20. Chung J. Lee W.K. Cha E.S. Lee J.E. Kim J.H. Ryu Y.H. Shear-wave elastography for the differential diagnosis of breast papillary lesions PLoS ONE 2016 11 e0167118 10.1371/journal.pone.0167118 27893857 21. Lee S.J. Wahab R.A. Sobel L.D. Zhang B. Brown A.L. Lewis K. Vijapura C. Mahoney M.C. Analysis of 612 benign papillomas diagnosed at core biopsy: Rate of upgrade to malignancy, factors associated with upgrade, and a proposal for selective surgical excision AJR Am. J. Roentgenol. 2021 217 1299 1311 10.2214/AJR.21.25832 34008998 22. Polat D.S. Knippa E.E. Ganti R. Seiler S.J. Goudreau S.H. Benign breast papillomas without atypia diagnosed with core needle biopsy: Outcome of surgical excision and imaging follow-up Eur. J. Radiol. 2020 131 109237 10.1016/j.ejrad.2020.109237 32905954 23. Ahn S.K. Han W. Moon H.G. Kim M.K. Noh D.Y. Jung B.W. Kim S.W. Ko E. Management of benign papilloma without atypia diagnosed at ultrasound-guided core needle biopsy: Scoring system for predicting malignancy Eur. J. Surg. Oncol. 2018 44 53 58 10.1016/j.ejso.2017.10.214 29174198 24. Wang L.J. Wu P. Li X.X. Luo R. Wang D.B. Guan W.B. Magnetic resonance imaging features for differentiating breast papilloma with high-risk or malignant lesions from benign papilloma: A retrospective study on 158 patients World J. Surg. Oncol. 2018 16 234 10.1186/s12957-018-1537-9 30558621 25. Liberman L. Tornos C. Huzjan R. Bartella L. Morris E.A. Dershaw D.D. Is surgical excision warranted after benign, concordant diagnosis of papilloma at percutaneous breast biopsy? AJR Am. J. Roentgenol. 2006 186 1328 1334 10.2214/AJR.05.0151 16632727 26. Wen X. Cheng W. Nonmalignant breast papillary lesions at core-needle biopsy: A meta-analysis of underestimation and influencing factors Ann. Surg. Oncol. 2013 20 94 101 10.1245/s10434-012-2590-1 22878621 27. Hong Y.R. Song B.J. Jung S.S. Kang B.J. Kim S.H. Chae B.J. Predictive factors for upgrading patients with benign breast papillary lesions using a core needle biopsy J. Breast Cancer 2016 19 410 416 10.4048/jbc.2016.19.4.410 28053629 28. Ciurea A. Calin A. Ciortea C. Dudea S.M. Ultrasound in the diagnosis of papillary breast lesions Med. Ultrason. 2015 17 392 397 10.11152/mu.2013.2066.173.aci 26343090 29. Zaky M.M. Hafez A. Zaky M.M. Shoma A. Soliman N.Y. Elmokadem A.H. MRI for assessment of pathologic nipple discharge: Is it mandatory? Egypt. J. Radiol. Nucl. Med. 2019 50 92 10.1186/s43055-019-0105-9 30. Rizzo M. Linebarger J. Lowe M.C. Pan L. Gabram S.G. Vasquez L. Cohen M.A. Mosunjac M. Management of papillary breast lesions diagnosed on core-needle biopsy: Clinical pathologic and radiologic analysis of 276 cases with surgical follow-up J. Am. Coll. Surg. 2012 214 280 287 10.1016/j.jamcollsurg.2011.12.005 22244207 |
PMC10000597 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050768 cells-12-00768 Article Lacticaseibacillus paracasei GM-080 Ameliorates Allergic Airway Inflammation in Children with Allergic Rhinitis: From an Animal Model to a Double-Blind, Randomized, Placebo-Controlled Trial Lin En-Kwang 12+ Chang Wen-Wei 34+ Jhong Jhih-Hua 5 Tsai Wan-Hua 6 Chou Chia-Hsuan 6 Wang I-Jen 78910*++ Rotondo John Charles Academic Editor Torreggiani Elena Academic Editor Mazziotta Chiara Academic Editor 1 Institute of Clinical Medicine, National Yang Ming Chiao Tung University, Taipei 112304, Taiwan 2 Division of Colorectal Surgery, Department of Surgery, Wanfang Hospital, Taipei Medical University, Taipei 110301, Taiwan 3 School of Biomedical Sciences, Chung Shan Medical University, Taichung 402306, Taiwan 4 Department of Medical Research, Chung Shan Medical University Hospital, Taichung 402306, Taiwan 5 Department of Medical Research, Hsinchu MacKay Memorial Hospital, Hsinchu 300044, Taiwan 6 Research and Development Department, GenMont Biotech Incorporation, Tainan 741014, Taiwan 7 Department of Pediatrics, Taipei Hospital, Ministry of Health and Welfare, New Taipei 242033, Taiwan 8 School of Medicine, National Yang Ming Chiao Tung University, Taipei 112304, Taiwan 9 College of Public Health, China Medical University, Taichung 406040, Taiwan 10 National Institute of Environmental Health Sciences, National Health Research Institutes, Miaoli 350401, Taiwan * Correspondence: [email protected]; Tel.: +886-2-2276-5566 (ext. 2532); Fax: +886-2-2998-8028 + These authors contributed equally to this work. ++ Current address: Department of Pediatrics, Taipei Hospital, Ministry of Health and Welfare, 127 Su-Yuan Road, Hsin-Chuang Dist., New Taipei City 242033, Taiwan. 28 2 2023 3 2023 12 5 76816 1 2023 16 2 2023 24 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Background: Probiotics may facilitate the clinical management of allergic diseases. However, their effects on allergic rhinitis (AR) remain unclear. We examined the efficacy and safety of Lacticaseibacillus paracasei GM-080 in a mouse model of airway hyper-responsiveness (AHR) and in children with perennial AR (PAR) by using a double-blind, prospective, randomized, placebo-controlled design. Methods: The production of interferon (IFN)-g and interleukin (IL)-12 was measured by using an enzyme-linked immunosorbent assay. GM-080 safety was evaluated via the whole-genome sequencing (WGS) of virulence genes. An ovalbumin (OVA)-induced AHR mouse model was constructed, and lung inflammation was evaluated by measuring the infiltrating leukocyte content of bronchoalveolar lavage fluid. A clinical trial was conducted with 122 children with PAR who were randomized to receive different doses of GM-080 or the placebo for 3 months, and their AHR symptom severity scores, total nasal symptom scores (TNSSs), and Investigator Global Assessment Scale scores were examined. Results: Among the tested L. paracasei strains, GM-080 induced the highest IFN-g and IL-12 levels in mouse splenocytes. WGS analysis revealed the absence of virulence factors or antibiotic-resistance genes in GM-080. The oral administration of GM-080 at 1 x 107 colony forming units (CFU)/mouse/day for 8 weeks alleviated OVA-induced AHR and reduced airway inflammation in mice. In children with PAR, the oral consumption of GM-080 at 2 x 109 CFU/day for 3 months ameliorated sneezing and improved Investigator Global Assessment Scale scores significantly. GM-080 consumption led to a nonsignificant decrease in TNSS and also nonsignificantly reduced IgE but increased INF-g levels. Conclusion: GM-080 may be used as a nutrient supplement to alleviate airway allergic inflammation. allergic airway inflammation animal model allergic rhinitis clinical trial probiotics Ministry of Science and TechnologyMOST 109-2628-B-192 - Biotech (Taiwan)This work was supported by a grant from the Ministry of Science and Technology (MOST 109-2628-B-192 -001-) and by GenMont Biotech (Taiwan). pmc1. Introduction Allergic rhinitis (AR) and asthma, which occur in the upper and lower airways, respectively, and their co-occurrence are noted in >40% of patients with allergic airway inflammation . T helper 2 (Th2) cells are considered a critical factor in the development of allergic airway inflammation. Animal models have revealed that the transfer of antigen-specific Th2 cells into naive mice followed by challenge with the specific antigen through inhalation induces asthmatic responses, such as airway hyper-responsiveness (AHR), eosinophilic inflammation, and mucus hyperproduction . The airway expression of interleukin (IL)-4, IL-5, and IL-13 released by Th2 cells is a crucial mediator . However, the major limitations related to biological drugs for allergic airway inflammation, such as dupilumab, include high costs and the loss of responsiveness over time . Thus, low-cost alternative agents for allergic airway inflammation should be developed. Although animal and human studies have indicated the benefits of probiotics in allergic disease treatment, medical societies hold a conservative attitude toward probiotic use. The World Allergy Organization states that despite the insufficient volume of data on the relevant probiotic strains and dosages, the application of probiotics in pregnant and breastfeeding women, as well as in infants with a family history of allergic disease, particularly eczema, may have considerable effects . The term "pharmabiotics" has been proposed, and the underlying immunomodulatory pathway is of interest to several researchers and clinicians . In a mouse model of house dust mite-induced asthma, Voo et al. found that combination treatment with Lacticaseibacillus rhamnosus and corticosteroid reduced AHR, serum IgE levels, and Th2 cytokines . Lan et al. reported that the oral administration of Lactiplantibacillus plantarum CQPC11 strain to an ovalbumin (OVA)-induced asthmatic mouse model decreased OVA-specific IgE or IgG1 in sera and proinflammatory cytokines in bronchoalveolar lavage fluid (BALF) . Yan et al. reported a meta-analysis of the beneficial effects of probiotics on AR by collecting 30 randomized controlled trials. Their results revealed that the consumption of probiotics improved the scores of Rhinitis Quality of Life and Rhinitis Total Symptom but not immunological parameters, including blood eosinophil count or total and antigen-specific serum IgE levels . Given the unequal effects of different probiotic species or strains in treating allergic AHR , investigating the efficacy of different probiotic strains in cell models, animal models, and clinical trials before their use as food supplements is essential. Here, we selected Lacticaseibacillus paracasei strain GM-080 (LP-33, BCRC 910220, CCTCC M 204012), which has been reported to alleviate allergic dermatitis in infants , grass pollen-induced persistent AR in adults , and perennial AR (PAR) in infants , as the main target probiotic. We examined its therapeutic effects in an asthma mouse model and a double-blind, prospective, randomized, placebo-controlled study on children with PAR. 2. Materials and Methods 2.1. L. paracasei Strain and Cell-Wall Component Preparation L. paracasei strains were obtained from GenMont Biotech (Tainan, Taiwan). Bacterial cells were cultured in MRS broth overnight, then resuspended in sterile phosphate-buffered saline (PBS), and subsequently diluted to 4 x 107 cells/mL. Lipoteichoic acid (LTA) and peptidoglycan (PGN) were purified from live concentrated GM-080 (1 x 1011 colony forming units [CFU]/mL) as previously reported . 2.2. Mouse Splenocyte Stimulation and Enzyme-Linked Immunosorbent Assay-Based Detection of Cytokines The mouse spleen was excised, passed through a 45 mm cell strainer (BD Biosciences, Franklin Lakes, NJ, USA) for conversion into a single-cell suspension, then subjected to red blood cell lysis by using RBC lysis buffer (eBiosciences Inc., San Diego, CA, USA). Isolated splenocytes were seeded into a 96-well-plate at a cell density of 4 x 105 cells/well and incubated with 2 mg/mL ConA, 1 mg/mL lipopolysaccharide (LPS), or GM-080 with different multiplicities of infection (MOIs) separately for 48 h. Cell culture supernatants were then harvested, and the presence of cytokines (IL-5, IL-12, and IFN-g) was detected by utilizing commercial enzyme-linked immunosorbent assay (ELISA) kits (Cat. No. 555256 for mouse IL-12[P70] and Cat. No. 555138 for mouse IFN-g; purchased from BD Biosciences, Franklin Lakes, NJ, USA). 2.3. Antimicrobial Susceptibility Profiling Antimicrobial susceptibility was determined by using the broth microdilution method and lymphocyte separation medium (LSM, including 90% IST medium [Cat. No. CM0473; Oxoid, Basingstoke, Hampshire, UK] and 10% MRS medium (Cat. No. 288130; Difco Laboratories Inc., Franklin Lakes, NJ, USA) in accordance with the guidelines of the Quality and Standards Authority of Ethiopia (ES ISO10932:2012). Twofold dilutions of clinically relevant antibiotics (clindamycin, chloramphenicol, erythromycin, gentamicin, kanamycin, streptomycin, tetracycline, and ampicillin, all from Sigma-Aldrich, Saint Louis, MO, USA) were prepared in LSM. Approximately 50 mL of 6 x 105 CFU/mL L. paracasei cells were loaded into a 96-well plate, followed by 50 mL of multiple antibiotics diluted in LSM. The plate was incubated under anaerobic conditions at 37 degC for 16-24 h. Minimum inhibitory concentrations (MICs) were defined as the lowest concentrations of antibiotics at which the growth of L. paracasei was completely inhibited. Strains were classified as susceptible or resistant by using the microbiological cutoffs established by the European Food Safety Authority (EFSA) . 2.4. DNA Extraction, Whole-Genome Sequencing, and Hybrid Genome Assembly We established a whole-genome sequencing (WGS) assembly pipeline in accordance with a previously reported procedure with minor modifications. In brief, genomic DNA was fragmented through ultrasonication by using Covaris S2 (Covaris, Woburn, MA, USA). Indexed polymerase chain reaction-free library construction was performed by using the multiplexed high-throughput sequencing TruSeq DNA Sample Preparation Kit (Illumina, Inc., San Diego, CA, USA) in accordance with the manufacturer's protocols with minor modifications. The GM-080 genome was deeply sequenced through Nanopore GridIon long-read sequencing with 175-fold coverage and whole-genome shotgun sequencing by using 2 x 250-bp paired-end sequencing at 125-fold coverage on a MiSeq platform and hybrid-assembled on a MaSuRCA v3.3.1 assembler . Benchmarking Universal Single-Copy Orthologs (version 4.0.0) was applied for genome completeness assessment through comparison with the lactobacillales_odb10 gene database. We deposited all the sequencing data in GenBank under BioProject ID no. PRJNA824946. 2.5. Annotation of Protein-Coding Genes, Virulence Factors, and Antibiotic Resistance The protein-coding genes in the GM-080 genome were annotated by using Prokka . On the basis of the risk assessment of potential genes of concern for microorganisms to be used in the food chain established by EFSA , virulence factors in the genome were separately identified by running BLAST against the virulence factor database (VFDB) by using the following criteria: query sequence hits with identity >=80%, alignment coverage > 70%, and E-value < 10-30. Antibiotic-resistance genes were predicted through a BLAST search against both the Comprehensive Antibiotic Resistant Database (CARD) and ResFinder (version 4.1) database by using the same criteria. Whole-genome average nucleotide identity was computed by OrthoANI . GM-080 phylogeny and other related genomes were reconstructed using MEGA X . After annotation, the circular genome atlas was generated using the Circos visualization tool . 2.6. OVA-Induced AHR Mouse Model Seven-week-old female BALB/c mice were purchased from The Experimental Animal Facility of the College of Medicine, National Taiwan University (Taipei, Taiwan). Allergic airway inflammation was induced with an intraperitoneal injection of 50 mg of OVA (purchased from Sigma-Aldrich) mixed with the Th2-adjuvant aluminum hydroxide on day 0 and was followed by the administration of 25 mg of OVA on days 14, 28, 42, and 56 to sensitize the mice and then by intranasal challenge with OVA (100 mg) on days 67 and 68. Sera were collected from the retro-orbital sinus on the day before the first OVA sensitization; on post sensitization days 35, 49, and 63; and on the day of sacrifice. 2.7. ELISA Determination of OVA-Specific Immunoglobulins A 96-well plate was coated with OVA at 1 mg/well and then incubated with blocking solution (1% bovine serum albumin in PBS buffer) at room temperature. The wells were added with mouse serum followed by antibodies diluted with the blocking solution (1:50, 1:1000, and 1:10,000 dilutions for IgE, IgG2a, and IgG1, respectively) and incubated at 4 degC overnight. Biotin-conjugated antibodies against mouse IgE, IgG1, or IgG2a (BD Biosciences) were added after the wells were washed with 0.05% Tween-20/PBS buffer. Then, the plate was incubated at room temperature for 2 h. Next, streptavidin-conjugated horseradish peroxidase was added to the wells. 3,3',5,5'-Tetramethylbenzidine substrate was used for color development. After the termination of the reaction by using 2 N H2SO4, absorbance was measured at 450 nm on a VERSAmax microplate reader (Molecular Devices, San Jose, CA, USA). 2.8. AHR Determination In mice, the development of AHR was determined on a Buxco system (Biosystem XA; Buxco Electronics, Sharon, CT, USA). The enhanced pause (Penh) values were calculated with the following formula: (pause x PIF)/PEF, with pause being (Te - Tr)/Tr. Here, PIF is the peak inspiratory flow, PEF is the peak expiratory flow, Te is the expiratory time, and Tr is the relaxation time. First, we recorded the Penh of the mice under awake conditions by using a whole-body plethysmograph chamber for 3 min with normal saline vaporization. Next, the mice received methacholine-containing aerosols at increasing concentrations of 6.25, 12.5, 25, and 50 mg/mL for 3 min. The Penh data are presented as the relative increase in Penh after normalization with Penh after PBS inhalation. 2.9. BALF Collection and White Blood Cell Count BALF was collected through lung flushing with HBSS buffer containing 2% bovine serum albumin three times by using a trachea cannula (Angiocath, BD Biosciences). BALF cells were then collected through centrifugation at 300x g for 5 min and spanned on slides by using a cytocentrifuge (Thermo Fisher Scientific, Waltham, MA, USA) and then subjected to Liu's staining. Monocyte, lymphocyte, neutrophil, and eosinophil numbers within the cell pellet were then counted with a 100x objective lens. The data are presented as the average cell count in at least five fields for each sample. BALF eotaxin and IL-5 levels were determined by using commercial ELISA kits (BD Biosciences). 2.10. Patient Recruitment PAR was diagnosed in accordance with the definition of Saleh and Durham, i.e., symptoms lasting more than 4 days per week and illness duration lasting more than 4 weeks . In total, 156 patients aged 5-16 years old were recruited from the pediatric outpatient clinics of two hospitals (MacKay Memorial Hospital, Taipei, Taiwan and Chang Gung Children's Hospital, Taipei, Taiwan) for eligibility assessment in our clinical trial. However, 34 patients were excluded because they withdrew without signing the written informed consent form or because they fulfilled the exclusion criteria. The inclusion criteria were (i) age = 5-16 years; (ii) PAR for >=1 year; (iii) positive on any one of the following tests within 12 months: skin-prick test reaction (wheal size at least 3 mm larger than that made in the diluent control group) or allergic reaction examined by the methods of Pharmacia-CAP or multiple allergens simultaneous test; and (iv) mean total nasal symptoms score (TNSS) of no less than 5 throughout the screening period (at least 4 days) and TNSS on the day before day 0 (first dosing day) visit of no less than 5. Subjects were excluded if they (i) had clinically significant abnormalities in laboratory results as determined during 14 days prior to visit 1 or during the baseline period by the investigator; (ii) had acute or significant chronic sinusitis, severe persistent asthma, congenital immunodeficiency, neuropsychiatric disorders, immune-compromised massive wounds in the oral cavity, use of rhinitis medications, and chronic use of tricyclic antidepressants; (iii) need to take prohibited medications during the study or took the medications within 30 days prior to the screening visit, including parenteral or oral corticosteroids, nasal corticosteroids, topical flurandrenolide, topical clobetasol propionate, topical halobetasol propionate, astemizole, ketotifene, nedocromil or sodium cromoglycate, loratadine, cetirizine, antileukotrienes, other H1 antihistamines, nasal decongestant, or any food supplements including L. paracasei; (iv) were undergoing desensitization therapy within 3 months prior to the screening visit or with vasomotor rhinitis; (v) participated in an investigational drug trial within 4 weeks before entering this study; (vi) were pregnant, lactating, or planning to become pregnant; and (vii) had any other serious disease considered by the investigator not in the condition to enter the trial. The trial was approved by the Joint Institutional Review Board (c/o Taipei Veterans General Hospital, Taipei City, Taiwan) with the reference number 05-016-A on the date of 13 November 2006 and complied with the principles of the Declaration of Helsinki. The trial was registered on ISRCTN Registry (ISRCTN14829274, accessed on 28 October 2022). All the eligible patients and their parents or guardians were provided verbal and written information regarding the study and provided written informed consent. 2.11. Randomized, Double-Blind, Placebo-Controlled Trial Design The randomized, double-blind, placebo-controlled trial on PAR was designed in accordance with the guidance for developing drug products for AR treatment published by the Division of Pulmonary, Allergy, and Rheumatology Products in the Center for Drug Evaluation and Research at the Food and Drug Administration (April 2000 version). All patients who met the eligibility criteria were randomized to either the probiotic-treated or control group. In total, 137 patients were double-blinded and randomized to four groups, namely, one placebo group and three groups that received live GM-080 at different dosages: group A (2 x 108 colony forming units [CFU]/day); group B (2 x 109 CFU/day), and group C (1 x 1010 CFU/day). All treatments were administered for 3 months. Considering the key factors that could potentially influence treatment outcomes, randomization was stratified by age, sex, and AR severity. All probiotic capsules were supplied and stored at less than 4 degC with cGMP facilities. The demographic characteristics of the patients were collected at baseline by using questionnaires. The primary outcome was the change in AR severity after 3 months of intervention compared with the AR severity at baseline. The patients visited the hospital for data collection six times: at screening (V1) and in treatment weeks 0 (V2), 2 (V3), 4 (V4), 8 (V5), and 12 (V6). TNSS was used to evaluate the severity of main AR symptoms; here, the higher TNSS, the more severe the AR symptoms. Generalized estimating equations (GEEs) were used for a within-subject covariance structure evaluation. The Investigator Global Assessment Scale score at V3-V6 was evaluated by the trial investigator to assess the overall improvement of the participants after the treatment, which was divided into four levels: complete relief (4 points), partial relief (3 points), no relief (2 points), or worse (1 point). All parents were contacted 1 month after treatment to determine whether they observed a relapse after treatment interruption. Changes in severity scores in the groups were evaluated at each visit. The secondary outcomes were changes in total serum IgE and INF-g levels observed in months 0 and 3. Only 122 patients (28, 31, 31, and 32 in group A, group B, group C, and the placebo group, respectively) were included in the final analysis due to loss of follow-up. 2.12. Skin Prick Tests and Serum Biomarkers Skin prick tests were performed with commercial allergen extracts of egg, milk, crab, mite, cockroach, and animal dander (ALK-Abell & Oacute, Round Rock, TX, USA). Skin reactivity to allergen sensitization was classified into four grades as previously described . Serum total IgE levels were measured by the Division of Laboratory Medicine of MacKay Memorial Hospital (Taipei City, Taiwan) and the Division of Laboratory Medicine of Chang Gung Memorial Hospital, Taipei (Taipei City, Taiwan). Serum IFN-g levels were measured by using ELISA kits (BD Biosciences). 2.13. Statistical Analysis The data from in vitro and OVA-induced airway inflammation mouse model experiments were presented as means +- standard deviations and were analyzed for differences between groups by using one-way analysis of variance (ANOVA), followed by Tukey's Honestly Significant Difference test. The chi-square test was conducted to compare clinical symptoms. Intragroup comparisons for TNSSs and blood biomarkers at baseline and 3 months after treatment commencement were carried out by using paired t-tests. Intergroup comparisons among the four groups were performed by using ANOVA. The differences in TNSSs at the six visits among the four groups were also evaluated by applying a mixed model with adjustment for potential confounders. All children who completed the study were included in an intention-to-treat analysis regardless of their compliance. All tests assumed a two-sided alternative hypothesis with a significance level of 0.05. All analyses were conducted by using SAS (version 9.1; SAS Institute, Cary, NC, USA). 3. Results 3.1. GM-080 Induces Th1 Cytokine Production in Mouse Splenocytes The Th1 cytokines of IFN-g or IL-12 have been demonstrated to suppress AHR in response to allergens . We used mouse splenocytes as a cell model and evaluated Th1 cytokine (IFN-g and IL-12) production in the cell supernatant after L. paracasei exposure to identify the potential L. paracasei strains with antiallergic effects. By using ConA or LPS as the positive control , we observed that out of all the 23 L. paracasei strains that were screened, GM-080 demonstrated the strongest IFN-g and IL-12 induction activity, whereas BCRC 16100 demonstrated the lowest induction activity. We examined the Th1 cytokine induction ability of live GM-080 and the derived LTA or PGN in mouse splenocytes. The best IFN-g and IL-12 induction activity was demonstrated by live GM-080 at MOI = 10 and 0.1, respectively; IL-12 induction was greater because of live GM-080 than because of LPS . LTA and PGN could induce IFN-g production with the highest extent of induction at a concentration of 1 mg/mL . These data suggest that GM-080 has the potential to alleviate allergic conditions. 3.2. WGS Analysis Revealed That GM-080 Is a Safe L. paracasei Strain We performed WGS analysis on GM-080 by using next-generation sequencing. The genetic organization of GM-080 was illustrated in a genome atlas created by using Circos . One CRISPR locus, one bacteriocin gene, and six prophage-like clusters (Table S1) were observed in the GM-080 genome, with no plasmid being identified . We next examined the determinants of putative antibiotic resistance and virulence factors to evaluate the safety of GM-080 on the basis of the WGS analysis guidelines for probiotics safety provided by EFSA . Our in silico analysis demonstrated no putative antibiotic resistance or virulence factor genes in the GM-080 chromosome. We further examined the safety of GM-080 by using antimicrobial susceptibility profiling in accordance with the guidelines established by QSEA (ET ISO 10932). All MICs of the tested antibiotics to GM-080 were below the cutoffs suggested by EFSA (Table S2). These data suggest that GM-080 is safe for consumption as a probiotic. 3.3. GM-080 Genome Contains Immunosuppressive Motifs and CpG-Containing Oligonucleotides That Induce Th1 Cytokines in Mouse Splenocytes We compared the genomes of GM-080 and BCRC 16100 to identify the potential functional genes in GM-080. As indicated by the phylogenetic tree , GM-080 was clustered with BCRC 16100 ; however, BCRC 16100 displayed a lower immunomodulation capability than GM-080 . Table 1 shows that the functional categories of genes in GM080 and BCRC 16100 were highly similar. The unique genes in GM-080 are summarized in Table S3. Immunosuppressive motifs (IMs) in Lactobacillus spp. may participate in ameliorating allergic conditions through their anti-inflammatory activity . In addition, the activation of toll-like receptor 9 by CpG-containing oligonucleotides (ODNs) can shift the immune dominance from Th2 to Th1, which may also facilitate the alleviation of allergic diseases . We analyzed well-known IM and CpG ODNs (Table S4) in two strains and next synthesized eight candidate IM sequences in accordance with the genome sequences of GM-080 (IM4-IM7) and BCRC 16100 (IM8-IM11) and used them to stimulate mouse splenocytes. Although none of the IMs induced IFN-g production, IM5 from the GM-080 genome exhibited capability for IL-12 induction activity (Table 2). We also synthesized six CpG-containing ODNs on the basis of the genome analysis of GM-080 (ODN1 or ODN2) and BCRC 16100 (ODN3-ODN6) and used them to stimulate mouse splenocytes. None of the BCRC 16100 ODNs exhibited IL-12 induction activity, but both ODNs from GM-080 (ODN1 and ODN2) induced IFN-g and IL-12 secretion (Table 2). Therefore, we hypothesized that among L. paracasei strains, GM-080 would be a probiotic with antiallergic activity. 3.4. Orally Gavage GM-080 Alleviates OVA-Induced Allergic Airway Inflammation in Mice We established an OVA-induced AHR mouse model by using the protocol shown in Figure 4A. Oral GM-080 treatment was administered from days 7 to 63 at the dose of either 2 x 106 CFU/mouse/day (low-dose group) or 1 x 107 CFU/mouse/day (high-dose group). Blood samples were collected on days 63 (4 days before intranasal challenge with OVA) and 70 (the day of sacrifice). After the intranasal challenge with OVA, the Penh value elevated as the methacholine dose was increased, indicating the occurrence of AHR. Oral GM-080 alleviated AHR in a dose-dependent manner . We also observed that the oral administration of high-dose GM-080 significantly reduced the numbers of eosinophils, neutrophils, lymphocytes, and monocyte numbers in BALF, whereas that of low-dose GM-080 significantly reduced the numbers of neutrophils and monocytes . The expression of eotaxin in BALF was reduced by oral GM-080 . In addition, the IL-5 levels in BALF were reduced by oral GM-080 at low and high dosages . Next, we found that oral GM-080 at a low dose significantly reduced anti-OVA IgE and anti-OVA IgG2a levels, whereas that at a high dose significantly reduced only anti-OVA IgG2a levels but not anti-OVA IgG1 levels . Moreover, oral GM-080 reduced ConA- and OVA- induced IL-5 production in mouse splenocytes. These data demonstrate that the oral administration of GM-080 can ameliorate allergic airway inflammation in OVA-induced AHR mice. 3.5. GM-080 Alleviates PAR in Children Finally, we conducted a double-blind, randomized, placebo-controlled trial to investigate the beneficial effects of GM-080 in children with PAR. The study design was summarized in Figure 7 and their baseline demographic characteristics were presented in Table 3; no significant differences were noted among the patient groups. We first evaluated the effects of probiotics on AR symptom severity scores. Only the sneezing subscale scores significantly decreased, particularly in group B (in which 2 x 109 CFU/day GM-080 was administered; Table 4). Scores for other symptoms such as rhinorrhea, nasal pruritus, and nasal congestion did not improve in groups A and C compared with the placebo group. By contrast, these scores significantly improved in group B over time (Table 4). We next examined the effects of GM-080 on patients' quality of life by using TNSSs. TNSSs are a convenient tool for symptom description and the assessment of functional problems (physical, emotional, social, and occupational) associated with AR. We noted a nonsignificant decrease in TNSSs over time in groups A, B, and C (Table 4). In our GEE model, TNSSs after the five visits among our four groups exhibited no significant differences (Table 5). Investigator Global Assessment Scale scores were higher in the groups that were administered GM-080 than in the placebo group after treatment (Table 6, p = 0.049). We also examined the effects of GM-080 administration on serum IgE production, skin sensitization, and serum IFN-g levels at baseline and at the end of treatment (month 3). No changes in sera total IgE levels were observed either at baseline or at the end of treatment after 12 weeks among the groups (Table 7). We observed an increasing trend in IFN-g levels of GM-080 consumption groups with middle and high doses (group B and group C, respectively) at the end of treatment compared to the placebo or low dose group (group A) (Table 7). However, the increases did not reach statistical differences. These data indicate the beneficial effects of GM-080 consumption in pediatric AR; however, these effects are not dosage-dependent but time-dependent. Additional well-designed clinical trials are warranted to identify the most effective dosage of GM-080. 4. Discussion Lactobacillus spp. alleviate allergic diseases through different mechanisms. In mice, the oral administration of Lactobacillus reuteri for 9 days increased CD4+CD25+FoxP3+ regulatory T (Treg) cell numbers in the spleen; moreover, the adoptive transfer of these Treg cells from L. reuteri-treated mice reduced airway inflammation induced by antigen challenge . Zhong et al. demonstrated that in OVA-sensitized rats, the administration of a mixture of probiotic genomic DNA derived from L. rhamnosus GG (LGG) and Bifidobacterium longum BB536 or that of a synthetic CpG-ODN reduced the production of Th2 cytokines and increased the Treg cell population in the spleen or mesenteric lymph nodes on the basis of increased toll-like receptor 9/nuclear factor kappa B activity . CpG-ODN and LGG DNA containing TTTCGTTT, which is an IM, also demonstrated antiallergic potential in mice; specifically, they downregulated OVA-specific IgE production and increased systemic Th1 responses . In the current study, the predicted IMs from BCRC 16100 did not display any Th1 cytokine induction capabilities and only IM5 in GM-080 exhibited IL-12 induction activity in mouse splenocytes (Table 2). Although the frequency of IMs in the genomes of probiotics may be associated with their antiallergic potential , the observed immunomodulatory activities of the predicted IMs warrant experimental examination. In the present study, we could not predict the relationship between the antiallergic activity of GM-080 and the associated changes in Th1 cytokine induction in vitro or the reduction in Th2 cytokine levels in vivo . Given the Th1 cytokine induction activity of GM-080 IM5 and GM-080 ODN1 and ODN2 (Table 2), CpG-ODN and IMs in the GM-080 genome may underlie the beneficial effects of GM-080 against AHR. Considering their IFN-g induction activity in vitro , LTA and PGN, two cell-wall components of GM-080, may be the key active ingredients involved in the improvement of AHR. Li et al. reported that PGN from Lactobacillus acidophilus inhibited IgE production and regulated Treg-Th17 balance, thus preventing b-lactoglobulin allergy . Mat et al. indicated that treatment with LTA from Staphylococcus aureus reduced IL-5 production in peripheral blood mononuclear cells from patients with asthma , suggesting that LTA might have AHR-alleviating effects. However, the detailed molecular mechanisms underlying IFN-g upregulation due to GM-080-derived LTA or PGN warrant further investigation. In our WGS of GM-080 and BCRC 16100 (Table 1 and Table S3), the undecaprenyl-phosphate galactose phosphotransferase gene (rfbp) and alpha-D-GlcNAc alpha-1,2-L-rhamnosyltransferase gene (rgpAc) involved in exopolysaccharide synthesis were noted only in GM-080. Wu et al. reported that exopolysaccharides from B. longum BCRC 14634 suppressed LPS-induced TNF-a production in J774A.1 macrophages . Whether the immunomodulatory effect of GM-080-derived exopolysaccharides is responsible for the antiallergic activity of GM-080 remains unclear. In Treg cells, glutathione (GSH) loss can reduce suppressor function, thus inducing multiorgan autoimmunity. GSH-enriched yeast has been demonstrated to alleviate CCl4-induced liver damage in rats. In our analysis of the GM-080 genome, we noted the presence of pepA, an aminopeptidase that may be involved in GSH biosynthesis . Considering that oxidative stress is a hallmark of asthma , GM-080 may be a GSH-enriched Lactobacillus that can improve the allergic condition. In our in vitro analysis of the effects of GM-080 on Th1 cytokine (IL-12 and IFN-g) production, the induction effect was unaffected by the MOI . This result indicates the presence of a therapeutic window of GM-080 dosage for the alleviation of airway allergic conditions. One animal study even confirmed that for GM-080, immunoregulatory function is determined by dehydrogenase activity . In particular, the effectiveness of probiotics in the human environment may be affected by factors related to the internal microbial ecosystem, including intestinal colonization duration and changes in intestinal microbiota, after the consumption of probiotics . The optimization of GM-080 dosage based on AHR severity warrants further investigation. In the OVA-induced asthma mouse model, we found that the oral consumption of GM-080 at the sensitization phase could reduce anti-OVA specific IgE at low doses or anti-OVA specific IgG2a at low and high doses but did not change anti-OVA specific IgG1 levels . As a result of the general role of IgG1 or IgG2a for representing Th2 or Th1 responses, respectively , the data on unchanged anti-OVA specific IgG1 may not reflect the downregulation of Th2 responses by GM-080. Thus, we further examined OVA-induced IL-5 production after GM-080 consumption and found that it could be reduced at both doses of GM-080 . Indeed, several studies related to the improvement in OVA-induced asthma in mouse models also reported the suppression of anti-OVA specific IgG2a . The data of reducing serum OVA-specific-IgE and OVA-induced IL-5 production by mouse splenocytes in this study suggests that GM-080, at an optimal dose, could inhibit Th2 responses. In the current trial, the single live strain GM-080 was used, and its PAR-alleviating effect appeared to be significant only in terms of symptom (sneezing) relief (Table 4) and quality of life improvement (Table 6). Moreover, the effects of GM-080 on symptom (sneezing) relief were not dosage-dependent. In the present study, the elevated serum IFN-g levels at the final visit were noted only in the patients who consumed a moderate amount of GM-080 at the dose of 2 x 109 CFU per day (Table 7), indicating that IFN-g induction may serve as a predictive factor for the selection of antiallergic probiotics. Lin et al. reported that in children aged 6-12 years, treatment with L. paracasei HF.A00232 at 5 x 109 CFU/capsule combined with 5 mg of levocetirizine significantly ameliorated AR symptoms, including sneezing, itching nose, and swollen, puffy eyes . Therefore, the effects of GM-080 in combination with antihistamines on AR symptoms should be investigated. 5. Conclusions L. paracasei GM-080 induced Th1 cytokine production in mouse splenocytes and improved airway inflammation in an OVA-induced asthma mouse model with Th2 cytokine downregulation. This prospective, double-blind, placebo-controlled, randomized clinical trial on PAR showed the ameliorating effects of GM-080 on symptoms, such as sneezing, TNSS, and Investigator Global Assessment Scale score. Therefore, when given as a food supplement, GM-080 can alleviate AR in children. Acknowledgments We thank for Shyh-Dar Shyur at Makay Memorial Hospital (Taipei, Taiwan) and Dah-Chin Yan at Chang Gung Memorial Hospital (Taoyuan, Taiwan) for their assistance with clinical data collection. Supplementary Materials The following supporting information can be downloaded at: Table S1: CRISPR loci, prophage-like clusters, and bacteriocin gene in the GM-080 genome; Table S2: Minimum inhibitory concentrations (MICs) of L. paracasei GM-080 toward eight antimicrobials and their microbiological cutoffs; Table S3: EggNOG functional annotations of L. paracasei GM-080; Table S4: Putative IM and CpG-containing ODN contents in L. paracasei strains. Click here for additional data file. Author Contributions W.-H.T. designed the in vitro and animal studies. C.-H.C. performed the in vitro and animal studies. J.-H.J. performed the WGS analysis. W.-H.T. and I.-J.W. analyzed the data. E.-K.L., W.-W.C., W.-H.T. and I.-J.W. wrote the article. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board of Taipei Veterans General Hospital (Taipei City, Taiwan) with a reference number of 05-016-A at the date of 13 November 2006, and it was registered on ISRCTN registry (ISRCTN14829274, accessed on 28 October 2022). The protocols of animal experiments were approved by the Animal Care and Utilization Committee of GenMont Biotech Incorporation, Inc. (Taiwan IACUC Approval No: 194; trial no: 104010 and 105001). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient(s) to publish this paper. Data Availability Statement The whole genome sequencing data of BCRC 16100 were available in GenBank(r) (National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, USA) with the accession number of GCA_022588775.1. The nucleotide sequences of GM-080 have been submitted into BioProject in NCBI with an accession number of PRJNA824946. Conflicts of Interest Taipei Hospital, Ministry of Health and Welfare Authority has a contract with GenMont Biotech, which provided funds for probiotics. W.H.T. and C.H.C. are the employee in GenMont Biotech. Figure 1 Th1 cytokine induction by GM-080 in mouse splenocytes. (A) List of L. paracasei strains used for coculture with mouse splenocytes. (B,C) Different strains of viable L. paracasei were used to coculture with mouse splenocytes at MOI = 10 for 48 h. The presence of IFN-g (B) and IL-12 (C) in the culture supernatants was determined using commercial ELISA kits. LPS (1 mg/mL) or ConA (2 mg/mL) was used as the positive control. Figure 2 Effects of GM-080 preparations and cell wall compartments in Th1 cytokines production of mouse splenocytes. We seeded 4 x 105 mouse splenocytes into a 48-well plate, followed by treatment with viable GM-080 preparation with the indicated MOI or cell wall compartments purified from GM-080 (LTA and PGN) for 48 h. The culture supernatants were then harvested, and INF-g (A-D) or IL-12 (E-H) production was determined using commercial ELISA kits. LPS (1 mg/mL) or ConA (1 mg/mL) was used as the positive control. N.D., not detected. Figure 3 WGS analysis of GM-080. (A) Circular genome graph of GM-080 illustrated using Circos. (B) Summary of general features of GM-080 genome. (C) Maximum-likelihood tree of GM-080 and BCRC 16100 with 32 other known Lactobacillus strains. Figure 4 Effects of oral GM-080 on respiratory resistance in our OVA-induced AHR mouse model. (A) Illustration of the experimental procedure used for induction of OVA-induced AHR mouse model. Two dosages of GM-080 (LP-L, 2 x 106 CFU/mouse/day; LP-H, 1 x 107 CFU/mouse/day) were force-fed to mice through gavage. (B) Respiratory resistance was evaluated on the basis of Penh values under the inhalation of the indicated concentration of methacholine. NC, nontreated control; H2O, H2O gavage. * p < 0.05 compared with the H2O group. Figure 5 Effects of oral GM-080 on immune cell infiltration and eosinophilic cytokines in lungs of mice with AHR. BALF was collected 2 days after the final OVA challenge followed by centrifugation, and the cell pellet was collected. (A-D) Eosinophil (A), neutrophil (B), lymphocyte (C), and monocyte (D) infiltration was evaluated through microscopy after staining cells with Liu's stain. (E,F) Eotaxin (E) or IL-5 (F) levels in BALF was determined using commercial ELISA kits. NC, nontreated control; H2O, H2O gavage; LP-L, GM-080 at 2 x 106 CFU/mouse/day through gavage; LP-H, GM-080 at 1 x 107 CFU/mouse/day through gavage. * p < 0.05, ** p < 0.01. Figure 6 Effects of GM-080 consumption in OVA-specific antibodies or OVA-induced IL-5 production in an OVA-induced AHR mouse model. (A-C) Mice sera were collected 2 days after final theOVA challenge, and the presence of OVA-specific IgE (A), IgG1 (B), and IgG2a (C) was determined using ELISA. (D-F) Splenocytes were collected 2 days after the final OVA challenge and then seeded into a 48-well-plate at a density of 5 x 106 cells/well to determine IL-5 production without treatment (D) or after treatment with 2 mg/mL ConA (E) or 100 mg/mL OVA (F). NC, nontreated control; H2O, H2O gavage; LP-L, GM-080 at 2 x 106 CFU/mouse/day through gavage; LP-H, GM-080 at 1 x 107 CFU/mouse/day through gavage. * p < 0.05, ** p < 0.01. Figure 7 Enrollment of children with AR in our double-blind, randomized, placebo-controlled trial. The inclusion and exclusion criteria are described in the Section 2. cells-12-00768-t001_Table 1 Table 1 COG functional categories of GM-080 and BCRC 16100. COG Functional Categories a Same Gene No. Unique Genes in GM-080 (Total No.) Unique Genes in BCRC 16100 (Total No.) Cell wall/membrane/envelope biogenesis (M) 72 rfbP, rgpAc, mprF (3) kdsD, tuaG, tagF, ywqC, tagE, gtf1 (6) Replication, recombination, and repair (L) 89 recT, pi112, tnpA1, tnp1216, cas2, cas1, cas9, int3, pi346, tnpR, is18, yqaJ, rusA (13) - Posttranslational modification, protein turnover, and chaperones (O) 44 gst (1) - Carbohydrate transport and metabolism (G) 139 agaD, kduI, kdgK, ahaA, xylP, lacE, lacG, lacF (8) pts32BC, gatY, mnaA (3) Amino acid transport and metabolism (E) 132 dppA, pepA, yxeO (3) - Coenzyme transport and metabolism (H) 53 - pdxA (1) Inorganic ion transport and metabolism (P) 92 feoA, ytmL (2) kdgT, sfuB, fbpC (3) Secondary metabolites biosynthesis, transport, and catabolism (Q) 15 kduD (1) - Intracellular trafficking, secretion, and vesicular transport (U) 36 chaT1, clpP (2) secY2, secA2 (2) a COG, Clusters of Orthologous Gene. The analysis of COG categories was done by National Center for Biotechnology Information access on 1 November 2021). cells-12-00768-t002_Table 2 Table 2 Immunoregulatory effects of putative IM and CpG-containing ODN a in GM-080 and BCRC 16100. Code Core Sequence Strain Sequence IFN-r (pg/mL) b IL-12 (pg/mL) c IM3 TCAAGCTTGA TCAAGCTTGA ND d ND IM4 TCAAGCTTGA GM-080 CAAGCGTCAAGCTTGAATGA ND ND IM5 GM-080 AAAAATTCAAGCTTGATAGT ND 609.8 +- 227.69 IM6 GM-080 CCATCGTCAAGCTTGACTTG ND ND IM7 GM-080 CCCTAATCAAGCTTGATTAA ND ND IM8 BCRC 16100 GCAGCTTCAAGCTTGAAAAA ND ND IM9 BCRC 16100 CCGGCCTCAAGCTTGAATTG ND ND IM10 BCRC 16100 TTTCATTCAAGCTTGACGCT ND ND IM11 BCRC 16100 CCTTAATCAAGCTTGATTAG ND ND ODN1 GACGATCGTC GM-080 GCTTGACGATCGTCTCTGGA ND 38.8 +- 65.2 ODN2 ACGACGTCGT GM-080 GGTCACGACGTCGTTTACAAA 490 +- 272.74 46.1 +- 32.88 ODN3 GACGATCGTC BCRC 16100 AATTGACGATCGTCTAATTC ND ND ODN4 BCRC 16100 TGTCGACGATCGTCGTCTGT ND ND ODN5 BCRC 16100 CAGAGACGATCGTCAAGCGA 77.381 +- 5.4 ND ODN6 ACGACGTCGT BCRC 16100 CGTCACGACGTCGTGACCGGC ND ND a ODN, oligonucleotides. b Treatment concentration was 0.125 mM. Data are presented as means +- standard deviations. c Treatment concentration was 1 mM. Data are presented as means +- standard deviations. d ND, not detected. cells-12-00768-t003_Table 3 Table 3 Demographic and baseline characteristics of patients a. Characteristics A Group (n = 28) B Group (n = 31) C Group (n = 31) PLC Group (n = 32) p Male b 20 (71.4%) 18 (58.1%) 18 (58.1%) 21 (65.6%) 0.660 Age (year) c 8.25 (2.78) 8.74 (2.77) 8.42 (2.55) 8.56 (2.95) 0.917 Sneezing c 1.69 (0.77) 1.61 (0.67) 1.49 (0.79) 1.75 (0.64) 0.527 Rhinorrhea c 1.83 (0.55) 1.81 (0.68) 1.69 (0.70) 1.95 (0.60) 0.435 Nasal pruritus c 1.73 (0.70) 1.72 (0.53) 1.69 (0.67) 1.70 (0.79) 0.997 Nasal congestion c 2.10 (0.63) 1.97 (0.66) 2.04 (0.71) 2.04 (0.61) 0.880 TNSS c 7.34 (1.44) 7.11 (1.74) 6.92 (1.57) 7.44 (1.56) 0.559 Total serum IgE (kU/L) c 652.58 (1366.19) 547.66 (530.99) 614.81 (661.08) 405.68 (384.19) 0.642 Combine with asthma b 11 (39.3%) 14 (45.2%) 11 (35.3%) 12 (37.5%) 0.878 Combine with atopic dermatitis b 12 (42.9%) 11 (35.5%) 8 (25.8%) 11 (34.4%) 0.590 Combine with conjunctivitis b 5 (17.9%) 6 (19.4%) 5 (16.1%) 9 (28.1%) 0.648 Allergic sensitization b Mite (Df) 23 (82.1%) 25 (80.6%) 20 (64.5%) 25 (78.1%) 0.350 Mite (Dp) 26 (92.9%) 30 (96.8%) 28 (90.3%) 31 (96.9%) 0.624 Cockroach 6 (21.4%) 7 (22.6%) 10 (32.3%) 5 (15.6%) 0.470 Animal dander (Cats) 6 (21.4%) 4 (12.9%) 7 (22.6%) 7 (21.9%) 0.749 Animal dander (Dogs) 7 (25.0%) 6 (19.4%) 7 (22.6%) 9 (28.1%) 0.869 Mold 1 (3.6%) 2 (6.5%) 2 (6.5%) 1 (3.1%) 0.887 a Severity of each symptom was measured on a 4-point scale: 0 = absent; 1 = mild; 2 = moderate; 3 = severe. Scores for each symptom were added to obtain the TNSS. b Data are presented as n (%). c Data are presented as mean (standard deviation). cells-12-00768-t004_Table 4 Table 4 AR symptom scores at baseline and follow-up visits among enrolled patients #. Subscale Examination A Group (n = 28) B Group (n = 31) C Group (n = 31) PLC Group (n = 32) p Sneezing Visit 2 1.69 (0.77) 1.61 (0.67) 1.49 (0.79) 1.75 (0.64) 0.527 Visit 3 1.61 (0.71) 1.20 (0.57) a 1.43 (0.83) 1.67 (0.67) 0.045 e Visit 4 1.68 (0.75) 1.19 (0.70) a 1.41 (0.83) 1.40 (0.65) 0.090 Visit 5 1.51 (0.74) 1.02 (0.66) a 1.35 (0.96) 1.50 (0.66) 0.049 e Visit 6 1.33 (0.76) 1.08 (0.68) a 1.06 (0.82) ab 1.56 (0.82) 0.033 e Rhinorrhea Visit 2 1.83 (0.55) 1.81 (0.68) 1.69 (0.70) 1.95 (0.60) 0.435 Visit 3 1.71 (0.67) 1.73 (0.83) 1.51 (0.86) 1.95 (0.75) 0.170 Visit 4 1.69 (0.71) 1.42 (0.80) a 1.39 (0.76) 1.65 (0.66) 0.252 Visit 5 1.37 (0.73) 1.20 (0.72) ab 1.24 (1.00) 1.59 (0.69) 0.203 Visit 6 1.42 (0.81) 1.11 (0.67) ab 1.24 (0.99) 1.44 (0.75) 0.320 Nasal pruritus Visit 2 1.73 (0.70) 1.72 (0.53) 1.69 (0.67) 1.70 (0.79) 0.997 Visit 3 1.54 (0.83) 1.42 (0.73) 1.35 (0.67) 1.65 (0.80) 0.415 Visit 4 1.43 (0.81) 1.34 (0.78) a 1.24 (0.59) a 1.51 (0.72) 0.466 Visit 5 1.24 (0.64) a 1.09 (0.64) a 1.11 (0.78) a 1.47 (0.75) 0.123 Visit 6 1.08 (0.77) a 1.08 (0.80) a 1.09 (0.75) a 1.33 (0.89) 0.520 Nasal congestion Visit 2 2.10 (0.63) 1.97 (0.66) 2.04 (0.71) 2.04 (0.61) 0.880 Visit 3 1.87 (0.74) 1.87 (0.82) 1.86 (0.88) 2.07 (0.76) 0.663 Visit 4 1.87 (0.81) 1.60 (0.78) 1.72 (0.64) 1.84 (0.72) 0.476 Visit 5 1.33 (0.71) ab 1.35 (0.88) a 1.47 (0.89) a 1.77 (0.84) 0.154 Visit 6 1.29 (0.73) abc 1.15 (0.88) abc 1.19 (0.90) abc 1.48 (0.84) a 0.414 TNSS Visit 2 7.34 (1.44) 7.11 (1.74) 6.92 (1.57) 7.44 (1.56) 0.559 Visit 3 6.74 (2.01) 6.23 (2.21) 6.15 (2.51) 7.35 (2.24) 0.135 Visit 4 6.68 (2.45) 5.54 (2.50) a 5.77 (1.95) a 6.40 (2.05) 0.174 Visit 5 5.44 (2.13) a 4.67 (2.41) ab 5.18 (2.97) a 6.33 (2.36) 0.066 Visit 6 5.12 (2.44) abc 4.42 (2.39) ab 4.57 (3.00) abc 5.56 (2.49) 0.286 # Severity of each symptom was measured on a 4-point scale: 0 = absent; 1 = mild; 2 = moderate; 3 = severe. Scores for each symptom were added to obtain the TNSS. Data were presented as mean (standard deviation). a p < 0.05, intragroup comparisons (visit 6 vs. visit 2, visit 5 vs. visit 2, visit 4 vs. visit 2, visit 3 vs. visit 2 in each group). b p < 0.05, intragroup comparisons (visit 6 vs. visit 3, visit 5 vs. visit 3, visit 4 vs. visit 3). c p < 0.05, intragroup comparisons (visit 6 vs. visit 4, visit 5 vs. visit 4). e p < 0.05, intragroup comparisons (difference between the four groups). cells-12-00768-t005_Table 5 Table 5 Differences in sneezing, rhinorrhea, nasal pruritus, nasal congestion, and TNSS among the four groups after five visits based on the GEE model. A Group (n = 28) B Group (n = 31) C Group (n = 31) PLC Group (n = 32) p Subscale n Value (95% CI a) n Value (95% CI) n Value (95% CI) n Value (95% CI) Sneezing 28 0.028 (-0.321, 0.377) 31 -0.229 (-0.536, 0.077) 31 -0.250 (-0.604, 0.104) 32 Referent 0.041 b Rhinorrhea 28 -0.014 (-0.404, 0.376) 31 -0.332 (-0.677, 0.012) 31 -0.202 (-0.627, 0.223) 32 Referent 0.104 Nasal pruritus 28 -0.258 (-0.671, 0.155) 31 -0.250 (-0.662, 0.161) 31 -0.242 (-0.643, 0.159) 32 Referent 0.414 Nasal congestion 28 -0.192 (-0.583, 0.199) 31 -0.330 (-0.750, 0.089) 31 -0.289 (-0.713, 0.136) 32 Referent 0.441 TNSS 28 -0.436 (-1.663, 0.791) 31 -1.142 (-2.328, 0.044) 31 -0.983 (-2.324, 0.357) 32 Referent 0.070 a CI, confidence interval. b p < 0.05, intergroup comparisons (difference between four groups). cells-12-00768-t006_Table 6 Table 6 Investigator Global Assessment Scale scores at baseline and follow-up visits among the four groups a. Subscale Examination A Group (n = 28) B Group (n = 31) C Group (n = 31) PLC Group (n = 32) p-Value 4 Groups Global Assessment Visit 3 2.43 (0.88) 2.42 (0.62) 2.58 (0.56) 2.47 (0.72) 0.795 Visit 4 2.61 (0.63) 2.61 (0.67) 2.71 (0.69) 2.44 (0.76) 0.471 Visit 5 3.00 (0.61) bc 2.94 (0.81) b 2.77 (0.76) 2.53 (0.84) 0.083 Visit 6 2.96 (0.74) c 2.97 (0.71) b 3.16 (0.86) d 2.59 (0.91) 0.049 e a Data are presented as mean (standard deviation). b p < 0.05, intragroup comparisons (visit 6 vs. visit 3, visit 5 vs. visit 3, visit 4 vs. visit 3 in each group). c p < 0.05, intragroup comparisons (visit 6 vs. visit 4, visit 5 vs. visit 4). d p < 0.05, intragroup comparisons (visit 6 vs. visit 5). e p < 0.05, intragroup comparisons (difference between the four groups). cells-12-00768-t007_Table 7 Table 7 Serum biomarker and sensitization levels at baseline and at the end of treatment #. Biomarker Examination A group (n = 28) B group (n = 31) C group (n = 31) PLC group (n = 32) p Total IgE (kU/L) a Baseline 652.58 (1366.19) 547.66 (530.99) 614.81 (661.08) 405.68 (384.19) 0.642 Visit 6 648.73 (1466.52) 449.69 (469.01) 599.39 (677.36) 419.52 (529.22) 0.833 IFN-g (ng/mL) a Baseline 436.96 (366.48) 494.73 (760.74) 302.30 (442.22) 383.25 (498.97) 0.331 Visit 6 467.10 (915.50) 1042.03 (3006.39) b 740.28 (2020.84) 515.89 (1830.42) 0.480 # Measurements were performed in month 3. a p < 0.05, intragroup comparisons (visit 6 vs. baseline in each group). b Data are presented as means (standard deviations). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Lipworth B.J. White P.S. Allergic inflammation in the unified airway: Start with the nose Thorax 2000 55 878 881 10.1136/thorax.55.10.878 10992543 2. Cohn L. Homer R.J. MacLeod H. Mohrs M. Brombacher F. Bottomly K. Th2-induced airway mucus production is dependent on IL-4Ralpha, but not on eosinophils J. Immunol. 1999 162 6178 6183 10.4049/jimmunol.162.10.6178 10229862 3. Cohn L. Homer R.J. Marinov A. Rankin J. Bottomly K. Induction of airway mucus production By T helper 2 (Th2) cells: A critical role for interleukin 4 in cell recruitment but not mucus production J. Exp. Med. 1997 186 1737 1747 10.1084/jem.186.10.1737 9362533 4. Brusselle G.G. Kips J.C. Tavernier J.H. van der Heyden J.G. Cuvelier C.A. Pauwels R.A. Bluethmann H. Attenuation of allergic airway inflammation in IL-4 deficient mice Clin. Exp. Allergy 1994 24 73 80 10.1111/j.1365-2222.1994.tb00920.x 8156448 5. Hogan S.P. Koskinen A. Foster P.S. Interleukin-5 and eosinophils induce airway damage and bronchial hyperreactivity during allergic airway inflammation in BALB/c mice Immunol. Cell Biol. 1997 75 284 288 10.1038/icb.1997.43 9243294 6. Taube C. Duez C. Cui Z.H. Takeda K. Rha Y.H. Park J.W. Balhorn A. Donaldson D.D. Dakhama A. Gelfand E.W. The role of IL-13 in established allergic airway disease J. Immunol. 2002 169 6482 6489 10.4049/jimmunol.169.11.6482 12444158 7. Strober B.E. Why Biologic Therapies Sometimes Lose Efficacy Semin. Cutan. Med. Surg. 2016 35 S78 S80 10.12788/j.sder.2016.022 27525443 8. McFarland L.V. From yaks to yogurt: The history, development, and current use of probiotics Clin. Infect. Dis. 2015 60 S85 S90 10.1093/cid/civ054 25922406 9. Das R.R. Singh M. Shafiq N. Probiotics in treatment of allergic rhinitis World Allergy Organ. J. 2010 3 239 244 10.1097/WOX.0b013e3181f234d4 23282801 10. Fiocchi A. Pawankar R. Cuello-Garcia C. Ahn K. Al-Hammadi S. Agarwal A. Beyer K. Burks W. Canonica G.W. Ebisawa M. World Allergy Organization-McMaster University Guidelines for Allergic Disease Prevention (GLAD-P): Probiotics World Allergy Organ. J. 2015 8 4 10.1186/s40413-015-0055-2 25628773 11. Sharma G. Im S.H. Probiotics as a Potential Immunomodulating Pharmabiotics in Allergic Diseases: Current Status and Future Prospects Allergy Asthma Immunol. Res. 2018 10 575 590 10.4168/aair.2018.10.6.575 30306743 12. Voo P.Y. Wu C.T. Sun H.L. Ko J.L. Lue K.H. Effect of combination treatment with Lactobacillus rhamnosus and corticosteroid in reducing airway inflammation in a mouse asthma model J. Microbiol. Immunol. Infect. 2022 55 766 776 10.1016/j.jmii.2022.03.006 35487817 13. Lan H. Gui Z. Zeng Z. Li D. Qian B. Qin L.Y. Dai L. Song J.L. Oral administration of Lactobacillus plantarum CQPC11 attenuated the airway inflammation in an ovalbumin (OVA)-induced Balb/c mouse model of asthma J. Food Biochem. 2022 46 e14036 10.1111/jfbc.14036 34981513 14. Yan S. Ai S. Huang L. Qiu C. Zhang F. He N. Zhuang X. Zheng J. Systematic review and meta-analysis of probiotics in the treatment of allergic rhinitis Allergol. Immunopathol. (Madr) 2022 50 24 37 10.15586/aei.v50i3.507 35527653 15. Forsythe P. Inman M.D. Bienenstock J. Oral treatment with live Lactobacillus reuteri inhibits the allergic airway response in mice Am. J. Respir. Crit. Care Med. 2007 175 561 569 10.1164/rccm.200606-821OC 17204726 16. Nawaz M. Ma C. Basra M.A. Wang J. Xu J. Amelioration of ovalbumin induced allergic symptoms in Balb/c mice by potentially probiotic strains of lactobacilli Benef. Microbes 2015 6 669 678 10.3920/BM2014.0141 25869278 17. Yan D.C. Hung C.H. Sy L.B. Lue K.H. Shih I.H. Yang C.Y. Chen L.C. Sun H.L. Lee M.S. Chambard J. A Randomized, Double-Blind, Placebo-Controlled Trial Assessing the Oral Administration of a Heat-Treated Lactobacillus paracasei Supplement in Infants with Atopic Dermatitis Receiving Topical Corticosteroid Therapy Skin Pharmacol. Physiol. 2019 32 201 211 10.1159/000499436 31117100 18. Costa D.J. Marteau P. Amouyal M. Poulsen L.K. Hamelmann E. Cazaubiel M. Housez B. Leuillet S. Stavnsbjerg M. Molimard P. Efficacy and safety of the probiotic Lactobacillus paracasei LP-33 in allergic rhinitis: A double-blind, randomized, placebo-controlled trial (GA2LEN Study) Eur. J. Clin. Nutr. 2014 68 602 607 10.1038/ejcn.2014.13 24569538 19. Ahmed M. Billoo A.G. Iqbal K. Efficacy of probiotic in perennial allergic rhinitis under five year children: A randomized controlled trial Pak. J. Med. Sci. 2019 35 1538 1543 10.12669/pjms.35.6.744 31777489 20. Tsai W.H. Chou C.H. Chiang Y.J. Lin C.G. Lee C.H. Regulatory effects of Lactobacillus plantarum-GMNL6 on human skin health by improving skin microbiome Int. J. Med. Sci. 2021 18 1114 1120 10.7150/ijms.51545 33526970 21. Additives E.P.o. Feed P.o.S.u.i.A. Guidance on the assessment of bacterial susceptibility to antimicrobials of human and veterinary importance EFSA J. 2012 10 2740 22. Jhong J.H. Tsai W.H. Yang L.C. Chou C.H. Lee T.Y. Yeh Y.T. Huang C.H. Luo Y.H. Heat-Killed Lacticaseibacillus paracasei GMNL-653 Exerts Antiosteoporotic Effects by Restoring the Gut Microbiota Dysbiosis in Ovariectomized Mice Front. Nutr. 2022 9 804210 10.3389/fnut.2022.804210 35187034 23. Zimin A.V. Puiu D. Luo M.C. Zhu T. Koren S. Marcais G. Yorke J.A. Dvorak J. Salzberg S.L. Hybrid assembly of the large and highly repetitive genome of Aegilops tauschii, a progenitor of bread wheat, with the MaSuRCA mega-reads algorithm Genome Res. 2017 27 787 792 10.1101/gr.213405.116 28130360 24. Seppey M. Manni M. Zdobnov E.M. BUSCO: Assessing Genome Assembly and Annotation Completeness Methods Mol. Biol. (Clifton N.J.) 2019 1962 227 245 10.1007/978-1-4939-9173-0_14 25. Seemann T. Prokka: Rapid prokaryotic genome annotation Bioinformatics (Oxford Engl.) 2014 30 2068 2069 10.1093/bioinformatics/btu153 26. European Food Safety A. EFSA statement on the requirements for whole genome sequence analysis of microorganisms intentionally used in the food chain EFSA J. 2021 19 e06506 10.2903/j.efsa.2021.6506 34335919 27. Liu B. Zheng D. Jin Q. Chen L. Yang J. VFDB 2019: A comparative pathogenomic platform with an interactive web interface Nucleic Acids Res. 2019 47 D687 D692 10.1093/nar/gky1080 30395255 28. Alcock B.P. Raphenya A.R. Lau T.T.Y. Tsang K.K. Bouchard M. Edalatmand A. Huynh W. Nguyen A.V. Cheng A.A. Liu S. CARD 2020: Antibiotic resistome surveillance with the comprehensive antibiotic resistance database Nucleic Acids Res. 2020 48 D517 D525 10.1093/nar/gkz935 31665441 29. Bortolaia V. Kaas R.S. Ruppe E. Roberts M.C. Schwarz S. Cattoir V. Philippon A. Allesoe R.L. Rebelo A.R. Florensa A.F. ResFinder 4.0 for predictions of phenotypes from genotypes J. Antimicrob. Chemother. 2020 75 3491 3500 10.1093/jac/dkaa345 32780112 30. Lee I. Ouk Kim Y. Park S.C. Chun J. OrthoANI: An improved algorithm and software for calculating average nucleotide identity Int. J. Syst. Evol. Microbiol. 2016 66 1100 1103 10.1099/ijsem.0.000760 26585518 31. Kumar S. Stecher G. Li M. Knyaz C. Tamura K. MEGA X: Molecular Evolutionary Genetics Analysis across Computing Platforms Mol. Biol. Evol. 2018 35 1547 1549 10.1093/molbev/msy096 29722887 32. Krzywinski M. Schein J. Birol I. Connors J. Gascoyne R. Horsman D. Jones S.J. Marra M.A. Circos: An information aesthetic for comparative genomics Genome Res. 2009 19 1639 1645 10.1101/gr.092759.109 19541911 33. Saleh H.A. Durham S.R. Perennial rhinitis BMJ 2007 335 502 507 10.1136/bmj.39304.678194.AE 17823191 34. Zeger S.L. Liang K.Y. Albert P.S. Models for longitudinal data: A generalized estimating equation approach Biometrics 1988 44 1049 1060 10.2307/2531734 3233245 35. Ring J. Behrendt H. de Weck A. History and classification of anaphylaxis Chem. Immunol. Allergy 2010 95 1 11 10.1159/000315934 20519878 36. Lack G. Bradley K.L. Hamelmann E. Renz H. Loader J. Leung D.Y. Larsen G. Gelfand E.W. Nebulized IFN-gamma inhibits the development of secondary allergic responses in mice J. Immunol. 1996 157 1432 1439 10.4049/jimmunol.157.4.1432 8759723 37. Kips J.C. Brusselle G.J. Joos G.F. Peleman R.A. Tavernier J.H. Devos R.R. Pauwels R.A. Interleukin-12 inhibits antigen-induced airway hyperresponsiveness in mice Am. J. Respir. Crit. Care Med. 1996 153 535 539 10.1164/ajrccm.153.2.8564093 8564093 38. Bouladoux N. Hall J.A. Grainger J.R. dos Santos L.M. Kann M.G. Nagarajan V. Verthelyi D. Belkaid Y. Regulatory role of suppressive motifs from commensal DNA Mucosal Immunol. 2012 5 623 634 10.1038/mi.2012.36 22617839 39. Montamat G. Leonard C. Poli A. Klimek L. Ollert M. CpG Adjuvant in Allergen-Specific Immunotherapy: Finding the Sweet Spot for the Induction of Immune Tolerance Front. Immunol. 2021 12 590054 10.3389/fimmu.2021.590054 33708195 40. Karimi K. Inman M.D. Bienenstock J. Forsythe P. Lactobacillus reuteri-induced regulatory T cells protect against an allergic airway response in mice Am. J. Respir. Crit. Care Med. 2009 179 186 193 10.1164/rccm.200806-951OC 19029003 41. Zhong Y. Huang J. Tang W. Chen B. Cai W. Effects of probiotics, probiotic DNA and the CpG oligodeoxynucleotides on ovalbumin-sensitized Brown-Norway rats via TLR9/NF-kappaB pathway FEMS Immunol. Med. Microbiol. 2012 66 71 82 10.1111/j.1574-695X.2012.00991.x 22612777 42. Iliev I.D. Tohno M. Kurosaki D. Shimosato T. He F. Hosoda M. Saito T. Kitazawa H. Immunostimulatory oligodeoxynucleotide containing TTTCGTTT motif from Lactobacillus rhamnosus GG DNA potentially suppresses OVA-specific IgE production in mice Scand J. Immunol. 2008 67 370 376 10.1111/j.1365-3083.2008.02080.x 18248528 43. Mazhary Z. Fard N.A. Minuchehr Z. Javanshir N. Package of anti-allergic probiotic Lactobacillus by focusing on the regulatory role of immunosuppressive motifs in allergy Inform. Med. Unlocked 2020 18 100280 10.1016/j.imu.2019.100280 44. Li A.L. Sun Y.Q. Du P. Meng X.C. Guo L. Li S. Zhang C. The Effect of Lactobacillus actobacillus Peptidoglycan on Bovine beta-Lactoglobulin-Sensitized Mice via TLR2/NF-kappaB Pathway Iran. J. Allergy Asthma Immunol. 2017 16 147 158 28601055 45. Mat Z. Grensemann B. Yakin Y. Knobloch J. Koch A. Effect of lipoteichoic acid on IL-2 and IL-5 release from T lymphocytes in asthma and COPD Int. Immunopharmacol. 2012 13 284 291 10.1016/j.intimp.2012.04.005 22542695 46. Wu M.H. Pan T.M. Wu Y.J. Chang S.J. Chang M.S. Hu C.Y. Exopolysaccharide activities from probiotic bifidobacterium: Immunomodulatory effects (on J774A.1 macrophages) and antimicrobial properties Int. J. Food Microbiol. 2010 144 104 110 10.1016/j.ijfoodmicro.2010.09.003 20884069 47. Gonzalez-Bacerio J. Izquierdo M. Aguado M.E. Varela A.C. Gonzalez-Matos M. Del Rivero M.A. Using microbial metalo-aminopeptidases as targets in human infectious diseases Microb. Cell 2021 8 239 246 10.15698/mic2021.10.761 34692819 48. Sahiner U.M. Birben E. Erzurum S. Sackesen C. Kalayci O. Oxidative stress in asthma World Allergy Organ. J. 2011 4 151 158 10.1097/WOX.0b013e318232389e 23268432 49. Aponte M. Murru N. Shoukat M. Therapeutic, Prophylactic, and Functional Use of Probiotics: A Current Perspective Front. Microbiol. 2020 11 562048 10.3389/fmicb.2020.562048 33042069 50. Stevens T.L. Bossie A. Sanders V.M. Fernandez-Botran R. Coffman R.L. Mosmann T.R. Vitetta E.S. Regulation of antibody isotype secretion by subsets of antigen-specific helper T cells Nature 1988 334 255 258 10.1038/334255a0 2456466 51. Yamaki K. Uchida H. Harada Y. Li X. Yanagisawa R. Takano H. Hayashi H. Taneda S. Mori Y. Yoshino S. Effect of methotrexate on Th1 and Th2 immune responses in mice J. Pharm. Pharmacol. 2003 55 1661 1666 10.1211/0022357022269 14738593 52. Hwang Y.H. Hong S.G. Mun S.K. Kim S.J. Lee S.J. Kim J.J. Kang K.Y. Yee S.T. The Protective Effects of Astaxanthin on the OVA-Induced Asthma Mice Model Molecules 2017 22 2019 10.3390/molecules22112019 29160801 53. Lin W.Y. Fu L.S. Lin H.K. Shen C.Y. Chen Y.J. Evaluation of the effect of Lactobacillus paracasei (HF.A00232) in children (6-13 years old) with perennial allergic rhinitis: A 12-week, double-blind, randomized, placebo-controlled study Pediatr. Neonatol. 2014 55 181 188 10.1016/j.pedneo.2013.10.001 24269033 |
PMC10000598 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050686 healthcare-11-00686 Article Validity and Reliability of a Non-Radiographic Postural Analysis Device Based on an RGB-Depth Camera Comparing EOS 3D Imaging: A Prospective Observational Study Lee Hyo Jeong Conceptualization Formal analysis Investigation Writing - original draft 1 Cho Han Eol Methodology Writing - review & editing 2 Kim Myungsang Methodology 2 Chung Seok Young Methodology 2 Park Jung Hyun Conceptualization Writing - original draft Supervision 234* Anwer Shahnawaz Academic Editor Wong Arnold Academic Editor Gok Kandasamy Academic Editor 1 Department of Rehabilitation Medicine, Bundang Jesaeng Hospital, Seongnam-si 13590, Gyeonggi-do, Republic of Korea 2 Department of Rehabilitation Medicine, Gangnam Severance Hospital, Rehabilitation Institute of Neuromuscular Disease, Yonsei University College of Medicine, Seoul 06229, Republic of Korea 3 Department of Integrative Medicine, Yonsei University College of Medicine, Seoul 06229, Republic of Korea 4 Department of Medical Device Engineering and Management, Yonsei University College of Medicine, Seoul 06229, Republic of Korea * Correspondence: [email protected]; Tel.: +82-2-2019-3490 25 2 2023 3 2023 11 5 68605 1 2023 21 2 2023 22 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). The posture-analyzing and virtual reconstructing device (PAViR) used a Red Green Blue-Depth camera as a sensor and skeleton reconstruction images were produced. This PAViR quickly analyzed the whole posture from multiple repetitive shots without radiation exposure in clothes and provided a virtual skeleton within seconds. This study aims to evaluate the reliability when shooting repeatedly and to assess the validity compared to parameters of full-body, low-dose X-rays (EOSs) when applied as diagnostic imaging. As a prospective and observational study, 100 patients with musculoskeletal pain underwent an EOS to obtain whole body coronal and sagittal images. The outcome measures were human posture parameters, which were divided by the standing plane in both EOSs and PAViRs as follows: (1) a coronal view (asymmetric clavicle height, pelvic oblique, bilateral Q angles of the knee, and center of seventh cervical vertebra-central sacral line (C7-CSL)) and (2) a sagittal view (forward head posture). A validation of the PAViR compared to the EOSs revealed that C7-CSL showed a moderate positive correlation with that of the EOS (r = 0.42, p < 0.01). The forward head posture (r = 0.39, p < 0.01), asymmetric clavicle height (r = 0.37, p < 0.01), and pelvic oblique (r = 0.32, p < 0.01) compared to those of the EOS had slightly positive correlations. The PAViR has excellent intra-rater reliability in people with somatic dysfunction. Except for both Q angles, the PAViR has fair-to-moderate validation when compared to EOS diagnostic imaging in the parameter representing coronal and sagittal imbalance. Although the PAViR system is not yet available in the medical field, it has the potential to become a radiation-free, accessible, and cost-effective postural analysis diagnostic tool after the EOS era. three-dimensional imaging posture skeleton musculoskeletal pain full-body X-ray Yonsei University College of Medicine6-2021-0090 This study was supported by a faculty research grant of Yonsei University College of Medicine (6-2021-0090). pmc1. Introduction Posture is defined as the alignment or orientation of the body in an upright position . Posture is related to the muscle length of the muscle activation rather than force . Clinically, posture is evaluated through the ideal gravitational line or vertical line that constitutes anatomical landmarks of the anterior, posterior, and lateral sides of the body . "Good" posture is considered to be a symmetrical alignment, upright, as well as effective at conserving energy ; it is important to reduce the risk of injury, certain prolonged static and awkward postures , and various cumulative traumatic disorders . Somatic dysfunction is a group of diseases of the musculoskeletal system and connective tissues, defined as an impaired or altered function of the related components of the somatic (body framework) system . Somatic dysfunction is aggravated by poor posture or results in an abnormal posture that leads to dysfunctional mechanics. It is the target of treatment in manual medicine , both chiropractic and osteopathic medicine, and has been classified as International Classification of Diseases 10th Revision, Clinical Modification (ICD-10-CM) Diagnosis Code M99.0 since 2016. Somatic dysfunction is usually assessed via palpatory investigation to evaluate four features: asymmetry (A), range of motion changes (R), tenderness (T), and tissue texture changes (T) . Among them, the evaluation of asymmetry is mainly made up of subjective inspections. If an objective evaluation can be made using quantitative equipment, a more accurate diagnosis will be possible. In addition, if the entire framework were to be imaged rather than a part of the body, it would be more helpful, biomechanically, to understand a patient's illness. Several studies have highlighted the importance of quantifying posture with radiographic or non-radiographic methods . Patients with musculoskeletal pain have a strong desire for diagnosis via imaging tests . Although radiographs using X-rays are the current gold standard protocol for diagnostic images, exposure to ionizing radiation may induce cancer . Therefore, these methods have limited use in sensitive populations, such as adolescents or pregnant women. Non-radiographic methods are available to monitor patient progress without repetitive radiation exposure . Biophotogrammetry, infra-red motion analysis, plumbline, spinal mouse, surface topography, and three-point ultrasound methods have been presented as viable alternatives; however, these methods are directly dependent on both mathematical methods and collection procedures, and few studies have systematically evaluated these methods . The EOS imaging system (EOS imaging, Paris, France) is a full-body, low-dose X-ray (EOS) that is used on patients in a weight-bearing position and is based on the X-ray detection technology of Charpak's chamber, which was awarded the 1992 Nobel Prize in Physics . Among the current radiographic methods, the EOS has relatively high accuracy. The dose of an EOS single micro-dose X-ray, 2.6 mSv, is much lower than the daily dose of natural background radiation , the global average of which is about 2.4 mSv annually from natural sources of radiation . Therefore, EOS has previously been used to diagnose abnormalities and malpositions of the spine, pelvis, and lower extremities using numerous parameters . However, high maintenance and labor costs make it less accessible for clinical applications . Additionally, patients are asked to assume an unnatural posture with both hands on the shoulder or zygomatic bones in a confined space that is usually present with the EOS. All of the previously proposed diagnostic imaging approaches, including the EOS, need one person to measure or derive parameters. Technologies using real-time, three-dimensional (3D) depth cameras have emerged with recent developments in camera and image processing technology. It is possible to repeatedly evaluate and obtain indices for body shape in real time and reconstruct the shape of the spine and skeleton with a specially designed human pose estimation algorithm . These processes can be performed with no radiation exposure, less space for equipment installation, and relatively lower costs. Additionally, it automatically shows the results without measuring by a physician. In this study, we propose a new device to evaluate human posture using a Red Green Blue-Depth (RGB-D) camera and an algorithm to reconstruct the virtual skeleton. This posture analyzing and virtual reconstructing device (PAViR) allows for a quick analysis of the person's whole posture from multiple repetitive shots without radiation exposure and provides a virtual skeleton within seconds. Therefore, the research sought to answer the following questions:In people with somatic dysfunction, is PAViR reliable when shooting repeatedly? When applied as diagnostic imaging, is PAViR valid compared to the parameters of EOS? 2. Materials and Methods 2.1. Participants Between January 2020 and June 2020, a prospective study was conducted on patients with somatic dysfunction who had been diagnosed with ICD-10 at the department of rehabilitation medicine in a tertiary hospital. Exclusion criteria included patients who were <19 years old; had a body mass index (BMI) > 35 kg/m2; had a history of metallic fixation devices after any spine surgery; or who were pregnant or could potentially be pregnant. Demographic data were shown in Table 1. A total of 100 patients (44 males and 56 females) with a mean age of 47.2 years were included in the study. The mean BMI was 23.1 +- 3.5 kg/m2. Informed consent for the publication of identifying information/images in an online open access publication was obtained from all subjects. This study was approved by the Institutional Review Board for Clinical Studies at our institution (3-2019-0305). This research was performed in accordance with the relevant guidelines and regulations. In addition, the study was certainly carried out in accordance with the Declaration of Helsinki. 2.2. EOS The full-body, low-dose EOS was performed on each patient. Participants were instructed to stand in the functionally loaded standing position (shoulders flexed to 45deg and hands resting on the zygomatic bones) and to stare forward. Standing positions were carefully monitored by a skilled operator who ensured that spinal segments did not compensate for movement while adapting the arms in a confined place. Whole-body coronal and sagittal images were obtained. 2.3. PAViR The PAViR (Moti Physio, MG solutions, Seoul, Republic of Korea) assessment was performed on each participant in clothes and was performed twice within 2 min by an experienced physician to determine the intra-rater reliability. When commanded by the PAViR system, a participant stands at the front, side, and back as if taking a picture. The physician only needs to ensure that the subject is standing in the position indicated by the laser. 2.3.1. RGB-D Camera Moti Physio systems use a real-time 3D RGB-D camera (Astra Pro, Orbbec 3D Technology International, Inc. Troy, MI, USA) as a sensor . To capture human motion, markerless approaches, convenient and accessible, typically use an RGB-D camera . The PAViR hardware system consists of a display unit, an input unit, an operation unit, and a positioning unit. The display unit serves to visually inform whether the subject is in the correct posture during measurement and to display the final result. The RGB-D camera of the input unit receives the data, and the operation unit processes the data received from the input unit to calculate the image to be displayed on the display unit. Finally, the positioning unit consists of a laser indicator and a floor mat, which illuminates a cross-laser line on the floor at a specified distance from the PAViR and fixes the position of the floor mat relative to the line. Thus, a specific distance and standard position are maintained for the real-time 3D RGB-D camera. 2.3.2. Support Vector Machine The RGB-D camera captures data in the front, side, and back while the subject stands in a comfortable position for 2-3 s without moving. The system produces a human silhouette from the depth frame data of the camera using the background subtraction method . Subsequently, as a processing algorithm, Simple Linear Iterative Clustering performs superpixel segmentation, and the parts are identified using the Support Vector Machine . 2.3.3. Geometric Method The coordinates of the user depth data are (x, y, z) in mm . It is necessary to adjust for sensitive clothing or hair and relatively unconcerned skeletal bones. To overcome this problem, we make some assumptions and use the geometric method. This operation is calculated every frame for 2-3 s, and the 28 points are used as the final calculated value based on the last average body part bone point (3D skeleton) and applied to the 3D model . The results derived from this algorithm are presented as an interactive virtual 3D model via a Liquid Crystal Display (LCD) screen, and the coronal and sagittal images are presented as print versions and email-based Portable Network Graphic (PNG) files. 2.4. Outcome Measures The primary outcome was human posture parameters, which were divided by the standing plane in both the EOS and PAViR as follows: (1) a coronal view (asymmetric clavicle height, the pelvic oblique, bilateral Q angles of the knee, and center of 7th cervical vertebra-central sacral line (C7-CSL)) and (2) a sagittal view (forward head posture). In the EOS, a physician measured the above values directly. Six outcomes were measured on each image: 1 the angle between the horizontal line and a line connecting the highest clavicle bones, 2 the horizontal distance between the midpoints of bilateral iliac crests, 3 and 4 the angle formed between the patella tendon and anterior superior iliac spine, 5 the vertical distance between the center of C7 and central sacral line, and 6 the angle between the center of C7 and audial canal in Figure 3. The unit of the asymmetric clavicle height, bilateral Q angles, and forward head posture is degrees, and the units of the pelvic oblique and C7-CSL are millimeters. However, with the PAViR, the device showed data in degrees immediately after shooting. Secondary outcomes included a data comparison between the PAViR and EOS for validation. 2.5. Data Analysis All statistical analyses were performed using SPSS for Windows, Version 25 (IBM Corp., Armonk, NY, USA). Intraclass correlation coefficients (ICCs) and their 95% confidence intervals were used to determine the intra-rater reliability for PAViR. ICC values > 0.75 represent excellent reliability, values between 0.4 and 0.75 represent fair-to-good reliability, and values < 0.4 represent poor reliability . The relationships between the PAViR measurements and EOS measurements were compared via a paired t-test and correlation analysis (Pearson correlation coefficient). Pearson correlation coefficients and the ICC were characterized as poor (0.00 to 0.20), fair (0.21 to 0.40), moderate (0.41 to 0.60), good (0.61 to 0.80), or excellent (0.81 to 1.00) . The level of significance was set at <0.05 for all statistical tests. 3. Results 3.1. Outcomes of Measuring with EOS and PAViR The descriptive outcomes of the coronal and sagittal parameters measured are shown in Table 2. Negative values mean that the left is raised or that the posture is tilted to the left. There is a significant difference between the two devices in pelvic oblique, bilateral Q angle, and C7-CSL. In the EOS, most participants had a position with their torso tilted to the left and head tilted forward. For the Q angle, only one patient had a negative value on the right knee. 3.2. Intra-Rater Reliability of PAViR All intra-rater correlation coefficients for the coronal (asymmetric clavicle height, pelvic oblique, bilateral Q angle of the knee, C7-CSL) and sagittal view (forward head posture) parameters were > 0.69, and the highest parameter of the PAViR was the C7-CSL (ICC= 0.84) (Table 3). 3.3. Validation of PAViR Compared to Parameters of EOS Primary outcomes were compared to validate the PAViR for each parameter. An analysis adjusted for age, height, weight, and BMI was calculated. Of the PAViR parameters, C7-CSL showed a moderate positive correlation (r = 0.42, p < 0.001) with that of the EOS. Forward head posture (r = 0.39, p < 0.002), asymmetric clavicle height (r = 0.37, p < 0.002), and pelvic oblique (r = 0.32, p < 0.002), compared to the results with the EOS, were fair positive correlations, as shown in Table 4. However, there was no significant correlation for the bilateral Q angles of the knee between PAViR and EOS. 4. Discussion To serve as a pilot study, among the asymmetry that is evaluated to diagnose somatic dysfunction, the PAViR values in the coronal (asymmetric clavicle height, the pelvic oblique, Q angles of the knee, C7-CSL) and sagittal views (forward head posture) were compared with the values obtained using the EOS. Intra-rater reliability was good-to-excellent in the newly developed PAViR system. In PAViR values compared to the EOS, all measured values except the Q angle showed fair-to-moderate correlation. Although we adjusted for sensitive clothing or hair and relatively unconcerned skeletal bones via the geometric method, the C7-CSL value, which is simple to capture the midline for the body frame, is thought to be the most consistent, rather than the Q angle with various knee creases. To clarify the validity, further study will be necessary to and should undertake analyses with different clothing or undressed participants. Additionally, there is a need for technological development. Somatic dysfunction is the impaired or altered function of related components of the somatic (body framework) system . Dysfunction is not defined by localization but by the result of the interplay of a whole chain of different structures. Clinically, patients may complain of pain after an awkward movement, prolonged posture, or overuse of muscles. Somatic dysfunction is aggravated by poor posture or results in an abnormal posture that leads to dysfunctional mechanics . Among features of somatic dysfunction, the evaluation of asymmetry is mainly made up of subjective inspections. If an objective evaluation can be made using quantitative equipment, a more accurate diagnosis will be possible. To evaluate the correlation of its links, it would be more helpful to understand a patient's condition if the whole body were imaged rather than a part of the body. Therefore, the EOS system is an outstanding equipment for measuring imbalances within current medical field . However, the EOS has also several limitations X-rays and EOS are highly accurate and essential for initial evaluation but are not suitable for routine continuous evaluations every few days or weeks due to radiation overexposure or cost concerns. In whole spine X-rays, the amount of radiation can exceed the annual amount of natural background radiation by 2.4 mSv annually. Therefore, given the risk-benefit analysis, it is difficult to recommend a whole spine X-ray to determine the effectiveness of treatment rather than for initial diagnostic purposes. In contrast, an effective dose of an EOS single micro-dose X-ray (2.6 mSv) is less than the amount from one day of natural background radiation . Nonetheless, the EOS is very expensive and can be difficult to access and maintain. Considering the cost-benefit ratio, the EOS may also not be suitable as a routine assessment tool for the follow-up of treatment effectiveness. However, with the PAViR, patients are not exposed to radiation, and the tests can be performed repeatedly without the risk of radiation, even in growing adolescents and young adults of a childbearing age. Since it is extremely cheaper, indeed, of a 100 times difference, there is no burden of regular multiple patient evaluations. Additionally, the installation and test area is less than 3 m2; therefore, there is almost no space limitation, and since it is possible to use without disrobing, there is no need for a separate changing space. Since the device automatically derives the measurements without supervision, the physiotherapist can immediately apply it to treatment without going through a doctor. Therefore, the PAViR can be used not only for medical purposes such as manual, chiropractic, and osteopathic therapy, but also for post-exercise performance analysis such as Pilates, yoga, and general workouts. The repeatability of thew measurement results is critical when using a simple evaluation tool such as the PAViR. The lack of repeatability would make it difficult to evaluate the effectiveness of manual, chiropractic, osteopathic, or exercise therapy. Other non-radiographic instruments, such as biophotogrammetry, infra-red motion analysis, plumbline, spinal mouse, surface topography, and free point ultrasound, rely on both mathematical methods and collection procedures . The PAViR intra-rater reliability results in this study were good-to-excellent for all parameters . Computed tomography (CT) scans and biplanar X-ray 3D reconstructions can be measured relatively accurately, but a reconstruction time of 10 min or more is required, and a skilled person is required for reconstruction. In addition, the risk of radiation exposure cannot be excluded . Optical methods such as Moire-Fringe topography, structured light techniques, the Integrated Shape Imaging System, Quantec system, and Orteliuss scanner can be used to detect spinal deformities . Some clinics have applied these tools to monitor scoliosis, but they are currently difficult to obtain due to the complicated manipulation of the equipment and the inconvenience of patients having to completely disrobe. Since these devices have been designed primarily to evaluate spinal deformities such as scoliosis, only the back view is taken; therefore, shoulder height asymmetry, pelvic obliques, Q angles of the knee, and forward head posture cannot be obtained. Conversely, the PAViR uses a 3D depth camera to collect body surface data, calibrates using human pose estimation (HPE), defines key specific points, and then reproduces the shape of the spine and skeletal system. The subjects can undergo the test without disrobing, and the results are received on screen or via e-mail within 1 min. It can be used conveniently without human effort like the previously mentioned technologies since all processes are automated. Therefore, the PAViR system could grant the potential to assess the skeletal posture of clothed participants without radiation exposure. Unlike conventional non-radiography, it is a technology that can analyze the whole body in real-time without taking off clothes via the markerless approach of the RGB-D camera. In addition, participants can stand in front of the device and follow the guide without a measurer, and it might be easily transmitted to the medical staff for an application. Further studies are needed to test whether our findings apply to patients with other musculoskeletal disorders, such as cobb angle of scoliosis, degenerative lumbar kyphosis, abnormal pelvic rotation, bowlegs, and knock knees. Study Limitations This study has some limitations. First, although power analysis revealed the statistical significance of our data; assuming a power of 80%, 29 curves were required for each subgroup according to the sample size calculator for the G-power 3.1.9.4 program; including greater numbers of subjects could further increase the power of our study. Second, the algorithm has not been perfected thus far; it is necessary to constantly update the algorithm through comparison with EOS and big data technology. Accumulated data will be helpful to increase the accuracy and correlation by using big data, thus increasing the prediction of risk for musculoskeletal disorders. Third, different positions may result in a variation in the value itself. Although the observed trend would be maintained, more research would be required to compare them in the same position in order to improve accuracy. 5. Conclusions The PAViR has excellent intra-rater reliability in people with somatic dysfunction. Except for both Q angles, the PAViR has fair to moderate validation when compared to EOS diagnostic imaging in the parameter representing coronal and sagittal imbalance. Although the PAViR system is not yet available in the medical field, it has the potential to become a radiation-free, accessible, and cost-effective postural analysis diagnostic tool after the EOS era. This study provides valuable insights to researchers interested in digital healthcare. Based on this technology, more research into improving algorithm accuracy using big data is required, and it will be necessary to apply to patients with musculoskeletal pain. Acknowledgments We would like to thank MG solutions for assisting with the figures and technical works. Author Contributions Conceptualization: J.H.P., H.J.L.; Methodology: M.K., S.Y.C., H.E.C.; Formal analysis and investigation: H.J.L.; Writing: J.H.P., H.J.L., H.E.C.; Supervision: J.H.P. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement This study was approved by the Institutional Review Board of Gangnam Severance Hospital, Seoul, Republic of Korea (Identifier: 3-2019-0305). Informed Consent Statement Informed consent for the publication of identifying information/images in an online open-access publication was obtained from all subjects. Data Availability Statement All data analyzed in this study are available from the corresponding author upon reasonable request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 An example of the measurement in the side (A). The flow of the human skeletal pose estimation method from the depth camera. After producing a human silhouette using the background-subtraction method (B), the human finally becomes normalized (C). The classification of body parts with various colors shows an already trained processing algorithm for segmentation via Simple Linear Iterative Clustering (D,E). Results of the skeletal point estimation on a human body are extracted by using an image processing algorithm. In a markerless object (E), white markers are created via a series of processes (F). Illustration provided by MG solutions. Figure 2 An example of a 'calibrated skeletal point of side shoulder' from the side. (A) Side user depth points with side shoulder (x, y, z) in the XY plane. (B) The set 'S' is defined as the depth point clouds on the shoulder line in the XZ plane; these points form a graph of quadratic equations. (C) The calibrated point of the shoulder was generated in the XY plane. Illustration provided by MG solutions. Figure 3 (A) The EOS and (B) PAViR data in the same subject: 1 Asymmetric clavicle height, 2 pelvic oblique, 3 right Q angles of the knee, 4 left Q angles of the knee, 5 the center of 7th cervical vertebra-central sacral line (C7-CSL), 6 forward head posture. All parameters are in degrees except when representing 2 and 5 in EOS, indicated instead as distance, millimeter. Negative values mean that the left is raised or that the posture is tilted to the left. Illustration provided by MG solutions. PAViR posture analyzing and virtual reconstructing device. healthcare-11-00686-t001_Table 1 Table 1 Demographic characteristics of participants. Variables (n = 100) Values Range Gender, n male/female 44/56 Age (y), (mean +- SD) 47.2 +- 16.5 19~81 Weight (kg), (mean +- SD) 63.3 +- 13.1 37.4~89.3 Body height (cm), (mean +- SD) 163.4 +- 19.3 143.0~183.0 Body mass index (kg/m2), (mean +- SD) 23.1 +- 3.5 15.6~30.8 SD standard deviation. healthcare-11-00686-t002_Table 2 Table 2 Descriptive statistics of coronal and sagittal parameters obtained via the EOS and PAViR. EOS PAViR Parameters Mean +- SD Range Mean +- SD Range Coronal view Asymmetric clavicle height (deg) 0.1 +- 2.8 -8.0 a~16.0 -1.0 +- 1.6 -5.0~3.8 Pelvic oblique (mm, deg) b -0.3 +- 5.0 -12.0~14.0 0.7 +- 1.6 * -2.6~6.4 Right Q angle (deg) 6.1+- 1.7 -1.8~10.4 0.9 +- 7.9 * -8.0~14.0 Left Q angle (deg) 5.6 +- 1.6 0.9~9.8 -3.1 +- 4.0 * -7.9~13.5 C7-CSL (mm, deg) b -3.0 +- 13.3 -59.0~36.0 -1.3 +- 2.2 * -8.1~4.3 Sagittal view Forward head posture (deg) 7.0 +- 6.9 -5.1~29.4 7.9 +- 6.3 -5.0~29.0 PAViR posture-analyzing and virtual reconstructing device, SD standard deviation, C7-CSL center of 7th cervical vertebra-central sacral line; a Negative values mean that the left is raised or that the posture is tilted to the left; b All parameters are in degrees except for the pelvic oblique and C7-CSL in the EOS, indicated as distance, millimeter; * p < 0.05. healthcare-11-00686-t003_Table 3 Table 3 Intra-rater reliability of PAViR. Parameters Coefficient Value p-Value Asymmetric clavicle height 0.69 0.005 Pelvic oblique 0.72 0.002 Right Q angle of knee 0.72 0.001 Left Q angle of knee 0.79 0.001 C7-CSL 0.84 0.002 Forward head posture 0.76 0.001 PAViR posture-analyzing and virtual reconstructing device, C7-CSL center of 7th cervical vertebra-central sacral line. healthcare-11-00686-t004_Table 4 Table 4 The Pearson correlation coefficient (r) for validity between PAViR and EOS. Parameters Correlation Coefficient p-Value Asymmetric clavicle height 0.37 <0.002 Pelvic oblique 0.32 <0.002 Right Q angle of knee -0.47 0.14 Left Q angle of knee -0.15 0.15 C7-CSL 0.42 <0.001 Forward head posture 0.39 <0.002 PAViR posture analyzing and virtual reconstructing device, C7-CSL center of 7th cervical vertebra-central sacral line. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Raine S. Twomey L. Attributes and qualities of human posture and their relationship to dysfunction or musculoskeletal pain Crit. Rev. Phys. Rehabil. Med. 1994 6 409 2. Kamper D.G. Fischer H.C. Cruz E.G. Impact of finger posture on mapping from muscle activation to joint torque Clin. Biomech. (Bristol Avon.) 2006 21 361 369 10.1016/j.clinbiomech.2005.11.005 16406192 3. Magee D.J. Orthopedic Physical Assessment (Musculoskeletal Rehabilitation) Saunders Toronto, Canada 2014 4. Pope P.M. Severe and Complex Neurological Disability: Management of the Physical Condition Butterworth-Heinemann Oxford, UK 2007 5. Laubli T. Karpilow C. Global Occupational Health Oxford university press Oxford, UK 2011 277 299 6. Takasaki H. May S. Mechanical diagnosis and therapy has similar effects on pain and disability as 'wait and see' and other approaches in people with neck pain: A systematic review J. Physiother. 2014 60 78 84 10.1016/j.jphys.2014.05.006 24952834 7. Rey-Matias R.R. 16--Manipulation, Traction, and Massage Braddom's Rehabilitation Care: A Clinical Handbook Cifu D.X. Lew H.L. Elsevier Amsterdam, The Netherlands 2018 111 118.e8 8. Chaitow L. Chapter 6--Osteopathic Assessment and Treatment of Thoracic and Respiratory Dysfunction Multidisciplinary Approaches to Breathing Pattern Disorders Chaitow L. Bradley D. Gilbert C. Ley R. Churchill Livingstone Edinburgh, Scotland 2002 131 172 9. Williams N. Managing back pain in general practice is osteopathy the new paradigm? Br. J. Gen. Pract. 1997 47 653 655 9474832 10. Muller R. Linz W. Buchmann J. Manual medicine--a powerful "hands on" facility to treat somatic and segmental dysfunction with musculosceletal pain, increased muscule tension, restrictions of fascia and posture asymmetries MMW Fortschr. Med. 2011 153 27 30 11. Fortin C. Ehrmann Feldman D. Cheriet F. Labelle H. Clinical methods for quantifying body segment posture: A literature review Disabil. Rehabil. 2011 33 367 383 10.3109/09638288.2010.492066 20568973 12. Lim Y.Z. Chou L. Au R.T.M. Seneviwickrama K.L.M.D. Cicuttini F.M. Briggs A.M. Sullivan K. Urquhart D.M. Wluka A.E. People with low back pain want clear, consistent and personalised information on prognosis, treatment options and self-management strategies: A systematic review J. Physiother. 2019 65 124 135 10.1016/j.jphys.2019.05.010 31227280 13. Legaye J. Follow-up of the sagittal spine by optical technique Ann Phys Rehabil Med 2012 55 76 92 10.1016/j.rehab.2011.12.002 22236627 14. Cohen L. Kobayashi S. Simic M. Dennis S. Refshauge K. Pappas E. Non-radiographic methods of measuring global sagittal balance: A systematic review Scoliosis Spinal Disord. 2017 12 30 10.1186/s13013-017-0135-x 29026895 15. Furlanetto T.S. Sedrez J.A. Candotti C.T. Loss J.F. Photogrammetry as a tool for the postural evaluation of the spine: A systematic review World J. Orthop. 2016 7 136 148 10.5312/wjo.v7.i2.136 26925386 16. Dubousset J. Charpak G. Dorion I. Skalli W. Lavaste F. Deguise J. Kalifa G. Ferey S. A new 2D and 3D imaging approach to musculoskeletal physiology and pathology with low-dose radiation and the standing position: The EOS system Bull. De L'academie Natl. De Med. 2005 189 287 297 287-297; discussion 297 17. Somoskeoy S. Tunyogi-Csapo M. Bogyo C. Illes T. Accuracy and reliability of coronal and sagittal spinal curvature data based on patient-specific three-dimensional models created by the EOS 2D/3D imaging system Spine J. 2012 12 1052 1059 10.1016/j.spinee.2012.10.002 23102842 18. Deschenes S. Charron G. Beaudoin G. Labelle H. Dubois J. Miron M.C. Parent S. Diagnostic imaging of spinal deformities: Reducing patients radiation dose with a new slot-scanning X-ray imager Spine (Phila Pa 1976) 2010 35 989 994 10.1097/BRS.0b013e3181bdcaa4 20228703 19. Hui S.C.N. Pialasse J.-P. Wong J.Y.H. Lam T.-p. Ng B.K.W. Cheng J.C.Y. Chu W.C.W. Radiation dose of digital radiography (DR) versus micro-dose x-ray (EOS) on patients with adolescent idiopathic scoliosis: 2016 "John Sevastic Award" Winner in Imaging Research Scoliosis Spinal Disord. 2016 11 46 10.1186/s13013-016-0106-7 28035336 20. Wall B.F. Ionising Radiation Exposure of the Population of the United States: NCRP Report no. 160 Oxford University Press Oxford, UK 2009 21. Metaxas V.I. Messaris G.A. Lekatou A.N. Petsas T.G. Panayiotakis G.S. Patient Doses in Common Diagnostic X-ray Examinations Radiat. Prot. Dosim. 2019 184 12 27 10.1093/rpd/ncy169 22. Kim H. Lee S. Lee D. Choi S. Ju J. Myung H. Real-time human pose estimation and gesture recognition from depth images using superpixels and SVM classifier Sensors 2015 15 12410 12427 10.3390/s150612410 26016921 23. Straka M. Hauswiesner S. Ruther M. Bischof H. Skeletal Graph Based Human Pose Estimation in Real-Time Proceedings of the British Machine Vision Conference Dundee, UK 29 August-2 September 2011 1 12 24. Achilles F. Ichim A.-E. Coskun H. Tombari F. Noachtar S. Navab N. Patient MoCap: Human Pose Estimation under Blanket Occlusion for Hospital Monitoring Applications Springer International Publishing Cham, Switzerland 2016 491 499 25. Faro F.D. Marks M.C. Pawelek J. Newton P.O. Evaluation of a functional position for lateral radiograph acquisition in adolescent idiopathic scoliosis Spine (Phila Pa 1976) 2004 29 2284 2289 10.1097/01.brs.0000142224.46796.a7 15480143 26. Garg B. Mehta N. Bansal T. Malhotra R. EOS(R) imaging: Concept and current applications in spinal disorders J. Clin. Orthop. Trauma 2020 11 786 793 10.1016/j.jcot.2020.06.012 32879565 27. Pumberger M. Schmidt H. Putzier M. Spinal Deformity Surgery: A Critical Review of Alignment and Balance Asian Spine J. 2018 12 775 783 10.31616/asj.2018.12.4.775 30060389 28. Vergari C. Skalli W. Clavel L. Demuynck M. Valentin R. Sandoz B. Similowski T. Attali V. Functional analysis of the human rib cage over the vital capacity range in standing position using biplanar X-ray imaging Comput. Biol. Med. 2022 144 105343 10.1016/j.compbiomed.2022.105343 35235874 29. Vinod Gutta E.D.L. Natalie Baddour, Pascal Fallavollita, A Comparison of Depth Sensors for 3D Object Surface Reconstruction The Canadian Medical and Biological Engineering Society CMBES/SCGB Ottawa, ON, Canada 2019 Volume 42 30. Jang C.W. Park J. Cho H.E. Park J.H. Appraisal of the new posture analyzing and virtual reconstruction device (PAViR) for assessing sagittal posture parameters: A prospective observational study Int. J. Env. Res. Public Health 2022 19 11109 10.3390/ijerph191711109 36078824 31. Camplani M. Salgado L. Background foreground segmentation with RGB-D Kinect data: An efficient combination of classifiers J. Vis. Commun. Image Represent. 2014 25 122 136 10.1016/j.jvcir.2013.03.009 32. Achanta R. Shaji A. Smith K. Lucchi A. Fua P. Susstrunk S. SLIC superpixels compared to state-of-the-art superpixel methods IEEE Trans. Pattern Anal. Mach. Intell. 2012 34 2274 2282 10.1109/TPAMI.2012.120 22641706 33. Fleiss J.L. Levin B. Paik M.C. Statistical Methods for Rates and Proportions John Wiley & Sons Hoboken, NJ, USA 2013 34. Campos S. Zhang L. Sinclair E. Tsao M. Barnes E.A. Danjoux C. Sahgal A. Goh P. Culleton S. Mitera G. The palliative performance scale: Examining its inter-rater reliability in an outpatient palliative radiation oncology clinic Support. Care Cancer 2009 17 685 690 10.1007/s00520-008-0524-z 18946683 35. Kasten K.M. Lewis D.D. High-Velocity, Low-Amplitude Management of Posterior Rib Somatic Dysfunction J. Am. Osteopath. Assoc. 2020 120 e1 e2 10.7556/jaoa.2020.013 31904779 36. Rousseau M.-A. Brusson A. Lazennec J.-Y. Assessment of the axial rotation of the pelvis with the EOS(r) imaging system: Intra-and inter-observer reproducibility and accuracy study Eur. J. Orthop. Surg. Traumatol. 2014 24 891 895 10.1007/s00590-013-1281-3 23942777 37. Cicchetti D.V. Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology Psychol. Assess. 1994 6 284 10.1037/1040-3590.6.4.284 38. Humbert L. De Guise J.A. Aubert B. Godbout B. Skalli W. 3D reconstruction of the spine from biplanar X-rays using parametric models based on transversal and longitudinal inferences Med. Eng. Phys. 2009 31 681 687 10.1016/j.medengphy.2009.01.003 19230743 39. Kosling S. Dietrich K. Steinecke R. Kloppel R. Schulz H.-G. Diagnostic value of 3 D CT surface reconstruction in spinal fractures Eur. Radiol. 1997 7 61 64 10.1007/s003300050110 9000399 40. Hocquelet A. Cornelis F. Jirot A. Castaings L. de Seze M. Hauger O. Patient-specific 3D models created by 3D imaging system or bi-planar imaging coupled with Moire-Fringe projections: A comparative study of accuracy and reliability on spinal curvatures and vertebral rotation data Eur. Spine J. 2016 25 3154 3161 10.1007/s00586-016-4659-1 27323963 41. Daruwalla J. Balasubramaniam P. Moire topography in scoliosis. Its accuracy in detecting the site and size of the curve J. Bone Joint Surg. Br. Vol. 1985 67 211 213 42. Berryman F. Pynsent P. Fairbank J. Disney S. A new system for measuring three-dimensional back shape in scoliosis Eur. Spine J. 2008 17 663 672 10.1007/s00586-007-0581-x 18247064 |
PMC10000599 | Approximately 40% of patients with cancer are eligible for check-point inhibitor (CPI) therapy. Little research has examined the potential cognitive impact of CPIs. First-line CPI therapy offers a unique research opportunity without chemotherapy-related confounders. The purpose of this prospective, observational pilot was to (1) demonstrate the feasibility of prospective recruitment, retention, and neurocognitive assessment for older adults receiving first-line CPI(s) and (2) provide preliminary evidence of changes in cognitive function associated with CPI(s). Patients receiving first-line CPI(s) (CPI Group) were assessed at baseline (n = 20) and 6 months (n = 13) for self-report of cognitive function and neurocognitive test performance. Results were compared to age-matched controls without cognitive impairment assessed annually by the Alzheimer's Disease Research Center (ADRC). Plasma biomarkers were measured at baseline and 6 months for the CPI Group. Estimated differences for CPI Group scores prior to initiating CPIs (baseline) trended to lower performance on the Montreal Cognitive Assessment-Blind (MOCA-Blind) test compared to the ADRC controls (p = 0.066). Controlling for age, the CPI Group's 6-months MOCA-Blind performance was lower than the ADRC control group's 12-months performance (p = 0.011). No significant differences in biomarkers were detected between baseline and 6 months, although significant correlations were noted for biomarker change and cognitive performance at 6 months. IFNg, IL-1b, IL-2, FGF2, and VEGF were inversely associated with Craft Story Recall performance (p < 0.05), e.g., higher levels correlated with poorer memory performance. Higher IGF-1 and VEGF correlated with better letter-number sequencing and digit-span backwards performance, respectively. Unexpected inverse correlation was noted between IL-1a and Oral Trail-Making Test B completion time. CPI(s) may have a negative impact on some neurocognitive domains and warrant further investigation. A multi-site study design may be crucial to fully powering prospective investigation of the cognitive impact of CPIs. Establishment of a multi-site observational registry from collaborating cancer centers and ADRCs is recommended. checkpoint inhibitors immunotherapy cognitive function cancer first-line therapy University of Kansas Cancer Center (KUCC)Kansas Institute for Precision Medicine COBREP20GM130423 This research was funded by the University of Kansas Cancer Center (KUCC) Lied Pilot Grant, a Children's Mercy Hospital KUCC Consortium Grant, and the Kansas Institute for Precision Medicine COBRE (P20GM130423). pmc1. Introduction The American Cancer Society estimates 1.9 million new diagnoses of cancer for 2022 . The rates of cancer and cancer-treatment-related cognitive impairment (CRCI) for survivors of non-central nervous system (CNS) malignancies ranges as high as 75%, with estimates of 30% of survivors experiencing long-lasting cognitive impairment . CRCI for non-CNS malignancies has a significant impact on a cancer survivor's quality of life at home, at work, and socially . Commonly reported issues of CRCI include trouble within the cognitive domains of short-term memory, attention and concentration, executive function, and visuospatial ability . These issues translate into difficulties with word finding, reading complex material, forgetting appointments, misplacing items, and, in some cases, issues with driving . Given the prevalence of these issues and the number of individuals who experience them, enhanced understanding of the treatments that contribute to CRCI is crucial for informed consent prior to administration of therapies, accurate patient/family education, and on-going assessment. Understanding the mechanisms involved in the development of CRCI is critical to the investigation of effective interventions. Evidence within the CRCI literature supports the association between prolonged production of inflammatory cytokines and the cognitive changes attributed to diagnosis and treatment of non-CNS malignancies . Specifically, the body's response to malignant cells and various cancer treatments, such as chemotherapy, induces the expression of a number of cytokines in the peripheral blood. Cytokines are thought to both actively and passively cross the blood-brain barrier and stimulate further cytokine release within the central nervous system inflammatory network. Cytokines are proposed to play a role in neuroprogenitor cell injury through interaction with cytokine-specific receptors in neuronal or endothelial brain cells . Cytokines of interest include: Interleukin-1 alpha, (IL-1a), IL-1ss, IL-2, IL-6, tumor necrosis factor alpha (TNFa), and interferon beta (INFss). Animal models of neuroinflammation in the context of cancer and treatment with radiation therapy and immunotherapy indicate changes also may occur in interferon gamma (IFNg) and fibroblast growth factor-basic (bFGF) levels . Neurotrophic factors (such as brain derived neurotrophic factor-BDNF), and other growth factors involved with neurogenesis (such as insulin-like growth factor 1 (IGF1) and vascular endothelial growth factor (VEGF)) also are of interest in preparation for future investigations of effective interventions to improve cognitive function . The inflammatory response to the cancer and cancer therapy has been postulated to have both a direct and indirect effect on neuroprogenitor cells, functional and structural connectivity, and cognitive function. Release of pro-inflammatory cytokines within the central nervous system and peripheral blood, and down-regulation of brain-derived neurotrophic factor are associated with CRCI. Other candidate pathways include biological pathways common to aging (e.g., hormonal changes in estrogen/testosterone, damage to DNA repair mechanisms, shortening of telomeres, and reduction of brain blood flow). Cancer and cancer therapy are postulated to accelerate cognitive aging. Thus, older adults treated for cancer may be at greater risk for cognitive impairment. At the time this study was being developed, the literature indicated that recent advances in immunotherapy with checkpoint inhibition had resulted in approvals for 7 drugs with indications for more than 15 tumor types . Initially, checkpoint inhibitors (CPIs) only were given as second-line or later therapies. However, recent approvals have been granted for first-line and combination regimens with both immunotherapy and chemotherapy or other targeted agents in addition to later-line therapies following recurrence after chemotherapy . CPIs have been developed to inhibit programmed cell death protein (PD-1), programmed death ligand (PD-L1), and cytotoxic T-lymphocyte antigen (CTLA-4) . Combination regimens with more than one CPI are able to target more than one receptor on the T cells. Other approved combined regimens include inhibitors of VEGF and various chemotherapy agents. The mechanism of action for checkpoint inhibition involves stimulation of an immune response to the cancer that results in the expression of inflammatory products, including cytokines. Immune-related adverse events (irAEs) are attributed to this inflammatory response . More severe irAEs have been associated with CTLA-4 inhibition and combination regimens . To date, scant research has been conducted to determine the impact of immunotherapy on cognitive function . What little research has been conducted has primarily involved patients who previously had received chemotherapy, making teasing out the cognitive impact for multiple lines of therapy difficult . First-line treatment with CPIs provides a unique and compelling opportunity to study the impact of this form of immunotherapy alone (i.e., without confounding by other cancer treatment) on cognitive function. A recent systematic review indicated that over 40% of patients with cancer now are considered eligible for CPI therapy, and this number will continue to increase . In addition to intravenous infusions investigation, the development of oral formulations of these drugs is also underway. Development of a prospective, observational registry for individuals receiving first-line checkpoint inhibition for cancer would contribute important information to the state of the science about the potential impact of checkpoint inhibition on cognitive function for a growing population. The purpose of this prospective, observational pilot study was to (1) demonstrate the feasibility of prospective recruitment, retention, and neurocognitive assessment for older adults with non-CNS malignancies receiving first-line treatment with CPI(s) and (2) provide preliminary evidence of changes in cognitive function associated with checkpoint inhibitor immunotherapy. Comparisons between baseline and 6-month assessments for changes in self-report of cognitive function and performance on neurocognitive tests were planned to generate an effect size to inform future prospective research with an observational registry. The pilot study objectives were to:Aim 1: Demonstrate the feasibility of recruiting, assessing, and retaining 20 older adults (>/= age 60) newly diagnosed with cancer who will receive first-line therapy with CPI(s) (CPI Group); Aim 2a: Estimate change and variability in participants' self-reports of cognitive function and objectively measured neurocognitive performance over time: Baseline (T1: within 1-2 weeks of initiation therapy with CPIs) and 6 months later (T2); Aim 2b: Estimate change and variability in inflammatory and neurotrophic biomarkers between T1 and T2; Aim 2c: Compare change and variability in CPI Group participants' objectively measured neurocognitive performances between T1 and T2 to existing control data available from the University of Kansas Alzheimer's Disease Research Center (ADRC) database for age-matched cognitively intact cohort participants (data recorded at baseline and 12 months). Pilot outcomes were expected to firmly establish successful recruitment procedures and identify potential barriers to retention. Information also was anticipated regarding the evolution of tumor types most likely to be represented within the institutional catchment area in addition to the previous projections based on retrospective Tumor Registry data. 2. Materials and Methods 2.1. Collaboration Alzheimer's Disease Research Centers across the United States contribute data to the National Alzheimer's Coordinating Center (NACC) Uniform Data Set (UDS). NACC has developed a standardized neuropsychologic battery to test the cognitive domains of executive function, episodic memory, attention/working memory, and language/semantic memory. ADRCs administer this standardized battery to hundreds of participants annually who have provided informed consent for the use of their de-identified data for research purposes. The UDS participants cover a broad range of neuropsychological performance, including a robust population of cognitively intact participants. As of December 2022, the UDS includes annual neuropsychological data for over 11,000 cognitively intact participants aged 35-84 years and over 3000 who are under age 65 (UDS Demographics and diagnoses | National Alzheimer's Coordinating Center (naccdata.org, accessed on 5 February 2023). The study design and implementation for this pilot study resulted from a collaborative effort between the National Cancer Institute-designated Comprehensive Cancer Center (University of Kansas Cancer Center-KUCC) and the National Institute on Aging funded by the Alzheimer's Disease Research Center (University of Kansas Medical Center-KUMC ADRC). The annual neuropsychological assessments are administered by the KUMC ADRC psychometricians. This collaboration leveraged the ADRC infrastructure for standardized neurocognitive assessment and provided the age-matched control data from the pool of cognitively intact cohort patients used for control comparisons in this pilot study. 2.2. Eligibility Control data were included from cognitively intact adults aged 60 and older with a Clinical Dementia Rating Score of zero and without clinically meaningful deficits in their cognitive performance. Control data were excluded for individuals with clinically meaningful depression or anxiety, Parkinson's disease, cancer within the last 5 years (except non-metastatic basal or squamous cell carcinoma), history of drug or alcohol abuse (DSM-IV criteria) in the last 2 years, and visual or auditory limitations or other systemic or neurological disease that may interfere with cognition. The eligible pool included 231 participants who were cognitively intact at baseline and at the 12-month follow up. The most recent fully extracted 12-month Tumor Registry data were used to estimate the potential pool of patients receiving first-line therapy with CPI(s). These data indicated that approximately 200 patients would meet eligibility criteria. Initial inclusion criteria required participants to be age 60 or older, scheduled to receive first-line treatment with CPI(s) (combination therapy with more than one CPI was accepted), diagnosed with any stage of non-CNS malignancy (without brain metastases), chemotherapy naive, and able to speak and read English. Patients initially were excluded for previous receipt of a BRAF or tyrosine kinase inhibitor, or comorbidities affecting cognitive function (such as Alzheimer's Disease or related dementias). Participation in other clinical trials was not automatically exclusionary but evaluated on a case basis by the research team. Concomitant participation in the institutional Biospecimen Repository Core Facility (BRCF) protocol was required, allowing the collection, storage, and analyses of extra blood sampling at the time of standard of care lab sampling. 2.3. Recruitment Recruitment was planned from the five KUCC sites located within the Kansas City metro area. Potential participants were to be identified by medical oncologists, urologists, nurse practitioners, and other advanced practice providers at the time of diagnosis and treatment planning. Clinical research coordinators were assigned to the study to assist with identification of eligible individuals. The principal investigator (PI) presented the study at all pertinent tumor-specific disease working groups, and the study synopsis and recruitment flyers were provided to all providers. 2.4. Data Collection Informed consent for this pilot was obtained from all CPI Group participants. CPI Group participants were asked to complete the study questionnaires (see Instruments, below) at baseline (within 1-2 weeks of initiating CPI therapy) and six months later during regularly scheduled clinic visits. CPI Group participants were also asked to complete a neurocognitive assessment within the same timeframes. The neurocognitive assessment was administered by the KUMC ADRC psychometricians under the oversight of the study team's clinical neuropsychologist. The neurocognitive tests were congruent with the annual battery (baseline and 12 months) administered to the longitudinal ADRC cohort of cognitively intact controls for the NACC UDS and are outlined below (see Table 1). 2.5. Lab Sampling and Processing The KUCC BRCF protocol permits collection of up to 6 tubes of blood from each participant at each visit. Lab sampling occurred during scheduled venipuncture for standard of care lab sampling in conjunction with participants' CPI treatment at baseline and six months. These samples were de-identified (marked only with participants' study identification numbers) and stored until samples from all timepoints from all participants were collected. At that time, the KUCC Biomarker Discovery Lab (BDL) staff obtained 1 mL of the deidentified plasma baseline and 6-month samples to quantify circulating levels of the study biomarkers using Luminex assays (Millipore-Sigma, Temecula, CA, USA) analyzed on a BioPlex 200 instrument. The mean-fluorescent intensity data generated by the Luminex assays for each marker was compared to standard curves and the resulting concentrations were supplied to the biostatistician for the analyses described below. 2.6. Self-Report Instruments CPI Group participants' self-reports of cognitive issues were measured with two instruments developed and validated by the National Institutes of Health . The Patient Reported Outcomes Management Information System (PROMIS) Cognitive Function and Cognitive Function Abilities 8-item short forms were developed to measure problems with cognitive function and the perception of cognitive ability, respectively. Items from these instruments are ranked from 1 to 5. T-scores are calculated and used for continuous variable comparisons. Potential confounders associated with changes in cognitive function include depression, anxiety, activity level, and sleep quality. Well-validated, psychometrically sound instruments to measure these potential confounders were included and outlined in Table 2. The National Alzheimer's Coordinating Center (NACC) Functional Assessment Scale also was administered to evaluate the potential impact of cognitive function on activities of daily living. 2.7. Neurocognitive Assessment An abbreviated 60 min neurocognitive assessment battery was selected from the NACC standardized neurocognitive battery administered annually to the KUMC ADRC cognitively intact cohort. All NACC tests were planned for inclusion with the exception of the Benson Complex Figure Copy and Multilingual Naming Test. This battery was further supplemented with four additional planned tests routinely administered as components of the KUMC-specific ADRC standard battery (Digit Symbol, Block Design, Stroop Test, and Letter Number Sequencing). These selections were made to specifically focus on the cognitive domains of attention/concentration, working memory, processing speed, and visuospatial ability (Table 1), and to minimize redundancy and participant burden. 2.8. Data Analyses To estimate the changes in self-report of cognitive function, performance on neurocognitive tests, and inflammatory markers, difference scores were calculated for each participant. Within-group changes in self-report of cognitive function and inflammatory markers were analyzed for the CPI Group participants. Between-group changes in performance on neurocognitive tests were analyzed for the CPI Group participants and the ADRC cognitively intact cohort. Mean difference scores were estimated, along with corresponding 95% confidence intervals. Pearson's correlation coefficients between mean difference scores for the biomarker levels and cognitive variables were examined for the CPI Group participants, or, if indicated, Spearman's correlation coefficients were used as a nonparametric alternative. For comparison to controls, linear mixed models were utilized to account for repeated measures comparing recruited versus control participants with data available from the KUMC ADRC cohort. Repeated measures using age as the time covariate enabled comparison in these measures with decline adjusted for age and allowed for comparison in slope between the recruited versus the control population to estimate and test for differences. Residual analyses were conducted, and nonparametric methods were utilized, if indicated, in place of the parametric approaches described. 2.9. Study and Recruitment Procedure Modification The pilot study was opened in September of 2020. Recruitment challenges were experienced due to a number of factors, including: reduction of in-person clinic visits due to the on-going COVID-19 pandemic, significant turnover in clinical research coordinators supporting the study, limitation to one KUCC recruitment site due to restrictions in clinical research coordinator coverage, and the prevalence of CPI therapy administered in conjunction with chemotherapy agents. To address these challenges, the eligibility criteria for the study were revised to remove the age-based restrictions and to allow previous therapy with a BRAF or tyrosine kinase inhibitor. Administration of the neurocognitive assessments was shifted from in-person to phone. This shift resulted in omission of three of the planned neurocognitive assessments: digit symbol, block design, and Stroop, and included use of the phone version of the Montreal Cognitive Assessment (MoCA-Blind) and Trail Making Tests without tasks requiring visual abilities). Total scores on the MoCA-Blind range from 0 to 22 instead of from 0 to 30 on the in-person test. The Oral Trail Making Tests require the participant to orally respond with sequential numbers (1-25, Test A) and alternating sequential numbers and letters (1-12, A-L, Test B). Scoring is based on the time in seconds required to complete the test and the number of correct responses. Despite the fact that the NACC had developed a phone version for the standardized neurocognitive battery, the KUMC ADRC implementation had not yet yielded 12-month phone assessments for the cognitively intact cohort. Thus, only the MoCA Blind assessment was available for between-group analyses. In March of 2021, the use of the Curated Clinical Outcomes Database (C3OD) was created by the University of Kansas Medical Center Department of Biostatistics and Data Science to facilitate translational cancer research in the patient population by combining disparate, complex data sources into a single, curated, referential source that is updated daily. C3OD data pulls were instituted to identify eligible patients based on the revised study outcome criteria. Refinements were made to the weekly data pulls provided to the clinical research coordinators and PI in May of 2021 and remained active throughout the remainder of study recruitment. 3. Results 3.1. Sample The cognitively intact control data were utilized for participants who had a documented MoCA-Blind score for at least two consecutive visits (about 12 months apart). To better align with the CPI Group demographics, patients older than 90 years at their baseline visit and those with more than 20 years of education were excluded. CPI Group participant recruitment was initiated in September of 2020. The first participant consented on 9/11/20 but was hospitalized prior to baseline assessment and not included in the analyses. No further participants were recruited until after the study modifications were approved in November of 2020. Upon institution and refinement of the C3OD database pulls, 21 participants were recruited between May and December 2021, bringing the number available for baseline assessments to 20 . Data collections for the 6-month assessments were completed by May of 2022. The majority of the cancer patient participants were diagnosed with melanoma (n = 12) (see Table 3). The remaining tumor types included renal cell, head and neck, urothelial, hepatocellular, and non-small cell lung cancers. Most participants received either pembrolizumab (n = 7, 33%) or nivolumab (n = 6, 28.5%). The remainder received combination therapy with either nivolumab/ipilumumab, pembrolizumab/axitinib, or atezolizumab/bevacizumab. The CPI Group participants were primarily white (95%), males (60%) with a mean of about 15 years of education (range 12-20 years) (Table 3). Little difference was noted between the cancer patient participants (at 6 months) and the ADRC controls (at 12 months) with the exception of age ranges (Table 4). Four of the CPI Group participants' ages were 56 or younger (32, 34, 43, 56 years). However, the mean age for both groups was >65. 3.2. Self-Report Instruments Descriptive statistics for the self-report instruments are listed in Appendix A (Table A1 and Table A2). No significant within-group changes were noted for the self-report instruments completed by the CPI Group participants (Table 5 and Table 6). Notably, CPI Group participants' scores for the PROMIS Cognitive Function and Cognitive Abilities 8-item short forms both decreased by 3 points. 3.3. Neurocognitive Tests Descriptive statistics for CPI Group participant scores on the cognitive tests are listed in Appendix A (Table A3). No significant within-group changes were noted for CPI Group participants' performances on the neurocognitive tests (Table 7). Comparisons of the MoCA-Blind test scores for both groups are depicted in Figure 2 and Table 8. Estimates for differences in MoCA-Blind scores after receiving CPI treatment (6-month assessments) compared to baseline (within 1-2 weeks of initiating CPI treatment) were not significant (p = 0.277). However, estimated differences for CPI Group participants' MoCA-Blind test scores compared to the ADRC cognitively intact controls approached significance at baseline (p = 0.066) and were significantly worse than the controls' scores after CPI treatment (6 months for CPI Group participants' scores compared to controls' scores at 12 months, p = 0.011). The trajectory of scores for the CPI Group participants was lower than that of the controls for participants of all ages. Given the fact that most CPI Group participants were diagnosed with melanoma (n = 12, 60%), we also conducted post facto analyses for this subgroup. Results mirrored that of the full CPI Group sample in that estimates for differences in MoCA-Blind scores after receiving CPI treatment compared to baseline were not significant (p = 0.2333). Estimated differences for the subgroup compared to the ADRC cognitively intact controls approached significance at baseline (p = 0.091) and were significantly worse than controls' scores at 12 months (p = 0.013). 3.4. Biomarkers Descriptive statistics for biomarker levels are depicted in Appendix A (Table A3). Scheduling issues prevented 6-month sampling for two participants. However, 3-month data were available for both and were included in the analyses. No significant change from baseline was noted for any of the biomarkers assessed (Table 9). Correlations for change from baseline between biomarkers and neurocognitive test scores are depicted in Table 10. Some significant correlations were noted. Inverse correlations were noted between IFNg, IL-1b, IL-2, and FGF2 and performance on the Craft story recall. Unexpected inverse correlation was noted between IL-1a and the total number of seconds needed for completion of Oral Trail Making Test B (lower time equals better performance). Positive correlation was noted between IGF-1 and the letter-number sequencing test performance. Positive correlation was noted between VEGF and digit-span backwards performance. However, inverse correlation was noted between VEGF and Craft story recall performance. 4. Discussion This pilot study yielded important information regarding the feasibility of investigating the potential cognitive impact of checkpoint inhibitor treatment. A number of challenges were experienced due to the ongoing pandemic, such as study staff attrition and reluctance of participants to attend in-person assessments. Recruitment was further complicated by limitations in study team availability to obtain consent from participants in all five of the KUCC clinical sites. Likewise, a majority of the participants screened for the cancer center site covered by the clinical research coordinators were not eligible due to planned combination therapy with chemotherapy. As a result, protocol modifications were made to relax the age restrictions and requirement for in-person study assessments. Implementation of the C30D database pulls for identification of eligible participants markedly enhanced the success of recruitment efforts. Despite the age disparity between the CPI Group participants (mean age 66) and the ADRC cognitively intact controls (mean age 80), and the fact that only the MoCA-Blind scores were available to compare between groups, a significant difference was estimated at the second assessment timepoint for patients who had received CPI treatment. Some significant correlations were demonstrated between various biomarkers and performance on neurocognitive tests. Most of these were in the expected direction, in that biomarkers known to be pro-inflammatory (IFNg, IL-1b, IL-2, and FGF2) were inversely correlated with neurocognitive performance. However, the inverse correlation between IL-1a and Oral Trail Making Test B performance was surprising. Likewise, we anticipated that VEGF levels would be positively correlated with neurocognitive performance given the known role VEGF plays in promotion of vascular endothelial cells and proliferation of neuronal precursors. One exception to this was the inverse correlation between VEGF and performance on the Craft Story Recall. A potential explanation may be that VEGF levels are known to increase in association with an inflammatory response due to its role in angiogenesis and vascular permeability. Up-regulation of VEGF is noted in conjunction with cytokine expression . Findings from this small pilot must be considered to be very preliminary, and a number of limitations must be acknowledged. Given available funding and the timing for research support, blood samples only were collected and analyzed for the CPI group, not allowing longitudinal comparisons for biomarker levels between the CPI Group and the cognitively intact controls. Neuropsychological testing was conducted at baseline and six-months for the CPI Group and compared to available data from baseline and twelve-month assessment for controls. The pilot study was not able to stratify results by tumor type, which may be an important future consideration due to the potential for variability in levels of cytokine production. In addition, the pilot study was not able to stratify by type of CPI regimen. As noted earlier, irAEs are more severe for patients receiving anti-CTLA-4 CPIs and for those receiving combination regimens. Since the initiation of this study, approvals for the promising drug class of CPIs continues to burgeon . A recent review indicates that in the classes of PD-1 and PD-L1 inhibitors approved treatment indications span 19 cancer types and 2tissue agnostic conditions (marker-based) . Given the increasing prevalence of use, determination of the cognitive impact of these agents, singly and in combination with other therapies, remains a critical need. Numerous other factors may contribute to changes in cognitive function for people with non-CNS malignancies, including comorbidities affecting oxygenation levels and pertinent nutritional deficiencies. Notably, elevation in inflammatory cytokine levels is also associated with a constellation of symptoms accompanying changes in cognitive function, sometimes referred to as sickness behavior, namely fatigue, sleep disturbance, anxiety, and depression . Patients experiencing some or all of these may report more significant cognitive changes. We did not control for these potential factors in this small feasibility pilot. The pilot study design did not include the use of neuroimaging to measure any related structural changes in the brain that may accompany treatment with CPIs and would be of interest in future studies. Future research with a larger sample is needed to address these limitations. Given the number of cancer diagnoses treated with CPIs and the variety of approved drug regimens, formation of a multi-site observational registry would be ideal to obtain the sample size needed for stratification by tumor type and CPI regimen. Collaboration among National Cancer Institute (NCI)-designated Cancer Centers whose parent institutions are also recognized as NACC ADRC centers would leverage existing infrastructure for neuropsychological assessments and access to both blood samples and neuroimaging. 5. Conclusions Results from this small observational pilot indicate that treatment with CPI(s) may have a negative impact on performance for some neurocognitive domains and contribute to the changes in cognitive function reported by individuals diagnosed with cancer. Further investigation is warranted. A multi-site study design may be critical to achieving the necessary power to critically examine the impact of CPI therapy on cognitive function. A potential solution may be the establishment of a multi-site observational registry. Partnerships with National Cancer Institute-designated Cancer Centers, whose parent institutions also are recognized as NACC ADRC centers, may provide the necessary standardization of infrastructure to conduct further investigation. Providers caring for patients receiving CPIs should be aware of this potential effect as they assess patients for treatment-related sequalae. Acknowledgments The authors' thanks to Suzanne Hunt, MS, MA for developing the electronic database for documentation of data collection. Author Contributions Conceptualization, J.S.M., A.C.P., J.D.M., H.B.P. and J.M.B.; methodology, J.S.M., A.C.P., H.B.P. and J.M.B.; software, D.S.; validation, A.C.P. and H.B.P.; formal analysis, J.D.M. and K.J.Y.; investigation, A.U., P.S., Y.A., S.S. and J.F.W.; resources, J.S.M., H.B.P. and J.M.B.; data curation, H.B.P., R.V.P., and K.J.Y.; writing--original draft preparation, J.M.B.; writing--review and editing, J.S.M., A.C.P., J.D.M., K.J.Y., H.B.P., R.V.P., A.U., P/S., Y.A., S.S., J.F.W., D.S. and J.M.B.; visualization, J.S.M. and K.J.Y.; supervision, J.S.M., A.C.P., J.D.M., H.B.P. and J.M.B.; project administration, J.S.M.; funding acquisition, J.S.M. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of the University of Kansas Medical Center (STUDY00146028, protocol code, 7 July 2020). Informed Consent Statement Informed consent was obtained from all participants involved in the study. Data Availability Statement The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy restrictions. Conflicts of Interest The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. Appendix A cancers-15-01615-t0A1_Table A1 Table A1 Self-report instruments: overall scores. Score Baseline (N = 20) Month 6 (N = 13) Change (N = 13) Geriatric Depression Scale N 20 13 13 Mean (SD) 3.3 (3.1) 3.9 (2.7) 1.2 (2.8) Median (IQR) 2.5 (1.0, 5.0) 4.0 (2.0, 6.0) 1.0 (0.0, 2.0) Min, Max 0, 11 0, 9 -4, 8 Geriatric Anxiety Scale N 20 13 13 Mean (SD) 4.1 (4.1) 5.6 (2.6) 1.0 (4.5) Median (IQR) 3.0 (1.0, 7.2) 6.0 (5.0, 7.0) 3.0 (-1.0, 4.0) Min, Max 0, 15 0, 10 -9, 6 PROMIS Cognitive Function T-Score N 20 13 13 Mean (SD) 51.08 (8.86) 47.65 (8.07) -2.45 (11.00) Median (IQR) 47.95 (44.95, 57.52) 49.10 (41.30, 53.30) -2.00 (-4.60, 3.70) Min, Max 35.8, 63.9 32.1, 59.3 -24.3, 17.2 PROMIS Cognitive Abilities T-Score N 20 13 13 Mean (SD) 54.51 (9.85) 51.62 (6.91) -3.74 (9.81) Median (IQR) 53.30 (50.05, 63.80) 51.70 (46.50, 55.00) -0.90 (-10.70, 1.00) Min, Max 37.5, 67.1 42.5, 67.1 -19.2, 13.2 NACC Functional Activities Scale N 20 13 13 Mean (SD) 2.0 (2.8) 2.5 (5.2) 0.5 (4.7) Median (IQR) 0.0 (0.0, 4.0) 1.0 (0.0, 2.0) 0.0 (-3.0, 1.0) Min, Max 0, 8 0, 19 -6, 13 Pittsburgh Sleep Quality Index N 20 13 13 Mean (SD) 7.4 (2.8) 7.8 (3.6) -0.3 (4.2) Median (IQR) 7.0 (5.8, 9.0) 9.0 (6.0, 9.0) 0.0 (-3.0, 1.0) Min, Max 3.0, 15.0 1.0, 16.0 -6.0, 8.0 IPAQ MET-minutes/week N 20 13 13 Mean (SD) 1957 (3149) 3402 (4473) 2385 (3978) Median (IQR) 346 (235, 2492) 758 (452, 6144) 461 (240, 3923) Min, Max 0, 12,852 254, 14,196 -3076, 11,805 IPAQ Activity Level, n (%) Low 11 (55.0%) 3 (23.1%) Moderate 3 (15.0%) 5 (38.5%) High 6 (30.0%) 5 (38.5%) cancers-15-01615-t0A2_Table A2 Table A2 PSQI scores--descriptive statistics. Score Baseline (N = 20) Month 6 (N = 13) Daytime Dysfunction Due to Sleepiness, n (%) 0 8 (40.0%) 3 (23.1%) 1 11 (55.0%) 10 (76.9%) 2 1 (5.0%) 0 (0.0%) Duration of Sleep, n (%) 0 10 (50.0%) 9 (69.2%) 1 7 (35.0%) 2 (15.4%) 2 2 (10.0%) 1 (7.7%) 3 1 (5.0%) 1 (7.7%) Need Meds to Sleep, n (%) 0 9 (45%) 5 (38%) 1 2 (10%) 2 (15%) 2 2 (10%) 2 (15%) 3 7 (35%) 4 (31%) Overall Sleep Quality, n (%) 0 3 (15%) 1 (7.7%) 1 16 (80%) 10 (77%) 2 0 (0%) 2 (15%) 3 1 (5.0%) 0 (0%) Sleep Disturbance, n (%) 1 8 (40%) 6 (46%) 2 11 (55%) 7 (54%) 3 1 (5.0%) 0 (0%) Sleep Efficiency, n (%) 0 8 (40%) 6 (46%) 1 7 (35%) 3 (23%) 2 3 (15%) 3 (23%) 3 2 (10%) 1 (7.7%) Sleep Latency, n (%) 0 7 (35%) 3 (23%) 1 5 (25%) 3 (23%) 2 6 (30%) 3 (23%) 3 2 (10%) 4 (31%) cancers-15-01615-t0A3_Table A3 Table A3 Biomarkers descriptive statistics. Test Baseline (N = 18) Month 3 (N = 2) Month 6 (N = 10) Change (N = 12) IFNg (pg/mL) N 16 1 8 9 Mean (SD) 281.32 (751.18) 0.42 (NA) 227.37 (539.92) 36.64 (412.08) Median (IQR) 8.52 (2.40, 40.55) 0.42 (0.42, 0.42) 16.05 (10.62, 96.47) 0.44 (-6.55, 4.71) Min, Max 0.28, 2969.00 0.42, 0.42 0.71, 1559.00 -632.00, 985.00 IL-1a (pg/mL) N 12 0 8 5 Mean (SD) 134.32 (234.47) 153.87 (286.97) 35.84 (78.58) Median (IQR) 7.67 (0.60, 156.00) 11.00 (3.09, 178.25) 1.31 (-0.10, 40.00) Min, Max 0.02, 666.00 0.49, 835.00 -31.00, 169.00 IL-1b (pg/mL) N 17 2 10 11 Mean (SD) 10.96 (31.06) 0.21 (0.16) 5.72 (12.26) 0.21 (4.85) Median (IQR) 0.76 (0.43, 3.37) 0.21 (0.15, 0.26) 0.60 (0.33, 4.77) -0.04 (-0.25, 0.07) Min, Max 0.18, 128.00 0.09, 0.32 0.14, 40.00 -9.82, 11.70 IL-2 (pg/mL) N 17 2 10 11 Mean (SD) 10.47 (29.67) 0.25 (0.19) 5.26 (10.20) 0.16 (4.38) Median (IQR) 0.89 (0.51, 3.82) 0.25 (0.18, 0.31) 0.71 (0.38, 5.33) -0.06 (-0.29, 0.07) Min, Max 0.22, 123.00 0.11, 0.38 0.16, 33.40 -9.00, 10.40 TNFa (pg/mL) N 18 2 10 12 Mean (SD) 18.95 (27.47) 9.07 (4.71) 32.19 (58.72) 9.21 (23.27) Median (IQR) 9.43 (7.46, 16.05) 9.07 (7.40, 10.74) 10.37 (7.28, 13.05) 0.23 (-1.06, 4.18) Min, Max 4.16, 122.00 5.74, 12.40 3.34, 193.00 -9.16, 71.00 VEGF (pg/mL) N 16 2 9 10 Mean (SD) 173.70 (238.96) 16.45 (5.59) 142.94 (191.27) -9.68 (146.34) Median (IQR) 74.85 (29.90, 207.75) 16.45 (14.47, 18.42) 51.70 (19.60, 236.00) 6.90 (-8.89, 13.97) Min, Max 4.26, 828.00 12.50, 20.40 1.92, 587.00 -380.00, 214.00 IL-6 (pg/mL) N 18 2 10 12 Mean (SD) 23.25 (51.21) 0.45 (0.21) 46.46 (81.00) 5.76 (16.83) Median (IQR) 3.05 (1.09, 7.68) 0.45 (0.37, 0.52) 6.60 (1.08, 60.15) 0.13 (-0.17, 2.30) Min, Max 0.18, 192.00 0.30, 0.59 0.07, 250.00 -3.00, 58.00 FGF2 (pg/mL) N 17 2 10 11 Mean (SD) 98.34 (194.08) 2.23 (1.44) 138.74 (308.81) 22.67 (113.09) Median (IQR) 9.73 (6.43, 72.00) 2.23 (1.72, 2.74) 19.30 (8.50, 102.38) 3.39 (-1.59, 4.95) Min, Max 0.30, 695.00 1.21, 3.25 3.70, 1006.00 -160.00, 311.00 IFG-1 (pg/mL) N 18 2 10 12 Mean (SD) 24,090.8 (12,976.1) 23,656.5 (12,414.7) 28,702.8 (15,486.7) 1811.1 (2367.9) Median (IQR) 23,510.5 (15,179.0, 30,270.5) 23,656.5 (19,267.2, 28,045.8) 26,859.0 (21,420.2, 37,906.2) 1811.0 (513.5, 3124.5) Min, Max 484, 56,998 14,878, 32,435 1317, 56,074 -2520, 6231 BDNF (pg/mL) N 18 2 10 12 Mean (SD) 17,816.4 (9096.7) 24,367.5 (3315.6) 18,398.7 (8882.2) 2787.6 (7174.0) Median (IQR) 14,954.5 (10,910.2, 21,800.0) 24,367.5 (23,195.2, 25,539.8) 18,626.0 (14,983.5, 21,157.2) 1934.0 (-3278.8, 8893.8) Min, Max 8553, 41,930 22,023, 26,712 6275, 37,604 -7446, 15,749 Figure 1 Consort diagram. * Curated Clinical Outcomes Database. Figure 2 MoCA-Blind scores by cohort and timepoint. cancers-15-01615-t001_Table 1 Table 1 Neurocognitive assessment tests. Measure Domain Brief Description Phone Adaptation Montreal Cognitive Assessment (MoCA) Global cognitive score for short-term memory recall, visuospatial ability, attention/concentration, working memory, and abstract reasoning Participants complete a series of tasks including recall of five nouns, a clock-drawing and three-dimensional cube copy test, target detection, serial subtraction, a three-item naming task, similarity description task, and orientation to time and place. Scores range from 0 to 30. MoCA Blind Assessment excludes clock-drawing and three-dimensional cube copy test Craft Story 21 Recall (immediate and delayed) Episodic memory Participants are read a short story and asked to repeat as much as they can remember immediately upon hearing the story and following a 20 min delay. Verbatim scores range from 0 to 44 correct words. Paraphrase scores range from 0 to 25 correct components. Digit Span Forward and Backward Working memory Number sequences are presented verbally in ascending order of length. Participants are asked to recall the numbers in both forward and backward sequences. Scored by number of correct trials and longest correct sequence span. Forward span trials range from 0 to 14 with spans of 3-9. Backward span trials range from 0 to 14 with spans of 2-8. Trail Making Test (TMT) A Processing speed Participants connect 25 circles containing numbers in numerical order as quickly as they can. Score includes number of seconds to complete, number of correct lines, and number of errors. Oral Trail Making A. Participants orally respond with sequential numbers 1-25. Trail Making Test (TMT) B Executive function Participants are asked to connect 25 circles containing either numbers or letters, alternating numerical with alphabetical order as quickly as they can. Scoring same as above. Oral Trail Making B. Participants orally respond with alternating sequential numbers and letters (1-12, A-L) Verbal Fluency (F and L) Verbal fluency Participants name as many words as they can in one minute starting with a specified target letter. Category Fluency (animals and vegetables) Category fluency Participants name as many words as they can in one minute within a specified target category. Digit Symbol Processing speed, attention, visuo-perceptual function, and executive function Participants are asked to use a key to match the randomly presented numbers 1-9 with easy to draw symbols in a timed test (90 or 120 s, up to 100 items). Scores equal number of correct answers Omit Block Design Visuospatial ability Participants are asked to use 3-dimensional cubes to replicate a series of up to 14 figures ascending in complexity in a timed test. Scores include the number of correct designs replicated within the time allowed. A total of 4 pts are awarded for each correct figure (maximum score = 48). Time bonuses also may be calculated if desired. Omit Stroop Test Selective attention, inhibition, and processing speed Participants are presented with a list of words that name common colors (e.g., red, blue, green). The words are printed in colors that differ from the meaning of the word (e.g the word red is written in blue ink). Participants must read the word as listed, not the color in which it is printed. Scores include response time and error rate. Omit Letter-Number Sequencing (LNS) Working memory Participants are verbally presented with random lists of letters and numbers in ascending length. Participants are asked to repeat the series listing the numbers in numerical order and the letters in alphabetical order. cancers-15-01615-t002_Table 2 Table 2 Self-report instruments. Measure Description PROMIS * Cognitive Function 8a An 8-item Likert-style short form to assess participants perceptions of cognitive problems. Items are ranked from 1 to 4. Raw scores are converted to T-scores and standard error. Higher scores indicate better cognitive function. PROMIS Cognitive Abilities 8a As above, but measures participants' perceptions of cognitive abilities. Geriatric Depression Scale (GDS) A 15-item short form, 1 point for each "yes" answer. Higher scores indicate more depression. Geriatric Anxiety Scale (GAS) A 10-item Likert style ranking from 0 to 3. Higher scores indicate more anxiety. NACC ** Functional Assessment Scale A 10-item Likert-style form ranked 0-3 with higher scores indicating higher levels of dependence in activities of daily living. Pittsburgh Sleep Quality Index (PSQI) Narrative and Likert-style instrument measuring 7 components of sleep quality. Yields a global score of sleep quality (0-21). Lower scores indicate better sleep quality International Physical Activity Questionnaire (IPAQ) A 7-item measure of low, moderate, and vigorous activity. Yields a metabolic equivalents (MET) total in minutes per week. * Patient Reported Outcomes Measurement Information System. ** National Alzheimer's Coordinating Center. cancers-15-01615-t003_Table 3 Table 3 CPI group demographics baseline vs. month 6. Baseline (N = 20) Month 6 (N = 13) Age (years) Mean (SD) 68.1 (15.3) 66.4 (13.9) Median [Min, Max] 73.5 [32.0, 88.0] 72.0 [34.0, 81.0] Sex Female 8 (40.0%) 5 (38.5%) Male 12 (60.0%) 8 (61.5%) Education (years) Mean (SD) 14.9 (3.03) 15.0 (2.94) Median [Min, Max] 14.5 [12.0, 20.0] 16.0 [12.0, 20.0] Ethnicity Not Hispanic 20 (100%) 13 (100%) Race Black or African American 1 (5.0%) 1 (7.7%) White 19 (95.0%) 12 (92.3%) Tumor Type Head and neck squamous cell carcinoma of hard palate 1 (5.0%) 1 (7.7%) Hepatocellular 2 (10.0%) 2 (15.4%) Melanoma 12 (60.0%) 7 (53.8%) Non-small cell lung cancer 1 (5.0%) 0 (0%) Renal Cell 2 (10.0%) 2 (15.4%) Squamous cell carcinoma of orbit 1 (5.0%) 0 (0%) Urothelial 1 (5.0%) 1 (7.7%) Number of Visits 1 7 (35.0%) 0 (0%) 2 13 (65.0%) 13 (100%) cancers-15-01615-t004_Table 4 Table 4 ADRC controls vs. CPI group at month 6--demographics. Control (N = 13) CPI (N = 13) Overall (N = 26) Age (years) Mean (SD) 80.8 (6.10) 66.4 (13.9) 73.6 (12.8) Median [Min, Max] 82.0 [70.0, 88.0] 72.0 [34.0, 81.0] 75.5 [34.0, 88.0] Sex Female 8 (61.5%) 5 (38.5%) 13 (50.0%) Male 5 (38.5%) 8 (61.5%) 13 (50.0%) Education (years) Mean (SD) 16.4 (2.22) 15.0 (2.94) 15.7 (2.65) Median [Min, Max] 16.0 [12.0, 20.0] 16.0 [12.0, 20.0] 16.0 [12.0, 20.0] Ethnicity Not Hispanic 13 (100%) 13 (100%) 26 (100%) Race Black or African American 2 (15.4%) 1 (7.7%) 3 (11.5%) White 11 (84.6%) 12 (92.3%) 23 (88.5%) Number of Visits 2 13 (100%) 13 (100%) 26 (100%) cancers-15-01615-t005_Table 5 Table 5 Overall scores for self-report instruments--change from baseline estimates. Test Estimate Standard Error 95% Confidence Interval p-Value PROMIS Cognitive Abilities T-Score -3.195 2.629 -8.981, 2.592 0.250 PROMIS Cognitive Function T-Score -3.239 2.789 -9.377, 2.899 0.270 Functional Activities Scale 0.509 1.177 -2.081, 3.099 0.673 Geriatric Anxiety Scale 1.408 1.150 -1.122, 3.939 0.246 Geriatric Depression Scale 0.922 0.763 -0.756, 2.601 0.252 IPAQ MET--minutes/week 1514.886 1062.387 -823.412, 3853.184 0.182 Pittsburgh Sleep Quality Index 0.248 1.015 -1.987, 2.483 0.812 cancers-15-01615-t006_Table 6 Table 6 PSQI scores--change from baseline estimates. Test Estimate Standard Error 95% Confidence Interval p-Value Daytime Dysfunction Due to Sleepiness 0.128 0.181 -0.271, 0.527 0.494 Duration of Sleep -0.198 0.285 -0.826, 0.430 0.502 Need Meds to Sleep -0.033 0.424 -0.966, 0.900 0.940 Overall Sleep Quality * 1.129 0.398 0.520, 2.452 0.737 Sleep Disturbance * 0.933 0.265 0.500, 1.742 0.812 Sleep Efficiency * 0.948 0.357 0.414, 2.172 0.889 Sleep Latency 0.390 0.342 -0.363, 1.143 0.279 * Modeled assuming a Poisson distribution, indicating a multiplicative change from baseline. cancers-15-01615-t007_Table 7 Table 7 Cognitive scores--change from baseline estimates. Test Estimate Standard Error 95% Confidence Interval p-Value 1a. MoCA-Blind TOTAL RAW SCORE--UNCORRECTED -0.260 0.831 -2.088, 1.569 0.760 2a. Total Craft story units recalled, verbatim scoring 2.833 1.960 -1.482, 7.148 0.176 2b. Total Craft story units recalled, paraphrase scoring 0.781 1.311 -2.104, 3.665 0.563 3a. Digit symbol forward number of correct trials 0.505 0.460 -0.507, 1.517 0.296 3b. Longest span forward 0.203 0.267 -0.384, 0.790 0.462 4a. Digit symbol backward number of correct trials 0.576 0.542 -0.616, 1.768 0.311 4b. Longest span backward 0.429 0.318 -0.270, 1.128 0.204 5a. Trail Making Test A: Total number of seconds to complete -0.722 0.860 -2.614, 1.169 0.419 5b. Trail Making Test B: Total number of seconds to complete -16.855 14.772 -49.369, 15.658 0.278 5b1. Number of commission errors -1.025 0.632 -2.415, 0.365 0.133 6a. Category Fluency number of animals -0.751 1.089 -3.148, 1.646 0.505 6b. Category Fluency number of vegetables 1.310 0.997 -0.885, 3.505 0.216 7a. Verbal Fluency number of correct F-words generated in 1 min 0.818 1.077 -1.552, 3.187 0.464 7b. Verbal Fluency number of F-words repeated in 1 min 0.231 0.371 -0.586, 1.048 0.547 7d. Verbal Fluency number of correct L-words generated in 1 min 0.632 1.000 -1.569, 2.832 0.541 7g. Verbal Fluency total number of correct F-words and L-words * 1.072 0.086 0.898, 1.280 0.405 7h. Verbal Fluency total number of F-words and L-words repetition errors * 1.343 0.454 0.638, 2.826 0.401 7i. Verbal Fluency total number of non-F/L words and rule violation errors * 0.099 0.103 0.010, 0.972 0.048 8a. Total Craft story delayed units recalled, verbatim scoring 1.374 1.765 -2.511, 5.258 0.453 8b. Total story delayed units recalled, paraphrase scoring -0.261 1.055 -2.583, 2.062 0.809 9a. Letter number sequencing 1.791 1.249 -0.991, 4.574 0.182 * Modeled assuming a Poisson distribution, indicating a multiplicative change from baseline value. Note: Change from baseline was not estimated for the following tests due to low variability in the responses: 5a1. Number of commission errors; 5a2. Number of correct lines; 5b2. Number of correct lines; 7c. Number of F-words and rule violation errors in 1 min; 7e. Number of L-words repeated in 1 min; 7f. Number of non-L-words and rule violation errors in 1 min; 8c. Delay time (minutes); 8d. Cue (boy) needed. cancers-15-01615-t008_Table 8 Table 8 Estimates for differences in MOCA-Blind scores. Cohort/Timepoint Estimate 95% Confidence Interval p-Value CPI Group After Treatment vs. Before Treatment -0.730 -2.085, 0.624 0.277 CPI Group Before Treatment vs. ADRC Controls -1.735 -3.591, 0.122 0.066 CPI Group After Treatment vs. ADRC Controls -2.465 -4.304, -0.627 0.011 cancers-15-01615-t009_Table 9 Table 9 Biomarkers--change from baseline estimates. Biomarker Estimate 95% Confidence Interval p-Value BDNF (pg/mL) 1.115 0.854, 1.454 0.385 FGF2 (pg/mL) 1.470 0.724, 2.985 0.250 IFG-1 (pg/mL) 1.180 0.984, 1.415 0.069 IFNg (pg/mL) 0.650 0.245, 1.723 0.331 IL-1a (pg/mL) 1.106 0.856, 1.430 0.299 IL-1b (pg/mL) 0.807 0.600, 1.087 0.139 IL-2 (pg/mL) 0.809 0.608, 1.077 0.128 IL-6 (pg/mL) 0.984 0.575, 1.683 0.947 TNFa (pg/mL) 1.130 0.823, 1.552 0.411 VEGF (pg/mL) 1.030 0.675, 1.572 0.875 Biomarkers were modeled assuming a log-normal distribution. Therefore, the estimates indicate a multiplicative change from baseline. cancers-15-01615-t010_Table 10 Table 10 Pearson correlations for change from baseline between biomarkers and cognitive scores. Test (pg/mL) IFN IL-1a IL-1b IL-2 TNFa VEGF IL-6 FGF2 IGF-1 BDNF 1a. MoCA-Blind TOTAL RAW SCORE--UNCORRECTED 0.064 -0.047 0.091 0.098 -0.119 0.388 -0.223 0.076 0.219 0.579 2a. Total Craft story units recalled, verbatim scoring -0.811 * -0.946 -0.820 * -0.820 * -0.412 -0.885 * -0.534 -0.716 * 0.217 -0.423 2b. Total Craft story units recalled, paraphrase scoring -0.835 * -0.945 -0.799 * -0.798 * -0.458 -0.910 * -0.519 -0.710 * 0.217 -0.473 3a. Digit symbol forward number of correct trials -0.645 -0.909 -0.524 -0.520 -0.213 -0.467 -0.475 -0.400 0.312 -0.571 3b. Longest span forward -0.538 -0.845 -0.468 -0.466 -0.110 -0.498 -0.456 -0.496 0.554 -0.462 4a. Digit symbol backward number of correct trials 0.344 0.576 0.314 0.317 0.167 0.716 0.166 0.273 -0.004 0.548 4b. Longest span backward 0.367 0.576 0.373 0.379 0.291 0.862 * 0.282 0.352 -0.219 0.660 5a. Trail Making Test A: Total number of seconds to complete -0.175 -0.685 -0.133 -0.126 0.037 -0.062 -0.053 -0.313 -0.222 -0.053 5b. Trail Making Test B: Total number of seconds to complete -0.192 -0.977 * -0.209 -0.209 -0.027 -0.352 -0.165 -0.424 0.412 -0.296 6a. Category Fluency number of animals 0.123 0.012 0.086 0.086 0.322 0.021 0.168 0.174 -0.133 0.315 6b. Category Fluency number of vegetables -0.749 -0.750 -0.334 -0.339 -0.405 -0.719 -0.372 -0.314 0.241 0.415 7a. Verbal Fluency total number of correct F-words generated in 1 min -0.212 -0.830 -0.181 -0.177 0.031 -0.086 -0.064 -0.073 -0.150 -0.033 7b. Verbal Fluency total number of F-words repeated in 1 min -0.055 -0.331 -0.025 -0.022 0.304 0.157 -0.131 0.006 0.232 -0.074 7d. Verbal Fluency total number of correct L-words generated in 1 min -0.375 -0.560 -0.385 -0.391 -0.027 -0.531 -0.141 -0.169 0.050 -0.391 7g. Verbal Fluency total number of correct F-words and L-words -0.336 -0.746 -0.334 -0.336 -0.006 -0.414 -0.127 -0.140 -0.042 -0.244 7h. Verbal Fluency total number of F-words and L-words repetition errors -0.078 -0.347 -0.047 -0.043 0.230 0.157 -0.248 -0.116 0.424 -0.086 7i. Verbal Fluency total number of non-F/L words and rule violation errors 0.081 0.331 0.005 0.000 -0.028 -0.161 0.032 -0.221 0.278 0.121 8a. Total Craft story delayed units recalled, verbatim scoring -0.420 -0.988 * -0.378 -0.370 -0.034 -0.451 -0.152 -0.284 -0.195 -0.040 8b. Total Craft story delayed units recalled, paraphrase scoring -0.571 -0.919 -0.582 -0.580 -0.309 -0.705 -0.298 -0.472 -0.083 -0.053 8c. Delay time (minutes) 0.162 0.295 0.184 0.192 0.158 0.222 0.242 0.114 -0.360 -0.287 13. Letter number sequencing 0.164 -0.070 0.135 0.132 0.296 0.215 -0.159 -0.035 0.707 * -0.068 * Correlation is significant (p-value < 0.05). Note: Correlations were not estimated for the following tests due to low variability in the responses: 5a1. Number of commission errors; 5a2. Number of correct lines; 5b1. Number of commission errors. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000600 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12050997 foods-12-00997 Article Assessing Sensory Attributes and Properties of Infant Formula Milk Powder Driving Consumers' Preference Xi Yanmei Conceptualization Methodology Investigation Data curation Writing - original draft Writing - review & editing 1 Zhao Tong Formal analysis Investigation 1 Liu Ruirui Writing - review & editing 1 Song Fuhang Investigation Writing - review & editing 1 Deng Jianjun Resources Writing - review & editing 2* Ai Nasi Methodology Writing - review & editing Supervision Project administration Funding acquisition 1* Santini Antonello Academic Editor 1 Beijing Advanced Innovation Center for Food Nutrition and Human Health, School of Food and Health, Beijing Technology & Business University, Beijing 100048, China 2 State Key Laboratory of Vegetable Biobreeding, Institute of Vegetables and Flowers, Chinese Academy of Agricultural Sciences, Beijing 100081, China * Correspondence: [email protected] (J.D.); [email protected] (N.A.) 26 2 2023 3 2023 12 5 99707 2 2023 18 2 2023 24 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Infant formula milk powder (IFMP) is an excellent substitute for breast milk. It is known that the composition of maternal food during pregnancy and lactation and exposure level to food during infancy highly influence taste development in early infancy. However, little is known about the sensory aspects of infant formula. Herein, the sensory characteristics of 14 brands of infant formula segment 1 marketed in China were evaluated, and differences in preferences for IFMPs were determined. Descriptive sensory analysis was performed by well-trained panelists to determine the sensory characteristics of evaluated IFMPs. The brands S1 and S3 had significantly lower astringency and fishy flavor compared to the other brands. Moreover, it was found that S6, S7 and S12 had lower milk flavor scores but higher butter scores. Furthermore, internal preference mapping revealed that the attributes fatty flavor, aftertaste, saltiness, astringency, fishy flavor and sourness negatively contributed to consumer preference in all three clusters. Considering that the majority of consumers prefer milk powders rich in aroma, sweet and steamed flavors, these attributes could be considered for enhancement by the food industry. IFMP sensory analysis K-means internal preference mapping National Natural Science Foundation of P. R. China32272460 Beijing Outstanding Young Scientist ProgramBJJWZYJH01201910011025 This study was supported by the National Natural Science Foundation of P. R. China (32272460) and Beijing Outstanding Young Scientist Program (BJJWZYJH01201910011025). pmc1. Introduction Breastfeeding is the best strategy for feeding newborn babies and provides a balanced supply of nutrients to ensure absorption and supports optimal infant growth and development. With advances in science and technology, the current understanding on the immunological, hormonal and nutritional properties of breast milk and how its composition can be tailored to the infant's unique needs and environment has increased . However, insufficient or lack of provision of breast milk may occur due to personal and social factors. Compared to other complementary foods, in the early stages of infant's incomplete digestion and absorption physiology, IFMP is often used as a substitute for breast milk to meet the infant's special nutritional needs while also reducing the stress of nursing mothers . Currently, the demand for infant formula has been increasing in the Chinese market along with the gradual implementation of the national three-child policy. The sensory profile is the combination of sensory impressions which directly determine the consumer terminal and corporate image. Previous studies have found that the development of newborn's gustatory functions during the typical forty-week gestation is highly related to the food consumed by the mother during pregnancy as well as later during infancy and childhood . Other studies revealed that baby facial expressions can be quantified to determine the development of the infant's taste preferences, which showed that taste development during infant's early development is closely related to the composition of foods consumed during pregnancy and lactation as well as to the degree of food contact in infants and the ability to receive food during childhood . In addition, it was found that taste receptivity of adult rats that were exposed to a certain taste during early postnatal development increased with age, which influenced subsequent physiology and behavior . Moreover, infants lack the ability to discriminate between actual breast milk and infant formula since their language function is not perfect; hence, prospective parents are particularly concerned about the taste of infant formula. However, current research is primarily concerned with nutrition and quality control of milk powder . Masum et al. comprehensively and critically reviewed the available information on the effects of product formulation, processing and storage processes on the physicochemical properties of end-of-life infant formulae. Phosanam et al. established correlations between storage environment, lactose crystallization, surface components of IMF powders and their agglomeration (including agglomeration intensity) to extend the shelf life of infant formula milk powder. Gredilla et al. propose a rapid, simple and conclusive analytical method based on chemometrics for assessing the toxicity of hazardous elements in commercial infant formula. In our research, we found that consumers pay more attention to the sensory quality of infant formula when buying infant formula. Only a few studies have considered the sensory quality of this product. Therefore, the aim of this study was to investigate the sensory attributes of fourteen popular brands of powdered milk currently found in the Chinese market by a sensory panel. In addition, the factors affecting buying behavior, descriptive sensory analysis and consumer testing of powdered milk were analyzed by multiple statistical analyses. Moreover, internal preference mapping was used to analyze the factors affecting buying of infant milk powder in terms of sensory preferences. The results discussed herein provide insights into the sensory attributes of milk powder marketed in China and highlight consumer's preferences as well as elucidate the preliminary sensory factors that block consumer acceptance of milk powder, thereby serving as a theoretical guide for both the manufacture and consumer choice related to milk powder. Distinct from previous studies, this paper highly focuses on the flavor quality of infant formula to promote the "flavor and health-oriented" development of the food industry. 2. Materials and Methods 2.1. Sample Collection IFMPs from 14 popular brands currently available in the Chinese market under various trademark names were purchased at different Yonghui supermarket branches (Beijing, China). These brands were selected for the extension of their market share in China as well as for their representation on a survey conducted in fifty mother-infant households for two months. Milk powder samples were purchased in bulk, numbered with a label within the range of S1-S14, and sent to the dairy flavor laboratory at the Beijing Technology and Business University for subsequent analysis. Milk powder from the following regions was evaluated herein: S1 and S2 samples came from the Heilongjiang region, while S3, S4 and S5 were produced in the Inner Mongolia region. The Beijing region produced S6 and S7, sample S8 was from Zhejiang, Guangdong company provided S9, S10, S11 and S12, while S13 and S14 originated from Hong Kong and Taiwan, respectively. Basic information was collected on the 14 samples (Table 1), including sample name, protein, fat and carbohydrate percentage. 2.2. Screening of Sensory Panelists Participants (aged 23-35 years) were enrolled in sensory analysis tests based on their ability to judge olfactory attributes, sensitivity to the evaluated attributes and verbal description. Participants were from the Dairy Flavor Chemistry Lab and had extensive experience in evaluating milk powder. The sensory evaluators are in good health, equipped with the ability to focus and remain free from outside influences. The candidate demonstrates interest and motivation in the sensory analysis and is able to attend sensory evaluations on time. In total, 56 panelists (45 females and 11 males) were recruited for further guidance and learning, who voluntarily participated in sensory evaluation of milk powder samples, and each panelist received an individual identification label. 2.3. Training of Sensory Panelists Sensory panelists were trained to improve their ability to perceive, identify and describe the sensory attributes of milk powder . Panelists were trained for 3 h each morning for 14 days. In total, nine sensory attributes were identified at the end of training, which included milk flavor, sweetness, aftertaste, saltiness, butter taste, astringent taste, fishy taste, sourness and steamed flavor. All samples were evaluated in two independent replicates, and three sensory attributes (i.e., butter taste, astringent taste, and aftertaste) during training were submitted to repeatability tests. 2.4. Descriptive Sensory Analysis Descriptive sensory analysis included the analysis of aroma and taste of infant formula milk powder. Descriptive sensory evaluation criteria are shown in Table 2. Sensory evaluation applied the principles of experimental design and statistical analysis to the use of human senses, thus aiming to isolate the sensory properties of foods and to provide information on sensory properties of food products . Formal sensory evaluation was conducted in a dairy flavor sensory evaluation laboratory at room temperature (25-28 degC) and under bright and soft LED lighting. Each sensory evaluator had an identical dedicated space free of disturbances. Each sample was dissolved in water at 25 degC, poured into a 50 mL evaluation cup labeled with a three-digit random number, and presented to each evaluator in random order. After identifying the sensory attributes of milk powder samples, their intensity was scored in a subsequent sensory evaluation . Participants had free sensory evaluation preference tests that were conducted based on a previous study using a nine-point scale to assess the intensity of sensory attributes of milk powder samples, in which 1 indicated the absence of the attribute, 5 indicated a moderately strong intensity, and 9 indicated very strong intensity. Participants rated the aroma and taste of milk powder samples, and they undertook a short (2-3 min) break after each sample evaluation. Participants had free access to water and sugar-free soda crackers for palate cleaning during sensory evaluation rounds. Data from 56 participants were recorded and summarized after evaluation. 2.5. Statistical Analysis Sensory evaluation data were analyzed with SPSS 16.0 (IBM Deutschland GmbH, Ehningen, Germany) for one-way ANOVA and significance analysis by Duncan's multiple-range tests. In addition, principal component analysis (PCA) was applied to explain the relationship between sensory attributes and sample characteristics using PanelCheck (Version 1.4.2). Furthermore, hierarchical cluster analysis (HCA) was performed on sample groups with similar sensory attributes and analytical characteristics based on Ward's linkage using Euclidean distances. Additionally, panelists were segmented via the k-means clustering method. To further understand consumer's preferences, the k-means clustering method was combined with PCA for external preference mapping. Finally, consumer preference scores were correlated with attribute intensity rated for descriptive analysis. 3. Results and Discussion 3.1. Sensory Evaluation Skills Assessment It is known that the results of sensory evaluation analysis are influenced by the evaluator's assessment ability . The results obtained herein from 56 sensory evaluators were recorded and imported into Panel Check for subsequent analysis. In the case of milk powder sensory data, it means that the MSE values for a total of three attributes (butter, astringency, and aftertaste) can be visualized in one graph. F values, mean squared error (MSE) values and p * MSE are usually used to assess the ability of sensory evaluators . F-value is the ratio of the difference between sample groups to the difference within groups; in general, if there is a larger F value on sensory attributes, it indicates that the evaluator has a better discrimination level of the sample. The MSE value indicates within-group variance and provides a measure for the reproducibility of evaluators' judgement; lower MSE values indicate better repeatability of evaluators' judgement . All 56 sensory evaluators indicated low MSE values (MSE < 3.5) for three sensory attributes, which revealed good repeatability of results after a fourteen-day training, thus ensuring the accuracy of experimental results. 3.2. Descriptive Sensory Analysis of Milk Powder Flavor is an important attribute which determines consumer's acceptance and preference of infant formula milk powder. Moreover, it has been shown that sensory properties of food products have been currently driving consumer liking . In addition, the manufacture and promotion of foods are intrinsically associated with sensory evaluation. Thus, descriptive sensory analysis was conducted to evaluate sensory properties of commercial milk powder, and the results were analyzed by ANOVA to determine significance of observed differences in sensory properties of descriptive sensory analysis. In total, the 56 trained panelists proposed nine sensory attributes to describe milk powder samples. The significance of differences in the nine sensory attributes--butter, sweetness, saltiness, sourness, aftertaste, astringent taste, milk flavor, steamed flavor and fishy flavor and preference--are shown in Table 3. Figure 2 depicts the results of quantitative descriptive analysis from the radar plot and evaluation scores. The results showed that the milk flavor and sweetness of sample S3 were more pronounced compared to other samples. However, no significant differences were observed for sourness. In addition, fishy flavor, butter and the astringent taste and sourness in samples M6 and M7 were significantly higher than others in samples (p < 0.05), which may directly affect the evaluator's preference. Interestingly, M6 and M7 were not significantly different from sample M12 in milk flavor scores, but they were significantly lower compared to the other milk powder samples (p < 0.05). Moreover, milk flavor scores were higher in samples S1, S3, S8 and S14 compared to other samples, which may resulted in higher preference. The sensory histogram of aroma and taste is shown in Figure 3. Steamed, fishy and astringent flavors were the main sources of off-flavors in milk powder samples, and butter was the most important flavor precursor of oxidized flavor. Lipids are the most important flavor precursors of oxidized flavors in dairy powders, and milk fat oxidation can lead to off-flavors during the processing and storage of dairy powders. Oxidized flavors can be described as fatty, greasy, soapy, and dyestuff . Hexanal and 2-heptanone are typical oxidative volatiles, which accelerate fat oxidation during preheating, concentration, drying and storage at 40 degC . Hexane, heptane and pentane correlated with "painty", "oxidized", "cooked", and "caramelized" attributes in dairy based powders during storage . In addition, steamed flavor is described as accompanied by a caramel fragrance. Lactose in cow's milk may undergo degradation, Maillard reaction and isomerization, which can lead to deepening its color as well as the generation of furanone, volatile furanones, furfural and malt phenols, which contribute to the development of steamed and caramel flavor in milk powder during processing or long-term storage . The heating of milk powders can trigger a Maillard reaction, producing furfural, furan derivatives, dicarbonyl compounds and other flavor substances, 2-ethylfuran, which produced the caramel aroma of milk powders . Furthermore, the endogenous enzymes found in cow's milk can degrade proteins and lead to the production of flavor precursors such as taste peptides and amino acids with sweet and sour flavors such as bitterness and astringency. The astringency is attributed to the interaction product of whey protein, calcium phosphate, and caseins during the high-temperature treatment , which are attributes that limit consumer's choice. Interestingly, The samples scored differently in terms of saltiness, and process variation may lead to differences in the degree of lactose degradation, which may contribute to differences in flavor. 3.3. Relationship between Sensory Attributes and Quality Grades of Milk Powder ANOVA failed to identify the interaction between sensory descriptors and the overall sensory characteristics and their degree of variation in milk powder samples. Thus, sensory data were further analyzed by PCA. Standardized PCA plots using correlation matrices are commonly used in sensory analysis to visualize the relationships between sensory properties, individual samples and the strength of correlation between samples and sensory attributes . PCA is a classification technique which consists of eigenvectors coupled with a correlation matrix, in which the maximum difference between data can be revealed by axes rotation. A new set of axes is then calculated by dimensionality reduction to capture the maximum difference between the entire data set. Thus, PCA enables data visualization by reducing the dimensionality of complex datasets, which increases interpretability and minimizes information loss . Pan compared the differences among skim milk samples processed under different preheating treatments by combining E-tongue with PCA and cluster analysis (CA). Moreover, Chi et al. observed variation trends both in volatile favor composition and aroma release in six skim milk products using PCA. Hence, PCA showed the relationship between the classification contributions of different milk powder brands and the internal association between samples and sensory attributes. Figure 4 shows the PCA plot, which revealed that the contribution of PC1 was 73.2%, whereas the contribution of PC2 was 9.2%, and these two dimensions described 82.4% of the variability in the data set. In the first dimension, samples S2, S6, S7, S9, S12, and S13 were clearly separated from samples S1, S3, S4, S5, S8, S10, S11 and S14. The first dimension (F1) enabled distinguishing milk powder samples by the intensity of the following attributes: sweetness, milk flavor, steamed flavor and preference. Conversely, the attributes associated with poorer milk powder quality were butter taste, fishy flavor, sourness, astringency, saltiness and aftertaste, which were found on the positive side of the plot. Furthermore, samples S6, S7, and S12 were strongly correlated with less favorable attributes such as saltiness, butter taste, fishy flavor and astringency. Conversely, milk flavor and sweetness contributed greatly to sensory descriptions of samples S3 and S1 in the first dimension, being thus preferred by sensory evaluators. Finally, the first dimension enabled distinguishing samples S6, S7, S12, S9, S13, and S2 from other samples. Moreover, CA was further used to analyze 14 samples based on the evaluation results of sensory evaluators. CA graph showed that samples S6, S7 and S12 clustered together, which revealed that evaluators were less receptive to these samples based on sensory attributes and preference analysis. Taken together, the PCA , CA and sensory analysis results were consistent and fully reflected the negative evaluation of sensory attributes of samples S6, S7, and S12. 3.4. Preference Analysis and External Preference Mapping K-means is an unsupervised clustering algorithm that allows deciding a priori on the number of cluster groups. In a previous study, extra virgin olive oil was classified into two categories based on the k-means algorithm . K-means clustering is a common statistical data analysis that captures consumer preferences in detail and accounts for the diversity of evaluators' preferences. CA enables the selection of three preference groups, with different groups of consumers with different preferences for milk powder samples . Figure 5 shows the results of ratings of milk powder samples for each preference group. The first preference group consisted of 17 members that contained more than five points for two samples, (i.e., S1, S3 and S11); members of this group had an average preference for most milk powder samples with an average preference score of 4.08, and this group was composed of female members. Hence, it can be considered that women have a lower preference for the milk powder samples evaluated herein. In addition, cluster 2 (n = 23) had higher preference scores for all milk powder samples compared to the members of cluster 1 and cluster 3. The average score was 5.11. Thus, members of this group showed a relatively high approval of the most milk powder samples. The preference for samples M6, M7, and M12 was below 4, which was consistent with the other cluster. Finally, the third preference group consisted of 16 members, which rated preference for samples S1, S2, S3, S4 and S5 above 5. Meanwhile, the scores of M6, M7, M12 were 2.6, 2.8, and 2.32, respectively. All groups have poor sensory impressions of M6, M7 and M12. Then, to further investigate the influence of personal liking on the sensory description of milk powder samples, PCA factor scores were linked to CA results to conduct external preference mapping . Cluster 1 (n = 17) and cluster 2 (n = 23) had similar preferences, since milk sweetness was appointed as an essential trait, which were distinct from cluster 3 (n = 16) when considering preference. Interestingly, cluster 1 scored below 4 for five milk powder samples. Moreover, most of the consumers (n = 23) associated milk powder flavor with the attribute sweetness, milk flavor and steamed flavor (S3 and S4). However, as identified in the present study, consumers tend to dislike saltiness, butter taste, astringency, fishy flavor, and aftertaste in milk powder. Therefore, different sensory attributes affect the degree of consumer preference for milk powder and may also affect brand choice. Moreover, cluster 3 was associated with milk flavor and cooking flavor in quadrant 3, since a higher consumers' acceptance was observed for the majority of samples. Cluster 3 reported preference for milk flavor, steamed flavor and preference scores above 4 for seven samples (Table 3). In addition, all clusters were located on the negative half-axis of PC1, which enabled distinction between samples S6, S7 and S12, which had fatty flavor, saltiness, fishy flavor, astringency and aftertaste. Thus, these odor attributes (i.e., fatty flavor, saltiness, fishy flavor, astringency and aftertaste) hindered consumers' choice of milk powder. Moreover, two sources of origin and milk powder brands, i.e., S6 and S7, were derived from Beijing, and S12 was obtained from Guangdong, and differences in origin and milk powder brand may be an important factor affecting the sensory quality and preference of milk powder. 4. Conclusions To date, no studies have been conducted on Chinese milk powder samples for specific sensory evaluation and preference descriptions. Herein, descriptive sensory analyses were conducted on fourteen milk powder samples currently available in the Chinese market. Nine sensory descriptors were outlined to describe the overall sensory characteristics and preferences of evaluated milk powder samples, which enabled a more valid and accurate assessment. In addition, k-means enabled classifying the sensory evaluators into three clusters based on preference. Moreover, external preference mapping showed that the three clusters were clearly separated from milk powder samples associated with saltiness, aftertaste, creamy flavor, astringent flavor, cow flavor, and sourness. Milk powder samples with a predominance of milk sweetness, milk flavor and steamed flavor were the preferred sensory attributes, which point toward the attributes that should be pursued by the industry. Finally, an in-depth understanding of flavor formation in IFMPs might improve the processing and quality of milk powder to reduce the formation of off-flavors. Acknowledgments We are very grateful to the volunteers who are willing to participate in the sensory evaluation, and Ai nasi from Beijing Technology & Business University for providing assistance in panel recruitment. We would like to acknowledge the reviewers for their helpful comments on this paper. Author Contributions Y.X. Contributed to conceptualization, methodology, data processing, plotting, and writing--original draft writing--review and editing. T.Z. Investigation and analysis. R.L. review and editing. F.S. investigation, review and editing. J.D. writing, review and resources. N.A. Guide, methodology, formal analysis, writing--review and editing, project administration and funding. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement There is no conflict of interest here. Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement Data is contained within the article. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Analysis of assessment capacity of sensory evaluators enrolled in this study of fourteen Chinese milk powder samples. Mean squared error (MSE) values indicate within-group variance of sensory evaluators. 1, 2, 3... 56 represent 56 sensory evaluators. Figure 2 Radar chart of the sensory evaluation of fourteen milk powder samples available in the Chinese market. Samples were individually assigned a label within the range of S1-S14. (a) Represent odor score, (b) represents taste score. Figure 3 Histogram of odor and taste attributes rating of fourteen milk powder samples (labeled individually from S1 to S14) available in the Chinese market. (a-d) Histograms of odor sensory scores. (e-i) Histograms of taste scores. Figure 4 Classification of milk power samples from different brands currently marketed in China. (a) Biplots based on the results of principal component analysis from the descriptive analysis of fourteen milk powder samples. The first two dimensions of the biplot explained 82.4% of the variation among the samples. (b) Dendrogram of cluster analysis. Graphical representation of samples from consumer preference on fourteen infant formula milk powder samples. Figure 5 K-means clustering analysis of preference scores for fourteen milk powder samples currently available in the Chinese market. Figure 6 External preference mapping based on results of principal component analysis and agglomerative hierarchical clustering analysis. Cluster1, cluster2, and cluster3 refer to the three cluster groups formed based on K-means clustering analysis of preference, respectively. foods-12-00997-t001_Table 1 Table 1 Sample Information. Sample Fat (Per 100 g Milk Powder) Protein (Per 100 g of Milk Powder) Carbohydrates (Per 100 g of Milk Powder) S1 25 11 55.2 S2 27 10.9 54.4 S3 28 10.8 53.2 S4 24 10.6 51.6 S5 28.1 10.2 50.2 S6 27 12 56 S7 26.1 11.7 56.4 S8 26.04 11.99 51.9 S9 26.8 11.5 54.6 S10 26.6 11.6 56.1 S11 25 11 55.2 S12 27 11.5 56.5 S13 27.9 11.1 52.8 S14 25 11.8 54.4 foods-12-00997-t002_Table 2 Table 2 Sensory attributes and standard of sensory evaluation . No. Attributes Evaluation Method 1 Milk flavor Vanillin added to milk powder base, inherent flavor of milk powder, the aroma is gentle and scented 2 Milk fat E, E-2,4-decadienal added to milk powder bases, intensity of greasy feeling in the mouth, harsh taste intensity of oil oxidation 3 Cooked Milk powder substrate in water bath at 85 degC for 30 min, steaming aroma on the nose, with a slight caramel aroma 4 Fishy flavor Freshly slice of mushrooms are added to the milk powder base, smell a metallic, mud-like fishy smell 5 Sweet taste 5% sucrose solution, intensity of the inherent sweetness for sample 6 Sour 1% citric acid solution, intensity of sourness was felt in the mouth 7 Saltiness 2% NaCI solution, intensity of salty taste perceived by the mouth 8 Astringent Soak 6 tea bags in water for 10 min, the intensity of an astringent sensation felt by the tongue 9 Aftertaste Sensory intensity of residual milk flavor, softness and fluidity remaining in the mouth and throat 10 Fullness Overall, degree of preference for test samples foods-12-00997-t003_Table 3 Table 3 Significance analysis of sensory evaluation results (mean and standard deviation, SD of milk powder samples). Sample Milk Flavor #,+,++ Milk Fat Cooked Fishy Flavor Sweet Taste Saltiness Sour Astringent Aftertaste Preference S1 5.10 +- 1.80 ab 3.99 +- 2.03 d 4.48 +- 1.89 abcd 2.67 +- 1.65 e 5.54 +- 1.84 a 2.21 +- 1.31 c 2.91 +- 1.52 bc 2.65 +- 1.53 d 4.29 +- 1.68 cde 6.03 +- 1.66 a S2 4.24 +- 1.99 d 4.35 +- 1.92 cd 4.61 +- 1.91 abc 4.17 +- 2.21 b 4.99 +- 1.84 bc 2.76 +- 1.81 a 3.46 +- 1.73 abc 3.04 +- 1.67 cd 4.33 +- 1.74 cde 4.73 +- 1.86 cd S3 5.55 +- 1.79 a 3.99 +- 1.94 d 5 +- 2.09 a 2.55 +- 1.77 e 5.54 +- 1.55 a 2.22 +- 1.35 bc 2.88 +- 1.60 a 2.62 +- 1.54 d 4.17 +- 1.70 cde 6.4 +- 1.46 a S4 3.92 +- 1.63 de 4.46 +- 1.77 abcd 3.88 +- 1.72 ef 3.08 +- 1.85 cde 5.18 +- 1.70 ab 2.5 +- 1.49 abc 3.49 +- 1.70 abc 3.08 +- 1.74 cd 4.17 +- 1.54 cde 5.14 +- 1.51 bc S5 3.79 +- 1.77 def 4.54 +- 1.89 abcd 3.82 +- 1.75 ef 2.82 +- 1.53 de 5.26 +- 1.73 ab 2.52 +-0.50 abc 3.62 +- 1.77 ab 2.68 +- 1.57 cd 4.3 +- 1.62 cde 4.95 +- 1.79 c S6 3.16 +- 1.92 h 5.02 +- 2.08 a 3.65 +- 2.06 f 5.04 +- 2.55 a 4.25 +- 1.90 de 2.56 +- 1.43 abc 3.67 +- 1.71 a 3.95 +- 1.94 a 5.27 +- 1.99 a 2.52 +- 1.74 f S7 3.24 +- 1.71 gh 4.85 +- 1.04 abc 3.66 +- 1.98 f 4.58 +- 2.46 ab 4.11 +- 1.77 e 2.71 +- 1.45 a 3.57 +- 1.68 ab 3.89 +- 1.85 a 5.04 +- 1.81 ab 2.88 +- 1.55 f S8 4.74 +- 1.94 cd 4.18 +- 1.81 d 4.15 +- 1.86 cdef 2.75 +- 1.79 de 4.29 +- 1.75 de 2.33 +- 1.26 abc 3.16 +- 1.46 bcde 2.99 +- 1.65 cd 3.97 +- 0.55 e 5.5 +- 2.03 b S9 3.69 +- 1.61 efg 4.96 +- 1.96 ab 4.04 +- 1.88 def 2.95 +- 1.85 cd 4.73 +- 1.91 bcd 2.33 +- 1.35 abc 3.35 +- 1.30 abde 2.97 +- 1.56 d 4.55 +- 1.69 cd 4.12 +- 1.55 e S10 4.14 +- 1.71 de 4.01 +- 1.82 d 3.71 +- 1.81 f 3.52 +- 2.11 c 4.31 +- 2.01 de 2.65 +- 1.43 abc 3.54 +- 1.68 abc 3.76 +- 1.85 ab 4.64 +- 1.78 bc 4.73 +- 1.70 cd S11 4.15 +- 1.64 de 3.98 +- 1.72 d 3.69 +- 1.62 f 2.94 +- 1.65 cde 4.71 +- 1.80 bcd 2.54 +- 1.55 abc 3.06 +- 1.45 cde 2.86 +- 1.50 cd 4.11 +- 1.50 de 4.88 +- 1.89 c S12 3.33 +- 1.71 fgh 4.83 +- 1.95 abc 3.88 +- 1.97 ef 4.3 +- 2.16 b 4.73 +- 1.82 bcd 2.49 +- 1.37 abc 3.24 +- 1.54 abcde 3.37 +- 1.91 bc 4.39 +- 1.62 cde 2.99 +- 1.99 f S13 4.3 +- 1.54 d 4.43 +- 1.93 bcd 4.34 +- 1.91 bcde 3.3 +- 1.74 cd 5.08 +- 1.87 ab 2.46 +- 1.39 abc 3.15 +- 1.60 bcde 2.62 +- 1.30 d 4.39 +- 1.64 cde 4.28 +- 1.87 de S14 4.75 +- 1.68 bc 4.02 +- 1.83 d 4.78 +- 1.97 ab 3.42 +- 1.94 c 4.52 +- 1.74 cde 2.66 +- 1.41 ab 3.39 +- 1.62 abcd 3.23 +- 1.67 c 4.13 +- 1.53 cde 5.21 +- 1.83 bc # n = 56, + 9-point scale 1 = absence of the attribute, 5 = moderately strong for an attribute, 9 = very strong for an attribute. ++ Means in each column with the same letter are not significantly difference (p > 0.05). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Bode L. Raman A.S. Murch S.H. Rollins N.C. Gordon J.I. Understanding the mother-breastmilk-infant "triad" Science 2020 367 1070 1072 10.1126/science.aaw6147 32139527 2. Almeida C.C. Mendonca Pereira B.F. Leandro K.C. Costa M.P. Spisso B.F. Conte-Junior C.A. Bioactive compounds in infant formula and their effects on infant nutrition and health: A Systematic Literature Review Int. J. Food Sci. 2021 2021 8850080 10.1155/2021/8850080 34095293 3. Johnson S.L. Children's Food Acceptance Patterns: The Interface of Ontogeny and Nutrition Needs Nutr. Rev. 2002 60 S91 S94 10.1301/00296640260130812 12035867 4. Mennella J.A. Jagnow C.P. Beauchamp G.K. Prenatal and postnatal flavor learning by human infants Pediatrics 2001 107 E88 10.1542/peds.107.6.e88 11389286 5. Witt M. Reutter K. Embryonic and early fetal development of human taste buds: A transmission electron microscopical study Anat. Rec. 1996 2464 507 523 10.1002/(SICI)1097-0185(199612)246:4<507::AID-AR10>3.0.CO;2-S 6. Beauchamp G.K. Cowart B.J. Moran M. Developmental changes in salt acceptability in human infants Dev. Psychobiol. 1986 19 17 25 10.1002/dev.420190103 3699249 7. Oster D.R.H. Differential Facial Responses to Four Basic Tastes in Newborns Child Dev. 1988 6 1555 1568 10.1111/j.1467-8624.1988.tb03683.x 8. Carlson S.E. Arachidonic acid status of human infants: Influence of gestational sge at birth and diets with very long chainn-3 and n-6 fatty acids J. Nutr. 1996 126 1092S 1098S 10.1093/jn/126.suppl_4.1092S 8642439 9. Smriga M. Kameishi M. Tori K. Brief exposureto NaCl during early postnatal development enhances adult intake of sweet and salty compounds Neuroreport 2002 13 2565 2569 10.1097/00001756-200212200-00037 12499869 10. Ge W.P. Yang X.J. Wu X.Y. Wang Z. Geng W. Guo C.F. Phthalate residue in goat milk-based infant formulas manufactured in China J. Dairy Sci. 2016 99 7776 7781 10.3168/jds.2016-11061 27522423 11. He Y. Zeng W. Zhao Y. Zhu X. Wan H. Zhang M. Li Z. Rapid detection of adulteration of goat milk and goat infant formulas using near-infrared spectroscopy fingerprints Int. Dairy J. 2023 137 105536 10.1016/j.idairyj.2022.105536 12. Liu Y. Ge W. Zhang J. Li X. Wu X. Li T. Li T. Zhang X. Wang X. Detection of Bacillus cereus sensu lato from environments associated with goat milk powdered infant formula production facilities Int. Dairy J. 2018 83 10 16 10.1016/j.idairyj.2018.02.005 13. Zhou J. Yang M. Li F. Wang M. Zhang Y. Wei M. Chai Y. Development of matrix certified reference material for accurate determination of docosahexaenoic acid in milk powder Food Chem. 2023 406 135012 10.1016/j.foodchem.2022.135012 36462352 14. Zhu R. Cheng H. Li L. Erichsen H.R. Petersen M.A. Soerensen J. Skibsted L.H. Temperature effect on formation of advanced glycation end products in infant formula milk powder Int. Dairy J. 2018 77 1 9 10.1016/j.idairyj.2017.09.005 15. Masuyer G. Zhang S. Barkho S. Shen Y. Henriksson L. Kosenina S. Dong M. Stenmark P. Structural characterisation of the catalytic domain of botulinum neurotoxin X--High activity and unique substrate specificity Sci. Rep. 2018 8 4518 10.1038/s41598-018-22842-4 29540745 16. Phosanam A. Chandrapala J. Huppertz T. Adhikari B. Zisu B. Changes in physicochemical and surface characteristics in model infant milk formula powder (IMF) during storage Dry Technol. 2021 39 2119 2129 10.1080/07373937.2020.1755978 17. Gredilla A. de Vallejuelo S.F. Arana G. de Diego A. Oliveira M.L.S. Da Boit K. A Rapid Routine Methodology Based on Chemometrics to Evaluate the Toxicity of Commercial Infant Milks Due to Hazardous Elements Food Anal. Methods 2022 15 2312 2322 10.1007/s12161-022-02267-6 18. Stone H. Sidel J. Oliver S. Woolsey A. Singleton R.C. Sensory Evaluation by Quantitative Descriptive Analysis Food Technol. 2004 23 34 10.1002/9780470385036 19. Margeta P. Kralik G. Grcevic M. Hanzek D. Kralik Z. Importance of sensory evaluation in assessment of egg quality Poljoprivreda 2019 25 56 63 10.18047/poljo.25.1.8 20. Drake M.A. Karagul-yuceer Y. Cadwallader K.R. Civille G.V. Tong P.S. Determination of the sensory attributes of dried milk powders and dairy ingredients J. Sens. Stud. 2003 18 199 216 10.1111/j.1745-459X.2003.tb00385.x 21. Jolliffe I.T. Cadima J. Principal component analysis: A review and recent developments Phys. Eng. Sci. 2016 374 20150202 10.1098/rsta.2015.0202 26953178 22. Torrico D.D. Tam J. Fuentes S. Gonzalez Viejo C. Dunshea F.R. Consumer rejection threshold, acceptability rates, physicochemical properties, and shelf-life of strawberry-flavored yogurts with reductions of sugar J. Sci. Food Agric. 2020 100 3024 3035 10.1002/jsfa.10333 32056214 23. Moskowitz H.B. Product testing and sensory evaluation of foods: Marketing and R&D approaches J. Mark. Res. 1984 21 473 474 10.1177/002224378402100413 24. Mcewan J.A. Hunter E.A. van Gemert L.J. Lea P. Proficiency testing for sensory profile panels: Measuring panel performance Food Qual. Prefer. 2002 13 181 190 10.1016/S0950-3293(02)00022-8 25. Tomic O. Nilsen A. Martens M. Naes T. Visualization of sensory profiling data for performance monitoring LWT--Food Sci. Technol. 2007 40 262 269 10.1016/j.lwt.2005.09.014 26. Barker S. Moss R. Mcsweeney M.B. Identification of sensory properties driving consumers' liking of commercially available kale and arugula J. Sci. Food Agric. 2022 102 198 205 10.1002/jsfa.11346 34061358 27. Li Y. Wang W.J. Short communication: Formation of oxidized flavor compounds in concentrated milk and distillate during milk concentration J. Dairy Sci. 2016 99 9647 9651 10.3168/jds.2016-11619 27743658 28. Li Y.H. Zhang L.W. Wang W.J. Han X. Differences in particle characteristics and oxidized flavor as affected by heat-related processes of milk powder J. Dairy Sci. 2013 96 4784 4793 10.3168/jds.2012-5799 23769359 29. Clarke H.J. O'Sullivan M.G. Kerry J.P. Kilcawley K.N. Correlating Volatile Lipid Oxidation Compounds with Consumer Sensory Data in Dairy Based Powders during Storage Antioxidants 2020 9 338 10.3390/antiox9040338 32326117 30. Solah V.A. Staines V. Honda S. Limley H.A. Measurement of Milk Color and Composition: Effect of Dietary Intervention on Western Australian Holstein-Friesian Cow's Milk Quality J. Food Sci. 2007 72 S560 S566 10.1111/j.1750-3841.2007.00491.x 17995621 31. Karagul-Yuceer Y. Drake M. Cadwallader K.R. Aroma-Active Components of Nonfat Dry Milk J. Agric. Food Chem. 2001 49 2948 2953 10.1021/jf0009854 11409991 32. Josephson R.V. Thomas E.L. Morr C.V. Coulter S.T. Relation of Heat-Induced Changes in Protein-Salt Constituents to Astringency in Milk Systems 1 J. Dairy Sci. 1967 50 1376 1383 10.3168/jds.S0022-0302(67)87638-0 33. Koch I.S. Muller M. Joubert E. van der Rijst M. Naes T. Sensory characterization of rooibos tea and the development of a rooibos sensory wheel and lexicon Food Res. Int. 2012 46 217 228 10.1016/j.foodres.2011.11.028 34. Pan M. Tong L. Chi X. Ai N. Cao Y. Sun B. Comparison of sensory and electronic tongue analysis combined with HS-SPME-GC-MS in the evaluation of skim milk processed with different preheating treatments Molecules 2019 24 1650 10.3390/molecules24091650 31035485 35. Chi X. Shao Y. Pan M. Yang Q. Yang Y. Zhang X. Ai N. Sun B. Distinction of volatile flavor profiles in various skim milk products via HS-SPME-GC-MS and E-nose Eur. Food Res. Technol. 2021 247 1539 1551 10.1007/s00217-021-03730-0 36. Cecchi L. Parenti A. Bellumori M. Migliorini M. Mulinacci N. and Guerrini L. Clustering monovarietal extra virgin olive oil according to sensory profile, volatile compounds, and k-Mean algorithm Eur. J. Lipid Sci. Technol. 2022 124 2 10 10.1002/ejlt.202200038 37. Cardello A.V. Food quallty: Relativity, contextand consume expectations Food Qual. Prefer. 1995 6 163 170 10.1016/0950-3293(94)00039-X |
PMC10000601 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050927 diagnostics-13-00927 Review Mental Health Experts as Objects of Epistemic Injustice--The Case of Autism Spectrum Condition Wodzinski Maciej 12* Moskalewicz Marcin 134 Stoyanov Drozdstoi Academic Editor 1 Institute of Philosophy, Maria Curie-Sklodowska University, M. Curie-Sklodowska sq. 4, 20-031 Lublin, Poland 2 Doctoral School of Humanities, Maria Curie-Sklodowska University, Weteranow 18, 20-038 Lublin, Poland 3 Philosophy of Mental Health Unit, Department of Social Sciences and the Humanities, Poznan University of Medical Science, Rokietnicka 7, 60-806 Poznan, Poland 4 Phenomenological Psychopathology and Psychotherapy, Psychiatric Clinic, Heidelberg University, Vossstrasse 4, 69115 Heidelberg, Germany * Correspondence: [email protected] 01 3 2023 3 2023 13 5 92731 12 2022 22 2 2023 28 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). This theoretical paper addresses the issue of epistemic injustice with particular reference to autism. Injustice is epistemic when harm is performed without adequate reason and is caused by or related to access to knowledge production and processing, e.g., concerning racial or ethnic minorities or patients. The paper argues that both mental health service users and providers can be subject to epistemic injustice. Cognitive diagnostic errors often appear when complex decisions are made in a limited timeframe. In those situations, the socially dominant ways of thinking about mental disorders and half-automated and operationalized diagnostic paradigms imprint on experts' decision-making processes. Recently, analyses have focused on how power operates in the service user-provider relationship. It was observed that cognitive injustice inflicts on patients through the lack of consideration of their first-person perspectives, denial of epistemic authority, and even epistemic subject status, among others. This paper shifts focus toward health professionals as rarely considered objects of epistemic injustice. Epistemic injustice affects mental health providers by harming their access to and use of knowledge in their professional activities, thus affecting the reliability of their diagnostic assessments. expert knowledge testimonial injustice hermeneutical injustice discrimination autism schizophrenia mental health diagnosis social epistemology Polish Ministry of Science and Higher EducationDI 2018 001348 Research Ethics Committee of Maria Sklodowska-Curie UniversityAlexander von Humboldt FoundationThe research received funding from the Polish Ministry of Science and Higher Education under Grant Agreement No DI 2018 001348 and was accepted by the Research Ethics Committee of Maria Sklodowska-Curie University. Marcin Moskalewicz was supported by the Alexander von Humboldt Foundation. pmc1. Introduction: Epistemic Injustice in Mental Health as concerning Both Service Users and Providers Can only 'patients' or 'service users' of mental healthcare systems be objects of epistemic injustice? Can health professionals (traditionally portrayed as the perpetrators of unfairness) also suffer from such injustice in their interactions with patients? If so, what might be the sources of epistemic injustice understood in such an unorthodox manner? By answering these questions, this paper presents the issue of epistemic injustice from a novel perspective. We hypothesize and argue that health professionals may also become objects of epistemic injustice and that recognizing this may significantly impact their practice. Miranda Fricker's influential work has brought much attention to the injustice and harm performed on people concerning knowledge production and distribution. Discussing epistemic injustice, researchers and public opinion focused on people typically associated with the 'weaker side' in different settings of power relations. In healthcare systems, these were the patients. It is typically pointed out that such harm is caused by the "stronger" side of the interaction. For example, doctors use stereotypes concerning various groups of patients, experts reject (without proper justification) the opinions of other experts about the patient's health, diagnosticians use biased diagnostic tools, or nurses, social workers, and even patients' families ignore the first-person viewpoint of their ill members. This paper shows that the 'stronger side' of these power relations might also be harmed. Considering the phenomenon of 'being an expert' from an epistemological and ethical perspective, one may notice that it involves not only the possession of specific competencies, but also the obligation to perform one's duties diligently. In the case of mental health, this is both extremely important and challenging. With this in mind, this review draws attention to the fact that those usually perceived as the 'stronger side' can also fall victim to epistemic injustice. Epistemic injustice, a concept developed by philosopher Miranda Fricker , is a situation in which, without adequate reason, someone suffers harm caused by or related to access to the process of producing and distributing knowledge. As Grim and colleagues point out, the notion of epistemic injustice can refer not only to the social practices involved in the production and transmission of knowledge, but also to the very status of subjects and their position in the social hierarchy of knowledge-making practices: "The concept of epistemic injustice refers to an injustice performed on individuals in their capacity as knowledge bearers, reasoners and questioners, in which their ability to take part in epistemic practices, such as giving knowledge to others (testifying) or making sense of their experiences (interpreting), is weakened". (p. 158) Table 1 presents the core arguments made further in this article. 2. Epistemic Injustice toward Patients When considering the issue of epistemic injustice, among other things, one can refer to phenomena, such as epistemic exploitation , epistemic oppression , systemic silencing , genocide denial , or even hermeneutical death . Although the literature describes many possible types of epistemic injustice, this paper focuses on its two main modes, described by Fricker, namely, testimonial and hermeneutical injustice. The former refers to a situation in which, without a sufficiently justified reason, the testimony of one subject is omitted or belittled (or its ability to carry out the relevant cognitive processes is denied): "Testimonial injustice" occurs when prejudice causes a hearer to ascribe a deflated level of credibility to a speaker's words or 'testimony' (p. 151). The second, hermeneutical injustice, refers to a situation in which the subject itself is 'epistemically wronged' due to not having resources sufficient for the ability to interpret its own experience adequately. In recent years, a number of researchers considered the occurrence of epistemic injustice in healthcare systems, including mental health . They pointed out that becoming ill exposes patients to the experience of such injustice and, consequently, to a lower chance of receiving appropriate help . Lakeman emphasized that the causes of testimonial injustice tend to have individual origins and essentially lie at the heart of the mental health care system: "Actual and potential testimonial injustice is endemic within mental health service delivery. For example, central to mental health legislation is the idea that some people lack the capacity to make decisions and it follows that what they might say, how they construe problems, their choices and preferences lack coherence, logic, or credibility. It is not surprising then that the testimony of all or most people who use mental health services might be considered suspect". (p. 151) The second type of epistemic injustice, hermeneutical injustice, is more socially grounded. It is a social situation in which an individual has difficulty making adequate sense of his or her experience and interpreting it due to the so-called "collective hermeneutical block" or bias. Explaining the 'social basis' of hermeneutical injustice, Grim and colleagues note that: 'Hermeneutical injustice occurs when there is a breach in shared conceptual, interpretative resources that puts individuals at a disadvantage when trying to make sense of their experiences'. (p. 158) An example given by Fricker is that women experienced sexual harassment at a time when the term did not yet exist. This is not, of course, just a lack of a label for certain behaviors but rather a set of social conditions (such as consent to certain behaviors, women thinking of themselves as inferior to men, cultural objectification of the body) that prevented women from thinking of themselves as victims of harassment. A structurally similar situation can be observed for people reaching out to mental healthcare. The dominant narratives (social, medial, academic, and scientific) very strongly shape how the participants in this system see themselves (both in terms of how they perceive their characteristics and in terms of their position in the web of relationships between the different parties involved in healthcare) . Epistemic inequalities result in ethnic minorities being forced to use crisis care and being less likely to receive primary mental health care . The differences between patients' first-person perspectives on illness and service providers' conceptualization of illness as a disease are growing . Patients' testimonies of perceived illness symptoms are downgraded , and somatic illnesses are often overshadowed by previously detected mental conditions leading to a decrease in the quality of medical care . Also, epistemic injustice impacts the policy-making level, resulting in the exclusion of specific patients or health professional groups from the legislative processes . Exploring the ideas such as epistemic solidarity among patients or increasing the role of biocommunities in knowledge production and circulation might reduce epistemic injustice in mental health . Equally important is the identification of causes and additional factors influencing the occurrence of epistemic injustice, such as the hubris phenomenon or low levels of intellectual humility among professionals , which increase the risk of systemic cognitive errors. 3. Epistemic Injustice toward People on the Autism Spectrum People on the autism spectrum, stereotypically perceived but also formally/diagnostically classified as having a mental disorder, are particularly vulnerable to testimonial injustice. Paul Crichton and colleagues point out a serious problem, which may in part cause testimonial injustice. The mere fact of having a mental illness or disorder seriously affects the credibility of a person's testimony. Psychiatrists then tend to interpret the behaviors or explanations in question as the effects of the illness rather than as phenomena resulting from another cause. Crichton gives interesting examples from his own clinical practice: "When one of the authors (P.C.) was a medical student in Munich, Germany, he saw a young man on an acute psychiatric ward who said he was a relative of the then Soviet leader. The responsible consultant took this to be a grandiose delusion, and therefore as evidence of a psychotic illness; it later turned out to be true". (p. 66) People on the autism spectrum are often denied epistemic authority. Their testimony is denied credibility without justification, and they are not seen as fully-fledged epistemic agents . This phenomenon occurs in two main ways in evaluations carried out by experts, such as medical examiners, diagnosticians, or forensic experts. It appears as the non-recognition or misinterpretation of the testimony of the service user and/or as the ignoring of the testimony of third parties involved (such as carers, who know the service user best or the opinions of the professional therapeutic teams working with the service user on a daily basis). Suppose a person on the autism spectrum under assessment (or his/her carer) declares that he/she has severe difficulties in social relationships. This information can be easily trivialized by the expert, who has a stereotypical image of a person with autism as incapable of interpersonal relationships. Such an assessment, unfortunately, results in overlooking problems of a relational nature or due to other causes, e.g., depression, and neglecting therapy/treatment tailored to the person's needs. Similar experiences of, inter alia, attributing social problems as resulting from diagnosed depression are described by Richard Lakeman in his studies on epistemic injustice in mental healthcare systems . On the other hand, people on the autism spectrum are also subject to hermeneutical injustice. The media-disseminated stereotypical representations of autism that construct the social space of meanings and understandings of the spectrum can have positive and negative impacts. Grim's observations can also be directly applied to the situation of people on the spectrum. Their accounts often mention that receiving a diagnosis allowed them to make sense of their behaviors or ways of thinking that they did not previously understand and for the interpretation of which they did not have good hermeneutic resources. The problem is that, if a negative or deficit message dominates socially, making sense of one's experience leads to seeing oneself as 'broken' or one's behavior as something shameful to be gotten rid of at all costs . For example, most media coverage portrays people on the autism spectrum as suffering victims of their condition and shows autism as inferior to a neurotypical state and needing correction. People with autism internalize these views, which then become essential to their identities. Another example is the academic discourse, which primarily treats people with autism as objects of research rather than equal partners in the research programs. As far as autism is concerned, the diagnosis and other kinds of expert assessment is problematic due to the lack of "hard data" (such as biological markers) and unequivocal findings as to its causes. The increasing public presence of autism, intensive scientific research, and growing media coverage often translate to the rise of numerous non-rational beliefs (such as stereotypes or prejudices) on this topic . These beliefs might constitute the essential discursive presuppositions of cognitive errors concerning autism--including those made by the experts. As previous research showed, these presuppositions might even work at the level of our background knowledge . The power relations between health experts and people with autism impact almost every area of autistic people's lives, from the construction of their identities (based, e.g., on social, media, and professional representations), through interpersonal relationships at all levels of social organization, to decisions made by third parties related to their economic and social situation . These discursive practices, which construe the social imagination and shape widely disseminated narrations, may lead to the reinforcing of hermeneutical epistemic injustice toward the users (and, as this paper argues further on, service providers) of the mental healthcare system: '(...) hermeneutical practices play an important role in health care because they allow service users sense-making reflectivity, which helps to turn a confusing and troubling set of symptoms into a more comprehensible and tenable context'. (p. 169) Thus, interpretive practices, based largely on socially available resources, are essential for autistic 'patients' meaning making concerning their lived experience. These practices are handicapped by hermeneutical injustice, e.g., when there are no other models for thinking of oneself than as a sick person or someone without a voice. Such tendencies are clearly visible in numerous media discourse analyses . 4. Mental Health Expertise and Its Fallibility The concept of expertise is widely and thoroughly developed. There are overview monographs presenting the history and cultural context of knowledge formation of expertise , monographs analyzing various types of expertise , and monographs on selected types of expertise and their transformations . Some studies show links of expertise to social and political issues , as well as the recent changes in this phenomenon (e.g., declining trust in expertise, the crisis of expertise in a democratic society) . One can distinguish two classical models of expert knowledge function, that is, 'knowledge how' and 'knowledge that' , and two aspects of expert knowledge, that is, the objective aspect (beliefs, assertions, rules for justifying them) and the subjective aspect (cognitive dispositions such as general cognitive abilities or epistemic virtues and expert competence). Both of these aspects are conditioned by several social factors. The concept of expertise and its applicable standards are shaped and transformed according to the social expectations and functions it is supposed to perform . The subjective dispositions of experts are also not suspended in an information vacuum and are partly shaped by the discourses (professional, scientific, and popular) in which the expert in question functions. From the perspective of social epistemology , expert knowledge is a set of primary and specialized information that comes from one or more scientific disciplines, remains at the expert's disposal, and relates to the issue that is the subject of the formulated expert opinion. Notably, scientific knowledge is indirectly related to expert knowledge constituting its basis. However, the latter shows a higher degree of flexibility regarding the conditions for creating and changing the standards of reliability or credibility of the statements contained therein. These standards are much more susceptible to the influence of social, political, or everyday life factors . The above is also true regarding specialized fields such as medicine and its subdisciplines, such as psychiatry. There is no ideal type of model of expert knowledge for all medical specializations. Instead, they function on account of a given practical situation, a specific problem situation, or a decision to be made. They express the demand of some group and respond to it, which all together determines a specific rather than a universal type of expertise. Speaking of medicine, one should shift emphasis from expert knowledge towards expertise, understood as conditioned cognitive dispositions, such as thinking, deduction, justification, and argumentation. As Majdik and Keith put it: '"Expertise" as a concept cannot be reconciled by only one shared principle. As a consequence, it is not comprehensible in a conceptual definition but only in its varied uses and enactments. We suggest, in other words, that there is not--even deep down on a conceptual level--one kind of expertise, but kinds of expertise that resonate with kinds of problems' . This is clearly visible in psychiatry, where different disorders have particular diagnostic demands. Experts in their fields might also be subject to stereotypes and simplified rules of reasoning . Draaisma and Hacking showed that even domain-specific experts are not fully stereotype-proof and might be influenced by simplified inferencing patterns (the so-called heuristics). Mental health experts, such as psychiatrists and clinical psychologists, are especially exposed to systemic cognitive errors as mental states, disorders, or illnesses are difficult to measure in an objectified way. Experts must contend with the co-existence of various conditions, including somatic ones, with considerable personal differences, lack of biomarkers, or non-specific symptoms. They also often base their judgments on intuitions whose reliability is unclear . Naturalistic decision-making theory states that heuristics allow experts to make complex decisions quickly when expert opinions, e.g., in a case-law process, must be issued mainly based on intuitive decisions . On the other hand, the heuristics and biases approach theory sees heuristics as a source of cognitive errors . Both approaches emphasise that heuristics are inherent parts of our decision-making processes. We cannot "turn them off". We may only try to become aware of them and control their effects. Damasio and Ericsson point out that despite the differences, both theories indicate three primary conditions required for correct intuitive decisions: (1) update knowledge based on credible sources in a specific field; (2) deliberate practice (systematically applying expertise continuously subjected to reflection); (3) feedback on the decisions made. If these conditions are not met, experts might be prone to mistakes and their elimination . 5. Epistemic Virtues of Experts Two modes of epistemic injustice can compromise the professional traits and qualities that experts are typically expected to possess. In performing their duties, experts form numerous opinions, make judgements or diagnoses based on their beliefs. Ideally, these beliefs should be (a) always true, (b) formulated based on a complete set of information obtained from all available sources, (c) supported by relevant evidence , (d) free from systemic biases, and (e) subjected to multi-stage control. Unfortunately, in the case of mental health phenomena, there is often (a) no single, universally valid description of a given reality that we can define as 'true', (b) sources of information are sometimes incomplete (some of them are also ignored by experts for various reasons) and the information itself is usually subject to a process of interpretation, (c) there is usually no 'hard' evidence, (d) experts are not free from biases, (e) often their decisions are not subject to automatic review--which is a result of the trust we place in those who have, sometimes rightly sometimes wrongly, the appropriate epistemic authority . In such a situation, to avoid more or less chaotic and 'luck-based' beliefs and decisions , expert knowledge theory should focus not solely on 'the outcome' but on the process of belief justification. According to the externalist view of Alvin Goldman's epistemological reliabilism , the most crucial feature of the knowledge creation process is the reliability of cognitive and epistemic processes. It means that 'what makes a belief epistemically justified is the cognitive reliability of the causal process via which it was produced' . Building on the findings of the reliabilists, virtue epistemologists (Ernst Sosa , John Greco John Turri ) take the view that whether the process of obtaining a belief has been carried out reliably is determined to the greatest extent by the characteristics of the subject himself (the notion of a reliable process is replaced by concepts such as 'competence', 'virtue', 'skill' or 'ability' ). Virtue epistemologists, however, strongly emphasize the subject's role in justifying beliefs and creating knowledge. The epistemic virtues of the subject, such as reliability, open-mindedness, understanding of another point of view, or intellectual humility, must form the basis for justifying the subject's beliefs. Reliable expert knowledge results from a cognitive process in which the subject can use his or her epistemic virtues in a controlled manner. For example, a list of such epistemic virtues includes intellectual conscientiousness, open-mindedness, attentiveness, curiosity, discernment, humility, objectivity, and understanding. The list of epistemic vices includes dogmatism, epistemic blindness, intellectual dishonesty, self-deception, the superficiality of thought, and superstition. Many such vices are responsible for the aforementioned cognitive biases . The knowledge basis for experts' decisions is 'true belief out of intellectual virtue, a belief that turns out right because of the virtue and not just by coincidence' (p. 277). Therefore, any process that might negatively influence such experts' virtues and hermeneutical abilities might be considered harmful. The phenomenon of epistemic injustice is such a case. Virtue theorists think cognitive biases result from over-reliance on intuitive, fast, non-discursive, and non-reflective forms of cognition. In their view, this deficiency might be controlled 'by cultivating responsibilist virtues, such as epistemic humility and self-vigilance' (p. 2). 6. Epistemic Injustice toward Mental Health Experts The up-to-date discussions have focused on epistemic injustice concerning service users. The sources of such epistemic injustice lie within the system of mental health experts and their judgments, and we have described its two subtypes regarding people on the autism spectrum. In the power dynamics of autistic and non-autistic people, the focus remained on how epistemically unjust expert decisions influence the patients' quality of life and self-esteem, among others. In short, it is related to the expert-to-patient transfer of "injustice". Taking into account the tasks set for experts (giving opinions and decisions that are factual, reliable, and conforming to the best standards) and their dispositions, abilities, and epistemic virtues, it is worth looking at experts, usually shown as perpetrators of epistemic injustice as its victims (see Table 1). In what sense and why are experts the object of epistemic injustice? Two additional questions are due. First, what is the harm resulting from cognitive injustice? Second, who perpetrates this injustice against experts? Following the reliabilist theories of expert knowledge, the harm consists primarily in the deformation and negative impact on experts' dispositions and competencies. Consequently, it hinders and even prevents the efficient and reliable performance of professional tasks. The harm is different in the case of testimonial injustice. The testimony of one expert may be undermined or rejected by another. It thus loses its justifying force regarding an opinion or diagnosis. Such situations occur when issuing interdisciplinary diagnoses or at the committee assessment of the degree of disability of a patient. An opinion-issuing expert should be guided by the patient's medical documentation (other 'experts' opinions). It may happen that, in whole or in part, such an opinion is not considered. This may also result from the internally assumed hierarchy of epistemic credibility within the health care system. For example, an opinion of a psychiatrist is usually perceived as more credible than a psychotherapist's. Another example of testimonial injustice harm is disregarding experts' intuitive decision-making, such as a rapid diagnosis of schizophrenia based on the so-called Praecox Feeling . It has been known and theorized since Ruemke that qualified psychiatrists can pronounce adequate judgments of schizophrenia based on intuition alone--a theory that has some experimental, although not conclusive, confirmation . These judgments are most likely based on immediate and pre-reflective intercorporal affective exchange with the patients, which does not present itself as a set of operationalized symptoms described by diagnostic manuals. Since current psychiatric diagnoses must be reliable, psychiatrists cannot openly speak of their intuitions, which would likely undermine their professional credibility. Instead, this sometimes vital aspect of the diagnostic process must remain publicly hidden. The psychiatrists' testimony regarding the source of their knowledge is, therefore, unjustly silenced by the dominant discourse of diagnostic tools operating within the "ticking boxes" framework . In both cases, others do not recognize or value the expertise. It can be dismissed by those in power, for example, psychiatrists over counselors, or may be blocked by the functional aspects of the medical discourse, such as formalized disregard of clinical intuition. The nature of the harm is different when experts are subject to hermeneutical injustice. Again, we speak of hermeneutical injustice "when there is a breach in shared conceptual, interpretative resources that puts individuals at a disadvantage when trying to make sense of their experiences" (p. 158). To use the above example once more, a psychiatrist himself may disregard one's diagnostic intuition due to being embedded in the dominant discourse that does not recognize its value. In the case of mental health experts, the hermeneutical injustice category includes biased or limited preconceptions about expert's own position and its capabilities and also extends to interpreting the experience of service users assessed. For example, an expert psychiatrist rooted firmly in the dominant biomedical tradition may find it difficult to perceive and understand the psycho-social difficulties of patients due to perceiving and interpreting oneself as normatively "normal". Moreover, not being entirely immune to the numerous cognitive errors based on popular stereotypes and simplifications present in various discourses, his/her ability to make a competent judgment about a patient's condition, and level of functioning may be compromised. In this sense, epistemic injustice is also dialectical and may operate between both parties. Hermeneutical injustice thus concerns medical experts when they are unable to fully understand or participate in the autism discourse due to a lack of shared interpretive resources, e.g., when they are unable to effectively communicate with or understand patients' perspectives due to a lack of shared language or cultural understanding. This can lead to misdiagnosis or inadequate treatment, as they may not fully understand the patient's symptoms or concerns. It can also lead to a lack of trust and credibility, as the patients may feel that their concerns are not being fully heard or understood. 7. The Sources of Epistemic Injustice toward Experts There are several answers to the question of who perpetrates epistemic injustice against experts. Firstly, testimonial injustice occurs when one expert refuses to consider another expert's opinion as relevant. This problem can be encountered, for example, during court hearings when expert witnesses are called upon to assess a case. It may also occur during committees deciding on the degree of disability and state assistance. An expert usually receives therapeutic documentation that includes the opinions of other experts working with the patient in question. These opinions may be ignored as not containing substantively relevant arguments. Perhaps this is due to acceptable ways of valuing evidence. As Crichton and colleagues put it: "Health professionals are trained to place higher value on 'hard' or objective evidence, namely the results of investigations, than on 'soft' or subjective evidence provided by patients. In psychiatry, there is virtually no hard evidence and diagnoses have to be made mainly on the basis of what patients say and how they behave". (p. 67) A similar problem occurs when ignoring not only the opinion of patients but also of other experts. In the case of mental conditions, one is usually not dealing with 'hard' evidence but with the opinions of others (even if they are other experts). The decision-making expert may put the observed condition of the patient above the written opinions received. However, it may be the case that the difficulties faced by people on the autism spectrum (such as difficulties of a relational and social nature) cannot be observed during a brief assessment, especially when dealing with individuals who camouflage their difficulties very well. This may lead to a cognitive error called WYSIATI--what you see is all that is. WYSIATI thus becomes a source of epistemic injustice towards another expert . As a result, some experts may suffer unequal opportunities to contribute to their field and not be equally represented in the decision-making bodies. In consequence, the voices of some experts may not be heard analogically to the voices of patients with autism. Furthermore, such a situation might exacerbate the lack of diversity within the medical profession and marginalization of community experts facing discrimination in their professional judgments. Secondly, more hermeneutic types of epistemic injustice may stem from traditions and cultural and systemic conditions in which experts gain knowledge, experience, and operate. The biomedical model, increasingly criticized but still dominant in psychiatry today, is one of the most important traditions shaping today's understanding of such phenomena as autism. Health professionals are attached to this model through education, subsequent practice, and the lack of viable alternatives on a systemic scale. It is through its prism that phenomena having a largely psycho-social background are assessed. In practice, this means that experts judge people on the autism spectrum through their somatic or physical difficulties (performing basic self-care activities or sensory hypersensitivity) instead of difficulties in social relations or the emotional sphere, which often are fundamental. Furthermore, as David Pena-Guzman and Joel Reynolds showed: '(...) medical experts are also part of an institution with a long and dark history concerning disability. Historically, medicine has played a central role in the construction of disability as both spectacle and tragedy (...)'. (p. 223) This grounds communicative difficulties between disabled or mentally ill patients' communities and medical professionals. This state of affairs persists and is still reproduced in many opinion-forming discourses at the media and government levels . An analysis of psychoanalytic discourse on autistic patients showed that, although professionals try to maintain a commendable intellectual humility and break with many popular narratives concerning the nature and origins of autism, they still return to the dominant rhetoric of deficit and unhappiness that autism advocacy communities fight today . Medical professionals' reliance on presupposed epistemic schemes (such as ableism belief in the rare occurrence of autism in females or overdiagnosis of autism due to media interest) may lead to misdiagnosis and disrupt the flow of information needed for effective diagnosis or treatment planning. These are examples of 'breach in shared conceptual and interpretative resources' that can be the basis for hermeneutical injustice. Rooted in certain intellectual traditions and the lack of alternatives, these schemes may significantly affect 'experts' self-understanding and hinder their reliable performance of their duties. Hermeneutical injustice concerning the diagnosis of autism thus has two dialectical sides, so to speak, and one does not exist without the other. It is crucial for autism experts to be mindful of this potential for hermeneutical injustice and to work to actively address it by engaging in inclusive and collaborative practices that seek to understand and respect the perspectives of autistic individuals. This can include consulting with and listening to people on the spectrum, seeking out diverse perspectives and experiences, and actively working to dismantle any barriers that may prevent autistic people from fully participating in and contributing to discussions and decisions that affect them. Pena-Guzman and Reynolds seem to capture the root of this problem with regard to the relationship between medical professionals and disability communities, and patients: 'We submit that at the root of these mechanisms is the medical community's lack of engagement with critical, non-medical modes of knowledge concerning disability, including and especially with respect to knowledge created by disability communities themselves, as well as bodies of work which draw directly on such knowledge, as literature in disability studies and philosophy of disability regularly does. In other words, a root cause of ableism in medicine is medicine's understanding of disability as an objective lack rather than as a diverse set of phenomena that are thoroughly socially mediated'. (p. 225) The third source of epistemic injustice is popular discourses, media, academic, or scientific, of which experts by experience are a part. As far as autism is concerned, the media disseminate the image of people deprived of the ability to reason and, to a significant extent, self-determination. It is the image of people suffering because of their autism and not from the social environment not understanding their specific needs. More and more, but still insufficiently, attention is paid to experts by experience in various discourses. The overlapping scientific and media discourses radically marginalize the first-person perspective and the voice of people on the autism spectrum. Even when this voice does appear, it is usually not accorded proper epistemic authority. Decision-making theories indicate that when experts make quick and complex, intuitive decisions, they are exposed to the influence of stereotypical depictions of various mental conditions from media or outdated scientific discourse. One of the epistemic schemes used is the systemic marginalization of the voice of the mentally ill. Our previous empirical study showed how deeply rooted in background knowledge the preconceptions concerning autism might be, despite the high declarative quality of knowledge . According to Heggen and Berg: 'Epistemic injustice can be a consequence of low disease prestige and negative stereotypes leading to bias against the knower or privileging certain epistemic and practice ideals like EBP, or privileging knowledge derived from medical training and theory. Health personnel might have the very best intentions to trust a patient and believe what the patient is telling them but nevertheless ignore the patient's testimony, for instance because it is not in accordance with medical expertise. Consequently, patients testimonies are not considered credible, and they are undermined as first-hand knowers. Their reports about their condition are marginalized during medical examination and they encounter difficulties in their efforts to make themselves understood'. (p. 3) It is difficult to consider the social role played by experts and the social conditions for producing and using expert knowledge in isolation from ethical issues. The epistemic authority that allows a person to be referred to as an expert is irrevocably connected with adequate fulfillment of such a role. Following virtue epistemology, an expert should justify his/her beliefs reliably to make it possible. S/he should use expert-appropriate dispositions, and precisely those, which are often disturbed resulting in epistemic injustice. Author Contributions M.W.--conceptualization, investigation, writing--original draft. M.M.--conceptualization, writing--review and editing, supervision. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors have no competing interest to declare that are relevant to the content of this article. diagnostics-13-00927-t001_Table 1 Table 1 Epistemic injustice as affecting both service users and mental health experts. Testimonial Injustice Hermeneutical Injustice Sources of epistemic injustice - low level of intellectual humility - epistemic ignorance - outdated but still dominating discourses - hierarchy of epistemic authorities within mental healthcare systems - outdated but still dominating discourses - lack of critical evaluation of possessed knowledge and paradigms utilized Service users as objects of epistemic injustice - ignored as knowers and denied the epistemic authority - insufficient degree of input in one's diagnostic or therapeutical process - limited ability to interpret one's own experience outside of dominative, often harming, narratives Mental health experts as objects of epistemic injustice - ignored as a reliable source of knowledge by other experts and thus excluded from knowledge creation processes - unable to utilize a full potential of information sources in the decision-making processes - limited ability to understand patient's point of view - biased preconceptions about patient's condition - biased or limited preconceptions about their own expert position and its capabilities - sanctioned by traditional background assumptions concerning the nature of mental health phenomena Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Fricker M. Epistemic Injustice Oxford University Press Oxford, UK 2007 9780198237907 2. Fricker M. Evolving Concepts of Epistemic Injustice The Routledge Handbook of Epistemic Injustice Routledge London, UK 2017 53 60 10.4324/9781315212043-5 3. Grim K. Tistad M. Schon U.-K. Rosenberg D. The Legitimacy of User Knowledge in Decision-Making Processes in Mental Health Care: An Analysis of Epistemic Injustice J. Psychosoc. Rehabil. Ment. Health 2019 6 157 173 10.1007/s40737-019-00145-9 4. Berenstain N. Epistemic Exploitation Ergo Open Access J. Philos. 2016 3 569 590 10.3998/ergo.12405314.0003.022 5. Wieseler C. Epistemic Oppression and Ableism in Bioethics Hypatia 2020 35 714 732 10.1017/hyp.2020.38 6. Dotson K. Conceptualizing Epistemic Oppression Soc. Epistemol. 2014 28 115 138 10.1080/02691728.2013.782585 7. Dotson K. Tracking Epistemic Violence, Tracking Practices of Silencing Hypatia 2011 26 236 257 10.1111/j.1527-2001.2011.01177.x 8. Melvern L. Moral Equivalence J. Int. Peacekeeping 2020 22 190 198 10.1163/18754112-0220104012 9. Berenstain N. Dotson K. Paredes J. Ruiz E. Silva N.K. Epistemic Oppression, Resistance, and Resurgence Contemp. Political Theory 2022 21 283 314 10.1057/s41296-021-00483-z 10. Medina J. Varieties of Hermeneutical Injustice The Routledge Handbook of Epistemic Injustice Routledge London, UK 2017 41 52 10.4324/9781315212043-4 11. Lakeman R. Epistemic Injustice and the Mental Health Service User Int. J. Ment. Health Nurs. 2010 19 151 153 10.1111/j.1447-0349.2010.00680.x 20550637 12. Heggen K.M. Berg H. Epistemic Injustice in the Age of Evidence-Based Practice: The Case of Fibromyalgia Humanit. Soc. Sci. Commun. 2021 8 235 10.1057/s41599-021-00918-3 13. Harper D.J. Vakili K. Mental Health Prejudice, Discrimination and Epistemic Injustice: Moving beyond Stigma and Biomedical Dominance The Routledge International Handbook of Discrimination, Prejudice and Stereotyping Routledge London, UK 2021 26 41 10.4324/9780429274558-3 14. Kurs R. Grinshpoon A. Vulnerability of Individuals with Mental Disorders to Epistemic Injustice in Both Clinical and Social Domains Ethics Behav. 2017 28 336 346 10.1080/10508422.2017.1365302 15. Crichton P. Carel H. Kidd I.J. Epistemic Injustice in Psychiatry BJPsych Bull. 2017 41 65 70 10.1192/pb.bp.115.050682 28400962 16. Carel H. Kidd I.J. Epistemic Injustice in Healthcare: A Philosophial Analysis Med. Health Care Philos. 2014 17 529 540 10.1007/s11019-014-9560-2 24740808 17. Carel H. Gyorffy G. The Art of Medicine: Seen but Not Heard: Children and Epistemic Injustice Lancet 2014 384 1256 1257 10.1016/S0140-6736(14)61759-1 25289422 18. Sanati A. Kyratsous M. Epistemic Injustice in Assessment of Delusions J. Eval. Clin. Pract. 2015 21 479 485 10.1111/jep.12347 25828924 19. Dohmen J. A Little of Her Language: Epistemic Injustice and Mental Disability Res. Philos. 2016 93 669 691 10.11612/resphil.1475 20. Spandler H. McKeown M. Exploring the Case for Truth and Reconciliation in Mental Health Services Ment. Health Rev. J. 2017 22 83 94 10.1108/MHRJ-01-2017-0011 21. Harcourt E. Epistemic Injustice, Children and Mental Illness J. Med. Ethics 2021 47 729 735 10.1136/medethics-2021-107329 34172522 22. Harris O. Andrews C. Broome M.R. Kustner C. Jacobsen P. Epistemic Injustice amongst Clinical and Non-Clinical Voice-Hearers: A Qualitative Thematic Analysis Study Br. J. Clin. Psychol. 2022 61 947 963 10.1111/bjc.12368 35466414 23. Hull L. Petrides K.V. Allison C. Smith P. Baron-Cohen S. Lai M.C. Mandy W. "Putting on My Best Normal": Social Camouflaging in Adults with Autism Spectrum Conditions J. Autism Dev. Disord. 2017 47 2519 2534 10.1007/s10803-017-3166-5 28527095 24. Kent T. Cooke A. Marsh I. "The Expert and the Patient": A Discourse Analysis of the House of Commons' Debates Regarding the 2007 Mental Health Act J. Ment. Health 2022 31 152 157 10.1080/09638237.2020.1818706 32930654 25. Riccio A. Kapp S.K. Jordan A. Dorelien A.M. Gillespie-Lynch K. How Is Autistic Identity in Adolescence Influenced by Parental Disclosure Decisions and Perceptions of Autism? Autism 2021 25 374 388 10.1177/1362361320958214 32969254 26. Macleod A. Lewis A. Robertson C. 'Why Should I Be like Bloody Rain Man?!' Navigating the Autistic Identity Br. J. Spec. Educ. 2013 40 41 49 10.1111/1467-8578.12015 27. Bansal N. Karlsen S. Sashidharan S.P. Cohen R. Chew-Graham C.A. Malpass A. Understanding Ethnic Inequalities in Mental Healthcare in the UK: A Meta-Ethnography PLoS Med. 2022 19 e1004139 10.1371/journal.pmed.1004139 36512523 28. Tosas M.R. The Downgrading of Pain Sufferers' Credibility Philos. Ethics Humanit. Med. 2021 16 8 10.1186/s13010-021-00105-x 34610826 29. Bueter A. Diagnostic Overshadowing in Psychiatric-Somatic Comorbidity: A Case for Structural Testimonial Injustice Erkenntnis 2021 1 21 10.1007/s10670-021-00396-8 30. Michaels J.A. Potential for Epistemic Injustice in Evidence-Based Healthcare Policy and Guidance J. Med. Ethics 2021 47 417 422 10.1136/medethics-2020-106171 32461243 31. Drozdzowicz A. Epistemic Injustice in Psychiatric Practice: Epistemic Duties and the Phenomenological Approach J. Med. Ethics 2021 47 e69 10.1136/medethics-2020-106679 32. Pot M. Epistemic Solidarity in Medicine and Healthcare Med. Health Care Philos. 2022 25 681 692 10.1007/s11019-022-10112-0 36045178 33. Wodzinski M. Rzadeczka M. Moskalewicz M. How to Minimize the Impact of Experts' Non-Rational Beliefs on Their Judgments on Autism Community Ment. Health J. 2022 1 14 10.1007/s10597-022-01062-1 34. Owen D. Hubris Syndrome Clin. Med. 2008 8 428 432 10.7861/clinmedicine.8-4-428 35. Vogelstein E. Professional Hubris and Its Consequences: Why Organizations of Health-Care Professions Should Not Adopt Ethically Controversial Positions Bioethics 2016 30 234 243 10.1111/bioe.12186 26307439 36. Bak W. Kutnik J. Domains of Intellectual Humility: Self-Esteem and Narcissism as Independent Predictors Pers. Individ. Differ. 2021 177 110815 10.1016/j.paid.2021.110815 37. Bak W. Wojtowicz B. Kutnik J. Intellectual Humility: An Old Problem in a New Psychological Perspective Curr. Issues Personal. Psychol. 2022 10 85 97 10.5114/cipp.2021.106999 38. Catala A. Metaepistemic Injustice and Intellectual Disability: A Pluralist Account of Epistemic Agency Ethical Theory Moral Pract. 2020 23 755 776 10.1007/s10677-020-10120-0 39. Lakeman R. Family and Carer Participation in Mental Health Care: Perspectives of Consumers and Carers in Hospital and Home Care Settings J. Psychiatr. Ment. Health Nurs. 2008 15 203 211 10.1111/j.1365-2850.2007.01213.x 18307649 40. Sarrett J.C. Trapped Children: Popular Images of Children with Autism in the 1960s and 2000s J. Med. Humanit. 2011 32 141 153 10.1007/s10912-010-9135-z 21225325 41. Reading A. Neurodiversity and Communication Ethics: How Images of Autism Trouble Communication Ethics in the Globital Age Cult. Stud. Rev. 2018 24 113 129 10.5130/csr.v24i2.6040 42. Douglas P. Rice C. Runswick-Cole K. Easton A. Gibson M.F. Gruson-Wood J. Klar E. Shields R. Re-Storying Autism: A Body Becoming Disability Studies in Education Approach Int. J. Incl. Educ. 2019 25 605 622 10.1080/13603116.2018.1563835 43. Baroutsis A. Eckert C. Newman S. Adams D. How Is Autism Portrayed in News Media? A Content Analysis of Australian Newspapers Articles from 2016-2018 Disabil. Soc. 2021 1 24 10.1080/09687599.2021.1971067 44. Billawala A. Wolbring G. Analyzing the Discourse Surrounding Autism in the New York Times Using an Ableism Lens Disabil. Stud. Q 2014 34 10.18061/dsq.v34i1.3348 45. Holton A.E. Farrell L.C. Fudge J.L. A Threatening Space?: Stigmatization and the Framing of Autism in the News Commun. Stud. 2014 65 189 207 10.1080/10510974.2013.855642 46. Huws J.C. Jones R.S.P. "They Just Seem to Live Their Lives in Their Own Little World": Lay Perceptions of Autism Disabil. Soc. 2010 25 331 344 10.1080/09687591003701231 47. Jones S.C. Harwood V. Representations of Autism in Australian Print Media Fac. Health Behav. Sci. Pap. 2009 24 5 18 10.1080/09687590802535345 48. Wodzinski M.C. Case Study: Cognitive Errors in Court Experts' Assessments in Autism Spectrum Cases Adv. Psychiatry Neurol. Postepy Psychiatr. I Neurol. 2021 29 259 264 10.5114/PPN.2020.103638 49. Bagatell N. Orchestrating Voices: Autism, Identity and the Power of Discourse Disabil. Soc. 2007 22 413 426 10.1080/09687590701337967 50. Anderson-Chavarria M. The Autism Predicament: Models of Autism and Their Impact on Autistic Identity Disabil. Soc. 2021 37 1321 1341 10.1080/09687599.2021.1877117 51. Bie B. Tang L. Representation of Autism in Leading Newspapers in China: A Content Analysis Health Commun. 2015 30 884 893 10.1080/10410236.2014.889063 25074820 52. Clarke J.N. Representations of Autism in US Magazines for Women in Comparison to the General Audience J. Child. Media 2012 6 182 197 10.1080/17482798.2011.587143 53. Robertson S.M. Neurodiversity, Quality of Life, and Autistic Adults: Shifting Research and Professional Focuses onto Real-Life Challenges Disabil. Stud. Q 2009 30 10.18061/dsq.v30i1.1069 54. Ericsson K.A. Hoffman R.R. Kozbelt A. Williams A.M. Cambridge Handbook Expertise and Expert Performance 2nd ed. Cambridge University Press Cambridge, UK 2018 9781107137554 55. Collins H. Evans R. Rethinking Expertise University of Chicago Press Chicago, IL, USA 2007 9780226113616 56. Collins H. Are Experts Right or Are They Members of Expert Groups? Soc. Epistemol. 2018 32 351 357 10.1080/02691728.2018.1546346 57. Collins H. Evans R. A Sociological/Philosophical Perspective on Expertise: The Acquisition of Expertise through Socialization The Cambridge Handbook of Expertise and Expert Performance Cambridge University Press Cambridge, UK 2018 21 32 58. Grundmann R. The Problem of Expertise in Knowledge Societies Minerva 2017 55 25 48 10.1007/s11024-016-9308-7 28239194 59. Eyal G. The Crisis of Expertise Polity Press Cambridge, UK 2019 978-0-745-66577-1 60. Turner S.P. The Politics of Expertise Routledge London, UK 2014 9781138929630 61. Nichols T.M. The Death of Expertise: The Campaign against Established Knowledge and Why It Matters Oxford University Press Oxford, UK 2019 0190469412 62. Ryle G. The Concept of Mind Penguin Books Sydney, Australia 1990 63. Hetmanski M. Expertise and Expert Knowledge in Social and Procedural Entanglement Eidos. A J. Philos. Cult. 2020 4 6 22 10.14394/eidos.jpc.2020.0013 64. Hetmanski M. Expert Knowledge: Its Structure, Functions and Limits Stud. Hum. 2018 7 11 20 10.2478/sh-2018-0014 65. Stewart C. Expertise and Authority Episteme 2020 17 420 437 10.1017/epi.2018.43 66. Goldman A. Whitcomb D. Social Epistemology: Essential Readings: Essential Readings Oxford University Press Oxford, UK 2010 67. Lavazza A. Farina M. Experts, Naturalism, and Democracy J. Theory Soc. Behav. 2022 52 279 297 10.1111/jtsb.12321 68. Lavazza A. Farina M. The Virtues Needed by Experts in Action Epistemol. Philos. Sci. 2021 58 142 157 10.5840/eps202158466 69. Majdik Z.P. Keith W.M. The Problem of Pluralistic Expertise: A Wittgensteinian Approach to the Rhetorical Basis of Expertise Soc. Epistemol. 2011 25 275 290 10.1080/02691728.2011.578307 70. Gigerenzer G. Simply Rational Oxford University Press Oxford, UK 2015 10.1093/ACPROF:OSO/9780199390076.001.0001 71. Kahneman D. Thinking, Fast and Slow 1st ed. Farrar, Straus and Giroux New York, NY, USA 2013 72. Tversky A. Kahneman D. Judgment under Uncertainty: Heuristics and Biases New Ser. 1974 185 1124 1131 10.1126/science.185.4157.1124 73. Meehl P.E. Clinical versus Statistical Prediction: A Theoretical Analysis and a Review of the Evidence University of Minnesota Press Minneapolis, MN, USA 1954 0963878492 74. Oskamp S. Overconfidence in Case-Study Judgments J. Consult. Psychol. 1965 29 261 265 10.1037/h0022125 14303514 75. Draaisma D. Stereotypes of Autism Philos. Trans. R. Soc. B Biol. Sci. 2009 364 1475 1480 10.1098/rstb.2008.0324 76. Hacking I. Kinds of People: Moving Targets: British Academy Lecture Proceedings of the British Academy Oxford University Press Oxford, UK 2007 151 10.5871/BACAD/9780197264249.003.0010 77. Hacking I. Autism Fiction: A Mirror of an Internet Decade? Univ. Tor. Q. 2010 79 632 655 10.3138/utq.79.2.632 78. Shanteau J. Competence in Experts: The Role of Task Characteristics Organ. Behav. Hum. Decis. Process. 1992 53 252 266 10.1016/0749-5978(92)90064-E 79. Moskalewicz M. Goze T. Clinical Judgment of Schizophrenia: Praecox Feeling and the Bizarreness of Contact--Open Controversies The Clinician in the Psychiatric Diagnostic Process Springer International Publishing Cham, Germany 2022 135 149 10.1007/978-3-030-90431-9_9 80. Moskalewicz M. Kordel P. Brejwo A. Schwartz M.A. Goze T. Psychiatrists Report Praecox Feeling and Find It Reliable. A Cross-Cultural Comparison Front. Psychiatry 2021 12 642322 10.3389/fpsyt.2021.642322 33746799 81. Chase W.G. Simon H.A. The Mind's Eye in Chess Visual Information Processing Academic Oxford, UK 1973 215 281 0121701506 82. Simon H.A. What Is an "Explanation" of Behavior? Psychol. Sci. 1992 3 150 161 10.1111/j.1467-9280.1992.tb00017.x 83. Damasio A.R. Descartes' Error: Emotion, Reason and The Human Brain Penguin Books London, UK 2015 9780099501640 84. Anders Ericsson K. Deliberate Practice and Acquisition of Expert Performance: A General Overview Acad. Emerg. Med. 2008 15 988 994 10.1111/j.1553-2712.2008.00227.x 18778378 85. Berg H. Virtue Ethics and Integration in Evidence-Based Practice in Psychology Front. Psychol. 2020 11 258 10.3389/fpsyg.2020.00258 32132960 86. Pietrini P. Lavazza A. Farina M. COVID-19 and Biomedical Experts: When Epistemic Authority Is (Probably) Not Enough J. Bioethical Inq. 2022 19 135 142 10.1007/s11673-021-10157-5 87. Bokros S.E. A Deference Model of Epistemic Authority Synthese 2021 198 12041 12069 10.1007/s11229-020-02849-z 88. Croce M. Expert-Oriented Abilities vs. Novice-Oriented Abilities: An Alternative Account of Epistemic Authority Episteme 2018 15 476 498 10.1017/epi.2017.16 89. BonJour L. Internalism and Externalism The Oxford Handbook of Epistemology Oxford University Press Oxford, UK 2002 234 260 10.1093/0195130057.003.0008 90. Sosa E. Knowledge in Perspective: Selected Essays in Epistemology Cambridge University Press Cambridge, UK 1991 9780521356282 91. Sosa E. Reflective Knowledge in the Best Circles J. Philos. 1997 94 410 430 10.2307/2564607 92. Goldman A.I. Social Process Reliabilism: Solving Justification Problems in Collective Epistemology Essays Collect. Epistemol. 2014 11 41 10.1093/ACPROF:OSO/9780199665792.003.0002 93. Goldman A.I. Epistemic Folkways and Scientific Epistemology Philos. Issues 1993 3 271 285 10.2307/1522948 94. Goldman A.I. What Is Justified Belief? Justif. Knowl. 1979 1 23 10.1007/978-94-009-9493-5_1 95. Sosa E. Goldman's Reliabilism and Virtue Epistemology Philos. Top. 2001 29 383 400 10.5840/philtopics2001291/214 96. Sosa E. Plantinga A. Proper Functionalism and Virtue Epistemology Nous 1993 27 51 65 10.2307/2215895 97. Greco J. Virtues and Vices of Virtue Epistemology Can. J. Philos. 1993 23 413 432 10.1080/00455091.1993.10717329 98. Turri J. From Virtue Epistemology to Abilism Character Oxford University Press Oxford, UK 2015 315 330 99. Stichter M. Ethical Expertise: The Skill Model of Virtue Ethical Theory Moral Pract. 2007 10 183 194 10.1007/s10677-006-9054-2 100. Sosa E. Virtue Epistemology: Character versus Competence Judgm. Agency 2015 34 61 10.1093/ACPROF:OSO/9780198719694.003.0002 101. Pritchard D. Extended Virtue Epistemology Inquiry 2018 61 632 647 10.1080/0020174X.2017.1355842 102. Moskalewicz M. Schwartz M.A. Goze T. Phenomenology of Intuitive Judgment: Praecox-Feeling in the Diagnosis of Schizophrenia AVANT J. Philos. Interdiscip. Vanguard 2018 9 63 74 10.26913/avant.2018.02.04 103. Goze T. Moskalewicz M. Schwartz M.A. Naudin J. Micoulaud-Franchi J.-A. Cermolacce M. Is "Praecox Feeling" a Phenomenological Fossil? A Preliminary Study on Diagnostic Decision Making in Schizophrenia Schizophr. Res. 2019 204 413 414 10.1016/j.schres.2018.07.041 30072280 104. Rumke H.C. Neeleman J. The Nuclear Symptom of Schizophrenia and the Praecoxfeeling Hist. Psychiatry 1990 1 331 341 10.1177/0957154X9000100304 11622458 105. Grube M. Towards an Empirically Based Validation of Intuitive Diagnostic: Rumke's 'Praecox Feeling' across the Schizophrenia Spectrum: Preliminary Results Psychopathology 2006 39 209 217 10.1159/000093921 16778451 106. Ungvari G.S. Xiang Y.-T. Hong Y. Leung H.C.M. Chiu H.F.K. Diagnosis of Schizophrenia: Reliability of an Operationalized Approach to 'Praecox-Feeling' Psychopathology 2010 43 292 299 10.1159/000318813 20639689 107. Stanghellini G. The Power of Images and the Logics of Discovery in Psychiatric Care Brain Sci. 2022 13 13 10.3390/brainsci13010013 36671995 108. Pena-Guzman D.M. Reynolds J.M. The Harm of Ableism: Medical Error and Epistemic Injustice Kennedy Inst. Ethics J. 2019 29 205 242 10.1353/ken.2019.0023 31656232 109. McKeever B.W. News Framing of Autism Sci. Commun. 2012 35 213 240 10.1177/1075547012450951 110. Georgiou K. Winter D. Davies S. Katsiana A. "But Who Knows What Autism Is?"; Negotiating the Notion of Autism During Free Associative Narrative Interviews with Psychoanalytic Psychotherapists Mater. Socio Med. 2021 33 138 10.5455/msm.2021.33.138-144 111. Wodzinski M. Golaska-Ciesielska P. The Construction of Autism: Between Reflective and Background Knowledge Constr. Found. 2021 17 33 47 112. Moskalewicz M. Is Heidegger's Fundamental Ontology an Antidote to Dominant Social Constructions of Autism? Constr. Found. 2021 17 62 64 |
PMC10000602 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050744 cells-12-00744 Article Heparan Sulfates Regulate Axonal Excitability and Context Generalization through Ca2+/Calmodulin-Dependent Protein Kinase II Song Inseon Methodology Formal analysis Investigation Writing - original draft Visualization 1 Kuznetsova Tatiana Formal analysis Investigation Writing - review & editing 2 Baidoe-Ansah David Formal analysis Investigation Writing - original draft Visualization 1 Mirzapourdelavar Hadi Methodology Data curation Writing - review & editing 1 Senkov Oleg Methodology Formal analysis Investigation Writing - review & editing 1 Hayani Hussam Investigation Writing - review & editing 1 Mironov Andrey Methodology Investigation Writing - review & editing 1 Kaushik Rahul Methodology Writing - review & editing Supervision 1 Druzin Michael Methodology Formal analysis Data curation Visualization Supervision 2 Johansson Staffan Methodology Resources Supervision Funding acquisition 2 Dityatev Alexander Conceptualization Methodology Validation Resources Writing - original draft Writing - review & editing Supervision Project administration Funding acquisition 134* Buday Laszlo Academic Editor 1 Molecular Neuroplasticity Group, German Center for Neurodegenerative Diseases (DZNE), 39120 Magdeburg, Germany 2 Department of Integrative Medical Biology, Umea University, 90187 Umea, Sweden 3 Medizinische Fakultat, Otto-von-Guricke-Universitat Magdeburg, 39120 Magdeburg, Germany 4 Center for Behavioral Brain Sciences (CBBS), 39106 Magdeburg, Germany * Correspondence: [email protected]; Tel.: +49-391-67-24526; Fax: +49-391-6724530 25 2 2023 3 2023 12 5 74401 2 2023 21 2 2023 23 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Our previous studies demonstrated that enzymatic removal of highly sulfated heparan sulfates with heparinase 1 impaired axonal excitability and reduced expression of ankyrin G at the axon initial segments in the CA1 region of the hippocampus ex vivo, impaired context discrimination in vivo, and increased Ca2+/calmodulin-dependent protein kinase II (CaMKII) activity in vitro. Here, we show that in vivo delivery of heparinase 1 in the CA1 region of the hippocampus elevated autophosphorylation of CaMKII 24 h after injection in mice. Patch clamp recording in CA1 neurons revealed no significant heparinase effects on the amplitude or frequency of miniature excitatory and inhibitory postsynaptic currents, while the threshold for action potential generation was increased and fewer spikes were generated in response to current injection. Delivery of heparinase on the next day after contextual fear conditioning induced context overgeneralization 24 h after injection. Co-administration of heparinase with the CaMKII inhibitor (autocamtide-2-related inhibitory peptide) rescued neuronal excitability and expression of ankyrin G at the axon initial segment. It also restored context discrimination, suggesting the key role of CaMKII in neuronal signaling downstream of heparan sulfate proteoglycans and highlighting a link between impaired CA1 pyramidal cell excitability and context generalization during recall of contextual memories. extracellular matrix synaptic plasticity context discrimination axon initial segment axonal excitability ankyrin G Karin and Harald Silvander FundInsamlingsstiftelsen at Umea University Medical FacultyThis study was supported by CRC 1436 A05 Project-ID 425899996, the Research Training Group 2413 SynAGE TP6, the Karin and Harald Silvander Fund, and Insamlingsstiftelsen at Umea University Medical Faculty. pmc1. Introduction Heparan sulfate proteoglycans (HSPGs) harbor long chains of variously sulfated polysaccharide residues. There are membrane-bound HSPGs, such as syndecans and glypicans, and secreted HSPGs, including agrin, perlecan, and collagen type XVIII. An increasing number of studies demonstrate that HSPGs have an important role in the nervous system during development and adulthood. In the mouse brain, syndecan-1 and glypican-4 are highly expressed in the neural tube, where the precursor cells are proliferating . These HSPGs are important for the proliferation of neural precursor cells and play a role as synaptic organizing molecules during synaptogenesis. Their heparan sulfate (HS) chains are essential for this role. Glypican 4 is bound to the presynaptic membrane via a GPI anchor and interacts with the postsynaptic protein, LRRTM4 (leucine-rich repeat transmembrane neuronal proteins), forming a trans-synaptic complex. This complex recruits other synaptic molecules to the synaptic cleft, contributing to the maturation of excitatory synapses. Mice deficient in glypican 4 exhibit a decreased number of synapses along with decreased expression of postsynaptic glutamate receptor subunit GluA1 and increased retention of presynaptic neuronal pentraxin 1 . Syndecans are differentially expressed in various neural cell types and exhibit differential subcellular localization in neurons . Contrary to glypicans lacking a cytoplasmic domain, transmembrane syndecans interact with specific cytoplasmic binding partners, such as Ca2+/calmodulin-dependent serine protein kinase (CASK), syntenin, synectin, and synbindin . Syndecan 2 is highly expressed in synapses and influences activities of postsynaptic scaffolding proteins, thereby contributing to filopodia and dendritic spine formation . Overexpression of full-length syndecan 2 in cultured immature hippocampal neurons accelerates dendritic spine formation, while a syndecan 2 deletion mutant that lacks the ability to bind to synthenin and CASK does not support spine maturation . The association of cortactin and fyn with syndecan is increased rapidly after induction of long-term potentiation (LTP), while inclusion of soluble syndecan 3 into the rat hippocampal slices inhibits high-frequency stimulation-induced LTP . Furthermore, syndecan 3 knockout mice exhibit strongly enhanced LTP and impaired hippocampus-dependent memory . Secreted HSPG agrin is also involved in filopodia and dendritic spine formation. While downregulation of agrin in the cultured neurons in vitro and in vivo reduces the number of dendritic filopodia, overexpression of agrin in rodent hippocampal neurons stimulates filopodia formation in vitro . An increasing number of structural, pharmacological, and genetic studies suggest a key role of the HS chains carried by HSPGs in mediating their activities. Interestingly, HSs bind to receptor protein tyrosine phosphatase sigma (RPTPs) at the same site as chondroitin sulfates. Crystallographic analyses of this site reveal conformational plasticity that can accommodate diverse glycosaminoglycans with comparable affinities. HSs induced RPTPs ectodomain oligomerization, stimulating neurite outgrowth. The oligomerization was inhibited by chondroitin sulfates, resulting in impaired neurite outgrowth . In acute hippocampal slices, treatment with a mixture of heparinases 1 and 3, which removes highly and low sulfated HSs, respectively, impaired LTP of synaptic transmission . This treatment also prevented the increase in the number of spines after induction of N-methyl-D-aspartate (NMDA) receptor-dependent LTP . Conditional ablation of Ext1, a gene involved in HS synthesis, in a subpopulation of pyramidal neurons leads to an autistic phenotype , providing genetic evidence for the importance of HSs in shaping brain function on many levels, from cellular properties to complex behaviors. More targeted ablation of HSs on neurexin-1 also revealed structural and functional deficits at central synapses. HS directly binds postsynaptic partners neuroligins and LRRTMs . Considering the high heterogeneity of HSs, we focused on a highly sulfated subset of HSs (HSHSs), which could be digested by heparinase 1. Such treatment of cultured hippocampal neurons resulted in a reduction in the mean firing rate of neurons , despite the upregulation of GluA1 protein expression . Acute treatment of hippocampal slices with heparinase 1 reduced CA1 pyramidal cellular excitability and impaired hippocampal LTP . Altered expression of ankyrin G (AnkG), as one of the major organizing proteins at the axon initial segment (AIS) in heparinase 1-treated hippocampal slices, led us to the hypothesis that HSHSs are involved in the modulation of neuronal activity through the changes in the AIS composition and function. Injection of heparinase 1 before fear conditioning impaired context discrimination , validating the importance of HSHSs at the systemic level. Based on previous in vitro findings of increased autophosphorylation levels of CaMKII a and b isoforms after heparinase 1 treatment , we hypothesized that CaMKII is the key molecule involved in the modulation of axonal excitability due to a loss of HSHSs and provided in vitro and in vivo evidence verifying this hypothesis biochemically, immunocytochemically, and electrophysiologically. Our studies show that an increased level of autophosphorylated CaMKII in heparinase-treated neurons is responsible for reduced neuronal excitability, altered expression of AnkG in the AIS of CA1 pyramidal neurons, and impaired contextual discrimination. 2. Materials and Methods 2.1. Immunoblot Analysis To access the level of endogenous CaMKII isoform expression and the level of its phosphorylation, murine hippocampal slices (treated with intact or heat-inactivated heparinase 1 in the same way as for electrophysiological recordings) were snap-frozen in isopropanol pre-cooled on dry ice. Later samples were homogenized in radioimmunoprecipitation assay (RIPA) buffer (ThermoFisher Scientific, Rockford, IL, USA) containing a protease inhibitor cocktail (Sigma-Aldrich P1860, St. Louis, MO, USA), a serine/threonine phosphatase inhibitor (Sigma-Aldrich P0044, St. Louis, MO, USA), and a tyrosine phosphatase inhibitor (Sigma-Aldrich P5726, St. Louis, MO, USA) using a glass tissue homogenizer. Non-soluble proteins were separated via centrifugation at 20,000 g for 15 min at 4 degC. The protein concentration of individual samples was measured using a DC Protein Assay (Bio-Rad, Hercules, CA, USA). A total of 10-30 mg of extract was resuspended in reducing (5.0% 2-mercaptoethanol) sample buffer (Bio-Rad, Hercules, CA, USA) and boiled at 100 degC for 5 min, separated via SDS-PAGE on 10% acrylamide gels, and transferred to the polyvinylidene difluoride membranes. Membranes were blocked for 1 h at room temperature with 5% Blotting-Grade Blocker (Bio-Rad, 1706404, Hercules, CA, USA) in Tris-buffered saline with Tween20 (TBS-T buffer), probed with appropriate primary antibody at 4 degC overnight and then for 1 h at room temperature with horseradish peroxidase (HRP)-conjugated secondary antibodies. To estimate the total expression of a and b forms of CaMKII, mouse anti-CaMKII (G1, sc-5306; 1:200-1:1000) from Santa-Cruz (Paso Robles, CA, USA) was used. To induce activation of CaMKII, rabbit anti-phospho Thr 286/287 CaMKII (p1005-286; 1:1000) from PhosphoSolutions (Aurora, CO, USA) was applied. To evaluate the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) level, mouse anti-GAPDH (MAB374; 1:15.000-20.000) from Millipore (Bedford, MA, USA) was used. HRP-conjugated secondary antibodies were donkey anti-rabbit (NA934V) from GE Healthcare (Buckinghamshire, UK), or goat anti-mouse (115-035-146) from Jackson ImmunoResearch (West Grove, PA, USA). Acquisition of chemiluminescent signal and densitometric analysis were performed using an Odyssey Fc Imaging System (LI-COR, NE, USA) and Image Studio Lite 5.2.5 software, respectively. The total levels of a or b forms of CaMKII and phospho Thr 286/287 levels were standardized to the level of loading control (GAPDH) in each sample. Standardized values were further normalized to the randomly chosen control sample (loaded in each gel). To evaluate CaMKII phosphorylation on Thr 286/287, the ratios of phospho-Thr 286/287 signal to the total amount of CaMKII protein were calculated. For statistical evaluation and the graphical representation of the data, the OriginPro 2022 9.9.0.225 software was used. The average +- SEM (standard error of mean) was calculated for control and experimental (heparinase-treated) groups, normalized to randomly chosen control samples. Statistical evaluation was carried out using a Mann-Whitney-Wilcoxon test. 2.2. Slice Preparation and In Vitro Electrophysiology Acute hippocampal slices were prepared as described elsewhere from 5-week-old male C57Bl/6 mice 1 day after injection with Heparinase 1, Ctrl (heat-inactivated Heparinase 1), or Heparinase 1 + autocamtide-2-related inhibitory peptide (AIP) into hippocampal CA1 area, which described below (Section 2.4). Transverse 350 mm thick hippocampal slices were obtained in ice-cold slice solution containing (in mM) 240 sucrose, 2 KCl, 2 MgSO4, 1.25 NaH2PO4, 26 NaHCO3, 1 CaCl2, 1 MgCl2, and 10 D-glucose. After slice recovery at room temperature, the slices were transferred to a submerged recording chamber and were perfused with ACSF (2-3 mL/min) containing (in mM) 124 NaCl, 2.5 KCl, 1.3 MgSO4, 1 NaH2PO4, 26.2 NaHCO3, 2.5 CaCl2, 1.6 MgCl2, and 11 D-glucose. The solution was saturated with 95% O2/5% CO2 (Osmolarity, 290 +- 5 mOsm). Whole-cell patch clamp recordings were obtained from visually identified CA1 pyramidal neurons with a glass electrode (4-5 MO, Hilgenberg, Germany) containing (in mM) 120 K-gluconate, 10 KCl, 3 MgCl2, 0.5 EGTA, 40 HEPES, 2 MgATP, and 0.3 NaGTP (pH 7.2 with KOH, 295 mOsm) for measuring neuronal excitability. In the current clamp configuration, cells were held at -70 mV and injected from -75 mV to +400 pA with 25 pA increments. For measuring miniature excitatory postsynaptic currents (mEPSCs), 5 mM QX314 was added into the intracellular pipette solution while GABAA receptor antagonist picrotoxin (PTX, 50 mM, Tocris, Bristol, UK), GABAB receptor antagonist CGP55845 (3 mM, Tocris, Bristol, UK), and Na+ channel blocker tetrodotoxin (TTX, 1 mM, Tocris, Bristol, UK) were added to ACSF. Miniature inhibitory postsynaptic currents (mIPSCs) were recorded with a glass electrode containing (in mM) 120 CsCl, 8 NaCl, 0.2 MgCl2, 10 HEPES, 2 EGTA, 0.3 Na3GTP, and 2 MgATP at pH 7.2 with CsOH, 290 mOsm. NBQX (25 mM, Tocris, Bristol, UK), D-APV (50 mM, Tocris, Bristol, UK), and TTX (1 mM) were added to ACSF to isolate action potential-independent mIPSCs. In vitro electrophysiological data were acquired using an EPC 10 amplifier (HEKA Elektronik, Germany) at a sampling rate of 10 kHz and low-pass-filtered at 2-3 kHz. The obtained data were analyzed offline using PatchMaster software v2X69 (HEKA Electronik, Germany), Clampfit 10 (Molecular Devices, U.S.A.), or MiniAnalysis (6.0.3 Synaptosoft, U.S.A.). The data were presented and analyzed using SigmaPlot 12 (Systat Software Inc, U.S.A.) and Prism 7 (GraphPad software, U.S.A.). 2.3. Immunocytochemistry in Hippocampal Cultured Neurons Hippocampal neurons from embryonic C57BL6/J mice (E18) were extracted and cultured as described earlier . The neuronal cells were plated on polyethyleneimine-coated (Sigma-Aldrich; 408727-100 mL) 18 mm coverslips in 12-well plates at a cell density of 150,000 per well. Neurons were maintained in 1 mL of neurobasal media (NB+ media) (Thermo Fisher Scientific, Waltham, MA, USA) containing 2% B27 and 1% L-glutamine and 1% Pen-Strep (Gibco, Grand Island, NY, USA). Cultured neurons were fed with 250 mL of NB+ media on days in vitro (DIV) 14 and 17. On DIV 21-23, cultured hippocampal neurons were incubated with Heparinase-1 (0.5 U/mL, Sigma-Aldrich, H2519, MO, USA), Ctrl (heat-inactivated Heparinase-1), or Heparinase-1 + AIP (100 nM), as previously described , for 2 h at 37 degC. After the treatment, hippocampal neurons were washed with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde (PFA) for 10 min, and then permeabilized with 0.1% Triton-X-100 in PBS for 10 min, washed 3 times, and blocked (0.1% Glycine + 0.1% Tween-20 + 10% Normal Goat Serum in PBS) for 60 min at room temperature. Then, the cells were stained with antibodies against AnkG (mouse monoclonal, 1:1000; Millipore, MABN466), pCaMKII (rabbit polyclonal, 1:1000; Phospho Solution, P-1005-286), MAP2 (chicken polyclonal, 1:500; Abcam, ab5543) and DAPI (Life Technologies, S36939), and finally mounted (Fluoromount; Sigma Aldrich, F4680-25ML, MO, USA) for imaging. Mounted coverslips were imaged using a Zeiss LSM 700 confocal microscope with a 63x/1.4 NA oil immersion objective. Image analysis was carried out as previously described . Using the microtubule-associated protein 2 (MAP2) and AnkG signals, the AISs were analyzed from the soma edge over a 40 mm long distance with a line profile (width = 3.0) using Fiji (ImageJ version 1.53c) . 2.4. In Vivo Intrahippocampal Injection Adult ( 4-month-old) male C57Bl/6j mice (Charles River) were used. At least 1 week before starting the experiments, mice were transferred to a small vivarium, where they were housed individually with food and water ad libitum on a reversed 12:12 light/dark cycle (light on at 9:00 p.m). All behavioral experiments were performed in the afternoons during the dark phase of the cycle when the mice were active, under constant temperature (22 +- 1 degC) and humidity (55 +- 5%). All treatments and behavioral procedures were conducted in accordance with ethical animal research standards defined by German law and approved by the Ethical Committee on Animal Health and Care of the State of Saxony-Anhalt, Germany, under the license numbers 42502-2-1159 and -1322 DZNE. Injection guide cannulas and electrodes were implanted as previously described , but electrophysiological analysis is not included in the present study due to an insufficient quality of recordings. The coordinates for bilateral cannulas were AP = 2.0 mm and L = +-2.2 mm from bregma and midline, respectively. For intrahippocampal injection, we used a digitally controlled infusion system (UltraMicroPump, UMP3, and Micro4 Controller, WPI, U.S.A.) fed with a 10 mL Hamilton syringe and a NanoFil (35 GA) bevel-tip needle, as previously described . The mouse was first anesthetized with 1-3% isoflurane and put into the stereotaxic frame. Heparinase 1 (Hep) from Flavobacterium heparinum (0.05 U/mL/site, Sigma-Aldrich, H2519), Heparinase 1 heat-inactivated at 100 oC for 30 min (Ctrl), or Heparinase 1 + autocamtide-2-related inhibitory peptide (0.17 mg/mL/site, Sigma-Aldrich, SCP0001) (Hep + AIP) was injected into the hippocampal CA1 area at a rate of 3 nl/s. After waiting for another 5 min, the injection needle was removed. 2.5. Fear Conditioning In this study, we used the previously described classical Pavlovian contextual fear conditioning paradigm in mice with a slight modification . In this study, on day 0 (d0), mice were initially placed in a 20 x 20 x 30 cm chamber with a neutral context (CC-), gray walls, and gray plastic floor for 5 min. Next, mice were exposed to the conditioned context (CC+), which includes patterned walls and a metal grid floor, for 5 min after an interval of 1 h. During the CC+ phase, mice's feet were shocked 3 times with mild intensity (0.5 mA, 1 s) with a 1 min inter-shock interval. Using a computerized fear conditioning system (Ugo Basile, Italy), the first memory retrieval session was carried out for 5 min for each mouse on day 1 (d1) with a 1 h interval following the sequence CC+. On day 2 (d2), mice were injected with the vehicle, Heparinase and Heparinase + AIP, into the hippocampus. Then, on day 3 (d3) the second memory retrieval test (test 2) was performed using a similar paradigm as that used on d1. A blinded trained observer used video recordings of each session for offline fear-conditioned behavioral analysis with the help of behavioral video acquisition and analysis software (ANY-maze, version 4.99, Stoelting Co., Wood Dale, IL). Finally, the overall context memory and discrimination performance for each mouse was estimated. 2.6. Statistics Numerical data are reported as mean +- SEM, with n being the number of samples. Student's t-test and multi-way ANOVA with suitable post hoc tests were used as indicated and performed in SigmaPlot or Prism. For non-Gaussian distributions, we used the Mann-Whitney-Wilcoxon test. Significance levels (p-values) are indicated in figures by asterisks. 3. Results 3.1. Heparinase Treatment Elevates CaMKIIb Autophosphorylation in the Mouse Hippocampus We previously observed an increase in the GluA1 expression and CaMKII activity in cultured mouse hippocampal neurons after heparinase 1 treatment . To investigate whether heparinase treatment also changes hippocampal CaMKII activity in vivo, Ctrl (heat-inactivated heparinase) or active heparinase 1 was injected into the dorsal hippocampus of 6-week-old mice. To investigate the level of endogenous CaMKII isoform protein expression and their activity, 24 h after injection of heparinase, hippocampal slices were acutely prepared and used for immunoblotting. CaMKIIa and CaMKIIb are major isoforms in the hippocampus, and these molecules are activated during memory formation. Activation of CaMKIIa and CaMKIIb was assessed via the analysis of phosphorylation at Thr286 and Thr287, respectively . Consistent with our previous observation for the cultured hippocampal neurons , the activity of CaMKIIb was strongly affected after heparinase injection in vivo. The ratio of phosphorylated to total CaMKIIb was increased after heparinase treatment, while the effect on CaMKIIa was less prominent . 3.2. Enzymatic Digestion of HSHSs Does Not Change Synaptic Transmission to CA1 Pyramidal Neurons Having found that heparinase treatment in vitro can up-scale mEPSCs, we measured glutamatergic transmission (mEPSC) and GABAergic transmission (mIPSC) to CA1 pyramidal neurons 1 day after heparinase injection in vivo. Unexpectedly, we did not observe changes in the mEPSCs' amplitude or frequency . Additionally, temporal parameters such as the rise and decay times were not affected . The properties of mIPSCs were also unchanged by heparinase treatment . 3.3. Impaired Neuronal Excitability after In Vivo Injection of Heparinase Is Rescued by CaMKII Inhibitor AIP Next, we investigated the excitability of CA1 pyramidal neurons. We previously reported that acute heparinase treatment of hippocampal slices reduced action potential (AP) probability during theta-burst stimulation and hence decreased Ca2+ influx to dendritic spines during the induction of LTP . Based on that study, we expected that one day of heparinase treatment may also result in reduced neuronal excitability in the CA1 pyramidal cells. Therefore, we performed patch clamp recordings in the current clamp configuration and measured the number of APs as a function of injected currents , the threshold of action potential generation , and other parameters characterizing the magnitude and shape of APs . To verify the role of CaMKII in shaping the effects of heparinase, we employed AIP as a selective and potent inhibitor of CaMKII, which has been used in slices and in vivo . We co-injected AIP with heparinase one day before recordings. Compared with the control group, fewer APs were evoked in response to depolarizing currents after injection of heparinase . Analysis of input-output curves showing the average number of APs for each intensity of stimulation revealed a significant reduction in cell excitability in the heparinase-treated neurons and restoration of excitability by AIP . Another indicator of excitability is the spike threshold (Scott et al., 2014). After the heparinase treatment, neurons started to fire at more positive membrane potential in the heparinase group as compared to the control, and this effect was abrogated by AIP . Analysis of two peaks in the second derivative of APs, which correspond to AP generation at the AIS and soma , revealed a tendency toward reduction in the magnitude of the first peak after heparinase treatment (but not the second peak or the interval between peaks), and a significant increase in the first peak after CaMKII inhibition, suggesting the modulation of AIS excitability . An axonal site of heparinase action is also indirectly suggested by the absence of heparinase effects on the peak spike voltage (AP amplitude), which represents an indicator of somatic sodium channel availability . Heparinase also reduced, in a CaMKII-dependent manner, the half-width and decay of the action potentials, suggesting some modulation of potassium channels . 3.4. Increased Activity of CaMKII and Impaired Expression of AnkG at the Axon Initial Segment after Heparinase Treatment Are Abrogated by AIP Our previous study revealed that the removal of HSHSs reduces AnkG expression at the AIS in vitro and in vivo . As we in the present study observed the increased autophosphorylation of CaMKII after heparinase treatment, we investigated if the reduction in AnkG at the AIS correlates with changes in CaMKII phosphorylation at the same subcellular domain and whether the pharmacological inhibition of the CaMKII autophosphorylation with AIP could abrogate the effects of heparinase treatment on AnkG expression. To facilitate the quantitative analysis of protein expression in the AIS, it was performed in vitro as previously described . We observed an increased level of pCaMKII at the AIS after digesting HSHSs, which was reduced by AIP to the control levels . Similar to our previous findings, the removal of HSHSs reduced the expression of AnkG along the 40 mm distance of the AIS relative to the control. In line with our electrophysiological recordings, co-incubating hippocampal neurons with heparinase and AIP restored the expression of AnkG to the control levels . Together with electrophysiological data, these results suggest that the reduction in AnkG expression at the AIS and reduced neuronal excitability after cleaving HSHSs are induced by the increased autophosphorylation of CaMKII. 3.5. Impaired Recall of Contextual Memories after Heparinase Treatment Is Rescued by Co-Administration of AIP In our previous study, we found that heparinase injected before contextual fear conditioning did not affect the level of spontaneous freezing/immobility before conditioning but impaired context discrimination 24 h after conditioning . This experiment, however, did not allow us to dissect whether HSHSs are essential for the acquisition, consolidation, or recall of contextual memories because re-expression of glycans is a slow process taking several weeks , and hence the removal of HSHSs before conditioning would result in impaired HSHS expression during acquisition, consolidation, and recall of memories for the next few days after conditioning. In the present study, we specifically tested if HSHSs are necessary for proper contextual memory recall by injecting heparinase on day 2 after contextual fear conditioning , i.e., after the acquisition and consolidation of memories were successfully completed. This was confirmed by normal freezing time in the conditioned context and normal context discrimination on day 1 in mice pre-assigned to all experimental groups, i.e., control, heparinase, and heparinase plus AIP . Additionally, on day 3 after conditioning, i.e., 24 h after heparinase injection, the freezing time in the conditioned context was normal in the control group, but heparinase-treated mice showed increased freezing in the neutral context CC- and impaired contextual discrimination . Co-administration of AIP restored normal context discrimination after heparinase treatment, not affecting freezing time in the conditioned context. 4. Discussion Our data show that enzymatic removal of HSHSs in the CA1 region of the hippocampus does not affect miniature postsynaptic currents but leads to reduced axonal excitability of pyramidal cells and impaired contextual discrimination, which correlate with increased activity of CaMKII in general, but particularly in the AIS. Inhibition of CaMKII with AIP normalizes excitability and expression of AnkG in the AIS after heparinase treatment, suggesting a causal link between HSPGs and regulation of axonal excitability via CaMKII autophosphorylation. Below, we discuss the functional importance and possible molecular mechanisms underlying these findings. Highly expressed in excitatory synapses in the hippocampus, CaMKII has been studied in many aspects of synaptic function, such as synaptic strength and synaptic plasticity. Overexpression of a and b isoforms of CaMKII in cultured neurons has the opposite effects on mEPSCs' strength frequency while CaMKIIb-overexpressing cells exhibit an increase . Thus, it is plausible to assume that in our experiments, the effects of increased CaMKIIb activation were counterbalanced by increased activity of synaptic CaMKIIa, but we cannot exclude the saturation of CaMKIIb effects under in vivo conditions of the present experiments. The autophosphorylation of CaMKII, especially that of CaMKIIa, on the other hand, has been shown to reduce the excitability of CA1 neurons, which may impact learning , while inhibition of CaMKIIa autophosphorylation by a point mutation at T286A increased CA1 neuron excitability. These data are in line with our finding that autophosphorylation of CaMKII was increased after heparinase treatment in the AIS, while expression of AnkG was impaired but could be rescued by the AIP co-administration. Studies show that AIP specifically inhibits CaMKII relative to other kinases, such as protein kinase C (PKC), CAMKI, and CaMKIV, in rat brain extracts and in mice . The degree of specificity of AIP effects on CaMKIIa versus CaMKIIb, however, has not been properly resolved. The AIS, located between axonal and somatodendritic domains, is a key structure for the initiation of action potential firing. AnkG, neuronal cell adhesion molecule (NrCAM), bIV-spectrin, and voltage-gated sodium and potassium channels are major structural/functional components of the AIS, and their alteration affects AIS assembly and function . bIV-spectrin, an AnkG interaction partner, serves as a bridge between AnkG and the actin-based cytoskeleton. Accordingly, animal models harboring AnkG gene deficiency exhibit abnormal animal behavior (such as ataxia) and neuronal excitability in the cerebellum, due to the mislocalization of sodium channels . Progressive ataxia and tremors are also observed in different bIV-spectrin mutant mice (qv3J and bIV-null mice) . Findings of the mislocalization of sodium channels in the AIS of cerebellar and hippocampal neurons in these mutant mice suggest that altered sodium channel expression is responsible for the neurological phenotypes of the mutants. In the cardiomyocyte, bIV-spectrin's interaction with CaMKII leads to sodium channel phosphorylation via bIV-dependent targeting of CaMKII . The abnormal kinetics of sodium channels and altered cellular excitability after a loss-of-function mutation in the bIV-spectrin gene in the qv3J mouse line suggest that the bIV-spectrin/CaMKII complex is an important component for Na+ channel regulation in cardiomyocytes. Interestingly, CaMKII is colocalized with bIV spectrin in the AIS of cerebellar Purkinje neurons as well, and qv3J mutant mice exhibit a relatively weak immunostaining signal of CaMKII in the AIS of Purkinje cells, implying that the bIV-spectrin/CaMKII complex would strongly affect cellular excitability in both the heart and the brain . Thus, further studies are warranted to study the distribution of bIV-spectrin and ion channels in the AIS after the targeting of HSHSs. Extracellularly, the secreted protein gliomedin is a key component at the nodes of Ranvier in the peripheral nerves. The deposition of gliomedin multimers at the nodal gaps facilitates the clustering of the axonodal cell adhesion molecules neurofascin and NrCAM and sodium channels by binding to HSPGs . In cortical neurons, agrin binds to a tyrosine kinase receptor, which results in the elevation of intracellular Ca2+ and subsequent activation of the CaMKII signaling pathway . Regarding potential protein carriers of HSHSs responsible for the regulation of CaMKII activity at the AIS, there are several candidates. Glypicans 1 and 2 are expressed axonally . Glypican-4 is also enriched on hippocampal granule cell axons and can bind to its partner orphan receptor, GPR158 . Additionally, syndecans are known to be localized at the nodes of Ranvier and axons . Syndecans 2 and 3 can directly bind to CASK protein via the PDZ domain that regulates CaMKII activity in neurons . Further studies on the AIS in mice deficient in these HSPGs could be instrumental to identify their role in AIS assembly and axonal excitability via regulation of CaMKII. Our behavioral experiments for the first time suggest the role of HSHSs in the proper recall of contextual memories and show that in vivo inhibition of CaMKII by AIP could abrogate the hypergeneralization induced by heparinase. Previously, AIP has been shown to significantly protect neurons from NMDA-induced neurotoxicity , fully restore contractility in cardiac muscles of diabetic rats , inhibit doxorubicin-induced apoptosis of cardiac cells , and prevent the reinstatement of morphine-seeking behavior in rats . As hypergeneralization is common for several conditions , targeting this mechanism might be of therapeutic value. A similar loss of context discrimination is found when contextual memories are transferred from the hippocampus to the anterior cingulate cortex via the retrosplenial cortex. Moreover, high-frequency stimulation of memory engrams in the retrosplenial cortex one day after learning produces a recent memory with features normally observed in consolidated remote memories, including contextual generalization and decreased hippocampal dependence . Thus, our data are consistent with the hypothesis that the recent contextual memory is distributed in several brain areas and, if the hippocampal engrams, in particular CA1, are not activated enough due to a loss of excitability induced by heparinase, another, presumably cortical, representation is used. In summary, our data make a stronger link between HSHSs and regulation of neuronal excitability and implicate CaMKII in this regulation. Aberrant expression or activity of HSPGs is associated with some pathological conditions, such as glioblastoma, Fragile X syndrome, neuroinflammation, and Parkinson's disease . Additionally, HSPGs are known to bind and co-aggregate with amyloid beta (Ab) . In light of reported neuronal hyperexcitability in Alzheimer's patients and models of Alzheimer's disease , our work suggests that Ab-HSPG interactions may affect the expression of HSPGs at the AIS, decreasing activation of CaMKII at the AIS and hence increasing neuronal excitability. At synapses, Ab is known to inhibit autophosphorylation of CaMKII at Thr286 and impair synaptic plasticity . Thus, our study suggests potential pathophysiological mechanisms and indicates an option to prevent these by targeting CaMKII signaling at the AIS. Acknowledgments We thank Katrin Boehm for her technical support. Author Contributions Conceptualization, O.S., I.S. and A.D.; Data curation, M.D.; Formal analysis, I.S., T.K., D.B.-A., H.M., O.S., H.H. and A.M.; Funding acquisition, S.J. and A.D.; Investigation, I.S., T.K., D.B.-A., O.S., H.H. and A.M.; Methodology, I.S. and O.S., R.K. and M.D.; Resources, S.J. and A.D.; Supervision, I.S., R.K., M.D., S.J. and A.D.; Visualization, I.S., T.K., M.D., D.B.-A. and O.S.; Writing--original draft, I.S., D.B.-A. and A.D.; Writing--review and editing, all co-authors. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement All treatments and behavioral procedures were conducted in accordance with ethical animal research standards defined by German law and approved by the Ethical Committee on Animal Health and Care of the State of Saxony-Anhalt, Germany, under the license numbers 42502-2-1159 and -1322 DZNE. Informed Consent Statement Not applicable. Data Availability Statement Data supporting reported results can be obtained upon a request to the corresponding author (A.D.). Conflicts of Interest A.D. is Editor-in-Chief of the section Cell Microenvironment in Cells. The other authors declare that they have no competing interest. Figure 1 Heparinase treatment increased CaMKII activity in vivo. (A) Heparinase-injected mouse hippocampi were lysed, and extracts were used for immunoblotting. Membranes were incubated with anti-phospho-Thr286/287 CaMKII antibody to measure autophosphorylated CaMKII a and b levels relative to the total protein expression levels. (B) Summarized graph showing the statistical evaluation of Western blotting experiments, as in (A). Data are presented as means +- SEMs. Note that there was a significant increase in CaMKIIb autophosphorylation one day after heparinase injection (** p < 0.01, Mann-Whitney-Wilcoxon test; Ctrl, n = 9; Hep, n = 9). Figure 2 Intact excitatory and inhibitory synaptic transmission onto CA1 pyramidal cells one day after heparinase (Hep) injection into the mouse hippocampal CA1 in vivo. (A) Representative traces of mEPSCs from CA1 pyramidal cells in heat-inactivated (Ctrl) and active heparinase 1-injected mice in the presence of PTX, CGP55845, and TTX. (B) Summarized bar graphs show that amplitude, frequency, rise time, and decay time of mEPSCs remained intact after heparinase injection (p > 0.05, two-tailed t-test; Ctrl, n = 15; Hep, n = 14). Data are presented as means +- SEMs. (C) Representative traces of mIPSCs recorded from CA1 pyramidal cells in heat-inactivated (Ctrl) and active heparinase 1-injected mice in the presence of NBQX (25 mM), D-AP5 (50 mM), CGP55845 (3 mM), and TTX (1 mM). (D) Summarized bar graphs show that amplitude, frequency, rise time, and decay time of mIPSCs remained intact after heparinase injection into the hippocampal CA1 area (p > 0.05, two-tailed t-test; Ctrl, n = 15; Hep, n = 14). Data are presented as means +- SEMs. Figure 3 CaMKII inhibitor (AIP) could rescue reduced neuronal excitability of CA1 pyramidal neurons after heparinase (Hep) injection in vivo. (A) Sample traces of action potentials (APs) elicited by depolarizing current injections in the CA1 pyramidal cells. (B) Phase plot (dV/dt versus V) of the first action potential in response to 300 pA current injection. Inset (left) shows the initial phase of the action potential from the phase plot (right). (C,D) Summary graphs showing the number of action potentials elicited by different depolarizing current injections and the AP threshold at 20 mV/ms in response to different current injections. * p < 0.05, *** p < 0.001, Holm-Sidak test after two-way RM ANOVA. (E) Analysis of the amplitude and interval (Delay) between two peaks of the second action potential derivative (d2/dt2 AP), which correspond to the activation of sodium channels in the AIS and soma, revealed facilitated AP generation at the AIS after heparinase treatment with co-administration of AIP. * p < 0.05, Holm-Sidak test after one-way ANOVA. (F,G) Summary graphs comparing Ctrl (heat-inactivated heparinase), heparinase, and heparinase+AIP-injected neurons in terms of the amplitude (F), half-width (G), and decay time (G) of the action potential (Ctrl, n = 22; Hep, n = 17; Hep+AIP, n = 10; *** p < 0.05, Holm-Sidak test after one-way ANOVA). Data are presented as means +- SEMs. Figure 4 The CaMKII inhibitor AIP abrogates heparinase (Hep)-induced reduction in AnkG expression in the AIS of hippocampal neurons. (A) Timeline showing the treatment of hippocampal neurons with heparinase and AIP for 2 h followed by immunocytochemistry (ICC). (B) Hippocampal neurons were stained with DAPI (blue), MAP2 (gray), pCaMKII (green), and Ankyrin G (AnkG, red) antibodies. Scale bars for both original and zoomed images are 20 mm. (C,D) The pCaMKII (C) and AnkG (D) distributions at the AIS were determined by computing a 40 mm long fluorescence intensity profile with a thickness of 3 pixels in the middle of AnkG-immunopositive areas starting from the edge of the soma (determined using the MAP2 signal). The average expression of AnkG and pCaMKII along the 40 mm line profile of the AIS was quantified from 165 (Ctrl), 183 (Hep), and 167 (Hep+AIP) neurons from three independent experiments. Line profiles were normalized to the peak AnkG and pCaMKII values from the control group for each independent experiment. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, Holm-Sidak post hoc test after two-way RM ANOVA for the line profiles binned with 10 mm interval. Figure 5 CaMKII inhibitor AIP rescues impaired context discrimination after heparinase (Hep) injection in vivo. (A) Scheme of the experiment. (B,C) Intrahippocampal heparinase injection 24 h after fear conditioning increases freezing in the neutral (CC-) context to the level measured in the conditioned (CC+) context (B) and impairs context discrimination (C), which can be rescued by inhibition of CaMKII with AIP. # p < 0.05, paired t-test to compare freezing in conditioned and neutral contexts. * p < 0.05, Holm-Sidak post hoc test after two-way RM ANOVA to compare treatments. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Ford-Perriss M. Turner K. Guimond S. Apedaile A. Haubeck H.D. Turnbull J. Murphy M. Localisation of specific heparan sulfate proteoglycans during the proliferative phase of brain development Dev. Dyn. 2003 227 170 184 10.1002/dvdy.10298 12761845 2. Farhy-Tselnicker I. van Casteren A.C.M. Lee A. Chang V.T. Aricescu A.R. Allen N.J. Astrocyte-Secreted Glypican 4 Regulates Release of Neuronal Pentraxin 1 from Axons to Induce Functional Synapse Formation Neuron 2017 96 428 445.e413 10.1016/j.neuron.2017.09.053 29024665 3. Hsueh Y.P. Sheng M. Regulated expression and subcellular localization of syndecan heparan sulfate proteoglycans and the syndecan-binding protein CASK/LIN-2 during rat brain development J. Neurosci. 1999 19 7415 7425 10.1523/JNEUROSCI.19-17-07415.1999 10460248 4. Hsueh Y.P. Yang F.C. Kharazia V. Naisbitt S. Cohen A.R. Weinberg R.J. Sheng M. Direct interaction of CASK/LIN-2 and syndecan heparan sulfate proteoglycan and their overlapping distribution in neuronal synapses J. Cell Biol. 1998 142 139 151 10.1083/jcb.142.1.139 9660869 5. Grootjans J.J. Zimmermann P. Reekmans G. Smets A. Degeest G. Durr J. David G. Syntenin, a PDZ protein that binds syndecan cytoplasmic domains Proc. Natl. Acad. Sci. USA 1997 94 13683 13688 10.1073/pnas.94.25.13683 9391086 6. Gao Y. Li M. Chen W. Simons M. Synectin, syndecan-4 cytoplasmic domain binding PDZ protein, inhibits cell migration J. Cell Physiol. 2000 184 373 379 10.1002/1097-4652(200009)184:3<373::AID-JCP12>3.0.CO;2-I 10911369 7. Ethell I.M. Hagihara K. Miura Y. Irie F. Yamaguchi Y. Synbindin, A novel syndecan-2-binding protein in neuronal dendritic spines J. Cell Biol. 2000 151 53 68 10.1083/jcb.151.1.53 11018053 8. Ethell I.M. Yamaguchi Y. Cell surface heparan sulfate proteoglycan syndecan-2 induces the maturation of dendritic spines in rat hippocampal neurons J. Cell Biol. 1999 144 575 586 10.1083/jcb.144.3.575 9971750 9. Zimmermann P. Zhang Z. Degeest G. Mortier E. Leenaerts I. Coomans C. Schulz J. N'Kuli F. Courtoy P.J. David G. Syndecan recycling [corrected] is controlled by syntenin-PIP2 interaction and Arf6 Dev. Cell 2005 9 377 388 10.1016/j.devcel.2005.07.011 16139226 10. Lauri S.E. Kaukinen S. Kinnunen T. Ylinen A. Imai S. Kaila K. Taira T. Rauvala H. Regulatory role and molecular interactions of a cell-surface heparan sulfate proteoglycan (N-syndecan) in hippocampal long-term potentiation J. Neurosci. 1999 19 1226 1235 10.1523/JNEUROSCI.19-04-01226.1999 9952400 11. Kaksonen M. Pavlov I. Voikar V. Lauri S.E. Hienola A. Riekki R. Lakso M. Taira T. Rauvala H. Syndecan-3-deficient mice exhibit enhanced LTP and impaired hippocampus-dependent memory Mol. Cell. Neurosci. 2002 21 158 172 10.1006/mcne.2002.1167 12359158 12. McCroskery S. Bailey A. Lin L. Daniels M.P. Transmembrane agrin regulates dendritic filopodia and synapse formation in mature hippocampal neuron cultures Neuroscience 2009 163 168 179 10.1016/j.neuroscience.2009.06.012 19524020 13. Coles C.H. Shen Y. Tenney A.P. Siebold C. Sutton G.C. Lu W. Gallagher J.T. Jones E.Y. Flanagan J.G. Aricescu A.R. Proteoglycan-specific molecular switch for RPTPsigma clustering and neuronal extension Science 2011 332 484 488 10.1126/science.1200840 21454754 14. Dityatev A. Dityateva G. Sytnyk V. Delling M. Toni N. Nikonenko I. Muller D. Schachner M. Polysialylated neural cell adhesion molecule promotes remodeling and formation of hippocampal synapses J. Neurosci. 2004 24 9372 9382 10.1523/JNEUROSCI.1702-04.2004 15496673 15. Irie F. Badie-Mahdavi H. Yamaguchi Y. Autism-like socio-communicative deficits and stereotypies in mice lacking heparan sulfate Proc. Natl. Acad. Sci. USA 2012 109 5052 5056 10.1073/pnas.1117881109 22411800 16. Zhang P. Lu H. Peixoto R.T. Pines M.K. Ge Y. Oku S. Siddiqui T.J. Xie Y. Wu W. Archer-Hartmann S. Heparan Sulfate Organizes Neuronal Synapses through Neurexin Partnerships Cell 2018 174 1450 1464 10.1016/j.cell.2018.07.002 30100184 17. Korotchenko S. Cingolani L.A. Kuznetsova T. Bologna L.L. Chiappalone M. Dityatev A. Modulation of network activity and induction of homeostatic synaptic plasticity by enzymatic removal of heparan sulfates Philos. Trans. R. Soc. London Ser. B Biol. Sci. 2014 369 20140134 10.1098/rstb.2014.0134 25225107 18. Garau G. Magotti P. Heine M. Korotchenko S. Lievens P.M. Berezin V. Dityatev A. Heparin/heparan sulfates bind to and modulate neuronal L-type (Cav1.2) voltage-dependent Ca(2+) channels Exp. Neurol. 2015 274 156 165 10.1016/j.expneurol.2015.08.006 26272754 19. Minge D. Senkov O. Kaushik R. Herde M.K. Tikhobrazova O. Wulff A.B. Mironov A. van Kuppevelt T.H. Oosterhof A. Kochlamazashvili G. Heparan Sulfates Support Pyramidal Cell Excitability, Synaptic Plasticity, and Context Discrimination Cereb. Cortex 2017 27 903 918 10.1093/cercor/bhx003 28119345 20. Baucum A.J. 2nd Shonesy B.C. Rose K.L. Colbran R.J. Quantitative proteomics analysis of CaMKII phosphorylation and the CaMKII interactome in the mouse forebrain ACS Chem. Neurosci. 2015 6 615 631 10.1021/cn500337u 25650780 21. De Carvalho Myskiw J. Furini C.R. Benetti F. Izquierdo I. Hippocampal molecular mechanisms involved in the enhancement of fear extinction caused by exposure to novelty Proc. Natl. Acad. Sci. USA 2014 111 4572 4577 10.1073/pnas.1400423111 24591622 22. Johnson C.N. Pattanayek R. Potet F. Rebbeck R.T. Blackwell D.J. Nikolaienko R. Sequeira V. Le Meur R. Radwanski P.B. Davis J.P. The CaMKII inhibitor KN93-calmodulin interaction and implications for calmodulin tuning of NaV1.5 and RyR2 function Cell Calcium 2019 82 102063 10.1016/j.ceca.2019.102063 31401388 23. Murakoshi H. Shin M.E. Parra-Bueno P. Szatmari E.M. Shibata A.C.E. Yasuda R. Kinetics of Endogenous CaMKII Required for Synaptic Plasticity Revealed by Optogenetic Kinase Inhibitor Neuron 2017 94 37 47 10.1016/j.neuron.2017.02.036 28318784 24. Meeks J.P. Mennerick S. Action potential initiation and propagation in CA3 pyramidal axons J. Neurophysiol. 2007 97 3460 3472 10.1152/jn.01288.2006 17314237 25. Romberg C. Yang S. Melani R. Andrews M.R. Horner A.E. Spillantini M.G. Bussey T.J. Fawcett J.W. Pizzorusso T. Saksida L.M. Depletion of perineuronal nets enhances recognition memory and long-term depression in the perirhinal cortex J. Neurosci. 2013 33 7057 7065 10.1523/JNEUROSCI.6267-11.2013 23595763 26. Thiagarajan T.C. Piedras-Renteria E.S. Tsien R.W. betaCaMKII. Inverse regulation by neuronal activity and opposing effects on synaptic strength Neuron 2002 36 1103 1114 10.1016/S0896-6273(02)01049-8 12495625 27. Sametsky E.A. Disterhoft J.F. Ohno M. Autophosphorylation of alphaCaMKII downregulates excitability of CA1 pyramidal neurons following synaptic stimulation Neurobiol. Learn. Mem. 2009 92 120 123 10.1016/j.nlm.2009.02.006 19245842 28. Giese K.P. Fedorov N.B. Filipkowski R.K. Silva A.J. Autophosphorylation at Thr286 of the alpha calcium-calmodulin kinase II in LTP and learning Science 1998 279 870 873 10.1126/science.279.5352.870 9452388 29. Pellicena P. Schulman H. CaMKII inhibitors: From research tools to therapeutic agents Front. Pharmacol. 2014 5 21 10.3389/fphar.2014.00021 24600394 30. Ishida A. Kameshita I. Okuno S. Kitani T. Fujisawa H. A novel highly specific and potent inhibitor of calmodulin-dependent protein kinase II Biochem. Biophys. Res. Commun. 1995 212 806 812 10.1006/bbrc.1995.2040 7626114 31. Huang C.Y. Rasband M.N. Axon initial segments: Structure, function, and disease Ann. N. Y. Acad. Sci. 2018 1420 46 61 10.1111/nyas.13718 29749636 32. Zhou D. Lambert S. Malen P.L. Carpenter S. Boland L.M. Bennett V. AnkyrinG is required for clustering of voltage-gated Na channels at axon initial segments and for normal action potential firing J. Cell Biol. 1998 143 1295 1304 10.1083/jcb.143.5.1295 9832557 33. Komada M. Soriano P. [Beta]IV-spectrin regulates sodium channel clustering through ankyrin-G at axon initial segments and nodes of Ranvier J. Cell Biol. 2002 156 337 348 10.1083/jcb.200110003 11807096 34. Hund T.J. Koval O.M. Li J. Wright P.J. Qian L. Snyder J.S. Gudmundsson H. Kline C.F. Davidson N.P. Cardona N. A beta(IV)-spectrin/CaMKII signaling complex is essential for membrane excitability in mice J. Clin. Investig. 2010 120 3508 3519 10.1172/JCI43621 20877009 35. Eshed Y. Feinberg K. Carey D.J. Peles E. Secreted gliomedin is a perinodal matrix component of peripheral nerves J. Cell Biol. 2007 177 551 562 10.1083/jcb.200612139 17485493 36. Hilgenberg L.G. Smith M.A. Agrin signaling in cortical neurons is mediated by a tyrosine kinase-dependent increase in intracellular Ca2+ that engages both CaMKII and MAPK signal pathways J. Neurobiol. 2004 61 289 300 10.1002/neu.20049 15389602 37. Ivins J.K. Litwack E.D. Kumbasar A. Stipp C.S. Lander A.D. Cerebroglycan, a developmentally regulated cell-surface heparan sulfate proteoglycan, is expressed on developing axons and growth cones Dev. Biol. 1997 184 320 332 10.1006/dbio.1997.8532 9133438 38. Litwack E.D. Ivins J.K. Kumbasar A. Paine-Saunders S. Stipp C.S. Lander A.D. Expression of the heparan sulfate proteoglycan glypican-1 in the developing rodent Dev. Dyn. 1998 211 72 87 10.1002/(SICI)1097-0177(199801)211:1<72::AID-AJA7>3.0.CO;2-4 9438425 39. Condomitti G. Wierda K.D. Schroeder A. Rubio S.E. Vennekens K.M. Orlandi C. Martemyanov K.A. Gounko N.V. Savas J.N. de Wit J. An Input-Specific Orphan Receptor GPR158-HSPG Interaction Organizes Hippocampal Mossy Fiber-CA3 Synapses Neuron 2018 100 201 215 10.1016/j.neuron.2018.08.038 30290982 40. Goutebroze L. Carnaud M. Denisenko N. Boutterin M.C. Girault J.A. Syndecan-3 and syndecan-4 are enriched in Schwann cell perinodal processes BMC Neurosci. 2003 4 29 10.1186/1471-2202-4-29 14622446 41. Luo N. Li H. Xiang B. Qiao L. He J. Ji Y. Liu Y. Li S. Lu R. Li Y. Syndecan-4 modulates the proliferation of neural cells and the formation of CaP axons during zebrafish embryonic neurogenesis Sci. Rep. 2016 6 25300 10.1038/srep25300 27143125 42. Lu C.S. Hodge J.J. Mehren J. Sun X.X. Griffith L.C. Regulation of the Ca2+/CaM-responsive pool of CaMKII by scaffold-dependent autophosphorylation Neuron 2003 40 1185 1197 10.1016/S0896-6273(03)00786-4 14687552 43. Laabich A. Cooper N.G. Neuroprotective effect of AIP on N-methyl-D-aspartate-induced cell death in retinal neurons Brain Res. Mol. Brain Res. 2000 85 32 40 10.1016/S0169-328X(00)00226-6 11146104 44. Daniels L.J. Wallace R.S. Nicholson O.M. Wilson G.A. McDonald F.J. Jones P.P. Baldi J.C. Lamberts R.R. Erickson J.R. Inhibition of calcium/calmodulin-dependent kinase II restores contraction and relaxation in isolated cardiac muscle from type 2 diabetic rats Cardiovasc. Diabetol. 2018 17 89 10.1186/s12933-018-0732-x 29903013 45. Ikeda S. Matsushima S. Okabe K. Ikeda M. Ishikita A. Tadokoro T. Enzan N. Yamamoto T. Sada M. Deguchi H. Blockade of L-type Ca(2+) channel attenuates doxorubicin-induced cardiomyopathy via suppression of CaMKII-NF-kappaB pathway Sci. Rep. 2019 9 9850 10.1038/s41598-019-46367-6 31285514 46. Liu Z. Zhang J.J. Liu X.D. Yu L.C. Inhibition of CaMKII activity in the nucleus accumbens shell blocks the reinstatement of morphine-seeking behavior in rats Neurosci. Lett. 2012 518 167 171 10.1016/j.neulet.2012.05.003 22579819 47. Duits P. Cath D.C. Lissek S. Hox J.J. Hamm A.O. Engelhard I.M. van den Hout M.A. Baas J.M. Updated meta-analysis of classical fear conditioning in the anxiety disorders Depress. Anxiety 2015 32 239 253 10.1002/da.22353 25703487 48. Vlaeyen J.W.S. Linton S.J. Fear-avoidance model of chronic musculoskeletal pain: 12 years on Pain 2012 153 1144 1147 10.1016/j.pain.2011.12.009 22321917 49. Dymond S. Dunsmoor J.E. Vervliet B. Roche B. Hermans D. Fear Generalization in Humans: Systematic Review and Implications for Anxiety Disorder Research Behav. Ther. 2015 46 561 582 10.1016/j.beth.2014.10.001 26459838 50. De Sousa A.F. Cowansage K.K. Zutshi I. Cardozo L.M. Yoo E.J. Leutgeb S. Mayford M. Optogenetic reactivation of memory ensembles in the retrosplenial cortex induces systems consolidation Proc. Natl. Acad. Sci. USA 2019 116 8576 8581 10.1073/pnas.1818432116 30877252 51. Kazanskaya G.M. Tsidulko A.Y. Volkov A.M. Kiselev R.S. Suhovskih A.V. Kobozev V.V. Gaytan A.S. Aidagulova S.V. Krivoshapkin A.L. Grigorieva E.V. Heparan sulfate accumulation and perlecan/HSPG2 up-regulation in tumour tissue predict low relapse-free survival for patients with glioblastoma Histochem. Cell Biol. 2018 149 235 244 10.1007/s00418-018-1631-7 29322326 52. Sears J.C. Broadie K. Fragile X Mental Retardation Protein Regulates Activity-Dependent Membrane Trafficking and Trans-Synaptic Signaling Mediating Synaptic Remodeling Front. Mol. Neurosci. 2017 10 440 10.3389/fnmol.2017.00440 29375303 53. Bussini S. Meda L. Scarpini E. Clementi E. Conti G. Tiriticco M. Bresolin N. Baron P. Heparan sulfate proteoglycan induces the production of NO and TNF-alpha by murine microglia Immun. Ageing 2005 2 11 10.1186/1742-4933-2-11 16022734 54. Chen M. Vincent J. Ezeanii A. Wakade S. Yerigenahally S. Mor D.E. Heparan sulfate proteoglycans mediate prion-like alpha-synuclein toxicity in Parkinson's in vivo models Life Sci. Alliance 2022 5 e202201366 10.26508/lsa.202201366 35790300 55. Liu C.C. Zhao N. Yamaguchi Y. Cirrito J.R. Kanekiyo T. Holtzman D.M. Bu G. Neuronal heparan sulfates promote amyloid pathology by modulating brain amyloid-beta clearance and aggregation in Alzheimer's disease Sci. Transl. Med. 2016 8 332ra344 10.1126/scitranslmed.aad3650 56. Van Horssen J. Wesseling P. van den Heuvel L.P. de Waal R.M. Verbeek M.M. Heparan sulphate proteoglycans in Alzheimer's disease and amyloid-related disorders Lancet Neurol. 2003 2 482 492 10.1016/S1474-4422(03)00484-8 12878436 57. Kazim S.F. Seo J.H. Bianchi R. Larson C.S. Sharma A. Wong R.K.S. Gorbachev K.Y. Pereira A.C. Neuronal Network Excitability in Alzheimer's Disease: The Puzzle of Similar versus Divergent Roles of Amyloid beta and Tau eNeuro 2021 8 ENEURO.0418-20.2020 10.1523/ENEURO.0418-20.2020 58. Lam A.D. Deck G. Goldman A. Eskandar E.N. Noebels J. Cole A.J. Silent hippocampal seizures and spikes identified by foramen ovale electrodes in Alzheimer's disease Nat. Med. 2017 23 678 680 10.1038/nm.4330 28459436 59. Zhao D. Watson J.B. Xie C.W. Amyloid beta prevents activation of calcium/calmodulin-dependent protein kinase II and AMPA receptor phosphorylation during hippocampal long-term potentiation J. Neurophysiol. 2004 92 2853 2858 10.1152/jn.00485.2004 15212428 |
PMC10000603 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051100 foods-12-01100 Article Decreasing the Crystallinity and Degree of Polymerization of Cellulose Increases Its Susceptibility to Enzymatic Hydrolysis and Fermentation by Colon Microbiota Thielemans Karel Conceptualization Methodology Investigation Writing - original draft 12 De Bondt Yamina Writing - review & editing 1 Comer Luke Formal analysis Writing - review & editing 3 Raes Jeroen Methodology Writing - review & editing 4 Everaert Nadia Methodology Writing - review & editing 3 Sels Bert F. Conceptualization Methodology Resources Funding acquisition 2 Courtin Christophe M. Conceptualization Methodology Resources Writing - review & editing Supervision Funding acquisition 1* Reis Sofia F. Academic Editor da Silva Jose A. Lopes Academic Editor Coelho Elisabete Academic Editor 1 Laboratory of Food Chemistry and Biochemistry, Department of Microbial and Molecular Systems (M2S), KU Leuven, Kasteelpark Arenberg 20, 3001 Leuven, Belgium 2 Center for Sustainable Catalysis and Engineering, Department of Microbial and Molecular Systems (M2S), KU Leuven, Celestijnenlaan 200F, 3001 Leuven, Belgium 3 Divison Animal and Human Health Engineering, Department of Biosystems, KU Leuven, Kasteelpark Arenberg 30, 3001 Leuven, Belgium 4 Department of Microbiology and Immunology, KU Leuven, Herestraat 49-Bus 1028, 3000 Leuven, Belgium * Correspondence: [email protected] 04 3 2023 3 2023 12 5 110023 1 2023 22 2 2023 02 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Cellulose can be isolated from various raw materials and agricultural side streams and might help to reduce the dietary fiber gap in our diets. However, the physiological benefits of cellulose upon ingestion are limited beyond providing fecal bulk. It is barely fermented by the microbiota in the human colon due to its crystalline character and high degree of polymerization. These properties make cellulose inaccessible to microbial cellulolytic enzymes in the colon. In this study, amorphized and depolymerized cellulose samples with an average degree of polymerization of less than 100 anhydroglucose units and a crystallinity index below 30% were made from microcrystalline cellulose using mechanical treatment and acid hydrolysis. This amorphized and depolymerized cellulose showed enhanced digestibility by a cellulase enzyme blend. Furthermore, the samples were fermented more extensively in batch fermentations using pooled human fecal microbiota, with minimal fermentation degrees up to 45% and a more than eight-fold increase in short-chain fatty acid production. While this enhanced fermentation turned out to be highly dependent on the microbial composition of the fecal pool, the potential of engineering cellulose properties to increased physiological benefit was demonstrated. cellulose depolymerization amorphization dietary fiber prebiotic colonic fermentation short chain fatty acids KU Leuven internal grantsFWO-SBO project BioWoodS003518N Internal Funds KU Leuven for PDM fundingPDMT1/21/022 Research Foundation-Flanders (FWO-Vlaanderen, Brussels, Belgium)12B3723N VIBKU LeuvenRega instituteThis work was supported by KU Leuven internal grants, and the FWO-SBO project BioWood (Grant number: S003518N). Yamina De Bondt acknowledges Internal Funds KU Leuven for PDM funding (PDMT1/21/022) and the Research Foundation-Flanders (FWO-Vlaanderen, Brussels, Belgium) for a position as postdoctoral research fellow (Grant number: 12B3723N). The Raes lab is funded by VIB, KU Leuven and the Rega institute. pmc1. Introduction Cellulose is the most abundant renewable material in nature, being the primary building block of the plant cell wall. It consists of long unbranched b-1,4-bound glucose polymers organized in long crystalline fibers with strong interactions between the different polymers . The high degree of crystallinity, high degree of polymerization and low specific surface give cellulose a very recalcitrant character. This recalcitrance is important in the plant cell wall, since cellulose provides the plant with mechanical strength and resilience against breakdown, but is a major drawback in valorization strategies of (ligno)cellulose, e.g., in the context of biorefineries . Furthermore, cellulose acts as an insoluble and recalcitrant dietary fiber in the human body when ingested. The use of cellulose as dietary fiber in foods could be very relevant since the food industry increasingly searches for dietary fiber enrichment strategies. While a sufficient daily intake of dietary fiber is correlated with various health benefits, such as a decreased risk of colorectal cancer, obesity, cardiovascular diseases and diabetes mellitus type II , the average daily intake of dietary fiber is too low in Western diets . Within dietary fiber fortification strategies, specific attention goes to (partially) fermentable dietary fiber. Fermentation of dietary fiber in the colon is correlated with different additional physiological benefits, linked to the production of short-chain fatty acids (SCFA), which are essential for colonic health h, glucose and cholesterol homeostasis and the regulation of the appetite . The fermentability of cellulose in the human colon is very low, however . Cellulolytic enzymes are produced by Ruminococcus, Enterococcus, Bacteroides or Prevotella species in the colon , but the highly ordered nature of cellulose limits the accessibility of the cellulosic fibers and the glucosidic b-1,4 bonds for enzymatic breakdown and results in very limited fermentability. We can assume that the fermentability and the physiological benefits of cellulose with it could be improved by breaking this recalcitrance before ingestion. Such accessible cellulose would remain insoluble and indigestible but could be fermented to a greater extent in the colon. Previous research already stated that the fermentability of cellulose depends on its physical appearance , and some attempts to improve cellulose fermentability by reducing the particle size were already successful in human in vitro experiments . However, the impact of the degree of polymerization and crystallinity has not been investigated in this context. At the same time, these structural parameters are known to affect cellulose accessibility greatly . Plenty of (ligno)cellulosic biomass pretreatment protocols, such as milling, irradiation, ultrasonication, hydrothermal treatment or solubilization in ionic liquids, have been developed and optimized to alter these cellulose characteristics . Moreover, several of them, such as ball milling, acid hydrolysis or solubilization in ionic liquid, are linked to an improved cellulose enzymatic accessibility as well . In this study, two effective pretreatment methods, ball milling and acid hydrolysis, are combined to make cellulose with a lowered degree of polymerization, a lowered degree of crystallinity and the combination of both. These samples are used to gain insight into the effect of these parameters on the enzyme accessibility of cellulose and its fermentability by colon microbiota, using batch in vitro fermentations. 2. Materials and Methods 2.1. Materials Microcrystalline cellulose (Avicel PH-101, 3.4% moisture), citric acid (analytical grade), the Cellic CTec2 cellulase enzyme blend and all other analytical chemicals and solvents were purchased from Sigma-Aldrich (Deurne, Belgium). 2.2. Production of Dietary Fiber Samples Starting from Microcrystalline Cellulose An overview of the production of the different dietary fiber samples is given in Figure 1 and Table A1. Microcrystalline cellulose (MC) was first depolymerized using a ball milling step and acid hydrolysis, similar to our previous work . MC was pretreated in a planetary ball mill (PM100, Retsch GmbH, Haan, Germany) in batches of 20 g with 6 zirconium oxide balls (O 10 mm) to induce para-crystalline zones in the cellulose fiber. These ball-milled cellulose fibers are called amorphized cellulose (AC). Milling time (60-360 min) and speed (400-500 rpm) were varied. Afterwards, the paracrystalline zones in the AC fibers were hydrolyzed with a 10% citric acid solution in water. Hydrolysis time (2-6 h) and temperature (90-130 degC) were varied (Table A1). The depolymerized insoluble cellulose samples were washed until neutral pH and dried for 45 h at 60 degC, yielding depolymerized cellulose (DC). After being dried, the DC sample was again treated in the planetary ball mill with 6 zirconium oxide balls (O 10 mm) at 500 rpm to produce amorphized depolymerized cellulose (ADC). Treatment times of 30, 60 and 360 min were used. All samples and respective production parameters are summarized in Table A1. 2.3. Characterization of Dietary Fiber Samples The average degree of polymerization (avDP) of cellulose was determined viscometrically in triplicate, based on the method of the French Institute for Normalisation . Fiber samples (0.075 g) were dissolved in a 0.5 M bis(ethylenediamine)copper(II)hydroxide solution (15 mL), and the viscosity of this solution at 25 degC was measured with a capillary viscometer, type nr. 509 04 (Schott Gerate, Jena, Germany). The avDP was calculated from the boundary viscosity of the solution (e), based on the empirical relation: Average DP^a = e/K, with a and K empirical constants, equal to 1 and 7.5 x 10-3, respectively. The boundary viscosity e was determined from ea = e.C.10^((0,14.e.C)), with ea the specific viscosity of the solution, and C the cellulose concentration (g/mL). The crystallinity of the fiber samples was determined by X-ray powder diffraction (XRD) on a high-throughput STOE STADI P Combi diffractometer (STOE & Cie GmbH, Darmstadt, Germany) in transition mode with Ge(111) monochromatic X-ray inlet beams (l = 1.5406 A, Cu Ka source). Crystallinity indexes were determined by the peak-height method of Segal and coworkers . 2.4. Enzymatic Digestibility Analysis The enzymatic digestibility of the dietary fiber samples was determined by calculating the enzymatic conversion (EC) after incubating samples with the Cellic CTec2 cellulase enzyme blend, as described by Chen and coworkers . Cellulose was suspended (1.0% w/v) in a 50 mM sodium acetate buffer (pH 4.8) with 20 U Cellic CTec2 cellulase enzyme blend per gram cellulose and stirred at 900 rpm. After 1 h of incubation at 40 degC, the enzymes were denatured by heating the solution (5 min, 110 degC). The solid fraction was separated from the supernatant by centrifuging at 5000 g. The amount of glucose and cellobiose in the supernatant from cellulose hydrolysis was determined by high-performance-anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD) on a Dionex ICS3000 chromatography system (Sunnyvale, CA, USA). Saccharides were separated on a Dionex CarboPac PA-100 column (4 x 250 mm), equilibrated with 90 mM NaOH. The enzymatic conversion was calculated from the amount of glucose (mg) and cellobiose (mcb) in the supernatant, and the amount of starting substrate (mc):(1) EC=mg+mcb1.1mc 2.5. In Vitro Fermentation of Dietary Fiber Samples Using Human Fecal Inoculum In vitro fermentation experiments (trial 1, 2 and 3) were performed as described by De Preter et al. . Fresh fecal samples of 8 healthy donors (consuming a mixed western diet, no history of antibiotic use in the last six months) were collected and pooled to make a 10 w/v% fecal slurry in phosphate-buffered saline. After intensive shaking, this fecal slurry was decanted, and the supernatant (referred to as the inoculum) was added to different fiber samples (25 mL to 100 mg cellulose) in triplicate. After being flushed with nitrogen gas, the tubes were incubated anaerobically for 48 h in a shaking water bath at 37 degC. At the end of incubation, the pH of the slurry was measured with a digital pH meter (Hanna Instruments HI 9025, Temse, Belgium). Aliquots were stored at -20 degC for the determination of short-chain fatty acid (SCFA) concentration and microbial analysis. 2.6. Short-Chain Fatty Acid Analysis The amounts of acetate, propionate and butyrate in the fecal inoculum were determined according to the gas-chromatographic method described by Bautil et al. . In this procedure, a 25% (w/v) NaOH solution was added to the inoculum to create sodium salts of the SCFA, which were neutralized by adding a 50% sulfuric acid solution afterwards. These salts were extracted to a diethyl ether phase, which was analyzed with an Agilent 6890 Series gas chromatograph with an EC-1000 Econo-Cap column (25 m x 0.53 mm, 130 degC, 1.2 mm film thickness) and helium (20 mL/min) as carrier gas. A flame ionization detector at 195 degC measured the different fatty acids. Within this analysis, 2-ethyl butyric acid was used as an internal standard. 2.7. Microbial Analysis Microbial profiling was done as described by Falony et al. . Nucleic acids were extracted from the aliquots using the RNeasy PowerMicrobiome kit (Qiagen, Venlo, The Netherlands). The manufacturer's protocol was modified by adding a heating step at 90degC for 10 min and excluding DNA removal steps. Afterwards, the extracted DNA was amplified in triplicate using 16S primers 515F (59-GTGYCAGCMGCCGCGGTAA-39) and 806R (59-GGACTACNVGGGTWTCTAAT-39) targeting the V4 region. Deep sequencing was performed on a MiSeq platform (2-by-250 paired-end [PE] reads; Illumina, San Diego, CA, USA). Initial quality assessment, sequence filtering and trimming of the FASTQ files were carried out using the FASTQC software (version 0.11.9) and the 'filterAndTrim' function of the DADA2 algorithm pipeline package. Analysis thereafter was performed using the 'mergePairs' function of the DADA2 package, which merges the forward and reverse sequences. Any chimeric sequences which were produced during aberrant PCR annealing were identified and removed. Taxonomy was assigned to the sequences using a naive Bayesian classifier method with the SILVA database (version 138.1) as a reference. 2.8. Statistics Significant differences were detected by performing a one-way analysis of variance (ANOVA) using JMP Pro 16 (SAS institute), with a comparison of the mean values using the Tukey test (a < 0.05). 3. Results and Discussion 3.1. Production of Samples with Different DP and Crystallinity from Microcrystalline Cellulose To investigate the impact of crystallinity and avDP on the enzymatic accessibility and fermentability by colon microbiota, a modification protocol using the combination of planetary ball mill treatments and acid hydrolysis was used . First, MC was treated in a ball mill to decrease the crystallinity of the cellulose by incorporation of paracrystalline zones. This decrease in crystallinity impacts the levelling-off degree of polymerization (LODP) of the cellulose, which represents the length of crystalline polymers that remain insoluble after a fast hydrolysis of the easily accessible paracrystalline zones . Second, the ball-milled, or amorphized cellulose (AC) was hydrolyzed with citric acid at elevated temperature (90-130 degC) to hydrolyse the polymers in the paracrystalline zones. After this hydrolysis, depolymerized cellulose (DC) with a decreased avDP is obtained. At last, this DC was treated in the planetary ball mill another time for 30-360 min to produce amorphized depolymerized cellulose (ADC), which is expected to be highly accessible. All samples are listed in Table A1. The influence of processing conditions (ball mill speed/time and acid hydrolysis time/temperature) on cellulose properties was extensively investigated in our previous study for the production of DC . The avDP of the DC fibers can be finely tuned, and also the crystallinity of the DC fibers can be controlled by applying varying process parameters. In short, when the acid hydrolysis is not performed for long enough to hydrolyse all the paracrystalline zones of the AC fibers, the LODP will not be reached and the crystallinity of the resulting DC fibers will remain low . Despite the extensive investigation of the impact of process parameters to produce DC, the impact of the second ball mill treatment on this DC was not yet investigated. For a sample with relatively high crystallinity (DC with avDP of 32 AGU and crystallinity index of 0.62), the effect of this additional ball mill treatment on avDP and crystallinity is shown in Figure 2. Figure 2a shows that the peaks from crystalline planes in the refractogram of the DC fibers indeed disappeared due to the ball mill treatment. Milling a DC for only 15 min already disrupted most of the crystalline structure, but the crystalline reflection at 2th of 22deg was still more prominent in these refractograms than in those of ADC fibers with longer milling times. After 30 min of milling of the DC at 500 rpm, an amorphous refractogram was detected, of which the shape did not change anymore upon longer milling times. Previously, it was shown that the crystallinity decrease during ball milling of unmodified MC was limited during the first 30 min of the milling process . The breakdown of crystallites, therefore, occurred more slowly for unmodified MC than for this DC . This faster decrease in crystallinity for the DC might be due to the different type of crystallites that need to be broken down. As visualized in Figure A1, 32% of the crystallites in DC fibers were cellulose II polymorphs, while it is known that no cellulose II is present in unmodified MC . We can hypothesize that cellulose II crystallites, formed during the first ball mill treatment and hydrolysis , are easier to decrystallize than cellulose I crystallites. The faster decrystallization of DC can also be caused by the lower avDP of the DC fibers. Depolymerization of the DC fibers does not seem to occur during the ball mill treatment since no significant decrease in avDP was detected for the different ADC fibers . Previous research stated that a ball mill treatment could not depolymerise cellulose shorter than 50 AGU . This theory seems to be confirmed here since no depolymerization of the DC fibers (DP 32) occurred. 3.2. Influence of Cellulose Structural Properties on Enzymatic Accessibility Figure 3 shows the enzymatic conversion of the modified cellulose into glucose or cellobiose after 1 h reaction with the commercial Cellic CTec2 enzyme blend under optimal conditions. Unmodified MC is compared with amorphized MC (AC124), DC and ADC with different avDP (Table A1). Only 30% of the long crystalline MC was converted into glucose and cellobiose by the cellulase blend within one hour (conversion degree of 0.30 +- 0.04). Decreasing the crystallinity by ball milling (260 min) improved the conversion degree slightly to 0.35 +- 0.01, but decreasing the avDP had the opposite effect. Surprisingly, the DC was all less accessible for the enzyme blend than unmodified MC or AC, while Kumar and Wyman showed that a shorter DP results in higher accessibility . We can hypothesize that a decrease in avDP from 168 to 28 AGU is not sufficient to compensate for the removal of para-crystalline zones and the presence of cellulose polymorph II in the DC fibers, two structural properties that lower enzymatic accessibility. This hypothesis can be confirmed by the positive association between avDP and enzymatic digestibility of the different DC samples. These various DC samples also slightly differed in crystallinity: the crystallinity of DC104 was lower than the crystallinity of DC28, since the LODP was not reached for the longer DC fibers (Table A1). This is because the mildest hydrolysis conditions were used for making DC104, resulting in the remaining of some easily accessible paracrystalline zones after drying. It seems that these small differences in crystallinity have a more significant impact on enzymatic conversion than the differences in avDP. Since the DC samples showed lower enzymatic digestibility for the cellulase blend than MC or AC, it can be concluded that a DP decrease to values lower than 100 AGU is not of interest to increase the enzymatic accessibility of cellulose. However, this DP decrease pays off once the short cellulose is made amorphous again in the ball mill. ADC with an avDP of 28 AGU had a conversion degree after 1 h of 0.52 +- 0.07, higher than the AC sample. Furthermore, there seems to be a negative correlation between the avDP and enzymatic digestibility for these amorphous samples. Even within the small DP range of 20 to 110 AGU, shortening the cellulose avDP enhances its enzymatic digestibility once a low crystallinity is assured. 3.3. Effect of Enhanced Accessibility of Cellulose on Fermentation in the Human Colon A correlation between the enzymatic accessibility of cellulose samples for the Cellic CTec2 enzyme blend and the fermentability by colon microbiota can be expected since the fermentation of complex carbohydrates starts with hydrolysis by excreted microbial hydrolytic enzymes as well . The behaviour of the fiber samples in the human large intestine was evaluated in three independent batch fermentation experiments using fecal inocula. In Figure 4, the production of linear SCFA and the pH evolution during each fermentation experiment are shown. In these experiments, MC, AC, DC and ADC with different avDP were added to the fecal inocula (Table A1). In trial one , only a limited amount of linear SCFA was produced in the fecal inoculum without cellulose addition (blank) during the incubation time of 24 h. Adding dietary fiber samples to the inoculum, however, resulted in enhanced production of SCFA during incubation. The majority of SCFA was only produced after the first 8 h of incubation had passed. As described by Mikkelsen et al., cellulose fermentation in a batch in vitro system is slow compared to other readily fermentable carbohydrates, such as arabinoxylans and glucans . In this secondary fermentation phase, it became clear that only a limited amount of MC was fermented within 24 h. During the incubation of MC, the linear SCFA concentration only increased from 10.83 +- 2.63 mmol/L to 19.99 +- 0.71 mmol/L. Breaking cellulose crystallinity by ball milling increased the fermentability already slightly. The average SCFA production after 24 h from the AC sample was 0.57 times higher than the SCFA production from MC. Decreasing the avDP of cellulose was a more effective way to improve the accessibility of cellulose for the gut microbiota: the linear SCFA concentration produced by fermentation of DC with DP 59 AGU and 32 AGU was 2.6 and 1.8 times higher compared to unmodified MC. Contrary to the breakdown by the CTec2 enzyme blend, the microbiota in this pooled inoculum could access the DC better than the AC. Furthermore, the slightly lower crystallinity of DC59 resulted in a slightly higher fermentation degree for DC59 than for DC32. The highest SCFA production, however, was obtained upon the addition of ADC to the fecal pool, with a linear SCFA concentration of 41.5 +- 6.4 mmol/L at the end of incubation. By reducing both the degree of polymerization and crystallinity of MC, the formation of linear SCFA by fermentation could be multiplied by a factor of 4.2. Based on the difference in mass of linear SCFA between the blank and ADC-enriched inoculum at 48 h, a minimal degree of fermentation (MDOF) of 45.8 +- 10.9% could be derived for the ADC25 sample, while this was only 7.6 +- 0.9% for MC. This MDOF is an underestimation of the actual fermentability since it only takes into account the mass of linear SCFA as a fermentation product. Furthermore, adding ADC to the fecal pool resulted in the largest pH drop, from 6.57 +- 0.01 to 5.67 +- 0.08 . In vivo, such a pH drop could be associated with different physiological benefits, such as the repression of pathogen growth and proteolytic fermentation . In a second in vitro fermentation experiment, two different chain lengths of ADC were investigated . Additionally, the ball mill posttreatment time for the ADC was reduced to 1 h, instead of 6 h. Although a different pool of human feces was used, the same trends could be observed for the fermentability of these modified celluloses: unmodified MC was only fermented to a minimal extent, while decreasing the DP of the cellulose resulted in higher production of linear SCFA from the cellulose, up to a factor of 5.4 for DC32 after 48 h. The highest linear SCFA production was found for ADC samples ADC27 and ADC37 (8.2 and 8.4 times higher than for MC, respectively). The small difference in avDP, 37 versus 27, did not induce a significant difference in the fermentability of the ADC sample. The ADC samples were fermented to at least 42.6 +- 3.6%, while the MDOF of MC was only 5.7 +- 0.2%. The enhanced fermentation resulted in a larger pH drop of the ADC37-enriched inoculum (pH 5.58) than the MC-enriched inoculum (pH 6.02) . Furthermore, it was demonstrated in this trial that this decreased pH resulted in a lowered production of branched SCFA as well . MC addition reduced the relative amount of branched SCFA from 8.5% to 7.9%, but the addition of ADC caused a further decrease to 6.0%. This is the first indication of a lowered protein fermentation in the inocula. Detailed analysis of the acetate, butyrate and propionate concentrations demonstrate that the relative amounts of butyrate and propionate also increased after 48 h upon the addition of ADC. The relative amount of butyrate in total linear SCFA was 13.1% for the blank fecal slurry, while this was 17.6% for the fecal slurry with ADC37 addition . This enhanced butyrate production suggests an additional physiological benefit since enhanced butyrate production is linked to a lower risk of colon inflammation and cancer . In this second trial, DC and ADC samples were mainly fermented between 24 and 48 h, while the main cellulose fermentation happened between 8 and 24 h in the first trial. The presence or absence of easily accessible fibers in the starting inoculum of the trials partly explains this difference. This was hypothesized since a fast production of linear SCFA after 4 h was observed in the second trial, while it was absent in the first trial. Therefore, the microbial community in the inoculum needed more time in this trial to switch to cellulolytic fermentation metabolism than in the absence of other (more easily fermentable) carbohydrate fibers in the first trial. A third trial showed a different fermentability behaviour for the ADC. No significant differences compared to the fermentability of unmodified MC were observed, and the pH decrease during the experiment was very limited for both the ADC-enriched inocula. Next to this trial, two other repetitions showed similar behaviour with no cellulose fermentation occurring for MC or ADC (data not shown). The starting microbial composition of the inoculum within every trial is different, of course, since other donors were used for the experiments. In Figure 5, the composition of the microbiome of the three in vitro trials is given at the genus level at the starting point of the experiment and after incubation with ADC. For in vitro fermentation trials 1 and 2, the microbiome composition was dominated by Bifidobacterium and Blautia species. The proportion of Bifidobacterium species was lower for the first trial than for the second. Surprisingly, these Bifidobacteria seemed to dominate the cellulose fermentation in this first trial. After 24 h, when the ADC fermentation had already taken place, the DNA proportion from Bifidobacterium species increased from 10.7% to 37.0%, while other genera seemed to be suppressed. The fermentation of ADC might be driven by Bifidobacteria, but this enrichment can also be the result of the presence of glucose, which is released from cellulose by others. In the second trial, a different evolution of the microbial community was observed. While the microbial community was also enriched in Bifidobacteria at 24 h (when the ADC was not fermented yet), Ruminococcus species took the upper hand between 24 h and 48 h, which is the period that the ADC fermentation took place. The microbiota in the third trial did not switch to a cellulolytic metabolism within the given time frame of 48 h. The microbial composition of the starting pool of this experiment was clearly less dominated by the Bifidobacterium and Blautia species than the ones from Trial 1 and Trial 2. Furthermore, during the experiment, Bacteroides dominated the medium instead of Bifidobacterium or Ruminococcus. Consequently, we can hypothesize that the switch to cellulose fermentation occurs only if specific specialized microorganisms are present, and the composition in the pool allows them to take the upper hand. Based on Trial 1 and 2, the authors hypothesize that this switch depends on the presence and abundance of specific Ruminococcus or Bifidobacterium species, but further research is needed to confirm this statement. Despite the comparable starting microbial community of those two trials, the fermentation of both ADC samples caused an enrichment of different microorganisms, demonstrating the complexity of this fermentation process. 4. Conclusions The combination of ball milling with acid hydrolysis was demonstrated to be a valuable strategy for increasing the enzymatic accessibility of microcrystalline cellulose since it can selectively decrease the avDP and crystallinity of cellulose simultaneously. These modifications effectively resulted in an enhanced digestibility by a commercially available cellulase blend. Within an avDP range of 20-110 AGU, the avDP impacted the hydrolysability by this enzyme blend, once a low crystallinity was ensured. Furthermore, the enhanced accessibility of such amorphized depolymerized cellulose resulted in a higher fermentation degree compared to unmodified cellulose upon incubation with a pooled fecal inoculum from human subjects. With this modification, the minimal degree of fermentability of cellulose (based on the mass of SCFA produced from cellulose) within 48 h could be enhanced from 5% to 45%. This could be observed in two independent studies. However, other efforts did not show this enhanced cellulose fermentation. Microbial analyses of the fecal inocula revealed the complexity of cellulose fermentation in batch systems. Performing a detailed analysis of the cellulose fermentation metabolism in the human colon is, therefore, key to fully revealing the effect of DP and crystallinity of cellulose on fermentation in batch conditions. Until this is investigated, the authors would like to stress that the interpretation of in vitro fermentation results always has to be performed with caution, and total characterization of the microbial pool is always encouraged. However, we can conclude that engineering the properties of cellulose to high accessibility can improve the fermentation in the colon as well, be it under specific circumstances. With this work, the first step is taken towards a highly functional cellulose-type dietary fiber additive. 5. Patents The use of amorphized depolymerized cellulose as partially fermentable dietary fiber is patented in EP2022/0784403 (not published). Acknowledgments The authors would like to thank Nick Pannecoucque and Justine Van Coillie for performing part of the experimental work. Author Contributions Conceptualization, K.T., B.F.S. and C.M.C.; Formal analysis, L.C.; Funding acquisition, B.F.S. and C.M.C.; Investigation, K.T.; Methodology, K.T., J.R., N.E., B.F.S. and C.M.C.; Resources, B.F.S. and C.M.C.; Supervision, C.M.C.; Writing--original draft, K.T.; Writing--review and editing, Y.D.B., L.C., J.R., N.E. and C.M.C. All authors have read and agreed to the published version of the manuscript. Data Availability Statement Data is available upon request via [email protected]. Conflicts of Interest The authors declare no conflict of interest. Appendix A foods-12-01100-t0A1_Table A1 Table A1 Ball mill and hydrolysis conditions for different AC, DC and ADC samples, and their corresponding average degree of polymerization (avDP) and crystallinity index. ACxx, DCxx or ADCxx represents amorphized cellulose, depolymerized cellulose or amorphized depolymerized cellulose with an average degree of polymerization of xx. Ball Mill Time (min) Ball Mill Speed (rpm) Hydrolysis Time (h) Hydrolysis Temperature (degC) Ball Mill Time after Hydrolysis (min) avDP (AGU) Crystallinity Index MC - - - - - 168 +- 1 0.72 AC124 260 400 - - - 124 +- 1 <0.30 AC97 360 500 - - - 97 +- 0 <0.30 DC 360 500 16 120 - 32 +- 0 0.69 DC28 260 400 16 130 - 28 +- 1 0.62 DC42 260 400 16 110 - 42 +- 1 0.56 DC85 260 400 2 110 - 85 +- 1 0.47 DC104 260 400 2 90 - 104 +- 2 0.51 DC32 360 500 16 120 - 32 +- 0 0.69 DC59 360 500 3 100 - 59 +- 1 0.56 ADC25 360 500 16 120 360 25 +- 1 <0.30 ADC27 360 500 16 120 60 27 +- 1 <0.30 ADC37 60 500 16 120 60 37 +- 2 <0.30 Appendix B Figure A1 Rietveld refinement analysis of an XRD refractogram of depolymerized cellulose (DC) with an avDP of 32AGU. Appendix C Figure A2 Acetate, butyrate and propionate concentrations, and the relative concentrations of branched SCFA (after 48 h) of cellulose-enriched fecal inocula in the second batch in vitro fermentation with pooled human feces, in function of incubation time. DCxx or ADCxx represents the fecal inoculum with the addition of a depolymerized cellulose or amorphized depolymerized cellulose with an average DP of xx. Different letters indicate significant differences between the samples at a certain timepoint (p < 0.05). Figure 1 Visual representation of the structural changes during the production process of the different fibers, starting from microcrystalline cellulose (MC), through amorphized cellulose (AC) and depolymerized cellulose (DC), to give amorphized depolymerized cellulose (ADC). Closely packed polymers represent crystalline zones, while paracrystalline zones are visualized by loosely bound polymers. Figure 2 Effect of the ball mill (BM) treatment time (500 rpm) on the (a) crystallinity and (b) average degree of polymerization (avDP) of depolymerized cellulose (DC). Error bars represent the standard deviation of the analysis. The starting material, microcrystalline cellulose (MC), is included in both figures as a reference. Figure 3 Enzymatic conversion degree after 1 h hydrolysis with Cellic CTec2 enzyme blend for microcrystalline cellulose (MC), amorphized cellulose (AC124), depolymerized cellulose (DC) and amorphized depolymerized cellulose (ADC) with a different average degree of polymerization. ACxx, DCxx or ADCxx represents AC, DC or ADC with an average degree of polymerization of xx. Different letters indicate significant differences (p < 0.05). Figure 4 Linear short-chain fatty acid (SCFA) concentration (a,c,e) and pH (b,d,f) of cellulose-enriched fecal inocula in three independent batch in vitro fermentations with pooled human feces, in function of incubation time. ACxx, DCxx or ADCxx represents the fecal inoculum with an amorphized cellulose, depolymerized cellulose or amorphized depolymerized cellulose with an average DP of xx added. Figure 5 Microbial composition of starting, intermediate and end inocula of the different in vitro fermentation trials. The fecal inocula were enriched with ADC25 (Trial 1) and ADC27 (Trial 2, Trial 3). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Klemm D. Heublein B. Fink H.-P. Bohn A. Cellulose: Fascinating Biopolymer and Sustainable Raw Material Angew. Chem. Int. Ed. 2005 44 3358 3393 10.1002/anie.200460587 2. Vincent J.F.V. From Cellulose to Cell J. Exp. Biol. 1999 202 3263 3268 10.1242/jeb.202.23.3263a 3. Dayib M. Larson J. Slavin J. Dietary Fibers Reduce Obesity-Related Disorders: Mechanisms of Action Curr. Opin. Clin. Nutr. Metab. Care 2020 23 445 450 10.1097/MCO.0000000000000696 32925180 4. Delzenne N.M. Olivares M. Neyrinck A.M. Beaumont M. Kjolbaek L. Larsen T.M. Benitez-Paez A. Romani-Perez M. Garcia-Campayo V. Bosscher D. Nutritional Interest of Dietary Fiber and Prebiotics in Obesity: Lessons from the MyNewGut Consortium Clin. Nutr. 2020 39 414 424 10.1016/j.clnu.2019.03.002 30904186 5. Harris P.J. Ferguson L.R. Dietary Fibre: Its Composition and Role in Protection against Colorectal Cancer Mutat. Res. Mol. Mech. Mutagen. 1993 290 97 110 10.1016/0027-5107(93)90037-G 7694104 6. Terry P. Giovannucci E. Michels K.B. Bergkvist L. Hansen H. Holmberg L. Wolk A. Fruit, Vegetables, Dietary Fiber, and Risk of Colorectal Cancer JNCI J. Natl. Cancer Inst. 2001 93 525 533 10.1093/jnci/93.7.525 11287446 7. EFSA Scientific Opinion on Dietary Reference Values for Carbohydrates and Dietary Fibre EFSA J. 2010 8 1462 8. Blaak E.E. Canfora E.E. Theis S. Frost G. Groen A.K. Mithieux G. Nauta A. Scott K. Stahl B. van Harsselaar J. Short Chain Fatty Acids in Human Gut and Metabolic Health Benef. Microbes 2020 11 411 455 10.3920/BM2020.0057 32865024 9. Den Besten G. van Eunen K. Groen A.K. Venema K. Reijngoud D.-J. Bakker B.M. The Role of Short-Chain Fatty Acids in the Interplay between Diet, Gut Microbiota, and Host Energy Metabolism J. Lipid Res. 2013 54 2325 2340 10.1194/jlr.R036012 23821742 10. Wong J.M.W. de Souza R. Kendall C.W.C. Emam A. Jenkins D.J.A. Colonic Health: Fermentation and Short Chain Fatty Acids J. Clin. Gastroenterol. 2006 40 235 243 10.1097/00004836-200603000-00015 16633129 11. Cummings J.H. Cellulose and the Human Gut Gut 1984 25 805 810 10.1136/gut.25.8.805 6378732 12. Chassard C. Delmas E. Robert C. Bernalier-Donadille A. The Cellulose-Degrading Microbial Community of the Human Gut Varies According to the Presence or Absence of Methanogens: Cellulolytic Microbiota and CH4 Production in the Human Gut FEMS Microbiol. Ecol. 2010 74 205 213 10.1111/j.1574-6941.2010.00941.x 20662929 13. Robert C. Bernalier-Donadille A. The Cellulolytic Microflora of the Human Colon: Evidence of Microcrystalline Cellulose-Degrading Bacteria in Methane-Excreting Subjects FEMS Microbiol. Ecol. 2003 46 81 89 10.1016/S0168-6496(03)00207-1 19719585 14. Mikkelsen D. Gidley M.J. Williams B.A. In Vitro Fermentation of Bacterial Cellulose Composites as Model Dietary Fibers J. Agric. Food Chem. 2011 59 4025 4032 10.1021/jf104855e 21417282 15. Dubey R. Toh Y.-R. Yeh A.-I. Enhancing Cellulose Functionalities by Size Reduction Using Media-Mill Sci. Rep. 2018 8 11343 10.1038/s41598-018-29777-w 30054552 16. Nsor-Atindana J. Zhou Y.X. Saqib M.N. Chen M. Douglas Goff H. Ma J. Zhong F. Enhancing the Prebiotic Effect of Cellulose Biopolymer in the Gut by Physical Structuring via Particle Size Manipulation Food Res. Int. 2020 131 108935 10.1016/j.foodres.2019.108935 32247486 17. Liao Y. de Beeck B.O. Thielemans K. Ennaert T. Snelders J. Dusselier M. Courtin C.M. Sels B.F. The Role of Pretreatment in the Catalytic Valorization of Cellulose Mol. Catal. 2020 487 110883 10.1016/j.mcat.2020.110883 18. Hallac B.B. Ragauskas A.J. Analyzing Cellulose Degree of Polymerization and Its Relevancy to Cellulosic Ethanol Biofuels Bioprod. Biorefining 2011 5 215 225 10.1002/bbb.269 19. Gomide F.T.F. da Silva A.S. da Silva Bon E.P. Alves T.L.M. Modification of Microcrystalline Cellulose Structural Properties by Ball-Milling and Ionic Liquid Treatments and Their Correlation to Enzymatic Hydrolysis Rate and Yield Cellulose 2019 26 7323 7335 10.1007/s10570-019-02578-8 20. Glas D. Paesen R. Depuydt D. Binnemans K. Ameloot M. De Vos D.E. Ameloot R. Cellulose Amorphization by Swelling in Ionic Liquid/Water Mixtures: A Combined Macroscopic and Second-Harmonic Microscopy Study ChemSusChem 2015 8 82 86 10.1002/cssc.201402776 25363520 21. Meng X. Ragauskas A.J. Recent Advances in Understanding the Role of Cellulose Accessibility in Enzymatic Hydrolysis of Lignocellulosic Substrates Curr. Opin. Biotechnol. 2014 27 150 158 10.1016/j.copbio.2014.01.014 24549148 22. Rinaldi R. Schuth F. Acid Hydrolysis of Cellulose as the Entry Point into Biorefinery Schemes ChemSusChem 2009 2 1096 1107 10.1002/cssc.200900188 19950346 23. Sumari S. Roesyadi A. Sumarno S. Effects of Ultrasound on the Morphology, Particle Size, Crystallinity, and Crystallite Size of Cellulose Sci. Study Res. Chem. Chem. Eng. Biotechnol. Food Ind. 2013 14 229 24. Wu Q. Xu J. Zhu S. Kuang Y. Wang B. Gao W. Crystalline Stability of Cellulose III Nanocrystals in the Hydrothermal Treatment and NaOH Solution Carbohydr. Polym. 2020 249 116827 10.1016/j.carbpol.2020.116827 32933674 25. Zhao X. Cheng K. Liu D. Organosolv Pretreatment of Lignocellulosic Biomass for Enzymatic Hydrolysis Appl. Microbiol. Biotechnol. 2009 82 815 827 10.1007/s00253-009-1883-1 19214499 26. Mais U. Esteghlalian A.R. Saddler J.N. Mansfield S.D. Enhancing the Enzymatic Hydrolysis of Cellulosic Materials Using Simultaneous Ball Milling Biotechnology for Fuels and Chemicals: The Twenty-Third Symposium Humana Press Totowa, NJ, USA 2002 815 832 27. Kuo C.-H. Lee C.-K. Enhancement of Enzymatic Saccharification of Cellulose by Cellulose Dissolution Pretreatments Carbohydr. Polym. 2009 77 41 46 10.1016/j.carbpol.2008.12.003 28. Zheng Y. Yu Y. Lin W. Jin Y. Yong Q. Huang C. Enhancing the Enzymatic Digestibility of Bamboo Residues by Biphasic Phenoxyethanol-Acid Pretreatment Bioresour. Technol. 2021 325 124691 10.1016/j.biortech.2021.124691 33461121 29. Thielemans K. De Bondt Y. Van den Bosch S. Bautil A. Roye C. Deneyer A. Courtin C.M. Sels B.F. Decreasing the Degree of Polymerization of Microcrystalline Cellulose by Mechanical Impact and Acid Hydrolysis Carbohydr. Polym. 2022 294 119764 10.1016/j.carbpol.2022.119764 35868746 30. AFNOR NF G06-037 norm Essais Des Fibre-Cellulose-Determination de l'indice de Viscosite Limite et Du Degre Du Polymerisation Moyen Viscosimetrique Available online: (accessed on 20 January 2023) 31. Segal L. Creely J.J. Martin A.E. Conrad C.M. An Empirical Method for Estimating the Degree of Crystallinity of Native Cellulose Using the X-Ray Diffractometer Text. Res. J. 1959 29 786 794 10.1177/004051755902901003 32. Chen H. Zhao J. Hu T. Zhao X. Liu D. A Comparison of Several Organosolv Pretreatments for Improving the Enzymatic Hydrolysis of Wheat Straw: Substrate Digestibility, Fermentability and Structural Features Appl. Energy 2015 150 224 232 10.1016/j.apenergy.2015.04.030 33. De Preter V. Falony G. Windey K. Hamer H.M. De Vuyst L. Verbeke K. The Prebiotic, Oligofructose-Enriched Inulin Modulates the Faecal Metabolite Profile: An in Vitro Analysis Mol. Nutr. Food Res. 2010 54 1791 1801 10.1002/mnfr.201000136 20568238 34. Bautil A. Buyse J. Goos P. Bedford M.R. Courtin C.M. Feed Endoxylanase Type and Dose Affect Arabinoxylan Hydrolysis and Fermentation in Ageing Broilers Anim. Nutr. 2021 7 787 800 10.1016/j.aninu.2020.11.015 34466683 35. Falony G. Joossens M. Vieira-Silva S. Wang J. Darzi Y. Faust K. Kurilshikov A. Bonder M.J. Valles-Colomer M. Vandeputte D. Population-Level Analysis of Gut Microbiome Variation Science 2016 352 560 564 10.1126/science.aad3503 27126039 36. Calvini P. The Influence of Levelling-off Degree of Polymerisation on the Kinetics of Cellulose Degradation Cellulose 2005 12 445 447 10.1007/s10570-005-2206-z 37. Agarwal U.P. Ralph S.A. Baez C. Reiner R.S. Detection and Quantitation of Cellulose II by Raman Spectroscopy Cellulose 2021 28 9069 9079 10.1007/s10570-021-04124-x 38. Ago M. Endo T. Hirotsu T. Crystalline Transformation of Native Cellulose from Cellulose I to Cellulose II Polymorph by a Ball-Milling Method with a Specific Amount of Water Cellulose 2004 11 163 167 10.1023/B:CELL.0000025423.32330.fa 39. Kumar R. Wyman C.E. Physical and Chemical Features of Pretreated Biomass That Influence Macro-/Micro-Accessibility and Biological Processing Aqueous Pretreatment of Plant Biomass for Biological and Chemical Conversion to Fuels and Chemicals John Wiley & Sons, Ltd. Hoboken, NJ, USA 2013 281 310 40. Oliphant K. Allen-Vercoe E. Macronutrient Metabolism by the Human Gut Microbiome: Major Fermentation by-Products and Their Impact on Host Health Microbiome 2019 7 91 10.1186/s40168-019-0704-8 31196177 41. Chen J. Vitetta L. Inflammation-Modulating Effect of Butyrate in the Prevention of Colon Cancer by Dietary Fiber Clin. Colorectal Cancer 2018 17 e541 e544 10.1016/j.clcc.2018.05.001 29866614 |
PMC10000604 | Glioblastoma (GBM) is the most common primary brain tumor, yet prognosis remains dismal with current treatment. Immunotherapeutic strategies have had limited effectiveness to date in GBM, but recent advances hold promise. One such immunotherapeutic advance is chimeric antigen receptor (CAR) T cell therapy, where autologous T cells are extracted and engineered to express a specific receptor against a GBM antigen and are then infused back into the patient. There have been numerous preclinical studies showing promising results, and several of these CAR T cell therapies are being tested in clinical trials for GBM and other brain cancers. While results in tumors such as lymphomas and diffuse intrinsic pontine gliomas have been encouraging, early results in GBM have not shown clinical benefit. Potential reasons for this are the limited number of specific antigens in GBM, their heterogenous expression, and their loss after initiating antigen-specific therapy due to immunoediting. Here, we review the current preclinical and clinical experiences with CAR T cell therapy in GBM and potential strategies to develop more effective CAR T cells for this indication. immunotherapy glioblastoma CAR T This research received no external funding. pmc1. Introduction Glioblastoma (GBM), histologically defined as grade IV astrocytoma, is a highly malignant brain tumor that affects over 17,000 people in the United States annually . GBM carries a particularly poor prognosis, with median survival of 15 months and a five-year survival rate of less than 5% . The standard-of-care therapy for GBM, established by Stupp et al. in 2005, demonstrated improved outcomes for patients treated with maximal safe surgical resection, radiotherapy, and temozolomide, resulting in an average survival of 14.6 months . Given the limited effectiveness of chemotherapy and radiation, alternative treatment strategies are needed that meet the specific challenges imposed by GBM, including early microscopic spread of tumor cells, impaired immunological response, and drug delivery limitations posed by the blood-brain barrier . Unfortunately, there has been little clinical progress in the treatment of this disease. Of the phase III clinical trials that have taken place since 2005, only one study has produced positive findings . Immunotherapy is being successfully applied to a growing number of advanced cancers; however, multiple phase III trials of immune checkpoint blockade have failed to meet their endpoints in GBM. Nevertheless, subgroup analyses indicate an immune-based approach may have activity in GBM, a notion that is supported by preclinical data. In this review, we will focus specifically on the potential of CAR T cell therapy, including the data available to date and potential future directions. Cancer immunotherapy broadly involves activating the immune system to eliminate cancer cells. The most widely studied approaches are immune checkpoint blockade, anti-tumor vaccines, oncolytic viruses, and cellular immunotherapy . Chimeric antigen receptor (CAR)-T cell therapy, a form of cellular immunotherapy, has garnered particular interest in recent years due to its remarkable activity in hematologic malignancies . Generally, CAR T therapy involves extracting T cells from a patient's blood, engineering these T cells to express synthetic antigen receptors, and infusing these cells to target tumor antigens. While this form of adoptive cell transfer has been performed with T cells harvested directly from tumors, these tumor-infiltrating T cells can be difficult to isolate and expand and may not actually be specific for the tumor . The synthetic chimeric receptors, on the other hand, include both extracellular domains that recognize tumor-associated antigens (TAAs) and intracellular co-stimulatory signaling domains to promote T cell activation . Initial successes with CAR T therapy have resulted in FDA approval of multiple therapeutics for liquid cancers, and prompted interest in investigating the use of CAR T therapy for solid tumors . Recent data demonstrating activity of CAR T cells targeting GAD2 in diffuse intrinsic pontine glioma demonstrate that achieving responses against primary brain tumors with CAR T cell therapy is possible . Here, we review current experiences with CAR T cell therapy in GBM and limitations posed by antigen expression, and discuss strategies to improve CAR T cell efficacy in this context. 2. CAR T Therapy Background Recent advances in understanding T-cell receptor structure and activation have paved the path for the development of numerous T-cell-based cancer therapies. CAR T-cell therapy is at the forefront of these therapies, currently dominating the clinical trials landscape for cell-based cancer therapy . This therapy involves inducing expression of a chimeric antigen receptor, which is engineered to target a specific antigen of interest, on autologous T cells. Thus, CAR T cell therapy bypasses the requirement for MHC-mediated antigen presentation, which is required for activation of endogenous T cells . This characteristic allows CAR T cells to target a wider array of tumor-associated antigens. In addition to the antigen-specific receptor, CARs contain the intracellular domains necessary for activation so they can perform cytotoxic functions independent of environmental signals. Specifically, CAR structure includes extracellular, transmembrane, and intracellular domains . The extracellular domain consists of the antigen-binding domain and a hinge domain. Generally, the MHC-independent binding domain is derived from a single-chain Fragment variant (scFv) of the variable portions of heavy and light chains of a monoclonal immunoglobulin. The transmembrane domain is made of an alpha-helix derived from CD4, CD8a, or CD28. The intracellular domain consists of a T-cell receptor CD3 complex that facilitates signal transduction through activation of downstream kinase pathways, ultimately leading to T-cell activation and cytokine production . More recent iterations of CAR T cells involve modifying the intracellular domain to improve activation, longevity, and effector function. The first generation of these engineered receptors included a CD3 zeta domain. Notably, they required IL-2 co-stimulation to exert their cytotoxic function. To circumvent this, a second generation of receptors was designed to incorporate co-stimulatory proteins such as CD28, 4-1 BB, or OX40. These engineered cells showed superior antitumor response, with less activation-induced cell death of the CAR T cells, and increased release of pro-inflammatory cytokines . Third-generation CAR T cells integrate multi-stimulatory pathways for T-cell activation, as opposed to a single pathway in the second generation. Finally, fourth-generation CAR T cells are being designed to produce cytokines that activate antitumor immune responses . In the last decade, CAR T cell therapy has garnered considerable attention for activity in acute lymphoblastic leukemia (ALL), chronic lymphocytic leukemia (CLL), and B cell lymphomas . Because CD19 is a highly expressed antigen in these hematologic malignancies, it has become the primary target for CAR T therapy in these cancers . The results of this therapy compared with current standard treatments for relapsing B cell lymphoma may soon elevate CAR T cell therapy to standard second-line treatment . Clinical implementation of these therapies has also highlighted some limitations. For instance, although rare, patients who have recurrence after this therapy often lack CD19 antigen due to immunoediting. Thus, other antigens, such as CD20 or CD22, can be targeted in combination with CD19 to counteract antigen escape and mutation in tumor cells . Additionally, this therapy has been associated with a rare but unique set of side effects. Cytokine release syndrome and neurological toxicity are both potentially life-threatening events resulting from systemic immune activation . Overall, however, these therapies have been safe, with ongoing clinical trials in treatment of ALL, CLL, and non-Hodgkin lymphoma (NHL) continuing to show promising results . Lessons from this work in hematologic malignancies may help inform treatment in solid tumors, including GBM. 3. CAR T Therapy in GBM The success of CAR T therapy in hematologic cancers has driven interest in application of this strategy to GBM. While GBM antigens have been targeted with antibodies/vaccines with modest results, CAR T therapy may have advantages over these therapeutics . For example, there is better blood-brain barrier penetration through utilization of immune cell trafficking . Additionally, this treatment can directly kill tumor cells once bound to receptor, and thus is not dependent on the endogenous immune response, which is severely suppressed in GBM . However, a major limitation in T cell-based therapies for GBM is the dearth of tumor-specific antigen expression and low mutational burden . As the genetic and protein profiles of these tumors continues to be further understood, more potential TAAs have been discovered and used as targets for CAR T cell therapy. In parallel, new approaches are being developed to overcome the challenges of targeting infrequent and intracellular tumor antigens . While many CAR T cells for solid tumors are still in preclinical development, several are now being tested in clinical trials (Table 1). The most targeted GBM antigen is mutant epidermal growth factor receptor (EGFR). EGFR is a transmembrane receptor tyrosine kinase that can normally be found throughout the body but is minimally expressed in the CNS . EGFR expression is amplified in over 50% of GBMs, making it a promising target . While overexpression of non-mutated EGFR is limited by systemic expression, it often develops mutations late in GBM development . One particular mutation, a deletion of exons encoding the ligand binding region, can be found in 30% of GBMs . This mutated receptor variant, called EGFRvIII, is constitutively active and contributes to tumor growth. The clinical relevance of this mutation is supported by the fact that its expression is a negative prognostic factor in patients who survive a year or longer . Because of its highly specific expression on GBM and absence of expression on healthy tissues, EGFRvIII has been explored extensively as a possible target for vaccines, and more recently in CAR T cell therapy . Multiple clinical trials have been performed with third-generation CAR T cells targeting EGFRvIII . The most common adverse events seen were related to lymphodepleting chemotherapy, which is given prior to infusion of the CAR T cells . Further, two patients developed severe dyspnea at the highest dose, one of whom died, and several patients developed new minor neurological changes or seizures, thus requiring steroids or antiepileptic medications. No patients were seen to develop severe immunological side effects . Although there are encouraging anecdotes, such as the patient with recurrent GBM who lived for 36 months after EGFRvIII-targeted CAR T cell infusion, reproducible responses have been elusive . The median progression-free survival was merely 1 month, only seen on first-follow up after infusion . The overall median survival after treatment was only 6.9 months. Another target of interest for GBM is interleukin-13 receptor a chain variant 2 (IL13Ra2). Under normal physiologic conditions, cells express variant 1 of this receptor, which binds IL-13 to activate downstream JAK-STAT signaling, which arrests cell proliferation and drives apoptosis . Variant 2, which lacks the downstream signaling pathway, is restricted to tumor cells (except for the testis), and acts as a decoy receptor that competitively sequesters IL-13, preventing apoptosis of tumor cells and driving tumor progression . IL13Ra2 is expressed on numerous solid tumors, including up to 50-80% of GBM cells . Importantly, there is no overlapping expression in the brain, making it a suitable target for antigen-specific therapies such as CAR T cells. Given that the similar variant IL13Ra1 is broadly expressed throughout the brain, CAR T cells are engineered to exclude cross-reactivity . These therapies have shown success in murine models and have been tested clinically . In a safety trial with three patients with recurrent GBM receiving a first-generation CAR T targeting IL13Ra1, no major adverse events were seen; headaches were attributed to high doses of infusion, and one patient developed neurological changes that improved with steroids . A notable result included a single patient with recurrent multi-focal GBM treated with CAR T cells who initially showed a complete response after infusion, only to have a recurrence 7.5 months later . While this example shows the potential benefit of this therapy, it was not curative, and just as with EGFRvIII CAR T cells, most patients had no significant benefit. On serial imaging, two patients appeared to have decreased expected tumor recurrence in periphery where intracranial injection occurred, and the patients lived 11-14 months after their first recurrence, which was prior to receiving this therapy . Human epidermal growth factor receptor 2 (HER2) has also been utilized as a target for T cell therapy. HER2 is a receptor tyrosine kinase that is expressed on normal epidermal tissue at low levels and overexpressed in several cancers. In GBM, upregulation has been noted in up to 80% of tumors, without any concomitant expression in healthy brain tissue . Unlike EGFRvIII and IL13Ra2, HER2 is expressed at low levels in healthy epidermal tissues, leading to a higher probability of off-target effects. Although most patients have tolerated these infusions well, one patient with colon cancer who received this therapy died from cytokine storm syndrome . Fortunately, no severe adverse events occurred in a clinical trial assessing the safety of a second-generation CAR T targeting HER2 in progressive GBM . As expected, the most common reported adverse events were sequelae of lymphodepletion, and a few patients developed seizures, headaches, or minor neurological deficits. There was possible benefit seen in some patients, as the median survival in 16 patients treated with this therapy was 24.5 months from diagnosis, which was 11.1 month survival after treatment . Three patients showed no progression as late as their 29 month follow-up. Further clinical trials are being conducted. 4. CAR T Therapy Evasion in GBM Despite intensive efforts to find highly specific and robustly expressed antigens in GBM cells, targeting these antigens with CAR T therapy in clinical trials has not yielded significant clinical benefit for patients. One of the major reasons for this failure is antigen escape. Even with the successes seen in hematologic malignancies, with up to 90% remission rates in some cancers, recurrences frequently harbor CD19-negative tumor cells . This is believed to be due to selective pressure by the CAR T cells for cancer cells with increasing expression of mutated or truncated variants of CD19, ultimately leading to a subpopulation of cancer cells not detected by the CAR T cells . In a GBM mouse model, even when CAR T therapy was given in conjunction with cytokine stimulation, tumors eventually lost their expression of IL13Ra2, which led to late recurrence . In a phase II clinical trial assessing the immunological effect of EGFRvIII-targeted peptide vaccine, it was shown that the vast majority (82%) of patients who went on to have tumor recurrence had lost EGFRvIII expression . Further, in patients who received CAR T therapy targeting EGFRvIII, there was loss of this antigen expression in five of seven patients who received surgery after infusion, along with increased expression of inhibitory molecules and T regulatory cells . Another patient treated with intraventricular IL13Ra2 CAR T therapy had initial resolution of multifocal GBM, until eventual relapse of disease with tumor cells showing decreased expression of the IL13Ra2 protein . These findings highlight antigen escape as a major challenge for achieving sustained responses to CAR T cell therapy. Contributing to antigen escape is the molecular heterogeneity of GBM . Not only can antigens have heterogeneous spatial expression throughout a tumor, but they can also vary across time . This variability in antigen expression may be related to a variety of changes in the tumor microenvironment, including immune infiltration, hypoxia, and metabolism . Further, these factors may have different effects on various antigens; for instance, even without the effects of treatment, EGFR expression may be inversely related to that of IL13Ra2 and HER2 . Thus, the antigenic composition of a tumor is influenced by treatment, tumor progression, and environmental factors. Unlike with liquid tumors, immune cells in GBM and other solid tumors may be particularly impacted by these macro and microenvironmental factors; the development of glioblastoma organoids that more accurately portray the heterogeneity of tumors is allowing for better exploration of this complexity, and is only just starting to be used in preclinical studies of CAR T cells . Furthermore, immunoediting can occur rapidly and has been documented as soon as 2.5 days in vitro . Thus, the intratumoral heterogeneity of GBM creates two challenges. First, only a subset of cells of the overall tumor may be targeted and, second, the targeted cells may shift their molecular repertoire, leading to further heterogeneity and rapidly dwindling targets. To improve treatment paradigms with this technique, antigen escape and heterogeneity must be taken into consideration. 5. Improving CAR T Therapy in GBM 5.1. Targeting Alternative Antigens Ideally, target antigens should be homogenously expressed on tumors without presence in healthy tissue. These characteristics are elusive in GBM. Some newly identified antigens are showing promise in preclinical studies, with a few now being evaluated in new clinical trials. While not an exhaustive list, several antigens are worth discussing (Table 2). B7-H3 is a transmembrane protein that may have immune checkpoint properties with mixed stimulatory/inhibitory effects on T cells . It is overexpressed in over half of GBM samples and invariably expressed in neurospheres, with minimal expression in normal tissues, making it a promising target . CAR T cells have thus been designed to target this antigen and have shown promising results in preclinical studies. Treatment of U-87, U-138, and GBM neurosphere cell lines has shown elimination of tumor cells in vitro, with induction of interferon gamma and IL-2 production . Further, this CAR T therapy has been shown to control tumor growth when U87 cell lines, patient-derived tumor lines, and neurospheres are implanted orthotopically in murine models, prolonging survival in 50% of these mice compared to control treatment . Interestingly, although some of these treated tumors eventually recur, the expression of B7-H3 is retained, which may suggest late recurrence is not due to antigen loss with this treatment. Several clinical trials are currently recruiting patients with recurrent GBM for this therapy. An antigen found in up to 90% of GBM samples and without expression on surrounding healthy brain is erythropoietin-producing hepatocellular carcinoma A2 (EphA2) . Thought to be involved in oncogenesis, these characteristics make this antigen a potentially useful target in immunotherapy. CAR T cells targeting EphA2 eliminated GBM neurospheres in culture along with preventing their formation . When treating multiple EphA2+ GBM cell lines with optimized CAR T cell design, treatment showed tumor cell lysis with significant upregulation of interferon gamma and IL-2, and only low levels of induced inhibitory cytokines IL-10 and IL-4 . Further, mice implanted with U373 tumor cells had a decrease in tumor burden and improved survival after treatment, with about 50% having complete regression of tumor . Notably, no adverse effects were seen in the treated mice. CD70 is a transmembrane protein that is the ligand for CD27, an immune cell receptor involved in co-stimulation of lymphocytes via tumor necrosis factor pathway . Expression of CD70 has been seen mostly in mesenchymal GBM cell subtypes and is negatively associated with survival. CAR T therapy targeting this molecule on tumor cells causes tumor cell death and increases expression of IFN gamma and TNF alpha in culture supernatant . Additionally, mice with implanted GBM cell lines had a survival advantage when treated with CD70 CAR T cells; complete resolution was seen in tumors derived from CD70+ clones, whereas improved survival with 38% cure was seen in tumors with heterogeneous (and thus more representative) CD70 expression. These results were seen in both xenograft and syngeneic murine models, without any significant adverse effects . One concern in targeting this protein is the potential to induce immunosuppression given its role in normal T cell activation; however, preclinical evidence does not suggest this effect is significant, with no inhibition of the adaptive immune response seen one month after administration . By further modifying these CAR T cells with an additional receptor that targets IL-8 secreted by tumor cells, there was improved tumor penetrance and persistence by these CAR T cells, leading to complete tumor regression in a U87 murine model . This modified CD70 CAR T cell will be assessed in a clinical trial that is being initiated soon to assess safety in GBM patients. Natural killer (NK) cells are one of several cytotoxic immune cells. They perform surveillance under normal conditions and become activated when their receptor, natural killer group 2-member D (NKG2D), binds various ligands expressed on cells during cellular stress . Tumors including GBM are known to express NKG2D ligands; however, the immunosuppressive microenvironment reduces the levels and efficacy of the receptor on NK cells, rendering them unable to target GBM . A unique and attractive quality of this receptor for CAR therapy is that it has several different stress-related ligands it binds to . This has a potential advantage over other single-antigen CAR T cells in preventing antigen escape. When engineered into CAR T cells, NKG2D-targeted therapy showed cytotoxicity toward T98, U251, and U87 GBM cell lines in vitro, measured by increased cytokine production as well as granzyme B and perforin . There was also activity noted against GBM neurospheres in vitro, and in murine orthotopic models, mice treated with NKG2D CAR T cells showed complete elimination of tumors without recurrence 42 days after treatment. These CAR T cells preferentially accumulated within the tumors in the mice, with little infiltration in and no noted immune effect on systemic organs. With this positive safety profile, several clinical trials assessing the safety and efficacy of this specific CAR T therapy in solid tumors will also assess its effect in GBM patients. Seen widely expressed in various solid tumors, GD2 is a disialoganglioside being tested as a target for CNS tumors, particularly childhood tumors . This carbohydrate-containing sphingolipid has been implicated in tumor development through various means . After confirmation that it is indeed expressed in GBM, mice were implanted with human tumors and then treated both intravenously and intracranially with CAR T cells targeting GD2 . The latter treatment group saw a robust increase in survival, while in vitro models showed increased cytokine release with treatment. Another study showed survival benefit with peripheral injection, and further improvement in tumor control with co-expression of transgenic IL-15 . Researchers in China have further tested a CAR T cell targeting this antigen in eight patients with GBM, which caused no severe adverse events and reduced tumor size in half of the patients after infusion; however, ultimately there was no survival benefit . Chondroitin sulfate proteoglycan 4 (CSPG4), also known as Neuroglia-2 (NG2), is a protein that is expressed in 67% of GBM specimens and in a more homogenous manner than other antigens . It has been shown to be involved in promoting tumor progression and inducing chemotherapy and radiotherapy resistance . CAR T therapy targeting CSPG4 in GBM neurospheres showed cytotoxic effects against tumor cells, induced IFN-g and IL-2, and showed proliferation in the presence of tumor cells . Further, multiple murine models of different GBM and neurosphere cell lines treated by intratumor injection showed significant tumor control, with prolonged survival in 42-97% of treated mice depending on cell line. A unique characteristic of this antigen is that it is inducible by TNF-a secretion from microglia, which may allow targeting of this antigen with CAR T cells even when little expression of CSPG4 is seen to begin with, as the immune cascade eventually yields its expression on GBM cells . These findings make targeting CSPG4 with CAR T cell technology a promising therapy that needs further study. Another potential target is chlorotoxin (CLTX), a peptide isolated from the venom of the death stalker scorpion. This molecule has previously has been shown to selectively bind to tumors of neuroectodermal origin, including gliomas, without binding to normal brain . In fact, surgeons have utilized this property by intraoperative use of fluorescently labeled CLTX to identify tumor from healthy tissue . This effect is due to binding of the toxin with matrix metalloproteinase-2 (MMP-2), a membrane-bound protein widely expressed in GBM . When engineered as a receptor for CAR T cells, this therapy showed broad treatment effect in a murine GBM model, and particularly notable was its effect in tumors not expressing other frequently targetable antigens . Treatment of tumors implanted in mouse flanks yielded complete tumor regression, while treatment of intracranial implanted tumors yielded about 50% survival. It was noted that different GBM cell lines responded differently to this treatment, and those with less response were seen to have upregulated immune checkpoints, suggesting exhaustion as the main driver in recurrent cases. This CAR T therapy is being investigated in a phase I clinical trial that is actively recruiting patients. 5.2. Targeting Multiple Antigens Targeting multiple antigens is another strategy to combat antigen escape, as tumor cells may be able to downregulate a single antigen without affecting critical functions but may not be able to simultaneously alter multiple pathways. A second advantage of this approach for heterogenous tumors such as GBM is that a higher percentage of tumor cells will be targeted, leaving fewer clones capable of escape. To meet this need, CAR T cells are being developed with either multiple different receptors, tandem receptors, or a single CAR with multiple antigen-binding domains. Much of this work has been done in hematologic malignancies. For instance, many patients with aggressive B cell lymphoma who are treated with CAR T therapy targeting CD19 may relapse, with 30% of those being CD19-negative . CAR T cell therapy targeting CD20 in combination with CD19 showed superior clinical results, with fewer CD19-negative recurrences . Similarly, in a GBM tumor model of kinase inhibitor resistance, dual therapy prevented the proliferation of subpopulations of cells not targeted with monotherapy . T cells bearing combination HER2 and IL13Ra2 CARs showed benefit both in vitro and in a murine glioma model . Not only was antigen loss absent, but there were also increased antitumor effects associated with use of the dual-targeted CAR T cell approach. Combining CD70 and B7-H3 in tandem showed preclinical evidence of improved anti-tumor response in multiple cancers, including glioma . Triple-antigen primed CAR T therapy is also being used. One trivalent therapy was developed to target HER2, IL13Ra2, and EphA2 in GBM. This trivalent approach showed superior benefit to single or dual antigen approaches, and predictive modeling suggested the ability to kill nearly all tumor cells from 15 different patients despite their antigen expression variability . Given the inherent limitation of tumor specific antigens without off-tumor cross reactivity and the heterogeneous nature of antigen expression on GBM, another strategy has been developed with a so-called prime-and-kill circuit. A synthetic Notch receptor can be engineered to recognize a certain antigen, leading to transcription of a CAR . Choe et al. used this technique to develop a synNotch CAR T cell against GBM, whereby first the synNotch receptor recognizes either a GBM-specific antigen (EGFRvIII) or CNS-specific antigen (MOG), priming the cell to then express a tandem CAR targeting GBM antigens, EphA2 and IL13Ra2 . In this way, it will only become active in the vicinity of the tumor, and then target various other antigens without concern for off-tumor cytotoxicity. This strategy has potential to subvert antigen loss associated with solely targeting single, highly specific TAAs, while also preventing adverse effects on healthy tissues that may be associated with less-specific multi-valent strategies. What is also interesting about this approach is that the cells can be primed by CNS-specific, rather than tumor-specific, antigens. This is highly applicable to GBM where there is limited highly specific TAA. 5.3. Targeting Antigens of Cancer Stem Cells While initial efforts with CAR T therapy for GBM targeted surface antigens on differentiated tumor cells, there is appeal in targeting cancer stem cells to limit growth and recurrence. It is thought that within a tumor, there are only a few cells that possess the ability to propagate uninhibited and lead to differentiated cancer cells . Without eliminating these precursor cells, tumors will recur even if the bulk of tumor cells are eliminated . Standard chemoradiation has limited activity against GBM stem cells . Thus, developing a therapy with activity against these cancer stem cells is critical, and CAR T therapy shows promise. Work with previously developed CAR T cells has demonstrated antitumor activity against GBM stem cells. For instance, expression of IL13Ra2 has been noted in some GBM stem cells, defined by CD133 expression, with their selective killing by CAR T cells targeting this molecule . However, its heterogeneous expression within this cell population does not allow for complete eradication of the multipotent tumor cells. Similarly, therapy targeting HER2 has shown efficacy in eliminating HER2/CD133 positive GBM stem cells but leaves the HER2-negative subpopulation untouched . Targeting of stem cell markers together with other TAAs may specifically eliminate these stem cells. The expression of EGFRvIII can be found on a subset of tumor cells expressing CD133 . A chimeric antibody with bispecific affinity for both EGFRvIII and CD133 showed increased efficacy in cytotoxicity and decreased self-renewal abilities of GBM cells and implanted tumors. Even more promising, B7-H3 is highly expressed in patient-derived GBM neurospheres, a cell population enriched in cancer stem cells and possibly more representative of primary GBM . CAR T cells targeting this antigen showed antitumor activity for both neurospheres as well as established GBM cell lines, suggesting antitumor effects against both stem cells and differentiated cells. Although still being explored, eliminating GBM stem cells has important implications for potential curative combinations. 5.4. Adjuvant Therapies to Increase Antigen Availability and CAR T Efficacy While improving CAR T cell design may increase efficacy against the limited number of antigens present on the tumor, inducing antigen expression may be another strategy to improve CAR T therapy results. Radiation therapy, which is part of the standard-of-care treatment for GBM, may have a beneficial role in CAR T cell therapy. For instance, irradiation of tumor cells may induce further expression of TAAs. GBM cell lines with CD70 positivity showed increased expression of this antigen after exposure to irradiation . Similarly, NKG2D has been shown to increase expression on GBM following irradiation; mice bearing glioma tumors had improved survival with combination irradiation and CAR T therapy targeting NKG2D, compared to monotherapy alone . This was shown to be due to two mechanisms, one being the direct induction of the antigen itself after irradiation, and another due to increased CAR T cell migration to the tumor. Interestingly, as seen in a pancreatic tumor model, low-dose radiation may also prime CAR T cells to attack tumor in a non-antigen-dependent manner . This effect could enhance CAR T cell efficacy even if there is escape of the primary targeted TAA. Thus, radiation may have multiple synergistic effects with CAR T therapy when performed in a coordinated manner. Similar results have been shown with chemotherapy. Chemotherapy can induce the expression of tumor antigens in solid tumors, leading to a more robust immune response . This has been shown with NKG2D CAR T cells in GBM models . These findings are especially relevant given that CAR T therapies for GBM will likely be tested in patients receiving chemotherapy and radiation. Oncolytic viruses, which can specifically infect tumor cells and have well-documented immune-mediated activity, can be used in conjunction with CAR T therapy. Their direct lytic effect on tumor cells not only can trigger inflammation but can also cause the tumor cells to release TAAs . Additionally, oncolytic viruses can also be used to induce expression of any desired de novo antigen by tumor cells; for instance, a virus was designed to induce expression of CD19 on solid tumor cells that otherwise would not express this protein, allowing for successful treatment with a CAR T cell targeting CD19 . In tumors that have few and heterogenous antigens, this could have significant benefits. Focused ultrasound is another emerging technology with potential immunotherapeutic potential in the treatment of brain tumors. Direct tumor ablation with acoustic energy can release tumor debris containing neoantigens, of which targeted CAR T therapy can take advantage . Additionally, focused ultrasound can also help with drug delivery, not only via blood-brain barrier disruption but also through acoustogenetic engineering whereby the CAR T cell is only activated after exposure to acoustic energy . Further work is needed to better understand the synergistic effects of these adjuvants with CAR T cells. 5.5. CAR T Cells and Immune Exhaustion Despite best efforts to optimize CAR T cell design, immune exhaustion is another limiting step with this therapy. Briefly, the immune system has built-in checks and balances to prevent immunotoxicity when fighting chronic infection . When T cells have prolonged exposure to an antigen, they can become exhausted, having decreased effect or function and increased inhibitory receptors called checkpoints. Cancer, including GBM, takes advantage of this by secreting immunoinhibitory signals to dampen the inherent immune system, which can also decrease the efficacy of CAR T cells . Thus, addressing exhaustion may improve the overall efficacy of CAR T therapy. Inhibiting immune checkpoints that are upregulated in GBM continues to be extensively explored as a way of reversing the exhaustive immune state in the tumor microenvironment, with the goal of improving native anti-tumor immune response . Combining checkpoint inhibitors with CAR T therapy may thus improve overall CAR T cell effect and has been assessed in preclinical studies . For instance, the blockade of immune checkpoints (PD-1, CTLA-4, and TIM-3), in addition to CAR T cell therapy, showed improved efficacy of the treatment in a D270 murine model, as seen by tumor growth control and survival . Which checkpoint inhibitor improved therapy the most was dependent on which CAR T cell was used, as CAR T cells targeting EGFRvIII and IL13Ra2 both induced different checkpoint milieus in their respective tumor microenvironments. Clinical trials are starting to look into the safety of this dual therapy in patients with GBM, using CAR T cells against EGFRvIII along with Pembrolizumab (NCT03726515) and CAR T cells targeting IL13Ra2 along with Nivolumab or Ipilimumab (NCT04003649). Additionally, CAR T cells can be engineered to express proinflammatory molecules. CAR T cells that secrete checkpoint inhibitors are not only as efficacious in slowing GBM growth in murine model as co-administration of the checkpoint inhibitor, but they also have more specific targeting, as seen by increased IFN gamma and TNF alpha secretion. In another study, the proliferative capacity and persistence of CAR T cells against IL13Ra2 in a murine model was improved with concomitant IL-15 transgenic expression . Another way to address exhaustion in CAR T therapy is through resting CAR T cells via transient downregulation of the receptor . As T cell exhaustion is driven by chronic antigen exposure, temporarily preventing CAR signaling may block epigenetic changes that lead to exhaustion. By designing CARs with drug-dependent destabilizing regions, the expression of the CAR can be regulated and thus transiently shut off. This cessation of CAR expression induced CAR T-cell progression toward a memory rather than exhausted state. Even more, CAR T cells already in an exhausted phenotype showed epigenetic reprogramming with function reinvigoration . Notably, this is not seen with checkpoint inhibition. Thus, addressing the exhaustive pressure the tumor microenvironment places on CAR T cells is one more way of hopefully improving the efficacy of this treatment modality in GBM. 6. Conclusions CAR T cell therapy is a promising tumor immunotherapy that has characteristics uniquely suited to the challenges posed by GBM. These cells can traffic into the CNS and eliminate tumor cells with precision and minimize collateral damage. Next-generation CAR T cells may also deliver payloads that increase the immunogenicity of the GBM microenvironment. To date, promising preclinical findings have not translated to clinical trials. These failures, however, are elucidating the obstacles hindering success, and new strategies are being developed in response. Antigen escape is being addressed by CAR T cells with multivalent receptors, induced antigens, and combination therapies. Adjuvants with other immunotherapies, chemotherapies, or mechanically ablative methods may improve results as well by creating a more favorable, pro-inflammatory microenvironment. Continued work to better understand how the tumor, tumor microenvironment, and host immune system interact with CAR T therapy will inform further CAR T cell development and bring us closer to an effective immunotherapy for GBM. Author Contributions Conceptualization, A.S.L. and C.M.J.; investigation, A.S.L., E.Y. and P.S.; writing--original draft preparation, A.S.L., E.Y., P.S. and C.M.J.; writing--review and editing, A.S.L. and C.M.J.; visualization, A.S.L. and E.Y.; supervision, C.M.J. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest C.M.J. is a scientific co-founder and shareholder of Egret Therapeutics and receives research support from Biohaven and InCephalo. He is an inventor of patents for using immune checkpoint agonists to treat neuro-inflammation and METRNL blockade to treat cancer. Johns Hopkins University has reviewed and approved these competing interests. The other authors declare no conflict of interest. Figure 1 Depiction of third-generation CAR T cell interacting with tumor antigen and the downstream intracellular signaling that ultimately leads to T-cell activation and cytokine production. Created with BioRender.com, accessed on 17 December 2022. cancers-15-01414-t001_Table 1 Table 1 Current list of clinical trials using CAR T therapy in adult patients with GBM, including those that have been completed and those that are actively or soon to be recruiting patients. Information obtained from ClinicalTrials.gov, accessed on 5 January 2023. Status Target Antigen Title Primary Site ID Completed EGFRvIII CART-EGFRvIII + Pembrolizumab in GBM Abramson Cancer Center of the University of Pennsylvania, Philadelphia, PA NCT03726515 Completed EGFRvIII CAR T Cell Receptor Immunotherapy Targeting EGFRvIII for Patients with Malignant Gliomas Expressing EGFRvIII National Institutes of Health Clinical Center, 9000 Rockville Pike, Bethesda, MD NCT01454596 Completed IL13Ra2 Cellular Adoptive Immunotherapy Using Genetically Modified T-Lymphocytes in Treating Patients with Recurrent or Refractory High-Grade Malignant Glioma City of Hope Medical Center, Duarte, CA NCT00730613 Completed HER2 CMV-specific Cytotoxic T Lymphocytes Expressing CAR Targeting HER2 in Patients With GBM Houston Methodist Hospital, Houston, TX; Texas Children's Hospital, Houston, TX NCT01109095 Active, not recruiting EGFRvIII The Efficacy and Safety of Brain-targeting Immune Cells (EGFRvIII-CAR T Cells) in Treating Patients with Leptomeningeal Disease from Glioblastoma. Administering Patients EGFRvIII -CAR T Cells May Help to Recognize and Destroy Brain Tumor Cells in Patients Jyvaskyla Central Hospital, Jyvaskyla, Finland; University of Oulu, Oulu, Finland; Apollo Hospital, New Delhi, India NCT05063682 Active, not recruiting IL13Ra2 Genetically Modified T-cells in Treating Patients with Recurrent or Refractory Malignant Glioma City of Hope Comprehensive Cancer Center, Duarte, CA NCT02208362 Recruiting B7-H3 Safety and Efficacy Study of Anti-B7-H3 CAR-T Cell Therapy for Recurrent Glioblastoma Beijing Tiantan Hospital, Beijing, China NCT05241392 Recruiting B7-H3 B7-H3 CAR-T for Recurrent or Refractory Glioblastoma Second Affiliated Hospital of Zhejiang University School of Medicine, Hangzhou, Zhejiang, China; Huzhou Central Hospital, Huzhou, Zhejiang, China; Ningbo Yinzhou People's Hospital, Ningbo, Zhejiang, China NCT04077866 Recruiting B7-H3 Autologous CAR-T Cells Targeting B7-H3 in Recurrent or Refractory GBM CAR.B7-H3Tc Lineberger Comprehensive Cancer Center, Chapel Hill, NC NCT05366179 Recruiting B7-H3 B7-H3 Chimeric Antigen Receptor T Cells (B7-H3CART) in Recurrent Glioblastoma Multiforme Stanford Cancer Institute, Palo Alto, CA NCT05474378 Recruiting Chlorotoxin Chimeric Antigen Receptor (CAR) T Cells with a Chlorotoxin Tumor-Targeting Domain for the Treatment of MMP2+ Recurrent or Progressive Glioblastoma City of Hope Medical Center, Duarte, CA NCT04214392 Recruiting HER2 Memory-Enriched T Cells in Treating Patients with Recurrent or Refractory Grade III-IV Glioma City of Hope Medical Center, Duarte, CA NCT03389230 Recruiting EGFRvIII Autologous CAR-T/TCR-T Cell Immunotherapy for Malignancies The First Affiliated Hospital of Zhengzhou University, Zhengzhou, Henan, China NCT03638206 Recruiting EGFRvIII Autologous CAR-T/TCR-T Cell Immunotherapy for Solid Malignancies Henan Provincial People's Hospital, Zhengzhou, Henan, China NCT03941626 Recruiting EGFRvIII, GD2, IL13Ra2, Her2, EphA2, CD133 Personalized Chimeric Antigen Receptor T Cell Immunotherapy for Patients with Recurrent Malignant Gliomas Xuanwu Hospital, Beijing, China NCT03423992 Recruiting IL13Ra2 A Clinical Study of IL13ROE +- 2 Targeted CAR-T in Patients with Malignant Glioma (MAGIC-I) National Cancer Center, Korea, Goyang-si, Gyeonggi, Republic of Korea NCT05540873 Recruiting IL13Ra2 IL13Ra2-CAR T Cells with or Without Nivolumab and Ipilimumab in Treating Patients With GBM City of Hope Medical Center, Duarte, CA NCT04003649 Recruiting IL13Ra2 Brain Tumor-Specific Immune Cells (IL13Ralpha2-CAR T Cells) for the Treatment of Leptomeningeal Glioblastoma, Ependymoma, or Medulloblastoma City of Hope Medical Center, Duarte, CA NCT04661384 Recruiting NKG2D NKG2D-based CAR T-cells Immunotherapy for Patient With r/r NKG2DL+ Solid Tumors Xunyang Changchun Shihua Hospital, Jiujiang, Jiangxi, China NCT05131763 Not yet recruiting CLTX A Phase 1 Study to Evaluate CHM-1101 CAR T Cells in Patients with MMP2+ Recurrent or Progressive Glioblastoma City of Hope Medical Center, Duarte, CA NCT05627323 Not yet recruiting Dual CD70, IL8 (CXCR2) Phase I Study of IL-8 Receptor-modified CD70 CAR T Cell Therapy in CD70+ and MGMT-unmethylated Adult Glioblastoma (IMPACT) University of Florida Health, Gainesville, FL NCT05353530 Not yet recruiting EGFRvIII Long-term Follow-up of Subjects Treated with CARv3-TEAM-E T Cells Massachusetts General Hospital, Boston, MA NCT05024175 Not yet recruiting IL7R1 (CD127) Tris-CAR-T Cell Therapy for Recurrent Glioblastoma Beijing Tiantan Hospital, Beijing, China NCT05577091 Not yet recruiting NKG2D Pilot Study of NKG2D CAR-T in Treating Patients with Recurrent Glioblastoma Not listed NCT04717999 Not yet recruiting NKG2D NKG2D CAR-T(KD-025) in the Treatment of Relapsed or Refractory NKG2DL+ Tumors The Affiliated Nanjing Drum Tower Hospital of Nanjing University Medical School, Nanjing, Jiangsu, China NCT04550663 cancers-15-01414-t002_Table 2 Table 2 List of most studied antigens used for CAR T therapy in GBM, and status of use in clinical trials. Antigen Full Name Description Clinical Trial Status EGFRvIII Epidermal growth factor receptor, variant 3 Transmembrane receptor tyrosine kinase that is amplified in half of GBM tumors, minimally expressed in CNS First trials completed, ongoing recruitment in multiple other trials IL13Ra2 Interleukin-13 receptor a chain variant 2 Receptor that binds IL-13 to activate downstream JAK-STAT signaling to promote apoptosis; acts as decoy receptor in GBM First trial completed, ongoing recruitment in multiple other trials HER2 Human epidermal growth factor receptor 2 Receptor tyrosine kinase normally expressed in epidermal tissue at low levels, upregulated in 80% of GBM tumors First trial completed, ongoing recruitment in multiple other trials B7-H3 B7 homolog 3 protein Transmembrane immune checkpoint protein with mixed stimulatory/inhibitory properties, expressed on over half of GBM tumors Active recruitment in multiple trials EphA2 Erythropoietin-producing hepatocellular carcinoma A2 Protein involved in oncogenesis, expressed in over 90% of GBM but not healthy CNS tissue Active recruitment in single trial CD70 Cluster of differentiation 70 Transmembrane ligand of CD27, a co-stimulatory immune cell receptor that activates TNF pathway, expressed in mesenchymal GBM cells Not yet recruiting NKG2D Natural killer group 2-member D Binds various ligands expressed during cellular stress, including ligands expressed by GBM tumor cells Active recruitment in single trial GD2 Disialoganglioside Carbohydrate-containing sphingolipid seen involved in tumor development, expressed in GBM Active recruitment in single trial CSPG4 Chondroitin sulfate proteoglycan 4 Protein homogeneously expressed in 67% of GBM, involved in tumor progression and chemotherapy/radiotherapy resistance No clinical trial submitted to date CLTX Chlorotoxin Peptide isolated from death stalker scorpion, selectively binds to GBM without binding to normal brain Active recruitment in single trial Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000605 | Most soft tissue sarcoma (STS) patients do not respond to traditional checkpoint inhibitor treatment, which may be due to infiltrating immunosuppressive tumour-associated macrophages. This study investigated the prognostic value of four serum macrophage biomarkers. Methods: Blood samples were taken from 152 patients with STS at the time of diagnosis; clinical data were prospectively collected. The concentrations of four macrophage biomarkers (sCD163, sCD206, sSIRPa, sLILRB1) were measured in serum, dichotomised based on median concentration, and evaluated either individually or when combined with established prognostic markers. Results: All macrophage biomarkers were prognostic of overall survival (OS). However, only sCD163 and sSIRPa were prognostic for recurrent disease (sCD163: hazard ratio (HR): 1.97 (95% CI: 1.10-3.51) and sSIRPa: HR: 2.09 (95% CI: 1.16-3.77)). A prognostic profile was made based on sCD163 and sSIRPa; it also included c-reactive protein and tumour grade. Patients with high-risk prognostic profiles (adjusted for age and tumour size) had a higher risk of recurrent disease compared to low-risk patients (HR: 2.64 (95% CI: 0.97-7.19)) and (HR 4.3 (95% CI: 1.62-11.47)), respectively. Conclusion: This study demonstrated that serum biomarkers of immunosuppressive macrophages were prognostic for OS; when combined with well-established markers of recurrence they allowed for a clinically relevant categorising of patients. sarcoma macrophage immune system biomarkers sCD163 sCD206 sSIRPa sLILRB1 Danish Cancer SocietyR248-Ai4683 This research was funded by a generous grant from the Danish Cancer Society, number R248-Ai4683. pmc1. Introduction Soft tissue sarcoma (STS) is a therapeutic challenge in oncology. Over the last several decades, advances in cancer treatment have improved overall survival in many types of cancer. However, a similar improvement has not been evident in STS. STS is a heterogeneous disease, comprising more than 80 histological subtypes, which has a grave overall prognosis . Today, the standard curative treatment for STS is wide surgical resection, often in combination with radiation therapy, depending on the location and size of the tumour. Adjuvant chemotherapy is not recommended as a standard treatment for sarcoma patients , but clinical benefit has been observed in a selective group of patients . However, no risk stratification has been implemented to select such patients for adjuvant chemotherapy. Despite a treatment strategy intended to be curative, the risk of local or distant recurrence is 40 percent depending on the histological subtype of STS . Moreover, if metastatic lesions appear, less than 40 percent of patients are expected to reach a two-year overall survival . Hence, a simple method for proper risk stratification is needed so patients at high risk of either having a relapse or dying of their disease can be separated from those at low risk. The introduction of immune-modulating therapy with checkpoint inhibitors has revolutionised the treatment of many cancers, including lung cancer and melanoma . Checkpoint inhibitors interrupt cancer cells' suppression of an activated immune system, which allows the immune system to eradicate cancer cells. However, traditional checkpoint inhibitors targeting the adaptive immune defence have been effective in only a limited number of sarcoma patients . Compared to lung cancer and melanomas, sarcomas are regarded as non-immunogenic due to sparse immune infiltration and low tumour mutational burden (TMB) , as sarcomas are primarily driven by translocations rather than mutations . Furthermore, the first clinical trials testing the effect of checkpoint inhibitors in sarcoma patients have yielded disappointing results so far . Therefore, research is now focused on targeted activation of the innate immune defence to supplement T-cell-based checkpoint inhibition. New data have indicated that tumour-associated macrophages (TAMs) in tumour tissue play an important role in tumour growth and metastasis, and that a large abundance of these cells constitutes an important prognostic factor in patients with sarcomas . Generally, TAMs are polarised in a so-called M2 direction, with high expression of CD163 and CD206. These cells suppress adaptive immunity, facilitate tumour progression and metastatic spread, regulate angiogenesis, and are immune-suppressive . Experimental immunotherapy now targets such TAMs, including a blockade of the SIRPalpha-CD47 phagocytosis checkpoint . The transmembrane inhibitory SIRPa on macrophages interacts with the integrin-associated protein (IAP) CD47, a transmembrane protein with high expression in malignant tumour cells. This interaction between SIRPa and CD47 inhibits the macrophages' phagocytose; hence, interruption of the SIRPa-CD47 interaction could re-establish the anti-neoplastic effect of macrophages. Recently, our laboratory has developed analyses to test the concentration of different macrophage biomarkers in peripheral blood samples . Both increased concentrations of soluble CD163 and CD206 are associated with a poor prognosis in various cancers . However, the prognostic significance of these serum biomarkers for sarcoma patients is not known. Investigating biomarkers related to the innate immune system, and TAMs in particular, may add new prognostic information and increase knowledge about the innate immune system in sarcomas. This study will focus on four macrophage surface receptors representing M2 macrophages: signal regulatory protein alpha (SIRPa), leukocyte immunoglobulin-like receptor B1 (LILRB1), CD163, and mannose receptor (Cluster of Differentiation 206, CD206); it is the first study to investigate the prognostic role of soluble macrophage markers in sarcoma patients. 2. Materials and Methods 2.1. Study Cohort and Sampling This study is a prospective, non-randomised, non-interventional explorative study investigating the prognostic value of macrophage biomarkers. Patients with soft tissue sarcoma referred to the Sarcoma Centre of Aarhus University Hospital from 4 September 2014 to 1 April 2020 were included in the study. Inclusion criteria: patients had soft tissue sarcoma; had grade I, II, or III tumours; were over 17 years of age; were able to understand the informed consent form; and were willing to donate blood for research use. In addition, patients referred to the Sarcoma Centre of Aarhus University Hospital for suspected sarcoma but who had not been diagnosed with cancer were included as a control group; 78 control patients were included. Detailed clinicopathological information was retrieved from the patient's electronic medical records. No information on comorbidity was available. The last follow-up was carried out in May 2022. Only patients with at least 2 years of follow-up were included in this study. Blood samples were obtained at the time of diagnosis before any treatment was given. Thirty mL of peripheral blood was collected from the STS patients in sodium citrate tubes and centrifuged at 2000x g or 2500x g for ten minutes; serum was isolated and stored at -80 degC until measurement. All blood samples were handled by the Danish Cancer Biobank, Bio-and GenomeBank, Denmark, according to their instructions. After inclusion and the first blood sampling, patients were treated according to national guidelines. For most patients, the primary treatment was surgery combined with radiation therapy, depending on histological subtype, stage, tumour size, location, and grade. The median time from diagnosis to surgery was 18 days. The grading system used was that of the Federation Nationale des Centres de Lutte Contre Le Cancer (FNCLCC). This grading system is based on tumour differentiation, mitotic count, and tumour necrosis, which results in three different gradings: low grade (I), intermediate grade (II), and high grade (III). 2.2. Enzyme-Linked Immunosorbent Assays (ELISA) for Macrophage Biomarkers Serum concentrations of sCD163, sCD206, and sSIRPa, were determined by in-house sandwich enzyme-linked immunosorbent assays (ELISA) essentially as previously described . A recently established in-house ELISA assay was used for sLILRB1. In brief, Microtitre plates were coated with polyclonal anti-human LILRB1 antibody (R&D systems AF2017) and incubated overnight. After blocking and washing, serum samples (diluted 1:50), controls, and standards were applied and incubated for 1 h. Subsequently, monoclonal anti-human LILRB1 antibody (R&D systems, MAB2017) was added and incubated for 1 h. After washing, polyclonal anti-mouse horse radish peroxidase-conjugated antibodies (Dako, P0447) were added and incubated for 1 h. After washing, TMB One (kementec, 4380 L) was added, and then the plates were incubated in the dark and stopped by 1 M phosphoric acid. The plates were read at 450/620 nm, and a standard curve ranging from 0.625-8 mg/L was prepared using recombinant human LILRB1 (R&D systems, 8989-T2). 2.3. Monocyte Count and C-reactive Protein Monocyte count and c-reactive protein (CRP) concentration were extracted from the laboratory information system. For each patient, the results of the analyses were retrieved along with information on analysis date. Any measurement performed up to 90 days before the sarcoma diagnosis was considered relevant. In the case of more than one measurement, the measurement analysis performed closest in time to the sarcoma diagnosis was extracted. The patient was excluded from further analysis if no measurement from the defined period was available. Monocyte count was performed on the Sysmex XN-10 analyser (Sysmex, Kobe, Japan), and CRP was performed by a turbidimetric method using a fully automated biochemical analyser system as part of routine laboratory assessment. The monocyte count was categorised into normal or high numbers, with a normal number defined as a monocyte count lower or equal to 0.7 x 109 cells/L. CRP was similarly categorised as normal or high, with a normal CRP defined as a value lower than or equal to 8 mg/L. 2.4. Data Analysis and Statistics Categorical variables are presented as frequencies and percentages, and continuous variables are expressed as medians with an interquartile range (IQR). The correlation between serum biomarkers and clinicopathological characteristics of STS was evaluated by either the chi-square test or Spearman's rank correlation coefficient. Due to the sample size, Kendall's correlations were used to investigate the correlations between age and the different serum markers. The quantile regression model tested differences in median serum concentration with respect to stage, histological subtypes, and tumour grade. A predictive profile was created using univariate analyses. The predictive profile included all significant categorical variables: sCD163, sSIRPa, CRP, and tumour grade. The values of sCD163 and sSIRPa were divided into low or high groups based on their median values. CRP was separated into two categories, low or high, based on a threshold of 8 mg/L. Each categorical variable was assigned a score, with low levels receiving one point and high levels receiving two points. The weight of each variable was equal except for tumour grade, which was assigned the following scores: low grade--one point, intermediate grade--two points, and high grade--three points. The final profile score was calculated by summing up the scores of all of the categorical variables, with a possible range of 4 to 9. The profile score was then divided into three risk stratification groups: low risk (score 4-5), intermediate risk (score 6-7), and high risk (score 8-9). In the final model, the continuous variables age and tumour size were also included. Time to recurrence was defined as the interval between the primary diagnosis and the first recurrent, local, or metastatic relapse. Overall survival was measured from the date of diagnosis until death from any cause. Patients still alive at the time of analyses were censored. Both time to recurrent disease and overall survival outcome were analysed using Kaplan-Meier curves, log-rank tests, and univariate/multivariate Cox regression analyses. The Akaike information criterion (AIC) and Harrell's concordance index were calculated to determine whether the new profile added prognostic value to the known prognostic factors. The model with the minimum AIC values was regarded as the best model. Likelihood ratio tests were used to evaluate whether the addition of a potential prognostic profile contributed significantly to the models' prognostic value. A p-value of <0.05 (two-sided) was considered statistically significant. All analyses were performed using Stata (version 15.1) software. 2.5. Ethics Written informed consent was obtained from each patient before blood sampling, and the study protocol was approved by the Ethics Committees (journal number 1-10-72-58-14) and the internal data inspectorate (journal number 1-16-02-112-14). 3. Results 3.1. Patients, Tumour and Treatment Characteristics A total of 152 patients with soft tissue sarcoma and 78 control patients were included in the study. There was an equal distribution of sex between the two groups of patients; however, control patients were younger than STS patients (median age: control patients: 55 years (IQR: 22-77), STS patients: 66 years (IQR: 27-85); p < 0.0001). The patient, tumour and treatment characteristics are shown in Table 1. Most patients were treated with surgery for high-grade tumours with curative intent. Of the 134 patients with localised disease, 61 were treated with postoperative radiation. The rest of the patients did not receive postoperative radiation treatment due to the following: a superficial location of the tumour (n = 30; of these, 6 patients also had tumours); intra-abdominal location (n = 20); low-grade tumour (n = 17, not including superficial tumours); amputation (n = 8); or wound complications (n = 3). Five additional patients either did not want postoperative radiation or had other complications which prohibited the use of such therapy. The median follow-up time for all patients was 4.5 years (p5-p95: 0.4-7.6 years). For patients still living, the median follow-up time was 6.3 years (p5-p95: 2.3-7.6 years). At the time of diagnosis, 18 patients had metastatic disease and 134 had localised disease. A total of 137 patients were without evidence of disease after the primary treatment. Of these patients, 52 patients had a relapse of the disease during the follow-up period. At the time of analysis, 57 patients had died from any cause. 3.2. Macrophages Biomarkers in Control and Sarcoma Patients The median concentrations of all four macrophage biomarkers for patients with localised disease at the time of diagnosis were not significantly different from those of control patients. (Supplementary Table S1). Furthermore, the median values of all four macrophage biomarkers did not differ between the different histological subtypes . The effect of sex and age on the Individual biomarkers was evaluated in the control group. There were no differences in median values between sexes in any of the four biomarkers. However, significant correlations between age and sCD163 (tau-b = 0.17, p = 0.03), sCD206 (tau-b = 2.89, p = 0.002) and sSIRPa (tau-b = 0.16, p = 0.036) were observed. The concentration levels of the four macrophage biomarkers were interrelated. The strongest association was seen between sCD163 and sCD206/LilRB1; in contrast, sSIRPa was only moderately associated with the other soluble macrophage biomarkers . The biomarkers were weakly associated with blood monocyte counts and moderately associated with CRP level . 3.3. Relation of Macrophage Biomarkers to Disease Severity Patients with localised disease had lower sCD163 compared to patients with metastatic disease (2.00 mg/L vs. 2.28 mg/L, p = 0.15); however, the difference was not significant. For sCD206, sSIRPa, and sLILRB1, there was no difference in median serum concentration levels between patients with localised or metastatic disease . Patients with low-grade tumours had lower sCD163, sCD206, and sSIRPa compared to patients with high-grade tumours (sCD163: 1.83 mg/L vs. 2.13 mg/L, p = 0.08; sCD206: 0.22 mg/L vs. 0.29 mg/L, p < 0.001; and sSIRPa: 24.85 mg/L vs. 28.6 mg/L, p = 0.01; see Supplementary Table S1). There was no correlation between tumour grade and sLILRB1 . A total of 28 patients were treated with wide surgical margin and radiation therapy for extremity sarcoma; of these, 5 patients had a local recurrence and 10 patients had a metastatic recurrence. The difference in macrophage biomarkers between local and metastatic recurrence was: for sCD163, 2.26 mg/L and 2.23 mg/L; and for sSIRPa, 29.8 mg/L and 26.77 mg/L. Only three patients were treated with chemotherapy in combination with the primary treatment for localised disease. 3.4. Relation of Macrophage Biomarkers to Disease Relapse and Overall Survival All four biomarkers were significant prognostic markers for overall survival in univariate analyses . Furthermore, sCD163 and sSIRPa, but not sCD206 and sLILRB1, were prognostic factors for disease relapse (Table 2). Evaluating other established prognostic markers of survival and recurrence, we found that CRP, tumour grade, tumour size, and age at diagnosis were all significant prognostic factors for both recurrent disease and overall survival. For patients with localised disease, the two markers included in the profile, sCD163 and sSIRPalfa, remained significant (sCD163: HR 2.15 (95% CI: 1.10-4.18, p = 0.024), sSIRPalfa: HR 2.71 (95% CI: 1.37-5.35, p = 0.004)). However, sCD206 was not significant after controlling for tumour grade, HR 1.37 (95% CI: 0.73-2.60, p = 0.33), and neither was LILRB1: HR 1.41 (95% CI: 0.76-2.63, p = 0.275). A prognostic profile was composed based on the categorical variables CRP and tumour grade, as well as macrophage biomarkers sCD163 and sSIRPa (prognostic markers for both overall survivals). The prognostic profile was an independent marker of recurrent disease, with a hazard ratio of 2.91 (95% CI: 1.11-7.64, p = 0.03) for patients in the intermediate-risk group and 6.20 (95% CI: 2.27-16.90, p < 0.001) for patients in the high-risk group when compared to patients in the low-risk group. After adjusting for age and tumour size, patients with intermediate risk showed an HR of 2.87 (95% CI: 1.10-7.55, p = 0.033) and patients with high risk showed an HR of 5.85 (95% CI: 2.08-16.45, p = 0.001) when compared to low-risk patients. The prognostic profile was also a prognostic marker of overall survival, with an HR of 6.29 (95% CI: 1.45-27.28, p = 0.014) for the intermediate-risk group and an HR of 18.65 (95% CI: 4.30-81.00, p < 0.001) for the high-risk group when compared to the low-risk group. After adjusting for age and tumour size, both the high-risk profile groups had significantly worse prognoses than the low-risk group. Figure 7 shows time to recurrent disease and overall survival according to risk profile. In the analysis, 46 out of 122 patients (38%) experienced a disease relapse (local or metastatic). Five out of 31 patients (16%) in the low-risk group had a local relapse. In the intermediate-risk group, 24 out of 61 patients (39%) had a relapse; of these, 10 were local relapses and 14 were metastatic relapses. In the high-risk group, 17 out of 30 patients (57%) had a metastatic recurrence. Based on the Akaike information criterion (AIC) and Harrell's concordance index (Table 3), the addition of the profile to a prognostic model containing the known prognostic factors age at diagnosis and tumour size significantly improved the prediction of both time to recurrence (p = 0.016) and overall survival (p = 0.008). The five-year overall survival rate for patients in the low-risk group was 97 percent (95% CI: 78-99%) compared to 38 percent (95% CI: 20-56%) in the high-risk group. 4. Discussion This study demonstrated that the serum macrophage biomarkers CD163 and SIRP1a were adverse prognostic factors for the risk of relapse and overall survival in sarcoma patients. Along with the known prognostic factors tumour grade and CRP, these biomarkers comprised an excellent prognostic profile, associated with almost no relapse or death in the low-risk group and a five-year survival rate of 45 percent in the high-risk group. Furthermore, patients in the low-risk group who had a relapse did not die as a result. Additionally, prognostication was significantly improved when the profile was included in a model containing known prognostic factors such as tumour size and age at diagnosis. This occurred despite sarcomas being regarded as non-immunogenic tumours. TAMs arise from monocytes entering a tumour through blood vessels . Usually, in healthy or inflamed tissue macrophages can kill microorganisms, present antigens, and produce high levels of T-cell stimulatory cytokines. However, exposure to anti-inflammatory stimuli in the tumour microenvironment (such as IL-4 and IL-10) induces a specific M2-like phenotype of macrophages that promotes tumour cell proliferation , invasion , angiogenesis , and metastatic spread . Most studies of TAMs have been conducted in cancers other than sarcoma and show that an increased number of TAMs is associated with poor prognosis . The number of TAMs in tissue from sarcoma patients indicates that M2-like macrophages expressing CD163 are correlated with poor prognosis in patients with leiomyosarcoma , myxoid liposarcoma , and osteosarcoma . However, these studies included only a few patients. When evaluating the risk of relapse, Smolle et al. showed that a high level of TAMs in tissue from 188 patients with soft tissue sarcoma was associated with an increased risk of local recurrence . Likewise, tissue samples from patients with either localised or metastatic osteosarcoma have shown a higher infiltration of CD163 macrophages in patients with metastatic disease than in patients with localised disease . Furthermore, a phase 2 clinical trial in sarcoma patients investigated the response of sequential chemotherapy in combination with checkpoint inhibitors. Here, only a subset of patients responded to the treatment; a lack of response was associated with macrophage infiltration . All these studies point towards an immunosuppressive effect of CD163 macrophages in sarcoma, as does the current study. This study is the first to investigate soluble forms of the macrophage markers SIRPa, LILRB1, CD163, and CD206 in sarcoma, and its conclusions are in accordance with a large study conducted by Dancsok showing that SIRPa in tissue samples is an adverse prognostic factor for soft tissue sarcomas . SIRPa is an inhibitory transmembrane macrophage receptor which interacts with the integrin-associated protein (CD47). CD47 is a transmembrane protein that is expressed on normal cells but increases in number on malignant tumour cells. Overexpression of CD47 allows tumour cells to evade phagocytosis. Dancsok et al. evaluated tissue samples from 1242 soft tissue sarcoma patients for the presence of CD68, CD163, CD47, and SIRPa across sarcoma types . Infiltrating CD163-positive macrophages outnumbered the tumour-infiltrating lymphocytes in all sarcoma types. Furthermore, CD47 was correlated with SIRPa score, with the highest expression observed in chordoma, angiosarcoma, and pleomorphic liposarcoma . Because a high expression of CD47 on tumour cells might be a new target in treating sarcoma patients, the use of CD47 antibodies has been tested using both in vitro and in vivo models of leiomyosarcoma , as well as a xenograft model of human osteosarcoma . Both studies showed reduced tumour growth with anti-CD47 treatment. Our study shows that high sSIRPa is a poor prognostic factor for relapse and overall survival. Therefore, inhibition of CD47-SIRPa-complex should be tested in sarcoma patients, as this treatment strategy has shown promising results in other cancers . Besides tumour cells themselves, TAMs are affected by the tumour stroma, where both CD163 and LILRB1 are present. In gastric cancer, immunofluorescence analyses have shown that M2 TAMs are the primary immune cell expressing LILRB1 . Furthermore, high LILRB1 expression has been associated with both more advanced stages of gastric cancer and infiltration of M2 tumour-associated macrophages. However, in this study, LILRB1 did not correlate with disease grade or risk of relapse, only overall survival. The major strength of this study is its unique cohort: a large number of sarcoma patients were included over a period when treatment modalities did not change significantly. Furthermore, we used thoroughly validated and robust ELISA assays for macrophage biomarkers that allowed for the detection of small but very important changes that occurred during sarcoma development. External validation of our results may pave the way for implementing these biomarkers in clinical risk stratification of soft tissue sarcoma patients. The cohort presented in this study comprises many different histological subtypes with expected differences in overall survival. However, we could not stratify on histology due to the low number of patients in each histological subgroup. Our new prognostic profile could allow clinicians to select sarcoma patients for adjuvant treatment or a more aggressive treatment strategy, and the presence of serum macrophage markers could serve as serum biomarkers for CD47 inhibitor immunotherapy in sarcoma. 5. Conclusions In conclusion, this study demonstrated that including serum biomarkers for M2-directed macrophages in a prognostic profile allowed us to differentiate patients with a very good prognosis; even if they experienced a relapse of the disease, they did not die from it. Additionally, we were also able to identify patients with a very poor prognosis who might need additional adjuvant treatment to lower their mortality risk. However, further studies are needed to determine the role of TAMs in the development and progression of sarcomas and in sarcoma patients' responses to chemotherapy. Acknowledgments We would like to thank the Department of Surgery Aarhus University Hospital Denmark, for the collaboration with this study and acknowledge the assistance the nurses in the department provided with handling the study logistics. Furthermore, we would like to thank the clinical research unit at Aarhus University Hospital for handling the logistics in the Department of Oncology at this site. The Danish CancerBiobank is acknowledged for handling and storing biological material. The study was supported by a generous grant from the Danish Cancer Society, number R248-Ai4683. Supplementary Materials The following supporting information can be downloaded at: Supplementary Table S1: The median serum concentration (mg/L) of the four different macrophage biomarkers for patients with sarcoma, as well as control patients. Range was measured as the 5 to 95 per cent percentile, and the rank-sum method was used to compare serum concentration values. Click here for additional data file. Author Contributions Conceptualisation, N.A.-P., H.N.F., T.B.-H., H.J.M. and B.S.-P.; methodology, N.A.-P., H.N.F., H.J.M. and B.S.-P.; formal analysis, N.A.-P.; writing--original draft preparation, N.A.-P. and B.S.-P.; writing--review and editing, N.A.-P., H.N.F., T.B.-H., H.J.M. and B.S.-P.; visualisation, N.A.-P.; project administration, N.A.-P.; funding acquisition, N.A.-P. and H.J.M. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study protocol was approved by the Ethics Committees (journal number 1-10-72-58-14) and the internal data inspectorate (journal number 1-16-02-112-14). Informed Consent Statement Informed consent was obtained from all participants involved in the study. Data Availability Statement The datasets in this study are not publicly available. This is in accordance with the rules concerning processing personal data described in the EU General Data Protection Regulation (GDPR) and the Danish Data Protection Act. However, should a researcher be interested in our data, they are welcome to contact the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 The concentration of the four different macrophage biomarkers according to histological subtype, as well as the control group. Only patients with localised disease at the time of diagnosis were included in the analysis. Figure 2 The correlations between the different macrophage biomarkers. The Kendall Tau-B coefficient determined the strength of the association. Tau-b for sCD206 vs. sCD163 = 0.34 (strong association); tau-b for sCD163 vs. sSIRPa = 0.27 (medium to strong association); tau-b sCD163 vs. sLILRB1 = 0.36 (strong association). Tau-b for sCD206 vs. sSIRPa = 0.20 (weak association to medium association); tau-b for sCD206 vs. sLILRB1 = 0.22 (medium association). Tau-b for sSIRPa vs. sLILRB1 = 0.25 (medium association). Figure 3 The correlations between the four macrophage biomarkers and monocyte count or CRP levels. The Kendall tau-b coefficient determined the strength of the association between the four different macrophage markers, monocyte level, and CRP levels in peripheral blood. Tau-b for monocytes vs. sCD163 = 0.09; tau-b for monocyte vs. sCD206 = 0.007; tau-b for monocyte vs. sSIRPa = 0.09; tau-b for monocyte vs. sLILRB1 = 0.08. All indicated a weak association between the four different macrophage markers and monocyte level. Tau-b for CRP vs. sCD163 = 0.22; tau-b for CRP vs. sCD206 = 0.14; tau-b for CRP vs. sSIRPa = 0.21; tau-b for CRP vs. sLILRB1 = 0.21. All indicated a medium association between the four different macrophage markers and CRP level. Figure 4 The concentration of the four different macrophage biomarkers for patients with localised or metastatic disease. Figure 5 The concentration of the four different macrophage biomarkers according to tumour grade, as well as the control group. Figure 6 Survival curves for the four different macrophage biomarkers. The macrophage biomarkers are divided into high-risk groups based on median values. Figure 7 Time to recurrent disease and overall survival for patients with soft tissue sarcoma according to serum biomarker profile. A total of 122 patients were included in the recurrent disease analysis; 123 patients were included in the overall survival analysis because a few patients did not have all the measurements included in the profile. cancers-15-01544-t001_Table 1 Table 1 Patient, tumour, and treatment characteristics. Number (n) Percentage Sex male 68 45 female 84 55 Age median (p5-p95) 66 (27-85) Stage at diagnosis localised 134 88 metastatic * 18 12 Histological subtype liposarcoma 33 22 UPS 25 16 leiomyosarcoma 21 14 myxofibrosarcoma 17 11 angiosarcoma 9 6 synovial sarcoma 8 5 others 39 26 Median tumour size (p5-p95) 7 (1-18) Tumour grade low 23 15 intermediate 43 28 high 86 57 Depth superficial 38 25 deep 96 63 n/a ** 18 12 Treatment surgery 139 91 radiation therapy 61 40 Treatment intent *** curative 137 90 palliative 15 10 relapse yes 52 38 no 85 62 UPS: Undifferentiated pleomorphic sarcoma. * Four of the patients with metastatic disease were treated with curative intent. ** Eighteen patients did not have a depth reported because of either intra-abdominal location or metastatic disease. *** One patient with localised disease was treated with palliative intent. cancers-15-01544-t002_Table 2 Table 2 Univariate analysis for 134 patients with localised disease. Risk of Relapse Overall Survival Hazard Ratio 95% CI p Hazard Ratio 95% CI p age 1.03 1.00 -1.05 0.01 1.06 1.03-1.09 <0.001 sex female 1 1 male 1.33 0.76-2.25 0.32 1.17 0.63-2.15 0.62 size 1.04 1.01 -1.08 0.01 1.05 1.01-1.09 0.009 tumour grade 1 1 1 2 1.28 0.38-4.24 0.69 2.60 0.54-12-52 0.234 3 4.25 1.51-11.94 <0.01 8.13 1.94-34.12 0.004 Serum biomarkers Monocytes 1.12 0.57-2.22 0.33 1.72 0.88-3.35 0.111 CRP 2.01 1.11-3.65 0.02 2.77 1.47-5.22 0.002 sCD163 1.66 0.94-2.93 0.08 2.80 1.45-5.40 0.002 sCD206 1.42 0.81-2.49 0.21 1.88 1.01-3.51 0.048 sSIRPa 1.75 1.00-3.07 0.05 3.42 1.74-6.72 <0.001 sLILRB1 1.32 0.75-2.30 0.33 1.64 0.89-3.04 0.116 CI: confidence interval, CRP: c-reactive protein. Monocytes were categorised into normal <=0.7 x 109 cells/L and high >0.7 x 109 cells/L levels. CRP was categorised into normal <=8 mg/L and high >8 mg/L levels. Serum biomarkers for sCD163, sCD206, sSIRPa and sLILRB1 were categorised into low and high groups based on median values. Significant results are marked as bold/italics. cancers-15-01544-t003_Table 3 Table 3 The AIC and concordance indexes for the different prognostic models. Predictive Accuracies of the Prognostic Models Relapse Survival Model AIC C-index AIC C-index Grade 450 0.66 369 0.67 Age 459 0.61 365 0.69 Tumour size 460 0.62 380 0.66 Grade + tumour size 445 0.66 364 0.74 Age + tumour size 455 0.67 363 0.72 sCD163 463 0.57 376 0.61 sCD163 + sSIRPa 462 0.60 370 0.67 sCD163 + sSIRPa + grade 451 0.69 361 0.73 sCD163 + sSIRPa + grade + age + tumour size 445 0.73 347 0.79 sCD163 + sSIRPa + grade + age + tumour size + CRP 402 0.73 311 0.81 sSIRPa + grade + age + tumour size + CRP 401 0.73 311 0.80 Profile 406 0.66 325 0.72 Profile + age 401 0.69 313 0.78 Profile + age + tumour size 403 0.71 308 0.81 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000606 | Over the past decade, the treatment landscape of CLL has vastly changed from the conventional FC (fludarabine and cyclophosphamide) and FCR (FC with rituximab) chemotherapies to targeted therapies, including inhibitors of Bruton tyrosine kinase (BTK) and phosphatidylinositol 3-kinase (PI3K) as well as inhibitors of BCL2. These treatment options dramatically improved clinical outcomes; however, not all patients respond well to these therapies, especially high-risk patients. Clinical trials of immune checkpoint inhibitors (PD-1, CTLA4) and chimeric antigen receptor T (CAR T) or NK (CAR NK) cell treatment have shown some efficacy; still, long-term outcomes and safety issues have yet to be determined. CLL remains an incurable disease. Thus, there are unmet needs to discover new molecular pathways with targeted or combination therapies to cure the disease. Large-scale genome-wide whole-exome and whole-genome sequencing studies have discovered genetic alterations associated with disease progression, refined the prognostic markers in CLL, identified mutations underlying drug resistance, and pointed out critical targets to treat the disease. More recently, transcriptome and proteome landscape characterization further stratified the disease and revealed novel therapeutic targets in CLL. In this review, we briefly summarize the past and present available single or combination therapies, focusing on potential emerging therapies to address the unmet clinical needs in CLL. CLL emerging therapies metabolism splicing whole-exome transcriptome National Institutes of Health (NCI)R01CA216273 R01CA21623 This work was supported by grants from the National Institutes of Health (NCI) R01CA216273 and R01CA21623 (to L.W.). pmc1. Introduction CLL is one of the most common forms of adult leukemia in the western world, characterized by the accumulation of CD19+CD5+ cells in bone marrow, lymph nodes, the spleen, and peripheral blood . Typically, CLL occurs in the older age group with a median age at diagnosis of 72. In 2020, it was estimated that 21,040 new cases and 4060 deaths in the USA per year . Despite all the treatment advancements in the past decade, CLL remains incurable. Approximately 10% of patients progress to an aggressive form of lymphoma called RT, and at least 20% develop chemorefractory disease or resistance to targeted therapies . Thus, novel treatment options are still needed to cure this disease. In the past decade, large-scale genome-wide whole-exome and whole-genome sequencing studies of primary CLL samples have revealed the mutational landscape of CLL and its vast genetic heterogeneity . Integration of somatic mutations and clinical annotation enabled the identification of genetic drivers and the improvement of the prognostication of CLL patients. Recurrent alterations have been identified in genes related to significant pathways such as RNA splicing and metabolism (SF3B1, U1, RPS15, DDX3X), DNA damage (ATM, TP53), MAPK-ERK (KRAS, BRAF, NRAS), B-cell receptor and Toll-like receptor signaling (MYD88, PAX5, BCOR, IKZF3), the cell cycle (CDKN1B, CDKN2A), NF-KB signaling (BIRC3, TRAF2, TRAF3), and chromatin modification (CHD2, SETD2, KMT2D, and ASXL1) . Genetic heterogeneity functions as a fuel for clonal evolution and is implicated in disease progression and poor prognosis . More in-depth studies uncovered that genetic heterogeneity is influenced by the cell of origin; for example, some mutations appear majorly in immunoglobulin heavy chain variable region gene (IGHV) unmutated CLL (U-CLL) (U1snRNA, NOTCH1, POT1), whereas others present predominantly in IGHV mutated CLL (M-CLL)(MYD88, PAX5), in the presence of sub-clonal mutations that are acquired during the disease evolution, according to the age of CLL patients (young patients have MYD88 mutations) , or according to the course of the disease (MYC amplification is acquired during transformation to aggressive lymphoma). Identification of mutations and copy number variations on independent cohorts of patients undergoing targeted therapies will unravel the predictive value of CLL management and develop personalized treatments to combat disease resistance and improve progression-free survival (PFS) outcomes. Studies of CLL biology have also revealed novel aspects of this disease. Splicing factor mutations were identified in ~20% of CLL samples, and splicing dysregulation is reviewed as a characteristic feature of CLL . Targeting of RNA splicing dysregulation was explored recently . Metabolic reprogramming is a hallmark of cancer and underlies disease progression and relapse. In the past few years, various studies have uncovered glucose, lipid, and glutamine metabolic dependencies in CLL samples . However, the exploitation of metabolic dependencies in clinical settings is still minimal. This review summarizes past, present, and emerging therapies to combat CLL in frontline and relapsed settings. 2. Chemoimmunotherapy 2.1. Fludarabine, Cyclophosphamide, and Rituximab (FCR) Observation without cytotoxic therapy is usually the approach for asymptomatic patients in earlier stages of CLL (Rai stage 0), while therapy is recommended for those with symptomatic CLL . For over a decade, chlorambucil, a purine analog, has been the preferred treatment for CLL patients. From 1997 to 2006, the German CLL study group conducted several trials to improve the survival of CLL patients . The CLL5 trial demonstrated that older patients did not benefit from first-line therapy with fludarabine, as their median overall survival (OS) was 46 months, whereas it was 64 months with chlorambucil. Nonetheless, fludarabine is more effective than chlorambucil . The CLL4 trial revealed that combining fludarabine with cyclophosphamide enhanced the quality and duration of response in younger patients under the age of 65 when compared to fludarabine alone . The FCR300 trial (fludarabine-cyclophosphamide combined with the CD20-monoclonal antibody, rituximab) showed an overall response rate (ORR) of 95% with a complete remission (CR) of 72% and a PFS of 6 years . Long-term outcomes from the FCR300 trial showed that one-half of the patients who had mutated IGHV and received FCR achieved negative status for minimal residual disease (MRD) with a PFS rate of 79.8 at 12.8 years . Overall, in patients with or without mutated IGHV, the PFS rate at 12.8 years was 53% and 8%, respectively . In phase III of the ECOG-ACRIN 1912 clinical trial, 529 treatment-naive CLL patients aged <=70 years were randomly assigned to receive ibrutinib-rituximab (IR)/six cycles of FCR . With a median follow-up of 70 months, the results showed superior PFS with IR compared to FCR in patients with mutated IGHV (HR, 0.27; 95% CI, 0.11-0.62). Of note, ibrutinib treatment was mainly discontinued in one out of five patients due to grade three adverse events, and 60% of patients were randomized to receive ibrutinib-based therapy with approximately six years of follow-up . FCR is also associated with side effects such as myelosuppression and infections. As a result, FCR therapy is not very well-tolerated in older patients aged >65 years. A phase III CLL10 trial showed that bendamustine and rituximab (BR) are tolerable in older patients with comparable efficacy. The results showed that FCR was superior to BR in terms of PFS at 55.2 months in the FCR arm vs. 41.7 months in the BR arm; however, patients > 65 years had increased infections and cytopenias in the FCR arm . 2.2. Monoclonal Antibody Treatment in CLL For over a decade, monoclonal antibodies, such as CD20-targeting rituximab and ofatumumab or CD52-targeting alemtuzumab, have been available for CLL patients. Ofatumumab has been shown to bind a specific epitope in CD20 , improving complement-dependent cytotoxicity over rituximab. In the PROLONG trial, single-agent ofatumumab improved the PFS and remission in relapsed patients when given maintenance therapy compared to the current standard of care (observation) . Combining ofatumumab with FC or pentostatin and cyclophosphamide has shown equivalent efficacy to FCR . Another monoclonal CD52-targeting antibody, alemtuzumab, has demonstrated superiority over chlorambucil in the frontline setting in CAM 307 trials . A combination of alemtuzumab with FCR was tested in 60 high-risk CLL patients as a frontline treatment and showed an OR of 92% of CR 3 grade and grade 3-4 myelotoxicity . Due to significant toxicities such as immunosuppression, cytokine storm, opportunistic infections, and neutropenia, alemtuzumab has limited use in CLL. Recently, a novel monoclonal CD20 antibody called obinutuzumab, humanized glycoengineered (lack of sugar moiety in the Fc region) antibody, has been added to the armamentarium . In the GAUGUIN monotherapy trial, obinutuzumab (Obi) demonstrated an ORR of 62% in phase I and 30% in phase II . In the phase III CLL11 trial conducted by GCLLSG, which involved 781 previously untreated patients with a cumulative illness rating scale score of >6, patients were divided into three arms: obinutuzumab + chlorambucil (Obi-Chl, arm I) compared to rituximab + chlorambucil (Rix-Chl, arm II) and chlorambucil alone (arm III). The median PFS was 26.7 months in arm I compared to 11.1 months in arm III and 16.3 months in arm II (p < 0.001). Overall, there was an improvement in ORR, PFS (p < 0.001), and OS (p = 0.002) in arm I as compared to arm III. Additionally, there was an improvement in PFS (p > 0.001) and CR rates (20.7% vs. 7%) when comparing arm I with arm II. Obi has improved efficacy in older patients with coexisting conditions . Based on the results of the CLL11 trial, Obi has been approved by the FDA with chlorambucil for treating previously untreated patients with comorbidities. The GAIA (CLL13) study assessed the effectiveness and safety of three frontline treatments consisting of venetoclax (BCL2 inhibitor) combined with CD20 antibody compared to standard CIT for patients with CLL who were fit and without the TP53 mutation/deletion. The study followed patients for a median of 38.8 months and found that the median PFS was not reached in patients treated with venetoclax (Ven), obinutuzumab (Obi), and ibrutinib (Ven-Obi-Ibr, HR = 0.32) or Ven and Obi (Ven-Obi, HR = 0.42). By contrast, the median PFS for standard chemotherapy was 52 months. The combination of Ven and rituximab had a similar PFS to CIT. Ven-Obi-Ibr reduced the risk of CLL progression by 68%, and Ven-Obi reduced the risk by 58%. The rates of three-year PFS were 90.5% and 87.7% for Ven-Obi-Ibr and Ven-Obi, respectively. The study also found that patients treated with venetoclax-based regimens had higher rates of MRD negativity. However, overall survival rates were similar across all treatment arms. The CLL14 randomized phase III trial evaluated the combination of Obi with Ven, particularly in elderly patients with coexisting conditions who cannot tolerate intensive CIT such as FCR . About 432 patients were enrolled, and patients were randomized to either 12 cycles of chlorambucil/obinutuzumab (Clb-Obi) or 12 cycles of Ven-Obi, and the main objective was improving PFS. Overall, patients treated with Ven-Obi showed longer PFS compared to those treated with Clb-Obi and had undetectable minimal residual disease (uMRD) levels of 76% at the end of treatment. The adverse events were minimal in both arms, and patients treated with Ven-Obi demonstrated improved global health status, insomnia, fatigue, and quality of life . 3. Targeted Therapies 3.1. Targeting B-Cell Receptor (BCR) Signaling in CLL The BCR is composed of a surface immunoglobulin (Ig) molecule non-covalently associated with an Ig-a and Ig-b (CD79a/CD79b) . In normal B-cells, antigenic stimulation leads to signalosome formation, a complex of scaffold proteins and kinases tethered at the plasma membrane at the sites of sIg activation . The antigenic binding leads to the phosphorylation of SRC family kinase LYN, which is further followed by a series of kinases such as SYK, which further recruit B-cell linker protein (BLNK) and other adaptor molecules. This signalosome complex activates phospholipase C-g2 (PLC-g2), and Ras. Ras binds to and leads to the activation of Raf, which activates ERK (Extracellular Regulated Kinase). PLC-g2 activation releases intracellular calcium, which activates PKC (Protein Kinase C). PKC activation leads to subsequent activation of various kinases--MAPK (Mitogen-activated Protein Kinases), c-JUN NH2-terminal kinase (JNK), and p38 MAPK and transcription factors such as MYC and NF-kB. BCR is negatively regulated by molecules such as CD22, CD5, CD72, and FcgRIIB that control the duration and intensity of the signaling . These negative regulators contain immunoreceptor tyrosine-based inhibitory motifs . Although CLL cells express low levels of surface immunoglobulin , these cells were found to be driven by antigen-independent autonomous signaling for growth and proliferation, which is dependent on the unique heavy-chain complementarity-determining region (CDR) and an internal epitope of the BCR . It is well-known that a striking aspect of CLL is that the immunoglobin heavy chain (IgVH) and immunoglobin light chain (IgVL) have a very limited repertoire with similar gene rearrangements. Approximately one-third of CLL patients carry quasi-identical BCR sequences that can be classified into stereotyped BCR subsets based on CDR structure. These findings support the idea that tonic BCR signaling is a critical pathogenic mechanism driving CLL . In addition, a strong correlation between BCR signaling and IGHV mutation status was observed . Strong down-modulation of BCR signaling is observed in M-CLL due to a lack of sIgM expression. Less dramatic reduction of sIgM in U-CLL leads to partial activation of downstream pathways. In addition to IgM modulation, other factors also contribute to BCR signaling. ZAP70 expression correlates with BCR signaling, and its overexpression augments BCR signaling capacity . In CLL, ZAP70 overexpression is associated with poor clinical outcomes and expresses unmutated IgHV . Additional evidence suggests that ZAP70 may modulate cell migration-associated pathways. CD38 is another prognostic indicator whose expression correlates with sIgM signaling capacity . Reports indicate that CD38-CD31 interactions contribute to cell migration and homing, enhancing CLL survival via inducing BCL2 and BCL XL . Overall, BCR signaling is essential for CLL pathogenesis; thus, targeting this pathway is vital for treating CLL patients. 3.2. Dasatinib and Ibrutinib Targeted tyrosine kinase inhibitors have been of particular interest in treating CLL due to the success of using tyrosine kinase inhibitors in CML. Dasatinib, which has been used to treat CML and Ph+ ALL acute lymphoblastic leukemia (ALL), can also inhibit the Src family of kinases, including Lyn, which is often dysregulated in CLL B-cells. Dasatinib induces apoptosis in CLL B-cells in vitro, with U-CLL cells being more sensitive , and has also been shown to sensitize tumor cells to chlorambucil and fludarabine to overcome CD40-mediated drug resistance in vitro . A phase II study of dasatinib in high-risk relapsed or refractory CLL patients has shown a partial response with side effects of myelosuppression in two-thirds of the cases . A recent in vitro study with 53 CLL patient samples treated with dasatinib showed that 17.7% of samples were apoptotic, indicating that dasatinib has anti-leukemic effects ; however, clinical use of dasatinib has been unclear. Ibrutinib, a covalent inhibitor for Bruton's tyrosine kinase (BTK), was approved by the FDA as one of the first inhibitors for treating relapsed or refractory (R/R) CLL after showing positive results in the RESONATE trial, a randomized phase III trial that compared ibrutinib with a single agent ofatumumab in R/R CLL patients . The 18-month PFS was similar regardless of the genetic mutations, including NOTCH1, BIRC3, and ATM, or other factors such as del(17p) and del(11q) . In addition, a phase Ib/II single-arm trial compared ibrutinib with single-agent PCYC-1102 treatment, where patients with R/R and TN settings were treated with ibrutinib as a single agent. A pooled analysis from the RESONATE and PCYC-1102 trials showed that patients who received ibrutinib and no prior treatment or bulky disease (lymph node >= 5 cm) achieved a CR. Another RESONATE-2 trial compared ibrutinib with chlorambucil in 269 treatment-naive older (65+) patients with CLL/SLL, where ibrutinib showed large improvements at ORR . RESONATE-17 was a single-arm study of ibrutinib in 144 R/R CLL or SLL del(17p). The results showed that the 24-month PFS, OS, and ORR were 63%, 75%, and 83%, respectively, after a median follow-up of 11.5 months . Notably, complex karyotype (CK) may be a strong predictor of inferior outcomes compared to del(17p) among patients with R/R CLL in the setting of ibrutinib therapy. The MD Anderson Cancer Center analyzed 88 patients with R/R CLL who received ibrutinib-based therapies between 2010 and 2013. Investigators found that only fludarabine refractoriness and complex karyotype were statistically significant prognostic factors associated with lower OS, with an HR of 6.9 and 5.9, respectively. Interestingly, del(17p) was not associated with a decreased OS . In light of the success of ibrutinib, second-generation BTK inhibitors (BTKi) such as acalabrutinib and zanubrutinib have been developed aiming to reduce grade 2-4 adverse events, bleeding, and cardiac toxicities. In the ASCEND trial, a phase III trial focused on cytogenetic high-risk R/R patients, acalabrutinib demonstrated superior PFS compared to CIT and the PI3Kd inhibitor idelalisib . Furthermore, the ELEVATE-TN trial examined first-line treatment in elderly CLL patients with comorbidities, which showed that single-agent acalabrutinib or in combination with the anti-CD20 mAb obinutuzumab can prolong PFS . In a recent ELEVATE-RR phase III-based clinical trial dedicated to R/R CLL patients, acalabrutinib treatment was non-inferior in terms of PFS (38.6 months) compared to ibrutinib (38.4 months PFS) and had an improved safety profile with fewer AF events . Zanubrutinib has demonstrated an ORR of 85% in phase II clinical trials with treatment-naive (TN) and R/R CLL with TP53 mutations . In a randomized phase III controlled clinical trial, zanubrutinib single treatment was compared with ibrutinib and zanubrutinib in R/R CLL patients . Zanubrutinib has shown a superior response rate and improved PFS with a lower rate of atrial fibrillation than ibrutinib . Similar to other kinase inhibitors, resistance was developed by either acquired mutations in the BTK domain, such as mutations in the drug binding site (BTK Cys481) PLCG2 mutations , or alterations, such as del(8p) and additional mutations in EP300, MLL2, and EIF2A. To overcome the BTKi resistance, pirtobrutinib has been developed as an orally available, highly selective reversible BTKi with equal potency against wildtype and Cys481 mutated BTK. It has shown to have an overall good safety profile with an ORR of 62% in phase I/II studies of R/R CLL patients . Mechanisms for non-genetic events-related resistance were under investigation in CLL and other B-cell malignancies with candidate treatment options proposed . 3.3. PI3 Kinase Inhibitors The p110d isoform of PI3K is responsible for transducing downstream signals of BCR signaling. Although idelalisib has been approved, many concerns exist regarding the toxicity and modest outcomes associated with treatment. In particular, compared to the relatively safer ibrutinib, the use of idelalisib in patients with CLL has been limited. Thus, it is critical to exercise a high degree of vigilance for these adverse events in patients receiving idelalisib. 3.3.1. Idelalisib (p110-PI3Kd Inhibitor) Idelalisib (IDEL) is a selective PI3Kd inhibitor that promotes the apoptosis of B-cells. Activating the BCR signaling increases the calcium, which increases the diacylglycerol and IP3, thus activating the PI3 kinase pathway. Hence, treatment of CLL cells with PI3Kd inhibitors causes a reduction in lymph node size and concomitant enhanced lymphocytosis, which can be mitigated using rituximab. In phase III, a randomized, double-blind, placebo-controlled trial (220 high-risk R/R CLL patients) of idelalisib + rituximab (IDEL-RIX) versus a placebo + rituximab, the IDEL-RIX arm had significantly improved ORR compared to the placebo arm (81% vs. 13% p < 0.001), as well as OS (92% vs. 80% at 12 months p = 0.02), irrespective of the presence of poor prognostic factors including del17p . In this phase III trial, the adverse events were similar in both the idelalisib and placebo groups; in the idelalisib group, there were five common adverse events like fatigue, pyrexia, nausea, chills, and diarrhea. Grade 3-4 elevation of hepatic transaminases occurred more in the idelalisib group. Gastrointestinal and skin disorders led to discontinuation in the idelalisib group . In another phase II trial involving 64 patients, IDEL-RIX showed an ORR of 97% with a CR of 19%. Despite its efficacies, grade 2-3 AEs (adverse events) such as transaminitis, rash, and diarrhea have been reported . All these AEs are managed with a drug hold; hence, the US FDA approved idelalisib in combination with rituximab for high-risk R/R CLL patients. 3.3.2. Duvelisib Duvelisib is an inhibitor that targets two forms of the phosphatidyl 3-kinase (PI3K) enzyme, specifically the PI3Kd and g forms, and has been used to manage high-risk CLL patients. In September 2018, the US FDA approved duvelisib for treating R/R CLL patients after having undergone two prior lines of therapy. In a phase I dose-escalation trial involving 52 R/R CLL patients and 15 older previously untreated patients with CLL, duvelisib showed an ORR of 48% with 89% nodal responses . Currently, duvelisib is undergoing testing in a phase III trial in comparison with ofatumumab for relapsed CLL patients. In a randomized phase III DUO trial, duvelisib was compared with ofatumumab monotherapy in patients with R/R CLL. Duvelisib showed a significantly higher ORR (74%) than the ofatumumab arm (45%; p = 0.0001). The median PFS was 13.3 months in the duvelisib arm versus 9.9 months in ofatumumab, respectively, among all patients. The median PFS of del(17p) or patients with the TP53 mutation was 12.7 months on treatment with duvelisib, while the median survival of patients without such lesions was not reported . The most common adverse events are diarrhea, neutropenia, pyrexia, nausea, anemia, and opportunistic infections such as P jirovecii pneumonia in patients who did not receive prophylactic treatment. These data indicate that duvelisib may be effective for R/R CLL patients. 3.3.3. Umbralisib Umbralisib is a novel PI3K inhibitor with efficacy similar to idelalisib is umbralisib, which has been shown to have low toxicity in R/R CLL patients. In the phase I-II trials comprising a triplet combination of umbralisib with ublituximab plus venetoclax, the ORR was 90% with a CR of 29% and a two-year PFS of 90%. In the peripheral blood and bone marrow, 58% of patients had undetectable MRD . Due to findings from the UNITY-CLL clinical trial (NCT02612311) continuing to show a possible increased risk of death in patients receiving umbralisib, TG Therapeutics withdrew it from the market in June 2022. 3.4. Targeting BCL2 B-cell leukemia/lymphoma-2 (BCL2) is a protein often overexpressed in CLL patients and plays a crucial role in regulating the apoptotic pathway. Venetoclax is a highly selective, second-generation small-molecule inhibitor of BCL2 that can effectively shift the balance of proteins in CLL cells towards apoptosis. In the phase I clinical trial that included 116 high-risk R/R CLL patients, venetoclax achieved an ORR of 79%, with 20% of patients experiencing a CR. Notably, patients with del(17p) and those who were refractory to fludarabine achieved even higher ORRs of 82% and 89%, respectively. The most common grade 3-4 adverse event reported was neutropenia, in 41% of patients . Currently, there are several combination trials underway involving venetoclax. In one study that examined the combination of venetoclax and rituximab (Ven-Rix) in high-risk R/R CLL patients, approximately half of the patients achieved CR, with 57% achieving uMRD activity . The estimated two-year PFS was 82% (95% CI 66-91); however, two patients progressed after 24 months of therapy . The phase III MURANO trial compared a combination of Ven-Rix versus bendamustine/rituximab (BR) in R/R CLL patients. The two-year PFS rate was 84.9% and 36.3% in the Ven-Rix arm and the BR arm, respectively, with a median PFS significantly higher in the Ven-Rix arm . Tumor lysis syndrome (TLS) is one of the most common adverse events associated with venetoclax treatment. Hence, appropriate patient selection, strategies to reduce TLS risk, and following the standard ramp-up will be critical for successful treatment with low TLS risk. Preclinical studies investigating the synergy of BTK and BCL2 inhibitors and single-center studies examining the effectiveness of ibrutinib in combination with venetoclax with or without obinutuzumab have shown promising results in CLL treatment. A phase II international study named CAPTIVATE (NCT02910583) examined the use of fixed-duration treatment with ibrutinib plus venetoclax (Ibr-Ven) in patients aged <=70 years who had not received any prior treatment for CLL . Despite a median follow-up of only 27.9 months, the results seen in this trial are remarkable. The treatment regimen demonstrated exceptional efficacy with a 56% CR, and with 76% and 62% of patients achieving uMRD in the blood and bone marrow. At the 24-month mark, 95% of patients were alive and free of progression. Notably, patients with high genomic risk diseases such as TP53 abnormalities had excellent outcomes. Additionally, patients with unmutated IGHV showed a trend toward achieving uMRD . A phase III trial named GLOW (n = 211) evaluated the efficacy and safety of Ibr-Ven in older patients and/or those with comorbidities with previously untreated CLL. After a median follow-up of 46 months, the Ibr-Ven treatment was found to reduce the risk of disease progression or death by 79% compared to chlorambucil with Obi (Clb-Obi) (Hazard Ratio (HR) 0.214; 95% Confidence Interval (CI), 0.138-0.334; p < 0.0001). This marks the first instance of a fixed-duration novel combination demonstrating an OS advantage compared to Clb-Obi as a first-line treatment for CLL (HR 0.487; 95% CI, 0.262-0.907; nominal p = 0.0205) with the Ibr-Ven treatment given once daily orally. Estimated data suggests that 74.6% of previously untreated older and/or comorbid patients remained alive and progression-free after 3.5 years of fixed-duration Ibr-Ven treatment, compared to 24.8% of patients in the Clb-Obi cohort. 4. Immunomodulatory Agents 4.1. Immune-Checkpoint Inhibitors CLL is a disease of the mature B-cells; however, recent reports indicate the role of T-cells in the disease pathogenesis and progression . In CLL, T-cell exhaustion is mediated by upregulation of co-inhibitory signals such as programmed death-1 (PD1), lymphocyte activation gene-3 (LAG-3), cytotoxic T-lymphocyte-associated protein-4 (CTLA4), and T-cell immunoglobin-3 (TIM-3). These findings have led to the development of immune-checkpoint inhibitors for managing CLL treatment . However, PD1 inhibitors used as single agents have failed to produce promising results in CLL . The anti-PD1 mAb pembrolizumab has shown selective efficacy in CLL patients progressing to RT. Recent trials in RT have demonstrated that combining ibrutinib with anti-PD1 mAb nivolumab has produced considerable results with acceptable levels of toxicity . Moreover, a triplet combination of umbralisib, ublituximab, and pembrolizumab has reported durable responses . As evidenced by high response rates in previously untreated RT patients, the combination of the Bcl2-inhibitor venetoclax, next-generation anti-CD20 mAb obinutuzumab, and anti-PDL1 mAb atezolizumab offer immunotherapy as a promising treatment approach. 4.2. Bispecific Antibodies Bispecific antibodies (bsAbs) are a promising approach that combines antibody therapies with cellular-mediated immunotherapy. BsAbs are antibodies with two binding sites directed at two different antigens or two different epitopes on the same antigen. There are two types of bsAbs: bispecific T-cell engagers (BiTEs) and dual-affinity targeting antibodies (DARTs). BiTEs are a subtype of bispecific antibodies that link two single-chain variable fragments with a flexible linker. One fragment binds to the tumor-associated antigen, and the other binds to a T-cell-specific antigen to activate the T-cell to kill the cancer cell to which it is linked. DART consists of two variable fragments that connect the opposite heavy chain variable regions through a disulfide bond, improving stability. Blinatumomab, a CD19/CD3 bsAb designed in a BiTE format, was one of the first bsAbs tested in CLL and has been shown to eliminate CLL cells in a mouse xenograft model . Blinatumomab was tested in vitro on 28 freshly treated naive CLL patients. The antibody treatment induced tumor cell death via T-cell activation and granzyme-mediated cytotoxicity . Clinical trials currently underway involve combining lenalidomide (NCT02568553) or blinatumomab-expanded T-cells (NCT03823365) in patients with a broad spectrum of NHL, including CLL. Promising results have emerged from preclinical studies of another bsAb called MGD011 CD3 X CD19 DART, which has shown the ability to effectively engage CLL-derived T-cells and promote the killing of tumor cells in vitro. Further, MGD011 also indicated an impact on eliminating CLL resistance to venetoclax . Recently, preclinical studies have explored the potential of a bispecific antibody that targets leukemic cells and Vg9Vd2 T-cells, a conserved T-cell subset with intrinsic anti-tumor activity. Furthermore, a CD40-bispecific gd T-cell engager has been found to trigger apoptosis through a powerful Vg9Vd2 T-cell-dependent anti-leukemic response . Altogether, bsAbs may be a potentially valid option for high-risk patients resistant to previous therapies. 4.3. Tri-Specific Killer Engagers Natural killer cells are hypofunctional in CLL, affecting target cell recognition and cellular toxicity . BiKEs (Bi-specific Killer Engagers) and TriKEs (Tri-specific Killer Engagers) recruit NK cells to target tumor antigens. The NKG2D receptor-ligand ULBP2 has been targeted by TriKEs (ULBP2/aCD19/aCD19 and ULBP2/aCD19/aCD33) and has demonstrated in vitro and in vivo activity against CLL . Further, a CD16/CD19 BiKE and a CD16/CD19/CD22 TriKE have been shown to trigger NK cell activation via CD16 signaling, for which CD16/CD19 TriKE induced better killing of CLL cells in vitro compared to rituximab . Overall, inducing an NK cell response against CLL cells is compelling to explore as a therapeutic option. 4.4. CAR T-Cells Chimeric antigen receptor T-cells (CAR T-cells) represent a promising area of investigation in adoptive cellular therapies, combining the strengths of T-cells and antibodies to boost T-cell anti-tumor activity. To date, there have been four generations of CAR T constructs developed. These constructs typically consist of an antigen binding domain, such as a single chain variable fragment (scFv) derived from immunoglobulin directed against the tumor antigen, as well as the intracellular domain from the CD3 chain, and a costimulatory domain, which is generally identified as the intracellular domain of a costimulatory molecule (CD28/4-1BB) . One of the key advantages of CARs is their ability to identify and target tumor antigens in an HLA-independent manner, thus targeting tumor cells in a tumor-evasive environment . CD19 CAR T-cells have been widely used in B-cell malignancies; however, their usage in CLL is controversial due to exhausted T-cell phenotype in CLL and the loss of CD19 upon CAR T therapy resistance. In the CLL 4 trial, CD19 CAR T-cells were used as a single agent, resulting in an ORR of 82%, with a 45% CR and a high rate of uMRD observed in heavily pre-treated CLL patients, including high-risk patients who were refractory to BTKi and venetoclax . Other studies of CD19 CAR T-cells in CLL have also shown an ORR of 50-70% and a CR of 20-30% . Kappa or lambda light chains can be attractive targets for CLL patients, as it allows high target specificity for the leukemia cells while avoiding the normal B-cells. Recently, a new CAR T has been investigated against the Ig light chain, as CLL cells mainly express the Ig light chain compared to their normal counterparts . Preclinical studies showed CAR T-cells have been effective against Igk or Igl in vitro or in vivo CLL models, and there is an ongoing clinical trial investigating anti-Ig kappa on CLL (NCT04223765). Combination therapies have also been tested. A recent study suggested increased viability and expansion in human CAR T-cells in the presence of ibrutinib. In line with this, administering ibrutinib with anti-CD19 CAR T has increased tolerability with a lower incidence of severe side effects. One of the impediments to CAR T treatment in CLL is decreased fitness and activity of CAR T-cells due to immune subversion. Thus, using allogenic CAR T-cells from healthy donors has been an attractive option; however, strategies are evolving, such as gene editing-based strategies to knock out endogenous abTCR to prevent graft-versus-host disease and donor-mediated rejection . Similarly, acalabrutinib has been shown to improve the in vitro and in vivo anti-tumor function of CD19 CAR T-cells . These studies' results will guide future treatment when including CAR T-cells. Even though CD19 as an antigen has been a promising target, there are resistance mechanisms, such as a CD19 loss, leading to exploration for novel targets. Another good target is CD20. Anti-CD20 CAR has been investigated in non-Hodgkin lymphomas. Among the three patients who received anti-CD20 CAR, two did not develop the evaluable disease with a progression-free survival of 12 and 24 months. The third patient achieved partial remission and relapsed after 12 months post infusions . An ongoing trial evaluates anti-CD20 CAR in R/R B-cell malignancies, including CLL (NCT0327779). Another attractive antigen expressed on CLL cells but not on normal B-cells is the ROR1 (Receptor tyrosine kinase-like orphan receptor 1) receptor . In vitro, promising data shows CAR T's effect on ROR1 and an ongoing trial evaluating anti-ROR1 CART against ROR1 malignancy, including CLL (NCT02706392). FcmR is expressed highly in CLL cells while at minor levels in normal healthy B-cells. Anti-FcmR CAR T has been investigated in CLL cells, which affects CLL cells without affecting normal healthy B-cells . 5. Emerging Novel Targets for CLL Treatment 5.1. Targeting RNA Splicing Dysregulation in CLL Splicing is a highly precise and stepwise process that converts pre-mRNA into mature RNA, facilitated by a group of proteins called the spliceosome. The spliceosome is composed of over 300 proteins, which includes more than 100 accessory proteins that process the U2-type introns. The core of the spliceosome includes U1, U2, U4, U5, and U6 small nuclear ribonucleoproteins (snRNPs), as well as seven Sm proteins or Lsm (U6-specific) proteins and other snRNP-specific factors . Each snRNP contains a small nuclear RNA (snRNA), enabling interactions between RNA-RNA and RNA-protein during the dynamic splicing process . Typically, cells generate various mRNA forms via alternative splicing, which occurs through multiple mechanisms such as alternative 5' or 3' splice sites, exon skipping, alternative promoter, intron retention, and alternative polyadenylation . Genome-wide cancer sequencing studies have identified recurrent mutations in RNA splicing factor proteins (SF3B1, U1 snRNA, SRSF2, U2AF1, ZRSR2) myeloid neoplasms, clonal hematopoiesis, mantle cell lymphoma, and CLL . All these mutations lead to transcriptome-wide RNA splicing dysregulation. Additionally, RNA sequencing studies of primary cancer cells across various cancer types have revealed that aberrant RNA splicing is a common feature of cancer. In TCGA (The Cancer Genome Atlas) analysis of 33 different cancer types, mutations in 119 splicing factors were reported, which comprise half of the splicing factor proteins . Moreover, 70% of splicing factors and 84% of RNA binding proteins are dysregulated at mRNA levels in various cancers, resulting in dysregulated splicing events. The discovery of splicing factor mutations has generated an interest in therapeutic targeting of the splicing factor mutant tumor cells. One interesting feature in splicing factor mutant cases is the solid mutual exclusivity. Several reports also suggest that co-expression of the most common splicing factor mutations in SF3B1, SRSF2, or U2AF1 is not tolerated in cells . Similarly, expression of a single wild type encoding these factors is tolerated, while deletion of the wild type in splicing factor mutant cell lines leads to cell death . This evidence further motivated the development of inhibitors to target the splicing catalysis function to kill splicing factor mutant tumors. Among the various inhibitors developed are a class of natural products and their synthetic analogs that bind to the Sf3b complex and prevent interaction with the branch point. Among the widely studied compounds, the pladienolide analogs (A-G and synthetic analog E7107) target Sf3b, which binds the U2 snRNP to disrupt splicing . PLAD-B and FD-895 have been shown to induce apoptosis and overcome the protective effect of the microenvironment in CLL in vitro, indicating that these inhibitors may work in R/R CLL patients. Another pladienolide analog, E7107, has been tested in phase I clinical trials and showed limited efficacy; however, due to adverse events, the trial was terminated . A recent report suggests the combination of E7107 with venetoclax sensitized both human and murine CLL cells in vitro and can overcome venetoclax resistance in vivo . Another class of inhibitors similar to pladienolides is FR901464 and its methylated derivative spliceostatin A (SSA), which inhibit the Sf3b subcomplex . Bcl2 family member Mcl1, an apoptosis regulator, is highly expressed in CLL samples with progressive disease. Reports suggest that spliceosome inhibitor spliceostatin (SSA) altered the splicing of Mcl1 and led to the downregulation of Mcl1, resulting in apoptosis . Notably, the microenvironmental signals such as CD40L and IL4 treatment of CLL cells offered resistance to spliceostatin. This resistance was reversed using a combination of spliceostatin with ABT-199/263-BCL2 family inhibitor, indicating that the combination may work for CLL cells resistant to spliceostatin . Another synthetic analog of FR901464, sudemycin, has been shown to induce apoptosis in CLL samples without affecting the normal B-cell counterparts in vitro and in vivo . Further, combining sudemycin with ibrutinib confers enhanced sensitivity to ibrutinib by modulating the loss of regulatory function of IBTK over BTK via alternative splicing regulation . Despite the promising results with splicing inhibitors in CLL, none of the inhibitors have been approved by the FDA for treatment, hence further understanding is needed to modulate splicing catalysis in vivo with an acceptable therapeutic index. 5.2. Targeting Metabolism in CLL Metabolic changes enable the tumor cells to sustain proliferation and adapt to stressful conditions. Hanahan and Weinberg noted that metabolic rewiring is a hallmark of cancer cells . Even though CLL cells are known to be quiescent, they have been shown to have high mitochondrial respiration and reactive oxygen species and enhanced antioxidant activity compared to normal B-cells . Accumulating evidence indicates that CLL cells undergo spontaneous apoptosis, and a gradual increase in the size of the CLL clone results from the newly proliferating lymphocytes. CLL cells are generally slowly proliferating, with approximately 0.1% to 1.75% of CLL cells proliferating daily compared to resting B-cells . Very few studies are investigating the role of metabolism in CLL and exploiting it as a therapeutic approach. We will briefly discuss the recent works related to metabolism in CLL. 5.2.1. Mitochondrial Metabolism Mitochondria play an important role in energy metabolism, as they regulate oxidative phosphorylation, reactive oxygen species, and ATP production via the TCA cycle. Mitochondria also regulate other metabolic processes, such as amino acid and fatty acid metabolism. CLL cells have higher mitochondrial mass, ROS, and activity than normal B-cells . Primary CLL samples have been shown to have high basal respiration via seahorse-based assays . Recent omics analyses indicate that CLL proliferation is linked to the mTOR-MYC-OXPHOS pathway . Overexpression of oxoglutarate dehydrogenase and isocitrate dehydrogenase was commonly found in CLL cells . In line with this, CLL cells are sensitive to the pharmacological inhibition of oxidative phosphorylation by OXPHOS inhibitors (PK11195, oligomycin A, and metformin) . 5.2.2. Glucose Metabolism Glucose can be converted to pyruvate via glycolysis and ribose via the pentose phosphate pathway. Glucose is further utilized to synthesize glycogen, fatty acid, and serine. An increase in glycolytic flux to produce ATP meets the energy demands of highly proliferating cells, rendering the cells addicted to glucose. CLL cells have been shown to have high glucose metabolism and uptake. Glucose uptake inhibitors (ritonavir) and glycolysis inhibitors (2-DG) were reported to induce in vitro cytotoxicity in CLL . ATM and TP53 have been shown to regulate central carbon metabolism. ATM deletion or 11q deletion in CLL cells led to increased insulin receptor expression and glucose uptake . Given this evidence, 11q-deleted CLL cells are more sensitive to glycolysis inhibition. Currently, there are clinical trials targeting the mitochondrial OXPHOS via metformin (NCT01750567) alone or in combination with GLUT4 (glucose transporter) via ritonavir (NCT02948283) . 5.2.3. Glutamine Metabolism Glutamine is the most abundant non-essential amino acid in the blood at a concentration of ~0.5-1 mM. Even though cells can synthesize glutamine via GLUL, many tumors are addicted to glutamine, especially KRAS and Myc-driven tumors . Glutamine is converted to glutamate by a rate-limiting step in glutamine catabolism via glutaminase (GLS1). Glutaminase is overexpressed in CLL samples, targeted in vitro by CB-839. Del11q CLL patients exhibit higher glutamine synthesis and metabolism than their negative counterparts. Del11q CLL samples are susceptible to glutaminase inhibitors indicating the pivotal role of glutamine metabolism in Del11q CLL . Glutamine is taken up by the cells via glutamine transporters such as SLC1A5, SLC38A1, and SLC38A2 . CLL cell proliferation is inhibited by targeting the glutamine transporters via the V9302 inhibitor, indicating the vital role of glutamine transport in CLL . To evaluate glutamine incorporation in CLL cells, there is an ongoing clinical trial (NCT04785989) in low disease burden CLL testing the glutamine incorporation via in vivo labeling. 5.2.4. Lipid Metabolism Metabolomic analysis of CLL samples shows a differential abundance of lipids compared to other metabolites in CLL, indicating the dependency of CLL on fatty acid metabolism . Lipoprotein lipase mRNA levels are highly expressed in CLL compared to normal B-cells . BCR stimulation further increases the LPL expression, indicating the function of BCR signaling in regulating fatty acid metabolism in CLL. Fatty acid synthesis and oxidation genes overexpressed in CLL are reported . Orlistat, a fatty acid synthesis inhibitor, is cytotoxic for CLL cells in vitro. CPT-1, a mitochondrial fatty acid transporter, is highly expressed in CLL . In line with this, CLL cells are sensitive to etomoxir, which targets fatty acid oxidation via CPT-1. Recent reports also suggest that ibrutinib affects fatty acid metabolism via inhibiting free fatty acid synthesis ; however, the mechanism is poorly understood. 6. Future Therapies 6.1. Combination Therapies 6.1.1. Targeted Therapy Coupled with Other Novel Chemotherapy Based on extensive CIT studies, in 2016 the European Medicine Agency accepted the use of uMRD, defined as <1 CLL cell per 10,000 leukocytes, as an intermediate endpoint and independent prognostic factor for PFS and OS . However, uMRD and CR are not commonly achieved with targeted agents such as ibrutinib, and several combinations are being tried to achieve increased efficacy. The HELIOS trial involved 578 R/R CLL patients who were randomly assigned to receive BR plus ibrutinib or BR plus a placebo. The results showed that adding ibrutinib led to a significantly higher ORR (83 vs. 68%), longer median PFS, and a higher rate of MRD negativity (13% vs. 5%). In total, ibrutinib improved the efficacy of the treatment in both naive patients with del(17p) and R/R patients compared to CIT treatment . However, some off-target ibrutinib may be responsible for its unique toxicities, such as atrial fibrillation and bleeding. Similarly, PI3Kd idelalisib has been tested with BR. The triple combination of PI3K with BR produced a significant increase in PFS (20.8 vs. 11.1 months) compared to BR alone; however, the triple combination was associated with adverse events such as increased infections, limiting its clinical use . Another trial tested the combination of ibrutinib with FC and obi in young treatment-naive patients with a favorable genetic profile (IGHV mutated and no TP53 aberrations). MRD was used to guide frontline therapy . The patients in the study received a quadruple combination for three courses, followed by either ibr plus obi for nine cycles or ibr plus obi for three cycles and ibrutinib for six cycles, based on their MRD status post-chemoimmunotherapy. Of the 28 patients who completed the 12-month treatment, all achieved undetectable MRD and a CR rate of 86%. Though these kinase inhibitors achieve impressive undetectable MRDs and CR, treating elderly patients with comorbidities becomes difficult due to increased toxicity rates. 6.1.2. Novel Agents in Combination with Anti-CD20 Antibodies The ALLIANCE randomized phase III trial compared ibrutinib and ibrutinib plus rituximab to bendamustine plus rituximab in older patients and found the efficacy to be almost identical for both the ibrutinib-containing arms . In the iLLUMINATE trial, Ibr-Obi was compared to Chl-Obi, with the former achieving superior PFS at 30 months (79% vs. 31% p < 0.0001). In the relapsed/refractory R/R setting, rituximab plus venetoclax, not ibrutinib, has produced improved MRD-negative rates, leading to its broad approval in the R/R setting . In the CLL2-BAG trial, a combination of bendamustine, venetoclax, and obinutuzumab was used, and an MRD-guided maintenance phase led to an 87% rate of MRD negativity . CLARITY is a phase II trial in which a combination of ibrutinib plus venetoclax was tested in 40 patients with R/R CLL. The results showed a CR rate of 58% (23/40) and no detectable MRD in peripheral blood after 12 months of treatment. In another trial of treatment-naive CLL patients (n = 80), the same combination was tested for 24 months. Ninety-six percent of patients treated with a combination of venetoclax and ibrutinib achieved a complete response (CR) after 12 months, and 69% had no undetectable MRD in the bone marrow . Combined treatment can achieve better efficacies; however, optimal treatment selection for each patient is challenging. 6.1.3. Targeted Therapy Coupled with an Immunomodulatory Agent Lenalidomide has been used as an immunomodulatory agent, which has been shown to induce T-cell activation resulting in CLL cell apoptosis. A combination of lenalidomide with idelalisib can potentially reduce lenalidomide-induced flare, as lenalidomide-induced cytokine release and immune activation are PI3K dependent . The most common toxicities are fatigue, thrombocytopenia, and neutropenia, noted in 83%, 78%, and 78% of patients, respectively . The ORR in different studies has been 32-54% with monotherapy and is better (66%) in combination with rituximab. In another trial, lenalidomide was evaluated for maintenance therapy post-chemotherapy in high-risk patients (NCT01556776). 6.2. Allogenic Transplant Allo-SCT has been considered an option for high-risk CLL patients in the CIT era and remains one of the potentially curative treatments for CLL. In this era of novel agents, allo-SCT remains an option for high-risk patients who progress after at least BTKi or venetoclax treatment. Due to substantial toxicities and morbidities, myeloablative-based treatment strategies were discontinued in CLL patients. However, recently large-scale prospective studies conducted with a median follow-up of 6 years have shown that RIC (reduced-intensity conditioning) allo-SCT can provide long-term disease control in 40% of patients and overcome TP53-based negative prognostication and refractoriness associated with fludarabine. In different studies, OS has been 50% with a PFS of ~40% . Non-relapse mortality (NRM) remains significant, affecting 15-25% of patients. In a prospective trial conducted on 55 patients, adding rituximab peri-transplant improved response rates compared to 157 historical control patients. The NRM rate in patients with no comorbidities was less than 12% . Limited data is available on allo-SCT efficacy; hence, more collaborative efforts are needed to understand the effectiveness of novel agents in this era. 6.3. Recent Targets and Trials in CLL Discussed in ASH2022 In the last year's ASH, several new targets were discussed. As patients treated with Ibr-Ven show resistance, these authors explored Protac-based BCL2/BCL-XL degrader PZ18753b to overcome Ibr-Ven resistance. The authors used OSU-CLL cells to generate BCL2 mutant (G101V, F104, and R107-110dup) cells. They showed rapid apoptosis with Protac-based degrader PZ118753b with efficient degradation of BCL-xL and partial degradation of BCL2, suggesting that venetoclax resistant cells retain dependency on BCL2/BCL-xL . As CDK9 is a master transcription regulator and a potential target in hematologic malignancies, the authors explored PRT2527 as a novel low nanomolar potent CDK9 inhibitor in B-ALL and CLL primary samples. Their studies showed that PRT2527 treatment has potent anti-leukemic activity in primary CLL and B-ALL samples evaluated ex vivo and two systemic models of B-ALL in vivo . Another group explored bromodomain-containing protein 9 (BRD9), a chromatin remodeling complex, as a potential target in CLL. The authors showed that BRD9i could improve the survival of orthotopic CLL xenograft mice with significantly reduced expression of targets related to the NRF2 pathway . CD70, a TNF family member, and its receptor CD27 were highly expressed in CLL cells compared to normal B-cells, suggesting immune deregulation in CLL. Using preclinical models, the authors showed that targeting CD70 with anti-CD70 antibodies in CLL altered BCR, CD40L, IL4, and TNFR signaling and delayed CLL disease progression, thus suggesting the use of anti-CD70 immunotherapy in combination therapies in CLL . As RT studies need more bona fide models, from our group, we reported the establishment of a new RT murine model driven by Mga KO and OXPHOS, where targeting the OXPHOS regulation along with CDK9i could improve RT mice survival . As RT has an unmet clinical need, the Phase Ib/II EPCORE CLL-1 trial explored Epcoritamab, a novel subcutaneously (SC) administered CD3/CD20 bispecific antibody in CD20+ RS LBCL (large B-cell lymphoma). Among the ten patients who received treatment, preliminary findings show that SC-Epcoritamab has encouraging single-agent activity with high overall and complete response rates. Most were seen at the first six-week assessment . This is an ongoing trial; updated data will be presented in future studies. 7. Conclusions The treatment of CLL has been transformed with the increasing availability of innovative agents that are now favored over traditional CIT in all settings. However, there are still many questions to be answered through well-designed studies. Areas of active research include RT metabolism, new targets, combinational treatments in CLL, splicing-related inhibitors, and the best treatment approach for high-risk patients. The cost-effectiveness of different treatments must also be taken into account. Although there are many obstacles, CLL therapy is still hopeful, with new agents recently approved by the FDA and more promising ones coming. Based on the ongoing recently completed accrual from several cooperative groups and international phase III trials (EA9161: NCT03701282; A041702: NCT03737981; CLL17: NCT04608318; MAJIC: NCT05057494), the proposed treatment can include treatment arms combining venetoclax and BTKi in the frontline setting. A viable strategy for relapsed or refractory CLL patients without identified BTK/PLCG2/BCL2 mutations may involve retreatment with either ibrutinib/non-covalent BTKi, venetoclax, or both. Lastly, advances in disease biology have revealed new targets and brought us closer to a cure for CLL by shifting from chemotherapy to more patient-friendly treatments. Author Contributions Conceptualization, writing, review and editing--P.I. and L.W. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest The authors declare no conflict of interest. Figure 1 BCR signaling and targeting agents for BCR signaling. Inhibitors of the BTK pathway (red line: left) and inhibitors of the PI3-kinase pathway (red line: right). SHP-1 modulates B-cell function by dephosphorylating various receptors, and downstream molecules (Src kinases-Pink) are recruited to the tyrosine phosphorylated motif of these receptors (dotted line). Created with Biorender.com. Figure 2 Summary of past, present, and emerging treatments in CLL. Created with Biorender.com. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000607 | Background: iron deficiency (ID) is frequent in older patients. Purpose: to evaluate the association between ID and survival in patients >= 75 years old with confirmed solid tumors. Methods: a retrospective monocentric study including patients between 2009 and 2018. ID, absolute ID (AID) and functional ID (FID) were defined according to the European Society for Medical Oncology (ESMO) criteria. Severe ID was defined by a ferritin level < 30 mg/L. Results: in total, 556 patients were included, the mean age was 82 (+-4.6) years, 56% were male, the most frequent cancer was colon cancer (19%, n = 104), and metastatic cancers were found in 38% (n = 211). Median follow-up time: 484 [190-1377] days. In anemic patients, ID and FID were independently associated with an increased risk of mortality (respectively, HR 1.51; p = 0.0065 and HR 1.73; p = 0.0007). In non-anemic patients, FID was independently associated with better survival (HR 0.65; p = 0.0495). Conclusion: in our study, ID was significantly associated with survival, and with better survival for patients without anemia. These results suggest that attention should be paid to the iron status in older patients with tumors and raise questions about the prognostic value of iron supplementation for iron-deficient patients without anemia. geriatric oncology iron deficiency geriatric medicine mortality This research received no external funding. pmc1. Introduction According to the World Health Organization (WHO), iron deficiency (ID) is the most common nutritional deficiency in the world. The consequences of ID in older patients are multiple, including anemia (in the United States, one out of five people aged 85 or over has anemia due to ID ), asthenia , immunodeficiency , and a reduction in cognitive functions . The diagnosis of ID can be difficult to establish as it is frequently associated with inflammatory conditions, such as cancer, which interfere with the iron status. Two types of ID have been described: absolute iron deficiency (AID), which is a lack of iron reserves, and functional iron deficiency (FID), which reflects the unavailability of iron. The effects of ID are particularly studied in cardiology settings. In chronic heart failure patients with reduced left ventricular ejection fraction (LVEF) (i.e., <50%), ID is associated with a significant increase in mortality, independently of anemia . In this population, intravenous iron supplementation has yielded a reduction in hospitalizations for heart failure and improvements in exercise tolerance and quality of life , but these interventional studies did not find a reduction in mortality. Since 2016, martial supplementation has been recommended by the European Society of Cardiology (ESC) in heart failure patients with reduced LVEF. The 2021 recommendations go further , recommending acute I.V. iron supplementation during hospitalization for acute heart failure . There are few data assessing the frequency of ID in the geriatric population, defined as aged 75 years or older. A recent large-scale European study with more than 12,000 participants from three cohorts in different countries with a median age of 59 years found AID and FID, respectively, in 60% and 64% of participants. FID and severe AID (ferritin < 30 mg/L) were significantly associated with all-cause mortality . Given that ID is a nutritional deficiency, and that malnutrition is a geriatric syndrome , ID may be more frequent in the older population with cancer . Studies have addressed ID prevalence in patients with cancer . In the CARENFER ONCO study, AID was reported in 20% and FID in 50% of patients with cancer . A recent study estimates that 64% of patients diagnosed with colorectal cancer have ID . However, few studies focus on older individuals. The aim of this study was to assess the relationship between ID, AID, FID, and survival in patients >= 75 years old with tumors. 2. Methods 2.1. Study Design and Population This was a retrospective, observational, single-center study that included patients >=75 years old, evaluated through a first geriatric oncology assessment at the University Hospital of Poitiers, France, between 1 January 2009 and 31 December 2018. 2.2. Data Collection The data were collected prospectively by the geriatric oncology physician during the assessment. A second-stage data collection of biological variables was completed retrospectively. The comprehensive geriatric assessment included several validated tests to evaluate geriatric domains : Activities of Daily Living (ADL), Instrumental Activities of Daily Living (IADL), Cumulative Illness Rating Scale (CIRS G), Mini Nutritional Assessment (MNA), Geriatric Depression Scale (GDS) with 15 questions, and Mini-Mental State Examination (MMSE). To assess the risk of falls, the following tests were administered: timed up and go, chair stand test, and the single-leg stance test. We have grouped cancers at risk of bleeding, which included urological, gynecological, and digestive cancers , and excluded breast and prostate cancers . We retrieved three comorbidities from medical records : anemia according to WHO criteria (i.e., hemoglobin concentration less than 12 g/dL in women and less than 13 g/dL in men), chronic heart failure, and chronic kidney disease, defined by Cockcroft and Gault clearance < 60 mL/minute. 2.3. Outcomes The primary outcome was to assess the relationship between ID, AID, FID, and survival in our population. We defined ID according to the 2018 European Society for Medical Oncology (ESMO) recommendations : ID is defined by a TSAT (transferrin saturation) < 20%, AID by a ferritin level < 100 mg/L and FID by the association of a ferritin level >= 100 mg/L, and a TSAT < 20%. Severe ID is defined by a blood ferritin value of less than 30 mg/L . The duration of follow-up was defined from the day of the consultation with the geriatric oncology physician to the date of last news. The cut-off date was 22 January 2022. This study was observational and was approved by the local ethics committee (2211250v0, reference: 23 January 2019), using existing data collected as part of the ANCRAGE cohort (Cancer and Age analysis). 2.4. Statistical Analysis Quantitative variables are described by means, standard deviations, minimums, and maximums for variables with a normal distribution, and by medians and interquartiles for non-normal distribution variables. Categorical variables are expressed as numbers and percentages. Survival analyses were performed by the Log-rank test with a graphical representation according to the Kaplan-Meier method, and by the Cox proportional hazards model test. We used a Chi2 test to compare qualitative variables. The comparison of survival events was expressed as the hazard ratio (HR) and 95% confidence interval (95% CI). Variables significantly associated (p < 0.05) in univariate analysis were retained for multivariate analysis, with a stepwise regression. We carried out a descriptive and then a statistical analysis using Statview software (SAS Institute, version 5.0). 3. Results 3.1. Patients During the study period, 556 patients were included in the study . Patients were predominantly male (n = 331, 56%), and the mean age was 82 years (Table 1). The prevalence of ID was 65%. ID was not associated with anemia in 36% of the cases. During the follow-up, 26% (n = 146) had at least one transfusion and 7% (n = 36) had at least one IV iron supplementation. Among these 36 patients, 4 had no anemia at the time of iron treatment. Lack of iron assessment was the main reason for non-inclusion; the median age of these patients (n = 792) was 81 (78-85) years and 48% were female (n = 350). The most common cancer was breast cancer: 19% (n = 142). In 672 of those patients with available data, 40% (n = 228) had anemia. 3.1.1. Iron Status In our population, the prevalence of ID was high (65%); 25% (n = 136) had ID without anemia. The most frequent type of ID was FID. The prevalence of FID was 16% (n = 86) for patients without anemia and 27% (n = 147) for patients with anemia. The prevalence of AID was 11% (n = 62) in the absence of anemia and 16% (n = 89) with anemia. The prevalence of severe ID was 7% (n = 41). 3.1.2. Oncological Characteristics The main tumor sites were the colon and rectum (19% (n = 104)), prostate (14% (n = 78)) and skin (12% (n = 64)). Cancer was metastatic at the time of the geriatric assessment in 38% (n = 211) of patients. Advanced stage cancer was not more frequent in patients with iron deficiency compared to those with normal iron status (n = 139 patients with metastatic cancer and ID (38,8%) vs. n = 72 patients with metastatic cancer with normal iron status (38,5%), p = 0.94). The most common treatment was chemotherapy in 45% (n = 249) of cases. 3.2. Survival Analysis 3.2.1. Survival According to General Characteristics Univariate analysis In univariate analysis (Table 2), survival was significantly associated with age (p = 0.036), anemia (p < 0.0001), chronic heart failure (p = 0.0314), chronic renal failure (p = 0.0035), cancers at risk of bleeding (p = 0.0150) and metastatic disease (p < 0.0001). Concerning geriatric domains, autonomy (ADL, p < 0.0001 and IADL, p < 0.0001), comorbidities (CIRS-G, p = 0.0154), nutritional status (MNA, p < 0.0001), mood status (GDS-15, p = 0.0318), cognition (MMSE, p = 0.0349) gait and balance disorders "timed up and go" (p = 0.0009), chair stand test (p = 0.0002), and single-leg stance test (p = 0.0075) were all significantly associated with survival. Multivariate analysis In multivariate analysis (Table 2), age (p = 0.0404), nutritional status (MNA, p < 0.0001), anemia (p = 0.0209), and metastases (p < 0.0001) remained significantly associated with survival. 3.2.2. Survival According to Iron Status Univariate analysis In univariate analysis, there was a significant association between survival and ID (p = 0.0106). Severe ID was significantly associated with survival (p = 0.0305). In the absence of anemia, ID, AID and FID were associated with better survival (respectively, ID (HR 0.70, p = 0.0024), AID (HR 0.72, p = 0.0393) and FID (HR 0.73, p = 0.0244)) (Table 3). Ferritin >= 100 mg/L and TSAT < 20% were significantly associated with survival . Multivariate analysis In anemic patients, ID (HR 1.51, p = 0.0065) and FID (HR 1.73, p = 0.0007) were associated with mortality. In non-anemic patients, only FID (HR 0.65; p = 0.0495) was associated with better survival. The nutritional status (MNA) and presence of metastases remained strongly associated with survival for both anemic and non-anemic patients (Table 4). 4. Discussion In our study, which included 556 older patients with cancer, the prevalence of ID was high (65%, classified as severe in 7%). In more than one third of patients, ID was not associated with anemia. Univariate and multivariate analyses have established a significant association between ID and survival, with better survival in non-anemic ID and increased mortality in subjects with both anemia and ID. The most frequent types of primary cancer in our population were colorectal, prostate and skin, which differ from French national statistics compiled in 2017 . The low number of patients with breast cancer in our study may be due to fewer referrals for this pathology by local gynecological surgeons and oncologists to the geriatrician. Women were likewise underrepresented in our population (44%). Cancers identified as being at risk of bleeding (i.e., urogynecological and digestive cancers, excluding breast and prostate cancers) were frequent (52%), but no association was found with mortality. As risk of bleeding may be linked to stage progression, it would have been interesting to analyze this data in our study. Nutritional status, assessed by the MNA score, was associated with better survival (HR 0.92 for each MNA point, p <= 0.0001), which is consistent with the literature . A significant association was found between chronic heart failure and mortality in our population. This result corroborates those of a recent study evaluating the risk of developing cancer at 10 years using statistical models in patients with heart failure . Data from the literature underline the frequent coexistence of cancer and heart failure, two conditions that have a major impact on life expectancy for older patients. Anemia is a known predictor of mortality in cancer patients and was significantly associated with mortality (HR 1.48; p = 0.0082), independently of oncological characteristics (type of primary cancer, cancers at risk of bleeding, and presence of metastasis). Since non-anemic ID is an independent mortality factor in chronic heart failure patients with altered LVEF, we wondered whether this was validated in patients with cancer. Contrary to our initial hypothesis, ID was not systematically associated with an increased rate of mortality. Our study demonstrates the beneficial effect of ID in older patients with cancer without anemia. Cellular and animal models have shown the effects of iron on tumor proliferation, through several signaling pathways and metabolic regulations . Excess iron may be related to an increased risk of solid cancers such as colon, liver, stomach, kidney and lung cancers . A pathophysiological explanation can be advanced, including several pathways. Cancer cells use iron as a growth factor . Tumor cells have a dysregulated iron metabolism and overexpress the transferrin receptor on their surface to increase the cellular uptake of iron. The tumor suppressor gene p53 induces cell cycle arrest, by decreasing intracellular free iron and inducing iron storage, via decreased expression of transferrin receptors to the cell membrane . Iron also contributes to the maintenance of cancer stem cells . Increased iron dependence has been reported in cancer cells and cancer stem cells in breast, ovarian, and prostate cancer cell models . Iron is also involved in several epigenetic and immunity processes related to tumor initiation and progression. The induction of ferroptosis could be an oncological therapeutic lead . Intracellular free iron leads to the synthesis of reactive oxygen species (ROS) via the Fenton reaction . It produces oxidative compounds, which can damage DNA, and therefore contributes to the instability of the genome . Hepcidin is a peptide synthesized by the liver. It regulates the concentration of iron in the blood. Its level is also dysregulated in cancer. Its secretion is stimulated in cases of inflammation, via interleukin 6 (IL6). It leads to the sequestration of iron in macrophages and enterocytes, resulting in a decreased level of serum iron. A study on breast cancer found that women with aggressive or chemo-resistant tumor phenotypes had higher levels of hepcidin . Cancer cells also have the ability to synthesize their own hepcidin . High hepcidin levels also lead to increased iron in cancer cells and activate Wnt and NF-kB signaling pathways, which are known to be linked to tumor progression . Iron chelators have an antiproliferative effect via the inhibition of ribonucleotide reductase. This results in cell cycle arrest but also inhibits several cell signaling pathways involved in tumor progression . Multiple in vitro studies evaluating the effects of iron deprivation observed cell cycle inhibition and apoptosis in colon, liver, and breast cancer cells . Two iron chelators DFO (deferoxamine) and DFX (deferasirox) have demonstrated their efficacy alone or concomitantly with adjuvant chemotherapy in mice with gastric , esogastric, pancreatic, liver, and mammary cancer xenografts. Several clinical trials are underway to validate the therapeutic effect of iron chelators, particularly in breast cancer . Other therapeutic strategies targeting iron metabolism are being evaluated, including monoclonal antibodies . To our knowledge, this is the first study to assess the relationship between ID, AID, FID and survival in older patients with cancer. Studies have demonstrated that ID, with or without anemia, was associated with a worse prognosis and mortality, but did not take into account the type of ID . A recent study demonstrated that FID seems to be independently associated with lymphatic invasion in colon cancers, with authors suggesting a potential relationship between FID and aggressive tumor characteristics . The mean age of our population was 82 years, representative of the French sample of older patients with cancer. Furthermore, our study was based on a prospective cohort with consecutive patient recruitment. The retrospective collection of iron biological data was carried out blinded to the geriatric assessment and vital status. Our study is also valuable with prolonged patient follow-up (median duration of follow-up 484 days) and a large proportion of advanced disease, as more than one third of our population had a metastatic disease at the time of geriatric assessment. This study has several biases. In terms of design, it is a retrospective and single-center study. There may be selection bias as patients referred to the geriatric oncology physician were previously screened by the oncologist, surgeon, or specialist according to a screening tool for geriatric frailty, as recommended by scientific societies, which excludes patients judged fit enough to undergo a typical cancer treatment plan . Referral by the specialist was not systematic and was based on local habits. Thereby, our results cannot be extrapolated to the overall population of aged patients with cancer. The results relied on a single biological assessment. Major biological variability is known for two biological parameters used to characterize ID, ferritin and TSAT. TSAT undergoes significant nychthemeral fluctuations ranging from 17 to 70% , reaching a maximum level in the morning. The level of ferritin may vary with inflammatory events. Iron status may also be closely correlated with bleeding events. Therefore, ferritin and TSAT could be poor markers for ID . In addition, many patients were excluded due to the lack of biological data available for the iron assessment. These individuals were more frequently diagnosed with breast cancer and presented a lower rate of anemia. We wonder whether those subjects who did not have an iron status evaluation had a geriatric profile comparable to our population. For these patients, physicians may have limited the biological tests and focused on palliative care. Given the current state of knowledge on the role of iron in oncology, it would be interesting to carry out a large prospective multicentric interventional randomized clinical trial to confirm those results and determine whether there is an interest in proposing intravenous iron supplementation in an iron-deficient patient with anemia or iron deficiency in the absence of anemia . 5. Conclusions We identified a significant association between ID and survival in older patients with cancer. The retrospective nature of our study did not allow us to establish the main judgement criteria, which are, nevertheless, essential in geriatric oncology, i.e., the evaluation of the quality of life and resistance to effort. These results need confirmation in a large-scale prospective multicenter study that includes a large number of patients with various tumor sites and diverse geriatric profiles. Acknowledgments The authors thank Jeffrey Arsham and Katherine Ryan for revising the manuscript. Author Contributions Conceptualization, S.V., A.J. and E.L.; methodology, E.L.; software, J.T. and E.L.; validation, E.L.; formal analysis, J.T. and E.L.; investigation, S.V., A.J. and E.L.; data procurement, J.T. and E.L.; original draft preparation, J.T. and E.L.; review and editing, V.R., P.-J.S., M.A., A.F., V.A., A.J., Y.M., S.V., J.T. and E.L.; visualization and supervision, E.L. and M.P. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was approved by the local ethics committee (reference 2211250v0, 23 January 2019). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are available on request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Flow chart. Figure 2 Kaplan-Meier curves according to transferrin saturation < 20% (A) and ferritin < 100 mg/L (B). cancers-15-01533-t001_Table 1 Table 1 Patient characteristics at baseline. Variables N, (%) Gender, n = 556 (female) 245 (44) Age, n = 556 (years), mean (S.D) 82 (+-4.6) Duration follow-up, n = 556, days (median) [interquartiles] 484 [190-1377] Geriatric assessment ADL, n = 555, mean (S.D) 5 (+-1) IADL, n = 342, mean (S.D) 5 (+-3.5) Place of living, n = 539 Community 472 (88) Nursing home 40 (7) Residential care home 10 (2) Service home 2 (0.4) Long-term care unit 1 (0.2) CIRS-G, n = 556, mean (S.D) 8 (+-5) MNA, n = 553, mean (S.D) 21 (+-6) GDS 15, n = 425, mean (S.D) 4 (+-3) MMSE, n = 536, mean (S.D) 26 (+-5) Timed up and go, n = 533 (pathological) 151 (28) Chair stand test, n = 459 (pathological) 170 (37) Single-leg stance test, n = 492 (pathological) 256 (52) Comorbidities Chronic heart failure, n = 556 159 (29) Chronic renal disease, n = 528 323 (61) Anemia, n = 545 302 (55) Iron measurements Ferritin (mg/L), n = 545, mean (S.D) 381 (+-366) TSAT, n = 545 18 (+-13) ADL = Activities of Daily Living ADL, IADL = Instrumental Activities of Daily Living, CIRS-G = Cumulative Illness Rating Scale, MNA = Mini-Nutritional Assessment, GDS = Geriatric Depression Scale, MMSE = Mini-Mental State Examination, TSAT = Transferrin saturation, S.D = Standard Deviation. cancers-15-01533-t002_Table 2 Table 2 General characteristics related to overall survival. Variables Univariate Analysis Multivariate Analysis HR [95% CI] p * HR [95% CI] p * Gender, female 0.83 [0.69-1.01] 0.0683 - - - Age, years 1.02 [1.00-1.04] 0.0306 1.03 [1.00-1.06] 0.0404 ADL (+1 point) 0.80 [0.74-0.86] <0.0001 0.89 [0.75-1.06] 0.1969 IADL (+1 point) 0.90 [0.86-0.95] <0.0001 0.97 [0.90-1.05] 0.5211 CIRS-G (+1 point) 1.02 [1.00-1.05] 0.0154 1.01 [0.98-1.05] 0.2798 MNA (+1 point) 0.89 [0.88-0.91] <0.0001 0.92 [0.89-0.96] <0.0001 GDS 15 (+1 point) 1.04 [1.00-1.07] 0.0318 1.01 [0.97-1.07] 0.4479 MMSE (+1 point) 0.97 [0.95-0.99] 0.0349 1.02 [0.99-1.06] 0.1404 Chronic renal failure 1.35 [1.11-1.65] 0.0027 1.04 [0.76-1.42] 0.7936 Chronic heart failure 1.24 [1.02-1.52] 0.0314 1.35 [1.00-1.84] 0.0496 Timed up and Go, pathological 1.42 [1.15-1.75] 0.0009 1.05 [0.64-1.73] 0.8323 Chair stand test, pathological 1.50 [1.21-1.86] 0.0002 0.95 [0.58-1.53] 0.8385 Single-leg stance test, pathological 1.31 [1.07-1.60] 0.0075 1.23 [0.88-1.70] 0.2112 Cancers at risk of bleeding ** 1.26 [1.04-1.52] 0.0150 1.28 [0.95-1.72] 0.1038 Metastases 1.86 [1.54-2.25] <0.0001 2.22 [1.65-2.98] <0.0001 Anemia 1.71 [1.41-2.08] <0.0001 1.41 [1.05-1.90] 0.0209 p * value for Cox model; Bold = significant p value at the threshold of 5%. ADL = Activities of Daily Living ADL, IADL = Instrumental Activities of Daily Living, CIRS-G = Cumulative Illness Rating Scale, MNA = Mini Nutritional Assessment, GDS = Geriatric Depression Scale, MMSE = Mini-Mental State Examination, TSAT = Transferrin saturation, I.V = Intravenous. ** Urogynecological cancers (excluding breast and prostate cancer) and digestive cancers. cancers-15-01533-t003_Table 3 Table 3 Iron status in relation to overall survival in univariate analysis. Univariate Analysis Variables Independently of Anemia Without Anemia With Anemia HR [95% CI] p * HR [95% CI] p * HR [95% CI] p * Severe ID 0.66 [0.45-0.96] 0.0305 0.70 [0.40-1.22] 0.2107 0.65 [0.40-1.07] 0.938 ID 1.29 [1.06-1.58] 0.0106 0.70 [0.56-0.88] 0.0024 1.69 [1.40-2.05] <0.0001 AID 0.77 [0.63-0.96] 0.0200 0.72 [0.53-0.98] 0.0393 0.87 [0.67-1.13] 0.3126 FID 1.44 [1.19-1.74] 0.0001 0.73 [0.56-0.96] 0.0244 2.07 [1.96-2.55] <0.0001 p * value for Cox model; Bold = significant p value at the threshold of 5%. ID = Iron Deficiency, AID = Absolute Iron Deficiency, FID = Functional Iron Deficiency. cancers-15-01533-t004_Table 4 Table 4 Iron status in relation to overall survival in multivariate analysis. Multivariate Analysis Variables Without Anemia With Anemia HR [95% CI] p * HR [95% CI] p * HR [95% CI] p * FID 0.65 [0.43-0.99] 0.0495 ID 1.51 [1.12-2.05] 0.0065 FID 1.73 [1.26-2.38] 0.0007 Age 1.03 [0.99-1.06] 0.0588 Age 1.03 [0.57-1.51] 0.0495 Age 1.03 [0.99-1.06] 0.0759 ADL 0.89 [0.75-1.06] 0.2132 ADL 0.90 [0.76-1.07] 0.2714 ADL 0.92 [0.77-1.09] 0.3457 IADL 0.96 [0.89-1.04] 0.3834 IADL 0.96 [0.89-1.03] 0.3147 IADL 0.95 [0.88-1.03] 0.2533 CIRS-G 1.01 [0.98-1.05] 0.3033 CIRS-G 1.02 [0.98-1.05] 0.2311 CIRS-G 1.02 [0.99-1.06] 0.1281 MNA 0.92 [0.88-0.95] <0.0001 MNA 0.93 [0.89-0.96] 0.0004 MNA 0.93 [0.89-0.97] 0.0006 GDS 15 1.02 [0.97-1.08] 0.3042 GDS 15 1.02 [0.97-1.07] 0.3756 GDS 15 1.02 [0.97-1.07] 0.3631 MMSE 1.03 [0.99-1.06] 0.0864 MMSE 1.02 [0.99-1.06] 0.1194 MMSE 1.02 [0.99-1.06] 0.1532 Timed up and go 1.07 [0.65-1.77] 0.7665 Timed Up and Go 1.06 [0.64-1.76] 0.8072 Timed Up and Go 1.01 [0.61-1.67] 0.9582 Chair stand test 0.93 [0.58-1.51] 0.7970 Chair stand test 0.93 [0.57-1.51] 0.7865 Chair stand test 1.00 [0.62-1.62] 0.9769 Single-leg stance test 1.23 [0.88-1.71] 0.2147 Single-leg stance test 1.22 [0.79-1.69] 0.2348 Single-leg stance test 1.21 [0.87-1.68] 0.2528 Chronic renal failure 1.13 [0.83-1.53] 0.4231 Chronic renal failure 1.08 [0.79-1.46] 0.6083 Chronic renal failure 1.08 [0.80-1.47] 0.5995 Chronic heart failure 1.44 [1.07-1.95] 0.0162 Chronic heart failure 1.33 [0.98-1.81] 0.0664 Chronic heart failure 1.41 [1.04-1.90] 0.0243 Cancers at risk of bleeding ** 1.34 [0.99-1.80] 0.0515 Cancers at risk of bleeding ** 1.21 [0.89-1.64] 0.2245 Cancers at risk of bleeding ** 1.29 [0.96-1.74] 0.0877 Metastases 2.25 [1.67-3.02] <0.0001 Metastases 2.16 [1.61-2.91] <0.0001 Metastases 2.12 [1.58-2.85] <0.0001 p * value for Cox model; Bold = significant p value at the threshold of 5%. ADL = Activities of Daily Living, IADL = instrumental Activities of Daily Living, CIRS-G = Cumulative Illness Rating Scale, MNA = Mini Nutritional Assessment, GDS = Geriatric Depression Scale, MMSE = Mini-Mental State Examination, TSAT = transferrin Saturation, FID = functional iron deficiency, ID = iron deficiency, ** urogynecological cancers (excluding breast and prostate cancer) and digestive cancers. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000608 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12050936 foods-12-00936 Article Single-View Measurement Method for Egg Size Based on Small-Batch Images Liu Chengkang Conceptualization Methodology Software Validation Formal analysis Investigation Data curation Writing - original draft Writing - review & editing 1 Wang Qiaohua Conceptualization Validation Resources Writing - review & editing Visualization Supervision Project administration Funding acquisition 123* Ma Meihu Conceptualization Writing - review & editing 4 Zhu Zhihui Writing - review & editing Visualization 1 Lin Weiguo Writing - review & editing Funding acquisition 1 Liu Shiwei Writing - review & editing Visualization 1 Fan Wei Writing - review & editing Visualization 1 Santini Antonello Academic Editor 1 College of Engineering, Huazhong Agricultural University, Wuhan 430070, China 2 Ministry of Agriculture Key Laboratory of Agricultural Equipment in the Middle and Lower Reaches of the Yangtze River, Wuhan 430070, China 3 National Research and Development Center for Egg Processing, Huazhong Agricultural University, Wuhan 430070, China 4 College of Food Science and Technology, Huazhong Agricultural University, Wuhan 430070, China * Correspondence: [email protected]; Tel.: +86-1870-2768-307 22 2 2023 3 2023 12 5 93624 11 2022 07 2 2023 08 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Egg size is a crucial indicator for consumer evaluation and quality grading. The main goal of this study is to measure eggs' major and minor axes based on deep learning and single-view metrology. In this paper, we designed an egg-carrying component to obtain the actual outline of eggs. The Segformer algorithm was used to segment egg images in small batches. This study proposes a single-view measurement method suitable for eggs. Experimental results verified that the Segformer could obtain high segmentation accuracy for egg images in small batches. The mean intersection over union of the segmentation model was 96.15%, and the mean pixel accuracy was 97.17%. The R-squared was 0.969 (for the long axis) and 0.926 (for the short axis), obtained through the egg single-view measurement method proposed in this paper. actual outline of eggs deep learning egg-carrying component image segmentation single-view metrology the National Natural Science Foundation of China31871863422 32072302 the Fundamental Research Funds for the Central Universities2662022GXYJ004 This research was funded by the National Natural Science Foundation of China (No. 31871863422 and No. 32072302) and the Fundamental Research Funds for the Central Universities (No. 2662022GXYJ004). pmc1. Introduction Eggs are an inexpensive and nutritious food, and egg quality grading is very important in commercialization. Furthermore, the measurement accuracy of egg size is essential in quality testing and affects the hatching rate of breeder eggs . However, manual labor cannot consistently perform the detection perfectly because of eye strain. A machine vision approach would not suffer from fatigue, but failure to capture an actual outline view results in measurement errors due to the ellipsoidal appearance of eggs, which affects the determination of the eggs' appearance quality. Therefore, it is crucial to study how to obtain the exact size of eggs through machine vision. Many scholars used machine vision for egg size detection, volume prediction, and freshness grading. Among them, some scholars used external parameters such as egg size to achieve the primary grading. Omid developed an intelligent system based on machine vision to achieve egg grading using defects and size . Vasileva analyzed the external parameters of eggs through machine vision to automate the primary grading stage . Some scholars achieved egg detection through quantifying volume and appearance parameters, where volume is estimated using images. Soltani proposed a new method for egg volume prediction and implemented an efficient image processing algorithm for egg size measurement . Quilloy developed an automatic single-line tabletop egg sorter that integrated machine learning and electromechanical principles; the sorter used egg projection area and weight estimation, which is more efficient than manual sorting . Harnsoongnoen performed egg grading and freshness assessment based on machine vision and load cells; they used the machine vision system to estimate the eggs' volume, and the results indicate that the system has potential in the poultry industry . In addition, machine vision has been applied to breeder eggs. Haixia detected egg size through machine vision and found that the detection values highly correlated with the measured values . Xiulian achieved the rejection of sperm-less eggs through machine-vision-extracted regional invariance parameters such as the long-to-short axis ratio . Weizhong screened breeder eggs based on machine vision, using egg size as indicators, and the screening results were highly consistent with manual results . Okinda used an infrared depth sensor to estimate the volume of eggs on the production line and developed a volume estimation model using feature variables . Nasiria achieved the classification of unwashed eggs using machine vision and deep learning methods . Aboonajmi discussed the importance of using non-destructive inspection methods such as machine vision in an in-line sorting system for the internal and external quality of eggs . Valeriy proposed a mathematical formula for describing the shape of the egg based on four parameters, namely egg length, maximum breadth, the shift of the vertical axis, and the diameter at one-quarter of the egg length . Liao proposed a clear definition of the standard egg shape curve in an analytical geometric sense and also discovered the equation for the segmental function of the curve . All the literature above used machine vision to achieve egg quality detection, but the two following issues are overlooked in acquiring egg images: first, the imaging plane's height differs for different egg sizes; second, due to eggs' ellipsoidal appearance and the component's structure, the imaging outline may not match the actual outline of the egg. Single-view metrology is convenient and effective for studying object dimensioning through images. Some scholars studied theoretical and algorithmic geometric information extraction for single uncalibrated images . These works focused on new single-view metrology, including new methods for obtaining vanishing points , off-surface point transfer , single-view metrology after lens distortion correction , and calculating camera matrix parameters using vanishing points . Other scholars studied applications for single-view metrology, including road horizontal and vertical measurements , distance estimation in the car driving , Google Street View building height estimation , and real-time target tracking and manipulation combined with robotics . Single-view metrology requires specific geometric information based on known scenes. The measurement accuracy depends heavily on the accuracy of image preprocessing, such as edge detection, line fitting, and vanishing point determination . This paper proposes a measurement method for eggs based on the above single-view metrology, which solves the problem of current machine vision methods being unable to capture the actual outline of eggs. In addition, this paper obtained an egg image segmentation model by combining small-batch egg samples with deep learning. Small-batch eggs refers to using a small number of egg samples to obtain the egg segmentation model, which can obtain the segmented image of a single egg; this model combined with single-view metrology can achieve egg size measurement, laying the foundation in egg commercialization lines. The goal of this study is to achieve size measurement of small-batch eggs through single-view metrology and deep learning. The specific studies are (1) designing the egg-carrying component to realize the egg's outline acquisition in a single view, (2) achieving image segmentation for eggs in small batches through deep learning, and (3) designing the geometric environment for vanishing point acquisition based on single-view metrology. Moreover, we completed the mathematical measurement method applicable to the egg's major and minor axes. 2. Materials and Methods 2.1. Experimental Materials A total of 30 eggs under the Sundaily Farm brand were purchased (Wuhan, China), and the mean and median weights of all samples measured using an electronic scale were 47.65 g and 46.71 g, respectively. The eggs' major axis and minor axis were measured using vernier calipers. In addition, egg sizes were measured five times in the morning and afternoon using vernier calipers to reduce random errors. After the vernier caliper measurement, a camera (IDS company, Saarbrucken, Germany, CCD industrial camera, model UI-2210RE-C-HQ, resolution 640 x 480; lens model M0814-MP2) and a surface lamp (Guiguang Instrument company, Guilin, China, DY30-1C, AC220V, DC24V, 30 W) were used to collect the egg images . As in Figure 2a,b, the egg deflected to different degrees in different views when placed on a flat surface. This placement causes the problem that the egg's major and minor axes cannot be measured using a single view. The egg-carrying component designed in this paper ensured that the major axis was parallel to the front view of the egg. Thus, the egg-carrying component designed in this paper can help in obtaining the actual outline of the egg from a single view. When the egg slides down from the top of the component, it adjusts its posture using gravity. The egg-bearing component was obtained via 3D printing using resin material; the component was 75 mm long, 64 mm wide, and 34 mm high. The principle is to use the tangency relationship between the egg's cross-sectional circles and the component's triangles boundary. Here, Table 1 shows the differences between this component and existing egg image capture components. 2.2. Single-View Measurement for Eggs (1) Single-View measurement principle The length, width, and height properties of three-dimensional objects cannot be measured from a single view alone, but the length ratio of two parallel objects can be obtained. Thus, if a reference object is added and the actual size is known, combined with the length ratio, the size of the measured object can be obtained. Criminisi described how the affine 3D geometry of a scene could be measured from a single perspective image . The proposed basic geometry of a single view is shown in Figure 3. The vanishing line l is the projection of the line at infinity of the reference plane in the image. The vanishing point v is the image of the point at infinity in the reference direction. Criminisi proposed interplanar distance measurement ; this method uses the cross-ratio with a known distance Zr (the distance between the planes p and p') to obtain the value of Z (the distance between the planes pr and pr'). First, the mathematical relation between the distance Zc and Z can be established according to the cross-ratio. Zc is the distance of the camera from the plane p. (1) Dx,cdx',c/dx,vdx',v=dX,CdX',C/dX,VdX',V X is a point on the plane p, X' is the parallel projection of the point X on the plane p', and d() is the distance between the two points. Since the vanishing point V is at infinity, dX,VdX',V=1, it follows that, (2) ZZ - Zc=dX,CdX',C=dx,cdx',vdx',cdx,v In order to compute the distance Z, it is necessary to determine Zc by the distance Zr. According to the cross-ratio, the mathematical relation between Zcr (the distance of the camera from the plane pr) and Zr (the distance between the planes pr and pr') can be obtained as follows, (3) ZrZcr=1 - dr1,vdr2,crdr2,vdr1,cr r1 and r2 are the points corresponding to R1 and R2 in the image, and cr is the intersection of the lines r1 and r2 with the fading line l. Where Zr is known, the distance Zcr can be determined. Further, based on the cross-ratio and the distance (Zc - Zcr) between the plane pr and p, the distance Zc can be determined as follows, (4) Zc=Zcrds1,csds1,vds2,csds2,v Further, the distance Z can be computed by combining Equation (4) with Equation (2). (2) Vanishing point determination and measurement principle of egg images In this paper, the vanishing point was determined using the multi-vanishing point determination method proposed by Nietoa ; the method is based on a robust version of the MLESAC (maximum likelihood estimation by sample and consensus) algorithm. Moreover, it is used similarly to the EM (expectation maximization algorithm), which allows a trade-off between accuracy and efficiency and enables real-time manipulation of video sequences. Nietoa provides the source code for download at (accessed on 5 July 2022) for the figures that were drawn in this paper. As shown in Figure 5, we prepared the cards with crossed parallel lines to obtain vanishing points. This paper used vanishing lines and points to measure the egg's major and minor axis based on the interplanar distance measurement principle proposed by Criminisi. First, as shown in Figure 6a, geometrizing the egg and the reference as a straight line, this paper assumed that AB is the egg of length Z and CD is the reference of length R, where the projection of CD on AB is FB. Next, as shown in Figure 6b, the parallel lines of the reference plane were used to obtain the vanishing points vl and vr. The vanishing lines were obtained by connecting the two vanishing points, and all the lines from the vanishing lines are parallel. Therefore, bd intersects the vanishing line at e, and ec then intersects ab to obtain the projection point f. Based on the cross-ratios, the mathematical relation between the geometrized egg and the reference can be obtained as follows, (5) d(B, F)d(B,A)/d(G,F)d(G,A)=d(b,f)d(b,a)/d(g,f)d(g,a) Since point G is at infinity, d(G,F)d(G,A)=1, Equation (5) can obtain the following equation. (6) RZ=db,fdg,adb,adg,f R is the reference height, a known quantity, so the length Z can be computed by combining the image processing results with Equation (6). 2.3. Egg Image Segmentation in Small Batches In order to achieve accurate and efficient image segmentation, this paper used Segformer to complete the egg image segmentation. Then, the egg size was measured based on single-view metrology. Segformer is a simple but powerful semantic segmentation method proposed by Xie; the method avoids the complex design commonly found in previous methods. Furthermore, it unifies the transformer encoder with a new hierarchy that does not require positional encoding and the lightweight multilayer perceptron (MLP) decoder. Thus, it achieves high efficiency and performance and shows zero-shot robustness . In this paper, the weights provided by Xie (pre-trained on ImageNet-1K) were used. Fifteen samples were randomly selected from thirty data samples as the test set to ensure independence. The remaining 15 samples were expanded to 75 for training by randomly flipping up and down, flipping left and right, and varying contrast and lightness. Moreover, the specific training parameters are shown in Table 2. Annotation was performed using the PixelAnnotationTool accessed on 15 June 2022) to obtain labels, and then trained and tested based on the code and usage provided by Enze accessed on 28 July 2022). 2.4. Single-View Measurement Framework and Evaluation Metrics The proposed measurement framework consists of three main steps to measure the egg's major and minor axis . The first step is to acquire a single-view image using the component that can be calibrated for the egg pose. The second step is to obtain segmentation images of eggs and references through a Segformer model trained with a few-shot. The third step is to obtain the vanishing point or the egg's major and minor axis through egg-size single-view metrology. In this paper, we used intersection over union (IoU) and pixel accuracy (PA) as the performance discriminators of the segmentation model. IoU = TP/(TP + FP + FN)(7) PA = (TP + TN)/(TP + TN + FP + FN)(8) Here, TP indicates that the model predicts a positive case and predicts correctly, FP indicates that the model predicts a positive case and predicts incorrectly, FN indicates that the model predicts a negative case and predicts incorrectly, and TN indicates that the model predicts a negative case and predicts correctly. The absolute error was used as an evaluation index to confirm the performance of the single-view measurement framework. Furthermore, R2 (R-squared) was used to assess the conformity degree between the single-view measured value and the actual value. In particular, to eliminate random errors in the measurement of actual values, each egg was measured by the same experimenter five times in the morning as well as the afternoon. Then, the average of the measured values was taken. 3. Results 3.1. Segmentation Result of Egg Images The mean intersection over union (MIoU) of the segmentation model was 95.35%, and MPA (pixel average accuracy) was 97.31%. IoU and PA for each category are shown in Table 3. Table 3 shows that the model has the best segmentation result for the background, with both IoU and PA reaching over 99%. The segmentation result of the egg is fine, with IoU and PA reaching 97% or more. The reference has the worst segmentation result, with a PA of 89.22% and Iou of 93.71%. It can be seen that the Segformer algorithm demonstrated strong segmentation performance. However, the small area in the image, the similar color to the background, and being occluded resulted in poor segmentation of the reference. Therefore, to reduce the measurement error caused by reference segmentation, the average of the reference coordinate values from multiple images was taken for the egg size calculation. 3.2. Vanishing Point Image Figure 9 shows the vanishing points obtained through the multiple vanishing point determination method (Section 2.2). Since the egg's ideal major and minor axes are perpendicular to different planes, the original image was rotated by 90 degrees to obtain the vanishing points of the right card. In addition, we cropped the images to prevent the two cards from interfering with each other in acquiring the vanishing points; we supplemented the vanishing point coordinate values in the measurement calculations. 3.3. Single-View Measurement Result of Egg Size Table 4 shows that the measurement errors of the major axis were all less than 1 mm, and the adj.R2 was 0.9725 , which was obtained from the regression analysis of the measured and actual values. For the minor axis, the measurement errors of four samples were greater than 1 mm, but they were all within 2 mm, and the adj.R2 was 0.8353 . In addition, this paper used an F-test for the regression models, and the results show that all models have a significant linear relationship (Table 5). The measurement results of the minor axis are undesirable. The possible reasons are as follows; first, the segmentation of the reference object could be more satisfactory, where the length and width determine the measurements of the major and minor axes of the egg. In addition, Figure 8 shows that there is more interference from the same color background towards the short side of the reference object, so it leads to a greater calculation error of the minor axis. Second, observing Figure 8, the contact position of the egg with the egg-bearing component is very dark, which also interferes with the segmentation of the short axis direction of the egg, which affects the calculation of the minor axis. 4. Discussion In this paper, we designed the egg-bearing component to acquire the egg's actual outline. This component uses the tangency of the egg's cross-sectional circle to the triangular boundary; we found that this component provides an excellent actual outline of eggs by comparing the actual and measured values of the egg's major and minor axis. However, due to the ellipsoidal shape of the egg, the component can only obtain the actual outline of the front view and not the actual outline of the other views. Most of the current studies into egg detection via image have been performed by processing and measuring directly through the image. Additionally, this research did not consider the consistency of the egg imaging plane. However, it is vital to ensure that the image plane of the egg is consistent. The literature also suggests that whether the major axis is parallel to the horizontal plane affects the measurement of eggs, and the literature agrees that a suitable egg roller can improve the accuracy of egg measurements. Compared with these studies , the method presented in this study can disregard the consistency of the egg imaging plane by applying single-view metrology. Using the vanishing point in the perspective view and the reference size, the difference in the imaging plane does not affect the measurement for different sizes of eggs. However, the measurement accuracy of the egg sizes depends on image processing. In this paper, an image segmentation model was developed using Segformer. Segformer can build a robust image segmentation model with few samples, but the image background is the same color as the reference; we obtain a worse reference segmentation. Therefore, we will use different colored backgrounds or references in future studies. This paper verified the feasibility of single-view metering eggs using an egg-carrying component and single-view metrology. Future research should explore a dynamic egg-carrying component. The dynamic online measurement component can provide a reliable and efficient basis for determining the eggs' appearance quality in the commercialization process. In addition, the component can provide researchers with a dynamic acquisition method for single-view dimensional metrology. Taken together, our results show that the actual outline of eggs can be obtained using the tangency of the egg cross-section circle to the triangular boundary; single-view metrology can solve the problem of inaccurate measurements caused by different egg imaging planes. However, this paper only verified the method's feasibility using the static egg-bearing component. Moreover, further improvements in the image segmentation method are yet to be made to improve the measurement accuracy of egg size. 5. Conclusions This paper verified the feasibility of single-view measurement of the egg's major and minor axes; this paper proposed the method of single-view egg major and minor axis determination by combining with single-view metrology. Due to the strong zero-shot robustness and high efficiency of Segformer, this paper achieved good image segmentation using a small batch of images. Moreover, we obtained excellent measurement results of the eggs' major and minor axes using the projection relation between parallel lines and the vanishing line property. In order to simplify model building and validation, this work chose a single type of egg for experiments, and subsequent exploration of universal models for multiple types will be undertaken. Future research will focus on multiple types of egg measurements and online dynamic egg measurements. Acknowledgments The authors thank laboratory colleagues for their help. Author Contributions Conceptualization, Q.W. and M.M.; methodology, C.L.; software, C.L.; validation, Q.W. and C.L.; formal analysis, C.L.; investigation, C.L.; resources, Q.W.; data curation, C.L.; writing--original draft preparation, C.L.; writing--review and editing, C.L., Q.W., M.M., Z.Z., W.L., S.L. and W.F.; visualization, Q.W., S.L., Z.Z. and W.F.; supervision, Q.W.; project administration, Q.W.; funding acquisition, Q.W. and W.L. All authors have read and agreed to the published version of the manuscript. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Schematic diagram of egg image acquisition. Figure 2 Different view of the egg (possible poses placed naturally on a flat surface, (a) front view; (b) top view. The stance using component, (c) top view; (d) front view). Figure 3 Basic geometry of single-view metrology. Figure 4 The geometry of calculating plane distance using cross-ratio, (a) in the world, (b) in the image. Figure 5 Cross-parallel reference card, (a) front view, (b) perspective view. Figure 6 The geometry of obtaining egg size based on cross-ratio and single-view metrology, (a) in the world, (b) in the image. Figure 7 The flow chart of egg single-view metrology. Figure 8 Results of image segmentation. (a) Original image; (b) image segmentation result; (c) mask image. Figure 9 The location of the vanishing points. (a) Original image; (b) rotated image. Figure 10 Comparison of measured and actual values of eggs' size. (a) Eggs' major axis; (b) eggs' minor axis. foods-12-00936-t001_Table 1 Table 1 Comparison of the egg-carrying component designed in this paper with common egg-carrying methods. Egg-Carrying Method Design Innovation Advantages Mahmoud Soltani The tangency between the cross-sectional circle of the fresh egg and the triangle of the component is used to obtain the true profile of the fresh egg. The design idea of this component can be used for the adjustment of the fresh egg posture of the production line. The components designed in this paper allow for a single view to obtain the real outline of the egg and for the adjustment of the egg's attitude using gravity. Erwin Harnsoongnoen This paper foods-12-00936-t002_Table 2 Table 2 Parameters used in training model. Parameter or Resource Value or Model Checkpoint_file pspnet_r50-d8_512x1024_40k_cityscapes_20200605_003338-2966598c.pth Optimizer SGD Batch size 8 Initial learning rate 0.01 Train, validate, and test images 60, 15 and 15 images Total iterations 5000 Backend Pytorch 1.7 Python 3.6 Operation system Windows 10 GPU NVIDIA RTX 2080 Ti Class ('background', 'egg', 'ref') Palette (255, 0, 0), (250, 250, 0), (255,255, 255) foods-12-00936-t003_Table 3 Table 3 Iou accuracy and PA of the segmentation model. Class IoU (%) PA (%) Background 99.72 99.88 Egg 97.11 98.34 Reference 89.22 93.71 foods-12-00936-t004_Table 4 Table 4 Eggs' size measurement results. Serial Number Minor Axis Size (mm) Major Axis Size (mm) Measured Value Actual Value Error Measured Value Actual Value Error 1 39.329 40.866 1.537 51.136 50.774 0.362 2 40.668 41.036 0.368 57.040 56.713 0.327 3 40.770 41.939 1.169 53.405 52.856 0.549 4 40.742 41.661 0.919 54.140 53.388 0.752 5 40.617 41.828 1.211 53.291 52.478 0.813 6 42.825 42.902 0.077 54.966 54.033 0.933 7 40.445 41.302 0.857 55.053 54.321 0.732 8 41.599 42.373 0.774 54.111 53.712 0.400 9 41.746 42.147 0.401 57.377 57.132 0.245 10 42.517 43.253 0.736 56.433 56.407 0.026 11 42.046 43.360 1.314 56.237 56.376 0.139 12 40.270 41.115 0.845 55.663 55.566 0.097 13 41.385 42.198 0.813 53.052 52.312 0.740 14 41.119 42.011 0.892 54.079 54.024 0.055 15 42.264 42.862 0.598 54.105 53.613 0.492 foods-12-00936-t005_Table 5 Table 5 F-test results for the regression models. Regression Models Significance Level F-Value Major axis 0.01 552.999 P{F >= 98.503} = 0.01 Minor axis 72.320 P{F >= 34.116} = 0.01 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Deng H. Liu Y.M. Wen Y.X. Wang S.C. Research on method of detecting egg size by machine vision J. Huazhong Agric. Univ. 2006 25 452 454 10.13300/j.cnki.hnlkxb.2006.04.027 2. Zhou W.Z. Feng X.H. Shape identification of fertile eggs by using computer vision Trans. CSAE 2000 16 126 130 10.3321/j.issn:1002-6819.2000.06.032 3. Omid M. Soltani M. Dehrouyeh M.H. Mohtasebi S.S. Ahmadi H. An expert egg grading system based on machine vision and artificial intelligence techniques J. Food Eng. 2013 118 70 77 10.1016/j.jfoodeng.2013.03.019 4. Vasileva A.V. Gorbunova E.V. Vasilev A.S. Peretyagin V.S. Chertov A.N. Korotaev V.V. Assessing exterior egg quality indicators using machine vision Br. Poult. Sci. 2018 59 636 645 10.1080/00071668.2018.1523535 30221543 5. Soltani M. Omid M. Alimardani R. Egg volume prediction using machine vision technique based on pappus theorem and artificial neural network J. Food Sci. Technol. 2015 52 3065 3071 10.1007/s13197-014-1350-6 25892810 6. Erwin P.Q. Delfin C.S. Pepito M.B. Single-line automated sorter using mechatronics and machine vision system for Philippine table eggs Afr. J. Agric. Res. 2018 13 918 926 10.5897/AJAR2018.13113 7. Harnsoongnoen S. Jaroensuk N. The grades and freshness assessment of eggs based on density detection using machine vision and weighing sensor Sci. Rep. 2021 11 16640 10.1038/s41598-021-96140-x 34404854 8. Ma X.L. Yi S.J. Unfertilized eggs verification system before hatching based on embedded system and machine vision Trans. Chin. Soc. Agric. Mach. 2011 42 187 192 9. Okinda C. Sun Y. Nyalala I. Korohou T. Opiyo S. Wang J. Shen M. Egg volume estimation based on image processing and computer vision J. Food Eng. 2020 283 110041 10.1016/j.jfoodeng.2020.110041 10. Nasiri A. Omid M. Taheri-Garavand A. An automatic sorting system for unwashed eggs using deep learning J. Food Eng. 2020 283 110036 10.1016/j.jfoodeng.2020.110036 11. Aboonajmi M. Mostafaei Z. Non-destructive quality assessment of table eggs for online sorting Informatics in Poultry Production Khaliduzzaman A. Springer Singapore 2022 35 51 12. Narushin V.G. Romanov M.N. Griffin D.K. Egg and math: Introducing a universal formula for egg shape Ann. N. Y. Acad. Sci. 2021 1505 169 177 10.1111/nyas.14680 34426991 13. Liao H.B. Egg shape standard curve and its piecewise function equation J. Math. Discov. 2022 2 1 31 10.55375/md.2022.2.1 14. Criminisi A. Single-view metrology: Algorithms and applications Proceedings of the Pattern Recognition: 24th DAGM Symposium Zurich, Switzerland 16-18 September 2002 15. Wang G. Wu Y. Hu Z. A novel approach for single view based plane metrology Proceedings of the 2002 International Conference on Pattern Recognition Quebec City, QC, Canada 11-15 August 2002 16. Peng K. Hou L. Ren R. Ying X.H. Zha H.B. Single view metrology along orthogonal directions Proceedings of the 20th International Conference on Pattern Recognition Istanbul, Turkey 23-26 August 2010 17. Mendonca P.R.S. Kaucic R. Single view metrology: A practical example Proceedings of the 2008 IEEE Workshop on Applications of Computer Vision Copper Mountain, CO, USA 7-9 January 2008 18. Chen Y. IP H.H.S. Single view metrology of wide-angle lens images Vis. Comput. 2006 22 445 455 10.1007/s00371-006-0014-6 19. Wang G.H. Hu Z.Y. Wu F.C. Tsui H.T. Single view metrology from scene constraints Image Vis. Comput. 2005 23 831 840 10.1016/j.imavis.2005.04.002 20. Hoo J.E. Lim K.C. Accuracy and error study of horizontal and vertical measurements with single view metrology for road surveying ARPN J. Eng. Appl. Sci. 2016 11 7872 7876 21. Ali A. Hassan A. Ali A.R. Khan H.U. Kazmi W. Zaheer A. Real-time vehicle distance estimation using single view geometry Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision Snowmass Village, CO, USA 1-5 March 2020 10.1109/WACV45572.2020.9093634 22. Diaz E. Arguello H. An algorithm to estimate building heights from Google street-view imagery using single view metrology across a representational state transfer system Dimensional Optical Metrology and Inspection for Practical Applications V SPIE Baltimore, MD, USA 2016 10.1117/12.2224312 23. Kansal S. Mukherjee S. Automatic single-view monocular camera calibration-based object manipulation using novel dexterous multi-fingered delta robot Neural Comput. Appl. 2019 31 2661 2678 10.1007/s00521-017-3221-3 24. Criminisi A. Reid I. Zisserman A. Single view metrology Int. J. Comput. Vis. 2000 40 123 148 10.1023/A:1026598000963 25. Nieto M. Salgado L. Real-time robust estimation of vanishing points through nonlinear optimization Real-Time Image and Video Processing 2010 SPIE Brussels, Belgium 2010 26. Xie E. Wang W. Yu Z. Anandkumar A. Alvarez J.M. Luo P. SegFormer: Simple and efficient design for semantic segmentation with transformers Adv. Neural Inf. Process. Syst. 2021 34 12077 12090 10.48550/arXiv.2105.15203 27. Zhang S.Q. Tu J. Sun L. Cai J.R. Gu H.F. Gu H. Online detection method of eggs major-minor-axis and mass Trans. Chin. Soc. Agric. Mach. 2014 45 204 209 10.6041/j.issn.1000-1298.2014.02.034 28. Widiasri M. Santoso L.P. Siswantoro J. Computer vision system in measurement of the volume and mass of egg using the disc method IOP Conf. Ser. Mater. Sci. Eng. 2019 703 012050 10.1088/1757-899X/703/1/012050 29. Zhu Z.H. Ye Z.F. Tang Y. Nondestructive identification for gender of chicken eggs based on GA-BPNN with double hidden layers J. Appl. Poult. Res. 2021 30 100203 10.1016/j.japr.2021.100203 |
PMC10000609 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050693 healthcare-11-00693 Article Morphological Characteristics of Proximal Ulna Fractures: A Proposal for a New Classification and Agreement for Validation Labronici Pedro Jose Conceptualization Methodology Software Validation Formal analysis Investigation Resources Data curation Writing - original draft Writing - review & editing Visualization Supervision Project administration Funding acquisition 12* Belangero William Dias Validation Formal analysis Writing - review & editing Visualization Project administration 3 Zublin Carlos Miguel Writing - review & editing Visualization 4 Goncalves Lucas Braga Jaques 5 Fajardo Humberto Validation Writing - review & editing Visualization 2 Pires Robinson Esteves Methodology Validation Formal analysis Writing - review & editing Visualization 6 Giordano Vincenzo Methodology Validation Formal analysis Investigation Writing - original draft Writing - review & editing Visualization Supervision 7* Giansanti Daniele Academic Editor 1 Departamento de Ortopedia e Traumatologia, Universidade Federal Fluminense (UFF), Av. Marques do Parana, 303, Niteroi 24220-000, RJ, Brazil 2 Servico de Ortopedia e Traumatologia Prof. Dr. Donato D Angelo, Hospital Santa Teresa, R. Paulino Afonso, 477, Petropolis 25680-003, RJ, Brazil 3 Departmento de Ortopedia e Traumatologia, Universidade Estadual de Campinas (UNICAMP), Cidade Universitaria Zeferino Vaz, Campinas 13083-970, SP, Brazil 4 Hospital de la Policia Federal Argentina Churruca-Visca, Uspallata 3400, Buenos Aires C1437 JCP, Argentina 5 Servico de Ortopedia, Hospital Madre Teresa, Av. Raja Gabaglia, 1002, Belo Horizonte 30441-070, MG, Brazil 6 Departamento do Aparelho Locomotor, Universidade Federal de Minas Gerais (UFMG), Servico de Ortopedia, Hospital Felicio Rocho, Av. do Contorno, 9530, Belo Horizonte 30110-934, MG, Brazil 7 Servico de Ortopedia e Traumatologia Prof. Nova Monteiro, Hospital Municipal Miguel Couto, Rua Mario Ribeiro 117, Rio de Janeiro 22430-160, RJ, Brazil * Correspondence: [email protected] (P.J.L.); [email protected] (V.G.); Tel.: +55-24-999619191 (P.J.L.); +55-21-997516859 (V.G.) 26 2 2023 3 2023 11 5 69329 12 2022 16 2 2023 23 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Historically, proximal ulna fractures have been simplistically diagnosed and treated as simple olecranon fractures, leading to an unacceptable number of complications. Our hypothesis was that the recognition of lateral, intermediate, and medial stabilizers of the proximal ulna and ulnohumeral and proximal radioulnar joints would facilitate decision-making, including the choice of approach and type of fixation. The primary aim was to propose a new classification for complex fractures of the proximal ulna based on morphological characteristics seen on three-dimensional computed tomography (3D CT). The secondary aim was to validate the proposed classification regarding its inter-rater agreement. Three raters with different levels of experience analyzed 39 cases of complex fractures of the proximal ulna using radiographs and 3D CT scans. We presented the proposed classification (divided into four types with subtypes) to the raters. In this classification, the medial column of the ulna involves the sublime tubercle and is where the anterior medial collateral ligament is inserted, the lateral column contains the supinator crest and is where the lateral ulnar collateral ligament is inserted, and the intermediate column involves the coronoid process of the ulna, olecranon, and anterior capsule of the elbow. inter-rater agreement was analyzed for two different rounds, and the results were evaluated according to Fleiss kappa, Cohen kappa, and Kendall coefficient. inter-rater agreement values were very good (0.82 and 0.77, respectively). Good inter-rater agreement attested to the stability of the proposed classification among the raters, regardless of the level of experience of each one. The new classification proved to be easy to understand and had very good inter-rater agreement, regardless of the level of experience of each rater. classification proximal ulna fracture elbow trauma validation AO TRAUMA Latin America Research OfficeThis study was fully supported by the AO TRAUMA Latin America Research Office. pmc1. Introduction Proximal ulna fractures account for approximately 10% of all upper limb fractures and 21% of all proximal forearm fractures . Historically, these injuries have been simplistically diagnosed as simple olecranon fractures or more complex olecranon fracture-dislocations. This misinterpretation has led to an unacceptable number of complications, mainly due to poorly reduced proximal ulna fractures, resulting in limb instability, stiffness, and dysfunction . A recent improvement in imaging techniques and a better understanding of the biomechanical role of the main determinants of elbow stability have been crucial to changing this situation, allowing proper identification of all morphological characteristics of the proximal ulna . On the medial side, the olecranon and coronoid process act as elbow stabilizers. In particular, the coronoid process, which functions as an important primary stabilizer of this joint, has two facets, separated by a ridge that runs along the greater sigmoid notch. While the anteromedial facet acts as a primary stabilizer, the anterolateral facet is a secondary stabilizer, sharing with the radial head the valgus stabilization of the elbow. In addition to these bony structures, the anterior bundle of the medial collateral ligament, which inserts into the sublime tubercle, is another fundamental stabilizer of the elbow joint on the medial side, resisting deforming forces in varus. Laterally, the radial head acts as a secondary restrictor to valgus deformation, so that the lateral ligament complex acts statically and dynamically to restrict valgus and varus forces. In addition, the lateral ulnar collateral ligament is of paramount importance in the posterior stability of the radial head. Finally, in the sagittal plane, both the olecranon and the triceps brachialis tendon and the coronoid process and anterior capsule of the elbow act as important restrictors of anterior and posterior translation of the ulna, respectively. Current treatment of proximal ulna fractures involves open repair or reconstruction of the osteoligamentous stabilizers of the ulnohumeral and proximal radioulnar joints, with early active mobilization postoperatively. In this context, full restoration of the anatomy of the proximal ulna and its anatomic relationships is essential, including the medial and lateral structures as well as the olecranon, coronoid process, and trochlear notch . Recent literature has highlighted the importance of identifying different patterns of injury to the proximal ulna, but there is no clear guidance on the role of 360deg stabilization of the elbow . Moreover, few studies have used this concept of treatment for proximal ulna fractures . Understanding the impact of injury to these osteoligamentous elements is the basis for avoiding post-traumatic elbow instability . In this scenario, the elbow joint should be divided into three columns as suggested by Watts et al. . These authors introduced the Wrightington classification of elbow fracture dislocation, describing the recognized injury patterns of the three columns to guide treatment decision-making. In their study , the radial head represents the lateral element, the anterolateral coronoid facet represents the middle element, and the anteromedial coronoid facet represents the medial element. Although these osseous structures are extremely important in the genesis of the post-traumatic instability of the dislocated elbow, other structures, such as the olecranon, supinator crest, anterior capsule of the elbow and the collateral ligaments, were not considered, which may lead to some misinterpretations in the definition of the treatment strategy for these challenging lesions. Thus, we propose the use of a true osteoligamentous three-column concept for all proximal ulna fractures to restore 360deg stability adequately and anatomically to the elbow. Our hypothesis is that recognition of the lateral, intermediate, and medial osteoligamentous stabilizers of the proximal ulna and ulnohumeral and proximal radioulnar joints that require repair will facilitate decision-making, including the choice of approach and type of internal fixation and/or ligament repair. The primary aim of our study was to propose a new classification for complex proximal ulna fractures based on morphological characteristics seen on three-dimensional computed tomography (3D CT). The secondary aim was to validate the proposed classification regarding its inter-rater agreement. 2. Materials and Methods 2.1. Patient Selection From 2018 to 2020, 142 fractures involving the proximal ulna treated in a Brazilian hospital were analyzed retrospectively. Patients aged 18 years or over with complex transolecranon or Monteggia-like fractures of the proximal ulna were included. By definition, transolecranon fracture-dislocation preserves joint congruency, whereas Monteggia fracture-dislocation presents proximal radioulnar joint incongruency. Patients with fracture-dislocations showing posterolateral and medial instability, those without a 3D CT scan, and those with previous trauma, infection sequela, or tumor or metabolic injury to the elbow region were excluded. This study was approved by a human research ethics committee with protocol number 49409121.2.1001.5127. All patients signed an Informed Consent form. The probability of making a Type II beta (b) error was calculated to determine the number of samples needed to detect significant changes. Thus, 39 complex fractures of the proximal ulna were randomly selected and analyzed by 3D CT. Ten (25.6%) patients were female and twenty-nine (74.4%) were male. Age ranged from 19 to 79 years with a mean of 44.4 years, a median of 43 years, and a standard deviation of 16.0 years (coefficient of variation was approximately 0.36). Twenty (51.3%) injuries were classified as Monteggia-like (Jupiter types IIA and IID) and nineteen (48.7%) as transolecranon. Of the 20 Monteggia-like injuries, 9 (45.0%) occurred in female patients and 11 (55.0%) in male patients. Of the 19 transolecranon injuries, 1 (5.3%) occurred in a female patient and 18 (94.7%) in male patients. No significant age difference was found between patients with Monteggia-like and transolecranon fractures (p = 0.899) (Table 1). Women were significantly older than men (p = 0.007, Mann-Whitney test). Figure 1 shows how the selection process used for defining the images was performed step by step. CT scans were obtained using a SOMATOM go. Up 32-slice system (Siemens, Erlangen, Germany), and slice thickness was 0.625 mm. Axial, oblique coronal, and sagittal images adapted to the elbow plane were generated, and 3D volume-rendered images and 3D reconstruction models were obtained for all patients. 2.2. Classification In the proposed classification, the medial column of the ulna involves the sublime tubercle and is where the anterior medial collateral ligament (MCL) is inserted; the lateral column of the ulna contains the supinator crest and is where the lateral ulnar collateral ligament (LUCL) is inserted; and the intermediate column involves the coronoid process of the ulna, olecranon, and anterior capsule of the elbow. Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7 and Figure 8 show the classification with the affected structures in the three columns, and the illustrated always show the medial view of the elbow on the left side and the lateral view of the elbow on the right side. Type IA fracture--transolecranon fracture of the proximal ulna with no involvement of the sublime tubercle and supinator crest . Type IB fracture--transolecranon fracture of the proximal ulna with associated fracture of the coronoid process and no involvement of the sublime tubercle and supinator crest . Type IIA fracture--transolecranon fracture of the proximal ulna with associated fracture of the sublime tubercle . Type IIB fracture--transolecranon fracture of the proximal ulna with associated fracture of the sublime tubercle and coronoid process . Type IIIA fracture--transolecranon fracture of the proximal ulna with associated fracture of the supinator crest . Type IIIB fracture--transolecranon fracture of the proximal ulna with associated fracture of the supinator crest and coronoid process . Type IV fracture--transolecranon fracture of the proximal ulna with associated fracture of the sublime tubercle, supinator crest, and coronoid process . Table 2 summarizes the classification of complex fractures of the proximal ulna according to the involvement of the three proposed columns. 2.3. Validation All cases were assessed by three raters with different levels of experience (rater 1 [R1], 9 years of graduation in orthopedics and traumatology; rater 2 [R2], 20 years; and rater 3 [R3], 30 years). Two assessment rounds were conducted with a 30-day interval between them. Between the first and second rounds, all images were allocated and identified by the main author (P.J.L.) and the positions of cases were randomly and manually changed by the principal researcher. The raters were not allowed to keep the images after the first round, and between rounds. In both rounds, the raters had access to the classification in both descriptive and visual forms, and there was no time limit for completing the assessment. 2.4. Statistical Analysis Inter-rater agreement was analyzed using Fleiss kappa weighted for agreement between the three raters and Kendall coefficient. Intra-rater agreement was analyzed using weighted Cohen kappa and Kendall coefficient. Overall analyses (full classification) and specific analyses (for each type of injury) were performed. The kappa value represents the degree of absolute agreement between classifications. Kappa values equal to or greater than 0.75 are considered good to excellent, while values lower than 0.40 indicate poor agreement. Kendall coefficient is a measure of agreement between raters who are assessing a given set of objects. Overall, coefficient values equal to or greater than 0.90 are considered very good. A high coefficient value means that raters applied essentially the same standards to assess the samples. 3. Results 3.1. Intra-Rater Agreement The three raters showed very good agreement between their repeated assessments. In absolute terms, R1 and R3 assigned the same classification in both assessments for 82.0% of the cases, while R2 assigned the same classification for 87% of the cases. For all raters, the significance tests attested for the significance of kappa and Kendall coefficient values (p < 0.05). Agreement in the two assessments by the three raters was found to be significantly different from zero, therefore very good (>0.8). Good intra-rater agreement attested to the stability of the proposed classification between raters, regardless of the level of experience of each one. Table 3 shows the results of intra-rater agreement analysis in the two study rounds. 3.2. Inter-Rater Agreement In absolute terms, inter-rater agreement was very good, as the three raters assigned the same classification for 77% of the cases in the first and second rounds. Agreement between two of the raters was greater than or equal to 0.77. All kappa and Kendall coefficient values in inter-rater analysis were greater than 0.75 (p < 0.001). Inter-rater agreement attested to the reproducibility of the proposed classification, regardless of the level of experience of each rater. Table 4 shows the results of inter-rater agreement analysis. 4. Discussion The proposed column-based classification for fractures of the proximal ulna showed very good inter-rater agreement. Our hypothesis was that the recognition of lateral stabilizers (fractures involving the supinator crest and injury to the LUCL), intermediate stabilizers (fractures involving the coronoid process and injury to the anterior capsule), and medial stabilizers (fractures involving the sublime tubercle and the anterior band of the MCL) of the proximal ulna and ulnohumeral and proximal radioulnar joints that require repair would facilitate decision-making, including the choice of approach and type of internal fixation and/or capsuloligamentous repair. This raises new possibilities in terms of identifying the structures that require repair by 360deg fixation of the elbow, especially in complex transolecranon and Monteggia-like fracture-dislocations of the proximal ulna. Elbow joint stability can be functionally divided into static and dynamic. Static stability is controlled by the osteoarticular architecture, capsule, and ligaments, while dynamic stability is determined by neuromuscular factors. In the elbow joint, this specifically means that static stability is primarily attributed to the congruence between the joint surfaces, anterior joint capsule, and medial and lateral collateral ligaments, particularly the LUCL . The intermediate column, formed by the olecranon, coronoid process, and anterior capsule of the elbow, surrounds the trochlea through an arc of approximately 170deg (trochlear notch), acting as a primary osteocapsular constraint to deforming forces in the sagittal plane and as a secondary constraint in the coronal plane, especially with the elbow at maximum extension. The medial column is formed by the medial facet of the ulna and extends proximally to the base of the coronoid process and distally to the sublime tubercle; it is the site of ulnar insertion of the anterior band of the MCL and acts as a primary constraint to coronal valgus instability . The lateral column is formed by the supinator crest, radial notch of the ulna, and LUCL, and this complex capsuloligamentous structure acts as a primary constraint to varus in the coronal plane . The radial head is relatively incongruent with a greater arc of curvature than the humeral capitellum, acting as a secondary constraint to both valgus deformity in the coronal plane and posterolateral translation. Several classifications have been proposed to describe fractures of the proximal ulna. Most of them have expanded knowledge to specific types of fracture, such as those of the olecranon and coronoid process, or associated fracture patterns, such as Monteggia and Monteggia-like injuries . Although reduction and anatomic fixation of these fractures are recommended, to our knowledge, there is a lack of attention to the importance of adequate reconstruction of all lateral, intermediate, and medial osteoligamentous stabilizers of the proximal ulna and ulnohumeral and proximal radioulnar joints. Melamed et al. reviewed the results of plating of various fracture patterns of the proximal ulna, including isolated olecranon fractures, olecranon fractures combined with a coronoid fracture, and olecranon fractures combined with a coronoid and radial head fracture . The authors presented a scheme that describes elbow stability in different fracture patterns and the treatment chosen for each subgroup. Although they mention the need for coronoid fixation and lateral ligament complex repair in some cases, there is no description of fixation of both the medial and lateral columns of the proximal ulna. Unlike those authors, we included fractures involving the medial facet and sublime tubercle as well as those involving the supinator crest and proximal radioulnar joint facet. In addition to radiographic evaluation, we regularly use 3D CT to facilitate the detection of all traces of fracture and joint incongruity, which ultimately correlate with capsuloligamentous injuries that are sometimes missed at first. Midtgaard et al. demonstrated in a biomechanical study that inferior humeral translation relative to the forearm on initial lateral radiograph of the elbow suggests injury to the collateral ligaments . When inferior translation exceeds 3 mm, the integrity of the lateral collateral ligament (LCL) or MCL should be questioned. If inferior translation exceeds 7.5 mm, a disruption of both ligaments (LCL and MCL) should be suspected. In the presence of a multifragmented fracture, a CT scan is necessary for preoperative planning and evaluation of the sublime tubercle and supinator crest. In the case of a capsuloligamentous injury due to a fracture at the bony insertions of the ligaments (LCL and MCL) or the anterior capsule, these fragments, even small ones, should be securely fixed. Watts et al. proposed a new comprehensive classification system based on the same three-column concept of elbow fracture dislocation. These authors described a lateral, a middle, and a medial column in the elbow joint, demonstrating the natural fulcrum in between the middle and medial columns. This classification system clearly showed the impact of injury to the osseous elements for elbow instability. However, these authors did not provide any information on the ligamentous structures and their role in the genesis in both acute and chronic elbow instability. In this context, the inclusion of all osteoligamentous structures of the elbow is probably the main strength of our study, as one can understand why minimal bone fragments can generate a major instability in this joint. Another strength of this classification for complex fractures of the proximal ulna is to help surgeons recognize areas of instability through 3D CT with a focus on the importance of the sublime tubercle, supinator crest, and coronoid process of the ulna, as well as the capsuloligamentous structures that insert into these structures. This facilitates indication for appropriate treatment based on the division of the proximal ulna into three columns (lateral, intermediate, and medial) for restoring 360deg stability to the elbow. Other authors have shown that the column-based division is useful for the interpretation and management of different skeletal trauma conditions . This study has some limitations. First, the new classification system was assessed by three orthopedic surgeons with different levels of expertise, which may have influenced our findings. However, we found good inter-rater agreement, which can be seen as a promising perspective for understanding and applying the proposed classification. Second, only transolecranon or Monteggia-like fractures of the proximal ulna were included, therefore one can ask about its validity in other complex injuries around the elbow involving the proximal region of the ulna, such as fracture-dislocations with posterolateral and/or medial instability. Although fractures of the proximal ulna vary in severity and in the diversity of associated injuries, all are characterized by presenting some degree of damage to the key stabilizing structures of the elbow . The treatment principle is based on adequate reconstruction of the ulna and elbow stabilizing structures, which have been proven to be factors related to better joint function and reduced risk of progressive joint degeneration due to chronic elbow instability . Thus, it seems logical that the 360deg understanding of the elbow should be a rule and the proposed classification proved to be useful for this. Finally, we did not validate the system in a clinical setting, including a report of the outcomes of our cases treated according to the three columns of the proximal ulna. The next step for this consortium of investigators is to carry out a multicenter prospective cohort study soon to validate the new classification in a clinical-therapeutic setting. 5. Conclusions The new classification for complex fractures of the proximal ulna was easily understood and can be considered reproducible, as it showed high inter-rater agreement in orthopedic surgeons with different levels of experience. Author Contributions Conceptualization, P.J.L., W.D.B., R.E.P. and V.G.; methodology, P.J.L., W.D.B. and V.G.; software, P.J.L. and L.B.J.G.; validation, P.J.L., L.B.J.G. and H.F.; formal analysis, P.J.L. and V.G.; investigation, P.J.L., W.D.B., C.M.Z., L.B.J.G., H.F., R.E.P. and V.G.; resources, P.J.L.; data curation, P.J.L., W.D.B., C.M.Z., R.E.P. and V.G; writing--original draft preparation, P.J.L.; writing--review and editing, V.G., W.D.B., C.M.Z. and R.E.P.; visualization, P.J.L., W.D.B., C.M.Z., L.B.J.G., H.F., R.E.P. and V.G.; supervision, P.J.L., W.D.B. and V.G.; project administration, P.J.L. and V.G.; funding acquisition, P.J.L. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of UFF with protocol number 49409121.2.1001.5127. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Selection process used for defining the images. Figure 2 Type IA fracture--transolecranon fracture of the proximal ulna with no involvement of the sublime tubercle and supinator crest. Figure 3 Type IB fracture--transolecranon fracture of the proximal ulna with associated fracture of the coronoid process and no involvement of the sublime tubercle and supinator crest. Figure 4 Type IIA fracture--transolecranon fracture of the proximal ulna with associated fracture of the sublime tubercle. Figure 5 Type IIB fracture--transolecranon fracture of the proximal ulna with associated fracture of the sublime tubercle and coronoid process. Figure 6 Type IIIA fracture--transolecranon fracture of the proximal ulna with associated fracture of the supinator crest. Figure 7 Type IIIB fracture--transolecranon fracture of the proximal ulna with associated fracture of the supinator crest and coronoid process. Figure 8 Type IV fracture--transolecranon fracture of the proximal ulna with associated fracture of the sublime tubercle, supinator crest, and coronoid process. healthcare-11-00693-t001_Table 1 Table 1 Age distribution for specific subgroups. Age Statistics Monteggia-like Transolecranon Overall Female Male Overall Female Male Minimum 22.0 30.0 22.0 19.0 62.0 19.0 Maximum 79.0 79.0 58.0 73.0 62.0 73.0 Median 48.0 55.0 31.0 40.0 62.0 40.0 Mean 45.2 55.7 36.5 43.7 62.0 42.7 Standard deviation 16.1 14.1 12.3 16.3 0.0 16.1 Coefficient of variation 0.36 0.25 0.34 0.37 0.00 0.38 Source: Servico de Ortopedia e Traumatologia Prof. Donato D'Angelo, Hospital Santa Teresa. healthcare-11-00693-t002_Table 2 Table 2 Affected column structures according to type in the new classification. Type Medial Column (Sublime Tubercle/Anterior Band of MCL) Intermediate Column (Coronoid Process/Anterior Capsule) Lateral Column (Supinator Crest/LUCL) IA - - - IB - + - IIA + - - IIB + + - IIIA - - + IIIB - + + IV + + + Source: Servico de Ortopedia e Traumatologia Prof. Donato D'Angelo, Hospital Santa Teresa. healthcare-11-00693-t003_Table 3 Table 3 Analysis of intra-rater agreement in the two study rounds. Rater Absolute Agreement Weighted Kappa (95% CI) Kendall Coefficient R1 0.82 0.82 (0.69; 0.90) ++ 0.82 ++ R2 0.87 0.91 (0.83; 0.95) ++ 0.84 ++ R3 0.82 0.92 (0.85; 0.96) ++ 0.90 ++ Source: Servico de Ortopedia e Traumatologia Prof. Donato D'Angelo, Hospital Santa Teresa. ++ p < 0.001. CI, confidence interval; R, rater. healthcare-11-00693-t004_Table 4 Table 4 Analysis of inter-rater agreement in the two study rounds. Assessment Round Agreement Measure R1-R2 R1-R3 R2-R3 R1, R2, and R3 1 Absolute agreement 0.79 0.77 0.77 0.77 Weighted kappa ++ 0.75 (0.45; 0.85) 0.75 (0.44; 0.82) 0.78 (0.49; 0.83) 0.76 (0.59; 0.86) Kendall coefficient ++ 0.78 0.77 0.72 - 2 Absolute agreement 0.82 0.77 0.79 0.77 Weighted kappa ++ 0.77 (057; 0.88) 0.75 (0.33; 0.81) 0.77 (0.44; 0.87) 0.78 (0.63; 0.88) Kendall coefficient ++ 0.78 0.77 0.78 - Source: Servico de Ortopedia e Traumatologia Prof. Donato D'Angelo, Hospital Santa Teresa. ++ p < 0.001. R, rater. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Sahajpal D. Wright T.W. Proximal ulna fractures J. Hand. Surg. Am. 2009 34 357 362 10.1016/j.jhsa.2008.12.022 19181239 2. Pereira B.P. Revisiting the anatomy and biomechanics of the anconeus muscle and its role in elbow stability Ann. Anat. 2013 195 365 370 10.1016/j.aanat.2012.05.007 22874649 3. Sanchez-Sotelo J. Morrey M. Complex elbow instability: Surgical management of elbow fracture dislocations EFORT Open Rev. 2016 1 183 190 10.1302/2058-5241.1.000036 28461946 4. An K.N. Morrey B.F. Chao E.Y. The effect of partial removal of proximal ulna on elbow constraint Clin. Orthop. Relat. Res. 1986 270 279 10.1097/00003086-198608000-00041 5. Bell T.H. Ferreira L.M. McDonald C.P. Johnson J.A. King G.J. Contribution of the olecranon to elbow stability: An in vitro biomechanical study J. Bone Jt. Surg. Am. 2010 92 949 957 10.2106/JBJS.H.01873 6. Closkey R.F. Goode J.R. Kirschenbaum D. Cody R.P. The role of the coronoid process in elbow stability. A biomechanical analysis of axial loading J. Bone Jt. Surg. Am. 2000 82 1749 1753 10.2106/00004623-200012000-00009 7. Pollock J.W. Brownhill J. Ferreira L. McDonald C.P. Johnson J. King G. The effect of anteromedial facet fractures of the coronoid and lateral collateral ligament injury on elbow stability and kinematics J. Bone Jt. Surg. Am. 2009 91 1448 1458 10.2106/JBJS.H.00222 8. Reichel L.M. Milam G.S. Hillin C.D. Reitman C.A. Osteology of the coronoid process with clinical correlation to coronoid fractures in terrible triad injuries J. Shoulder Elbow Surg 2013 22 323 328 10.1016/j.jse.2012.10.038 23333172 9. Reichel L.M. Morales O.A. Gross anatomy of the elbow capsule: A cadaveric study J Hand Surg. Am. 2013 38 110 116 10.1016/j.jhsa.2012.09.031 23200220 10. Ring D. Guss D. Jupiter J.B. Reconstruction of the coronoid process using a fragment of discarded radial head J. Hand Surg. Am. 2012 37 570 574 10.1016/j.jhsa.2011.12.016 22305734 11. Sanchez-Sotelo J. O'Driscoll S.W. Morrey B.F. Medial oblique compression fracture of the coronoid process of the ulna J. Shoulder Elb. Surg 2005 14 60 64 10.1016/j.jse.2004.04.012 12. Sanchez-Sotelo J. O'Driscoll S.W. Morrey B.F. Anteromedial fracture of the coronoid process of the ulna J. Shoulder Elb. Surg 2006 15 e5 e8 10.1016/j.jse.2005.05.009 16979044 13. Schneeberger A.G. Sadowski M.M. Jacob H.A. Coronoid process and radial head as posterolateral rotatory stabilizers of the elbow J. Bone Jt. Surg. Am. 2004 86 975 982 10.2106/00004623-200405000-00013 15118040 14. von Ruden C. Woltmann A. Hierholzer C. Trentz O. Buhren V. The pivotal role of the intermediate fragment in initial operative treatment of olecranon fractures J. Orthop. Surg. Res. 2011 6 9 10.1186/1749-799X-6-9 21310049 15. Ring D. Jupiter J.B. Fracture-dislocation of the elbow Hand Clin. 2002 18 55 63 10.1016/S0749-0712(02)00004-5 12143418 16. Watts A.C. Singh J. Elvey M. Hamoodi Z. Current concepts in elbow fracture dislocation Shoulder Elb. 2021 13 451 458 10.1177/1758573219884010 17. Doornberg J. Ring D. Jupiter J.B. Effective treatment of fracture-dislocations of the olecranon requires a stable trochlear notch Clin. Orthop. Relat. Res. 2004 429 292 300 10.1097/01.blo.0000142627.28396.cb 15577501 18. Giannicola G. Greco A. Sacchetti F.M. Cinotti G. Nofroni I. Postacchini F. Complex fracture-dislocations of the proximal ulna and radius in adults: A comprehensive classification J. Shoulder Elb. Surg. 2011 20 1289 1299 10.1016/j.jse.2011.06.003 19. Jupiter J.B. Leibovic S.J. Ribbans W. Wilk R.M. The posterior Monteggia lesion J. Orthop. Trauma 1991 5 395 402 10.1097/00005131-199112000-00003 1761999 20. Ring D. Jupiter J.B. Sanders R.W. Mast J. Simpson N.S. Transolecranon fracture-dislocation of the elbow J. Orthop. Trauma 1997 11 545 550 10.1097/00005131-199711000-00001 9415859 21. Hamoodi Z. Singh J. Elvey M.H. Watts A.C. Reliability and validity of the Wrightington classification of elbow fracture-dislocation Bone Joint J. 2020 102 1041 1047 10.1302/0301-620X.102B8.BJJ-2020-0013.R1 32731824 22. de Haan J. Schep N.W. Eygendaal D. Kleinrensink G.J. Tuinebreijer W.E. den Hartog D. Stability of the elbow joint: Relevant anatomy and clinical implications of in vitro biomechanical studies Open Orthop. J. 2011 5 168 176 10.2174/1874325001105010168 21633722 23. Regan W. Morrey B. Fractures of the coronoid process of the ulna J. Bone Jt. Surg. Am. 1989 71 1348 1354 10.2106/00004623-198971090-00011 24. Jackson J.D. Steinmann S.P. Radial head fractures Hand Clin. 2007 23 185 193 185-193, vi 10.1016/j.hcl.2007.01.009 17548010 25. O'Driscoll S.W. Jupiter J.B. Cohen M.S. Ring D. McKee M.D. Difficult elbow fractures: Pearls and pitfalls Instr. Course Lect. 2003 52 113 134 12690844 26. O'Driscoll S.W. Morrey B.F. Korinek S. An K.N. Elbow subluxation and dislocation. A spectrum of instability Clin. Orthop. Relat. Res. 1992 186 197 27. Regan W.D. Korinek S.L. Morrey B.F. An K.N. Biomechanical study of ligaments around the elbow joint Clin. Orthop. Relat. Res 1991 170 179 10.1097/00003086-199110000-00023 28. Ring D. Jupiter J.B. Fracture-dislocation of the elbow J. Bone Jt. Surg. Am. 1998 80 566 580 10.2106/00004623-199804000-00014 29. Chick G. Nordin J.Y. Fractures complexes de l'extremite superieure des deux os de l'avant-bras chez l'adulte: Proposition d'une nouvelle classification Rev. Chir. Orthop. Reparatrice Appar. Mot. 2001 87 765 772 11845080 30. Egol K.A. Tejwani N.C. Bazzi J. Susarla A. Koval K.J. Does a Monteggia variant lesion result in a poor functional outcome?: A retrospective study Clin. Orthop. Relat. Res. 2005 438 233 238 10.1097/01.blo.0000168806.79845.8b 16131896 31. Xiao R.C. Chan J.J. Cirino C.M. Kim J.M. Surgical management of complex adult Monteggia fractures J. Hand Surg. Am. 2021 46 1006 1015 10.1016/j.jhsa.2021.07.023 34507868 32. Melamed E. Danna N. Debkowska M. Karia R. Liporace F. Capo J.T. Complex proximal ulna fractures: Outcomes of surgical treatment Eur. J. Orthop. Surg. Traumatol. 2015 25 851 858 10.1007/s00590-015-1628-z 25869104 33. Midtgaard K.S. Nolte P.C. Miles J.W. Tanghe K.K. Peebles L.A. Provencher M.T. Adolfsson L. Biomechanical significance of the collateral ligaments in transolecranon fracture dislocations J. Shoulder Elbow Surg. 2021 30 1245 1250 10.1016/j.jse.2020.09.010 33010439 34. Kfuri M. Schatzker J. Revisiting the Schatzker classification of tibial plateau fractures Injury 2018 49 2252 2263 10.1016/j.injury.2018.11.010 30526924 35. Park S.G. Ko Y.J. Medial buttress plating for humerus fractures with unstable medial column J. Orthop. Trauma 2019 33 e352 e359 10.1097/BOT.0000000000001515 31188258 36. Rikli D.A. Babst R. Jupiter J.B. Distale Radiusfraktur: Neue Konzepte als Basis fur die operative Therapie Handchir. Mikrochir. Plast. Chir. 2007 39 2 8 10.1055/s-2007-964921 17402134 37. Sameer M. Bassetty K.C. Singaravadivelu V. Fixation of tibial pilon fractures based on column concept: A prospective study Acta Orthop. Belg. 2017 83 568 573 30423663 38. Zhang R. Yin Y. Li A. Wang Z. Hou Z. Zhuang Y. Fan S. Wu Z. Yi C. Lyu G. Three-column classification for acetabular fractures: Introduction and reproducibility assessment J. Bone Jt. Surg. Am. 2019 101 2015 2025 10.2106/JBJS.19.00284 39. Gonzalez L.J. Shields C.N. Leucht P. Konda S.R. Egol K.A. Fracture-dislocations of the elbow: A comparison of Monteggia and terrible triad fracture patterns Orthopedics 2022 2 1 6 10.3928/01477447-20221129-07 40. Siebenlist S. Buchholz A. Braun K.F. Fractures of the proximal ulna: Current concepts in surgical management EFORT Open Rev. 2019 4 1 9 10.1302/2058-5241.4.180022 30800474 |
PMC10000610 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12050967 foods-12-00967 Article Application of Three-Dimensional Digital Photogrammetry to Quantify the Surface Roughness of Milk Powder Ding Haohan Methodology Software Validation Formal analysis Investigation Resources Data curation Writing - original draft Writing - review & editing Visualization 12 Wilson David I. Conceptualization Methodology Software Writing - review & editing Visualization 3* Yu Wei Conceptualization Formal analysis Writing - review & editing Supervision Project administration 4 Young Brent R. Validation Writing - review & editing Supervision Project administration 4 Cui Xiaohui Investigation Writing - review & editing 15 Bhunia Arun K. Academic Editor 1 Science Center for Future Foods, Jiangnan University, Wuxi 214122, China 2 School of Artificial Intelligence and Computer Science, Jiangnan University, Wuxi 214122, China 3 Electrical and Electronic Engineering Department, Auckland University of Technology, Auckland 1010, New Zealand 4 Department of Chemical & Materials Engineering, University of Auckland, Auckland 1010, New Zealand 5 School of Cyber Science and Engineering, Wuhan University, Wuhan 430072, China * Correspondence: [email protected] 24 2 2023 3 2023 12 5 96713 12 2022 15 2 2023 22 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). The surface appearance of milk powders is a crucial quality property since the roughness of the milk powder determines its functional properties, and especially the purchaser perception of the milk powder. Unfortunately, powder produced from similar spray dryers, or even the same dryer but in different seasons, produces powder with a wide variety of surface roughness. To date, professional panelists are used to quantify this subtle visual metric, which is time-consuming and subjective. Consequently, developing a fast, robust, and repeatable surface appearance classification method is essential. This study proposes a three-dimensional digital photogrammetry technique for quantifying the surface roughness of milk powders. A contour slice analysis and frequency analysis of the deviations were performed on the three-dimensional models to classify the surface roughness of milk powder samples. The result shows that the contours for smooth-surface samples are more circular than those for rough-surface samples, and the smooth-surface samples had a low standard deviation; thus, milk powder samples with the smoother surface have lower Q (the energy of the signal) values. Lastly, the performance of the nonlinear support vector machine (SVM) model demonstrated that the technique proposed in this study is a practicable alternative technique for classifying the surface roughness of milk powders. 3D image analysis surface roughness milk powder contour slice analysis National Key Research and Development Program of China2022YFF1101100 Food Safety Data Sharing and Event Alert1507-250071440 Fundamental Research Funds for the Central UniversitiesJUSRP123053 This work was supported by the National Key Research and Development Program of China (2022YFF1101100), the Food Safety Data Sharing and Event Alert (1507-250071440) and the Fundamental Research Funds for the Central Universities (JUSRP123053). pmc1. Introduction Quality control and product consistency are key properties for any food industry, and especially for the dairy industry . Experience shows that the surface appearance usually influences consumers' assumptions about the organoleptic and functional performance of milk powders, particularly for those consumers who use the product regularly and have become accustomed to the surface appearance of the product . Additionally, it is inevitable for consumers to assess the quality of milk powders through visual perceptions of the product, so, consequently, variations in the surface appearance may lead to the customers considering that these milk powder products are counterfeit or even unsafe . Therefore, the surface appearance is particularly important for milk powders, and visual consistency is necessary for milk powder plants. However, since different process conditions may affect the properties and appearance of milk powders , it is problematic for different plants to maintain the visual consistency of milk powders. In addition, the traditional way to grade the roughness of milk powders is to use sensory panelists which is empirical and laborious. Consequently, it is essential to find an efficient and reliable way to assess the surface roughness of milk powders and to maintain the visual consistency of milk powder products. Computer vision techniques that have the ability to quantify the surface environments were used by many studies to acquire texture features from various images , and were used to measure the surface roughness of components used in engineering . Three-dimensional laser scanning methods that can generate geometrical triangulated data using a non-contact active method have been used previously in the texture analysis of some photographs involving metal and concrete . Alternatively, photogrammetry that can reconstruct three-dimensional (3D) models by stitching together numerous images from various positions is a cheaper alternative to 3D laser scanning and has become a practicable method in 3D reconstructions or surface texture analysis . However, many studies used photogrammetry to analyze the texture of soil and sediments . Moret-Fernandez et al. utilized photogrammetry to evaluate the bulk density of small soil aggregates, while Merel and Farres also used photogrammetry to measure the surface evolution and microrelief caused by erosion, and found that this method is sufficiently accurate. In addition, Moret-Fernandez et al. stated that it is better for the novice to use automated software to process photogrammetry, and the time required for processing data will be higher if the system is manual . Furthermore, the fast Fourier transform (FFT) spectra were used in many studies to extract the textural feature from images . To date, standard laboratory tests, such as flowability, water activity, bulk density, and particle size distribution, have been used to measure the functional properties of milk powders . For example, a kinetic pulse nuclear magnetic resonance (NMR) technique that can measure the rehydration of milk powders was proposed in , Nijdam and Langrish calculated the bulk density by computing the volume variation of milk powders in a graduated cylinder after tapping, while Lee et al. measured the dispersibility, as well as the wettability of milk powders by calculating the variance between the electrical resistance of water and the electrical resistance of air. However, since these instrumental measures are different from human perception, these instrument measures may misjudge the sensory quality of products . Additionally, the sensory quality of products has been analyzed by many texture analysis techniques. For instance, the sensory and texture properties of cholesterol-removed and whole milk cream cheese had been compared by Jeon et al. throughout four weeks of storage, Gosselin et al. analyzed the texture of polymer powders by using the GLCM technique, while Lille et al. evaluated the flavor and appearance of snacks made from wholegrain rye flour and whole milk powder by using the sensory analysis. In addition, particle texture analysis (PTA), powder electron diffraction, and scanning electron micrographs are utilized to assess the surface of various products . Furthermore, Traill et al. utilized a trained sensory panel and a Rate-All-That-Applies technique to distinguish the roughness of milk powders, and determined that the significant distinction between milk powder samples is the size of milk powder lumps. Thus, Traill et al. designed a photographic standard that can be utilized by grading assessors to classify commercial dairy powders into different lumpiness (roughness) groups, and demonstrated that this technique can grade the milk powders according to the level of visual lumpiness. However, using the sensory panel to categorize the appearance of milk powders is subjective. Additionally, all the human evaluators need to be trained. In a previous related study, Ding et al. proposed a purely geometric algorithm based on the 3D mesh which computes the area of the triangle formed by the three adjacent surface normals to classify the local surface smoothness of the milk powder samples. This study is based on the previous study , and the aim of the previous study is to explore the feasibility of classifying the visual appearances of different milk powders by using photogrammetry, while the aim of this continuing study is to improve on this single local method, to propose 3D image processing quantification algorithms, and to explore the reliability and feasibility of using this method to grade the surface appearance of different milk powder samples. In addition, the surface normal analysis which compares the area of triangle formed by the three adjacent surface normals as well as the angle between the adjacent surface normals was used in the previous study , and principal component analysis (PCA) was used to reduce the set of variables for the classifier, while a third-order polynomial nonlinear support vector machine (SVM) classifier was developed to classify the surface smoothness of milk powder samples in the previous study . However, this work introduces a strategy based on analyzing contours and additionally, performing a frequency analysis of these curves to extract the high frequency components which are related to the lumpiness of the sample, and a second-order polynomial nonlinear support vector machine (SVM) model was chosen to grade the samples. Consequently, the methods used in this work are entirely different from the methods used in the previous work. Furthermore, the results in the previous work showed that the surface normal analysis is effective for quantifying the surface appearance of milk powders, while it is expected that the results in this study will prove that the 3D digital photogrammetry techniques proposed in this work can effectively distinguish the visual appearance of milk powders and is a practicable alternative technique for classifying the surface roughness of milk powders. 2. Materials and Methods 2.1. Milk Powder Samples The method of preparing milk powder samples in this study is the same as the method used in a previous study . To duplicate the experience of a representative customer, the Fonterra Co-operative milk powders used were purchased off the shelf from a local Auckland, NZ, superstore. These milk powders comprise both instant whole and instant trim milk powders which allows the investigation of the effects of moisture level on the surface texture of various types of milk powders. The milk powders bought have similar moisture levels, though the surface texture properties of the milk powder samples may be different due to different moisture levels . Therefore, to artificially create milk powders purchased at different locations and seasons, with corresponding differing moisture levels and surface appearances, varying amounts of water were sprayed on the samples. To assess the actual moisture level of each milk powder sample, three replicates of samples were dried in an oven (Cole-Parmer, Vernon Hills, IL, USA) after photogrammetry tests, where the mass difference before and after drying is the weight of moisture contained, and the mean of three values was used. Four moisture levels were manually made in both samples (instant whole and instant trim milk powders), and these surface roughness grades are denoted as original, smooth, medium, and rough, where the first surface roughness grade (original) are the original milk powders. Three replicates of analysis were made for each moisture class for repeatability. Figure 1 shows the calibration chart with a somewhat arbitrarily chosen background image which is subsequently in the image processing algorithm to robustly differentiate specific reference points. The squares of different colors in Figure 1 are used to distinguish the location of the model when building 3D models of milk powder samples. 2.2. Sample Preparation As noted in Ding , it is important for the subsequent geometrically based analysis strategy that the milk powder cone is constructed in a consistent manner. Figure 2a shows the milk powder delivery device which contains a funnel with a stopper, a box, and a sample holder . To maintain consistency in the sample preparation, the size of the funnel and the relative position of the funnel to the sample holder are the same. Various sizes of the funnel can make milk powder cones with different shapes and sizes. For each sample, 80 g of milk powders was released onto the sample holder below the funnel. The shape of all the milk powder cones is similar in order to increase the comparability. Additionally, Figure 2b shows the diagram of the photogrammetry equipment . A Nikon D810 camera (Nikon, Tokyo, Japan), was used for the image capture and was fixed on a tripod. The distance from the camera to the samples was constant. In addition, to emphasize the boundary of the milk powder cones, and to eliminate the effects of excess shadow on the milk powder cones, four floodlights were turned on and placed on both sides of the samples. A black backing cardboard was used to maintain a constant background and lighting. Furthermore, over 60% overlap between spatially continuous photos is needed for photogrammetry to develop 3D digital models of milk powder cones. Therefore, the milk powder cones on the turntable were photographed about every 11deg, so that each sample has 33 images. Detailed figures of the milk powder delivery device and photogrammetry equipment are shown in Ding et al. , and an example picture of a cone is presented in Figure 3. 2.3. 3D Digital Models Building The computer used for building 3D digital models was a Lenovo computer (Lenovo, Beijing, China), with an Intel (R) Core (TM) i9-10900 CPU @ 2.80 GHz with 32 GB installed RAM, and the graphics card of the computer is NVIDIA GeForce RTX 3060 (12 GB). After achieving all the photos of milk powder samples from photogrammetry equipment, the 33 images of each milk powder sample were separately imported into the software RealityCapture 1.0 (EPIC GAMES, Carrytown, NC, United States), which builds the 3D triangular mesh of milk powder cones. Subsequently, the whole three-dimensional reconstructions with the texture of milk powder cones were created. The 3D triangular meshes were processed in Matlab R2019b (MathWorks, Natick, MA, USA). 3. Three-Dimensional Image Analysis In previous work, Ding et al. built the surface normals of each triangle mesh for milk powder samples, and used the differences (area and angle formed by the adjacent surface normals) between the 3D milk powder models to measure the surface smoothness of milk powder samples. However, this work aims to slice the 3D milk powder samples into equally spaced contours, and use the frequency response of the deviations for each contour to quantify the surface roughness of milk powder samples, which is completely different from the method used in the previous study. To achieve this, we first extract and unwarp the contours from the cone models, then compute the differences between the unwrapped contours and the best-fit perfect circles. Finally, we can then analyze the roughness of milk powder samples by comparing the dominant frequencies of the wavelengths of expected lumpiness. 3.1. Contour Slice Analysis In order to make the thickness between each layer as small as possible and to ensure that the data between each layer are enough for the subsequent analysis, around 40 equally spaced contours from the bottom of the cone to the top of the cone were used. The first step is to extract the contours of the cone, as shown in Figure 4. The contours will be circles if the milk powder sample is perfectly smooth and fall in a cone-shaped heap, while the roughness (lumpiness) of the sample is indicated by the deviations from the circles. 3.2. Frequency Analysis of the Deviations For each contour, this technique was used to measure the features (dominant frequencies and shape) of the differences from circularity, and it is assumed that the lumpiness will deform the circularity of the contour. For each contour, the first step is to extract the boundary (x, y) at the given height. For an ideal cone, the extracted boundary will be a circle, but for the milk powder cone, the extracted boundary will diverge to some extent from a circle. Subsequently, the data of the contour was used to fit the least-squares circle, and the radius (R) of this circle was calculated. Given the true circle radius, the difference between the contour radius (r) and R as a function of angle around the circle can be calculated by Equation (1):(1) e(th)=R-r(th) In order to take advantage of the efficient FFT, this trend will need to be interpolated on an equal grid spacing. The dominant frequencies (or alternatively the dominant wavelengths) and the variance of the difference between the unwrapped contour and the least-squares fit mean can be calculated if the deviations (at each altitude) are given. To better focus on the wavelength regions, the l-axis can be abridged because the estimated sizes of the lumpiness are known. In addition, the frequency component is a spatial frequency (typically denoted b) since the basic measurement is a distance. Generally, longer data sets are needed to acquire a better resolution at the higher wavelengths, and it is necessary to sample at a finer resolution to better measure the higher frequencies. However, in this study, the finer resolution is not essential to better quantify the lumpiness, and furthermore, since the cone is circular, it is inappropriate to simply sample more data since one will then traverse again around the cone a second time. Lastly, the energy of the signal can be calculated by Equation (2): (2) Q=|FFT(e)|dl 3.3. Support Vector Machine (SVM) The SVM is a kernel-based method that is widely used to address pattern recognition problems and binary classification problems . To prevent overfitting, cross validation was used in this study. The data were separated into four subsets (each subset has 25% of the data). After trying different classifiers, a second-order polynomial nonlinear SVM classifier outperformed the others. Therefore, this SVM model was chosen to categorize the surface roughness of the samples. In addition, the confusion matrix , which is a classification assessment method , was utilized to assess the performance of the SVM classifier. The performance indicators, including specificity, overall accuracy, and sensitivity, are computed, and the detailed definition of these indicators is described in . 4. Results and Discussion 4.1. Milk Powder Cones Table 1 shows the mean moisture values of all the moisture levels with standard deviation , while Figure 5 presents the front views of each sample . Traill et al. used photo standards to grade the surface roughness of milk powders by a trained sensory panel. Compared with the photo standards of lumpiness grades classified in , the original (first class) milk powder samples have a similar surface appearance to the level 0 dairy powders, and the rough (fourth class) milk powder samples have a similar appearance to the extreme appearance (level 14) of milk powders. Consequently, the original and rough milk powder samples are, respectively, referred to as Class 0 and Class 3. In addition, the smooth (second class) samples and the medium (third class) samples have a similar surface appearance to the moderate clumping samples (level 4-9) and the high clumping samples (level 9-13), separately. Thus, the smooth milk powder samples and the medium milk powder samples are separately denoted as Class 1 and Class 2. From the figures, since the oblique vertical views of samples are clearly visually very different from each other, it is reasonable to assume that there is a strong correlation between the surface texture properties and the moisture level of powder samples. However, it is vital to discover a mathematical and robust method to quantify these relations. For each moisture level, three milk powder cones were fabricated to ensure the repeatability of the method. Figure 6 shows an example 3D reconstruction triangular mesh model which has around 500,000 triangles of a sample. From Figure 6, it is notable that if the milk powder sample has no lumpiness (perfectly smooth), the milk powder cone will be a near-perfect circular-based cone, which means that the top view of the contour extracted from the three-dimensional milk powder model will be a perfectly concentric circle. On the other hand, the contours will show fluctuations around the perfect circles if the sample's surface is rough (with some lumpiness). Figure 7a shows the unwrapped contour of a Class 0 milk powder sample at the middle layer (layer 20), while Figure 7b presents an unwrapped contour of the Class 3 milk powder sample at the middle layer (layer 20). The red circles in the figures are the least-squares circles fitted by the data of the 3D digital models in this layer. From these trends, it is clear that the contour of the smooth milk powder sample shows a near-perfect circle, and the unwrapped contour of the smooth sample shows a correspondingly near-straight horizontal line. Conversely, the contour of the Class 3 sample appears irregular, and the unwrapped contour of the rough milk powder sample exhibits considerable fluctuations. 4.2. Contour Slice Analysis The 3D digital model top views of Class 0-3 instant whole and trim milk powder samples are shown in Figure 8. It is obvious that the contours of the Class 0 milk powder samples are relatively circular showing that the Class 0 samples exhibit little lumpiness, while the contours of the Class 3 samples presented in Figure 8 are relatively irregular showing that there are many lumps in the Class 3 milk powder samples. Furthermore, although the Class 0 instant whole milk powder sample does not show any obvious lumpiness, on the whole, the surface of this sample is rougher than the surface of the Class 0 instant trim milk powder sample. All contours (40 layers) for a Class 0 sample and a Class 3 sample are separately shown in Figure 9a,b. Note that the radius of the circle becomes increasingly smaller from the bottom to the top, which means that the bottom contour slices have more data than the top contour slices. Additionally, from Figure 9b, it is clear that there are more lumps on the bottom contour slices, which proves that the lumpiness tends to fall down the milk powder cone. It is also notable that all the contours of the Class 0 sample are more circular than the contours of the Class 3 sample, illustrating that the smoother the sample, the more circular the contours. 4.3. Variance of the Contours The unwrapped contours of the Class 0-3 instant whole and trim milk powder samples are shown in Figure 10. It is worth mentioning that the higher the cone (the smaller of the radius), the fewer the data points of the contours. Additionally, the contours extracted from the smooth samples have fewer oscillations than the contours extracted from the rough samples, and the large oscillations represent the lumpiness. From Figure 10, it is clear that the Class 3 milk powder cones have the most lumps, while the Class 0 milk powder cones have almost no lumps. In addition, the Class 1 milk powder cones have a small number of lumps and the Class 2 milk powder cones slightly more lumps than the Class 1 milk powder cones. Furthermore, all the oscillations are concentrated in the contour slices with a small radius (bottom). The deviations between the contours and the least-squares circles for each layer of Class 0-3 instant whole and trim samples are presented as a linear plot in Figure 11, and Figure 11 also plots the standard deviation for each altitude. A rough surface will have a higher standard deviation while the low standard deviation will be shown in the smooth surface, which is obvious in both the rough and smooth samples. For example, the standard deviations of the Class 0 cones are the lowest, while the standard deviations of the Class 3 cones are the highest. In addition, the Class 1 milk powder cones have a larger standard deviation than the Class 0 cones but a smaller standard deviation than the Class 2 milk powder cones. Additionally, the trends of the standard deviations in each Class milk powder cones are similar. Furthermore, since the lumpiness tends to fall down the milk powder cone, a higher variance is shown in the longer contours nearer the bottom. 4.4. Comparing the Frequency Responses Figure 12a,b separately show the results of the contours, deviation, and frequency response of a Class 0 and Class 3 samples. For each sample, the absolute value of the frequency response was calculated, and the log of the wavelength against the log of the frequency response (which is easier to interpret) was presented in the figures. The linear perimeter distances for each layer (in the deviation plot) are simply computed by using the nominal radius (the following numbers) and the angle around the cone, and it is clear that the computed values are the approximation of the true distances. The independent scale in frequency response plots is shown in wavelength, and measured in distance units which are referred to as 'meters' for convenience. In addition, the DC component of the signal should be near zero because the mean value was subtracted. Consequently, the magnitude of the FFT is not adjusted at the Nyquist or DC frequency. Additionally, it is obvious that the unwrapped contours are comparatively flat for the smooth milk powder cone (Class 0 milk powder sample), but there is slight fluctuation for the shorter contours (higher layers). The top of the frequency response in Figure 12a also strengthens the assumption. However, the unwrapped contour trends are far more variable for the rough (Class 3) milk powder cone in Figure 12b. It is expected that the lumpiness might drop to the bottom of the milk powder sample. However, the deviation plot in Figure 12b tends not to demonstrate this. Figure 13a,b separately compare the contours, deviations, and frequency response of Class 0-3 instant whole and trim milk powder samples at layer 20 (approximately the middle of the cone). To clearly show the shape of each contour, the radii of Class 1-3 are enlarged appropriately (the radii of these four samples are very similar) in the contour plots of Figure 13. Additionally, the axis is truncated in the frequency response plots of Figure 13 since the high frequency (low wavelength) has little useable information. From Figure 13a,b, the Class 3 milk powder sample shows more energy at wavelengths (l) around 10 distance units, and the milk powder samples with the rougher surface show more energy at the higher wavelengths, which is expected. Furthermore, Figure 14a,b, respectively, compare the frequency response of Class 0-3 samples with a linear scale because the differences in frequency response are reduced due to the logarithmic scale in Figure 13a,b. The Q values (the energy of the signal) of each sample in Figure 14a,b are computed by Equation (2), and as expected, the milk powder samples with smoother surfaces have lower Q values. This result proves that it is appropriate to use Q values as the tool to categorize the roughness of the milk powder samples. 4.5. Classification of the Surface Roughness Since the Q value of each layer can be calculated, the Q values of 40 layers for each milk powder sample were used to build the classifier to categorize the surface roughness of milk powders. Additionally, since each moisture level has 6 milk power cones (three instant trim cones and three instant whole cones), this study has 24 milk powder cones in total. Figure 15 presents the results of the nonlinear SVM classifier. The blue parts in the confusion chart represent the true positive and true negative predictions, and the false negatives and false positives from the predictions are shown as reddish cells. Table 2 lists the specificities and sensitivities of the developed SVM model. For the performance of the model developed in , only one milk powder sample was wrongly predicted, while from these results, all the smooth milk powder samples (Class 0) and the extremely rough-surface milk powder samples (Class 3) were predicted correctly. However, one Class 1 sample was incorrectly classified as a Class 0 sample, and two Class 2 milk powder samples were incorrectly classified as a Class 0 and a Class 1 sample, respectively. This may be because the lumps (roughness) in these three samples were not too distinct. Additionally, the overall accuracy of this classifier is 87.5% which is close to the overall accuracy of the classifier developed in (the overall accuracy is about 88%). Furthermore, since the sample size (only eight samples) of the classifier developed in is smaller than the sample size (24) of the classifier developed in this study, the reliability of this classifier is better than the reliability of the classifier developed in . In addition, the loss (mean squared error) obtained by the cross-validated regression model is about 0.17, so that the accuracy (1--loss) of the cross validation is approximately 0.83, which means that this classifier performs well in the four-fold cross validation. Consequently, it is feasible to use this technique as a preliminary means to classify the milk powders into various surface roughness grades. 5. Conclusions This study investigated the application of three-dimensional digital photogrammetry to classify the surface roughness of milk powder. The technique proposed in this study is objective and is an alternative to the traditional manual surface roughness classification methods. Different from the 3D image analysis methods used in , which classify the surface smoothness of milk powder samples by comparing the area of triangles formed by the three adjacent surface normals as well as the angles between the adjacent surface normals, the 3D digital photogrammetry techniques proposed in this study were used to classify the surface roughness of milk powder samples by comparing the variances and frequency responses of each contour slices between different milk powder samples. However, this research only considered four surface roughness classes. To improve the robustness of the classifier, more milk powder samples and surface roughness classes are recommended. From the results of the method proposed in this study, a higher standard deviation was observed on the rough surface, while the smooth surface had a low standard deviation. Furthermore, the milk powder samples with rougher surfaces had higher Q values (the energy of the signal), while the milk powder samples with smoother surfaces had lower Q values. Finally, the performance of the nonlinear SVM classifier demonstrated that the 3D image processing technique developed is a practicable alternative technique for classifying the surface roughness of milk powders. Acknowledgments The authors would like to acknowledge Irina Boiarkina and Rachel Traill from Fonterra Co-Operative Group Limited, for their advice and support. Author Contributions Conceptualization, W.Y. and D.I.W.; methodology, H.D. and D.I.W.; software, H.D. and D.I.W.; validation, H.D. and B.R.Y.; formal analysis, H.D. and W.Y.; investigation, H.D. and X.C.; resources, H.D.; data curation, H.D.; writing--original draft preparation, H.D.; writing--review and editing, H.D., B.R.Y., D.I.W., X.C. and W.Y.; visualization, H.D. and D.I.W.; supervision, W.Y. and B.R.Y.; project administration, W.Y. and B.R.Y. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement All related data and methods are presented in this paper. Additional inquiries should be addressed to the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 A calibration chart used to differentiate specific reference points. Figure 2 (a) The diagram of the milk powder delivery device; (b) A plan view of the photogrammetry equipment and sample stage. Figure 3 An example picture of a cone. Figure 4 An oblique view of the triangulated model of a milk powder cone sliced at 40 equally spaced contours (only 10 slices are diagrammatically shown in this figure for clarity). Figure 5 The front views of Class 0 (a), Class 1 (b), Class 2 (c), Class3 (d) instant trim milk powder samples and Class 0 (e), Class 1 (f), Class 2 (g), Class 3 (h) instant whole milk powder samples. Figure 6 An 3D reconstruction triangular mesh model of a milk powder sample. Figure 7 (a) An unwrapped contour of Class 0 milk powder sample at layer 20; (b) an unwrapped contour of Class 3 milk powder sample at layer 20. Figure 8 The 3D digital model top views of Class 0-3 milk powder samples. Figure 9 (a) All contours for a Class 0 milk powder sample; (b) all contours for a Class 3 milk powder sample. Figure 10 The unwrapped contours of Class 0-3 milk powder samples. Figure 11 The deviations and the standard deviation of Class 0-3 instant whole and trim milk powder samples. Figure 12 (a) The contours, deviations, and frequency response of a Class 0 sample; (b) the contours, deviations, and frequency response of a Class 3 sample. Figure 13 (a) Comparing the contours, deviations, and frequency response at layer 20 across the Class 0-3 instant trim samples; (b) comparing the contours, deviations, and frequency response at layer 20 across the Class 0-3 instant whole samples. Figure 14 (a) Comparing the frequency response of the Class 0-3 instant trim samples; (b) comparing the frequency response of the Class 0-3 instant whole samples. Figure 15 The confusion chart for the developed SVM classifier. foods-12-00967-t001_Table 1 Table 1 The actual moisture level. Milk Powder Type Class 0 Class 1 Class 2 Class 3 Instant Trim Milk Powder 5.91 +- 0.27% 8.19 +- 0.45% 9.79 +- 0.66% 11.98 +- 0.96% Instant Whole Milk Powder 5.18 +- 0.23% 7.74 +- 0.51% 9.44 +- 0.64% 11.02 +- 0.87% foods-12-00967-t002_Table 2 Table 2 The specificities and sensitivities of the developed SVM model. Class Sensitivity Specificity Class 0 100% 75% Class 1 83.3% 83.3% Class 2 66.7% 100% Class 3 100% 100% Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Ding H. Li B. Boiarkina I. Wilson D.I. Yu W. Young B.R. Effects of Morphology on the Bulk Density of Instant Whole Milk Powder Foods 2020 9 1024 10.3390/foods9081024 32751793 2. Traill R. Thomasen L. Turnbull R. Coolbear T. The visual measurement of the degree of clumping of dairy powders using photo standards Int. Dairy J. 2022 127 105198 10.1016/j.idairyj.2021.105198 3. Traill R.M. Luckman M.S. Fisk N.A. Peng M. Application of the Rate-All-That-Apply (RATA) method to differentiate the visual appearance of milk powders using trained sensory panels Int. Dairy J. 2019 97 230 237 10.1016/j.idairyj.2019.05.013 4. Ding H. Yu W. Boiarkina I. Depree N. Young B.R. Effects of morphology on the dispersibility of instant whole milk powder J. Food Eng. 2020 276 109841 10.1016/j.jfoodeng.2019.109841 5. Pisecky J. Handbook of Milk Powder Manufacture GEA Process Engineering A/S Skanderborg, Denmark 2012 6. Crowley S.V. Desautel B. Gazi I. Kelly A.L. Huppertz T. O'Mahony J.A. Rehydration characteristics of milk protein concentrate powders J. Food Eng. 2015 149 105 113 10.1016/j.jfoodeng.2014.09.033 7. Massot-Campos M. Oliver-Codina G. Ruano-Amengual L. Miro-Julia M. Texture analysis of seabed images: Quantifying the presence of posidonia oceanica at palma bay Proceedings of the 2013 MTS/IEEE OCEANS-Bergen Bergen, Norway 10-13 June 2013 1 6 8. Garcia P. Petrou M. Kamata S.-I. The use of Boolean model for texture analysis of grey images Comput. Vis. Image Underst. 1999 74 227 235 10.1006/cviu.1999.0760 9. Tsai D.-M. Hsieh C.-Y. Automated surface inspection for directional textures Image Vis. Comput. 1999 18 49 62 10.1016/S0262-8856(99)00009-8 10. Weszka J.S. Rosenfeld A. An application of texture analysis to materials inspection Pattern Recognit. 1976 8 195 200 10.1016/0031-3203(76)90039-X 11. Al-Kindi G. Baul R. Gill K. An application of machine vision in the automated inspection of engineering surfaces Int. J. Prod. Res. 1992 30 241 253 10.1080/00207549208942892 12. Gupta M. Raman S. Machine vision assisted characterization of machined surfaces Int. J. Prod. Res. 2001 39 759 784 10.1080/00207540010011045 13. Kiran M. Ramamoorthy B. Radhakrishnan V. Evaluation of surface roughness by vision system Int. J. Mach. Tools Manuf. 1998 38 685 690 10.1016/S0890-6955(97)00118-1 14. Lee B. Juan H. Yu S. A study of computer vision for measuring surface roughness in the turning process Int. J. Adv. Manuf. Technol. 2002 19 295 301 10.1007/s001700200038 15. Liang J. Gu X. Deng H. Ni F. Detecting device and technology of pavement texture depth based on high precision 3D laser scanning technology IOP Conference Series: Materials Science and Engineering IOP Publishing Bristol, UK 2019 012063 16. Hola J. Sadowski L. Reiner J. Stach S. Usefulness of 3D surface roughness parameters for nondestructive evaluation of pull-off adhesion of concrete layers Constr. Build. Mater. 2015 84 111 120 10.1016/j.conbuildmat.2015.03.014 17. Li K. Wang J. Qi D. Damage Diagnosis of Reactive Powder Concrete under Fatigue Loading Using 3D Laser Scanning Technology Algorithms 2019 12 260 10.3390/a12120260 18. Ohno K. Date H. Kanai S. Study on Real-Time Point Cloud Superimposition on Camera Image to Assist Environmental Three-Dimensional Laser Scanning Int. J. Autom. Technol. 2021 15 324 333 10.20965/ijat.2021.p0324 19. Schenk T. Introduction to photogrammetry Ohio State Univ. Columb. 2005 106 2005 20. Waltenberger L. Rebay-Salisbury K. Mitteroecker P. Three-dimensional surface scanning methods in osteology: A topographical and geometric morphometric comparison Am. J. Phys. Anthropol. 2021 174 846 858 10.1002/ajpa.24204 33410519 21. James M.R. Robson S. Smith M.W. 3-D uncertainty-based topographic change detection with structure-from-motion photogrammetry: Precision maps for ground control and directly georeferenced surveys Earth Surf. Process. Landf. 2017 42 1769 1788 10.1002/esp.4125 22. Abadie A. Boissery P. Viala C. Georeferenced underwater photogrammetry to map marine habitats and submerged artificial structures Photogramm. Rec. 2018 33 448 469 10.1111/phor.12263 23. Stathopoulou E. Remondino F. Semantic photogrammetry: Boosting image-based 3D reconstruction with semantic labeling Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2019 42 W9 10.5194/isprs-archives-XLII-2-W9-685-2019 24. Caravaca G. Le Mouelic S. Rapin W. Dromart G. Gasnault O. Fau A. Newsom H.E. Mangold N. Le Deit L. Maurice S. Long-Distance 3D Reconstructions Using Photogrammetry with Curiosity's ChemCam Remote Micro-Imager in Gale Crater (Mars) Remote Sens. 2021 13 4068 10.3390/rs13204068 25. Bertin S. Friedrich H. Field application of close-range digital photogrammetry (CRDP) for grain-scale fluvial morphology studies Earth Surf. Process. Landf. 2016 41 1358 1369 10.1002/esp.3906 26. Peter Heng B. Chandler J.H. Armstrong A. Applying close range digital photogrammetry in soil erosion studies Photogramm. Rec. 2010 25 240 265 10.1111/j.1477-9730.2010.00584.x 27. Rieke-Zapp D.H. Nearing M.A. Digital close range photogrammetry for measurement of soil erosion Photogramm. Rec. 2005 20 69 87 10.1111/j.1477-9730.2005.00305.x 28. Moret-Fernandez D. Latorre B. Pena C. Gonzalez-Cebollada C. Lopez M. Applicability of the photogrammetry technique to determine the volume and the bulk density of small soil aggregates Soil Res. 2016 54 354 359 10.1071/SR15163 29. Merel A. Farres P. The monitoring of soil surface development using analytical photogrammetry Photogramm. Rec. 1998 16 331 345 10.1111/0031-868X.00129 30. Belmonte A. Biong M. Macatulad E. DEM generation from close-range photogrammetry using extended python photogrammetry toolbox Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2017 42 11 19 10.5194/isprs-archives-XLII-4-W5-11-2017 31. Omar H. Mahdjoubi L. Kheder G. Towards an automated photogrammetry-based approach for monitoring and controlling construction site activities Comput. Ind. 2018 98 172 182 10.1016/j.compind.2018.03.012 32. Fabris M. Pesci A. Automated DEM extraction in digital aerial photogrammetry: Precisions and validation for mass movement monitoring Ann. Geophys. 2005 48 Available online: (accessed on 12 February 2023) 10.4401/ag-3247 33. Geladi P. Some special topics in multivariate image analysis Chemom. Intell. Lab. Syst. 1992 14 375 390 10.1016/0169-7439(92)80119-O 34. Indahl U.G. Naes T. Evaluation of alternative spectral feature extraction methods of textural images for multivariate modelling J. Chemom. A J. Chemom. Soc. 1998 12 261 278 10.1002/(SICI)1099-128X(199807/08)12:4<261::AID-CEM513>3.0.CO;2-Z 35. Tomita F. Tsuji S. Computer Analysis of Visual Textures Springer Science & Business Media Berlin, Germany 2013 Volume 102 36. Pugliese A. Cabassi G. Chiavaro E. Paciulli M. Carini E. Mucchetti G. Physical characterization of whole and skim dried milk powders J. Food Sci. Technol. 2017 54 3433 3442 10.1007/s13197-017-2795-1 29051638 37. Sharma A. Jana A.H. Chavan R.S. Functionality of milk powders and milk-based powders for end use applications--A review Compr. Rev. Food Sci. Food Saf. 2012 11 518 528 10.1111/j.1541-4337.2012.00199.x 38. Davenel A. Schuck P. Mariette F. Brule G. NMR relaxometry as a non-invasive tool to characterize milk powders Le Lait 2002 82 465 473 10.1051/lait:2002024 39. Nijdam J. Langrish T. An investigation of milk powders produced by a laboratory-scale spray dryer Dry. Technol. 2005 23 1043 1056 10.1081/DRT-200060208 40. Lee J. Chai C. Park D.J. Lim K. Imm J.-Y. Novel convenient method to determine wettability and dispersibility of dairy powders Korean J. Food Sci. Anim. Resour. 2014 34 852 10.5851/kosfa.2014.34.6.852 26761684 41. Carrasco A. Siebert K.J. Human visual perception of haze and relationships with instrumental measurements of turbidity. Thresholds, magnitude estimation and sensory descriptive analysis of haze in model systems Food Qual. Prefer. 1999 10 421 436 10.1016/S0950-3293(99)00029-4 42. Jeon S.-S. Ganesan P. Lee Y.-S. Yoo S.-H. Kwak H.-S. Texture and sensory properties of cream cheese and cholesterol-removed cream cheese made from whole milk powder Food Sci. Anim. Resour. 2012 32 49 53 10.5851/kosfa.2012.32.1.49 43. Gosselin R. Duchesne C. Rodrigue D. On the characterization of polymer powders mixing dynamics by texture analysis Powder Technol. 2008 183 177 188 10.1016/j.powtec.2007.07.021 44. Lille M. Kortekangas A. Heinio R.-L. Sozer N. Structural and textural characteristics of 3D-printed protein-and dietary fibre-rich snacks made of milk powder and wholegrain rye flour Foods 2020 9 1527 10.3390/foods9111527 33114227 45. Gemmi M. Voltolini M. Ferretti A.M. Ponti A. Quantitative texture analysis from powder-like electron diffraction data J. Appl. Crystallogr. 2011 44 454 461 10.1107/S0021889811012106 46. Oghazi P. Palsson B. Tano K. Applying traceability to grinding circuits by using Particle Texture Analysis (PTA) Miner. Eng. 2009 22 710 718 10.1016/j.mineng.2009.01.017 47. Pang Z. Deeth H. Sopade P. Sharma R. Bansal N. Rheology, texture and microstructure of gelatin gels with and without milk proteins Food Hydrocoll. 2014 35 484 493 10.1016/j.foodhyd.2013.07.007 48. Ding H. Wilson D.I. Yu W. Young B.R. Assessing and Quantifying the Surface Texture of Milk Powder Using Image Processing Foods 2022 11 1519 10.3390/foods11101519 35627090 49. De Knegt R. Van Den Brink H. Improvement of the drying oven method for the determination of the moisture content of milk powder Int. Dairy J. 1998 8 733 738 10.1016/S0958-6946(97)00110-6 50. Yang J. Huang M. Peng J. Shi J. Rapid determination of the moisture content of milk powder by microwave sensor Measurement 2016 87 83 86 10.1016/j.measurement.2016.03.012 51. Reljic I. Dunder I. Application of Photogrammetry in 3D Scanning of Physical Objects TEM J. 2019 8 94 52. Fritsch D. Klein M. 3D preservation of buildings-Reconstructing the past Multimed. Tools Appl. 2018 77 9153 9170 10.1007/s11042-017-4654-5 53. Hellmuth R. Wehner F. Giannakidis A. Datasets of captured images of three different devices for photogrammetry calculation comparison and integration into a laserscan point cloud of a built environment Data Brief 2020 33 106321 10.1016/j.dib.2020.106321 33088869 54. Reljic I. Dunder I. Seljan S. Photogrammetric 3D scanning of physical objects: Tools and workflow TEM J. 2019 8 383 55. Huang M.-W. Chen C.-W. Lin W.-C. Ke S.-W. Tsai C.-F. SVM and SVM ensembles in breast cancer prediction PLoS ONE 2017 12 e0161501 10.1371/journal.pone.0161501 28060807 56. Pavithra S. Janakiraman S. Enhanced polynomial kernel (EPK)-based support vector machine (SVM) (EPK-SVM) classification technique for speech recognition in hearing-impaired listeners Concurr. Comput. Pract. Exp. 2021 33 e5210 10.1002/cpe.5210 57. Kohavi R. Provost F. Glossary of terms journal of machine learning Mach. Learn 1998 30 271 274 58. Jiang N. Liu H. Understand system's relative effectiveness using adapted confusion matrix Proceedings of the International Conference of Design, User Experience, and Usability Las Vegas, NV, USA 21-26 July 2013 294 302 59. Ding H. Wilson D.I. Yu W. Young B.R. An investigation of the relative impact of process and shape factor variables on milk powder quality Food Bioprod. Process. 2021 126 62 72 10.1016/j.fbp.2020.12.010 60. Khan A. Munir M.T. Yu W. Young B. Wavelength Selection FOR Rapid Identification of Different Particle Size Fractions of Milk Powder Using Hyperspectral Imaging Sensors 2020 20 4645 10.3390/s20164645 32824764 |
PMC10000611 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050824 diagnostics-13-00824 Systematic Review Application of Artificial Intelligence Techniques for Monkeypox: A Systematic Review Chadaga Krishnaraj 1 Prabhu Srikanth 1* Sampathila Niranjana 2* Nireshwalya Sumith 3 Katta Swathi S. 4 Tan Ru-San 56 Acharya U. Rajendra 789 El-Baz Ayman Academic Editor 1 Department of Computer Science and Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal 576104, India 2 Department of Biomedical Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal 576104, India 3 Department of Information and Communication Technology, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal 576104, India 4 Manipal Institute of Management, Manipal Academy of Higher Education, Manipal 576104, India 5 Department of Cardiology, National Heart Centre Singapore, Singapore 168752, Singapore 6 Duke-NUS Medical School, Singapore 169857, Singapore 7 Ngee Ann Polytechnic, Department of Electronics and Computer Engineering, Singapore 599489, Singapore 8 Department of Biomedical Engineering, School of Science and Technology, SUSS University, Singapore 599494, Singapore 9 Department of Biomedical Informatics and Medical Engineering, Asia University, Taichung 40444, Taiwan * Correspondence: [email protected] (S.P.); [email protected] (N.S.) 21 2 2023 3 2023 13 5 82431 1 2023 17 2 2023 20 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Monkeypox or Mpox is an infectious virus predominantly found in Africa. It has spread to many countries since its latest outbreak. Symptoms such as headaches, chills, and fever are observed in humans. Lumps and rashes also appear on the skin (similar to smallpox, measles, and chickenpox). Many artificial intelligence (AI) models have been developed for accurate and early diagnosis. In this work, we systematically reviewed recent studies that used AI for mpox-related research. After a literature search, 34 studies fulfilling prespecified criteria were selected with the following subject categories: diagnostic testing of mpox, epidemiological modeling of mpox infection spread, drug and vaccine discovery, and media risk management. In the beginning, mpox detection using AI and various modalities was described. Other applications of ML and DL in mitigating mpox were categorized later. The various machine and deep learning algorithms used in the studies and their performance were discussed. We believe that a state-of-the-art review will be a valuable resource for researchers and data scientists in developing measures to counter the mpox virus and its spread. monkeypox mpox artificial intelligence machine learning deep learning This research received no external funding. pmc1. Introduction Monkeypox or Mpox disease, a smallpox-like illness caused by the mpox virus , originated primarily in the rainforests of Central and West Africa . However, since the recent outbreak in May 2022 , it has spread to many countries, threatening to be a global epidemic. The virus is transmitted by contact with bodily fluids and respiratory droplets ; infected persons typically manifest symptoms for a few weeks, which include fever, swollen lymph nodes, and body rash . The disease is self-limiting in most affected individuals and requires only symptomatic management. However, severe medical complications can occur in some patients , which leads to fatality in 3-5% of cases . Currently, no specific drug targets the mpox virus. Instead, therapies developed for treating smallpox in adults, including antivirals and vaccinia immune globin, are used to manage severe mpox infection . The emergence of mpox amid the devastating coronavirus disease 2019 pandemic has triggered a global alert and galvanized efforts toward a scientific reappraisal of the mpox virus. Machine learning (ML) is a subdomain of artificial intelligence (AI) and has significantly altered the healthcare infrastructure. ML and deep learning (DL) algorithms can be used on the acquired data to identify previously hidden information, monitor the patient's health to detect, and alert about life-threatening conditions . They employ scientific and mathematical methodologies for generating new information from the data. This helps to obtain accurate and robust diagnosis systems. X-rays, magnetic resonance imaging, computed tomography, and other modalities generate high-resolution images, which can sometimes be challenging to understand, even for an experienced radiologist. ML has already demonstrated that it can be highly beneficial for pathologists and radiologists as it always yields highly accurate and efficient diagnosis models . Such automated systems have been used for cardiovascular abnormalities , fractures , neurological diseases , cancers , and many more . ML and DL models are also used in drug and vaccine design. The pathogenesis, etiology, transmission, clinical features, diagnosis, and management of mpox infection have been comprehensively reviewed recently . Other authors focused on the science of mpox disease prevention and treatment , including candidate antiviral drugs and immunological strategies . Artificial intelligence (AI)-based approaches are increasingly harnessed to amplify research efficiency to develop countermeasures against mpox. Patel et al. reviewed various AI approaches employed for detecting mpox virus, viral genome characterization, monitoring of spread, and prognosticating mortality risk. Gul et al. reviewed various diagnostic methods for mpox, including image recognition, immunodiagnostics, nucleic acid, and whole-particle detection. Our article reviewed various AI applications for mpox diagnosis, forecasting, drug discovery, and other use cases. Figure 2 shows a summary of reviews published on AI for mpox-related applications, but they were not systemic reviews. Our study is a systematic study focusing on the application of applying machine learning (ML) and deep learning (DL) techniques for the automated detection of mpox. We believe a state-of-the-art review will be a valuable resource for researchers and data scientists in developing measures to counter the mpox virus and its spread. The contributions of our systematic review are as follows:Various diagnostics techniques using machine learning and deep learning for mpox detection are systematically reviewed. Articles that used AI to discover new vaccines and drugs for mpox are also reviewed. Other AI studies, such as epidemiological modeling of mpox infection spread and web management of information on mpox, are also included. A thorough discussion regarding the above applications for mpox is provided. Challenges and directions for future mpox research using AI are also discussed. Due to the sudden advent of rampant infections, such as COVID-19 and Mpox, quick and rapid identification of patients without contact with hospital personnel is required . Hence, various automated diagnostic strategies for mpox viruses are compiled in this review. These detection techniques can prevent the transmission of mpox by avoiding contact between patients and healthcare workers. Furthermore, several drugs and vaccines are being created to combat the virus . Hence, we also reviewed various articles which used AI for drug and vaccine development. Other applications of AI, such as forecasting mpox cases and Twitter sentiment analysis, are also included in this review. The remainder of the article is structured as follows. Section 2 contains the review methodology. AI applications in mpox diagnosis, which comprises gene/proteomic analyses, image recognition, and clinical data-mining approaches, are presented in Section 3. Section 4 provides other AI applications, vaccine/drug discovery, forecasting of mpox cases, and countering of social engineering. The results are discussed in Section 5. Section 6 describes the limitations and future trends. Lastly, the paper concludes in Section 7. The structure of the article is described in Figure 3. 2. Review Methodology A systematic literature review gathers and summarizes several research findings to analyze the work conducted by others by following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines . It involves steps such as identifying the related literature, synthesizing the findings, tabulating similar studies, and drawing various conclusions from the research. Using PRISMA methodology , we searched for all full-text English language articles related to AI, ML, and DL applications in human mpox research published up to 1 January 2023 in standard databases (Scopus, Google Scholar, Web of Science, PubMed, Medline, Crossref, Arxiv, and Embase) and archives of publishers (Elsevier, Springer Nature, IEEE, MDPI, Taylor and Francis, Wiley, and Arxix). We used the keyword strings "artificial intelligence and monkeypox", "machine learning and monkeypox", "deep learning and monkeypox", "natural language processing and monkeypox", "data science and monkeypox", "regression analysis and monkeypox", "monkeypox cases forecasting", and their permutations; we also included substitutions of the term "monkeypox" by "monkeypox" and "mpox". All preselected articles were screened by the first author (K.C.), who excluded animal research, case reports, letters to editors, commentaries, short reports, methods papers without results, and articles that were assessed to not actually employ AI, ML, or DL methods. A total of 1523 articles were preselected, later pruned to 34 articles after eliminating duplicate works, filtering by article titles, keywords, and abstracts, and rejection of full-text content unrelated to AI, ML, or DL . Since the mpox outbreak is relatively new and has lasted for a short period, only a few articles have been published. The final selection comprised 20, five, five, and four articles on mpox diagnosis, drug/vaccine discovery, forecasting of mpox cases, and sentiment analysis, respectively . All the above papers were carefully chosen after reading the entire content. The number of articles that included mpox detection using AI were comparatively higher than other applications. 3. Artificial Intelligence in Mpox Diagnosis Polymerase chain reaction amplification of viral genetic material from skin vesicular fluid samples can provide diagnostic confirmation of mpox infection. Nevertheless, manual analysis of the genetic sequence readouts requires expertise and is time-consuming. In , deoxyribonucleic acid sequences of mpox and human papilloma viruses (n = 55 each), which manifest similar skin lesions, were numerically encoded and standardized to equal lengths using zero padding before being input into a bidirectional long short-term memory-based model for binary classification. After optimizing model parameters by trial and error, accuracy, precision, recall, and F1-score of 99.5%, 100%, 100%, and 99.95%, respectively, were attained. In , the authors studied the plasma proteome of a group of patients with clinical symptoms suggestive of mpox infection. There was some overlap of protein biomarkers with coronavirus infection 2019, but differences also existed that offered diagnostic clues to mpox infection. In the absence of laboratory polymerase chain reaction-based examination, mpox infection is clinically diagnosed by expert examination of the characteristic skin lesions. Many researchers have published the use of AI to facilitate real-time image recognition of mpox skin lesions (Table 1) . To address the dearth of publicly available training image datasets for mpox research, Ahsan et al. built a dataset named "Monkeypox2022", comprising images of skin lesions caused by infection with mpox, chickenpox, and measles viruses, as well as normal controls, which were collated from multiple open-source online portals. Applying transfer learning to "Monkeypox2022", the authors used a pretrained modified VGG-16, a type of convolutional neural network (CNN), to classify mpox from other classes, attaining accuracy rates ranging from 78% to 97%. Similarly, Ali et al. created the "Monkeypox skin lesion dataset", containing images of skin lesions in mpox, chickenpox, and measles infections. They used data augmentation to increase the sample size. Among four models--InceptionV3, ResNet50, VGG-16, and an ensemble of all three--ResNet50 outperformed the rest, attaining average accuracy of 82.96%. Abdelhamid et al. developed an image classification algorithm, "AI-Biruni-Earth-Radius". Using GoogLeNet deep neural network to extract features, they attained a maximum accuracy of 98.8% for mpox detection in a multiclass dataset. In , among multiple classifiers used to distinguish mpox from other skin lesions, the naive Bayes algorithm attained optimal 91% accuracy, outperforming other CNNs and shallow classifiers: GoogLeNet, VGG-16, AlexNet, k-nearest neighbor, support vector machine, random forest, and decision tree. Sitaula et al. compared 13 different pretrained models with the help of transfer learning on a mpox dataset. They developed a CNN-based model to classify the skin lesion into eight disease classes and compared their solution with the help of the VGG-16 pretrained models. The accuracy, average precision, recall, and F1-score obtained for classifying mpox images were 87%, 85%, 85%, and 85%, respectively. Sahin et al. developed a mobile application to detect mpox from video images of skin lesions captured and uploaded using Android phones. The CNN-based model, which was trained and tested using Matlab software embedded with TensorFlow and TensorFlow Lite, attained the best accuracy of 91.11%. In , the authors studied seven advanced transfer learning models for classifying 804 digitized skin images obtained from cases of mpox, smallpox, chickenpox, cowpox, and measles, as well as healthy subjects. They reported a mean precision of 85% but a poor recall of 58%. It was observed that deep learning models tend to overfit or underfit, which may be attributed to the tradeoff between training sample size and the number of trainable model parameters. In , various CNN-based models such as ResNet50, MobileNet-V2, EfficientNet-B0, and VGG-19, as well as an ensemble of these models were trained on images of mpox skin lesions, other rashes, and normal subjects. The ensemble model attained the best accuracy of 98.33%. Explainable AI (XAI) offers insights into diagnostic model behavior and performance that enhance the interpretation , which can potentially garner a better clinical understanding. For example, Sizikova et al. used DL based on non-negative matrix factorization mutation to classify tuberculosis and mpox skin images. They incorporated XAI in the form of automatic concept discovery techniques, which made the model more accessible and interpretable. A maximum accuracy of 88.64% was obtained. XAI techniques were used in to diagnose mpox on skin lesion images. A total of 572 images and 12 different models were studied. The MobileNetV2 model obtained a maximum accuracy of 98.25% among all the classifiers. Ahsan et al. studied six deep learning models and obtained the best accuracy of 93% and 99% for MobileNetV2 and InceptionResNetV2, respectively, for discriminating mpox from normal skin images in an imbalanced dataset. The local interpretable model-agnostic explanations (LIME) method was incorporated into the model to comprehend and verify the model's predictions. The insights obtained with LIME could facilitate refining the model and developing and evaluating other deep learning models, including those trained using imbalanced datasets. Alcala-Rmz et al. used a CNN model for mpox diagnosis. They implemented the MiniGoogLeNet for various epochs. The dataset comprised 2062 images, of which 1439 were chickenpox and measles images, and 1168 were mpox images. A maximum accuracy of 97.08% was attained when 50 epochs were used. Khafaga et al. used a deep CNN to classify mpox images. The dataset from Kaggle consisted of 293 normal, 279 mpox, 107 chickenpox, and 91 measles cases. A maximum accuracy of 98.83% was attained. Haque et al. used attention mechanisms and deep learning models to classify mpox in humans. Five deep learning models were implemented, and a maximum accuracy of 84% was attained. Saleh et al. used AI and data mining techniques to classify mpox. An improved binary chimp optimization was used for feature selection. The maximum accuracy, precision, and recall of 98.48%, 91.11%, and 89%, respectively, were obtained. The transfer of sensitive medical images raises privacy concerns. In , a secure data aggregation scheme was proposed to minimize cyber threats. Instead of transmitting raw data, blockchain-based data acquisition and federated learning were employed to assemble the data in the form of trained models. Using ResNet18 to perform binary classification of trained models constructed from images of skin lesions of mpox versus other conditions, the scheme attained 99.81% accuracy, which demonstrated the feasibility of fully secure and remote automated image-based mpox diagnosis. Joshua et al. used a neuro-fuzzy model to detect the mpox virus. Their architecture, "MDiNFIS", embodied both hardware and software. Leveraging the uncertainty-handling capability of fuzzy logic systems and the learning capability of artificial neural networks, they built a diagnostic system for mpox detection. diagnostics-13-00824-t001_Table 1 Table 1 Summary of published artificial intelligence-based mpox classification models using skin images. Paper Images (n) Classifier (s) Acc (%) Pre (%) Rec (%) F1 (%) Ahsan Mpox, chickenpox, measles, normal VGG-16 97 97 97 97 Ali Mpox (102), others (126) VGG-16, ResNet50, Inception-V3, ensemble 82.96 85 81 83 Abdelhamid Mpox (279), chickenpox (107), measles (91), normal (293) AlexNet, VGG-19, GoogLeNet, ResNet50 98.8 - 63 74 Kumar - AlexNet, GoogLeNet, VGG-16, support vector machine, k-nearest neighbor, naive Bayes, decision tree, random forest 91.11 - - - Sitaula Mpox, chickenpox, measles, normal 13 deep learning models. 87.13 85.44 85.47 85.4 Sahin Mpox (102), others (126) Modified MobilNetV2 91.11 - - - Islam Mpox, chickenpox, smallpox, cowpox, measles, normal ResNet50, DenseNet121, Inception-V3, SqueezeNet, MobileNet-V2, ShuffleNet-V2, ensemble 83 - 58 - Saavedra Mpox (100), other rashes (100), normal (100) VGG-16, VGG-19, ResNet-50, MobileNet-V2, EfficientNet, ensembles 98.33 - - - Sizikova Mpox, tuberculosis VGG-16, EfficientNet-B3 88.64 - - - Akin Mpox (252), others (264) 12 deep learning algorithms 98.25 - - 98.25 Ahsan Mpox, normal 6 deep learning models 99 - - - Alcala-Rmz Mpox, control MiniGoogLeNet 97.08 - - - Khafaga Mpox, control Al-Biruni Earth radius optimization-based stochastic fractal search 98.83 - 85 80 Haque Mpox, others (total 2142) 5 deep learning algorithms 84 90 89 90 Saleh Mpox (296), others (204) Various machine learning and deep learning models 98.48 91.11 89 - Islam Mpox (1428), others (1764) ResNet18, block-chain federated learning for privacy protection 99.81 - - - Acc, accuracy; F1, F1-score; Pre, precision; Rec, recall. 4. Other Applications of AI in Combating Mpox 4.1. Epidemiological Modeling of Mpox Infection Spread Researchers have used AI models to predict mpox outbreaks (Table 2). Arotolu et al. used a maximum entropy algorithm to model the environmental variables in 116 spatially unique cases of prior mpox infections from 2017 to 2021 in Nigeria. The model--the top five features being precipitation, human population density, elevation, and maximum and minimum temperature--accurately predicted (area under the curve of 92%) conditions and geographies conducive to mpox spread, facilitating resource distribution to at-risk regions in the country. Majumder et al. trained a polynomial neural network on mpox case data collected between 6 May 2022 and 28 July 2022 to develop a predictive model that could forecast mpox cases developing over the next 100 days. In , a proposed "BER-LSTM" model based on long short-term memory (LSTM) network with the hyperparameters tuned using the "AI-Biruni Earth Radius" algorithm was proposed to predict mpox disease spread. Incorporating statistical methods prior to training, such as analysis of variance, regression, and Wilcoxon tests, the hybrid algorithm attained a low mean bias error of 0.06%. Yasmin et al. developed a prediction model that used regression analysis to forecast mpox outbreak. Among nine algorithms, the model yielded the best performance with mean absolute error and root-mean-square error of 146.29 and 204.75, respectively. For forecasting mpox cases, Qureshi et al. compared various time series AI models such as autoregressive integrated moving average (ARIMA), extreme machine learning, support vector machine, and multilayer perceptron. The latter was found to be the most reliable. 4.2. Candidate Vaccine and Drug Design Multi-epitope vaccines have the potential to generate specific immunogenic responses based on conserved epitopes in complete antigenic sequences , thus avoiding responses against unfavorable epitopes that might induce immunopathogenic or immune-modulating responses against the host. Using the Virus Pathogen Database and Analysis Resource Database to retrieve mpox proteins, nine overlapping epitopes were chosen to create multi-epitope vaccination constructs associated with appropriate linkers and various adjustments to improve the immune system responses. In , 176 genome-encoded protein structures were screened as candidate mpox vaccines using immunoinformatics. The final model possessed excellent binding energy of 98.37% kcal/mol to the mpox virus. In , three extracellular antigenic proteins were studied using AI simulations of B and T lymphocyte immune response. The silico analysis was used to select the best candidate for further vaccine development against mpox. Altayb et al. focused on three mpox proteins that play crucial roles in viral replication; they used various in silico techniques such as molecular docking, computational modelling, and molecular dynamics simulations to digitally screen 1615 FDA-approved drugs for activity against these proteins . Fludarabine, an anticancer therapeutic, was discovered to have the optimal docking score (7.53 kcal/mol) for the mpox protein. Lam et al. used a DL server, AlphaFold, to define the structure of the 304-amino-acid mpox E8 protein. They discovered that diosmin and flavin adenine dinucleotide, both commercially available drugs, could potentially be repurposed to target the E8 protein, with maximum binding energies being 9.69 kcal/mol and 7.65 kcal/mol, respectively. 4.3. Web Management of Information on Mpox Natural language processing (NLP) is an important tool for collating accruing web information generated on specific topic labels. Kolluri et al. developed a browser extension named "POXVERIFI" to verify information about mpox on the internet. Their bidirectional encoder representations from the transformers (BERT) model attained 96% validation accuracy. As more users installed the extension and rate articles, crowd-sourced votes automatically generated accurate labels for new sources. The labeled data could then be leveraged to measure mpox-related misinformation. Mohbey et al. used a hybrid CNN-LSTM model to classify mpox tweets into positive, negative, and neutral classes. Their model attained an absolute accuracy of 94% for assessing the sentiments of Twitter users on mpox. Unsupervised ML was used to analyze 352,182 (duplicate tweets and retweets were excluded) Twitter posts on mpox in . The topics were clustered into three themes (stigmatization of minorities, safety concerns, and lack of faith in institutions), and transformers such as BERT and BERTopic were used to perform sentiment analysis. A total of 15,936 mpox tweets in German were analyzed in using a mixed-methods research methodology and ML. The authors opined that a multidisciplinary strategy could be used to minimize and prevent mpox-related misinformation. A summary of articles related to AI mpox and web information is given in Table 3. 5. Discussion Mpox emerged in early 2022 and has spread to many different countries. Polymerase chain reaction (PCR) is the gold standard for diagnosing this virus. However, PCR tests are prone to false-negative and erroneous results. They also consume substantial time and require trained medical professionals to conduct the test. In the first part of this comprehensive review, various ML and DL techniques used to diagnose mpox were screened. Most of the researchers used skin lesion images to detect the virus. Rashes generally appear on the patient skin, and these classifiers can utilize these features for accurate diagnosis. Similar diseases which cause skin rashes, such as measles, chickenpox, and smallpox, were present in the datasets. Hence, the AI algorithm diagnosis can be subjected to thorough investigation in the hospitals. The AI-based system can be deployed in various healthcare facilities to assist clinicians. This automated diagnosis can be useful in countries that lack infrastructure and medical technology. It can also be used to confirm the PCR tests. The DNA sequences were also used to diagnose mpox in one study. Most of the algorithms were able to obtain high accuracy in detecting mpox. The most widely used algorithm was CNN and its variants. Conventional ML algorithms such as support vector machine, K nearest neighbors, decision tree, and random forest were also utilized. We believe that all three methods should be used to diagnose mpox if available. Mpox detection using skin lesion images and deep learning models were highly effective since the models could easily distinguish between mpox and other similar diseases. The deep learning models can be used in parallel to the PCR tests to prevent false negative results. The most promising drugs were diosmin, flavin adenine dinucleotide, and fludarabine. Many studies used explainable artificial intelligence (XAI) for mpox detection. This enables the clinicians to understand and convince the decisions taken by the deep learning classifiers. LIME and gradient-weighted class activation map (Grad-CAM) can also be used to visualize the lesions in the images . Furthermore, we plan to explore the possibility of the model's authenticity when employed in a real-world scenario, as it can be influenced by noise. Hence, computing the uncertainty of the model will help AI specialists understand the reliability in a noisy environment . For trustworthy deep learning detection, progress in several areas, such as software engineering, XAI, and ML ops, is required. Medical validation must be performed by doctors to establish the reliability of the models. The algorithm's robustness must be measured, and the AI systems must be continuously audited. Cyberattacks in deep learning have also become common. Neural architectures must be protected from malicious users and threats. Model and data uncertainty issues must also be addressed. Furthermore, we reviewed other ML studies that can help us manage mpox. Much wrong information is being published online related to mpox, such as its infection rate and duration, symptoms, and preventive measures. NLP techniques were used to mitigate this problem. The public sentiments of users on mpox were also analyzed. The majority of the researchers used the BERT technique to perform sentiment analysis. Currently, there are no drugs or vaccines to treat mpox. Various drug designing and repurposing techniques using ML were also included in this review. Drugs such as fludarabine, diosmin, and flavin adenine dinucleotide (FAD) have shown promising results. It is important to discover new drugs to prevent severe virus symptoms. ML and DL models were also used to forecast the number of cases in various countries. LSTM was the most preferred algorithm for prediction. Models such as ARIMA and MLP were also utilized. Evaluation metrics such as mean absolute error (MAE), mean squared error (MSE), and R-squared were considered to understand the efficiency of the models. In total, 34 published articles on mpox and AI were reviewed. This systematic literature review can help researchers from AI, computer science, data science, and medicine understand various applications that can be used for a potential mpox outbreak. Figure 7 shows the various ML and DL algorithms in the reviewed studies for the different applications, with the use frequencies indicated. MobileNet was the most widely used DL algorithm for automated detection of mpox. Models such as VGG-16, ResNet, InceptionNet, DenseNet, and EfficientNet were also frequently used. ML classifiers such as k-nearest neighbor and support vector machine were used in a few studies. For specific applications such as mpox case forecast, drug discovery, and public sentiment analysis, models such as BERT, LSTM, ARIMA, and regression methodologies were employed. 6. Limitations and Future Directions 6.1. Limitations First, there is a shortage of open-source data for training AI models. Few public datasets are available for mpox, and most are small, limiting the use of data-hungry DL algorithms. AI applications require high data volume to avoid bias and overfitting. Hence, mpox data collection should be encouraged and publicly shared to facilitate research. The training of developed models on common open-source datasets will also allow a meaningful comparison of model performance. Second, data filtering is necessary to remove wrong mpox information and ensure data input fidelity into AI models. Third, sharing, especially of sensitive raw data, can potentially compromise privacy. Therefore, protecting sensitive information from malicious users and hackers is imperative. Data must be structured well for effective model training. With most of the online data being unstructured, it is hard to derive useful information despite abundant data. Deep learning algorithms are also prone to false negative and erroneous results. Hence, deep learning algorithms can be combined with other machine learning techniques such as unsupervised learning and reinforcement learning. From the above studies, we can see that most researchers used skin lesion images to diagnose mpox. However, it is important to look into other modalities such as blood tests and laboratory markers. When multiple modalities are used, the results are generally more reliable. The strains of some viruses continue to mutate after a certain amount of time. Hence, the models must also be tested on data consisting various mutations. The models must also be tested on various datasets from different geographical territories. 6.2. Future Directions We see the potential for a secure cloud-based system for comprehensive mpox management . Patient data, wearable devices, smartphone data, and AI models can be deployed on the cloud. Doctors can access these via the cloud infrastructure, including uploaded skin lesion images for remote mpox diagnosis. Clinical, laboratory, epidemiological, demographic, and other parameters can also be incorporated to allow systems-wide analysis for healthcare resource allocation. To prevent the infection from spreading, teleconsultation via remote video conferencing can also be integrated into the system, and appropriate symptomatic treatment can be dispensed remotely for mild cases. Such a system is scalable, and most of the researchers use datasets from a single source or geographical regions. With a cloud system, data from multiple locations can be gathered and analyzed to detect trends for high-level strategic planning. Furthermore, the correlation between dataset size and classifier performance can be quantified via sensitivity analysis . We may then analyze the study results to determine how much information is required and how small a dataset is needed to accurately anticipate performance on larger datasets. XAI methods (e.g., Shapley additive explanation, LIME, Eli5, and QLattice) can be incorporated to enhance the interpretability of the model outputs, which will help researchers and clinicians better understand the model predictions using different visualization techniques . Computational costs of the applications can also be computed and compared. 7. Conclusions The current Mpox outbreak is a cause for global concern. Although not as fatal as the coronavirus infection in 2019, it would be wise to prepare for worsening outcomes. In recent years, AI has significantly accelerated adoption in health science research and applications. Here, we comprehensively reviewed the latest AI-related methods applied to combat the Mpox virus. AI models for diagnosing Mpox, predicting Mpox outbreaks, vaccine and drug discovery, and other related applications were explored in depth. We believe that this review can help researchers and medical professionals understand the various AI applications in place, which can be further expanded or refined if the Mpox outbreak worsens. Key issues and future directions were also discussed. Acknowledgments We would like to thank Manipal Academy of Higher Education for facilitating this research. Author Contributions Conceptualization, K.C. and N.S.; methodology, S.P. and K.C.; software, K.C.; validation, U.R.A.; formal analysis, R.-S.T.; investigation, S.N.; resources, S.S.K.; data curation, K.C.; writing--original draft preparation, K.C.; writing--review and editing, N.S., R.-S.T. and U.R.A.; visualization, S.N. and K.C.; supervision, N.S. and S.P.; project administration, S.S.K. and U.R.A.; funding acquisition, N.S. and U.R.A. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Data will be made available on request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Mpox rash in the face (a), hand (b), shoulder (c), and feet (d) . Figure 2 Illustration of comparison with existing review papers published on mpox detection. Siegrist et al. , Rizk et al. , Lum et al. , Patel et al. and Gul et al. . Figure 3 Overview of the structure of this review article. Figure 4 Article selection procedure using PRISMA methodology. Figure 5 Articles selected in various categories related to mpox and AI. Figure 6 Deep learning-based system which can be deployed to diagnose mpox virus. Figure 7 Sunburst plot of AI studies using Mpox data. The inner layers depict various categories of applications; the outer layers depict the algorithms used, with each corresponding use frequency indicated by a numbered suffix. For example, MobileNet-7 indicates that MobileNet has been used for mpox diagnosis in seven studies. Figure 8 Cloud-based system for mpox management. diagnostics-13-00824-t002_Table 2 Table 2 Published artificial intelligence-based models for predicting the number of mpox cases. Paper Dataset Model Results Arotolu. 116 mpox patients from Nigeria Maximum entropy algorithm Area under curve 92% Majumder Mpox cases from 6 May to 28 July 2022 Polynomial neural network Predicted mpox cases would decrease after 20 October 2022 Eid Global mpox cases dataset, Kaggle BER-LSTM Mean absolute error 15.25 Yasmin Global mpox cases dataset, Kaggle Nine forecasting models Mean absolute error 146.29 Quereshi From website "Our World in Data" between 6 May to 28 July 2022 Multi-layer perceptron, ARIMA model Mean absolute error 32.59 diagnostics-13-00824-t003_Table 3 Table 3 Published artificial intelligence-based models for mpox-related web information. Paper Use Case Dataset Model Results Kolluri Handling mpox misinformation using a browser extension 170 actual facts and 55 wrong facts collected from the internet BERT-based machine learning model. 96% accuracy Mohbet Sentiment analysis of Twitter users Twitter posts about mpox Hybrid CNN-LSTM 94% accuracy Ng Sentiment analysis of Twitter users 352,182 Twitter posts BERT, BERTopic - Al Ahdal Sentiment analysis of Twitter users in Germany 15,936 Twitter posts from Germany Latent Dirichlet allocation - Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Gessain A. Nakoune E. Yazdanpanah Y. Monkeypox N. Engl. J. Med. 2022 387 1783 1793 10.1056/NEJMra2208860 36286263 2. Mileto D. Riva A. Cutrera M. Moschese D. Mancon A. Meroni L. Giacomelli A. Bestetti G. Rizzardini G. Gismondo M.R. New challenges in human monkeypox outside Africa: A review and case report from Italy Travel Med. Infect. Dis. 2022 49 102386 10.1016/j.tmaid.2022.102386 35738529 3. Hemati S. Farhadkhani M. Sanami S. Mohammadi-Moghadam F. A review on insights and lessons from COVID-19 to the prevent of monkeypox pandemic Travel Med. Infect. Dis. 2022 50 102441 10.1016/j.tmaid.2022.102441 36084881 4. Hernaez B. Munoz-Gomez A. Sanchiz A. Orviz E. Valls-Carbo A. Sagastagoitia I. Ayerdi O. Martin R. Puerta T. Vera M. Monitoring monkeypox virus in saliva and air samples in Spain: A cross-sectional study Lancet Microbe 2022 4 e21 e28 10.1016/S2666-5247(22)00291-9 36436538 5. Beeson A.M. Haston J. McCormick D.W. Reynolds M. Chatham-Stephens K. McCollum A.M. Godfred-Cato S. Mpox in Children and Adolescents: Epidemiology, Clinical Features, Diagnosis, and Management Pediatrics 2023 151 e2022060179 10.1542/peds.2022-060179 36471498 6. WHO Monkeypox Outbreak 2022 2022 Available online: (accessed on 30 January 2023) 7. Bala D. Monkeypox Skin Images Dataset (MSID) 2022 Available online: (accessed on 30 December 2022) 8. Ahsan M.M. Uddin M.R. Farjana M. Sakib A.N. Momin K.A. Luna S.A. Image Data collection and implementation of deep learning-based model in detecting Monkeypox disease using modified VGG16 arXiv 2022 10.48550/arXiv.2206.01862 2206.01862 9. Ciotti M. Ciccozzi M. Terrinoni A. Jiang W.C. Wang C.B. Bernardini S. The COVID-19 pandemic Crit. Rev. Clin. Lab. Sci. 2020 57 365 388 10.1080/10408363.2020.1783198 32645276 10. Jahmunah V. Ng E.Y. San T.R. Acharya U.R. Automated detection of coronary artery disease, myocardial infarction and congestive heart failure using GaborCNN model with ECG signals Comput. Biol. Med. 2021 134 104457 10.1016/j.compbiomed.2021.104457 33991857 11. Raghavendra U. Bhat N.S. Gudigar A. Acharya U.R. Automated system for the detection of thoracolumbar fractures using a CNN architecture Future Gener. Comput. Syst. 2018 85 184 189 10.1016/j.future.2018.03.023 12. Yildirim O. Baloglu U.B. Acharya U.R. A deep convolutional neural network model for automated identification of abnormal EEG signals Neural Comput. Appl. 2020 32 15857 15868 10.1007/s00521-018-3889-z 13. Acharya U.R. Sree S.V. Krishnan M.M. Molinari F. Garberoglio R. Suri J.S. Non-invasive automated 3D thyroid lesion classification in ultrasound: A class of ThyroScanTM systems Ultrasonics 2012 52 508 520 10.1016/j.ultras.2011.11.003 22154208 14. Yildirim O. Talo M. Ay B. Baloglu U.B. Aydin G. Acharya U.R. Automated detection of diabetic subject using pre-trained 2D-C.N.N models with frequency spectrum images extracted from heart rate signals Comput. Biol. Med. 2019 113 103387 10.1016/j.compbiomed.2019.103387 31421276 15. Siegrist E.A. Sassine J. Antivirals with activity against monkeypox: A clinically oriented review Clin. Infect. Dis. 2022 76 155 164 10.1093/cid/ciac622 35904001 16. Rizk J.G. Lippi G. Henry B.M. Forthal D.N. Rizk Y. Prevention and treatment of monkeypox Drugs 2022 82 957 963 10.1007/s40265-022-01742-y 35763248 17. Lum F.M. Torres-Ruesta A. Tay M.Z. Lin R.T. Lye D.C. Renia L. Ng L.F. Monkeypox: Disease epidemiology, host immunity and clinical interventions Nat. Rev. Immunol. 2022 22 597 613 10.1038/s41577-022-00775-4 36064780 18. Patel M. Surti M. Adnan M. Artificial intelligence (A.I.) in Monkeypox infection prevention J. Biomol. Struct. Dyn. 2022 41 1 5 10.1080/07391102.2022.2134214 19. Gul I. Liu C. Yuan X. Du Z. Zhai S. Lei Z. Chen Q. Raheem M.A. He Q. Hu Q. Current and Perspective Sensing Methods for Monkeypox Virus Bioengineering 2022 9 571 10.3390/bioengineering9100571 36290539 20. Sanchez P. Voisey J.P. Xia T. Watson H.I. O'Neil A.Q. Tsaftaris S.A. Causal machine learning for healthcare and precision medicine R. Soc. Open Sci. 2022 9 220638 10.1098/rsos.220638 35950198 21. Bojkova D. Bechtel M. Rothenburger T. Steinhorst K. Zoller N. Kippenberger S. Schneider J. Corman V.M. Uri H. Wass M.N. Drug Sensitivity of Currently Circulating Mpox Viruses N. Engl. J. Med. 2022 388 279 281 10.1056/NEJMc2212136 36577096 22. Sarkis-Onofre R. Catala-Lopez F. Aromataris E. Lockwood C. How to properly use the PRISMA Statement Syst. Rev. 2021 10 117 10.1186/s13643-021-01671-z 33875004 23. Alakus T.B. Baykara M. Comparison of Monkeypox and Wart, DNA Sequences with Deep Learning Model Appl. Sci. 2022 12 10216 10.3390/app122010216 24. Wang Z. Tober-Lau P. Farztdinov V. Lemke O. Schwecke T. Steinbrecher S. Muenzner J. Kriedemann H. Sander L.E. Hartl J. The human host response to monkeypox infection: A proteomic case series study EMBO Mol. Med. 2022 14 e16643 10.15252/emmm.202216643 36169042 25. Ali S.N. Ahmed M. Paul J. Jahan T. Sani S.M. Noor N. Hasan T. Monkeypox skin lesion detection using deep learning models: A feasibility study arXiv 2022 10.48550/arXiv.2207.03342 2207.03342 26. Abdelhamid A.A. El-Kenawy E.S. Khodadadi N. Mirjalili S. Khafaga D.S. Alharbi A.H. Ibrahim A. Eid M.M. Saber M. Classification of Monkeypox Images Based on Transfer Learning and the Al-Biruni Earth Radius Optimization Algorithm Mathematics 2022 10 3614 10.3390/math10193614 27. Kumar V. Analysis of CNN features with multiple machine learning classifiers in diagnosis of monkepox from digital skin images medRxiv 2022 10.1101/2022.09.11.22278797 28. Sitaula C. Shahi T.B. Monkeypox virus detection using pre-trained deep learning-based approaches J. Med. Syst. 2022 46 78 10.1007/s10916-022-01868-2 36201085 29. Sahin V.H. Oztel I. Yolcu Oztel G. Human Monkeypox Classification from Skin Lesion Images with Deep Pre-trained Network using Mobile Application J. Med. Syst. 2022 46 79 10.1007/s10916-022-01863-7 36210365 30. Islam T. Hussain M.A. Chowdhury F.U. Islam B.R. Can artificial intelligence detect Monkeypox from digital skin images? bioRxiv 2022 10.1101/2022.08.08.503193 31. Munoz-Saavedra L. Escobar-Linero E. Civit-Masot J. Luna-Perejon F. Civit A. Dominguez-Morales M. Monkeypox Diagnostic-Aid System with Skin Images Using Convolutional Neural Networks SSRN 2022 SSRN4186534 10.2139/ssrn.4186534 32. Loh H.W. Ooi C.P. Seoni S. Barua P.D. Molinari F. Acharya U.R. Application of Explainable Artificial Intelligence for Healthcare: A Systematic Review of the Last Decade (2011-2022) Comput. Methods Programs Biomed. 2022 226 107161 10.1016/j.cmpb.2022.107161 36228495 33. Sizikova E. Vendrow J. Cao X. Grotheer R. Haddock J. Kassab L. Kryshchenko A. Merkh T. Madushani R.W. Moise K. Automatic Infectious Disease Classification Analysis with Concept Discovery arXiv 2022 10.48550/arXiv.2209.02415 2209.02415 34. Akin K.D. Gurkan C. Budak A. Karatas H. Classification of Monkeypox Skin Lesion using the Explainable Artificial Intelligence Assisted Convolutional Neural Networks Avrupa Bilim Teknol. Derg. 2022 40 106 110 10.31590/ejosat.1171816 35. Ahsan M.M. Abdullah T.A. Ali M.S. Jahora F. Islam M.K. Alhashim A.G. Gupta K.D. Transfer learning and Local interpretable model agnostic based visual approach in Monkeypox Disease Detection and Classification: A Deep Learning insights arXiv 2022 10.48550/arXiv.2211.05633 2211.05633 36. Alcala-Rmz V. Villagrana-Banuelos K.E. Celaya-Padilla J.M. Galvan-Tejada J.I. Gamboa-Rosales H. Galvan-Tejada C.E. Convolutional Neural Network for Monkeypox Detection Proceedings of the International Conference on Ubiquitous Computing and Ambient Intelligence 2022 Cordoba, Spain 30 November-2 December 2022 Springer Cham, Switzerland 2022 89 100 10.1007/978-3-031-21333-5_9 37. Khafaga D.S. Ibrahim A. El-Kenawy E.S. Abdelhamid A.A. Karim F.K. Mirjalili S. Khodadadi N. Lim W.H. Eid M.M. Ghoneim M.E. An Al-Biruni Earth Radius Optimization-Based Deep Convolutional Neural Network for Classifying Monkeypox Disease Diagnostics 2022 12 2892 10.3390/diagnostics12112892 36428952 38. Haque M. Ahmed M. Nila R.S. Islam S. Classification of Human Monkeypox Disease Using Deep Learning Models and Attention Mechanisms arXiv 2022 10.48550/arXiv.2211.15459 2211.15459 39. Saleh A.I. Rabie A.H. Human monkeypox diagnose (H.M.D) strategy based on data mining and artificial intelligence techniques Comput. Biol. Med. 2022 152 106383 10.1016/j.compbiomed.2022.106383 36481764 40. Islam A. Shin S.Y. A Blockchain-Based Privacy Sensitive Data Acquisition Scheme during Pandemic Through the Facilitation of Federated Learning Proceedings of the 13th International Conference on Information and Communication Technology Convergence (ICTC) Jeju Island, Republic of Korea 19-21 October 2022 IEEE Piscataway, NJ, USA 2022 83 87 10.1109/I.C.T.C.55196.2022.9952440 41. Tom J.J. Anebo N.P. A Neuro-Fussy Based Model for Diagnosis of Monkeypox Diseases Available online: (accessed on 30 January 2023) 42. Arotolu T.E. Afe A.E. Wang H. Lv J. Shi K. Huang L. Wang X. Spatial modeling and ecological suitability of monkeypox disease in Southern Nigeria PLoS ONE 2022 17 e0274325 10.1371/journal.pone.0274325 36126054 43. Majumder P. Analyses of polynomial neural networks for prediction of the prevalence of monkeypox infections in Asia and around the world Electron. J. Gen. Med. 2022 19 em410 10.29333/ejgm/12400 44. Eid M.M. El-Kenawy E.S. Khodadadi N. Mirjalili S. Khodadadi E. Abotaleb M. Alharbi A.H. Abdelhamid A.A. Ibrahim A. Amer G.M. Meta-Heuristic Optimization of LSTM-Based Deep Network for Boosting the Prediction of Monkeypox Cases Mathematics 2022 10 3845 10.3390/math10203845 45. Yasmin F. Hassan M.M. Zaman S. Aung S.T. Karim A. Azam S. A Forecasting Prognosis of the Monkeypox Outbreak Based on a Comprehensive Statistical and Regression Analysis Computation 2022 10 177 10.3390/computation10100177 46. Qureshi M. Khan S. Bantan R.A. Daniyal M. Elgarhy M. Marzo R.R. Lin Y. Modeling and Forecasting Monkeypox Cases Using Stochastic Models J. Clin. Med. 2022 11 6555 10.3390/jcm11216555 36362783 47. Aiman S. Alhamhoom Y. Ali F. Rahman N. Rastrelli L. Khan A. Farooq Q.U. Ahmed A. Li C. Multi-epitope chimeric vaccine design against emerging Monkeypox virus via reverse vaccinology techniques-a bioinformatics and immunoinformatics approach Front. Immunol. 2022 13 985450 10.3389/fimmu.2022.985450 36091024 48. Bhattacharya M. Chatterjee S. Nag S. Dhama K. Chakraborty C. Designing, characterization, and immune stimulation of a novel multi-epitopic peptide-based potential vaccine candidate against monkeypox virus through screening its whole genome encoded proteins: An immunoinformatics approach Travel Med. Infect. Dis. 2022 50 102481 10.1016/j.tmaid.2022.102481 36265732 49. Pickett B.E. Sadat E.L. Zhang Y. Noronha J.M. Squires R.B. Hunt V. Liu M. Kumar S. Zaremba S. Gu Z. ViPR.: An open bioinformatics database and analysis resource for virology research Nucleic Acids Res. 2012 40 D593 D598 10.1093/nar/gkr859 22006842 50. Ullah A. Shahid F.A. Haq M.U. Qamar M.T. Irfan M. Shaker B. Ahmad S. Alrumaihi F. Allemailem K.S. Almatroudi A. An integrative reverse vaccinology, immunoinformatic, docking and simulation approaches towards designing of multi-epitopes based vaccine against monkeypox virus J. Biomol. Struct. Dyn. 2022 1 4 10.1080/07391102.2022.2125441 36129135 51. Altayb H.N. Fludarabine, a potential D.N.A-dependent R.N.A polymerase inhibitor, as a prospective drug against monkeypox virus: A computational approach Pharmaceuticals 2022 15 1129 10.3390/ph15091129 36145351 52. Lam T.P. Tran V.H. Mai T.T. Lai N.V. Dang B.T. Le M.T. Tran T.D. Trinh D.T. Thai K.M. Identification of Diosmin and Flavin Adenine Dinucleotide as Repurposing Treatments for Monkeypox Virus: A Computational Study Int. J. Mol. Sci. 2022 23 11570 10.3390/ijms231911570 36232872 53. Kolluri A. Vinton K. Murthy D. PoxVerifi: An Information Verification System to Combat Monkeypox Misinformation arXiv 2022 10.48550/arXiv.2209.09300 2209.09300 54. Mohbey K.K. Meena G. Kumar S. Lokesh K. A CNN-LSTM-based hybrid deep learning approach to detect sentiment polarities on Monkeypox tweets arXiv 2022 10.48550/arXiv.2208.12019 2208.12019 55. Ng Q.X. Yau C.E. Lim Y.L. Wong L.K. Liew T.M. Public sentiment on the global outbreak of monkeypox: An unsupervised machine learning analysis of 352,182 twitter posts Public Health 2022 213 1 4 10.1016/j.puhe.2022.09.008 36308872 56. AL-Ahdal T. Coker D. Awad H. Reda A. Zuratynski P. Khailaie S. Improving Public Health Policy by Comparing the Public Response during the Start of COVID-19 and Monkeypox on Twitter in Germany: A Mixed Methods Study Vaccines 2022 10 1985 10.3390/vaccines10121985 36560395 57. Jahmunah V. Ng E.Y. Tan R.S. Oh S.L. Acharya U.R. Explainable detection of myocardial infarction using deep learning models with Grad-CAM technique on ECG signals Comput. Biol. Med. 2022 146 105550 10.1016/j.compbiomed.2022.105550 35533457 58. Jahmunah V. Ng E.Y. Tan R.S. Oh S.L. Acharya U.R. Uncertainty quantification in DenseNet model using myocardial infarction ECG signals Comput. Methods Programs Biomed. 2023 229 107308 10.1016/j.cmpb.2022.107308 36535127 59. Kapusuzoglu B. Mahadevan S. Information fusion and machine learning for sensitivity analysis using physics knowledge and experimental data Reliab. Eng. Syst. Saf. 2021 214 107712 10.1016/j.ress.2021.107712 |
PMC10000612 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050817 diagnostics-13-00817 Article Deep Learning-Based Auto-Segmentation of Spinal Cord Internal Structure of Diffusion Tensor Imaging in Cervical Spondylotic Myelopathy Fei Ningbo Methodology Software Formal analysis Investigation Data curation Writing - original draft Visualization 12 Li Guangsheng Conceptualization Methodology Formal analysis Resources 12 Wang Xuxiang Validation 1 Li Junpeng Validation 1 Hu Xiaosong Validation 13 Hu Yong Conceptualization Investigation Resources Writing - original draft Writing - review & editing Supervision Funding acquisition 123* Kurkin Semen A. Academic Editor 1 Spinal Division, Orthopedic and Traumatology Center, The Affiliated Hospital of Guangdong Medical University, Zhanjiang 524013, China 2 Department of Orthopaedics and Traumatology, The University of Hong Kong, Hong Kong SAR, China 3 Department of Orthopaedics and Traumatology, The University of Hong Kong-Shenzhen Hospital, Shenzhen 518009, China * Correspondence: [email protected]; Tel.: +852-29740359 21 2 2023 3 2023 13 5 81708 12 2022 14 2 2023 19 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Cervical spondylotic myelopathy (CSM) is a chronic disorder of the spinal cord. ROI-based features on diffusion tensor imaging (DTI) provide additional information about spinal cord status, which would benefit the diagnosis and prognosis of CSM. However, the manual extraction of the DTI-related features on multiple ROIs is time-consuming and laborious. In total, 1159 slices at cervical levels from 89 CSM patients were analyzed, and corresponding fractional anisotropy (FA) maps were calculated. Eight ROIs were drawn, covering both sides of lateral, dorsal, ventral, and gray matter. The UNet model was trained with the proposed heatmap distance loss for auto-segmentation. Mean Dice coefficients on the test dataset for dorsal, lateral, and ventral column and gray matter were 0.69, 0.67, 0.57, 0.54 on the left side and 0.68, 0.67, 0.59, 0.55 on the right side. The ROI-based mean FA value based on segmentation model strongly correlated with the value based on manual drawing. The percentages of the mean absolute error between the two values of multiple ROIs were 0.07, 0.07, 0.11, and 0.08 on the left side and 0.07, 0.1, 0.1, 0.11, and 0.07 on the right side. The proposed segmentation model has the potential to offer a more detailed spinal cord segmentation and would be beneficial for quantifying a more detailed status of the cervical spinal cord. diffusion tensor imaging (DTI) image segmentation deep learning fractional anisotropy (FA) cervical spondylotic myelopathy (CSM) National Natural Science Foundation of China82072507 This research was funded by the National Natural Science Foundation of China (82072507). pmc1. Introduction Cervical spondylotic myelopathy (CSM) is characterized by chronic spinal degeneration causing structural modifications to the intervertebral discs, ligaments, etc. . Magnetic resonance imaging (MRI) is the gold standard for diagnosing cervical spinal cord dysfunction . Conventional MRIs, including T1-weighted and T2-weighted MRIs, are commonly used to obtain morphological information about the spinal cord, such as intramedullary or extramedullary abnormalities, spinal cord compression, disk herniation, etc., for identification of spinal cord injury . Based on signal abnormalities on T2-weighted MRI, the Brain and Spinal Injury Center (BASIC) score is used to classify acute traumatic spinal cord injuries . However, conventional MRI findings and clinical presentation in CSM appear disconnected, and conventional MRI cannot highlight the microstructural spinal cord abnormalities associated with CSM . Diffusion tensor imaging (DTI) is a type of multi-parametric MRI and is considered a promising imaging technique for studying the ultrastructure of the spinal cord. The DTI parameter, fractional anisotropy (FA), has been used in several studies to investigate the diagnosis and prognosis of cervical spondylotic myelopathy . To detect detailed neurological deficits in the spinal cord, regions of interest (ROI) can be segmented on the cord to measure the corresponding features in different tracts from the FA map . A greater number of ROIs within a spinal cord leads to precise information about myelopathy. Numerous studies have demonstrated the feasibility and effectiveness of ROI-based features for providing details on the internal state of the spinal cord . Due to its manual voxel selection and user-dependent nature, ROI delineation involves the user drawing ROIs based on perceptions of the location of underlying anatomical structures, such as the gray matter or the corticospinal tract . Hand animation ROIs are prone to inter-rater and intra-rater variability and bias and are not easily transferable to large-scale studies . Consequently, the manual drawing of describing ROI has serious limitations. Some methods have been proposed for segmenting the entire spinal cord. EI Mendili et al. used T2-weighted MR images as input and employed a dual-threshold-based approach for spinal cord segmentation . De Leener et al. proposed a PropSeg method based on multi-resolution propagation of tubular deformation models . Chen et al. proposed a method to segment the spinal cord using deformable atlases and topological constraints that are robust to noise and artifacts . In all of the above techniques, manual intervention is required, or a large database needs to be established to create the spinal cord segmentation map, and T1-weighted or T2-weighted images are the primary input. Several research studies proposed further segmenting the internal structures, gray matter (GM) and white matter (WM). Antal et al. use variational formulation to automatically detect cerebrospinal fluid, WM, and GM and combine them with shape prior to segment GM and WM . Ferran proposed a two-stage pipeline using the Optimized Patch Match Label fusion (OPAL) method for segmentation of the whole spinal cord and the Similarity and Truth Estimation for Propagated Segmentations (STEPS) for the further extraction of GM and WM . However, these methods can be applied to specific spinal cord levels, limiting their application to other segments. In DTI image analysis, there are some applications of segmentation methods. Marek used the semiautomatic algorithm provided by ITK-SNAP to segment the cervical spine GM and WM . This method takes a T2-weighted image as input and registers the segmentation results with the DTI image. Because of the deviations between T2-weighted images and DTI images, a manual check is still necessary to guarantee the registration result. Richu applied six commonly used automatic thresholding algorithms for segmentation and vote to obtain the final segmentation results of GM and WM . Although this method uses DTI images as input, the segmentation of GM and WM given by this method cannot reveal the details of each white matter column. Therefore, it should be pointed out that the existing methods of spinal cord segmentation based on DTI images provide limited ROIs within the spinal cord. In recent years, great attention has been paid to deep learning in the medical field. One of the most widely used deep learning networks in medical image segmentation is UNet, which is based on an encoder-decoder architecture . Several deep learning-based models applied UNet or UNet-like structures on spinal cord segmentation. For instance, Xiaoran et al. proposed a UNet-based, fully automatic method on 2D axial-view MRI slices for the whole spinal cord segmentation of patients with CSM . Alhanouf et al. used a pre-trained MobileNet-V3 CNN model as the backbone for feature extraction, which was augmented by a set of up-sampling layers and employing skip connection similar to the UNet architecture used for spinal cord GM segmentation . The deep-learning model achieved better performance on the spinal cord gray matter segmentation challenge dataset compared to Spinal Cord Toolbox (SCT), the Variational Bayesian Expectation Maximization (VBEM) method, and the Gray Matter Segmentation Based on Maximum Entropy (GSBME) method. This paper proposes a deep learning model based on the UNet architecture for segmenting multiple ROIs within the spinal cord in DTI images. Segmentation results can be obtained simultaneously for both sides' ventral, lateral, dorsal columns, and gray matter. Specifically, heatmap-distance loss (HDL) is proposed to train with the UNet model to make the model have better performance on the small area of column-based ROI and gray matter. This study hopes to provide a more detailed segmentation than the gray and white matter segmentation of the spinal cord and more details of the internal status of the spinal cord. 2. Materials and Methods 2.1. Study Population An overall sample size of 89 patients with CSM was recruited for this study. All of the CSM patients are symptomatic. Clinical examination by the Japanese Orthopedic Association (JOA) produced scores of 9.78 +- 3.48, while the healthy JOA score is 17 with motor function of upper and lower limb (8 scores), sensory (6 scores) and sphincter function (3 scores). Inclusion criteria were a clinical diagnosis of CSM without a history of spinal surgery. Those with neurological disorders or prior neurological trauma were excluded. Every participant in this study completed a written informed consent form approved by the institutional review board. DTI scanning protocol was performed with a Philips 3T Achieva scanner (Philips Medical System, Best, Netherlands). Specifically, the one-shot echoplanar imaging sequence was used. We carried out diffusion encoding in 15 nonlinear and noncoplanar directions with a b value of 600 s/mm2. The scanning parameters are listed as follows: field of view = 80 x 80 mm2, thickness of slices = 7 mm, gap between slices = 2.2 mm, fold-over direction = anteroposterior, reconstruction resolution = 0.63 x 0.63 x 7 mm3, and voxel resolution = 1.0 x 1.26 x 7 mm3, and TTE/TR = 60 ms/5 heartbeats. For the purpose of sup-pressing the fold-over effect, spatial saturation was inverted and spectral presaturation was applied. To minimize the effects of cerebrospinal fluid on cardiac vector cardiograms, cardiac vector cardiograms were triggered. 2.2. Manual Segmentation The manual delineation of the ROI is an essential step for feature extraction to provide more information about the inside of the spinal cord. FA is one of the most popular parameters in DTI studies to evaluate the microstructural abnormality of the spinal cord . Each CSM patient extracted 12 axial slices from three stacks covering the vertebrae between C2 and C7/T1. Spinal Cord Toolbox, Version 2.3 was used to preprocess the DTI images, extract the B0 images, and calculate the FA map. B0 and FA images were used as input for spinal cord ROI auto-segmentation in the following experiments. The auto-segmentation ROIs and manual segmentation ROIs were superimposed on the FA images to extract the mean FA value for each ROI. An experienced researcher in the cervical spinal cord manually drew ROIs on B0 images supplemented with FA images. The ROIs have covered more specific columns within the entire cord region, as showed in Figure 1. A total of eight regions were identified for each image to obtain detailed information about the spinal cord: lateral column (LC), dorsal column (DC), ventral column (VC), and gray matter (GM) on both sides of the spinal cord. We acquired 96 segmentation masks for each subject (8 ROIs x 12 slices). A total of 8472 segmentation masks have been created for all patients with CSM (8 ROIs x 1059 slices) except for images that are difficult to segment manually. The manual delineated ROIs were defined as the ground truth. A second expert manually segmented these ROIs to enhance confidence in identifying these eight ROIs. Whenever there was a discrepancy, a third expert was consulted until a consensus was reached. 2.3. Segmentation Models The UNet was introduced for its popularity in medical image segmentation. The model structure is similar to a fully convolutional network, which consists of two parts. The first part consists of layers of convolution and max pooling (known as the encoder). B0 images will be used as input to extract multi-scale features. The second part is mirror-symmetric to the first part (also known as the decoder), consisting of transposed convolution and up-sampling layers for feature expansion of a feature vector of an image of sizes corresponding to the original medical image. Several skip connections between the encoder and decoder were added to further utilize the multi-scale feature of the encoder for feature expansion. Figure 2 demonstrates the detailed parameters of the UNet model. Generally, the feature extracted from the encoder highly affects the segmentation performance. Two different depth general network structures, VGG-16 and ResNet50 , were used as the backbone for feature extraction. Two segmentation tasks corresponded to two models with different backbones for feature extraction, including vgg_9classes for nine classes (plus background) segmentation using VGG16 as the backbone and resnet_9classes for nine classes (plus background) segmentation using ResNet50 as the backbone. A challenge that needs to be noticed in this study is the small object of manual segmentation. In our dataset, all ROIs are small areas, and the largest ROI only takes about 1/1000 of the whole DTI images from Figure 3. The challenge for model training comes from small-shape annotations that contribute less to the loss than background annotations, which makes the model prone to classify a pixel as background because image segmentation is a pixel-wise classification. Wentao et al. use the Dice loss combined with the focal loss to segment small-volume organs in the Organs-At-Risk delineation problem and achieve good segmentation accuracy, while the Dice loss learns the imbalanced class distribution, and the focal loss trains the model to learn hard segmented annotations for these small-volume organs. Hence, this hybrid loss was selected for the training of two models. The total loss can be formulated as follows:(1) TP(i)=n=1Npn(i)gn(i) (2) FN(i)=n=1N(1-pn(i))gn(i) (3) FP(i)=n=1Npn(i)(1-gn(i)) (4) Ltotal=LDICE+LFocal=C-i=1C-1TP(i)TP(i)+FN(i)+FP(i)-1Nc=1C-1n=1Ngn(i)(1-pn(i))2log(pn(i)) where TP(i), FN(i), and FP(i) are the true positives rate, false negatives rate, and false positives rate for the classes. pn(i) is the predicted output value for the pixel n class i, gn(i) is the ground truth label for pixel n class i, and C is the total number of anatomies included in the background. N is the total number of pixels in one slice of a B0 slice image. However, the performance of the segmentation model using B0 images as input on the test dataset was not satisfactory. Although we tried to use the FA images as input, the segmentation performance was also unsatisfactory. Except for the small object, there is another challenge in this dataset: the difference in pixel values between different classes is minimal, making it difficult to distinguish the boundaries of different ROIs. Based on the above challenges and performance on the test dataset, we believe the Dice loss combined with focal loss is unsuitable for this research. Inspired by the landmark localization based on two-dimensional (2D) heatmaps and the boundary-weighted map from Qikui's research , we proposed the heatmap-distance loss (HDL). The loss can be formulated as follows:(5) HDL=i=19||Hi(X)-Hi^(X)||2 (6) Hi^(x)={1, xXlandmarkexp(-||max(x-Xlandmark)||222s2),xXlandmark where X denotes the set of pixels in 2D space from any class ground truth mask, and x represents any pixels from X. Let Xlandmark denote the set of landmark points whose pixel value is equal to one. The notation ||||2 denotes the L2 or Euclidean distance. Hi^(x) refers to the ground truth heatmap for the ith class of ground truth masks, while Hi(X) is the predicted heatmap for the ith class. The pixel value of the heatmap Hi^(x) is created based on formulation (6). When the point from X has the same coordinate with the point in Xlandmark, the value of Hi^(x) is equal to one. Otherwise, each point will calculate the L2 or Euclidean distance between the point and each landmark point. The max distance between each pixel and the set of landmark points will be used as input to a Gaussian function with variance s2 to generate the value of Hi^(x). The rationale behind the design is that the small area of segmentation can be expanded to a large area, and the pixels created by Gaussian based on distance could make the small area contribute to the training process. Based on the test dataset performance and the model structure's complexity, a better encoder between VGG16 and ResNet50 would be used as a backbone trained with the proposed loss. 2.4. Training and Evaluation We use 81% (857 images) of our dataset for training, 9% (96 images) of our dataset for validation, and 10% (106 images) as a test dataset. The UNet model is trained on the training dataset and validated on the validation dataset. The final performance is exclusively reported on the test set. The experiments were performed in the Pytorch environment, and we trained the segmentation model on NVIDIA RTX3080Ti. The batch size was set to 16 based on the memory of the GPU and computer. We used the Adam optimizer with a momentum of 0.9 and a cosine learning rate strategy with a learning rate range from 1 x 10-4 to 1 x 10-6; the number of epochs is 500. The Dice loss combined with focal loss and the HDL loss was trained under the same hyperparameters. The threshold definition is an essential step that transforms the model output into segmentation results. The threshold values for different segmentation classes are selected from 0 to 1 with intervals of 0.001 on validation datasets, and the threshold value of the best performance was to evaluate the final performance on the test dataset. For evaluation of the proposed method, three metrics for each class are used, including the Recall (also known as True Positive Rate), which represents a method's ability to segment pixels as a proportion of all correctly labeled pixels, the Precision, which measures the ratio of correctly segmented pixels and the Dice coefficient, which is a similarity index between two masks. The following equations define these three metrics: GT refers to ground truth, and PM refers to predict mask. Symbols FN, FP, and TP are explained as follows: true positive (TP) if it was a pixel in GT mask and it was segmented in PM; false positive (FP) if it was not a pixel in GT mask and it was segmented in PM; and finally, false negative (FN) if it was a pixel in GT mask and it was not segmented in PM:(7) DICE=2|GT PM||GT|+|PM| (8) Recall=TPTP+FN (9) Precision=TPTP+FP To further evaluate the accuracy of the mean FA value from the segmentation model, a metric FAerrori, the percentage of absolute error of FA was utilized to compare the mean value of the ground truth mask superimposing on the DTI metric map (FA map) and the prediction segmentation mask superimposing on the DTI metric map (FA map), which is formulated as follows:(10) FAerrori=|FAGTi-FAPMi|FAGTi where FAGTi refers to the FA mean value from the ground truth mask of the ith segmentation class and FAPMi refers to the FA mean value from the predicted mask of the ith segmentation class. The performance of the algorithm was evaluated using basic statistics, such as the mean and standard deviation of FAerrori. Outliers were defined as slices of DTI images that did not obtain the segmentation results from the model. After removing these outliers, the performance was calculated. The author performed an outlier assessment by calculating the percentage of outliers within this dataset. 3. Results 3.1. Encoder Comparison Figure 4 demonstrates the performance of the UNet with different backbones on the test dataset (B0 as input images). We used the VGG16 and ResNet50 as the encoder for UNet training with a hybrid loss of Dice loss and focal loss. The UNet with VGG16 as the backbone achieves the highest Dice value of 0.65 on the right dorsal column and reaches the lowest Dice value of 0.48 on the left ventral column. Figure 4 demonstrates that the UNet model with VGG16 as the backbone performs better than using ResNet50 as the backbone on all segmentation classes. Therefore, VGG16 will be used as a backbone for UNet trained with the proposed loss function. 3.2. Model Performance Figure 5 shows the segmentation result of several samples using the proposed model. In these pictures, the line in blue color is by GT and the line in red color is by PM. Figure 6 demonstrates the performance on test datasets (FA as input images) of UNet training with hybrid loss and HDL loss. The UNet with VGG16 as the backbone achieves the highest Dice value of 0.69 on the left dorsal column and the lowest Dice value of 0.54 on the left gray matter. Figure 6 demonstrates that the UNet model trained with the proposed HDL loss performs better than the hybrid loss on all segmentation classes. Table 1 presents the overall performance of UNet using the proposed HDL loss. To eliminate the effect of different kinds of images, Figure 7 demonstrates the performance of the UNet model training under different loss functions and using different types of DTI images. UNet model training with the proposed HDL loss using FA as input images achieves the highest performance. There is a situation in which several images could not obtain the segmentation result from the model defined as the outlier slices. Table 2 illustrates the statistics of performance removing these outlier slices. The Dice coefficients of the left lateral column, ventral column, and gray matter were improved by 1%, 4%, and 3%, respectively. The Dice coefficients of the right lateral column, ventral column, and gray matter were improved by 2%, 4%, and 6%, respectively. The "Outliers" column demonstrated the percentage of DTI slices that cannot obtain output from segmentation. 3.3. Accuracy of FA Prediction Furthermore, segmentation aims to extract the mean FA value within ROIs from the spinal cord. Therefore, we calculate the mean FA values from auto-segmented and manual-segmented ROIs. A series of higher intra-class correlation coefficients were found in Figure 8, indicating excellent agreement between the FA from segmented ROIs and the FA from ground truth ROIs. Table 3 demonstrates the percentage of the absolute error between the mean FA value from the ground truth mask and the mean FA value from the prediction mask. The percentage of the mean absolute error for lateral, dorsal, ventral, and gray matter was 0.07, 0.07, 0.11 and 0.08 on the left side as well as 0.07, 0.1, 0.1 and 0.07 on the right side, respectively. 4. Discussion Microstructural spinal cord abnormalities are vital for CSM patients. DTI can reveal the microstructural change in the spinal cord compared to conventional MRI. However, the DTI analysis based on the whole cord region cannot reveal the details of the spinal cord. The subtle abnormality of DTI parameters caused by minor microstructure impairment in a small ROI may be missed due to the whole cord analysis dilution effect. ROI-based DTI analysis could avoid the drawback and enable the quantitative evaluation of specific regions with the spinal cord. This study combines UNet with the proposed heatmap distance loss to segment multiple ROIs within the spinal cord on DTI-related images from CSM patients. Many previous studies in spinal cord segmentation mainly focus on T1-weighted and T2-weighted images instead of DTI-related images. The related studies in spinal cord segmentation approaches on DTI images are semi-automatic or diffusion-tensor-tracking-based instead of fully automatic. The segmentation region of associated studies is the whole cord or gray matter/white matter instead of multiple ROIs within the spinal cord. To the authors' knowledge, this is the first study using the deep learning method to fully automate and segment the multiple ROIs within the spinal cord on DTI images from CSM patients. We focused on developing an approach to help clinicians provide a more detailed description of spinal status of CSM patients by providing ROI-based features to facilitate ROI-specific analysis of DTI images. We used two backbone networks to extract the feature within DTI-related images. VGG16 and ResNet50 are well-known convolutional neural networks in many natural-images-related tasks, and these networks achieve good performance on feature extraction. We hypothesize that the higher-level feature would be beneficial for segmentation. Figure 4 illustrates that the hypothesis is wrong. Figure 3 demonstrates that the percentages of pixels for all classes are small than 0.0011. Hence, the advanced features would not be suitable for this task because the convolutional process from a deeper layer would eliminate the efficient information for the training of the segmentation model. The obstacle to fulfilling the segmentation in this dataset is the ROIs within the spinal cord. Figure 5 demonstrates several samples with the auto-segmentation result. Because the spinal medulla is misshaped in CSM patients, the ROIs contain variable sizes. To deal with the above issue, several loss functions have been proposed. The Dice loss and focal loss are commonly used for segmentation. Wentao used focal loss combined with Dice loss for organ-at-risk (OAR) segmentation, especially for small-volume organ segmentation, achieving better segmentation accuracy. In this research, we use the focal loss combined with Dice loss to train the UNet to fulfill the segmentation of ROIs within the spinal cord. Figure 4 demonstrates that the segmentation performance of sub-ROIs with the spinal cord is not satisfied. The possible reason is that the region of sub-ROIs from the spinal cord is smaller than the volume of OARs in Wentao's research, and the boundary of these ROIs in DTI-related images is not as clear as in computed tomography (CT) images. We proposed a new heatmap distance loss function inspired by landmark localization. Figure 6 demonstrates that the UNet with the proposed HDL achieves the highest segmentation performance for all classes. The basic principle of HDL is that we manually expand the ground truth area by designing the pseudo-probability, which makes the small area significantly contribute to the model training. However, the higher performance in Figure 6 was based on the FA images, while the lower performance in Figure 6 was based on the B0 images. Regardless of the image types, Figure 7 illustrates that the UNet model training using HDL performs better than hybrid loss. The B0 images have a wide range, as well as the CT image in Wentao's research, while the value range of FA images is from 0 to 1. A wide range of values would be beneficial for the convolutional process to extract the high-level feature and the hybrid loss, which could be the reason that VGG_B0_hybird_loss has better performance than VGG_FA_hybird_loss. Figure 7 also illustrates that the pixel value of images has less effect on the model training using proposed HDL loss. Figure 8 demonstrates the correlation between the mean FA of auto-segmented ROIs and the mean FA of ground truth. The high intra-class correlation on the diagonal position of the correlation matrix demonstrated the high agreement between the mean FA value from auto-segmented ROIs and ground truth ROIs. Several values on the correlation matrix need to be explored. It should be noted that the correlations from several pairs are high, including {L_DC, R_DC}, {R_DC, L_DC}, {L_VC, R_VC}, {R_VC, L_VC}, {L_GM, R_GM}, and {R_GM, L_GM}. This result illustrated that the ROI of Pred_L_DC or Pred_R_VC intersects with GT_L_DC or GT_R_VC, and there is a high agreement between the left and right of DC, VC, and GM, which could be the factor that affects the segmentation performance. Table 3 demonstrates the absolute error percentage between the FA of auto-segmented ROIs and ground truth ROIs on the test dataset. The mean distance between the mean FA value obtained from the segmentation model and the mean FA value obtained from manually segmented ROIs is small, demonstrating that the mean FA value extracted from the segmentation model has the potential for detailed spinal cord status evaluation, such as the mean FA value extracted from the ground truth. The spinal medulla misshaped in CSM patients still significantly affects segmentation performance. However, most of the segmentation result is close to the manual segmentation, and most of the auto-segmentation results are on the related region, which could be why the distance between mean FA values from the auto-segmentation and mean FA values from manual segmentation are small. Clinically, DTI has three applications. Firstly, it has the ability to identify spinal cord structures and injuries, localization, and diagnosis of CSM. Cui et al. 's research demonstrated that the orientation entropy from DTI analysis is a valuable tool for identifying the pathological level in multilevel CSM patients. Shu-Qiang Wang combined the eigenvalues from DTI and machine learning methods to achieve satisfactory performance in recognizing the levels of CSM. According to Monika Skotarczak , the DTI metrics can be used as a biomarker to illustrate the microstructural disorder of the spinal cord not visible on conventional MRI. Secondly, there is the indication of surgery. As demonstrated by Karsten Scholler et al. , DTI metrics achieve higher sensitivity in identifying levels requiring decompression surgery than increased MRI signals. Thirdly, there is dynamic spinal function examination and prognosis assessment. For instance, Ellingson et al. 's study demonstrated that DTI has the potential to monitor symptomatic patients and asymptomatic patients, and Chun Yi Wen et al. 's research revealed that FA is a biomarker for surgical outcomes. The auto-segmentation of ROIs with the spinal cord DTI has the potential to reduce inter-rater and intra-rater variability in manually drawn ROIs and help the entire DTI analysis workflow be automated. There are several limitations. Firstly, there is space for improvement in segmentation performance. Based on the findings from Figure 5, the deep learning model with fewer layers will be explored, which will be regarded as part of the future extension of this research. Secondly, the intersection between the ROI of Pred_L_DC and the ROI of GT_R_DC, the ROI of Pred_R_DC and the ROI of GT_L_DC, the ROI of Pred_L_VC and the ROI of GT_R_VC or the ROI of Pred_R_VC and the ROI of GT_L_VC is a problem that could be regarded as one reason to improve the segmentation performance of DC(L & R) and VC(L&R). A specified preprocessing or postprocessing method needs to be explored in the future. Thirdly, the diffusion tensor for calculation of the DTI metric is reconstructed based on a series of diffusion-weighted images and the b-matrix that integrates the parameters of diffusion-sensitizing gradients. However, many factors influence the accuracy or reliability of the diffusion tensor evaluation, including image noise , eddy currents , diffusion gradient nonlinearity , and others. For example, in the circumstance of low anisotropy, imaging noise could lead to the wrong estimation of eigenvalues, therefore causing an overestimation of DTI metrics (such as FA) due to the incorrect order of eigenvalues , and diffusion gradient inhomogeneities would cause distortion of the diffusion tensor's eigenvalues as well as rotation of the eigenvectors . The diffusion gradient inhomogeneity is an important source of systematic error, and several methods have been proposed to correct the error due to diffusion gradient inhomogeneity . They are either based on a calibration that uses an anisotropic phantom as a reference for estimation of the diffusion tensor, mapping the actual magnetic field, or using the coil system's manufacturer-provided specifications. The b-matrix spatial distribution in DTI (BSI-DTI) is a frequently employed technique for correcting this kind of inaccuracy and has the capacity to minimize the impact on the assessment of DTI metrics. To reduce the effects of inhomogeneous magnetic field gradients and make the data as accurate as possible, we intend to adopt the BSI-DTI approach in the future. Finally, different types of input images potentially affect the performance of segmentation, and many factors of MRI sequences can affect the DTI images. The MRI scanner involves the measurement of DTI metrics (such as FA) . Except for the MRI scanner, several studies demonstrated that the parameters of MRI sequence, such as the b-value , echo time , the number of DTI directions , and the number of signal acquisitions , could influence the diffusion quantification. In order to expand the scope of application of the model, it is necessary to extend the diversity of the DTI image dataset: for instance, acquiring the DTI images of CSM patients with different b-value and different numbers of DTI directions. In future work, we will consider the continuation path for this research in the following steps. Firstly, we will continue to expand the size of our dataset, as the sample size in this research is insufficient for deep learning training. Secondly, weak-supervised training will be used on the new dataset. A new and helpful segmentation model will be developed using the manual segmentation result in this research. Thirdly, DTI images of CSM patients will be classified into several categories based on the types of CSM compression or the classification result from unsupervised learning. The segmentation model will be trained on the specified cases to verify the algorithm's effectiveness in the specific category of CSM cases. 5. Conclusions This paper proposed using UNet and the heatmap distance loss to automatically segment the sub-ROIs within the spinal cord on DTI images from CSM patients. The performance of the segmentation model and the agreement between the mean FA value of auto-segmented ROIs and the mean FA value of manually segmented ROIs demonstrates that the deep learning model has the potential to provide more details of the internal status of the spinal cord. Author Contributions Conceptualization, Y.H. and G.L.; methodology, N.F. and G.L.; software, N.F.; validation, J.L., X.W. and X.H.; formal analysis, N.F. and G.L.; investigation, N.F. and Y.H.; resources, Y.H. and G.L.; data curation, N.F.; writing--original draft preparation, N.F. and Y.H.; writing--review and editing, Y.H.; visualization, N.F.; supervision, Y.H.; funding acquisition, Y.H. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster ([HKU/HA HKW IRB UW 18-394]). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement Because the limits of ethic approval, the original DTI dataset is not opened to public. The post-process data can be asked from the corresponding author by request, which cannot be used for commercial purpose. Conflicts of Interest The authors declare no conflict of interest. Figure 1 ROIs covering the whole spinal cord. Figure 2 Structure of the segmentation network. Figure 3 The percentage of pixels for each class on DTI images. Figure 4 The comparison between different encoders. Figure 5 Several segmentation results with the proposed model. Figure 6 The performance comparison of UNet with different losses. Figure 7 The performance comparison of UNet with different losses using different DTI images. Figure 8 The correlation between FA of predicted segmentation ROIs and FA of ground truth ROIs. Pred_: the ROIs from the proposed segmentation model; GT_: the ROIs from the manual segmentation. diagnostics-13-00817-t001_Table 1 Table 1 The performance of UNet with proposed HDL loss. Class Dice Recall Precision Left dorsal column 0.69 +- 0.19 0.78 +- 0.23 0.64 +- 0.2 Left lateral column 0.67 +- 0.24 0.74 +- 0.29 0.63 +- 0.23 Left ventral column 0.57 +- 0.29 0.62 +- 0.33 0.55 +- 0.28 Left gray matter 0.54 +- 0.28 0.59 +- 0.33 0.53 +- 0.28 Right dorsal column 0.68 +- 0.21 0.74 +- 0.25 0.65 +- 0.2 Right lateral column 0.67 +- 0.22 0.74 +- 0.26 0.65 +- 0.22 Right ventral column 0.59 +- 0.26 0.63 +- 0.29 0.58 +- 0.27 Right gray matter 0.55 +- 0.31 0.56 +- 0.34 0.57 +- 0.32 diagnostics-13-00817-t002_Table 2 Table 2 The performance of UNet with proposed loss removed outlier slices. Class Dice Recall Precision Outliers Left dorsal column 0.69 +- 0.19 0.78 +- 0.23 0.64 +- 0.2 0 Left lateral column 0.68 +- 0.23 0.76 +- 0.27 0.64 +- 0.21 2% Left ventral column 0.61 +- 0.25 0.67 +- 0.3 0.59 +- 0.25 7% Left gray matter 0.57 +- 0.25 0.63 +- 0.3 0.56 +- 0.25 7% Right dorsal column 0.68 +- 0.2 0.74 +- 0.24 0.66 +- 0.2 1% Right lateral column 0.69 +- 0.18 0.76 +- 0.23 0.67 +- 0.19 3% Right ventral column 0.63 +- 0.23 0.67 +- 0.26 0.61 +- 0.24 6% Right gray matter 0.61 +- 0.27 0.63 +- 0.3 0.63 +- 0.27 10% diagnostics-13-00817-t003_Table 3 Table 3 The percentage of absolute error between two types of FA value. Class FAerrori Left dorsal column 0.07 +- 0.07 Left lateral column 0.07 +- 0.11 Left ventral column 0.11 +- 0.17 Left gray matter 0.08 +- 0.14 Right dorsal column 0.07 +- 0.08 Right lateral column 0.1 +- 0.11 Right ventral column 0.1 +- 0.23 Right gray matter 0.07 +- 0.09 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Wang K. Chen Z. Zhang F. Song Q.-x. Hou C.-l. Tang Y.-x. Wang J. Chen S. Bian Y. Hao Q. Evaluation of DTI Parameter Ratios and Diffusion Tensor Tractography Grading in the Diagnosis and Prognosis Prediction of Cervical Spondylotic Myelopathy Spine 2017 42 E202 E210 10.1097/BRS.0000000000001784 28207659 2. Shabani S. Kaushal M. Budde M.D. Wang M.C. Kurpad S. Diffusion tensor imaging in cervical spondylotic myelopathy: A review J. Neurosurg. Spine 2020 1 8 10.3171/2019.12.SPINE191158 3. Quencer R.M. Cervical myelopathy: Imaging/clinical correlations and the application of evolving magnetic resonance imaging techniques in cervical spinal cord abnormalities Spine (Phila Pa 1976) 2001 26 2411 2413 10.1097/00007632-200111150-00003 11707701 4. Fehlings M.G. Martin A.R. Tetreault L.A. Aarabi B. Anderson P. Arnold P.M. Brodke D. Burns A.S. Chiba K. Dettori J.R. A Clinical Practice Guideline for the Management of Patients With Acute Spinal Cord Injury: Recommendations on the Role of Baseline Magnetic Resonance Imaging in Clinical Decision Making and Outcome Prediction Glob. Spine J. 2017 7 221S 230S 10.1177/2192568217703089 5. Fehlings M.G. Tetreault L.A. Riew K.D. Middleton J.W. Aarabi B. Arnold P.M. Brodke D.S. Burns A.S. Carette S. Chen R. A Clinical Practice Guideline for the Management of Patients With Degenerative Cervical Myelopathy: Recommendations for Patients With Mild, Moderate, and Severe Disease and Nonmyelopathic Patients With Evidence of Cord Compression Glob. Spine J. 2017 7 70S 83S 10.1177/2192568217701914 6. Kurpad S. Martin A.R. Tetreault L.A. Fischer D.J. Skelly A.C. Mikulis D. Flanders A. Aarabi B. Mroz T.E. Tsai E.C. Impact of Baseline Magnetic Resonance Imaging on Neurologic, Functional, and Safety Outcomes in Patients With Acute Traumatic Spinal Cord Injury Glob. Spine J. 2017 7 151S 174S 10.1177/2192568217703666 7. Ahuja C.S. Wilson J.R. Nori S. Kotter M.R.N. Druschel C. Curt A. Fehlings M.G. Traumatic spinal cord injury Nat. Rev. Dis. Prim. 2017 3 21 10.1038/nrdp.2017.18 8. Talbott J.F. Whetstone W.D. Readdy W.J. Ferguson A.R. Bresnahan J.C. Saigal R. Hawryluk G.W.J. Beattie M.S. Mabray M.C. Pan J.Z. The Brain and Spinal Injury Center score: A novel, simple, and reproducible method for assessing the severity of acute cervical spinal cord injury with axial T2-weighted MRI findings J. Neurosurg.-Spine 2015 23 495 504 10.3171/2015.1.SPINE141033 26161519 9. Lee E. Lee J.W. Bae Y.J. Kim H.J. Kang Y. Ahn J.M. Reliability of pre-operative diffusion tensor imaging parameter measurements of the cervical spine in patients with cervical spondylotic myelopathy Sci. Rep. 2020 10 17410 10.1038/s41598-020-74624-6 33060848 10. Cui J.L. Cui J.L. Wen C.Y. Wen C.Y. Hu Y. Hu Y. Mak K.C. Mak K.C. Mak K.H.H. Mak K.H.H. Orientation entropy analysis of diffusion tensor in healthy and myelopathic spinal cord Neuroimage 2011 58 1028 10.1016/j.neuroimage.2011.06.072 21777679 11. Jin R. Luk K.D. Cheung J.P.Y. Hu Y. Prognosis of cervical myelopathy based on diffusion tensor imaging with artificial intelligence methods NMR Biomed. 2019 32 e4114 10.1002/nbm.4114 31131933 12. Li X. Cui J.-L. Mak K.-C. Luk K.D.-K. Hu Y. Potential Use of Diffusion Tensor Imaging in Level Diagnosis of Multilevel Cervical Spondylotic Myelopathy Spine 2014 39 E615 E622 10.1097/BRS.0000000000000288 24583723 13. Chan T.Y. Li X. Mak K.C. Cheung J.P. Luk K.D. Hu Y. Normal values of cervical spinal cord diffusion tensor in young and middle-aged healthy Chinese Eur. Spine J. 2015 24 2991 2998 10.1007/s00586-015-4144-2 26208941 14. Jin R.C. Hu Y. Effect of segmentation from different diffusive metric maps on diffusion tensor imaging analysis of the cervical spinal cord Quant. Imaging Med. Surg. 2019 9 292 303 10.21037/qims.2019.02.08 30976553 15. Fonov V.S. Troter A.L. Taso M. Leener B.D. Leveque G. Benhamou M. Sdika M. Benali H. Pradat P. Collins D.L. Framework for integrated MRI average of the spinal cord white and gray matter: The MNI-Poly-AMU template Neuroimage 2014 102 817 827 10.1016/j.neuroimage.2014.08.057 25204864 16. El Mendili M.-M. Chen R. Tiret B. Villard N. Trunet S. Pelegrini-Issac M. Lehericy S. Pradat P.-F. Benali H. Fast and Accurate Semi-Automated Segmentation Method of Spinal Cord MR Images at 3T Applied to the Construction of a Cervical Spinal Cord Template PLoS ONE 2015 10 e0122224 10.1371/journal.pone.0122224 25816143 17. Leener B.D. Cohen-Adad J. Kadoury S. Automatic Segmentation of the Spinal Cord and Spinal Canal Coupled With Vertebral Labeling IEEE Trans. Med. Imaging 2015 34 1705 1718 10.1109/TMI.2015.2437192 26011879 18. Chen M. Carass A. Oh J. Nair G. Pham D.L. Reich D.S. Prince J.L. Automatic magnetic resonance spinal cord segmentation with topology constraints for variable fields of view Neuroimage 2013 83 1051 1062 10.1016/j.neuroimage.2013.07.060 23927903 19. Horvath A. Pezold S. Weigel M. Parmar K. Bieri O. Cattin P.C. Variational Segmentation of the White and Gray Matter in the Spinal Cord Using a Shape Prior Computational Methods and Clinical Applications for Spine Imaging Springer Cham, Switzerland 2016 20. Prados F. Cardoso M.J. Yiannakas M.C. Hoy L.R. Tebaldi E. Kearney H. Liechti M.D. Miller D.H. Ciccarelli O. Wheeler-Kingshott C.A.M. Fully automated grey and white matter spinal cord segmentation Sci. Rep. 2016 6 36151 10.1038/srep36151 27786306 21. Dostal M. Kerkovsky M. Korit Akova E. Nemcova E. Stulik J. Stankova M. Bernard V. Analysis of diffusion tensor measurements of the human cervical spinal cord based on semiautomatic segmentation of the white and gray matter J. Magn. Reson. Imaging JMRI 2018 48 1217 1227 10.1002/jmri.26166 29707834 22. Ronneberger O. Fischer P. Brox T. U-net: Convolutional networks for biomedical image segmentation Proceedings of the International Conference on Medical Image Computing and Computer-Assisted Intervention Munich, Germany 5-9 October 2015 234 241 23. Zhang X. Li Y. Liu Y. Tang S.X. Liu X. Punithakumar K. Shi D. Automatic spinal cord segmentation from axial-view MRI slices using CNN with grayscale regularized active contour propagation Comput. Biol. Med. 2021 132 104345 10.1016/j.compbiomed.2021.104345 33780869 24. Alsenan A. Youssef B.B. Alhichri H. A Deep Learning Model based on MobileNetV3 and UNet for Spinal Cord Gray Matter Segmentation Proceedings of the 2021 44th International Conference on Telecommunications and Signal Processing (TSP) Virtual 26-28 July 2021 244 248 25. De Leener B. Levy S. Dupont S.M. Fonov V.S. Stikov N. Louis Collins D. Callot V. Cohen-Adad J. SCT: Spinal Cord Toolbox, an open-source software for processing spinal cord MRI data Neuroimage 2017 145 24 43 10.1016/j.neuroimage.2016.10.009 27720818 26. Simonyan K. Zisserman A. Very deep convolutional networks for large-scale image recognition arXiv 2014 1409.1556 27. He K. Zhang X. Ren S. Sun J. Deep residual learning for image recognition Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Las Vegas, NV, USA 26 June-1 July 2016 770 778 28. Zhu W. Huang Y. Zeng L. Chen X. Liu Y. Qian Z. Du N. Fan W. Xie X. AnatomyNet: Deep learning for fast and fully automated whole-volume segmentation of head and neck anatomy Med. Phys. 2019 46 576 589 10.1002/mp.13300 30480818 29. Tsai A. Anatomical landmark localization via convolutional neural networks for limb-length discrepancy measurements Pediatr. Radiol. 2021 51 1431 1447 10.1007/s00247-021-05004-z 33885918 30. Zhu Q.K. Du B. Yan P.K. Boundary-Weighted Domain Adaptive Neural Network for Prostate MR Image Segmentation IEEE Trans. Med. Imaging 2020 39 753 763 10.1109/TMI.2019.2935018 31425022 31. Leener B.D. Kadoury S. Cohen-Adad J. Robust, accurate and fast automatic segmentation of the spinal cord Neuroimage 2014 98 528 536 10.1016/j.neuroimage.2014.04.051 24780696 32. Yiannakas M.C. Mustafa A.M. De Leener B. Kearney H. Tur C. Altmann D.R. De Angelis F. Plantone D. Ciccarelli O. Miller D.H. Fully automated segmentation of the cervical cord from T1-weighted MRI using PropSeg: Application to multiple sclerosis NeuroImage. Clin. 2016 10 71 77 10.1016/j.nicl.2015.11.001 26793433 33. Fouladivanda M. Kazemi K. Helfroush M.S. Shakibafard A. Morphological active contour driven by local and global intensity fitting for spinal cord segmentation from MR images J. Neurosci. Methods 2018 308 116 128 10.1016/j.jneumeth.2018.07.015 30036546 34. Sabaghian S. Dehghani H. Batouli S.A.H. Khatibi A. Oghabian M.A. Fully automatic 3D segmentation of the thoracolumbar spinal cord and the vertebral canal from T2-weighted MRI using K-means clustering algorithm Spinal Cord 2020 58 811 820 10.1038/s41393-020-0429-3 32132652 35. Cadotte A. Cadotte D.W. Livne M. Cohen-Adad J. Fleet D. Mikulis D.J. Fehlings M.G. Spinal Cord Segmentation by One Dimensional Normalized Template Matching: A Novel, Quantitative Technique to Analyze Advanced Magnetic Resonance Imaging Data PLoS ONE 2015 10 e0139323 10.1371/journal.pone.0139323 26445367 36. van Hecke W. Leemans A. Sijbers J. Vandervliet E. Van Goethem J.W. Parizel P.M. A tracking-based diffusion tensor imaging segmentation method for the detection of diffusion-related changes of the cervical spinal cord with aging J. Magn. Reson. Imaging 2008 27 978 991 10.1002/jmri.21338 18425838 37. Gros C. Leener B.D. Badji A. Maranzano J. Eden D. Dupont S.M. Talbott J.F. Zhuoquiong R. Liu Y. Granberg T. Automatic segmentation of the spinal cord and intramedullary multiple sclerosis lesions with convolutional neural networks Neuroimage 2019 184 901 915 10.1016/j.neuroimage.2018.09.081 30300751 38. Datta E. Papinutto N. Schlaeger R. Zhu A.H. Carballido-Gamio J. Henry R.G. Gray matter segmentation of the spinal cord with active contours in MR images Neuroimage 2017 147 788 799 10.1016/j.neuroimage.2016.07.062 27495383 39. Dupont S.M. Leener B.D. Taso M. Troter A.L. Nadeau S. Stikov N. Callot V. Cohen-Adad J. Fully-integrated framework for the segmentation and registration of the spinal cord white and gray matter Neuroimage 2017 150 358 372 10.1016/j.neuroimage.2016.09.026 27663988 40. Prados F. Ashburner J. Blaiotta C. Brosch T. Carballido-Gamio J. Cardoso M.J. Conrad B.N. Datta E. David G. Leener B.D. Spinal cord grey matter segmentation challenge Neuroimage 2017 152 312 329 10.1016/j.neuroimage.2017.03.010 28286318 41. Perone C.S. Calabrese E. Cohen-Adad J. Spinal cord gray matter segmentation using deep dilated convolutions Sci. Rep. 2018 8 5966 10.1038/s41598-018-24304-3 29654236 42. Wang S.Q. Li X. Cui J.L. Li H.X. Luk K.D.K. Hu Y. Prediction of myelopathic level in cervical spondylotic myelopathy using diffusion tensor imaging J. Magn. Reson. Imaging 2015 41 1682 1688 10.1002/jmri.24709 25044870 43. Skotarczak M. Dzierzanowski J. Kaszubowski M. Winklewski P.J. Romanowski A. Szurowska E. Szarmach A. Diagnostic value of diffusion tensor imaging in patients with clinical signs of cervical spondylotic myelopathy Neurol. Neurochir. Pol. 2022 56 341 348 10.5603/PJNNS.a2022.0031 35471632 44. Scholler K. Siller S. Brem C. Lutz J. Zausinger S. Diffusion Tensor Imaging for Surgical Planning in Patients with Cervical Spondylotic Myelopathy J. Neurol. Surg. Part A 2020 81 1 9 10.1055/s-0039-1691822 45. Ellingson B.M. Salamon N. Woodworth D.C. Yokota H. Holly L.T. Reproducibility, temporal stability, and functional correlation of diffusion MR measurements within the spinal cord in patients with asymptomatic cervical stenosis or cervical myelopathy J. Neurosurg.-Spine 2018 28 472 480 10.3171/2017.7.SPINE176 29424671 46. Wen C.Y. Cui J.L. Liu H.S. Mak K.C. Cheung W.Y. Luk K.D.K. Hu Y. Is Diffusion Anisotropy a Biomarker for Disease Severity and Surgical Prognosis of Cervical Spondylotic Myelopathy? Radiology 2014 270 197 204 10.1148/radiol.13121885 23942607 47. Borkowski K. Krzyzak A.T. Assessment of the systematic errors caused by diffusion gradient inhomogeneity in DTI-computer simulations NMR Biomed. 2019 32 e4130 10.1002/nbm.4130 31343807 48. Ahn C.B. Cho Z.H. ANALYSIS OF THE EDDY-CURRENT INDUCED ARTIFACTS AND THE TEMPORAL COMPENSATION IN NUCLEAR-MAGNETIC-RESONANCE IMAGING IEEE Trans. Med. Imaging 1991 10 47 52 10.1109/42.75610 18222799 49. Pierpaoli C. Basser P.J. Toward a quantitative assessment of diffusion anisotropy Magn. Reson. Med. 1996 36 893 906 10.1002/mrm.1910360612 8946355 50. Borkowski K. Krzyzak A.T. Analysis and correction of errors in DTI-based tractography due to diffusion gradient inhomogeneity J. Magn. Reson. 2018 296 5 11 10.1016/j.jmr.2018.08.011 30195248 51. Bammer R. Markl M. Barnett A. Acar B. Alley M.T. Pelc N.J. Glover G.H. Moseley M.E. Analysis and generalized correction of the effect of spatial gradient field distortions in diffusion-weighted imaging Magn. Reson. Med. 2003 50 560 569 10.1002/mrm.10545 12939764 52. Tan E.T. Marinelli L. Slavens Z.W. King K.F. Hardy C.J. Improved correction for gradient nonlinearity effects in diffusion-weighted imaging J. Magn. Reson. Imaging 2013 38 448 453 10.1002/jmri.23942 23172675 53. Krzyzak A.T. Olejniczak Z. Improving the accuracy of PGSE DTI experiments using the spatial distribution of b matrix Magn. Reson. Imaging 2015 33 286 295 10.1016/j.mri.2014.10.007 25460327 54. Rogers B.P. Blaber J. Welch E.B. Ding Z. Anderson A.W. Landman B.A. Stability of Gradient Field Corrections for Quantitative Diffusion MRI Proceedings of the SPIE Medical Imaging Orlando, FL, USA 11-16 February 2017 Volume 10132 10.1117/12.2254609 55. Klodowski K. Krzyzak A.T. Innovative anisotropic phantoms for calibration of diffusion tensor imaging sequences Magn. Reson. Imaging 2016 34 404 409 10.1016/j.mri.2015.12.010 26707852 56. Zhu T. Hu R. Qiu X. Taylor M. Tso Y. Yiannoutsos C. Navia B. Mori S. Ekholm S. Schifitto G. Quantification of accuracy and precision of multi-center DTI measurements: A diffusion phantom and human brain study Neuroimage 2011 56 1398 1411 10.1016/j.neuroimage.2011.02.010 21316471 57. Melhem E.R. Itoh R. Jones L. Barker P.B. Diffusion tensor MR imaging of the brain: Effect of diffusion weighting on trace and anisotropy measurements Am. J. Neuroradiol. 2000 21 1813 1820 11110532 58. Qin W. Yu C.S. Zhang F. Du X.Y. Jiang H. Yan Y.X. Li K.C. Effects of Echo Time on Diffusion Quantification of Brain White Matter at 1.5T and 3.0T Magn. Reson. Med. 2009 61 755 760 10.1002/mrm.21920 19191286 59. Ni H. Kavcic V. Zhu T. Ekholm S. Zhong J. Effects of number of diffusion gradient directions on derived diffusion tensor imaging indices in human brain Am. J. Neuroradiol. 2006 27 1776 1781 16971635 60. Widjaja E. Mahmoodabadi S.Z. Rea D. Moineddin R. Vidarsson L. Nilsson D. Effects of Gradient Encoding and Number of Signal Averages on Fractional Anisotropy and Fiber Density Index In Vivo at 1.5 Tesla Acta Radiol. 2009 50 106 113 10.1080/02841850802555646 19052934 |
PMC10000613 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050712 healthcare-11-00712 Article Are Health Information Systems Ready for the Digital Transformation in Portugal? Challenges and Future Perspectives Teixeira Leonor Conceptualization Methodology Validation Formal analysis Investigation Resources Data curation Writing - original draft Writing - review & editing 1* Cardoso Irene Conceptualization Methodology Validation Formal analysis Investigation Resources Data curation Writing - original draft Writing - review & editing 2 Oliveira e Sa Jorge Conceptualization Methodology Validation Formal analysis Investigation Resources Data curation Writing - original draft Writing - review & editing 3 Madeira Filipe Conceptualization Methodology Validation Formal analysis Investigation Resources Data curation Writing - original draft Writing - review & editing 4 Alor-Hernandez Giner Academic Editor Mejia-Miranda Jezreel Academic Editor Sanchez-Cervantes Jose Luis Academic Editor Rodriguez-Gonzalez Alejandro Academic Editor 1 Department of Economics, Management, Industrial Engineering and Tourism (DEGEIT), Institute of Electronics and Informatics Engineering of Aveiro (IEETA)/Intelligent Systems Associate Laboratory (LASI), University of Aveiro, 3810-193 Aveiro, Portugal 2 Associacao Portuguesa de Sistemas de Informacao (APSI), 4800-058 Guimaraes, Portugal 3 Department of Information Systems, Centro ALGORITMI, University of Minho, 4800-058 Guimaraes, Portugal 4 Department of Informatics and Quantitative Methods, Research Centre for Arts and Communication (CIAC)/Pole of Digital Literacy and Social Inclusion, Polytechnic Institute of Santarem, 2001-904 Santarem, Portugal * Correspondence: [email protected] 28 2 2023 3 2023 11 5 71221 12 2022 19 2 2023 24 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Purpose: This study aimed to reflect on the challenges of Health Information Systems in Portugal at a time when technologies enable the creation of new approaches and models for care provision, as well as to identify scenarios that may characterize this practice in the future. Design/methodology/approach: A guiding research model was created based on an empirical study that was conducted using a qualitative method that integrated content analysis of strategic documents and semi-structured interviews with a sample of fourteen key actors in the health sector. Findings: Results pointed to the existence of emerging technologies that may promote the development of Health Information Systems oriented to "health and well-being" in a preventive model logic and reinforce the social and management implications. Originality/value: The originality of this work resided in the empirical study carried out, which allowed us to analyze how the various actors look at the present and the future of Health Information Systems. There is also a lack of studies addressing this subject. Research limitations/implications: The main limitations resulted from a low, although representative, number of interviews and the fact that the interviews took place before the pandemic, so the digital transformation that was promoted was not reflected. Managerial implications and social implications: The study highlighted the need for greater commitment from decision makers, managers, healthcare providers, and citizens toward achieving improved digital literacy and health. Decision makers and managers must also agree on strategies to accelerate existing strategic plans and avoid their implementation at different paces. digital transformation health information systems emerging technologies Health 4.0 empirical study Foundation for Science and TechnologyUIDB/00127/2020 This research was funded by Foundation for Science and Technology, in the context of the project UIDB/00127/2020. pmc1. Introduction Globalization, associated with many rapidly evolving factors, such as the COVID-19 pandemic, leads to an ever-increasing need to share health data outside of the physical space where they are generated . Demographic changes, increased chronic diseases, rising health spending, and fairer healthcare access are global challenges , which, if associated with the increase in people's average life expectancy and the growth in their literacy, show the greater importance of new Health Information Systems (HISs) that allow efficient communication between the Health Systems (HS) and their stakeholders. This has resulted in recent years in a new generation of emerging technologies that offer new opportunities for healthcare delivery and the practice of medicine while also ensuring greater efficiency of HIS and more responsive communication channels. HIS includes mechanisms for capturing, processing, analyzing, and transmitting any needed information in health services whilst also having an important role in care planning, management, and even in research for public health . With a growing need for decentralized and remote work, HIS also plays a key role since they support communication among geographically dispersed actors, promote solutions to value chain management, and support new business models. In addition, digital health offers a valuable opportunity to handle health issues, such as the pandemic situation, with near real-time responsiveness . This trend is transversal to other areas of knowledge, with different types of applications, as demonstrated in the study of Epizitone et al. . Information and Communication Technologies (ICT) applied to the health context have been the subject of several research works, with a greater incidence in the Digital Transformation (DT) of this sector , such as processes related to healthcare delivery models and medical practices . Some studies point to digitalization as a future priority in the health and public sector, reinforcing the need to adopt intelligent technological applications and connectivity mechanisms . Cavallone and Palumbo , for example, stated that Industry 4.0 (I4.0), Artificial Intelligence (AI) and Digitalization are revolutionizing the design and the delivery of care. Despite the progress observed in recent decades, the gap that emphasizes the need to create solutions with responses to extreme events is demonstrated by some authors . In addition, Gehring and Eulenfeld argued that there is still a pressing need to significantly improve the infrastructure and functionalities of the HIS for the benefit of users and also for research in areas such as biomedical sciences, health sciences, and also computer and information sciences. Knowledge of the current HISs' development state is a requirement when researching future directions. An analysis of current HISs identifies five different groups of obstacles that limit these systems' application and development--(i) technical problems, (ii) usage problems, (iii) quality problems, (iv) operational functionality, and issues related to (v) maintenance and support. In addition, the study of Khubone et al. discussed a set of challenges that should not be ignored when adopting HIS, which are mainly related to the lack of technical consensus, poor leadership and limited human resource, staff resistance and lack of management, and non-engagement of the users. In the area of telemedicine, Tabaeeian et al. compiled a set of barriers that should not be ignored when implementing such solutions, concluding that "the future of telemedicine depends on consistency in system usage and minimizing problems, increasing system compatibility with users and learning how to use". In turn, the development of a HIS requires several connections between local systems, which can be small county or regional systems or national platforms, a connection between different sectors (such as public, private, and other), and can even need the articulation at an international level (e.g., between European countries, or USA and Canada). There is a generally recognized importance of HISs by health policymakers. In Portugal, this is a reality reflected by the creation of the National Strategy for the Health Information Ecosystem 2020 (ENESIS-2020) . This working framework elected the main goal to have more efficient information processes that would (i) increase the general sharing of information and knowledge between all actors to promote the literacy and general health of citizens; (ii) offer greater efficiency for healthcare providers; (iii) offer greater rationalization of resources with an impact on global efficiency and health management; and (iv) offer an alignment of HIS strategies with other European countries (i.e., standards and interoperability). Considering the impact that these systems have on healthcare management and the medical practice, whilst also considering the current technological trends currently reported in the literature, this study aimed to understand the current state of HIS in Portugal and its main challenges, as well as foresee future trends about the use and impact of this type of systems that can benefit with emerging I4.0 technologies. The authors adopted a two-phase methodological approach. First, the authors carried out a literature review, covering topics related to the role of HIS with existing technologies, namely the ones that could enhance the development of these systems. A focus on digitalization technologies was pursued. Secondly, the results from the literature review were compared to the current Portuguese HIS situation. This was achieved by conducting an empirical study based on interviews with some representative HIS actors in Portugal while also considering some official documents, legislation, and strategies/policies currently in force in Portugal. Theoretically, this work aimed to advance some likely future trends and scenarios for the Portuguese's HIS based on the Industry 4.0 drivers. In practical terms, it was expected to provide some recommendations to practitioners and decision makers on the opportunities of DT and the expected main impacts on the Portuguese HIS. As an additional contribution, it was also intended to identify some aspects that would enable societal health empowerment through the adoption of emerging technologies. This article is structured into six sections. The current section presents the gaps, the motivations for this study, and the main objectives to be achieved. In Section two, a literature review is conducted. The third section describes the methodology adopted in this study. Section four presents the main results, challenges, and future scenarios for HISs in Portugal obtained from the point of view of the study participants (i.e., the interviewees) and from documents, legislation, and strategies/policies analyzed. In Section five, the authors present their views on future challenges and scenarios for the Portuguese HISs. Finally, the last section is devoted to some final remarks. 2. Theoretical Framework 2.1. Evolution of HIS to Date The practice of medicine has changed significantly in the last 50 years, with ICT making a strong contribution to this change. There have also been changes in the information paradigm itself, moving from institution-centered information to a more patient-centered approach . Due to globalization and other circumstantial factors, such as the COVID-19 pandemic, there is an increasing need to share health data outside of the physical space where they are generated . Furthermore, the change in the clinical information consumption patterns, where the citizen assumes an increasingly significant role, is also a reason that highlights the importance of HIS. The final use given to the data, which before was focused on responses to clinical practice, has gradually assumed greater importance also in health planning and clinical and epidemiological research. Over the past decades, HIS has had various stages, as briefly described in Figure 1, which presents HIS usage in health in four digital eras--from Health 1.0 to Health 4.0. The first era, also called Health 1.0, began in the 1960s and was associated with the introduction of patients' records. During this period, HIS was exclusively designed to store patients' information locally, in paper or digital format. Access was limited and only available in each service, department, or institution. Afterward, in the Health 2.0 era, which started in the late 1980s, HIS was further developed, so to allow the grouping of patients' data in digital repositories with private access to authorized users and necessary services, materializing the Electronic Health Record (EHR). The information was then citizen-centered , increasing the role of patients/citizens in HISs as they started to have limited access to the information recorded by health professionals. The Health 3.0 era, which began in the early 2000s, was characterized by the development of Personal Health Records (PHR). During this time, the main goal of HISs was to support the citizen's life cycle, with data introduced by both healthcare providers and the citizen himself. This further advancement of HIS allowed patients to engage proactively and collaboratively in their care. Thus, data were co-created and maintained by both providers and patients , and society moved from institution-centered information to a more patient-centered approach . PHR quickly became attractive, as it allowed the centralization of each citizen's health data in digital platforms widely available, engaging both providers and receivers in healthcare deliverance. Nevertheless, most countries and providers still felt that there was a need to have a connected decentralized health system. This was made possible by new communication platforms and emerging technologies, as well as the use of Artificial Intelligence (AI), developed during this past decade. Personalized Health Information Systems prevail, but with a vastly improved connectivity between healthcare actors, in what is called the Health 4.0 era. 2.2. The Fourth Industrial Revolution and the HIS The creation of digital ecosystems through a set of tools that enable the connection between the digital and the physical worlds is paramount in Industry 4.0 (I4.0) technologies. This last concept, which is a product of the fourth Industrial Revolution (4IR), is intricately related to principles such as interoperability, decentralization, real-time responsiveness, data-based services, virtualization, and modularity . I4.0 uses, for the technologies developed, a variety of concepts such as automation and data exchange through cloud computing, Big Data, Internet of Things (IoT), Robotics, 5G Technologies, Virtual and Augmented Reality, Additive Manufacturing, Cyber-Physical Systems, and AI, among others . With the promise of empowering the creation of better healthcare services, 4IR enhances the personalization and individualization of the services provided, the optimization of resources associated with the practice of medicine, as well as the promotion of health based on preventive models . Concepts include Health 4.0 Healthcare 4.0 , Medicine 4.0 or Care 4.0 , Hospital 4.0 , or more specific applications such as Surgery 4.0 , represent only some approaches/applications that make use of I4.0 emerging technologies to create new models of medical practice and health promotion. These concepts, in their different terminologies, represent just some extensions of the I4.0 principles applied to the medical/health area. Several authors considered that Health 4.0 could not be dissociated from the Digital Transformation (DT) concept, as the latter is used not only in the deliverance of care but also in the governance processes of all the value chains . Health 4.0 makes possible the future virtualization of healthcare delivery and medical practice . Connectivity and computing power, enhanced by emerging technologies, are crucial factors in the deliverance of need-oriented care, considering individualized approaches based on preventive and predictive models. The emerging technologies of I4.0, when applied to healthcare, can greatly enhance the productivity of the providers, as well as promote the creation of preventive care models since they allow for early detection of health-related anomalous situations . It is then possible to avoid future health issues (and costs) for both citizens and society alike. Some other applications include, for example, additive manufacturing or 3D printing, which allows faster and more personalized creation of health products, such as implants, tools, and specific devices, according to the different needs and requirements of each patient ; or robotics, which can be used in surgery and physiotherapy services, fostering improvements in performance, movement, and control. IoT allows connectivity with mobile and other devices, enabling the automatic collection of human data. Other examples include Big Data which, in addition to storing a large amount of data, allows, through Data Analytics, the identification of patterns and trends, enabling the decision-making on predicted problems of future health events. Finally, AI can help manage and analyze data, make decisions, and identify and forecast upcoming health trends or issues . Recently, Large Language Models (LLMs) with billions of parameters have brought a big boost to the base models of deep learning, and several architectures have emerged, such as ChatGPT and BERT (which are among the best known). The research and interest around these deep learning technologies are huge and promising. ChatGTP, launched in late 2022, presents both great potential and challenges for the Medicine and Healthcare sectors. AI-assisted technologies have for some time (and with different degrees of success) been employed in several aspects' areas of healthcare. For example, in 2016, IBM launched Watson for Oncology , an AI clinical decision-support for cancer treatment, that achieved moderate success before being discontinued for failure to achieve the same clinical marks as real-time physicians . Other examples come from radiology or pathology, where AI-assisted tools are being employed to identify tumors, differentiate between healthy and abnormal tissue samples, and provide clinicians with diagnostic suggestions, which lead to faster and more efficient results and prompt treatment . So, if AI is not new to the healthcare sector, it has mostly been used in specific areas and as support for clinical decisions. ChatGTP is different! Firstly, it was not designed with a specific medical intention in mind. Secondly, it is widely available. Finally, it has shown a considerate level of clinical accuracy , e.g., achieving the passing threshold of the USMLE (the United States Medical Licensing Examination). These characteristics can lead to important innovations in the healthcare sector, namely in developed countries, where this sector is struggling to deliver good healthcare in an aging society:Triage of patients--LLMs, such as ChatGTP, can be used as primary points of contact between the patient and the healthcare system, triaging patients and decreasing the burden on the healthcare system. Moreover, these tools can also reduce clinical biases, providing a standard of consideration to every patient, independent of personal characteristics. Medical scribe functions--modern healthcare systems require the input by the physicians or their assistants of large amounts of data. LLMs can be used to help or reduce this workload, performing note-taking tasks and writing brief patient summaries and presentations. One such example is a recent Microsoft announcement that Teams would provide note-taking features for meetings . Diagnosis assistance--LLMs can become important tools to help clinicians to make an evidence-based differential diagnosis as unbiased tools that can be trained not only with large amounts of medical information but which can also be updated with the latest relevant data, including innovative academic studies or clinical trials. LLMs can also drive innovation and competition in the healthcare sector. Medical Sciences are, by nature, an uneven market field, where the provider (physicians) have all the knowledge, and the user (patient) does not know what he/she is getting before the service is complete. The digital revolution of the past few decades has reduced this gap, both empowering patients with information and promoting health literacy in general. Just in Europe, it is estimated that half of the patients look for health information online . LLMs, such as ChatGTP, can increase this trend, promoting more informed choices and leading to more demanding customers (i.e., patients). This has the chance to drive innovation and promote excellence across the medical field. There are, nevertheless, some challenges. ChatGTP, for example, remains an imperfect tool, with the CEO of OpenAI, the company behind this tool, recently twittered that ChatGPT remains "incredibly limited", and that "It's a mistake to rely on it for anything important right now" . Obstacles such as misinformation, artificial hallucination, data protection, or ethical questions remain relevant and should, if not limit, at least warrant some cautions in the use and dissemination of these tools. In conclusion, LLMs present significant opportunities for the healthcare sector, but a careful approach involving practitioners, patients, policymakers, and other relevant field professionals are needed before they become mainstream. With all these developments, the Health 4.0 concept is closer than humanity imagines. Thus, the main pillars of Health 4.0 (and its derivations) are framed within digital ecosystems and are focused on people, technologies, and co-design because it presupposes a change in hospital business models to an ever-increasing citizen-centered care provision. Technology insofar represents the drives that are at the basis of the Health 4.0 concept itself, and without which its implementation would not be possible. Finally, to co-design patients' involvement is a requirement, not only in the HISs as active actors but also in the design and development of these systems, to allow their future participation . 3. Materials and Methods Starting from the objective that supported this research (to understand the current state of HIS in Portugal and its main challenges, as well as foresee future trends), an empirical and exploratory study was carried out, supported by a qualitative methodological approach, whose research design is presented below. So, the research protocol starts with a comprehensive literature review to define the boundaries of the research subject and the research questions, as well as to build the interview guidelines. Next, an intentional sample was selected from a population of HIS users in the private and public Healthcare organizations and Governmental entities involved in the definition of the HIS strategy in Portugal. In the third phase, the fieldwork that consisted of the execution of the interviews, as well as the selection of strategic documents, was conducted. The fourth stage comprised the processing and analysis of data, and finally, the analysis of findings and the consequent production of conclusions. Figure 2 presents the research design followed in this study, where the main research question (Q1) was broken down into two sub-questions (Q1.1 and Q1.2), and these in other more specific questions. To collect data, different sources were used, which are broadly categorized as (i) analysis of strategic documents and legislation; and (ii) interviews. Strategic documents report a set of initiatives launched by government entities that can regulate and promote approaches for modernization in this sector. These documents included digital platforms, official documents, and legislation produced by the entities responsible for the definition of strategy and management of information on the health ecosystem at the national level, i.e., ENESIS-2020 and ENESIS 20/22 . The interviews were selected insofar, as they are considered one of the most proper methods to explore participants' experiences and/or reconstruct past events. 3.1. Data Collection Methods and Procedures As mentioned, the qualitative approach was adopted in the data collection, combining: (i) the analysis of strategic documents and legislation; and (ii) semi-structured interviews conducted with different entities involved both in the definition of the HIS strategy in Portugal and as users of these HISs. The interviews (audio-recorded) were applied between August and December 2019, following previously developed scripts, oriented and adapted to the interviewees' profiles, and structured according to the specific goals mentioned in Figure 2. Each script included ten questions in addition to those that anonymously characterized the interviewee. Three types of interviewees were found, namely: (i) managers, which included health professionals with management or coordination positions; (ii) health professionals (physicians and nurses); and, also, (iii) users of HS. To conduct the interviews, the participant's consent was obtained, and confidentiality and anonymity were guaranteed. All participants were interviewed in person by the researchers. 3.2. Data Analysis Methods Due to the nature of the data obtained, qualitative analysis methods were used. Both the strategic reports and the transcribed interviews were subjected to a thematic-categorical content analysis, which represents a technique to capture the meaning of texts relating to a particular phenomenon under study . For this purpose, the typical phases of content analysis were followed, which are based on: (i) the organization of the material and the definition of the procedures (pre-analysis); (ii) the identification of the categories that arise with the interpretation of the text (exploration); and finally, (iii) the treatment of the results, where we sought to interpret the data around the categories created. A manual coding procedure was used in this process. 3.3. The Sample Profile of Respondents Given the exploratory nature of the study, intentional sampling was chosen. To minimize the limitations caused by the reduced sample size, particularly those related to the replicability and reliability of the study, we sought to diversify the demographic regions from where the participants originated. Table 1 presents the sample, composed of fourteen participants distributed as five health professionals (coded with the suffix P), six health professionals with management positions (suffix M), two users (suffix U), and one member of a governmental entity (suffix GE). Table 1 also presents other data that allow for a better characterization in terms of the region and organization to which they belong, the regime (public or private) in which they work, their profiles and positions, as well as their occupations and age groups. All interviewees mentioned being users of the HIS, although to different degrees. 4. Results Based on the analysis of the interviews and some strategic documents , it was intended to answer the major research question Q1, see Figure 2. To achieve that goal, it is equally important to obtain answers to questions Q1.1 and Q1.2., i.e., to understand how the current state of the HIS in Portugal is as well as their trends in terms of future development. 4.1. Analysis and Reflection on the Current HIS 4.1.1. Current HIS Situation: Document Analysis From document analysis, which includes online platforms, and legal documents that approved strategies for HIS, for the last two three-year-olds , it was perceived that police decision-makers consider that IS Healthcare can act in any organization that involves healthcare (public and private hospitals, clinics, clinics, pharmacies, nursing services, and primary healthcare, among others). According to the Resolution of the Council of Ministers of 26 July 2017 , the HIS seems to include "all local and central information subsystems, in the entities of the National Health Service (NHS) and third parties integrated with it, to make available to several users all useful information to health literacy and health self-management (citizens), providing healthcare (health professionals), system management (local and central managers), health research and cross-cutting needs for public administration" . The main strategy for HIS, entitled Health Information Ecosystem Strategy 2020 adopted by Council of Ministers Resolution No. 62/2016, replaced by ENESIS 20/22 , was approved on 7 January 2020 after public consultation. The ENESIS 2020 , as well as the following (ENESIS 20/22) , assume the objective of promoting the Digital Transformation of the health sector in Portugal and creating the conditions that allow the evolution of the Health Information Ecosystem (eSIS). They seek to respond to the priorities defined in terms of health policies, extending to the entire Health System and ensuring a common vision for the area of IS/IT. Analyzing the strategy still in place, ENESIS 20/22, the authors considered in general terms that it promotes a citizen-centered approach, ensuring simple and timely access to healthcare and improving his/her experience with the system . The implementation of the strategy was structured in a set of six axes, namely: (i) access to healthcare throughout the life cycle of the citizen; (ii) training and empowerment of citizens; (iii) efficiency and sustainability of the health system; (iv) quality and safety of healthcare; (v) prevention, protection, and promotion of health; and (vi) training of professionals in organizations. Some of these axes (e.g., access to healthcare throughout the life cycle of the citizen; training and empowerment of citizens) seem in line with some literature which says that the healthcare sector begins to adopt a perspective less and is less based on hospital space and health professionals, and more focused on the citizen and their needs. Yet, in relation to the aspects such as efficiency and sustainability of the health system, quality and safety of healthcare, prevention, protection, and promotion of health, and training of professionals in organizations, they can find in other authors. Regarding the Health Information Ecosystem in Portugal and considering the results obtained from documental analysis (placing references to the websites and the legislation/strategy), the authors could verify that there was a comprehensive evolution of the HIS in Portugal, which followed the evolution of the HIS in general and in several countries of Europe, which is reflected in its strategy. Adopting a holistic and citizen-centered view, the HIS in Portugal increasingly tries to respond to the information needs around the life cycle of the person, from birth to death, being visible in the definition of the Health Information Ecosystem (eSIS), "a set of technologies, people and processes that intervene in the life cycle of information related to all dimensions of the health of citizens (...) regardless of the place of care and/or organizational barriers (p. 3736)". It is precisely in this vision that the authors make the description of the HIS existing and/or under development in Portugal. In Portugal, many HISs can be found in various areas (Administrative and patient management, Clinical, Financial, Management and planning, Informative and IT, and Communication), according to SPMS . However, the authors focused their analysis primarily on HIS connected with the life cycle of citizens. Thus, following a citizen-centric perspective, Table 2 presents the main HIS classified according to the different stages of people's life cycle, i.e., birth; (ii) health and well-being; (iii) disease, which may be acute and/or chronic; (iv) aging; and (v) Death. Some of the HISs are transversal to the various stages of the life cycle, serving the citizen from birth to death, such as the 'National Register of Users' (RNU) and the 'Electronic Health Record (RSE). 4.1.2. Findings from Interviews The results obtained from these data are presented according to the following themes that have already emerged from the literature review with the unfolding of the research question: (i) the state of digitalization of clinical information processes; (ii) the impact of DT on the HS; (iii) difficulties and benefits of operating with digitalized processes; (iv) conditions for dematerialization; and (v) reduction of info exclusion. The State of Processes' Digitalization The results pointed to two distinct groups of respondents' thoughts. The first was satisfied when asked about the state-of-the-art of HIS in their organizations and claims to have achieved a significant digitalization through DT, as can be read in the following statements: The process of digitalization is almost consolidated in the public sector, (...), we have problems with very large hospitals with very old and poorly computerized systems, both at the regional and national levels that delay the process as a whole. (E13-GE) We are taking steps towards full integration in terms of system development (...). Portugal is far ahead, compared to many countries, in Europe. (E11-M) We are even implementing the paperless hospital project which is a project that has had good and referenced results, our hospital can get 70% of patients to leave the hospital without paper. (E10-M) The second was more cautious in their statements, being more skeptical, saying that although there has been a significant advance in recent years, there is still a lack of integration and communication between systems, as interpreted by the following comments: (...) it has evolved into a strategic concept, but there is a long way to go, there is a lack of clinical information in the interaction between private and public institutions (...) (E6-P) (...) there is still a lot of lack of communication between hospitals and health centers. A lot of time is spent transcribing the analyses (...) it is necessary to continuously improve the software (...) (E3-M) Impact of Digital Transformation on the Healthcare In general, respondents saw digital health as a way to promote health. One interviewee referred that DT supports another way of doing medicine and promoting citizens' health with an impact on the creation of new business models in clinical practice. Digital health is another type of (...) health service, and another way of doing health" (...) to telehealth, but is connected for example with preventive medicine, with precision drugs. (E13-GE) However, and in general, despite the potential benefits that the interviewees see in the eventual digitalization of processes, they were also unanimous about the challenges that DT brings. Thus, some benefits and challenges emerged, as can be seen in the following comments: These platforms can complement and help diagnosis (...) (E1-U) The potential to do good, to change health (...) the transformation of health, from one-on-one health practice to a population health practice (E11-M) There are already positive impacts, but the centrality of the patient in the system does not exist (E6-P) Difficulties and Benefits of Operating with Digitalized Processes As regards difficulties in the operationalization and use of systems, some respondents with management responsibilities reported difficulties in the human resources area related to resistance to change. Additionally, difficulties associated with the functioning of the HIS, namely: slowness, redundancy, lack of response to clinical practice, and lack of interoperability (communication and integration) between public and private institutions, were also pointed out by some respondents: (...) it has more to do with people than with technology (resistance to change, stability of teams, continuous training) (E11-M) (...) lack of integration of systems between institutions (E2-P) (...) there is no standardization of the systems themselves, they are always different systems (E4-P) Regarding the expected benefits of digitalization, respondents (E3-M) (E4-P) (E5-M) (E10-M) perceived some benefits in the digitalization process, such as (i) greater agility in information flows, with access to information in almost real-time; (ii) more security, in terms of access to data only by authorized users; (iii) less propensity to human error; (iv) greater capacity to share information between services and organizations providing of healthcare; (v) easier of access to health services without geographical restrictions; (vi) reducing paper circulation and consumption, and (viii) cost savings often associated with the repetition of exams to better understand everything that surrounds the activity of healthcare. Conditions for Digital Transformation The results pointed to the need to work on the aspects related to legislation, namely with governmental entities of each country and even outside the country. For example, in the sharing and use of health data from wearable devices, the greatest trouble is with the issues of use and legitimacy. With the technologies currently available, citizens can generate data to complement their health records through smart devices. To support the use of these technologies for health, legislation and some joint work with health regulators is needed. The results of the interviews also suggested the need for standards to achieve the interoperability of systems at the international level. (...) it is necessary to create legitimacy to define the use of digital technology [wearables] because it is different if I use it to record my health data, or if the data generated by these technologies can be used to make diagnoses or suggest therapy. This is too important to be seen at an international level to define interoperability standards and rules (E13-EG) The ENESIS-2020 and ENESIS-20/22 assume the objective of promoting the DT of the health sector in Portugal and creating the conditions that allow responding to the priorities defined in terms of health policies, extending to the entire HS and also privileges a citizen-centered approach, ensuring simple and timely access to healthcare and improving his experience with the system. Conditions Required for the Reduction of Info Exclusion When it comes to ensuring the digitalization of processes, the question arises about citizens' digital literacy, which may represent an enabling factor or an obstacle in the functioning of processes in that ecosystem. The reduction of info exclusion was referred to by 12 (86%) of interviewees as an important challenge to overcome. On this subject, the interviewees reported that the reduction of illiteracy depends on several factors, namely: (i) the opening of the user to this type of knowledge and innovation; (ii) the development of accessible systems; and, also, (iii) the monitoring and training of older people and/or those with greater difficulty to ensure health literacy and digital literacy. The responsibility for this training and monitoring would be on the government, health organizations, educational institutions, and community entities, such as the town halls and parish councils. Supporting these findings, we have the comments presented below: (...) the current population is very ageing (...) it does not easily adapt to IT. The state should ensure minimal training, monitoring, and simplify the development of these technologies (...) (E1-U) (...) we have pioneering projects such as Citizen HOSP that, through our social workers, support users to take advantage of the use of IT in access to health services (...) (E10-M) The implementation of ENESIS-2020 and ENESIS20/22 , developed in recent years, is still in the consolidation phase, presenting, however, a significant advance in terms of the number of new digital services, namely digital platforms such as the Health Web Portal and other applications that can be accessed from the Citizen Web Portal. These systems, in turn, tend to respond to certain stages of a citizen's life, ranging from the simple registration of citizen follow-up to systems that accommodate the entire clinical history when in situations of illness (severe or chronic). This concept appears implicitly in ENESIS-2020 , defining it as a set of technologies, people, and processes that intervene in the life cycle of information related to all dimensions of citizen health and related, regardless of the place of care and/or organizational barriers. 4.2. Prospects for HIS in Portugal and Scenarios Looking at future scenarios and based on the analysis of the results, three categories were shown that allowed to outline some scenarios for the future of the HIS in Portugal (i) Medicine Practices; (ii) Technologies; and (iii) Fears and Challenges. 4.2.1. Medicine Practices The way the interviewees saw the evolution of the practice of medicine associated with technological evolution was dual. If, on the one hand, they understand that technologies combined with the greater use of AI will make medicine richer, more dematerialized, advanced, and effective (namely by speeding and accuracy in diagnosis, self-learning, and knowledge because it is based on predictability and allows personalized treatments), with an impact on the increase in quality average life expectancy, on the other hand, they consider that these benefits cannot imply the loss of the doctor-patient relationship nor can the data overlap to the psychological reality of the patient in the interpretation of his disease. (...) Reduction of doctor-patient contact because computer solutions will compare certain standards by AI and allow diagnostics, without the patient presence. (...) a great combination of general medical knowledge with computer knowledge (E1-U) The practice of medicine in the future will be more dematerialized, remote[telemedicine], a preventive and precision medicine (...) the citizen will be more involved in his/her health/disease and the decisions about it, he/she will now his/her test results, and he will already bring the data stored in digital media. (E13-GE) The last comment highlighted the core of current health strategies and policies, which focus on the citizen/patient, and the centrality of the patient with his/her information would enhance preventive models in health and accuracy in personalized diagnostics and treatment. However, it should be noted that the results also identified a gap between the progression of technology and the way healthcare is organized, the latter being associated with strategy, management, and legislation: (...) technology is evolving and the way we organize ourselves to supply healthcare is not advancing at the same pace. (E13-GE) 4.2.2. Technologies On this subject, some health professionals were skeptical about the adoption of technologies in medical practices, believing in a worsening of the social part, with a negative impact on the patient-doctor relationship, contrary to other professionals who had an optimistic view of the future of HIS. (...) there has been a decrease in doctor-patient confidence and (...), this system, although useful, can aggravate even more this situation (E9-P) (...) we need robust systems to treat this information, such as Business Intelligence or Data Mining, which are being implemented in our hospital (...) technologies allow us to innovate health, better manage resources, know patients (...) (E10-M) In ten years, I think it will be possible to computerize almost total medical information (E9-P) Some interviewees, such as (E13-GE) and (E1-U), went further in what they think is the future of HIS, referring to the decrease in interfaces between technology and users, the existence of speech recognition software to support health professionals filling EHRs, the increasing existence of intra-devices, and the storage of data by patients themselves for reasons of cybersecurity, privacy, and data sharing. 4.2.3. Fears and Challenges Faced with the idea of a fully digitalized reality, most respondents mentioned concerns about the confidentiality of information (who accesses the data and with what intention), as well as other types of threats (e.g., cyberattacks). The dehumanization in the provision of care also arises as an apprehension that stems from the reduction of physical contact between doctors and patients (for example, health professionals can make decisions based on a set of data without the need for the physical presence of the patient), thus losing the emotional aspects characteristic of human interaction, typical of traditional medicine. The comments presented below reflect these fears: (...) lose the patients' data (...) (E2-P) (...) who has access to this information and what are you going to do with it? (E11-M) Some of the pointed-out fears and negative opinions about ICT evolution can be seen as challenges by HIS developers. The following comments can be illustrative: (...) doctors must rediscover themselves, as coachers, people who guide reading (E13-GE) (...) the data still cannot sustain the psychological reality of the patient in the interpretation of his disease (...) we must not forget that we have biological complexity and that the Human being is not purely data. We do not treat data, we treat people (E6-P) 4.2.4. Possible Scenarios Based on the previous categories and following the three dimensions identified and described above, some relevant concepts were found, which allowed the design of future scenarios for the HIS. (i) Medical Practices--the following concepts were found:Precision Medicine/Individualized--a medicine whose treatment is specific to a particular patient. Preventive Medicine--in which the focus is to keep healthy instead of curing the disease. Point-of-Care (Telemedicine)--allowing citizens to be physically distant from medical centers to have access to expert diagnoses. Assisted Medical Practices--in which machines (robots) with AI embedded start to help or even replace health professionals in various medical acts. (ii) Technologies--the concepts found are:Interoperability--integration of HIS with inter-organizational information exchange between public and private and national and international entities. Digital Health Transformation--health processes are aided by technologies. Technology to Assist Medical Practices--such as robots and AI, among others, helping health professionals. Use of electronic devices (wearables), robotics, and intra-devices--used by citizens to monitor their health, with the possibility of collecting data and sending these data to health entities/health professionals. (iii) Challenges and Risks--the concepts are:Resistance to change (people, users, and health professionals) Information exclusion (Training/Monitoring) Information (privacy, quality, and security--access and loss) Given these perspectives, three HIS scenarios were developed for the future, i.e.,: (i) realistic; (ii) pessimistic; and (iii) optimistic scenarios. Table 3 describes the scenarios for Medical Practices. Table 4 shows the concepts related to Technologies. Table 5 outlines the scenarios related to Challenges and Risks. 5. Discussion and Reflection The health sector has, from early on, incorporated the benefits of the use of information technologies, with the first era of HIS that supported caregivers in their practice arising in the 1960s. The evolution has been remarkable, and the EHR, initially held by healthcare providers, is now run by citizens, true owners, and interested parties of their health data. However, this change requires an effort from all agents in terms of the design of "future" HIS, where aspects such as interoperability, standardization, privacy, security, and actions to address info exclusion appear as striking challenges. These factors and challenges are shared by the respondents, as shown in Table 6. However, a greater or lesser sharing should not be understood as a greater or lesser importance of each of the challenges but rather demonstrates the degree of awareness of these factors and challenges by the group of respondents. This change has been driven by social factors, such as increased life expectancy and mobility, but also by a set of ambitious responses and strategies created by healthcare decision-makers and managers. The reinforcement and commitment to HIS are expected to provide an adequate response to the health of citizens and guarantee the overall sustainability of the system. Thus, the technological advances brought by the 4IR have had a significant impact and acceptance in this sector, and the pandemic has further accelerated its adoption. The Health 4.0 concept incorporates innovative technologies which promote substantial improvements in health services and facilitates a more citizen-focused model concerning health. The implementation of HIS in Portugal takes place at different paces, depending on the areas of activity, the type of sectors involved, and the legal regime of the organization, among other factors. It should also be noted that existing strategic documents specifically cover the public health sector and therefore do not consider private health entities; although the proposed ENESIS-20/22 strategy refers to a wider health ecosystem, in practice, the approaches presented are very much public-sector-focused. In line with Ciasullo et al. , to make the HS sustainable, it is necessary for citizens to actively take part in their health process, and to do so, they need to have adequate means and knowledge, as well as the ability to interpret their health data, which requires an investment in their digital and health literacy. Likewise, health organizations, public or private, together with industry regulators and those responsible for health strategy, need to work together to enable (digital) communication and integrated sharing of health data, particularly in the context of healthcare. In addition, emerging I4.0 technologies have enhanced the creation of better conditions for data collection and information sharing, as reported by some studies . Nowadays, there are already several types of equipment that can be used or applications that can be installed on personal mobile devices, which allow monitoring the parameters of health and quality of life . It is, therefore, a pressing thing to evaluate and classify these types of devices and applications in terms of quality and reliability and to legitimize them for this purpose so that citizens and health professionals can see their usefulness, have confidence in their use, and can use them by increasing preventive and predictive health models. Given the current technological context, the citizen, in addition to standing as a fundamental actor in data creation, can also play a relevant role as a consumer of information. As such, it is also important to provide citizens with access to health-related information, for example, through the EHR, thereby enhancing better decisions about their health while promoting a preventive health model, concerns already seen in other studies that seek citizen centricity . Another important aspect that was highlighted in this study is the need to raise awareness and training of the citizen so that they are the main promoters of their health which confirms the results presented by Rahi et al. . Thus, it is necessary to ensure that the citizen has conditions for this, such as (i) empowering the citizen to use the systems; (ii) raising citizens' awareness to manage their health data, as well as sharing them with the professionals responsible for monitoring them, in a digital, holistic, and integrated ecosystem; and (iii) empowering citizens on the correct use of digital health solutions. The existence of health data, part of them collected by the citizens themselves, as well as the later exchange of this data between patients and health professionals, would certainly allow an increase in the value of the services, enhancing benefits for both parties. It is essential to ensure reliable and quality data sources, as well as their protection, privacy, and security. In a more social and human aspect, and in line with what has been proposed in several other studies , it is important to highlight the importance of health professionals with adequate skills to implement DT in the health sector and, consequently, develop new processes in terms of healthcare and/or restructure existing ones, as well as the training of citizen so that they are the main responsible and promoters of their health. To conclude, several new trends, described around three scenarios, which may emerge in the future with the adoption of emerging I 4.0 technologies, should be noted, although these require a new strategic approach, with appropriate action plans to promote accessibility for all citizens and professionals. 5.1. Theoretical Implications This study reinforced the literature's long-held view about the importance of future health promotion strategies through I4.0 emerging technologies. The findings of this study confirmed that the digitalization of processes in healthcare can bring benefits to stakeholders while also bringing some challenges that should be properly addressed beforehand to maximize positive results. Furthermore, these findings could be used by researchers in Business Management and Information Systems areas to advance novel solutions to e-health-related sectors. 5.2. Managerial and Societal Implications This study has several implications that are useful not only for health providers and receivers but also for society in general. Our findings showed that there is a need for greater commitment from managers and decision makers to invest in solutions that allow a more equal DT approach between different health institutions/sectors. In addition, decision makers should promote processes that not only support interoperability, respecting data privacy and security but also increase the digital literacy of all healthcare providers and the health literacy of system users. 6. Conclusions and Recommendations This study assessed the current state of HIS in Portugal, verifying that there is a strategic alignment with our European partner countries in terms of legislation and definitions of health policies. Although there has been continuous redesign and the emergence of several applications/solutions for different processes, Portuguese health systems remain incomplete, with gaps to be filled in legislation and the adoption of innovative health-care processes by organizations. Moreover, struggles with integration and interoperability between solutions from the public and other sectors (private and social), or even from the same institution, not only lead to substantial costs in terms of redundancy and consistency of information but also reduce the interaction with users (healthcare providers, managers, and citizens). There were some other issues found, namely the need to improve digital literacy in all actors involved, as well as the urgency to increase citizens' health literacy, both tasks requiring significant educational effort. Additionally, there remains some resistance to change. Nevertheless, the benefits expected (some already verified) by the different parties involved, such as the dematerialization, digitalization, and incorporation of emerging technologies, showed that there is an effective process of health DT in Portugal. When it comes to the future of HIS, three possibilities which include a pessimistic, optimistic as well as a more realistic scenario, were outlined based on the Portuguese case. These fell into three main categories: (i) Medical Practices; (ii) Technologies; and (iii) Fears and Challenges. Limitations and Future Work Firstly, a limited number of interviews were included. Even though an attempt was made to have individuals from different professional backgrounds and geographical areas, the results stood for a fraction of each sector's professionals, and carefulness was warranted when generalizing the results. Moreover, important, the interviews were performed before the COVID-19 pandemic, so the consequent increase in health's DT was not considered, meaning that the results might not mirror the post-pandemic reality. Another limitation assumed by the authors is the fact that the results achieved and reported here came from qualitative research only, and there was no data to quantify the results. As such, the authors would like, in the future, to re-evaluate the Portuguese HISs, checking the impact of the pandemic on the users' health literacy and HISs' development, using a more comprehensive method of data collection, with emphasis on quantitative approaches to data collection and analysis. In addition, it would be interesting to understand if the coronavirus outbreak forced a greater articulation of HISs between the public and private sectors. Finally, it should be noted that considering the utmost importance of issues related to data protection, privacy, and security of health data, it would be interesting to extend this study to evaluate questions related to these themes. Author Contributions Conceptualization, L.T., I.C., J.O.e.S. and F.M.; methodology, L.T., I.C., J.O.e.S. and F.M.; formal analysis, L.T., I.C., J.O.e.S. and F.M.; investigation, L.T., I.C., J.O.e.S. and F.M.; resources, L.T., I.C., J.O.e.S. and F.M.; data curation, L.T., I.C., J.O.e.S. and F.M.; writing--original draft preparation, L.T., I.C., J.O.e.S. and F.M.; writing--review and editing, L.T., I.C., J.O.e.S. and F.M. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement Data sharing not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Evolution of HIS in health according to technological "eras". Figure 2 Research Design. healthcare-11-00712-t001_Table 1 Table 1 Characterization of the sample profile of respondents. Interv. Region Organization Regime Profile (Position) Profession Age E1-U Azores Praia and Vitoria Health Centre Public User Computer Technician -- E2-P Azores Praia and Vitoria Health Centre Public Professional Doctor 41-51 E3-M Azores Praia and Vitoria Health Centre Public Professional (Clinical Director) Doctor <40 E4-P Azores Praia and Vitoria Health Centre Public Professional Physician <40 E5-M Algarve Family Health Unit Sol Nascente Public Professional (Hospital Coordinator) Doctor >51 E6-P Algarve Lusiadas Hospital Private Professional Doctor >51 E7-M Aveiro Finess Medical Clinic Private Professional (Clinical Director) Doctor >51 E8-P Aveiro Finess Medical Clinic Private Professional Nurse <40 E9-P Aveiro Tamega e Sousa Hospital Centre Public Professional Doctor <40 E10-M Aveiro Ovar Hospital Public Professional (Chairman of the Board of Directors) Manager 41-51 E11-M Lisbon Cascais Hospital PPP Professional (Chief Medical Information Officer) Doctor 41-51 E12-M Lisbon Cascais Hospital PPP Professional (Chief Nursing Information Officer) Nurse 41-51 E13-GE Lisbon Shared Services of the Ministry of Health Public Government Entity (President) Manager/Doctor 41-51 E14-U Faro Retired Public User Nurse >51 healthcare-11-00712-t002_Table 2 Table 2 Health Information Ecosystem in Portugal. Categories of Ecosystem of His (ESIS) Designation Description Transversal of all NHS (National health service) RNU; SER; SNS24; SINAVE These HIS have the role of centralizing and distribution of information for NHS users Life Cycle--Birth Birth News; to Birth Citizen, Health Child's Bulletin and Youth's Health Bulletin and, eBulletin of Vaccines. To receive a new citizen in society and, to monitor him/she in terms of surveillance and/or monitoring of public health Life Cycle--Health and Wellness Daily of My Health, SISO, SIIMA; SClinic CSP, RENTEV and others This cycle comprises the systems that accompany citizens in a perspective of prevention and promotion of health Life Cycle--Acute or Chronic Disease ICC, Sclinic Hospital, SClinic CSP; RCU2, SINAVE, PEM, SI VIDA; RNCCI; and, CNTS These HIS serve to accompany the user in his/her disease process, allowing the recording, diagnosis, and treatments, in all clinical episodes Life Cycle--Aging RECM; RNCCI. HIS is intended to support clinical practice in the adoption and maintenance of healthy life models by the elderly Life Cycle--Death SICO The main objective is dematerializing the process of certification of deaths and better articulation between the entities involved in the process. healthcare-11-00712-t003_Table 3 Table 3 Scenarios for Medical Practices. Medical Practices Pessimist Realist Optimistic Precision Medicine/Individualized The costs (financial and adaptation) are enormous, and for this reason, it will not be the usual practice. It will be used to solve serious and critical diseases, where the cost/benefit justifies it. Medicine will be fully focused on the citizen, with better accuracy in the personalized diagnoses and treatments. Preventive medicine Those responsible for the health area still have difficulties adapting to a reality focused on prevention. Health officials will try to make health digital, with a focus on health rather than a disease, optimizing the entire HS. The practice of medicine is focused on prevention and health promotion. Point-of-Care (Telemedicine) It is already a current practice when distance obliges. One should bet on its development. Telemedicine will be used regularly, regardless of distance, and more focused on solving the problems of the citizen. Telemedicine will be used frequently, facilitating the sharing of information between professionals for cases of complex diagnosis, and the citizen will have privileged consultations with healthcare professionals through Telemedicine and Telehealth. Assisted Medical Practices Healthcare professionals will have digital assistants who will help make diagnoses, but the presence of the health professional will be required. Healthcare professionals, in some diagnoses, will be replaced by machines. The use of machines (robots) to help some medical practices (e.g., surgeries) will be more common. The diagnoses will be made by machines, and these machines (robots) will replace healthcare professionals in clinical practices, such as surgeries. healthcare-11-00712-t004_Table 4 Table 4 Scenarios for Technologies. Technologies Pessimist Realist Optimistic Interoperability (integration) Health organizations (public and private), due to the existence of legacy systems or heterogeneous HISs, do not allow interoperability of the systems. Thus, the sharing of a citizen's health data between several entities will be a distant reality. Health organizations (public and private) collaborate in defining a set of shared services that allows the integration and access of a citizen's EHRs. HIS providers adopt international, European, and national recommendations, enabling interoperability between existing HISs and facilitating the sharing of EHR between different health organizations, respecting existing (legislation) standards. Digital Health Transformation It will occur when health organizations/entities (public and private) can change/innovate their processes, improve their leadership, and reduce resistance. There are health organizations/entities (public and private) that innovate their processes, achieving significant efficiency gains. These cases will be examples to follow by other entities. Health organizations/entities (public and private) present advanced dematerialization with significant gains in process performance. Success stories are shared and replicated. Technology to Assist Medical Practices Gradually technology that incorporates intelligence will be applied in the HIS; there is a need to create legitimacy for this to happen. Health professionals will resist but will eventually adopt the technologies. The technology is currently able to assist professionals in medical practices. However, there are still obstacles to overcome: legitimation (legislation) and acceptance by all involved (health professionals and citizens) of the existed possibilities and limitations. Health organizations and professionals perceive the positive side of incorporating intelligent technology and force legitimation (legislation) to occur. Intelligent technology, being incorporated into all medical processes and practices, leads to a huge efficiency gain and cost reduction. Use of wearables There are more and more devices able to collect data on citizens' health. However, these data will not be used without regulation to process it. On the other hand, the existent healthcare services do not have the capacity to treat such data. The collection of device-generated data is already a reality, and it does not raise technical issues; it is a matter of work, standards, and interoperability. The legitimacy of these systems and devices will occur, and healthcare models will adapt to this reality. In the short term, legislation will be created to enable the collection and use of health data from electronic devices. Clinical practices will already use these data to promote models of healthy living for citizens. healthcare-11-00712-t005_Table 5 Table 5 Scenarios for Challenges and Risks. Challenges and Risks Pessimist Realist Optimistic Changing the existing culture (resistance to change among users and health professionals) It will only occur when all stakeholders can understand the benefits to be obtained and realize that they will have to change/innovate their processes, and this will be a time-consuming process. Health and care processes need to be innovated. There is little research and literature in this area. It is necessary to study the way care is organized and identify advantages and benefits causing changes in culture. Technology is a means and not the solution. Healthcare delivery models will be studied and changed by accommodating emerging technologies with a positive and high impact on citizens' health. Info Exclusion (Training) There is a need to simplify and disseminate the HIS, mainly those that are in place and those that will appear in the future, and the advantage of their use (to health professionals and citizens). It will be necessary to reduce digital illiteracy, mainly among older people. Copying good practices successfully implemented by some organizations (training, monitoring, and involving all stakeholders), showing the advantages/benefits of using it. All entities realize the advantages/benefits of using technological solutions and increasingly seek technological solutions to solve their problems. Information (privacy, quality, security) Legislation is needed to regulate the collection, access, treatment, and security of health information. This will be one of the biggest challenges of the next years. The question of legitimacy (legislation) will be resolved quickly (by national or European directives). The next step will be to ensure the quality and security of this information so that all stakeholders maintain confidence in it. The collection, access, and sharing of health data will already be sufficiently regulated to maintain high standards of data security and privacy. All stakeholders (within their legitimacy) can add and share health information with confidence with other entities. healthcare-11-00712-t006_Table 6 Table 6 Important factors in HIS. Key Factors to Address Number of Interviewers Who Referenced It Interoperability 4 (29%) Standardization 3 (21%) Privacy 6 (43%) Security 7 (50%) Actions to address info exclusion 12 (86%) Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Gokalp E. Kayabay K. Gokalp M.O. Leveraging Digital Transformation Technologies to Tackle COVID-19: Proposing a Privacy-First Holistic Framework Emerging Technologies During the Era of COVID-19 Pandemic Arpaci I. Al-Emran M. Al-Sharafi M.A. Marques G. Studies in Systems, Decision and Control Springer Cham, Switzerland 2021 Volume 348 149 166 10.1007/978-3-030-67716-9_10 2. Loeza-Mejia C.-I. Sanchez-DelaCruz E. Pozos-Parra P. Landero-Hernandez L.-A. The potential and challenges of Health 4.0 to face COVID-19 pandemic: A rapid review Health Technol. 2021 11 1321 1330 10.1007/s12553-021-00598-8 3. Nomura S. Siesjo V. Tomson G. Mohr W. Fukuchi E. Shibuya K. Tangcharoensathien V. Miyata H. Contributions of information and communications technology to future health systems and Universal Health Coverage: Application of Japan's experiences Health Res. Policy Syst. 2020 18 73 10.1186/s12961-020-00585-x 32586326 4. WHO Global Observatory for eHealth, World Health Organization 2019 Available online: (accessed on 22 March 2019) 5. El-Sherif D.M. Abouzid M. Elzarif M.T. Ahmed A.A. Albakri A. Alshehri M.M. Telehealth and Artificial Intelligence Insights into Healthcare during the COVID-19 Pandemic Healthcare 2022 10 385 10.3390/healthcare10020385 35206998 6. Epizitone A. Moyane S.P. Agbehadji I.E. Health Information System and Health Care Applications Performance in the Healthcare Arena: A Bibliometric Analysis Healthcare 2022 10 2273 10.3390/healthcare10112273 36421597 7. Massaro M. Digital transformation in the healthcare sector through blockchain technology. Insights from academic research and business developments Technovation 2023 120 102386 10.1016/j.technovation.2021.102386 8. Haux R. Health Information Systems--From Present to Future? The German Medical Informatics Initiative Methods Inf. Med. 2018 57 e43 e45 10.3414/ME18-03-0004 30016816 9. Can O. Sezer E. Bursa O. Unalir M.O. Comparing Relational and Ontological Triple Stores in Healthcare Domain Entropy 2017 19 30 10.3390/e19010030 10. Haarbrandt B. Schreiweis B. Rey S. Sax U. Scheithauer S. Rienhoff O. Knaup-Gregori P. Bavendiek U. Dieterich C. Brors B. HiGHmed--An Open Platform Approach to Enhance Care and Research across Institutional Boundaries Methods Inf. Med. 2018 57 e66 e81 10.3414/ME18-02-0002 30016813 11. Al-Jaroodi J. Mohamed N. Abukhousa E. Health 4.0: On the Way to Realizing the Healthcare of the Future IEEE Access 2020 8 211189 211210 10.1109/ACCESS.2020.3038858 34976565 12. Marinelli S. Basile G. Zaami S. Telemedicine, Telepsychiatry and COVID-19 Pandemic: Future Prospects for Global Health Healthcare 2022 10 2085 10.3390/healthcare10102085 36292533 13. Haddara M. Staaby A. RFID Applications and Adoptions in Healthcare: A Review on Patient Safety Procedia Comput. Sci. 2018 138 80 88 10.1016/j.procs.2018.10.012 32288898 14. Prasser F. Kohlbacher O. Mansmann U. Bauer B. Kuhn K.A. Data Integration for Future Medicine (DIFUTURE) Methods Inf. Med. 2018 57 e57 e65 10.3414/ME17-02-0022 30016812 15. Cavallone M. Palumbo R. Debunking the myth of industry 4.0 in health care: Insights from a systematic literature review TQM J. 2020 32 849 868 10.1108/TQM-10-2019-0245 16. Nayak S. Patgiri R. A Vision on Intelligent Medical Service for Emergency on 5G and 6G Communication Era EAI Endorsed Trans. Internet Things 2020 6 10.4108/eai.17-8-2020.166293 17. Nayak S. Patgiri R. 6G Communication Technology: A Vision on Intelligent Healthcare Health Informatics: A Computational Perspective in Healthcare Patgiri R. Biswas A. Roy P. Studies in Computational Intelligence Springer Singapore 2021 Volume 932 1 18 10.1007/978-981-15-9735-0_1 18. Gehring S. Eulenfeld R. German Medical Informatics Initiative: Unlocking Data for Research and Health Care Methods Inf. Med. 2018 57 e46 e49 10.3414/ME18-13-0001 30016817 19. Tummers J. Tekinerdogan B. Tobi H. Catal C. Schalk B. Obstacles and features of health information systems: A systematic literature review Comput. Biol. Med. 2021 137 104785 10.1016/j.compbiomed.2021.104785 34482198 20. Khubone T. Tlou B. Mashamba-Thompson T.P. Electronic Health Information Systems to Improve Disease Diagnosis and Management at Point-of-Care in Low and Middle Income Countries: A Narrative Review Diagnostics 2020 10 327 10.3390/diagnostics10050327 32443856 21. Tabaeeian R.A. Hajrahimi B. Khoshfetrat A. A systematic review of telemedicine systems use barriers: Primary health care providers' perspective J. Sci. Technol. Policy Manag. 2022 10.1108/JSTPM-07-2021-0106 22. SPMS ENESIS-2020 2017 Available online: (accessed on 29 December 2019) 23. Marutha N.S. Landscaping health-care system using functional records management activities Collect. Curation 2020 40 9 14 10.1108/CC-03-2020-0006 24. Hawthorne K.H. Richards L. Personal health records: A new type of electronic medical record Rec. Manag. J. 2017 27 286 301 10.1108/RMJ-08-2016-0020 25. Carolina A. Monteiro B. Franca R.P. Estrela V.V. Iano Y. Khelassi A. Razmjooy N. Health 4.0: Applications, Management Technologies and Review Med. Technol. J. 2019 2 262 276 10.26415/2572-004X-vol2iss1p262-276 26. Alcacer V. Cruz-Machado V. Scanning the Industry 4.0: A Literature Review on Technologies for Manufacturing Systems Eng. Sci. Technol. Int. J. 2019 22 899 919 10.1016/j.jestch.2019.01.006 27. Karatas M. Eriskin L. Deveci M. Pamucar D. Garg H. Big Data for Healthcare Industry 4.0: Applications, challenges and future perspectives Expert Syst. Appl. 2022 200 116912 10.1016/j.eswa.2022.116912 28. Javaid M. Haleem A. Industry 4.0 applications in medical field: A brief review Curr. Med. Res. Pract. 2019 9 102 109 10.1016/j.cmrp.2019.04.001 29. Ciasullo M.V. Orciuoli F. Douglas A. Palumbo R. Putting Health 4.0 at the service of Society 5.0: Exploratory insights from a pilot study Socio-Econ. Plan. Sci. 2022 80 101163 10.1016/j.seps.2021.101163 30. Caceres C. Rosario J.M. Amaya D. Towards Health 4.0: e-Hospital Proposal Based Industry 4.0 and Artificial Intelligence Concepts Artificial Intelligence in Medicine Riano D. Wilk S. ten Teije A. Lecture Notes in Computer Science Springer International Publishing Cham, Switzerland 2019 Volume 11526 10.1007/978-3-030-21642-9_12 31. Liu Z. Ren L. Xiao C. Zhang K. Demian P. Virtual Reality Aided Therapy towards Health 4.0: A Two-Decade Bibliometric Analysis Int. J. Environ. Res. Public Health 2022 19 1525 10.3390/ijerph19031525 35162546 32. Wehde M. Healthcare 4.0 IEEE Eng. Manag. Rev. 2019 47 24 28 10.1109/EMR.2019.2930702 33. Schnurr H.-P. Aronsky D. Wenke D. MEDICINE 4.0--Interplay of Intelligent Systems and Medical Experts Knowledge Management in Digital Change North K. Maier R. Haas O. Springer Cham, Switzerland 2018 51 63 10.1007/978-3-319-73546-7_3 34. Chute C. French T. Introducing Care 4.0: An Integrated Care Paradigm Built on Industry 4.0 Capabilities Int. J. Environ. Res. Public Health 2019 16 2247 10.3390/ijerph16122247 31242687 35. Afferni P. Merone M. Soda P. Hospital 4.0 and Its Innovation in Methodologies and Technologies Proceedings of the 2018 IEEE 31st International Symposium on Computer-Based Medical Systems (CBMS) Karlstad, Sweden 18-21 June 2018 333 338 10.1109/CBMS.2018.00065 36. Unterhofer M. Rauch E. Matt D.T. Hospital 4.0 roadmap: An agile implementation guideline for hospital manager Int. J. Agil. Syst. Manag. 2021 14 635 10.1504/IJASM.2021.120230 37. Feussner H. Ostler D. Kranzfelde M. Kohn N. Koller S. Wilhelm D. Thuemmler C. Schneider A. Surgery 4.0 Health 4.0: How Virtualization and Big Data Are Revolutionizing Healthcare Thuemmler C. Bai C. Springer International Publishing Cham, Switzerland 2017 91 107 10.1007/978-3-319-47617-9_5 38. Kickbusch I. Health promotion 4.0 Health Promot. Int. 2019 34 179 181 10.1093/heapro/daz022 30939200 39. Hou Y. Wang W. Bartolo P. Application of additively manufactured 3D scaffolds for bone cancer treatment: A review Bio-Des. Manuf. 2022 in press 10.1007/s42242-022-00182-7 40. Siuly S. Aickelin U. Kabir E. Huang Z. Zhang Y. Guest Editorial: Special issue on "Artificial Intelligence in Health Informatics" Health Inf. Sci. Syst. 2021 9 23 10.1007/s13755-021-00155-7 34194729 41. Lohr S. What Ever Happened to IBM's Watson? The New York Times 17 July 2021 Available online: (accessed on 18 February 2023) 42. Zou F.W. Tang Y.F. Liu C.Y. Ma J.A. Hu C.H. Concordance Study between IBM Watson for On-cology and Real Clinical Practice for Cervical Cancer Patients in China: A Retrospective Analysis Front. Genet. 2020 11 200 10.3389/fgene.2020.00200 32265980 43. Wong A.N.N. He Z. Leung K.L. To C.C.K. Wong C.Y. Wong S.C.C. Yoo J.S. Chan C.K.R. Chan A.Z. Lacambra M.D. Current Developments of Artificial Intelligence in Digital Pathology and Its Future Clinical Applications in Gastrointestinal Cancers Cancers 2022 14 3780 10.3390/cancers14153780 35954443 44. Koh D.M. Papanikolaou N. Bick U. Illing R. Kahn C.E. Kalpathi-Cramer J. Matos C. Marti-Bonmati L. Miles A. Mun S.K. Artificial intelligence and machine learning in cancer imaging Commun. Med. 2022 2 133 10.1038/s43856-022-00199-0 36310650 45. Oren O. Gersh B.J. Bhatt D.L. Artificial intelligence in medical imaging: Switching from radiographic pathological data to clinically meaningful endpoints Lancet Digit. Health 2020 2 e486 e488 10.1016/S2589-7500(20)30160-6 33328116 46. Kung T.H. Cheatham M. Medenilla A. Sillos C. De Leon L. Elepano C. Madriaga M. Aggabao R. Diaz-Candido G. Maningo J. Performance of ChatGPT on USMLE: Potential for AI-Assisted Medical Education Using Large Language Models medRxiv 2022 10.1371/journal.pdig.0000198 47. Clayton B.J. Microsoft Unveils New Bing with ChatGPT Powers BBC News 7 February 2023 Available online: (accessed on 18 February 2023) 48. Eurostat One in Two EU Citizens Look for Health Information Online. 6 April 2021 Available online: (accessed on 18 February 2023) 49. Sam Altman on. (11 December 2022). Twitter Available online: (accessed on 18 February 2023) 50. SPMS ENESIS-20/22 2019 Available online: (accessed on 29 December 2022) 51. Krippendorf K. Content Analysis: An Introduction to Its Methodology 4th ed. Sage Publication Thousand Oaks, CA, USA 2018 52. PCM Resolucao do Conselho de Ministros 62/2016, 2016-10-17--DRE 2016 Available online: (accessed on 18 February 2023) 53. PCM Resolucao do Conselho de Ministros 108/2017, 2017-07-26--DRE 2017 Available online: (accessed on 18 February 2023) 54. SPMS Sobre os Sistemas de Informacao 2023 Available online: (accessed on 18 February 2023) 55. Rahi S. Khan M.M. Alghizzawi M. Factors influencing the adoption of telemedicine health services during COVID-19 pandemic crisis: An integrative research model Enterp. Inf. Syst. 2021 15 769 793 10.1080/17517575.2020.1850872 |
PMC10000614 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050955 diagnostics-13-00955 Interesting Images A Familial Case of Multiple Endocrine Neoplasia 2A: From Morphology to Genetic Alterations Penetration in Three Generations of a Family Chen Lan 1* Zhang Jing-Xin 1 Liu Dong-Ge 1 Liu Hong-Gang 2* Tar-Choon Aw Academic Editor 1 Pathology Department, Beijing Hospital, National Center of Gerontology, Institute of Geriatric Medicine, Chinese Academy of Medical Sciences, Beijing 100730, China 2 Pathology Department, Beijing Tongren Hospital, Capital Medical University, Beijing Key Laboratory of Head and Neck Molecular Diagnostic Pathology, Beijing 100730, China * Correspondence: [email protected] (L.C.); [email protected] (H.-G.L.); Tel.: +86-10-85133913 (L.C.); +86-18811612227 (H.-G.L.) 02 3 2023 3 2023 13 5 95509 1 2023 21 2 2023 22 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). This paper illustrates a rare syndrome of multiple endocrine neoplasia type 2A (MEN2A) in a family of three generations. In our case, the father, son and one daughter developed phaeochromocytoma (PHEO) and medullary thyroid carcinoma (MTC) over a period of 35 years. Because of the metachronous onset of the disease and lack of digital medical records in the past, the syndrome was not found until a recent fine needle aspiration of an MTC-metastasized lymph node from the son. All resected tumors from the family members were then reviewed and supplemented with immunohistochemical studies, previously wrong diagnoses were then corrected. Further molecular study of targeted sequencing also revealed a RET germline mutation (C634G) in the family tree including the three members with onset of the disease and one granddaughter who had no disease at the time of testing. Despite the syndrome being well-known, it may still be misdiagnosed because of its rarity and long disease onset. A few lessons can be learned from this unique case. Successful diagnosis requires high suspicion and surveillance and a tri-level methodology including a careful review of family history, pathology and genetic counselling. medullary thyroid carcinoma pheochromocytoma multiple endocrine neoplasia type 2A RET germline mutation Beijing Key Laboratory of Head and Neck Molecular Diagnostic Pathology, Capital Medical UniversityOpen Grant 2016-2017 This research was funded by the Beijing Key Laboratory of Head and Neck Molecular Diagnostic Pathology Open Grant, Capital Medical University, grant number Beijing Key Laboratory of Head and Neck Molecular Diagnostic Pathology Open Grant 2016-2017. pmcFigure 1 The cytology of a metastasized medullary thyroid carcinoma (MTC) in the neck lymph node and its primary MTC in the son born in 1968. A 49-year-old male in the outpatient clinic had a right lymphadenopathy, 1.2 cm in diameter, in area III. He had ultrasound-guided fine needle aspiration. The smear was composed of a large amount of monomorphic, small, round cells in large aggregates or dispersed ((a), HE, 100x). The cells were plasmacytoid with eccentric nuclei, which had fine chromatin or small nucleoli. There was no mitosis or necrosis ((b), HE, 400x). Immunohistochemistry on smears showed strong immunopositivity for chromogranin A ((c), 400x), TTF-1 and CEA (not shown). There was moderate positivity for calcitonin ((d), 400x) and no expression of thyroglobulin ((e), 400x). It was consistent with metastasized thyroid medullary carcinoma. He was diagnosed with bilateral multiple MTC with right neck lymph node metastasis in 1/2 in our hospital 8 years ago when he was 41. Tumors infiltrated thyroid tissues in a solid and trabecular growth pattern and there was amyloid deposition in the background ((f), 50x). It was composed of small monomorphic cells with fine chromatin or small nucleoli in a syncytial pattern ((g), 200x). Immunohistochemistry showed a strong expression of chromogranin A ((h), 100x) and calcitonin ((i), 100x) and absence of thyroglobulin ((j), 25x). The cell morphology of his recent lymph node aspiration was consistent with his previous pathology. The patient was then admitted into the ward for further treatment. A family syndrome of multiple endocrine neoplasia type 2A (MEN2A) surfaced when we dug into his family history and previous pathology. His father and one of his elder sisters used to have intermittent hypertension, cardiac palpitation, dizziness and epileptic seizure, etc. He was 31 when he first suffered similar symptoms and was then hospitalized and diagnosed with left adrenal pheochromocytoma (PHEO) and right thyroid adenoma. We suspected the thyroid adenoma might have been misdiagnosed as application of immunohistochemistry was in its early stage in the late 1990s during his first onset of the disease. His pathology materials were not kept due to the long time period in a different local hospital where he received surgical resections. We proceeded to review the pathology of his father and sister who had surgeries in our hospital. Figure 2 The pathology of thyroid and adrenal gland tumor of his elder sister born in 1964. One of his elder sisters, who was 4 years older than him, was diagnosed with left adrenal PHEO and right thyroid follicular carcinoma at the age of 28 and left thyroid MTC at 45. As per her previous pathology, her PHEO diagnosed at 28 ((a), HE, 200x) was composed of zellballen-like nests of tumor cells separated by vascular sinuses. The tumor cells were spindle-like with amphophilic granular cytoplasm. The small and monomorphic nuclei were vacuolar with small nucleoli. Immunohistochemistry showed that tumor cells were positive for chromogranin A ((b), 50x). The small spindle-like sustentacular cells were outlined by S100 positivity at the periphery of tumor nests ((c), 200x). The proliferative index was less than 1% as indicated by Ki67 ((d), 100x). Her left MTC diagnosed at 45 ((e), HE, 200x) was composed of trabecular solid growth patterns of spindle or polygonal tumor cells. The tumor cells had round nuclei that varied in size with amyloid substance deposition in the background. They were strongly positive for chromogranin A ((f), 100x) and calcitonin ((g), 100x) and negative for thyroglobulin ((h), 100x). When we reviewed her right thyroid follicular carcinoma diagnosed at 28 ((i), 200x), it was also composed of a similar pathology as her left MTC showing nests of syncytial cell aggregates with monomorphic, round nuclei. There was also nuclei variation in size ((j), 200x). Supplementary immunohistochemistry demonstrated immunopositivity of chromogranin A and calcitonin (not shown). The thyroid follicular carcinoma diagnosis was wrong and it should be MTC instead. Therefore, the elder sister had both MTC and PHEO in a time span of 17 years. The misdiagnosis was not recognized earlier as her first medical record in early 1990s was not digitalized and was disconnected with her second record until we manually discovered it this time. The syndrome was missed then due to the metachronous onset of the disease that was misdiagnosed as thyroid adenoma or follicular carcinoma in the brother and sister. MTC can mimic any thyroid tumor and requires immunohistochemistry for accurate diagnosis; however, immunohistochemistry back then was not comprehensive. Figure 3 The pathology of left MTC and left adrenal PHEO in the father diagnosed in our hospital. Born in 1940, the father had surgeries for PHEO and MTC at 42. His previous medical file and pathology were searched manually, reviewed carefully and supplemented with immunohistochemistry. The left thyroid tumor was 4 x 3.5 x 2.5 cm in size and composed of a trabecular growth pattern of spindle-like cells with monomorphic, round or spindle nuclei ((a), HE, 200x). They were strongly positive for calcitonin ((b), 200x) and CEA ((c), 200x), and weakly positive for chromogranin A ((d), 200x) and the diagnosis was consistent with MTC. His PHEO was composed of nested polygonal cells with abundant amphophilic cytoplasm separated by rich vascular sinuses ((e), HE, 100x). They were strongly positive for chromogranin A ((f), 100x) and sparsely positive for Ki67 ((g), 200x). The sustentacular cells at the periphery of tumor cells were strongly positive for S100 ((h), 200x). The syndrome was missed because our knowledge for this rare syndrome was limited in early 1980s. His children had not had onset of the disease back then. Figure 4 RET germline mutation detected in the family of three generations. In 1968, Steiner first reported multiple endocrine neoplasia type 2. The syndrome is classified into MEN2A, 2B and familial MTC depending on the presence and type of additional symptoms. Virtually all MEN2A patients have MTC and 50% of them have concurrence of PHEO, and 15-30% have hyperparathyroidism. It accounts the most common variation of MEN2 with an incidence of 1/1,973,500 . The major molecular alterations of RET in MEN2A lie in codons 609, 611, 618 or 620 of exon 10 or codon 634 of exon 11. The latter accounts for 80% of germline mutations of MEN2A . For our cases, 4 mL blood was taken for each person from the three generations of the family as shown in the family tree. It included the parents, their four children and two grandchildren outlined in red. DNA was purified using a QIAmp Blood DNA kit (Qiagen, Germantown, MD, USA). Targeted next-generation sequencing was successfully performed for the RET proto-oncogene by Gene Plus, Beijing, China, according to manufacturer's protocol. A missense mutation c.T1900G in exon 11 causing protein alteration of p.C634G was discovered in all three members with onset of the disease and one granddaughter who had no disease at the time of testing. The family members with the mutation are labeled as RET in the figure, while those with onset of MEN2A are labeled in blue. The granddaughter with the RET mutation, was 18 years old and is the daughter of the son. A close monitoring was undertaken for the girl according to the American Thyroid Association guidelines for adults, that requires surveillance of serum calcitonin . The family tragedy might be stopped thereafter. MEN2-associated RET mutations have a gain of function effect to promote activation of the kinase and oncogenic conversion via downstream signaling pathways . RET tyrosine kinase inhibition will provide targeted therapy for the family members with established RET-positive cancers . Selpercatinib, one kinase inhibitor, has been approved by the U.S. Food and Drug Administration for patients with advanced or metastatic RET-mutant MTC . It inhibits wild-type RET and multiple mutated RET isoforms and may benefit the son who had repeated recurrence of the cancer. In conclusion, misdiagnosis of thyroid tumor and metachronous onset of the disease in the family hampered our recognition of this well-documented but rare disease. A few lessons can be learned from this case. High suspicion was a critical starting point in our diagnosis. A careful review of previous medical records helped connect the dots. Molecular screening then played a key role in reaching our eventual diagnosis. Acknowledgments The authors thank the technicians in the Pathology Department of Beijing Hospital for their technical support. Author Contributions Conceptualization, L.C.; methodology, J.-X.Z.; validation, D.-G.L.; formal analysis, L.C.; investigation, L.C.; resources, J.-X.Z.; data curation, J.-X.Z.; writing--original draft preparation, L.C. and J.-X.Z.; writing--review and editing, L.C. and D.-G.L.; visualization, J.-X.Z.; supervision, H.-G.L.; project administration, H.-G.L.; funding acquisition, L.C. and H.-G.L. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Ethical review and approval were waived for this study because the study was under a normal procedure to perform clinical pathology, molecular diagnosis and genetic counseling to the patients. Genetic diagnosis was paid for by a government grant due to rarity of the disease and the poor economic conditions of the patients. Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient(s) to publish this paper. Data Availability Statement The authors declare that all datasets on which the conclusions of the paper rely are available to editors and reviewers without unnecessary restriction. The datasets can be obtained from the corresponding author Lan Chen on reasonable request. Conflicts of Interest The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Wells Samuel A. Jr. AsaSylvia L. EvansDouglas B. GagelRobert F. MoleyJeffrey F. Sara R. WaguespackSteven G. Revised American Thyroid Association guidelines for the management of medullary thyroid carcinoma Thyroid 2015 25 567 610 10.1089/thy.2014.0335 25810047 2. Marx S.J. Simonds W.F. Hereditary hormone excess: Genes, molecular pathways, and syndromes Endocr. Rev. 2005 26 615 661 10.1210/er.2003-0037 15632315 3. Runeberg-Roos P. Saarma M. Neurotrophic factor receptor RET: Structure, cell biology, and inherited diseases Ann. Med. 2007 39 572 580 10.1080/07853890701646256 17934909 4. Santoro M. Melillo R.M. Carlomagno F. Vecchio G. Fusco A. Minireview: RET: Normal and abnormal functions Endocrinology 2004 145 5448 5451 10.1210/en.2004-0922 15331579 5. Bradford D. Larkins E. Mushti S.L. Rodriguez L. Skinner A.M. Helms W.S. Price L.S.L. Zirkelbach J.F. Li Y. Liu J. FDA Approval Summary: Selpercatinib for the Treatment of Lung and Thyroid Cancers with RET Gene Mutations or Fusions Clin. Cancer Res. 2021 27 2130 2135 10.1158/1078-0432.CCR-20-3558 33239432 |
PMC10000615 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050812 cells-12-00812 Communication Phosphorylation of LKB1 by PDK1 Inhibits Cell Proliferation and Organ Growth by Decreased Activation of AMPK Borkowsky Sarah 1 Gass Maximilian 1 Alavizargar Azadeh 2 Hanewinkel Johannes 1 Hallstein Ina 1 Nedvetsky Pavel 1 Heuer Andreas 2 Krahn Michael P. 1* Momeny Majid Academic Editor Babu Vishnu Suresh Academic Editor Majumder Avisek Academic Editor 1 Medical Cell Biology, Medical Clinic D, University Hospital of Munster, Albert-Schweitzer Campus 1-A14, 48149 Munster, Germany 2 Institute of Physical Chemistry, University of Munster, Corrensstr. 28/30, 48149 Munster, Germany * Correspondence: [email protected]; Tel.: +49-251-8357052 06 3 2023 3 2023 12 5 81219 8 2022 24 2 2023 03 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). The master kinase LKB1 is a key regulator of se veral cellular processes, including cell proliferation, cell polarity and cellular metabolism. It phosphorylates and activates several downstream kinases, including AMP-dependent kinase, AMPK. Activation of AMPK by low energy supply and phosphorylation of LKB1 results in an inhibition of mTOR, thus decreasing energy-consuming processes, in particular translation and, thus, cell growth. LKB1 itself is a constitutively active kinase, which is regulated by posttranslational modifications and direct binding to phospholipids of the plasma membrane. Here, we report that LKB1 binds to Phosphoinositide-dependent kinase (PDK1) by a conserved binding motif. Furthermore, a PDK1-consensus motif is located within the kinase domain of LKB1 and LKB1 gets phosphorylated by PDK1 in vitro. In Drosophila, knockin of phosphorylation-deficient LKB1 results in normal survival of the flies, but an increased activation of LKB1, whereas a phospho-mimetic LKB1 variant displays decreased AMPK activation. As a functional consequence, cell growth as well as organism size is decreased in phosphorylation-deficient LKB1. Molecular dynamics simulations of PDK1-mediated LKB1 phosphorylation revealed changes in the ATP binding pocket, suggesting a conformational change upon phosphorylation, which in turn can alter LKB1's kinase activity. Thus, phosphorylation of LKB1 by PDK1 results in an inhibition of LKB1, decreased activation of AMPK and enhanced cell growth. LKB1 PDK1 AMPK mTOR cell proliferation the German Research Foundation (DFG)CRC1348-A01 CRC1348-A05 the Interdisciplinary Center of Clinical Research (IKF) MunsterKr-A-031.21 This research was funded by the German Research Foundation (DFG) grant number CRC1348-A01 and CRC1348-A05 and the Interdisciplinary Center of Clinical Research (IKF) Munster grant number Kr-A-031.21. pmc1. Introduction The serine-threonine kinase LKB1 was originally identified in Caenorhabditis elegans as the "Partitioning defective protein" 4 (Par4) and described to be essential for asymmetric division in C. elegans zygotes . Similar, Drosophila LKB1 (DmLKB1) determines anterior-posterior polarity of the oocyte and apical-basal polarity in epithelial cells of the follicular epithelium and in the compound eye . In Drosophila neural stem cells (neuroblasts, NBs), DmLKB1 regulates spindle formation and asymmetric cell division and has been identified as an upstream regulator of the Hippo pathway effector Yorkie . LKB1 is a master kinase, activating several downstream kinases, in particular of the AMPK (AMP-dependent kinase)-family, e.g., AMPK, MARKs, SAD-Kinases and NUAKs . Furthermore, LKB1 has been shown to activate the tumor suppressors PTEN and p53 . Thereby, LKB1 regulates various cellular processes, including cell polarity, cell migration, cell cycle control, apoptosis and energy metabolism (reviewed by ). Mutations in STK11, the gene encoding human LKB1 (hLKB1) are the cause of the Peutz-Jeghers Syndrome (PJS), a rare autosomal dominant disease, which manifests in intestinal benign polyps and mucocutaneous mispigmentations/lentigines . Patients suffering from PJS exhibit a strongly increased risk to develop intestinal and extraintestinal cancer. Moreover, LKB1 has been demonstrated to be mutated or downregulated in various tumor types, in particular in non-small cell lung cancer, prostate cancer and cervix carcinoma (reviewed by ). In mice models, inactivation of LKB1 together with mutations in PTEN results in a strongly enhanced tumorigenic potential (reviewed by ). Although LKB1 has been supposed to be a constitutively active kinase, its export from the nucleus and its activity are enhanced upon binding to the pseudokinase STRADa and to the adaptor protein Mo25 . Apart from formation of this trimeric complex, several posttranslational modifications have been found to regulate the activity of the LKB1 complex . We recently found that LKB1 binds directly to phosphatidic acid (PA) via a C-terminal polybasic motif. This motif is essential for stable membrane recruitment of LKB1 in cultured cells and in vivo. Furthermore, binding to PA is essential for efficient kinase activity and for the tumor suppressor function of LKB1 . Activation of protein kinases by phospholipids of the (plasma) membrane has been described for several enzymes, including mTOR, SGK3, PKCs and Raf1 . Phosphoinositide-dependent kinase 1 (PDK1) contains a pleckstrin homology (PH) domain, which binds to Phosphatidylinositol(3,4,5)-tris-phosphate (PI(3,4,5)P3), thereby releasing autoinhibition of the enzyme, resulting in an activation of PDK1 . PDK1 activates several downstream serine/threonine kinases of the AGC family , in particular Akt/PKB, which is further activated by PIP3 , controlling mTOR signaling and other important cellular signaling pathways . In this study, we now describe a new regulatory mechanism of LKB1 in Drosophila: The binding to and direct phosphorylation of a conserved motif within its kinase domain by PDK1 inhibits the activity of LKB1, resulting in decreased activation of AMPK. This was approved using molecular dynamics (MD) simulations, showing that the size of the binding pocket shrinks in the case of the phosphorylated protein. Cells expressing phospho-deficient LKB1 consequently display enhanced AMPK activation and decreased mTOR activity, resulting in reduced cell size. 2. Materials and Methods 2.1. Drosophila Stocks and Genetics Fly stocks were cultured on standard cornmeal agar food and maintained at 25 degC. Knockin of LKB1 variants was established using CRISPR/Cas9 technique. In short, a plasmid (pU6-Bbs-chiRNA) encoding the guide-RNAi (GTTATCATGAAGTGCAATCA) targeting Cas9 to the third intron of LKB1 was injected into vasa::Cas9 transgenic flies (#51323 obtained from Bloomington stock center) together with a donor plasmid containing ca. 1 kbp 5' and 3' homology arms and an eye-driven (3xP3 promoter) dsRed (pHD-dsRed) . Point mutations (T353A and T353D) were introduced by site-directed mutagenesis. MARCM (mosaic analysis with a repressible cell marker) clones were produced by crossing LKB1 FRT82B flies with hsFlp, tub::GAL4, UAS::nGFP;;FRT82B, tubP::GAL80 (obtained from Bloomington Stock Center, Bloomington, IN, USA). GFP-marked LKB1-variant-mutant clones in imaginal discs were induced by heat shock in first instar larvae. 2.2. Immunohistochemistry Imaginal discs of third instar larvae were dissected in PBS and fixed for 20 min in 4% PFA/PBS. Subsequently, discs were washed three times in PBS + 0.2% Triton X-100 and blocked with 1% BSA for 1 h, incubated over night with primary antibodies in PBS + 0.2% Triton X-100 + 1% BSA, washed three times and incubated for 2 h with secondary antibodies. After three washing steps and DAPI-staining, discs were mounted with Mowiol. Embryos were fixed and stained as described before . Primary antibodies used were as follows: anti Baz (1:500) , goat anti GFP (1:500, #600-101-215, Rockland, Pottstown, PA, USA), mouse anti Disc Large (1:100, 4F3, Developmental Studies Hybridoma Bank (DSHB), Iowa City, IA, USA) and mouse anti Histone-3 phospho-S10 (1:500, Cell Signaling #9706, Danvers, MA, USA). Secondary antibodies conjugated with Alexa 488, Alexa 568 and Alexa 647 (Life Technologies, Carlsbad, CA, USA) were used at 1:400. Images were taken on a Leica SP8 confocal microscope (Leica Microsystems, Wetzlar, Germany) using lightning program and processed using ImageJ version 1.53t. 2.3. Coimmunoprecipitation and Western Blot For coimmunoprecipitation, Schneider S2R+ cells, which were isolated from late-stage Drosophila embryos and commonly used for protein expression were cotransfected with GFP-LKB1 variants and PDK1-HA. Three days after transfection, cells were lysed and GFP-LKB1 was immunoprecipitated using GFP-binder (Chromotek, Planegg, Germany). Embryonic lysates were made of Drosophila embryos collected from an overnight plate with Laemmli buffer. SDS PAGE and Western blotting was performed according to standard procedures. The following primary antibodies were used: mouse anti ss Actin (1:1000, Santa Cruz #47778, Dallas, TX, USA), rabbit anti phospho-AMPK T172 (1:500, Cell Signaling #2535), mouse anti AMPK (1:500, Santa Cruz #sc-74461), mouse anti HA (1:500, Santa Cruz #7392), mouse anti GFP (1:500, Santa Cruz #9996), guinea pig anti LKB1 (1:500), mouse anti Myc (1:100, 9E10, DSHB) and mouse anti pT389 S6K (1:500, Cell Signaling #9206). 2.4. In Vitro Kinase Assay MBP-LKB1 and MBP-LKB1 E253A proteins were expressed in E. coli BL21 cells and purified using Amylose resin. In total, 2 mg of recombinant protein was incubated together with 1 mg recombinant PDK1 (ProQuinase #0367-0000-1, Malvern, PA, USA) and 0.3 mCi[g-32ATP] in kinase buffer (10 mM HEPES pH 7.5, 100 mM NaCl, 10 mM MgCl2, 1 mM DTT) for 1 h at 30 degC. The reaction was terminated by addition of SDS sample buffer and samples were subjected to SDS-PAGE. Phosphorylation was detected by exposure to X-ray films. 2.5. Molecular Dynamics Simulations Chain C of PDB 2WTK, corresponding to the LKB1 protein , was extracted using VMD . The missing amino acids (75-77) were modeled using Modeller version 9.16 . Then, CHARMM-GUI membrane builder was used to prepare the phosphorylated and unphosphorylated proteins, which were then amidated and acetylated and finally solvated by water molecules with the addition of neutralizing ions. The MD simulations were performed using version 2019.6 of GROMACS and the CHARMM36 force field , as well as the TIP3P model for water molecules. Periodic boundary conditions were applied in all directions. The long-range electrostatic interactions were treated using particle mesh Ewald method , with a cutoff distance of 1.2 nm and a compressibility of 4.5 x 10-5. For the van der Waals (vdW) interactions, cutoff schemes with a cutoff distance of 1.2 nm were utilized, smoothly truncated between 1.0 and 1.2 nm. Constant pressure was controlled by coupling the system to the Parrinello-Rahman barostat with an isotropic pressure of 1 bar. The temperature was controlled at 310 K by coupling the system to the Nose-Hoover thermostat . The LINCS algorithm was employed to constrain the bonds . The systems were first minimized in 10,000 steps and were subsequently equilibrated using initially the NVT (500 ps) and then the NPT (16 ns) protocol in multiple steps. During the course of equilibration, restraints (starting with 4000 kJ/mol-1.nm-2) were applied on the heavy atoms of the protein, which were then gradually decreased to zero. The production simulations for both phosphorylated and unphosphorylated proteins were performed for 3 ms using a time step of 2 fs. To judge the statistical relevance of our results, we performed two independent runs, denoted sample 1 and sample 2. The simulations data were analyzed using in-house codes, incorporating the MDAnalysis package . VMD was used to visualize the structures and trajectories as well as preparing snapshots . 2.6. Statistics Statistical significance was determined by one way ANOVA with Bonferroni multiple comparison test: **** p < 0.0001, *** p < 0.001, ** p < 0.01, * p < 0.05, n.s. is not significant. 3. Results 3.1. LKB1 Contains a Canonical PDK1-Binding and -Consensus Motif LKB1 is localized to the plasma membrane of epithelial cells and neural stem cells . In contrast to many other kinases, LKB1 is not activated by T-loop phosphorylation but was supposed to be constitutively active. However, we have shown that direct binding of LKB1 to phosphatidic acid in the plasma membrane is essential for membrane recruitment and activation of LKB1 . We now investigated whether plasma membrane-bound LKB1 gets modified by kinases, which localized to the plasma membrane, too. Checking the amino acid sequence of LKB1 for kinase consensus motifs, we identified a canonical PDK1-binding motif as well as a PDK1 consensus motif in the kinase domain of LKB1 . Notably, both motifs are highly conserved from fly to men, but not in C. elegans. Using transfected Schneider 2R+ (S2R+) cells , we verified that LKB1 co-immunoprecipitates with PDK1 and that mutation of the PDK1-binding motif (E253A) strongly decreases the interaction of the two proteins . Mutation of the phosphorylated Threonine 353 to Alanine increases binding of PDK1 , suggesting that LKB1-PDK1 binding is reinforced if phosphorylation cannot occur. Finally, in vitro kinase assay with recombinant MBP (maltose binding protein)-LKB1 demonstrates that PDK1 efficiently phosphorylates LKB1, but not LKB1 T353A . 3.2. Phosphorylation of LKB1 by PDK1 Does Not Affect Protein Localization In Vivo In order to test whether phosphorylation of LKB1 by PDK1 alters the localization of LKB1, we established phosphorylation-deficient (T353A) and phospho-mimetic (T353D) rescue constructs of GFP-LKB1 expressed from its endogenous promoter . As depicted in Figure 2A-C, both mutant variants localize to the lateral plasma membrane, colocalizing with Discs large (Dlg) and indistinguishable from wild type GFP-LKB1. Bazooka (Baz) was used as a marker for the apical cell-cell contacts . 3.3. T353 Phosphorylation Is Not Essential for Survival of the Fly, but Regulates Organism Size Knockout of LKB1 is pupal lethal and deletion of LKB1 results in strong polarity defects in various cell types . Therefore, we investigated whether PDK1-mediated phosphorylation of LKB1 is lethal and produces any (polarity) phenotypes. For that, we established wild type, phosphorylation-deficient and phosphomimetic knockins using CRISPR-Cas9 gene editing. Notably, we found no increased lethality and normal (or even better) adult flies hatching rates in both mutant variants . Moreover, apical-basal polarity in epithelial cells and neuroblasts as well as anterior-posterior polarity in oocytes is not affected in mutant knockins (data not shown). However, quantification of body size of LKB1-alleles revealed a significant reduction in phospho-deficient knockin flies, suggesting changes in the function of LKB1 which affects cell proliferation, and thus, organ/organism growth . 3.4. Modeling of T353 Phosphorylation Reveals a Narrowed ATP-Binding Pocket As no crystal structure of Drosophila LKB1 has been resolved so far, we used human LKB1 to model the impact of LKB1 phosphorylation by PDK1. The kinase domain of Drosophila and human LKB1 is well conserved (67.9% identical acids, 82.5% similarity). Although located in the kinase domain, analysis of the published crystal structure of human LKB1 suggests that the PDK1-phosphorylation site is not engaged in the catalytic center but that it is rather exposed at the surface of the LKB1/STRADa/Mo25 complex. In order to identify conformational changes upon T353 phosphorylation (T230 in human LKB1), we performed two independent atomistic simulations for the non-phosphorylated (LKB1) and phosphorylated (pLKB1) form of LKB1, with a total simulation time of 3 ms for each system. The root-mean-square deviations (RMSDs) of the backbone atoms of the protein shows that the structure of the protein remains stable throughout the simulation in both simulations . The RMSD for LKB1 and pLKB1 in the first sample are nearly constant between 0.7 and 2.3 ms. In the second sample, the RMSD is more stable for both LKB1 and pLKB1 proteins compared to the first sample. However, the LKB1 undergoes more structural changes between 1 and 1.7 ms . This is mainly due to the structural changes in the activation loop (A-loop) region, which is comparatively much more distorted compared to the first sample. Therefore, the main changes in the protein structure in the second sample is due to the changes in the A-loop region. The root-mean-square-fluctuations (RMSFs) or atomic positional fluctuations describe how flexible the individual residues are. The average RMSFs of protein Ca atoms over the two samples shows that the residues 118-125 and 203-211 (in the A-loop region) are more flexible in the LKB1 compared to pLKB1 , revealing distinct differences between the two proteins flexibility, in particular in the A-loop part. We further calculated the average structure from the trajectory between 1 and 3 ms for both LKB1 and pLKB1 systems by overlaying the structures in each frame on the crystal structure and then averaging over all frames. The overlay of the average structures for both systems is shown in Figure 3B. Quantification of the distance between the Ca atoms after optimum mapping reveals small but significant effects in amino acids 57-63 and 118-125 on the N-lobe, 203-211 on the A-loop and 223-230 and 253-273 on the C-lobe . Notably, most of these regions are not close to the T353 phosphorylation. These results suggest that there is a small but significant effect of phosphorylation propagating throughout the protein. As part of the structural changes in the N-lobe region, the distance between S60 and A195 (located on the A-loop) become smaller in pLKB1, which occurs between 0.9 and 1.9 ms of the simulation . Accordingly, the volume of the ATP-binding pocket, which was calculated using CASTp , is temporarily decreased in pLKB1 . In the second simulation, the distance between these two residues remains nearly similar . Instead, the distance between K78 and E98, which point towards the binding pocket, is smaller for pLKB1 , again narrowing the pocket . This shows that the ATP-binding pocket can shrink in different ways upon T353 phosphorylation, suggesting a decreased kinase activity of LKB1. 3.5. LKB1 T353 Phosphorylation Regulates Kinase Activity and Cell Growth In Vivo Next, we investigated whether the prediction drawn from the modeling simulations is recapitulated in vivo and accounts for the decreased body size of T353A knockin flies. Indeed, activation of AMPK (by phosphorylation of the T-loop by LKB1) is increased in lysates from T353A knockin embryos, whereas it is decreased in case of T353D . Consequently, phospho-S6K, a downstream target of mTOR is decreased in T353A, indicating that an enhanced activation of phospho-deficient LKB1 results in increased AMPK activity, which inhibits mTOR signaling. Of note, in in vitro kinase assays, LKB1 T353A and LKB1 E253A did not display an altered kinase activity towards AMPK , suggesting that the phosphorylation of LKB1 T353 by PDK1 is indeed essential for regulating the kinase activity. To further validate this finding in vivo, we generated clone mutants for T353wt, T353A and T353D, respectively, in an otherwise wild type background in wing imaginal discs using the MARCM (mosaic analysis with a repressible cell marker ) technique. Indeed, the cell size of mutant (GFP-marked) cells was reduced in case of T353A . In addition, the percentage of proliferating cells (quantified by Histone 3 phospho-S10) was lower in T353A clones (0.9%), compared to wild type (1.2%) and T353D clones (1.4%). These results support our hypothesis derived from modeling simulations that phosphorylation of LKB1 by PDK1 at T353 inhibits the kinase activity of LKB1, thus resulting in decreased activation of AMPK and its downstream target mTOR, which leads to increased cell size and proliferation. We finally tested whether impaired LKB1 T353 phosphorylation results in changes in binding of the LKB1 cofactors Stlk and Mo25, which might explain differences in the activation of LKB1. However, no differences in Stlk/Mo25 binding were detectable in co-immunoprecipitation assays , which is in line with the prediction from LKB1's crystal structure that the binding interface to its cofactors does not involve T353. 4. Discussion In this study, we describe the phosphorylation of the tumor suppressor kinase LKB1 by PDK1 as a new regulatory mechanism to control LKB1's activity. LKB1 exhibits a conserved PDK1-binding motif, which is essential for binding to PDK1 as well as a conserved PDK1-phosphorylation motif within its kinase domain. Our modeling results suggested a decrease in LKB1 kinase activity due to narrowing of the activation loop in the catalytic center upon phosphorylation by PDK1. Indeed, we confirmed that a phosphorylation-deficient variant of LKB1 exhibits an increased activation of AMPK, the major substrate of LKB1, which results in increased mTOR activation and decreased cell proliferation. Notably, C. elegans Par4, the homologue of human and Drosophila LKB1 exhibits a well conserved PDK1-binding and -phosphorylation motif but lacks the phosphorylated residue itself (T353 in Drosophila, F365 in C. elegans). This indicates a conserved regulatory mechanism, which was partly lost during evolution of nematodes. However, PDK1 has also been described to exhibit kinase-independent functions ; thus, the PDK1/LKB1 interaction might also be of importance in C. elegans. In Drosophila in vivo, flies can obviously scope with impaired T353 phosphorylation, as overall development and hatching rates are not impaired. However, adult flies exhibit a reduced body size due to decreased cell proliferation upon overactivation of the AMPK/mTOR axis, which controls cell growth and proliferation . Thus, in their physiological environment, the regulation of LKB1 activity by PDK1 might turned out to be a selection advantage during evolution. Of note, mimicking a constitutive phosphorylation of LKB1 (T353D) displays identical phenotypes as wild type LKB1. This suggests that under cellular conditions of PDK1 activation, e.g., by PI3K activation upon growth factor stimulation, LKB1 is mainly phosphorylated by PDK1, thus providing a negative feedback mechanism, inhibiting aberrant mTOR activation by PI3K/PDK1 via the LKB1-AMPK axis. Up to now, several upstream phosphorylation sites of LKB1 have been described to fine-tune the activity of LKB1: conserved residues at the very C-terminus (S562 in Drosophila, S428 in human LKB1) can be phosphorylated by PKA as well as aPKC, thereby promoting nuclear export, and thus, activation of LKB1 . LKB1 Tyrosine phosphorylation by Fyn results in a similar activation , whereas phosphorylation of LKB1 by Aurora-A blocks binding and activation of AMPK, thus inhibiting the LKB1/AMPK axis . The phosphorylation of LKB1 by PDK1 described in this study adds another upstream regulatory mechanism of LKB1, which is likely to be important for the finetuning of LKB1's activity during development. Furthermore, this signaling pathway might serve as a backup to compensate increased PI(3,4,5)P3 levels in the plasma membrane, e.g., in PI3K gain of function or PTEN loss of function mutations, which occur in various types of cancer: increased PI(3,4,5)P3 enhances the activation of PDK1/Akt, which in turn results in the activation of mTOR , which is counterbalanced by simultaneous activation of LKB1/AMPK and subsequent inhibition of mTOR by AMPK . In tumors, aberrant PDK1 activation by enhanced production of PI(3,4,5)P3 due to mutations in PI3K or PTEN frequently coincidences with downregulation of mutation of LKB1 , resulting in a decreased activation of AMPK, further enhancing Akt/mTOR activation, and thus, cancer progression. PDK1 has been well characterized to phosphorylate and, thereby, activate kinases of the AGC family . To our knowledge, LKB1 is only the third non-AGC kinase substrate, apart from the kinases p21-activated kinase (PAK1 ) and polo-like kinase and Integrin b3 to be phosphorylated by PDK1. Both kinases, LKB1 and PDK1, are recruited to the plasma membrane by direct binding to phospholipids: LKB1 binds to phosphatidic acid via its C-terminal polybasic motif , while PDK1 contains a PH domain, which preferentially recognizes PI(3,4,5)P3 . Thus, the plasma membrane might serve as a platform for the LKB1-PDK1 interaction and the regulation of LKB1 by PDK1 phosphorylation in a similar mechanism as described for PDK1/Akt --although in that case, both enzymes bind to the same phospholipid (PI(3,4,5)P3). Acknowledgments We thank the Bloomington Drosophila stock center at the University of Indiana (USA) and the Developmental Studies Hybridoma Bank at the University of Iowa (USA) for providing reagents. This work was supported by grants of the German research foundation (DFG) to M. P. K. (CRC1348-A05) and A.H. (CRC1348-A01) and the Interdisciplinary Center for Clinical Research (IZKF) Munster (Kr-A-031.21). Supplementary Materials The following supporting information can be downloaded at: Figure S1: RMSD of the proteins, Figure S2: Distances between pairs of residues in the binding pocket. Click here for additional data file. Author Contributions S.B., J.H., A.A., I.H. and M.G. performed the experiments, analyzed the data and revised the manuscript; P.N., A.H. and M.P.K. supervised the experiments and wrote the manuscript. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement All data are available in main and supplemental figures. Conflicts of Interest The authors declare no conflict of interest. Figure 1 LKB1 displays conserved PDK1-binding and -phosphorylation motifs and is phosphorylated by PDK1 in vitro. (A) Scheme of Drosophila LKB1 and sequence alignment of PDK1-binding and consensus-motif. DmLKB1--Drosophila melanogaster LKB1, HsLKB1--Homo sapiens LKB1, DrLKb1--Danio rerio LKB1 (zebrafish), CeLKB1--C. elegans LKB1. (B,C) S2R+ cells were co-transfected with HA-PDK1 and either wild type GFP-LKB1 and GFP-LKB1 E253A (B) or wild type GFP-LKB1 and GFP-LKB1 T253A (C) and GFP alone. GFP-proteins were immunoprecipitated and GFP(-LKB1) and HA-PDK1 (B) or Myc-PDK1 (C) were detected by immunoblotting. (D) Recombinant wild type or mutant MBP-LKB1 produced in E. coli was used in a 32P radioactive kinase assay with recombinant PDK1. CCB is colloidal coomassie blue, which was used to visualize the LKB1 input proteins. Figure 2 GFP-LKB1 variants localize normally and do not impair fly survival but affect body size (A-C) Immunostainings of embryonic epidermis epithelial cells expressing GFP-LKB1 variants from its endogenous promoter. Discs large (Dlg) was used as marker for the lateral membrane and Bazooka (Baz) marks the apical junctions. (D) Flies with CRISPR/Cas9-mediated knockin of wild type LKB1, LKB1 T353A or LKB1 T353D display comparable survival rates (n = 100, N = 3), but different body sizes (E), n >= 35. Scales bars are 5 mm in (A-C). Error bars are standard error of the means. Significance was determined by one way ANOVA with Bonferroni multiple comparison test: * p < 0.05, n.s. not significant. Figure 3 Simulation of LKB1 phosphorylation reveals changes in the ATP-binding pocket. (A) The RMSF of the protein Ca atoms averaged over the two samples is shown for human LKB1 (Blue) and human phospho-LKB1 (pLKB1, red). (B) Overlay of the two average structures for human LKB1 (blue) and human pLKB1 (orange). The average structures were obtained from 1 to 3 ms of the simulations. (C) The difference between human LKB1 and human pLKB1 average structures are shown for the two samples. For each Ca atom a shift distance is determined. The atoms with the 50% largest distances are described by a shifted value of 1. The remaining distances are ordered and linearly mapped on shift values between 0 and 1. The shift values are translated into the respective color codes. This non-linear procedure avoids that the parts, which are dramatically changed, blur the other parts. On the right-hand side, the cartoon representation of the phosphorylated protein is shown, colored based on the right-hand side figure, showing the degree of change in different parts of the protein between the two systems. (D) The distance between S60 and A194 residues over the simulation time. The protein is drawn and these two residues, along with the pT230 (corresponding to T353 in Drosophila LKB1) residue, are shown as stick representations. (E) The free volume available inside the ATP-binding pocket, obtained from 0.9 to 1.9 ms (represented with a dashed rectangle) of the first sample simulations, was calculated using CASTp and is shown in red spheres. (F) The distance between K78 and E98 residues over the simulation time along with the pT230 residue are shown in stick representations. (G) The free volume available inside the protein for the average structures, obtained from 2.0 to 2.4 ms (represented with a dashed rectangle) of the second sample simulations. Figure 4 LKB1 T353 phosphorylation regulates AMPK activation, cell size and proliferation in vivo. (A) Western blots of embryonic lysates of LKB1 wt, T353A and T353D knockin flies against the indicated proteins. (B) In vitro kinase assay with recombinant GST-AMPK (aa 108-280) and the indicated LKB1 variants plus its cofactor Stlk. CCB is colloidal coomassie blue, which was used to visualize the GST-AMPK input proteins. (C-F) GFP-marked MARCM (mosaic analysis with a repressible cell marker) clones of LKB1 knockin-variants in otherwise wild type tissue in wing imaginal discs stained with a proliferation marker (Histone 3 phospho-S10, pH3) and Dlg to label cell boundaries. The size of GFP-marked cells was quantified (n > 200). Scale bars are 20 mm. Error bars are standard error of the means. Significance was determined by one way ANOVA with Bonferroni multiple comparison test: *** p < 0.001, n.s. not significant. (G) S2R+ cells were co-transfected with Mo25-HA, Stlk-Myc and either wild type GFP-LKB1 or GFP-LKB1 T353A. GFP alone was used as control. GFP proteins were immunoprecipitated and GFP(-LKB1), Mo25-HA and Stlk-Myc were detected by immunoblotting. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Kemphues K.J. Priess J.R. Morton D.G. Cheng N.S. Identification of genes required for cytoplasmic localization in early C. elegans embryos Cell 1988 52 311 320 10.1016/S0092-8674(88)80024-2 3345562 2. Watts J.L. Morton D.G. Bestman J. Kemphues K.J. The C. elegans par-4 gene encodes a putative serine-threonine kinase required for establishing embryonic asymmetry Development 2000 127 1467 1475 10.1242/dev.127.7.1467 10704392 3. Amin N. Khan A. St Johnston D. Tomlinson I. Martin S. Brenman J. McNeill H. LKB1 regulates polarity remodeling and adherens junction formation in the Drosophila eye Proc. Natl. Acad. Sci. USA 2009 106 8941 8946 10.1073/pnas.0812469106 19443685 4. Martin S.G. St Johnston D. A role for Drosophila LKB1 in anterior-posterior axis formation and epithelial polarity Nature 2003 421 379 384 10.1038/nature01296 12540903 5. Bonaccorsi S. Mottier V. Giansanti M.G. Bolkan B.J. Williams B. Goldberg M.L. Gatti M. The Drosophila Lkb1 kinase is required for spindle formation and asymmetric neuroblast division Development 2007 134 2183 2193 10.1242/dev.02848 17507418 6. Gailite I. Aerne B.L. Tapon N. Differential control of Yorkie activity by LKB1/AMPK and the Hippo/Warts cascade in the central nervous system Proc. Natl. Acad. Sci. USA 2015 112 E5169 E5178 10.1073/pnas.1505512112 26324895 7. Lizcano J.M. Goransson O. Toth R. Deak M. Morrice N.A. Boudeau J. Hawley S.A. Udd L. Makela T.P. Hardie D.G. LKB1 is a master kinase that activates 13 kinases of the AMPK subfamily, including MARK/PAR-1 EMBO J. 2004 23 833 843 10.1038/sj.emboj.7600110 14976552 8. Karuman P. Gozani O. Odze R.D. Zhou X.C. Zhu H. Shaw R. Brien T.P. Bozzuto C.D. Ooi D. Cantley L.C. The Peutz-Jegher gene product LKB1 is a mediator of p53-dependent cell death Mol. Cell 2001 7 1307 1319 10.1016/S1097-2765(01)00258-1 11430832 9. Mehenni H. Lin-Marq N. Buchet-Poyau K. Reymond A. Collart M.A. Picard D. Antonarakis S.E. LKB1 interacts with and phosphorylates PTEN: A functional link between two proteins involved in cancer predisposing syndromes Hum. Mol. Genet. 2005 14 2209 2219 10.1093/hmg/ddi225 15987703 10. Vaahtomeri K. Makela T.P. Molecular mechanisms of tumor suppression by LKB1 FEBS Lett. 2011 585 944 951 10.1016/j.febslet.2010.12.034 21192934 11. Hemminki A. Markie D. Tomlinson I. Avizienyte E. Roth S. Loukola A. Bignell G. Warren W. Aminoff M. Hoglund P. A serine/threonine kinase gene defective in Peutz-Jeghers syndrome Nature 1998 391 184 187 10.1038/34432 9428765 12. Jenne D.E. Reimann H. Nezu J. Friedel W. Loff S. Jeschke R. Muller O. Back W. Zimmer M. Peutz-Jeghers syndrome is caused by mutations in a novel serine threonine kinase Nat. Genet. 1998 18 38 43 10.1038/ng0198-38 9425897 13. Sanchez-Cespedes M. A role for LKB1 gene in human cancer beyond the Peutz-Jeghers syndrome Oncogene 2007 26 7825 7832 10.1038/sj.onc.1210594 17599048 14. Chen J. Zhang X.D. Proud C. Dissecting the signaling pathways that mediate cancer in PTEN and LKB1 double-knockout mice Sci. Signal. 2015 8 pe1 10.1126/scisignal.aac8321 26329580 15. Boudeau J. Baas A.F. Deak M. Morrice N.A. Kieloch A. Schutkowski M. Prescott A.R. Clevers H.C. Alessi D.R. MO25alpha/beta interact with STRADalpha/beta enhancing their ability to bind, activate and localize LKB1 in the cytoplasm EMBO J. 2003 22 5102 5114 10.1093/emboj/cdg490 14517248 16. Boudeau J. Scott J.W. Resta N. Deak M. Kieloch A. Komander D. Hardie D.G. Prescott A.R. van Aalten D.M. Alessi D.R. Analysis of the LKB1-STRAD-MO25 complex J. Cell Sci. 2004 117 6365 6375 10.1242/jcs.01571 15561763 17. Dorfman J. Macara I.G. STRADalpha regulates LKB1 localization by blocking access to importin-alpha, and by association with Crm1 and exportin-7 Mol. Biol. Cell 2008 19 1614 1626 10.1091/mbc.e07-05-0454 18256292 18. Kullmann L. Krahn M.P. Controlling the master-upstream regulation of the tumor suppressor LKB1 Oncogene 2018 37 3045 3057 10.1038/s41388-018-0145-z 29540834 19. Dogliotti G. Kullmann L. Dhumale P. Thiele C. Panichkina O. Mendl G. Houben R. Haferkamp S. Puschel A.W. Krahn M.P. Membrane-binding and activation of LKB1 by phosphatidic acid is essential for development and tumour suppression Nat. Commun. 2017 8 15747 10.1038/ncomms15747 28649994 20. Fang Y. Park I.H. Wu A.L. Du G. Huang P. Frohman M.A. Walker S.J. Brown H.A. Chen J. PLD1 regulates mTOR signaling and mediates Cdc42 activation of S6K1 Curr. Biol. 2003 13 2037 2044 10.1016/j.cub.2003.11.021 14653992 21. Ghosh S. Strum J.C. Sciorra V.A. Daniel L. Bell R.M. Raf-1 kinase possesses distinct binding domains for phosphatidylserine and phosphatidic acid. Phosphatidic acid regulates the translocation of Raf-1 in 12-O-tetradecanoylphorbol-13-acetate-stimulated Madin-Darby canine kidney cells J. Biol. Chem. 1996 271 8472 8480 10.1074/jbc.271.14.8472 8626548 22. Limatola C. Schaap D. Moolenaar W.H. van Blitterswijk W.J. Phosphatidic acid activation of protein kinase C-zeta overexpressed in COS cells: Comparison with other protein kinase C isotypes and other acidic lipids Biochem. J. 1994 304 Pt 3 1001 1008 10.1042/bj3041001 7818462 23. Toschi A. Lee E. Xu L. Garcia A. Gadir N. Foster D.A. Regulation of mTORC1 and mTORC2 complex assembly by phosphatidic acid: Competition with rapamycin Mol. Cell Biol. 2009 29 1411 1420 10.1128/MCB.00782-08 19114562 24. Virbasius J.V. Song X. Pomerleau D.P. Zhan Y. Zhou G.W. Czech M.P. Activation of the Akt-related cytokine-independent survival kinase requires interaction of its phox domain with endosomal phosphatidylinositol 3-phosphate Proc. Natl. Acad. Sci. USA 2001 98 12908 12913 10.1073/pnas.221352898 11606732 25. Zelasko J. Czogalla A. Selectivity of mTOR-Phosphatidic Acid Interactions Is Driven by Acyl Chain Structure and Cholesterol Cells 2021 11 119 10.3390/cells11010119 35011681 26. Levina A. Fleming K.D. Burke J.E. Leonard T.A. Activation of the essential kinase PDK1 by phosphoinositide-driven trans-autophosphorylation Nat. Commun. 2022 13 1874 10.1038/s41467-022-29368-4 35387990 27. Mora A. Komander D. Van Aalten D.M. Alessi D.R. PDK1, the master regulator of AGC kinase signal transduction Semin. Cell Dev. Biol. 2004 15 161 170 10.1016/j.semcdb.2003.12.022 15209375 28. Truebestein L. Hornegger H. Anrather D. Hartl M. Fleming K.D. Stariha J.T.B. Pardon E. Steyaert J. Burke J.E. Leonard T.A. Structure of autoinhibited Akt1 reveals mechanism of PIP3-mediated activation Proc. Natl. Acad. Sci. USA 2021 118 e2101496118 10.1073/pnas.2101496118 34385319 29. Fruman D.A. Chiu H. Hopkins B.D. Bagrodia S. Cantley L.C. Abraham R.T. The PI3K Pathway in Human Disease Cell 2017 170 605 635 10.1016/j.cell.2017.07.029 28802037 30. Manning B.D. Toker A. AKT/PKB Signaling: Navigating the Network Cell 2017 169 381 405 10.1016/j.cell.2017.04.001 28431241 31. Gratz S.J. Cummings A.M. Nguyen J.N. Hamm D.C. Donohue L.K. Harrison M.M. Wildonger J. O'Connor-Giles K.M. Genome engineering of Drosophila with the CRISPR RNA-guided Cas9 nuclease Genetics 2013 194 1029 1035 10.1534/genetics.113.152710 23709638 32. Kullmann L. Krahn M.P. Redundant regulation of localization and protein stability of DmPar3 Cell. Mol. Life Sci. CMLS 2018 75 3269 3282 10.1007/s00018-018-2792-1 29523893 33. Schneider I. Cell lines derived from late embryonic stages of Drosophila melanogaster J. Embryol. Exp. Morphol. 1972 27 353 365 10.1242/dev.27.2.353 4625067 34. Zeqiraj E. Filippi B.M. Deak M. Alessi D.R. van Aalten D.M. Structure of the LKB1-STRAD-MO25 complex reveals an allosteric mechanism of kinase activation Science 2009 326 1707 1711 10.1126/science.1178377 19892943 35. Humphrey W. Dalke A. Schulten K. VMD: Visual molecular dynamics J. Mol. Graph. 1996 14 33 38 10.1016/0263-7855(96)00018-5 8744570 36. Fiser A. Do R.K.G. Sali A. Modeling of loops in protein structures Protein. Sci. 2000 9 1753 1773 10.1110/ps.9.9.1753 11045621 37. Jo S. Kim T. Iyer V.G. Im W. CHARMM-GUI: A web-based graphical user interface for CHARMM J. Comput. Chem. 2008 29 1859 1865 10.1002/jcc.20945 18351591 38. Van Der Spoel D. Lindahl E. Hess B. Groenhof G. Mark A.E. Berendsen H.J. GROMACS: Fast, flexible, and free J. Comput. Chem. 2005 26 1701 1718 10.1002/jcc.20291 16211538 39. Klauda J.B. Venable R.M. Freites J.A. O'Connor J.W. Tobias D.J. Mondragon-Ramirez C. Vorobyov I. MacKerell A.D. Jr. Pastor R.W. Update of the CHARMM all-atom additive force field for lipids: Validation on six lipid types J. Phys. Chem. B 2010 114 7830 7843 10.1021/jp101759q 20496934 40. Essmann U. Perera L. Berkowitz M.L. Darden T. Lee H. Pedersen L.G. A Smooth Particle Mesh Ewald Method J. Chem. Phys. 1995 103 8577 8593 10.1063/1.470117 41. Parrinello M. Rahman A. Polymorphic Transitions in Single-Crystals-a New Molecular-Dynamics Method J. Appl. Phys. 1981 52 7182 7190 10.1063/1.328693 42. Nose S. A Unified Formulation of the Constant Temperature Molecular-Dynamics Methods J. Chem. Phys. 1984 81 511 519 10.1063/1.447334 43. Hess B. P-LINCS: A Parallel Linear Constraint Solver for Molecular Simulation J. Chem. Theory Comput. 2008 4 116 122 10.1021/ct700200b 26619985 44. Michaud-Agrawal N. Denning E.J. Woolf T.B. Beckstein O. MDAnalysis: A toolkit for the analysis of molecular dynamics simulations J. Comput. Chem. 2011 32 2319 2327 10.1002/jcc.21787 21500218 45. Gao X. Harris T.K. Steady-state kinetic mechanism of PDK1 J. Biol. Chem. 2006 281 21670 21681 10.1074/jbc.M602448200 16737971 46. Lee J.H. Koh H. Kim M. Park J. Lee S.Y. Lee S. Chung J. JNK pathway mediates apoptotic cell death induced by tumor suppressor LKB1 in Drosophila Cell Death Differ. 2006 13 1110 1122 10.1038/sj.cdd.4401790 16273080 47. O'Farrell F. Lobert V.H. Sneeggen M. Jain A. Katheder N.S. Wenzel E.M. Schultz S.W. Tan K.W. Brech A. Stenmark H. Class III phosphatidylinositol-3-OH kinase controls epithelial integrity through endosomal LKB1 regulation Nat. Cell Biol. 2017 19 1412 1423 10.1038/ncb3631 29084199 48. Tian W. Chen C. Lei X. Zhao J.L. Liang J. CASTp 3.0: Computed atlas of surface topography of proteins Nucleic Acids Res. 2018 46 W363 W367 10.1093/nar/gky473 29860391 49. Lee T. Luo L. Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis Neuron 1999 22 451 461 10.1016/S0896-6273(00)80701-1 10197526 50. Pinner S. Sahai E. PDK1 regulates cancer cell motility by antagonising inhibition of ROCK1 by RhoE Nat. Cell Biol. 2008 10 127 137 10.1038/ncb1675 18204440 51. Gagliardi P.A. Di Blasio L. Puliafito A. Seano G. Sessa R. Chianale F. Leung T. Bussolino F. Primo L. PDK1-mediated activation of MRCKalpha regulates directional cell migration and lamellipodia retraction J. Cell Biol. 2014 206 415 434 10.1083/jcb.201312090 25092657 52. Raimondi C. Chikh A. Wheeler A.P. Maffucci T. Falasca M. A novel regulatory mechanism links PLCgamma1 to PDK1 J. Cell Sci. 2012 125 3153 3163 10.1242/jcs.100511 22454520 53. Saxton R.A. Sabatini D.M. mTOR Signaling in Growth, Metabolism, and Disease Cell 2017 169 361 371 10.1016/j.cell.2017.03.035 28388417 54. Song P. Xie Z. Wu Y. Xu J. Dong Y. Zou M.H. Protein kinase Czeta-dependent LKB1 serine 428 phosphorylation increases LKB1 nucleus export and apoptosis in endothelial cells J. Biol. Chem. 2008 283 12446 12455 10.1074/jbc.M708208200 18321849 55. Yamada E. Pessin J.E. Kurland I.J. Schwartz G.J. Bastie C.C. Fyn-dependent regulation of energy expenditure and body weight is mediated by tyrosine phosphorylation of LKB1 Cell Metab. 2010 11 113 124 10.1016/j.cmet.2009.12.010 20142099 56. Zheng X. Chi J. Zhi J. Zhang H. Yue D. Zhao J. Li D. Li Y. Gao M. Guo J. Aurora-A-mediated phosphorylation of LKB1 compromises LKB1/AMPK signaling axis to facilitate NSCLC growth and migration Oncogene 2018 37 502 511 10.1038/onc.2017.354 28967900 57. Sarbassov D.D. Guertin D.A. Ali S.M. Sabatini D.M. Phosphorylation and regulation of Akt/PKB by the rictor-mTOR complex Science 2005 307 1098 1101 10.1126/science.1106148 15718470 58. Shaw R.J. Bardeesy N. Manning B.D. Lopez L. Kosmatka M. DePinho R.A. Cantley L.C. The LKB1 tumor suppressor negatively regulates mTOR signaling Cancer Cell 2004 6 91 99 10.1016/j.ccr.2004.06.007 15261145 59. Corradetti M.N. Inoki K. Bardeesy N. DePinho R.A. Guan K.L. Regulation of the TSC pathway by LKB1: Evidence of a molecular link between tuberous sclerosis complex and Peutz-Jeghers syndrome Genes Dev. 2004 18 1533 1538 10.1101/gad.1199104 15231735 60. Pullen N. Dennis P.B. Andjelkovic M. Dufner A. Kozma S.C. Hemmings B.A. Thomas G. Phosphorylation and activation of p70s6k by PDK1 Science 1998 279 707 710 10.1126/science.279.5351.707 9445476 61. Kobayashi T. Cohen P. Activation of glucocorticoid-regulated protein kinase by agonists that activate phosphatidylinositide 3-kinase is mediated by 3-phosphoinositide-dependent protein kinase-1 (PDK1) and PDK2 Biochem. J. 1999 339 Pt 2 319 328 10.1042/bj3390319 10191262 62. Jensen C.J. Buch M.B. Krag T.O. Hemmings B.A. Gammeltoft S. Frodin M. 90-kDa ribosomal S6 kinase is phosphorylated and activated by 3-phosphoinositide-dependent protein kinase-1 J. Biol. Chem. 1999 274 27168 27176 10.1074/jbc.274.38.27168 10480933 63. Le Good J.A. Ziegler W.H. Parekh D.B. Alessi D.R. Cohen P. Parker P.J. Protein kinase C isotypes controlled by phosphoinositide 3-kinase through the protein kinase PDK1 Science 1998 281 2042 2045 10.1126/science.281.5385.2042 9748166 64. Dutil E.M. Toker A. Newton A.C. Regulation of conventional protein kinase C isozymes by phosphoinositide-dependent kinase 1 (PDK-1) Curr. Biol. 1998 8 1366 1375 10.1016/S0960-9822(98)00017-7 9889098 65. King C.C. Gardiner E.M. Zenke F.T. Bohl B.P. Newton A.C. Hemmings B.A. Bokoch G.M. p21-activated kinase (PAK1) is phosphorylated and activated by 3-phosphoinositide-dependent kinase-1 (PDK1) J. Biol. Chem. 2000 275 41201 41209 10.1074/jbc.M006553200 10995762 66. Tan J. Li Z. Lee P.L. Guan P. Aau M.Y. Lee S.T. Feng M. Lim C.Z. Lee E.Y. Wee Z.N. PDK1 signaling toward PLK1-MYC activation confers oncogenic transformation, tumor-initiating cell activation, and resistance to mTOR-targeted therapy Cancer Discov. 2013 3 1156 1171 10.1158/2159-8290.CD-12-0595 23887393 67. Kirk R.I. Sanderson M.R. Lerea K.M. Threonine phosphorylation of the beta 3 integrin cytoplasmic tail, at a site recognized by PDK1 and Akt/PKB in vitro, regulates Shc binding J. Biol. Chem. 2000 275 30901 30906 10.1074/jbc.M001908200 10896934 68. di Blasio L. Gagliardi P.A. Puliafito A. Sessa R. Seano G. Bussolino F. Primo L. PDK1 regulates focal adhesion disassembly by modulating endocytosis of alphavbeta3 integrin J. Cell Sci. 2015 128 863 877 10.1242/jcs.149294 25588838 69. Deak M. Casamayor A. Currie R.A. Downes C.P. Alessi D.R. Characterisation of a plant 3-phosphoinositide-dependent protein kinase-1 homologue which contains a pleckstrin homology domain FEBS Lett. 1999 451 220 226 10.1016/S0014-5793(99)00556-6 10371193 70. Milburn C.C. Deak M. Kelly S.M. Price N.C. Alessi D.R. Van Aalten D.M. Binding of phosphatidylinositol 3,4,5-trisphosphate to the pleckstrin homology domain of protein kinase B induces a conformational change Biochem. J. 2003 375 531 538 10.1042/bj20031229 12964941 |
PMC10000616 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051017 foods-12-01017 Article Antioxidant Capacity, Nitrite and Nitrate Content in Beetroot-Based Dietary Supplements Brzezinska-Rojek Joanna 1 Sagatovych Svitlana 2 Malinowska Paulina 2 Gadaj Kamila 1 Prokopowicz Magdalena 3 Grembecka Malgorzata 1* Kovacevic Danijela Bursac Academic Editor Pavlic Branimir Academic Editor Misan Aleksandra Academic Editor Putnik Predrag Academic Editor 1 Department of Bromatology, Faculty of Pharmacy, Medical University of Gdansk, Al. Gen. J. Hallera 107, 80-416 Gdansk, Poland 2 Student Scientific Circle at the Department of Bromatology, Faculty of Pharmacy, Medical University of Gdansk, Al. Gen. J. Hallera 107, 80-416 Gdansk, Poland 3 Department of Physical Chemistry, Faculty of Pharmacy, Medical University of Gdansk, Al. Gen. J. Hallera 107, 80-416 Gdansk, Poland * Correspondence: [email protected] 27 2 2023 3 2023 12 5 101701 2 2023 20 2 2023 23 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Due to the high content of bioactive substances, beetroot and its preserves might be a valuable constituent of a diet. Research into the antioxidant capacity and content of nitrate (III) and (V) in beetroot-based dietary supplements (DSs) worldwide is limited. The Folin-Ciocalteu method, CUPRAC, DPPH, and Griess methods were used to determine total antioxidant capacity, total phenolic content, nitrites, and nitrates content in fifty DSs and twenty beetroot samples. Moreover, the safety of products was evaluated because of the concentration of nitrites, nitrates, and the correctness of labelling. The research showed that a serving of fresh beetroot provides significantly more antioxidants, nitrites, and nitrates than most daily portions of DSs. Product P9 provided the highest dose of nitrates (169 mg/daily dose). However, in most cases, the consumption of DSs would be associated with a low health value. The acceptable daily intake was not exceeded in the cases of nitrites (0.0015-0.55%) and nitrates (0.056-48%), assuming that the supplementation followed the manufacturer's recommendation. According to European and Polish regulations, 64% of the products tested did not meet all the requirements for labelling food packaging. The findings point to the need for tighter regulation of DSs, as their consumption might be dangerous. antioxidant capacity beetroot nitrate nitrite dietary supplement CUPRAC DPPH Folin-Ciocalteu European Social FundOperational Programme Knowledge Education DevelopmentThe study was supported by the project POWR.03.02.00-00-I014/17-00 co-financed by the European Union through the European Social Fund under the Operational Programme Knowledge Education Development 2014-2020. pmc1. Introduction Beetroot (Beta vulgaris L.) is a rich source of nutrients and bioactive substances such as fibre, carbohydrates, and phenolic compounds. In addition, this vegetable contains microelements such as potassium, iron, calcium, copper, sodium, and zinc, as well as vitamins B1, B2, B3, B6, biotin, and B12. Red beetroot owes its characteristic intense colour to betalain pigments-betacyanins: betanin (the dominant pigment), isobetanin, betanidin, isobetanidin, vulgaxanthin I and II, and indixanthin . Beetroot peel has the highest betanin content. A correlation between antioxidant activity and the content of betacyanins has been found . Betacyanins, along with phenolic acids, flavonoids, and ascorbic acid, are responsible for the antioxidant properties of beetroot. Furthermore, this vegetable is rich in nitrites and nitrates . The oral bioavailability of nitrates from plants is 100% . Beetroots might lead to several health-promoting effects, such as a stimulating effect on the circulatory and immune systems; improving the functioning of the endothelium; regulating the level of blood pressure; protecting the liver, the intestines, and the kidneys against toxic compounds; protecting against radiation consequences; and strengthening the gastric mucosa . Due to these effects, the consumption of beetroot products may be beneficial in the cases of diabetes , post-menopausal women , diseases of the cardiovascular system , and athletes' support . Moreover, beetroot products with an appropriate concentration of inorganic nitrites can be an effective ergogenic agent, acting faster than a product containing only nitrate salts . An excessive supply of nitrates may pose a health risk; therefore, in the interest of the health of consumers, a maximum acceptable daily intake (ADI) has been established that does not harm health when consumed throughout life. For nitrates, it is 0-5 mg/kg b.w. NO3 ions (corresponding to 0-3.7 mg NaNO3), while for nitrites, it is 0-0.2 mg/kg b.w. NO2 ions (corresponding to 0-0.07 mg of NaNO2) . The main source of nitrates in the diet is vegetables. It is estimated that they provide 80-85% of the nitrates consumed. The supply of these compounds in drinking water, meat, or processed foods is much less important . The total amount of nitrates consumed from all sources should be monitored, as there is a risk of exceeding the ADI, especially in children who have a lower body weight. Poisoning with nitrites may lead to methemoglobinemia or the development of neoplasms due to the formation of N-nitrosamines, which are carcinogenic. N-nitrosamines can be formed in the acidic environment of the stomach from nitrites in their reaction with secondary and tertiary amines . Beetroot is consumed in various forms, such as fresh vegetables, juice, pickles, chips, and gel . Dietary supplements containing Beta vulgaris L., manufactured in the form of tablets, lozenges, capsules, juices, powder, and many others, are also popular. However, producers often do not standardise products, which casts doubt on their effectiveness. Consequently, several potential risks for consumers appear, such as exposure to an excessive supply of nitrates or nitrites and loading the body with a product without health-promoting properties due to the lack of data on effectiveness compared to a fresh vegetable or the content of bioactive substances. A potential risk is also associated with the mislabelling of finished products. Total antioxidant capacity (TAC) describes the antioxidant properties of a complex material (such as beetroot and beetroot preserves) consisting of numerous compounds. It is not just the sum of the antioxidant capacities of individual bioactive compounds. The TAC is the result of the synergistic effects of the different bioactive substances, trace elements, metals, vitamins, and other food constituents . It was decided to determine the TAC instead of the concentrations of individual antioxidant substances because both the DSs and the vegetables are complex matrices, and their biological effect will be the result of the interaction of various components. The research aimed to assess the quality and safety of beetroot-based DSs in comparison with beetroot samples. The TAC, total phenolic content (TPC), nitrites, and nitrate contents of fifty beetroot-based DSs (in the form of tablets, capsules, and powders) and twenty samples of beetroots available on the Polish market were determined for this purpose. Vegetables were divided into three subgroups: peeled, unpeeled, and skins. Reference was made to the average values for conventional and organic beetroots to compare DSs with vegetables. Manufacturers usually do not provide information on how the beetroot has been processed before manufacturing beetroot-based DSs. On several products, there was information that whole beetroot was used, which is why we also included vegetables with skins in our analysis. DSs are concentrated forms, so they can potentially provide a significant amount of bioactive substances, especially antioxidants and nitrogen compounds. As a result, the DS results were compared to vegetables to determine which are better sources of antioxidants, nitrites, and nitrates. The health risk was assessed because of the realisation of ADI for nitrites and nitrates. Furthermore, the correctness of the labelling of finished products was assessed based on Polish and European regulations because misinformation might also be dangerous for consumers. Statistical analyses were applied to verify the potential correlation between different methods of antioxidant capacity assessment and between TAC and nitrate and nitrite content. Moreover, it was assessed whether the content of antioxidants, nitrites, and nitrates differed statistically significantly in individual subgroups of beetroots. Despite the growing popularity of beetroot supplements, there is a lack of research on the quality and safety of their use. 2. Materials and Methods 2.1. Materials 2.1.1. Sample Preparation The process of collection and sample preparation has been shown in Figure 1. Tables S1 and S2 provided in-depth details about beetroots and dietary supplements (DSs), respectively. Beetroot samples were lyophilized in an Alpha 1-4 LD plus freeze dryer (Christ, Osterode am Harz, Germany). Every DS was signed according to the alphanumeric code, including the formulation and the sequence number. Moreover, the letters (A, B, and C) were used to mark the same DSs with other serial numbers. To be more specific, ten of the examined products did not meet the requirements for labelling the category of dietary supplements, but they were tentatively included in the group of DSs in the following section of the work. At the purchase stage, they were described by sellers as "dietary supplements". Only the verification of the labelling showed that they do not legally belong to this group, they are just traditional food products. They were marked in orange in Table S2. The exclusion criteria for beetroot-based DSs were: other forms of products (such as juice, shot, gel and bar); beetroot was not a main ingredient but an auxiliary substance; and unavailability for Polish consumers in the mentioned time period. Only ceramic tools were used for sample preparation. In total, seventy samples of DSs and vegetables were analysed in triplicate. 2.1.2. Reagents and Standards Reagents for the Folin-Ciocalteu assay were as follows: anhydrous sodium carbonate (purity > 99.5%, Chempur(r), Piekary Slaskie, Poland), Folin-Ciocalteu reagent (analytical grade, Chempur(r), Piekary Slaskie, Poland), gallic acid (Sigma-Aldrich(r), Darmstadt, Germany). Reagents for the CUPRAC assay were as follows: ethanol 96% (LiChrosolv(r), Darmstadt, Germany), 6-Hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid (Sigma-Aldrich(r), Buchs, Switzerland), copper (II) chloride (Sigma-Aldrich(r), St. Louis, MO, USA), ammonium acetate (Sigma-Aldrich(r), Darmstadt, Germany), neocuproine (Sigma-Aldrich(r), Darmstadt, Germany). Reagents for the DPPH assay were as follows: methanol (LiChrosolv(r), Darmstadt, Germany), 2,2-diphenyl-1-picrylhydrazyl (Sigma-Aldrich(r), Darmstadt, Germany). Reagents for the nitrites and nitrates determination were as follows: hydrochloric acid 35-38% (Chempur(r), Piekary Slaskie, Poland), acetic acid min. 99.5% (Chempur(r), Piekary Slaskie, Poland), sodium nitrite (Chempur(r), Piekary Slaskie, Poland), sodium tetraborate (POCH, Gliwice, Poland), sulfanilamide (Sigma-Aldrich(r), Darmstadt, Germany), N-naphthylethylenediamine dihydrochloride (Sigma-Aldrich(r), St. Louis, MO, USA), zinc acetate dihydrate (POCH, Gliwice, Poland), Carrez solution I--Potassium hexacyanoferrate (II) 0.25 mol/L (aqueous solution) (VWR, Leuven, Belgium), ammonium buffer pH 9.6 (obtained from ammonia (POCH, Gliwice, Poland) and hydrochloric acid 37% (VWR, Leuven, Belgium), cadmium sulphate (VI) (Chempur(r), Piekary Slaskie, Poland), zinc sticks >=99.99% (Merck Millipore, Darmstadt, Germany). Ultrapure water (18.2 MO*cm, Millipore Simplicity System, Billerica, MA, USA) was used for all aqueous solutions. 2.2. The Total Antioxidant Capacity (TAC) and Total Phenolic Content (TPC) The assessment of the TAC was carried out using the CUPRAC and DPPH assays. Moreover, the Folin-Ciocalteu (FC) method was applied to determine TPC. 2.2.1. Extract Preparation for TAC and TPC Determination Based on the conditions described by Capanoglu et al. and other literature reports on the extraction of beetroot products , optimisation of the extraction was performed using seven combinations of solvents in three variants of extraction . Each variant was examined in triplicate with a threefold measurement. Variant C was chosen as the most optimal (based on results of ANOVA test): two-stage ultrasound-assisted extraction with 50% MeOH + 0.1% FA . Table S3 summarises the obtained results. The extraction was carried out on the selected lyophilizate; therefore, the results are expressed in mg GAE/g of lyophilizate. Figure 3 depicts the extraction procedure. Each sample was examined in triplicate with a threefold measurement. Lyophilizates and DSs were homogenized in a mortar straight before analyses. The samples were kept in the freezer all of the time and were thawed at room temperature before starting the individual analyses. 2.2.2. TPC Determination The total phenolic content (TPC) in the extracted samples was determined using the Folin-Ciocalteau reagent (FCR), according to the optimised and validated method developed by the authors based on literature research . The mutual ratio of the reagents used (FCR and Na2CO3) and the incubation time before the measurement were optimised according to the scheme shown in Figure 2. Each variant was examined in triplicate with a threefold measurement. The obtained results are summarised in Tables S4 and S5. Model 1 was found to be the most efficient (ANOVA; p < 0.05), thus, 5 mL of FCR was mixed with 10 mL of Na2CO3, and 30 min of incubation was applied. It was assumed that the composition of supplements in the form of tablets may differ the most from pure lyophilizate due to the presence of auxiliary substances enabling the tabletting process; thus, the time of incubation was also optimised for the product in a tablet. Sample extracts (x mL) and Mili-Q water ((1.0 - x) mL) were placed in a centrifuge tube to have a volume of 1 mL. Next, 5 mL of FCR was added, the sample was mixed, and it was left for 3 min. Then, 10 mL of saturated sodium carbonate solution (150 g/L) was added. The test tubes were carefully blended after the addition of each reagent using a vortex (Lab dancer, VWR(r), Gdansk, Poland). The absorbance was measured threefold at 760 nm (Genesys 10S, Thermo Fisher Scientific, Waltham, MA, USA) after incubation (30 min at room temperature without light). Results were expressed in gallic acid equivalents (mg GAE/g of product and mg GAE/daily dose of product for DSs or mg GAE/g dry weight (d.w.) and mg GAE/100 g fresh weight (f.w.) of beetroot). Analogously, a calibration curve was prepared to range from 0.1 to 10 mg/mL. The calibration curve was made in three independent replications with threefold measurements. 2.2.3. CUPRAC The CUPRAC assay was carried out as described by Apak et al. . Analogously, a calibration curve was prepared in the range of 0.0005 to 0.07 uM/mL of Trolox. Three independent replications with threefold measurements were used to create the calibration curve. Results were expressed in Trolox acid equivalents (TE) (mmol TE/g of product and mmol TE/daily dose of the product, or mmol TE/g d.w. and mmol TE/100 g f.w. of beetroot). The incubation time of the samples was previously optimised. Results after 30 min and 60 min of incubation did not differ significantly, so the first one was applied. 2.2.4. DPPH The DPPH assay was carried out as described by Ravichandran et al. . The absorbance of the sample was measured threefold at 515 nm (Genesys 10S, Thermo Fisher Scientific, Waltham, MA, USA) after incubation (30 min at room temperature). Results were expressed as a percentage of the antioxidant activity, which was calculated as follows:(1) Activity (%)=Ac - AsAcx100% Ac--absorbance of control; As--absorbance of a sample. 2.3. The Nitrite and Nitrate Determination Quantification of nitrites and nitrates in beetroot (Beta vulgaris L.) and beetroot products was carried out by spectrophotometry using Griess reagents I, II and III according to ISO 6635-1984 (E) . 2.3.1. Extraction for Nitrites and Nitrates Determination A total of 1.0 to 10 g of the test sample were weighed, according to the expected nitrite content. Then, 3.0 g of activated carbon, 5 mL of disodium tetraborate solution, and 100 mL of hot, purified water were added to each sample. The flasks were shaken for 15 min at 80 degC. Next, 2 mL of potassium hexacyanoferrate (II) and 2 mL of zinc acetate solution were added to the samples. The solutions, after cooling to room temperature, were transferred to 200 mL volumetric flasks, made up to the mark, and shaken. Finally, solutions were filtered into conical flasks through paper filters. 2.3.2. Nitrites Determination At least 10 mL of solution was transferred to the 50 mL volumetric flask and diluted into 30 mL with purified water. Then, 5 mL of solution I (sulfanilamide dissolved in water with hydrochloric acid) and 3 mL of solution III (hydrochloric acid) were added. The content of the flask was thoroughly mixed and left for 1 min at ambient temperature, protected from light. Next, 1 mL of solution II (0.1% solution of N-(1-naphthyl)ethylenediamine dihydrochloride) was added, mixed carefully, and left for 3 min at ambient temperature, protected from light. After making up to the mark with water, the solution was mixed. The absorbance at a wavelength of 538 nm was measured within 15 min using the spectrometer. Results were expressed as mg/g of NO2 and mg/daily dose of NO2 or mg/g d.w. of NO2 and mg/100 g f.w. of NO2 of beetroot, which is calculated as follows:(2) (NO2-)=m1x200V1xm0 m0--the mass, in grams, of the test portion; m1--the mass, in micrograms, of nitrite ion (NO2 ) contained in the aliquot portion (V1) of filtrate taken, read from the calibration graph; V1--the volume, in millilitres, of the aliquot portion of filtrate taken. Analogously, a calibration curve was prepared to range from 0.0 to 0.06 mg/mL of nitrites. The calibration curve was made in three independent replications with threefold measurements. 2.3.3. Nitrates Determination About 2 g of the cadmium and 5 mL of the buffer solution, and an aliquot portion of the filtrate (10 mL or less) were placed in a 25 mL conical flask. The flask was agitated for 5 min. Next, the solution was filtered into a 50 mL one-mark volumetric flask and made up to the mark. The determination proceeded analogously to total nitrites (Section 2.3.2) using 10 mL of the test solution. Results of nitrate determination were expressed as mg/g of NO3 and mg/daily dose of NO3 or mg/g d.w. of NO3 and mg/100 g f.w. of NO3 of beetroot, which was calculated as follows:(3) (NO3-)=1.348(m2x10 000V3xV2xm0-m1x200V1xm0) m2--the total mass of nitrite, in micrograms of nitrite ion (NO2 ), contained in the volume (V2) of test solution taken, read from the calibration graph; V2--the volume, in millilitres, of the test solution taken for the spectrometric measurement; V3--the volume, in millilitres, of the aliquot portion of the filtrate taken for the preparation of the test solution; m0, m1, V1--have the same meanings as in Equation (2). The ratio between the relative molecular masses of the nitrate ion (NO3 ) and nitrite ion (NO2 ) is 1.348. 2.4. Validation The following validation parameters were determined for all methods: linearity range, precision, accuracy, the limit of determination (LOD), and the limit of quantification (LOQ). The LOD and LOQ were computed as described by Huber : (4) LOD=3.3SDab SDa--standard deviation of the intercept for the calibration curve; b--slope for the calibration curve. (5) LOD=3xLOD Table 1 shows the results of the validation. Due to the lack of reference material corresponding to the analysed material, accuracy was determined using the method of standard addition (GA in the FC assay and Trolox in the CUPRAC assay) to the chosen DS and lyophilizate and was expressed as recovery. DPPH assay was validated based on gallic acid standard solutions. The average recovery for the selected methods was in the range of 80-120% which was an acceptable level for such an analysis. The precision was computed as the coefficient of variation for all the results obtained in all the analysed samples during validation. The signal obtained for standards (Sexpected) and the signal calculated from the calibration equation (Scalculated) were applied for the calculation of recovery for calibration curves (Rcc): (6) Rcc= 2.5. Labelling Assessment Thirty-four packages of DSs and eight traditional food products were assessed. Before evaluating packaging and labelling, each product's registration in the register of products subject to notification of first market placement was checked . As a result, eight products could not be included in the "dietary supplement" category due to the lack of appropriate labelling on the packaging and registration with the Chief Sanitary Inspectorate (GIS). Requirements on food and DS labelling are specified in Regulation (EU) No 1169/2011 and the Act on Food and Nutrition Safety of 25 August 2006 . As food, DSs have been assessed because of the requirements specified in the Regulation of the Minister of Health of 9 October 2007 . The correctness of the labelling was assessed according to the following criteria : Labelling in Polish; The name of the food; The list of ingredients; The net amount of food; The date of minimum durability or best-before date; The presence of the term "dietary supplement"; Indication of the recommended daily portion of the product; The presence of a warning regarding not exceeding the recommended daily portion; A statement that dietary supplements cannot be used as a substitute for a varied diet; A statement that they should be kept out of the reach of small children. In addition, the manufacturer should provide information on the content of active ingredients per recommended daily portion and information on the content of vitamins and minerals in percentages concerning the reference daily intake (RDI). Particular attention was paid to the health claims on the packaging of the tested products, which were compared with the list of permitted health claims defined in Regulation No. 1924/2006 and with the statements contained in the register of the European Food Safety Authority . The difference in the number of analysed packages versus the total number of analysed DSs is due to the fact that some products were purchased in multiple repetitions (different lot numbers), resulting in the same package design. 2.6. Statistical Analyses The data were reported as the mean +- standard deviation of three independent samples, each measured three times. Statistical analyses such as the ANOVA Kruskal-Wallis test, the U Mann-Whitney test, or Spearman's rank correlation coefficient preceded by an analysis of the normality (the Shapiro-Wilk test) of the distribution were used to compare the treatments. They were performed by the Statistica for Windows (version 13, Statsoft, Cracow, Poland) software package. Differences at p < 0.05 were deemed significant. The validation parameters for spectrophotometric assays, the overall mean, and the standard error values were calculated using the Microsoft Office Excel software (version 2007 12.0.6787.5000 SP3 MSO, Microsoft Corporation, Redmond, WA, USA). 3. Results 3.1. Total Phenolic Content and the TAC The averaged values of TAC (CUPRAC, DPPH) and TPC (FC) in the analysed beetroot and DSs, divided into groups (tablets, capsules, powders), are shown in Table 2 and Table 3, respectively. For the FC, results are expressed in gallic acid equivalents (GAE), for CUPRAC in Trolox equivalents (TE), and for DPPH as a percentage reduction in DPPH. Tables S6 and S7 show the full characteristics of the beetroot and beetroot-based DSs studied due to their TAC, TPC, nitrate, and nitrite contents. Powders were characterised by significantly higher TPC (mg GAE/d. d.; Table S8) than tablets (U Mann-Whitney test, p = 9.9 x 10-5) and capsules (U Mann-Whitney test, p = 4.4 x 10-5). However, lyophilizates showed the highest TPC compared to any group of DSs. There was no statistically significant difference between tablets and capsules (U Mann-Whitney test, p = 0.99). In the case of TPC per gram of a product, a statistically significant difference was found only between tablets and lyophilizates (U Mann-Whitney test, p = 0.0049), which could be caused by the presence of excipients in tablets used in tabletting processes. In the FC assay, the product marked as T7 (41 mg GAE/d. d.) was characterised by the highest TPC among the tablets, C8A (42 mg GAE/d. d.) and C8B (41 mg GAE/d. d.) among the capsules, and product P9 (251 mg GAE/d. d.) among the powders. Powders showed higher TAC than tablets (U Mann-Whitney test, p = 1.6 x 10-4) and capsules (U Mann-Whitney test, p = 2.3 x 10-5). Lyophilizates provided a higher TAC than DSs. Considering TAC expressed as TE/g, lyophilizates showed significantly more antioxidants than all DS formulations. The highest TAC among the tablets was found in the T7 product (350 mmol TE/d. d.), the C13 (363 mmol TE/d. d.) among capsules, and the P9 product (3520 mmol TE/d. d.) among the powders. It is worth noting that product P9 exhibited a higher TPC (251 mg GAE/d. d.) and TAC (3520 mmol TE/d. d.) than average beetroots (Table 2). Different trends between the FC and CUPRAC methods may result from the variability of the conditions under which the tests were conducted and the reaction mechanisms. In the CUPRAC method, for example, the antioxidant potential is tested at pH = 7, which is close to the pH of human blood, as opposed to the FC method, which tests at pH 8-9. These changes in pH can influence the development of various antioxidant capacities of products, especially considering the complex matrix of DSs. In addition, the reaction with the DPPH radical is specific to individual antioxidants; they can react at different rates. Despite the concentrated DS formula as dried material, the daily portion of fresh beetroot (100 g f.w.) was richer in antioxidants. In the DPPH method, the highest TAC was shown by the product T7 (90%) among the tablets and the products C6 and C8B among the capsules--the activities of which were both 90%--while the activity of the product C8A was 74%. Among the powders, the product P11 (90%) showed the highest TAC. The content of antioxidants in the dietary supplements T2A, T2B, and T2C was greatly varied, despite the use of the standard material standardisation procedure. Moreover, all three supplements were sold as the same commercial product. Three other products were characterised by a high RSD (30% for T6; 41% for T10; and 33% for C2). In all cases, the analysis was repeated, but the analogous results were obtained, and the Q-Dixon test did not show a significant error. A high RSD could be caused by the heterogeneity of the supplement, as the method has been validated and the analysis conditions have not changed. In a study conducted by Guldiken et al. , in which the content of antioxidants was measured using colorimetric methods in fresh beetroot, the following results were obtained: 255 mg GAE/100 g of fresh weight in the FC method and 15,538 mmol TE/100 g (3889 mg TE/100 g) in the CUPRAC method. The values obtained by the FC method are comparable to those obtained in this work for the majority of whole and peeled beetroot lyophilizates (Table S7). However, the values obtained by the CUPRAC method in this work are lower for most samples. Only the samples of skins 6Sk and 7Sk, which were from organic farming, can be considered comparable (9779 and 8932 mmol TE/100 g, respectively). This may be due to the differences in the profile of compounds and antioxidant capacity between different varieties and the freshness of the material analysed. In this study, the vegetables were processed and freeze-dried immediately after purchase. However, there is no way to trace the storage conditions of fresh material before purchase. There is a lack of reports in the literature regarding the TPC and TAC of DSs made from beetroot. Comparing the results obtained for supplements per gram in capsules (0.68 to 33 mg GAE/g), tablets (2.0 to 41 mg GAE/g) or powders (4-61 mg GAE/g) with the results obtained by Guldiken et al. for dried beetroot 3.3 mg GAE/g (347 mg GAE/100 g), there can be observed a lower TPC in dried beetroot compared to our supplements calculated per g of d.w. Moreover, Spearman's rank correlation coefficient was performed to test for a potential correlation between the antioxidant potential results obtained by different methods. A fairly strong relationship (0.7-0.9) or a very strong relationship (>0.9) was observed between the results obtained by the FC, CUPRAC, and DPPH methods in all the groups analysed (beetroots, capsules, tablets, and powders) (Table S9). In the study by Apak et al. , the correlation between the CUPRAC and FC methods was comparable and amounted to r = 0.966. Another study by Guclu et al. showed a high correlation (r = 0.93) between the FC and CUPRAC methods . 3.2. Nitrate and Nitrite Content In general, significantly lower levels of nitrite ions (0.21-78 mg/d. d.) than nitrate ions (0.197-169 mg/d. d.) were found in DS samples. Due to the lack of a normal distribution (the Shapiro-Wilk test, p < 0.05), the U Mann-Whitney test was applied to check for statistically significant differences. It was found that supplements in tablets (p = 0.000295) and capsules (p = 0.014038) contained significantly fewer nitrite ions, as well as supplements in tablets (p = 0.262612) had statistically significantly fewer nitrate ions than lyophilized vegetables, considering their content in 1 g of product (dry weight for beetroot). The nitrite ion content of powders and beetroots did not differ significantly, nor did the nitrate ion content of beetroots, capsules, and powders. Moreover, individual parts of beetroot (peeled beetroot, skins) did not differ significantly in the content of nitrites (p = 0.133615) and nitrates (p = 0.830324) (the U Mann-Whitney test, p < 0.05). All results of the U Mann-Whitney test were shown in Table S10. An average portion of conventional beetroots provided more nitrites (49 mg/100 g f.w.) and nitrates (90 mg/100 g f.w., Table 3) than most of the other products analysed. Only products P9 (169 mg/d.d.), P10 (99 mg/d.d.), and P13 (131 mg/d.d.) contained more nitrates than beetroots. The highest content of nitrite ions was found in supplement number C4 (8.4 mg/g) in the case of capsules, T3A (2.48 mg/g) for tablets, and P13 (6.4 mg/g) for powders. The highest level of nitrate ions was found in T11 (3746 mg/kg), C16 (15,186 mg/kg), C7B (11,924 mg/kg), P13 (13,110 mg/kg), and P9 (10,224 mg/kg) DSs. In all tested vegetable samples, a significantly lower content of nitrite ions (0.702 mg/g-15 mg/g) than nitrate ions (423 mg/kg-8801 mg/kg) was determined. The highest content of nitrite ions among vegetables was found in skin samples, except for group 7, where the highest level of these ions was determined in a sample of peeled beetroot. In the case of nitrate ions, the situation was the opposite: the skin samples were characterised by the lowest content of these ions, except for sample no. 1, where their level was the highest in the batch. Moreover, individual subgroups of beetroot (peeled, unpeeled, skins) differed in terms of TAC and TPC, regardless of the method used to assess the potential (ANOVA Kruskal-Wallis test: FC p = 0.0022, CUPRAC p = 0.0016, DPPH p = 0.006). The skins were the richest in antioxidants (Dunn's test, p < 0.05). There were no statistically significant differences in the content of nitrites and nitrates between the individual subgroups (ANOVA, p > 0.05). For the comparison of vegetable supplements, reference was made to the average values for conventional and organic beetroots (Table 2). Manufacturers usually do not provide information on how the beetroot has been processed before preparing supplements from it. Several products contained information that whole beetroot was used, which is why we also included vegetables with skin in our analysis. The content of these compounds in beetroot depends primarily on the amount of nitrogen fertilization, agrotechnical treatments, and the plant growth phase . Goscinna and Czapski observed higher contents of nitrates in the middle parts of the root compared to its outer parts. Although they used a different division of the tuber (into 4 parts), it can be considered that the conclusions from our study and their research are consistent-beetroot skin is characterised by a lower content of nitrates. Only four DS had a nitrate content declaration. Supplements C10 and C11 did not contain nitrates (<LOQ) despite their presence being declared by the manufacturer. Both products were produced by the same manufacturer but were available under different trade names and with different graphic designs. Product C1 contained a negligible amount of nitrates compared to the declaration (4.2%). Product P9 contained nitrates in amounts close to the declared one (85%). Simultaneously, it is the product that contains the most nitrates per daily portion of all the tested foods, as well as more than the average portion of fresh beetroot. In the years 2003-2004, research on the content of nitrites and nitrates was carried out on certain vegetables purchased in random shops in Olsztyn . Among these vegetables, beetroot was included, which was classified as a plant with a high content of nitrate--an average of 1408.17 mg/kg. A high level of nitrates (III) was determined in the analysed beetroots (on average 11.4 mg/kg), which differed from the average values for this vegetable. The content of nitrite and nitrate ions in the beetroot samples in this study was 0.120-2.935 mg/kg for nitrites and 102.30-1619.80 mg/kg for nitrates, respectively. Health Risk Assessment In terms of nitrite content (2.1% ADI for NO2 ), none of the products tested posed a risk (Table 4). Fresh beetroot (100 g) provided more NO3 (15-20.1% ADI for NO3 , conventional and organic, respectively) than any of the analysed DSs in the form of tablets (3.2% ADI for NO3 ) or capsules (5.1% ADI for NO3 ). DSs in powders provided a similar amount of the substance as a serving of beetroot (based on the realisation of the ADI). Product P9, marketed by the manufacturer as having an "increased dose of nitrates", had the highest nitrate dose (48% of the ADI for NO3 ) and the lowest nitrite levels (0.21% of the ADI for NO2 ). That is why the manufacturer advertised it as a product for athletes to be consumed before and after training "to increase the body's efficiency, accelerate regeneration after training, and reduce accumulated lactic acid" (information on the packaging). In comparison, the recommended daily dose of products in capsules provided a maximum of 3.2% ADI for NO3 and 5.1% ADI for NO3 in tablets. Keller et al. analysed eighteen DSs in terms of the content of nitrites and nitrate and determined the percentage of ADI to evaluate the exposure to these compounds through the intake of the recommended portion. The ADI for nitrate amounted to 22% in the case of the Neo40 supplement, described by the manufacturer as "improving the functioning of the cardiovascular system", and 97% in BeetElite--advertised as "improving exercise endurance and increasing oxygen supply in the body. The ADI for nitrites amounted to 450% and 225%, respectively. The rest of the DSs were characterised by values within the range of 0.01-47.26% for nitrates and 0.00-21.11% for nitrites . The reason for such a high content of nitrites and nitrates in some supplements was most likely their composition--rich in nitrates dehydrated or concentrated forms of vegetables or concentrated vegetable juices. The supplements analysed for this work did not pose a risk to the consumer because of the ADI. Studies have shown that consumption of beetroot products is more beneficial than supplementation with nitrate salts because the flavonoids and vitamin C present in beetroot reduce the risk of nitrosamine formation. Haem iron may increase the risk of the formation of these compounds . However, beetroot supplements, if fortified with iron, are in the non-haem form (usually iron gluconate or fumarate; see T2, T5, and T6). Because of ADI limits, all sources of nitrates in the diet should be considered when estimating daily nitrate and nitrite consumption and performing a safety assessment. Green leafy vegetables and root vegetables such as beetroot constitute rich sources of these substances . Moreover, nitrates are added to meat and meat products to prevent Clostridium botulinum, Listeria monocytogenes, Bacillus cereus, Clostridium perfringens, and Staphylococcus aureus growth, improve their colour, and develop their characteristic flavour . According to some studies, their high concentrations in water (>50 mg/L) can cause methemoglobinemia and gastrointestinal carcinogenesis . It should be mentioned that hypotensive effects were observed after the nitrate dose corresponding with the upper limit of the WHO ADI. Ashor et al. described the hypotensive effect after using beetroot juice rich in NO3 (70 mL containing 400 mg). Similarly, Mills et al. discovered a hypotensive effect after 6 months of drinking beetroot juice rich in NO3 (70 mL containing 694 mg of NO3 ). Furthermore, Kapil et al. reported that 4 weeks of supplementation with beetroot juice containing 450 mg of NO3 had beneficial therapeutic effects on endothelium and arterial stiffness. However, no therapeutic effects were observed with additional daily administration of 300 mg of NO3 . Considering these values, the tested supplements are probably not able to exert a hypotensive effect even with long-term use, as the highest content of nitrates found in product P9 amounted to 48% of ADI. 3.3. The Correlation between the Antioxidant Potential and the Content of Nitrites and Nitrates in Beetroot and Beetroot-Based Products A statistically significant negative correlation (Spearman's rank correlation analysis) was found only in beetroot samples, both between the results of the antioxidant potential obtained by the FC and CUPRAC methods and the content of nitrates (Table S11). The group of beetroot samples was more homogeneous in terms of composition than DSs. Some of the DSs were enriched in various substances such as nitrates, iron compounds, and vitamin C, which could have disturbed the existence of a potential correlation. The Spearman's rank correlation coefficient equals -0.54 for the FC method and -0.62 for CUPRAC, which means that dependence is moderate. 3.4. Labelling Assessment An assessment of thirty-four packages of dietary supplements and eight traditional food products was carried out because of Polish and European food labelling regulations. It was found that 64% of packaging did not meet the legal requirements for food labelling, 12% were not reported to the Chief Sanitary Inspectorate, and 6% did not have the term "dietary supplement" on the packaging, despite having registration in the GIS in this category. Furthermore, 26% of products were not fully labelled in Polish, as a result of which the consumer is not able to get acquainted with the information presented on the packaging in detail and its content is not formulated understandably. It is worth emphasising that 21% of the tested products contained prohibited health or non-registered claims, which means that 15% of the products suggested that they had the properties of preventing or treating diseases. It is also worth noting that the product P9, which contained the highest amount of nitrates and was sold as a supplement, contained significant labelling deficiencies, including the lack of the wording "dietary supplement". Moreover, part of the labelling was only in English, and the dosage was not precisely defined, which poses a risk of nitrate overdose. The detailed results of the analysis are summarised in Table 5. 4. Conclusions The importance of this study was to determine and compare TAC, TPC, nitrite, and nitrate content in beetroot-based DSs and beetroot. Moreover, the safety of consumption of DSs because of nitrites, nitrates, and the correctness of labelling were assessed. The research revealed that TAC, TPC, nitrate, and nitrite concentrations expressed per unit of product weight (g or kg), DSs in capsules, and DSs in powders were comparable to the average beetroot. Tablets contained notably fewer of these substances, which might result from the presence of auxiliary substances used for tabletting. However, the average portion (100 g) of conventional or organic beetroot provided significantly more nitrates, nitrites, and substances with antioxidant properties than most of the DSs in capsules, tablets, and powders dosed according to the manufacturer's recommendations. Only P9 (48% ADI for NO3 ), P10 (28% ADI for NO3 ), and P13 (37% ADI for NO3 ) delivered higher doses than beetroots. Most of the products did not have the declared content of nitrates. The antioxidant content in a serving of tablets or capsules was negligible, so their use has a low health value. In many of the samples studied, the nitrate content was not correlated to the antioxidant potential. A statistically significant negative correlation was found only in beetroot samples between the results of the FC and CUPRAC methods and the content of nitrates. The labelling assessment has shown that 64% of packaging did not meet the legal requirements for food labelling. Some DSs contained illegal health claims that suggested healing properties or were misleading. This situation might result in reduced effectiveness or withdrawal from conventional therapies by consumers who would choose adulterated DSs. There were significant deficiencies in labelling, including a lack of full labelling in Polish, unclear dosage and others. Such deficiencies, combined with unknown nitrite and nitrate content, may result in consumers overdosing on these substances as a result of incorrect product intake. In addition, the unknown content of nitrates and nitrites may pose a threat to the consumer because the content of these compounds in vegetables varies depending on the place of origin and growing conditions. The conducted research indicated a strong need for more rigorous control when launching DSs, both in terms of composition and labelling. Although production and labelling guidelines are published, there is a lack of decisive action by national and European authorities related to the control and withdrawal of defective products. Supplementary Materials The following supporting information can be downloaded at: Table S1. Collected information about the analysed beetroot samples. Table S2. Collected information about the analysed beetroot-based DSs. Table S3. Optimisation of extraction of beetroot products on the example of a selected lyophilizate. Table S4. Optimisation of the FC method because of model and incubation time. Table S5. Optimisation of the FC method because of incubation time for lyophilizate and supplement in tablets according to method 1. Table S6. Full characteristics of the analysed beetroot-based dietary supplements because of TAC, nitrate, and nitrite content. Table S7. Full characteristics of the analysed beetroots samples because of TAC, nitrate, and nitrite content. Table S8. The results of the Mann-Whitney U test check the existence of differences between the individual groups of the analysed products (values p < 0.05 are marked in red). Table S9. The results of the correlation between the TAC and TPC results obtained by different methods of beetroot and beetroot-based products (Spearman's rank correlation coefficient). Values which were statistically significant (p < 0.05) are marked in red. Table S10. Results of U Mann Whitney test for all analysed samples because of nitrites and nitrates content [mg/kg] (values p < 0.05 are marked in red). Table S11. The results of the correlation between the antioxidant potential and the content of nitrites and nitrate in beetroot and beetroot-based products (Spearman's rank correlation coefficient). Values which were statistically significant (p < 0.05) are marked in red. Click here for additional data file. Author Contributions Conceptualization, J.B.-R. and M.G.; methodology, J.B.-R. and M.G.; software, J.B.-R.; validation, J.B.-R. and S.S.; formal analysis, J.B.-R.; investigation, J.B.-R., S.S., P.M. and K.G.; resources, J.B.-R. and M.G.; data curation, J.B.-R.; writing--original draft preparation, J.B.-R. and S.S.; writing--review and editing, M.P. and M.G.; visualization, J.B.-R.; supervision, M.P. and M.G.; project administration, J.B.-R. and M.G.; funding acquisition, J.B.-R., M.P. and M.G. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Data is contained within the article or supplementary material. Conflicts of Interest The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Figure 1 Scheme of collecting and preparation of samples. Figure 2 Diagram showing the optimisation of the samples' extraction (a) and the Folin-Ciocalteu method (b). The selected conditions are marked in yellow in the figure. Figure 3 The procedure of extraction for TAC and TPC determination. foods-12-01017-t001_Table 1 Table 1 The validation parameters of the applied methods. CUPRAC FC DPPH Nitrates and Nitrites Standard substance Trolox (TE) gallic acid (GA) gallic acid (GA) sodium nitrite Calibration curve equation y = 17.3x + 0.00234 y = 0.131x + 0.000874 y = -0.481x + 0.786 y = 1.0973x + 0.0037 The determination coefficient (R2) 0.9993 0.9991 0.9991 0.9998 Linearity range 0.0005-0.070 mmol/mL 0.10-10 mg/mL 0.13-1 mg/mL 0.0027-0.6 mg/mL LOD 0.000187 mmol/mL 0.074 mg/mL 0.040 mg/mL 0.009 mg/mL LOQ 0.000562 mmol/mL 0.22 mg/mL 0.12 mg/mL 0.0027 mg/mL supplement lyophilizate supplement lyophilizate gallic acid supplement lyophilizate Precision 1.7-3.1% 0.34-1.4% 1.1-3.9% 1.1-5.5% 2.0-4.6% I stage: 1.26-1.97% II stage: 4.73-4.91% I stage: 0.22-4.28% II stage: 2.11-4.95% Recovery 95-115% 99-109% 96-106% 86-105% 91-113% I stage: 86.17-93.47%II stage: 99.06-104.07% I stage: 80.77-94.08%II stage: 93.99-104.72% foods-12-01017-t002_Table 2 Table 2 Results of TAC, TPC, nitrite, and nitrate content in analysed conventional and organic beetroots. Beetroot Method Unit Conventional Organic n Mean SD Min Median Max n Mean SD Min Median Max FC mg GAE/g d.w. 12 15 0.702 6.8 12 26 8 14 0.47 9.7 13 34 CUPRAC mmol TE/g d.w. 12 171 11 93 150 365 8 196 7.1 123 179 413 Nitrites mg/g d.w. 12 2.4 0.25 0.702 1.8 7.1 8 5.1 0.298 1.96 3.7 15 Nitrates mg/kg d.w. 12 4980 111 2101 4912 8801 8 2612 88 423 2509 6606 FC mg GAE/100 g f.w. 12 211 16 91 218 595 8 299 13 170 279 794 CUPRAC mmol TE/100 g f.w. 12 2743 233 1320 2884 7988 8 4064 192 2464 3905 9779 Nitrites mg/100 g f.w. 12 49 6.9 12 38 196 8 109 7.5 34 81 293 Nitrates mg/100 g f.w. 12 90 1.97 45 90 153 8 54 1.997 11 49 134 DPPH % 12 42 1.7 28 40 63 8 43 1.3 31 41 64 SD--standard deviation, Min--minimum, Max--maximum, n--number of products where analytes were determined >LOQ, d.w.--dry weight, f.w.--fresh weight, GAE--gallic acid equivalent, TE--Trolox equivalent. foods-12-01017-t003_Table 3 Table 3 Results of TAC, TPC, nitrite, and nitrate content in the analysed DSs expressed per mass unit (g or kg) of product and daily dose (d. d.). Beetroot-Based Dietary Supplements Method Unit Capsules Tablets Powders n Mean SD Min Median Max n Mean SD Min Median Max n Mean SD Min Median Max FC mg GAE/g 21 14 1.2 1.8 6.2 41 16 7.8 0.77 0.68 3.4 33 12 13 0.68 2.5 10 61 CUPRAC mmol TE/g 21 126 11 13 62 312 16 76 3.5 1.3 42 278 12 139 5.9 27 119 467 Nitrites mg/g 21 2.3 0.21 0.81 1.2 8.9 16 1.3 0.079 0.29 1.2 2.5 12 2.9 0.25 0.95 2.4 6.36 Nitrates mg/kg 21 4230 155 383 2373 15,186 16 2099 67 504 1979 3746 12 4161 241 91 2265 13,110 FC mg GAE/d. d. 21 15 1.3 0.73 8.5 42 16 17 1.4 1.5 18 41 12 99 9.6 4.1 92 251 CUPRAC mmol TE/d. d. 21 138 10 5.2 103 363 16 167 8.1 5.5 174 350 12 1228 97 48 1065 3520 Nitrites mg/d. d. 21 3.3 0.56 0.33 1.85 23.78 16 4.2 0.36 0.21 2.6 14 12 28 3.4 1.7 18 78 Nitrates mg/d. d. 21 6.1 0.26 0.20 3.9 18 16 4.6 0.35 1.3 3.8 11 12 43 2.7 0.90 22 169 DPPH % 21 39 1.67 7.2 22 90 16 35 2.4 4.9 17 90 12 37 1.3 12 37 90 SD--standard deviation, Min--minimum, Max--maximum, n--number of products where analytes were determined >LOQ, GAE--gallic acid equivalent, TE--Trolox equivalent, d. d.--daily dose recommended by the manufacturer. foods-12-01017-t004_Table 4 Table 4 The realisation of ADI for NO2 and NO3 by DSs and beetroots. %ADI for NO2 %ADI for NO3 Product n Mean SD Min Median Max n Mean SD Min Median Max DSs and food products Tablets 14 0.030 0.030 0.0015 0.018 0.098 15 1.3 0.90 0.38 1.1 3.2 Capsules 18 0.023 0.039 0.0023 0.012 0.17 17 1.7 1.5 0.056 0.87 5.1 Powders 11 0.203 0.18 0.012 0.13 0.55 11 12 16 0.26 6.2 48 Beetroot Conventional 12 0.35 0.35 0.087 0.27 1.4 12 20.1 13 1.84 16.1 42.7 Organic 8 0.78 0.62 0.24 0.58 2.1 8 15 11 3.02 14 38 ADI for NO2 equals 0.2 mg of NO2 /kg/day (14 mg of NO2 /70 kg/day); ADI for NO3 equals 5 mg of NO3 /kg/day (350 mg of NO2 /70 kg/day); n--number of the analysed products with the determined content of analytes >LOQ. foods-12-01017-t005_Table 5 Table 5 Summary of the labelling assessment of the analysed DSs. The Analysed Feature of Product Marking Results Registration in the register of products is subject to notification of first placing on the market 88% of analysed products were reported to the Register and 12% did not Labelling in Polish 26% of products had a complete or partial lack of markings in Polish List of ingredients 5% of the packaging was missing the word "ingredients" before the list of ingredients The net amount of food 26% of packages did not declare the net weight of the product Date of minimum durability or best-before date 23% of manufacturers used incorrect wording preceding the date of minimum durability in the labelling The presence of the term "dietary supplement" 6% of products were not marked with the term "dietary supplement" Indication of the recommended daily portion 3% of products did not have the recommended daily portion for consumption specified The presence of a warning regarding not exceeding the recommended daily portion 12% of packages lacked such a warning The statement that DSs cannot be used as a substitute (replacement) for a varied diet 12% did not include this statement The statement that DSs should be kept out of the reach of small children 9% did not include this statement The content of vitamins and minerals and other substances with nutritional or other physiological effects present in the dietary supplement in numerical form, calculated per recommended daily portion of the product 21% of the packaging did not contain the content of vitamins, minerals and other substances per recommended daily portion The information on the content of vitamins and minerals in percentage concerning the reference daily intake (RDI) 3% of the packaging did not contain information on the content of vitamins and minerals in percentage concerning RDI Labelling may not suggest that the food has effects or properties that it does not have, or attribute to the food the property of preventing or treating disease 15% of the products suggested that the food had the properties of preventing or treating diseases Health claims 21% of products contained health claims on the packaging that were not included in the EFSA Health Claims Register or were not allowed to be used due to a lack of scientific evidence Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Grembecka M. Szefer P. Dybek K. Gurzynska A. Ocena zawartosci wybranych biopierwiastkow w warzywach Rocz. Panstwowego Zakladu Hig. 2008 59 179 186 2. Fu Y. Shi J. Xie S.Y. Zhang T.Y. Soladoye O.P. Aluko R.E. Red beetroot betalains: Perspectives on extraction, processing, and potential health benefits J. Agric. Food Chem. 2020 68 11595 11611 10.1021/acs.jafc.0c04241 33040529 3. Czapski J. Mikolajczyk K. Kaczmarek M. Relationship between antioxidant capacity of red beet juice and contents of its betalain pigments Polish J. Food Nutr. Sci. 2009 59 119 122 4. Chhikara N. Kushwaha K. Sharma P. Gat Y. Panghal A. Bioactive compounds of beetroot and utilization in food processing industry: A critical review Food Chem. 2019 272 192 200 10.1016/j.foodchem.2018.08.022 30309532 5. van Velzen A.G. Sips A.J.A.M. Schothorst R.C. Lambers A.C. Meulenbelt J. The oral bioavailability of nitrate from nitrate-rich vegetables in humans Toxicol. Lett. 2008 181 177 181 10.1016/j.toxlet.2008.07.019 18723086 6. Bonilla Ocampo D.A. Paipilla A.F. Marin E. Vargas-Molina S. Petro J.L. Perez-Idarraga A. Dietary nitrate from beetroot juice for hypertension: A systematic review Biomolecules 2018 8 134 10.3390/biom8040134 30400267 7. Chawla H. Parle M. Sharma K. Yadav M. Beetroot: A health promoting functional food Inven. Rapid Nutraceuticals 2016 1 0976 3872 8. Clifford T. Howatson G. West D.J. Stevenson E.J. The potential benefits of red beetroot supplementation in health and disease Nutrients 2015 7 2801 2822 10.3390/nu7042801 25875121 9. Gilchrist M. Winyard P.G. Fulford J. Anning C. Shore A.C. Benjamin N. Dietary nitrate supplementation improves reaction time in type 2 diabetes: Development and application of a novel nitrate-depleted beetroot juice placebo Nitric Oxide-Biol. Chem. 2014 40 67 74 10.1016/j.niox.2014.05.003 10. Shepherd A.I. Gilchrist M. Winyard P.G. Jones A.M. Hallmann E. Kazimierczak R. Rembialkowska E. Benjamin N. Shore A.C. Wilkerson D.P. Effects of dietary nitrate supplementation on the oxygen cost of exercise and walking performance in individuals with type 2 diabetes: A randomized, double-blind, placebo-controlled crossover trial Free Radic. Biol. Med. 2015 86 200 208 10.1016/j.freeradbiomed.2015.05.014 25998421 11. Bock J.M. Ueda K. Feider A.J. Hanada S. Casey D.P. Combined inorganic nitrate/nitrite supplementation blunts a-mediated vasoconstriction during exercise in patients with type 2 diabetes Nitric Oxide 2022 118 17 25 10.1016/j.niox.2021.10.006 34718145 12. Lundberg J.O. Feelisch M. Bjorne H. Jansson E.A. Weitzberg E. Cardioprotective effects of vegetables: Is nitrate the answer? Nitric Oxide-Biol. Chem. 2006 15 359 362 10.1016/j.niox.2006.01.013 16563818 13. Hughes W.E. Treichler D.P. Ueda K. Bock J.M. Casey D.P. Sodium nitrate supplementation improves blood pressure reactivity in patients with peripheral artery disease Nutr. Metab. Cardiovasc. Dis. 2022 32 710 714 10.1016/j.numecd.2021.12.002 35090799 14. Lotfi M. Azizi M. Tahmasbi W. Bashiri P. The effects of consuming 6 weeks of beetroot juice (Beta vulgaris L.) on hematological parameters in female soccer players J. Kermanshah Univ. Med. Sci. 2018 22 e82300 10.5812/jkums.82300 15. Ranchal-Sanchez A. Diaz-Bernier V.M. De La Florida-Villagran C.A. Llorente-Cantarero F.J. Campos-Perez J. Jurado-Castro J.M. Acute effects of beetroot juice supplements on resistance training: A randomized double-blind crossover Nutrients 2020 12 1912 10.3390/nu12071912 32605284 16. Garnacho-Castano M.V. Palau-Salva G. Serra-Paya N. Ruiz-Hermosel M. Berbell M. Vinals X. Bataller M.G. Carbonell T. Vilches-Saez S. Cobo E.P. Understanding the effects of beetroot juice intake on CrossFit performance by assessing hormonal, metabolic and mechanical response: A randomized, double-blind, crossover design J. Int. Soc. Sports Nutr. 2020 17 56 10.1186/s12970-020-00388-z 33187518 17. Broeder C.E. Flores V. Julian B. Wojan F. Tauber R. Schubert L. Salacinski A. Ivy J.L. Nitric oxide enhancement supplement containing beet nitrite and nitrate benefits high intensity cycle interval training Curr. Res. Physiol. 2021 4 183 191 10.1016/j.crphys.2021.07.002 34746837 18. Mortensen A. Aguilar F. Crebelli R. Di Domenico A. Dusemund B. Frutos M.J. Galtier P. Gott D. Gundert-Remy U. Lambre C. Re-evaluation of potassium nitrite (E 249) and sodium nitrite (E 250) as food additives EFSA J. 2017 15 e04786 10.2903/j.efsa.2017.4786 32625504 19. Mortensen A. Aguilar F. Crebelli R. Di Domenico A. Dusemund B. Frutos M.J. Galtier P. Gott D. Gundert-Remy U. Lambre C. Re-evaluation of sodium nitrate (E 251) and potassium nitrate (E 252) as food additives EFSA J. 2017 15 e04787 10.2903/j.efsa.2017.4787 32625505 20. Guldiken B. Toydemir G. Nur Memis K. Okur S. Boyacioglu D. Capanoglu E. Home-processed red beetroot (Beta vulgaris L.) products: Changes in antioxidant properties and bioaccessibility Int. J. Mol. Sci. 2016 17 858 10.3390/ijms17060858 27258265 21. Sadowska-Bartosz I. Bartosz G. Evaluation of the antioxidant capacity of food products: Methods, applications and limitations Processes 2022 10 2031 10.3390/pr10102031 22. Bartosz G. Total antioxidant capacity Adv. Clin. Chem. 2003 37 219 292 12619709 23. Capanoglu E. Beekwilder J. Boyacioglu D. Hall R. De Vos R. Changes in antioxidant and metabolite profiles during production of tomato paste J. Agric. Food Chem. 2008 56 964 973 10.1021/jf072990e 18205308 24. Sawicki T. Baczek N. Wiczkowski W. Betalain profile, content and antioxidant capacity of red beetroot dependent on the genotype and root part J. Funct. Foods 2016 27 249 261 10.1016/j.jff.2016.09.004 25. Kujala T. Loponen J. Pihlaja K. Betalains and phenolics in red beetroot (Beta vulgaris) peel extracts: Extraction and characterisation Zeitschrift Fur Naturforsch.-Sect. C J. Biosci. 2001 56 343 348 10.1515/znc-2001-5-604 26. Delgado-Vargas F. Jimenez A.R. Paredes-Lopez O. Natural pigments: Carotenoids, anthocyanins, and betalains--Characteristics, biosynthesis, processing, and stability Crit. Rev. Food Sci. Nutr. 2000 40 173 289 10.1080/10408690091189257 10850526 27. Azeredo H.M.C. Betalains: Properties, sources, applications, and stability-a review Int. J. Food Sci. Technol. 2009 44 2365 2376 10.1111/j.1365-2621.2007.01668.x 28. Strack D. Vogt T. Schliemann W. Recent advances in betalain research Phytochemistry 2003 62 247 269 10.1016/S0031-9422(02)00564-2 12620337 29. Kavalcova P. Bystricka J. Tomas J. Karovicova J. Kovarovic J. Lenkova M. The content of total polyphenols and antioxidant activity in red beetroot Potravinarstvo 2015 9 77 83 10.5219/441 30. Kujala T.S. Loponen J.M. Klika K.D. Pihlaja K. Phenolics and betacyanins in red beetroot (Beta vulgaris) root: Distribution and effect of cold storage on the content of total phenolics and three individual compounds J. Agric. Food Chem. 2000 48 5338 5342 10.1021/jf000523q 11087483 31. Sentkowska A. Pyrzynska K. Determination of selenium species in beetroot juices Heliyon 2020 6 e04194 10.1016/j.heliyon.2020.e04194 32566793 32. Lamuela-Raventos R.M. Folin-Ciocalteu method for the measurement of total phenolic content and antioxidant capacity Measurement of Antioxidant Activity & Capacity: Recent Trends and Applications 1st ed. Apak R. Capanoglu E. Shahidi F. John Wiley & Sons Ltd. Oxford, UK 2017 Volume 1 107 115 10.1002/9781119135388.ch6 33. Agbor G.A. Vinson J.A. Donnelly P.E. Folin-Ciocalteau reagent for polyphenolic assay Int. J. Food Sci. Nutr. Diet. 2014 3 147 156 10.19070/2326-3350-1400028 34. Musci M. Yao S. Optimization and validation of Folin-Ciocalteu method for the determination of total polyphenol content of Pu-erh tea Int. J. Food Sci. Nutr. 2017 68 913 918 10.1080/09637486.2017.1311844 28399674 35. Apak R. Guclu K. Ozyurek M. Karademir S.E. Novel total antioxidant capacity index for dietary polyphenols and vitamins C and E, using their cupric ion reducing capability in the presence of neocuproine: CUPRAC method J. Agric. Food Chem. 2004 52 7970 7981 10.1021/jf048741x 15612784 36. Ravichandran K. Ahmed A.R. Knorr D. Smetanska I. The effect of different processing methods on phenolic acid content and antioxidant activity of red beet Food Res. Int. 2012 48 16 20 10.1016/j.foodres.2012.01.011 37. ISO 6635:1984 Fruits, Vegetables and Derived Products--Determination of Nitrite and Nitrate Content--Molecular Absorption Spectrometric Method ISO International Organization for Standardization Geneva, Switzerland 1984 1 8 38. Huber L. Validation of analytical methods and processes Pharmaceutical Process Validation Nash R.A. Wachter A.H. CRC Press Boca Raton, FL, USA 2003 9780203912119 39. The Chief Sanitary Inspectorate Register of Products Subject to the Notification of the First Placing on the Market Available online: (accessed on 14 November 2022) 40. The European Parliament and the Council of the European Union Regulation (EU) No 1169/2011 of the European Parliament and of the Council of 25 October 2011 on the Provision of Food Information to Consumers, Amending Regulations (EC) No 1924/2006 and (EC) No 1925/2006 of the European Parliament and of the Council, and Repealing Commission Directive 87/250/EEC, Council Directive 90/496/EEC, Commission Directive 1999/10/EC, Directive 2000/13/EC of the European Parliament and of the Council, Commission Directives 2002/67/EC and 2008/5/EC and Commission Regulation (EC) No 608/2004 Off. J. Eur. Union 2011 304 18 63 Available online: (accessed on 25 January 2023) 41. The Seym of the Republic of Poland Act of August 25, 2006 on Food and Nutrition Safety (as Amended) J. Law 2006 171 1225 Available online: (accessed on 25 January 2023) 42. The Minister of Health Regulation of the Minister of Health of 9 October 2007 on the composition and labeling of dietary supplements J. Law 2007 196 1425 43. The European Parliament and the Council Regulation (EC) No 1924/2006 of the European Parliament and of the Council of 20 December 2006 on Nutrition and Health Claims Made on Foods Off. J. Eur. Union 2006 404 9 25 Available online: (accessed on 25 January 2023) 44. The European Commission Commission Regulation (EU) No 432/2012 of 16 May 2012 Establishing a List of Permitted Health Claims Made on Foods, Other than Those Referring to the Reduction of Disease Risk and to Children's Development and Health Off. J. Eur. Union 2012 136 1 40 Available online: (accessed on 25 January 2023) 45. European Commission EU Register of Nutrition and Health Claims Made on Foods (v.3.6) Available online: (accessed on 14 November 2022) 46. Apak R. Guclu K. Ozyurek M. Esin Karademir S. Ercg E. The cupric ion reducing antioxidant capacity and polyphenolic content of some herbal teas Int. J. Food Sci. Nutr. 2006 57 292 304 10.1080/09637480600798132 17135020 47. Guclu K. Altun M. Ozyurek M. Karademir S.E. Apak R. Antioxidant capacity of fresh, sulphited-dried Malatya apricot (Prunus armeniaca) assayed by CUPRAC, ABTS/TEAC and folin methods Int. J. Food Sci. Technol. 2006 41 76 85 10.1111/j.1365-2621.2006.01347.x 48. Goscinna K. Czapski J. Wplyw wielkosci i czesci korzenia buraka cwiklowego odmiany Bonel na zawartosc wybranych zwiazkow i aktywnosc przeciwutleniajaca soku Zesz. Probl. Postepow Nauk Rol. 2018 588 149 157 10.22630/ZPPNR.2017.588.14 49. Murawa D. Banaszkiewicz T. Majewska E. Blaszczuk B. Sulima J. Zawartosc azotanow (III) i (V) w wybranych gatunkach warzyw i ziemniakach dostepnych w handlu w Olsztynie w latach 2003-2004 Bromatol. Chem. Toksykol. 2008 41 67 71 50. Keller R.M. Beaver L. Prater M.C. Hord N.G. Dietary nitrate and nitrite concentrations in food patterns and dietary supplements Nutr. Today 2020 55 218 226 10.1097/NT.0000000000000253 51. Hord N.G. Tang Y. Bryan N.S. Food sources of nitrates and nitrites: The physiologic context for potential health benefits Am. J. Clin. Nutr. 2009 90 1 10 10.3945/ajcn.2008.27131 19439460 52. McDonagh S.T.J. Wylie L.J. Thompson C. Vanhatalo A. Jones A.M. Potential benefits of dietary nitrate ingestion in healthy and clinical populations: A brief review Eur. J. Sport Sci. 2019 19 15 29 10.1080/17461391.2018.1445298 29529987 53. Parthasarathy D.K. Bryan N.S. Sodium nitrite: The "cure" for nitric oxide insufficiency Meat Sci. 2012 92 274 279 10.1016/j.meatsci.2012.03.001 22464105 54. Shakil M.H. Trisha A.T. Rahman M. Talukdar S. Kobun R. Huda N. Zzaman W. Nitrites in cured meats, health risk issues, alternatives to nitrites: A Review Foods 2022 11 3355 10.3390/foods11213355 36359973 55. Fewtrell L. Drinking-water nitrate, methemoglobinemia, and global burden of disease: A discussion Environ. Health Perspect. 2004 112 1371 1374 10.1289/ehp.7216 15471727 56. Ashor A.W. Jajja A. Sutyarjoko A. Brandt K. Qadir O. Lara J. Siervo M. Effects of beetroot juice supplementation on microvascular blood flow in older overweight and obese subjects: A pilot randomised controlled study J. Hum. Hypertens. 2015 29 511 513 10.1038/jhh.2014.114 25471613 57. Mills C.E. Govoni V. Faconti L. Casagrande M.L. Morant S.V. Crickmore H. Iqbal F. Maskell P. Masani A. Nanino E. A randomised, factorial trial to reduce arterial stiffness independently of blood pressure: Proof of concept? The VaSera trial testing dietary nitrate and spironolactone Br. J. Clin. Pharmacol. 2020 86 891 902 10.1111/bcp.14194 31833569 58. Kapil V. Khambata R.S. Robertson A. Caulfield M.J. Ahluwalia A. Dietary nitrate provides sustained blood pressure lowering in hypertensive patients: A randomized, phase 2, double-blind, placebo-controlled study Hypertension 2015 65 320 327 10.1161/HYPERTENSIONAHA.114.04675 25421976 59. Blekkenhorst L.C. Lewis J.R. Prince R.L. Devine A. Bondonno N.P. Bondonno C.P. Wood L.G. Puddey I.B. Ward N.C. Croft K.D. Nitrate-rich vegetables do not lower blood pressure in individuals with mildly elevated blood pressure: A 4-wk randomized controlled crossover trial Am. J. Clin. Nutr. 2018 107 894 908 10.1093/ajcn/nqy061 29868911 60. Sundqvist M.L. Larsen F.J. Carlstrom M. Bottai M. Pernow J. Hellenius M.L. Weitzberg E. Lundberg J.O. A randomized clinical trial of the effects of leafy green vegetables and inorganic nitrate on blood pressure Am. J. Clin. Nutr. 2020 111 749 756 10.1093/ajcn/nqaa024 32091599 |
PMC10000617 | Introduction The purpose of this study was to introduce and test a simple, individualized carbohydrate counting tool designed for persons with Type 1 Diabetes Mellitus (T1DM) in order to determine whether the tool improved A1C levels for participants with age, education or language barriers. Methods In a randomized controlled trial, 85 participants were offered six diabetes instructional sessions free of charge over a six-month period. Forty-one received guidance using the regular carbohydrate counting (RCC) method. Forty-four received guidance using an individualized 'Simple Carb Counting' (SCC), involving two customized tables prepared for participants. Results The simple, individualized SCC tool for carbohydrate counting was non-inferior to the standard method of RCC. The SCC tool was more effective among participants aged 40 and older, while no differences were found when comparing participants by education level. Irrespective of intervention group, all participants improved their A1C level (9.9% = 13.2 mmol/L vs 8.6% = 11.1 mmol/L, p = .001). A greater improvement in A1C level was seen in newly diagnosed participants (-6.1 vs -0.7, p = .005, -3.4 vs 0.9, p = .032) in both the RCC and SCC groups. All participants expressed improved emotional level per their PAID5 questionnaires (Problem Areas in Diabetes Scale-PAID), (10.6 (+-5.7) vs 9.5 (+-5.7), p = .023), with women reporting greater improvement than men. Conclusions SCC is a simple, individualized, feasible, low-tech tool for carbohydrate counting, which promotes and enables accurate insulin dosing in people with T1DM. It was found more effective among participants aged 40 and older. Additional studies are needed to corroborate these findings. The Scc tool is a simple tool enabling accurate insulin dosing to all diabetes patients treated with basal and bolus insulin. The SCC tool has the potential to apply to all diabetes patients, particular to those who are uncomfortable with the use of advanced technology, or who do not have access to such technology due to age, education or language barriers. carbohydrate counting patient instruction type 1 diabetes source-schema-version-number2.0 cover-dateMarch 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.2.6 mode:remove_FC converted:10.03.2023 Witkow S , Liberty IF , Goloub I , et al. Simplifying carb counting: A randomized controlled study - Feasibility and efficacy of an individualized, simple, patient-centred carb counting tool. Endocrinol Diab Metab. 2023;6 :e411. doi:10.1002/edm2.411 Shulamit Witkow and Idit F. Liberty were equal contributors to the preparation and execution of this study. pmc1 INTRODUCTION Achieving glycaemic control in Type 1 diabetes mellitus (T1DM) is vital in order to reduce diabetes-related complications. 1 , 2 Treatment recommendations include multiple daily injections of prandial and basal insulin, or continuous insulin infusion (CSII). As carbohydrates are the major nutrient affecting the post-prandial response, it is important to educate individuals on matching prandial insulin doses to carbohydrate intake by means of carbohydrate counting (CC). The checking of pre-meal blood glucose levels and co-ordination of anticipated physical activity 3 are also recommended for optimizing glucose control. CC takes into account the carbohydrate intake per meal, and enables adjustment of the prandial insulin dosage necessary to achieve individual post-prandial glucose targets. 4 To manage CC, the patient has many variables to consider: (i) the personal insulin/carb ratio (I:C) - the amount of insulin the patient needs to overcome a 15-gram portion of carbohydrates; (ii) the accurate assessment of the amount of carbohydrates consumed; (iii) the insulin sensitivity response (IS), that is, the decrease in blood glucose after injecting one unit of insulin to counter an elevated glucose level; (iv) the individual target glucose level post-meal; and (v) the level of glucose prior to the meal. This means that in order to normalize post-prandial glucose levels, individuals need to be highly motivated to maintain wellbeing through on-going compliant conduct, and to be possessed of high mathematical and literacy skills. Studies 5 have shown the difficulties of successfully carrying out CC, especially concerning the underestimation of carbohydrate content in various foods that can lead to post prandial hypo or hyperglycaemia respectively. In a systematic review 6 of CC efficacy in managing T1DM, Bell et al 6 found only partial benefit in achieving glycaemic control through CC, notwithstanding the difficulty of complying with CC instructions. These difficulties are added to other barriers patients with T1DM cope with following dietary recommendations such as, frequent glucose measurements and importantly their feelings about having diabetes. In recent years, new technologies have been developed to simplify CC. 7 While various commercial applications and bolus calculators may lead to better glucose control, such options are not available to all populations including lower socioeconomic status, older people and those with lower technological skills. In an attempt to overcome disparities in access and other barriers, the Diabetes Clinic of Soroka University Medical Center (SUMC) developed a simple, easy-to-use tool - Simple Carb Counting 'SCC.' This tool includes adjustments for those of differing educational backgrounds, cultures and cognitive abilities. It affords persons with diabetes continued enjoyment of their personal eating habits and food preferences through the successful adoption of the SCC method into their daily routines with ease and accuracy. The aim of this randomized controlled study was to test the feasibility and efficacy of SCC, compared to the regular CC method. Our Hypothesis is that this simple tool will enable all people with T1DM improve diabetes control. 2 METHODS We performed an open-label randomized controlled trial at the Diabetes Clinic of SUMC, a tertiary 1200-bed hospital that treats a diverse population, including Bedouin and other Arabs, and Jews from the general, ultra-Orthodox and Ethiopian sectors. Patients with T1DM were eligible for this study if they were (i) over 18 years old, (ii) treated either with an insulin pump or with multiple daily injections of insulin and (iii) had a hemoglobinA1c (A1C) level equal to, or >8.5. The study excluded pregnant or lactating women, patients with severe renal failure, heart failure or under active treatment for cancer. All participants signed an informed consent statement. The study was approved by the SUMC Helsinki Committee on 8 October 2015, approval number 0320-15, and is registered at ClinicalTrials.gov, ID NCT04132128, and conducted from November 2015 to July 2017. Using simple randomization, we assigned participants to one of the two groups, (RCC) Regular carbohydrate counting, or (SCC) simplified individualized tool for carbohydrate counting. The selection process was purely random for every assignment made by Diabetes Clinic staff. All participants were allowed up to six instructional sessions with a registered dietitian who is also a diabetes care and education specialist during a period of about 6 months. All sessions were free of charge and lasted at least 60 min. During the first session, participants were introduced to the carbohydrate counting method to which they were assigned at randomization. Subsequent sessions were dedicated to reinforcing and practicing their method and changes in insulin dosage parameters were made if needed. All patients treated with insulin pumps were encouraged to use the bolus calculator. The primary end-point was level of A1C 6 months after the intervention. Additional data parameters were collected including those of demographics, education and duration of diabetes. Weight was measured and blood studies were conducted before and after the intervention to determine baseline A1C and lipid levels. In order to identify depression and diabetes-related distress, the participants were asked to complete the PAID5 questionnaire (Problem Areas in Diabetes Scale-PAID) 8 at baseline and post-intervention. 2.1 Regular carbohydrate counting (CC) During the instructional sessions, the rationale for carbohydrate counting was explained. Commercial booklets containing a list of the carbohydrate content of foods were provided, and participants were introduced to websites or cellular phone applications designed to assist the public with determining the carbohydrate content of various foods. The participants were taught to calculate the amount of insulin needed using their personal I:C ratio, IS, correction factor, and the glucose target goals prescribed by the Diabetes Clinic team. Participants were encouraged to keep a food diary to assist them with carbohydrate counting. 2.2 Simplified individualized carbohydrate counting (SCC) The SCC tool consisted of two tables written in the participant's native language and adjusted to the participant's specific requirements. Insulin doses were calculated by professional staff using personalized I:C ratios and IS. First Table: The first table, derived from patients' personal IS, listed the number of units that participants needed to administer in order to correct every pre-meal blood glucose level so as to reach their target glucose. Second Table: The second table contained a list of food items derived from participants' personal eating habits, as recorded in their food diaries. The list consisted mainly of the most common foods they regularly consumed, the carbohydrate content of those foods and the number of insulin units needed, as calculated by the personal I:C ratio per usual portion of each food item. High carbohydrate content foods that participants included in their diet were listed, not for healthy nutrition education, but for purposes of facilitating carb counting. Foods that did not contain carbohydrates, such as protein or fat items, were also listed to ensure that the patient realized that these foods contained no carbs. Patients treated with insulin pumps received a personalized table containing the carbohydrates in their food list as grams to accurately calculate the amount of carbs entered to the calculator . FIGURE 1 SCC carb counting chart for MDI users. FIGURE 2 SCC carb counting chart for insulin pump users. At each instructional session, participants' tables were reviewed, personal dosing was tested, food items were added to, or deleted from participants' lists as warranted, and the logic of the method was reiterated for the purpose of reinforcing participants' understanding and compliance. 2.3 Statistical analysis On the basis of the expected A1C difference of 0.8% between groups with 90% power providing an SD of 0.4% and a significance level of 0.05% we estimated that the sample size should consist at least 37 participants in each group. We used descriptive statistics to summarize the data, reporting results as means and standard deviations. Categorical variables were summarized as counts and percentages. Paired t-test was used to examine within changes in A1C between baseline and follow-up, and Student's t-test was used to examine between differences as to the two intervention groups. We further analysed the data after stratifying the study population by sex, education level (above and below 12 years of school), age (above and below age 40) and duration of diabetes (more or <5 years since diagnosis). p values are shown. All analyses were performed with IBM SPSS. 3 RESULTS In total, 107 men and women were recruited for the study, of whom 22 were excluded, as shown in Figure 3. Of the 85 people who were deemed eligible, 48.2%, n = 41, of participants (23 women, 18 men) were assigned to RCC method group, and 51.8%, n = 44, of participants, (17 women, 27 men) to the SCC method group. The mean age of participants was 43.1 (18-74). About 43% of the RCC method group were treated with an insulin pump versus 36% of the SCC group (p = .48). All patients monitored their glucose by self-monitoring of blood glucose (SMBG) . FIGURE 3 Flow diagram - Trial of standard carb counting versus simple individualized carb counting. TABLE 1 Baseline characteristics of study population by carb counting method. Variable RCC (n = 41) SCC (n = 44) p value Age, Years (+-SD) 40.9 (15.5) 45.3 (18) .18 Women, N (%) 23 (56) 17 (39) .08 Diabetes duration, Years (+-SD) 15.2 (12) 16.3 (13.6) .7 Use of insulin pump n (%) 18 (44) 16 (36) .48 Education above 12 years n (%) 17 (42) 21 (48) .36 BMI kg/m2 (+-SD) 26.5 (5.0) 25.6 (6.0) .43 A1C % (+-SD) 10.0 (1.9) 9.9 (1.6) .70 Cholesterol mg/dL (+-SD) 178.4 (39.4) 177.8 (42.7) .95 LDL mg/dL (+-SD) 104.8 (31.0) 100.8 (32.2) .56 HDL mg/dL (+-SD) 55.8 (14.2) 50.7 (15.6) .11 Triglycerides mg/dL (+-SD) 106.4 (70.8) 128.3 (57.1) .12 Note: RCC = The regular method of carbohydrate counting (CC); SCC = The individualized simple tool for carbohydrate counting. Abbreviation: BMI, Body mass index. The participants were followed for a mean period of 6 months, during which all participants in the study improved their A1C level between baseline and follow-up (9.9% = 13.2 mmol/L vs 8.6% = 11.1 mmol/L, p = .001). Other biomarkers did not change from baseline in either of the groups. We stratified the study population by participant age (older and younger than age 40). Among older participants (mean age 55.2 (+-9.4)), only those who used the SCC method exhibited significant improvement in A1C level, from baseline (9.6 (+-1.3) = 12.7 mmol/L to 8.6 (+-1.1) = 11.1 mmol/L, p = .002). While those using the RCC method showed some improvement, it did not reach statistical significance (9.7 (+-1.3) = 12.9 mmol/L to 9.2 (+-1.5) = 12.1 mmol/L, p = .09). . FIGURE 4 HbA1c results SCC vs RCC. Among younger participants (mean age 29.6 (+-6.3)), a significant improvement was found in both the RCC group (10.2 (+-2.3) = 13.7 mmol/L to 8.3 (+-1.4) = 10.6 mmol/L, p = .001) and SCC group (10.3 (+-1.9) = 13.8 mmol/L to 8.3 (+-1.1) = 10.6 mmol/L, p = .002). We stratified the study population by level of education (above and below 12 school years). Both higher and lower educated participants in the SCC group demonstrated a significant improvement in A1C level. The results for those with above and below 12 school years were -1.4% (+-2.0) vs -1.3% (+-1.9) respectively, p = .7. We found a greater improvement in A1C levels when we compared participants with more recently diagnosed diabetes (<5 years from diagnosis, n = 13) to those whose diabetes was of longer duration (n = 72). (-6.1 vs -0.7, span >-3.4 vs 0.9, p = .032) in both the RCC and SCC groups. We measured the degree of patient compliance by the number of instructional sessions attended during the study period. Compliance was fairly good for all participants, with a mean of 4.8 visits (out of the six allowed). No differences were found in compliance between participants in the RCC group and those in the SCC group, but women were more compliant than men, with a mean of 5.3 (1.3) visits, compared to 4.4 (1.9) for the men (p = .01). Compliance tended to be higher among participants with more than 12 years of education, compared to those with fewer years of education (4.5 visits (+-1.9) vs 5.3 (+-1.3), p = .08). The degree of compliance correlated with decreased A1C level post-intervention. Sixty-three participants (30 in RCC and 33 in SCC) completed the PAID5 questionnaire at baseline and post-intervention. All participants reported increased satisfaction, as exhibited by a decreased PAID5 score (10.6 (+-5.7) vs 9.5 (+-5.7), p = .023). Based on the questionnaire responses, there were no differences in diabetes-related emotional distress between participants in the RCC group and those in the SCC group. Yet, when the investigators stratified the data by sex, they found a significant improvement among women, who reported a decrease in diabetes-related emotional distress from 11.5 to 9.9, p = .04, compared to men, 9.9 to 9.18, p = .27 (Table 2). TABLE 2 PAID 5 score results. Baseline PAID Score (+-SD) Post-intervention PAID Score (+-SD) p-value p-value (SCC vs RCC) All participants (N = 63) 10.6 (5.7) 9.5 (5.7) .02 All SCC (N = 33) 12.5 (5.4) 11.1 (5.5) .03 All RCC (N = 30) 9.3 (5.4) 8.42 (5.6) .09 .69 All women (N = 30) 11.5 (5.8) 9.9 (5.6) .02 SCC women (N = 18) 13.11 (5.5) 11.28 (5.6) .04 RCC women (N = 12) 9 (5.6) 7.75 (5.1) .15 .7 All men (N = 33) 9.9 (5.7) 9.18 (5.9) .13 SCC men (N = 12) 10.67 (6.2) 9.83 (6.1) .23 RCC men (N = 21) 9.48 (5.5) 8.81 (5.9) .21 .9 Education >12 y (N = 31) 8.65 (4.4) 7.87 (4.7) .12 SCC education >12 y (N = 15) 9.5 (4.9) 8.3 (5.2) .11 RCC education >12 y (N = 16) 7.9 (3.9) 7.4 (4.4) .32 .6 Education <12 y (N = 32) 12.6 (6.3) 11.1 (6.2) .02 SCC education <12 y (N = 15) 14.8 (5.5) 13.1 (5.5) .08 RCC education <12 y (N = 17) 10.6 (6.4) 9.3 (6.5) .08 .77 Age > 40 (N = 33) 9.1 (6.0) 8.1 (5.3) .07 SCC age > 40 (N = 13) 11.3 (6.6) 9.6 (5.4) .08 RCC age > 40 (N = 20) 7.75 (5.2) 7.2 (5.2) .2 .4 Age < 40 (N = 30) 12.3 (5.1) 11 (5.9) .04 SCC age < 40 (N = 17) 12.8 (5.2) 12 (6.1) .11 RCC age < 40 (N = 13) 11.7 (5.6) 10.3 (5.8) .11 .9 During the 6 month follow-up, two patients from the RCC group were hospitalized with diabetic ketoacidosis (DKA). There were no hospitalizations or ER visits due to hypoglycaemia in both groups. 4 DISCUSSION The findings showed that the SCC simple, individualized tool for carbohydrate counting was non-inferior to the standard method of RCC. The SCC tool was more effective among participants aged 40 and older, while no differences were found when comparing participants above and below 12 school years. However, significant improvement in A1C level was observed in all participants. Participants in both RCC and SCC groups who were diagnosed with diabetes within the previous 5 years exhibited significantly greater improvement in A1C level, compared to participants with diabetes of longer duration. The SCC method presented in our study was developed in order to overcome difficulties and barriers that the diabetes clinic patients encountered in implementing CC, as described in several studies. Kawamura et al 5 tested the errors in carbohydrate content estimation among 37 paediatric patients, their parents, and their health care professionals, including physicians and dietitians. In all groups studied, they found overestimation of the carb content in foods with small amounts of carbs, and underestimation in foods with high carb content. While past experience in CC was important, some foods, such as rice, were hard to estimate even by experienced participants. In the qualitative study of Gurkan et al 9 investigators interviewed adolescents with diabetes, finding multiple barriers to effective treatment. Among the barriers were patients' negative feelings about having diabetes, as well as personal and environmental barriers. Personal barriers included lack of knowledge about the disease, trouble with glucose measurement, and difficulty following dietary recommendations. These findings were corroborated by Ahola et al 10 who found that many patients experienced difficulty managing their post-prandial glucose, and were subject to a high percentage of time in a hyperglycaemic state. In this study, we present an option to overcome some of the barriers described in the above studies. Through personalization, flexibility, and a departure from a restrictive diet paradigm, SCC affords persons with diabetes an opportunity to continue eating their usual diet, including the customary dishes of their culture, and to go on with their life-long social dining habits. The study showed superiority in reaching glycaemic control in participants older than age 40 who used the SCC method, compared to those in the RCC group. Treating older patients with T1DM is complicated by the combined challenges of insulin-dependent diabetes, age-related complications, and possible comorbidities, all of which negatively affect the older population's ability to self-manage diabetes. 11 The SCC tool presented in the study simplifies the tasks needed for carbohydrate control, and consequently leads to better glycaemic control especially for older age group. Contrary to these findings, the study demonstrates that SCC was non-inferior in people with various levels of education. In a cross-sectional multicentre study of 768 subjects under age 18 with T1DM, Gesuita et al 12 found that only 28.1% of participants reached target A1C values (<7.5%). A strong correlation was found between higher socio-economic status (SES), higher level of education and higher ability to follow ordinary CC. Significantly, Gesuita et al. highlighted the need for an accessible tool for non-privileged populations. Recently diagnosed participants in both RCC and SCC groups showed the greatest benefit in improving glycaemic control. This may be explained in two possible ways, one psychological and one physiological. When first diagnosed, many patients are highly motivated to do well. In addition, in the early period after diagnosis, sometimes called the 'honeymoon period,' the pancreas seems to do better and secretes more insulin, although this phenomenon decreases with time and differs with each patient. All participants exhibited significant decreases in their PAID5 scores. Studies have shown 13 , 14 that people with diabetes suffer from higher levels of psychological distress than does the healthy population. People with T1DM are three times more likely to develop depression than those without diabetes. 15 Moreover, psychological distress has been shown to be associated with hyperglycaemia, complications and higher mortality rate. 16 , 17 Thus, there is a consensus that treating psychological stress and achieving psychological wellbeing ought to be one of the treatment goals of diabetes care. 18 A study by Zagarins et al 19 revealed a correlation between improvement of glycaemic control and alleviation of overall psychological stress, but not in depression. Our study corroborates these findings, and underscores the need for on-going diabetes education, better understanding and treatment of diabetes and promoting a greater sense of self-efficacy among patients in controlling the disease, as means of improving not only metabolic control, but also mental health. 4.1 Limitations The intervention tool was introduced at a single diabetes clinic in a tertiary teaching hospital with one registered dietitian/diabetes care and education specialist. The method was not tested on paediatric patients, a population that has more difficulties with glycaemic control than others. A larger population of people from the lower socio-economic and more culturally diverse backgrounds should be studied in order to corroborate the results and establish generalizability across populations. 5 CONCLUSIONS AND IMPLICATIONS In large measure, the research into T1DM treatment is focused on advanced technologies including insulin pumps, continuous glucose monitoring, the artificial pancreas and various applications to support CC and diabetes management. Studies have shown 20 that although advanced applications are accessible and improve glycaemic control, only a small percentage of the population with T1DM chooses to use them. This may be explained by the human factor, that is, personal expectations, perceptions of the burden of new technologies, user-friendliness and long-term cost. The SCC tool tested in this study has the potential to apply to all diabetes patients, and in particular to those who are uncomfortable with the use of advanced technology, or who do not have access to such technology. In conclusion this study presents a simple, feasible, low-tech tool that simplifies carbohydrate counting and which promotes and enables accurate insulin dosing in people with T1DM. Additional studies are needed to corroborate these findings. AUTHOR CONTRIBUTIONS Shulamit Witkow: Conceptualization (lead); data curation (lead); investigation (equal); methodology (equal); resources (lead). Idit F Liberty: Data curation (equal); formal analysis (supporting); investigation (equal); methodology (equal); project administration (equal); supervision (lead); validation (equal); writing - original draft (lead); writing - review and editing (lead). Irina Goloub: Data curation (supporting); resources (supporting). Malka Kaminsky: Conceptualization (supporting); data curation (equal); investigation (supporting); resources (supporting). Olga Otto: Data curation (equal); investigation (supporting). Yones Abu Rabia: Data curation (supporting); investigation (supporting); resources (equal). Ilana Harman Boehm: Conceptualization (equal); data curation (equal); investigation (equal); methodology (equal); resources (equal). Rachel Golan: Formal analysis (lead); writing - review and editing (supporting). CONFLICT OF INTEREST STATEMENT The Authors declare that there are no conflicts of interest. ACKNOWLEDGEMENTS The Authors declare there are no conflicts of interest and no financial support for this study. We thank all the diabetes unit staff for all technical assistance. DATA AVAILABILITY STATEMENT Raw data were generated at the diabetes centre of SUMC. Derived data supporting the findings of this study are available from the corresponding author [IFL] on request |
PMC10000618 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050733 cells-12-00733 Article Blocking EREG/GPX4 Sensitizes Head and Neck Cancer to Cetuximab through Ferroptosis Induction Jehl Aude 1 Conrad Ombline 1 Burgy Mickael 12 Foppolo Sophie 1 Vauchelles Romain 1 Ronzani Carole 3 Etienne-Selloum Nelly 14 Chenard Marie-Pierre 5 Danic Aurelien 6 Dourlhes Thomas 6 Thibault Claire 6 Schultz Philippe 6 Dontenwill Monique 1 Martin Sophie 1* Bar Julia Academic Editor Zidar Nina Academic Editor Cierpikowski Piotr Academic Editor 1 Laboratory of Bioimaging and Pathology, University of Strasbourg, UMR7021 CNRS, 67401 Illkirch, France 2 Department of Medical Oncology, Institute of Cancerology Strasbourg Europe, 67200 Strasbourg, France 3 Laboratory of Design and Application of Bioactive Molecules, University of Strasbourg, UMR7199, CNRS, 67400 Illkirch, France 4 Department of Pharmacy, Institute of Cancerology Strasbourg Europe, 67200 Strasbourg, France 5 Department of Pathology, Strasbourg University Hospital, 67200 Strasbourg, France 6 Department of Otolaryngology and Cervico-Facial Surgery, Strasbourg University Hospital, 67200 Strasbourg, France * Correspondence: [email protected]; Tel.: +33-36-885-4197; Fax: +33-36-885-4313 24 2 2023 3 2023 12 5 73302 12 2022 13 2 2023 20 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). (1) Background: Epiregulin (EREG) is a ligand of EGFR and ErB4 involved in the development and the progression of various cancers including head and neck squamous cell carcinoma (HNSCC). Its overexpression in HNSCC is correlated with short overall survival and progression-free survival but predictive of tumors responding to anti-EGFR therapies. Besides tumor cells, macrophages and cancer-associated fibroblasts shed EREG in the tumor microenvironment to support tumor progression and to promote therapy resistance. Although EREG seems to be an interesting therapeutic target, no study has been conducted so far on the consequences of EREG invalidation regarding the behavior and response of HNSCC to anti-EGFR therapies and, more specifically, to cetuximab (CTX); (2) Methods: EREG was silenced in various HNSCC cell lines. The resulting phenotype (growth, clonogenic survival, apoptosis, metabolism, ferroptosis) was assessed in the absence or presence of CTX. The data were confirmed in patient-derived tumoroids; (3) Results: Here, we show that EREG invalidation sensitizes cells to CTX. This is illustrated by the reduction in cell survival, the alteration of cell metabolism associated with mitochondrial dysfunction and the initiation of ferroptosis characterized by lipid peroxidation, iron accumulation and the loss of GPX4. Combining ferroptosis inducers (RSL3 and metformin) with CTX drastically reduces the survival of HNSCC cells but also HNSCC patient-derived tumoroids; (4) Conclusions: The loss of EREG might be considered in clinical settings as a predictive biomarker for patients that might undergo ferroptosis in response to CTX and that might benefit the most from the combination of ferroptosis inducers and CTX. head and neck cancers EREG ferroptosis tumoroid biomarkers metabolism autophagy Interdisciplinary Thematic Institute program of the University of StrasbourgANR-10-IDEX-0002 ANR-20-SFRI-0012 Ligue Contre le Cancer, Conference de coordination interregionale du Grand Est programme-inter regionS19R417B URPS Chirurgiens Dentistes Grand EstThis work was supported by the Interdisciplinary Thematic Institute program of the University of Strasbourg (France, ANR-10-IDEX-0002 and ANR-20-SFRI-0012) in the frame of the InnoVec Institute. This research was funded by Ligue Contre le Cancer, Conference de coordination interregionale du Grand Est programme-inter region (France, No S19R417B) and the URPS Chirurgiens Dentistes Grand Est. pmc1. Introduction With nearly 932,000 new cases and 450,000 deaths in 2018 worldwide, head and neck cancers, and, more particularly, head and neck squamous cell carcinoma (HNSCC), rank sixth among the most frequently observed cancers in the world . Locally advanced HNSCC (LA-HNSCC, stage III/IV) represents 70% of patients at diagnosis. They require primary surgery followed by adjuvant (chemo)-radiotherapy or definitive chemoradiotherapy including cetuximab (CTX) . The chimeric antibody CTX targets the epidermal growth factor receptor (EGFR), which happens to be overexpressed in more than 90% of HNSCC. CTX prevents ligand binding and dimerization with other HER family members . Once bound, CTX blocks EGFR phosphorylation and signal transduction and promotes EGFR internalization, thus turning down the oncogenic EGFR signaling . Unfortunately, some patients do not benefit from CTX treatment, and others show recurrences soon after the end of the treatment. Intrinsic or therapeutically acquired resistance to CTX is extensively studied to understand the mechanisms involved. We have recently shown the involvement of the caveolin-1/epiregulin/YAP axis in the resistance to CTX and irradiation therapy . Epiregulin (EREG), encoded by the EREG gene located on chromosome 4q13.3, belongs to the ErbB family of ligands. The 162 amino acids transmembrane proform of EREG is proteolytically cleaved by ADAM17 to release a soluble form of 46 amino acids . EREG shares 24-50% of its sequence with those of other members of the EGF family . EREG binds to EGFR (extracellular domain I and III that partially overlaps the EGF binding site ) and ErbB4. It stimulates homodimers of EGFR and ErbB4 in addition to heterodimers of ErbB2 and ErbB3, leading to the activation and the transduction of downstream signaling pathways . In contrast to EGF, EREG leads to complete EGFR recycling and not to lysosomal degradation . EREG induces less stable EGFR dimers than EGF, but unexpectedly, this weakened dimerization elicits more sustained EGFR signaling than EGF . Low or non-existent in most human tissues, the elevation of EREG expression is observed in the early stages of cancer development, in which EREG induces epithelial-mesenchymal transition . EREG is a transcriptional target of the oncogenic KRAS and is also overexpressed in cells with an oncogenic mutation of EGFR and BRAF . EREG also promotes tumorigenicity, metastasis, drug resistance and cell plasticity and modulates the tumor microenvironment and metabolism . The increased expression of EREG is associated with short overall survival in patients with HNSCC/OSCC . Elevated levels of EREG appear to be a potential predictive biomarker of anti-EGFR therapies in several cancer types including HNSCC . We recently reported that the overexpression of caveolin-1 in HNSCC is associated with the total loss of EREG as well as the oncogenic addiction to EGFR. Silencing EREG activates the YAP/TAZ pathway, which enables cells to resist CTX/radiotherapy. The resistance of HNSCC cells to therapy is linked to the protection of the mitochondria and drives the recurrence of caveolin-1-expressing HNSCC tumors . How the loss of EREG affects the cellular metabolism and how it might influence the response to anti-EGFR therapies is only sparsely documented in HNSCC. Here, we aimed to clarify this point and highlighted that the loss of EREG sensitizes cells to CTX through the induction of ferroptosis. Our study reveals the glutathione metabolism as a targetable vulnerability of HNSCC that should be exploited in CTX-based therapies. 2. Materials and Methods 2.1. Cell Culture, Transfection and Drugs The CAL27, CAL33 and SCC9 cell lines were purchased from the ATCC(r) and DSMZ (authenticated by STR profiling). All cell lines tested negative for mycoplasma contamination. CAL27 and CAL33 were grown in DMEM (PAN Biotech, Aidenbach, Germany) supplemented with 2 mM ultra-glutamine, 0.5 mM sodium pyruvate and 10% heat-inactivated FBS (Gibco, Dutscher SAS, Brumath, France). SCC9 were grown in DMEM-F12 (PAN Biotech) supplemented with 2.5 mM ultra-glutamine, 15 mM HEPES, 400 ng/mL hydrocortisone (Sigma-Aldrich, St Quentin Fallavier, France) and 10% FBS (Gibco). EREG expression was downregulated by transfecting the cell lines with 25 nM siRNAEREG (Human EREG and the respective control siRNACtrl, SMARTPool from Dharmacon) using Lipofectamine 2000TM (Invitrogen, Thermo Fischer Scientific, Illkirch, France). Efficient EREG silencing was determined by Western blot. When indicated, cells were treated with solvent, 30 nM of CTX (ErbituxTM, 5 mg/mL, Merck, ICANS, Strasbourg, France), 5 mM RSL3 (MedChemExpress, Clinisciences, Nanterre, France) or 1 mM metformin (MedChemExpress) alone or in combination with CTX. 2.2. Sphere Evasion Assay After the treatments, 500,000 cells were resuspended in 1 mL of regular culture medium supplemented with 20% methylcellulose. Spheroids were formed using the hanging drop culture method. Drops of 20 mL cell suspension were placed onto the lids of 60 mm dishes, which were inverted over the dishes. The dishes were cultured in humidified chambers for 48 h to allow the formation of round aggregates. The spheroids were harvested and seeded in plastic 24-well plates (6 spheres/well) for an additional 24 h to allow for the evasion of cells from the attached spheres. Pictures were taken using the Evos XI Core microscope (AMG, Thermo Fischer Scientific, Illkirch, France), with 10x magnification. The results were expressed, in pixels, as the evasion area of the cells relative to the area of the attached sphere (total area - area of the sphere), determined using ImageJ 1.53t). 2.3. IncuCyte(r) Assay After transfection, the cells were seeded (4000 for CAL33 and 8000 for CAL27 and SCC9 cells/200 mL/well) in 96-well plates. The plates were placed at 37 degC, and the confluence, growth, cell health and morphology were monitored for 164 h/7 days. The percentage of confluence was determined using the IncuCyte(r) analysis software after normalization to day 0 (Essen BioScience, Sartorius, Goettingen, Germany). 2.4. Clonogenic Survival Assay A total of 72 h after the transfection and/or an additional 48 h of treatment with the solvent, 30 nM CTX, 5 mM RSL3, 1 mM metformin or co-treatments, the cells were seeded (500 for CAL27 and SCC9 and 1000 for CAL33 cells/2 mL/well) in 6-well plates and allowed to grow for 10 days. The cells were colored with crystal violet at 0.1% (Sigma-Aldrich, St Quentin Fallavier, France). The colonies were counted to determine the plating efficiency (PE) and the surviving fraction (SF). PE = number of surviving cells/number of cells plated. SF = PE of the experimental group/PE of the control group. 2.5. Metabolic Assays After the treatment, 20,000 cells were plated in a Seahorse XF Cell Culture microplate in XF growth medium (non-buffered DMEM containing 10 mM glucose, 4 mM L-glutamine and 2 mM sodium pyruvate). The OCR (oxygen consumption rate), ECAR (extracellular acidification rate) and ATP consumption were measured using the ATP rate assay procedure under basal conditions and in response to 1.5 mM oligomycin and 0.5 mM rotenone/antimycin A with the XFp Extracellular Flux Analyzer (Seahorse Bioscience, Agilent, Les Ulis, France). The metabolic profiles were analyzed using the online Seahorse analytics platform. 2.6. Western Blot A total of 72 h after the transfection and/or an additional 48 h of treatment with the solvent, 30 nM CTX, 5 mM RSL3, 1 mM metformin or co-treatments, the cells were lysed with the lysis buffer (1% Triton, 100 nM NaF, 10 mM Na4O7P2, 1 mM Na3VO4, protease inhibitor cocktail (Roche, Meylan, France) in PBS) for 30 min at 4 degC and then sonicated. The supernatant was recovered by centrifugation at 20,000x g for 10 min at 4 degC. In total, 5 to 20 mg of proteins were separated on a 4-20% TGX-denaturing polyacrylamide gel (SDS-PAGE Bio-Rad Marnes-La-Coquette, France) and transferred to polyvinylidene difluoride (PVDF) membrane (Amersham, Sigma-Aldrich, St. Quentin Fallavier, France). After 1 h of blocking at room temperature, the membranes were probed with appropriate primary antibodies (see Supplementary Table S1) overnight at 4 degC. The membranes were subsequently incubated with anti-rabbit or anti-mouse antibodies conjugated to horseradish peroxidase (Promega, Charbonnieres les-Bains, France), developed using chemoluminescence (ECL, Bio-Rad, Marnes-La-Coquette, France) and visualized with an Las4000 image analyzer (GE Healthcare, Tremblay-en-France France). The quantification of non-saturated images was carried out using ImageJ software (National Institutes of Health, Bethesda, MD, USA). GAPDH was used as the loading control. The results were expressed as histograms representing the mean +- SEM of the ratios protein/GAPDH normalized against the controls. 2.7. Quantification of Intracellular Fe2+ Accumulation A total of 72 h after the transfection and/or an additional 48 h of treatment with 30 nM CTX, the cells were seeded at 20,000 cells for CAL27 and CAL33 and at 15,000 cells for SCC9 for 24 h in 96-well plates with opaque walls. The intracellular accumulation of Fe2+ was determined using the intracellular probe FerroOrange at 1 mM for 30 min (Dojindo, TebuBio, LePerray en Yvelines, France). The fluorescence intensity was measured with a Varioskan LUX (Thermo Scientific, Illkirch, France) plate reader. In parallel, the cells were seeded at 30,000 cells for CAL27 and at 20,000 cells for CAL33 and SCC9 in an eight-well LabTek for imaging with an LEICA TCS SPE II confocal microscope (Leica Microsystems SA, Nanterre Cedex, France), with a x20 magnification objective, and analyzed with ImageJ software. 2.8. Detection of the Accumulation of Lipid Peroxides A total of 72 h after the transfection and/or an additional 48 h of treatment with 30 nM CTX, the cells were seeded for 24 h at 30,000 for CAL27 and at 20,000 for CAL33 and SCC9 in an eight-well LabTek. The accumulation of lipid peroxides was determined using the Liperfluo kit at 1 mM for 30 min (Dojindo, China). The cells were also seeded into LabTek wells for imaging with an LEICA TCS SPE II confocal microscope (Leica Microsystems SA, Nanterre Cedex, France), with a x20 magnification objective, and analyzed with ImageJ software 1.53t). 2.9. Tumoroids Culture The study was approved by the Scientific Committee of the tumor bank and the Department of Pathology of the CHU Strasbourg-Hautepierre (France). The patients have signed an informed consent form. Tumor extractions were carried out in the Department of Cervico-Facial Surgery of the CHU Strasbourg-Hautepierre (France). The resected pieces were histologically diagnosed. The tumoroids were extracted from head and neck cancer surgical resection following the protocol developed by Driehus et al. and cultured in advanced DMEM/F12 supplemented with GlutaMax, Penicilin/streptomycin, 10 mM HEPES, 10 mM Y-27632 (Euromedex, Souffelweyersheim, France), 0.5 mg/mL Capsofungin (Sigma), 1x B27 supplement (Thermo Fisher Scientific), 1.25 mM N-acetyl-L-cysteine (Sigma-Aldrich), 10 mM Nicotinamide (Sigma-Aldrich), 500 nM A83-01 (Sigma-Aldrich), 0.3 mM CHIR99021 (Sigma-Aldrich), 50 ng/mL human EGF (PeproTech, Thermo Scientific, Illkirch, France), 10 ng/mL human FGF10 (PeproTech), 5 ng/mL human FGF2 (PeproTech), 1 mM Prostaglandin E2 (Bio-techne, R&D Systems, Noyal Chatillon sur Seiche, France) and 1 mM Forskolin (Bio-techne), 4% (col/col) RSPO3-Fc fusion protein conditioned medium (ImmunoPrecise, IPATherapeutics, Utrecht, Netherlands) and 4% (vol/vol) Noggin-Fc fusion protein conditioned medium (ImmunoPrecise). Quality control of the tumoroids was performed. The tumoroids were plated at 2500 cells/10 mL of 70% Cultrex UltiMatrix reduced growth factor basement membrane Extract (R&D Systems, Noyal Chatillon sur Seiche, France) in 24-well plates. The tumoroids were treated with the solvent, 30 nM CTX, 5 mM RSL3, 1 mM metformin or co-treatments 7 days after plating for an additional 10 days. The cell viability was assessed after the exposure of the cells to trypan blue (Bio-Rad) and reading via a TC20 Automated Cell Counter (Bio-Rad). Moreover, this culture was monitored by imaging at x4 and x20 magnification via an Evos XI Core microscope (AMG). 2.10. Immunohistochemistry on Tumoroids Following the recovery, the tumoroids were fixed in PFA 4% for 20 min and washed in PBS. After a 15 min permeabilization step in PBS/0.1% Tween-20 and a 60 min blocking step in PBS/0.1% Triton X-100/2% BSA/5% NGS, the tumoroids were incubated overnight at 4 degC with primary antibodies (Rabbit anti-EREG, #CSB-PA007779NA01HU, Cusabio Technology, dilution 1/300; Mouse anti-Caveolin-1, #66067-1-lg, Proteintech, dilution 1/1000). After washing in PBS/0.1% Triton X-100/0.2% BSA, the cells were incubated for 3 h at room temperature with appropriate secondary antibodies (Life Technologies; dilution 1/500) and DAPI (#D9542; Sigma-Aldrich, St Quentin Fallavier, France; 1 mg/mL). After washing twice in PBS/0.1% Triton X-100/0.2% BSA and twice in PBS, the slides were mounted using FUnGI medium (50% (v/v) glycerol, 9.4% (v/v) dH2O, 10.6 mM Tris base, 1.1 mM EDTA, 2.5 M fructose and 2.5 M urea). Images were acquired using an LEICA TCS SPE II confocal microscope (Leica Microsystems SA, Nanterre Cedex, France), with a 20x magnification objective, and analyzed with ImageJ software version 1.53r, access on 3 May 2021) or Imaris software (Imaris x64 9.3.1--22 May 2019). 2.11. Statistical Analysis Quantitative variables are presented as their mean and standard deviations and were compared to univariate analyses with a Student's t-test if they followed a Gaussian distribution (Shapiro-Wilk tests were used to assess the Gaussian distribution) or with a Wilcoxon's rank test if they followed a non-Gaussian distribution. 3. Results 3.1. Silencing EREG Prevents Survival and Growth and Sensitizes to CTX We recently reported that decreased EREG expression conferred Cav1-overexpressing cells resistance to CTX/radiotherapy . We postulated that it was the result of a decrease in EREG-driven oncogenic addiction to EGFR. To go further, EREG was silenced in a panel of three basal-like HNSCC cell lines using siRNA. The basal expression of Cav1, EREG and EGFR was not altered by the transfection of siRNActrl. Silencing EREG does not alter Cav1 expression and exhibits a significant reduction in EGFR, which is in contrast with the molecular alterations observed previously . EREG-silenced cells (siRNAEREG) show reduced clonogenic survival . Although CTX significantly reduces the survival of control (siRNActrl-transfected) cells, its effect is even more pronounced in EREG-silenced cells . Reduced clonogenic survival is associated with an altered growth capacity, reflected here by the inability of cells to reach confluency. Again, CTX is more prone to blocking the growth of cells silenced for EREG . The cleavage of PARP, reflecting apoptosis induction, could only barely be detected and only in cells exposed to CTX . No additional cleavage could be observed in siRNAEREG-transfected cells when compared to the controls. Thus, apoptosis induction could not account for the reduction in survival and growth observed in siRNAEREG-cells remaining untreated or treated with CTX. Taken together, these data show that the concomitant silencing of EREG and EGFR targeting would be more effective in inhibiting tumor growth and survival. 3.2. Silencing EREG Promotes Mitochondrial Dysfunction and Inhibits Autophagy in Reponse to CTX The metabolic reprogramming of cancer cells has a beneficial effect not only on tumor growth and survival but also on metastasis and chemoresistance. We therefore investigated how EREG might affect the mitochondrial metabolism. The ATP Rate assays reveal that all three cell lines exhibit different basal oxygen consumption (OCR), extracellular acidification (ECAR) and ATP production. CAL33 cells are the most energetic . EREG-silencing significantly reduces OCR and ECAR and ATP production in all three cell lines, with the highest efficiency in the most energetic cell line, CAL33 . Although CTX significantly reduces the production of ATP in both siRNAEREG-transfected cells, SCC9 seem less sensitive to it . EREG-silenced cells treated with CTX appear less metabolically active. Thus, silencing EREG and blocking EGFR with CTX cause mitochondria dysfunction, which is more important in highly metabolic cells. Autophagy is a critical protective mechanism against mitochondrial dysfunction. It maintains cellular homeostasis by removing damaged macromolecules and organelles, including mitochondria. The expression of ULK-1, a kinase regulating the early stages of the autophagosome formation, is only induced in cells exposed to CTX, and no differences are observed between siRNAEREG-transfected cells. In contrast, the silencing EREG reduces the expression of Beclin1 and LC3B in CAL33 and the expression of LC3B in SCC9 without affecting CAL27 cells . Although CTX does not affect Beclin1 and LC3B by itself in any cell line tested, it reduces their expression even further in siRNAEREG-transfected CAL33 and SCC9 cells without affecting CAL27 cells . The data show that silencing both EREG and EGFR signaling inhibits autophagy. 3.3. Silencing EREG Promotes Ferroptosis in Response to CTX We next focused on ferroptosis, a different class of cell death, characterized by the accumulation of ferrous ions (Fe2+) and the increase in the production of lipid reactive oxygen species (ROS). The accumulation of Fe2+ was monitored using the FerroOrange probe. Silencing EREG significantly reduces Fe2+ staining in CAL27 and CAL33 cells . Although CTX does not affect Fe2+ accumulation in siRNActrl-cells, it significantly induces Fe2+ staining in siRNAEREG-transfected CAL27 and CAL33 without affecting SCC9 . Lipid peroxides were monitored using the LiperFluo probe and revealed similar staining profiles to the ones observed in Figure 3B. Thus, silencing EREG reduces the accumulation of lipid peroxides in CAL27 and CAL33 cells . Turning to CTX, it significantly induces lipid peroxides staining in siRNAEREG-transfected CAL27 and CAL33 without affecting SCC9 . Altogether, EREG-silencing reprograms cells to induce ferroptosis in the presence of CTX. 3.4. EREG-Silencing Uncovers the Vulnerability of Cells to GPX4 Inhibition It has been shown that ferroptosis is initiated either by the loss of glutathione peroxidase 4 (GPX4, an enzyme involved in lipid repair) or the depletion of cystine. GPX4, together with its co-factor glutathione (GSH), catalyzes the inhibition of lipid peroxides. Its loss is concomitant with the accumulation of lipid peroxides in membranes, which leads to ferroptosis. Silencing EREG does not affect GPX4 expression in any of the cell lines tested. The exposure of siRNActrl-transfected cells to CTX does not affect it either . However, the treatment of siRNAEREG cells with CTX results in a significant inhibition of GPX4 expression in all three cell lines . The cystine/glutamate antiporter system (also called x-c or xCT (coded by the genes SLC7A11 and SLC3A2)) imports extracellular cystine that will be further reduced into cysteine. Cysteine acts as a precursor for the synthesis of GSH, the cofactor of GPX4. GPX4 is also a direct transcriptional target of NRF2. SLC7A11 (as well as SLC1A5 and SLC7A5) and NRF2 are under the control of the oncogene c-Myc ; we investigated how EREG and/or CTX might affect c-Myc expression in our system. Silencing EREG or exposing siRNActrl-transfected cells to CTX does not affect c-Myc expression in any of the cell lines tested . However, the treatment of siRNAEREG cells with CTX results in a significant inhibition of c-Myc expression in all three cell lines . Pharmacological inhibitors such as RSL3 have been reported to either degrade GPX4 or inhibit its function. RSL3 reduces the surviving fraction of siRNAEREG-transfected cells to similar levels as CTX . The surviving fraction of both siRNAEREG-transfected cells is even further inhibited when RSL3 is combined with CTX . Metformin, already used in the treatment of diabetes, was recently described as promoting ferroptosis in different ways, including by the inhibition of SLC7A11 . We therefore studied the effects of this non-specific inducer of ferroptosis in our model. Metformin reduces the surviving fraction of siRNAEREG-transfected cells to similar levels as CTX in CAL33, but it was far more potent in CAL27 and SCC9 cells . The surviving fraction of both siRNAEREG-transfected cells is even further inhibited when metformin is combined with CTX . The exposure of the cells to RSL3 or metformin alone or in combination with CTX results in a significant inhibition of GPX4 expression, which is even more pronounced in EREG-silenced cells . Altogether, the data show that GPX4 is crucial for cell survival and that its disappearance sensitizes to CTX. 3.5. GPX4 Inhibition Sensitizes the Patient-Derived Tumoroid to CTX In order to validate our hypothesis, we exposed patient-derived tumoroids to CTX, RSL3, metformin and a combination of drugs for 7 days. Tumoroids are treated 7 days after plating in 3D drops of basement membrane extract to allow for formation. After 7 days of treatment, the tumoroids were photographed (pictures only shown for T1) at a low magnification to follow the growth in the 3D BME drops (dotted circle) characterized by an increase in the size and at a high magnification to observe the variations in the morphology related to the different treatments. CTX and RSL3 alone do not affect the growth or the viability of tumoroids when compared to the untreated tumoroids. In contrast, the combination of CTX and RSL3 clearly reduces the size and the viability to 60 +- 4% and 59 +- 5% in tumoroids 1 and 2, respectively . The non-specific inducer of ferroptosis, metformin by itself, reduces the size and the viability of tumoroids 1 and 2 to 48+-6% and 31 +- 5%, respectively . The combination of CTX and metformin reduces the viability even further in tumoroid 1 but not in tumoroid 2 and is more efficient than CTX and RSL3. Reduced viability is associated with the appearance of debris in 3D BME drops . Finally, the exposure of tumoroids to CTX does not affect GPX4 expression . GPX4 is significantly reduced by RSL3 alone (30 +- 10%) and even further when RSL3 is combined with CTX (80 +- 9%). Similar results were obtained in T2 and T3 (data not shown). It is also significantly reduced by metformin alone (76 +- 14%) and totally lost when metformin is combined with CTX . Altogether, the data confirm the efficacy of targeting xCT and GPX4 in CTX-resistant tumors. 4. Discussion The dysregulation of EREG may contribute to the progression of various cancers including HNSCC and may be a putative mechanism of resistance to EGFR-targeted therapies. EREG is usually overexpressed in HNSCC and correlates with short overall survival and progression-free survival . EREG conducts an even more potent mitogenic signal than EGF in HNSCC mimicking EGFR oncogenic mutations . Job et al. recently described a subgroup of HPV-negative HNSCC named "basal" sharing molecular similarities such as the upregulation of genes involved in the EGFR signaling pathway including EREG and AREG . Cells sharing these characteristics appear to be more sensitive to EGFR-targeted therapy, with CTX being the least efficient. Because the suppression of EREG expression reduces cell survival, the authors suggested that cells may be addicted to an EREG feedback loop and that EREG should be considered as a functional biomarker for HNSCC sensitivity to EGFR blockade . In line with this study, we observed that HNSCC tumor cells expressing caveolin-1 could use EREG silencing, but not AREG silencing, to overcome oncogenic dependence on EGFR and develop resistance to CTX/irradiation combination therapy . The resistance was due, at least in part, to the silencing of the HIPPO pathway, leading to the activation of YAP/TAZ . The cross-suppression of both AREG and EREG has also been reported to lead to the emergence of CTX resistance, which is related to the loss of cell addiction to EGFR, compensated by the hyperactivation and addiction to FGFR3 in melanoma . We show here that the direct suppression of EREG expression reduces both EGFR expression and HNSCC basal cell survival. Rather than driving resistance to CTX, the loss of EREG reduces survival even further. While it cannot be excluded that long-term EREG silencing may lead to the emergence of CTX resistance, the acute targeting of EREG combined with CTX is effective in reducing cell survival and could be a feasible antitumor strategy for HNSCC. Fepixnebart (LY3016859, developed by Eli Lilly and Co.) is a monoclonal antibody that binds epiregulin and TGF-a and is well tolerated and efficient in neutralizing both targets . It is currently in phase II for back pain and neuropathic pain and in phase III for diabetic neuropathies. It would be interesting to determine its anti-tumor effect and adjuvant effect for EGFR-targeting therapies. Besides its autocrine feedback loop, EREG is also secreted by the component of the tumor microenvironment such as macrophages and cancer-associated fibroblasts (CAF) . Macrophages-derived EREG induces EGFR-TKI resistance in NSCLC, and CAF-derived EREG promotes OSCC invasion and metastasis through the induction of EMT. Thus, targeting EREG might not only prevent therapy resistance but also HNSCC progression. Aberrant metabolism and metabolism reprogramming represent malignant tumor hallmarks that are required for cancer cells to proliferate and progress. The metabolism of cancer cells is mainly based on nonoxidative glycolysis, followed by the fermentation of lactic acid to produce ATP, a phenomenon known as the Warburg effect. EREG/EGFR signaling enhances glycolysis by increasing glucose consumption, lactate production, extracellular acidification (ECAR) and the intracellular levels of ATP as well as by activating several glycolytic genes . However, HNSCC also depend on glutamine for producing energy , which is imported in cells by the Na+-glutamine/Na+-cysteine exchanger ASCT2. Besides serving as a source of carbon and nitrogen for macromolecule synthesis, glutamine provides a-ketoglutarate for the tricarboxylic acid (TCA) cycle and contributes to the production of the most powerful antioxidant, glutathione (GSH) (for a review, see ). The production of GSH also requires cysteine, which is imported into cells through the x-c or xCT cystine/glutamate antiporter. GSH serves as a cofactor of the glutathione peroxidase 4 (GPX4) to suppress destructive lipid reactive oxygen species (ROS). This pathway plays a key role in the regulation of ferroptosis, which is a regulated cell death triggered by an iron-dependent lipid peroxidation . Both GPX4 and x-c antiporter are crucial regulators of ferroptosis. Here, we show that silencing EREG as well as blocking EGFR lead to mitochondrial defects characterized by a reduction in ATP production, oxygen consumption (OCR) and ECAR. Combining EREG silencing with an EGFR blockade shifts cells from an energetic state toward a less metabolic phenotype. If the dysfunction of the mitochondria could affect cell survival through energetic stress, death could neither be attributed to apoptosis, which was undetectable, or to autophagy, which was inhibited. Mitochondria play a major role in regulating oxidative metabolism, are the main source of reactive oxygen species (ROS) and are the primary site of Fe2+ iron storage and utilization. Therefore, the dysregulation of mitochondria induced by the silencing of EREG and the inhibition of EGFR might alter the iron metabolism and generate a redox imbalance strong enough to trigger ferroptosis. Here, we show that the silencing of EREG combined with the blockade of EGFR lead to the accumulation of Fe2+ and lipid peroxides associated with the downregulation of GPX4. No ferroptosis could be achieved by the loss of EREG alone or by the treatment with CTX. Accordingly, although CTX downregulates ASCT2 via a CTX-dependent EGFR endocytosis, it does not alter survival by itself. However, it decreases the intracellular uptake of glutamine and the levels of GSH, which sensitizes HNSCC to ROS-induced death . In colorectal cancer cells, CTX neither affected proliferation or survival by itself, even though it inhibited NRF2 signaling (known to promote GPX4 and HO-1 transcription). By targeting NRF2/HO-1, CTX enhances RSL-3-induced ferroptosis . Further studies will be needed to determine if those molecular targets are also altered by the loss of EREG. However, we did show the inhibition of c-Myc and GPX4 expression in EREG-silenced cells treated by CTX. We previously reported that the oncogene c-Myc, a known target of EREG/EGFR, is downregulated by CTX in EREG-/caveolin-1-expressing cells . As c-Myc regulates ASCT2, LAT1, x-c antiporter as well as NRF2 , it deserves further investigations. Inducing ferroptosis seems to be an attractive potential anti-cancer strategy with broad clinical implications. Several preclinical studies show that ferroptosis inducers can synergize with traditional chemotherapeutics . They either target the depletion of the cellular antioxidant GSH through the x-c antiporter (Erastin) or directly target GPX4 (RSL3). Here, the loss of GPX4 is associated with the induction of ferroptosis in EREG-silenced cells exposed to CTX. Inhibiting GPX4 by using RSL3 reduces the survival of control cells exposed to CTX to levels equivalent to those observed in EREG-silenced cells exposed to CTX. However, RSL3 also sensitizes EREG-silenced cells to CTX to an even greater extent. The maximum decrease in cell viability corresponds to a steep decrease in GPX4. The data uncover the value of targeting GPX4 to effectively sensitize tumor cells to CTX. Accordingly, the overexpression of GPX4 was described in EGFR-TKI-resistant lung adenocarcinoma and colorectal cancers. RSL3 restores their sensitivity to EGFR-TKI . Similar results could be obtained using metformin, which is widely used for the treatment of type 2 diabetes mellitus (T2DM). Its use in T2DM has been associated with cancer incidence and mortality decreases, including in HNSCC (for a review, see ). This effect seems to be due to the reduction in circulating insulin, since both the insulin-IGF system and hyperglycemia have been associated with cancer risk. However, metformin also has a direct anti-tumor effect via the induction of energetic stress. It inhibits the mitochondrial respiratory chain complex I, leading to mitochondrial dysfunctions, changes in the levels of ROS and the iron homeostasis (for a review, see ). Acting independently of GPX4, metformin also downregulates SLC7A11 (the catalytic unit of the x-c cytine/glutamate antiporter), protein stability and expression by inhibiting UFM1 expression and the subsequent UFMylation of SLC7A11 . Metformin increases intracellular total ROS and lipid ROS levels and reduces intracellular GSH, which ultimately leads to ferroptosis . Metformin was also recently reported to induce ferroptosis in breast cancer by inhibiting autophagy . In our hands, metformin was more effective in reducing cell survival than RSL3, which is probably due to its multiple targets. Its co-administration with CTX in control cells demonstrated an adjuvant effect of metformin that is even more pronounced in EREG-silenced cells. Again, the decrease in cell viability is associated with the loss of GPX4. Although the targeting of the x-c antiporter or GPX4 sensitizes control cells to CTX treatment, the most striking effects are seen in cells where EREG is lost. Tumoroids are 3D tumor-resembling cellular clusters generated from primary patient material. They closely recapitulate the 3D tissue architecture, cellular composition and characteristics (including genetic and cellular intratumor heterogeneity as well as resistance to therapy) of the tumor from which they were derived, offering useful benefits over conventional 2D cell culture and 3D multicellular spheroids. They can be grown long-term without genetic or functional changes . The results obtained to date indicate that tumoroids respond in a largely consistent manner to the patients they were derived from , show heterogeneous sensitivities to standard treatments and might predict a patient's clinical outcome. Tumors hold promise for biomarker identification, drug discovery and aiding personalized therapy. For all these reasons, we have chosen to generate HNSCC tumoroids to validate our therapeutic approaches combining ferroptosis inducers with CTX. As a proof of concept, patient-derived HNSCC tumoroids showing resistance to CTX were co-treated with RSL3 or metformin. Tumoroids survival was drastically decreased with RSL3-CTX co-treatment and was almost completely abrogated in response to metformin-CTX. Metformin combined with CTX was therefore more effective in reducing viability than RSL3/CTX. As stated above, it might be related to the fact that RSL3 only targets GPX4, whereas metformin has a multitude of targets, some of which, such as SCL7A11, act further upstream in antioxidant signaling. The data also underline a heterogeneity in the response of tumoroids. Indeed, if metformin sensitizes tumoroid 1 to CTX, this is not the case for the second. An inhibition of EGFR expression by metformin could lead to this desensitization to CTX, as previously observed . Further studies will be necessary to understand this heterogeneity. The maximum diminution of cell viability is associated with the strongest reduction in GPX4 levels. 5. Conclusions To our knowledge, this is the first study reporting that a loss of EREG might sensitize HNSCC to CTX through the induction of ferroptosis. To date, only a high expression of EREG was considered to predict the response of a patient to anti-EGFR therapies. However, care should be taken, since emerging studies report that secreted EREG in the microenvironment might support therapy resistance and tumor progression . Thus, using EREG expression levels to identify patients likely to benefit from EGFR-TKI therapies could lead to the exclusion of some who would be better responders. Here, we propose combining ferroptosis inducers with CTX. Our data clearly show that the combination of both reduces the survival of tumors expressing EREG and that the effect is even more pronounced in tumors where EREG is lost. This is also the first study validating the efficacy of using ferroptosis inducers in combination with CTX to inhibit survival in a patient-derived tumoroid model resistant to CTX. In conclusion, a loss of EREG might be considered in clinical settings as a predictive biomarker for patients that might benefit the most from the combination of ferroptosis inducers and CTX. Acknowledgments We acknowledge the PIQ-Quest imaging platform. Supplementary Materials The following supporting information can be downloaded at: Table S1: Antibodies. Click here for additional data file. Author Contributions Participated in the research design: S.M. and A.J. Conducted experiments: A.J., S.F., O.C., M.B. and C.R. Performed data analysis: A.J., S.M. and S.F. Wrote or contributed to the writing of the manuscript: S.M., A.J., O.C. and M.D. Provided and processed samples: P.S., M.-P.C., T.D., A.D. and C.T. Provided drugs: N.E.-S. Data analysis: A.J., S.M., S.F. and R.V. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Ethics Committee of Heidelberg University (S-519/2019). The study was approved by the Scientific Committee of the tumor bank and the Department of Pathology of the CHU Strasbourg-Hautepierre (France). The patients have signed an informed consent form. Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement Not applicable. Conflicts of Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Figure 1 (A) Expression of EREG, EGFR, CAV1 and GAPDH was determined by Western blot in CAL27, CAL33 and SCC9 cells untransfected and transfected with siRNACtrl or siRNAEREG. Histograms represent the mean +- SEM (CAL27 and CAL33 n = 6-11 and SCC9 n = 6-13, with * p < 0.05 and *** p < 0.001) of the protein expression in siRNAEREG-transfected cells normalized with GAPDH. (B) Histograms show the surviving fraction of CAL27, CAL33 and SCC9 cells transfected with siRNACtrl or siRNAEREG and treated with solvent or 30 nM CTX. Data represent the mean +- SEM surviving fraction at day 8 post-transfection and post-treatment (n = 4 with * p < 0.05, ** p < 0.01 and *** p < 0.0001). (C) Curves show the percentage of confluence after normalization to day 0 of CAL27, CAL33 and SCC9 cells transfected with siRNACtrl or siRNA EREG and treated with solvent or 30 nM CTX. Data are represented as the mean +- SEM (small dots) of confluence at 168 h post-transfection and post-treatment (CAL27 and CAL33 n = 4-5 and SCC9 n = 6-5 with * p < 0.05, ** p < 0.01 and *** p < 0.001). (D) Expression of cleaved PARP and GAPDH was determined by Western blot in CAL27, CAL33 and SCC9 cells transfected with siRNACtrl or siRNAEREG and treated with solvent or CTX (30 nM). Figure 2 (A) Energetic map of CAL27, CAL33 and SCC9 cells transfected with siRNACtrl or siRNA EREG and treated with solvent or 30 nM CTX. Histograms show the ATP production of CAL27, CAL33 and SCC9 cells transfected with siRNACtrl or siRNAEREG and treated with solvent or 30 nM CTX. Data are represented as the mean +- SEM of ATP produced 144 h post-transfection and post-treatment (n = 8-9, with * p < 0.05, ** p < 0.001 and *** p < 0.0001). (B) Expression of ULK-1, Beclin1, LC3B and GAPDH was determined by Western blot in CAL27, CAL33 and SCC9 cells not transfected or transfected with siRNACtrl or siRNAEREG and treated with solvent or 30 nM CTX. Histograms represent the mean +- SEM (CAL27 n = 11, CAL33 n = 9 and SCC9 n = 10, with * p < 0.05, ** p < 0.01 and *** p < 0.001) of the protein expression normalized with GAPDH. Figure 3 (A) Histograms showing the levels of intracellular Fe2+ measured in CAL27, CAL33 and SCC9 cells transfected with siRNACtrl or siRNAEREG and treated with solvent or 30 nM CTX. Data are represented as the mean +- SEM of intracellular Fe2+ intensity (n = 8, with * p < 0.05, ** p < 0.001 and *** p < 0.0001). Pictures show (B) the intracellular Fe2+ production (in red, staining of individual CAL27 and SCC9 cells or clustered CAL33) and (C) the lipid peroxidation (in green) acquired by confocal microscopy in CAL27, CAL33 and SCC9 cells transfected with siRNACtrl or siRNAEREG and treated with solvent or 30 nM CTX (scale bar: 100 mm). Figure 4 Expression of GPX4 and GAPDH (A) and c-Myc and GAPDH (B) was determined by Western blot in CAL27, CAL33 and SCC9 cells untransfected or transfected with siRNACtrl or siRNAEREG and treated with solvent or 30 nM CTX. Histograms represent the mean +- SEM (CAL27 and CAL33 n = 10 and SCC9 n = 9, with * p < 0.05, ** p < 0.01 and *** p < 0.001) of the protein expression normalized with GAPDH. (C) Histograms show the surviving fraction of CAL27, CAL33 and SCC9 cells transfected with siRNACtrl or siRNAEREG and treated with 30 nM CTX, 5 mM RSL3 or a combination of both. Data represent the mean +- SEM surviving fraction at day 8 post-transfection and post-treatment (n = 4, with *** p < 0.0001). (D) Histograms show the surviving fraction of CAL27, CAL33 and SCC9 cells transfected with siRNACtrl or siRNAEREG and treated with 30 nM CTX, 1 mM metformin (METF) or a combination of both. Data represent the mean +- SEM surviving fraction at day 8 post-transfection and post-treatment (n = 4, with ** p < 0.01 and *** p < 0.0001). (E) Expression of GPX4 and GAPDH was determined by Western blot in CAL27 transfected with siRNACtrl or siRNAEREG and treated with 30 nM CTX, 5 mM RSL3 or a combination of both (upper panel) or with 30 nM CTX, 1 mM metformin (METF) or a combination of both (lower panel). Histograms represent the mean +- SEM (n = 4, with * p < 0.05, ** p < 0.01 and *** p < 0.001) of the protein expression normalized with GAPDH. Figure 5 (A) Patient-derived tumoroids T1 were treated with solvent, 30 nM CTX, 5 mM RSL3, 1 mM metformin (METF), a combination of CTX + RSL3 or CTX + metformin. Pictures were before treatment (left panel) and 7 days after treatment with 4x (middle panel) and 10x (right panel) magnification. (B) Viability was determined after 7 days of the culture. Each bar represents the mean +- SEM of the percentage of viability (n = 3 for T1 and T2, with * p < 0.05, ** p < 0.001, *** p < 0.0001. (C) Expression of GPX4 and GAPDH was determined by Western blot in T1 tumoroids treated with solvent, 30 nM CTX, 5 mM RSL3, 1 mM metformin (METF), a combination of CTX + RSL3 or CTX + metformin. Histograms represent the mean +- SEM (n = 4, with * p < 0.05, ** p < 0.01 and *** p < 0.001) of the protein expression normalized with GAPDH. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Sung H. Ferlay J. Siegel R.L. Laversanne M. Soerjomataram I. Jemal A. Bray F. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries CA Cancer J. Clin. 2021 71 209 249 10.3322/caac.21660 33538338 2. Bonner J.A. Harari P.M. Giralt J. Azarnia N. Shin D.M. Cohen R.B. Jones C.U. Sur R. Raben D. Jassem J. Radiotherapy plus cetuximab for squamous-cell carcinoma of the head and neck N. Engl. J. Med. 2006 354 567 578 10.1056/NEJMoa053422 16467544 3. Bonner J.A. Harari P.M. Giralt J. Cohen R.B. Jones C.U. Sur R.K. Raben D. Baselga J. Spencer S.A. Zhu J. Radiotherapy plus cetuximab for locoregionally advanced head and neck cancer: 5-year survival data from a phase 3 randomised trial, and relation between cetuximab-induced rash and survival Lancet Oncol. 2010 11 21 28 10.1016/S1470-2045(09)70311-0 19897418 4. Goldstein N.I. Prewett M. Zuklys K. Rockwell P. Mendelsohn J. Biological efficacy of a chimeric antibody to the epidermal growth factor receptor in a human tumor xenograft model Clin. Cancer Res. 1995 1 1311 1318 9815926 5. Masuelli L. Budillon A. Marzocchella L. Mrozek M.A. Vitolo D. Di Gennaro E. Losito S. Sale P. Longo F. Ionna F. Caveolin-1 overexpression is associated with simultaneous abnormal expression of the E-cadherin/alpha-beta catenins complex and multiple ErbB receptors and with lymph nodes metastasis in head and neck squamous cell carcinomas J. Cell. Physiol. 2012 227 3344 3353 10.1002/jcp.24034 22213373 6. Brand T.M. Iida M. Wheeler D.L. Molecular mechanisms of resistance to the EGFR monoclonal antibody cetuximab Cancer Biol. Ther. 2011 11 777 792 10.4161/cbt.11.9.15050 21293176 7. Ciardiello F. Tortora G. EGFR antagonists in cancer treatment N. Engl. J. Med. 2008 358 1160 1174 10.1056/NEJMra0707704 18337605 8. Burgy M. Jehl A. Conrad O. Foppolo S. Bruban V. Etienne-Selloum N. Jung A.C. Masson M. Macabre C. Ledrappier S. Cav1/EREG/YAP Axis in the Treatment Resistance of Cav1-Expressing Head and Neck Squamous Cell Carcinoma Cancers 2021 13 3038 10.3390/cancers13123038 34207120 9. Sahin U. Weskamp G. Kelly K. Zhou H.M. Higashiyama S. Peschon J. Hartmann D. Saftig P. Blobel C.P. Distinct roles for ADAM10 and ADAM17 in ectodomain shedding of six EGFR ligands J. Cell Biol. 2004 164 769 779 10.1083/jcb.200307137 14993236 10. Toyoda H. Komurasaki T. Uchida D. Takayama Y. Isobe T. Okuyama T. Hanada K. Epiregulin. A novel epidermal growth factor with mitogenic activity for rat primary hepatocytes J. Biol. Chem. 1995 270 7495 7500 10.1074/jbc.270.13.7495 7706296 11. Liu S. Wang Y. Han Y. Xia W. Zhang L. Xu S. Ju H. Zhang X. Ren G. Liu L. EREG-driven oncogenesis of Head and Neck Squamous Cell Carcinoma exhibits higher sensitivity to Erlotinib therapy Theranostics 2020 10 10589 10605 10.7150/thno.47176 32929368 12. Komurasaki T. Toyoda H. Uchida D. Morimoto S. Epiregulin binds to epidermal growth factor receptor and ErbB-4 and induces tyrosine phosphorylation of epidermal growth factor receptor, ErbB-2, ErbB-3 and ErbB-4 Oncogene 1997 15 2841 2848 10.1038/sj.onc.1201458 9419975 13. Roepstorff K. Grandal M.V. Henriksen L. Knudsen S.L. Lerdrup M. Grovdal L. Willumsen B.M. van Deurs B. Differential effects of EGFR ligands on endocytic sorting of the receptor Traffic 2009 10 1115 1127 10.1111/j.1600-0854.2009.00943.x 19531065 14. Freed D.M. Bessman N.J. Kiyatkin A. Salazar-Cavazos E. Byrne P.O. Moore J.O. Valley C.C. Ferguson K.M. Leahy D.J. Lidke D.S. EGFR Ligands Differentially Stabilize Receptor Dimers to Specify Signaling Kinetics Cell 2017 171 683 695.e18 10.1016/j.cell.2017.09.017 28988771 15. Wang Y. Jing Y. Ding L. Zhang X. Song Y. Chen S. Zhao X. Huang X. Pu Y. Wang Z. Epiregulin reprograms cancer-associated fibroblasts and facilitates oral squamous cell carcinoma invasion via JAK2-STAT3 pathway J. Exp. Clin. Cancer Res. 2019 38 274 10.1186/s13046-019-1277-x 31234944 16. Sunaga N. Kaira K. Epiregulin as a therapeutic target in non-small-cell lung cancer Lung Cancer 2015 6 91 98 10.2147/LCTT.S60427 28210154 17. Cheng W.L. Feng P.H. Lee K.Y. Chen K.Y. Sun W.L. Van Hiep N. Luo C.S. Wu S.M. The Role of EREG/EGFR Pathway in Tumor Progression Int. J. Mol. Sci. 2021 22 12828 10.3390/ijms222312828 34884633 18. Shigeishi H. Higashikawa K. Hiraoka M. Fujimoto S. Mitani Y. Ohta K. Takechi M. Kamata N. Expression of epiregulin, a novel epidermal growth factor ligand associated with prognosis in human oral squamous cell carcinomas Oncol. Rep. 2008 19 1557 1564 18497965 19. Kogashiwa Y. Inoue H. Kuba K. Araki R. Yasuda M. Nakahira M. Sugasawa M. Prognostic role of epiregulin/amphiregulin expression in recurrent/metastatic head and neck cancer treated with cetuximab Head Neck 2018 40 2424 2431 10.1002/hed.25353 30302873 20. Hu K. Li S.L. Gan Y.H. Wang C.Y. Yu G.Y. Epiregulin promotes migration and invasion of salivary adenoid cystic carcinoma cell line SACC-83 through activation of ERK and Akt Oral Oncol. 2009 45 156 163 10.1016/j.oraloncology.2008.04.009 18620900 21. Job S. Reynies A. Heller B. Weiss A. Guerin E. Macabre C. Ledrappier S. Bour C. Wasylyk C. Etienne-Selloum N. Preferential Response of Basal-Like Head and Neck Squamous Cell Carcinoma Cell Lines to EGFR-Targeted Therapy Depending on EREG-Driven Oncogenic Addiction Cancers 2019 11 795 10.3390/cancers11060795 31181806 22. Driehuis E. Kretzschmar K. Clevers H. Establishment of patient-derived cancer organoids for drug-screening applications Nat. Protoc. 2020 15 3380 3409 10.1038/s41596-020-0379-4 32929210 23. Bhutia Y.D. Babu E. Ramachandran S. Ganapathy V. Amino Acid transporters in cancer and their relevance to "glutamine addiction": Novel targets for the design of a new class of anticancer drugs Cancer Res. 2015 75 1782 1788 10.1158/0008-5472.CAN-14-3745 25855379 24. Wu X.Y. Xu W.W. Huan X.K. Wu G.N. Li G. Zhou Y.H. Najafi M. Mechanisms of cancer cell killing by metformin: A review on different cell death pathways Mol. Cell. Biochem. 2022 478 197 214 10.1007/s11010-022-04502-4 35771397 25. Yang J. Zhou Y. Xie S. Wang J. Li Z. Chen L. Mao M. Chen C. Huang A. Chen Y. Metformin induces Ferroptosis by inhibiting UFMylation of SLC7A11 in breast cancer J. Exp. Clin. Cancer Res. 2021 40 206 10.1186/s13046-021-02012-7 34162423 26. Oliveras-Ferraros C. Cufi S. Queralt B. Vazquez-Martin A. Martin-Castillo B. de Llorens R. Bosch-Barrera J. Brunet J. Menendez J.A. Cross-suppression of EGFR ligands amphiregulin and epiregulin and de-repression of FGFR3 signalling contribute to cetuximab resistance in wild-type KRAS tumour cells Br. J. Cancer 2012 106 1406 1414 10.1038/bjc.2012.103 22491422 27. Sloan-Lancaster J. Raddad E. Deeg M.A. Eli M. Flynt A. Tumlin J. Evaluation of the Safety, Pharmacokinetics, Pharmacodynamics, and Efficacy after Single and Multiple Dosings of LY3016859 in Healthy Subjects and Patients with Diabetic Nephropathy Clin. Pharmacol. Drug Dev. 2018 7 759 772 10.1002/cpdd.436 29385323 28. Ma S. Zhang L. Ren Y. Dai W. Chen T. Luo L. Zeng J. Mi K. Lang J. Cao B. Epiregulin confers EGFR-TKI resistance via EGFR/ErbB2 heterodimer in non-small cell lung cancer Oncogene 2021 40 2596 2609 10.1038/s41388-021-01734-4 33750895 29. Poliakova M. Aebersold D.M. Zimmer Y. Medova M. The relevance of tyrosine kinase inhibitors for global metabolic pathways in cancer Mol. Cancer 2018 17 27 10.1186/s12943-018-0798-9 29455660 30. He M. Jin Q. Chen C. Liu Y. Ye X. Jiang Y. Ji F. Qian H. Gan D. Yue S. The miR-186-3p/EREG axis orchestrates tamoxifen resistance and aerobic glycolysis in breast cancer cells Oncogene 2019 38 5551 5565 10.1038/s41388-019-0817-3 30967627 31. Zhang Z. Liu R. Shuai Y. Huang Y. Jin R. Wang X. Luo J. ASCT2 (SLC1A5)-dependent glutamine uptake is involved in the progression of head and neck squamous cell carcinoma Br. J. Cancer 2020 122 82 93 10.1038/s41416-019-0637-9 31819178 32. Yang W.S. Stockwell B.R. Ferroptosis: Death by Lipid Peroxidation Trends Cell Biol. 2016 26 165 176 10.1016/j.tcb.2015.10.014 26653790 33. Lu H. Li X. Lu Y. Qiu S. Fan Z. ASCT2 (SLC1A5) is an EGFR-associated protein that can be co-targeted by cetuximab to sensitize cancer cells to ROS-induced apoptosis Cancer Lett. 2016 381 23 30 10.1016/j.canlet.2016.07.020 27450723 34. Tao X. Lu Y. Qiu S. Wang Y. Qin J. Fan Z. AP1G1 is involved in cetuximab-mediated downregulation of ASCT2-EGFR complex and sensitization of human head and neck squamous cell carcinoma cells to ROS-induced apoptosis Cancer Lett. 2017 408 33 42 10.1016/j.canlet.2017.08.012 28823958 35. Yang J. Mo J. Dai J. Ye C. Cen W. Zheng X. Jiang L. Ye L. Cetuximab promotes RSL3-induced ferroptosis by suppressing the Nrf2/HO-1 signalling pathway in KRAS mutant colorectal cancer Cell Death Dis. 2021 12 1079 10.1038/s41419-021-04367-3 34775496 36. Zhou R.P. Chen Y. Wei X. Yu B. Xiong Z.G. Lu C. Hu W. Novel insights into ferroptosis: Implications for age-related diseases Theranostics 2020 10 11976 11997 10.7150/thno.50663 33204324 37. Roh J.L. Kim E.H. Jang H.J. Park J.Y. Shin D. Induction of ferroptotic cell death for overcoming cisplatin resistance of head and neck cancer Cancer Lett. 2016 381 96 103 10.1016/j.canlet.2016.07.035 27477897 38. Zhang C. Wang C. Yang Z. Bai Y. Shukuya T. Poh M.E. Ekman S. Li J. Xu Y. Deng S. Identification of GPX4 as a therapeutic target for lung adenocarcinoma after EGFR-TKI resistance Transl. Lung Cancer Res. 2022 11 786 801 10.21037/tlcr-22-318 35693278 39. Vilaseca I. Fuster G. Aviles-Jurado F.X. The impact of diabetes in head and neck cancer Curr. Opin. Otolaryngol. Head Neck Surg. 2020 28 107 111 10.1097/MOO.0000000000000606 32022731 40. Figueiredo R.A. Weiderpass E. Tajara E.H. Strom P. Carvalho A.L. de Carvalho M.B. Kanda J.L. Moyses R.A. Wunsch-Filho V. Diabetes mellitus, metformin and head and neck cancer Oral Oncol. 2016 61 47 54 10.1016/j.oraloncology.2016.08.006 27688104 41. Ogunsakin A. Infield J. Zuber J. Solomon S.S. Metformin Associated with Improved Outcomes in Diabetic Patients with Laryngeal and Oropharyngeal Carcinoma Am. J. Med. Sci. 2018 356 574 575 10.1016/j.amjms.2018.09.002 30342717 42. Tsou Y.A. Chang W.C. Lin C.D. Chang R.L. Tsai M.H. Shih L.C. Staniczek T. Wu T.F. Hsu H.Y. Chang W.D. Metformin Increases Survival in Hypopharyngeal Cancer Patients with Diabetes Mellitus: Retrospective Cohort Study and Cell-Based Analysis Pharmaceuticals 2021 14 191 10.3390/ph14030191 33652909 43. Pernicova I. Korbonits M. Metformin-mode of action and clinical implications for diabetes and cancer Nat. Rev. Endocrinol. 2014 10 143 156 10.1038/nrendo.2013.256 24393785 44. Chen J. Qin C. Zhou Y. Chen Y. Mao M. Yang J. Metformin may induce ferroptosis by inhibiting autophagy via lncRNA H19 in breast cancer FEBS Open Bio 2022 12 146 153 10.1002/2211-5463.13314 34644456 45. Fatehullah A. Tan S.H. Barker N. Organoids as an in vitro model of human development and disease Nat. Cell Biol. 2016 18 246 254 10.1038/ncb3312 26911908 46. Sato T. Vries R.G. Snippert H.J. Van De Wetering M. Barker N. Stange D.E. Van Es J.H. Abo A. Kujala P. Peters P.J. Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche Nature 2009 459 262 265 10.1038/nature07935 19329995 47. Wensink G.E. Elias S.G. Mullenders J. Koopman M. Boj S.F. Kranenburg O.W. Roodhart J.M. Patient-derived organoids as a predictive biomarker for treatment response in cancer patients NPJ Precis. Oncol. 2021 5 30 10.1038/s41698-021-00168-1 33846504 48. Driehuis E. Kolders S. Spelier S. Lohmussaar K. Willems S.M. Devriese L.A. de Bree R. de Ruiter E.J. Korving J. Begthel H. Oral Mucosal Organoids as a Potential Platform for Personalized Cancer Therapy Cancer Discov. 2019 9 852 871 10.1158/2159-8290.CD-18-1522 31053628 49. Tanaka N. Osman A.A. Takahashi Y. Lindemann A. Patel A.A. Zhao M. Takahashi H. Myers J.N. Head and neck cancer organoids established by modification of the CTOS method can be used to predict in vivo drug sensitivity Oral Oncol. 2018 87 49 57 10.1016/j.oraloncology.2018.10.018 30527243 50. Wang W.M. Yang S.S. Shao S.H. Nie H.Q. Zhang J. Su T. Metformin Downregulates the Expression of Epidermal Growth Factor Receptor Independent of Lowering Blood Glucose in Oral Squamous Cell Carcinoma Front. Endocrinol. 2022 13 828608 10.3389/fendo.2022.828608 35222283 |
PMC10000619 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050813 cells-12-00813 Article 17-Estradiol Protects against HIV-1 Tat-Induced Endolysosome Dysfunction and Dendritic Impairments in Neurons Datta Gaurav Methodology Formal analysis Writing - original draft + Miller Nicole M. Methodology Formal analysis Writing - original draft + Chen Xuesong Conceptualization Writing - review & editing * Levi Giovanni Academic Editor Department of Biomedical Sciences, University of North Dakota School of Medicine and Health Sciences, Grand Forks, ND 58203, USA * Correspondence: [email protected] + These authors contributed equally to this work. 06 3 2023 3 2023 12 5 81330 1 2023 03 3 2023 04 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). HIV-1 Tat continues to play an important role in the development of HIV-associated neurocognitive disorders (HAND), which persist in 15-55% of people living with HIV even with virological control. In the brain, Tat is present on neurons, where Tat exerts direct neuronal damaging effects by, at least in part, disrupting endolysosome functions, a pathological feature present in HAND. In this study, we determined the protective effects of 17a-estradiol (17aE2), the predominant form of estrogen in the brain, against Tat-induced endolysosome dysfunction and dendritic impairment in primary cultured hippocampal neurons. We demonstrated that pre-treatment with 17aE2 protected against Tat-induced endolysosome dysfunction and reduction in dendritic spine density. Estrogen receptor alpha (ERa) knockdown impairs the ability of 17aE2 to protect against Tat-induced endolysosome dysfunction and reduction in dendritic spine density. Furthermore, over-expressing an ERa mutant that fails to localize on endolysosomes impairs 17aE2's protective effects against Tat-induced endolysosome dysfunction and reduction in dendritic spine density. Our findings demonstrate that 17aE2 protects against Tat-induced neuronal injury via a novel ERa-mediated and endolysosome-dependent pathway, and such a finding might lead to the development of novel adjunct therapeutics against HAND. HIV-1 Tat endolysosomes 17a-estradiol estrogen receptors alpha dendritic spine National Institute of Mental HealthMH100972 MH105329 MH119000 This work was supported by the National Institute of Mental Health (MH100972, MH105329, and MH119000). pmc1. Introduction The effective suppression of HIV-1 replication by combined antiretroviral therapy (ART) has significantly decreased mortality rates and improved the lifespan and quality of life of people living with HIV (PLWH). However, HIV-associated neurocognitive disorders (HAND) persist in 15-55% of PLWH even with ART . One of the key pathological features of HAND that correlate closely with neurocognitive impairment is synaptodendritic impairments that occur in discrete brain regions such as the prefrontal cortex and hippocampus . Although the underlying mechanisms of synaptodendritic impairments in HAND remain elusive in the ART era, various HIV-related factors are involved, including the presence of low levels of HIV-1 from the reactivation of latent infected cells , HIV-1 viral proteins , ART drugs , and substance use disorders . Reversible synaptodendritic impairment resulting from these HIV-related factors contribute to the development of mild cognitive impairment in PLWH . At the cellular level, many of these HIV-1 related factors could induce endolysosome dysfunction, a pathological feature that is present in the post-mortem brains of HAND patients . As acidic organelles that are responsible for degradation and membrane trafficking, endolysosomes are especially critical for neuronal function because neurons are long-lived post-mitotic cells with extensive processes. As long-lived post-mitotic cells, neurons cannot remove non-degraded or partially degraded and potentially toxic constituents through cell division, and thus quality control over a large volume of cytoplasm and membrane relies on the efficient proteolytic degradation capability of endolysosomes . Furthermore, endolysosomes in neurons serve as a robust and dynamic intracellular vesicular trafficking machinery, which is required for establishing and maintaining the axonal and somatodendritic membrane domains. As such, endolysosomes play an important role in modulating dendritic and synaptic plasticity . Thus, endolysosome dysfunction could lead to synaptodendritic impairments. Conversely, enhancing endolysosome function and/or alleviating endolysosome dysfunction represents a promising therapeutic strategy against HAND. Among various HIV-related factors, HIV-1 Tat has been consistently demonstrated to play an important role in the pathogenesis of HAND. As an essential protein for HIV-1 viral transcription, Tat can be actively secreted from infected cells . Significantly, current anti-HIV strategies do not block the secretion of Tat , and brain levels of Tat remain elevated despite virological control . In the brain, Tat has been detected on neurons , where Tat can exert neurotoxic effects directly and induce synaptodendritic impairment . Such Tat-mediated direct neuronal damaging effects depend, at least in part, on the internalization of Tat into endolysosomes . We have shown that Tat induces endolysosome dysfunction in neurons . Thus, Tat-induced endolysosome dysfunction could contribute to the direct neuronal damaging effect of Tat. Based on our recent findings that 17a-estradiol (17aE2), the predominant form of estrogen in the brain, exerts enhancing effects on endolysosome function and dendritic spine formation via estrogen receptor alpha (ERa) , we determined the extent to which 17aE2 protects against HIV-1 Tat-induced endolysosome dysfunction and dendritic impairment in primary cultured hippocampal neurons. 2. Materials and Methods Cell Cultures: ERa-expressing immortalized mouse E-18 neurons CLU199 (Cellutions Biosystems, Cedarlane, ON, Canada) were grown and maintained in 1x DMEM with 25 mM glucose, 10% fetal bovine serum (FBS), and 1% penicillin/streptomycin and maintained in a 37 degC incubator with 5% CO2. Cells from passages 3-7 were used for all experiments in this study. Primary cultured mouse hippocampal neurons were obtained from BrainBits LLC (C57EHP, Springfield, IL, USA), and the neurons were grown as per the manufacturer's instructions. The neurons, at a density of 100,000 cells per well, were plated on either poly-D-lysine-coated 12 mm coverslips (GG-12-PDL, Neuvitro Corporation, Vancouver, WA, Canada) or on ploy-D-lysine-coated 35 mm glass bottom dishes (P35GC-0-10-C, MatTek Life Sciences, Ashland, MA, USA). NbActiv medium (BrainBits) was used for both the plating and maintenance, and half the medium was replaced with new medium every 3-4 days. The neurons at DIV 12-17 were used for experiments. Dendritic Spine Live Imaging: Mouse hippocampal neurons expressing cytoplasmic GFP (BacMam GFP, B10383, Thermo Fisher, Waltham, MA, USA) were used, and images were acquired with a Zeiss LSM800 Confocal microscope using the 63X objective. The neurons were treated with recombinant HIV-1 Tat (ImmunoDx, Woburn, MA, USA) and/or 17aE2 (10 nM, Tocris, Minneapolis, MN, USA), according to experimental design, with heat-inactivated Tat (95 degC for 1 h) used as a control. Images were acquired at 1 min intervals for 10 min with a z-stack set at 0.5 mm. From the acquired confocal images at 0 (t-0) and 10 min (t-10), the dendritic spines were reconstructed in 3D with Imaris 9.5. using the filaments module. The percentage of spines lost/gained over a 10 min period was plotted. For each treatment, at least 7-10 neurons were imaged, and more than 5000 spines were analyzed. Experiments were repeated three times with different cultures of neurons. Spine classification criteria published in an earlier study were used. Immunostaining: Following treatment as per the experimental design, the neurons were fixed with 4% PFA at RT for 20 min, and 0.05-0.1% Triton X-100 treatment for 10 min was used to permeabilize the membranes. Following blocking (3% BSA with 1% normal goat or donkey serum in PBS) for 1 h, the neurons were incubated with MAP2 antibody (1:500, ab32454, Abcam, Waltham, MA, USA) overnight at 4 degC. After washing, the neurons were incubated with Alexa Fluor 488 goat anti-rabbit (1:500, Thermo Fisher) for 1 h. The cells on the coverslips were mounted onto microscope slides (Fisher Scientific) and the cells in the 35 mm dishes were used directly with ProLong Gold Antifade (P36930, Thermo Fisher). Cells stained only with primary antibodies (background controls) or only with secondary antibodies were used as controls for eliminating autofluorescence and bleed-through between channels. Dendritic Spine Density Measurement in Fixed Neurons: Neurons stained with MAP2 were used for assessing dendritic morphology and length. Neurons stained with phalloidin were used for the measurement of dendritic spines as phalloidin labels F-actin, which is enriched at spine heads and serves as a marker for dendritic spines. Following the acquisition of confocal images under a 63X objective using a Zeiss LSM800 confocal microscope, the dendrites and spines were reconstructed in 3D with Imaris 9.5 using the filaments module. Multiple dendrites from different neurons were reconstructed and the spines were classified using a previously defined criterion. The number of dendritic spines/10 mm of primary and secondary neurites was calculated and served as a measure of dendritic spine density. Endolysosome pH Measurement: The luminal pH of the endolysosomes in the CLU199 cells was measured ratiometrically using a pH sensitive and pH insensitive dextran, as described previously . CLU199 cells plated on 35 mm glass bottom dishes were incubated with a pH insensitive dextran Texas Red (10 mg/mL, D1863, Thermo Fisher) and a pH sensitive pHrodo Green dextran (10 mg/mL, P35368, Thermo Fisher). Following a 24 h incubation period, the dextran-containing medium was replaced with warm Hibernate E Low Fluorescence Medium (HELF, Brainbits, Springfield, Illinois, USA) for imaging under a Zeiss LSM800 confocal microscope. Images were acquired with z-stacks at 1 mm intervals and fluorescence emission at 533 nm for Green dextran and 615 nm for Texas Red dextran. For pH calibration, a standard curve was determined in another set of cells incubated with buffers of different pH with the addition of 10 mM nigericin and 20 mM Monensin in Hibernate E Flow Fluorescence (HELF) Medium, which was performed as per the intracellular pH calibration kit (P35379, Thermo Fisher) instructions. Using this standard curve, the fluorescence ratio at 615/533 was converted to pH. In these experiments, a 40X objective was used, a total of 5 fields with at least 5-10 cells per field were imaged, and the experiments repeated independently three times. Active Cathepsin D Staining: Active cathepsin D was identified in cells stained with BODIPY-FL Pepstatin A (P12271, Thermo Fisher), and total endolysosomes were identified with LysoTracker Red DND-99. Briefly, the cells were incubated with LysoTracker Red DND-99 (10 nM) and with BODIPY-FL Pepstatin A (1 mM) for 30 min at 37 degC. Following washing, fresh, warm Hibernate E low fluorescence (HELF) medium was added for imaging under the 63X objective of a Zeiss LSM 800 confocal microscope using 0.5 mm z-stack intervals. Twenty-five to thirty cells per treatment group were imaged, and the experiments were repeated independently three times. Total endolysosomes (LysoTracker Red) and active cathepsin D positive endolysosomes (BODIPY-FL Pepstatin A) were reconstructed as spots using Imaris 9.5 software, and the percentage of active cathepsin D positive endolysosomes vs. total endolysosomes was calculated. ERa over-expression and siRNA knockdown: ERa-HA and ERa C451A-HA cloned into pCMV6-AC-3HA vectors were obtained from Origene Technologies (Rockville, MD, USA). Lipofectamine 2000 transfection Reagent (11668019, ThermoFisher) was used for transient transfections, and cells plated on either 35 mm glass bottom dishes or 12 mm coverslips were transfected with 1-2 mg of plasmid DNA/well in Opti-MEM Reduced Serum medium (31985062, Thermo Fisher) for 48 h. On-Target plus mouse Esr1 (Entrez Gene ID-13982) siRNA-SMART pool (Horizon Discovery Biosciences Limited) was used for the siRNA knockdown of ERa. The following target sequences were used: CCUACUACCUGGAGAACGA, GAAAGGCGGCAUACGGAAA, GUCCAGCAGUAACGAGAAA, and GGGCUAAAUCUUGGUAACA. Accell1 transfection media (B-005000, Dharmacon/Horizon Discovery Biosciences Limited, Waterbeach, Cambridge, UK) and DharmaFECT 1 (T-2001-02, Dharmacon) were used as transfection reagents, and siRNA at a concentration of 50 nM was used. Following 48 h of siRNA treatment, the transfection efficiency was determined using immunoblotting. Immunoblotting: CLU199 cells or primary mouse hippocampal neurons plated on Poly D-lysine-coated 6-well plates at 1 x 106/well or 0.5 x 106/well, respectively, were used for immunoblotting. Following washing and harvesting, the cells were lysed in ice-cold 1 x IP lysis buffer (Thermo Fisher) with Protease Inhibitor Cocktail (Pierce) for 30 min on ice. Cell lysates were cleared with centrifugation (13,000x g for 10 min at 4 degC), and the protein concentrations of the collected supernatants were determined using Precision Red Advanced Assays (Cytoskeleton Inc., Denver, CO, USA). The proteins (50 mg) were separated with 4-20% SDS-PAGE gel and transferred to PVDF membranes (Thermo Fisher). The membranes were kept in blocking (LiCor) for 1 h, followed by overnight incubation with N-terminal anti-ERa antibody (1:500, sc-8002 Santa Cruz) at 4 degC, with GAPDH as a loading control (1:2000, ab8245, Abcam). Following washing with TBS, the membranes were incubated with LiCor IR secondary antibodies (1:5000). After washing, the blots were imaged, and the densitometries of the blots were quantified and analyzed using Li-Cor Odyssey Fc Imager. Statistical Analysis: All data were expressed as mean +- SEM. For each experiment, the "n" was specified in the figure legend. GraphPad Prism 9.0 (GraphPad Software, Inc., Boston, MA, USA) was used for data analysis and the preparation of all the graphs. For the statistical analysis, Student's t-tests (two-tailed) or one-way or two-way ANOVA with Tukey's post-hoc tests were used to calculate the statistical significance between the groups, and an alpha of 0.05 was used as the cutoff for significance. 3. Results 3.1. Tat Induces Dendritic Spine Impairment and Endolysosome Dysfunction As a secreted HIV-1 protein, Tat can enter all CNS cells via endocytosis by interacting with cell surface receptors or proteins . Tat has been detected on neurons , astrocytes , and microglia in the brain, where Tat can exert neurotoxic effects directly and indirectly . Here, we determined the extent to which Tat affects dendritic length and dendritic spines in primary hippocampal neurons. Morphological changes to the structure or configuration of dendrites were assessed with MAP2 staining, and morphological changes in F-actin-laden dendritic spines were assessed with Alexa Fluor-488 Phalloidin in neurons treated with different concentrations of HIV-1 Tat (10 nM, 50 nM, and 100 nM) for 48 h. Although Tat did not significantly reduce dendritic length, Tat (100 nM) induced a significant decrease in dendritic spine density , a finding that is consistent with those of other studies . Such Tat-mediated direct neuronal damaging effects depend, in part, on the internalization of Tat into endolysosomes . Thus, internalized Tat can disturb endolysosome function directly. Indeed, we have shown that Tat induces endolysosome dysfunction in rat neurons . Here, we determined the effects of Tat on endolysosome function in mouse hippocampal neurons. We demonstrated that Tat (100 nM for 10 min) induced endolysosome de-acidification . Endolysosome de-acidification can lead to the abnormal accumulation of undegraded materials, and subsequently to endolysosome enlargement . We therefore measured changes in endolysosome size and demonstrated that Tat (100 nM for 48 h) significantly increased the size of the endolysosomes . It has been shown that endolysosome dysfunction in neurons affects dendritic spine dynamics , and thus Tat-induced endolysosome dysfunction could lead to the observed dendritic spine impairments. 3.2. 17aE2 Prevents Tat-Induced Dendritic Damage As the predominant form of estrogen in the brain , 17aE2 has been shown to promote dendritic spine and synapse formation . Here, we evaluated the ability of 17aE2 to protect against Tat-induced dendritic damage. First, EGFP-expressing neurons pre-treated with 17aE2 (10 nM for 10 min) were treated with Tat (100 nM) for 10 min, and dynamic changes in the dendritic spines were monitored using time-lapse imaging. The same dendrite at 0 min (t-0) and 10 min (t-10) of Tat treatment was imaged. The net gain/loss in various types of spines over the 10 min treatment was calculated to assess dendritic spine turnover. We demonstrated that Tat treatment resulted in a rapid decrease in mushroom, stubby, and long/thin types of spines . Pre-treatment with 17aE2 significantly prevented the Tat-induced decreases in mushroom and stubby types of spines, but not long/thin types of spines . These findings show that 17aE2 rapidly modulates the plasticity of dendritic spines in primary hippocampal neurons and that 17aE2 blocks Tat-induced dendritic impairment. Next, we determined whether this neuroprotective effect of 17aE2 would persist over a longer period. We demonstrated that pre-treatment with 17aE2 (10 nM, 6h) significantly prevented Tat (100 nM for 48 h)-induced reductions in dendritic spine density . 3.3. 17aE2 Prevents Tat-Induced Endolysosome Dysfunction Because dendritic spine remodeling is intrinsically linked to lysosomal function in neurons , we determined the ability of 17aE2 to prevent Tat-induced endolysosome dysfunction. We checked the levels of active cathepsin D (CatD) in neuronal endolysosomes (a functional outcome of endolysosome de-acidification) using the dye BODIPY FL-Pepstatin A, which binds to the active site of cathepsin D when its active site is exposed in an acidic environment. The proportion of active endolysosomes (active CatD positive) to total endolysosomes identified with LysoTracker Red (LTR) decreased upon the addition of Tat at 100 nM for 30 min . We demonstrated that 17aE2 pre-treatment (10 nM, 10 min) increased the percentage of active endolysosomes (CatD positive) and significantly prevented Tat-induced decreases in the percentage of active endolysosomes (CatD positive) . This is consistent with our previous findings that 17aE2 acidifies endolysosomes and enhances endolysosome function . 3.4. 17aE2 Enhances Endolysosome Function via ERa In the hippocampus and hippocampal neurons, where the nuclear presence of estrogen receptors (ERs) is sparse , extranuclear membrane-bound ERs have been implicated in the enhancing effects of estrogen on cognition and synaptic function . Such extranuclear membrane-bound ERs have been shown to have distinct subcellular distribution, with ERa expression on endolysosomes , ERb on mitochondria , and G-protein coupled estrogen receptor 1 on the endoplasmic reticulum . We have consistently demonstrated that that ERa is primarily expressed on endolysosomes, with ERa co-localizing with LAMP1-positive endolysosomes in hippocampal neurons . Thus, it is possible that 17aE2 could activate endolysosome-localized ERa and initiate endolysosome-dependent actions. To explore such a possibility, we first knocked down the expression of ERa with siRNA in mouse hippocampal neurons (CLU199) and observed a 70% reduction in ERa protein levels . We then measured endolysosome pH using a ratio-metric method which combines the use of the pH-sensitive pHrodo-dextran and the pH-insensitive Texas-Red-dextran. We demonstrated that the endolysosome pH of ERa knockdown (ERaKD) neurons was significantly increased compared with that of scrambled siRNA-treated neurons (ERa scr) . Given that 17aE2 can be endogenously produced by neurons , our findings suggest ERa knockdown could impair the endolysosome-enhancing effects of endogenously produced 17aE2. Furthermore, we demonstrated that exogenously added 17aE2 was less able to induce endolysosome acidification in ERa KD than in ERa scr cells . Together, our findings suggest that ERa is necessary for the endolysosome-enhancing effects of 17aE2. 3.5. 17aE2 Protects against Tat-Induced Endolysosome Dysfunction and Impairment in Dendritic Spines via ERa Having shown that 17aE2 acidifies endolysosomes via ERa, we explored the extent to which ERa knockdown affects the protective effects of 17aE2 against Tat-induced endolysosome dysfunction. Here, endolysosome function was assessed using alterations in the percentage of active CatD-positive endolysosomes in both ERa scr and ERa KD CLU199 cells . Pre-treatment with 17aE2 significantly prevented Tat-induced reduction in the percentage of active endolysosomes (CatD-positive) in ERa scr cells, but not in ERa KD cells . Our findings suggest that 17aE2 protects against Tat-induced endolysosome dysfunction via ERa. Because ERa is present in dendritic spines , and because lysosomal activity and its mobility along the dendrites is involved in dendritic spine dynamics , we next determined whether ERa knockdown affects the protective effects of 17aE2 against Tat-induced dendritic spine impairment. Here, we measured the dendritic spine density in ERa scr and ERa KD neurons treated with Tat (100 nM for 30 min) in the presence or absence of 17aE2 pre-treatment . We demonstrated that 17aE2 pre-treatment resulted in significantly increased dendritic spine density in Tat-treated neurons, but only in neurons treated with scrambled siRNA (ERa scr), not in ERa knockdown (ERa KD) neurons . Thus, our findings suggest that ERa mediates the protective effects of 17aE2 against Tat-induced dendritic spine impairment. Furthermore, ERa-mediated endolysosome enhancement could underlie the protection afforded by 17aE2 against Tat-induced dendritic impairments. 3.6. 17aE2 Protects against Tat-Induced Endolysosome Dysfunction and Impairment in Dendritic Spines via Endolysosome Localization of ERa We have shown that ERa is localized predominantly on Rab7-postive endolysosomes and, to a lesser extent, on LAMP1-positive endolysosomes . Given that the membrane localization of ERa depends on palmitoylation , we generated a mouse-specific ERa mutant (C451A) which lacks the known palmitoylation site. We found that this ERa mutant failed to localize on Rab7-positive endolysosomes in mouse neuronal cells . Furthermore, we demonstrated that 17aE2 exerts its enhancing effect on endolysosomes via endolysosome-localized ERa . Because we had demonstrated that ERa is essential for the protective effect of 17aE2 against Tat-induced endolysosome dysfunction and damage of dendritic spines, we determined further whether the endolysosome localization of ERa mediates the above mentioned protective effects of 17aE2. To investigate whether the endolysosome localization of ERa mediates the protective effect of 17aE2 against Tat-induced endolysosome dysfunction, we overexpressed ERa C451A and determined the percentage of active endolysosomes (CatD-positive) in CLU199 mouse neuronal cells treated with Tat (100 nM for 30 min), both in the absence and presence of 17aE2 pre-treatment . Because we have demonstrated that 17aE2 treatment increases the localization of ERa on endolysosomes , we reasoned that the over-expressed ERa C451A mutant, which fails to localize on endolysosomes, would compete the binding of 17aE2 to WT ERa. As such, the over-expressed ERa C451A mutant would block the 17aE2-induced localization of WT ERa to the endolysosomes, and thus attenuate the enhancing effect of 17aE2 on endolysosomes. As expected, we demonstrated that in cells expressing wild-type ERa (ERa WT), pre-treatment with 17aE2 increased the percentage of CatD-positive active endolysosomes when challenged with Tat . In contrast, in cells overexpressing ERa C451A, 17aE2 pre-treatment failed to increase the percentage of active endolysosomes (CatD-positive) when challenged with Tat . To further explore whether the endolysosome localization of ERa contributes to the protection afforded by 17aE2 against Tat-induced dendritic damage, we measured the density of the dendritic spines in the ERa C451A over-expressing primary hippocampal neurons that were treated with Tat, both with and without 17aE2 pre-treatment. As with our earlier results with ERa KD, over-expressing the ERa C451A mutant impaired the ability of 17aE2 to increase the density of the dendritic spines in the Tat-treated neurons . Thus, our findings suggest that ERa localized on endolysosomes is critical for 17aE2-mediated protection against Tat-induced endolysosome dysfunction and impairment in dendritic spines. 4. Discussion The present study explores the protective role of 17aE2 against HIV-1 Tat-induced endolysosome dysfunction and dendritic impairment. The prominent findings of the present study are that Tat induces endolysosome dysfunction and reduction in dendritic spines, that 17aE2 protects against Tat-induced endolysosome dysfunction and reduction in dendritic spines, and that ERa and its endolysosome localization mediates the protective effects of 17aE2. Endolysosomes, which include endosomes, lysosomes, and autolysosomes, form a dynamic and interconnected network inside the cell. A hallmark feature of endolysosomes is their acidic luminal environment , which is critical for the proper function of up to 60 pH-sensitive hydrolytic enzymes . As terminal degradation centers, endolysosomes mediate the degradation of extracellular materials internalized by endocytosis and/or phagocytosis, as well as intracellular components delivered to lysosomes via autophagy. The acidic luminal pH is also critical for the sorting and trafficking of molecules and/or membranes to their proper destinations for proper processing and physiological functions . Endolysosomes are especially critical for the proper function of neurons because neurons are long-lived and post-mitotic cells with extreme polarity and extensive processes. As long-lived and post-mitotic cells, neurons cannot remove non-degraded or partially degraded products and potentially toxic constituents through cell division, and thus quality control over a large volume of cytoplasm and membrane relies on an efficient proteolytic system, such as that of endolysosomes, with proteasomes and autolysosomes responsible for quality control over cytosolic macromolecules and organelles and endolysosomes for plasma membrane components . Furthermore, neurons are extremely polarized cells, and they undergo extensive processes, complex dendritic arbors representing the information input centers and axons with lengths ranging from tens to hundreds of centimeters representing the information output center . Such extreme polarity of neurons makes it challenging to regulate the trafficking of membrane components, which requires highly dynamic endolysosomes . It is not surprising that mutations to endolysosomal genes preferentially affect the nervous system and contribute to neurodegenerative diseases . Indeed, endolysosome dysfunction has been said to play a central role in a range of age-related neurodegenerative disorders . Endolysosome dysfunction has been implicated by post-mortem brain tissues from HAND patients . Experimentally, we and other have shown that various HIV-1-related factors, such as HIV-1 , HIV-1 proteins (including Tat , gp120 , Nef , and Vpr ), a subset ART drugs , morphine , and methamphetamine , induce endolysosome dysfunction. As a secreted HIV protein that is not reduced or blocked by current ART drugs , Tat is present on neurons and is known to enter neurons via endocytosis . Thus, internalized Tat could affect endolysosomes directly . We have demonstrated that exogenously added Tat de-acidifies endolysosome pH, reduces the activity of endolysosome enzymes, and dramatically increases the size of endolysosomes. Such morphological and functional changes in endolysosomes could result from Tat-induced endolysosome de-acidification, which not only impairs the degradation ability of endolysosomes (leading to an abnormal accumulation of undegraded materials ), but also leads to the impaired fusion and/or trafficking of endolysosomes . Endolysosomes contribute to membranes and the transportation of membrane proteins to synapses and dendritic spines. Endolysosomes also play an important role in the degradation of dendritic cargo, and it not surprising that endolysosomes play a critical role in the modulating of the dynamics of dendritic spines and synaptic plasticity . As such, Tat-induced endolysosome de-acidification and dysfunction could reduce their degradative capacity and impair their trafficking along dendrites , thus disrupting dendritic spine remodeling and leading to a reduction in dendritic spine density . Given that endolysosome dysfunction contributes to dendritic impairment, enhancing endolysosome function and/or alleviating endolysosome dysfunction represents a therapeutic strategy against Tat-induced dendritic spine impairment. As the major form of estrogen in the brain , 17aE2 promotes dendritic spine and synapse formation . Significantly, we have recently shown that the neuroprotective effect of 17aE2 depends, at least in part, on its enhancing effect on endolysosomes . In the present study, we established that 17aE2 prevented Tat-induced reductions in dendritic spines. Based on our recent findings that 17aE2 enhances endolysosome function and increases the density of dendritic spines via endolysosome-localized ERa , we further determined whether ERa plays a role in the protective effect of 17aE2 against Tat-induced endolysosome dysfunction and impairment in dendritic spines. We demonstrated that siRNA knockdown of ERa or over-expressing an ERa mutant (C451A in mice) that fails to localize on endolysosomes impairs the protective effects of 17aE2 against Tat-induced endolysosome dysfunction and impairment in dendritic spines. Thus, our findings demonstrate clearly that 17aE2 can protect against Tat-induced direct synaptodendritic damage. In neurons, endolysosomes are responsible for transporting cargo both towards and away from the dendritic spines, their activity is required for the remodeling of dendritic spines , and ERa is localized in dendritic spines . Thus, we speculated that activating endolysosome-localized ERa would mediate the endolysosome-acidifying effect of 17aE2 . Such endolysosome-acidifying effects of 17aE2 could lead to enhanced proteolysis and the proper travelling of endolysosomes along dendrites, which could in turn result in increased dendritic spine formation. However, our findings do not exclude the possibility that ERa localization on other membranes could also affect dendritic spine formation, and this warrants further investigation. In summary, our findings demonstrate clearly that Tat induces endolysosome dysfunction and reduction in dendritic spines, that 17aE2 protects against Tat-induced endolysosome dysfunction and impairment in dendritic spines, and that ERa and its endolysosome localization mediates the protective effects of 17aE2. Neurons are highly polarized cells, and the compartmentalized signaling of neurons plays a critical role in the formation and maintenance of dendritic spines and synaptic plasticity. Thus, our findings provide novel insights into the neuroprotective effects of 17aE2 which may lead to the development of new adjunct therapeutics against HAND. Author Contributions Conceptualization, X.C.; methodology, G.D. and N.M.M.; data acquisition and analysis, G.D. and N.M.M.; writing--original draft preparation, G.D. and N.M.M.; writing--review and editing, X.C. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The data presented in this study are available in this article. Conflicts of Interest The authors declare no competing financial interest. Figure 1 HIV-1 Tat induces dendritic impairment and endolysosome dysfunction. (A) Effects of Tat (10-100 nM for 48 h) on the morphology of dendrites (MAP2) and dendritic spines (phalloidin-actin) with heat-inactivated Tat as a control. (B) The quantitative data of (A) show that Tat (100 nM for 48 h) slightly decreased dendritic length and significantly reduced the density of dendritic spines (n = 15-30 neurons from 2 repeats, *** p < 0.001). (C) Tat (100 nM for 10 min)-induced endolysosome de-acidification (n = 3 repeats, **** p < 0.0001). (D) Tat (100 nM for 48 h) increased the size of the endolysosomes, as determined using LysoTracker Red (n = 15 neurons from 2 repeats, **** p < 0.0001). Figure 2 17aE2 protects against HIV-1 Tat-induced dendritic impairment. (A) Dendritic spine turnover in EGFP-expressing neurons. Spine growths are indicated by solid arrows, and spine reductions by hollow arrows (Scale = 5 mm). The quantitative data show dendritic spine turnover between 0 and 10 min. Spine formation is indicated by positive values, while spine elimination is indicated by negative values. Tat treatment (100 nM) results in a reduction in stubby, mushroom, and long/thin types of dendritic spines (n = 15, 180 neurons, * p < 0.05). (B) In the presence of Tat (100 nM), 17aE2 (10 nM) increases Tat-induced reductions in stubby, mushroom, and long/thin spines (n = 15, 180 neurons, * p < 0.05). (C) Representative confocal images of dendritic spines (phalloidin-actin). The quantitative data show that 17aE2 (10 nM, pre-treatment of 6 h) prevented a Tat (100 nM for 48 h)-induced reduction in total dendritic spine density (n = 3, 20 neurons, * p < 0.05, *** p < 0.001, **** p < 0.0001). Figure 3 17aE2 prevents HIV-1 Tat-induced endolysosome dysfunction. The images and quantitative data show alterations in the percentage of active endolysosomes (active CatD, green) vs. total endolysosomes (LysoTracker, red). 17aE2 (10 nM, pre-treatment for 10 min) increased the percentage of active endolysosomes and prevented Tat-induced decreases in the percentage of active endolysosomes (n = 3, 20-30 neurons, ** p < 0.01, *** p < 0.001). Figure 4 ERa knockdown attenuates 17aE2-induced acidifying effects on endolysosomes. (A) ERa protein levels were knocked down in a mouse hippocampal cell line (CLU-199) treated with ERa siRNA (n = 4, **** p < 0.0001). (B) ERa KD resulted in a greater de-acidification of endolysosomes compared with ERa scr (n = 3, * p < 0.05). Endolysosome pH was measured ratiometrically with the use of a pH-sensitive (pHrodo) and pH-insensitive (Texas Red) dextran. (C) 17aE2 (10 nM for 10 min)-acidified endolysosomes in ERa scr and ERa KD cells (n = 3, ** p < 0.01, **** p < 0.0001). (D) The magnitude of the 17aE2-induced decrease in endolysosome pH was significantly reduced in ERa KD cells (n = 3, ** p < 0.01). Figure 5 siRNA knockdown of ERa prevents the protective effects of 17aE2 against HIV-1 Tat-induced endolysosome dysfunction and dendritic impairment. (A) Tat (100 nM for 30 min)-induced alterations in the percentage of active endolysosomes (active CatD, green) vs. total endolysosomes (LTR, LysoTracker, red) in CLU199 cells treated with scrambled (scr) or targeted (KD) siRNA against ERa with or without 17aE2 (10 nM pre-treatment for 10 min). (B) In the presence of Tat (100 nM for 30 min), 17aE2 significantly increased the percentage of active endolysosomes in ERa scr cells, but not in ERa KD cells (n = 5, ** p < 0.01). (C) Tat (100 nM for 30 min)-induced changes in dendritic spines in EGFP expression neurons treated with scrambled siRNA (ERa scr) or siRNA against ERa (ERa KD). (D) In the presence of Tat (100 nM for 30 min), 17aE2 (10 nM, pre-treatment for 10 min) increased the density of dendritic spines in ERa scr neurons, but not in ERa KD neurons (n = 5, 20-30 neurons ** p < 0.01). Figure 6 17aE2 protects against HIV-1 Tat-induced endolysosome dysfunction and impairment in dendritic spines via endolysosome-localized ERa. (A) Tat (100 nM for 30 min)-induced alterations in the percentage of active endolysosomes (active CatD, green) vs. total endolysosomes (LTR, LysoTracker, red) in CLU199 neuronal cells expressing wildtype ERa (ERa-HA) or ERa C451A-HA (ERa C451A) with and without 17aE2 (10 nM pre-treatment for 10 min). (B) In the presence of Tat (100 nM for 30 min), 17aE2 significantly increased the percentage of active endolysosomes in wild type cells, but not in ERa C451A over-expressing cells (n = 2, 11-20 neurons **** p < 0.0001). (C) Tat (100 nM for 30 min)-induced changes in dendritic spines in BacMAM EGFP-transduced wildtype (ERa-HA) neurons and ERa C451A-HA over-expressing neurons. (D) In the presence of Tat (100 nM for 30 min), 17aE2 (10 nM, pre-treatment for 10 min) increased the density of dendritic spines in wildtype (WT) neurons, but not in ERa C451A over-expressing neurons (n = 5, 20-30 neurons, * p < 0.05, ** p < 0.01). Figure 7 Proposed model according to which 17aE2 protects against HIV-1 Tat-induced endolysosome dysfunction and impairment in dendritic spines via endolysosome-localized ERa. Internalized Tat-induced endolysosome de-acidification and dysfunction could reduce their degradative capacity and impair their trafficking along dendrites, thus disrupting dendritic spine remodeling and leading to a reduction in dendritic spine density. Such damaging effects of Tat can by attenuated by 17aE2, which acidifies endolysosomes by activating endolysosome-localized ERa. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Wang Y. Liu M. Lu Q. Farrell M. Lappin J. Shi J. Lu L. Bao Y. Data from: Global prevalence and burden of HIV-associated neurocognitive disorder: A meta-analysis Neurology 2020 95 e2610 e2621 10.1212/WNL.0000000000010752 32887786 2. Sacktor N. Skolasky R.L. Seaberg E. Munro C. Becker J.T. Martin E. Ragin A. Levine A. Miller E. Prevalence of HIV-associated neurocognitive disorders in the Multicenter AIDS Cohort Study Neurology 2016 86 334 340 10.1212/WNL.0000000000002277 26718568 3. Maschke M. Kastrup O. Esser S. Ross B. Hengge U. Hufnagel A. Incidence and prevalence of neurological disorders associated with HIV since the introduction of highly active antiretroviral therapy (HAART) J. Neurol. Neurosurg. Psychiatry 2000 69 376 380 10.1136/jnnp.69.3.376 10945813 4. Crews L. Patrick C. Achim C.L. Everall I.P. Masliah E. Molecular Pathology of Neuro-AIDS (CNS-HIV) Int. J. Mol. Sci. 2009 10 1045 1063 10.3390/ijms10031045 19399237 5. McArthur J.C. HIV dementia: An evolving disease J. Neuroimmunol. 2004 157 3 10 10.1016/j.jneuroim.2004.08.042 15579274 6. Masliah E. Ellis R.J. Bs M.M. Heaton R.K. Marcotte T.D. Ba J.A.N. Grant I. Atkinson J.H. Wiley C.A. Achim C.L. Dendritic injury is a pathological substrate for human immunodeficiency virus-related cognitive disorders. HNRC Group. The HIV Neurobehavioral Research Center Ann. Neurol. 1997 42 963 972 10.1002/ana.410420618 9403489 7. Albright A.V. Soldan S.S. Gonzalez-Scarano F. Pathogenesis of human immunodeficiency virus-induced neurological disease J. Neurovirol. 2003 9 222 227 10.1080/13550280390194073 12707852 8. Archibald S.L. Masliah E. Fennema-Notestine C. Marcotte T.D. Ellis R. McCutchan J.A. Heaton R.K. Grant I. Mallory M. Miller A. Correlation of In Vivo Neuroimaging Abnormalities With Postmortem Human Immunodeficiency Virus Encephalitis and Dendritic Loss Arch. Neurol. 2004 61 369 376 10.1001/archneur.61.3.369 15023814 9. Everall I.P. Heaton R.K. Marcotte T.D. Ellis R. McCutchan J.A. Atkinson J.H. Grant I. Mallory M. Masliah E. HNRC Group Cortical Synaptic Density is Reduced in Mild to Moderate Human Immunodeficiency Virus Neurocognitive Disorder Brain Pathol. 2006 9 209 217 10.1111/j.1750-3639.1999.tb00219.x 10. Sa M.J. Madeira M.D. Ruela C. Volk B. Mota-Miranda A. Paula-Barbosa M.M. Dendritic changes in the hippocampal formation of AIDS patients: A quantitative Golgi study Acta Neuropathol. 2004 107 97 110 10.1007/s00401-003-0781-3 14605830 11. Ellis R. Langford D. Masliah E. HIV and antiretroviral therapy in the brain: Neuronal injury and repair Nat. Rev. Neurosci. 2007 8 33 44 10.1038/nrn2040 17180161 12. Marban C. Forouzanfar F. Ait-Ammar A. Fahmi F. El Mekdad H. Daouad F. Rohr O. Schwartz C. Targeting the Brain Reservoirs: Toward an HIV Cure Front. Immunol. 2016 7 397 10.3389/fimmu.2016.00397 27746784 13. Ajasin D. Eugenin E.A. HIV-1 Tat: Role in Bystander Toxicity Front. Cell. Infect. Microbiol. 2020 10 61 10.3389/fcimb.2020.00061 32158701 14. Avdoshina V.M. Mocchetti I. Recent Advances in the Molecular and Cellular Mechanisms of gp120-Mediated Neurotoxicity Cells 2022 11 1599 10.3390/cells11101599 35626635 15. Sviridov D. Mukhamedova N. Makarov A.A. Adzhubei A. Bukrinsky M. Comorbidities of HIV infection: Role of Nef-induced impairment of cholesterol metabolism and lipid raft functionality AIDS 2020 34 1 13 10.1097/QAD.0000000000002385 31789888 16. Li G. Makar T. Gerzanich V. Kalakonda S. Ivanova S. Pereira E.F.R. Andharvarapu S. Zhang J. Simard J.M. Zhao R.Y. HIV-1 Vpr-Induced Proinflammatory Response and Apoptosis Are Mediated through the Sur1-Trpm4 Channel in Astrocytes Mbio 2020 11 e02939-20 10.1128/mBio.02939-20 33293383 17. Underwood J. Robertson K.R. Winston A. Could antiretroviral neurotoxicity play a role in the pathogenesis of cognitive impairment in treated HIV disease? AIDS 2015 29 253 261 10.1097/QAD.0000000000000538 25426811 18. Hidalgo M. Atluri V.S.R. Nair M. Drugs of Abuse in HIV infection and neurotoxicity Front. Microbiol. 2015 6 217 10.3389/fmicb.2015.00217 25852673 19. Fitting S. Xu R. Bull C. Buch S.K. El-Hage N. Nath A. Knapp P.E. Hauser K.F. Interactive Comorbidity between Opioid Drug Abuse and HIV-1 Tat: Chronic Exposure Augments Spine Loss and Sublethal Dendritic Pathology in Striatal Neurons Am. J. Pathol. 2010 177 1397 1410 10.2353/ajpath.2010.090945 20651230 20. Soontornniyomkij V. Translational Methamphetamine AIDS Research Center (TMARC) Group Kesby J.P. Morgan E.E. Bischoff-Grethe A. Minassian A. Brown G.G. Grant I. Effects of HIV and Methamphetamine on Brain and Behavior: Evidence from Human Studies and Animal Models J. Neuroimmune Pharmacol. 2016 11 495 510 10.1007/s11481-016-9699-0 27484318 21. Irollo E. Luchetta J. Ho C. Nash B. Meucci O. Mechanisms of neuronal dysfunction in HIV-associated neurocognitive disorders Cell. Mol. Life Sci. 2021 78 4283 4303 10.1007/s00018-021-03785-y 33585975 22. Saylor D. Dickens A. Sacktor N. Haughey N. Slusher B. Pletnikov M. Mankowski J.L. Brown A. Volsky D.J. McArthur J.C. HIV-associated neurocognitive disorder--pathogenesis and prospects for treatment Nat. Rev. Neurol. 2016 12 234 248 10.1038/nrneurol.2016.27 26965674 23. Heaton R.K. Franklin D.R. Deutsch R. Letendre S. Ellis R.J. Casaletto K.B. Marquine M.J. Woods S.P. Vaida F. Atkinson J.H. Neurocognitive Change in the Era of HIV Combination Antiretroviral Therapy: The Longitudinal CHARTER Study Clin. Infect. Dis. 2014 60 473 480 10.1093/cid/ciu862 25362201 24. Ru W. Tang S.J. HIV-associated synaptic degeneration Mol. Brain 2017 10 40 10.1186/s13041-017-0321-z 28851400 25. Bandera A. Taramasso L. Bozzi G. Muscatello A. Robinson J.A. Burdo T.H. Gori A. HIV-Associated Neurocognitive Impairment in the Modern ART Era: Are We Close to Discovering Reliable Biomarkers in the Setting of Virological Suppression? Front. Aging Neurosci. 2019 11 187 10.3389/fnagi.2019.00187 31427955 26. Gelman B.B. Soukup V.M. Holzer C.E. Fabian R.H. Schuenke K.W. Keherly M.J. Richey F.J. Lahart C.J. Potential Role for White Matter Lysosome Expansion in HIV-Associated Dementia J. Acquir. Immune Defic. Syndr. 2005 39 422 425 10.1097/01.qai.0000164250.41475.f2 16010164 27. Spector S.A. Zhou D. Autophagy: An overlooked mechanism of HIV-1 pathogenesis and NeuroAIDS? Autophagy 2008 4 704 706 10.4161/auto.6105 18424919 28. Zhou D. Spector S.A. Human immunodeficiency virus type-1 infection inhibits autophagy AIDS 2008 22 695 699 10.1097/QAD.0b013e3282f4a836 18356598 29. Cysique L.A. Hewitt T. Croitoru-Lamoury J. Taddei K. Martins R.N. Chew C.S. Davies N.N. Price P. Brew B.J. APOE epsilon4 moderates abnormal CSF-abeta-42 levels, while neurocognitive impairment is associated with abnormal CSF tau levels in HIV+ individuals--A cross-sectional observational study BMC Neurol. 2015 15 51 10.1186/s12883-015-0298-0 25880550 30. Nixon R.A. Cataldo A.M. The endosomal-lysosomal system of neurons: New roles Trends Neurosci. 1995 18 489 496 10.1016/0166-2236(95)92772-I 8592758 31. Nixon R.A. Cataldo A.M. Lysosomal system pathways: Genes to neurodegeneration in Alzheimer's disease J. Alzheimer's Dis. 2006 9 277 289 10.3233/JAD-2006-9S331 16914867 32. Goo M.S. Sancho L. Slepak N. Boassa D. Deerinck T.J. Ellisman M.H. Bloodgood B.L. Patrick G.N. Activity-dependent trafficking of lysosomes in dendrites and dendritic spines J. Cell Biol. 2017 216 2499 2513 10.1083/jcb.201704068 28630145 33. Padamsey Z. McGuinness L. Bardo S.J. Reinhart M. Tong R. Hedegaard A. Hart M.L. Emptage N.J. Activity-Dependent Exocytosis of Lysosomes Regulates the Structural Plasticity of Dendritic Spines Neuron 2017 93 132 146 10.1016/j.neuron.2016.11.013 27989455 34. Nikoletopoulou V. Tavernarakis N. Regulation and Roles of Autophagy at Synapses Trends Cell Biol. 2018 28 646 661 10.1016/j.tcb.2018.03.006 29731196 35. Yap C.C. Digilio L. McMahon L.P. Garcia A.D.R. Winckler B. Degradation of dendritic cargos requires Rab7-dependent transport to somatic lysosomes J. Cell Biol. 2018 217 3141 3159 10.1083/jcb.201711039 29907658 36. Farias G.G. Guardia C.M. Britt D.J. Guo X. Bonifacino J.S. Sorting of Dendritic and Axonal Vesicles at the Pre-axonal Exclusion Zone Cell Rep. 2015 13 1221 1232 10.1016/j.celrep.2015.09.074 26527003 37. Farias G.G. Guardia C.M. De Pace R. Britt D.J. Bonifacino J.S. BORC/kinesin-1 ensemble drives polarized transport of lysosomes into the axon Proc. Natl. Acad. Sci. USA 2017 114 E2955 E2964 10.1073/pnas.1616363114 28320970 38. Ferguson S.M. Axonal transport and maturation of lysosomes Curr. Opin. Neurobiol. 2018 51 45 51 10.1016/j.conb.2018.02.020 29529416 39. Agostini S. Ali H. Vardabasso C. Fittipaldi A. Tasciotti E. Cereseto A. Bugatti A. Rusnati M. Lusic M. Giacca M. Inhibition of Non Canonical HIV-1 Tat Secretion Through the Cellular Na+, K+-ATPase Blocks HIV-1 Infection EBioMedicine 2017 21 170 181 10.1016/j.ebiom.2017.06.011 28645727 40. Ensoli B. Barillari G. Salahuddin S.Z. Gallo R.C. Wong-Staal F. Tat protein of HIV-1 stimulates growth of cells derived from Kaposi's sarcoma lesions of AIDS patients Nature 1990 345 84 86 10.1038/345084a0 2184372 41. Chang H.C. Samaniego F. Nair B.C. Buonaguro L. Ensoli B. HIV-1 Tat protein exits from cells via a leaderless secretory pathway and binds to extracellular matrix-associated heparan sulfate proteoglycans through its basic region AIDS 1997 11 1421 1431 10.1097/00002030-199712000-00006 9342064 42. Rayne F. Debaisieux S. Yezid H. Lin Y.-L. Mettling C. Konate K. Chazal N. Arold S.T. Pugniere M. Sanchez F. Phosphatidylinositol-(4,5)-bisphosphate enables efficient secretion of HIV-1 Tat by infected T-cells EMBO J. 2010 29 1348 1362 10.1038/emboj.2010.32 20224549 43. Mediouni S. Darque A. Baillat G. Ravaux I. Dhiver C. Tissot-Dupont H. Mokhtari M. Moreau H. Tamalet C. Brunet C. Antiretroviral therapy does not block the secretion of the human immunodeficiency virus Tat protein Infect. Disord. Drug Targets 2012 12 81 86 10.2174/187152612798994939 22280310 44. Johnson T.P. Patel K. Johnson K.R. Maric D. Calabresi P.A. Hasbun R. Nath A. Induction of IL-17 and nonclassical T-cell activation by HIV-Tat protein Proc. Natl. Acad. Sci. USA 2013 110 13588 13593 10.1073/pnas.1308673110 23898208 45. Henderson L.J. Johnson T.P. Smith B.R. Reoma L. Santamaria U.A. Bachani M. DeMarino C. Barclay R.A. Snow J. Sacktor N. Presence of Tat and transactivation response element in spinal fluid despite antiretroviral therapy AIDS 2019 33 (Suppl. S2) S145 S157 10.1097/QAD.0000000000002268 31789815 46. Chang J.R. Mukerjee R. Bagashev A. Del Valle L. Chabrashvili T. Hawkins B.J. He J.J. Sawaya B.E. HIV-1 Tat Protein Promotes Neuronal Dysfunction through Disruption of MicroRNAs J. Biol. Chem. 2011 286 41125 41134 10.1074/jbc.M111.268466 21956116 47. Donoso M. D'Amico D. Valdebenito S. Hernandez C.A. Prideaux B. Eugenin E.A. Identification, Quantification, and Characterization of HIV-1 Reservoirs in the Human Brain Cells 2022 11 2379 10.3390/cells11152379 35954221 48. Li W. Huang Y. Reid R. Steiner J. Malpica-Llanos T. Darden T.A. Shankar S.K. Mahadevan A. Satishchandra P. Nath A. NMDA Receptor Activation by HIV-Tat Protein Is Clade Dependent J. Neurosci. 2008 28 12190 12198 10.1523/JNEUROSCI.3019-08.2008 19020013 49. Sabatier J.M. Vives E. Mabrouk K. Benjouad A. Rochat H. Duval A. Hue B. Bahraoui E. Evidence for neurotoxic activity of Tat from human immunodeficiency virus type 1 J. Virol. 1991 65 961 967 10.1128/jvi.65.2.961-967.1991 1898974 50. Bertrand S.J. Mactutus C.F. Aksenova M.V. Espensen-Sturges T.D. Booze R.M. Synaptodendritic recovery following HIV Tat exposure: Neurorestoration by phytoestrogens J. Neurochem. 2014 128 140 151 10.1111/jnc.12375 23875777 51. Fitting S. Ignatowska-Jankowska B.M. Bull C. Skoff R.P. Lichtman A.H. Wise L.E. Fox M.A. Su J. Medina A.E. Krahe T.E. Synaptic Dysfunction in the Hippocampus Accompanies Learning and Memory Deficits in Human Immunodeficiency Virus Type-1 Tat Transgenic Mice Biol. Psychiatry 2013 73 443 453 10.1016/j.biopsych.2012.09.026 23218253 52. Hargus N.J. Thayer S.A. Human Immunodeficiency Virus-1 Tat Protein Increases the Number of Inhibitory Synapses between Hippocampal Neurons in Culture J. Neurosci. 2013 33 17908 17920 10.1523/JNEUROSCI.1312-13.2013 24198379 53. Nath A. Steiner J. Synaptodendritic injury with HIV-Tat protein: What is the therapeutic target? Exp. Neurol. 2014 251 112 114 10.1016/j.expneurol.2013.11.004 24246278 54. Debaisieux S. Rayne F. Yezid H. Beaumelle B. The ins and outs of HIV-1 Tat Traffic 2012 13 355 363 10.1111/j.1600-0854.2011.01286.x 21951552 55. Tyagi M. Rusnati M. Presta M. Giacca M. Internalization of HIV-1 Tat Requires Cell Surface Heparan Sulfate Proteoglycans J. Biol. Chem. 2001 276 3254 3261 10.1074/jbc.M006701200 11024024 56. Liu Y. Jones M. Hingtgen C.M. Bu G. Laribee N. Tanzi R.E. Moir R.D. Nath A. He J.J. Uptake of HIV-1 Tat protein mediated by low-density lipoprotein receptor-related protein disrupts the neuronal metabolic balance of the receptor ligands Nat. Med. 2000 6 1380 1387 10.1038/82199 11100124 57. Frankel A.D. Pabo C.O. Cellular uptake of the Tat protein from human immunodeficiency virus Cell 1988 55 1189 1193 10.1016/0092-8674(88)90263-2 2849510 58. Mann D.A. Frankel A.D. Endocytosis and targeting of exogenous HIV-1 Tat protein EMBO J. 1991 10 1733 1739 10.1002/j.1460-2075.1991.tb07697.x 2050110 59. Gaskill P.J. Miller D.R. Gamble-George J. Yano H. Khoshbouei H. HIV, Tat and dopamine transmission Neurobiol. Dis. 2017 105 51 73 10.1016/j.nbd.2017.04.015 28457951 60. Hui L. Chen X. Haughey N.J. Geiger J.D. Role of Endolysosomes in HIV-1 Tat-Induced Neurotoxicity ASN Neuro 2012 4 243 252 10.1042/AN20120017 22591512 61. Chen X. Hui L. Geiger N.H. Haughey N.J. Geiger J.D. Endolysosome involvement in HIV-1 transactivator protein-induced neuronal amyloid beta production Neurobiol. Aging 2013 34 2370 2378 10.1016/j.neurobiolaging.2013.04.015 23673310 62. Datta G. Miller N.N. Du W. Geiger J.D. Chang S. Chen X. Endolysosome Localization of ERalpha Is Involved in the Protective Effect of 17alpha-Estradiol against HIV-1 gp120-Induced Neuronal Injury J. Neurosci. 2021 41 10365 10381 10.1523/JNEUROSCI.1475-21.2021 34764157 63. Christensen A. Dewing P. Micevych P. Membrane-Initiated Estradiol Signaling Induces Spinogenesis Required for Female Sexual Receptivity J. Neurosci. 2011 31 17583 17589 10.1523/JNEUROSCI.3030-11.2011 22131419 64. Nash B. Tarn K. Irollo E. Luchetta J. Festa L. Halcrow P. Datta G. Geiger J.D. Meucci O. Morphine-Induced Modulation of Endolysosomal Iron Mediates Upregulation of Ferritin Heavy Chain in Cortical Neurons Eneuro 2019 6 ENEURO.0237-19.2019 10.1523/ENEURO.0237-19.2019 31300544 65. Chauhan A. Turchan J. Pocernich C. Bruce-Keller A. Roth S. Butterfield D.A. Major E.O. Nath A. Intracellular Human Immunodeficiency Virus Tat Expression in Astrocytes Promotes Astrocyte Survival but Induces Potent Neurotoxicity at Distant Sites via Axonal Transport J. Biol. Chem. 2003 278 13512 13519 10.1074/jbc.M209381200 12551932 66. Gupta S. Knight A.G. Gupta S. Knapp P.E. Hauser K.F. Keller J.N. Bruce-Keller A.J. HIV-Tat elicits microglial glutamate release: Role of NAPDH oxidase and the cystine-glutamate antiporter Neurosci. Lett. 2010 485 233 236 10.1016/j.neulet.2010.09.019 20849923 67. Khan N. Halcrow P.W. Afghah Z. Baral A. Geiger J.D. Chen X. HIV-1 Tat endocytosis and retention in endolysosomes affects HIV-1 Tat-induced LTR transactivation in astrocytes FASEB J. 2022 36 e22184 10.1096/fj.202101722R 35113458 68. Saxena S. Bucci C. Weis J. Kruttgen A. The Small GTPase Rab7 Controls the Endosomal Trafficking and Neuritogenic Signaling of the Nerve Growth Factor Receptor TrkA J. Neurosci. 2005 25 10930 10940 10.1523/JNEUROSCI.2029-05.2005 16306406 69. Ikeda T. Makino Y. Yamada M.K. 17a-Estradiol is generated locally in the male rat brain and can regulate GAD65 expression and anxiety Neuropharmacology 2015 90 9 14 10.1016/j.neuropharm.2014.10.019 25446575 70. Toran-Allerand C.D. Tinnikov A.A. Singh R.J. Nethrapalli I.S. 17alpha-estradiol: A brain-active estrogen? Endocrinology 2005 146 3843 3850 10.1210/en.2004-1616 15947006 71. Sengupta D.C. Lantz C.L. Rumi M.A.K. Quinlan E.M. 17a Estradiol promotes plasticity of spared inputs in the adult amblyopic visual cortex Sci. Rep. 2019 9 19040 10.1038/s41598-019-55158-y 31836739 72. Heron P.M. Turchan-Cholewo J. Bruce-Keller A.J. Wilson M.E. Estrogen Receptor Alpha Inhibits the Estrogen-Mediated Suppression of HIV Transcription in Astrocytes: Implications for Estrogen Neuroprotection in HIV Dementia AIDS Res. Hum. Retroviruses 2009 25 1071 1081 10.1089/aid.2009.0065 19886840 73. Lu Y.P. Zeng M. Hu X.-Y. Xu H. Swaab D. Ravid R. Zhou J.-N. Estrogen receptor alpha-immunoreactive astrocytes are increased in the hippocampus in Alzheimer's disease Exp. Neurol. 2003 183 482 488 10.1016/S0014-4886(03)00205-X 14552888 74. Milner T.A. Ayoola K. Drake C.T. Herrick S.P. Tabori N.E. McEwen B.S. Warrier S. Alves S.E. Ultrastructural localization of estrogen receptor beta immunoreactivity in the rat hippocampal formation J. Comp. Neurol. 2005 491 81 95 10.1002/cne.20724 16127691 75. Milner T.A. McEwen B.S. Hayashi S. Li C.J. Reagan L.P. Alves S.E. Ultrastructural evidence that hippocampal alpha estrogen receptors are located at extranuclear sites J. Comp. Neurol. 2001 429 355 371 10.1002/1096-9861(20010115)429:3<355::AID-CNE1>3.0.CO;2-# 11116225 76. Lai Y.J. Yu D. Zhang J.H. Chen G.-J. Cooperation of Genomic and Rapid Nongenomic Actions of Estrogens in Synaptic Plasticity Mol. Neurobiol. 2016 54 4113 4126 10.1007/s12035-016-9979-y 27324789 77. Srivastava D.P. Waters E.M. Mermelstein P.G. Kramar E.A. Shors T.J. Liu F. Rapid Estrogen Signaling in the Brain: Implications for the Fine-Tuning of Neuronal Circuitry J. Neurosci. 2011 31 16056 16063 10.1523/JNEUROSCI.4097-11.2011 22072656 78. Frick K.M. Molecular mechanisms underlying the memory-enhancing effects of estradiol Horm. Behav. 2015 74 4 18 10.1016/j.yhbeh.2015.05.001 25960081 79. Mukai H. Tsurugizawa T. Murakami G. Kominami S. Ishii H. Ogiue-Ikeda M. Takata N. Tanabe N. Furukawa A. Hojo Y. Rapid modulation of long-term depression and spinogenesis via synaptic estrogen receptors in hippocampal principal neurons J. Neurochem. 2007 100 950 967 10.1111/j.1471-4159.2006.04264.x 17266735 80. Hojo Y. Murakami G. Mukai H. Higo S. Hatanaka Y. Ogiue-Ikeda M. Ishii H. Kimoto T. Kawato S. Estrogen synthesis in the brain--Role in synaptic plasticity and memory Mol. Cell. Endocrinol. 2008 290 31 43 10.1016/j.mce.2008.04.017 18541362 81. Sheppard P.A.S. Choleris E. Galea L.A.M. Structural plasticity of the hippocampus in response to estrogens in female rodents Mol. Brain 2019 12 22 10.1186/s13041-019-0442-7 30885239 82. Sampayo R.G. Toscani A.M. Rubashkin M.G. Thi K. Masullo L.A. Violi I.L. Lakins J.N. Caceres A. Hines W.C. Leskow F.C. Fibronectin rescues estrogen receptor a from lysosomal degradation in breast cancer cells J. Cell Biol. 2018 217 2777 2798 10.1083/jcb.201703037 29980625 83. Liao T.L. Tzeng C.-R. Yu C.-L. Wang Y.-P. Kao S.-H. Estrogen receptor-beta in mitochondria: Implications for mitochondrial bioenergetics and tumorigenesis Ann. N. Y. Acad Sci. 2015 1350 52 60 10.1111/nyas.12872 26301952 84. Yang S.H. Liu R. Perez E.J. Wen Y. Stevens S.M. Jr. Valencia T. Brun-Zinkernagel A.-M. Prokai L. Will Y. Dykens J. Mitochondrial localization of estrogen receptor beta Proc. Natl. Acad. Sci. USA 2004 101 4130 4135 10.1073/pnas.0306948101 15024130 85. Revankar C.M. Cimino D.F. Sklar L.A. Arterburn J.B. Prossnitz E.R. A Transmembrane Intracellular Estrogen Receptor Mediates Rapid Cell Signaling Science 2005 307 1625 1630 10.1126/science.1106943 15705806 86. Azcoitia I. Mendez P. Garcia-Segura L.M. Aromatase in the Human Brain Androg. Clin. Res. Ther. 2021 2 189 202 10.1089/andro.2021.0007 35024691 87. Kretz O. Fester L. Wehrenberg U. Zhou L. Brauckmann S. Zhao S. Prange-Kiel J. Naumann T. Jarry H. Frotscher M. Hippocampal Synapses Depend on Hippocampal Estrogen Synthesis J. Neurosci. 2004 24 5913 5921 10.1523/JNEUROSCI.5186-03.2004 15229239 88. Acconcia F. Ascenzi P. Bocedi A. Spisni E. Tomasi V. Trentalance A. Visca P. Marino M. Palmitoylation-dependent estrogen receptor alpha membrane localization: Regulation by 17beta-estradiol Mol. Biol. Cell 2005 16 231 237 10.1091/mbc.e04-07-0547 15496458 89. Meitzen J. Luoma J.I. Boulware M.I. Hedges V.L. Peterson B.M. Tuomela K. Britson K.A. Mermelstein P.G. Palmitoylation of Estrogen Receptors Is Essential for Neuronal Membrane Signaling Endocrinology 2013 154 4293 4304 10.1210/en.2013-1172 24008343 90. Adlanmerini M. Solinhac R. Abot A. Fabre A. Raymond-Letron I. Guihot A.-L. Boudou F. Sautier L. Vessieres E. Kim S.H. Mutation of the palmitoylation site of estrogen receptor a in vivo reveals tissue-specific roles for membrane versus nuclear actions Proc. Natl. Acad. Sci. USA 2014 111 E283 E290 10.1073/pnas.1322057111 24371309 91. Mindell J.A. Lysosomal Acidification Mechanisms Annu. Rev. Physiol. 2012 74 69 86 10.1146/annurev-physiol-012110-142317 22335796 92. Huotari J. Helenius A. Endosome maturation EMBO J. 2011 30 3481 3500 10.1038/emboj.2011.286 21878991 93. McGuire C. Stransky L. Cotter K. Forgac M. Regulation of V-ATPase activity Front. Biosci.-Landmark 2017 22 609 622 94. Prasad H. Rao R. Histone deacetylase-mediated regulation of endolysosomal pH J. Biol. Chem. 2018 293 6721 6735 10.1074/jbc.RA118.002025 29567836 95. de Duve C. The lysosome turns fifty Nat. Cell Biol. 2005. 7 847 849 10.1038/ncb0905-847 96. Paroutis P. Touret N. Grinstein S. The pH of the Secretory Pathway: Measurement, Determinants, and Regulation Physiology 2004 19 207 215 10.1152/physiol.00005.2004 15304635 97. Kellokumpu S. Golgi pH, Ion and Redox Homeostasis: How Much Do They Really Matter? Front. Cell Dev. Biol. 2019 7 93 10.3389/fcell.2019.00093 31263697 98. Weisz O.A. Organelle Acidification and Disease Traffic 2003 4 57 64 10.1034/j.1600-0854.2003.40201.x 12559032 99. Winckler B. Faundez V. Maday S. Cai Q. Almeida C.G. Zhang H. The Endolysosomal System and Proteostasis: From Development to Degeneration J. Neurosci. 2018 38 9364 9374 10.1523/JNEUROSCI.1665-18.2018 30381428 100. Lee S. Sato Y. Nixon R.A. Lysosomal proteolysis inhibition selectively disrupts axonal transport of degradative organelles and causes an Alzheimer's-like axonal dystrophy J. Neurosci. 2011 31 7817 7830 10.1523/JNEUROSCI.6412-10.2011 21613495 101. Harris J.J. Attwell D. The Energetics of CNS White Matter J. Neurosci. 2012 32 356 371 10.1523/JNEUROSCI.3430-11.2012 22219296 102. Price J.C. Guan S. Burlingame A. Prusiner S.B. Ghaemmaghami S. Analysis of proteome dynamics in the mouse brain Proc. Natl. Acad. Sci. USA 2010 107 14508 14513 10.1073/pnas.1006551107 20699386 103. Heo S. Diering G.H. Na C.H. Nirujogi R.S. Bachman J.L. Pandey A. Huganir R.L. Identification of long-lived synaptic proteins by proteomic analysis of synaptosome protein turnover Proc. Natl. Acad. Sci. USA 2018 115 E3827 E3836 10.1073/pnas.1720956115 29610302 104. Nikoletopoulou V. Papandreou M.-E. Tavernarakis N. Autophagy in the physiology and pathology of the central nervous system Cell Death Differ. 2015 22 398 407 10.1038/cdd.2014.204 25526091 105. Root J. Merino P. Nuckols A. Johnson M. Kukar T. Lysosome dysfunction as a cause of neurodegenerative diseases: Lessons from frontotemporal dementia and amyotrophic lateral sclerosis Neurobiol. Dis. 2021 154 105360 10.1016/j.nbd.2021.105360 33812000 106. Bonam S.R. Wang F. Muller S. Lysosomes as a therapeutic target Nat. Rev. Drug Discov. 2019 18 923 948 10.1038/s41573-019-0036-1 31477883 107. Nixon R.A. The role of autophagy in neurodegenerative disease Nat. Med. 2013 19 983 997 10.1038/nm.3232 23921753 108. Nixon R.A. The aging lysosome: An essential catalyst for late-onset neurodegenerative diseases Biochim. Biophys. Acta Proteins Proteom. 2020 1868 140443 10.1016/j.bbapap.2020.140443 32416272 109. Cinti A. Le Sage V. Milev M.P. Valiente-Echeverria F. Crossie C. Miron M.-J. Pante N. Olivier M. Mouland A.J. HIV-1 enhances mTORC1 activity and repositions lysosomes to the periphery by co-opting Rag GTPases Sci. Rep. 2017 7 5515 10.1038/s41598-017-05410-0 28710431 110. Moorjani H. Craddock B.P. Morrison S.A. Steigbigel R.T. Impairment of Phagosome-Lysosome Fusion in HIV-1-Infected Macrophages J. Acquir. Immune Defic. Syndr. Hum. Retrovirol. 1996 13 18 22 10.1097/00042560-199609000-00003 8797681 111. Fields J. Dumaop W. Elueteri S. Campos S. Serger E. Trejo M. Kosberg K. Adame A. Spencer B. Rockenstein E. HIV-1 Tat Alters Neuronal Autophagy by Modulating Autophagosome Fusion to the Lysosome: Implications for HIV-Associated Neurocognitive Disorders J. Neurosci. 2015 35 1921 1938 10.1523/JNEUROSCI.3207-14.2015 25653352 112. Bae M. Patel N. Xu H. Lee M. Tominaga-Yamanaka K. Nath A. Geiger J. Gorospe M. Mattson M.P. Haughey N.J. Activation of TRPML1 clears intraneuronal Abeta in preclinical models of HIV infection J. Neurosci. 2014 34 11485 11503 10.1523/JNEUROSCI.0210-14.2014 25143627 113. Datta G. Miller N.M. Afghah Z. Geiger J.D. Chen X. HIV-1 gp120 Promotes Lysosomal Exocytosis in Human Schwann Cells Front. Cell. Neurosci. 2019 13 329 10.3389/fncel.2019.00329 31379513 114. Halcrow P.W. Lakpa K.L. Khan N. Afghah Z. Miller N. Datta G. Chen X. Geiger J.D. HIV-1 gp120-Induced Endolysosome de-Acidification Leads to Efflux of Endolysosome Iron, and Increases in Mitochondrial Iron and Reactive Oxygen Species J. Neuroimmune Pharmacol. 2021 17 181 194 10.1007/s11481-021-09995-2 33834418 115. Campbell G.R. Rawat P. Bruckman R.S. Spector S.A. Human Immunodeficiency Virus Type 1 Nef Inhibits Autophagy through Transcription Factor EB Sequestration PLoS Pathog. 2015 11 e1005018 10.1371/journal.ppat.1005018 26115100 116. Santerre M. Arjona S.P. Allen C.N. Buch S. E Sawaya B. HIV-1 Vpr protein impairs lysosome clearance causing SNCA/alpha-synuclein accumulation in neurons Autophagy 2021 17 1768 1782 10.1080/15548627.2021.1915641 33890542 117. Hui L. Ye Y. Soliman M.L. Lakpa K.L. Miller N.M. Afghah Z. Geiger J.D. Chen X. Antiretroviral Drugs Promote Amyloidogenesis by De-Acidifying Endolysosomes J. Neuroimmune Pharmacol. 2021 16 159 168 10.1007/s11481-019-09862-1 31338753 118. Cubells J. Rayport S. Rajendran G. Sulzer D. Methamphetamine neurotoxicity involves vacuolation of endocytic organelles and dopamine-dependent intracellular oxidative stress J. Neurosci. 1994 14 2260 2271 10.1523/JNEUROSCI.14-04-02260.1994 8158268 119. Nara A. Aki T. Funakoshi T. Unuma K. Uemura K. Hyperstimulation of macropinocytosis leads to lysosomal dysfunction during exposure to methamphetamine in SH-SY5Y cells Brain Res. 2012 1466 1 14 10.1016/j.brainres.2012.05.017 22613348 120. Kawai A. Uchiyama H. Takano S. Nakamura N. Ohkuma S. Autophagosome-Lysosome Fusion Depends on the pH in Acidic Compartments in CHO Cells Autophagy 2007 3 154 157 10.4161/auto.3634 17204842 121. Johnson D.E. Ostrowski P. Jaumouille V. Grinstein S. The position of lysosomes within the cell determines their luminal pH J. Cell Biol. 2016 212 677 692 10.1083/jcb.201507112 26975849 122. Tsuruta F. Dolmetsch R.E. PIKfyve mediates the motility of late endosomes and lysosomes in neuronal dendrites Neurosci. Lett. 2015 605 18 23 10.1016/j.neulet.2015.07.021 26232680 |
PMC10000620 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050987 diagnostics-13-00987 Review Mechanisms of Resistance to CDK4/6 Inhibitors and Predictive Biomarkers of Response in HR+/HER2-Metastatic Breast Cancer--A Review of the Literature Stanciu Ioana-Miruna 12 Parosanu Andreea Ioana 12* Orlov-Slavu Cristina 12 Iaciu Ion Cristian 12 Popa Ana Maria 12 Olaru Cristina Mihaela 12 Pirlog Cristina Florina 12 Vrabie Radu Constantin 12 Nitipir Cornelia 12 Lim Sung Chul Academic Editor 1 Department of Oncology, "Carol Davila" University of Medicine and Pharmacy, 020021 Bucharest, Romania 2 Elias University Emergency Hospital, 011461 Bucharest, Romania * Correspondence: [email protected]; Tel.: +40-725-683-118 05 3 2023 3 2023 13 5 98709 2 2023 25 2 2023 03 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). The latest and newest discoveries for advanced and metastatic hormone receptor-positive (HR+) and human epidermal growth factor receptor 2-negative (HER2-) breast cancer are the three cyclin-dependent kinases 4 and 6 inhibitors (CDK4/6i) in association with endocrine therapy (ET). However, even if this treatment revolutionized the world and continued to be the first-line treatment choice for these patients, it also has its limitations, caused by de novo or acquired drug resistance which leads to inevitable progression after some time. Thus, an understanding of the overview of the targeted therapy which represents the gold therapy for this subtype of cancer is essential. The full potential of CDK4/6i is yet to be known, with many trials ongoing to expand their utility to other breast cancer subtypes, such as early breast cancer, and even to other cancers. Our research establishes the important idea that resistance to combined therapy (CDK4/6i + ET) can be due to resistance to endocrine therapy, to treatment with CDK4/6i, or to both. Individuals' responses to treatment are based mostly on genetic features and molecular markers, as well as the tumor's hallmarks; therefore, a future perspective is represented by personalized treatment based on the development of new biomarkers, and strategies to overcome drug resistance to combinations of ET and CDK4/6 inhibitors. The aim of our study was to centralize the mechanisms of resistance, and we believe that our work will have utility for everyone in the medical field who wants to deepen their knowledge about ET + CDK4/6 inhibitors resistance. CDK4/6 inhibitors advanced/metastatic breast cancer biomarkers of response progression on CDK4/6 inhibitors resistance mechanisms endocrine therapy Romanian National Society of Medical OncologyThe cost for the publication of this manuscript will be supported by the Romanian National Society of Medical Oncology. pmc1. Introduction Breast cancer has the highest number of new cases for both sexes and all ages, according to GLOBOCAN 2020. It is the second leading cause of mortality among women, and it has become a global health challenge. It is estimated that about 7.8 million women were diagnosed in 2021 . Unfortunately, the global burden of breast cancer is increasing both in developed countries and in developing ones . Breast cancer is grouped into four categories based on the immunohistochemical expression of hormone receptors: estrogen receptor positive (ER+), progesterone receptor positive (PR+), human epidermal growth factor receptor positive (HER2+), and triple-negative (TNBC), which is characterized by the lack of expression of any of the above receptors . We found it of great interest and intriguing that one of the latest studies on the regulation on signaling pathways, which highlighted that even natural products obtained from plants, fruits and vegetables (such as viridiflorol, verminoside, novel phloroglucinol derivatives, genistein, vulpinic acid, calcitrinone A, kaempferol, protopanaxadiol, thymoquinone, arctigenin, glycyrrhizin, 25-OCH3-PPD, oridonin, apigenin, wogonin, fisetin, curcumin, berberine, cimigenoside, and resveratrol) show anticancer activities against breast cancer through the inhibition of angiogenesis, cell migrations, proliferations, and tumor growth, as well as cell cycle arrest by inducing apoptosis and cell death, the downstream regulation of signaling pathways (such as Notch, NF-kB, PI3K/Akt/mTOR, MAPK/ERK, and NFAT-MDM2), and the regulation of EMT processes . The investigators actually concluded that natural products also act synergistically to overcome the drug resistance issue, thus improving their efficacy as an emerging therapeutic option for breast cancer therapy. However, in this review we stay focused on molecular resistance to the treatment of HR+/ cancer. One of the most common subtypes (20-25% of all breast cancers) is HR+/ cancer . Endocrine therapy (ET) is the main treatment for the HR+ luminal subtype of breast cancer, in association with targeted therapy. Cyclin-dependent kinases 4 and 6 inhibitors (CDK4/6i) restore the cell cycle by selectively inhibiting cyclin-dependent kinases 4 and 6, and block cell proliferation in a variety of tumor cells, including those of breast cancer . There are three CDK4/6 inhibitors approved by the US Food and Drug Administration that are transforming the treatment landscape nowadays: palbociclib, ribociclib, and abemaciclib (Table 1). They all have similar mechanisms of action and properties, with few differences in their preclinical and pharmacological settings and toxicity profiles . There is a need for a personalized approach to overcome the growing financial burden for health care systems through more effective patient selection. Palbociclib, ribociclib and abemaciclib are expensive anticancer drugs because they are currently protected by drug patents, and hence the need for predictive biomarkers of response beyond estrogen receptor positivity . 2. Aims and Objectives The management of breast cancer CDK4/6 inhibitor resistance is one of the most important clinical issues to be overcome, indicating a clear need for continuous discovery-based preclinical and clinical approaches. In order to assess these issues, we performed a systematic review of the published literature. The two key objectives were to identify resistance biomarkers and to understand molecular mechanisms underpinning drug resistance for CDK4/6 inhibition in breast cancer patients. Every single biomarker and signaling pathway was taken and discussed in separate paragraphs, highlighting the mechanism of possible resistance and its clinical and therapeutical implication. 3. Materials and Methods The databases used to gather information for this review include Pubmed.gov and Clinicaltrials.gov. We reviewed the PubMed database from January 2013 to January 2023 and selected all relevant articles. The inclusion criteria for this literature review encompassed studies that examined resistance to CDK4/6 inhibitors. The inclusion criteria were studies that evaluated and validated biomarkers of predictive response to therapy and potential mechanisms of resistance. Studies that addressed future directions after the progression of inhibitors were also assessed. Exclusion criteria were articles with unavailable abstracts, non-English-written articles, and conference presentations. Keywords used to search for references included CDK4/6 inhibitor, biomarker, progression, and resistance in order to achieve the most specific results. The search generated 75 results, but only 25 articles met our criteria. 4. Review 4.1. How Do CDK4/6 Inhibitors Work? The malignant transformation of normal cells begins with chaotic cellular proliferation, which takes place due to cell cycle dysregulation . The cell cycle has four important stages: G1 (cells grow, increasing in size), S (synthesis of the DNA), G2 (cells grow more and make proteins), and M (mitosis). In the end, the cell splits into two daughter cells . One of the most important cell cycle malfunctions starts right at the beginning of the cell cycle, which is controlled by the retinoblastoma protein (pRb). When in its active state, it stops the cell from progressing in the S phase by binding and suppressing E2F transcription factors. Phosphorylation of the Rb protein, which can be undertaken by the cyclin D-CDK4/6 complex, leads to E2F release. Thus, the cell can enter the S phase, and the cell cycle continues . In turn, the complex is activated through the PI3K/AKT/mTOR and RAS/MAPK pathways by the activation of hormone receptors (including the estrogen receptor (ER)) and growth factors . Obviously, the complex itself is downregulated by endogenous CDK inhibitors: the INK4 and Cip/Kip protein families . A schematic representation of how CDK4/6 inhibitors work can be found in Figure 1. There are several resistance mechanisms and potential biomarkers of response to CDK4/6 inhibitor regimens, which we will review in the upcoming paragraphs. 4.2. Cyclin D-CDK4/6 Abnormal Activation The most frequently encountered resistance to CDK4/6i is the upregulation of the Cyclin D-CDK4/6-pRb pathway . In a study conducted by Yang et al. in 2017, the majority of cells that were resistant to abemaciclib contained an amplification of CDK6 . While CDK6 amplification was demonstrated to have an impact on potential treatment resistance, both high and low levels of CDK4 have been seen in resistance models . In the same year, Gong et al. demonstrated that cells with the highest sensitivity to abemaciclib showed increased cyclin D activity, which promotes cyclin D1 turnover . Additionally, the overexpression of Cyclin D1 in breast cancer cells showed higher sensitivity to palbociclib . However, many studies demonstrated that the overexpression of Cyclin D, with or without Cyclin D1 gene amplification, occurred in more than 50% of breast cancer cells . Cyclin D1 is also a direct transcriptional target of ER , so the activation of the Cyclin D-CDK4/6 complex also contributes to endocrine therapy resistance . Another important down-regulatory component of the complex is the p16 protein (a member of the CDKN2/INK family), whose inactivity could also contribute to aggressive breast cancer . It is a tumor-suppressor protein that inhibits the activity of CDK4/6, and its expression correlates with a better prognosis in breast cancer patients. Low activity of p16 is correlated with increased CDK4/6 activity and increased sensitivity to palbociclib . 4.3. Loss of pRb The loss of G1/S control is a hallmark of cancer, and is often caused by the inactivation of the retinoblastoma pathway . As shown above, the integrity of the retinoblastoma protein is an important condition for the cells to be sensitive to CDK4/6 inhibitors, as it is at the center of the action mechanism. RB1 is the gene that encodes pRb, one of the most studied and reported biomarkers to date. Its loss or mutation is one of the most observed resistance mechanisms for CDK4/6i . However, pRb function loss prior to CDK4/6i treatment is uncommon in metastatic breast cancer with HR+/HER2- . In the PALOMA-3 study, only six out of 127 patients developed an RB1 loss of function after treatment with palbociclib and fulvestrant . Another study conducted by Li et al. found a statistically significant difference in progression-free survival (PFS) regarding treatment with CDK4/6i; 3.6 months for patients who had a loss of the RB1 gene, compared to 10.1 months for patients with intact RB1 . The first examples of acquired resistance were reported by Condorelli et al. , where acquired RB1 mutations were detected in ER-positive breast cancer patients treated with palbociclib and fulvestrant or ribociclib and letrozole. To determine the function of Rb phosphorylation by cyclin D-CDK4/6, Topacio and colleagues sought to generate variants of Rb that could no longer interact with cyclin D-Cdk4,6 while preserving all the other interactions with other cyclin-Cdk complexes . They analyzed the docking interactions between Rb and cyclin D-CDK4/6 complexes and found that cyclin D-CDK4/6 targets the Rb family of proteins for phosphorylation, primarily by docking a C-terminal alpha-helix, which is not recognized by the other major cell-cycle cyclin-CDK complexes cyclin E-CDK2, cyclin A-CDK2, and cyclin B-CDK1 . Their results showed that cyclin D-CDK4/6 phosphorylates and inhibits Rb via a C-terminal helix, and that this interaction is a major driver of cell proliferation . 4.4. Cyclin E-CDK2 Pathway Activation During a normal cell cycle, cyclin E1 and cyclin E2 can bind to and activate CDK2 in order to phosphorylate pRb, but only after it has already been phosphorylated by the cyclin D-CDK4/6 complex as a second wave of signaling . The activation of the cyclin E1/cyclin E2-CDK2 complex permits cells to bypass the inhibiting activity of CDK4/6 and encourage growth and proliferation . Therefore, the overexpression of cyclin E1, cyclin E2, and CDK2 can subvert the CDK4/6 inhibition . An interesting study conducted by Guarducci et al. showed that the ratio of cyclin E1 to RB1 level (not only cyclin E1 amplification and RB1 loss) is a poor prognostic factor and predicts palbociclib de novo resistance in HR+ breast cancer . Herrera-Abreu et al. demonstrate in a study from 2016 that cyclin E1 is upregulated via CDK2 activation in palbociclib-resistant cells (that were generated via chronic exposure to the drug and named palbociclib-resistant MCF-7 breast cancer cells) . In a phase II study (the NeoPalAna trial), researchers studied palbociclib resistance in patients with high levels of cyclin E1 . Cyclin E1 overexpression was also predictive of an abemaciclib response to targeted therapy, as shown in the study conducted by Gong X et al. . Next, gene expression analysis of 302 ER+ breast cancer samples from PALOMA-3 trial revealed that lower Cyclin E1 (CCNE1) mRNA levels were associated with a better response to palbociclib . This association was confirmed in a preoperative setting, in the cohort of POP (PreOperative Palbociclib) trial . Taking all this together, cyclin E1, cyclin E2, and CDK2 are upregulated in the CDK4/6 inhibitor resistance models . 4.5. PI3K/AKT/mTOR Pathway Activation This signaling pathway activation is another mechanism for both de novo and acquired resistance to CDK4/6i, with the hyperactivity of PI3K playing a role in endocrine-resistant mechanisms . PIK3CA mutations could be identified in almost 40% of breast cancers with hormonal receptors . Activating PIK3CA mutations could be a biomarker of either intrinsic resistance or acquired resistance. However, PI3KCA mutations have not been associated with resistance to CDK inhibitors in clinical studies to date . One study identified that the PI3K pathway kinase (PDK1) was overexpressed in ribociclib-resistant cells . Not only in ribociclib-resistant cell lines, but also in palbociclib-resistant cell lines, PIK3CA loss led to reduced proliferation of all cell lines regardless of RB status, as shown by Attia and colleagues in a 2020 study . There are works in the literature that suggest adding a PI3K inhibitor, such as alpelisib, to CDK4/6i in order to circumvent the resistance mechanisms that develop for CDK4/6. It could be added after progression on CDK4/6i and ET (endocrine therapy), or from the start in triple combination to prevent the onset of resistance to the combination of CDK4/6i and ET (via modulation of early adaptive response) . The mammalian target of rapamycin (mTOR) is implicated in cell cycle processes such as cell growth, size control, division, and proliferation, and it could be one of the reasons for CDK4/6i resistance. mTORC1 and mTORC2 are two different complexes that are formed by the mTOR kinase. A study conducted by Michaloglou and colleagues demonstrates that an mTORC1/mTORC2 inhibitor (vistusertib) could prevent early adaptive resistance to palbociclib in HR-positive breast cancer cells . According to the specialty literature, the most frequent therapy used after progression on CDK4/6i is the mTOR inhibitor (everolimus) . The AKT (serine/threonine kinase of the AGC kinase family) is activated via phosphorylation, which induces growth and survival. For this process, PDK1 (3-phosphoinositide dependent kinase 1) has an important role in the PI3K-AKT pathway. A low level of PDK1 makes tumor cells more sensitive to CDK4/6i . On the other hand, a high level of AKT1 activity was seen in palbociclib-resistant cells . 4.6. FGFR1 Activation (FGF/FGFR Signaling Pathway Activation) Fibroblast growth factor receptor 1 (FGFR1) is a protein of the tyrosine kinase family that plays an important role in the cell cycle, being implicated in the migration, proliferation, differentiation, and survival of the cells. In more than 15% of breast cancers with hormone receptors present, a mutation of FGFR1 is found . Thus, the causal relationship between FGFR1 mutations and endocrine therapy resistance has already been explained and demonstrated . It is also important to find out if there is a connection between these mutations and resistance to CDK4/6 inhibitors. In order to do this, Formisano and colleagues showed that the cells that overexpressed FGFR1 were resistant to ribociclib and fulvestrant, and they also demonstrated that the cells that received an FGFR1 tyrosine kinase inhibitor (lucitanib) reversed the resistance. Moreover, the study highlighted a shorter PFS rate in those with FGFR overexpression among patients enrolled in the MONALEESA-2 clinical trial . Surprisingly, the patients enrolled in the PALOMA-2 trial with FGFR2 amplification in the palbociclib + letrozol arm benefited from a longer PFS than those who were given placebo and letrozole . In a study from 2019, Drago and colleagues evaluated the clinical response to endocrine and targeted therapies in a cohort of 110 patients with HR+/ breast cancer and validated the functional role of FGFR1-amplification in mediating response/resistance to hormone therapy in vitro. The investigators concluded that, while FGFR1 amplification confers broad resistance to ER, PI3K, and CDK4/6 inhibitors, mTOR inhibitors might have a unique therapeutic role in the treatment of patients with ER+/FGFR1+ metastatic breast cancer . Another study conducted by Mouron et al. included 251 patients with HR+ breast cancer and studied the role of ER, CDK4/6, and/or FGFR1 blockade alone or in combinations in Rb phosphorylation, cell cycle, and survival. They showed how hormonal deprivation leads to FGFR1 overexpression, thus being associated with resistance to hormonal monotherapy or in combination with palbociclib. Both resistances have been reverted with triple ER, CDK4/6, and FGFR1 blockade . 4.7. RAS Activation The RAS family of protooncogenes encodes three oncogenes, KRAS, NRAS, and HRAS, each with important roles in the cell cycle, such as apoptosis, growth, and differentiation. Kirsten Rat Sarcoma Viral Oncogene Homolog (KRAS) is the most frequently mutated RAS gene . Many studies over the last years have revealed how the engagement of RAS function might result in mandatory downstream varied oncogenic alterations for progression, metastatic dissemination, and therapy resistance in breast cancers . In this direction, we found a review from 2019 conducted by Galie where he underlined the major studies over the last 30 years which have explored the role of RAS proteins and their mutation in breast cancer patients . An overexpression of KRAS has been associated over the years with many types of cancer growth and development, including breast cancer resistance to CDK4/6i. A study from 2021 by Raimondi et al., who enrolled 106 patients with HR+ metastatic breast cancer, showed resistance to palbociclib and fulvestrant in the cells that developed a KRAS amplification. Moreover, the PFS was just three months for the subjects with KRAS mutations, whereas, in the other arm, the PFS had not even been reached by the 18-month follow-up . Cells with KRAS, NRAS, and HRAS activating mutations are, therefore, susceptible to CDK4/6 inhibitor resistance . 4.8. FAT1 Loss Another important and well-studied biomarker of possible resistance to CDK4/6i is the loss of FAT1. FAT atypical cadherin 1 (FAT1) is among the most frequently mutated genes in many types of cancer . This is a tumor suppressor gene, a member of the cadherin superfamily, which interacts with beta-catenin and Hippo signaling pathways. It is found in 6% of metastatic HR+ breast cancers . Chen and colleagues performed a literature review on the diverse functions of FAT1 in cancer progression and presented the phenotypic alterations due to FAT1 mutations, several signaling pathways and tumor immune systems known or proposed to be regulated by this protein . A study conducted by Li et al. on 348 patients treated with CDK4/6 inhibitors highlighted, after genetic sequencing, that patients with loss of FAT1 had a lower PFS compared to those with intact FAT1 (2.4 months and 10.1 months, respectively) and rendered cells resistant to all three CDK4/6i. The investigators highlighted that FAT1 loss is also associated with CDK6 overexpression via downregulation of the Hippo signaling pathway (through YAP and TAZ transcription factors) . The role of FAT1 deleterious mutations was then confirmed in vivo. Cells with FAT1 knockout or knockdown did not stop cell growth upon exposure to abemaciclib, and MCF7-implanted xenografts experienced much less sensitivity to abemaciclib than mice with a non-mutated FAT1 gene . 4.9. PTEN Loss PTEN is a tumor suppressor gene and one of the frequently mutated genes in human cancers . The increased expression of PTEN leads to the inactivation of CDK, which enables the Rb1 to keep dephosphorylating, while binding to transcription factor E2F, which ultimately inhibits cell proliferation . The overexpression of AKT could reduce PTEN expression and render breast cancer cells resistant to CDK4/6i . Costa and colleagues performed an analysis of serial biopsies, which uncovered both RB and PTEN loss as mechanisms of acquired resistance to CDK4/6i. The investigators demonstrated that, in breast cancer cells, the ablation of PTEN through increased AKT activation was sufficient to promote resistance to CDK4/6 inhibition (ribociclib and letrozole) in vitro and in vivo; PTEN loss resulted in the exclusion of p27 from the nucleus, leading to increased activation of both CDK4 and CDK2 . PTEN loss is rare in treatment-naive ER-positive tumors . The loss of PTEN confers resistance to PI3K inhibitors (alpelisib) , as well as cross-resistance to CDK4/6i and PI3K inhibitors . Lee and colleagues conducted a retrospective analysis using real-world data, molecular biomarkers such as FGFR1 amplification, PTEN loss, and DNA repair pathway gene mutations, and showed a significant association of shorter PFS with CDK4/6i therapy . 4.10. S6K1 Amplification S6K1 is a conserved serine/threonine protein kinase that belongs to the family of protein kinases, being the principal kinase effector downstream of the mammalian target of rapamycin complex 1 (mTORC1) . S6K1 is an important regulator of cell size control, protein translation and cell proliferation . S6K1 is one of the best-characterized downstream targets of mTORC1, and rapamycin treatment results in rapid dephosphorylation and inactivation of S6K1 . The hyperactivation of mTORC1/S6K1 signaling may be closely related to ER-positive status in breast cancer, and may be utilized as a marker for prognosis and a therapeutic target . A study from 2012 highlights that the S6K1-ER relationship creates a positive feed-forward loop in the control of breast cancer cell proliferation and, furthermore, the co-dependent association between S6K1 and ERa may be exploited in the development of targeted breast cancer therapies . During the literature review, we found of interest a recent research article from August 2022 conducted by Mo and colleagues regarding S6K1 amplification. The Chinese investigators demonstrated that S6K1 amplification confers innate resistance to palbociclib and ET through activating c-Myc pathway in 36 patients with ER+ breast cancer. In those who had received palbociclib, patients with high-expressed S6K1 had significantly worse progression-free survival and significantly worse relapse-free survival than those with low S6K1 expression. S6K1 overexpression was sufficient to promote resistance to palbociclib. S6K1 overexpression increased the expression levels of G1/S transition-related proteins and the phosphorylation of Rb, mainly through the activation of the c-Myc pathway. Mo et al. showed that this resistance could be abrogated by the addition of the mTOR inhibitor, which blocked the upstream of S6K1, in vitro and in vivo . 4.11. AURKA Amplification Aurora kinase A (AURKA) belongs to the family of serine/threonine kinases, whose activation is necessary for cell division processes via the regulation of mitosis. AURKA shows significantly higher expression in cancer tissues than in normal control tissues for multiple tumor types . The amplification of the mitotic kinase AURKA has been identified in 11 out of 41 HR+ breast cancer biopsies from tumors resistant to CDK4/6 inhibitors, including examples of both intrinsic and acquired resistance, with no alterations detected in sensitive samples . Aurora A has been previously shown to mediate endocrine resistance through the downregulation of ER expression in an SMAD5-dependent manner . Two studies have shown that Aurora kinase A/B inhibition is synthetically lethal with RB1 deficiency in breast cancer and small-cell lung cancer cell lines , suggesting alternative therapeutic strategies for RB1-null tumors or new combinatorial strategies to prevent acquired resistances to CDK4/6 inhibitors . 4.12. c-Myc Upregulation c-Myc is a member of a family of protooncogenes that code for transcription factors, and is often overexpressed in cancer . It is activated by phosphorylation, and in this form c-Myc is stable and allows cells to escape senescence. CDK2 and CDK4/6 inhibition decreases the phosphorylation of c-Myc, which destabilizes the gene and allows cells to enter the apoptosis process . Mateyak et al. performed a comprehensive analysis and found that the largest defect observed in c-myc-/- cells was a 12-fold reduction in the activity of cyclin D1-CDK4/6 complexes during the G0 to S transition. The investigators suggested that c-Myc affects the cell cycle at multiple independent points, because the restoration of the CDK4 and 6 defect does not significantly increase growth rate . Pandey et al. concluded in a study from 2020 that overexpression of c-Myc leads to palbociclib-resistant cells . In the MONARCH-3 trial, 5% of patients with newly acquired c-Myc mutations were associated with resistance to abemaciclib + ET, and 9% of patients treated with abemaciclib alone in the MONARCH-1 trial acquired new Myc alterations . 4.13. miR Downregulation MicroRNAs are non-coding RNA molecules involved in the post-transcriptional regulation of gene expression and regulate 30-60% of the human genome. MicroRNAs regulate the cell cycle through cyclin-dependent kinases and cyclins. The downregulation of miRNAs negatively regulates CDK6, which leads to CDK6 activation. CDK6 activation results in palbociclib resistance, as shown by Li and colleagues in a study from 2020 . Moreover, in a retrospective analysis of 44 patients treated with CDK4/6i, microRNA levels were higher in those with intrinsic or acquired CDK4/6i resistance . Krasniqi et al. summarized in their study that some miRNAs (such as miR-326, miR-29b-3p, miR-126, and miR3613-3p) are associated with sensitivity to CDK4/6 inhibitors, whereas others (such as miR-432-5p, miR-223, and miR-106b) appear to confer treatment resistance . Identifying specific expression patterns of miRNAs could be a promising approach to study tumor response to CDK 4/6 inhibitors and exploit them as novel biomarkers . Non-coding RNAs have been demonstrated to be strictly lineage-specific; their expression may therefore determine cell phenotype, allowing for the identification of specific tumor sub-populations resistant to CDK inhibitors . 4.14. TK1 Activity Thymidine kinase 1 (TK1) is a DNA salvage pathway enzyme involved in regenerating thymidine for DNA synthesis and DNA damage . It catalyzes the conversion of thymidine to deoxythymidine monophosphate, which is further phosphorylated to triphosphates before its use for DNA synthesis . In resting cells, observable TK1 activity is low to absent, increasing during G1/S transcription and peaking at S phase . In healthy subjects, levels of TK1 are low to absent, with contrastingly elevated levels observed in patients with a range of malignancies, including breast cancer . TK1 is a phosphotransferase that plays a role in DNA replication, is regulated by the E2F pathway, and is downstream of CDK4/6. Its activity is a marker of tumor proliferation. TKs' activity has been shown to be a prognostic marker in patients with metastatic breast cancer, both when measured at baseline and during treatment. There are some clinical studies that support this statement . A prospective monitoring trial (ClinicalTrials.gov NCT01322893) from Sweden, in which 156 metastatic breast cancer patients planned to start first-line systemic therapy, has reported that the TK1 activity level is prognostic for survival (decreases in TK1 levels from 3 to 6 months correlate to improved survival PFS and OS) in patients with newly diagnosed metastatic breast cancer . McCartney and colleagues reported that intense TK1 activity is seen in cell lines resistant to palbociclib. The phase II TRend study also reported a shorter PFS for patients with high levels of TK1 than in the other arm (3 months vs. 9 months) . Another study (the ECLIPS trial) reported progressive disease in patients with metastatic breast cancer treated with palbociclib . In the NeoPalAna trial, investigators observed an important reduction in TK1 activity after the initiation of palbociclib, suggesting a reduction in tumor proliferation . 4.15. Endocrine Resistance and CDK4/6i Sensitivity--An Association Worthy of Consideration Endocrine treatment is one of the most important approaches when it comes to ER+ breast cancers, and for metastatic disease it becomes the physician's first choice, along with other targeted therapies (except in the case of a visceral crisis scenario, when chemotherapy should be the first choice). To date, some endocrine-resistant mechanisms have been described, including the upregulation of ER coactivators (e.g., FOXA1), cyclins (cyclin D and E), CDK proteins (CDK2 and CDK6), mitogen signaling pathways (PI3K and RAS pathways), or the downregulation of CDK inhibitor proteins (p16) . As already known, CDK4/6 inhibition acts downstream of endocrine therapy; therefore, some resistance mechanisms are common to both types of treatments (endocrine therapy and CDK4/6 inhibitors) . Among these resistance mechanisms, many studies and clinical trials have found a connection between estrogen receptor 1 (ESR1) mutations and acquired resistance to endocrine therapy. ESR1 mutations are the most important alterations resulting in resistance to aromatase inhibitor treatment, and can be found in almost 40% of metastatic breast cancer patients and in approximately 20% of patients with endocrine-resistant breast cancer . However, no association was found between ESR1 and CDK4/6i resistance. In the MONALEESA-2 trial, there was no correlation between ESR1 levels and response to ribociclib , and neither was there in the PALOMA-3 trial, where there was no link between ESR1 mutations and response to palbociclib . Moreover, the PFS was improved both for patients with ESR1 mutations and for patients with non-mutated ESR1, demonstrating that this mutation does not affect treatment response. However, in the PALOMA-3 trial, at the end of the treatment 12.8% of patients developed new mutations in the ESR1 gene, with the Y537S mutation in particular . Different results were observed in MONARCH-2, in which patients with ERS1 mutations showed an overall survival benefit . O'Leary and colleagues also investigated PIK3CA mutations and concluded that both PIK3CA and ESR1 mutations were evenly distributed in both arms of the study, which leads to the idea that these mutations are more likely to affect the response to fulvestrant than to palbociclib . The PALOMA-3 trial highlights the idea that ET resistance should be taken into consideration when talking about resistance to combination regimens in HR+/HER2-breast cancer. There is also an ongoing trial from Johns Hopkins University (NCT03439735) that studies the association between ESR1 mutations and clinical outcomes in patients treated with palbociclib and aromatase inhibitor as a first-line treatment regimen; its reported results should be available in June 2024. However, all three pivotal clinical trials (PALOMA-3, MONARCH-2, and MONALEESA-3) demonstrated that CDK4/6 inhibitors prolong PFS even after ET resistance, which demonstrates that CDK4/6i maintain effectiveness regardless of the endocrine-resistant disease. Additionally, endocrine-resistant tumors maintain sensitivity to CDK4/6 inhibitors, particularly when they are used in association with ET . 5. Discussions CDK4/6 inhibitors remain a landmark for the treatment of hormone receptor-positive and human epidermal growth factor receptor 2-negative metastatic breast cancer, being the most significant advance in the last decade. Various preclinical and translational research efforts have begun to shed light on the genomic and molecular landscape of resistance to these agents . As we showed above, it is important to understand the mechanism of action of CDK4/6 inhibitors in order to target specific signaling pathways and predictive biomarkers of response, taking into consideration that intrinsic and acquired resistance could limit the activity of these inhibitors. In addition, one of the greatest challenges is distinguishing between mechanisms causing resistance to CDK4/6 inhibition and endocrine resistance. Approximately 10% of patients will have primary resistance to CDK4/6 inhibitors . For instance, patients with evidence of functional Rb loss at baseline are not likely to benefit from CDK4/6 inhibition, or from increased cyclin E1/E2 expression. A rise in TK1 activity may also provide a marker of early resistance . Mutations in RB1, resulting in the activation of other cell cycle factors, such as E2F and the Cyclin E-CDK2 axis, have been demonstrated in cases of acquired resistance . In the table below (Table 2), we summarized the main resistance mechanisms and biomarkers of resistance, which we have previously reviewed. Following progression, no prospective randomized data exist to help guide second-line treatment . While prospective data are needed, analysis of real-world data suggests a survival benefit for the continuation of CDK4/6i beyond a frontline progression for patients with HR+/ breast cancer . Several ongoing Phase 1 and 2 trials (MAINTAIN NCT02632045, PACE NCT03147287, NCT01857193, NCT 02871791, and TRINITI-1 NCT 02732119) are investigating the potential benefit of continuing CDK4/6i beyond progression . For more successful treatment, biomarkers are of potential interest in order to identify patients who might be responsive or not to CDK4/6 inhibitors, facilitating an early switch to a more efficacious treatment. 6. Conclusions To date, no biomarker has been studied enough to be approved as a predictor of response to treatment or a targeted signaling pathway. Personalized treatment based on an individual's response and tumor genomics represents the future of oncology. Therefore, it is a justification for future clinical trials because the identification of biomarkers of resistance is still a problem universally, and there is still more to be discovered about CDK4/6 inhibitor resistance. The optimum management of HR+/HER2-metastatic breast cancer is essential for patients as they might have only one more card to play, so future therapeutic targets should be examined in clinical trials to delay or overcome treatment resistance to combinations of ET and CDK4/6 inhibitors. In conclusion, we strongly believe that the validation of proposed biomarkers should be an option to consider before starting treatment with CDK4/6 inhibitors and hormonal therapy. This can be carried out via whole exome and targeted sequencing of solid and liquid biopsies, in order to reveal several possible genomic alterations that could change the course of treatment. In Romania, unfortunately there are few patients who can afford the costs of this type of testing. After doing such exhaustive research for this review, our personal opinion is that some biomarkers are worth testing more than others, such as loss of retinoblastoma protein. Some mechanisms of resistance, such as PI3K/AKT/mTOR or Cyclin E-CDK2 pathway activation, have already had their implication validated in resistance to CDK4/6i + ET; therefore it would be a worthy idea to take into consideration before starting the treatment. Breast cancer patients, maybe more than any other patients, are susceptible to depression and self-esteem loss, thus making any kind of treatment more difficult. We believe that a good start is always a better start and we do hope that in the near future breast cancer patients would benefit from the best personalized treatment. Author Contributions I.-M.S. performed the literature review and drafted the manuscript; C.F.P., A.I.P., C.M.O. and R.C.V. revised and improved the manuscript; resources, C.O.-S.; supervision, I.C.I., A.M.P. and C.N.; manuscript revision: I.-M.S. and C.F.P. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Key mechanisms of action of CDK4/6 inhibitors in HR+/ cancer. diagnostics-13-00987-t001_Table 1 Table 1 Pivotal studies on the three approved CDK4/6 inhibitors. Study Description Phase Number of Patients Enrolled Median PFS Identifier NCT Status Treatment 1st line PALOMA-1 Palbociclib + Letrozol vs. Placebo + Letrozole II 177 18.1 m vs. 11.1 m NCT00721409 Completed Palbociclib 125 mg/d orally for 3 weeks, followed by 1 week off Letrozole 2.5 mg/d orally on a continuous regimen Placebo 125 mg/d orally for 3 weeks, followed by 1 week off PALOMA-2 Palbociclib + Letrozol vs. Placebo + Letrozole III 666 24.8 m vs. 14.5 m NCT01740427 Active, not recruiting Palbociclib 125 mg/d orally for 3 weeks, followed by 1 week off Letrozole 2.5 mg/d orally on a continuous regimen Placebo 125 mg/d orally for 3 weeks, followed by 1 week off MONALEESA-2 Ribociclib + Letrozol vs. Placebo + Letrozole III 668 25.3 m vs. 16.0 m NCT01958021 Active, not recruiting Ribociclib 200 mg x 3/d orally for 3 weeks, followed by 1 week off Letrozole 2.5 mg/d orally on a continuous regimen Placebo 200 mg x 3/d orally for 3 weeks, followed by 1 week off MONALEESA-7 Ribociclib + Goserelin + Tamoxifen/Letrozole/Anastrozole vs. Placebo +Goserelin +Tamoxifen/Letrozole/Anastrozole III 672 23.8 m vs. 13.0 m NCT02278120 Active, not recruiting Ribociclib 200 mg x 3/d orally for 3 weeks, followed by 1 week off Goserelin 3.6 mg subcutaneous injection once every 28 days Letrozole 2.5 mg/d orally or Anastrozole 1 mg/d orally or Tamoxifen 20 mg/d orally on a continuous regimen Placebo 200 mg x 3/d orally for 3 weeks, followed by 1 week off MONARCH-3 Abemaciclib + Letrozole/Anastrozole vs. Placebo + Letrozole/Anastrozole III 493 28.18 m vs. 14.76 m NCT02246621 Active, not recruiting Abemaciclib 150 mg x 2/d orally Letrozole 2.5 mg/d orally or Anastrozole 1 mg/d orally Placebo 150 mg x 2/d orally 2nd line PALOMA-3 Palbociclib + Fulvestrant vs. Placebo + Fulvestrant III 521 9.50 m vs. 4.60 m NCT01942135 Completed Palbociclib 125 mg/d orally for 3 weeks, followed by 1 week off Fulvestrant 500 mg intramuscular injection on day 1 and day 15 of cycle 1 and then on day 1 of each cycle Placebo 125 mg/d orally for 3 weeks, followed by 1 week off MONARCH-2 Abemaciclib + Fulvestrant vs. Placebo + Fulvestrant III 669 16.40 m vs. 9.30 m NCT02107703 Active, not recruiting Abemaciclib 150 mg x 2/d orally Fulvestrant 500 mg intramuscular injection on day 1 and day 15 of cycle 1 and then on day 1 of each cycle Placebo 150 mg x 2/d orally Later line MONARCH-1 Abemaciclib alone (one arm clinical trial) II 132 5.95 m NCT02102490 Completed Abemaciclib 200 mg x 2/d orally diagnostics-13-00987-t002_Table 2 Table 2 The main resistance mechanisms and biomarkers of resistance to CDK 4/6 inhibitors. Activation Downregulation ET Resistance Cyclin E1, E2 RB ESR1 mutations PI3K/AKT/mTOR FAT1 AKT1 PTEN CDK2 miR CDK6 FGFR1 MYC RAS AURKA S6K1 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Jha V. Devkar S. Gharat K. Kasbe S. Matharoo D.K. Pendse S. Bhosale A. Bhargava A. Screening of Phytochemicals as Potential Inhibitors of Breast Cancer using Structure Based Multitargeted Molecular Docking Analysis Phytomedicine Plus 2022 2 100227 10.1016/j.phyplu.2022.100227 2. Nassif A.B. Talib M.A. Nasir Q. Afadar Y. Elgendy O. Breast cancer detection using artificial intelligence techniques: A systematic literature review Artif. Intell. Med. 2022 127 102276 10.1016/j.artmed.2022.102276 35430037 3. Shaath H. Elango R. Alajez N.M. Molecular Classification of Breast Cancer Utilizing Long Non-Coding RNA (lncRNA) Transcriptomes Identifies Novel Diagnostic lncRNA Panel for Triple-Negative Breast Cancer Cancers 2021 13 5350 10.3390/cancers13215350 34771513 4. Naeem M. Iqbal M.O. Khan H. Ahmed M.M. Farooq M. Aadil M.M. Jamaludin M.I. Hazafa A. Tsai W.-C. A Review of Twenty Years of Research on the Regulation of Signaling Pathways by Natural Products in Breast Cancer Molecules 2022 27 3412 10.3390/molecules27113412 35684353 5. Ran R. Ma Y. Wang H. Yang J. Yang J. Treatment strategies for hormone receptor-positive, human epidermal growth factor receptor 2-positive (HR+/HER2+) metastatic breast cancer: A review Front. Oncol. 2022 12 975463 10.3389/fonc.2022.975463 36620573 6. Huang J. Zheng L. Sun Z. Li J. CDK4/6 inhibitor resistance mechanisms and treatment strategies (Review) Int. J. Mol. Med. 2022 50 128 10.3892/ijmm.2022.5184 36043521 7. George M.A. Qureshi S. Omene C. Toppmeyer D.L. Ganesan S. Clinical and Pharmacologic Differences of CDK4/6 Inhibitors in Breast Cancer Front. Oncol. 2021 11 693104 10.3389/fonc.2021.693104 34327137 8. Asghar U.S. Kanani R. Roylance R. Mittnacht S. Systematic Review of Molecular Biomarkers Predictive of Resistance to CDK4/6 Inhibition in Metastatic Breast Cancer JCO Precis. Oncol. 2022 6 e2100002 10.1200/PO.21.00002 35005994 9. Hanahan D. Weinberg R.A. Hallmarks of Cancer: The Next Generation Cell 2011 144 646 674 10.1016/j.cell.2011.02.013 21376230 10. Phases of the Cell Cycle (Article)|Khan Academy Available online: (accessed on 8 February 2023) 11. Portman N. Alexandrou S. Carson E. Wang S. Lim E. Caldon C.E. Overcoming CDK4/6 inhibitor resistance in ER-positive breast cancer Endocr. Relat. Cancer 2019 26 R15 R30 10.1530/ERC-18-0317 30389903 12. Sharifi M.N. Anandan A. Grogan P. O'Regan R.M. Therapy after cyclin-dependent kinase inhibition in metastatic hormone receptor-positive breast cancer: Resistance mechanisms and novel treatment strategies Cancer 2020 126 3400 3416 10.1002/cncr.32931 32426848 13. Scheidemann E.R. Shajahan-Haq A.N. Resistance to CDK4/6 Inhibitors in Estrogen Receptor-Positive Breast Cancer Int. J. Mol. Sci. 2021 22 12292 10.3390/ijms222212292 34830174 14. Yang C. Li Z. Bhatt T. Dickler M. Giri D. Scaltriti M. Baselga J. Rosen N. Chandarlapaty S. Acquired CDK6 amplification promotes breast cancer resistance to CDK4/6 inhibitors and loss of ER signaling and dependence Oncogene 2017 36 2255 2264 10.1038/onc.2016.379 27748766 15. Gong X. Litchfield L.M. Webster Y. Chio L.-C. Wong S.S. Stewart T.R. Dowless M. Dempsey J. Zeng Y. Torres R. Genomic Aberrations that Activate D-type Cyclins Are Associated with Enhanced Sensitivity to the CDK4 and CDK6 Inhibitor Abemaciclib Cancer Cell 2017 32 761 776.e6 10.1016/j.ccell.2017.11.006 29232554 16. Li Z. Razavi P. Li Q. Toy W. Liu B. Ping C. Hsieh W. Sanchez-Vega F. Brown D.N. Da Cruz Paula A.F. Loss of the FAT1 Tumor Suppressor Promotes Resistance to CDK4/6 Inhibitors via the Hippo Pathway Cancer Cell 2018 34 893 905.e8 10.1016/j.ccell.2018.11.006 30537512 17. Zhang J. Wang Q. Wang Q. Cao J. Sun J. Zhu Z. Mechanisms of resistance to estrogen receptor modulators in ER+/ breast cancer Cell. Mol. Life Sci. 2020 77 559 572 10.1007/s00018-019-03281-4 31471681 18. Watts C.K.W. Sweeney K.J.E. Warlters A. Musgrove E.A. Sutherland R.L. Antiestrogen regulation of cell cycle progression and cyclin D1 gene expression in MCF-7 human breast cancer cells Breast Cancer Res. Treat. 1994 31 95 105 10.1007/BF00689680 7981461 19. Schoninger S.F. Blain S.W. The Ongoing Search for Biomarkers of CDK4/6 Inhibitor Responsiveness in Breast Cancer Mol. Cancer Ther. 2020 19 3 12 10.1158/1535-7163.MCT-19-0253 31909732 20. van Harn T. Foijer F. van Vugt M. Banerjee R. Yang F. Oostra A. Joenje H. te Riele H. Loss of Rb proteins causes genomic instability in the absence of mitogenic signaling Genes Dev. 2010 24 1377 1388 10.1101/gad.580710 20551164 21. Gomatou G. Trontzas I. Ioannou S. Drizou M. Syrigos N. Kotteas E. Mechanisms of resistance to cyclin-dependent kinase 4/6 inhibitors Mol. Biol. Rep. 2021 48 915 925 10.1007/s11033-020-06100-3 33409716 22. Xi J. Ma C.X. Sequencing Endocrine Therapy for Metastatic Breast Cancer: What Do We Do After Disease Progression on a CDK4/6 Inhibitor? Curr. Oncol. Rep. 2020 22 57 10.1007/s11912-020-00917-8 32415339 23. O'Leary B. Cutts R.J. Liu Y. Hrebien S. Huang X. Fenwick K. Andre F. Loibl S. Loi S. Garcia-Murillas I. The Genetic Landscape and Clonal Evolution of Breast Cancer Resistance to Palbociclib plus Fulvestrant in the PALOMA-3 Trial Cancer Discov. 2018 8 1390 1403 10.1158/2159-8290.CD-18-0264 30206110 24. Condorelli R. Spring L. O'Shaughnessy J. Lacroix L. Bailleux C. Scott V. Dubois J. Nagy R.J. Lanman R.B. Iafrate A.J. Polyclonal RB1 mutations and acquired resistance to CDK 4/6 inhibitors in patients with metastatic breast cancer Ann. Oncol. 2018 29 640 645 10.1093/annonc/mdx784 29236940 25. Topacio B.R. Zatulovskiy E. Cristea S. Xie S. Tambo C.S. Rubin S.M. Sage J. Koivomagi M. Skotheim J.M. Cyclin D-Cdk4,6 Drives Cell-Cycle Progression via the Retinoblastoma Protein's C-Terminal Helix Mol. Cell 2019 74 758 770.e4 10.1016/j.molcel.2019.03.020 30982746 26. Guarducci C. Bonechi M. Benelli M. Biagioni C. Boccalini G. Romagnoli D. Verardo R. Schiff R. Osborne C.K. De Angelis C. Cyclin E1 and Rb modulation as common events at time of resistance to palbociclib in hormone receptor-positive breast cancer Npj Breast Cancer 2018 4 38 10.1038/s41523-018-0092-4 30511015 27. Herrera-Abreu M.T. Palafox M. Asghar U. Rivas M.A. Cutts R.J. Garcia-Murillas I. Pearson A. Guzman M. Rodriguez O. Grueso J. Early Adaptation and Acquired Resistance to CDK4/6 Inhibition in Estrogen Receptor-Positive Breast Cancer Cancer Res. 2016 76 2301 2313 10.1158/0008-5472.CAN-15-0728 27020857 28. Ma C.X. Gao F. Luo J. Northfelt D.W. Goetz M. Forero A. Hoog J. Naughton M. Ademuyiwa F. Suresh R. NeoPalAna: Neoadjuvant Palbociclib, a Cyclin-Dependent Kinase 4/6 Inhibitor, and Anastrozole for Clinical Stage 2 or 3 Estrogen Receptor-Positive Breast Cancer Clin. Cancer Res. 2017 23 4055 4065 10.1158/1078-0432.CCR-16-3206 28270497 29. Miller T.W. Rexer B.N. Garrett J.T. Arteaga C.L. Mutations in the phosphatidylinositol 3-kinase pathway: Role in tumor progression and therapeutic implications in breast cancer Breast Cancer Res. 2011 13 224 10.1186/bcr3039 22114931 30. Turner N.C. Liu Y. Zhu Z. Loi S. Colleoni M. Loibl S. DeMichele A. Harbeck N. Andre F. Bayar M.A. Cyclin E1 Expression and Palbociclib Efficacy in Previously Treated Hormone Receptor-Positive Metastatic Breast Cancer J. Clin. Oncol. 2019 37 1169 1178 10.1200/JCO.18.00925 30807234 31. Arnedos M. Bayar M.A. Cheaib B. Scott V. Bouakka I. Valent A. Adam J. Leroux-Kozal V. Marty V. Rapinat A. Modulation of Rb phosphorylation and antiproliferative response to palbociclib: The preoperative-palbociclib (POP) randomized clinical trial Ann. Oncol. 2018 29 1755 1762 10.1093/annonc/mdy202 29893769 32. Attia Y.M. Shouman S.A. Salama S.A. Ivan C. Elsayed A.M. Amero P. Rodriguez-Aguayo C. Lopez-Berestein G. Blockade of CDK7 Reverses Endocrine Therapy Resistance in Breast Cancer Int. J. Mol. Sci. 2020 21 2974 10.3390/ijms21082974 32340192 33. O'Brien N.A. McDermott M.S.J. Conklin D. Luo T. Ayala R. Salgar S. Chau K. DiTomaso E. Babbar N. Su F. Targeting activated PI3K/mTOR signaling overcomes acquired resistance to CDK4/6-based therapies in preclinical models of hormone receptor-positive breast cancer Breast Cancer Res. 2020 22 89 10.1186/s13058-020-01320-8 32795346 34. Michaloglou C. Crafter C. Siersbaek R. Delpuech O. Curwen J.O. Carnevalli L.S. Staniszewska A.D. Polanska U.M. Cheraghchi-Bashi A. Lawson M. Combined Inhibition of mTOR and CDK4/6 Is Required for Optimal Blockade of E2F Function and Long-term Growth Inhibition in Estrogen Receptor-positive Breast Cancer Mol. Cancer Ther. 2018 17 908 920 10.1158/1535-7163.MCT-17-0537 29483206 35. Migliaccio I. Bonechi M. McCartney A. Guarducci C. Benelli M. Biganzoli L. Di Leo A. Malorni L. CDK4/6 inhibitors: A focus on biomarkers of response and post-treatment therapeutic strategies in hormone receptor-positive HER2-negative breast cancer Cancer Treat. Rev. 2021 93 102136 10.1016/j.ctrv.2020.102136 33360919 36. Jansen V.M. Bhola N.E. Bauer J.A. Formisano L. Lee K.-M. Hutchinson K.E. Witkiewicz A.K. Moore P.D. Estrada M.V. Sanchez V. Kinome-Wide RNA Interference Screen Reveals a Role for PDK1 in Acquired Resistance to CDK4/6 Inhibition in ER-Positive Breast Cancer Cancer Res. 2017 77 2488 2499 10.1158/0008-5472.CAN-16-2653 28249908 37. Lenihan C. Bouchekioua-Bouzaghou K. Shia A. Wilkes E. Casado-Izquierdo1 P. Cutillas P. Schmid P. Abstract P3-06-02: Characterization of resistance to the selective CDK4/6 inhibitor palbociclib in ER positive breast cancer Cancer Res. 2016 76 P3-06-02 10.1158/1538-7445.SABCS15-P3-06-02 38. Finn R. Liu Y. Martin M. Rugo H. Dieras V. Im S.-A. Gelmon K. Harbeck N. Zhu Z. Lu D. Abstract P2-09-10: Comprehensive gene expression biomarker analysis of CDK 4/6 and endocrine pathways from the PALOMA-2 study Cancer Res. 2018 78 P2-09-10 10.1158/1538-7445.SABCS17-P2-09-10 39. Drago J.Z. Formisano L. Juric D. Niemierko A. Servetto A. Wander S.A. Spring L.M. Vidula N. Younger J. Peppercorn J. FGFR1 Amplification Mediates Endocrine Resistance but Retains TORC Sensitivity in Metastatic Hormone Receptor-Positive (HR+) Breast Cancer Clin. Cancer Res. 2019 25 6443 6451 10.1158/1078-0432.CCR-19-0138 31371343 40. Raimondi L. Raimondi F.M. Pietranera M. Di Rocco A. Di Benedetto L. Miele E. Lazzeroni R. Cimino G. Spinelli G.P. Assessment of Resistance Mechanisms and Clinical Implications in Patients with KRAS Mutated-Metastatic Breast Cancer and Resistance to CDK4/6 Inhibitors Cancers 2021 13 1928 10.3390/cancers13081928 33923563 41. Mouron S. Manso L. Caleiras E. Rodriguez-Peralto J.L. Rueda O.M. Caldas C. Colomer R. Quintela-Fandino M. Bueno M.J. FGFR1 amplification or overexpression and hormonal resistance in luminal breast cancer: Rationale for a triple blockade of ER, CDK4/6, and FGFR1 Breast Cancer Res. 2021 23 21 10.1186/s13058-021-01398-8 33579347 42. Galie M. RAS as Supporting Actor in Breast Cancer Front. Oncol. 2019 9 1199 10.3389/fonc.2019.01199 31781501 43. Chen Z.G. Saba N.F. Teng Y. The diverse functions of FAT1 in cancer progression: Good, bad, or ugly? J. Exp. Clin. Cancer Res. 2022 41 248 10.1186/s13046-022-02461-8 35965328 44. Myc Wikipedia 2023 Available online: (accessed on 27 January 2023) 45. Pandey K. Park N. Park K.-S. Hur J. Cho Y.B. Kang M. An H.-J. Kim S. Hwang S. Moon Y.W. Combined CDK2 and CDK4/6 Inhibition Overcomes Palbociclib Resistance in Breast Cancer by Enhancing Senescence Cancers 2020 12 3566 10.3390/cancers12123566 33260316 46. Dosil M.A. Mirantes C. Eritja N. Felip I. Navaridas R. Gatius S. Santacana M. Colas E. Moiola C. Schoenenberger J.A. Palbociclib has antitumour effects on endometrial neoplasias: Inhibition of cyclin D-CDK4/6 axis in Pten-deficient neoplasias J. Pathol. 2017 242 152 164 10.1002/path.4896 28349562 47. Wang B. Li R. Wu S. Liu X. Ren J. Li J. Bi K. Wang Y. Jia H. Breast Cancer Resistance to Cyclin-Dependent Kinases 4/6 Inhibitors: Intricacy of the Molecular Mechanisms Front. Oncol. 2021 11 651541 10.3389/fonc.2021.651541 34123801 48. Juric D. Castel P. Griffith M. Griffith O.L. Won H.H. Ellis H. Ebbesen S.H. Ainscough B.J. Ramu A. Iyer G. Convergent loss of PTEN leads to clinical resistance to a PI(3)Ka inhibitor Nature 2015 518 240 244 10.1038/nature13948 25409150 49. Costa C. Wang Y. Ly A. Hosono Y. Murchie E. Walmsley C.S. Huynh T. Healy C. Peterson R. Yanase S. PTEN Loss Mediates Clinical Cross-Resistance to CDK4/6 and PI3Ka Inhibitors in Breast Cancer Cancer Discov. 2020 10 72 85 10.1158/2159-8290.CD-18-0830 31594766 50. The Cancer Genome Atlas Network Comprehensive molecular portraits of human breast tumours Nature 2012 490 61 70 10.1038/nature11412 23000897 51. Pereira B. Chin S.-F. Rueda O.M. Vollan H.-K.M. Provenzano E. Bardwell H.A. Pugh M. Jones L. Russell R. Sammut S.-J. The somatic mutation profiles of 2,433 breast cancers refine their genomic and transcriptomic landscapes Nat. Commun. 2016 7 11479 10.1038/ncomms11479 27161491 52. Razavi P. Dickler M.N. Shah P.D. Toy W. Brown D.N. Won H.H. Li B.T. Shen R. Vasan N. Modi S. Alterations in PTEN and ESR1 promote clinical resistance to alpelisib plus aromatase inhibitors Nat. Cancer 2020 1 382 393 10.1038/s43018-020-0047-1 32864625 53. Lee J.S. Yost S.E. Li S.M. Cui Y. Frankel P.H. Yuan Y.-C. Schmolze D. Egelston C.A. Guo W. Murga M. Genomic Markers of CDK 4/6 Inhibitor Resistance in Hormone Receptor Positive Metastatic Breast Cancer Cancers 2022 14 3159 10.3390/cancers14133159 35804935 54. Holz M.K. The role of S6K1 in ER-positive breast cancer Cell Cycle 2012 11 3159 3165 10.4161/cc.21194 22895181 55. Fingar D.C. Blenis J. Target of rapamycin (TOR): An integrator of nutrient and growth factor signals and coordinator of cell growth and cell cycle progression Oncogene 2004 23 3151 3171 10.1038/sj.onc.1207542 15094765 56. Hay N. Sonenberg N. Upstream and downstream of mTOR Genes Dev. 2004 18 1926 1945 10.1101/gad.1212704 15314020 57. Maruani D.M. Spiegel T.N. Harris E.N. Shachter A.S. Unger H.A. Herrero-Gonzalez S. Holz M.K. Estrogenic regulation of S6K1 expression creates a positive regulatory loop in control of breast cancer cell proliferation Oncogene 2012 31 5073 5080 10.1038/onc.2011.657 22286763 58. Mo H. Liu X. Xue Y. Chen H. Guo S. Li Z. Wang S. Li C. Han J. Fu M. S6K1 amplification confers innate resistance to CDK4/6 inhibitors through activating c-Myc pathway in patients with estrogen receptor-positive breast cancer Mol. Cancer 2022 21 171 10.1186/s12943-022-01642-5 36042494 59. Du R. Huang C. Liu K. Li X. Dong Z. Targeting AURKA in Cancer: Molecular mechanisms and opportunities for Cancer therapy Mol. Cancer 2021 20 15 10.1186/s12943-020-01305-3 33451333 60. Wander S.A. Cohen O. Gong X. Johnson G.N. Buendia-Buendia J. Lloyd M.R. Kim D. Luo F. Mao P. Helvie K. The genomic landscape of intrinsic and acquired resistance to cyclin-dependent kinase 4/6 inhibitors in patients with hormone receptor positive metastatic breast cancer Cancer Discov. 2020 10 1174 1193 10.1158/2159-8290.CD-19-1390 32404308 61. Opyrchal M. Salisbury J.L. Zhang S. McCubrey J. Hawse J. Goetz M.P. Lomberk G.A. Haddad T. Degnim A. Lange C. Aurora-A Mitotic Kinase Induces Endocrine Resistance through Down-Regulation of ERa Expression in Initially ERa+ Breast Cancer Cells PLoS ONE 2014 9 e96995 10.1371/journal.pone.0096995 24816249 62. Gong X. Du J. Parsons S.H. Merzoug F.F. Webster Y. Iversen P.W. Chio L.-C. Van Horn R.D. Lin X. Blosser W. Aurora A Kinase Inhibition Is Synthetic Lethal with Loss of the RB1 Tumor Suppressor Gene Cancer Discov. 2019 9 248 263 10.1158/2159-8290.CD-18-0469 30373917 63. Oser M.G. Fonseca R. Chakraborty A.A. Brough R. Spektor A. Jennings R.B. Flaifel A. Novak J.S. Gulati A. Buss E. Cells Lacking the RB1 Tumor Suppressor Gene Are Hyperdependent on Aurora B Kinase for Survival Cancer Discov. 2019 9 230 247 10.1158/2159-8290.CD-18-0389 30373918 64. Goetz M.P. Hamilton E.P. Campone M. Hurvitz S.A. Cortes J. Johnston S.R.D. Jerusalem G.H.M. Graham H. Wang H. Litchfield L. Acquired genomic alterations in circulating tumor DNA from patients receiving abemaciclib alone or in combination with endocrine therapy J. Clin. Oncol. 2020 38 3519 10.1200/JCO.2020.38.15_suppl.3519 65. Mateyak M.K. Obaya A.J. Sedivy J.M. c-Myc Regulates Cyclin D-Cdk4 and -Cdk6 Activity but Affects Cell Cycle Progression at Multiple Independent Points Mol. Cell. Biol. 1999 19 4672 4683 10.1128/MCB.19.7.4672 10373516 66. Ji W. Zhang W. Wang X. Shi Y. Yang F. Xie H. Zhou W. Wang S. Guan X. c-myc regulates the sensitivity of breast cancer cells to palbociclib via c-myc/miR-29b-3p/CDK6 axis Cell Death Dis. 2020 11 760 10.1038/s41419-020-02980-2 32934206 67. Cornell L. Wander S.A. Visal T. Wagle N. Shapiro G.I. MicroRNA-Mediated Suppression of the TGF-b Pathway Confers Transmissible and Reversible CDK4/6 Inhibitor Resistance Cell Rep. 2019 26 2667 2680.e7 10.1016/j.celrep.2019.02.023 30840889 68. McCartney A. Bonechi M. De Luca F. Biagioni C. Curigliano G. Moretti E. Minisini A.M. Bergqvist M. Benelli M. Migliaccio I. Plasma Thymidine Kinase Activity as a Biomarker in Patients with Luminal Metastatic Breast Cancer Treated with Palbociclib within the TREnd Trial Clin. Cancer Res. 2020 26 2131 2139 10.1158/1078-0432.CCR-19-3271 31937617 69. Del Re M. Bertolini I. Crucitta S. Fontanelli L. Rofi E. De Angelis C. Diodati L. Cavallero D. Gianfilippo G. Salvadori B. Overexpression of TK1 and CDK9 in plasma-derived exosomes is associated with clinical resistance to CDK4/6 inhibitors in metastatic breast cancer patients Breast Cancer Res. Treat. 2019 178 57 62 10.1007/s10549-019-05365-y 31346846 70. Krasniqi E. Goeman F. Pulito C. Palcau A.C. Ciuffreda L. Di Lisa F.S. Filomeno L. Barba M. Pizzuti L. Cappuzzo F. Biomarkers of Response and Resistance to CDK4/6 Inhibitors in Breast Cancer: Hints from Liquid Biopsy and microRNA Exploration Int. J. Mol. Sci. 2022 23 14534 10.3390/ijms232314534 36498861 71. Pulito C. Cristaudo A. Porta C.L. Zapperi S. Blandino G. Morrone A. Strano S. Oral mucositis: The hidden side of cancer therapy J. Exp. Clin. Cancer Res. 2020 39 210 10.1186/s13046-020-01715-7 33028357 72. Bitter E.E. Townsend M.H. Erickson R. Allen C. O'Neill K.L. Thymidine kinase 1 through the ages: A comprehensive review Cell Biosci. 2020 10 138 10.1186/s13578-020-00493-1 33292474 73. Welin M. Kosinska U. Mikkelsen N.-E. Carnrot C. Zhu C. Wang L. Eriksson S. Munch-Petersen B. Eklund H. Structures of thymidine kinase 1 of human and mycoplasmic origin Proc. Natl. Acad. Sci. USA 2004 101 17970 17975 10.1073/pnas.0406332102 15611477 74. Sherley J.L. Kelly T.J. Regulation of human thymidine kinase during the cell cycle J. Biol. Chem. 1988 263 8350 8358 10.1016/S0021-9258(18)68484-4 3372530 75. Finn R.S. Crown J.P. Lang I. Boer K. Bondarenko I.M. Kulyk S.O. Ettl J. Patel R. Pinter T. Schmidt M. The cyclin-dependent kinase 4/6 inhibitor palbociclib in combination with letrozole versus letrozole alone as first-line treatment of oestrogen receptor-positive, HER2-negative, advanced breast cancer (PALOMA-1/TRIO-18): A randomised phase 2 study Lancet Oncol. 2015 16 25 35 10.1016/S1470-2045(14)71159-3 25524798 76. Larsson A.-M. Bendahl P.-O. Aaltonen K. Jansson S. Forsare C. Bergqvist M. Jorgensen C.L.T. Ryden L. Serial evaluation of serum thymidine kinase activity is prognostic in women with newly diagnosed metastatic breast cancer Sci. Rep. 2020 10 4484 10.1038/s41598-020-61416-1 32161278 77. Bagegni N. Thomas S. Liu N. Luo J. Hoog J. Northfelt D.W. Goetz M.P. Forero A. Bergqvist M. Karen J. Serum thymidine kinase 1 activity as a pharmacodynamic marker of cyclin-dependent kinase 4/6 inhibition in patients with early-stage breast cancer receiving neoadjuvant palbociclib Breast Cancer Res. 2017 19 123 10.1186/s13058-017-0913-7 29162134 78. Schiavon G. Hrebien S. Garcia-Murillas I. Cutts R.J. Pearson A. Tarazona N. Fenwick K. Kozarewa I. Lopez-Knowles E. Ribas R. Analysis of ESR1 mutation in circulating tumor DNA demonstrates evolution during therapy for metastatic breast cancer Sci. Transl. Med. 2015 7 313ra182 10.1126/scitranslmed.aac7551 79. Razavi P. Chang M.T. Xu G. Bandlamudi C. Ross D.S. Vasan N. Cai Y. Bielski C.M. Donoghue M.T.A. Jonsson P. The Genomic Landscape of Endocrine-Resistant Advanced Breast Cancers Cancer Cell 2018 34 427 438.e6 10.1016/j.ccell.2018.08.008 30205045 80. Hortobagyi G.N. Stemmer S.M. Burris H.A. Yap Y.S. Sonke G.S. Paluch-Shimon S. Campone M. Petrakova K. Blackwell K.L. Winer E.P. Updated results from MONALEESA-2, a phase III trial of first-line ribociclib plus letrozole versus placebo plus letrozole in hormone receptor-positive, HER2-negative advanced breast cancer Ann. Oncol. 2018 29 1541 1547 10.1093/annonc/mdy155 29718092 81. Fribbens C. O'Leary B. Kilburn L. Hrebien S. Garcia-Murillas I. Beaney M. Cristofanilli M. Andre F. Loi S. Loibl S. Plasma ESR1 Mutations and the Treatment of Estrogen Receptor-Positive Advanced Breast Cancer J. Clin. Oncol. 2016 34 2961 2968 10.1200/JCO.2016.67.3061 27269946 82. Tolaney S.M. Toi M. Neven P. Sohn J. Grischke E.-M. Llombart-Cussac A. Soliman H. Wang H. Wijayawardana S. Jansen V.M. Clinical Significance of PIK3CA and ESR1 Mutations in Circulating Tumor DNA: Analysis from the MONARCH 2 Study of Abemaciclib plus Fulvestrant Clin. Cancer Res. 2022 28 1500 1506 10.1158/1078-0432.CCR-21-3276 35121623 83. Lloyd M.R. Spring L.M. Bardia A. Wander S.A. Mechanisms of Resistance to CDK4/6 Blockade in Advanced Hormone Receptor-positive, HER2-negative Breast Cancer and Emerging Therapeutic Opportunities Clin. Cancer Res. 2022 28 821 830 10.1158/1078-0432.CCR-21-2947 34725098 84. McCartney A. Migliaccio I. Bonechi M. Biagioni C. Romagnoli D. De Luca F. Galardi F. Risi E. De Santo I. Benelli M. Mechanisms of Resistance to CDK4/6 Inhibitors: Potential Implications and Biomarkers for Clinical Practice Front. Oncol. 2019 9 666 10.3389/fonc.2019.00666 31396487 85. Martin J.M. Handorf E.A. Montero A.J. Goldstein L.J. Systemic Therapies Following Progression on First-line CDK4/6-inhibitor Treatment: Analysis of Real-world Data Oncologist 2022 27 441 446 10.1093/oncolo/oyac075 35552450 |
PMC10000621 | Background: The impact of several non-clinical factors on cancer survival is poorly understood. The aim of this study was to investigate the influence of travel time to the nearest referral center on survival of patients with cancer. Patients and methods: The study used data from the French Network of Cancer Registries that combines all the French population-based cancer registries. For this study, we included the 10 most common solid invasive cancer sites in France between 1 January 2013 and 31 December 2015, representing 160,634 cases. Net survival was measured and estimated using flexible parametric survival models. Flexible excess mortality modelling was performed to investigate the association between travel time to the nearest referral center and patient survival. To allow the most flexible effects, restricted cubic splines were used to investigate the influence of travel times to the nearest cancer center on excess hazard ratio. Results: Among the 1-year and 5-year net survival results, lower survival was observed for patients residing farthest from the referral center for half of the included cancer types. The remoteness gap in survival was estimated to be up to 10% at 5 years for skin melanoma in men and 7% for lung cancer in women. The pattern of the effect of travel time was highly different according to tumor type, being either linear, reverse U-shape, non-significant, or better for more remote patients. For some sites restricted cubic splines of the effect of travel time on excess mortality were observed with a higher excess risk ratio as travel time increased. Conclusions: For numerous cancer sites, our results reveal geographical inequalities, with remote patients experiencing a worse prognosis, aside from the notable exception of prostate cancer. Future studies should evaluate the remoteness gap in more detail with more explanatory factors. cancer outcomes survival travel time geographical accessibility Ligue Nationale Contre le CancerInstitut National du CancerSante Publique FranceThis work was supported by the "Ligue Nationale Contre le Cancer" and the cancer registries are funded by the "Institut National du Cancer" and by "Sante Publique France". No grant number is applicable. pmc1. Introduction Factors related to the tumor and demographic characteristics are the main predictors of cancer prognosis. However, the influence of numerous non-clinical factors such as socioeconomic environment and health care organization, including hospital characteristics, on cancer survival has been gaining increasing interest. Health care accessibility is defined by the capacity to achieve the best medical outcome in a timely manner, being mainly driven by two components, social access and geography. The influence of social condition on access to health care has been widely studied. Patients with a low socioeconomic status have a lower probability of participating in screening and early detection, leading to a more advanced cancer stage at clinical presentation, as well as a lower probability of being treated in a high volume hospital, both resulting in a lower cancer survival compared to those with a higher socioeconomic status . These inequalities have been reported in all industrialized countries. In France, where patients can choose their preferred hospital whatever its location or level of specialization, studies have shown that numerous factors such as patient's age, sex and distance between the hospital and place of residence have an influence on this choice . Socio-economic factors can also intervene, even though in France integral reimbursement of all health-related expenditures is guaranteed from cancer diagnosis. However, leading up to diagnosis, the outstanding amount to be paid by the patient can be substantial if the patient relies only on universal insurance health care coverage. This financial burden may restrict access to certain physicians in the private sector, which in turn could delay diagnosis for patients with less financial resources. Moreover, distance to medical facilities can also present a financial burden for patients due to travel costs. To date, geographic inequalities have been little documented and studies have been limited to a few cancer sites. There are some published studies on the influence of travel time for patients on survival in Europe. The majority of these studies measured the theoretical travel time to the nearest hospital. Kravdal et al. in 2006 in their study on Norwegian data, showed no significant effect of travel time to hospital on survival. In England, Jones et al. also showed no significant difference, with only the journey to the GP having an influence on survival. For colorectal cancer, Sjostrom et al. in a Swedish study in 2020 found no evidence of an association between travel time and colorectal cancer survival. In contrast, Murage in 2016 showed that longer average travel times to cancer services were associated with poorer survival. In their study in 2014 , Dejardin et al. showed the same phenomenon on French data. Kelly et al. in 2016 reviewed 108 papers on studies of data from global north countries. In summary, 77% of the included studies found an association between poorer health status and distance from the health facilities the patient needs to attend. The authors conclude that a distance effect cannot be excluded, and distance and travel time should be taken into account when configuring the location of health facilities and treatment options for patients. Since the factors that determine where people live depend on both social and geographical criteria, social and geographic inequalities in health are closely linked. It therefore seems essential to integrate, in the analysis of geographic inequalities in cancer survival, information on the social environment of patients in order to identify the specific effect of travel time on survival of patients with cancer. Some of the above papers have used either old data, or data that are not population based, or focused on very few tumor types. In addition, only a few studies have considered net survival , i.e., net from other causes of death and only related to the cancer. In order to study the effect of geographical inequalities on cancer survival, use of data from population-based cancer registries guarantees the completeness of cases in the study areas, unlike hospital series or hospital discharge data. Prior to explaining the mechanisms of geographical inequalities, which requires much medical information specific to each cancer site, the illustration and the comparison of such geographical inequalities across cancer sites using the same methodology is a mandatory prerequisite. To our knowledge, no study has conducted a systematic analysis of the geographical inequalities in cancer survival adjusted on social environment for the 10 most frequent cancers. The aim of this study was to investigate the influence of travel time to the nearest referral center on net survival of patients with one of the ten most common cancers between 2013 and 2015 in France, using data from French cancer registries. 2. Patients and Methods 2.1. Study Design and Patient Population The study used data from the French Network of Cancer Registries (FRANCIM), which combines all the French population-based cancer registries, between 1 January 2013 and 31 December 2015; it included the 10 commonest solid invasive cancer types in France according to the 3rd edition of the International Classification of Diseases for Oncology (ICD-O-3); the included cancers were: breast (C50), colon-rectum and anal canal (C18-C21), lung (C33-C34), pancreas (C25), prostate (C61), skin melanoma (C44--morphology 87203 to 87803), bladder (C67), head and neck (C01-C06; C09-C14), kidney (C64), and liver (C22). Our study excluded hematological malignancies, male breast cancers (because of the limited sample size, N = 309) and pediatric cancers (aged under 15 years, covered by a national pediatric registry). In total, this represented 160,634 cases during the study period. The area covered by the registries included in our study represents approximatively 20% of the French population. The French departments covered (fully or partially) by a register and included in this study are the following: Calvados, Charente, Charente-Maritime, Cote-d'Or, Doubs, Finistere, Gironde, Herault, Isere, Loire-Atlantique, Manche, Nord, Haut-Rhin, Saone-et-Loire, Deux-Sevres, Somme, Tarn, Vendee, Vienne, Haute-Vienne and Territoire-de-Belfort. The completeness and data quality of the included registries are regularly assessed by the International Agency for Research on Cancer (IARC) and the European Network of Cancer Registries (ENCR). This study was approved by the Commission Nationale Informatique et Libertes (CNIL--ndeg921057). 2.1.1. Variables Data collected included sex, year of diagnosis, age, and site of primary tumor. Age at diagnosis was separated into groups: 15-54, 55-64, 65-74, and 75+ years. For all diagnosed cancers, patient addresses were geolocalized using Geographic Information Systems (ArcGIS; Esri, Redlands, CA, USA); the exact geographic position (X, Y coordinates) of the patient address and the residential IRIS (Ilots Regroupes pour l'Information Statistique) were available. 2.1.2. Travel Time For each patient, the travel time in minutes from residential address to the reference cancer care center was calculated. The reference cancer care center was defined by the closest University Hospital (Centre Hospitalier Universitaire) or Cancer Control Center (Centre de Lutte Contre le Cancer) in terms of travel time. Although not as specialised as the centers of reference, non-reference care centers (local and private hospitals) provide comprehensive care for a patient. These travel times were estimated using a road-network database (Multinet TeleAtlas). Travel speeds, computed in minutes, were estimated according to legal speeds for the different road classes. Travel times were separated into four classes: <30, 30-59, 60-89 and 90+ minutes for descriptive purposes, but used in a continuous form in a parametric survival model. 2.1.3. Social Deprivation Social deprivation was measured by the French version of the European Deprivation Index (EDI); the principles and methods for building this are detailed in previous papers . Briefly, it is an ecological index that measures relative poverty in small geographic areas, based on information from the European Union Survey of Income and Living Conditions (EU-SILC) and census information. The IRIS is the smallest geographic unit in France defined by the "Institut National de la Statistique et des Etudes Economiques" (INSEE); the French version of the EDI is associated with the residential IRIS available. Due to the suspected association between travel time and geographical distribution of socioeconomic deprivation, this index was systematically integrated to the analyses. 2.1.4. Outcomes Survival time was defined as the difference between date of diagnosis and date of last contact for vital status. Follow-up ended on 30 June 2018, i.e., patients alive on that date had their survival time censored. Lost to follow-up accounted for about 3.5% of cases overall. The information on vital status was collected through an active standardized search procedure by the French Network of Cancer Registries, based on requests to the "Repertoire National d'Identification des Personnes Physiques" [RNIPP] and, if necessary, other sources of information (including medical records or birthplace public services). Survival time was defined as the difference between the date of last information and the date of cancer diagnosis. When the date of death was the same as the date of cancer diagnosis, those cases were included in the analyses with a survival time equal to 0.5 days, following standard procedure. 2.1.5. Statistical Analysis All analyses were conducted in the framework of net survival. This method assumed that patients would only die of their cancer. Thus, as shown in equation 1, the observed mortality for one patient is the sum of the expected mortality and of the mortality in excess (i.e., only related to the cancer):(1) mobserved=mexpected+mexcess Expected mortality could be known either by the registration of cause of death or by the use of a life table to estimate the expected mortality for a comparable patient in terms of age, sex, year of diagnosis, and department. We calculated excess mortality using life tables. In our study, a flexible parametric survival model was used. Net survival is based on the estimated risk of excess mortality. Net survival probabilities and 95% confidence intervals (CIs) by class of travel time to the nearest referral center were estimated. This method estimates cancer-specific survival and, from these calculated net survivals, the remoteness gap (RGap) was calculated. This corresponds to the difference in net survival between patients living closer to referral centers (between 0 and 30 min) and those living further away (more than 90 min). The 95% CI of the remoteness gap was derived using the variance of the net survival of the patients living closer to referral centers and the variance of the patients living further away. We considered that the difference in net survival between the two groups was statistically significant when the 95% CI of the corresponding remoteness gap did not include zero. Finally, the percentage of variation of net survival was calculated between the two groups. To allow the most flexible effects, restricted cubic splines were used to investigate the influence of travel times to the nearest cancer center on excess hazard ratio. These restricted cubic splines were adjusted for year of diagnosis, age, and EDI. Concerning the model selection, the inclusion of age in a survival model is obvious. As all survival analysis produced by FRANCIM were stratified by sex, the same strategy was adopted here . The inclusion of EDI is trickier. Previous publications using FRANCIM data have clearly established the association between deprivation and survival for the vast majority of cancer sites. To study the proper effect of travel times, deprivation index was systematically included in all analysis. The choice of the number of nodes for the splines was made using the minimization of Akaike information criteria (AIC). The use of restricted cubic splines allowed us to obtain detailed information on the pattern of the effects of travel time. However, before the first node and after the last node, the splines are linear by definition. The position of each node is represented on the graphs. The interpretation of the curves after the last node should be made with caution. Finally, a non-proportional effect of travel time was systematically tested. A significant non-proportional effect denotes that the association between travel time and survival was not constant throughout the natural course of the disease. Cases with missing values for travel time were excluded from these analyses, corresponding to complete cases analysis. This represented 0.82% of the data. Our analyses were all stratified by sex and cancer site. All statistical analyses were performed using the stmp2, mkpline and xblc packages in STATA SE 16 software (StataCorp LLC, College Station, TX, USA). 3. Results The study included a total of 160,634 cases. Characteristics of the population are presented in Table 1. More than half of the cases in this study were men (57.12%). The median age for men was 68.67 years while for women it was 66.57 years. Table 2 and Table 3 provide net survival probabilities and 95% confidence intervals in each class of travel time and the remoteness gap (RGap) for each cancer site in men and women, respectively. 5-year net survival were higher in patients living closer to referral centers than patients living further away for all cancer sites except head and neck cancer (statistically significant) and pancreas cancer in men, and except head and neck cancer in women. In men, the remoteness gap in relation to 5-year net survival was in favor of better survival for those living closest to referral centers and statistically significant for lung cancer (RGap1-year = 3.75 [0.82;6.69]; RGap5-year = 3.26 [0.85;5.66]) and skin melanoma (RGap1-year = 3.54 [0.69;6.38]; RGap5-year = 9.82 [2.66;16.99]). In women, the remoteness gap regarding 1-year net survival was statistically significant for lung cancer (RGap1-year = 7.31 [2.80;11.82]) and pancreas cancer (RGap1-year = 7.97 [2.32;13.62]). At 5 years, the remoteness gap was statistically significant for lung cancer (RGap5-year = 7.27 [3.27;11.27]), pancreas cancer (Rgap5-year = 4.47 [1.78;7.15]), and skin melanoma (RGap5-year = 6.50 [0.14;12.87]). The tables also show the variation in 5-year net survival between patients living closer to referral centers and patients living further away for each cancer site in men and women. The largest decline in 5-year net survival was observed for lung cancer in men (D = 7.83% and 16.45%, respectively) and for pancreas cancer in women (D = 20.54% and 41.97%, respectively). The restricted cubic splines of the effect of travel time on excess mortality ratio highlighted 4 patterns: 3.1. Linear Pattern Figure 1 shows the restricted cubic splines of the effect of travel time on excess mortality for cancer sites where the excess hazard ratio was higher with increasing travel time. This trend was observed for lung cancer and skin melanoma in both sexes, and for breast cancer in women. For lung cancer, the excess hazard ratio reached a maximum of 1.3 in men and 1.35 in women. For skin melanoma, the excess hazard ratio reached a maximum of 2.2 in men and 2.1 in women. For breast cancer in women, the excess hazard ratio reached a maximum of 1.3. 3.2. Reverse U-Shape Pattern Figure 2 shows the restricted cubic splines with reverse u-shape trends. These trends were observed for liver cancer in both sexes, for colon-rectum and anal canal cancer in men, and for head and neck cancer in women. These trends meant that the excess mortality rate increased and then decreased. For liver cancer in men, the excess hazard ratio reached a maximum of 1.3 at 50 min of travel time, and 1.45 at 50 min in women, and then decreased. For colon-rectum and anal canal cancer in men, the excess hazard ratio reached a maximum of 1.2 at 50 min of travel time. For head and neck cancer in women, the excess hazard ratio reached a maximum of 1.4 at 50 min of travel time. 3.3. No Association Figure 3 shows the restricted cubic splines for cancer types with no significant results: pancreas cancer, kidney cancer, and bladder cancer in both sexes; head and neck cancer in men; and colon-rectum and anal canal cancer in women. 3.4. Better Prognosis from Remote Patients Figure 4 shows the restricted cubic splines for prostate cancer, which were associated with a better prognosis for remote patients (excess hazard ratio was lower for patients living further from referral centers). For all patterns of excess mortality, the corresponding net survival curves are available in supplementary material . 4. Discussion Through flexible modelling of the effect of travel time to care center on excess mortality, our study shows that half of cancer types studied were subject to geographical inequalities in survival (4/9 cancer sites in men and 5/9 cancer sites in women demonstrating linear or reverse U-shape patterns, and 1 type with a better prognosis in patients living in the most remote areas). The remoteness gap in net survival for the linear pattern was estimated to be almost 10% at 5 years for skin melanoma in men and 7% for lung cancer in women. This geographical gap in survival is directly linked to cancer-specific mortality and not the consequence of other causes of death. This remoteness gap constitutes a real loss of opportunity for these remote patients and should be considered in future studies and clinical practice. An interesting finding was that the pattern of the effect of travel time is highly different by cancer type, being either linear, reverse U-shape, non-significant, or better for more remote patients. The explanation of such patterns cannot be addressed in a low-resolution study like this. On one hand, the centralization of care for some tumor types may play a role in the linear pattern, and the opposite may be true for the reverse U-shape patterns. The mechanism of such geographical inequalities should be studied in more detailed population-based studies according to tumor types. A potential explication for the worse survival for patients living further out could be that those patients have a greater delay between symptom onset and cancer diagnosis. The influence of diagnostic delay has been highlighted in numerous publications . Unfortunately, date of first symptom is very hard to collect even in dedicated surveys and, thus, is not collected in cancer registries. Moreover, previous publications in France or elsewhere , have provided evidence that patients living further from referral centers are mainly treated in local hospitals. This preference for proximity could also explain in part our results. Using comparable data (all cancer registries in France) and comparable statistical modelling (flexible net survival models), Tron et al. published a systematic study on the association between socioeconomic environment and excess mortality. Deprivation gaps were calculated between the most deprived and the least deprived. When compared to our remoteness gap, we found similarities for pancreas cancer at 1 year in women (Deprivation Gap = 7.2 [1.6;12.8] vs Remoteness Gap = 7.97 [2.32;13.62]) and for lung cancer at 5 years in men (Deprivation Gap = 2.9 [0.7;5.0] vs Remoteness Gap = 3.26 [0.85;5.66]). Our study reveals that the magnitude and direction of both effects would be similar for some cancer types. Kelly et al. highlighted 3 different patterns for the effect of travel time on survival. Firstly, "distance decay association", which is an association between patients living closer to a health facility and having better health outcomes/access to health services, compared to those living farther away. Thanks to our flexible modelling of travel time, we can highlight the heterogeneity according to cancer types inside the "distance decay association" group. Secondly, "distance bias association" which, in contrast to the first pattern, is an association between patients living farther away from the health care facility and having better health outcomes/access rates to health services compared to those living closer. Prostate cancer is the only cancer type in which this pattern was detected in our study. The third pattern found in this review is the absence of association between travel time and survival. We confirmed this absence of association for numerous cancer sites using a large study population and flexible statistical models. An early study in Scotland showed strong evidence that increasing distance from a cancer center is associated with worse survival. A subsequent study in England showed that longer average travel times are associated with worse survival . Illustrating a distance bias association, another study in Scotland from the same group showed that patients living in a rural area and travelling farther to a GP have a lower likelihood of emergency admissions and a better survival . Other studies have examined the effect of living in rural or urban areas on survival, which is another way to attempt to understand geographical inequalities. A study in Denmark in 2018 showed that there was a better survival rate for pancreatic cancer patients for those living in urban areas compared to those living in rural areas. Two systematic literature reviews in 2018 and 2019 have also shown this association for cancer patients. Carriere et al. , in their international review and meta-analysis, report that there is strong evidence of an association between rural residence and poor cancer survival outcomes. Afshar et al. conducted a systematic review of studies from Australia, the USA, Europe, Canada and New Zealand. In this review they showed that cancer patients living in rural and remote areas had poorer survival than those living in urban areas. Our study has a number of strengths. First, this study is based on data from French cancer registries, which guarantee high quality data and the exhaustiveness of all cases in the studied area, and thus a high statistical power. For survival analyses, in our study we used excess base rate modelling from Royston models , unlike most other studies that use the Cox model, and we used the maximum likelihood facilities to fit our models without the unrealistic clinical hypotheses required by the Cox model. Moreover, thanks to the use of flexible modelling of the effect of distance, our study shows that the pattern of the effect of travel time is notably different according to cancer type. Four patterns of association were identified in our study: linear pattern, reverse U-Shape pattern, no association, and better prognosis for more isolated patients. To our knowledge, only one study limited to colorectal cancer used a comparable methodology (net survival and spline modelling for travel times). Dejardin et al. showed that the effect of travel-time to the nearest reference cancer center for colorectal cancer patients was a reverse U-shape in France and not significant in England . This previous study highlights the impact of individuality of health care systems, with no association found for England, potentially due to GP gate keeping, and with an association found in France, potentially due to free hospital choice. Our study confirms this influence of geographical accessibility on colorectal cancer on a bigger population, with a similar reverse U-shape found for colorectal cancer. Our study has some limitations. First, there was a lack of data about stage at diagnosis and the actual care facility attended by the patients. At a national scale, detailed data on stage and hospital facility attended were not available in an exhaustive way in all French cancer registries. However, the aim of this study was to describe the effect of travel-time to the nearest cancer center and survival using the most extensive data available in France. It is therefore necessary to first establish the dynamics of cancer survival according to travel time to cancer centres, then to add potential mediating factors to explain these dynamics. Cancer stage at diagnosis appears essential to help explain the associations found. Information on the actual center attended by patients would also be important to check that there is no underestimation of travel times for some patients who may travel further than to the nearest center. Stage and place of treatment, which are important to account for, will be analyzed in a future study to explain the different patterns of remoteness gap in survival highlighted in our study. Another limitation was the use of travel time to the nearest cancer center. As aforementioned, place of treatment was not known in this study. Most European studies calculated travel time from the patient's home (home postal code or specific address) to the nearest cancer referral center or hospital . However, as mentioned by Kelly et al. in their review, the use of the nearest referral center is only a proxy of geographic isolation. Due to the variety of public transport available depending on the area, alongside the modalities such as timetables, etc., the integration of these parameters in a GIS would have been extremely complicated. Thus, we made the assumption that road car travel-time is a good proxy for accessibility . Recently indexes of isolation such as APL or Scale have been developed to capture the heterogeneity of spatial accessibility. Nonetheless, such indexes mainly refer to spatial access to primary health care rather than access to specialized cancer centers. Depending on health care system organization, previous publications have shown that both components of spatial accessibility could be associated with cancer prognosis. In the UK, access to primary health care seems more critical rather than access to reference cancer centers in contrast to France . Another limitation of this study is the representativeness of the data. We have a large population but this represents only a little more than 20% of the French population, and there are no large French cities in the areas covered by cancer registries. Finally, concerning statistical analysis, we estimated survival using additive net survival models and general population life tables. Under the hypothesis of a geographical gradient in background mortality, this implies that we might have overestimated the gradient because geographical life tables do not exist. This problem also arises with studies on social inequalities that use the EDI. However, a previous publication highlights that, even if causes of death are available, the use of a life table is recommended . 5. Conclusions Our results suggest that, for a non-negligible number of cancer types, the travel time by road between a patient's home and the nearest referral center has an effect on net survival. This geographical gap in survival is directly linked to cancer-specific mortality and not the consequence of other causes of death, and the pattern of the effect of remoteness on cancer survival varies among cancer sites. Geographical inequalities have been previously underestimated relative to social inequalities. However, our study shows that the magnitude of both effects could be comparable. From a clinical and public health perspective, French authorities decided in 2016 to encourage reference hospitals to share medical expertise with local hospitals to provide better medical support for remote patients (Groupe Hospitalier Territorial). Further studies in the coming years should evaluate the effectiveness of this strategy. Acknowledgments The authors thank the staff of each member registry of the FRANCIM network who participated in the collection of data. We also thank all of the contributors to the registry data, that is, general practitioners, pathologists, and specialists belonging to private and public health systems. Supplementary Materials The following supporting information can be downloaded at: Figure S1: Survival probability--cases with linear pattern; Figure S2: Survival probability--cases with reverse U-shape pattern; Figure S3: Survival probability--cases with no association; Figure S4: Survival probability--cases with better prognosis for remote patients. Click here for additional data file. Author Contributions J.G., S.W., J.B., O.D. designed the study, obtained ethics approval and wrote the manuscript. J.G., S.W., J.B., O.D. designed and performed analysis. J.G., S.W., A.-V.G., V.B., L.T., L.L., A.A., Francim Group, G.L., F.M., J.B., O.D. reviewed and approved the manuscript. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement This study is based on data from cancer registries gathered in the French network of cancer registries (FRANCIM). Each cancer registry involved in this study has received the approval of the French regulatory authorities for the collection and analysis of medical data: the Comite Consultatif sur le Traitement de l'Information en matiere de Recherche dans le Domaine de la Sante (ethics approval) and the Commission Nationale Informatique et Libertes (legal framework and data protection). In conformity with the French law, patients were individually informed before the start of data collection of the nature of the information provided, the purpose of data processing, and their right of access, rectification or objection. The ethics committee, in accordance with French law, did not request informed consent. However, due to geolocalization used for this work, this study received an additional approval from French Data Protection Authority (CNIL, authorization ndeg921057). Informed Consent Statement No applicable. Data Availability Statement The data that supports the findings of our study are available from the corresponding author upon reasonable request. Conflicts of Interest The authors have declared no conflict of interest. Appendix A Members of the French Network of Cancer Registries (FRANCIM) included in the study: Arnaud ALVES (Registre des cancers digestifs du Calvados), Simona BARA (Registre des cancers de la Manche), Anne-Marie BOUVIER (Registre Bourguignon des cancers digestifs), Marc COLONNA (Registre general des cancers de l'Isere), Gaelle COUREAU (Registre general des cancers de la Gironde), Sandrine DABAKUYO YONLI (Registre des cancers du sein et des cancers gynecologiques de Cote d'Or), Tania D'ALMEIDA (Registre general des cancers en region Limousin), Gautier DEFOSSEZ (Registre des cancers de Poitou-Charentes), Pascale GROSCLAUDE (Registre des cancers generaux du Tarn), Anne-Valerie GUIZARD (Registre general des tumeurs du Calvados), Karima HAMMAS (Registre des cancers du Haut-Rhin), Benedicte LAPOTRE-LEDOUX (Registre general des cancers de la Somme), Florence MOLINIE (Registre des tumeurs de Loire-Atlantique/Vendee), Jean-Baptiste NOUSBAUM (Registre Finisterien des tumeurs digestives), Sandrine PLOUVIER (Registre des cancers de Lille et de sa region), Brigitte TRETARRE (Registre general des tumeurs de l'Herault), Anne-Sophie WORONOFF (Registre des tumeurs du Doubs et du Territoire de Belfort). Figure 1 Excess mortality rates as a function of travel time, using restricted cubic splines--cases with linear pattern. EHR: excess hazard ratio; CI: confidence Interval; travel time: travel time in minutes. Adjusted for year of diagnosis, age, and European Deprivation Index. Figure 2 Excess mortality rates as a function of travel time, using restricted cubic splines--cases with reverse U-shape pattern. EHR: excess hazard ratio; CI: confidence Interval; travel time: travel time in minutes Adjusted for year of diagnosis, age, and European Deprivation Index. Figure 3 Excess mortality rates as a function of travel time, using restricted cubic splines--cases with no association. EHR: excess hazard ratio; CI: confidence Interval; travel time: travel time in minutes. Adjusted for year of diagnosis, age, and European Deprivation Index. Figure 4 Excess mortality rates as a function of travel time, using restricted cubic splines--cases with better prognosis for remote patients. EHR: excess hazard ratio; CI: confidence Interval; travel time: travel time in minutes. Adjusted for year of diagnosis, age, and European Deprivation Index. cancers-15-01516-t001_Table 1 Table 1 Characteristics of population. Travel Time in Min [0;30] [30;60] [60;90] [90;++] Missing Total n = 63,636 n = 55,850 n = 32,586 n = 7248 n = 1314 n = 160,634 Sex Male Female Male Female Male Female Male Female Male Female Male Female n = 34,741 % n = 28,895 % n = 32,866 % n = 22,984 % n = 18,984 % n = 13,602 % n = 4356 % n = 2892 % n = 810 % n = 504 % n = 91,757 % n = 68,877 % Median age (year) 68 66 69 67 69 68 70 68 71 71 69 67 Age in classes [15;55] 3942 11.35 7516 26.01 3218 9.79 5415 23.56 1662 8.75 2992 22.00 345 7.92 536 18.53 75 9.26 102 20.24 9242 10.07 16,561 24.04 [55;65] 9251 26.63 6448 22.32 8415 25.60 4894 21.29 4706 24.79 2883 21.20 1002 23.00 656 22.68 168 20.74 91 18.06 23,542 25.66 14,972 21.74 [65;75] 11,236 32.34 6464 22.37 11,152 33.93 5235 22.78 6488 34.18 3130 23.01 1553 35.65 703 24.31 243 30.00 94 18.65 30,672 33.43 15,626 22.69 [75;++] 10,312 29.68 8467 29.30 10,081 30.67 7440 32.37 6128 32.28 4597 33.80 1456 33.43 997 34.47 324 40.00 217 43.06 28,301 30.84 21,718 31.53 EDI's quintile Q1 9597 27.62 7788 26.95 4287 13.04 2917 12.69 2327 12.26 1634 12.01 345 7.92 215 7.43 1 0.12 2 0.40 16,557 18.04 12,556 18.23 Q2 6387 18.38 5235 18.12 6672 20.30 4480 19.49 4344 22.88 2916 21.44 969 22.25 598 20.68 1 0.12 0 0.00 18,373 20.02 13,229 19.21 Q3 5051 14.54 4443 15.38 7606 23.14 5229 22.75 4588 24.17 3253 23.92 1324 30.39 932 32.23 2 0.25 1 0.20 18,571 20.24 13,858 20.12 Q4 5410 15.57 4708 16.29 7620 23.19 5318 23.14 4515 23.78 3315 24.37 1202 27.59 762 26.35 1 0.12 0 0.00 18,748 20.43 14,103 20.48 Q5 7379 21.24 6074 21.02 6137 18.67 4637 20.17 3167 16.68 2466 18.13 515 11.82 385 13.31 0 0.00 1 0.20 17,198 18.74 13,563 19.69 Missing 917 2.64 647 2.24 544 1.66 403 1.75 43 0.23 18 0.13 1 0.02 0 0.00 805 99.38 500 99.21 2310 2.52 1568 2.28 Year of diagnosis 2013 11,306 32.54 9451 32.71 10,764 32.75 7456 32.44 6511 34.30 4583 33.69 1381 31.70 972 33.61 297 36.67 180 35.71 30,259 32.98 22,642 32.87 2014 11,445 32.94 9569 33.12 11,025 33.55 7978 34.71 6493 34.20 4748 34.91 1487 34.14 953 32.95 233 28.77 142 28.17 30,683 33.44 23,390 33.96 2015 11,990 34.51 9875 34.18 11,077 33.70 7550 32.85 5980 31.50 4271 31.40 1488 34.16 967 33.44 280 34.57 182 36.11 30,815 33.58 22,845 33.17 Solid tumor sites 1 Bladder 2218 6.38 539 1.87 2177 6.62 468 2.04 1208 6.36 265 1.95 283 6.50 66 2.28 50 6.17 16 3.17 5936 6.47 1354 1.97 Breast 14,864 51.44 10,973 47.74 6311 46.40 1367 47.27 138 27.38 33,653 48.86 Colon-rectum 5711 16.44 5247 18.16 5715 17.39 4769 20.75 3617 19.05 2929 21.53 713 16.37 543 18.78 80 9.88 97 19.25 15,836 17.26 13,585 19.72 Head and neck 2202 6.34 683 2.36 1963 5.97 542 2.36 1012 5.33 306 2.25 225 5.17 69 2.39 31 3.83 16 3.17 5433 5.92 1616 2.35 Kidney 1758 5.06 855 2.96 1580 4.81 748 3.25 905 4.77 454 3.34 201 4.61 116 4.01 61 7.53 34 6.75 4505 4.91 2207 3.20 Liver 2088 6.01 546 1.89 1933 5.88 446 1.94 1143 6.02 296 2.18 245 5.62 53 1.83 58 7.16 18 3.57 5467 5.96 1359 1.97 Lung 6571 18.91 2965 10.26 6108 18.58 2360 10.27 3567 18.79 1424 10.47 847 19.44 334 11.55 134 16.54 56 11.11 17,227 18.77 7139 10.36 Pancreas 1585 4.56 1546 5.35 1418 4.31 1345 5.85 925 4.87 891 6.55 185 4.25 175 6.05 42 5.19 41 8.13 4155 4.53 3998 5.80 Prostate 11,004 31.67 10,701 32.56 5889 31.02 1464 33.61 277 34.20 29,335 31.97 Skin melanoma 1604 4.62 1650 5.71 1271 3.87 1333 5.80 718 3.78 726 5.34 193 4.43 169 5.84 77 9.51 88 17.46 3863 4.21 3966 5.76 1 Except hematological malignancies. cancers-15-01516-t002_Table 2 Table 2 One and five-year net survival probabilities and 95% confidence intervals and remoteness gap (RGap), by cancer site, in men. 1-Year Net Survival [0;30] [30;60] [60;90] [90;++] Remoteness Gap [95% CI] Percentage of Variation at 1 Year Solid Tumor Sites 1 Bladder (C67; all morphology) 77.12 (75.56-78.70) 76.66 (75.09-78.26) 75.94 (73.90-78.04) 74.54 (70.51-78.79) 2.58 (-1.73;6.89) -3.35 Colon-rectum (C18-C21; all morphology) 83.79 (82.92-84.66) 82.01 (81.11-82.93) 81.59 (80.49-82.71) 82.55 (80.28-84.87) 1.24 (-1.18;3.67) -1.48 Head and neck (C01-C06,C09-C14; all morphology) 72.24 (70.70-73.81) 71.76 (70.15-73.41) 70.78 (68.61-73.02) 76.82 (72.82-81.04) -4.58 (-8.86;-0.29) 6.34 Kidney (C64; all morphology) 88.19 (86.79-89.61) 85.50 (83.90-87.12) 83.97 (81.85-86.15) 85.85 (81.74-90.16) 2.34 (-1.99;6.67) -2.65 Liver (C22; all morphology) 50.88 (49.00-52.82) 46.72 (44.79-48.74) 44.73 (42.30-47.30) 49.15 (44.14-54.73) 1.72 (-3.63;7.07) -3.40 Lung (C33-C34; all morphology) 47.91 (46.86-48.99) 45.76 (44.68-46.87) 43.84 (42.46-45.26) 44.16 (41.42-47.08) 3.75 (0.82;6.69) -7.83 Pancreas (C25; all morphology) 37.59 (35.50-39.81) 35.62 (33.46-37.92) 35.83 (33.21-38.66) 40.44 (34.95-46.79) -2.85 (-8.72;3.03) 7.58 Prostate (C61; all morphology) 98.11 (97.83-98.39) 97.67 (97.35-97.99) 98.20 (97.85-98.55) 98.07 (97.39-98.75) 0.04 (-0.69;0.77) -0.04 Skin melanoma (C44; 87203 to 87803) 96.81 (96.00-97.62) 95.31 (94.24-96.38) 95.09 (93.73-96.46) 93.27 (90.54-96.08) 3.54 (0.69;6.38) -3.66 5-year Net Survival [0;30 ] [30;60 ] [60;90 ] [90;++ ] Remoteness Gap [95% CI] Percentage of Variation at 5 years Solid tumor sites 1 Bladder (C67; all morphology) 51.82 (49.31-54.46) 51.04 (48.54-53.67) 49.84 (46.59-53.32) 47.55 (41.34-54.68) 4.27 (-2.42;10.97) -8.24 Colon-rectum (C18-C21; all morphology) 62.69 (61.12-64.31) 59.25 (57.66-60.88) 58.45 (56.49-60.47) 60.27 (56.05-64.80) 2.42 (-2.07;6.92) -3.86 Head and neck (C01-C06,C09-C14; all morphology) 39.69 (37.45-42.06) 38.94 (36.62-41.41) 37.45 (34.34-40.83) 47.26 (40.65-54.94) -7.57 (-14.55;-0.59) 19.07 Kidney (C64; all morphology) 76.77 (74.33-79.29) 71.93 (69.22-74.75) 69.26 (65.68-73.03) 72.55 (65.47-80.39) 4.22 (-3.26;11.7) -5.50 Liver (C22; all morphology) 20.32 (18.56-22.25) 16.62 (14.99-18.44) 15.00 (13.10-17.19) 18.74 (14.53-24.16) 1.59 (-2.97;6.15) -7.78 Lung (C33-C34; all morphology) 19.82 (18.82-20.86) 17.91 (16.94-18.93) 16.29 (15.14-17.53) 16.56 (14.37-19.09) 3.26 (0.85;5.66) -16.45 Pancreas (C25; all morphology) 10.33 (8.92-11.95) 9.11 (7.76-10.69) 9.24 (7.66-11.13) 12.23 (8.68-17.23) -1.91 (-5.72;1.91) 18.39 Prostate (C61; all morphology) 93.62 (92.80-94.44) 92.18 (91.31-93.06) 93.92 (92.82-95.02) 93.49 (91.32-95.71) 0.13 (-2.19;2.45) -0.14 Skin melanoma (C44; 87203 to 87803) 90.46 (88.40-92.58) 86.21 (83.59-88.90) 85.59 (82.11-89.22) 80.64 (73.77-88.14) 9.82 (2.66;16.99) -10.86 1 Except hematological malignancies. cancers-15-01516-t003_Table 3 Table 3 One and five-year net survival probabilities and 95% confidence intervals and remoteness gap (RGap), by cancer site, in women. 1-Year Net Survival [0;30] [30;60] [60;90] [90;++] Remoteness Gap [95% CI] Percentage of Variation at 1 Year Solid tumor sites 1 Bladder (C67; all morphology) 65.36 (61.71-69.23) 61.31 (57.38-65.52) 63.74 (58.70-69.22) 60.83 (51.39-71.99) 4.53 (-5.58;14.65) -6.93 Breast (C50; all morphology) 97.81 (97.60-98.02) 97.34 (97.08-97.60) 97.29 (96.98-97.60) 97.25 (96.65-97.86) 0.56 (-0.08;1.19) -0.57 Colon-rectum (C18-C21; all morphology) 82.99 (82.07-83.93) 81.32 (80.32-82.33) 81.06 (79.83-82.30) 81.15 (78.53-83.86) 1.85 (-0.93;4.63) -2.22 Head and neck (C01-C06,C09-C14; all morphology) 79.20 (76.69-81.79) 75.89 (72.97-78.93) 75.51 (71.80-79.41) 81.20 (74.56-88.42) -2.00 (-9.09;5.10) 2.53 Kidney (C64; all morphology) 86.94 (84.86-89.08) 85.12 (82.80-87.50) 85.08 (82.14-88.13) 83.41 (77.58-89.68) 3.53 (-2.66;9.72) -4.06 Liver (C22; all morphology) 47.11 (43.49-51.03) 41.55 (37.70-45.79) 38.42 (33.84-43.62) 45.42 (35.23-58.57) 1.68 (-9.14;12.50) -3.59 Lung (C33-C34; all morphology) 55.60 (54.03-57.21) 54.69 (52.97-56.46) 54.42 (52.27-56.67) 48.29 (44.06-52.93) 7.31 (2.80;11.82) -13.15 Pancreas (C25; all morphology) 38.80 (36.66-41.05) 34.66 (32.45-37.01) 34.52 (31.89-37.38) 30.83 (25.59-37.13) 7.97 (2.32;13.62) -20.54 Skin melanoma (C44; 87203 to 87803) 96.12 (95.29-96.96) 95.02 (93.96-96.09) 94.35 (92.92-95.81) 93.03 (89.92-96.24) 3.10 (-0.12;6.31) -3.21 5-year Net Survival [0;30 ] [30;60 ] [60;90 ] [90;++ ] Remoteness Gap [95% CI] Percentage of Variation at 5 years Solid tumor sites 1 Bladder (C67; all morphology) 44.30 (39.70-49.43) 39.20 (34.45-44.59) 42.22 (36.05-49.46) 38.60 (27.90-53.40) 5.70 (-5.95;17.34) -12.87 Breast (C50; all morphology) 90.46 (89.79-91.13) 88.50 (87.68-89.33) 88.29 (87.21-89.38) 88.14 (85.81-90.55) 2.31 (-0.12;4.74) -2.56 Colon-rectum (C18-C21; all morphology) 62.27 (60.67-63.92) 59.13 (57.43-60.87) 58.64 (56.53-60.84) 58.81 (54.16-63.86) 3.46 (-1.46;8.38) -5.56 Head and neck (C01-C06,C09-C14; all morphology) 54.72 (50.65-59.12) 49.01 (44.55-53.91) 48.37 (42.66-54.84) 58.36 (46.94-72.57) -3.64 (-15.77;8.49) 6.65 Kidney (C64; all morphology) 76.02 (72.61-79.59) 72.93 (69.21-76.86) 72.87 (68.14-77.93) 70.09 (60.90-80.68) 5.93 (-3.88;15.74) -7.80 Liver (C22; all morphology) 18.04 (14.93-21.79) 13.55 (10.75-17.09) 11.34 (8.41-15.29) 16.60 (9.27-29.73) 1.43 (-6.53;9.40) -7.98 Lung (C33-C34; all morphology) 26.88 (25.24-28.64) 25.91 (24.13-27.82) 25.63 (23.42-28.05) 19.61 (15.97-24.09) 7.27 (3.27;11.27) -27.05 Pancreas (C25; all morphology) 10.65 (9.18-12.35) 8.15 (6.85-9.70) 8.08 (6.59-9.91) 6.18 (3.93-9.72) 4.47 (1.78;7.15) -41.97 Skin melanoma (C44; 87203 to 87803) 91.48 (89.80-93.19) 89.13 (87.02-91.30) 87.73 (84.86-90.69) 84.97 (78.83-91.60) 6.50 (0.14;12.87) -7.12 1 Except hematological malignancies. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000622 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050999 diagnostics-13-00999 Article The Prevalence and Risk Factors of Chronic Heart Failure in the Mongolian Population Sukhbaatar Pagmadulam Methodology Investigation Data curation Writing - original draft 1 Bayartsogt Batzorig Methodology Formal analysis Data curation Writing - original draft 2 Ulziisaikhan Ganchimeg Investigation Data curation Writing - original draft 3 Byambatsogt Bolortuul Writing - original draft Visualization 1 Khorloo Chingerel Investigation Data curation Writing - original draft 1 Badrakh Burmaa Investigation Data curation Writing - original draft 1 Tserendavaa Sumiya Investigation Data curation Writing - original draft 1 Sodovsuren Naranchimeg Writing - original draft Writing - review & editing Supervision 4 Dagva Mungunchimeg Resources Writing - original draft Project administration 3 Khurelbaatar Mungun-Ulzii Resources Writing - original draft Project administration 5 Tsedensodnom Sodchimeg Data curation Writing - original draft 6 Nyamsuren Bat-Erdene Investigation Data curation Writing - original draft 1 Myagmardorj Rinchyenkhand Data curation Writing - original draft Writing - review & editing 7 Unurjargal Tsolmon Conceptualization Methodology Writing - original draft Writing - review & editing Supervision 1* Yakushin Sergey S. Academic Editor 1 Department of Cardiology, School of Medicine, Mongolian National University of Medical Sciences, Ulaanbaatar 14210, Mongolia 2 Department of Epidemiology and Biostatistics, School of Public Health, Mongolian National University of Medical Sciences, Ulaanbaatar 14210, Mongolia 3 National Cardiovascular Center of Mongolia, The Third State Central Hospital, Ulaanbaatar 16081, Mongolia 4 Department of Communication Skill, Bio-Medical School, Mongolian National University of Medical Sciences, Ulaanbaatar 14210, Mongolia 5 Cardiac Rhythmology Center of the Third State Central Hospital Mongolia, Ulaanbaatar 16081, Mongolia 6 Nobel Pharmaceutical Company, Ulaanbaatar 16100, Mongolia 7 Cardiovascular Department, University Hospital of Mongolian National University of Medical Sciences, Ulaanbaatar 13270, Mongolia * Correspondence: [email protected] 06 3 2023 3 2023 13 5 99921 12 2022 23 2 2023 02 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Background: The prevalence of heart failure in the Mongolian population is unknown. Thus, in this study, we aimed to define the prevalence of heart failure in the Mongolian population and to identify significant risk factors for heart failure among Mongolian adults. Methods: This population-based study included individuals 20 years and older from seven provinces as well as six districts of the capital city of Mongolia. The prevalence of heart failure was based on the European Society of Cardiology diagnostic criteria. Results: In total, 3480 participants were enrolled, of which 1345 (38.6%) participants were males, and the median age was 41.0 years (IQR 30-54 years). The overall prevalence of heart failure was 4.94%. Patients with heart failure had significantly higher body mass index, heart rate, oxygen saturation, respiratory rate, and systolic/diastolic blood pressure than patients without heart failure. In the logistic regression analysis, hypertension (OR 4.855, 95% CI 3.127-7.538), previous myocardial infarction (OR 5.117, 95% CI 3.040-9.350), and valvular heart disease (OR 3.872, 95% CI 2.112-7.099) were significantly correlated with heart failure. Conclusions: This is the first report on the prevalence of heart failure in the Mongolian population. Among the cardiovascular diseases, hypertension, old myocardial infarction, and valvular heart disease were identified as the three foremost risk factors in the development of heart failure. heart failure prevalence risk factor diagnostic criteria This research received no external funding. pmc1. Introduction Heart failure (HF) is one of the major public health concerns with an increasing incidence over the years and it remains to be one of the leading causes of mortality among cardiovascular (CV) diseases . As of 2020, 64.3 million people were suffering from chronic HF worldwide . In recent years, chronic HF has become more prevalent because of an aging population, increased cardiovascular risk factors caused by economic trends promoting unhealthy lifestyle behaviors, and modern therapeutic advances that have been extending the lifespan of patients with CV diseases. In the future, the number of patients with HF will continue to rise globally due to the above-mentioned reasons as well as the rise in related comorbidities . HF is becoming an epidemic which has significant epidemiological variations among different geographic regions and countries . Geographically, the prevalence of HF varies; the Middle East, North Africa, and Central Europe have the highest prevalence rates, whereas Southeast Asia and Eastern Europe have lower rates. In the United States, the prevalence of HF has been reported to range from 2.4 to 3.0%; in European countries such as Italy, England, France, and Germany it is between 1.2 and 3.9% of the total population; in some Asian countries (Indonesia and Taiwan) the prevalence of chronic HF is 4-6% . According to the EPICA (Epidemiology of Heart Failure and Learning) research conducted in late 1990 in Portugal, the prevalence of chronic HF was 1.36% in the age group 25-49 years, 2.93% in the age group 50-59 years, 7.63% in the age group 60-69 years, 12.67% in the age group 70-79 years, and 16.14% for those aged over 80 years . The Rotterdam Study revealed that the prevalence of symptomatic HF increased from 0.5% in participants aged 55-64 years to over 14% in participants aged between 85 and 94 years . Chronic HF prevalence increased from 0.7% in the relatively younger age group (45-54 years) to more than 8% in participants over the age of 74 years, according to a population-based study from the USA . Moreover, HF is no longer a disease only for elders, since the prevalence of HF is also rising in younger individuals . The prevalence of chronic HF has been reported to be similar between the genders, although there are differences in the characteristics between women and men with HF. Among people aged 65-85 years, the incidence rate of HF in men roughly doubles, whereas in women the HF incidence rate triples. In female patients, HF tends to develop later in life and they have a longer life expectancy compared to men . According to studies from the USA, the main risk factors for HF development include ischemic heart disease, hypertension, diabetes mellitus, older age (>65 years), and obesity . Study results from European countries have also denoted similar risk factors but added smoking as a main risk factor for HF . Interestingly, rheumatic heart disease along with coronary heart disease and hypertension are the most common causes of HF in South Asian developing countries . Additionally, poorly controlled diabetes (HbA1c >= 8%), uncontrolled hypertension (SBP >= 160), and severe obesity (BMI >= 35 kg/m2) are the main risk factors contributing to HF incidence . In Mongolia, the prevalence of CV diseases has increased 2.5-fold during the past decade and has become a significant concern of the health sector . Identification of prevalence and risk factors for HF will support development of strategies for early detection and prevention ; however, there is no comprehensive study on the prevalence and risk factors of HF in Mongolia. The aim of this study is to determine the prevalence and risk factors of HF among the adult population of Mongolia. 2. Materials and Methods 2.1. Data Source and Sampling This population-based cross-sectional study was based on subjects attending primary health care centers and was conducted between January and May 2022. Since Mongolia is one of the least densely populated countries in the world with a certain diversity among its population, we used cluster combined with stratified and three-stage random sampling. Based on their geographical region, the country was divided into four distinct sampling zones. The sampling zones included selected districts from the capital city Ulaanbaatar and 7 provinces from the Western, Eastern, Central, and Mountainous regions of the country . 2.1.1. Study Sample Size The sample size was calculated based on the total population aged 20 years and older (n = 2,157,011) and an average prevalence from previous international studies (1.5%) assuming a 95% confidence interval (Z = 1.96) with a 2% acceptable margin of error (e = 0.02), which gave a sample size of 3600 subjects. 2.1.2. Sample Selection Study clusters (n = 75) and subjects were randomly selected from 7 provinces in the Western, Mountain, Eastern, and Central regions according to geographical zoning and 6 districts of the Ulaanbaatar city. At each primary health care center, the subjects were enrolled in this study by using systematic sampling and were stratified into 10-year-interval age groups. The target sample size including 3600 subjects composed of 900 subjects in the 20-29 age group, 900 subjects in the 30-39 age group, 675 subjects in the 40-49 age group, 600 subjects in the 50-59 age group, 375 subjects in the 60-69 age group, and 150 subjects in the 70 years and over age group. Because the exclusion criterion was subjects with incomplete data, 120 subjects (3.3%) were excluded from the final analysis. 2.2. Data Collection We performed quantitative survey methodology using standard questionnaires. Prior to the data collection, all research staff were provided with detailed instructions and trained for conducting interviews using study questionnaires. The questionnaire included subject's demographics, social characteristics, presence of CV risk factors, comorbidities, and HF-related symptoms. Educational level was divided into 3 groups, low, medium, and high. Marital status was categorized into 2 groups including married or cohabiting and divorced or single. Lifestyle characteristics such as smoking and alcohol consumption were classified as dichotomous variables: smoker or non-smokers, never or normal/abnormal use of alcohol. Diabetes mellitus was defined as self-reported physician-diagnosed diabetes and/or use of insulin and/or oral hypoglycemic medications. Coronary disease was defined as a prior myocardial infarction or revascularization (coronary bypass surgery or angioplasty). Physical examinations were performed by well-trained physicians in order to identify HF-related signs. Blood pressure, heart rate, oxygen saturation, respiratory rate, and weight were measured by physicians. Body mass index (BMI, kg/m2) was calculated by dividing the weight (kg) and height (m2). Obesity was defined as a BMI of 30.0 kg/m2 or greater. Hypertension was defined by a physician's diagnosis, systolic blood pressure >=130 mmHg or diastolic blood pressure >=80 mmHg, or use of antihypertensive medication. We defined atrial fibrillation as a history or the presence of atrial fibrillation on an electrocardiography. CKD was defined as kidney damage or glomerular filtration rate (GFR) <60 mL/min/1.73 m2. Anemia was defined as serum hemoglobin levels <13.0 g/dL (<130 g/L) for men and <12.0 g/dL (<120 g/L) for women. The International Classification of Disease (ICD) 10 codes was used for the following comorbidities: COPD(J44), sleep apnea (G47.3), and thyroid disorders (E03 and E05). Heart failure was defined as a syndrome recognized by the physician based on symptoms of exercise intolerance, signs of fluid retention, and response to therapy, according to the Guidelines of the ESC Working Group on Heart Failure. In our study, a diagnosis of chronic HF was based on the ESC clinical diagnostic criteria, including if the participant had chronic HF-related symptoms both at rest and during exercise (breathlessness, ankle swelling, and fatigue) or chronic HF-related signs (peripheral oedema, hepatomegaly, neck vein distention, third heart sound (S3) gallop rhythm, and pulmonary crepitations) and, in cases where there was doubt, the patient's response to diuretic treatment. An HF diagnosis was considered when the first and second criteria were both met . 2.3. Statistical Analysis Patients' demographic characteristics and clinical variables were analyzed in the whole sample using descriptive statistics. Continuous variables were expressed as means +- standard deviations (for normal distribution) or medians with interquartile range (non-normal distribution). Categorical variables were shown as absolute numbers and percentages. The distribution of normality was based on visual assessment of a histogram and the Kolmogorov-Smirnov test. Categorical data were compared using a chi-square test. while continuous variables were compared using an independent sample t-test. Correlations between cardiovascular risk factors and both HF and non-HF groups were assessed using the Pearson's correlation coefficient. In addition, a logistic regression analysis was performed to calculate the odds ratio to assess associations between risk factors and covariates. Statistical significance was considered for two-sided p-values less than <0.05. The Statistical Package for the Social Sciences (SPSS 24.0) was used for data analysis. 3. Results In total, 3480 patients aged 18-87 years were enrolled in this study, of which 1345 (39%) were men, and the median age was 41.0 years (IQR 30-54 years). Demographic characteristics of the study population are summarized according to gender in Table 1. Compared to men, women had significantly higher education and intellectual labor than men. Age groups, marital status, and proportion of population in the main administrative groups were comparable in terms of gender. The prevalence of CV risk factors according to age groups are included in Table 2. The frequencies of some risk factors increased along with an increase in age group; hypertension, diabetes mellitus, and CAD were more prevalent in patients aged 70 years and over. Whereas middle-aged patients (40-49 and 50-59 years, some included 60-69 years) had significantly higher prevalences of the remaining risk factors including valvular heart disease (3%, 5%, and 4%), abnormal alcohol consumption (11% and 13%), smoking (24% and 23%), and obesity (30%, 29%, and 30%). Comparing the above-mentioned risk factors and comorbidities according to the main administrative groups, in the urban population there was significantly more smoking (22% vs. 19%) and diabetes mellitus (7% vs. 5%) than in the rural population (Table 3), while in the rural population there was significantly more abnormal alcohol consumption (10% vs. 7%) and obesity (25% vs. 21%). The overall prevalence of chronic HF was 4.94% for the total study population based on the ESC diagnostic criteria for HF. The prevalence of chronic HF strongly increased with age; it was 0.7% for the 20-29 year age group, while its frequency was 21.0% for the 70-87-year age group . Figure 3 shows the age-specific prevalence of overall HF for the 10-year-interval age groups in men and women. The prevalence of heart failure increased from 1.2% for men aged 20-29 years to 17.7% for men aged >=70 years. For women, the prevalence increased from 0.5% in the lowest age group to 23.3% in the highest age group. In the 40-49 and 50-59 year age groups, men and women showed comparable point prevalences. The prevalences of HF in the urban and rural populations are shown in Figure 4. The prevalences of HF for males and females in the rural population were higher than those in the urban population. The demographic and social characteristics of the study participants are shown in Table 4. The patients with HF were significantly older (median age 57 years), fewer had higher level education (18% vs. 36%), and more had low level education (30% vs. 13%) and were unemployed (58% vs. 40%) compared to subjects without HF. The remaining variables including sex and marital status were comparable. Demographic and social characteristics of the study participants are shown in Table 4. For the logistics regression analysis, cardiovascular risk factors were included for analyzing the correlations between the variables and HF (Table 5). Among the cardiovascular risk factors, CAD, hypertension, valvular heart disease, abnormal alcohol consumption, and obesity significantly increased the risk of HF. There were significant differences between the non-HF and HF groups regarding cardiovascular risk factors and clinical characteristics based on a physical examination (Table 6). Patients with HF experienced significantly more cardiovascular risk factors. The physical examination for HF showed that patients with HF, as compared with patients without HF, had significantly higher body mass index (28.8 kg/m2 vs. 26.2 kg/m2), heart rate (84.3 bpm vs. 79.6 bpm), respiratory rate (19.2 vs. 18.2), and systolic (137.7 mmHg vs. 121.2 mmHg) and diastolic blood pressure (86.2 mmHg vs. 78.3 mmHg), and lower oxygen saturation (95.0% vs. 96.2%). There were significant differences between the non-HF and HF groups regarding use of medications (Table 7). Diuretics were the most used medication, followed by renin-angiotensin system (RAAS) inhibitors and b-blockers. Atrial fibrillation (24%), sleep apnea (20%), and COPD (16%) were common comorbidities in patients with HF . 4. Discussion First, this is the first study that revealed the prevalence of HF including both urban and rural populations using different clusters. Secondly, we found that Mongolian patients with HF had significantly higher frequencies of comorbidities and risk factors and poorer physical characteristics. Thirdly, hypertension, coronary heart disease, and valvular heart disease were leading CV causes of HF in Mongolian patients. Based on our study results, the prevalence of HF (4.94%) in Mongolian adults is higher than that reported by studies from USA, some European countries (such as Italy, England, France, and Germany), some Asian countries (such as China and Japan), relatively comparable to Singapore, and lower than the prevalence in Malaysia . The present study findings suggest that the reason for the prevalence of HF was higher in the rural population than in the urban population, could be explained by disparities in economic levels (types of occupation), lifestyles (abnormal alcohol usage and obesity), education levels (p < 0.0001), and clinical conditions between urban and rural areas. These findings are consistent with those of a study in India . Moreover, our findings show that unemployed and low education increase the risk of HF compared to participants without HF. A recent meta-analysis of 11 studies found that low socioeconomic status assessed by all common measures (education, income, occupation, and area) independently increased the incidence risk of heart failure by 62%, overall . Based on our data, the prevalences of HF were 0.7%, 1.6%, 4.1%, 8.2%, 11.3%, and 21.0% for subjects who were 20-29, 30-39, 40-49, 50-59, 60-69, and >=70 years of age, respectively. These findings are consistent with previous studies that have demonstrated an increased prevalence of HF following advanced age . Researchers from the same region have reported that the prevalences of HF were 0.57%, 3.86%, and 7.55% for individuals who were 25-64, 65-79, and >=80 years of age, respectively . The current study showed that the prevalence of HF increased with age from 2.0% among 20-49-year-old subjects to 11.0% among 50-87-year-old subjects. Another study showed the prevalence of HF strongly increased with age from 3.0% among 45-54-year-old subjects to 22.0% among 75-83-year-old subjects . We observed that HF was more prevalent in men compared to women, despite a significantly higher prevalence of HF in women aged 70 years and older compared to men. These findings agreed with the results of a Chinese study . HF is known primarily as a disease of the elderly. However, recent studies have indicated that the HF burden may be increasing in young individuals. Thus, the mean age for HF onset has been declining and the incidence of patients with HF aged below 50 years has increased by two-fold, particularly increasing from 3% to 6% . In a Swedish study that linked national hospital discharge and death registries between 1987 and 2006, HF incidence increased in the last 5-year period by 50% among people aged 18-34 years and 43% among those aged 35-44 years . In our study, the median age for a diagnosis of HF was 50 years, while the mean age at HF diagnosis was 73.7 +- 14.3 years in a UK population aged >=30 years . Overall, despite the South Asian and African ethnicity groups being significantly younger at HF onset than the Caucasian ethnicity group, they had similar or better cardiovascular risk profiles, which agreed with those previously reported in a younger UK general population . The results of our study showed a highly age-specific prevalence of HF compared to other studies that have mostly included and examined populations aged 45 years and over and used various diagnostic approaches and criteria (see Table 8). Although medical records were reviewed for the definition of HF to identify HF diagnosis according to the Framingham Criteria in the Olmsted County Study , subjects in the Rotterdam Study were clinically examined to identify symptoms and signs suggestive of HF (e.g., shortness of breath, ankle oedema, and pulmonary crepitations) . There is an ongoing debate regarding the definition of heart failure and there is a lack of a gold standard for assessing the presence of the syndrome in population-based studies. According to the ESC Guidelines on the diagnosis of HF, to establish the presence of heart failure, objective evidence of cardiac dysfunction must be present in addition to symptoms or medication for HF . In the Framingham study, the overall prevalence of HF was 0.7% for those aged between 50 and 89 years, varying between 0.1% and 7.9% with age . In the Rochester study, in 1986, the prevalence of HF in those over 35 years was 1.9%, increasing from 1% to 7.6% with age . A recent randomized controlled trial suggested that the most common risk factors for HF were CAD, hypertension, and diabetes mellitus . More specifically, the risk factors highly correlated with HF incidence included poorly controlled diabetes (HbA1c >= 8%), uncontrolled hypertension (SBP >= 160), and advanced obesity (BMI >= 35) . Likewise, our study demonstrated that a previous history of CAD, hypertension, valvular heart disease, obesity, and abnormal alcohol consumption were main risk factors of HF. Because these findings were different compared to the NHANES study , we assume that it could be caused by the disparities of living standards and cultural differences. The present study demonstrated that coronary artery disease (CAD) is the strongest risk factor for the development of HF among other risk factors with a prevalence of 29.1% (n = 50) in the total HF population. Secondly, having hypertension was also viewed as a major factor in the progression of HF with a prevalence of 84.9% (n = 149) in the HF population in this study. These results were in line with those of former studies such as a cardiovascular health study and a Spanish study (81.8%) . Our analysis also supports that valvular heart disease and obesity are important risk factors in the development of HF. This could be because of a higher incidence of rheumatic heart disease, lack of health education, and the cultural point of view among the Mongolian population. The strength of our study is the population size which is representative enough for the overall population of Mongolia. Therefore, our study results could be generalizable to patients with and without HF in the general adult population. A limitation is, however, that the symptoms suggestive of HF as well as different disease prevalences (e.g., CHD, COPD, and diabetes) were self-reported, and therefore we were not able to validate this information. Another limitation of this study is that the participants who met the criteria for the clinical diagnosis of chronic HF were not further evaluated through echocardiography and natriuretic peptide testing for a confirmation of the diagnosis. Moreover, the diagnosis of HF was not validated in this study, which increases the risk of observer bias. 5. Conclusions This is the first investigation in Mongolia that describes the prevalence of HF among the general population. The prevalence of HF appears high (4.94%) in the Mongolian population compared with other studies. Our study revealed that coronary heart disease, hypertension, and valvular heart disease are the three foremost risk factors in the development chronic heart failure. Acknowledgments The study team would like to express our sincere gratitude to the Mongolian Society of Cardiologists, provincial and district primary care practitioners for their cooperation. Author Contributions Conceptualization, T.U.; Methodology, P.S., B.B. (Batzorig Bayartsogt) and T.U.; Formal analysis, B.B. (Batzorig Bayartsogt); Investigation, P.S., G.U., C.K., B.B. (Burmaa Badrakh), S.T. (Sumiya Tserendavaa) and B.-E.N.; Resources, M.D. and M.-U.K.; Data curation, P.S., B.B. (Batzorig Bayartsogt), G.U., C.K., B.B. (Burmaa Badrakh), S.T. (Sumiya Tserendavaa), S.T. (Sodchimeg Tsedensodnom), B.-E.N. and R.M.; Writing--original draft, P.S., B.B. (Batzorig Bayartsogt), G.U., B.B. (Bolortuul Byambatsogt), C.K., B.B. (Burmaa Badrakh), S.T. (Sumiya Tserendavaa), N.S., M.D., M.-U.K., S.T. (Sodchimeg Tsedensodnom), B.-E.N., R.M. and T.U.; Writing--review & editing, N.S., R.M. and T.U.; Visualization, B.B. (Bolortuul Byambatsogt); Supervision, N.S. and T.U.; Project administration, M.D. and M.-U.K. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Research ethical permission was obtained from the Ethics Committee of the Mongolian National University of Medical Sciences (Code: 21/02/02, 22 October 2021). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement Raw data that support the findings of this study are available from the corresponding author, upon reasonable request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Sampling clusters. Figure 2 Age-specific prevalence of HF. Note: The vertical axis shows prevalence rate expressed as a percentage. The horizontal axis shows age groups. Figure 3 Age-specific prevalence of overall HF by gender. Figure 4 The prevalences of HF in the urban and rural populations. Note: The horizontal axis shows the total, urban, and rural population distribution. The vertical axis shows prevalence rate expressed as a percentage. Figure 5 Sex-specific comorbidities in HF. Note: AF, atrial fibrillation; COPD, chronic obstructive pulmonary disease, CKD, chronic kidney disease. The horizontal axis shows comorbidities. The vertical axis shows prevalence rate expressed as a percentage. diagnostics-13-00999-t001_Table 1 Table 1 Demographic and social characteristics of the study population. Characteristics Total Study Population (n = 3480) Male (n = 1345) Female (n = 2135) p-Value Age group 0.484 20-29, n (%) 807 (23.2) 286 (21.3) 521 (24.4) 30-39, n (%) 863 (24.8) 335 (24.9) 528 (24.7) 40-49, n (%) 682 (19.6) 268 (19.9) 414 (19.4) 50-59, n (%) 595 (17.1) 236 (17.6) 359 (16.8) 60-69, n (%) 381 (10.9) 158 (11.8) 223 (10.5) >=70, n (%) 152 (4.4) 62 (4.6) 90 (4.2) Education <0.001 Higher, n (%) 1229 (35.3) 418 (31.1) 811 (38.0) Medium, n (%) 1784 (51.3) 696 (51.7) 1088 (51.0) Lower, n (%) 467(13.4) 231 (17.2) 236 (11.0) Occupation <0.001 Manual labor, n (%) 830 (23.9) 398 (29.6) 432 (20.2) Intellectual labor, n (%) 1238 (35.7) 457 (34.0) 781 (36.6) Unemployed, n (%) 1412 (40.4) 490 (36.4) 922 (43.2) Marital status 0.432 Married, n (%) 2661 (76.5) 1032 (76.7) 1629 (76.3) Divorced, n (%) 132 (3.8) 45 (3.3) 87 (4.1) Unmarried, n (%) 687 (19.7) 268 (20.0) 419 (19.6) 0.459 Administrative region Urban area, n (%) 1686 (48) 641 (47.6) 1045 (49.9) Rural area, n (%) 1794 (52) 704 (52.3) 1090 (51.1) diagnostics-13-00999-t002_Table 2 Table 2 Risk factors for HF according to the age groups. Variables Total Subjects (n = 3480) 20-29 Years (n = 807) 30-39 Years (n = 863) 40-49 Years (n = 682) 50-59 Years (n = 595) 60-69 Years (n = 381) >=70 Years (n = 152) p-Value CAD, n (%) 90 (3) 1 (0) 6 (1) 20 (3) 26 (4) 23 (6) 14 (9) <0.0001 Hypertension, n (%) 1358 (39) 91 (11) 210 (24) 295 (43) 392 (66) 246 (65) 124 (82) <0.0001 DM, n (%) 198 (6) 12 (2) 33 (4) 37 (5) 59 (10) 36 (9) 21 (14) <0.0001 VHD, n (%) 82 (2) 8 (1) 12 (1) 17 (3) 28 (5) 16 (4) 1 (1) <0.0001 Obesity, n (%) 1358 (39) 86 (11) 175 (20) 207 (30) 173 (29) 114 (30) 35 (23) <0.0001 Smoking, n (%) 712 (21) 117 (15) 191 (22) 164 (24) 135 (23) 78 (21) 27 (18) <0.0001 Abnormal alcohol consumption, n (%) 300 (9) 30 (4) 82 (10) 74 (11) 76 (13) 28 (7) 10 (7) <0.0001 Note: CAD, coronary heart disease; DM, diabetes mellitus; VHD, valvular heart disease. diagnostics-13-00999-t003_Table 3 Table 3 Risk factors for HF according to the main administrative groups. Variables Total Participants n = 3480 Population in Urban Area (n = 1686) Population in Rural Area (n = 1794) p-Value Coronary heart disease, n (%) 90 (3) 46 (3) 44 (3) 0.609 Hypertension, n (%) 1358 (39) 661 (39) 697 (39) 0.831 Diabetes mellitus, n (%) 198 (6) 113 (7) 85 (5) 0.012 Valvular heart disease 82 (2) 35 (3) 47 (2) 0.290 Obesity, n (%) 790 (23) 350 (21) 440 (25) 0.008 Smoking, n (%) 712 (21) 373 (22) 339 (19) 0.018 Abnormal alcohol consumption, n (%) 300 (9) 118 (7) 182 (10) 0.001 diagnostics-13-00999-t004_Table 4 Table 4 Comparison of demographic and social characteristics between HF and non-HF groups. Characteristics Total Study Population n = 3480 Non-HF Group n = 3308 (95%) HF Group n = 172 (5%) p-Value Sex 0.227 Male 1345 (39) 1271 (94.5) 74 (5.5) Female 2135 (61) 2037 (95.4) 98 (4.6) Age group <0.0001 20-29 807 (23) 801(24) 6 (4) 30-39 863 (25) 849 (26) 14 (8) 40-49 682 (20) 654 (20) 28 (16) 50-59 595 (17) 546 (16) 49 (28) 60-69 381 (11) 338 (10) 43 (25) >=70 152 (4) 120 (4) 32 (19) Average age 41.0 (30.0-54.0) 40.0 (30.0-53.0) 57.0 (49.0-65.8) <0.0001 Education <0.0001 Higher 1229 (35) 1198 (36) 31 (18) Medium 1784(51) 1695 (51) 89 (52) Lower 467(14) 415 (13) 52 (30) Occupation <0.0001 Manual labor work 830 (24) 793 (24) 37 (21) Intellectual labor work 1238 (51) 1202 (36) 36 (21) Unemployed 1412 (40) 1313 (40) 99 (58) Marital status 0.479 Married 2661 (76) 2523 (76) 138 (80) Divorced 132 (4) 127 (4) 5 (3) Unmarried 687 (20) 658 (20) 29 (17) diagnostics-13-00999-t005_Table 5 Table 5 Logistic regression analysis adjusted to age and gender. Variable OR Min Value Max Value p-Value Hypertension 4.855 3.127 7.538 <0.0001 CAD 5.117 3.040 8.614 <0.0001 Valvular heart disease 3.872 2.112 7.099 <0.0001 Abnormal alcohol consumption 1.861 1.155 2.998 0.011 Smoking 1.391 0.918 2.109 0.120 Obesity 2.136 1.542 2.959 <0.0001 Diabetes mellitus 1.440 0.865 2.397 0.161 CAD, coronary heart disease diagnostics-13-00999-t006_Table 6 Table 6 Comparison of risk factors and clinical charactheristics between the HF and non-HF groups. Variables Total Study Population (n = 3480) Non-HF Group n = 3308 (95%) HF Group n = 172 (5%) p-Value Risk factors Hypertension, n (%) 1358 (39) 1213 (37) 145 (84) <0.0001 Valvular heart disease, n (%) 82 (2) 67 (2) 15 (9) <0.0001 Abnormal alcohol consumption, n (%) 300(9) 275 (8) 25 (15) 0.005 Smoking, n (%) 712 (20) 667 (20) 45 (26) 0.057 Diabetes mellitus, n (%) 198 (6) 178 (5) 20 (12) 0.001 Obesity, n (%) 790 (23) 719 (22) 71 (41) <0.0001 Coronary artery disease, n (%) 90 (3) 63 (2) 27 (16) <0.0001 Clinical charactheristics Body mass index, (kg/m2) 26.4 +- 5.1 26.2 +- 5.0 28.8 +- 6.2 <0.0001 Heart rate, per minute 80.0 +- 11.1 79.6 +- 10.8 84.9 +- 15.7 <0.0001 Respiratory rate, per minute 18.2 +- 5.1 18.2 +- 5.2 19.2 +- 3.1 0.012 Oxygen saturation, % 96.2 +- 3.7 96.2 +-3.8 95.0 +- 3.1 <0.0001 Systolic blood pressure, mmHg 122.0 +- 19.4 121.2 +- 18.9 137.7 +- 22.4 <0.0001 Diastolic blood pressure, mmHg 78.7 +- 12.6 78.3 +- 12.4 86.2 +- 15.2 <0.0001 diagnostics-13-00999-t007_Table 7 Table 7 Comparison of medications between the HF and non-HF groups. Medications Non-HF Group (n = 3308) HF Group (n = 172) p-Value Diuretics, n (%) 181 (5.4) 88 (51.2) <0.0001 RAAS inhibitors, n (%) 210 (6.3) 86 (50.0) <0.0001 Beta blockers, n (%) 190 (5.7) 75 (43.6) <0.0001 Sacubitril/valsartan, n (%) 42 (1.3) 44 (25.6) <0.0001 Digoxin, n (%) 26 (0.8) 26 (15.1) <0.0001 Ivabradine, n (%) 27 (0.8) 22 (12.8) <0.0001 Note: RAAS, renin-angiotensin-aldosterone system diagnostics-13-00999-t008_Table 8 Table 8 Prevalence of heart failure in Mongolia compared with other population-based studies. Age Group Present Study (Mongolia) CARLA Study (Germany) Rotterdam Study (The Netherlands) Olmsted County Study (USA) 45-54 years 4.3% 3.0% 0.7% - 55-64 years 9.4% 6.0% 0.7% 1.3% 65-74 years 13.1% 10.4% 2.7% 1.5% 75-84 years 25.7% 22.0% 13% 8.4 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Savarese G. Lund L.H. Global Public Health Burden of Heart Failure Card. Fail. Rev. 2017 3 7 11 10.15420/cfr.2016:25:2 28785469 2. GBD 2017 Disease and Injury Incidence and Prevalence Collaborators Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990-2017: A systematic analysis for the Global Burden of Disease Study 2017 Lancet 2018 392 1789 10.1016/S0140-6736(18)32279-7 30496104 3. Cho D.H. Yoo B.S. Current Prevalence, Incidence, and Outcomes of Heart Failure with Preserved Ejection Fraction Heart Fail. Clin. 2021 17 315 326 10.1016/j.hfc.2021.03.002 34051964 4. Ziaeian B. Fonarow G.C. Epidemiology and aetiology of heart failure Nat. Rev. Cardiol. 2016 13 68 378 10.1038/nrcardio.2016.25 5. Savarese G. Becher P.M. Lund L.H. Seferovic P. Giuseppe Rosano M.C. Andrew Coats J.S. Global burden of heart failure: A comprehensive and updated review of epidemiology Cardiovasc. Cardiovasc. Res. 2022 118 3272 3287 10.1093/cvr/cvac013 35150240 6. Ceia F. Fonseca C. Mota T. Morais H. Matias F. Sousa A. Antonio Oliveira A. Prevalence of chronic heart failure in Southwestern Europe: The EPICA study Eur. J. Heart Fail. 2002 4 531 539 10.1016/S1388-9842(02)00034-X 12167394 7. Mosterd A. Hoes A.W. de Bruyne M.C. Deckers J.M. Linker D.T. Hofman A. Grobbee D.E. Prevalence of heart failure and left ventricular dysfunction in the general population; The Rotterdam Study 1999 Eur. Heart J. 2000 20 447 455 10.1053/euhj.1998.1239 8. Redfield M.M. Jacobsen S.J. Burnett J.C. Jr. Mahoney D.W. Bailey K.R. Rodeheffer R.J. Burden of systolic and diastolic ventricular dysfunction in the community: Appreciating the scope of the heart failure epidemic J. Amer. Med. Assoc. 2003 289 194 202 10.1001/jama.289.2.194 12517230 9. Christiansen M.N. Kober L. Weeke P. Vasan R.S. Jeppesen J.L. Smith J.G. Gislason G.H. Torp-Pedersen C. Andersson C. Age-Specific Trends in Incidence, Mortality, and Comorbidities of Heart Failure in Denmark, 1995 to 2012 Circulation 2017 135 1214 1223 10.1161/CIRCULATIONAHA.116.025941 28174193 10. Barasa A. Schaufelberger M. Lappas G. Swedberg K. Dellborg M. Rosengren A. Heart failure in young adults: 20-year trends in hospitalization, aetiology, and case fatality in Sweden Eur. Heart J. 2014 35 25 32 10.1093/eurheartj/eht278 23900697 11. Jencks S.F. Williams M.V. Coleman E.A. Rehospitalizations among patients in the Medicare fee-for-service program N. Engl. J. Med. 2009 360 418 428 10.1056/NEJMsa0803563 12. Komanduria S. Jadhaoa Y. Gudurua S.S. Cheriyatha P. Wertb Y. Prevalence and risk factors of heart failure in the USA: NHANES 2013-2014. Epidemiological follow-up study J. Community Hosp. Intern. Med. Perspect. 2017 7 15 20 10.1080/20009666.2016.1264696 28634519 13. Pfeffer M.A. Shah A.M. Borlaug B.A. Heart failure with preserved ejection fraction in perspective Circ. Res. 2019 24 1598 1617 10.1161/CIRCRESAHA.119.313572 31120821 14. Pillai H.S. Ganapathi S. Heart failure in South Asia Curr. Cardiol. Rev. 2013 9 102 111 10.2174/1573403X11309020003 23597297 15. Hamo C.E. Kwak L. Wang D. Florido R. Echouffo-Tcheugui J.B. Blumenthal R.S. Loehr L. Matsushita K. Nambi V. Ballantyne C.M. Heart failure risk associated with severity of modifiable heart failure risk factors: The ARIC Study J. Am. Heart Assoc. 2022 11 e021583 10.1161/JAHA.121.021583 35156388 16. Center for Health Development Health Indicators Center for Health Development Ulaanbaatar, Mongolia 2021 17. Schocken D.D. Benjamin E.J. Fonarow G.C. Krumholz H.M. Levy D. Mensah G.A. Narula J. Shor E.S. Young J.B. Hong Y. Prevention of heart failure: A scientific statement from the American Heart Association Councils on Epidemiology and Prevention, Clinical Cardiology, Cardiovascular Nursing, and High Blood Pressure Research; Quality of Care and Outcomes Research Interdisciplinary Working Group; and Functional Genomics and Translational Biology Interdisciplinary Working GroupPrevention of Heart Failure Circulation 2008 117 2544 2565 10.1161/CIRCULATIONAHA.107.188965 18391114 18. Swedberg K. Cleland J. Dargie H. Drexler H. Follath F. Komajda M. Tavazzi L. Smiseth O.A. Gavazzi A. Haverich A. Guidelines for the diagnosis and treatment of chronic heart failure: Executive summary (update 2005): The task force for the diagnosis and treatment of chronic heart failure of the European Society of Cardiology Eur. Heart J. 2005 26 1115 1140 10.1093/eurheartj/ehi204 15901669 19. Sakata Y. Shimokawa H. Epidemiology of heart failure in Asia Circ. J. 2013 7 2209 2217 10.1253/circj.CJ-13-0971 23955345 20. Hu S.S. Kong L.Z. Gao R.L. Zhu M.L. Wang W. Wang Y.J. Wu Z.S. Chen W.W. Liu M.B. Outline of the report oncardiovascular disease in China, 2010 Biomed. Env. Sci. 2012 25 251 256 10.3967/0895-3988.2012.03.001 22840574 21. Ponikowski P. Anker S.D. AlHabib K.F. Cowie M.R. Force T.L. Hu S. Jaarsma T. Krum H. Rastogi V. Rohde L.E. Heart failure: Preventing disease and death worldwide ESC Heart Fail. 2014 1 4 25 10.1002/ehf2.12005 28834669 22. Lam C.S. Heart failure in Southeast Asia: Facts and numbers ESC Heart Fail. 2015 2 46 49 10.1002/ehf2.12036 28834655 23. Chaturvedi V. Parakh N. Seth S. Bhargava B. Ramakrishnan S. Roy A. Saxena A. Gupta N. Misra P. Rai S.K. Heart Failure in India: The INDUS (INDia Ukieri Study) Study J. Pract. Cardiovasc. Sci. 2016 2 28 35 10.4103/2395-5414.182988 24. Potter E.L. Hopper I. Sen J. Salim A. Marwick T.H. Impact of socioeconomic status on incident heart failure and left ventricular dysfunction: Systematic review and meta-analysis Eur. Heart J. Qual. Care Clin. Outcomes. 2019 5 169 179 10.1093/ehjqcco/qcy047 30295783 25. Wang H. Chai K. Du M. Wang S. Prevalence and Incidence of Heart Failure Among Urban Patients China Circ. Heart Fail. 2021 14 e008406 10.1161/CIRCHEARTFAILURE.121.008406 34455858 26. Tiller D. Russ M. Greiser K.H. Nuding S. Ebelt H. Kluttig A. Kors J.A. Thiery J. Bruegel M. Haerting J. Prevalence of Symptomatic Heart Failure with Reduced and with Normal Ejection Fraction in an Elderly General Population-The CARLA Study PLoS ONE. 2013 8 e5922 10.1371/journal.pone.0059225 27. Cleland J.G. Erdmann E. Ferrari R. Hess O.M. Poole-Wilson P.A. Remme W.J. Guidelines for the diagnosis of heart failure Eur. Heart J. 1995 16 741 751 7588917 28. McKee P.A. Castelli W.P. McNamara P.M. Kannel W.B. The natural history of congestive heart failure: The Framingham study N. Engl. J. Med. 1971 26 441 446 10.1056/NEJM197112232852601 29. Phillips S.J. Wishnant J.P. O'Fallon W.M. Frye R.L. Prevalence of cardiovascular disease and diabettes mellitus in residents of Rochester, Minnesota Mayo. Clin. Prod. 1990 65 344 359 10.1016/S0025-6196(12)62535-X 30. Banerjee A. Pasea L. Chung S. Direk K. Asselbergs F.W. Grobbee D.E. Kotecha D. Anker S.A. Dyszynski T. Tyl B. A population-based study of 92 clinically recognized risk factors for heart failure: Co-occurrence, prognosis and preventive potential Eur. J. Heart Fail. 2022 24 466 480 10.1002/ejhf.2417 34969173 31. Gottdiener J.S. Arnold A.M. Aurigemma G.P. Polak J.F. Tracy R.P. Kitzman D.W. Gardin J.M. Rutledge J.E. Boineau R.C. Predictors of congestive heart failure in the elderly: The Cardiovascular Health Study J. Am. Coll. Cardiol. 2000 35 1628 1637 10.1016/S0735-1097(00)00582-9 10807470 32. Lawson C.A. Zaccardi F. Squire I. Okhai H. Davies M. Huang W. Mamas M. Lam C.S.P. Khunti K. Umesh T. Risk Factors for Heart Failure Circ. Heart Fail. 2020 13 e006472 10.1161/CIRCHEARTFAILURE.119.006472 32059630 |
PMC10000623 | Background Hyperphenylalaninemias (HPA) are due to several gene mutations, of which the PAH gene is the most frequently involved. Prevalence and incidence of disease vary between populations, with genotype/phenotype correlations not always capable to correctly predict disease severity. The aim of this study was to give an overview of PAH mutations among one of the largest cohort of patients among Europe, born in Lombardy (Italy) starting from late 1970 s and including over a 60 years of activity; furthermore, to evaluate and discuss identified genotype/phenotype correlations and related reliability. Patients/Methods Eight hundred and twenty-six HPA patients in current follow-up at the San Paolo Hospital in Milan (Italy) were retrospectively reviewed, including molecular results and allelic phenotype and genotype values (attributed on the basis of the APV/GPV system) to verify genotype-phenotype correlations. Results A total of 166 different PAH variants were reviewed; of those, seven variants were identified as not previously described in literature. Most frequently reported variant was p.Ala403Val, followed by p.Arg261Gln, p.Val245Ala, IVS10-11 g>a, p.Tyr414Cys and p.Leu48Ser. Phenotype prediction, based on APV/GPV, matched the actual phenotype in most cases, but not always. Conclusion/Discussion The cohort of patients included in this study constitute a representative sample of the HPA population worldwide. Studies on this sample may allow to improve clinical and genetic evaluation performances for affected patients, consequently to develop personalized medicine interventions and provide more precise indications on the correct treatment approach based on the accumulated evidence, also in light of a prognostically reliable but not always conclusive APV/GPV system. An overview of PAH mutations among one of the largest cohort of patients among Europe to evaluate and discuss identified genotype/phenotype correlations and related reliability. APV/GPV system is a prognostically reliable but not always conclusive system. genotype hyperphenylalaninemia PAH phenotype phenylketonuria PKU source-schema-version-number2.0 cover-dateMarch 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.2.6 mode:remove_FC converted:10.03.2023 Rovelli V , Cefalo G , Ercoli V , et al. Hyperphenylalaninemias genotyping: Results of over 60 years of history in Lombardy, Italy. Endocrinol Diab Metab. 2023;6 :e396. doi:10.1002/edm2.396 pmc1 INTRODUCTION Hyperphenylalaninemias (HPAs) are heterogeneous group of autosomal recessive inborn errors of metabolism characterized by the inability to metabolize phenylalanine (Phe) due to enzyme defects of either phenylalanine hydroxylase (PAH) or its cofactors, resulting in possible consequent neurological damage. 1 , 2 The diagnosis is based on the finding of mutations in the PAH gene, with more than 1500 pathogenic variants reported to date (most frequently missense mutations), or other genes related to cofactor deficiencies (such as QDPR, PTS, GCH1 and PCBD1). 13 Variants in DNAJC12 gene have also been recently described as causing possible hyperphenylalaninemia. 14 , 15 , 16 Usually, patients with HPA are detected through newborn screening for phenylketonuria (PKU) and defined as affected by milder to more severe forms (PKU/HPA) depending on serum phenylalanine levels, 3 which is due to a remarkable allelic variability accounting for different related patterns of phenylalanine hydroxylase residual enzymatic activity (EC 1.14.16.1). 4 , 5 , 6 Prompt and correct identification/prediction of the hyperphenylalaninemia subtype represent the most important task on a clinical standpoint in order to set appropriate treatments, including in most cases a tailor-made dietary intervention according to patients' needs, metabolic control and Phe tolerance. 7 , 8 PAH molecular pattern represent one of the mostly recognized predictive elements for possible metabolic outcomes. Furthermore, genotype/phenotype correlations have been reported with the expansion of a system of arbitrary values that, based on the assignment of an Allelic Phenotype Value (APV) to each PAH variant (GPV), seems to be capable of predicting the clinical phenotype with a validity index between 80% and 90%. 9 , 10 , 11 , 12 In this paper, we review the mutation spectrum of the hyperphenylalaninemic patients born since late '70 s in Lombardy, Italy, in follow-up at the Clinical Department of Paediatrics, San Paolo Hospital, ASST Santi Paolo e Carlo, University of Milan, Italy, analysing allelic phenotype and genotype values (attributed on the basis of the APV/GPV system) to verify genotype-phenotype correlations and discussing potential clinical implications. 2 PATIENTS AND METHODS After full anonymization, medical records of patients diagnosed with any form of hyperphenylalaninemia between 1955 and 2020 and in follow-up at the Clinical Department of Paediatrics - San Paolo Hospital, ASST Santi Paolo e Carlo, University of Milan, Italy, were retrospectively reviewed and studied, including demographic details, indices of metabolic control and genetic analysis results. Data regarding PAH, QDPR, PTS, GCH1 and PCBD1 analysis were collected considering genomic DNA previously extracted from whole blood samples with different techniques; DNAJC12 analysis results were collected when available. Used techniques for genetic analysis included variably DGGE (Denaturing Gradient Gel Electrophoresis), direct sequencing, HRM (High Resolution Melt), Sanger sequencing, copy number deletions or duplications with Multiple Ligation Probe Amplification (MLPA, MRC Holland, Resnova) or Next-Generation Sequencing (NGS), following technological advances over years (MiSeq platform - Illumina and custom TrueSeq Custom Amplicon kit 2016-2018, Nextera Rapid Capture Custom enrichment chemistry starting from 2019-2020). Variants were described according to Human Genome Variation Society (HGVS) nomenclature guidelines (using Mutalyzer Name Checker tool) and classified as disease causing when described as pathogenic or likely pathogenic in Clinvar, PAHvdb, LOVD or HGMD databases. As PAHvdb is linked to the genotype-phenotype BIOPKU database (BIOPKU; ), this was also used for analyses, providing additional information including allelic phenotype values (APV). The accession number for PAH was RefSeq: ENSG00000171759; GeneBank: NM_000277.1. For newly identified variants, VarSite was used as data source 13 with reference PAH protein sequence UNIPROT P00439. All novel variants were evaluated using ACMG classification and Varsome. 14 , 15 , 16 New variants were considered 'pathogenic' or 'likely pathogenic' and reported after evaluating classification of functional Prediction tools (Mutation Taster, ACMG classification, Varsome, FATHMM, Varsite). To predict phenotype in patients with PAH deficiency, results of Phe concentrations at the newborn screening were considered, including either dosages in dried blood spots (DBS) or venous puncture. For cases in which newborn screening details could not be retrieved due to incompleteness of documentation (including late diagnosed patients), available results of Phe concentrations prior to the start of any type of treatment were considered (either on DBS or by venous puncture). For classification purposes, the BIOPKUdb proposed model (BIOPKU; ) was used, as follows: classicPKU (cPKU >= 1200 mmol/L), mildPKU (mPKU 600-1199 mmol/L) and mildHPA (mHPA < 599 mmol/L). Based on the European guidelines on phenylketonuria, 17 , 18 the distinction between forms 'requiring treatment' (that is pre-treatment blood Phe levels >=360 mmol/L) and 'not requiring treatment' (that is pre-treatment blood Phe levels 120-360 mmol/L), was also applied to add more insights defining disease type and severity. 18 The Allelic Phenotype Values (APV) system proposed by Garbade 5 , 11 was calculated and used to perform genotype/phenotype correlations thus, for each patient, APVs related to both pathogenic variants of PAH, as proposed in the BIOPKUdb and PAHvdb databases, were reported (APVs range between 0 and 10, identifying cPKU for APV 0-2.7, mPKU for APV 2.8-6.6 and mHPA for APV 6.7-10). 11 Based on APVs, Genotypic Phenotype Values (GPVs) were also calculated and intended as the higher found APV (APVmax) between the two variants; this was considered applicable since the milder variant (that is with a higher APV) is contemplated as always dominant over the severe one. 11 , 19 Such type of evaluations were not performed for patients with no reported variants and/or with only a single variant reported (e.g., simple heterozygotes with a biochemical diagnosis of hyperphenylalaninemia), with the exception of patients in which the reported mutation was related to a GPV > 6.7 because, being the mildest one, the expected phenotype could not have been different since always dominant over the severe one. 11 , 21 Linear discriminant analyses (LDA) were thus computed to predict the clinical phenotype (cPKU, mPKU, and mHPA) starting from individuals' GPV: GPVs <= 2.7 were considered predictive for cPKU, GPVs 2.7-6.6 for mPKU and GPVs >= 6.7 for mHPA. Thus, a PKU-variant (APV < 6.7), associated with an HPA-variant (APV >= 6.7) should predict a clinical phenotype of mildHPA (GPV >= 6.7). To ensure GPV test's ability to generate reliable results, validity index was used for each individual to compare expected phenotype (based on GPV) with presenting phenotype; patients were thus considered 'true positives' when matching and 'false positives' when differing. Possible effects of interallelic complementation, epigenetic factors and other possible elements which may influence phenotype, 20 , 21 were not considered in this study. Statistical analysis was performed and graphed using R version 4.0.3, an open-source software and flexible programming language used for the statistical data analysis as well as graphic creations. All data in this study are reported as means +- SD or numbers (%) unless otherwise indicated. Phe values are reported as means +- SD, median, maximum and minimum for each group. The validity index was calculated as the ability of the GPV-based test to generate results corresponding to the truth, thus as the percentage of patients with clinical phenotype matching the predicted phenotype, out of the total number of patients with a given predicted phenotype. Informed written consent was provided from patients to have data from their medical records used in this research and data submissions and procedures were in accordance with the ethical standards while approved by local institutional review boards, where applicable. 3 RESULTS 3.1 Demographics Eight hundred and twenty-six patients (n = 826) affected by different forms of hyperphenylalaninemia were actually in follow up at the Clinical Department of Paediatrics (ASST Santi Paolo e Carlo, San Paolo Hospital, University of Milan, Italy) at the initiation time of the study. Out of them, 814 were either PKU or HPA affected (99%), with genders nearly equally distributed and average age 18.75 +- 12.63 years, while the remaining part was constituted of nine patients affected by PTPS deficiency, one patient affected by DHPR deficiency, and two patients affected by PCD deficiency. Distribution of patients and demographics characteristics are summarized in Figure 1. Ethnicity was predominantly Caucasian, for both populations. FIGURE 1 Distribution of patients based on type of hyperphenylalaninemia and demographics characteristics. Distribution of patients affected by any type of HPA and in follow up at the Clinical Department of Paediatrics, ASST Santi Paolo e Carlo, San Paolo Hospital, University of Milan, Italy, along with indications about gender and age distribution. Abbreviations: HPAs = hyperphenylalaninemias; PTPS = 6-Pyruvoyl-Tetrahydropterin Synthase Deficiency; DHPR = Dihydropteridine Reductase Deficiency; PCD = Pterin-4 alpha-carbinolamine dehydratase (PCD) deficiency. 3.2 Molecular results 3.2.1 PAH gene Ninety three percent of patients (n = 758) in the evaluated cohort are affected either by phenylketonuria or hyperphenylalaninemia, for a total of 1519 PAH genotype records retrieved and 166 different genetic variants highlighted, distributed in all 13 exons and 3'UTR region. p.Ala403Val is the most frequently observed variant, present in 149 alleles (9.81%) out of the total genetic records collected. Other very frequently observed variants included: p.Arg261Gln in 126 alleles (8.29%); p.Val245Ala in 121 alleles (7.96%); IVS10-11 g > a in 103 alleles (6.78%); p.Tyr414Cys and p.Leu48Ser in 84 alleles (5.53%) (TableS1). In a minor percentage of cases, variants in PAH seem absent (4%, n = 31) or appear to identify only heterozygosity (3%, n = 25), thus are not yet fully defined on a genetic basis in patients that still are clinically affected. Especially in such cases, MLPA copy number analysis technique resulted preferably used to elaborate on the case and in most cases concluding identifying deletions. As an example, the case of a particular deletion affecting exon 4 and initially escaping because under the primer and visible only in the forward sequence, thus not with the reverse primer. This was not in fact a true deletion, rather a mutation that fell into the splicing site just below the MLPA probe which, for this reason, could not be linked. After carrying out both PCR and sequencing according to Sanger of exon 4, the c.441 + 5G > T (IVS4 + 5 g > t) variant could be found (a well-known and pathogenetic variant whose identification allowed to conclude the case as a double mutate with an already found mutation found in previous analysis). Seven novel variants in seven different patients were identified (see Table 1), not found in gnomAD browser (minor allele frequency MAF =0): (1) c.272 T > C (p.Leu91Pro), (2) c.569 T > A (p.Val190Glu), (3) c.947 T > A (p.Glu316Val), (4) c.978G > C (p.Trp326Cys), (5) c.1075 T > C (p.Ser359Pro), (6) c.583A > G (p.Lys195Glu), except for (7) c.1013A > C (p.Asp338Ala) (MAF = 0.000003981). All variants are predicted as pathogenic/likely pathogenic in ACMG classification/Varsome. In all cases, the second associated variant was already known and described. Each of these patients was either diagnosed with HPA (n = 3) or PKU (n = 4) based on Phe levels. TABLE 1 New PAH gene variants identified in our cohort of patients. Specifications about newly identified variants (on both alleles) and related effects in patients affected by HPAs in current follow-up at the Clinical Department of Paediatrics, San Paolo Hospital, ASST Santi Paolo e Carlo, University of Milan, Italy. Allele 1 variant Allele 1 effect Allele 2 variant (Effect) Predicted effect (ACMG classification/varsome) 1 c.272 T > C p.Leu91Pro c.688G > A (p.Val230Ile) pathogenic 2 c.569 T > A p.Val190Glu nt45delCT (p.Ser16Ter) pathogenic 3 c.947 T > A p.Glu316Val c.782G > A (p.Arg261Gln) pathogenic 4 c.978G > C p.Trp326Cys c.1222C > T (p.Arg408Trp) likely pathogenic 5 c.1075 T > C p.Ser359Pro c.1222C > T (p.Arg408Trp) likely pathogenic 6 c.583A > G p.Lys195Glu c.1157A > G (p.Tyr386Cys) pathogenic 7 c.1013A > C p.Asp338Ala c.842C > T (p.Pro281Leu) likely pathogenic The c.272 T > C (p.Leu91Pro) variant was found in one patient in association with a second variant (p.Val230Ile) with an APV = 10, thus cannot predict alone the expected phenotype. This patient has pre-treatment Phe levels only mildly elevated (199 mmol/L) and is currently not requiring dietary treatment. This variant is predicted as 'pathogenic' in ACMG classification/Varsome. The c.569 T > A (p.Val190Glu) variant was found in one patient in association with a second variant (p.Ser16Ter). This patient has pre-treatment Phe levels considerably elevated (2578 mmol/L) and is currently on dietary regimen with a Phe tolerance of 283 mg Phe/day up to date (3 years of age). This new variant is predicted as 'pathogenic' in ACMG classification/Varsome. As the second variant in this case has an APV = 0, p.Val190Glu may be suggestive for cPKU. The c.947 T > A (p.Glu316Val) variant was found in one patient in association with a second variant c.782G > A (p.Arg261Gln). In this patient, pre-treatment Phe level were only slightly elevated (164 mmol/L) and a dietary intervention has never been required up to date. This variant is predicted as 'pathogenic' in ACMG classification/Varsome. Since the second variant has an APV = 1.6, p.Glu316Val may be predictive for mildHPA. The c.978G > C (p.Trp326Cys) variant was found in association with a second variant c.1222C > T (p.Arg408Trp) in one patient that showed high pre-treatment Phe levels (942 mmol/L) and is currently undergoing dietary treatment, with a Phe tolerance of 340 mg Phe/day up to date (3 years of age). This variant is predicted as 'likely pathogenic' in ACMG classification/Varsome. Since the second variant in this case is linked to an APV = 0, it can be expected that this new variant may be linked to mild PKU. The c.1075 T > C (p.Ser359Pro) variant was found in a patient that demonstrated very high pre-treatment Phe levels (2488 mmol/L) and is currently on dietary treatment (Phe tolerance = 250 mg Phe/day at 3 years of age). This variant is predicted as 'likely pathogenic' in ACMG classification/Varsome. In this patient, the second found variant is c.1222C > T (p.Arg408Trp), which is linked to an APV = 0 thus this new variant may be indicative of cPKU. The c.583A > G (p.Lys195Glu) variant was found in one patient with high pre-treatment Phe levels (596 mmol/L) in association with a second variant c.1157A > G (p.Tyr386Cys). This patient is currently on dietary treatment with a Phe tolerance of 550 mg Phe/day at 3 years of age. p.Lys195Glu is predicted as 'pathogenic' in ACMG classification/Varsome. Since the second variant is linked to an APV = 0, it is possible that this new variant is predictive of mildHPA. The c.1013A > C (p.Asp338Ala) variant was found in one patient with only slightly elevated pre-treatment Phe levels (143 mmol/L) and without the need for any type of medical intervention. This new variant is predicted as 'pathogenic' in ACMG classification/Varsome. The second variant in this case is c.842C > T (p.Pro281Leu), which is linked to an APV = 0 thus this new variant may be indicative for mildHPA. All novel variants with predicted effects explanation are presented in Table 1. For all these novel variants, as they were found in compound heterozygosis with other common variants reported as pathogenic in ClinVar and related to clinically affected patients, pathogenicity is strongly suggested. 3.2.2 PAH gene associated phenotypes In this study, APV and GPV could be calculated for a total of 748 of patients (92% of sample size) differentiating our cohort in classicPKU (cPKU, n = 223, for GPV <= 2.7), mildPKU (mPKU, n = 89, for GPV 2.7 - 6.7) and mildHPA (mHPA, n = 436, for GPV >= 6.7) based on expected phenotype, as represented in Figure 2. FIGURE 2 Distribution of according to GPV. Distribution of patients is represented with data expressed as absolute values, based on identified GPV value and classified in different forms of hyperphenylalaninemia (mild PKU, hyperphenylalaninemia = HPA and classic PKU). Calculation of expected phenotype was obtained from the site www.biopku.org. Predictions were reliable with the presenting phenotype only in 85% of cases (n = 638) according to Blau et al.'s classification, 10 with an identified relevant discrepancy between expected results and actual ones . FIGURE 3 Discrepancies in 'expected' vs. 'presenting' phenotype based on type of HPAs. Evidence of relevant discrepancies between type of HPA prediction and reality: on the left, the distribution of the expected phenotypes according to the GPV; on the right, the distribution of actual presenting phenotype based on pre-treatment Phe levels, divided according to Blau's classification. Specifically: 1. GPV <= 2.7 (expected phenotype = cPKU). In this category, phenotype could be reliably predicted in 76% of cases (n = 169) infact presenting with a cPKU form (pre-treatment blood Phe > 1200 mmol/L). In the remaining 24% of cases (n = 54), GPV was not reliable: 17% of cases (n = 38) presented with a mildPKU form (pre-treatment blood Phe 600-1200 mmol/L) and 7% (n = 16) with a mildHPA one (pre-treatment blood Phe < 600 mmol/L). For those presenting with a mildPKU form, a dietary treatment was needed, as expected based on pre-treatment levels (mean blood-Phe levels 745.45 +- 112.08). Besides that, this was not as restricted as expected for a classic form in terms of tolerance. Particularly, for one patient, found homozygous for p.Ala104Asp, a dietary treatment was needed but related dietary tolerance was 650 mg Phe/day at 6 years, with corresponding mean blood Phe levels demonstrating good metabolic control (<360 mmol/L). In 2 other patients, both homozygous for p.Leu48Ser, mean blood Phe levels were 854.97 +- 126.13 mmol/L and, for one of them, pharmacological treatment with sapropterin dihydrochloride could be unexpectedly applied resulting in optimal metabolic control (<360 mmol/L) and a dietary tolerance of 900 mg Phe/day to date. With regards to patients presenting instead with a mild HPA form, 11 patients had pre-treatment Phe levels of 360-600 mmol/L and 5 patients <360 mmol/L (mean 167.92 mmol/L, min 140.20 mmol/L max 215.50 mmol/L); of the latter, no patients required dietary intervention at least up to date. Following variants were identified in these patients: compound heterozygosity (p.Arg261Gln;p.Tyr386Cys) for 2 patients, compound heterozygosity (p.Arg408Trp; IVS10-11 g > a) for 1 patient and homozygosity (p.Leu48Ser) for the remaining 2 patients. All these variants are reported in literature as normally linked to a cPKU phenotype. p.Leu48Ser variant was frequently observed in cases with phenotypes differing from those expected; in particular, this occurred in 40% of the cases of homozygosity for p.Leu48Ser, suggesting uncertainty for its definition as 'classic' or else a possible variable expressiveness. 2. GPV 2.7-6.7 (expected phenotype = mild PKU). In this category, phenotype could be reliably predicted only in a minor percentage of patients (48%, n = 43). In the remaining 52% of patients (n = 46), 7% (n = 6) unexpectedly presented with a cPKU form and 45% (n = 40) with a mild HPA. Focusing on patients that presented with a mildHPA form, pre-treatment blood Phe levels were 360-600 mmol/L in 31 patients, while <360 mmol/L in the remaining nine patients. Five of the latter never required any kind of intervention (mean blood Phe levels 212.3 mmol/L; min 144.50 mmol/L, max 262.00 mmol/L) and the following genetic variants were found: patient (1) p.Pro119Ser in homozygosity, patient (2) p.Ala395Gly in homozygosity, patient (3) compound heterozygosity for p.Pro281Leu and p.Lys396Arg, patient (4) compound heterozygosity for p.Arg158Trp and Ala309Val, patient (5) compound heterozygosity for p.Tyr414Cys and p.Leu41Phe. With the aim to best address genetic insights, we deepened the analysis with regards to variants found in homozygosity, for whom a clear connection with presenting phenotype was more applicable. p.Pro119Ser and p.Ala395Gly, when in homozygosity, presented with pre-treatment blood-Phe levels of 144 and 192 mmol/L, respectively, still remaining stable without need of dietary intervention. The p.Tyr414Cys variant was also frequently observed in cases with phenotypes differing from those expected (mildHPA instead of mild PKU); in particular, this occurred in 64% of the cases of homozygosity for p.Tyr414Cys, suggesting uncertainty for its definition as 'mild PKU' or else a possible variable expressiveness. With regards to patients expected to carry a mildPKU form and instead presenting with a cPKU one, p.Leu333Phe was found in 2 siblings, both homozygous, with pre-treatment blood-Phe levels of 1612.00 +- 520 mmol/L and requiring dietary intervention. 3. GPV >= 6.7 (expected phenotype = mild HPA). In this category, phenotype could be reliably predicted in 98% of cases; remaining part was 8 patients with a mild PKU phenotype and 2 patients with cPKU. p.Glu390Gly was observed in all cases presenting with a mild PKU phenotype, variably associated with other variants all reported as linked to classic forms. This observation suggest uncertainty for the definition of p.Glu390Gly as a 'mild HPA' variant or else a possible variable expressiveness. With regards to the two latter patients reported with a unexpected cPKU phenotype, these were both females with compound heterozygosity (patient 1: p.Arg158Gln and p.Ala300Ser; patient 2: p.Arg261Gln and p.Ala300Ser); pre-treatment blood Phe levels were 1368 and 1400 mmol/L, respectively. The first subject is confirming her clinical phenotype overtime and is on dietary treatment still. Her dietary Phe tolerance was 270 mg/day at 1 year of life and reached 350 mg Phe/day at 6 years of life; as expected based on blood pre-treatment Phe levels and BH4 loading test performed, no responsiveness to sapropterin could be demonstrated though both variants are currently described as BH4-responsive in literature. 22 The second subject, on the contrary, despite high pre-treatment blood Phe levels but carrying possibly responsive variants, was confirmed responsive at the BH4 loading test with a much higher Phe tolerance than expected, increasing overtime (400 mg Phe/day at 1 year of life reaching 1200 mg Phe/day at 3 years old, when put on sapropterin); still, a liberalization of the diet has not been gained. Despite a highly positive predictive value for GPV >= 6.7, it has to be considered that this specific category includes, based on Blau's classification, 10 both patients who need some type of treatment and patients who do not, and that even if high GPV seem to be very predictive for a mild HPA form, it cannot discriminate alone who is going to need treatment and who's not. As this can result in possible relevant difficulties managing the patient at the newborn screening, we felt that this population could be better assessed in defining prediction discrepancies based on a different classification, thus dividing patients into those 'not requiring treatment' (90%, n = 394) and those 'requiring treatment' (10%, n = 42), in accordance with European guidelines. 18 When divided as illustrated, it became even more evident how much GPV is reliable defining mild forms but that, in numerical terms, this category of patients has decidedly higher possibilities of not needing diet therapy than needing it. We then tried to identify possible GPV cut offs that could help differentiate among the two differently identified subgroups. For this purpose, we tried to apply the use of a ROC curve for the identification of a GPV cut-off value capable of discriminating between patients requiring treatment and not but failed. It was in fact not possible to identify this specific threshold value, as the only effective identified cut off (calculated on a Youden's Index basis 23 ) was associated to a GPV < 9, thus not useful on a clinical standpoint (the aim was to help discriminate among patient with GPV >= 6.7 in a more strict way). Even when we tried to construct GPV percentiles based on phenotype (dividing into the three different identified categories), it was not possible to delimit the various categories with a 95% confidence interval. This made it impossible to identify with a certainty range the delimitation, therefore the prediction of the clinical phenotype only according to the GPV. Aiming to add more insights on this specific subgroup of patients, we can state that mean Phe concentration among patients with GPV >= 6.7 was 241.46 +- 136.25 mmol/L (min 102.8; max 1014.0). Dividing patients based on Phe plasma levels (cut-off 360 mmol/L) we could find that for patients 'requiring treatment' (10%, n = 42), mean Phe levels were 513.85 +- 145.23 mmol/L and predominant genotypes were: Y414C (allele frequency 49.12%), L48S (allele frequency 38.6%), R261Q (allele frequency 10.53%), R408W (allele frequency 7.02%), R158Q (allele frequency 5.26%) and IVS12 + 1 g- > a (allele frequency 5.26%). With regards to patients 'not requiring treatment' (90%, n = 394), mean Phe levels were 195.80 +- 60.20 mmol/L and predominant genotypes were: A403V (allele frequency 36.47%), V245A (allele frequency 31.17%), A300S (allele frequency 13.52%), R261Q (allele frequency 10%) and Y414C (allele frequency 7.12%). 3.3 Other molecular results 3.3.1 >BH4 deficiencies associated genes Out of the total of patients reviewed over a time frame of 65 years, 12 are carrying mutations in genes involved in some type of BH4 deficiency (1% of patients, which is in line with reported literature where tetrahydrobiopterin deficiencies account for 1%-3% of all cases of elevated phenylalanine levels). In these cases, a different diagnosis was expected based on previously collected pterins results. Patients were diagnosed as follows: nine patients affected by PTPS deficiency (and identified PTS variants), two patients affected by Pterin-four alpha-carbinolamine dehydratase deficiency (and identified PCBD1 variants) and one patient affected by DHPR deficiency (and identified QDPR variants). There were no patients tested for GCH1 gene mutations, as none showed compatible alterations on a pterins level. Variants identified in our cohort of BH4 deficient patients, along with details regarding values presented at screening (both in terms of phenylalanine and pterins), are indicated in Table 2 and have already been described in current literature. 24 , 25 TABLE 2 Variants identified in patients with BH4 deficiencies in our cohort. Variants identified in patients affected by BH4 deficiency in our cohort, including patients affected by PCD, DHPR or PTPS deficiencies. Patient # Enzyme defect (gene) Allele 1 variant (Effect) Allele 2 variant (Effect) 1 DHPR (QDPR) c. 529A > G (p. Tyr150Cys) c. 741C > T (p.Arg221Ter) 2 PTPS (PTS) c.393del (p.Lys131fs*64) g.3760_3816del (IVS2-762-718del) 3 PTPS (PTS) c.393del (p.Lys131fs*64) g.3760_3816del (IVS2-762-718del) 4 PTPS (PTS) c.260C > T (p.Pro87Leu) c.308 T > C (p.Val103Ala) 5 PTPS (PTS) c.260C > T (p.Pro87Leu) c.308 T > C (p.Val103Ala) 6 PTPS (PTS) c.317C > T (p.Thr106Met) c.308 T > C (p.Val103Ala) 7 PTPS (PTS) c.240G > T (p.Met80Ile) c.164-37dup (IVS2-37dup) 8 PTPS (PTS) c.308 T > C (p.Val103Ala) c.379C > G (p.Leu127Val) 9 PTPS (PTS) c.370G > T (p.Val124Leu) c.164-1G > C (IVS2-1 g > c) 10 PTPS (PTS) c.146A > G (p.His49Arg) c.244-8G > C (IVS4-8 g > c) 11 PCD (PCBD1) c.172G > A (p.Glu58Lys) c.179 T > C (p.Leu60Pro) 12 PCD (PCBD1) c.172G > A (p.Glu58Lys) c.179 T > C (p.Leu60Pro) 3.4 >DNAJC12 gene DNAJC12 analysis resulted as performed for those cases were copy number deletion/duplication analysis nor NGS sequencing could lead to conclusive results despite evidence of higher-than-normal Phe levels. None of our patients re-tested to date did highlight any variant in DNAJC12. 4 DISCUSSION This study aims to investigate the genotypic characteristics of the population of patients affected by hyperphenylalaninemia in follow-up at the reference clinical centre of the Lombardy region, Italy. This cohort of patients appears to be one of the numerically largest in our nation as well as in Europe, therefore it constitutes a representative sample of the PKU affected population worldwide. Eight hundred and twenty-six patients affected by different forms of hyperphenylalaninemia are in follow up at the Clinical Department of Paediatrics (ASST Santi Paolo e Carlo, San Paolo Hospital, University of Milan, Italy) and included in this study. Eight hundred fourteen are either PKU or HPA affected (49% of patients currently on dietary treatment, thus PKU), while the remaining part includes patients affected by various types of BH4 deficiency (nine patients affected by PTPS deficiency, one patient affected by DHPR deficiency and two patients affected by PCD deficiency). With regards to PAH genotype, 166 different genetic variants have been identified in this study. p.Ala403Val is the most frequent, found in 149 alleles (9.81%) out of the total of genetic records collected; this result is in line with what already found in current literature with regards to the Italian population. 25 , 26 Other very frequently observed variants includes: p.Arg261Gln in 126 alleles (8.29%); p.Val245Ala in 121 alleles (7.96%); IVS10-11 g > a in 103 alleles (6.78%); p.Tyr414Cys and p.Leu48Ser in 84 alleles (5.53%). These latter results are in contrast with actual known genotype distribution in European PKU patients, among whom the most common variant, as reported in a recent study by Hillert et al., 5 is p.Arg408Trp, with an allelic frequency of about 64%, followed by c.106611G > A and p.Arg261Gln. Patients are compound heterozygotes in 82% of cases, homozygotes in 13% and simple heterozygotes in the remaining 5% of cases. Seven novel variants have been also identified in our cohort. The identification of these new variants is relevant for the scientific field, as they can be included in current used database and improve and increase knowledge about patients affected by hyperphenylalaninemia. In our study, we tried to predict phenotype based on the Allelic Phenotype Values and Genotypic Phenotype Values (APVs-GPVs system). 10 , 11 According to Blau et al.'s classification, actual cases were matching with related prediction in most cases, but not all of them (85% of cases, n = 638), identifying a relevant discrepancy between expected results and presenting ones. Aiming to identify possible explanations, we proceeded evaluating patients after dividing them in different cohorts based on GPV, further clarifying that the APVs-GPVs system may be helpful for clinical purposes but it may not be as representative as expected to when taken alone. We could actually identify, for example, patients with very low GPVs (<2.7) expected to be very severe clinically, instead presenting with low blood-Phe levels that did not require dietary intervention. There were some variants more associated with unexpected results. This included p.Leu48Ser. If imagined as in a functional hemizygote, this variant would be assigned to a phenotype of classic phenylketonuria (GPV < 2.7) and would seem to have a residual enzymatic activity of almost zero if associated with a 'null allele'. Besides that, among our population this variant was found in two patients with mild PKU and in two patients with mild HPA, probably indicating a higher residual enzymatic activity than expected. Again, other variants were frequently observed in cases with phenotypes other than those expected. This is the case of the p.Tyr414Cys variant, that presented with a mildHPA phenotype instead of mild PKU, and the p.Glu390Gly variant, that emerged variably associated with other mutations defined as mild HPA based on GPV but instead presenting with a mild PKU form based on pre-treatment Phe levels. Such observations suggest uncertainty for a strict definition of these specific variants as done to date and suggest on the counterpart variable expressiveness due to possible intrinsic characteristics. We also tried to identify possible GPV cut-off values capable of ameliorate discrimination abilities in order to better predict prognosis, but this was not possible as the only found threshold was associated to a GPV not useful on a clinical standpoint. These limits said, we can only speculate that, considering GPVs, patients with pre-treatment blood Phe levels < 360 mmol/L had higher GPV value (mean 9.7, range 7-10) than patients with pre-treatment blood Phe levels >=360 mmol/L (mean 8.9, range 7-10). We can add that when presenting phenotypes are differing from expected ones, this can also be due to the possibility that some mutations may be in deep intronic regions or in sequences that are difficult to study with the actual available means, thus the direct study of the RNA by looking for alternative splicing or considering interactions with a specific noncoding RNA should be considered; alternatively, the 'switching off' of one of the two alleles by transcriptional regulatory mechanisms may be supposed. This may be considered also for such patients where only one variant could be identified and yet are presenting with severe phenotypes. Furthermore, positive or negative interallelic complementation or epigenetic factors should be taken into account, such as non-genetic factors that affects phenotypic expression that may also have come into play, although, to date, they are still yet to be fully known. This includes considerations about recent works that stress about mechanistic heterogeneity of PAH dysfunction as a key factor when dealing with phenotype expression. For example, the sequence/structure predictive methods rely too heavily on the one sequence-one structure-one function assumption that lies at the basis for much bioinformatics analysis, explaining the 'why' of the frequent failure of current predictions. 27 Furthermore, PAH structure is multimeric, usually a tetramer, with different mole fractions of the different PAH variants and even possible regulation processes for protein degradation, 28 , 29 , 30 thus some predictive methods may ignore such tertiary and quaternary possible interactions. 31 , 32 , 33 Mechanism is also linked to PAH allostery, which involves high levels of Phe causing an activation associated with major conformational changes, thus allosteric failure may explain why PKU-associated variations can occur throughout the PAH sequence. 34 With regards to the latter, our paper may already even more confirm this assumption, as it identified p.Leu48Ser (based in the region of the protein that repositions and forms the binding site for allosteric Phe) and p.Tyr414Cys (whose environment is predicted to change in the transition from the auto-inhibited resting state to the activated state) as having clinical outcomes poorly predictable, which can be thus related to the structural basis of PAH allosteric activation. Taken all together, these limitations of yet to be fully known underlying mechanisms should be accounted when considering possible genotype-phenotype correlations. Among our population there are still some patients currently on dietary treatment but lacking to identify any kind of mutation within analysed genes, including DNAJC12. These patients are expected to be carrying causal variants not identified with the current sequencing technologies or, alternatively, to be carrying mutations in other genes yet to be identified. 5 CONCLUSIONS Hyperphenylalaninemia is a very complex family of metabolic disorders, characterized by a multitude of genetic variations. This paper summarizes past 65 years genetic analysis of 826 patients with HPA/PKU from a single metabolic centre in Lombardy, Italy, identifying 7 new variants in the PAH gene and comparing metabolic phenotype with the APV genotype-prediction. Many differences and discrepancies were pointed out in different categories, addressing the value and limitations of a variety of predictive tools that are currently used to direct clinicians in the management of such patients. GPV, although capable of defining predicted phenotype with reasonable appropriateness based on its numerical value, seems not sufficient to characterize the patient with certainty into a defined category and, taken alone, it may not be as representative as expected to. Even within the same GPV category in fact, there are patients who differ significantly one another regardless of the same calculated GPV. This could be connected to the presence of some specific mutations, recurring with relation to the category, which could result in a decisive alteration of the phenotype, thus in need of much more clinical attention when occurring. Other variants in the same patient, not identifiable with current available systems, should also be accounted as possible responsible for mismatching. Obviously, the deepening of knowledge in this sense could result in further improvements as well as making significant changes in the approach to the patient to better define his subsequent treatment needs. Based on our results, we can conclude that no current predictive method is appropriate in more than 50% of the patients, thus there is still a growing need to identify other elements that could help redefine and better characterize such cases for conclusive diagnostic/therapeutic approaches. Further studies are needed to improve and collect increasing evidence regarding this topic. AUTHOR CONTRIBUTIONS Conceptualization: Valentina Rovelli, Davide Bassi and Juri Zuvadelli; Data curation: Juri Zuvadelli, Alice Re Dionigi, Daniela Graziani, Olivia Turri, Luisella Alberti and Valentina Rovelli; Formal analysis: Vittoria Ercoli, Daniela Graziani and Juri Zuvadelli; Investigation: Valentina Rovelli, Juri Zuvadelli, Daniela Graziani and Vittoria Ercoli; Methodology: Vittoria Ercoli and Daniela Graziani; Project administration: Valentina Rovelli; Supervision: Valentina Rovelli and Giuseppe Banderali; Validation: Valentina Rovelli, Vittoria Ercoli and Juri Zuvadelli; Visualization: Valentina Rovelli, Vittoria Ercoli and Juri Zuvadelli; Writing - original draft: Valentina Rovelli, Davide Bassi, Daniela Graziani, Olivia Turri; Writing - review & editing: Sabrina Paci, Elisabetta Salvatici, Raed Selmi, Graziella Cefalo, Alice Re Dionigi, Valentina Rovelli and Giuseppe Banderali. FUNDING INFORMATION This research received no external funding. CONFLICT OF INTEREST The authors declare no conflict of interest. ETHICAL APROVAL The study was conducted according to the guidelines of the Declaration of Helsinki and with D.L. 196/2003 and the guidelines for the processing of personal data of our clinical institution. Supporting information Table S1 Click here for additional data file. DATA AVAILABILITY STATEMENT The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request. |
PMC10000624 | Introduction Hyperglycaemic hyperosmolar state (HHS) is a known complication of type 2 diabetes mellitus; however, carbonated carbohydrate fluid intake may precipitate a more severe presentation of type 1 diabetes mellitus with hyperosmolar state. The management of these patients is not easy and can lead to severe complications such as cerebral venous thrombosis. Methods We present the case of a 21-month-old boy admitted for consciousness disorders revealing a hyperglycaemic hyperosmolar state on a new-onset type 1 diabetes and who developed cerebral venous thrombosis. Results and Conclusion Emergency physicians should be aware of HHS in order to start the appropriate treatment as early as possible and to monitor the potential associated acute complications. This case highlights the importance of decreasing very gradually the osmolarity in order to avoid cerebral complications. Cerebral venous thrombosis in HHS paediatric patients is rarely described, and it is important to recognize that not all episodes of acute neurological deterioration in HHS or diabetic ketoacidosis are caused by cerebral oedema. Emergency physicians should be aware of HHS in order to start the appropriate treatment as soon as possible and to think about its acute complication. This case highlights the importance to restore euvolemia before insulin treatment and decrease very gradually the osmolarity in order to avoid cerebral complication. Cerebral venous thrombosis in HHS paediatric patients is rarely described, and it is important to recognize that not all episodes of acute neurological deterioration in HHS or DKA are caused by cerebral oedema. case report cerebral thrombophlebitis diabetes hyperglycaemic hyperosmolar state source-schema-version-number2.0 cover-dateMarch 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.2.6 mode:remove_FC converted:10.03.2023 Injeyan M , Baron S , Lauzier B , Gaillard-Le Roux B , Denis M . Hyperglycaemic hyperosmolar state and cerebral thrombophlebitis in paediatrics: A case report. Endocrinol Diab Metab. 2023;6 :e389. doi:10.1002/edm2.389 pmc1 BACKGROUND Diabetic ketoacidosis (DKA) and hyperglycaemic hyperosmolar state (HHS) are acute metabolic complications that occur in children with type 1 diabetes mellitus. Diabetic ketoacidosis is the most common hyperglycaemic crisis 1 but HHS is associated with higher mortality 2 , 3 , 4 and its incidence is increasing. 5 , 6 , 7 HHS is a known complication of type 2 diabetes mellitus 8 ; however, carbonated carbohydrate fluid intake may precipitate a more severe presentation of type 1 diabetes mellitus with hyperosmolar state. 4 , 9 , 10 The main objective is to correct electrolyte disorders and avoid neurological complications. The common cerebral complication associated with HHS is cerebral oedema. While the risk of developing peripherical venous thromboembolism has already been described 11 , 12 , 13 in DKA or HHS, cerebral venous thrombosis has been rarely described. 14 , 15 We present the case of a 21-month-old boy admitted for consciousness disorders revealing a hyperglycaemic hyperosmolar state for a new-onset type 1 diabetes and who developed cerebral venous thrombophlebitis (CVT). 2 CASE PRESENTATION A 21-month-old boy, weight 11 kg, presented at the emergency room with an unconscious and hypotonic state. He was born at 39 weeks of gestational age, with a birth weight of 3440 g and birth length of 51 cm. There was no family history of diabetes mellitus. He was previously healthy, presenting only a slight delay in psychomotor development as he could not sit up at 9 months nor stand alone at 21 months. He presented a rhinopharyngitis treated with amoxicillin one week before the hospitalization. He had a 3-day history of polyuria and lethargy and a great alteration of the general condition for 2 days. His parents hydrated him only with sugar water. On the morning of the third day, he was found to be unresponsive in his bed. Upon admission, he presented normal temperature (36.8degC), tachycardia (heart rate: 150 bpm), hypotension (blood pressure: 61/37 mm Hg), cold extremities, tachypnoea (respiratory rate: 40 cycles/min, SaO2: 100%) and dehydration signs objectified by dried mucous membranes. He presented unconscious (Glasgow coma scale [GCS]: 5) with no signs of localization and his pupils were in reactive mydriasis. Rapid glucose test was higher than 33 mmol/L. Blood tests reported hyperglycaemia (138 mmol/L), metabolic acidosis (pH 7.30, bicarbonate 16 mmol/L), hyperlactatemia (5.20 mmol/L), ketosis (0.9 mmol/L) and ketonuria. Electrolyte test revealed hypernatremia (sodium chloride 128 mmol/L and corrected sodium chloride 167 mmol/L) and hyperkalaemia (7.9 mmol/L). Creatinine and blood urea nitrogen were 159 mmol/L and 39 mmol/L, respectively, and serum osmolarity was 511 mOsm/L. Prothrombin rate and fibrinogen were 80% and 1.6 g/L, respectively. The electrocardiogram showed a regular sinus rhythm without any conduction disorder but hyperkalaemia signs including very large and sharp T waves and a diffuse ST shift. An injected brain computed tomography scan (CT-scan) was performed and showed a permeable veinous sinus and a hypodensity of the white matter in front of the temporal horn and the right ventricular carrefour. The patient received a volume expansion of 250 ml of isotonic fluid NaCl 0.9% followed by a rehydration with 110 ml/kg/day of NaCl 0.9%. Hyperkalaemia was treated by calcium gluconate 10% (4 ml), Terburtaline aerosol (1 aerosol of 2.5 mg), and sodium bicarbonate 8.4% (10 ml). His mental status required intubation under sedation by 2 mg/kg of ketamine and 1 mg/kg of rocuronium. He was then referred to the paediatric intensive care unit (PICU) and regular insulin (Insulin analogue [Lispro] 0.02 UI/kg/h) was started. During PICU hospitalization, volume expansion was continued intravenous by 150 ml of NaCl 0.9% and 50 ml of ringer lactate and maintain intravenous fluid was changed for Glucose 5% with NaCl 20% 4 g/L and KCl 10% 3 g/L. Laboratory examinations were performed in order to check and modulate electrolyte changes . The serum sodium concentration and the serum osmolarity increased to a maximum of 167 mmol/L (corrected natremia) and to 511 mOsmol/L, respectively. Serum glucose level decreased and normalized to 7.6 mmol/L in 24 h, which led to a gradual decreased in serum osmolarity and sodium. This corresponded to an average drop of 1 mmol/L/h of sodium in the first 24 h, with a maximum of 3.5 mmol/L/h between H10 and H12, and of 7.37 mOsm/L/h of osmolarity with a maximum of 13.5 mOsm/L/h between H10 and H14. His hemodynamic status remained stable, without any vasopressor support requirement. The echocardiography showed normal cardiac function. The transcranial Doppler (TCD) performed did not show any anomalies. The patient was sedated by morphine and midazolam which was stopped on Day 2. At the initial examination without sedation and before extubation, the child presented a rolling of the lower and upper limbs, a pyramidal syndrome including bilateral Babinsky's sign, sharp reflexes with increased reflexogenic area and a withdrawal reaction to pain. Seizures were detected on Day 2 and treated by clonazepam (0.05 mg/kg) and then levetiracetam (40 mg/kg/day). No clinical signs of intracranial hypertension (ICH) were found and TCD was normal (pulsatility index (PI): 0.9 and diastolic velocity (Vd): between 30 and 40 cm/s). Electroencephalogram showed a marked encephalopathy pattern without epileptic activity, possibly of metabolic origin and related to sedation. His brain magnetic resonance imaging (MRI) reported an extensive thrombophlebitis of the superior sagittal sinus and of several cortical veins in the vertex, the torcular, the left transverse sinus as well as presence of small thrombi in the right transverse sinus. Significant bilateral and symmetrical signal abnormalities of the fronto-parieto-temporo-occipital subcortical white matter and of the external capsules were also visible in diffusion restriction, which seemed rather to be related to metabolic disorders . He was immediately treated by anticoagulant (unfractionated heparin, 20 UI/kg/h, target antiXa: 0.3-0.5). The mental status improved and extubation was possible on Day 3. FIGURE 1 Changes in biological parameters and medical interventions over time upon admission. H, Hours FIGURE 2 Magnetic resonance imaging performed on Day 2 and at 3 months. (A) MRI performed on Day 2, extensive thrombophlebitis of the superior sagittal sinus and of several cortical veins in the vertex, the torcular, the left transverse sinus and presence of small thrombi in the right transverse sinus. Significant bilateral and symmetrical signal abnormalities of the fronto-parieto-temporo-occipital subcortical white matter and of the external capsules were also visible in diffusion restriction, which seemed rather to be related to metabolic disorders. (B) MRI at 3 months, partial permeabilization of the superior sagittal sinus with persistence of a heterogeneous aspect of the sinus, seat of some linear images probably corresponding to residual filiform thrombi, a complete permeabilization of the transverse sinuses and the torcular and cortical veins, and an appearance of FLAIR signal abnormalities of the periventricular white matter, probably sequelae of the abnormalities visualized in diffusion on the first MRI. Five days after PICU admission, the neurological examination showed no focal signs, symmetrical osteo-tendon reflexes, no Babinski's sign, no peripheral hypertonia, no epileptoid tremor and some upper limb stereotypies. The brain CT-Scan showed an absence of haemorrhagic remodelling and a stability of the extent of the thrombophlebitis reaching the superior sagittal sinus, torcular, left transverse sinus and some cortical veins. The white matter signal abnormalities described on the MRI were not visible on the CT scan. A Doppler ultrasound of the lower limbs showed a deep venous thrombosis in superficial femoral vein, to the location of the central catheter. Serology revealed positive antibody against insulin (4 unit/ml; normal high <0.4 unit/ml) and no antibody against GAD, IA2, ICA, Zn8. A central hypothyroidism was found (T4 3.7 pg/ml [9.4-18 pg/ml], TSH 0.46 mUI/ml [0.2-4 mUI/ml]). We did not find any abnormality on the other pituitary axes except a low IGF1 level (27 ng/ml [49-297 ng/ml]), controlled at 38 ng/ml two days later, which may be related to dehydration and acute malnutrition. The infant was transferred to the general paediatric ward on Day 5 and discharged home on Day 26. Initially, he was treated with a regimen of basal and rapid insulin and then transitioned to an insulin pump. L-Thyroxine treatment was stopped at 1 month with normalization of the thyroid balance (TSH 0.21 mUI/ml; T4 10.6 pg/ml); without any antibodies stating a transient acute central hypothyroidism; and diabetes was well balanced. Chromosomal analysis by CGH-Array was normal and the search for fragile X syndrome was negative. The metabolic workup was unremarkable (mucopolysaccharides, urinary oligosaccharides, plasma lysosphingolipids and Gaucher disease). In addition, the mitochondrial DNA study was normal. He was treated with anticoagulants (antivitamin K: coumadin 2 mg per day) for 3 months. The brain MRI performed at 3 months showed partial repermeabilization of the superior sagittal sinus with persistence of a heterogeneous aspect of the sinus, seat of some linear images probably corresponding to residual filiform thrombi, a complete repermeabilization of the transverse sinuses and the torcular and cortical veins, and an appearance of FLAIR signal abnormalities of the periventricular white matter, probably sequelae of the abnormalities visualized in diffusion on the first MRI . The neurological examination returned to baseline, and he is followed up in a specialized neurodevelopment centre. 3 DISCUSSION We report the first case of HHS revealing a new-onset type 1 diabetes complicated by cerebral thrombophlebitis in a young boy. In France, diabetes mellitus incidence increases by 4% each year since 1988. 16 HHS remained a rare diagnosis in paediatrics but hyperosmolar events had higher rates of complications. 4 In this case, the little boy was thirsty, and his parents gave him only sugar water. That probably contributed to an increase in blood sugar and the hyperosmolar state. Several case reports have shown that carbonated carbohydrate fluid intake may precipitate a more severe presentation of type 1 diabetes mellitus in adolescents, 10 but this presentation is rarely described in children. The mortality rate in patient with HHS varies between 20% and 60% and timely diagnosis as appropriate treatment prevent complications and death. 2 , 3 , 17 The child presented with hyperglycaemia without acidosis, which could have led to a delayed diagnosis of diabetes 1. Contrary to diabetic ketoacidosis, HHS is characterized by development of a severe hyperglycaemia without acidosis, these patients maintain a low insulin production leading to the absence or underproduction of ketones via lipolysis. The pathophysiology of HHS is further compounded by a disordered renal response. HHS commonly occurs after polyuria and polydipsia, resulting in profound dehydration. This is accompanied by severe electrolyte imbalance, greater than in DKA because of the longer duration of osmotic diuresis. The hypertonicity of the hyperosmolar state preserves intravascular volume, which contributes to masking clinical signs of dehydration which can lead to a difficult and delayed diagnosis. 7 , 18 HHS is a life-threatening emergency. The treatment of HHS requires rehydration, electrolyte balance, intravenous insulin and management of complication. The patient received a volume expansion of 250 ml of isotonic fluid NaCl 0.9% followed by a rehydration with 110 ml/kg/day of NaCl 0.9% and then insulin after his transfer in PICU. The priority is to restore euvolemia before the use of insulin. Fluid replacement should be started with 0.9% saline in order to maintain circulatory volume, restore renal perfusion and gradually correct hyperosmolarity. 19 It is commonly admitted that hypotonic maintenance intravenous fluids lead to a greater risk of developing cerebral oedema in children, and thus its use should be avoided. 3 We monitored for signs of ICH through clinical examination and DTC during the first 48 h. Adequate fluid replacement must begin before insulin administration for many reasons. First, early insulin is unnecessary in HHS, because ketosis is usually minimal. 20 Secondly, insulin bolus could induce shifts of potassium to the intracellular space and hypokalaemia 20 ; move intravascular water into the cells, then exacerbating hypotension and increasing the risk of brain oedema. 19 Furthermore, fluid administration alone promotes dilution and decreases serum glucose. Finally, the osmotic pressure exerted by glucose may fall rapidly then leading to circulatory compromise and venous thrombosis. 20 HHS can cause many complications. The main risks of severe dehydration and hyperviscosity are brain oedema, rhabdomyolysis, pancreatitis, renal failure, malignant hyperthermia and thromboembolism. This state can lead to death with refractory arrhythmia or multisystem failure (cardiac, pulmonary oedema and renal failure). 21 The severity of altered mental status correlates with the level of hyperosmolarity. 3 In our case, the patient presented a severe dehydration with hyperosmolarity and hyperviscosity that altered mental status. The brain MRI and doppler ultrasound showed an extensive cerebral thrombophlebitis and thrombosis in superficial femoral vein, respectively. Adult literature reported that patients with diabetes and hyperosmolarity had a significantly higher risk of venous thromboembolism (VTE). 11 Several studies showed that young children (<3 years old) with DKA have an increased incidence of clinical VTE associated with the placement of femoral central venous catheters (CVC) 12 , 13 and these patients had significantly higher glucose and corrected sodium concentrations. 13 Cerebral venous thrombosis in the course of diabetic hyperglycaemia seems to be extremely rare. Only two reports were described in the literature. In the first one, venous thrombosis was interpreted as a consequence of the combination of dehydration and iron deficiency anaemia but no hyperosmolarity was reported. 14 The second one reported an 8-year-old male patient with CVT at early stage of management and was associated with hyperosmolarity. 15 In the context of cerebral thrombophlebitis, anticoagulation is effective in reducing the risk of death, sequelae in acute phase and of recurrence. 22 There is no recommendation for prophylactic anticoagulation for patients with HHS. Anticoagulation treatment should be reserved for children who require CVC and are immobile for more than 24 to 48 hours. 23 4 CONCLUSION HHS is an increasingly common complication of T1D but its diagnostic and management remains difficult. Emergency physicians should be aware of HHS in order to start the appropriate treatment as soon as possible and to think about its acute complication. This case highlights the importance to restore euvolemia before insulin treatment and decrease very gradually the osmolarity in order to avoid cerebral complication. Cerebral venous thrombosis in HHS paediatric patients are rarely described, and it is important to recognize that not all episodes of acute neurological deterioration in HHS or DKA are caused by cerebral oedema. Thromboprophylaxis is not recommended but early anticoagulation in case of cerebral thrombophlebitis may improve prognosis. AUTHOR CONTRIBUTIONS Maud Injeyan: Conceptualization (equal); data curation (equal); investigation (equal); resources (equal); writing - original draft (equal). Sabine Baron: Writing - review and editing (supporting). Benjamin Lauzier: review and editing (supporting); validation (supporting). Benedicte Gaillard-Le Roux: Data curation (supporting); resources (supporting); writing - original draft (supporting); writing - review and editing (supporting). Manon Denis: Conceptualization (equal); data curation (equal); supervision (lead); validation (lead); writing - original draft (lead); writing - review and editing (lead). CONFLICT OF INTEREST The authors declare that they have no competing interest to disclose. CONSENT FOR PUBLICATION Written informed consent was obtained from the patient's legal guardians for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal. DATA AVAILABILITY STATEMENT All data supporting the findings of this article are included in the article. |
PMC10000625 | Adenoid cystic carcinoma (ACC) is an aggressive malignancy that most often arises in salivary or lacrimal glands but can also occur in other tissues. We used optimized RNA-sequencing to analyze the transcriptomes of 113 ACC tumor samples from salivary gland, lacrimal gland, breast or skin. ACC tumors from different organs displayed remarkedly similar transcription profiles, and most harbored translocations in the MYB or MYBL1 genes, which encode oncogenic transcription factors that may induce dramatic genetic and epigenetic changes leading to a dominant 'ACC phenotype'. Further analysis of the 56 salivary gland ACC tumors led to the identification of three distinct groups of patients, based on gene expression profiles, including one group with worse survival. We tested whether this new cohort could be used to validate a biomarker developed previously with a different set of 68 ACC tumor samples. Indeed, a 49-gene classifier developed with the earlier cohort correctly identified 98% of the poor survival patients from the new set, and a 14-gene classifier was almost as accurate. These validated biomarkers form a platform to identify and stratify high-risk ACC patients into clinical trials of targeted therapies for sustained clinical response. oral cancer biomarker MYB oncogene transcriptome analysis bioinformatics survival analysis UNM Comprehensive Cancer CenterNIH NCI P30CA118100 the Adenoid Cystic Carcinoma Research FoundationNIH NIDCR R01DE023222 Region Zealand's Research FundThe RNA-seq analysis was partially supported by UNM Comprehensive Cancer Center Support Grant from NIH NCI P30CA118100. K.J.B. and S.A.N. were partially supported by grants from the Adenoid Cystic Carcinoma Research Foundation and from NIH NIDCR R01DE023222 (to S.A.N.). The DK studies were substantially supported by Region Zealand's Research Fund. pmc1. Introduction Adenoid cystic carcinoma (ACC) is one of the most common salivary gland malignancies, arising mainly in minor and major salivary glands, but ACC also occurs less frequently in other organs, and the clinical behavior of non-salivary ACC varies widely. This suggests that the ACC tumors arising in different organs may be biologically distinct or that they are affected by different host factors. Molecular analyses have shown that most ACC tumors have recurrent chromosomal translocations that activate the MYB oncogene or the related MYBL1 gene , resulting in characteristic gene expression changes . The translocations frequently relocate a distant, salivary gland-specific enhancer in proximity to the MYB or MYBL1 genes, leading to their overexpression . Many of the translocations occur within the MYB or MYBL1 genes, leading to truncation and overexpression of the genes and their gene products . The MYB and MYBL1 genes encode the DNA-binding transcription factor proteins Myb (c-Myb) and A-Myb, which are important for normal development . Relatively small changes in these proteins, such as truncations of the C-terminal domains, can lead to profound differences in the genes they regulate , suggesting that the proteins perform complex regulatory functions. Indeed, the Myb protein can function as a 'pioneer' transcription factor capable of initiating the formation of new enhancers that can modify the expression of distant genes . Thus, Myb or A-Myb proteins activated by C-terminal truncations may induce a specific ACC tumor phenotype , similar to the actions of Myb proteins in other types of cells and malignancies . Identifying the ACC-specific regulatory mechanisms that induce an 'ACC phenotype' could lead to new types of therapies. ACC patients often have a slow clinical course with a poor long-term prognosis . However, clinical outcomes can vary dramatically; unpredictable aggressive and progressive disease is not uncommon. Post-surgical survival ranges from just a few months to 15 years or longer. The protracted temporal progression of ACC tumors necessitates using archived samples at least 5-10 years old for studies linking genomic changes to outcomes. However, standard genomic methods are largely unsuitable for reliable RNA-sequencing (RNA-seq) analysis of archived samples, because the recovered RNA is often highly fragmented, necessitating the use of specialized approaches . Despite these complications, several studies have identified subgroups of ACC patients with distinct molecular features linked to differences in prognosis and survival , suggesting that applying these approaches to well-structured retrospective cohorts of ACC tumors could produce biomarkers for identifying poor prognosis patients and recommending them for targeted therapy. In previous studies, we were able to use optimized RNA-seq approaches to successfully analyze the transcriptomes of archived, formalin-fixed, paraffin-embedded ACC tumor samples up to 25 years old, which led to the identification of the first ACC tumors with MYBL1 translocations . Extending those studies to a larger cohort of 68 samples (the TX cohort) led to the identification of several subgroups of ACC tumors with unique gene expression signatures, including one subgroup with poor survival and a 'No Myb' group that expressed neither MYB nor MYBL1 . Although we identified gene expression signatures that correlated with poor survival, it was not possible to validate the results using only one cohort of samples. Here we describe the analysis of a new cohort of ACC tumor samples, from Denmark and Florida (the DK cohort), primarily from salivary gland but also including some ACC tumors from the lacrimal gland, breast, and skin. The availability of the large cohorts allowed us to designate a training set for defining a gene expression classifier to distinguish poor survival samples, which we validated using the second cohort. These results set the stage for using clinical RNA-seq assays for identifying patients who are likely to be in the poor survival subgroup, so they can be offered clinical trials or additional treatment to improve their outcomes. 2. Materials and Methods 2.1. Human Salivary Gland ACC Samples De-identified adenoid cystic carcinoma tumor samples were obtained from several institutions: the Department of Otorhinolaryngology and Maxillofacial Surgery, Zealand University Hospital; the Department of Otorhinolaryngology, Head and Neck Surgery and Audiology, Rigshospitalet; the Department of Pathology, Rigshospitalet, University of Copenhagen; and the Department of Ophthalmology, Rigshospitalet-Glostrup, University of Copenhagen, Copenhagen, Denmark. Some lacrimal gland samples were obtained from the Dr. Nasser Al-Rashid Orbital Vision Research Center and the Bascom Palmer Eye Institute, Department of Ophthalmology, University of Miami Miller School of Medicine. All samples were provided Formalin-Fixed and Paraffin-Embedded (FFPE) as 5-micron sections baked onto glass slides. Salivary gland samples with survival information had at least 5-year follow-up. All samples were collected in accordance with the principle of the Declaration of Helsinki and with Institutional Review Board-approved protocols: Danish Regional Ethics Committee (H-6-2014-086) and the Danish Data Protection Agency (Journal no. REG-94-2014). 2.2. RNA Isolation and Sequencing Total RNA was isolated from one or two 5-micron slide-mounted FFPE sections using the RNeasy FFPE kit (Qiagen, 19300 Germantown Rd Germantown MD 20874, USA). cDNA synthesis and library preparation were performed using the SMARTer Universal Low Input RNA Kit for Sequencing (Takara 1290 Terra Bella Avenue Mountain View, CA 94043, USA) and the Ion Plus Fragment Library Kit (ThermoFisher, 168 Third Avenue, Waltham, MA 02451, USA), as previously described . Ion Torrent sequencing using Ion S5/XL systems (ThermoFisher) was performed in the Analytical and Translational Genomics Shared Resource at the University of New Mexico Comprehensive Cancer Center. Data are available for download from the NCBI BioProject database using study accession number PRJNA287156. The TX cohort of samples has been described previously . 2.3. Data Analysis Low quality and non-human RNA-seq reads were filtered and removed using the Kraken2 suite . High-quality reads were aligned to the human genome (hg38) using TMAP (v5.10.11), and gene counts were calculated using HT-Seq, as previously described . Poor quality samples with fewer than 10% of the median number of reads of all samples were removed. Samples that failed other quality control tests were also removed. The same parameters were used when the data from the new (DK) cohort were combined with the previously described samples in the TX cohort . (Software versions are provided in File S1). 2.4. Unsupervised Hierarchical Clustering For identifying clusters, analyses were limited to genes that were expressed above a threshold level in a number of samples (e.g., 250 reads in at least 10 samples). These thresholds were reduced (e.g., to 50 reads in at least 10 samples) to generate the final heatmaps, to include as many relevant genes as possible, while retaining the clusters and the sample order. Multi-dimensional scaling was performed using plotMDS from the limma package version 3.48.0. Hierarchical clustering was performed using hclust from the stats package (R/Bioconductor version 4.1.0) . 2.5. Statistical Analysis Overall survival (OS), defined as time from the date of diagnosis to the date of death, was the primary endpoint for outcome. Subjects who were lost to follow-up or alive within the follow-up period were censored at the date of the last contact. OS was estimated using the Kaplan-Meier method. Differences in OS were examined using the log-rank test. All the statistical analyses were performed using statistical software R (version 4.1.0) . Gene Ontology analyses were performed with Bioconductor package GO.db version 3.13.0, as previously described . 2.6. Development and Evaluation of Prognostic Classifiers Personalized logistic regression with the elastic net and LASSO regulations implemented in R package glmnet was used to develop the classifiers that distinguish between the poor prognosis group and the remaining samples. ROC curve and the area under the curve (AUC) were used to evaluate the performance of the classifiers. The prognostic models were built using the training data set, i.e., the TX cohort, through a 10-fold cross-validation (CV) procedure. An unbiased estimate for each of the final models was obtained by performing a nested CV procedure that included the full cycle of the 10-fold inner loop CV followed by a 100 x 6-fold outer loop CV using the training set. The prediction accuracy of each model was further validated using an independent test set, i.e., the DK cohort. 3. Results 3.1. ACC Tumors from Different Organ Sites Share a Common Transcriptional Profile Although they most commonly occur in major and minor salivary glands, ACC tumors can also arise in lacrimal, bronchial, mammary, or skin adnexal glands. To assess the similarities and differences in gene expression patterns in ACC tumors from different tissues, we performed RNA-sequencing (RNA-seq) on a cohort of 113 ACC tumors, comprised of 17 samples from breast tissue, 24 from cutaneous tissue, 16 from lacrimal glands, and 56 from salivary glands (Table 1). Most of the samples came from Denmark (DK cohort), with the exception of 6 lacrimal gland samples from Florida (FL). We used optimized methods for RNA analysis and Ion Proton sequencing that we developed previously . For the DK cohort, 91 samples (81%) clearly expressed MYB, while 13 (12%) expressed MYBL1 and 9 (8%) expressed neither MYB nor MYBL1. We first performed Multi-Dimensional Scaling (MDS, i.e., principal component analysis) for the ACC samples. As shown in Figure 1A, the dots representing ACC tumor samples from different organs are shaded with different colors (see legend). For comparison, we included RNA-seq results from several normal salivary gland tissues (shaded black) and several from a histologically different salivary gland tumor, acinic cell carcinoma (shaded gray), that have been described previously . All the ACC tumor samples clustered into a large group at the right, suggesting that the transcription profiles of the ACC tumors are more similar to each other than they are to normal tissue or other salivary gland tumors, despite the different tissues of origin. Next, we used unsupervised hierarchical clustering to group samples that were similar and generated the heatmap comparing the transcriptional profiles shown in Figure 1B. Interestingly, as shown by the dendrogram at the top of the heatmap, the ACC tumor samples formed several large subgroups, but each cluster contained samples from all the tissue types: salivary gland, lacrimal gland, breast and cutaneous (dark blue, orange, pink and green in the color bar at top, respectively). Thus, although the ACC samples in this cohort are heterogeneous and formed distinct subgroups, the groups are not defined by the tissue of origin. Instead, the subgroups could represent biological differences amongst the ACC samples, irrespective of the tissue from which they were derived. These results differ somewhat from a previous report showing that microRNA expression profiles could distinguish ACC tumors from different tissue types , suggesting that the biological mechanisms leading to the formation of ACC tumors may have a more profound impact on the overall mRNA transcriptional profile than on microRNAs, which may be more tissue-specific. Several genes that are known to be important in ACC tumors are highlighted with black dots at right, including (from top to bottom) EN1, GABRP, MYB, MYBL1 and NFIB, all of which are expressed more highly in the ACC tumors than in the normal salivary gland or acinic cell carcinoma samples (at left). 3.2. Many ACC Tumors Fail to Express Tissue-Specific Gene Expression Markers Since our initial hierarchical clustering did not separate ACC tumors by tissue type, we reanalyzed the data to see if we could identify tissue-specific gene expression patterns in the tumors. We specifically selected genes that were differentially expressed in the ACC tumors originating in different tissues. The heatmap in Figure 2 summarizes the results when the most tissue-specific genes are chosen for display (and the normal salivary gland and acinic cell carcinoma samples are left out). This type of analysis led to better clustering of the ACC samples from lacrimal gland (orange, left), salivary gland (blue), cutaneous (green) and breast (pink). There were specific genes that were up-regulated in some ACC samples compared to the others. For example, at the far left of the heatmap (labeled A at bottom, orange color bar at top) is a group of ACC tumors, mostly from the lacrimal gland, that overexpress the OPRPN gene, which encodes the Opiorphin Proline-Rich Lacrimal Protein 1 (previously named PROL1). The next group (labeled B) are salivary ACC (blue color bar at top) and overexpress several salivary-gland specific genes including MUC19, CA6, MUC7, SMR3B, LPO, BPIFA2, CST5, CST2, CST1 and CST4 (marked by blue dots at lower right of figure). Group C also contains salivary ACC samples with a few from other tissues, and several overexpress KRT4 and KRT13. Most of the cutaneous samples clustered in a group (D, green color bar) and are identified by overexpression of FLG, KRT10, FLG2, KRT6A and KRT1. However, the remaining ACC tumors formed a large cluster (labeled E), which included most of the breast ACC tumors as well as samples from salivary, lacrimal and cutaneous adnexal glands. The last group was notable because the samples failed to express the gland-specific marker genes that defined the other groups, suggesting that they had a more de-differentiated or perhaps more stem cell-like phenotype. Thus, while we were able to identify tissue-specific marker genes in some ACC tumors, the specificity was not absolute and there remained significant heterogeneity in the gene expression patterns of different samples. Also, since tumor samples always contain some normal cells, the tissue-specific differences that were detected could be due to the non-tumor cells in the samples. We conclude that an ACC-specific gene expression pattern dominated the tumors, apparently overriding the tissue-specific differences. 3.3. The New DK Cohort of ACC Tumors Also Contains a Poor-Survival Subgroup of Patients Previous studies of ACC tumor samples identified subgroups of tumors with distinct gene expression and survival characteristics . We carefully evaluated the new cohort of salivary gland ACC tumors for evidence of subgroups with distinct gene expression patterns. Figure 3A shows a multi-dimensional scaling plot of the 56 DK cohort salivary gland samples. Most of the tumors (shaded light blue) form a large cluster but a small group of tumor samples (brown) formed a separate group at the upper left corner of the plot. Figure 3B shows a Kaplan-Meier survival analysis: the samples in the brown group had a median survival of only 8 months, compared to the main group (light blue), which showed a median survival of 80 months (p-value = 0.006). This is reminiscent of our previous results with the TX cohort, where a subgroup of ACC tumors displayed similarly poor survival . Some ACC tumors display a 'solid form' morphology, which has been associated with worse prognosis . Although 'solid morphology' ACC tumors were excluded from the TX cohort, the DK cohort contains 11 such samples: 5 in the poor prognosis group and 6 in other group. This suggests that the poor prognosis group is not defined simply by solid tumor morphology. The differential gene expression analysis identified 273 genes at least 2-fold down-regulated in the brown subgroup (adjusted p-value < 0.05). The results for 85 of the genes are summarized in the heatmap in Figure 3C (The poor-survival subgroup cluster is at the left side of the heatmap, marked by the brown color bar at the top). In the heatmap, all the genes that were regulated in similar directions both in this new DK cohort and also in the previously described TX cohort (e.g., up-regulated in both poor survival groups) are marked by bars along the right edge of the figure. Genes that were up-regulated in both poor-survival subgroups include (from bottom of the heatmap) CD37, SERPINE2, CDK19, PRLR, and RPL23. Down-regulated genes include AQP3, SCNN1A, LTF, ELL2, and DKK3. These results further indicate that a subgroup of ACC tumors from patients with poor survival display a unique gene expression profile that could potentially be used for prognostication . 3.4. Combining the TX and DK Cohorts Provides Additional Details about Subgroups of ACC Tumors The results described above suggest that the new DK cohort of ACC tumor samples contains subgroups of patients that are very similar to the subgroups we identified previously in the TX cohort . To compare the subgroups we combined the RNA-seq results of the two independent cohorts and performed a unified analysis of 124 salivary gland ACC samples (56 from DK and 68 from TX). Figure 4A shows the multi-dimensional scaling plot of the combined data sets, which form three main groups. The largest group, in the middle of the plot, have been shaded dark blue or cyan to indicate that they overexpress MYB or the related MYBL1 gene, respectively. A group at the upper left is shaded red and contains the poor survival samples from both cohorts, all of which express MYB. Finally, a group of samples at the right, shaded orange, express neither MYB nor MYBL1 ('no MYB'). This group was described previously, and the 'driver' oncogenes or mutations responsible for that group remain unknown . Although these cohorts of ACC samples were completely independent and the patients came from different countries, both cohorts formed similar major subgroups when analyzed together. A Kaplan-Meier survival analysis of these groups is shown in Figure 4B. The overall survival for patients in the orange 'no Myb' group was similar to the main group of samples expressing either MYB or MYBL1. As described above, the red group displayed much worse survival compared to the other patients. While the median survival for most patients exceeded 120 months, including the orange 'no MYB' group, median survival for the red group was only 16.8 months (p-value < 1 x 10-6). These groups were segregated using only their different gene expression characteristics, suggesting that biomarkers could be developed to identify the patients in the poor survival group at the time of surgery. The MDS plot in Figure 4C shows the large overlap in the DK and TX cohorts, despite being analyzed separately and several years apart. 3.5. ACC Tumors That Do Not Express MYB or MYBL1 Have a Unique Transcription Profile Most ACC tumors have recurrent chromosomal translocations that activate the MYB oncogene or the related MYBL1 gene , but this raises questions about the underlying biology and driver genes active in the remaining ACC tumors that do not express MYB or MYBL1. As shown in Figure 4B, the 'no MYB' subgroup of tumors (orange line) had survival similar to the bulk of ACC samples (blue and cyan lines). To further explore the potential driver genes in these samples, we compared them to the rest of the ACC samples in the combined cohort and performed a differential gene expression analysis. In the heatmap shown in Figure 5, the dendrogram at the top shows the hierarchical clustering that was used to arrange the samples from left to right. The 'no MYB' samples are at the far right (marked by orange at the top). The heatmap summarizes the gene expression differences for 124 of the 881 genes that were differentially expressed (at least 2-fold down-regulated, adjusted p-value < 0.05) when the 'no MYB' samples were compared to all the others. The top 10 down-regulated genes are listed in Table 2, and the full list is provided in Supplementary Data (Table S1). There are several important conclusions from this analysis. First, as described previously , the ACC samples that express MYBL1 do not form their own subgroup, but mix in with the samples expressing MYB, suggesting that the two oncogenes have similar effects on gene expression patterns . Second, each of the three main groups (orange, red, blue) contains samples from both the TX and DK cohorts, suggesting that these subgroups are consistent in ACC tumors and are not a characteristic unique to just one cohort or one analysis. Several interesting genes are marked along the right side of the heatmap, including AFF1, EBF1, EMP1, ZFP36, FOXO1, and SFRP2, which are all up-regulated in the 'no MYB' tumors (marked by orange dots). The SHANK2, NFIB, GABRP, MEX3A, PRLR, and MYB genes were down-regulated in the 'no MYB' tumors (marked by blue dots). A gene set enrichment analysis identified a number of Gene Ontology Cellular Process categories that were over-represented in the differentially expressed genes. The top six categories are described in Table 3. The finding that the 'no MYB' samples have such a dramatically different gene expression profile reinforces the conclusion that the ACC phenotype can be achieved through different regulatory pathways. 3.6. Poor Survival ACC Samples Have a Unique Transcription Profile For the combined cohorts, the poor survival subgroup, marked by red at the top of the heatmap in Figure 5 and in the survival plot in Figure 4B, displayed a median survival of only 22 months, compared to greater than 123 months for the other patients. This suggests that a gene expression panel could be developed to identify patients at highest risk of poor survival. To characterize the poor survival subgroup in more detail, we performed in depth analysis of the gene expression patterns of tumors from these patients. The heatmap in Figure 6 summarizes the results of a differential gene expression analysis comparing the poor survival subgroup samples to all the other ACC tumor samples in the combined cohort. The samples are arranged in the same left-to-right order as in Figure 5, using the dendrogram generated by hierarchical clustering, and the poor-survival samples are indicated by the red color bar at the top. The samples from the TX and DK cohorts are indicated by the gray and purple color bar at the bottom, respectively. The heatmap summarizes the relative expression of the 124 most differentially expressed genes out of the 729 genes that were at least 2-fold down-regulated (with adjusted p-values > 0.05). Several notable down-regulated genes are marked by red or gray dots, respectively, along the right side of the heatmap. The genes that were up-regulated in the poor survival samples include EZH2, HDAC2, PRLR, SOX8, NFIB, SHANK2, and ADARB1. The down-regulated genes include CND2, TP63, AQP3, NTRK3, and ADARB2. The top 10 down-regulated genes are listed in Table 4 and the full list is provided in Supplementary Data (Table S2). Interestingly, there is no single gene that is specifically down-regulated only in the poor survival samples, or that could be used to identify either the poor survival or better survival patients, suggesting that a multi-gene biomarker could be developed to identify the patients in the poor-survival subgroup. 3.7. A Multi-Gene Classifier to Identify Poor Survival Patients As shown in Figure 3, the DK cohort of ACC samples contained a subgroup of patients with poor survival, similar to one that was originally identified in the TX cohort . Having patients from two independent cohorts allowed us to use the original TX cohort as a training set to develop a multi-gene biomarker panel, which could be validated with the DK cohort. Starting with expression data for 3597 genes expressed above a threshold level in the training set (TX cohort), we used an elastic net type of penalized logistic regression model to identify genes that could distinguish the poor prognosis cohort from the rest of the patients. The model selection was performed with a 10-fold cross-validation and yielded a 49-gene classifier developed solely with data from the TX cohort (Table 5). The ROC curve analysis was used to evaluate the classifier's accuracy. As shown in Figure 7, the elastic net classifier could distinctly separate the poor prognosis samples from others in the training set (TX cohort, left panel) as AUC = 1. An unbiased estimate for the AUC (AUC = 1) was also achieved through a double loop nested cross-validation, which showed a perfect classification performance. However, the accuracy achieved was expected because the same gene expression data were used to develop and test the classifier. Importantly, the classifier developed with the TX cohort also gave nearly perfect (AUC = 0.984) separation on the independent DK cohort test set (right panel). We also generated a 14-gene subset of the classifier using a Least Absolute Shrinkage Selection Operator (LASSO) approach. The 14-gene classifier containing genes A2M, ACTA2, ANO1, APOL6, DMD, IPO9, LIMCH1, MAMLD1, MIR205HG, PLAT, RASSF6, SEMA3C, SLPI, and TP63 (shown in bold in Table 4) was only slightly less accurate, with AUC = 0.976. These results suggest that a gene classifier can be used to identify ACC patients with poor prognosis. To illustrate their usefulness, we tested the genes in the classifiers on the combined DK and TX cohorts of salivary gland ACC samples. We limited the data sets to only the 49-gene or 14-gene lists (except that MYB, MYBL1, and NFIB were added back for comparison), and performed hierarchical clustering, which identified the two major clusters shown in the dendrograms in Figure 8A,B. The heatmaps display the differences in gene expression. Interestingly, most of the classifier genes were down-regulated in the poor prognosis tumors compared to the other samples. As an example, the TP63 tumor suppressor gene is significantly down-regulated in the poor prognosis group. The poor prognosis tumors appear to lack the expression of specific genes that are expressed by the other ACC samples. Notably, only 4 of the 11 'solid form' morphology samples from the DK cohort were in the poor prognosis subgroup, suggesting that solid morphology is insufficient to classify samples as poor prognosis . As shown by the color bar at the top of each heatmap, some of the samples that were in the poor prognosis subgroup described in Figure 4 (marked red) did not cluster with the poor survival samples identified by the gene classifiers, and a few samples that were not included above did cluster in the poor survival group in this analysis. However, as shown in the Kaplan-Meier survival plots in Figure 8C,D, the classifiers did identify a poor-prognosis group with median survival of less than 20 months, compared to a median of 125 months for the rest of the samples. In addition, none of the poor-prognosis patients identified by the classifiers survived 10 years, while more than half of the other patients survived at least 10 years. Thus, the multi-gene classifiers identified using the TX cohort samples were able to identify a subset of ACC patients in the independent DK cohort, which validates the classifiers and suggests that adapting them to the clinic could be useful. To examine whether the gene classifier provides more information for survival outcomes beyond that contained in the clinical covariates, we performed univariate and multivariate Cox regression analyses, with the gene classifier and clinical covariables deemed to be the risk factors as predictors. (Details of the analysis are in File S1). The available clinical covariables include Margins (free or close), Vascular Invasion (yes or no), Radiotherapy (yes or no), Cribriform (tubular or solid), and Stage (I-II or III-IV). The analyses were restricted to 56 samples; a union of the subsets to which the data of each variable are available. However, the number of samples used by each Cox regression analysis varied subject to data availability. The univariate and bivariate analyses are in the Supplementary Materials, and the multivariate analysis is reported in Table 6. The univariate analysis (see Supplementary Materials) showed that the two variables, Vascular Invasion and Cribriform, were significantly associated with survival outcomes (p < 0.05), while the variable Stage was marginally significant (p = 0.084). We compared these three variables with the gene classifier through bivariate Cox regression (see Supplementary Materials). The result showed a remarkable association between our gene classifier and survival after adjusting for each clinical covariate's effect. We further performed a multivariate Cox regression (Table 6), and our gene classifier was still significantly correlated with the survival outcomes after adjusting for Vascular Invasion and Stage effects. Note that the Cox regression with three or more variables will not converge if we include Cribriform in the model, which limits our ability to conduct further investigation in this respect. However, the results have given sufficient statistical evidence that our gene classifier provided more information about the survival outcome than the available clinical parameters. 4. Discussion We compared the transcription profiles of ACC tumor samples that arose in very different tissues: salivary gland, lacrimal gland, breast, and skin. Despite being from different tissues, all ACC tumors had markedly similar gene expression profiles. Indeed, the ACC samples were much more similar to each other than they were to normal salivary gland tissue or another type of salivary gland tumor, acinic cell carcinoma . These results demonstrate that ACC tumors arising in different tissues are highly related and are difficult to distinguish using gene expression patterns alone. Interestingly, different types of ACC tumors were shown previously to have distinct patterns of microRNA expression --a result that we could not reproduce using gene expression results. This suggests that the activated MYB or MYBL1 oncogenes may induce an ACC-specific gene expression pattern that affects protein-coding genes much more than microRNAs. This is a fascinating biological difference that could be important for explaining tumor phenotypes and some aspects of tissue differentiation. Having RNA-seq data from a new set of ACC samples provided us with the opportunity to perform a validation cohort analysis. Despite the challenges that exist for translating RNA sequencing (RNA-seq) results into widely used clinical assays , several types of gene expression signatures have been developed for clinical use . In this study, we used RNA-seq data from a previous cohort of 68 salivary gland ACC samples to develop a 49-gene expression classifier for identifying a subgroup of patients with poor survival. We then validated the result using results from the new cohort of 56 salivary ACC samples, finding that the biomarker was able to distinguish 98% of the poor survival patients. A smaller 14-gene classifier achieved similar results with slightly less accuracy. Salivary gland ACC patients display widely variable outcomes, with some patients surviving decades after surgery and others succumbing after only a few months . It seems clear that the development of new clinical trials should be targeted to the ACC patients that are most likely to have a recurrence and die from the disease. The validated biomarker we have described should have important utility, if it can be developed into an assay suitable for clinical laboratory use. Although our results do not suggest a new or modified treatment for ACC patients, they do suggest that developing a suitable biomarker assay to identify the worse prognosis patients is worthwhile so a new therapeutic strategy could be developed for them. Clinical RNA-seq is fast becoming a routine assay for cancer patients, so these biomarkers should be adaptable to clinical laboratories. Some ACC tumors display a 'solid form' morphology, which has been associated with worse prognosis . Other clinical features, such as advanced tumor stage, lack of clean margins during surgery, or vascular invasion, might also be used to identify higher risk patients. However, in our analysis, none of these other markers were able to identify the poor prognosis group of patients that we identified using gene expression patterns. Therefore, we conclude that the gene classifiers provide a novel and independent means of distinguishing poor prognosis ACC patients that should be pursued and studied further. The next step will be to develop assays that work in clinical laboratories so that these classifiers can be used to identify patients that should be targeted for clinical trials or more aggressive therapy to improve their survival. In addition to identifying and validating a multi-gene classifier for ACC patients, analyzing a new cohort of 56 salivary gland ACC samples from Denmark (DK) also validated important biological results that we described previously using 68 ACC patient samples from the Salivary Gland Tumor Bank in Texas (TX) . The main result is that ACC patients can be divided into at least three distinct groups based on gene expression signatures. These groups are easily discernable in the multi-dimensional scaling plots . The samples in the main group, comprising 76% of the total, express either MYB or MYBL1 and have a median survival of more than 10 years after surgery. A second group, about 10% of the samples with survival similar to the main group, express neither MYB nor MYBL1. These samples have a unique gene expression signature, suggesting a different mechanism driving the malignancy. The samples in the final group, about 14% of the total, are the focus of the multi-gene classifier because they have much worse survival than the rest of the ACC patients. Although the detailed transcriptome analyses that we performed were able to discern distinct subgroups of ACC tumor types, the bulk RNA-sequencing does not provide information on cell lineage composition within tumors. Thus, it is not clear if the different subgroups result from the unique features of different types of ACC tumor cells or whether the subgroups are due to differences in cellular composition in the tumors. Addressing those questions will require using single-cell genomics assays or spatial genomics approaches that can discern different cell types in the tumors. 5. Conclusions Our somewhat surprising result is that ACC tumors arising in different tissues or organs have remarkably similar transcriptional profiles. Indeed, we were unable to identify gene signatures that distinguished the ACC tumors from different organs. This may point to an important underlying biology in ACC tumors that makes them so similar. Since the majority of ACC tumors overexpress the MYB or MYBL1 genes, the dominant ACC phenotype may be induced by the activated Myb transcription factors. A second, but very important finding is that RNA sequencing analysis can be used to identify a subgroup of MYB-expressing salivary gland ACC patients with poor prognosis. We were able to use the new DK cohort of ACC samples to validate a biomarker developed with an earlier (TX) cohort. This is especially important for diseases like ACC, in which many patients survive more than 10 years post-surgery. Our results provide a tool for identifying the patients that should be enrolled in clinical trials of targeted therapies to improve their outcomes. Acknowledgments S.A.N. and K.J.B. acknowledge the outstanding technical support from J. Padilla, J. Woods, and M. Cyphery in the Ness laboratory and the Analytical and Translational Genomics Shared Resource. Supplementary Materials The following supporting information can be downloaded at: Figures S1-S6: Larger versions of the heatmaps in Figure 1, Figure 2 and Figure 3, Figure 5, Figure 6 and Figure 8, respectively. Table S1: No Myb group differential gene expression analysis results. Table S2: Poor survival differential gene expression analysis results; File S1: Clinical data analysis details with univariate and bivariate Cox regression results and details of software packages used. Click here for additional data file. Author Contributions Conceptualization, S.A.N. and S.A.; methodology, K.J.B. and S.A.N.; software, S.A.N.; validation, K.J.B. and S.A.N.; formal analysis, H.K. and Y.G.; resources, S.A., A.K.E.-N., P.H., K.K., L.M., S.H., D.P., A.M. and D.T.T.; data curation, K.J.B. and S.A.N.; writing--original draft preparation, S.A.N.; writing--review and editing, K.J.B., A.K.E.-N., D.Y.L., P.H. and K.K.; visualization, S.A.N. and K.J.B.; supervision, S.A.N.; project administration, K.J.B. and S.A.N.; funding acquisition, S.A.N. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of University of New Mexico (protocol code 20-120 approved 4 June 2020) and the Danish Regional Ethics Committee (H-6-2014-086) and the Danish Data Protection Agency (Journal no. REG-94-2014). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement RNA sequencing data is available for download from the NCBI BioProject database using study accession number PRJNA287156. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Common features of ACC tumors from different tissues. Differential gene expression analysis was performed on the DK cohort of 113 ACC tumor samples from different tissues, including breast (n = 17, pink), cutaneous (n = 24, green), lacrimal gland (n = 16, orange), and salivary gland (n = 56, blue) as well as normal salivary gland (n = 5, black) and salivary gland acinic cell carcinoma samples (n = 7, gray). (A) Multidimensional scaling shows that the acinic cell carcinoma samples clustered at the upper left, the normal salivary gland samples in the lower left, and all of the ACC tumor samples at the right, no matter what tissue they were derived from. (B) The heatmap summarizes the gene expression differences. The color bar at the top identifies the tissue type and tissue of origin. The normal salivary gland and acinic cell carcinoma samples are at the left. Several genes important for ACC tumors are marked by dots at the right. . Figure 2 Tissue-specific gene expression differences in ACC tumors. The ACC tumors from the DK cohort were analyzed for tissue-specific gene expression by specifically selecting genes that marked tumors derived from different tissues. A total of 1089 differentially expressed genes were identified by comparing all the tissue groups to each other (at least 2-fold down-regulated with adjusted p-value < 0.05). The heatmap summarizes the gene expression differences for 123 of the most highly expressed genes. The tissues of origin are indicated in the color bar at top: lacrimal gland, salivary gland, cutaneous, and breast are indicated by orange, blue, green and pink, respectively. Notable genes mentioned in the text are marked by dots along the right edge. . Figure 3 The DK cohort of ACC tumors contains a poor-outcome subgroup. (A) Multi-dimensional scaling plot of the gene expression data from the DK cohort of salivary gland ACC tumor samples. (B) Kaplan-Meier survival analysis shows that the samples in the brown subgroup had a significantly (p-value = 0.006) worse survival. (C) The heatmap summarizes the differential gene expression analysis comparing the poor survival (brown) subgroup to the rest of the samples. Genes marked by bars at the right were also identified previously in a poor-survival subgroup from the TX cohort . The color bars at the bottom summarize the available clinical information for gender, solid or tubulocribiform morphology, tumor stage, margins, vascular invasion, and radiotherapy. A larger version of the heatmap is provided in Supplementary Figure S3. Figure 4 Combined analysis of the DK and TX cohorts of ACC tumor samples. The RNA-seq data from the new DK cohort was combined with previously described (Frerich et al., 2018 ) data from the TX cohort for a combined gene expression analysis. (A) Combined multi-dimensional scaling plot. Orange shading indicates 'No Myb' samples that express neither MYB nor MYBL1, blue and cyan indicate MYBL1-expressing samples in the main group, red indicates the MYB-expressing samples with poor survival. (B) Kaplan-Meier survival plot of samples in the four groups of ACC samples. (C) Multi-dimensional scaling plot similar to panel A, but with samples from DK or TX cohorts labeled purple or gray, respectively. Figure 5 Differential gene expression analysis: 'No Myb' samples. The heatmap summarizes the differential gene expression analysis using the combined cohorts of ACC samples from DK and TX, comparing the 'no Myb' group (orange color bar at top) to the rest of the samples, all of which express either MYB or MYBL1. Notable genes mentioned in the text are marked by dots at the right. The purple and white color bar at the bottom indicates samples from the DK and TX cohorts, respectively. A larger version of this heatmap is provided in the Supplementary Materials, as Figure S4. Figure 6 Differential gene expression analysis: Poor survival samples. The heatmap summarizes the differential gene expression analysis using the combined cohorts of ACC samples from DK and TX, comparing the poor survival group (red color bar at the top) to the rest of the samples. Notable genes mentioned in the text are marked by dots at the right. Red and blue dots indicate genes down-regulated in the poor survival samples. The purple and white color bar at the bottom indicates samples from the DK and TX cohorts, respectively. A larger version of this heatmap is provided in the Supplementary Materials, as Figure S5. Figure 7 Elastic net ROC curves. The elastic net classifier developed with the TX cohort training set (A, Left) produced an Area Under the Curve (AUC) of 0.984 in the DK cohort test set (B, Right). Figure 8 Classifier Groups. The 49-gene (A) or 14-gene (B) classifiers were used to separate ACC samples into groups by hierarchical clustering, as illustrated in the dendrograms at the top of each heatmap, which compare the gene expression profiles of the two groups. (The MYB, MYBL1, and NFIB genes were added to the analysis for comparison. They are marked by black dots at the right). The samples from the original poor survival group are marked by red in the color bars. At the bottom, Kaplan-Meier plots compare the survival of patients in the groups defined by the 49-gene (C) or 14-gene (D) classifiers. Larger versions of these heatmaps are in Supplementary Figure S6. cancers-15-01390-t001_Table 1 Table 1 ACC Tumor Cohorts. Tissue Number Form Gender Stage Age at Surgery Source Oncogene * New Cohort (DK) Breast 17 NA NA NA NA DK MYB: 16 MYBL1: 1 Cutaneous 24 NA NA NA NA DK MYB: 17 MYBL1: 3 No_MYB: 4 Lacrimal 16 NA NA NA NA DK: 10 FL: 6 MYB: 11 MYBL1: 2 No_MYB: 3 Salivary 56 Solid: 11 Tubulocribiform: 43 NA: 2 F: 29 M: 27 Stage I-II: 30 Stage III-IV: 26 Age: 32-87 NA: 17 DK MYB: 47 MYBL1: 7 No_MYB: 2 Cohort Total 113 MYB: 91 MYBL1: 13 No_MYB: 9 Previous Cohort (TX, Frerich et al. 2018 ) Salivary 68 NA F: 30 M: 38 NA NA TX MYB: 52 MYBL1: 7 No_MYB: 9 Combined 181 MYB: 143 MYBL1: 20 No_MYB: 18 * No MYB indicates no expression of either MYB or MYBL1. cancers-15-01390-t002_Table 2 Table 2 Differentially Expressed Genes in No MYB Samples vs. All Others. DE Genes Top 10 Up-Regulated FC Adj p-Value Top 10 Down-Regulated FC Adj p-Value 881 TG 136.26 1.41 x 10-32 LINC01833 0.02 1.39 x 10-5 HMGA2 42.11 1.66 x 10-14 MUC7 0.03 1.25 x 10-2 IGFN1 14.91 7.41 x 10-8 LOC643201 0.03 2.04 x 10-3 LYZ 12.38 4.19 x 10-6 MUC19 0.03 1.93 x 10-3 FLG2 12.08 5.87 x 10-5 CTNND2 0.04 6.12 x 10-5 COL1A1 9.62 9.72 x 10-20 FIRRE 0.04 4.73 x 10-10 FLG 8.80 5.38 x 10-6 LOC105378521 0.04 8.12 x 10-6 LTF 8.05 1.37 x 10-6 ART3 0.04 7.41 x 10-8 HSPB8 7.97 5.59 x 10-6 LOC107984390 0.04 1.42 x 10-11 TNNT1 7.74 2.29 x 10-3 SIX3 0.04 5.12 x 10-4 cancers-15-01390-t003_Table 3 Table 3 Gene Ontology Categories of DE Genes from No MYB Group Analysis. GO.ID Term Annotated Signif Expected Adj p-Value GO:0043062 extracellular structure organization 218 87 31.6 2.40 x 10-19 GO:0045229 external encapsulating structure organization 219 87 31.74 3.50 x 10-19 GO:0045765 regulation of angiogenesis 128 47 18.55 2.40 x 10-10 GO:0022610 biological adhesion 658 166 95.37 3.10 x 10-10 GO:0008284 positive regulation of cell population prolif. 354 93 51.31 7.80 x 10-8 GO:0009617 response to bacterium 178 54 25.8 2.10 x 10-7 cancers-15-01390-t004_Table 4 Table 4 Poor Survival Samples vs. All Others. DE Genes Top 10 Up-Regulated FC Adj p-Value Top 10 Down-Regulated FC Adj p-Value 729 ANKRD1 10.28 3.99 x 10-5 CST4 2.2 x 10-04 3.46 x 10-3 LINC02275 8.57 2.77 x 10-5 CST5 3.0 x 10-03 3.82 x 10-3 LINC02515 5.66 4.90 x 10-5 CST2 3.0 x 10-03 8.83 x 10-3 NCAN 5.31 5.79 x 10-3 CST1 3.1 x 10-03 2.66 x 10-3 HPSE2 4.54 4.08 x 10-3 SPRR3 3.3 x 10-03 6.00 x 10-3 NPY5R 4.51 9.50 x 10-7 KRT13 4.2 x 10-03 3.94 x 10-4 LINC01833 4.22 2.95 x 10-3 SPRR1A 5.3 x 10-03 2.05 x 10-3 SOX8 4.04 8.42 x 10-7 KRT6C 5.5 x 10-03 6.69 x 10-4 ASL 3.88 3.82 x 10-5 SMR3B 5.9 x 10-03 2.82 x 10-3 PRLR 3.81 1.99 x 10-8 BPIFA2 7.1 x 10-03 1.67 x 10-2 cancers-15-01390-t005_Table 5 Table 5 Elastic Net 49-Gene Classifier. Symbol Value Symbol Value Symbol Value A2M -0.11 HMCN1 -0.065 PLA2R1 -0.0434 ABCA8 -0.0129 IPO9 0.376 PLAT -0.0448 ACTA2 -0.1264 ITPR1 -0.1121 PLD1 -0.0356 ADAMTS9 -0.0986 KRT14 -0.0359 PPARGC1A -0.0378 ALDOA -0.018 LDLRAD4 0.0354 PRUNE2 -0.0026 ANO1 -0.099 LGR4 -0.0391 RASSF6 -0.1461 APOL6 -0.0977 LIMA1 -0.0793 SEMA3C -0.0582 CARMN -0.0091 LIMCH1 -0.145 SH3D19 -0.0562 CD9 -0.0886 LOC107987158 0.0363 SLPI -0.1399 CFH -0.0226 LTF -0.0048 SVIL -0.168 COL17A1 -0.0151 MAMLD1 0.0339 SYNPO2 -0.0159 COL7A1 -0.0095 MIR205HG -0.0511 TAGLN -0.0321 COL9A2 -0.0072 MLPH -0.013 TNFRSF19 -0.0054 DMD -0.3178 MTUS1 -0.0917 TP63 -0.1442 EFS 0.0074 PARP14 -0.0746 TPM2 -0.0175 EGFR -0.0288 PCYOX1 -0.1327 FRMD4B -0.0279 PIK3R1 -0.0097 Note: Genes used in the 14-gene classifier are shown in bold. cancers-15-01390-t006_Table 6 Table 6 Multivariate Cox Regression Analysis. Clinical Covariates & Gene Classifier Values Hazard Ratio 95% Confidence Interval p-Value Vascular Invasion No 1 0.097 Yes 2.703 0.83-8.76 Stage I-II 1 0.271 III-IV 1.601 0.69-3.70 Gene Classifier Group 1 1 0.010 Group 2 26.01 2.19-309.3 Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000626 | Introduction De-escalated treatment of hemithyroidectomy without radioactive iodine (RAI) is now accepted for patients with low-risk, well-differentiated thyroid cancer (WDTC). The benefit of long-term follow-up care remains controversial. This study aims to describe parameters associated with less than total thyroidectomy, and discharge from specialist follow-up in patients with low-risk WDTC in Australia. Methods An online survey was distributed to Australian members of Endocrine Society of Australia, Australian and New Zealand Endocrine Surgeons, and Australian Society of Otolaryngology, Head and Neck Surgery. Clinicians completed a survey of management and follow-up care preferences for four clinical vignettes (all low-risk WDTC). Results 119 clinicians (48% endocrinologists, 55% male) answered at least one question. The majority (59%) of respondents recommended less than total thyroidectomy and omission of RAI in patients with WDTC <2 cm. Most (62%) would discharge a patient with micropapillary thyroid cancer within 1 year following total thyroidectomy. In contrast, for WDTC 1-4 cm, >90% of clinicians would continue specialist follow-up for at least 5 years. The majority of clinicians felt that patients experienced disproportionate fear of recurrence and were reassured by follow-up. After multivariable analysis, clinicians who participated in multidisciplinary teams (MDTs) were more likely to choose de-escalated care for both initial treatment (p = .005) and follow-up care (>5 years, p = .05). Conclusion Clinician attitudes captured by this survey reflect recent changes in guidelines towards hemithyroidectomy for low-risk WDTC, particularly amongst MDT attendees. There is a need to further examine the impact of de-escalated care on fear of recurrence and quality of life in thyroid cancer survivors. This survey of Australian thyroid clinicians, demonstrates that de-escalation of surgical care (less than total thyroidectomy) of patients with low-risk thyroid cancer has become common since the 2015 American Thyroid Association Guidelines were published. Clinicians feel that patients fear of cancer recurrence is disproportionately excessive compared to the true risk; but that they are reassured by clinician follow-up. thyroid cancer health-related quality of life thyroidectomy cancer survivorship fear of cancer recurrence multi-disciplinary team source-schema-version-number2.0 cover-dateMarch 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.2.6 mode:remove_FC converted:10.03.2023 Widjaja W , Rowe CW , Oldmeadow C , et al. Current patterns of care in low-risk thyroid cancer--A national cross-sectional survey of Australian thyroid clinicians. Endocrinol Diab Metab. 2023;6 :e398. doi:10.1002/edm2.398 pmc1 INTRODUCTION Well-differentiated thyroid cancer (WDTC) accounts for more than 95% of thyroid cancer diagnoses. Generally, it carries an excellent prognosis with a 10-year survival of >97%. 1 , 2 This is partly due to early detection and effective surgical and medical treatment, but it is also due to the over-diagnosis of indolent cancers. 3 Despite a good prognosis, it has been recently shown that health-related quality of life (HRQoL) detriments are common in thyroid cancer survivors. 4 , 5 The 2015 American Thyroid Association (ATA) management guidelines advocated for a paradigm shift towards the de-escalation and individualisation of management of low-risk WDTC. 6 There is a need for contemporary data regarding the uptake of these guidelines, particularly outside of the United States. The appropriate length and frequency of specialist follow-up for WDTC remain uncertain, and few guidelines exist. 7 This may result in significant heterogeneity in practice, possibly with detriment to both patients and health services. For example, for those with low-risk disease and excellent response to treatment, longer term follow-up can be associated with false positive ultrasound results and is unlikely to be cost effective or an appropriate use of healthcare resources. 8 , 9 It is also unclear what the impact of de-escalation of initial care might have on follow-up practices. Finally, recent evidence has highlighted high rates of anxiety, depression and fatigue in low-risk thyroid cancer patients, driven in part by high rates of fear of cancer recurrence (FCR). 10 , 11 The impact of longer-term follow-up on patient HRQoL and FCR is unknown. 12 The aim of this study was to explore specialists' preferences for uptake of de-escalation and follow-up in patients with low-risk WDTC, with particular reference to the group of patients identified as suitable for treatment with less than total thyroidectomy by the 2015 ATA thyroid cancer guidelines. 6 We aimed to identify important factors that impact treatment decisions in low-risk thyroid cancer. These data can facilitate the development of standardised guidelines, highlight potential implementation challenges to de-escalated care and inform future research into HRQoL in thyroid cancer. 2 METHODS 2.1 Survey development This study was prospectively approved by the Hunter New England Human Research Ethics Committee (2020ETH01892) and the University of Newcastle Human Research Ethics Committee (H-2021-0058). A 50-item study-specific survey was developed by a team of endocrine surgeons; ear, nose and throat (ENT) surgeons, endocrinologists and behavioural scientists (Appendix 1). Survey elements included clinician demographic characteristics (age, gender, geography, medical speciality, years in practice, annual thyroid cancer caseload and MDT attendance). Structured responses to four clinical vignettes (Table 1) investigated initial treatment and follow-up preferences for patients with very low and low-risk WDTC according to ATA guidelines. 6 The scenarios were presented to survey participants in order of increasing risk and aimed to explore clinician's preferences for surgical intervention, long-term specialist follow-up, neck ultrasound and time to discharge back to primary care according to patient and tumour characteristics (gender, age, tumour size, lymph node involvement, vascular invasion, BRAF V600E status and extent of surgery). The survey also asked clinicians about their perception of patients' FCR and whether follow-up care might exacerbate or ameliorate this. The survey was piloted by a group of ten clinicians from different geographical locations across Australia. TABLE 1 Clinical vignettes description. Scenario 1 A 45-year-old female patient has an incidental papillary thyroid carcinoma in the pathology specimen after a total thyroidectomy for a multinodular goitre, the tumour measures 4 mm and there are no other concerning features. Scenario 2 A 40-year-old female patient presents with an incidentally discovered papillary thyroid carcinoma measuring 15 mm. The tumour is completely intra-thyroidal, the contralateral lobe is normal on ultrasound and there are no suspicious lymph nodes. Pre-operative TSH is <2 mIU/L. Scenario 3 A 65-year-old female patient has a solitary 2 cm left thyroid nodule with Bethesda IV cytology. The contralateral lobe is normal, and there are no suspicious lymph nodes either clinically or on ultrasound. She undergoes a hemithyroidectomy. The pathology report reveals an 18 mm minimally invasive follicular thyroid cancer with capsular but not vascular invasion. Scenario 4 A 48-year-old male patient presents with a solitary right thyroid nodule. It measures 35 mm and has suspicious sonographic features. Pre-operative cytology is Bethesda VI, suggestive of papillary thyroid cancer. There are no suspicious cervical lymph nodes either clinically or on ultrasound. The contralateral lobe is normal. 2.2 Study population and administration The cross-sectional online survey was administered on the REDcap(r) platform from October 2020 to January 2021. 13 The survey was sent as an email invitation with a weblink and was distributed to the Australian members of Endocrine Society of Australia (ESA), Australian and New Zealand Endocrine Surgeons (ANZES) and Australian Society of Otolaryngology, Head and Neck Surgery (ASOHNS). These organisations are the professional bodies of Australian endocrinologists, endocrine surgeons, and ear, nose and throat (ENT) surgeons, respectively. A reminder e-mail was sent four weeks later to non-responders. 2.3 Statistical analysis Descriptive statistics were evaluated across clinical vignettes. Univariate and multivariable mixed effects regression models were used to examine the associations between management decisions (for each scenario, giving repeated measures at the clinician level), and clinician's demographic data and practice patterns. The model included fixed effects for the demographic and practice pattern variables, and a random intercept for the clinician to model the lack of independence induced by repeated measures of responses for each scenario. Variables were removed from the full model in a sequential manner, starting with the highest p-value, provided their removal reduced the Akieke information criterion and did not alter the regression coefficients by more than 15%. Statistical analyses were programmed using SAS v9.4 (SAS institute, Cary, North Carolina, USA); alpha level of 0.05 was used for statistical significance. 3 RESULTS 3.1 Survey respondents The survey was sent to 400 members of ESA, 123 members of ANZES and 488 members of ASOHNS. A total of 119 participants accessed the survey link, 109 answered at least 1 question and 87 completed at least 1 clinical scenario. It can be presumed that all 123 members of ANZES perform thyroid cancer surgery (response rate in this group of 24%) but it is unknown how many of the clinicians in other speciality groups are actively involved in thyroid cancer management. The data that support this article are available from the corresponding author upon request. Demographic characteristics for the participants are included in Appendix 2. There was no statistically significant difference in gender, age or geographical location between respondents and non-respondents (defined as those who opened the survey but did not complete any scenarios). However, there was a significant difference with responders reporting higher levels of MDT participation versus non-responders (63% vs 38%, p = .04). 3.2 Initial management of WTDC The summary of the participants' responses on the management of WDTC are presented in Table 2. Clinicians were asked to nominate their preferred operation for patients with low-risk WDTC. The majority selected hemithyroidectomy for scenarios 2 and 3, 69% and 59%, respectively, while only 13% were comfortable with hemithyroidectomy with scenario 4. In scenario 2, 60% of surgeons would omit prophylactic central lymph node clearance, but 38% would perform a unilateral prophylactic central neck clearance; a higher proportion than the 26% who would offer total thyroidectomy. Despite there being no difference in the extent of surgery preferred by endocrine surgeons compared to ENT surgeons (76% vs 75% hemithyroidectomy, respectively) in this scenario; the approach of hemithyroidectomy with unilateral prophylactic central neck clearance was particularly prevalent amongst endocrine surgeons with 13 (50%) preferencing this approach compared with 2 (10%) of ENT surgeons. TABLE 2 Characteristics of surgical decision-making in the scenarios. Scenario 2 n = 87 Scenario 3 n = 75 Scenario 4 n = 71 Hemithyroidectomy 56 (69%) N/A a 9 (13%) b Total thyroidectomy 21 (26%) N/A a 61 (87%) b Active surveillance 2 (2.5%) N/A a N/A a Clinician equipoise 2 (2.5%) N/A a N/A a No further surgery N/A a 44 (59%) N/A a Completion thyroidectomy N/A a 16 (21%) N/A a Completion thyroidectomy and radioactive ablation N/A a 15(20%) N/A a No prophylactic central neck dissection 48 (60%) N/A a 16 (23%) Yes (unilateral prophylactic central neck dissection) 30 (38%) N/A a 39 (55%) Yes (bilateral prophylactic central neck dissection) 2 (2.5%) N/A a 16 (23%) No radioiodine ablation 74(91%) 60 (80%) 27 (38%) Yes, low dose (1 GBq, 25 mCI) 7 (8.6%) 15 (20%) 34 (48%) Yes, high dose (4 GBq, 100 mCi) N/A a 0 10 (14%) Note: The patient in scenario 1 had an incidental, post-operative diagnosis of micropapillary thyroid cancer; thus, there was no surgical decision-making in this vignette. a Factor was not listed as an option in this scenario. b One missing. For the higher risk scenario 4, the chosen management was more aggressive with only 23% of clinicians being comfortable to omit prophylactic central neck clearance, 55% would offer unilateral and 23% bilateral neck clearance. Endocrine surgeons were again more aggressive in their surgical approach to the central neck with only 4 (8%) electing to omit central neck clearance compared to 5 (42%) of ENT surgeons. When respondents were asked if they would offer radioactive iodine with the assumption that there is no biochemical or structurally persistent disease following surgery, 9% of clinicians in scenario 2, 20% in scenario 3% and 62% in scenario 4 would recommend radioactive iodine. In scenario 4, 48% would advise low dose (approx. 1 GBq, 25 mCi) and 14% would advise high dose (3-4 GBq, 100 mCi). 3.3 Follow-up of WDTC Clinicians were asked to indicate their preferred specialist care, follow-up protocol for each scenario, assuming that all patients remained free of disease and that patients in scenarios 1, 2 and 4 had undergone a total thyroidectomy, while the patient in scenario 3 only underwent hemithyroidectomy. Of the 52 surgeons surveyed, 56% indicated that they would follow-up the patients themselves, while the remainder would follow-up in conjunction with or to refer back to an endocrinologist for follow-up (these were included as specialist follow-up). While 97% of respondents would offer long-term follow-up to the patient in scenario 4, 88%, 75% and 24%, respectively, would offer long-term follow-up to patients in scenario 3, 2 and 1. The five-year mark was the most common time point for discharge to non-specialist care; however, there was some preference towards 10-year follow-up in scenario 4 (Table 3). Univariate analysis showed that surgeons were significantly less likely to suggest specialist follow-up for their patients for more than 5 years post-operatively, p < .05. TABLE 3 Intention, frequency and modality of follow-up. Scenario 1 n = 87 b Scenario 2 n = 61 Scenario 3 n = 66 Scenario 4 n = 71 c 'When would you discharge this patient back to their GP?' Post-operatively 29 (33%) NA a NA a NA a 1 year after surgery 25 (29%) 1 (1.6%) 1 (1.5%) 2 (2.8%) 1-5 years after surgery 32 (19%) 38 (63%) 41 (62%) 35 (49%) >5 years after surgery 8 (9%) 22 (36%) 24 (36%) 34 (48%) Modality of follow-up Ultrasound Not at all 3 (14%) 9 (15%) 3 (4.6%) 2 (2.9%) 6 monthly 1 (4.8%) 4 (6.6%) 10 (15%) 8 (12%) 12 monthly 13 (62%) 36 (59%) 43 (66%) 48 (70%) 24 monthly 2 (9.5%) 6 (9.8%) 7 (11%) 4 (5.8%) Other 2 (9.5% 6 (9.8%) 2 (3.1%) 7(10%) Thyroglobulin Not at all 1 (4.8%) 0 9 (14%) 0 6 monthly 4(19%) 15 (25%) 13 (20%) 22 (32%) 12 monthly 15 (71%) 45 (74%) 44 (67%) 46 (67%) Other 1 (4.8%) 1 (1.6%) 0 1 (1.4%) Note: Clinicians were asked to nominate follow-up on the assumption that patients in Scenario and 1, 2, and 4 had undergone total thyroidectomy, while patient in Scenario 3 had only undergone hemithyroidectomy. a Factor was not listed as an option in this scenario. b Modality of follow-up was only asked to participants who elected to follow-up patients beyond 1 year in Scenario 1, there were 21 respondents. c 2 missing responses to modality of follow-up. The majority of the participants agreed that specialist follow-up should be performed at 12 monthly intervals. Across all scenarios (Table 3), the majority of clinicians recommended 12 monthly neck ultrasound and serum thyroglobulin. A summary of factors which would prompt more intensive follow-up are presented in Table 4. The common view was that larger tumour size, the presence of vascular invasion, an abnormal contralateral thyroid lobe after hemithyroidectomy, aggressive histopathology and positive BRAF V600E were factors that would increase the intensity of follow-up. TABLE 4 Factors leading to more intensive follow-up. Factors Scenario 1 n = 86 Scenario 2 n = 80 Scenario 3 n = 73 Scenario 4 n = 68 Abnormal contralateral lobe (if patient chose hemithyroidectomy) 70 (81%) 57 (71%) NA a 53 (78%) Multi-focal tumour 55 (64%) 47 (59%) NA a 30 (44%) BRAF 600E IHC positive 45 (52%) 44 (55%) NA a 44 (65%) Microscopic extra-thyroidal extension 65 (76%) 59 (74%) 70 (96%) 39 (57%) Vascular invasion c 80 (93%) 73 (91%) 71 (97%) 61 (90%) Aggressive histological variant c 77 (90%) 71 (89%) NA a 63 (93%) Family History of thyroid cancer 47 (55%) 44 (55%) 31 (42%) 26 (38%) Male gender 18 (22%) 17 (21%) 14 (19%) NA a Age >65 years 19 (17%) 26 (33%) NA a 23 (34%) Central neck nodal metastasis NA a NA a NA a 62 (91%) Larger tumour size b 8 vs 4 mm 6 (7%) 10-20 vs 4 mm 72 (84%) 30 vs 15 mm 59 (74%) 32 vs 18 mm 56 (77%) >40 mm 55 (81%) a Factor was not listed as an option in this scenario. b Size was defined differently for each scenario. c These factors would increase papillary thyroid cancer to ATA intermediate risk. 3.4 Participant self-evaluation of WDTC treatment Clinicians were asked 'do you feel that your patients' perception of their fear of recurrence equates to their actual risk profile of recurrence?' . The majority of clinicians felt that 77% of patients had disproportionately high FCR, 19% were realistic and 4% feared recurrence less than the actual risk. Most clinicians (76%) felt that patients were reassured by their follow-up. The intensity of follow-up was deemed appropriate by 57% but the remaining 43% felt it was excessive. There was no significant difference in perceptions around FCR, the reassurance value of follow-up or the appropriateness of follow-up based on gender, speciality or MDT attendance. Review of free-text comments suggested that some clinicians continue to follow-up patients to manage their anxiety and fear of cancer recurrence. Others commented that their ability to discharge patients related to a lack of confidence that these patients would be well managed in a primary care setting. 3.5 WDTC treatment compared to clinician characteristics Univariate and multivariable analysis was undertaken for treatment (total thyroidectomy as reference) and follow-up (<5 years as reference). Parameters with p < .1 are displayed in Table 5; full data are provided in Appendix 3. There was no statistically significant difference in responses based on gender; geography (capital vs non-capital city, state vs state); years in practice or volume of practice; clinician opinions of the appropriateness of follow-up; or their perception of follow-up as reassuring for patients. Both univariate and multivariable analysis showed that participants who attended a MDT were less likely to suggest total thyroidectomy for WDTC (p < .001 and p = .005, respectively). Univariate analysis suggested that surgeons were significantly less likely to suggest total thyroidectomy when compared to endocrinologists, OR 0.39, p < .05; however, these differences did not persist once multivariable analysis was undertaken. TABLE 5 Univariable and multivariable analysis of treatment decision (total thyroidectomy) and follow-up decision (>5 years follow-up). Variable Univariate OR (95% CI) Univariate p value Multivariate OR (95% CI) Multivariate p value Total thyroidectomy MDT attendance 0.29 (0.15-0.54) <.001 0.29 (0.12-0.68) .005 Surgeon (endocrine or ENT) 0.39 (0.21-0.71) .002 0.65 (0.33-1.25) .19 Follow-up >5 years MDT attendance 0.34 (0.10-1.11) .072 0.23 (0.05-0.99) .049 Surgeon (endocrine or ENT) 0.28 (0.09-0.93) .037 0.35 (0.10-1.2) .094 4 DISCUSSION Since the publication of the 2015 ATA guidelines, there has been a move towards de-escalation of treatment for low-risk thyroid cancer. This study aimed to evaluate preferences for uptake of de-escalation of initial management and follow-up care of low-risk WDTC by specialist clinicians in Australia. Our results suggest that many specialist thyroid cancer clinicians are comfortable to offer hemithyroidectomy and omit radioactive iodine for patients with low-risk thyroid cancer, particularly where the primary tumour measures <2 cm and there are no higher risk features. This practice is consistent with international data confirming that in appropriately selected patients with low-risk WDTC, oncological outcomes are not compromised with hemithyroidectomy. 14 The philosophy behind de-escalation is to minimise harm from overdiagnosis and the complications of treatment in those with low-risk disease, while still maintaining excellent disease outcomes with appropriate treatment and surveillance for higher risk patients. 6 , 15 Overdiagnosis or overtreatment can result in costly use of unnecessary healthcare services and distress to patients and their support networks. There is evidence that complications such as vocal cord paralysis and hypoparathyroidism may be increased in lower-volume settings; yet in this study, the uptake of de-escalation was associated with MDT attendance, a likely surrogate marker of higher-volume practice. 16 Within the broader cancer field, MDTs have been associated with evidence-based guideline awareness and application. 17 It is probable that MDT attendance may also indicate exposure to discussion about de-escalation and increased confidence with de-escalation due to alignment with colleagues. The association between de-escalation amongst surgeons in univariate analysis is interesting. Although it did not reach statistical significance in multivariable analysis, it is worthy of further investigation. The high prevalence of prophylactic unilateral central neck clearance in association with hemithyroidectomy amongst endocrine surgeons in Scenario 2 is interesting. Neither the British Thyroid Association nor ATA Thyroid Cancer Guidelines recommend prophylactic unilateral central neck dissection in patients under the age of 45, with papillary thyroid cancer under 4 cm, with no other adverse features. 6 , 18 These guidelines, however, suggest that a personalised decision approach might be appropriate in patients who do not meet all these low-risk criteria. It is hypothesised that unilateral central neck clearance might be preferred at the time of hemithyroidectomy to decrease rates of recurrence in the central neck and to obtain lymph node staging which might lead to increased rates of completion thyroidectomy and radioactive iodine treatment in this group of patients. 19 , 20 The hypothetical nature of these scenarios limits further investigation of prophylactic central lymph node clearance in the context of de-escalated care; further research in this area is required. Follow-up for thyroid cancer aims to manage treatment side effects, detect recurrence early and provide supportive care to patients and their families. There is general consensus from this survey that the majority of the clinicians provide up to five years follow-up post-operatively for low-risk WDTC. Long-term follow-up is associated with substantial cost to the healthcare with only a 2% risk of recurrence. 8 , 9 A recent Cochrane review suggested that there was little or no difference to overall survival with less intensive versus more intensive follow-up. 21 Interestingly, 43% of clinicians in this study felt that follow-up protocols in their institution were excessive, yet 76% felt that patients were reassured by their follow-up. Further study, including understanding patient perceptions of follow-up and patient coping styles, is required to evaluate the true value of follow-up in terms of clinical, psychosocial benefit and cost-effectiveness. The effect of de-escalated care on HROoL remains unanswered with some studies suggesting that HRQoL is improved in patients treated with hemithyroidectomy. 22 , 23 In contrast, FCR has been shown to be a prominent driver of HRQoL detriments in WDTC, 24 , 25 and some studies have suggested that HRQoL is improved with total thyroidectomy due to a decrease in FCR with a more aggressive surgical approach. 22 , 23 , 26 In this Australian study, 75% of clinicians felt that thyroid cancer survivors experienced FCR in excess of the true clinical risk of recurrence. This is in line with data from international patient studies, where FCR was found to be the most prominent concern amongst thyroid cancer survivors. 26 , 27 There are emerging data to suggest that clinician follow-up visits provide substantial psychosocial support to patients and may decrease FCR in this group. 28 There is also need for further research to investigate the effect of follow-up on FCR and whether integrating interventions aimed at ameliorating FCR into clinician follow-up may benefit patients. The main weakness of this survey is the low response rate, with an overall response rate of 11.8% and only 87 of the 119 respondents completing at least one clinical scenario. Furthermore, the response rate for endocrine surgeons was 24%, for other specialists a clear response rate cannot be calculated; this is a potential source of selection bias in our analysis and limited our ability to analyse factors such as central neck dissection by speciality. Despite this low response rate, there were a large number of high-volume clinicians in this survey, suggesting that the results may be broadly indicative of current practice in Australia. The definition and structure of MDT were not clearly defined in the questionnaire but left open to individual clinician interpretation; this is a possible source of heterogeneity in this study. Within Australia, there is no formal guidance regarding thyroid cancer MDT composition. However, the majority of MDT committees adhere to the standardised membership and governance as outlined in international guidelines. 29 MDT in Australia is generally defined as a meeting with at least four medical practitioners from different medical specialities and includes allied health professionals. This broad definition reflects the flexibility required in cancer care due to the vast geography of Australia where MDTs must meet the needs of both metropolitan and regional cancer care centres. The clinical vignettes were carefully designed to capture a variety of low-risk thyroid cancer scenarios and were piloted by a group of surgeons and endocrinologists prior to wider dissemination. It is acknowledged that any group of scenarios cannot fully represent all the nuances of clinical care; particularly in the era of individualised patient care and shared decision-making. Active surveillance was only listed as an option in Scenario 2 and was preferred by a minority of clinicians; we do not consider that this study has adequately assessed perceptions of active surveillance in the Australian setting. Finally, although clinical perceptions of FCR were surveyed, it is obviously critical that the patient perspective is fully understood. This is beyond the scope of this research but is the subject of other research by our group. 5 CONCLUSION The uptake of de-escalation (hemithyroidectomy and omission of radioactive iodine) in low-risk thyroid cancer is likely to be commonplace in Australia, particularly by clinicians involved in thyroid cancer MDTs. Most patients are offered at least 5 years of follow-up with annual review being most common. The majority of clinicians felt that patients' FCR was excessive, but that clinician follow-up provided reassurance. These findings are critical to understanding patterns of thyroid cancer management within Australia and will be useful to any researcher wishing to study thyroid cancer survivorship. 6 AUTHOR CONTRIBUTIONS Winy Widjaja: Formal analysis (equal); validation (supporting); writing - original draft (lead); writing - review and editing (equal). Christopher Rowe: Conceptualization (equal); data curation (equal); formal analysis (equal); investigation (equal); methodology (equal); project administration (equal); supervision (equal); validation (equal); writing - review and editing (equal). christopher oldmeadow: Formal analysis (lead); resources (equal); supervision (equal); writing - review and editing (equal). Daron Cope: Data curation (equal); investigation (equal); methodology (equal); project administration (equal); resources (equal); supervision (equal); writing - review and editing (equal). Elizabeth fradgley: Conceptualization (equal); data curation (equal); investigation (equal); resources (equal); supervision (equal); writing - review and editing (equal). Christine Paul: Conceptualization (equal); formal analysis (equal); investigation (equal); methodology (equal); supervision (equal); writing - review and editing (equal). Christine O'Neill: Conceptualization (equal); data curation (equal); formal analysis (equal); investigation (equal); methodology (equal); project administration (equal); resources (equal); supervision (lead); writing - original draft (supporting); writing - review and editing (lead). 8 CONFLICT OF INTEREST STATEMENT None declared. 9 7 ACKNOWLEDGEMENTS Elvina Wiadji, Surgical Registrar, John Hunter Hospital; Nicholas Blefari, Surgical Registrar, John Hunter Hospital; Rosemary Carroll, Research Nurse and co-ordinator, John Hunter Hospital; Conjoint A/Prof Stephen Smith, Colorectal Surgeon, John Hunter Hospital. 9.1 DATA AVAILABILITY STATEMENT The data that supports this article are available from the corresponding author upon request. APPENDIX 1 Confidential 1.1 Follow-up of Patients with Well-Differentiated Thyroid Cancer: a Survey of Clinician Preferences 25th August 2020 Dear Colleague, It has recently been identified that patients with thyroid cancer may have significant quality of life detriments. We are exploring these issues in Australian patients and as part of this research we seek to understand the current treatment and follow-up preferences of specialist clinicians in Australia. This survey will explore current practice in the follow-up of patients with well differentiated thyroid cancer through four case vignettes. It is not a test of your knowledge, but we do aim to explore your perception of risk of recurrence with each case and how this might influence your decisions in the follow-up of individual patients. We hope to record a wide variety of responses. On average this survey will take 10 min to complete. This survey is being sent to all Australian Members of the Endocrine Society of Australia, ANZ Endocrine Surgeons and the Australian Society of Otolaryngology, Head and Neck Surgery. If you treat any patients with thyroid cancer at any stage of their treatment, we would ask you to complete this study. It is important for our future research that we have has a high a participation rate as feasible. The results of this study will be used to inform future research in designing interventions to ameliorate quality of life detriments in patients with thyroid cancer. Research data from this study forms part of the PhD of Dr Christine O'Neill and will also be prepared for publication. By completing this survey you are indicating your consent to participate in this project. Participation is voluntary, responses will be anonymous. If you do not want to participate the close this window. Survey responses once submitted cannot be reviewed, edited, or withdrawn. Thank you and kind regards, Dr Christine O'Neill Endocrine Surgeon John Hunter Hospital and University of Newcastle, Newcastle NSW Dr Christopher Rowe Endocrinologist John Hunter Hospital and University of Newcastle, Newcastle NSW Dr Daron Cope 23/02/2021 8:59 ENT Surgeon John Hunter Hospital and University of Newcastle, Newcastle NSW This research has been approved by the Hunter New England Human Research Ethics Committee of Hunter New England Local Health District, reference 2020/ETH01892. Should you have concerns about your rights as a participant in this research, or you have a complaint about the manner in which the research is conducted, it may be given to the researcher, or, if an independent person is preferred, to Dr Nicole Gerrand, Manager Research Ethics and Governance, HNE Research Office Level 3, POD, Hunter Medical Research Institute, Lot 1 Kookaburra Circuit, New Lambton Heights NSW 2305, telephone (02) 49214950, email [email protected] APPENDIX 2 Clinicians demographic and practice characteristics 2.1 Demographic Category Participants who completed at least one scenario (%) (N = 87) Participants who did not complete any scenarios (%) (N = 22) p-Value Age (years) n = 107 <45 37 (43%) 8 (38%) .82 46-55 31 (36%) 8 (38%) 55-65 13 (15%) 3 (14%) >65 5 (5.8%) 2 (9.5%) Gender n = 107 Male 48 (56%) 11 (52%) .835 Medical Speciality Endocrinologist 40 (46%) 12 (55%) .002 Endocrine (General) surgeon 28 (32%) 2 (9.1%) ENT, head and neck surgeon 18 (21%) 4 (18%) Other 1 (1.1%) 4 (18%) Practice experience (years) n = 108 1-5 16 (18%) 6 (29%) .60 6-10 17 (20%) 2 (9.5%) 11-15 18 (21%) 4 (19%) >15 36 (41%) 9 (43%) Volume of WDTC cases (per year) n = 108 <10 26 (30%) 11 (52%) .23 10-20 30 (34%) 6 (29%) 21-50 19 (22%) 3 (14%) >50 12 (14%) 1 (4.8%) MDT attendance n = 107 Yes 45 (63%) 8 (38%) .04 Geography n = 102 Capital city 56 (68%) 16 (80%) .30 APPENDIX 3 Univariate and multivariable analysis of treatment decision (total thyroidectomy) and follow-up decision (>5 years follow-up) 3.1 Variable Univariate OR (95% CI) Univariate p value Multivariate OR (95% CI) Multivariate p value Total thyroidectomy MDT attendance 0.29 (0.15-0.54) <.001 0.29 (0.12-0.68) .005 Surgeon (endocrine or ENT) 0.39 (0.21-0.71) .002 0.65 (0.33-1.25) .19 Female 1.80 (0.95-3.43) .073 1.21 (0.62-2.34) .58 Case volume per year >10 patients 0.57 .108 1.34 .484 Practicing Experience of >10 years 0.76 .418 1.51 .251 Follow-up >5 years MDT Attendance 0.34 (0.10-1.11) .072 0.23 (0.05-0.99) .049 Surgeon (endocrine or ENT) 0.28 (0.09-0.93) .037 0.35 (0.10-1.2) .094 Female 3.11 (0.92-10.54) .069 2.16 (0.64-7.23) .21 Case volume per year >10 patients 0.98 .975 2.49 .251 Practicing Experience of >10 years 1.64 .433 2.37 .239 |
PMC10000627 | After a short introduction about the history of liquid biopsy, aimed to noninvasively replace the common tissue biopsy as a help for cancer diagnosis, this review is focused on extracellular vesicles (EVs), as the main third component, which is now coming into the light of liquid biopsy. Cell-derived EV release is a recently discovered general cellular property, and EVs harbor many cellular components reflecting their cell of origin. This is also the case for tumoral cells, and their cargoes might therefore be a "treasure chest" for cancer biomarkers. This has been extensively explored for a decade, but the EV-DNA content escaped this worldwide query until recently. The aim of this review is to gather the pilot studies focused on the DNA content of circulating cell-derived EVs, and the following five years of studies about the circulating tumor EV-DNA. The recent preclinical studies about the circulating tEV-derived gDNA as a potential cancer biomarker developed into a puzzling controversy about the presence of DNA into exosomes, coupled with an increased unexpected non vesicular complexity of the extracellular environment. This is discussed in the present review, together with the challenges that need to be solved before any efficient clinical transfer of EV-DNA as a quite promising cancer diagnosis biomarker. extracellular vesicles (EVs) exosomes (EXs) liquid biopsy (LB) EV-associated DNA (EV-DNA) cancer diagnosis This research received no external funding. pmc1. Introduction Cancer is a major burden on humanity, as recapitulated by Globocan, the Global Cancer Statistics 2020, concerning 36 cancers, with regard to their respective incidence and mortality in men and women from 185 countries worldwide accessed on 31 October 2021). Human cancer is still a mysterious multiform disease, and each organ-specific cancer has to be considered as a unique disease, with some common hallmarks. They also share in common the necessity of an early diagnosis for an optimal outcome for the patient. Besides the many sophisticated technologies now available for asserting a cancer diagnosis, liquid biopsy, first in blood (serum, plasma) and now in many other body fluids (urine, cerebrospinal fluid, saliva), has brought the hope of an efficient cancer signature for significantly helping an early diagnosis. Many components released from the tumor cell machinery during life and death might be candidates for being noninvasive cancer "whistleblowers", which explain the already long-lasting query about the most promising liquid biopsy biomarkers. Liquid biopsy in blood already has a long history as a promising substitute for tissue biopsy for cancer diagnosis. Cancer biomarkers were first focused on rare circulating tumor cells (CTCs), followed by cell-free tumor DNAs (cf-tDNAs). Recently, circulating tumor extracellular vesicles (cir-tEVs) became the third most interesting resource for cancer liquid biopsy . The high EVs heterogeneity has been classified into three main EV categories, according to their size, biogenesis, composition and biological properties : apoptotic bodies (ABs) (50 nm-5 mm in diameter), microvesicles (MVs) (100 nm-1 mm) and exosomes (EXs) (30 nm-150 nm). Due to the lack of specific vesicle biomarkers and the EVs overlapping size properties, it is presently difficult to efficiently discriminate the different EVs; therefore, they currently share the generic name of extracellular vesicles (EVs) . The release of different types of extracellular vesicles (EVs) is recognized as a new important common cell property, extending each cell influence well beyond its plasma membrane. For about a decade, EVs have been recognized as important messengers of intercellular communication, and nowadays, their major biological functions in human health and disease are highly investigated. With regard to their recent involvement as circulating EVs (cirEVs) in many body fluids for diagnosis of human diseases including cancers, the smallest vesicles, i.e., mainly exosomes, are the most considered. As recently reviewed , an increasing worldwide search is focused on finding the most relevant biomarkers to achieve early diagnosis of different human cancers among the many macromolecular components, which are specifically carried inside the rich cargoes of the numerous circulating tumor EVs (cir-tEVs). Although cf-tDNAs was the second important resource for cancer liquid biopsy, EV-DNA remained long ignored as a tumoral biomarker. The aim of the present review is to point out the recent studies which shed light on the potential capacity of cir-tEV-DNA as a new, interesting biomarker candidate for early diagnosis and prognosis of human cancers. The current knowledge evolution about the composition of the extracellular medium will also be discussed, as well as the challenges to solve before any usable routine clinical transfer. 2. Pilot Studies Focused on the DNA Content of Circulating Cell-Derived EVs (2011-2016) After the mere observation that tumor cells release more EVs (tEVs) than their normal cell counterparts (nEVs), it was obviously interesting to check the comparative cargo composition of tEVs and nEVs. This EV cargo comparison was first focused on EV proteins, then on EV-RNAS (coding messenger RNAs (mRNAs), and mainly noncoding RNAs (microRNAs (miRNAs), long noncoding RNAs (lncRNAs), and circular RNAs (cirRNAs)). At first, small EVs were not supposed to harbor DNAs, which was only assumed as a known property of apoptotic bodies. However, in 2011, Balaj et al. asserted that tumor cells release an abundance of microvesicles containing a selected set of proteins and RNAs. However, they also carry DNA, which reflects the genetic status of the tumor, including a significant sequence amplification of the c-Myc oncogene for three medulloblastoma cell lines compared with normal fibroblasts and other tumor cell types. ExoDNA appeared to be primarily single stranded (ssDNA). Tumor microvesicles contain genetic information available for horizontal gene transfer and provide a potential source of tumor biomarkers. In 2012, Waldenstrom et al. , after having previously revealed that human prostasomes contain chromosomal DNA, successfully searched DNA in microvesicles/exosomes derived from a murine cardiomyocite cell line; they also showed that these EVs, containing DNA/RNA, could transfer chromosomal DNA sequences to the cytosol or nuclei of target fibroblasts. These two pioneering works on MVs initiated the interest in EV-DNA. In 2013, six years after the noticeable observation of Valadi et al. , claiming that "exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells", Cai et al. showed that "extracellular vesicle-mediated transfer of donor genomic DNA to recipient cells is a novel mechanism for genetic influence between cells". They first examined the existence of genomic DNA (gDNA) in EVs derived from human plasma and from vascular smooth muscle cells (VSMCs) in culture. They found at least 16,434 gene fragments in the human plasma, ranging in size from 1 to 20 kilobases (kb), but mostly around 17 kb. They showed with VSMCs that apoptosis was not the source of EV-DNA. Moreover, they observed that DNA was present only inside the thoroughly washed EVs and that EVs contain double-stranded DNA (dsDNA). Then, they investigated the function of transferable DNA in the recipient cells. To determine the pathophysiological significance of EV-gDNA transfer into cells, they further examined the transfer of BCL/ABL hybrid gene in EVs from K562 cells to normal human neutrophils isolated from human peripheral blood. They found that the numerous gDNA fragments in EVs are transportable between the same or different types of cells and increase the gDNA-coding mRNA and protein expressions in the recipient cells. This immediately boosted the interest of circulating EV-DNA as a new cancer biomarker in liquid biopsy. In 2014, independently of the three pilot studies detailed above, Kahlert et al. investigated whether exosomes from two pancreatic cancer cell lines and serum from (a few) patients with pancreatic ductal adenocarcinoma (PDAC) contain gDNA. They provided evidence that exosomes contain >10 kb fragments of ds-gDNA spanning all chromosomes. They showed that the known specific KRAS and p53 DNA mutations found in the pancreatic tumor cells were recovered in the serum exosomes of patients with pancreatic cancer. Therefore, serum exosomes might be used to determine gDNA mutations for cancer prediction. Moreover, their data suggested that the majority of circulating DNAs from the serum samples may come from inside the exosomes and are not present as free-floating circulating DNA. This important preliminary study opened the way to a preclinical liquid biopsy study, involving a larger number of PDAC patients, compared to the appropriate healthy controls. At the same time, Thakur et al. used a quite interesting approach to evidence, for the first time, dsDNA in exosomes derived from two human cancer cell lines (myeloid leukemia K562 and colorectal carcinoma HCT116) and one murine melanoma cell model. They extracted DNA from exosomes either intact or pretreated with DNAse. However, instead of using the nonspecific DNAse I, they used either S1 nuclease, specific for ssDNA, or shrimp-dsDNAse, specific for dsDNA. Thus, they convincingly demonstrated that the majority of exosome-dsDNA with a size greater than 2.5 kb is associated with the outer membrane, whereas internal exosome-dsDNA depicts a size between 100 bp and 2.5 kb. This first observation of exosome-associated dsDNA was confirmed by atomic force microscopy (AFM) and extended a broad panel of human tumor cell lines to the analysis. A predominant dsDNA form of internal exoDNA was detected in all exosomes, but with lower amounts in most pancreatic and lung cancer cell lines. It is noticeable that exosomes from two normal fibroblast stromal cell lines exhibited about 20-fold less exoDNA than the one isolated from tumor cells. Furthermore, exosomes derived from murine B16-F10 melanoma revealed that only a (10%) subset of exosomes contained DNA, suggesting a specific targeting of DNA into exosomes. Another very important point of this work is that exoDNA represents the entire genome and might then mirror the tumor state. Focusing on a major modification of nuclear DNA, i.e., the methylation of 5'-cytosine, exoDNA was found methylated to a similar level to gDNA. ExoDNAwas also tested for some cancer-specific mutations such as the BRAF (V600E) mutation, present in 50% of malignant melanoma. They detected the mutant alleles in exoDNA of all cell lines containing the mutation and only the wild type (WT) in exoDNA originating from the cell lines with non mutated BRAF. The same search was performed with the epidermal growth factor receptor (EGFR), which is mutated in several cancers, including non small cell lung cancer (NSCLC), and respective EGFR mutations were also detected in 100% of exoDNA isolated from the NSCLC cell lines, harboring EGFR mutations. Thus, Thakur et al. showed that double-stranded DNA in exosomes reflects the mutational status of parental tumor cells, illustrating its significant translational potential as a novel circulating biomarker candidate in cancer detection. Lee et al. used (RAT-1), an immortalized nontumorigenic rat intestinal epithelial cell line (IEC-18) and its tumoral derivative (RAS-3) transfected with the V12 mutant c-H-ras human oncogene. By whole genome sequencing (WGS), these EVs, containing chromatin-associated dsDNA large fragments (777 bp, 2200 bp), were shown to cover the entire rat genome, including the full-length H-ras oncogene (3308 bp). Moreover, these EVs could transfer this oncogene to nontumorigenic cells and induce their increased proliferation. After evidencing gDNA inside microvesicles/exosomes , the presence of DNA was also questioned in other EVs . Shelke et al. claimed that the EV-DNA released by human mast cells is mostly associated with the outside of EVs and cause their aggregation. Fisher et al. showed that EVs (50-150 nm in size) released from human bone marrow-derived mesenchymal stromal cells (BM-hMSC) also carry high-molecular DNA. This DNA, which is not derived from apoptotic or necrotic cells, was mainly associated with the outer EV membrane and, to a smaller degree, inside the EVs. The DNA isolated from EVs was not organized in nucleosomes. The EV-gDNA amount was sufficient for next-generation sequencing (NGS) and virtually covered the complete human genome. After transducing a plant-DNA into BM-hMSCs, the released EVs were tagged with the Arabidopsis thaliana-DNA (A.t.-DNA) and able to rarely perform the (A.t.-DNA) EV-mediated transfer to naive BM-hMSCs. As previously observed with rat cells , this is a confirmation of the EV-mediated horizontal DNA gene transfer to recipient cells as a new important EV biological function. In 2016, Kalluri and Lebleu summarized the discovery of double-stranded genomic DNA in circulating exosomes , focusing on studies related to the origin of gDNA in exosomes and its utility in cancer diagnosis and disease monitoring. Lastly, Jin et al. proved that EVs extracted from serum are stable under different storage conditions (at 4 degC for 24 h, 72 h, 168 h; at room temperature for 6 h, 12 h, 24 h, 48 h; and after one-time-, three-time-, five-time-repeated freeze-thaw cycles). DNA in serum EVs is also stable under different storage conditions. Serum DNA is mainly present in exosomes, and EVs-DNA stayed stable for 1 week at 4 degC, 1 day at room temperature, and after fewer than three-time-repeated freeze-thaw cycles. The observed stability of serum EVs and EVs-DNA is the premise for using cirEVs for the search of new potential genetic DNA biomarkers for cancer diagnostics. A summary of these precursor studies on the DNA content of cell-released EVs is given in Table 1. 3. Following Studies about the Circulating Tumor EV-DNA (2014-2019) First, Lazaro-Ibanez et al. showed different gDNA fragments in the subpopulations of EVs (Abs, MVs, and EXs) with prostate cancer (PCa) cell lines (LNCaP, PC-3, and R92a/hTERT) in vitro. Derived from morphologically heterogeneous cancer cells, their respective MVs and EXs had comparable sizes and concentrations (1.36-2.52 x 108 particles/mL per million cells) for MVs (n = 16) and (0.56-1.93 x 108 particles/mL per million cells) for EXs (n = 16). However, for each of the three cell lines, the MVs' total protein content (6 mg protein/106 cells) was about twice that of EXs (3.2 mg protein/106 cells). Besides very rare MLH1 mutations in prostate cancer (PCa), TP53 and PTEN were the only significantly mutated genes in both localized PCa and castration-resistant (CRPC) tumors. The number of amplified gDNA fragments of MLH1 (108 bp), PTEN (225 bp), and TP53 (316 bp) were almost double between MVs and EXs (n = 12), showing that different types of EVs carried different gDNA contents, which suggests a selective gDNA package into the different PCa cell-derived EV subtypes. Moreover, they demonstrated that the EV-derived gDNA fragments from the LNCaP cells had no MLH1 mutation but a frame-shift PTEN mutation and a (C > G) TP53 mutation, showing that EV-gDNA could even harbor specific gDNA mutations of the parent cells. Then, they provided evidence that plasma-derived EVs are more abundant in PCa patients (n = 4) than in healthy donors (n = 4) and that human plasma-derived EVs also carry double-stranded gDNA fragments. However, they did not observe any significant differences in the MVs and EXs or in the total EV population isolated from human plasma samples of PCa patients compared with healthy controls. Moreover, the previously described gDNA mutations for the LNCaP cell-derived EVs were not detected from the small studied cohort of plasma EVs. Thus, the promising in vitro observations are to be confirmed by other extended preclinical studies, before asserting EV-DNA as a valuable biomarker for PCa diagnostics. After previous isolation from tumor cells with high migratory and invasive abilities of new, unusually large (1 mm in diameter) EVs (L-EVs), also named large oncosomes (LO), Vagner et al. first characterized the DNA in large L-EVs (LO surrogate) and small S-EVs (EX surrogate) from the same PC3 (PCa) or U87 (glioblastoma) cancer cell lines, as well as from plasma of a PCa mouse model. L-EVs emerged as the EV subpopulation containing most of the circulating DNA, which was quantified as a high molecular weight (up to 2 Mb) chromatinized DNA. Then, they isolated L-EVs and S-EVs from human plasma of patients (n = 40) with metastatic castration-resistant PCa (mCRPC). As observed in vitro, and despite a pronounced interpatient variability in the amount of EV DNA, L-EVs contained significantly more DNA than S-EVs, whereas DNA was totally absent from both L-EVs and S-EVs in controls. Moreover, L-EVs isolated from human mCRPC patients contained large-size dsDNA, covering the entire tumor genome, with reported cancer-specific genomic alterations (MYC/PTEN imbalance). It is noticeable that, in line with their in vitro and in vivo results, the ssDNA/dsDNA ratio was 5/1 in three out of four patients and the amount of EV-free DNA was comparable or higher than the amount of DNA in L-EVs in two patients. This points out the necessity for further preclinical studies to shed light on the relationship between disease progression and the composition of the DNA cargo in L-EVs. Pancreatic cancer, in urgent need of early diagnosis, was also considered under the light of DNA biomarker. Allenson et al. compared exosome-derived DNA (exoDNA) to cfDNA in liquid biopsies of patients with pancreatic ductal adenocarcinoma (PDAC) on 263 individuals, including a discovery cohort of 68 PDAC patients of all stages, 20 PDAC patients with localized tumor after curative resection, and 54 healthy controls. A validation cohort of 39 cancer patients and 82 healthy controls was studied to validate KRAS detection rates in early-stage PDAC patients. KRAS mutations were more detectable in exoDNA than in cfDNA. However, mutant KRAS was also detected in a substantial minority of healthy samples, which limits its utility as a cancer-screening method. Yang et al. added to the search of KRASG12D mutation in serum exosomal DNA, the associated search of TP53R273H mutation from patients with pancreatic cancer and healthy individuals. The minimal exosomal DNA used for digital PCR analyses was 0.663 ng. A sufficient amount of exosomal DNA for the KRASG12D and TP53R273H mutations search was obtained for 49% (76/156) of patients and 66% (114/171) of healthy serum samples. In 39.6% of the serum samples of PDAC patients (n = 48), the KRASG12D mutation was identified, whereas the TP53R273H mutation appeared in 4.2% of the serum samples, leaving 27 samples without these two specific mutations. With the frequency of the KRASG12D mutation being measured as about 40-50% in PDAC tumor tissue, this exosomal DNA study likely captures most of the KRASG12D mutation in PDAC patients. This also appears to be the case for the TP53R273H mutation. Thus, this study showed that exosomal DNA can be used as a substitute for less convenient tissue biopsy to identify mutations using digital PCR. Moreover, whereas KRASG12D mutation was detected in 2.6% of a large cohort (n = 114) of healthy individuals, TP53R273H mutation was never detected in healthy subjects. On the other hand, in vitro and in vivo studies showed the interest of engineered exosomes (iExosomes) to carry short interfering RNA (siRNA) or short hairpin RNA (shRNA), specific to oncogenic KrasG12D, for efficiently targeting KRAS. Mendt et al. reported a bioreactor-based generation and testing of large-scale production of clinical-grade iExosomes for targeting KRAS in pancreatic cancer. These iExosomes were thoroughly tested in vitro with many cell lines and in vivo on several mouse models with pancreatic cancer. These studies confirmed the suppression of oncogenic Kras and an increase in the survival of mouse with pancreatic cancer, illustrating their therapeutic potentialities. Garcia-Romero et al. showed that all three types of EVs (Abs, MVs, and EXOs) secreted by human glioma cells contained gDNA sequences. Some sequences appeared in all EVs, whereas a few sequences appeared exclusively in one type of EVs. IDH1, harboring the most relevant mutation for human glioma diagnostic, was detected only in MVs and EXOs. Moreover, in vivo studies demonstrated that all types of tumor-derived EVs cross the intact blood-brain barrier and can be detected in the peripheral blood. In a small cohort of glioma patients, they demonstrated that the IDH1G395A mutation could be successfully detected in the peripheral blood EVs cargo as a minimally invasive method compared to liquid biopsy from cerebrospinal fluid. In 2019, Kahlert wondered whether the exosomal gDNA, discovered only some years ago, might be a better choice as a cancer biomarker in liquid biopsy than the cfDNA discovered six decades before. After recapitulating the origin of both DNAs and their respective advantages and disadvantages, he concluded that both are currently complementary. Whereas cfDNA can be detected in healthy individuals and patients with nonmalignant or malignant disease, mutated cfDNA is more tumor-specific and enriched in smaller fragments between 90 and 150 bp and in the size range 250 to 320 bp, originating from cell death remnants insufficiently cleared by infiltrating phagocytes; therefore, cfDNA, with an easier amount of accessible DNA and higher copy numbers of some cancer-specific mutations, is more efficient for prognosis of late tumor stages. By contrast, exosomal gDNA can be found less fragmented (with a size range between 2.5-10 kb ) not only in exosomes, but in all EV types, more frequently in MVs and EXOs and sometimes only in some specific EXO subsets , with an apparent distribution depending on the tumor type. Although in a smaller amount, exosomal gDNA, spanning all the chromosomes, is sufficient to obtain the significant tumor-specific mutated DNA sequences by using the most recent PCR technologies. Thus, exosomal DNA might be a better potential biomarker for early cancer diagnosis than cfDNA. However, the clinical translation of exosomal DNA as a cancer biomarker is greatly hampered by the urgent need for finding a valuable substitute to the "gold standard" of differential ultracentrifugation for EVs extraction from human body fluids. The greatest promise for using the tumoral EV-specific gDNAs as an early cancer diagnosis biomarker might be to specifically extract tumor EVs from the whole circulating EV population by capture on "lab-on-chip" solutions, for example, by targeting some tumor-specific EV outer membrane proteins, such as glypican-1, followed by the use of the new PCR technologies for reaching the cancer-specific mutation(s) of interest. To define EV component(s) as potential biomarker(s) for a given human cancer diagnosis by liquid biopsy, three steps are generally undertaken: in vitro studies with specific tumor cell lines, in vivo studies with murine tumor models, and preclinical studies on circulating tumor-derived EVs from a few patients' plasma or serum. Whereas two-dimensional (2D) cell cultures are generally used as "gold standard" in vitro models, Thippabhotla et al. intended to compare the EVs respectively released by an immortalized HeLa (2D) cell culture, issued from a cervical cancer patient, and a three-dimensional (3D) organoid culture, elaborated on peptide hydrogel with the same HeLa cells. They found that the EV secretion dynamics were significantly different for both culture types. Moreover, their respective EV-RNA and EV-DNA compositions were also quite different. The 3D-culture-derived EV-small RNA profile (<200 nt) showed a much higher similarity (about 96%) than the 2D culture-derived EVs to plasma EV-small RNA profile from two cervical cancer patients with one healthy control. In contrast with RNA, analysis of the cir-tEV-DNA sequencing data showed that culture or growth conditions do not affect the genomic DNA information carried by EV secretion. Therefore, at least for cervical cancer, 2D culture seems to remain a valuable in vitro tool for the search of human cir-tEV-gDNA cancer biomarker, whereas the 3D culture system may constitute a more useful in vitro model for the search of cir-tEV-RNA cancer biomarkers. Yokoi et al. were the first to question the mechanisms of nuclear content loading to exosomes. Upon induction of genomic instability with genotoxic drugs, they identified a link between micronuclei (MN) formation and the generation of some specific exosomal loading with gDNA and other nuclear contents. On the other hand, Lazaro-Ibanez et al. , using two human mast (HMC-1) and erythroleukemic (TF-1) cell lines, prepared, by ultracentrifugation, exosome-enriched small extracellular vesicles (sEVs). The amount of sEVs for TF-1 cells was over 2.5-fold more than that for HMC-1. By further using a high-resolution iodixanol density gradient on the two sEVs populations, the authors discriminated two novel heterogeneous subpopulations with different DNA content and topology. Each sEVs fraction was separated in nine 1 mL fractions (F1-F9) with measured densities from top to bottom. For both cell lines, the respective (F1 = F7) fractions were clustered in two low-density (LD) (F1-F3) and high-density (HD) (F4-F7) sEV subsets. The majority of the classical exosome-like sEVs were contained in the LD fractions. DNA was less abundant than RNA, and DNA was mainly present as ssDNA in the HD fractions for both cell types. The (HMC-1) HD fraction had a DNA-to-RNA ratio of 2.2/1, while the (TF-1) HD fraction was enriched in RNAs with a 1/2.9 DNA-to-RNA ratio. The LD fractions had the most prominent rRNA peaks and least DNA, while the HD fractions had most of the DNA cargo and small RNAs with no ribosomal rRNA peaks. DNA was predominantly localized on the outside or surface of sEVs, with only a small portion inside the vesicles. The entire human genome was represented both on the inside and outside of the sEVs. When sEVs were analyzed in bulk, whole-genome sequencing identified gDNA fragments of various lengths (from 500 to 10,000 bp), spanning both mitochondrial DNA and all chromosomes. These interesting and somewhat amazing observations have to be further explained, especially the cell mechanisms for the sEV specific loading before release and the curious DNA topology. In 2019, Jeppesen et al. questioned the heterogeneity of the exosome-enriched crude sEVs sample. From their in-depth studies published in Cell, the authors claimed the necessary reassessment of the "classical" exosome composition both with regard to their assumed biogenesis and to their widely admitted global composition. The most "iconoclast" assertion for the topic of the present review was that extracellular dsDNA was not associated with exosomes or any other types of sEVs. Reviewing the ongoing studies from 2020 might perhaps clarify this pending question concerning exosomal DNA, which is important for keeping the current assumed interest of EVs as a potential rich tumor DNA resource for early cancer diagnosis (cf. detailed discussion in part 5.). A Summary of these (2014-2019) studies about circulating tumor EVs-DNAs can be found in Table 2. 4. Preclinical Studies about the cirtEV-Derived gDNA as a Potential Cancer Biomarker (2020-2021) In line with the prestigious, newly reassessed exosome description , Hoshino and 116 coauthors brought, also in Cell, a more-medical insight by investigating the proteomic profile of potential new liquid biopsy cancer biomarkers in 426 human cancer and non cancer samples derived from various cells, tissues, and body fluids. However, instead of using two-pooled LD and HD density fractions of the crude sEVs , the authors categorized the crude sEVs into three prominent subpopulations: small exosomes (Exo-S 50-70 nm), large exosomes (Exo-L 90-120 nm), and exomeres (non vesicular (NV) particles <50 nm), collectively referred to as extracellular vesicles and particles (EVPs), with the aim of defining EVP protein signatures that distinguish cancer patients from healthy individuals. Exomeres were identified in 2018 as nanoparticles distinct from EVs by using asymmetric flow field-flow fractionation (AF-4) for EV analysis . Among their in-depth studies , the authors analyzed 120 plasma-derived EVP proteomes from 77 cancer patients with 16 different cancer types and 43 healthy controls (HC). They highlighted the identification of EVP markers, characterized EVP markers in human tissues and plasma, and suggested that EVP proteins can be useful for cancer detection and determination of cancer type. For the present review focused on exosomal DNAs, it is noticeable that not only is the choice of the EV-transported components (proteins/RNAs/DNAs) as the best type of cancer biomarkers still widely questioned, but even the more appropriate nature of the circulating extracellular transporter (EVs and/or NV materials) is also becoming a matter of debate. Although aware of the recent reassessment of the composition of EVs and the overturn of some previous findings , Teng and Fussenegger kept the EV common classification in three main types (Exos, MVs, and ABs) for extensively reviewing the EV biogenesis, focusing mainly on exosomes and microvesicles. They detailed the current knowledge about the three distinct steps concerning exosomes biogenesis and release, initiated from the endosomal pathway, with further intracellular transport of the multivesicular bodies (MVBs) containing intraluminal vesicles, and fusion of some MVBs with the plasma membrane for exosomes release. Likewise, they detailed the mechanisms of biogenesis and release of microvesicles and discussed the current knowledge upon EV uptake and cell-cell communication, as well as upon the cargo sorting into EVs. Lastly, with all this accumulated knowledge, they concluded by recapitulating the many possible EV bioengineering methodologies for therapy improvements in the future. Besides the increasing knowledge about EVs' biogenesis and composition, some recent reviews were focused on potential EV-derived DNAs as liquid biopsy biomarkers applied on a few specific cancers. Thus, Kim et al. were concerned with lung adenocarcinoma. After summarizing older liquid biopsy approaches to overcome the small tissue availability in lung cancer patients, they advocated for EVs as ideal carriers of cancer biomarkers. They recalled that, contrary to the passively released fragmented cfDNAs (about 200 bp), cirEV-DNAs consist of both large-sized ds-gDNAs (up to 10 kb) and fragmented mutated DNAs, giving an active image of both the viable and dying tumor cells. Moreover, a higher sensitivity can be achieved by using EV-DNAs obtained from bronchoalveolar lavage fluid (BALF) than those from blood. Compared with the short half-life (2-2.5 h) of cfDNAs, the membrane-protected EV-DNAs also have a high stability. In conclusion, cirEV-DNAs are expected to be more widely used in the future, when their current sophisticated isolation methods will become clinically adapted. By contrast, Sun et al. claimed an improved detection of cell-free tumor DNAs (cf-tDNAs) in EVs-depleted plasma of cancer patients. It is to be stressed that exosomes were prepared either by mere precipitation using ExoQuick (System Biosciences, CA, USA) or fractionated by using five sequential centrifugations and ExoQuick instead of ultracentrifugation. However, preparing exosomes by a precipitation method might not be a guarantee for keeping the exosomal DNA cargo intact, and it is noticeable that, in this case, the exosomal fraction 5 was dominated by small (~160 bp) nucleosome-like DNAs . It is also noticeable that an older research article , using two different methods for exosomes isolation, brought contradictory evidence that more than 90% of cfDNA in human blood plasma is localized in exosomes. However, agarose gel electrophoresis of DNA isolated from plasma exosomes showed two prominent bands, one high intensity and high molecular weight, and the other of low molecular weight (less than 200 bp in length). By RNase treatment, the first band turned out to be exosome copurified RNA, with a 5-fold higher amount than the exosomal dsDNA, corresponding to the second band. It would be worth performing some in vitro studies about the exosomal DNA yield and size as a function of the methods used for collecting the exosomes. Cambier et al. aimed to identify circulating nucleic acid sequences associated with serum EVs as a step toward an osteosarcoma (OS) early detection assay. qPCR analysis of PEG-precipitated EVs revealed the over-representation of some repetitive element DNAs in OS patient versus control sera. Taken into account that, in these serum EVs the OS-associated repetitive element DNAs were sensitive to DNase I, they were not in a protected EV cargo. Moreover, the repetitive DNA elements were copurified with EVs in PEG precipitation and size exclusion chromatography (SEC), but not in CD81 or CD9 EV immunoaffinity capture. These observations were taken as supporting the recent exosome reassessment , claiming that exosomes do not contain DNA, or tightly associate with other non vesicular entities containing dsDNAs that are extruded from cancer cells. Ruhen et al. aimed to use low-pass whole-genome sequencing to identify copy number variants (CNVs) in serial samples of both cf-tDNA and EV-DNA from plasma of a patient with metastatic breast cancer. Of the 52 CNVs identified in tDNA, 36 (69%) were detected in at least one cf-tDNA sample and 13 (25%) in at least one EV-DNA sample. Variants ranged in size from 0.3 to 106.5 Mb and were distributed randomly throughout the genome. Both kinds of noninvasive liquid biopsy depicted a CNV increase with disease progression, but this case study demonstrated that cf-tDNA, shed from apoptotic tumor cells, had a greater sensitivity for serial monitoring of breast cancer than EV-DNA actively secreted from viable neoplastic cells. Elzanowska et al. summarized the biological and clinical aspects of EV-DNA and examined the current role of EV-DNA specifically in cancer. Overall, they emphasized that EV-DNA as a biomaterial for liquid biopsies is a new but definitely promising area of study, but its study in the clinical context is still quite open for further validation. Lee et al. performed targeted NGS of DNA derived from bronchoalveolar lavage fluid (BALF-EV DNA) of 20 patients with EGFR-mutated non small cell lung cancer (NSCLC) and DNA from matched formalin-fixed paraffin-embedded (FFPE) tissue samples. EVs from BALF were heterogeneous (100-300 nm in size); EV-DNAs from the BALF existed in short and long sizes, but mostly in about 11 kb; and EVs contained DNAs from both vesicle surface and inside. The DNA yield from BALF-EVs was 100-times less than tissue DNA but had enough tumor-specific DNA for use in NGS analysis for the identification of actionable genetic alterations. This approach has a high potential clinical feasibility and utility. Kim et al. , also enrolling NSCLC patients after tyrosine kinase inhibitor therapy, compared different technological tools to detect EGFR mutations in 54 plasma samples and 13 pleural fluids. They demonstrated that combined tumor nucleic acid analysis (exoTNA+cfTNA) in the plasma and exoTNa in the pleural fluid allowed for the detection of target mutations more sensitively than that using cfDNA or total DNA alone. Amintas et al. , claiming that "dsDNA in EVs might be the latest most promising biomarker of tumor presence and complexity", focused on the recent knowledge on the DNA inclusion in vesicles, the technical aspects of EV-DNA detection and quantification, and the use of EV-DNA as a clinical biomarker. They recapitulated the cell-free DNA cell sources by active or passive mechanisms and summarized the tumor genome hallmarks reflected by EV-DNA, as well as the results of the main clinical studies assessing the performance of EV-DNA biomarkers (cf. their Table 1). Although suggesting EV-tDNA as an alternative to reach the promise of cftDNA, they concluded by enumerating the many challenging questions remaining to be solved before reaching this goal. Maire et al. investigated whether the DNA in glioblastoma cell-derived EVs reflects genome-wide tumor methylation and mutational profiles and allows noninvasive tumor subtype classification. They found that DNA is present in the vast majority of EVs, with a major localization to the EV surface. Genome-wide methylation profiling identified with high accuracy in EV-DNA the methylation of the parental tumor-specific mutations and copy number variations (CNVs). Interestingly, the methylation profiling and CNV results were not affected by the EV isolation techniques. This showed that EV-DNA reflects the genome methylation, CNV, and mutational status of glioblastoma cells. Likewise, Baris et al. compared epigenetic alterations in the target gene Enhancer of Zeste Homolog-2 (EZH-2) between plasma-derived exosomes and matched primary tumor tissues of 21 patients with aggressive diffuse large B cell lymphoma (DLBCL). They showed, for the first time, the presence of DNA in plasma exosomes of DLBCL patients and found that CDKN2A and CDKN2B were methylated in both plasma exosomes and primary tumor tissue samples. Compared to 21 healthy individuals, exosome concentration was approximately six-times higher in DLBCL patients, but the exosomal dsDNA content was extremely low compared to RNA contents. Zavridou et al. were also the first to perform a direct comparison of gene expression and DNA methylation markers in CTCs and paired plasma-derived exosomes. This revealed a remarkable heterogeneity on gene expression and DNA methylation markers between EpCAM-positive CTCs and paired plasma-derived exosomes in metastatic castration-resistant prostate cancer (mCRPC) patients, with a significantly higher positivity in CTCs. Lastly, Hur and Lee extensively reviewed the properties of extracellular vesicle-derived DNA for future clinical applications. They examined the biogenesis of DNA-containing EVs, their DNA methylation, and the use of next-generation sequencing (NGS). They questioned the use of EV-DNA as a biomarker in clinical settings, the modality of EV-DNA gene transfer, and its therapeutic potential. They hypothesized that DNA might exist inside an EV in a protected nucleosome or supercoiled form, which would enable the packaging of long dsDNA. Taking into account the nucleosome's 11 nm size, long dsDNA would more likely be present in larger EVs. However, the presence and topology of DNA in extracellular EVs will continue to be controversial until the development of a method for isolating pure EV subsets. Nonetheless, the authors recalled that the (100 bp to 20 kbp) EV-dsDNA fragments can represent the entire genome and reflect the mutational status of tumor parental cells. Lastly, mentioning several recent liquid biopsy studies in different body fluids of EVs associated-dsDNA for cancer patients, they also expressed the strong interest of EV-DNA as a new potential cancer biomarker. A summary of the discussed preclinical studies (2020-2021), about the circulating tumor-derived as a potential cancer biomarker, is given in Table 3. 5. Evolution of the Knowledge about the Composition of the Extracellular Environment (2019-2021) During many years, the exosome concept was "the tree that hid the forest of EVs". However, EVs "came on stage" about one decade ago and have been studied worldwide since 2012, reaching a huge, still-uncontrolled complexity in heterogeneity. An EV classification into three main categories as a function of their size and biogenesis, i.e., apoptotic bodies (ABs), microvesicles (MVs), and exosomes (EXOs), obtained a general long-lasting consensus until Jeppesen et al. recently proposed a complete reassessment of exosome composition, with a new classification of low-density (LD) "exosome-like" small extracellular vesicles (sEVs), without any DNAs in their cargos, and a much more significant high-density (HD) non-EV extracellular mixed component associated with DNAs. Therefore, they used different cell lines, human plasma, and tissue for preparing sEVs samples by the commonly used differential centrifugations. Then, they further used a density discrimination by means of a discontinuous iodixanol gradient density. Being aware of the ultracentrifugation-induced aggregation artefacts, they kept, in parallel, parts of the 15,000x g filtered supernatants as precleared media. These media were submitted to direct immunoaffinity-capture (DIC) of exosomes by means of magnetic beads conjugated to exosomes-specific tetraspanins antibodies. The crude sEVs, their different density fractions, and the scarce directly captured CD63-, CD81-, or CD9-specific EVs were submitted to the same immunoblots. Different studies aimed to give an insight upon the proteins, RNA, and DNA composition of two-pooled low-density (LD) and high-density (HD) fractions of the crude sEVs samples. Surprisingly, many of the presumed components of exosomes were absent from the "classical" exosomes expressing CD63, CD81, and CD9. Many of the most abundant miRNAs were associated with extracellular nonvesicular (NV) fractions rather than with purified sEVs. Moreover, extracellular dsDNA was stressed as being not associated with exosomes or any other types of sEVs. An multivesicular endosome-related pathway was supposed to be the driver of extracellular DNA secretion instead of the exosome-dependent pathway. These assertions were sufficiently "iconoclast" to be seriously questioned before entering into the many details suggested for supporting the new exosome model. The results in were indeed "interesting and amazing", but when compared with those recalled in , both taken together were quite perturbing. The methods used for flotation, although with the same technology of discontinuous iodixanol gradient density, were not exactly the same (12-36% and 6-30% for the gradients and 20-45% for the one). Neither was on the same samples, and they used different means of sample deposit to the bottom of the centrifugation tube, i.e., 1 mL of crude sEVs suspension in PBS was mixed with iodixanol to a final 36% concentration or mixed with 3 mL of a 60% iodixanol solution . Moreover, the resulting increasing densities from top to bottom were differently measured, either with a refractometer on a mock identical gradient without sample or directly on all the 1 mL collected fractions by absorbance at 340 nm . The final results were indeed analogous for the LD fractions covering the "classical" exosomal sEVs. However, they were so different with regard to the HD fractions, corresponding either to a sum of many nonvesicular extracellular materials or to another "non-classical" exosome subset , that it would be worth further questioning the properties of the discontinuous iodixanol gradient density method as a function of the chosen parameters on the same crude sEV sample. Although quite new and highly cited by further publications , the conclusions asserted by Jeppesen et al. were only poorly confirmed . Their claimed absence of exosomal DNA did not appear to be quite convincing , especially when compared with the observations of Lazaro-Ibanez et al. . These authors, also using a discontinuous iodixanol gradient density separation of crude sEVs, reached only two heterogeneous (LD and HD) subpopulations of sEVs and only a small discarded heavier fraction of non-EV material. Sun et al. , taking into account the suggestion that extracellular DNA may not be associated with exosomes, but copurifies with the sEV fraction during standard isolation protocols , elaborated a clinically feasible protocol to analyze the cirEVs influence on the whole plasma cf-tDNAs' measurements. The authors selected nine small-cell lung cancer (SCLC) patients with a known relatively high cf-tDNA content; for each patient, they prepared, from a 1 mL blood sample, a platelet-poor conventional-plasma and, from another ml of the same blood sample, four pelleted fractions by successive light centrifugations, with replacement of the usual last ultracentrifugation by an ExoQuick exosome precipitation. Thus, the fractionated plasma corresponded, respectively, to "cells and larger debris" (fraction 1); crude "large microvesicles" (fraction 3); exosomes (fraction 5), which were characterized by transmission electron microscopy (TEM); nanoparticle tracking analysis (NTA); and by CD63/CD81 ratio using flow cytometry. The last supernatant (fraction 6) corresponded to the conventional plasma depleted of EVs. Then, the DNA yield and size distribution were compared in the whole plasma and in the different fractions for the nine (SCLC) patients. From 1 mL starting plasma, the average DNA yield was 5.3 ng in fraction 1, 1.73 ng in fraction 2, 0.99 ng in fraction 3, 0.68 ng in fraction 4, 4.17 ng in fraction 5, and 4.28 ng in fraction 6, and the average summed DNA yields in fractions 1, 5, and 6 accounted for 79.9% of the total DNA yields. Comparatively, whole plasma showed an average of 23.84% cftDNA in the same group of patients. The DNA size distribution was also measured in each DNA sample and showed a peak size of 7000-10,000 bp in fraction 1 and gradually reduced in fractions 2-3, while smaller fragments (about 160 bp) gradually increased from fractions 3 to 6. They also estimated cir-tDNA content in the different fractions and showed that the copy number variations (CNVs) were more detectable in fractions 3 (large EVs), 5 (exosomes), and 6 (EV-depleted plasma). Interestingly, the authors "were not able to remove any DNA copurified with exosomes", as previously suggested , and therefore, they questioned the origin of cir-tDNA detected in fraction 5. Maire et al. observed, in glioblastoma cell-derived EVs, that even after robust digestion of surface-associated DNA and any possibly contaminating free-floating DNA, they still detected DNA in 76.4% of the CD63/CD81-positive vesicles, strongly supporting the notion that EVs contained DNA inside. Some others tried reserved contradictory comments toward the suggested reassessment of exosome composition . Thus, Elzanowska et al. pointed out "the unreported amount of exosomes used in the study, as well as a limited cell lines included in the report". For Hur et al. , the inconsistency about the presence or absence of exosomal DNA can be attributed to the preparation method and size of the isolated EVs. Zhou et al. advocated exosomal DNA as possessing more abundant biological information and higher accuracy for prognosis prediction than cf-DNA in liquid biopsy. However, they recognized that it is unclear whether gDNA exists in exosomes in all mentioned studies with different DNA detection methods. Shen et al. asserted that "too strict an exosome isolation strategy may result in the loss of DNA-containing vesicles". Kalluri and Lebleu , summarizing the hallmarks of exosomes as being "a cell-to-cell transit system in the human body with pleiotropic functions", mentioned the current controversy about exosomal DNA and gave a negative appreciation of ref. , which "did not specify the quantity of exosomes used in its analytical assays, leading to ambitious conclusions". However, the pioneering studies of Jeppesen et al. , stressing the importance of extracellular nonvesicular particles as DNA biomarkers, was highly comforted by the discovery of exomeres, using the new technology of asymmetric flow-field fractionation (AF-4) for identification of subsets of extracellular vesicles . Furthermore, Zhang et al. demonstrated the exosome-like ability of exomeres to transfer functional cargoes . Malkin and Bratman , focusing on the increasing huge heterogeneity of the extracellular medium, brought an outstanding review article about "the nomenclature of EVs and extracellular particles (EPs), the physical and structural characteristics of EV/EP DNA, the physiological roles of EV/EP DNA in health and disease and the emerging potential of EV/EP DNA as a molecular biomarker." Interestingly, they extended the consensual long-lasting EV classification to nonvesicular EPs and modified the current nomenclature of extracellular components into large EVs (100 to >1000 nm), including apoptotic bodies (ABs), large oncosomes (LOs), microvesicles (MVs), originating from the plasma membrane; small EVs, including 50 to 130 nm exosomes (EXOs) of endosomal origin; and extracellular particles (<50 nm), including exomeres, with mean diameter of 35 nm, and chromatimers, both of yet unknown origin. Thus DNA, the overlooked component of EV/EPS is now becoming the central actor of many pending unanswered questions . 6. Challenging Questions to Solve before Clinical Use of cirEV-tDNAs and Technological State of the Art about EVs Isolation and Characterization The assets of circulating EV-DNAs, as a new promising biomarker for cancer diagnosis and prognosis, have been convincingly demonstrated. However, the clinical transfer of the accumulated preclinical knowledge that began about one decade ago is highly hampered by some important challenging questions needing to be solved as a priority. The suppression of the main "bottlenecks", both biological and medical, in the present knowledge about the extremely heterogeneous tumor-derived EVs/EPs populations, is highly dependent upon the future technological advances about their specific isolation and characterization. All the cells present in a human body, whether procaryotes or eukaryotes, are potentially equipped with the general cell property of releasing extracellular material, aimed either to remove no-longer-employed cell components or to send important epigenetic messengers into blood and/or into the many other minor subpopulation body fluids for modifying the fate of some specific recipient cells. Among this newly discovered "stellar" complexity of active extracellular material, it is not yet possible to precisely define the few tumor-specific subpopulations. At a smaller level of complexity, a given tumor cell population releases a quasicontinuum of EVs, with partly overlapping sizes and some common outer membrane protein markers. Therefore, the necessary classification of EV subsets without any specific biomarker is currently out of reach, which precludes further evidence for any of their specific biological functions. Moreover, the mechanisms used for specifically loading the multi components (proteins, lipids, nucleic acids, metabolites) into each EV cargo are almost completely unknown. This is also true for the EV-transported DNA, with some supplementary controversial questions about its topological localization inside the EV, outside on the EV membrane, or in both positions, and also on its size and nature as ssDNA, dsDNA, gDNA, or nucleosomes. The same questions will probably arise with the more recently discovered EPs, together with the one about the part played by the older known proteins such as Argonaute in the protected intercellular DNA transport. This detailed picture is aimed to show the huge problem of EV/EP heterogeneity, which has to be at least partly solved before efficiently facing the medical validation of a few promising EV-derived biomarkers for cancer diagnosis by well-standardized protocols for a given cancer, undertaken with important patient cohorts in different cancer centers worldwide. Some recent technological reviews have been selected to give an insight into the current state of the art concerning EV isolation and characterization . Valencia and Montuenga focused on the biological properties of exosomes and especially on their heterogeneity, which is due to the association of five factors: the cell of EVs origin, the EVs size and number, their molecular composition, and their functionality transferred in recipient cells. Different combinations of these factors result in highly complex EV heterogeneity. Moreover, in an oncologic patient, tumor-derived exosomes are estimated to be no more than 10% of all the circulating exosomes. Nevertheless, the authors suggest that exosomal DNA might become the future liquid biopsy gold standard. However, to become a clinical reality, every single procedure (EV isolation and characterization and all analytical protocols) remains to be standardized for a valid comparison of the different EV-DNA biomarker studies. Saad et al. detailed eight exosome-isolation methods and discussed the advantages and disadvantages associated with each method (cf. their Table 1). They also discussed the physical and chemical characterization and the detection techniques for exosomal samples. Widely studied since 2012, exosomes/EVs, with their potential to develop new clinical approaches of modern medicine, are also progressively entering the medical field, especially in cancer, cardiovascular disease, and central nervous system defects . Hirata et al. summarized the assets of liquid biopsy as a distinctive approach to the diagnosis and prognosis of cancer. They strongly advocated liquid biopsy compared with the usual tissue biopsy and its drawbacks. Although mentioning exosomes, they only actualized the comparison between the two older liquid biopsy circulating biomarkers CTCs and cell-free tDNAs as cancer diagnostic and prognostic tools. With regard to compared EV characterization between tumors and normal controls, Western blots and all the "omics" technologies, i.e., proteomic, transcriptomic, metabolomic, lipidomic, and genomic, gave, at each level, an interesting global insight of the tumor-induced modifications. Recently, Shaba et al. reviewed the EV multiomics integrated approach and summarized the state of the art of EVs omic studies. The abundant information reached for each omic level has to be correctly deciphered, and this is even more necessary if the different omics levels interact together. One essential requisite for multiomics integration is, beyond the generation of different omic datasets from the same biological samples, the development of statistical and annotation tools, which is essential for the interpretation of data. Still, many issues are encountered in each step of EV multiomic analysis, starting from EV isolation to the data integration methods, suggesting that this field is at its early state and requires further improvements. However, considering the complex EVs as optimal targets for omic sciences, the authors predicted a future challenging milestone for a multiomic integrative approach, which might contribute to explore EV functions, their tissue-specific origin, and their potentiality. On the other hand, it is feasible that each cancer-related global EV description might be "blurring" minute but important EV subsets, specifically linked with the tumor processes. Therefore, the new recent analyses at the single extracellular vesicle level (SVA), summarized by Bordanaba-Florit et al. , seem to be a quite interesting complementary approach to unravel the heterogeneity of extracellular vesicles. The authors extensively described some of the current methods so far developed for single-vesicle analysis (SVA). They reviewed the assets of SVA methods on recent advances in the EV field of research. They also focused on prostate cancer (PCa) diagnostics, showing the important improvements brought by the SVA of EVs. Ultimately, they concluded that an entirely new cell-to-cell EV-mediated communication network will be founded by single-vesicle techniques. SVA is also bridging the "omic" studies, carried for deciphering the global EV properties and their further functional studies, to the clinical world, by participating in the elaboration of simpler and less time-consuming technologies for EV isolation, such as microfluidics. Recently, Mousavi et al. extensively reviewed microfluidics for detection of exosomes and microRNAs in cancer. First introduced in the early 1990s, microfluidics manipulates microliter volumes in microchannels ranging in size only from 1 to 1000 mm. When compared with conventional studies, microfluidics platforms have many advantages, including enhanced reliability, sensitivity, accessibility, lower consumption of samples and reagents, reduced costs, quicker processing and response times, and the possibility of automated multiplexing. The authors summarized the microfluidic technologies used for exosome isolation and analysis and specifically applied for cancer studies. They also focused on microfluidic-based miRNA detection in human cancer. An increased interest has been shown in microfluidics use for biomarker discovery, but many challenges are yet to be faced, such as standardization and validation at a large scale, before any routine clinical application for cancer diagnosis. Recently, Campos-Silva et al. described a simple immunoassay for extracellular vesicle liquid biopsy in microliters of unprocessed plasma. They demonstrated that many EVs in solution, being like stable colloidal suspensions, are therefore unable to interact with a stationary functionalized surface. A more efficient capture on antibody-coated surfaces was obtained by using flocculation methods with cationic polymers. This led to the optimization of a protocol allowing effective immunocapture of EVs in bead-assisted flow cytometry. Only a few microliters of plasma were necessary for easy detection of tumor markers without previous ultracentrifugation. This easily adaptable method has been validated using plasma from lung cancer patients, with detection of the epithelial cell marker EpCAM on EVs. This radically improves the efficiency of clinical EV detection in immunocapture assays, opening new possibilities for the validation of EV biomarkers with large cohorts of patients. 7. Conclusions To gain a new step toward the clinical practice, it is mandatory to deeply investigate the still controversial nature and topology of the EV-associated DNA and the largely unknown EP-associated DNA. Moreover, microfluidics should focus on new technologies for discriminating circulating tumor EVs/EPs from the wide panel of the numerous other circulating EVs/EPs, blurring the tumoral message. As observed in this review focused on EV-DNA, the extracellular world is now becoming even more complex by the recent introduction of extracellular particles (EPs) , competing with EVs for assuming the many important intercellular messenger functions involved in human health and disease. It stresses the fundamental importance of deeply deciphering the extracellular environment composition and functions to complement the current knowledge slowly accumulated during two centuries about the cell machinery. As already mentioned , using a multiomics integrative approach at the single EV level is probably the ultimate goal to elucidate the most challenging EVs/EPs complexity in the far future. Therefore, it is probably only the very beginning of a long-standing scientific query, highly dependent on many future technological advances to control the EV/EP-epigenetic extracellular heterogeneity governing their putative, specific intercellular functions. Besides overcoming these major challenges, it will be necessary to define a standardized protocol for analyzing each given promising liquid biopsy biomarker for a given cancer type. Finally, the essential large-scale intercenter clinical validation might bring the putative biomarker to the long-awaited clinical practice. To give an optimistic insight into the huge interest in these hard future steps, one can mention a recent editorial about exosomes in cancer therapy and a hopeful commentary upon liquid biopsy of extracellular biomarkers for prostate cancer personalized treatment decision . Moreover, a recently published new EV data base (EV-ADD), the first one to be concerned with EV-associated DNA in human liquid biopsy samples , corroborates the current potential interest of this long-neglected EV component, not only for early cancer diagnosis but also possibly in the future for prognosis and disease monitoring after treatment, and even for EV-mediated therapy and resistance to therapy. Conflicts of Interest The author declares no conflict of interest. cancers-15-01456-t001_Table 1 Table 1 Precursor studies on the DNA content of cell-derived EVs (2011-2016). Cell Lines/Samples Main Results Reference Three medulloblastoma cell lines. MVs carry DNA which reflects the genetic status of the tumor with a significant amplification of the c-Myc oncogene; exoDNA is primarily single stranded. Murine cardiomyocite muscle cell line. MVs/exos containing DNA/RNA could transfer chromosomal DNA sequences to target fibroblasts. Human VSMCs culture and plasma. K562s and human neutrophils. EV-mediated transfer of gDNA to recipient cells: a novel mechanism for intercellular genetic influence. Transfer of BCL/ABL hybrid gene from K562s-EVs to normal human neutrophils. Two pancreatic cancer cell lines. Serum from PDAC patients. Exos contain >10kb fragments of ds-gDNA spanning all chromosomes. Specific KRAS and p53DNA mutations found in serum exosomes of PDAC patients. Three cancer model cell lines: human myeloid leukemia; human colorectal carcinoma; and murine melanoma. The majority of DNA associated with tumor exos is double stranded either externally (>8.5 kb), larger than internal ExoDNA, or extended to a broad panel of tumor cell lines; in murine melanoma, only a 10% sExo subset contained DNA; exo-dsDNA reflects the mutational status of parental cells. Two (IEC) rat cell lines, (nontumorigenic (RAT-1) and tumoral (RAS-3). The RAS-3 EVs contained dsDNA large fragments, covering the entire rat genome, including the transferable full-length H-RAS oncogene (3308 bp). Human mast cells. The EV-DNA released by human mast cells is usually associated with the outside of EVs. Human BM-hMSCs-/+ transduction with a plant DNA The cell-derived EVs also carry high molecular DNA not originating from dying cells, mainly associated to the outer EV membrane, and not organized in nucleosomes. Confirmation of the EV-mediated horizontal gene transfer. Review Summary of ds-gDNA in circulating exosomes. Serums Stability of EVs extracted from serums under different storage conditions. cancers-15-01456-t002_Table 2 Table 2 Further studies on circulating tumor EVs-DNAs (2014-2019). Samples/Aims Main Results Reference Three prostate cancer (PCa) cell lines. Plasma of human (PCa) patients (n = 4). Different gDNA fragments in the subpopulations of EVs (Abs, MVs, and EXOs). EV-gDNA could harbor specific gDNA mutations of the parent cells. Plasma EVs also carry double-stranded gDNA with no differences in MVs/EXOs. Glioblastoma, PC3 prostate cancer, or U87 cancer cell lines. Plasma of a PCa mouse model; human plasma of mCRPC patients (n = 40). Large EVs (oncosomes) contain most of the circulating chromatinized DNA (up to 2 Mb). L-EVs from human mCRPC patients contained large-sized dsDNA, covering the entire tumor genome, with reported cancer-specific (MYC/PTEN) genomic alterations. Whole blood samples of pancreatic cancer (PDAC) patients (n = 127) and controls. KRAS mutations were more detectable in exoDNA than in cell-free DNA, but mutant KRAS was also detected in a substantial minority of healthy samples. Serum from patients with (PDAC) pancreatic cancer or pancreatic disease and from healthy individuals. The minimal exosomal DNA used for digital PCR analyses was 0.663 ng. Potential clinical utility of circulating exosomal DNA for identification of KRASG12D and TP53R273H mutations in patients with pancreas-associated pathologies. Engineered exosomes from fibroblasts-like mesenchymal cells (iEXosomes). Compared to liposomes, iExosomes facilitate therapeutic targeting of oncogenic KRAS in pancreatic cancer. Bioreactor-based generation of clinical-grade iExosomes. Large-scale production of clinical-grade iExosomes for targeting KRAS in pancreatic cancer. Xenotransplant mouse model of human glioma-cancer stem cells featuring an intact blood-brain barrier (BBB). The three types of glioma-derived EVs (ABs, MVs, and EXOs) contained gDNA sequences. Some sequences appeared in all EVs, whereas a few sequences appeared exclusively in one type of EVs. All tumor-derived EVs cross the intact BBB and can be detected in the peripheral blood. Comparison of circulating cfDNA and EV-DNA, their origins, and their respective advantages and disadvantages for cancer diagnostic. Mutated cfDNA, more tumor-specific and enriched in smaller fragments, is more efficient for prognosis of late tumor stages. Exosomal gDNA (between 2.5-10 kb) might be a better potential biomarker for early cancer diagnosis. An immortalized HeLa cervical cancer (2D) cell culture and a three-dimensional (3D) organoid culture. The EV secretion dynamics were significantly different for both culture types: 2D culture remains a valuable tool for the search of human cir-tEV-gDNA cancer biomarker, whereas the 3D culture seems more useful for searching cir-tEV-RNA. Mechanisms of nuclear content loading to exosomes. A link between micronuclei (MN) formation and the generation of some specific exosomal loading with gDNA was identified by inducing genomic instability. Human mast (HMC-1) cell line and (TF-1) erythroleukemic cell line. Exosome-enriched small extracellular vesicles (sEVs) were discriminated by a high resolution iodixanol density gradient into two novel heterogeneous EV subpopulations of low density (LD) and high density (HD) with different RNA/DNA EV cargoes.DNA was predominantly localized on the outside or surface of sEVs. Human colon (DKO1) and glioblastoma (Gli36) cell lines; normal primary kidney epithelial cells and human plasma. Necessary reassessment of the "classical" exosome composition and biogenesis: extracellular dsDNA is not associated with exosomes or any other types of small EVs, but with extracellular particles (EPs). cancers-15-01456-t003_Table 3 Table 3 Preclinical studies about the circulating tumor-derived EV-gDNA as a potential cancer biomarker (2020-2021). Aims/Samples Main Results Reference Proteomic profile of potential cancer biomarkers in 426 human cancer and noncancer samples derived from various cells, tissues, and body fluids. Crude sEVs categorized into (EVPs) three subpopulations: small exosomes (Exo-S 50-70 nm), large exosomes (Exo-L 90-120 nm), and exomeres (non vesicular (NV) particles <50 nm). Analysis of 120 plasma-derived EVP proteomes from 77 cancer patients with 16 different cancer types and 43 healthy controls (HC) suggested that EVP proteins can be useful for cancer detection and determinization of cancer type. Extensive review on the EV biogenesis, focusing mainly on EXOs and MVs. Discussion about the current knowledge upon EV-uptake and cell-cell communication, as well as upon the cargo sorting into EVs. Possible EV bioengineering methodologies for therapy improvements. Comparison of EV-mediated liquid biopsy with older liquid biopsies for lung adenocarcinoma diagnosis. EVs are advocated for as ideal carriers of cancer biomarkers. Contrary to the passively released fragmented cfDNAs (about 200 bp), cEV DNAs consist of both large-sized ds-gDNAs (up to 10 kb) and fragmented, mutated DNAs. The membrane-protected EV-DNAs also have a high stability A higher sensitivity can be achieved by using EV-DNAs obtained from bronchoalveolar lavage fluid (BALF) than those from blood. Nine small-cell lung cancer (SCLC) patients and twenty-two (SCLC) patients with known tumor EGFR mutation. Platelet-poor plasma was fractionated by five sequential centrifugations and ExoQuick for preparing the exosomal fraction 5, which was then dominated by small (~160 bp) nucleosome-like DNAs.Improved detection of cell-free tumor DNAs (cf-tDNAs) is claimed in EV-depleted plasma (fraction 6), and higher mutation detection rates (14/22) are observed than in whole plasma (10/22). Blood samples from healthy human donors. This older study contradicts the previous one by showing the association of dsDNA inside the plasma exosomes and stating that "more than 93% of amplifiable cfDNA in plasma is located in plasma exosomes". Human osteosarcoma (OS) serum samples. Copurification of OS-associated repetitive element DNAs with EVs in size exclusion chromatography but not in exosome immunoaffinity capture. Repetitive element DNAs showed a high sensitivity and specificity for sera of patients with an OS diagnosis but were not tightly bound to CD9+ or CD81+ exosomes, supporting that exosomes either do not contain DNA or are tightly associated with particles with DNA. Comparison of cf-tDNA and EV-DNA in serial plasma samples of a metastatic breast cancer patient. Of the 52 copy number variants (CNVs) (from 0.3 to 106.5 Mb) in tDNA, 36 were detected in at least one cf-tDNA and 13 in one EV-DNA sample and were distributed randomly throughout the genome. cf-tDNA, shed from apoptotic tumor cells, had a greater sensitivity for serial monitoring of breast cancer than EV-DNA actively secreted from viable neoplastic cells. Summary of the biological and clinical aspects of EV-DNA and role of EV-DNA in cancer. EV-DNA as a biomarker for liquid biopsy is a new but definitely promising area of study, but its study in the clinical context is still quite open for further validation. Bronchoalveolar lavage fluid (BALF) of 20 (NSCLC) patients with EGFR-mutations and matched fixed-tissue samples. Heterogeneous (100-300 nm) EVs from BALF contained mostly ~11kb DNAs from both vesicle surface and inside. The DNA yield from BALF-EVs was 100 times less than tissue DNA but had enough tumor-specific DNA for the identification of actionable genetic alterations with a high potential clinical utility. 54 plasma samples and 13 pleural fluids of (NSCLC) patients after tyrosine kinase inhibitor therapy. By comparison of different technological tools to detect EGFR mutations, combined tumor nucleic acid analysis (exoTNA+cfTNA) in the plasma and exoTNa in the pleural fluid allowed for the detection of target EGFR mutations more sensitively than using cfDNA or total DNA alone. Focus on the DNA inclusion in EVs, the techniques of EV-DNA detection and quantification, and the clinical use of EV-DNA. Recapitulation of the cell-free DNA cell sources by active or passive mechanisms and summary of the tumor genome hallmarks reflected by EV-DNA as well as the results of the main clinical studies assessing the performance of EV-DNA biomarkers.Enumeration of the many challenging questions remaining to be solved before reaching the clinics. Cell lines and glioblastoma stem-like (GS) cell cultures.Human glioma patients' tissue and nontumoral tissue. The vast majority of EVs carry DNA, which localizes more to the EV surface than inside EVs. Proof of principle that glioblastoma-derived EV-DNA reflects the genome-wide methylation, CNVs, and mutational status of glioblastoma cells with high accuracy and enables their molecular classification. Plasma and matched primary tumor tissues of 21 patients with aggressive diffuse large B cell lymphoma (DLBCL). First study to show the presence of DNA in plasma exosomes of DLBCL patients.CDKN2A and CDKN2B were methylated in both plasma exosomes and primary tumor tissue samples.Compared to 21 healthy individuals, exosome concentration was approximately 6 times higher in DLBCL patients, but the exosomal dsDNA content was extremely low compared to RNA contents. First direct comparison on gene expression and DNA methylation markers in CTCs and paired plasma-derived exosomes. Remarkable heterogeneity on gene expression and DNA methylation markers between EpCAM-positive CTCs and paired plasma-derived exosomes in metastatic castration-resistant prostate cancer (mCRPC) patients with significantly higher positivity in CTCs. Extensive review of the characteristics and clinical applications of extracellular vesicle-derived DNA. The presence of DNA in excreted exosomes will continue to be controversial until the development of a method for isolating pure exosomes or microvesicles. Nonetheless, the size of dsDNA found in EVs (from ~100 bp to ~20 kbp) can represent the entire genome and reflects the mutational status of tumor parental cells. With DNA extracted from all categories of EVs, EV-DNA is the latest and most promising biomarker for identifying tumor presence and complexity. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000629 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050776 cells-12-00776 Article CFTR Inhibitors Display In Vitro Antiviral Activity against SARS-CoV-2 Lagni Anna Conceptualization Methodology Software Formal analysis Data curation Writing - original draft Writing - review & editing 1+ Lotti Virginia Conceptualization Methodology Software Formal analysis Data curation Writing - original draft Writing - review & editing 1*+ Diani Erica Methodology Data curation Writing - review & editing 1 Rossini Giada Methodology Writing - review & editing 2 Concia Ercole Conceptualization Writing - review & editing 3 Sorio Claudio Conceptualization Resources Data curation Writing - original draft Writing - review & editing Supervision Funding acquisition 4++ Gibellini Davide Conceptualization Formal analysis Resources Data curation Writing - original draft Writing - review & editing Supervision Funding acquisition 1++ Castagnola Elio Academic Editor 1 Microbiology Section, Department of Diagnostic and Public Health, University of Verona, 37134 Verona, Italy 2 Microbiology Unit, IRCCS Azienda Ospedaliero-Universitaria di Bologna, 40138 Bologna, Italy 3 Department of Diagnostic and Public Health, University of Verona, 37134 Verona, Italy 4 General Pathology Section, Department of Medicine, University of Verona, 37134 Verona, Italy * Correspondence: [email protected] + These authors contributed equally to this work. ++ These authors contributed equally to this work. 28 2 2023 3 2023 12 5 77628 12 2022 24 2 2023 27 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Several reports have indicated that SARS-CoV-2 infection displays unexpected mild clinical manifestations in people with cystic fibrosis (pwCF), suggesting that CFTR expression and function may be involved in the SARS-CoV-2 life cycle. To evaluate the possible association of CFTR activity with SARS-CoV-2 replication, we tested the antiviral activity of two well-known CFTR inhibitors (IOWH-032 and PPQ-102) in wild type (WT)-CFTR bronchial cells. SARS-CoV-2 replication was inhibited by IOWH-032 treatment, with an IC50 of 4.52 mM, and by PPQ-102, with an IC50 of 15.92 mM. We confirmed this antiviral effect on primary cells (MucilAirTM wt-CFTR) using 10 mM IOWH-032. According to our results, CFTR inhibition can effectively tackle SARS-CoV-2 infection, suggesting that CFTR expression and function might play an important role in SARS-CoV-2 replication, revealing new perspectives on the mechanisms governing SARS-CoV-2 infection in both normal and CF individuals, as well as leading to potential novel treatments. SARS-CoV-2 cystic fibrosis CFTR CFTR inhibitors antiviral IOWH-032 PPQ-102 Fondazione CariveronaLega Italiana Fibrosi Cistica onlusThe present study was supported by Fondazione Cariverona, ENACT project VIRO-COVID and by Lega Italiana Fibrosi Cistica onlus. pmc1. Introduction SARS-CoV-2 is an enveloped, positive-sense, single-stranded RNA virus belonging to the genus Betacoronavirus, and it is the etiological agent of coronavirus disease 2019 (COVID-19). This virus first appeared in Wuhan, China, in December 2019 and, despite containment attempts, expanded rapidly worldwide, creating an urgent need to identify effective antiviral drugs . Intense fatigue, headaches, dyspnea, myalgia, and gastrointestinal symptoms such as vomiting, stomach pain, loss of appetite, and diarrhea are all common symptoms of COVID-19. Moreover, certain other symptoms, including hyposmia, anosmia, ageusia, maculopapular rash, and urticarial lesions, have also been reported. Cases with severe clinical evolution are related to cytokine storm syndrome, a derangement of the regulation of certain cytokines such as IFN, MCP1, IP-10, TNF-a, and IL-10. This dysregulation causes local tissue damage and systemic non-protective inflammation, which may lead to sepsis and lung injury, often involving acute respiratory distress syndrome (ARDS), pneumonitis, respiratory failure, sepsis shock, organ failure, and possibly death . Several molecules were considered in in vitro studies to determine their efficacy and safety as potential agents for COVID-19 treatment. Corticosteroids, antivirals, interferons, and monoclonal antibodies directed against immune system proteins, such as interleukin-6 (IL-6), or against certain SARS-CoV-2-specific targets were investigated in order to tackle either the viral replication cycle or the inflammation related to infection . Currently, a few treatments are recommended against SARS-CoV-2; specifically, in the case of non-severe disease but a high risk of hospital transmission, the WHO recommends the use of Nirmatrelvir (Paxlovid), an active protease inhibitor associated with Ritonavir. In addition, the ribonucleoside analog Molnupinavir and the viral polymerase inhibitor Remdesivir are effective in the inhibition of SARS-CoV-2 replication, with a favorable safety and tolerability profile , and both molecules are conditionally recommended by the WHO in cases of non-severe and severe COVID-19 . Although the onset of new variants has partially counteracted the antiviral activity of several monoclonal antibodies towards S viral protein, they are still consistently used for SARS-CoV-2 infection treatment. Recent reports have shown that SARS-CoV-2 infection does not result in worse outcomes in people with cystic fibrosis (pwCF); in a multinational cohort study of 40 CF patients, the incidence of COVID-19 in pwCF was 0.07%, compared to 0.15% in the general population . Moreover, an international cohort analysis of 181 CF patients from 19 different countries demonstrated how the majority of pwCF may experience milder outcomes from SARS-CoV-2 infection, despite the much larger number of hospitalized patients who underwent organ transplantation than patients who did not . Confirming these retrospective observations, a recent study found that SARS-CoV-2 replication is reduced in cystic fibrosis transmembrane conductance regulator (CFTR)-mutated bronchial cells with respect to functional CFTR bronchial epithelial cells . These observations indicate an important role of CFTR protein in the regulation of SARS-CoV-2 replication; thus, CFTR was suggested as a potential novel molecular target for innovative antiviral treatment. In this study, we evaluated the effect of short-term (up to 48 hpi) incubation with the specific pharmacological inhibitors of CFTR (IOWH-032 and PPQ-102) in bronchial epithelial cells expressing native CFTR. 2. Materials and Methods 2.1. Cells, Virus, and Treatments The cell line used herein is a subclone obtained from transfection of the parental line with an HIV-based lentiviral vector containing the wild-type CFTR gene. The parental a CF human bronchial epithelial cell line, derived from a CF patient homozygous for the DF508 CFTR mutation, which was immortalized using the SV40 plasmid with a defective origin of replication (pSVori-). To culture these cells, minimum essential medium (MEM, Gibco, Thermo Fischer, Monza, Italy) supplemented with 10% FBS (Euroclone, Milan, Italy), 1% Glutamine (GlutaMAX, Gibco, Thermo Fischer), and the appropriate amount of selective factor (Puromycin) was used. MucilAirTM (Epithelix Sarl, Geneva, Switzerland) is an in vitro cell model of the human airway epithelium cultured at the air-liquid interface and reconstituted using human primary cells at low passage (P1). It was kept in MucilAirTM culture medium (Epithelix Sarl) at the air-liquid interface (ALI) for 1 week after receipt. We used this fully differentiated bronchial epithelial 3D model from primary human cells derived from healthy donors' (wt/wt-CFTR MucilAirTM) (n = 3; age mean 46 +- 16) and cystic fibrosis patients' homozygotes for DF508 (F508del/F508del-CFTR MucilAirTM) (n = 3; age mean 28 +- 10). Cells were infected with two different SARS-CoV-2 strains: SARS-CoV-2 B.1 strain (hCoV-19/Italy/BO-VB12/2020|EPI_ISL_16978127), replicated in Vero E6 cells as described by Ogando and colleagues , and SARS-CoV-2 BA.5.1 (Omicron) strain, expanded in Calu-3 cells. Cells were treated with IOWH-032 (MedChemExpress LLC., Monmouth Junction, NJ, USA), a synthetic small hydrazide molecule designed to selectively inhibit the CFTR channel with a reported IC50 of 8 mM, acting at the external surface of CFTR . This molecule, with antisecretory activity, was first studied to treat diarrhea; it entered Phase II clinical trials in 2013, but did not progress further in clinical development, despite being proven to be safe and well-tolerated in healthy volunteers . To confirm that the antiviral effects were due to CFTR inhibition, PPQ-102 (Selleck Chemicals LLC, Houston, TX, USA), a well-known CFTR inhibitor with an IC50 of 90 nM, was tested . This drug targets the intracellular nucleotide binding domain of CFTR and inhibits the CFTR-mediated chloride current in a reversible and voltage-independent manner. It showed efficacy in a mouse model of ADPKD . Both IOWH-032 and PPQ-102 were initially prepared as 10 mM stocks diluted in DMSO. 2.2. Cytotoxicity Assay In vitro cell viability was assessed after IOWH-032 and PPQ-102 treatment on CFBE WT cells. The assay was carried out according to the CellTiter 96(r) AQueous One Solution Cell Proliferation Assay (Promega, Madison, WI, USA) at 1 h, 24 h, and 48 h post-treatment. Briefly, 104 cells were seeded in 96-well plates and incubated at 37 degC, humidified, and 5% CO2. After 24 h of incubation, the different dilutions of IOWH-032 (0.1 mM, 1 mM, 5 mM, 7 mM, 10 mM, 20 mM, 30 mM, 50 mM, and 100 mM) and PPQ-102 (0.1 mM, 1 mM, 5 mM, 10 mM, 20 mM, 30 mM, 50 mM, and 100 mM) were added for a final volume of 100 mL. In parallel, we also considered cells treated with pure DMSO at the same concentrations as the two compounds. Cells were then incubated at 37 degC, humidified, and 5% CO2, and after 1 h, 24 h, and 48 h post-treatment incubation, 20 mL of CellTiter 96(r) AQueous One Solution Reagent was added to each well. After 2 h of incubation, absorbance was recorded at 490 nm using a 96-well plate reader. DMSO was used as a blank. The 50% cytotoxic concentration (CC50) was determined. 2.3. TEER Measurement In order to assess the integrity of the tissue monolayer, transepithelial electrical resistance (TEER) variations were measured, both in the wt/wt-CFTR MucilAirTM cells and wt/wt-CFTR MucilAirTM cells treated with 10 mM IOWH-032, using a volt-ohm meter (EVOM2, Epithelial Volt/Ohm Meter for TEER) and STX 2 electrodes (World Precision Instruments, Sarasota, FL, USA), according the manufacturer's instructions. Briefly, on the apical side of the insert, 200 mL of pre-warmed MucilAirTM culture media was applied. The electrode was cleaned in 70% ethanol and equilibrated in saline solution (0.9% NaCl; 1.25 mM CaCl2; 10 mM HEPES) until the volt-ohm meter registered 0.00. The electrode was inserted, with the long stem entering through the gap of the insert and leaning on the bottom of the well, and the short stem was immersed in the culture media, above the apical surface, for measurement. Following the measurement, the medium was immediately removed from the apical side. The values were calculated and represented as O*cm2, based on the surface area of the inserts (0.33 cm2). 2.4. Antiviral Activity Evaluation T25 cell flasks were prepared with CFBE WT cells and, an hour before infection, increasing concentrations of IOWH-032 (0.1 mM, 1 mM, 5 mM, 7 mM, 10 mM, 20 mM, and 30 mM) and PPQ-102 (0.1 mM, 1 mM, 5 mM, 10 mM, 20 mM, and 30 mM) were added. An untreated flask, an uninfected flask, and a DMSO-treated flask were also prepared. After 1 h of incubation at 37 degC and 5% CO2, the SARS-CoV-2 B.1 strain was inoculated into the cells at a multiplicity of infection (MOI) of 1, and the cells were incubated for 1 h at 37 degC and 5% CO2. After incubation, the media was changed with fresh medium with the restored IOWH-032 concentrations. The supernatant and cell samples were collected at 48 h post-infection (hpi). For MucilAirTM infection experiments, the apical sides were gently washed twice with pre-warmed OptiMEM medium (GIBCO, Thermo Fisher Scientific, Monza, Italy); then, inserts were treated, adding 10 mM IOWH-032 on wt/wt-CFTR MucilAirTM, and incubated for 1 h at 37 degC and 5% CO2 before infection. Next, SARS-CoV-2, diluted in OptiMEM at MOI 1, was inoculated in both wt/wt-CFTR MucilAirTM and F508del/F508del-CFTR MucilAirTM and incubated for 1 h at 37 degC and 5% CO2. After incubation, the medium on the basolateral side was changed to a fresh medium containing IOWH-032 10 mM, and the inoculum from the apical side was removed to restore the ALI state. At 48 hpi, cells were harvested in lysis buffer (Promega, Madison, WI, USA) while supernatants from apical washes or basolateral media were collected. The supernatant's SARS-CoV-2 load was detected with the multiplex real-time PCR Allplex 2019-nCoV assay kit targeting the E, RdRp/S and N genes (Seegene, Seoul, Republic of Korea) following the manufacturer's instructions, while RNA was extracted from cells with ReliaPrepTM RNA Miniprep Systems (Promega), retrotranscribed into cDNA with iScriptTM Reverse Transcription Supermix for RT-qPCR (Bio-rad, Hercules, CA, USA), and then analyzed by real-time qPCR on a CFX96 Real-Time System (Bio-Rad) using the primers indicated in Table 1. The percentage of inhibition of SARS-CoV-2 replication was estimated for each drug concentration, and the half-maximal inhibitory concentration (IC50) was determined. To verify that CFTR inhibitors' antiviral activity is not SARS-CoV-2 strain-specific, the supernatant analyses described above were performed on treated WT infected with the BA.5.1 strain of SARS-CoV-2. 2.5. Inhibition Stage Determination In addition to the "full-time" treatment (described above), two different treatment protocols were employed in order to shed light on the replication cycle phase targeted by CFTR inhibitors to inhibit viral replication. The first treatment was based on treatment performed with either 10 mM IOWH-032 or 20 mM PPQ-102 on WT 1 h before viral infection and maintained during the viral incubation. The medium was changed after the 1 h of viral incubation with a fresh medium without a CFTR inhibitor. The second treatment was carried out with either 10 mM IOWH-032 or 20 mM PPQ-102 added after the 1 h of viral incubation, and then maintained until the collection of the supernatant at 48 hpi. The supernatant's viral load was determined as previously described. 2.6. Electrophysiological Measurements cells were seeded on Costar Transwell(r) inserts (Corning, NY, USA) and cultured in the air-liquid interphase (ALI) for two weeks. Once ready, the inserts were mounted on a slider in an Ussing chamber, and the transepithelial short-circuit current (Isc) was monitored using an EVC4000 multi-channel voltage/current clamp (WPI, World Precision Instruments). The two half chambers were filled with Quinton saline buffer solution. Subsequently, the ENaC blocker Amiloride (10 mM) was added to the apical side, and the cAMP analog CPT (100 mM), as well as the CFTR inhibitors IOWH-032 (10 mM) and PPQ-102 (10 mM), were added to both the apical and the basolateral sides. The tracings were recorded with PowerLab (8/35, AD Instruments, Sydney, Australia), and data analyses were performed using Labchart v8 software (AD Instruments). 3. Results 3.1. Cytotoxicity Evaluation In order to exclude the possibility that the antiviral effects recorded in this experimental model may be related to cell toxicity, we evaluated the cell cytotoxicity of different concentrations of IOWH-032 and PPQ-102 in the cell line . We assessed several IOWH-032 concentrations (0.1, 1, 5, 7, 10, 20, 30, 50, and 100 mM) at different time points (1, 24, and 48 h post-treatment) and evaluated cell cytotoxicity with the CellTiter 96(r) AQueous One Solution Cell Proliferation Assay (Promega, Madison, WI). The analysis of cytotoxicity reported a CC50 > 50 mM at 48 h post-treatment . No cytotoxicity was detected using the vehicle (DMSO) at the same concentration used for the preparation of the tested drugs. No significant cytotoxicity of IOWH-032 was reported at the previously reported CFTR inhibitory concentration IC50 . Similar results were obtained for PPQ-102, with no significant cytotoxicity detected below 20 mM . We then selected the 10 mM concentration and tested the effect of the treatment with IOWH-032 on wt/wt-CFTR MucilAirTM, a fully differentiated 3D model bronchial epithelium. Given the limited number of 3D models available, we assessed tissue integrity by measuring transepithelial electrical resistance (TEER) at 0, 24, and 48 h after treatment to confirm the lack of cytotoxicity which was also seen in this experimental model. TEER response after the IOWH-032 treatment of wt/wt-CFTR MucilAirTM , expressed as a percentage of the control, demonstrates that there is no significant reduction in tissue integrity after treatment. 3.2. Anti-SARS-CoV-2 Activity of CFTR Inhibitors To assess the antiviral activity of CFTR inhibitors IOWH-032 and PPQ-102, we firstly analyzed SARS-CoV-2 B.1 RNA content using quantitative real-time RT-PCR, targeting the E viral gene at 48 hpi. We recorded a concentration-dependent viral inhibition in the cell culture supernatant by IOWH-032 , reaching nearly 100% viral inhibition at 30 mM, with an IC50 of 4.52 mM (95% CI 3.24-5.71) . To determine which step of the viral life cycle is inhibited by the treatment, we also tested the intracellular SARS-CoV-2 content . IOWH-032 treatment caused a viral inhibition of nearly 100% at 30 mM with an IC50 of 6.66 mM (95% CI 5.72-8.67), slightly higher than the concentration causing 50% inhibition in the supernatant fraction . PPQ-102 antiviral activity was also detected in the supernatant following a concentration-dependent increase in the percentage of inhibition, approaching 100% at 30 mM , and an IC50 of 15.92 mM (95% CI 11.99-45.87) . At the intracellular level, the inhibition trend is still concentration-dependent, with an IC50 of 12.1 mM (95% CI 4.48-14.01) , even though viral inhibition does not reach a value close to 100% at 30 mM . As no significant cell cytotoxicity was recorded at drug concentrations around 10 mM , we can conclude that the antiviral effect is not related to the cytotoxic effect that occurs at higher concentrations. We then tested the antiviral activity of IOWH-032 at a concentration of 10 mM on wt/wt-CFTR MucilAirTM, analyzing both the supernatant and intracellular SARS-CoV-2 RNA content. F508del-CFTR cells were used as a reference control of CFTR inhibition. Even for this cellular model, we detected strong viral inhibition in the supernatant , while at the intracellular level, the viral inhibition, although greater than 50%, was less consistent . It is important to note that the viral inhibition associated with treatment in the supernatant is comparable to the effect of a dysfunctional CFTR . On the other hand, there is a significant difference in viral inhibition at the intracellular level between these two cell types . It is noteworthy that the reduced viral titer induced by IOWH-032 in both the WT wt/wt-CFTR MucilAirTM cells was significantly more consistent in the supernatant than intracellularly, suggesting that CFTR inhibition can negatively affect the final steps of the viral cycle and the viral release from infected cells . 3.3. CFTR Inhibitors Affect Post-Entry Stages of SARS-CoV-2 Infection To assess the antiviral activity of CFTR inhibitors and viral replication stages, we treated the cell cultures with two different protocols . In the first protocol, we infected the cell culture in the presence of either IOWH-032 or PPQ-102 for 1 h only to investigate the consequence of CFTR inhibition in the first phase of the viral replication cycle. It is noteworthy that viral inhibition was very low. In the second protocol, IOWH-032 or PPQ-102 were added to the cells one hour after exposure to the virus. In this case, both CFTR inhibitors significantly impacted the viral replication. 3.4. CFTR Inhibitors' Antiviral Activity Is Not SARS-CoV-2 Strain-Dependent To verify that the antiviral activity of CFTR inhibitors is not SARS-CoV-2 strain-specific, we repeated the supernatant analysis on treated WT infected with the SARS-CoV-2 BA.5.1 strain . It is noteworthy that CFTR inhibitors' antiviral activity is still important, even against the SARS-CoV-2 Omicron variant. The additional infection protocols confirm the data obtained with SARS-CoV-2 B.1 . 3.5. Electrophysiological Measurements To confirm the capability of the CFTR inhibitors to efficiently inhibit CFTR channel function, in Ussing chambers, we assessed the transepithelial ion transport in the same setting, i.e., a polarized monolayer derived from cells . Firstly, the epithelial sodium channel (ENaC) was transiently blocked by amiloride in order to measure only anion transport. The following addition of CFTR inhibitors IOWH-032 and PPQ-102 at a concentration of 10 mM induced a block in anion secretion through the monolayers, causing a current lowering (32.9 +- 14 and 48.8 +- 14 mA, respectively). To assess the residual CFTR functionality after the inhibition, a CFTR activator, CPT (100 mM), was added to a control monolayer (without adding the inhibitor) and to both the IOWH- PPQ-102-treated monolayers. No anion secretory response was elicited by the PPQ-102-treated monolayers, confirming a complete blockade of the CFTR channel. While the cAMP agonist induces an increase in CFTR function in untreated cells, only a limited signal was detectable in IOWH-032-treated monolayers (16.6 +- 4 and 7.8 +- 5 mA, respectively, n = 3). These comparative results suggest a lower CFTR-inhibitory capacity of IOWH-032 compared to PPQ-102, supporting the limited efficacy of this inhibitor when used in vivo for the treatment of cholera . 4. Discussion Several studies reported a significantly lower severity, number of infection cases, and viral spread of SARS-CoV-2 in pwCF than in the normal population, even though pwCF are expected to be at an increased risk of developing severe manifestations of COVID-19, since they are considered "fragile" patients . Several factors can be associated with lower SARS-CoV-2 infection levels, such as preventive measures with which pwCF are familiar and the use of CF therapies. However, a recent study showed that SARS-CoV-2 infection is less productive in cells with CFTR defects than in healthy ones, suggesting that CFTR dysfunction may play a direct or indirect role in the viral replication cycle or its maturation . It is known that chloride channels are involved in several steps of the viral life cycle, such as viral entry, membrane fusion processes, endosomal trafficking, and viral replication . The CFTR channel has been reported to play a role in several respiratory viral infections; some studies have shown that the influenza virus M2 matrix protein decreased CFTR activity and expression in bronchial epithelial cells. Panou et al. reported a reduced Polyomavirus BK (BKPyV) infection associated with the inhibition of CFTR function, with CFTR being required during BKPyV transport to the endoplasmic reticulum. Interestingly, TMEM16A and TMEM16F, calcium-activated , are involved in viral replication; the inhibition of TMEM16A leads to reduced respiratory syncytial virus (RSV) gene expression and can be activated during SARS-CoV-2 infection. Their inhibition with niclosamide causes the inhibition of the SARS-CoV-2 spike protein-driven fusion of the syncytia . Epidemiological and early mechanistic studies revealed a role for CFTR in SARS-CoV-2 replication/function. As CFTR-defective cells have a complex array of dysfunctions associated with altered CFTR processing and a lack of activity, we focused on the short-term chemical inhibition of CFTR function in native cells to evaluate whether an altered CFTR function, more than alteration of additional pathways present in chronically CFTR defective cells, was involved. We chose two highly selective CFTR inhibitors which have been widely used in the literature, namely IOWH-032 and PPQ-102, to evaluate their antiviral activity. We found that the antiviral activity of IOWH-032 and PPQ-102 occurred at IC50 values of 4.52 mM and 15.92 mM, respectively, which are lower than those causing significant cytotoxicity. Importantly, CFTR inhibitor antiviral activity was also confirmed against two viral strains, including the Omicron SARS-CoV-2 variant. The disruption of the SARS-CoV-2 replication cycle and alteration in protein structure and function may be caused by a shift in ionic balance and intracellular pH variation. According to a recent study, the endolysosome proteases TMPRSS-2 and cathepsins B and L activate the SARS-CoV-2 S protein in an acidic environment, which is why the deacidification of this organelle has been found to inactivate proteases and prevent viral infection . The CFTR protein is necessary to maintain an intracellular ionic and pH balance and, based on the results reported, its dysfunction might represent an important factor driving reduced SARS-CoV-2 replication . In particular, SARS-CoV-2 controls autophagosomal machinery to facilitate the creation of double-membrane vesicles, effectively subverting autophagy to increase replication . Interestingly, CFTR dysfunction leads to BECN1 inactivation, perturbing endosomal fusion/maturation and trafficking, with a negative impact on intracellular trafficking. This leads to defective autophagosome formation , preventing SARS-CoV-2 from being able to take advantage of this pathway. In a study by Merigo et al., it was found that SARS-CoV-2 infection in malfunctioning CFTR cells caused the onset of autophagosomal structures containing cellular recycling material rather than replicative structures, as it did in wild-type cells . As we have just observed, many pathways could be involved in reduced SARS-CoV-2 replication. In an attempt to determine which stage of the SARS-CoV-2 life cycle is involved in this reduced replication, we assessed the impact of treatments at different time-points. According to our results, CFTR inhibitors did not significantly impact SARS-CoV-2 cell entry; however, they were significantly more effective in the post-entry phase. In addition, comparing the intracellular and supernatant viral loads after CFTR inhibition, we found a stronger inhibitory effect in the supernatant, thus suggesting a major SARS-CoV-2 suppression effect during the intracellular phases of the SARS-CoV-2 replication cycle. SARS-CoV-2 virions can be released from an infected cell via the Golgi compartment , the function and structure of which seem to be altered to facilitate viral trafficking , or by incorporation into deacidified lysosomes ; both are pathways that can be adversely impacted by pH deregulation, as in the case with CFTR alteration . An ultrastructural study revealed that small secretory vesicles carrying a single virus particle are the main mechanism of SARS-CoV-2 egress . Recent findings have demonstrated that SARS-CoV-2 exploits extracellular vesicles (EVs) for cellular exit and intercellular communications ; however, in this scenario, environmental pH stress and altered lipogenesis, possibly caused by CFTR disfunction, can cause a decrease in the number of EVs . Interestingly, treatment with IOWH-032 reaches an almost 100% inhibition of SARS-CoV-2 mRNA levels, but at a concentration that displays significant cytotoxicity. Hence, whether an incomplete limitation of SARS-CoV-2 replication is enough to determine a significant clinical impact remains to be determined. In conclusion, in this study, we demonstrated the significant anti-SARS-CoV-2 activity of CFTR inhibitors occurring during intracellular viral replication and below cytotoxic concentrations. Given that both CFTR inhibitors demonstrated strong and similar antiviral activity in two models of native bronchial epithelial cells and against two different SARS-CoV-2 strains, we can safely conclude that the loss of CFTR function may be considered an important factor in SARS-CoV-2 replication, suggesting a role for CFTR in SARS-CoV-2 infection. Although further studies are necessary to understand whether the molecular pathways governing CFTR-dependent SARS-CoV-2 replication may be considered as a clinically relevant target for antiviral treatment, this study supports the repurposing of CFTR inhibitors from diarrheal disease to antiviral application. Author Contributions Conceptualization, A.L., V.L., E.C., C.S. and D.G.; methodology, A.L., V.L., E.D. and G.R; software, A.L. and V.L.; formal analysis, A.L., V.L. and D.G.; resources, C.S. and D.G.; data curation, A.L., V.L., E.D., C.S. and D.G.; writing--original draft preparation, A.L., V.L., C.S. and D.G.; writing--review and editing, A.L., V.L., E.D., G.R., E.C., C.S. and D.G.; supervision, C.S. and D.G.; funding acquisition, C.S. and D.G. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Toxicity evaluation of IOWH-032 and PPQ-102 (0.1, 1, 5, 10, 20, 30, 50, and 100 mM) at 1, 24, and 48 h post-treatment, evaluated on WT . (a) Graphical representation of IOWH-032 concentration-dependent toxicity, represented as a percentage of viable cells. The red dashed line represents the CC50; (b) graphical representation of PPQ-102 concentration-dependent toxicity as a percentage of viable cells. The red dashed line represents the CC50. (n = 3, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001). Data are presented as the mean +- SD from three independent experiments, each of which was performed in triplicate. Figure 2 TEER response, presented as a percentage of the control, of wt/wt-CFTR MucilAirTM after treatment with 10 mM IOWH-032, measured after 0, 24, and 48 h post-treatment. Data are shown as the mean and SD of three independent experiments (n = 3). Figure 3 IOWH032 (0.1, 1, 5, 7, 10, 20, and 30 mM) anti-SARS-CoV-2 activity was studied in CFBE41oTM WT cells. The data are presented as the mean +- SD from three independent experiments. (a) Cellular supernatant's SARS-CoV-2 inhibition at different IOWH-032 concentrations, presented as a percentage of the control CFBE41oTM WT untreated cells; (b) graphical representation of supernatant IOWH-032 IC50; (c) intracellular SARS-CoV-2 inhibition at different IOWH-032 concentrations, presented as a percentage of the control CFBE41oTM WT untreated cells; (d) graphical representation of the intracellular IOWH-032 IC50. Figure 4 PPQ-102 (0.1, 1, 5, 10, 20, and 30 mM) anti-SARS-CoV-2 activity in . Data are presented as the mean +- SD from three independent experiments. (a) Supernatant's SARS-CoV-2 inhibition at different PPQ-102 concentrations, presented as a percentage of the control untreated cells; (b) graphical representation of the supernatant IC50 determination of PPQ-102; (c) intracellular SARS-CoV-2 inhibition at different PPQ-102 concentrations, presented as a percentage of the control untreated cells; (d) graphical representation of the intracellular IC50 determination of PPQ-102. Figure 5 Combined data of WT ' viability and viral inhibition after CFTR inhibitor treatment 48 hpi. (a) Comparison between the percentage of viable cells (red line), percentage of supernatant viral inhibition (blue line, empty circles), and percentage of intracellular viral inhibition (black line, triangle) after IOWH-032 treatment; (b) comparison between the percentage of viable cells (red line), percentage of supernatant viral inhibition (blue line), and percentage of intracellular viral inhibition (black line) after PPQ-102 treatment. Figure 6 Comparison of SARS-CoV-2 replication between wt/wt-CFTR, wt/wt-CFTR with 10 mM IOWH-032, and F508del/F508del-CFTR MucilAirTM. (a) Supernatant's viral load of wt/wt-CFTR, wt/wt-CFTR treated with 10 mM IOWH-032, and F508del/F508del-CFTR MucilAirTM, expressed as a percentage (* p < 0.05); (b) supernatant's viral inhibition, presented as a percentage of the control untreated wt/wt-CFTR MucilAirTM; (c) intracellular E gene expression, normalized to ACTB (b-Actin) of SARS-CoV-2 infected cells (* p < 0.05); (d) intracellular viral inhibition, presented as a percentage of the control untreated wt/wt-CFTR MucilAirTM (** p < 0.01). Data are presented as the mean +- SD from three independent experiments. Figure 7 Comparison between intracellular and supernatant viral inhibition in both and wt/wt-CFTR MucilAirTM treated with IOWH-032 10 mM (n = 3, * p < 0.05, ** p < 0.01). Figure 8 Antiviral activities of CFTR inhibitors after "full", "entry", and "post-entry" treatment in WT . Data are presented as the mean +- SD from three independent experiments. (a) IOWH-032 10 mM viral inhibition at different stages of SARS-CoV-2 infection; (b) PPQ-102 20 mM viral inhibition at different stages of SARS-CoV-2 infection. (n = 3, ** p < 0.01, *** p < 0.001). Figure 9 Antiviral activity of CFTR inhibitors against BA.5.1 variant on WT CFBE41o-. The data are presented as the mean +- SD from three independent experiments. (a) IOWH032 (0.1, 1, 5, 7, 10, 20, and 30 mM) anti-SARS-CoV-2 activity; (b) graphical representation of supernatant IOWH-032 IC50; (c) PPQ-102 (0.1, 1, 5, 10, 20, and 30 mM) anti-SARS-CoV-2 activity; (d) graphical representation of PPQ-102 IC50. Figure 10 Antiviral activities of CFTR inhibitors against SARS-CoV-2 BA.5.1 after "full", "entry", and "post-entry" treatment in WT . Data are presented as the mean +- SD from three independent experiments. (a) IOWH-032 10 mM viral inhibition at different stages of SARS-CoV-2 infection; (b) PPQ-102 20 mM viral inhibition at different stages of SARS-CoV-2 infection. (n = 3, ** p < 0.01, **** p < 0.0001). Figure 11 Short-circuit current (Isc) measurement tracing was used to prove the efficacy of IOWH-032 and PPQ-102 in inhibiting the CFTR channel in WT CFBE41o-. ENaC blocker amiloride (10 mM) was added on the apical side; the cAMP analog CPT (100 mM) and the CFTR inhibitors IOWH-032 (10 mM) and PPQ-102 (10 mM) were added to both the apical and the basolateral sides. A representative experiment of the three performed experiments is shown. cells-12-00776-t001_Table 1 Table 1 Primers used for intracellular RNA RT-qPCR. Primers Sequences (5'-3') E gene F ACAGGTACGTTAATAGTTAATAGCGT E gene R ATATTGCAGCAGTACGCACACA GAPDH F TCAAGAAGGTGGTGAAGCAGG GAPDH R CAGCGTCAAAGGTGGAGGAGT b-Actin F CCCTGGACTTCGAGCAAGAG b-Actin R ACTCCATGCCCAGGAAGGAA Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Zhu N. Zhang D. Wang W. Li X. Yang B. Song J. Zhao X. Huang B. Shi W. Lu R. A Novel Coronavirus from Patients with Pneumonia in China, 2019 N. Engl. J. Med. 2020 382 727 733 10.1056/NEJMoa2001017 31978945 2. Huang C. Wang Y. Li X. Ren L. Zhao J. Hu Y. Zhang L. Fan G. Xu J. Gu X. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China Lancet 2020 395 497 506 10.1016/S0140-6736(20)30183-5 31986264 3. Guan W. Ni Z. Hu Y. Liang W. Ou C. He J. Liu L. Shan H. Lei C. Hui D.S.C. Clinical Characteristics of Coronavirus Disease 2019 in China N. Engl. J. Med. 2020 382 1708 1720 10.1056/NEJMoa2002032 32109013 4. Gendrot M. Andreani J. Boxberger M. Jardot P. Fonta I. Le Bideau M. Duflot I. Mosnier J. Rolland C. Bogreau H. Antimalarial drugs inhibit the replication of SARS-CoV-2: An in vitro evaluation Travel Med. Infect. Dis. 2020 37 101873 10.1016/j.tmaid.2020.101873 32916297 5. Wang M. Cao R. Zhang L. Yang X. Liu J. Xu M. Shi Z. Hu Z. Zhong W. Xiao G. Remdesivir and chloroquine effectively inhibit the recently emerged novel coronavirus (2019-nCoV) in vitro Cell Res. 2020 30 269 271 10.1038/s41422-020-0282-0 32020029 6. Holwerda M. V'kovski P. Wider M. Thiel V. Dijkman R. Identification of an Antiviral Compound from the Pandemic Response Box that Efficiently Inhibits SARS-CoV-2 Infection In Vitro Microorganisms 2020 8 1872 10.3390/microorganisms8121872 33256227 7. Najjar-Debbiny R. Gronich N. Weber G. Khoury J. Amar M. Stein N. Goldstein L.H. Saliba W. Effectiveness of Paxlovid in Reducing Severe COVID-19 and Mortality in High Risk Patients Clin. Infect. Dis. 2022 76 342 349 10.1093/cid/ciac443 8. Cox R.M. Wolf J.D. Plemper R.K. Therapeutically administered ribonucleoside analogue MK-4482/EIDD-2801 blocks SARS-CoV-2 transmission in ferrets Nat. Microbiol. 2021 6 11 18 10.1038/s41564-020-00835-2 33273742 9. Beigel J.H. Tomashek K.M. Dodd L.E. Mehta A.K. Zingman B.S. Kalil A.C. Hohmann E. Chu H.Y. Luetkemeyer A. Kline S. Remdesivir for the Treatment of Covid-19--Final Report N. Engl. J. Med. 2020 383 1813 1826 10.1056/NEJMoa2007764 32445440 10. Agarwal A. Appiah J.A. Arabi Y. Blumberg L. Calfee C.S. Cao B. Cecconi M. Cooke G. Dunning J. Geduld H. A living WHO guideline on drugs for COVID-19 BMJ 2020 370 3379 10.1136/BMJ.M3379 11. Cosgriff R. Ahern S. Bell S.C. Brownlee K. Burgel P.R. Byrnes C. Corvol H. Cheng S.Y. Elbert A. Faro A. A multinational report to characterise SARS-CoV-2 infection in people with cystic fibrosis J. Cyst. Fibros. 2020 19 355 358 10.1016/j.jcf.2020.04.012 32376098 12. McClenaghan E. Cosgriff R. Brownlee K. Ahern S. Burgel P.R. Byrnes C.A. Colombo C. Corvol H. Cheng S.Y. Daneau G. The global impact of SARS-CoV-2 in 181 people with cystic fibrosis J. Cyst. Fibros. 2020 19 868 871 10.1016/j.jcf.2020.10.003 33183965 13. Lotti V. Merigo F. Lagni A. Di Clemente A. Ligozzi M. Bernardi P. Rossini G. Concia E. Plebani R. Romano M. CFTR Modulation Reduces SARS-CoV-2 Infection in Human Bronchial Epithelial Cells Cells 2022 11 1347 10.3390/cells11081347 35456026 14. Ogando N.S. Dalebout T.J. Zevenhoven-Dobbe J.C. Limpens R.W.A.L. van der Meer Y. Caly L. Druce J. de Vries J.J.C. Kikkert M. Barcena M. SARS-coronavirus-2 replication in Vero E6 cells: Replication kinetics, rapid adaptation and cytopathology J. Gen. Virol. 2020 101 925 940 10.1099/jgv.0.001453 32568027 15. De Hostos E.L. Choy R.K.M. Nguyen T. Developing novel antisecretory drugs to treat infectious diarrhea Future Med. Chem. 2011 3 1317 1325 10.4155/fmc.11.87 21859305 16. Erdem R. Ambler G. Al-Ibrahim M. Fraczek K. Dong S.D. Gast C. Mercer L.D. Raine M. Tennant S.M. Chen W.H. A Phase 2a randomized, single-center, double-blind, placebo-controlled study to evaluate the safety and preliminary efficacy of oral iOWH032 against cholera diarrhea in a controlled human infection model PLoS Negl. Trop. Dis. 2021 15 e0009969 10.1371/journal.pntd.0009969 34793441 17. Tradtrantip L. Sonawane N.D. Wan N. Verkman A.S. Nanomolar potency pyrimido-pyrrolo-quinoxalinedione CFTR inhibitor reduces cyst size in a polycystic kidney disease model J. Med. Chem. 2009 52 6447 6455 10.1021/jm9009873 19785436 18. Jin B.J. Thiagarajah J.R. Verkman A.S. Convective washout reduces the antidiarrheal efficacy of enterocyte surface-targeted antisecretory drugs J. Gen. Physiol. 2013 141 261 272 10.1085/jgp.201210885 23359285 19. Filkins L.M. O'Toole G.A. Cystic Fibrosis Lung Infections: Polymicrobial, Complex, and Hard to Treat PLOS Pathog. 2015 11 e1005258 10.1371/journal.ppat.1005258 26719892 20. Stelzer-Braid S. Johal H. Skilbeck K. Steller A. Alsubie H. Tovey E. Van Asperen P. McKay K. Rawlinson W.D. Detection of viral and bacterial respiratory pathogens in patients with cystic fibrosis J. Virol. Methods 2012 186 109 112 10.1016/j.jviromet.2012.08.008 22940004 21. Wat D. Gelder C. Hibbitts S. Cafferty F. Bowler I. Pierrepoint M. Evans R. Doull I. The role of respiratory viruses in cystic fibrosis J. Cyst. Fibros. 2008 7 320 328 10.1016/j.jcf.2007.12.002 18255355 22. Corvol H. de Miranda S. Lemonnier L. Kemgang A. Gaubert M.R. Chiron R. Dalphin M.L. Durieu I. Dubus J.C. Houdouin V. First Wave of COVID-19 in French Patients with Cystic Fibrosis J. Clin. Med. 2020 9 3624 10.3390/jcm9113624 33182847 23. Mondejar-Lopez P. Quintana-Gallego E. Giron-Moreno R.M. Cortell-Aznar I. Ruiz de Valbuena-Maiz M. Diab-Caceres L. Prados-Sanchez C. Alvarez-Fernandez A. Garcia-Marcos P.W. Penalver-Mellado C. Impact of SARS-CoV-2 infection in patients with cystic fibrosis in Spain: Incidence and results of the national CF-COVID19-Spain survey Respir. Med. 2020 170 106062 10.1016/j.rmed.2020.106062 32843180 24. Charlton F.W. Pearson H.M. Hover S. Lippiat J.D. Fontana J. Barr J.N. Mankouri J. Ion Channels as Therapeutic Targets for Viral Infections: Further Discoveries and Future Perspectives Viruses 2020 12 844 10.3390/v12080844 32756358 25. Londino J.D. Lazrak A. Jurkuvenaite A. Collawn J.F. Noah J.W. Matalon S. Influenza matrix protein 2 alters CFTR expression and function through its ion channel activity Am. J. Physiol.-Lung Cell. Mol. Physiol. 2013 304 582 592 10.1152/ajplung.00314.2012 26. Londino J.D. Lazrak A. Noah J.W. Aggarwal S. Bali V. Woodworth B.A. Bebok Z. Matalon S. Influenza virus M2 targets cystic fibrosis transmembrane conductance regulator for lysosomal degradation during viral infection FASEB J. 2015 29 2712 2725 10.1096/fj.14-268755 25795456 27. Panou M.M. Antoni M. Morgan E.L. Loundras E.A. Wasson C.W. Welberry-Smith M. Mankouri J. Macdonald A. Glibenclamide inhibits BK polyomavirus infection in kidney cells through CFTR blockade Antiviral Res. 2020 178 104778 10.1016/j.antiviral.2020.104778 32229236 28. Pearson H. Todd E.J.A.A. Ahrends M. Hover S.E. Whitehouse A. Stacey M. Lippiat J.D. Wilkens L. Fieguth H.G. Danov O. TMEM16A/ANO1 calcium-activated chloride channel as a novel target for the treatment of human respiratory syncytial virus infection Thorax 2021 76 64 72 10.1136/thoraxjnl-2020-215171 33109690 29. Braga L. Ali H. Secco I. Chiavacci E. Neves G. Goldhill D. Penn R. Jimenez-Guardeno J.M. Ortega-Prieto A.M. Bussani R. Drugs that inhibit TMEM16 proteins block SARS-CoV-2 spike-induced syncytia Nature 2021 594 88 93 10.1038/s41586-021-03491-6 33827113 30. Cesar-Silva D. Pereira-Dutra F.S. Giannini A.L.M. de Almeida C.J.G. The Endolysosomal System: The Acid Test for SARS-CoV-2 Int. J. Mol. Sci. 2022 23 4576 10.3390/ijms23094576 35562967 31. Padmanabhan P. Desikan R. Dixit N.M. Targeting TMPRSS2 and Cathepsin B/L together may be synergistic against SARS-CoV-2 infection PLoS Comput. Biol. 2020 16 e1008461 10.1371/journal.pcbi.1008461 33290397 32. Antigny F. Norez C. Becq F. Vandebrouck C. CFTR and Ca2+ signaling in cystic fibrosis Front. Pharmacol. 2011 2 67 10.3389/fphar.2011.00067 22046162 33. Anderson M.P. Gregory R.J. Thompson S. Souza D.W. Paul S. Mulligan R.C. Smith A.E. Welsh M.J. Demonstration That CFTR Is a Chloride Channel by Alteration of Its Anion Selectivity Science 1991 253 202 205 10.1126/science.1712984 1712984 34. Banerjee A.K. Blanco M.R. Bruce E.A. Honson D.D. Chen L.M. Chow A. Bhat P. Ollikainen N. Quinodoz S.A. Loney C. SARS-CoV-2 Disrupts Splicing, Translation, and Protein Trafficking to Suppress Host Defenses Cell 2020 183 1325 1339.e21 10.1016/j.cell.2020.10.004 33080218 35. Li J.Y. Liao C.H. Wang Q. Tan Y.J. Luo R. Qiu Y. Ge X.Y. The ORF6, ORF8 and nucleocapsid proteins of SARS-CoV-2 inhibit type I interferon signaling pathway Virus Res. 2020 286 198074 10.1016/j.virusres.2020.198074 32589897 36. Randhawa P.K. Scanlon K. Rappaport J. Gupta M.K. Modulation of Autophagy by SARS-CoV-2: A Potential Threat for Cardiovascular System Front. Physiol. 2020 11 1560 10.3389/fphys.2020.611275 37. Luciani A. Villella V.R. Esposito S. Brunetti-Pierri N. Medina D. Settembre C. Gavina M. Pulze L. Giardino I. Pettoello-Mantovani M. Defective CFTR induces aggresome formation and lung inflammation in cystic fibrosis through ROS-mediated autophagy inhibition Nat. Cell Biol. 2010 12 863 875 10.1038/ncb2090 20711182 38. Villella V.R. Esposito S. Bruscia E.M. Maiuri M.C. Raia V. Kroemer G. Maiuri L. Targeting the Intracellular Environment in Cystic Fibrosis: Restoring Autophagy as a Novel Strategy to Circumvent the CFTR Defect Front. Pharmacol. 2013 4 1 9 10.3389/fphar.2013.00001 23346057 39. Marat A.L. Haucke V. Phosphatidylinositol 3-phosphates--At the interface between cell signalling and membrane traffic EMBO J. 2016 35 561 579 10.15252/embj.201593564 26888746 40. Liang C. Lee J.S. Inn K.S. Gack M.U. Li Q. Roberts E.A. Vergne I. Deretic V. Feng P. Akazawa C. Beclin1-binding UVRAG targets the class C Vps complex to coordinate autophagosome maturation and endocytic trafficking Nat. Cell Biol. 2008 10 776 787 10.1038/ncb1740 18552835 41. Merigo F. Lotti V. Bernardi P. Conti A. Di Clemente A. Ligozzi M. Lagni A. Sorio C. Sbarbati A. Gibellini D. Ultrastructural Characterization of Human Bronchial Epithelial Cells during SARS-CoV-2 Infection: Morphological Comparison of Wild-Type and CFTR-Modified Cells Int. J. Mol. Sci. 2022 23 9724 10.3390/ijms23179724 36077122 42. Chen D. Zheng Q. Sun L. Ji M. Li Y. Deng H. Zhang H. ORF3a of SARS-CoV-2 promotes lysosomal exocytosis-mediated viral egress Dev. Cell 2021 56 3250 3263.e5 10.1016/j.devcel.2021.10.006 34706264 43. Pizzato M. Baraldi C. Boscato Sopetto G. Finozzi D. Gentile C. Gentile M.D. Marconi R. Paladino D. Raoss A. Riedmiller I. SARS-CoV-2 and the Host Cell: A Tale of Interactions Front. Virol. 2022 1 46 10.3389/fviro.2021.815388 44. Zhang J. Kennedy A. Xing L. Bui S. Reid W. Joppich J. Ahat E. Rose M. Tang Q. Tai A.W. SARS-CoV-2 triggers Golgi fragmentation via down-regulation of GRASP55 to facilitate viral trafficking bioRxiv 2022 10.1101/2022.03.04.483074 45. Ghosh S. Dellibovi-Ragheb T.A. Kerviel A. Pak E. Qiu Q. Fisher M. Takvorian P.M. Bleck C. Hsu V.W. Fehr A.R. b-Coronaviruses Use Lysosomes for Egress Instead of the Biosynthetic Secretory Pathway Cell 2020 183 1520 1535.e14 10.1016/j.cell.2020.10.039 33157038 46. Liu J. Walker N.M. Ootani A. Strubberg A.M. Clarke L.L. Defective goblet cell exocytosis contributes to murine cystic fibrosis-associated intestinal disease J. Clin. Investig. 2015 125 1056 1068 10.1172/JCI73193 25642775 47. Edlund A. Barghouth M. Huhn M. Abels M. Esguerra J.S.E. Mollet I.G. Svedin E. Wendt A. Renstrom E. Zhang E. Defective exocytosis and processing of insulin in a cystic fibrosis mouse model J. Endocrinol. 2019 241 45 57 10.1530/JOE-18-0570 48. Eymieux S. Uzbekov R. Rouille Y. Blanchard E. Hourioux C. Dubuisson J. Belouzard S. Roingeard P. Secretory Vesicles Are the Principal Means of SARS-CoV-2 Egress Cells 2021 10 2047 10.3390/cells10082047 34440816 49. Sbarigia C. Vardanyan D. Buccini L. Tacconi S. Dini L. SARS-CoV-2 and extracellular vesicles: An intricate interplay in pathogenesis, diagnosis and treatment Front. Nanotechnol. 2022 4 66 10.3389/fnano.2022.987034 50. Nakase I. Ueno N. Matsuzawa M. Noguchi K. Hirano M. Omura M. Takenaka T. Sugiyama A. Bailey Kobayashi N. Hashimoto T. Environmental pH stress influences cellular secretion and uptake of extracellular vesicles FEBS Open Bio 2021 11 753 10.1002/2211-5463.13107 |
PMC10000630 | Introduction The management of obesity is difficult with many failures of lifestyle measures, hence the need to broaden the range of treatments prescribed. The aim of our work was to study the influence of pre and probiotics on weight loss psychological profile and metabolic parameters in obese patients. Methods It is a clinical trial involving 45 obese patients, recruited from the Obesity Unit of the National Institute of Nutrition between March and August 2022 divided into three groups: diet only (low-carbohydrate and reduced energy diet), prebiotics (30 g of carob/day) and probiotics (one tablet containing Bifidobacterium longum, Lactobacillus helveticus, Lactococcus lactis, Streptococcus thermophilus/day). The three groups were matched for age, sex and BMI. Patients were seen after 1 month from the intervention. Anthropometric measures, biological parameters, dietary survey and psychological scores were performed. Results The average age of our population was 48.73 +- 7.7 years, with a female predominance. All three groups showed a significant decrease in weight, BMI and waist circumference with p < .05. Only the prebiotic and probiotic group showed a significant decrease in fat mass (p = .001) and a significant increase in muscle strength with p = .008 and .004, but the differences were not significant between the three groups. Our results showed also a significant decrease in insulinemia and HOMA-IR in the prebiotic group compared to the diet-alone group (p = .03; p = .012) and the probiotic group showed a significant decrease in fasting blood glucose compared to the diet alone group (p = .02). A significant improvement in sleep quality was noted in the prebiotic group (p = .02), with a significant decrease in depression, anxiety and stress in all three groups. Conclusions The prescription of prebiotics and probiotics with the lifestyle measures seems interesting for the management of obesity especially if it is sarcopenic, in addition to the improvement of metabolic parameters and obesity-related psychiatric disorders. The supplementation with prebiotics and probiotics showed an improvement in lean mass, glycaemic profile, insulin resistance and uric acid more than diet alone. obesity prebiotics probiotics weight loss source-schema-version-number2.0 cover-dateMarch 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.2.6 mode:remove_FC converted:10.03.2023 Ben Othman R , Ben Amor N , Mahjoub F , et al. A clinical trial about effects of prebiotic and probiotic supplementation on weight loss, psychological profile and metabolic parameters in obese subjects. Endocrinol Diab Metab. 2023;6 :e402. doi:10.1002/edm2.402 pmc1 INTRODUCTION Today, obesity is a global epidemy according to the World Health Organization, given the increase in its frequency in the world, and its responsibility in the appearance of several chronic pathologies, such as type 2 diabetes, hypertension, cardiovascular diseases, respiratory diseases, osteoarticular diseases, cancer and other pathologies. In 2021, the WHO announced that more than 40% of men and women, or 2.2 billion people, are overweight and that an unbalanced diet was responsible for at least 8 million deaths per year. It is estimated that by 2025, 167 million people would be at risk of impaired health due to obesity. 1 In Tunisia, the prevalence of obesity was 26.2% in 2016 according to the results of the "Tunisian Health Examination Survey-2016". 2 This disease is multifactorial, among the contributing factors of obesity are: a high-fat diet, a sedentary lifestyle, but also the imbalance of the intestinal flora, "the gut microbiota" 3 which today represents the focus of several publications. Gut microbiota is defined by all the beneficial microorganisms that live and grow in the intestine. It is set up from birth and evolves according to different factors such as antibiotic treatments or diet (presence of fibres, richness of foods in pre and probiotics). Probiotics are living microorganisms, they are bacteria such as Lactobacilli, Bifidobacteria, Streptococci and many others or yeasts. They can be present naturally in our diet, especially in fermented foods such as certain yoghurts or fermented milks, whereas prebiotics represent substrates for these bacteria which allow them to ensure their growth and thus exercise their beneficial roles, they are also provided by our diet, from the dietary fibres present in vegetables and fruits, such as carob, chicory and others. Today, the microbiota is considered a therapeutic revolution, where researchers use its enrichment to prevent or treat certain diseases including obesity, 4 such as faecal transplantation, 5 but also the enrichment of the microbiota by prebiotics and probiotics to treat obesity. 6 , 7 Hence, our interest in transposing these theoretical results to clinical practice. Aim: The objective of this interventional clinical trial was to evaluate the effects of a probiotic supplement containing Bifidobacteruim, Lactobacillus strains and a prebiotic supplement by carob on the changes in body composition and metabolic biomarkers in subjects with obesity (main purpose), we also checked the psychological profile of the population (quality of sleep, stress, anxiety and depression) as secondary purpose. 2 MATERIALS AND METHODS We conducted a prospective interventional study at the obesity unit at the Zouhair El Kallel National Institute of Nutrition and Food Technology of Tunis, from March 2022 to August 2022. We included in our study obese patients (BMI >=30 kg/m2) aged over 18 years. Patients with: renal failure, hypothyroidism, cancer, diabetic patients on insulin, on long-term corticosteroid therapy, former patients of the obesity unit were not included. No participants dropped out of the study during the intervention period. Forty-five patients were recruited on their first visit to the obesity unit (T0) and were randomly assigned to three groups matched for age, sex and BMI. All participants were enrolled in the weight loss program at the beginning of the study and followed a low-carbohydrate, reduced-energy intake eating plan provided by the same dietician. First group called "diet only": on low-calorie diet alone without any intervention (15 patients). Second group: 15 patients on the same diet plan but additionally received prebiotic supplementation (2 carob beans/day about 30 g) called "prebiotic group". Third group: same diet with probiotic supplementation (n = 15). The probiotic component used in the study was one tablet containing an association of four microbiological strains which are: Bifidobacteruim longum, Lactobacillus helveticus, Lactococcus lactis, Streptococcus thermophilus (1 tablet (10.109 UFC/capsule)/day) called "probiotic group". The probiotic supplement was produced by Pileje Labs. Patients were reassessed after 1 month (T1) and we track adherence by regular phone calls. All subjects gave their informed consent for participating in the study. The study was approved by the ethical committee of the national institute of nutrition of Tunis and the clinical trial was registered under number PACTR202210705998795 in the Pan African Clinical Trial Registry. Body mass index (BMI) was calculated using body weight and height measured with bare feet and in minimal clothing according to the World Health Organization definition and classification. 8 Body composition parameters (body fat mass and percentage and body lean mass) were acquired before and after 1 month of intervention by impedance meter TANITA BC418MA. We took the waist circumference of the patients. Muscle strength was measured by the handgrip. Sarcopenia was defined by muscle strength lower than 27 kg for men and 16 for women. A biological assessment was carried out at T0 and T1 including: fasting glycaemia, HbA1C, Cholesterol, triglycerides, HDL, calculated LDL Friedwald formula, 9 insulinemia, calculated HOMA-IR (HOMA-IR = (insulin (mU/l) x glycaemia (mmol/l))/22.5), AST, ALT, GGT, creatinine and calculated eGFR. Blood glucose results were interpreted according to American diabetes association guidelines. 10 We looked at the physical examination for blood pressure and other complications of obesity such as hernia, sleep apnoea syndrome, osteoarthritis and NASH and if necessary we completed with the necessary radiological examinations. All patients benefited from an interview including food survey, stress questionnaire (Cunji), sleep questionnaire (Epworth), symptoms of depression and anxiety (HADS). For the evaluation of stress, we used the brief stress evaluation scale, this is the scale of Cungi 1997. 11 This scale is made up of 11 items, and for each the response is from 1 to 6. The evaluation of the quality of sleep was carried out using the Epworth Sleepiness Scale, 12 this questionnaire assesses the level of daytime sleepiness of the patient. It is composed of eight items, and for each situation, the patient must select an answer from (0 to 3). The interpretation is as follows: A total of less than 10 suggests that there is no excessive daytime sleepiness. A total of 10 and above suggests excessive daytime sleepiness. To assess the depressive state of the patients, we used the "HAD" scale (Hospital Anxiety and Depression Scale). 13 This is a structured questionnaire of 14 items. This questionnaire consists of two subscales, each having 7 items, one for anxiety, the other for depression. Each item is rated on a 4-point scale, that is from 0 to 3, evaluating the intensity of symptoms over the past week. The scores therefore range from 0 to 21 and the highest scores correspond to the presence of more severe symptoms. The addition of the scores obtained for each item allows the following interpretation: Less than 7 points: no symptoms of depression. Eight to 10 points: doubtful symptomatology. Eleven and over: certain symptomatology. 2.1 Statistical analysis The three-variable ANOVA with Student's t test for paired series were used for group comparison of the body composition and metabolic parameters at T1 and T0 (SPSS Statistics, v. 25). The results were expressed as mean +- SD, and mean differences were considered significant at p < .05. 3 RESULTS The average age of our population was 48.73 +- 7.7 years with extremes ranging from 33 to 63 years. Half of the population (51%) was over 50 years old. The majority of participants were female 93.3% (n = 42) against 6.7% (n = 3) of men. Past medical history, complications and lab test results are present in Table 1. TABLE 1 Past medical history, complications of obesity and lab test results Diet only (%) Prebiotic (%) Probiotic (%) p Past medical history Diabetes 6.7 26.7 13.3 .3 Hypertension 6.7 20 33.3 .2 Dyslipidaemia 6.7 26.7 13.3 .3 Active smokers (%) 6.7 13.3 6.7 .7 Osteoarthritis (%) 33.3 20 26.7 .6 Sleep apnoea syndrome (%) 26.7 66.7 33.3 .07 Hernia (%) 13.3 6.7 13.3 .6 NASH (%) 24 30 24 .42 Diabetes (%) 13.3 53.3 20 .06 Prediabetes (%) 13.3 6.7 33.3 .06 Insulin resistance (%) 6.7 0 7.1 .65 High TG levels (%) 46.7 66.7 26.7 .18 Low HDL levels(%) 46.7 46.7 33.3 .27 High LDL levels(%) 26.7 33.3 33.3 .4 Blood pressure values are comparable in the three groups. Our three groups were matched for BMI. There was no statistically significant difference for anthropometric measurements (weight, height, IMC, fat mass, muscle mass and waist circumference) between the three groups. In addition, the majority of patients in all three groups had normal muscle strength. Sarcopenia at T0 was noted in 20% in the diet-only group, 6.7% in the prebiotic group and 13.3% in the probiotic group. In each group, 93.3% of patients were sedentary. At recruitment, we performed a frequency questionnaire consumption of foods rich in prebiotics and probiotics such as coffee, tea, garlic, onion, fermented foods, cacao, yoghurts and fruits. There were no differences between groups. No patient reported alcohol consumption and none had a regular consumption of carob. Most of the patients of the three groups had a high level of anxiety, depression and stress but without statistically significant difference. The result of the intervention after 1 month are in Table 2. TABLE 2 Results of the intervention on different parameters for the three groups Diet only Prebiotic Probiotic T0 T1 p T0 T1 p T0 T1 p Weight (kg) 103.7 101.2 .001 103.5 101.6 .003 106.09 104.4 .02 Fat mass (kg) 46.7 44.8 .07 47.3 44.3 .001 47.5 45.01 .001 Lean mass (kg) 54.06 53.5 .3 55.1 55.6 .2 55.5 56.4 .08 Waist circumference (cm) 119 117.3 .01 124 120 .03 122 119 .001 Muscle strength (kg) 24.4 24.3 .8 27.4 28.8 .008 24.8 26.5 .004 Systolic blood pressure (mmHg) 13 12.8 .6 13 12.2 .03 13.3 12.6 .01 Fasting glucose (mmol/l) 5.3 5.5 .27 7.5 5.1 .2 5.66 5.6 .6 HbA1c (%) 5.6 5.5 .03 6.6 6.3 .3 5.8 5.6 .003 Insulin (mUI/l) 18.4 15.2 .07 23.8 14.5 .002 17.5 13.7 .005 HOMA-IR 4.3 3.8 .2 9.1 3.8 .009 4.5 3.4 .009 Cholesterol (mmol/l) 5.2 4.8 .03 5.3 4.9 .005 5.2 4.6 .08 HDL (mmol/l) 1.6 1.05 .9 1.07 1.08 .8 1.2 1.22 .7 LDL (mmol/l) 3.2 2.9 .05 3.3 2.9 .003 3.2 2.8 .004 Triglycerides (mmol) 1.7 1.6 .4 1.9 1.4 .001 1.6 1.4 .03 ALAT (UI/l) 21.2 18.4 .03 21.3 20.8 .7 21.1 17.8 .01 Uric acid 277.2 289.4 .4 346.3 365.3 .3 295.7 284.2 .1 Epworth 9.8 8.6 .06 8.7 7 .02 10.2 7.9 .03 Anxiety 13.3 11.2 .02 11.4 9.4 .01 13.3 11.6 .06 depression 12.4 9.9 .001 11.2 8.06 .01 11.5 8.9 .001 Stress 40.1 33.4 .01 36.2 31.3 .001 35.6 29.3 .002 Bold value indicates statistically significant p < 0.05. The results of anthropometric measurements after the intervention in the three groups showed a statistically significant decrease in weight, BMI and WC, but muscle strength has increase only with pre and probiotics. The population has significantly decreased energy and macronutrient (protein, carbohydrate and lipid) intake, with a significant decrease in sugar and sodium intake. A significant increase in fibre intake was noted in the diet and prebiotic group but not in the probiotic group. The quality of sleep was not improved by the diet only and probiotics did not enhance anxiety. Taking probiotics was associated with the occurrence of diarrhoea in 20% of cases (p < .001). Then we compared the diet alone versus prebiotics group for all the parameters listed in Table 3. The difference was not significant. Then it was the diet alone group versus probiotics and finally prebiotics versus probiotics. TABLE 3 Comparison of biological assessments according to the intervention Mean difference (T0-T1) p Mean difference (T0-T1) p Mean difference (T0-T1) p Diet only Prebiotics Diet only Probiotics Prebiotics Probiotics Fasting glycaemia (mmol/l) -0.18 1.6 .016 -0.18 0.06 .02 1.6 0.06 .3 HbA1c (%) 0.12 a 0.2 .17 0.12 a 0.18 a .3 0.2 0.18 a .4 Insulin (mUI/l) 2.9 9.3 a .03 2.9 3.8 a .3 9.3 a 3.8 a .2 HOMA-IR 0.5 5.3 a .012 0.5 1.04 a .1 5.3 1.04 a .2 Uric acid -12.2 -19.07 .75 -12.2 11.6 .001 -19.07 11.6 .02 Bold value indicates statistically significant p < 0.05. a psignificant between T0 and T1 month. Our conclusion is that the different therapeutic means are equal for the dietary survey, the different scores (stress, sleep, anxiety and depression). The influence of the three means on weight loss is equivalent even if it is the diet alone group which reduced the weight more except for the lean mass which was clearly increased by probiotics compared to diet (p = .05). On the other hand, significant differences between the three means were found in the results of the blood tests represented in Table 3. Prebiotics and probiotics were better than diet for the reduction of fasting glycemia and insulin resistance but probiotics did not lower uric acid as much as others. 4 DISCUSSION This study was an interventional clinical trial designed to examine the effects of a combination of probiotic bacteria B. longum, L. helveticus, L. lactis, S. thermophilus and a prebiotic supplement by 30 g/day of carob on changes in body composition, metabolic biomarkers and psychological profile in obese human subjects enrolled on a weight loss program. The weight loss program was a low-carbohydrate, energy-restricted eating plan. The study has confirmed that a low-carbohydrate, restricted-energy diet can be effectively used for weight loss in obese individuals. Our work has some strength--to our knowledge in Tunisia no one studied the association between prebiotics or probiotics and obesity, the only Tunisian study that has worked on the microbiota has studied the imbalance of the microbiota in diabetic patients. 14 The use of carob as a prebiotic for weight loss is an innovation that fits into abandoned Tunisian habits. Carob is available at a nominal cost less than some fruits and vegetables. Our study focused on several parameters apart from anthropometry, such as biology and other assessment tests such as the Epworth score, the HAD and the Cungi stress score but it has some limitations like the small number of patients for each group and microbiological analysis for the gut microbia was not performed. In addition, the study was conducted over a month; perhaps a longer duration of intervention would show other results. Many studies have shown the effect of pre or probiotic on the weight loss. Sergeev et al., 15 compared the effect of symbiotic supplementation (prebiotic and probiotic) on the body composition of obese patients against a placebo group which received only a low-calorie diet, they found a significant decrease in weight in both groups. However, the study of Hiel et al., 16 using inulin as prebiotic compared to placebo, found a significant reduction in weight in the prebiotic group. This difference may be due to the difference in the prescribed diet and also to the difference in the number of patients. In addition, the study by Stenman et al., 17 which is a study that compared the effect of prebiotic alone, probiotic alone and prebiotic+probiotic to a placebo group, found that only the probiotic alone group presented weight loss compared to the other groups. Some other studies did not found a difference between groups. 18 , 19 This difference may be due to the difference in the diet given and also the type of prebiotic and probiotic used. Similarly, Rodriguez in their studies showed that there were responders and non-responders in obese patients treated with prebiotics depending on the initial species of intestinal flora present in the host during the intervention. 20 Indeed, the microbiota intervenes in the regulation of energy expenditure by acting on specific hormones, thanks to a bidirectional signalling between the brain and the intestine, the gut microbiota regulates appetite and energy expenditure then follows a weight regulation. 21 Prebiotics act on the microbiota by increasing the production of short-chain fatty acids, which in turn causes a cascade of modifications leading to weight reduction and improved metabolic parameters. 22 Our study showed a significant increase in muscle strength in both the prebiotic group and the probiotic group. As well as Zahao and Kang in their studies. 23 , 24 Alteration of the gut microbiota has been shown to directly affect muscle strength. Probiotics, prebiotics and short-chain fatty acids are potential new therapies to improve lean mass and physical performance. Strains of Lactobacillus and Bifidobacterium (present in Lactibiane*) can restore age-related muscle loss. The pathways by which microbiota influence muscle are diverse and complex. 25 Our results showed a beneficial effect of prebiotics and probiotics on carbohydrate metabolism. These results were in agreement with the study conducted by Miller et al., 26 which found that the symbiotic yoghurt protected mice against diabetes by significantly improving fasting blood glucose levels versus unenriched control yoghurt. In addition, a preparation rich in fibre and lactulose as prebiotics used in an old clinical study, 27 showed a decrease in blood sugar in 10 obese patients. Oral supplementation with prebiotics and probiotics acts on the regulation of glycaemia, the mechanism of action consists in reducing the secretion of inflammatory markers such as IFN-g and IL-1b by increasing the production of IL-10 anti-inflammatory. In addition, probiotics stimulate the secretion of the neurotransmitter GABA which decreases the production of glucagon and stimulates the production of insulin. 28 , 29 Our study showed a decrease in uric acid in the probiotic group with a significant difference compared to the diet-alone group and the prebiotic group. To study the effect of probiotics on uric acid, there was first the pilot study of Garcia-Arroyo carried out in 2018 on six rats which affirmed this hypothesis. 30 Then other studies followed with the same results. 31 , 32 The decrease in energy intake found after prebiotic and probiotic supplementation is explained by the stimulation of leptin secretion and the decrease in ghrelin secretion, which increase satiety and consequently decrease in intake. In addition, the reduction of microbiome lipopolysaccharides by pre and probiotics promotes reduced appetite by increasing satiety. 33 A decrease in Epworth score was found in all three groups. Our study was consistent with others. 34 , 35 However, the study by Buigues et al. 36 did not show conclusive results of prebiotics on sleep quality. Following the fermentation of fibres from prebiotics by microbiota, there will be production of butyrate which improves sleep quality 37 but the mechanisms involved are more complex than that. 38 The three means were comparable in their influence on depression and anxiety. Other studies proved a good improvement of these symptoms when patients took probiotic. 39 , 40 It has been shown that probiotics stimulate the production of inhibitory neurotransmitters such as the neurotransmitter GABA, which causes a reduction in anxiety and depression. 41 On the other hand, the imbalance of the gut microbiota is responsible for the occurrence of depression by the decrease in the production of some lipid metabolites (endogenous cannabinoids). 42 As for the stress, prebiotics and probiotics increase the production of serotonin, which is a molecule involved in mood regulation, by stimulating the synthesis of tryptophan 43 which improves the symptoms of stress. 5 CONCLUSION The imbalance in the functioning of the body is due on the one hand to the imbalance of the gut microbiota because of obesity which alters the beneficial microorganisms and on the other hand this alteration which further promotes obesity by several mechanisms and signalling pathways. 44 The intestinal microbiota, as it is called the second brain, intervenes in the regulation of the functioning of the organism, which has been demonstrated by several studies. Hence the importance of modulating the gut microbiota with prebiotics and probiotics to treat obesity and improve related metabolic parameters. In the light of this study and other studies, it is advisable to take certain measures to treat obesity: Follow a diet balanced in energy intake to prevent the alteration of the gut microbiota. Enrich the diet with foods rich in prebiotics and probiotics, either to prevent the onset of obesity or to treat it. Treatment with pre and probiotics should be considered in case of sarcopenic obesity. Adopt treatment with prebiotics and probiotics, especially if obesity is linked to a glycaemic disorder. Prescription of prebiotics and probiotics can Improve the quality of sleep, anxiety and stress in some cases. AUTHOR CONTRIBUTIONS Nadia Ben Amor: Visualization (equal). Faten Mahjoub: Visualization (equal). Olfa Berriche: Visualization (equal). Chaima El Ghali: Investigation (equal). Amel Gamoudi: Project administration (equal). Henda Jamoussi: Writing - review and editing (equal). FUNDING INFORMATION This research received no funding. CONFLICT OF INTEREST The authors declare no conflict of interest. DATA AVAILABILITY STATEMENT Data sharing is not applicable to this article as no new data were created or analyzed in this study. |
PMC10000631 | Introduction Type 2 diabetes mellitus (T2DM) is among the world's top 10 leading causes of death. Additionally, prediabetes is a major risk factor for diabetes. Identifying diabetes co-occurring disorders can aid in reducing adverse effects and facilitating early detection. In this study, we evaluated dyslipidaemia, metabolic syndrome (MetS), and liver enzyme levels in pre-diabetic and T2DM patients in the Persian cohort compared to a control group. Materials and Methods In this cross-sectional study, 2259 pre-diabetes, 1664 T2DM and 5840 controls (35-70 years) who were selected from the Hoveyzeh cohort centre were examined. Body mass index, blood pressure, fasting blood glucose (FBG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), triglyceride (TG) and liver enzymes: g-glutamyltransferase (GGT), alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were determined using the standard protocols. MetS subjects were also identified based on the National Cholesterol Education Program guidelines. Results Prediabetes and T2MD were closely correlated with the lipid profile, MetS, and liver enzymes (ALT, GGT, ALT/AST). MetS increases the risk of T2DM by 12.45 [95% CI: 10.88-14.24] fold, while an increase in ALT/AST ratio increases the risk of T2DM by 3.68 [95% CI: 3.159-4.154] fold. ROC curve analysis also revealed the diagnostic roles of GGT, ALT, AST and the ALT/AST ratio among pre-diabetics, diabetics and the control group. The GGT level corresponds to the highest AUCs (0.685) with the highest sensitivity (70.25%). Conclusions Our results indicated a significant increase in liver enzymes, lipid profile and MetS status in both pre-diabetic and T2MD subjects, with the differences being more pronounced in diabetic individuals. Consequently, on the one hand, these variables may be considered predictive risk factors for diabetes, and on the other hand, they may be used as diagnostic factors. In order to confirm the clinical applications of these variables, additional research is required. Our results indicated a significant increase in liver enzymes, lipid profile and MetS status in both pre-diabetic and T2MD subjects, with the differences being more pronounced in diabetic individuals. Hoveyzeh lipid profile liver enzymes metabolic syndrome Persian cohort T2DM Ahvaz Jundishapur University of Medical Sciences 10.13039/501100005001 HCS-9820 The Iranian Ministry of Health and Medical Education700/534 source-schema-version-number2.0 cover-dateMarch 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.2.6 mode:remove_FC converted:10.03.2023 Dinarvand N , Cheraghian B , Rahimi Z , Salehipour Bavarsad S , Bavarsad A , Mohammadtaghvaei N . Examining dyslipidaemia, metabolic syndrome and liver enzyme levels in patients with prediabetes and type 2 diabetes in population from Hoveyzeh cohort study: A case-control study in Iran. Endocrinol Diab Metab. 2023;6 :e401. doi:10.1002/edm2.401 pmc1 INTRODUCTION Diabetes mellitus, as a metabolic disorder, 1 is one of the most prevalent global public health issues 2 and contributes to a rise in morbidity and mortality. 3 According to estimates from the International Diabetes Federation (IDF), 1 in 11 individuals between the ages of 20 and 79 had type 2 diabetes mellitus (T2DM) in 2015, 4 which could reach 629 million by 2045. 2 Diabetes is hyperglycaemia resulting from insulin deficiency, insulin resistance or both. 1 , 3 Prediabetes is a major diabetes risk factor. 2 It is a hyperglycaemic condition marked by impaired fasting glucose (IFG), impaired glucose tolerance (IGT) or glycated haemoglobin (A1C) of 6.0%-6.4%, or a combination of these. 1 , 2 Both dyslipidaemia and hypertension are significant risk factors for T2DM. According to the American Diabetes Association, patients with T2DM who have dysregulated levels of lipids such as total cholesterol, triglycerides, very-low-density lipoprotein (VLDL), low-density lipoprotein (LDL) and high-density lipoprotein (HDL) are diagnosed with diabetic dyslipidaemia. Alternatively, lipid markers may be a useful predictor of risk in diabetic patients. 5 In addition, prediabetes and T2DM are common metabolic syndrome (MetS) manifestations. 1 Some studies indicate that individuals with metabolic syndrome are four times more likely to develop T2DM. 6 MetS are characterized by hypertriglyceridemia, low HDL cholesterol, abdominal obesity or a high BMI ratio, glucose intolerance or insulin resistance, hypertension and microalbuminuria. 7 Insulin resistance syndrome may result in hepatic dysfunction, resulting in T2DM. 6 Therefore, patients with advanced liver disease have a higher incidence of diabetes than the general population. 8 Conversely, releasing free fatty acids (FFAs) due to T2DM decreases hepatic mitochondrial function. In turn, this causes further triglyceride storage in the hepatocyte and, ultimately, liver damage. 8 Serum levels of liver enzymes, such as alanine aminotransferase (ALT), aspartate aminotransferase (AST), and to a lesser extent g-glutamyltransferase (GGT), are frequently used as indicators of liver damage. 9 In the past decade, several studies have linked serum concentrations of these enzymes to multiple metabolic syndrome symptoms, including hepatic insulin resistance, T2DM and dyslipidaemia. 9 , 10 , 11 Since then, little research has been conducted on the relationship between dyslipidaemia, metabolic syndrome and liver enzyme levels in pre-diabetic and T2DM patients. In order to determine the relationship between these risk factors and the development of prediabetes and diabetes in the adult population of Hoveyzeh cohort centre, this study was conducted on three groups: healthy, pre-diabetic and T2DM. 2 MATERIALS AND METHODS We conducted a cross-sectional study in men and women aged 35-70 who underwent a comprehensive health screening exam at the Hoveyzeh cohort centre for Prospective Epidemiological Research Studies in Iran (PERSIAN), a region in Iran's Southwest Khuzestan province, between 1 May 2016 and 31 August 2018. Therefore, 10,009 people were recruited in Hoveyzeh cohort centre. Patients with any of the following conditions at baseline were excluded from the study: a history of cancer, renal failure, known liver disease, ALT>3 times normal, alcohol consumption, recent (1 year) MI, acute coronary syndrome, stroke and weight loss of more than 5 kg in a month as well as microvascular complications. Finally, 9763 of 10,009 cases had the criteria for this study. All participants then completed questionnaires, including demographic information, cigarette smoking, opium use, consumable drugs, disease history and physical activity. First, blood samples for analysis were obtained from the antecubital vein of patients and subjects who had fasted for 10 to 12 h. In the central laboratory of the Hoveyzeh cohort centre, all biochemical parameters were measured using standardized protocols on automated equipment. Fasting serum glucose was assayed using the hexokinase/glucose-6-phosphate dehydrogenase method. Diabetes was defined as FBS levels >=126 mg/dl or receiving anti-diabetic drugs or self-reported diagnosis of diabetes. Standard enzymatic colorimetric techniques were used to measure serum total cholesterol (TC), triacylglycerol (TG) and high-density lipoprotein cholesterol (HDL-C) levels. The level of low-density lipoprotein cholesterol (LDL-C) was determined using the Friedewald et al. formula (LDL-C = TC - HDL - VLDL cholesterol). 9 The levels of AST, ALT and GGT were determined using the International Federation of Clinical Chemistry's method. All these analyses were done using commercial kits (Pars Azmon Inc.). MetS is defined by three or more of the following National Cholesterol Education Program criteria: high TG (>=150 mg/dl); low HDL-C (<=40 mg/dl) for men and <50 for women; high fasting blood sugar (>=100 mg/dl) or known type 2 diabetes; hypertension (at least 135/85 mmHg or receiving antihypertensive medication); and a waist circumference greater than 102 cm for men and 88 for women. 6 , 12 , 13 2.1 Statistical analysis The statistical analyses were conducted using SPSS (v. 15.0). For quantitative variables, data were presented as mean +- standard deviation; for qualitative variables, data are expressed as frequency (number (%)), The normality of data was determined using the Kolmogorov-Smirnov test, and the chi-square test was used to determine the association between qualitative variables. Differences between the two groups were calculated by Mann-Whitney tests for skewed data. In addition, the Kruskal-Wallis test was used to compare variables in three groups. Moreover, logistic regression analysis was employed to calculate studied risk factors for prediabetes and diabetes vs. control group. Then, multivariable model was performed for adjusting of age, gender and BMI. Receiver operating characteristic (ROC) curve analysis was used to determine the prognostic relationship of liver enzymes and lipid profile in prediabetes and diabetes. All p-values were two-tailed, and p < .05 were considered statistically significant. 3 RESULTS 3.1 Characteristics of the study participants according to FBS tertiles The final database contained 9763 subjects (3809 males and 5954 females); subjects were divided into three groups based on FBS levels. Table 1 illustrates the characteristics of three distinct groups. T2DM prevalence was 17.0% (18.1% in males and 16.4% in females), prediabetes prevalence was 23.1% (21.0% in males and 24.5% in females), and control prevalence was 59.8% (60.9% in males, 59.1% in females). Participants with prediabetes and T2DM were older and had a higher BMI, waist circumference, diastolic blood pressure (DBP) and systolic blood pressure (SBP) than control subjects. TABLE 1 Anthropometrics and biochemical characteristics of the study participants according to the tertiles of FBS Variables Fasting glucose (mg/dl) FBS <=100 5840 (59.82%) FBS: 100-125 2259 (23.14%) FBS>=126 1664 (17.04%) p-Value Anthropometrics Gender Male 2321 (60.9%) A 799 (21.0%) B 689 (18.1%) C <.0001** Female 3519 (59.1%) A 1460 (24.5%) B 975 (16.4%) C Age (year) 47.03 +- 8.79 A 50.53 +- 9.27 B 52.81 +- 8.89 C <.0001* Waist Circumference (cm) 97.62 +- 11.83 A 102.45 +- 12.09 B 103.15 +- 11.46 B <.0001* BMI (kg/m2) 28.14 +- 5.15 A 30.03 +- 5.52 B 29.63 +- 5.23 C <.0001* Diastolic blood pressure (mmHg) 70.32 +- 10.97 A 72.60 +- 11.38 B 73.16 +- 11.51 B <.0001* Systolic blood pressure (mmHg) 110.46 +- 16.95A 115.49 +- 18.90 B 118.27 +- 20.19 C <.0001* Metabolic syndrome No 4409 (75.5%) A 630 (27.9%) B 330 (19.8%) C <.0001** Yes 1413 (24.5%) A 1629 (72.1) B 1334 (80.2%) C Biochemicals FBS (mg/dl) 88.97 +- 6.54 A 108.25 +- 6.96 B 201.48 +- 67.98 C <.0001* LDL (mg/dl) 105.62 +- 31.17A 109.51 +- 33.83A 106.86 +- 37.33B <.0001* TG (mg/dl) 147.6 +- 84.2 A 170.06 +- 107.3 B 202.02 +- 135.4 C <.0001* Total Cholesterol (mg/dl) 185.76 +- 37.1 A 193.97 +- 40.8 B 196.34 +- 47.9 B <.0001* HDL (mg/dl) 50.68 +- 12.24 A 50.34 +- 11.75 A 49.21 +- 11.61 B <.0001* Hepatic enzymes AST (units/L) 18.06 +- 7.62 A 19.30 +- 9.19 A 17.36 +- 9.05 B <.0001* ALT (units/L) 20.52 +- 13.72 A 22.03 +- 14.88 B 22.25 +- 13.35 C <.0001* GGT (units/L) 24.14 +- 16.61 A 27.05 +- 17.68 B 34.73 +- 34.01 C <.0001* ALT/AST 1.06 +- 0.384 A 1.10 +- 0.380 B 1.28 +- 0.514 C <.0001* Note: The total 9763 subjects were divided into tertiles according to FBS distribution and the significance of any differences in means or proportions were tested with analysis of Kruskal-Wallis* and chi-squared** tests, respectively. The results are presented as the means +- SD. Similar letters (A, B and C) indicate that there is no significant difference between the mean in the groups (p-value > .05). Different letters (A, B and C) indicate that there is a significant difference between the mean in the groups (p-value < .05). Abbreviations: ALT, Alanine aminotransferase; AST, Aspartate aminotransferase; BMI, Body mass index; FBS, Fasting blood sugar; GGT, Gamma-glutamyltransferase; HDL, High density lipoprotein; LDL, Low density lipoprotein. In prediabetes and T2DM, biochemical variables, including TG, were significantly higher than in the control group. Compared to the control group, prediabetes and diabetes had significantly higher mean total cholesterol levels, whereas there was no significant difference between prediabetes and diabetes. In addition, the mean LDL in diabetes and normal groups was significantly higher than in the prediabetes group, but there was no significant difference between diabetes and normal groups. In contrast, the HDL level was significantly lower in T2DM compared to prediabetes and the control group, whereas there was no significant difference between prediabetes and the control group. Those who developed prediabetes and T2DM had significantly higher levels of hepatic enzymes, including GTT and ALT, compared to the control group. In contrast, the mean AST was significantly lower in T2DM than in prediabetes and the control group, and there was no significant difference between prediabetes and the control group (Table 1). 3.2 ROC curve analysis Receiver operating characteristic curve analysis revealed the significance of GGT, ALT, AST and the ALT/AST ratio in identifying prediabetes or diabetes . The ROC curve analysis is presented in Table 2. TABLE 2 Receiver operating characteristic curve analysis of GGT, ALT, AST and ALT/AST ratio in diabetes and pre-diabetes, respectively Variable AUC 95% CI p-Value Cut-off value Sensitivity(%) Specificity(%) Diabetes GGT 0.685 0.673-0.694 <0.0001 21.36 70.25% 57.76% ALT 0.564 0.553-0.575 <0.0001 14 70.85% 39.88% AST 0.412 0.395-0.428 <0.0001 14 43.33% 71.64% ALT/AST 0.669 0.658-0.679 <0.0001 1.06 68.87% 57.29% Pre-diabetes GGT 0.573 0.562-0.583 <0.0001 20.33 57.10% 53.89% ALT 0.537 0.526-0.548 <0.0001 15 60.69% 45.26% AST 0.513 0.502-0.523 0.082 25 14.52% 88.48% ALT/AST 0.542 0.531-0.553 <0.0001 0.95 61.66% 46.25% Note: Data were reported as area under curve (AUC) (95% confidence interval). FIGURE 1 Receiver operating characteristic curve analysis of GGT, ALT, AST and ALT/AST ratio in diabetes and pre-diabetes (data from Table 2). Roc curve analysis of GTT, ALT, AST and ALT/AST in diabetes vs. control group was as follows: AUC = 0.685; (95% CI: 0.673-0.694; p < .0001; Cut-off value: >21.36; Sensitivity: 70.25%; Specificity: 57.76%), AUC = 0.564; (95% CI: 0.553-0.575; p < .0001; Cut-off value: >14; Sensitivity: 70.85%; Specificity: 39.88%), AUC = 0.588; (95% CI: 0.577-0.600; p < .0001; Cut-off value: <14; Sensitivity: 43.33%; Specificity: 71.64%), AUC = 0.669; (95% CI: 0.658-0.679; p < .0001; Cut-off value: >1.06; Sensitivity: 68.87%; Specificity: 57.29%), respectively. Similar results were also observed in the case of prediabetes vs. control group, including: AUC = 0.573; (95% CI: 0.562-0.583; p < .0001; Cut-off value: >20.33; Sensitivity: 57.10%; Specificity: 53.89%), AUC = 0.537; (95% CI: 0.526-0.548; p < .0001; Cut-off value:>15; Sensitivity: 60.69%; Specificity: 45.26%), AUC = 0.513; (95% CI: 0.526-0.548; p = .082; Cut-off value: >25; Sensitivity: 14.52%; Specificity: 88.48%), AUC = 0.542; (95% CI: 0.531-0.553; p < .0001; Cut-off value: >0.95; Sensitivity: 61.66%; Specificity: 46.25%). 3.3 Logistic regression analysis According to logistic regression analysis, some liver enzymes, lipid profiles and metabolic syndrome were associated with an increased odds of developing prediabetes or diabetes (Table 3). The estimated ORs for metabolic syndrome in the prediabetes and diabetes groups were 7.966 (95% CI: 7.139-8.889; p < .0001) and 12.45 (95% CI: 10.88-14.24; p < .0001), respectively. In the case of AST, however, the odds ratio (0.976) indicated a reduction in diabetes odds (95% CI: 0.968-0.984; p < .0001). On the contrary, the ALT/AST ratio increases the odds of prediabetes and diabetes development by 1.347 (95% CI: 0.968-0.984; p < .0001) and 3.623 (95% CI: 3.159-4.154; p < .0001, respectively). After the adjustment for age, sex and BMI, there was almost no difference with the results obtained from univariate analysis. However, both analyses display significantly positive relationships between ALT/AST ratio and metabolic syndrome with prediabetes and diabetes. TABLE 3 Logistic regression analysis of liver enzymes, lipid profile, and MetS in prediabetes and diabetes vs. control group Prediabetes Diabetes Variable Odds Ratios 95% CI p-Value Odds Ratios 95% CI p-Value model 1 a GGT 1.009 1.006-1.012 <.0001 1.024 1.021-1.027 <.0001 AST 1.010 1.004-1.016 <.0001 0.976 0.968-0.984 <.0001 ALT 1.007 1.004-1.010 <.0001 1.008 1.004-1.012 <.0001 ALT/AST 1.347 1.190-1.525 <.0001 3.623 3.159-4.154 <.0001 LDL 1.003 1.002-1.005 <.0001 1.001 0.999-1.002 <.0001 HDL 0.997 0.993-1.001 <.0001 0.989 0.984-0.993 <.0001 TG 1.002 1.001-1.002 <.0001 1.005 1.004-1.005 <.0001 TC 1.005 1.004-1.006 <.0001 1.006 1.005-1.007 <.0001 MetS 7.966 7.139-8.889 <.0001 12.45 10.88-14.24 <.0001 model 2 b GGT 1.010 1.007-1.013 <.0001 1.024 1.021-1.027 <.0001 AST 1.015 1.009-1.021 <.0001 0.979 0.970-0.988 <.0001 ALT 1.012 1.009-1.016 <.0001 1.014 1.010-1.018 <.0001 ALT/AST 1.595 1.382-1.842 <.0001 5.632 4.776-6.640 <.0001 LDL 1.002 1.000-1.003 .040 0.999 0.997-1.000 .151 HDL 0.996 0.992-1.001 .087 0.988 0.983-0.993 <.0001 TG 1.002 1.002-1.003 <.0001 1.005 1.004-1.006 <.0001 TC 1.004 1.002-1.005 <.0001 1.005 1.003-1.006 <.0001 MetS 6.833 6.100-7.654 <.0001 10.67 9.268-12.28 <.0001 a Model 1: Crude model. b Model 2: Adjusted for age, gender and body mass index (BMI). Abbreviation: CI, confidence interval. 4 DISCUSSION In the current study, we observed a significant increase in all metabolic risk factors and liver enzymes, except for HDL-C and AST, in both prediabetic and T2MD subjects, with the differences being more pronounced in diabetic individuals. In subjects with prediabetes and T2DM, the mean LDL, TG and TC levels were higher. Consistent with these findings, Dhoj et al. 14 demonstrated that diabetes is associated with a high prevalence of dyslipidaemia characterized by elevated levels of cholesterol, TG and LDL. Additionally, Jasim et al. 5 identified TG as one of the promising biomarkers for predicting prediabetes and T2DM. These findings support that diabetes patients are more susceptible to co-occurring diseases such as hyperglycaemia, chronic renal failure, hypothyroidism and polypharmacy, with drugs known to have adverse effects on lipid profiles. Patients with diabetes must therefore be treated to prevent coronary artery disease. 15 Individual metabolic syndrome characteristics (such as higher BMI, waist circumference, DBP and SBP levels, among others) were associated with the prevalence of prediabetes and T2DM, according to the findings of this study. Thus, 80% of subjects with T2DM and 72% in the prediabetes group had MetS, whereas only 24% of the control group exhibited metabolic syndrome symptoms. In addition, Ogedengbe et al. 16 found that the prevalence of MetS among T2DM patients is extremely high. This study revealed that liver enzymes, including ALT and GGT but not AST, and the ALT/AST ratio were significantly elevated in prediabetes and T2MD cases. However, some studies have found no correlation between elevated ALT and diabetes, possibly due to the ethnic diversity of the study populations. 6 Forlani et al. 17 reported a high prevalence of elevated ALT, AST and GGT levels in T2DM, which is consistent with our findings. Although there are no clear biological explanations for the relationships between liver indicators and glucose metabolism, one possible mechanism is that MetS and T2DM increase the risk of liver damage, increasing liver enzyme levels. 9 To reduce the risk of liver damage, prediabetics and diabetic patients may require a comprehensive clinical, laboratory and histological examination. In addition, GGT, ALT and the ALT/AST ratio, but not AST, can be used to identify prediabetes and diabetes based on ROC results. Among prediabetic and diabetic subjects, the GGT level has the highest areas under the curve (AUC) and the highest sensitivity compared to the control group. In contrast, logistic regression analysis revealed that higher levels of ALT, GGT and ALT/AST were independent risk factors for prediabetics and diabetics and that an increase in the ALT/AST ratio increased the risk of T2MD by 3.68-fold, whereas lower AST levels were associated with the risk of diabetes. Sun-Hye et al. 18 observed that higher levels of GGT and ALT and a lower AST/ALT ratio were independent risk factors for diabetes and impaired fasting glucose (IFG). Additionally, Zhao et al. 19 evidenced that the ALT/AST ratio may be a useful indicator of insulin resistance (IR) in the Chinese population. According to several studies, elevated GGT and ALT levels are also beneficial for identifying early markers of dysregulated glucose metabolism, which strongly correlate with prediabetes and diabetes. 20 A second proposed mechanism for the relationship between hepatic indices and glucose metabolism is that elevated serum ALT and GGT levels indicate hepatic steatosis, resulting in hepatic insulin resistance (IR). 18 IR is a risk factor for T2DM. 19 Therefore, it is unknown whether T2DM increases liver enzyme levels or whether elevated liver enzyme levels increase the risk of developing T2DM. Therefore, additional research is required to clarify these theories. In contrast to our findings, some studies have found that elevated GGT levels, but not ALT or AST, can be used to predict the onset of T2DM. 9 Sattar et al. 21 also demonstrated that elevated ALT levels within the 'normal' range predict diabetes independently of elevated AST levels. Although we did not examine the role of gender in transaminase levels in this study, a possible explanation for these contradictory findings may be that transaminase levels are gender-specific, according to the findings of some studies. 22 Consequently, it appears that using the ratio of variables, such as ALT/AST, rather than each variable individually may be more effective in evaluating diabetes patients. 5 CONCLUSION Our results indicated a significant increase in liver enzymes except AST, lipid profile except HDL-C, and MetS status in both prediabetic and T2MD subjects, with the differences being more pronounced in diabetic individuals. On the one hand, these variables or their ratio may be considered predictive risk factors for diabetes, and on the other hand, they may be utilized as diagnostic factors. However, it is unknown whether T2DM increases liver enzyme levels or whether elevated liver enzyme levels increase the incidence of T2DM, and the pathophysiologic pathways underlying this association are unclear. Therefore, additional research is required to clarify these theories and validate their clinical applications. AUTHOR CONTRIBUTIONS N. M. designed and supervised the study. N. D. wrote the paper. S. SP., Z. R. and B. C. analysed data. All authors read and approved the final manuscript. CONFLICT OF INTEREST The authors declare no conflict of interest. ETHICAL APPROVAL This study was approved by the Ethics Committee of Ahvaz Jundishapur University of Medical Sciences (Ethical code: IR. AJUMS. REC.1398.455), and the informed consent was taken from all patients who participated in Hoveyzeh Cohort. ACKNOWLEDGEMENTS This research has been financially supported by The Iranian Ministry of Health and Medical Education (Grant No. 700/534) and Ahvaz Jundishapur University of Medical Sciences (Grant No. HCS-9820). DATA AVAILABILITY STATEMENT Data will be made available on request. |
PMC10000632 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050741 healthcare-11-00741 Article Additional Active Movements Are Not Required for Strength Gains in the Untrained during Short-Term Whole-Body Electromyostimulation Training Stephan Holger Validation Formal analysis Resources Data curation Writing - original draft Writing - review & editing Visualization * Wehmeier Udo Frank Conceptualization Methodology Validation Formal analysis Resources Data curation Writing - review & editing Supervision Project administration Forster Tim Conceptualization Validation Investigation Resources Data curation Writing - review & editing Project administration Tomschi Fabian Formal analysis Writing - review & editing Hilberg Thomas Formal analysis Resources Writing - review & editing Visualization Supervision * Oliveira Rafael Academic Editor Brito Joao Paulo Academic Editor Department of Sports Medicine, University of Wuppertal, Moritzstrasse 14, 42117 Wuppertal, Germany * Correspondence: [email protected] (H.S.); [email protected] (T.H.) 03 3 2023 3 2023 11 5 74126 1 2023 24 2 2023 27 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Recommendations for conventional strength training are well described, and the volume of research on whole-body electromyostimulation training (WB-EMS) is growing. The aim of the present study was to investigate whether active exercise movements during stimulation have a positive effect on strength gains. A total of 30 inactive subjects (28 completed the study) were randomly allocated into two training groups, the upper body group (UBG) and the lower body group (LBG). In the UBG (n = 15; age: 32 (25-36); body mass: 78.3 kg (53.1-114.3 kg)), WB-EMS was accompanied by exercise movements of the upper body and in the LBG (n = 13; age: 26 (20-35); body mass: 67.2 kg (47.4-100.3 kg)) by exercise movements of the lower body. Therefore, UBG served as a control when lower body strength was considered, and LBG served as a control when upper body strength was considered. Trunk exercises were performed under the same conditions in both groups. During the 20-min sessions, 12 repetitions were performed per exercise. In both groups, stimulation was performed with 350 ms wide square pulses at 85 Hz in biphasic mode, and stimulation intensity was 6-8 (scale 1-10). Isometric maximum strength was measured before and after the training (6 weeks set; one session/week) on 6 exercises for the upper body and 4 for the lower body. Isometric maximum strength was significantly higher after the EMS training in both groups in most test positions (UBG p < 0.001-0.031, r = 0.88-0.56; LBG p = 0.001-0.039, r = 0.88-0.57). Only for the left leg extension in the UBG (p = 0.100, r = 0.43) and for the biceps curl in the LBG (p = 0.221, r = 0.34) no changes were observed. Both groups showed similar absolute strength changes after EMS training. Body mass adjusted strength for the left arm pull increased more in the LBG group (p = 0.040, r = 0.39). Based on our results we conclude that concurring exercise movements during a short-term WB-EMS training period have no substantial influence on strength gains. People with health restrictions, beginners with no experience in strength training and people returning to training might be particularly suitable target groups, due to the low training effort. Supposedly, exercise movements become more relevant when initial adaptations to training are exhausted. WB-EMS electric stimulation strength training strength test This research received no external funding. pmc1. Introduction Whole-body electromyostimulation (WB-EMS) is a training method that can complement or to some extent replace traditional resistance training, as it can be used alone, superimposed, or combined (different training time points). Since several electrodes are used , different muscles can be stimulated at the same time . Strength improvements can be achieved with both high-intensity resistance training and WB-EMS . Previous studies have shown that WB-EMS is applicable in healthy people and also in patients, e.g., in people suffering from Parkinson or sarcopenic obesity . In conventional resistance training, the one repetition maximum (1-RM) is used to describe the training intensity . Since it represents the maximal voluntary contraction, a comparison between electromyostimulation (EMS) and normal contraction is possible . A low-cost way to determine the intensity of strength training is to capture the perceived exertion using the Borg scale , which is also used in WB-EMS training . In contrast to the 1-RM, where voluntary force production under external load is recorded, the perceived exertion reflects the internal load. For beginners in conventional strength training, at least 2 training sessions per week are recommended. Both multi-joint and single-joint exercises can be performed, using a variety of equipment and the own body weight. Per set, 8 to 12 repetitions should be completed at 60-70% of the repetition maximum . To provide a safe and effective application of WB-EMS, guidelines recommend restricting the duration of one session to a maximum of 20 min. Moreover, the frequency should be limited to one session a week for at least the first eight weeks or a minimum interval of four days should be maintained thereafter . Perceived exertion should be rated approximately as "hard" to "hard+" (lower during initial training) , corresponding to 5 to 6 on the Borg CR 10 scale . Nevertheless, in some trials, the training frequencies were higher , and sometimes lower with one session a week compared to the aforementioned recommendation after familiarization. The aggregated training stimulus consists of the number of sessions a week and the length of the training period. Usually, eight sessions or more have been conducted in strength related WB-EMS studies with healthy subjects . Early strength improvements due to strength training can be attributed mainly to neural factors. From the third to fifth week on, strength development is mainly caused by hypertrophy . Increases after very few sessions (as seen after three training sessions) are supposedly attributable to lower antagonist activity or motoric improvements of synergists . Elgueta-Cancino and colleagues elicited less inhibitory activity in the cortex, higher corticospinal excitability, and altered motor unit activation as assumed mechanisms of initial strength gain. Muscle growth and strength gain can also be achieved by compact training (eight weeks with three sessions a week) with neuromuscular electrical stimulation . Similar to conventional strength training early strength gains owing to EMS-training are achieved without muscle growth . The body of research on WB-EMS training is growing . EMS can be superimposed on maximum or sub-maximum voluntary dynamic or isometric contractions or applied without any concomitant voluntary contraction. Nevertheless, little is known about the importance of active exercise movements during stimulation. Strength gains due to EMS with exercise movements were previously shown and some authors addressed the impact of EMS superimposed on intense strength training . To our knowledge, only Kemmler and colleagues investigated the effects of smaller, WB-EMS accompanying movements. In this randomized controlled trial (RCT), participants trained once a week for 12 weeks. However, only older females with little muscle mass were included for the comparison between dynamic use (movements during stimulation) and passive use (only isometric contractions during stimulation) limiting the generalizability of the results obtained. Therefore, the present study aims to investigate whether active exercise movements during stimulation have a positive effect on strength gains of selected upper and lower body muscles in young healthy subjects of both sexes in training sessions using mobile, easily accessible fitness equipment, or the own body mass. We hypothesized that WB-EMS combined with concurrent exercise movements will result in higher strength gains than WB-EMS alone. Hence, this study was designed to clarify whether movement sequences are necessary for strength gains during WB-EMS or, whether the electrostimulation alone induces strength gains. The results might help fitness professionals and EMS-users to optimize recommendations for WB-EMS training depending on individual goals and requirements. 2. Materials and Methods 2.1. Subjects The number of subjects to be included in the study was determined using an a priori sample size calculation for statistical comparison of the means of two unpaired groups (using the program GPower 3.1) based on the mean of the effect sizes (D strength leg extension: d = 1.67; D strength leg flexion: d = 0.79) reported by Kemmler and colleagues . This study is similar to the present study. A predefined lower limit of statistical power of 80% and ana error probability of 0.05 were assumed. A dropout rate of 20% was further added. Based on the results of this calculation, a total of 30 subjects were initially recruited for participation. Subjects were included when being aged between 20 and 40 years and having abstained from physical activity for at least six months prior to the start of the study. Access was possible for both sexes. Subjects were excluded when acute injuries or physical complaints were reported or when contraindications as listed by Kemmler and colleagues or Stollberger and Finsterer were present (e.g., epilepsy, bleeding disorders). No other exclusion criteria were defined (e.g., BMI, VO2max). The study was conducted in accordance with the principles of the Declaration of Helsinki and approved by the ethics committee of the University of Wuppertal (MS/BBL 200114 Wehmeier). All subjects signed a written consent to participate in the study. 2.2. Experimental Design The procedure was based on a randomized controlled trial design . Subjects were randomly assigned to two training groups (with the program RandList 1.2), with the number of subjects in both groups being equal. In the upper body group (UBG), WB-EMS was accompanied by exercise movements of the upper body only and in the lower body group (LBG) by exercise movements of the lower body only. Therefore, the UBG served as a control when lower body strength is considered, and the LBG served as a control when upper body strength is considered. With this design, WB-EMS without exercises and WB-EMS with exercises could be compared. Intervention duration was set to six weeks, training frequency to one session/week, and the duration of the training session to 20 min. Before and after the training period, maximum force was determined during various exercises. Blinding of subjects was not possible because the intervention is identifiable. Blinding of the investigator was not applicable because the training instructions and the test instructions were given by the same person, a professional EMS trainer with a bachelor's degree in sports science. Subjects were asked to maintain their dietary habits and to keep their physical activity levels constant, which also meant avoiding additional physical activity. All interventions and measurements were conducted in an EMS studio (go!Orange--Studio fur EMS, Remscheid, Germany). 2.3. WB-EMS Procedure Both the UBG and the LBG received the same WB-EMS application (miha bodytec II; miha bodytec GmbH, Gersthofen, Germany) once a week. Subjects wore thin tight-fitting underwear. The vest with wetted electrodes was placed on the upper body and the wetted electrode bands on the arms, buttocks, and legs (miha bodytec). During the 20-min training, the upper and lower back, abdominal muscles, buttocks, muscles around the thigh, chest, and muscles around the upper arm were stimulated with 85 Hz of 350 ms wide rectangular pulses in biphasic mode. Both the duration of the pulse interval (stimulation on) and the pulse pause (stimulation off) were set to 4 s. The pulses were ramped up to the targeted intensity without delay (full intensity directly available) and similarly ramped down to zero (direct interruption of the stimulation) at the end of the stimulation phase. To maintain the same conditions, the stimulation intensity was adjusted to 6-8 on a scale of 1 (hardly noticeable) to 10 (painful) . Regardless of group affiliation, muscles were voluntarily tensed during the stimulation episode. 2.4. Exercise Procedure Both groups received WB-EMS and performed exercises meanwhile. . The UBG used upper body exercise movements (chest and upper back including shoulders and arms) and the LBG used lower body exercise movements (buttocks and thigh muscles including abductors and adductors). The UBG training consisted of rowing, butterfly reverse, latissimus pulls, pushups, butterfly, biceps curls, and triceps pulldowns. The LBG training consisted of squats, lunges, adductions, abductions, hip lifts, and leg raises. Both groups exercised the trunk (abdomen and lower back) with back extensions, crunches, and oblique crunches. Selected exercises were performed with additional fitness equipment (fitness tubes and elastic bands, each with varying resistance, and a Swiss Ball). During the first 1 to 2 sessions (depending on the training level), subjects maintained the position over the period of stimulation that they had taken at the onset of the stimulus. One set of 12 repetitions was performed per exercise, with each repetition beginning with the onset of the pulse. To maintain the same physical load level, i.e.,16 to 17 on the Borg RPE scale , the number of movements during an impulse interval could be increased up to three. If the training stimulus was not sufficient after the aforementioned customization, the originally targeted static exercise position should be maintained during the interval break. However, overexertion led to a backward correction. Another way to increase the intensity to the desired level was to increase the resistance either by giving an additional fitness tube or rubber band, or by using a version that offered more resistance. 2.5. Isometric Strength Testing Procedure Isometric maximum strength (N) was determined during 10 different exercises (arm adduction, arm pull, leg extension, and leg curl, each unilateral left and unilateral right, as well as during biceps curl and triceps pulldown, each bilateral) in standardized positions pre (initial measurement) and post (final measurement) intervention using a mobile device (KD 9363 including DMS measuring amplifier GVS-2; ME-measuring systems GmbH, Hennigsdorf, Germany), which was more practicable than the determination of the 1-RM. Reliability of the isometric maximum strength measurement method was verified by Runkel and colleagues for several test positions (triceps pulldown, biceps curl, arm pull, sit-up, leg curl, leg extension) in healthy subjects with a comparable body mass index by a high interclass correlation coefficient (r = 0.764 to 0.934). At both time points, the tests were performed three times in each position. The pause was set to 10 seconds between individual tests. In each case, the maximum value was used for analysis. The whole testing procedure lasted approximately 20 min. 2.6. Statistical Analysis Due to the presence of some discordant values (see box plots), skewed distribution in some cases (Shapiro-Wilk test), partial heterogeneity of error variances (Levene's test), and partial heterogeneity of covariances (Box test), nonparametric statistical tests were employed. The differences between the initial and the final maximum isometric strength were determined separately for each group using the Wilcoxon test. The initial and the final values were compared between the groups using the Mann-Whitney U test. Absolute differences were calculated by subtracting the initial values from the final values, and relative differences were calculated by dividing the final values by the initial values (the initial value was set to 100%). Group comparisons were performed using the Mann-Whitney U test for absolute and relative differences. The significance level was set to < 0.05. Two-tailed analyses were used. The results of the non-parametric tests were used to calculate the effect sizes . A distinction was made between large effects (r >= 0.5), medium effects (< 0.5 to 0.3), and small effects (< 0.3 to 0.1) . Statistics were calculated using SPSS (IBM SPSS Statistics for Windows, Version 28.0., IBM Corp., Armonk, NY, USA) and Excel (Microsoft Excel for Windows, 16.0., Microsoft Corp., Redmond, WA, USA). An intention-to-treat analysis was not possible due to dropouts occurring at baseline. 3. Results Of the included subjects, 28 completed the study. The dropouts occurred due to personal reasons. The characteristics of the groups did not differ significantly from each other (Table 1) and the total training volume was similar in both groups. Most subjects (n = 9 in each group) completed five sessions and no adverse effects occurred. The body mass remained unchanged in both the UBG and the LBG (Table 1). Neither the initial nor the final values differed significantly between the two groups. Isometric maximum strength was significantly higher after EMS training in both groups, both in absolute terms (Table 2 UBG; Table 3 LBG) and body mass adjusted (N/kg), except for left leg extension in the UBG and biceps curl in the LBG. The changes in absolute strength were similar in both groups (Table 4). Body mass adjusted strength during left arm pull showed a higher increase in the LBG . In the other test positions, group affiliation made no difference . Furthermore, the LBG achieved a higher percentage strength gain in left arm pull, both absolute (Table 4) and body mass adjusted (UBG median 114.25% vs. LBG median 137.05%; p = 0.020; r = 0.44). 4. Discussion 4.1. Overview Significant strength changes were observed in both groups after about five weeks training (one session per week). The percentage differences between the initial and final tests were higher than those found in the reliability analysis of the test device by Runkel and colleagues . Therefore, the changes could be attributed to training. LBG training improved left arm pull strength more than UBG training. However, there were no group differences in the other exercises. Initial values between the two groups were not significantly different, but possibly at clinically relevant levels. If the higher initial values had been due to differences in training history, a lower ability to further increase strength would have be needed to be considered . However, subjects should have abstained from intense physical activity for at least six months before starting the study. 4.2. Accompanying Voluntary Activity Little is known about the effects of movements for strength gain during EMS. During local application, movements are usually avoided and isometric contractions are performed. Maffiuletti summarized that there are no differences in strength increase between EMS and EMS superimposed on voluntary contractions. However, the conclusion is based on the results of isometric interventions. Although movements are thought to promote the activity of stimulated muscles , our results failed to show a consistent influence of active exercise movements on strength gains. Furthermore, strength gains from conventional resistance training depend, among others, on the range of motion used . However, isometric contractions at multiple joint angles might cover at least in part the physiological range of motion. For EMS training, Maffiuletti recommends changing the joint position and furthermore, changing the electrode positioning to increase recruitment. Admittedly, Kemmler and colleagues demonstrated the benefit of movement during WB-EMS use, with participants exercising in supine position. In contrast, our participants performed exercises in different positions. Therefore, any movements of body parts that were not primarily intended for the exercises and possible differences in resistance to gravity might have influenced the results. Furthermore, it needs to be considered that additional fitness equipment (fitness tubes and elastic bands with different resistance as well as a Swiss Ball) was used for selected exercises. However, exercise movements using additional fitness equipment did not affect the results. In addition, both the UBG and LBG performed exercises for the trunk. Therefore, both groups received partially similar dynamic training stimuli (three exercises). Movements inevitably lead to changes in muscle length and shape (e.g., biceps muscle during curl). Hence, changes in the electrode contact were very likely to occur. Furthermore, training that aims to enhance endurance and strength at the same time, such as EMS superimposed on cycling , requires movements. However, stimulation intensity must be considered to ensure the range of motion . 4.3. Training Models and Adaptations Supraspinal mechanisms appear to be responsible for the initial strength development through EMS training . Bezerra and colleagues showed increased strength after EMS superimposed onto maximum isometric quadriceps contractions, not only of the exercised leg but also of the unexercised leg, confirming neural contribution. The potential to use EMS for rapid strength gains was demonstrated by Deley and colleagues , who reported that maximum dynamic leg extension torque in prepubertal girls could be increased by up to 50.6% with three weekly isometric applications over a three-week period. According to Adams , atrophic patients as well as casualties are target groups for the use of EMS. After 5 to 6 weeks, a 10 to 15% enhancement of muscle function can be achieved, but three sessions a week are recommended. Several studies confirmed the impact of WB-EMS on strength . However, to our knowledge, only Kemmler and colleagues have studied the effects of exercise during WB-EMS to date. In most cases, the lower body was investigated. Von Stengel and Kemmler showed that leg/hip strength can be improved with 1.5 WB-EMS training sessions (with unloaded, low effort exercises) per week over a 14 to 16 week period, regardless of age. Furthermore, strength gains due to unloaded WB-EMS were similar compared to a HIT training after 16 weeks with three sessions in two weeks . An increase in strength was also observed after shorter training periods. For example, WB-EMS superimposed on jumps twice a week over seven weeks significantly improved leg strength in contrast to normal jump training . In the study by Wirtz and colleagues , leg flexors strength increased only after combining stimulation of multiple body parts with loaded squats (100% 10 RM) twice per week and it was higher three weeks after the six-week training compared to the same training without stimulation. Dormann and colleagues showed significant improvements in leg strength after a four-week, eight-session WB-EMS training program that were similar to those seen in the control group, which performed the same training that included strength exercises, without additional stimulation, and in which intensification was accomplished using other training tools. However, not only leg muscles but also upper body muscles could benefit from dynamic WB-EMS. Reljic and colleagues observed improvements throughout the entire body after a 12-week WB-EMS program with slight motions, consisting of two sessions per week. Our results suggest that even fewer training sessions are beneficial than previously described, whether or not exercise movements are performed during stimulation, which appears to be due to neural factors. Therefore, not only locally applied EMS training regimens have the potential to increase strength, but also WB-EMS training regimens without additional exercise movements. 4.4. Transferability Benefits from WB-EMS can also be expected, for example, for patients suffering from sarcopenia, sarcopenic obesity, and low back pain . It might be useful especially for beginners to start WB-EMS training with a five-week training period without additional exercise movements to improve basic strength before starting a more challenging exercise program. WB-EMS without additional exercise movements can be a first access to training when health conditions do not allow conventional exercises or when a lack of compliance exists. Relative to WB-EMS, local application appears to be superior in gaining strength . However, the lack of focus on selected zones owing to stimulation of the entire body is a suggested explanation for the difference . Therefore, only target muscles could be stimulated and not all available electrodes could be used, even if an electrode suit is worn, or zones could be stimulated in an individual order. 4.5. Limitations We have shown that the effect of WB-EMS on strength gains is independent of the concomitant exercise movements. Nevertheless, some limitations need to be acknowledged. A test of core strength would have been useful, as both groups performed core strength exercises under the same conditions and a higher strength can be expected as observed in the study by Berger and colleagues , although they used a more extensive training program. Owing to two dropouts, the group sizes were slightly different, which affected the comparison. Furthermore, the strength gains of the dynamically trained muscles might have been underestimated, since only isometric strength was tested. It must also be mentioned that the increase in strength might have been influenced by deviations from the predefined number of training sessions. To evaluate the intensity of the movement sequences, an unstimulated group could have been used. Furthermore, an inactive group could have been used as a reference for the interventions. However, the study focused on the comparison between the EMS application without and the application with concurring exercise movements. When using WB-EMS training technology, the load parameters must be set with care to avoid unintended side effects, particularly during the first sessions of novices when adaptation to the load has not yet occurred in the form of the "repeated bout effect" . 5. Conclusions WB-EMS training without accompanying movement exercises leads to substantial strength gains even during a short WB-EMS training period. At the beginning of WB-EMS training, electromyostimulation is more important for strength gains than active exercise movements. Therefore, future studies should examine the effects of exercise movements during long-term training periods, or consider individuals already adapted to WB-EMS training or strength training. The transferability of the results to a collective experienced with WB-EMS or strength training should be questioned, as movements (and maybe other approaches, e.g., additional mass or complicating tasks) may become more relevant when initial adaptations to training are exhausted. Since the training effort with WB-EMS is low, people with health restrictions, beginners without experience in strength training, and those returning to training might benefit from these results. These groups could refrain from exercise movements during the first WB-EMS training sessions and integrate them during the course of the subsequent training. Acknowledgments We would like to thank go!ORANGE-Studio fur EMS for providing the WB-EMS equipment and the studio. We acknowledge the support of the Open Access Publication Fund of the University of Wuppertal. Supplementary Materials The following supporting information can be downloaded at: Figure S1: Exercise movements for the upper body performed by the upper body group (UBG); Figure S2: Exercise movements for the lower body performed by the lower body group (LBG); Figure S3: Exercises for the trunk performed by both groups; Figure S4: Isometric strength tests; Table S1: Data overview. Click here for additional data file. Author Contributions Conceptualization, U.F.W. and T.F.; methodology, U.F.W.; validation, H.S., U.F.W. and T.F.; formal analysis, H.S., U.F.W., F.T. and T.H.; investigation, T.F.; resources, H.S., U.F.W., T.F. and T.H.; data curation, H.S., U.F.W. and T.F.; writing--original draft preparation, H.S.; writing--review and editing, H.S., U.F.W., T.F., F.T. and T.H.; visualization, H.S. and T.H.; supervision, U.F.W. and T.H.; project administration, U.F.W. and T.F. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the ethics committee of the University of Wuppertal (MS/BBL 200114 Wehmeier). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement Data are available (Table S1). Conflicts of Interest The authors declare no conflict of interest. Figure 1 Overview of the study design and procedure. UBG: upper body group; LBG: lower body group; WB-EMS: whole-body electromyostimulation (dashed frame); UBM: upper body exercise movements; LBM: lower body exercise movements. Figure 2 Differences between the final and the initial body weight adjusted maximum upper body strength (arm adduction and arm pull) in the upper body group and in the lower body group. D: differences; UBG (n = 15): upper body group; LBG (n = 13): lower body group; white box: right arm; grey box: left arm; circles represent discordant values; means are displayed by crosses and medians by crossbars; * significant difference between LBG and UBG, p < 0.040 (effect size r = 0.39). Figure 3 Differences between the final and the initial body weight adjusted maximum arm strength (biceps curl and triceps pulldown) in the upper body group and in the lower body group. D: differences; UBG (n = 15): upper body group; LBG (n = 13): lower body group; white box: triceps pulldown; grey box: biceps curl; the circle represents a discordant value; means are displayed by crosses and medians by crossbars; no significant differences occurred. Figure 4 Differences between the final and the initial body weight adjusted maximum leg strength (leg extension and leg curl) in the upper body group and in the lower body group. D: difference; UBG (n = 15): upper body group; LBG (n = 13): lower body group; white box: right leg; grey box: left leg; circles represent discordant values; means are displayed by crosses and medians by crossbars; no significant differences occurred. healthcare-11-00741-t001_Table 1 Table 1 Characteristics presented as medians (ranges) of the total collective (n = 28) and the two groups. Total (m 11; f 17) UBG (m 6; f 9) LBG (m 5; f 8) Age [years] 28 (20-36) 32 (25-36) 26 (20-35) Height [cm] 173.0 (159-186) 174.0 (159.0-186.0) 171.0 (160.0-186.0) Body mass pre [kg] 74.1 (47.4-114.3) 78.3 (53.1-114.3) 67.2 (47.4-100.3) Body mass post [kg] 74.4 (48.0-112.9) 78.2 (52.8-112.9) 68.0 (48.0-99.7) BMI pre [kg/m2] 25.33 (18.21-40.98) 25.68 (19.27-40.98) 23.88 (18.21-30.35) BMI post [kg/m2] 25.06 (18.08-38.57) 25.63 (19.16-38.57) 23.74 (18.08-30.65) Number of sessions 5 (3-6) 5 (3-6) 5 (3-6) m: male; f: female; UBG: upper body group; LBG: lower body group; BMI: body mass index. healthcare-11-00741-t002_Table 2 Table 2 Initial and final median maximum strength values (ranges) of the upper body group (UBG). Test Position Strength (N) Initial Strength (N) Final Significance Effect Size r Arm adduction right 83.3 (44.8-143.0) 116.2 (42.6-178.3) 0.002 ** 0.81 Arm adduction left 83.0 (44.8-124.4) 124.4 (54.2-196.0) <0.001 ** 0.88 Arm pull right 173.0 (117.0-293.7) 232.6 (124.7-331.6) 0.001 ** 0.84 Arm pull left 196.0 (114.1-331.7) 242.2 (132.4-378.3) 0.006 ** 0.70 Triceps pulldown 253.0 (142.7-461.2) 279.2 (149.6-510.4) 0.012 * 0.65 Biceps curl 308.5 (117.6-512.7) 331.7 (143.6-528.2) 0.008 ** 0.69 Leg extension right 377.2 (196.7-697.4) 404.8 (237.4-766.0) 0.015 * 0.63 Leg extension left 373.4 (130.9-682.5) 403.3 (218.6-769.0) 0.100 0.43 Leg curl right 184.6 (40.9-296.1) 200.1 (71.2-447.5) 0.005 ** 0.72 Leg curl left 185.5 (48.0-296.5) 186.9 (85.5-396.0) 0.031 * 0.56 n = 15; * significant difference p < 0.05; ** highly significant difference p < 0.01. healthcare-11-00741-t003_Table 3 Table 3 Initial and final median maximum strength values (ranges) of the lower body group (LBG). Test Position Strength (N) Initial Strength (N) Final Significance Effect Size r Arm adduction right 61.7 (28.8-127.4) 84.9 (43.1-155.5) 0.007 ** 0.75 Arm adduction left 56.2 (29.0-111.3) 85.1 (46.1-170.0) 0.002 ** 0.84 Arm pull right 140.0 (80.9-281.7) 170.4 (145.0-362.2) 0.001 ** 0.88 Arm pull left 131.0 (94.9-216.8) 167.9 (136.8-378.2) 0.001 ** 0.88 Triceps pulldown 178.0 (125.0-474.0) 203.0 (140.2-474.2) 0.039 * 0.57 Biceps curl 215.3 (137.0-559.5) 212.7 (167.9-531.0) 0.221 0.34 Leg extension right 330.5 (218.2-725.0) 385.6 (281.0-787.8) 0.002 ** 0.86 Leg extension left 304.0 (184.8-612.0) 348.5 (228.1-704.9) 0.001 ** 0.88 Leg curl right 139.4 (104.6-268.7) 170.0 (140.0-311.9) 0.001 ** 0.88 Leg curl left 127.0 (105.0-261.4) 166.2 (132.0-287.9) 0.002 ** 0.86 n = 13; * significant difference p < 0.05; ** highly significant difference p < 0.01. healthcare-11-00741-t004_Table 4 Table 4 Median differences (ranges) between final and initial maximum strength values in both groups (UBG and LBG). Test Position D UBG (N) D UBG (%) D LBG (N) D LBG (%) Arm adduction right 26.5 (-11.9-68.3) 137.01 (83.43-175.63) 12.6 (-13.9-44.3) 120.68 (86.47-222.92) Arm adduction left 27.5 (5.6-83.3) 131.70 (104.61-183.05) 29.1 (-15.2-58.7) 152.74 (84.85-244.67) Arm pull right 29.4 (-6.3-85.4) 114.74 (96.14-149.36) 32.5 (10.4-96.1) 128.46 (105.28-182.82) Arm pull left 21.6 (-27.2-90.2) 115.07 (84.68-144.24) 41.4 (4.6-164.0) 131.67 (102.47-176.56) * Triceps pulldown 11.7 (-9.0-89.5) 105.88 (97.20-161.89) 19.6 (-22.5-74.5) 110.17 (93.72-130.56) Biceps curl 24.5 (-31.3-69.4) 107.45 (88.79-139.86) 21.2 (-45.1-40.7) 108.82 (88.02-122.55) Leg extension right 52.0 (-90.7-141.2) 117.40 (86.99-125.25) 55.1 (-4.4-130.3) 116.67 (98.94-135.90) Leg extension left 41.8 (-94.5-120.3) 110.05 (80.11-167.00) 56.0 (4.7-215.7) 121.68 (101.14-144.09) Leg curl right 30.3 (-27.8-151.4) 117.45 (79.41-185.79) 37.1 (15.5-126.3) 124.87 (111.12-168.34) Leg curl left 13.9 (-27.3-99.5) 107.42 (88.28-178.13) 39.7 (-1.7-70.7) 131.41 (98.79-160.84) D: differences; UBG (n = 15): upper body group; LBG (n = 13): lower body group; * significant difference compared to UBG, p = 0.020 (effect size r = 0.44). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Berger J. Ludwig O. Becker S. Backfisch M. Kemmler W. Frohlich M. Effects of an Impulse Frequency Dependent 10-Week Whole-body Electromyostimulation Training Program on Specific Sport Performance Parameters J. Sports Sci. Med. 2020 19 271 281 32390720 2. Jee Y.S. The efficacy and safety of whole-body electromyostimulation in applying to human body: Based from graded exercise test J. Exerc. Rehabil. 2018 14 49 57 10.12965/jer.1836022.011 29511652 3. Kemmler W. Teschler M. Weissenfels A. Froehlich M. Kohl M. von Stengel S. Whole-Body Electromyostimulation Versus High Intensity (Resistance Exercise) Training--Impact on Body Composition and Strength Dtsch. Z. Sportmed. 2015 66 321 327 10.5960/dzsm.2015.209 4. Pano-Rodriguez A. Beltran-Garrido J.V. Hernandez-Gonzalez V. Reverter-Masia J. Effects of Whole-Body Electromyostimulation on Physical Fitness in Postmenopausal Women: A Randomized Controlled Trial Sensors 2020 20 1482 10.3390/s20051482 32182674 5. Fiorilli G. Quinzi F. Buonsenso A. Casazza G. Manni L. Parisi A. Di Costanzo A. Calcagno G. Soligo M. di Cagno A. A Single Session of Whole-Body Electromyostimulation Increases Muscle Strength, Endurance and proNGF in Early Parkinson Patients Int. J. Environ. Res. Public Health 2021 18 5499 10.3390/ijerph18105499 34065571 6. Kemmler W. Teschler M. Weissenfels A. Bebenek M. von Stengel S. Kohl M. Freiberger E. Goisser S. Jakob F. Sieber C. Whole-body electromyostimulation to fight sarcopenic obesity in community-dwelling older women at risk. Resultsof the randomized controlled FORMOsA-sarcopenic obesity study Osteoporos. Int. 2016 27 3261 3270 10.1007/s00198-016-3662-z 27289534 7. Kraemer W.J. Ratamess N.A. Fundamentals of resistance training: Progression and exercise prescription Med. Sci. Sports Exerc. 2004 36 674 688 10.1249/01.MSS.0000121945.36635.61 15064596 8. Filipovic A. Kleinoder H. Dormann U. Mester J. Electromyostimulation--a systematic review of the influence of training regimens and stimulation parameters on effectiveness in electromyostimulation training of selected strength parameters J. Strength Cond. Res. 2011 25 3218 3238 10.1519/JSC.0b013e318212e3ce 21993042 9. Tiggemann C.L. Korzenowski A.L. Brentano M.A. Tartaruga M.P. Alberton C.L. Kruel L.F. Perceived exertion in different strength exercise loads in sedentary, active, and trained adults J. Strength Cond. Res. 2010 24 2032 2041 10.1519/JSC.0b013e3181d32e29 20634752 10. Filipovic A. DeMarees M. Grau M. Hollinger A. Seeger B. Schiffer T. Bloch W. Gehlert S. Superimposed Whole-Body Electrostimulation Augments Strength Adaptations and Type II Myofiber Growth in Soccer Players during a Competitive Season Front. Physiol. 2019 10 1187 10.3389/fphys.2019.01187 31607944 11. Filipovic A. Bizjak D. Tomschi F. Bloch W. Grau M. Influence of Whole-Body Electrostimulation on the Deformability of Density-Separated Red Blood Cells in Soccer Players Front. Physiol. 2019 10 548 10.3389/fphys.2019.00548 31156450 12. American College of Sports Medicine ACSM's Guidelines for Exercise Testing and Prescription 11th ed. Wolters Kluwer Philadelphia, PA, USA 2021 13. Kemmler W. Froehlich M. von Stengel S. Kleinoder H. Whole-Body Electromyostimulation--The Need for Common Sense! Rationale and Guideline for a Safe and Effective Training Dtsch. Z. Sportmed. 2016 67 218 221 10.5960/dzsm.2016.246 14. Kemmler W. Weissenfels A. Willert S. Shojaa M. von Stengel S. Filipovic A. Kleinoder H. Berger J. Frohlich M. Efficacy and Safety of Low Frequency Whole-Body Electromyostimulation (WB-EMS) to Improve Health-Related Outcomes in Non-athletic Adults. A Systematic Review Front. Physiol. 2018 9 573 10.3389/fphys.2018.00573 29875684 15. Micke F. Kleinoder H. Dormann U. Wirtz N. Donath L. Effects of an Eight-Week Superimposed Submaximal Dynamic Whole-Body Electromyostimulation Training on Strength and Power Parameters of the Leg Muscles: A Randomized Controlled Intervention Study Front. Physiol. 2018 9 1719 10.3389/fphys.2018.01719 30568596 16. Schuhbeck E. Birkenmaier C. Schulte-Gocking H. Pronnet A. Jansson V. Wegener B. The Influence of WB-EMS-Training on the Performance of Ice Hockey Players of Different Competitive Status Front. Physiol. 2019 10 1136 10.3389/fphys.2019.01136 31551812 17. Ludwig O. Berger J. Schuh T. Backfisch M. Becker S. Frohlich M. Can A Superimposed Whole-Body Electromyostimulation Intervention Enhance the Effects of a 10-Week Athletic Strength Training in Youth Elite Soccer Players? J. Sports Sci. Med. 2020 19 535 546 32874107 18. Dormann U. Wirtz N. Micke F. Morat M. Kleinoder H. Donath L. The Effects of Superimposed Whole-Body Electromyostimulation during Short-Term Strength Training on Physical Fitness in Physically Active Females: A Randomized Controlled Trial Front. Physiol. 2019 10 728 10.3389/fphys.2019.00728 31316389 19. Moritani T. deVries H.A. Neural factors versus hypertrophy in the time course of muscle strength gain Am. J. Phys. Med. 1979 58 115 130 453338 20. Coburn J.W. Housh T.J. Malek M.H. Weir J.P. Cramer J.T. Beck T.W. Johnson G.O. Neuromuscular responses to three days of velocity-specific isokinetic training J. Strength Cond. Res. 2006 20 892 898 10.1519/R-18745.1 17194247 21. Elgueta-Cancino E. Evans E. Martinez-Valdes E. Falla D. The Effect of Resistance Training on Motor Unit Firing Properties: A Systematic Review and Meta-Analysis Front. Physiol. 2022 13 817631 10.3389/fphys.2022.817631 35295567 22. Natsume T. Ozaki H. Kakigi R. Kobayashi H. Naito H. Effects of training intensity in electromyostimulation on human skeletal muscle Eur. J. Appl. Physiol. 2018 118 1339 1347 10.1007/s00421-018-3866-3 29679248 23. Hortobagyi T. Maffiuletti N.A. Neural adaptations to electrical stimulation strength training Eur. J. Appl. Physiol. 2011 111 2439 2449 10.1007/s00421-011-2012-2 21643920 24. Rodrigues-Santana L. Adsuar J.C. Denche-Zamorano A. Vega-Munoz A. Salazar-Sepulveda G. Contreras-Barraza N. Galan-Arroyo C. Louro H. Bibliometric Analysis of Studies on Whole Body Electromyostimulation Biology 2022 11 1205 10.3390/biology11081205 36009832 25. von Stengel S. Kemmler W. Trainability of leg strength by whole-body electromyostimulation during adult lifespan: A study with male cohorts Clin. Interv. Aging 2018 13 2495 2502 10.2147/CIA.S185018 30573954 26. Reljic D. Konturek P.C. Herrmann H.J. Neurath M.F. Zopf Y. Effects of whole-body electromyostimulation exercise and caloric restriction on cardiometabolic risk profile and muscle strength in obese women with the metabolic syndrome: A pilot study J. Physiol. Pharmacol. 2020 71 89 98 10.26402/jpp.2020.1.08 27. Evangelista A.L. Teixeira C.V.S. Barros B.M. de Azevedo J.B. Paunksnis M.R.R. Souza C.R. Wadhi T. Rica R.L. Braz T.V. Bocalini D.S. Does whole-body electrical muscle stimulation combined with strength training promote morphofunctional alterations? Clinics 2019 74 e1334 10.6061/clinics/2019/e1334 31721936 28. Wirtz N. Zinner C. Doermann U. Kleinoeder H. Mester J. Effects of Loaded Squat Exercise with and without Application of Superimposed EMS on Physical Performance J. Sports Sci. Med. 2016 15 26 33 26957923 29. Kemmler W. Teschler M. von Stengel S. Role of whole body-electromyostimulation--"A series of studies" Osteologie 2015 24 20 29 10.13140/RG.2.1.4135.6008 30. Kemmler W. Weissenfels A. Willert S. Frohlich M. Ludwig O. Berger J. Zart S. Becker S. Backfisch M. Kleinoder H. Recommended Contraindications for the Use of Non-Medical WB-Electromyostimulation Dtsch. Z. Sportmed. 2019 70 278 282 10.5960/dzsm.2019.401 31. Stollberger C. Finsterer J. Side effects of and contraindications for whole-body electro-myo-stimulation: A viewpoint BMJ Open Sport Exerc. Med. 2019 5 e000619 10.1136/bmjsem-2019-000619 32. Association W.M. World Medical Association Declaration of Helsinki: Ethical principles for medical research involving human subjects JAMA 2013 310 2191 2194 10.1001/jama.2013.281053 24141714 33. Morishita S. Tsubaki A. Nakamura M. Nashimoto S. Fu J.B. Onishi H. Rating of perceived exertion on resistance training in elderly subjects Expert. Rev. Cardiovasc. Ther. 2019 17 135 142 10.1080/14779072.2019.1561278 30569775 34. Runkel B. Kappelhoff M. Hilberg T. Complex strength performance in patients with haemophilia A. Method development and testing Hamostaseologie 2015 35 (Suppl. S1) S12 S17 10.1055/s-0037-1619824 26540124 35. Tomczak M. Tomczak E. The need to report effect size estimates revisited. An overview of some recommended measures of effect size Trends Sport Sci. 2014 1 19 25 36. Cohen J. Statistical Power Analysis for the Behavioral Sciences Erlbaum Hillsdals, NJ, USA 1988 Volume 2 37. Hughes D.C. Ellefsen S. Baar K. Adaptations to Endurance and Strength Training Cold Spring Harb. Perspect. Med. 2018 8 a029769 10.1101/cshperspect.a029769 28490537 38. Maffiuletti N.A. Physiological and methodological considerations for the use of neuromuscular electrical stimulation Eur. J. Appl. Physiol. 2010 110 223 234 10.1007/s00421-010-1502-y 20473619 39. Martinez-Cava A. Hernandez-Belmonte A. Courel-Ibanez J. Moran-Navarro R. Gonzalez-Badillo J.J. Pallares J.G. Bench Press at Full Range of Motion Produces Greater Neuromuscular Adaptations Than Partial Executions after Prolonged Resistance Training J. Strength Cond. Res. 2022 36 10 15 10.1519/JSC.0000000000003391 31567719 40. Pallares J.G. Cava A.M. Courel-Ibanez J. Gonzalez-Badillo J.J. Moran-Navarro R. Full squat produces greater neuromuscular and functional adaptations and lower pain than partial squats after prolonged resistance training Eur. J. Sport Sci. 2020 20 115 124 10.1080/17461391.2019.1612952 31092132 41. Omoto M. Matsuse H. Hashida R. Takano Y. Yamada S. Ohshima H. Tagawa Y. Shiba N. Cycling Exercise with Electrical Stimulation of Antagonist Muscles Increases Plasma Growth Hormone and IL-6 Tohoku J. Exp. Med. 2015 237 209 217 10.1620/tjem.237.209 26522057 42. Masayuki O. Matsuse H. Takano Y. Yamada S. Ohshima H. Tagawa Y. Shiba N. Oxygen Uptake during Aerobic Cycling Exercise Simultaneously Combined with Neuromuscular Electrical Stimulation of Antagonists J. Nov. Physiother. 2013 3 185 10.4172/2165-7025.1000185 43. Paillard T. Training Based on Electrical Stimulation Superimposed Onto Voluntary Contraction Would be Relevant Only as Part of Submaximal Contractions in Healthy Subjects Front. Physiol 2018 9 1428 10.3389/fphys.2018.01428 30369886 44. Bezerra P. Zhou S. Crowley Z. Brooks L. Hooper A. Effects of unilateral electromyostimulation superimposed on voluntary training on strength and cross-sectional area Muscle Nerve 2009 40 430 437 10.1002/mus.21329 19705423 45. Deley G. Cometti C. Fatnassi A. Paizis C. Babault N. Effects of combined electromyostimulation and gymnastics training in prepubertal girls J. Strength Cond. Res. 2011 25 520 526 10.1519/JSC.0b013e3181bac451 20224450 46. Adams V. Electromyostimulation to fight atrophy and to build muscle: Facts and numbers J. Cachexia Sarcopenia Muscle 2018 9 631 634 10.1002/jcsm.12332 30028092 47. Filipovic A. Kleinoder H. Pluck D. Hollmann W. Bloch W. Grau M. Influence of Whole-Body Electrostimulation on Human Red Blood Cell Deformability J. Strength Cond. Res. 2015 29 2570 2578 10.1519/JSC.0000000000000916 26308832 48. Filipovic A. Grau M. Kleinoder H. Zimmer P. Hollmann W. Bloch W. Effects of a Whole-Body Electrostimulation Program on Strength, Sprinting, Jumping, and Kicking Capacity in Elite Soccer Players J. Sports Sci. Med. 2016 15 639 648 27928210 49. Filipovic A. Kleinoder H. Dormann U. Mester J. Electromyostimulation--a systematic review of the effects of different electromyostimulation methods on selected strength parameters in trained and elite athletes J. Strength Cond. Res. 2012 26 2600 2614 10.1519/JSC.0b013e31823f2cd1 22067247 50. Kemmler W. Teschler M. Bebenek M. von Stengel S. [(Very) high Creatinkinase concentration after exertional whole-body electromyostimulation application: Health risks and longitudinal adaptations] Wien. Med. Wochenschr. 2015 165 427 435 10.1007/s10354-015-0394-1 26498468 |
PMC10000633 | Introduction There are limited data on the real-world management of diabetes in the Indian population. In this 2-year analysis of the LANDMARC study, the management of type 2 diabetes mellitus (T2DM) and related complications were assessed. Method This multicenter, observational, prospective study included adults aged >=25 to <=60 years diagnosed with T2DM (duration >=2 years at enrollment) and controlled/uncontrolled on >=2 anti-diabetic agents. This interim analysis at 2 years reports the status of glycaemic control, diabetic complications, cardiovascular (CV) risks and therapy, pan-India including metropolitan and non-metropolitan cities. Results Of the 6234 evaluable patients, 5318 patients completed 2 years in the study. Microvascular complications were observed in 17.6% of patients (1096/6234); macrovascular complications were observed in 3.1% of patients (195/6234). Higher number of microvascular complications were noted in patients from non-metropolitan than in metropolitan cities (p < .0001). In 2 years, an improvement of 0.6% from baseline (8.1%) in mean glycated haemoglobin (HbA1c) was noted; 20.8% of patients met optimum glycaemic control (HbA1c < 7%). Hypertension (2679/3438, 77.9%) and dyslipidaemia (1776/3438, 51.7%) were the predominant CV risk factors in 2 years. The number of patients taking oral anti-diabetic drugs in combination with insulin increased in 2 years (baseline: 1498/6234 [24.0%] vs. 2 years: 1917/5763 [33.3%]). While biguanides and sulfonylureas were the most commonly prescribed, there was an evident increase in the use of dipeptidyl peptidase-IV inhibitors (baseline: 3049/6234, 48.9% vs. 2 years: 3526/5763, 61.2%). Conclusion This longitudinal study represents the control of T2DM, its management and development of complications in Indian population. Clinical Trial Registration Number CTRI/2017/05/008452. This 2-year interim analysis of the 3-year long LANDMARC study highlights the T2DM burden, management practices and related complications, across metropolitan and non-metropolitan cities of India. It indicates that the burden of uncontrolled T2DM in India is high with 20.8% of participants achieving glycaemic control in 2 years (HbA1c < 7%; 53 mmol/mol); and 17.6% having microvascular and 3.1% having macrovascular complications. There is need for effective diabetes management to meet glycaemic targets and prevent CV risk and vascular complications in a developing country like India with high prevalence of T2DM. cardiovascular diseases diabetes complications diabetic nephropathies glycaemic control Sanofi 10.13039/100004339 source-schema-version-number2.0 cover-dateMarch 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.2.6 mode:remove_FC converted:10.03.2023 Das AK , Kalra S , Joshi S , et al. Two-year trends from the LANDMARC study: A 3-year, pan-India, prospective, longitudinal study on the management and real-world outcome in patients with type 2 diabetes mellitus. Endocrinol Diab Metab. 2023;6 :e404. doi:10.1002/edm2.404 pmc1 INTRODUCTION India has been severely affected by the global diabetes epidemic. As per the 10th edition of International Diabetes Federation's (IDF) diabetes atlas (2021), India has 74.2 million people living with diabetes currently, with an age-adjusted prevalence of 9.6% among adults. India is expected to have 124.9 million people in the age range of 20-79 years living with diabetes by 2045. 1 The American Diabetes Association (ADA) recommends a combination of modified lifestyle and pharmacological treatment to achieve good metabolic control in diabetes and long-term maintenance. 2 , 3 Earlier studies have demonstrated that baseline glycated haemoglobin (HbA1c) and body mass index (BMI) can serve as important biomarkers to understand the disease aetiology and to identify suitable treatment options. 4 , 5 Long-term uncontrolled diabetes can cause cardiovascular (CV) diseases and damage kidneys, nerves and other vital organs. If optimum diabetes control is achieved, these serious complications can be delayed or prevented altogether. 1 This has also been substantiated by a study demonstrating that the occurrence of variations in glycaemic levels was associated with microvascular and macrovascular complications. 6 The INSPIRED study, which was conducted in India, included 19,084 individuals (aged 10-97 years) with type 2 diabetes mellitus (T2DM) and varying phenotypic characteristics. The findings of this study emphasized the association between increased hazards of retinopathy and nephropathy with a rise in blood glucose levels. 7 The Million death study also conducted in India showed that diabetes was associated with a significantly increased odds of stroke mortality (odds ratio, 95% confidence intervals [CI]: 1.6, 1.4-1.7, p < .0001). 8 The American College of Cardiology (ACC)/American Heart Association (AHA)/Heart Failure Society of America (HFSA) guideline for the management of heart failure states that diabetes and heart failure often occur concomitantly, and each disease independently increases the risk of the other. 9 Additionally, a review of clinical evidence-based research on the association between T2DM and myocardial infarction (MI) showed that not only T2DM is strongly associated with MI, but it also increases the risk of developing MI and related complications. 10 The review further discusses that in people with T2DM, MI is the primary cause of death and that T2DM leads to an increase in the risk of coronary events in individuals both with or without previous history of coronary events. 10 Hence, tight glycaemic control is essential in the early stages of diabetes. The LANDMARC study (LongitudinAl Nationwide stuDy on Management And Real-world outComes of diabetes in India) was a 3-year comprehensive, robust, longitudinal and prospective study. It aimed to collect data on glycaemic therapy and diabetes complications in people with diabetes living in different regions of India (including metropolitan and non-metropolitan cities). The study protocol, 11 baseline data 12 and 1-year results 13 have been published earlier. The 1-year results of the LANDMARC study indicate the progression of vascular complications and accumulation of CV risk among Indian patients with T2DM. Hypertension and dyslipidaemia, the most common CV risk factors reported, were pronounced in those who were overweight/had obesity. About one-fifth of the patients had optimal glycaemic control (HbA1c < 7%). Patients from both non-metropolitan and metropolitan cities were comparable in terms of improvement in glycaemic status and having optimum control. 13 The aim of this 2-year interim analysis was to further evaluate and understand the diabetic complications and T2DM management pattern in adult patients with T2DM across India (including a sub-analysis of metropolitan and non-metropolitan cities). 2 MATERIALS AND METHODS 2.1 Study design This was a multicenter, observational, prospective study conducted over 3 years (conducted between March 2017 and July 2021). The study was divided into seven visits with an interval of 6 months each. The present manuscript includes results from the second year (within a window period of +-90 days) of the 3-year evaluation period. 2.2 Study patients Adults aged between >=25 years and <=60 years with T2DM for >=2 years and who were controlled/uncontrolled on >=2 anti-diabetic agents at the time of enrollment were included in the study. The details of the study design, methodology, inclusion/exclusion criteria and statistical analysis have been published previously. 11 , 12 , 13 2.3 Study assessments At the end of the second year (visit 5), data related to glycaemic control status (fasting plasma glucose [FPG], post-prandial glucose [PPG] and HbA1c) were collected. The proportion of patients with macrovascular complications (non-fatal MI, non-fatal stroke, CV death and peripheral vascular disease [PVD]), microvascular complications (retinopathy, nephropathy and neuropathy) and CV risk factors (hypertension, dyslipidaemia and albuminuria) was assessed. The glycaemic parameters and complications in patients from metropolitan and non-metropolitan cities were assessed. The proportion of patients taking oral anti-diabetic drugs (OADs) and injectable glucose-lowering drugs was also assessed. Data related to anthropometry (weight) and frequency and severity of hypoglycaemia episodes were collected. 2.4 Data collection The data related to the study end-points were collected prospectively every 6 months up to the end of the study at 36 months. The 450 sites that were selected for this study represent the four geographical regions (East, West, North and South), urban/rural practice across India. The study design was planned to mirror real-life management of patients with T2DM; hence, none of the assessments were mandated. The available data were recorded in electronic-Case Report Forms (e-CRFs). Data quality control was performed by qualified designated personnel. An adverse drug reaction related to any Sanofi product (clinical signs, laboratory values or other) was reported and followed up until the clinical recovery was complete and laboratory results (if clinically significant) had returned to normal or until progression had been stabilized. This was a planned interim analysis to assess the changes in the disease characteristics from baseline and may require modification in the assessment parameters for subsequent interim and final analyses. 2.5 Statistics A minimum sample size of 4387 was decided assuming that the percentage of patients with composite incidence of non-fatal MI, stroke and CV death after 3 years would have been 3%. The study planned to evaluate 6300 patients to estimate the composite incidence percentage with a precision of at least 1%, considering a 30% rate of patients dropping out from the study before the end of the 3 years. 2.6 Ethics The protocol complies with the Declaration of Helsinki and this study was conducted in accordance with the principles laid by the 18th World Medical Assembly (Helsinki, 1964) and all subsequent amendments. The study was also in accordance with the guidelines for Good Epidemiology Practice (US & European) 14 , 15 and aligned to the local regulations, ethics committee(s) (institutional review board/independent ethics committee) and competent authorities. The study was approved by the ethics committees of all participating sites (or a central ethics committee, where applicable). All the patients provided written informed consent before data collection/documentation. 3 RESULTS 3.1 Demographics and baseline characteristics Among the 450 sites, data from 382 sites were analysed for the 2-year visit. Of the 6279 patients recruited, 6234 patients were evaluated; of these, 5318 patients completed 2 years in the study . At baseline, the mean +- standard deviation (SD) age of the patients was 52.1 +- 9.2 years with 57.0% (3552/6234) of the study population in the age range of 50 to 65 years; more than half of the patients (56.6%, 3526/6234) were men. The mean +- SD baseline BMI was 27.2 +- 4.6 kg/m2, and the majority of the patients were obese (66.8%, 4149/6215) (Table 1). Most of the patients (74.4%, 4640/6234) were taking only OADs. (Table S1). FIGURE 1 Patient disposition. *Reasons for CV death were sudden death (n = 19), myocardial infarction (n = 9), stroke (n = 1) and coronary artery procedure (n = 1). CV, cardiovascular; n, number of patients. TABLE 1 Demographics and baseline characteristics. Parameters Baseline values Age (years), n 6234 Mean +- SD (years) 52.1 +- 9.2 <=30 years 61 (1.0) 31-49 years 2192 (35.2) 50-65 years 3552 (57.0) >=66 years 429 (6.9) Gender, n 6234 Men 3526 (56.6) Women 2708 (43.4) Body mass index, n 6215 Mean +- SD (kg/m2) 27.2 +- 4.6 Underweight (<18.0) 44 (0.7) Normal (18.0-22.9) 903 (14.5) Overweight (23.0-24.9) 1119 (18.0) Obese (>=25.0) 4149 (66.8) Treatment at baseline, n 6234 Insulin 1549 (24.8) Insulin-naive 4685 (75.2) HbA1c (%), n 4477 Mean +- SD (%) 8.1 +- 1.6 <6.5% 528 (11.8) 6.5%-6.9% 593 (13.2) <7% 1121 (25.0) 7%-7.9% 1420 (31.7) 8%-8.9% 922 (20.6) >=9% 1014 (22.6) Fasting plasma glucose, n 5013 Mean +- SD (mg/dl) 142.8 +- 50.4 Postprandial glucose, n 4908 Mean +- SD (mg/dl) 205.7 +- 72.3 Duration of T2DM (years) Mean +- SD 8.6 +- 5.6 Median (range) 7.1 (2.0, 40.7) Duration of T2DM by treatment (years), mean +- SD Insulin, (n = 1549) 11.3 +- 6.6 Insulin-naive, (n = 4685) 7.7 +- 5.0 Note: Values are presented as n (%) unless specified otherwise. HbA1c was not measured for all patients and, hence, the percentage may not add up to 100%. Abbreviations: HbA1c, glycated haemoglobin; N, number of patients analysed; n, number of patients with non-missing results at the visit; SD, standard deviation; T2DM, type 2 diabetes mellitus. 3.2 Glycaemic status In 2 years, all the glycaemic parameters improved (decreased) significantly from baseline (mean change +- SD: HbA1c: -0.6 +- 1.7%; FPG: -14.6 +- 54.5 mg/dl; and PPG: -22.0 +- 79.0 mg/dl; p < .0001) . When patients were stratified by HbA1c levels, a significant (p < .0001) reduction in the number of patients in the HbA1c 8%-8.9% and HbA1c >= 9% subgroups were noted at the 2-year visit compared with baseline; while a significant (p < .0001) increase in patient numbers in the HbA1c 7%-7.9% subgroup was noted. Overall, 20.8% (1297/6234) of the patients met optimum glycaemic control (HbA1c < 7%) in 2 years . FIGURE 2 Change in glycaemic parameters at the end of 2 years. Values are presented as mean +- standard deviation. For change from baseline, HbA1c: n = 3020, FPG: n = 3668, PPG: n = 3454. FPG, fasting plasma glucose; HbA1c, glycated haemoglobin; n, number of patients analysed; PPG, postprandial glucose. FIGURE 3 Proportion of patients across HbA1c categories (N = 6234). Data presented as n (%) from baseline (N = 6234). HbA1c was not measured for all patients and hence the percentage may not add up to 100%.p-values are reported using McNemar's test with the null hypothesis that the proportion of paired samples is equal. Patients who met each criterion and those who did not meet the criteria are considered as binary outcomes for the test. The p-values reported are not adjusted for inflation in type I error. *p < .0001. HbA1c, glycated haemoglobin; N, number of patients analysed; n, number of patients with non-missing results at the visit. 3.3 Microvascular and macrovascular complications in 2 years Microvascular complications were noted in 17.6% of patients (1096/6234) (Table 2); while new macrovascular complications were noted in 3.1% of patients (195/6234) (Table 3). The most frequently noted microvascular complication was neuropathy, which was reported in 0.6% of patients (32/6234), followed by nephropathy in 0.3% of patients (19/6234) and retinopathy in 0.1% of patients (6/6234) (Table 2). Overall, 57 new events of microvascular complications were reported in 55 patients at the 2-year visit. Nephropathy, neuropathy and retinopathy were significantly (p < .0001) higher in patients with CV risk factors, while retinopathy was found to be significantly (p = .0351) higher in the HbA1c >= 7% subgroup. (Table S2). TABLE 2 Proportion of patients with microvascular complications at baseline and 2 years (N = 6234). Microvascular complications Baseline n (%) 2 years n (%) Patients at risk with new complications until 2 years, n Patients with new complications at 2 years n (%) Total microvascular complications 902 (14.5) 1096 (17.6) 6234 55 (0.9) Neuropathy 737 (81.7) 898 (81.9) 5368 32 (0.6) Nephropathy 154 (17.1) 221 (20.2) 6032 19 (0.3) Retinopathy 141 (15.6) 165 (15.1) 6075 6 (0.1) Note: This is an interim analysis and possible modifications on variables, and data could be performed for the subsequent interim analyses and final analysis. Abbreviations: N, number of patients analysed; n, number of patients with non-missing results at the visit. TABLE 3 Proportion of patients with macrovascular complications at baseline and 2 years (N = 6234). Macrovascular complications Baseline n (%) 2 years n (%) Patients at risk with new complications until 2 years, n Patients with new complications at 2 years n (%) Total macrovascular complications 145 (2.3) 195 (3.1) 6234 15 (0.2) Myocardial infarction 74 (51.0) 80 (41.0) 6234 1 (0.0) Peripheral vascular disease 45 (31.0) 61 (31.3) 6234 4 (0.1) Stroke 30 (20.7) 33 (16.9) 6234 1 (0.0) Cardiovascular death 0 30 (15.4) 6213 9 (0.1) Note: Values are presented as n (%) unless specified otherwise. This is an interim analysis and possible modifications on variables, and data could be performed for the subsequent interim analyses and final analysis. At 2 years, the newly documented macrovascular complications are reported. Abbreviations: N, number of patients analysed; n, number of patients with non-missing results at the visit. In 2 years, the most reported new macrovascular complications included CV death (0.1%, 9/6234) and PVD (0.1%, 4/6234) (Table 3). A total of 41 deaths were reported; of which, 30 deaths were attributed to CV causes (sudden death [n = 19], MI [n = 9], stroke [n = 1] and coronary artery procedure [n = 1]) and remaining 11 deaths were due to other causes . Three patients were hospitalized between the 1-year to 18-month period due to MI (one patient) and acute coronary syndrome (ACS; two patients). One patient was hospitalized due to ACS, heart failure and unstable angina between the 18-month to 2-year period (Table S3). 3.4 Cardiovascular risk factors An increasing trend in the CV risk profile of patients was observed (baseline: 52.6%, 3281/6234 vs. 2-year: 55.1%, 3438/6234). Among the 63 new cases (69 events) of CV risk factors, dyslipidaemia and hypertension (33 cases, each) were the most commonly reported (Table 4). Hypertension was noted more in men than women (p = .0019). Patients with hypertension and dyslipidaemia were greater in the subgroup having BMI >= 23 kg/m2 versus BMI < 23 kg/m2 (p < .0001 and p = .0525, respectively). Patients having hypertension and dyslipidaemia were higher in the subgroup having uncontrolled HbA1c levels (>=7%) versus controlled HbA1c levels (<7%) (Table 4). TABLE 4 Summary of cardiovascular risk factors at 2-year visit, by HbA1c, BMI and gender (N = 6234). CV risk factors Total N = 6234 Baseline 2 years Patients at risk with new CV risk factors until 2 years, n Patients with new CV risk factors at 2 years Total number of CV risk factors, Ne 4419 4698 69 Patients with CV risk factors 3281 (52.6) 3438 (55.1) 6234 63 (1.0) Hypertension b 2566 (78.2) 2679 (77.9) 3588 33 (0.9) Dyslipidaemia b 1635 (49.8) 1776 (51.7) 4491 33 (0.7) Albuminuria b 153 (4.7) 169 (4.9) 6234 3 (0.0) Family history of PCD b 65 (2.0) 65 (1.9) - - No complications 2562 - - - Unknown c 391 - - - CV risk factors by HbA1c, BMI and gender Risk factors HbA1c < 7% HbA1c >= 7% p-value a BMI < 23 kg/m2 BMI >= 23 kg/m2 p-value a Men Women p-value a Hypertension 539 (8.7) 1135 (18.2) .1787 264 (4.2) 1808 (29.0) <.0001 1455 (23.3) 1224 (19.6) .0019 Dyslipidaemia 352 (5.7) 807 (13.0) .0098 191 (3.1) 1192 (19.1) .0525 991 (15.9) 785 (12.6) .4441 Albuminuria 2 (0.0) 0 (0.0) .1113 0 (0.0) 3 (0.1) >.9999 0 3 (0.1) .0819 F/H of PCD 9 (0.1) 29 (0.5) .2033 5 (0.1) 44 (0.7) .3112 40 (0.6) 25 (0.4) .4157 Note: Values are presented as n (%) unless specified otherwise. Abbreviations: BMI, body mass index; CV, cardiovascular; F/H, family history; HbA1c, glycated haemoglobin; N, number of patients analysed; n, number of patients with non-missing results at the visit; Ne, number of events; PCD, premature coronary disease. a p-values are reported from Fisher's test if the cell frequency is lesser than 5. p-values are reported using the Chi-square test otherwise. The null hypothesis is that there is no difference between the two population proportions. The p-values reported are not adjusted for inflation in type I error. b Percentages are calculated at baseline based on N = 3281 and at 2-year visit based on N = 3438. c Patients who had chosen "No" and "Unknown" for multiple complications are counted under "Unknown". 3.5 Glycaemic trends and vascular complications in metropolitan and non-metropolitan cities The baseline age, disease duration and HbA1c parameters were comparable across patients of non-metropolitan and metropolitan cities (Table S4A). In 2 years, an improvement (decrease) was noted in all the glycaemic parameters (HbA1c, FPG and PPG) in patients from both metropolitan and non-metropolitan cities (Table S4A). The microvascular complications (neuropathy, nephropathy and retinopathy) were significantly (p < .0001) higher in patients from non-metropolitan than in metropolitan cities (Table S4B). The number of CV deaths was higher in patients from non-metropolitan than in metropolitan cities (19.7%, 25/135 vs. 7.4%, 5/70). Of the newly reported cases of macrovascular complications in the second year, in non-metropolitan cities, PVD was reported in four patients, MI in one patient, CV death in eight patients and stroke in one patient; while in metropolitan cities, one new case was reported (CV death) (Table S4C). Overall, the number of diabetes-related complications in metropolitan and non-metropolitan cities increased from baseline over 2 years (Table S4C). 3.6 Anti-diabetic treatment therapies In 2 years, the total proportion of patients taking OAD + insulin increased (baseline: 24.0%, 1498/6234 vs. 2 years: 33.3%, 1917/5763), while the proportion of those taking only OADs, decreased (baseline: 74.4%, 4640/6234 vs. 2 years: 64.8%, 3735/5763) (Table S1). Biguanides and sulfonylureas were the most prescribed OADs at baseline and 2 years (biguanides, baseline: 93.0%, 5798/6234 and in 2 years: 92.7%, 5340/5763; sulfonylureas, baseline: 76.3%, 4759/6234 and in 2 years: 77.7%, 4480/5763). The highest increase in OAD addition was seen for dipeptidyl peptidase 4 (DPP4) inhibitors (baseline: 48.9%, 3049/6234 vs. 2-years: 61.2%, 3526/5763) followed by sodium-glucose cotransporter-2 inhibitors (baseline: 10.5%, 654/6234 vs. 2 years: 21.3%, 1227/5763) (Table S5). The mean (95% CI) change in HbA1c from baseline was -0.5 (-0.5, -0.4) in <=3 OAD subgroup and -0.4 (-0.6, -0.2) in >3 OAD subgroup. Improvement in the glycaemic parameters in 2 years was more in the <=3 OAD versus >3 OAD subgroup (p values were 0.8193, 0.1139 and 0.5541 for HbA1c, FPG and PPG, respectively) (Table 5). TABLE 5 Comparison of glycaemic assessments between additional therapy groups for subgroup of patients who attended all visits (N = 6234). Change in HbA1c (%) a Change in FPG (mg/dl) a Change in PPG (mg/dl) a n; mean (95% CI) n; mean (95% CI) n; mean (95% CI) Insulin-naive 1454; -0.4 (-0.5, -0.4) 1821; -9.9 (-12.1, -7.8) 1705; -17.2 (-20.6, -13.8) Insulin 505; -1.0 (-1.2, -0.9) 721; -25.8 (-30.5, -21.0) 697; -38.7 (-45.2, -32.2) p-value b <.0001 <.0001 <.0001 Basal (with/without prandial) insulin 234; -1.1 (-1.4, -0.9) 323; -26.0 (-32.9, -19.2) 323; -38.3 (-47.5, -29.0) Premix (with/without prandial) 198; -1.0 (-1.3, -0.8) 298; -27.9 (-35.5, -20.4) 282; -37.3 (-47.9, -26.8) p-value b .5758 .7125 .8915 Basal long-acting (without prandial) insulin 174; -1.3 (-1.6, -1.0) 238; -23.8 (-31.5, -16.0) 230; -40.0 (-51.4, -28.7) Premix (without prandial) insulin 205; -1.1 (-1.3, -0.8) 304; -27.7 (-35.3, -20.2) 289; -39.7 (-50.2, -29.3) p-value b .2520 .4747 .9696 <=3 OAD 843; -0.5 (-0.5, -0.4) 1110; -10.9 (-13.5, -8.3) 1024; -17.4 (-21.6, -13.3) >3 OAD 218; -0.4 (-0.6, -0.2) 262; -6.0 (-11.8, -0.1) 247; -14.6 (-23.1, -6.0) p-value b .8193 .1139 .5541 Abbreviations: CI, confidence interval; FPG, fasting plasma glucose; HbA1c, glycated haemoglobin; OAD, oral anti-diabetic; PPG, postprandial glucose. a Represents change from baseline to 2 years. b p-value is calculated between the treatment subgroups using independent t-test. In 2 years, the commonly prescribed injectables were basal and premix insulins (basal insulin, baseline: 13.5%, 839/6234 and 2 years: 20.6%, 1188/5763; premix insulin, baseline: 11.0%, 683/6234 and 2 years: 14.7%, 849/5763) (Table S5). The change from baseline in all three glycaemic parameters in 2 years was significantly more in the insulin-receiving subgroup than in the insulin-naive subgroup (p < .0001). The mean (95% CI) change in HbA1c from baseline was -1.0 (-1.2, -0.9) in insulin subgroup and -0.4 (-0.5, -0.4) in insulin-naive subgroup (p < .0001, both) (Table 5). 3.7 Adverse drug reactions A total of 19 events (18 patients) of hypoglycaemia were recorded between the 1-year to 18-month period. In the latter 6 months of the 2-year visit, 0.3% of patients (17/6234) reported hypoglycaemic events (documented symptomatic, n = 11; asymptomatic, n = 2; nocturnal, n = 4). (Table S3). One adverse drug reaction (hypoglycaemia) by one patient was noted until the end of 2 years. 4 DISCUSSION This pan-India, real-world, large-scale, longitudinal study is designed to assess glycaemic control, treatment trends, CV risks and development of microvascular complications over 3 years in Indian adults with T2DM. Herewith, we report an interim analysis done at the 2-year time point. Over 2 years, while there was an overall improvement in glycaemic status, only 1 in 5 patients achieved HbA1c < 7%. Approximately one-third of the patients in metropolitan (30.6%) and non-metropolitan (35.4%) cities had HbA1c < 7% at the end of 2 years. While also highest at baseline among microvascular complications, the proportion of patients with neuropathy showed an increase at the end of 2 years. Hypertension and dyslipidaemia were the most reported CV risks. The 2-year results show that the majority of the patients with T2DM are treated with OADs. Biguanides and sulfonylureas are the most commonly used OADs in Indian routine clinical practice. In this study, most of the patients (~90%) were aged between 31 and 65 years and had obesity (67%). With age it is difficult to reduce weight as the deposition of central fat becomes more pronounced and obesity sets in. Obesity paves the way for lifestyle disorders, one of them being T2DM. 16 Previous reports demonstrate that obesity is a well-established risk factor for chronic illnesses like T2DM. 17 , 18 Worsening of T2DM leads to CV risks and vascular complications. 6 , 7 , 8 , 9 , 10 A possible mechanism linking T2DM and obesity with subsequent CV complications is inflammation and lipid accumulation due to overexpression of cytokines (tumour necrosis factor-a, interleukin (IL)-1, IL-6, leptin, resist in monocyte chemoattractant protein (MCP)-1, plasminogen activator inhibitor (PAI)-1, fibrinogen and angiotensin) by adipose tissue, which have a deleterious effect on blood vessels and can lead to the development of MI and cardiomyopathy. 19 Current treatment recommendations instate close monitoring and control of glycaemic levels to improve cardiac outcomes. 2 However, a considerable gap exists between diabetes care followed in real practice versus that recommended by evidence-based guidelines. 20 The results of this study shed light on the real-world burden of uncontrolled diabetes in India. In this 2-year analysis, 20.8% of the study population had optimal glycaemic control (HbA1c < 7%). In the GOAL study, 29.7% of patients had glycaemic control after 12 months, and in the wave -7 of IDMPS study, 25.2% of patients had optimal glycaemic control. 21 , 22 The proportion of patients having optimum glycaemic control worsened over 2 years despite an increase in the use of OADs, reiterating the need for early control. Despite improvement in the HbA1c levels in 0.6% of patients in 2 years, there was an overall increase in the number of patients with microvascular and macrovascular complications. Those patients who were overweight or had suboptimal glycaemic control or CV risk factors had more complications versus those without these comorbidities, thus, substantiating an established fact that high BMI and poor glycaemic control lead to vascular complications. 18 , 19 , 23 The UKPDS 38 study examined the effect of tight control of blood pressure on macrovascular and microvascular complications in patients with T2DM. After 9 years of follow-up, the results showed a 34% reduction in macrovascular complications (MI, sudden death, stroke and PVD) and a 37% reduction in the risk of microvascular complications (retinopathy requiring photocoagulation, vitreous haemorrhage and fatal or non-fatal renal failure) in tightly controlled blood pressure group compared with the less tightly controlled group. 24 In this study, neuropathy was the most reported complication in 2 years. These results are consistent with the observation of the 1-year LANDMARC study 13 and A1chieve study. 25 It is a well-known fact that in people with T2DM, uncontrolled high blood sugar for a long duration degenerates the neurons, leading to a loss of sensory function or diabetic neuropathy. 26 Previous reports have established the observation that hypertension and dyslipidaemia are generally prevalent in people with diabetes. 27 , 28 The 2-year data in the present study also showed that all three microvascular complications (neuropathy, nephropathy and retinopathy) and heart failure were reported in more patients with CV risks than without CV risks. As anticipated with the progressive nature of the disease after 2 years in the study, half of the patients who had diabetes for >10 years at baseline were taking OAD + insulin and those who had diabetes for <=10 years at baseline were predominantly on OADs. Biguanides and sulfonylureas continued to remain the most used OAD classes in 2 years. Similar to the 1-year results, the highest addition was seen in patients on DPP4 inhibitors. 13 There was a shift from the use of OADs towards the introduction of insulin as the need for injectables is common in people with longer duration of diabetes. 2 , 30 In 2 years, improvement in glycaemic parameters was significantly higher in the insulin receiving subgroup than in the insulin-naive subgroup (p < .0001), which is in alignment to the 1-year results of the LANDMARC study. 13 The ORIGIN study showed that the progression of diabetes was substantially reduced with timely insulin treatment in comparison with standard of care. 29 There is also evidence that in patients who are diagnosed with severe hyperglycaemia (HbA1c > 9%-10%), insulin can control lipo-toxicity within a few days of therapy. 30 Therefore, timely initiation and intensification of insulin treatment can help achieve glycaemic control and improve the treatment outcomes. There was an improvement in the HbA1c levels in patients from both metropolitan and non-metropolitan cities. The microvascular complications were significantly more in patients from non-metropolitan cities than in metropolitan cities (p < .0001). The number of CV deaths and newly reported macrovascular complications were also higher in patients from non-metropolitan cities than in metropolitan cities. Difference in the disease management between metropolitan and non-metropolitan cities could result in variable outcomes. Further studies can help understand the diabetes management patterns in metropolitan and non-metropolitan cities. LANDMARC is one of the first-of-its-kind large-scale longitudinal studies from India involving 6234 patients from 382 centers to investigate microvascular and macrovascular complications, glycaemic control and therapy pattern in patients with T2DM over 3 years across India. This real-world data analysis provides a longitudinal course of the T2DM burden, management practices and related complications across the nation including a sub-analysis across metropolitan and non-metropolitan cities of India. This study is inherent to the limitations associated with a real-world study. Moreover, being observational in nature, any study-specific procedures or screening for complications or CV risks were not possible. Furthermore, this study does not capture data on factors such as financial status, educational qualification of the patients and access to treatment facilities that warranted a better understanding, if investigated. 5 CONCLUSIONS The results of the second year of the LANDMARC study showed high burden of uncontrolled diabetes in patients with T2DM from India. In 2 years, 17.6% of the study population had microvascular complications, predominantly neuropathy. A higher number of complications were observed in patients from non-metropolitan versus metropolitan cities. Hypertension was the most reported CV risk. In 2 years, an increase in number of injectables was also observed. These 2-year trends are similar to those observed in the 1-year results of the LANDMARC study. This pan-India, real-world study highlights the need for effective diabetes management including enhanced awareness among patients and providers to meet glycaemic targets and prevent CV risk and vascular complications in a developing country like India with high prevalence of T2DM. AUTHOR CONTRIBUTIONS Ashok Kumar Das: Conceptualization (equal); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (lead); methodology (equal); project administration (supporting); resources (supporting); software (supporting); supervision (lead); validation (lead); visualization (lead); writing - original draft (equal); writing - review and editing (equal). Sanjay Kalra: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Shashank R Joshi: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Ambrish Mithal: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Prasanna Kumar KM: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Ambika G Unnikrishnan: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Hemant Thacker: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Bipin Sethi: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Subhankar Chowdhury: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Amarnath Sugumaran: Conceptualization (equal); data curation (equal); formal analysis (supporting); funding acquisition (equal); investigation (equal); methodology (equal); project administration (equal); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Senthilnathan Mohanasundaram: Conceptualization (equal); data curation (equal); formal analysis (supporting); funding acquisition (equal); investigation (equal); methodology (equal); project administration (equal); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Shalini Kesav Menon: Conceptualization (equal); data curation (equal); formal analysis (supporting); funding acquisition (equal); investigation (equal); methodology (equal); project administration (equal); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Vaibhav Salvi: Conceptualization (supporting); data curation (equal); formal analysis (equal); funding acquisition (supporting); investigation (equal); methodology (equal); project administration (equal); resources (supporting); software (equal); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Deepa Chodankar: Conceptualization (equal); data curation (equal); formal analysis (lead); funding acquisition (supporting); investigation (equal); methodology (lead); project administration (equal); resources (lead); software (lead); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Saket Thaker: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Chirag Trivedi: Conceptualization (equal); data curation (equal); formal analysis (equal); funding acquisition (equal); investigation (equal); methodology (equal); project administration (equal); resources (equal); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Subhash Kumar Wangnoo: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Abdul Zargar: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). Nadeem Rais: Conceptualization (supporting); data curation (supporting); formal analysis (supporting); funding acquisition (supporting); investigation (equal); methodology (supporting); project administration (supporting); resources (supporting); software (supporting); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). FUNDING INFORMATION This study was funded by Sanofi. CONFLICT OF INTEREST AKD, AM, AGU and NR received honoraria from Sanofi and other pharmaceutical companies. KMPK is on the advisory board of Sanofi and received an honorarium for his talks. SJ received speaker/advisory/research grants from Abbott, AstraZeneca, Biocon, Boehringer Ingelheim (BI), Eli Lilly, Franco Indian, Glenmark, Lupin, Marico, MSD, Novartis, Novo Nordisk, Roche, Sanofi, Serdia, Twinhealth and Zydus. SK received honoraria/speaker fees from Eli Lilly, Novo Nordisk and Sanofi. HT received honoraria from MSD, Novartis, Sanofi and other companies for advice and lectures. BS received an honorarium from Aventis, Novo Nordisk, Eli Lilly, BI and MSD. SC received honoraria/grants from Biocon, BI, Intas, Novartis, Sanofi and Serdia. SKW has nothing to declare. AHZ received honoraria from Novo Nordisk, Eli Lilly, Johnson & Johnson, AstraZeneca, BI and Sanofi. AS, SM, SKM, DC, VS, ST and CT are employees of Sanofi and may hold stock options. PREVIOUS PRESENTATIONS AND PUBLICATIONS Part of the data from this paper was presented at the 82nd Scientific Sessions of American Diabetes Association (ADA) 2022, New Orleans, LA, and the 58th Annual Meeting of the European association for the study of Diabetes, Stockholm, 19-23 September 2022. The protocol of this study is published in Diabetic Medicine; DOI: 10.1111/dme.14171. The baseline data and 1-year data of this study have been published in Endocrinology, Diabetes & Metabolism; their DOIs are 10.1002/edm2.231 and 10.1002/edm2.316, respectively. Supporting information Table S1. Table S2. Table S3. Table S4. Table S5. Click here for additional data file. ACKNOWLEDGEMENTS The authors would like to thank the study investigators and study patients as well as their families/caregivers who were involved in this study. The medical writing support was provided by Rukhsar Wasta (Tata Consultancy Services, India) and paid by Sanofi. Editorial support was also provided by Anahita Gouri of Sanofi, India. The authors are responsible individually and collectively for all content and editorial decisions and did not receive any payment from Sanofi directly or indirectly (through a third party) related to the development or presentation of this publication. The authors acknowledge the role of DignoSearch for site management and monitoring activities, JSS Medical Research India Pvt. Ltd and Tech Observer Pvt. Ltd for site management and coordination support and Zifo R&D Solutions for data management services. DATA AVAILABILITY STATEMENT Qualified researchers may request access to person-level data and related study documents including the clinical study report, study protocol with any amendments, blank case report form, statistical analysis plan and data set specifications. Person-level data will be anonymized, and study documents will be redacted to protect the privacy of study patients. Further details on Sanofi's data sharing criteria, eligible studies and the process for requesting access can be found at |
PMC10000634 | Background Metabolic syndrome (MetS) is a global public health concern. Chronic inflammation plays a role in MetS; haematological inflammatory parameters can be used as MetS predicting factors. Objective Hereditary and environmental factors play an important role in the development of MetS. This study aimed to determine the relationship between haematological parameters and MetS in the adult population of southeastern Iran, Kerman. Methods This cross-sectional study was a sub-analysis of 1033 subjects who participated in the second phase of the Kerman Coronary Artery Disease Risk Factor Study (KERCADRS). Metabolic syndrome was diagnosed according to Adult Treatment Panel III (ATP III) definition. Pearson correlation coefficient was used to investigate the relationship between haematological parameters with age and components of metabolic syndrome. The role of WBC, neutrophil, lymphocyte and monocyte in predicting metabolic syndrome was evaluated using the receiver operating characteristic (ROC) curve. Results White blood cell (WBC) and its subcomponent cells count, red cell distribution width (RDW), monocyte to HDL ratio (MHR) and Neutrophil to HDL ratio (NHR) had a significant positive correlation with the severity of MetS. The cut-off value of WBC was 6.1 (x103/mL), the sensitivity was 70%, the specificity was 52.9% for females, the cut-off value of WBC was 6.3 (x103/mL), the sensitivity was 68.2% and the specificity was 46.7%, for males. Conclusion WBC and its subcomponent count, RDW, MHR and NHR parameters are valuable biomarkers for further risk appraisal of MetS in adults. These markers are helpful in early diagnoses of individuals with MetS. Chronic inflammation plays a role in Metabolic syndrome (MetS); hematologic inflammatory parameters can be used as MetS predicting factors.Adults with high WBC count, RDW, MHR, and NHR without any associated underlying chronic disease must be screened because they are at high risk of developing MetS. complete blood count Haematological parameters leukocytes metabolic syndrome obesity Kerman University of Medical Sciences 10.13039/501100004621 95000008 source-schema-version-number2.0 cover-dateMarch 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.2.6 mode:remove_FC converted:10.03.2023 Najafzadeh MJ , Baniasad A , Shahabinejad R , Mashrooteh M , Najafipour H , Gozashti MH . Investigating the relationship between haematological parameters and metabolic syndrome: A population-based study. Endocrinol Diab Metab. 2023;6 :e407. doi:10.1002/edm2.407 Mohammad Javad Najafzadeh and Amir Baniasad contributed equally to this work. pmc1 INTRODUCTION Metabolic syndrome (MetS) has had different definitions since 1988, which was first introduced by Reaven. 1 Based on the latest definition of this syndrome, MetS includes at least three factors from the following disorders: central obesity, hypertension, elevated fasting glucose and dyslipidemia (reduced high-density lipoprotein (HDL) or elevated triglycerides (TG)), 2 which increases the risk of insulin resistance, diabetes mellitus, cerebrovascular disease, cardiovascular disease, common cancers, osteoporosis and total mortality. 3 , 4 , 5 Prevalence and incidence of MetS have increased significantly following the increase in urbanization, improper nutrition and lack of physical activity, and it has become a global health concern. 6 , 7 Although the underlying mechanism of MetS has not been known yet, oxidative stress, chronic inflammation and insulin resistance seem to be the most likely mechanism. 8 A growing number of studies emphasize the association of MetS components and haematological parameters, including white blood cell (WBC), red blood cell (RBC), platelet (PLT) count and haematocrit (HCT) level as potential indicators markers of thrombotic and inflammatory states. 9 , 10 , 11 , 12 Meng et al. demonstrated that leukocyte was a good marker for assessing the risk of MetS and cardiovascular disease. 13 Some studies reported that WBC and PLT counts were significantly correlated with the numbers of MetS components. 14 , 15 Ahmadzadeh et al. pointed out that high haemoglobin (HB) levels and HCT can also indicate MetS development. 16 Since inflammation plays a role in MetS, these haematological inflammatory parameters can be used as MetS predicting factors. Performing cost-effective CBC tests can easily measure haematological parameters from peripheral blood. Hereditary and environmental factors play an important role in the development of MetS. Currently, no study has investigated the characteristics of MetS and its relationship with blood parameters in the population of southeastern Iran. Therefore, this study aimed to determine the relationship between haematological parameters and MetS in southeastern Iran, Kerman. 2 METHODS 2.1 Study design and participants This cross-sectional study was a sub-analysis of 1033 subjects who participated in the second phase of the Kerman Coronary Artery Disease Risk Factor Study (KERCADRS). 17 The sampling method was a cluster from the entire population of Kerman residents. In the first phase of the KERCADRS, according to the post-office list of city residents, 250 postal codes were selected randomly. We invited people over 15 to participate in the study. In the first phase, 24 people were collected in each cluster. In the second phase (420 clusters including 24 participants), people were contacted again, and 1033 who met the inclusion criteria from February 2017 to October 2018 were included in our study . None of the included participants had a history of chronic infectious or inflammatory diseases or the use of any drugs known to affect haematological parameters or lipoprotein metabolism. More details about the data collection method have been published in the study of Najafipour et al. 17 FIGURE 1 The flowchart of included participants. 2.2 Data collection After obtaining informed consent forms from the subjects, demographic data (age, gender) and anthropometric information were collected. A trained interviewer asked participants about cigarette smoking and opium use. People who routinely smoked cigarettes or consumed opium at the time of data collection were considered cigarette smokers and opium addicts, respectively. Height in the standing position without shoes, from heel to head, with an error of 0.5 cm error, weight without shoes and extra clothing with an error of 100 g on a digital scale, body mass index (BMI) which the weight (kg) of the patients was divided by the square of their height (m2), waist circumference (WC) in the standing position with 20-30 cm distance between the feet were measured. WC (cm) was measured at the umbilical level. Hip circumference (HC) (cm) was measured based on the largest circumference around the buttocks. Waist-to-hip ratio (WHR) was calculated by dividing WC by HC. After 10 min of rest, blood pressure (BP) was measured with a standard manometer from the right arm in the sitting position according to the World Health Organization (WHO) standards, and the blood samples were taken after 12-14 hours of fasting and kept at room temperature. CBC, fasting plasma glucose (FPG) and serum lipids (HDL cholesterol and TG) were tested by routine laboratory methods. According to Adult Treatment Panel III (ATP III) definition, the presence of at least two of the following five factors is required for the diagnosis of metabolic syndrome: blood pressure over 130/80 mm Hg or consumption of antihypertensive drugs, TG level over 150 mg/dl, FPG over 100 mg/dl or consumption of anti-diabetic medication like insulin, HDL cholesterol level less than 40 mg/dL (men) or 50 mg/dl (women) and WC over 102 cm (men) or 88 cm (women). 2.3 Sample size estimation In the study of Oda and Kawai, the mean WBC in women with three components of metabolic syndrome was 5416 +- 1163 and in women with only two components of metabolic syndrome was 5077 +- 1358. 18 The minimum sample size required based on the mentioned numbers and considering the power of 0.8 and alpha of 0.05 for each group was considered to be at least 275 people. 2.4 Statistical analysis Statistical analysis was performed using SPSS version 16 software (SPSS Inc.). Quantitative variables were reported as mean +- standard deviation, and qualitative variables were reported as numbers and percentages. Qualitative variables were compared between the two groups using the Pearson chi-square or Fisher's exact test. Quantitative variables were compared separately between individuals with and without metabolic syndrome in male and female groups using the independent samples t test. Pearson correlation coefficient was used to investigate the relationship between haematological parameters with age and components of metabolic syndrome. The relationship between haematological parameters and the variable number of components of metabolic syndrome was investigated using Spearman's correlation coefficient. The role of WBC, neutrophil, lymphocyte and monocyte in predicting metabolic syndrome was evaluated using the receiver operating characteristic (ROC) curve and with MedCalc(r) Statistical Software version 20.013 (MedCalc Software Ltd). The optimal cut-off point was determined using the Youden index. 3 RESULTS A total of 1033 individuals (660 women, 373 men) were included in this study, and the sociodemographic, laboratory parameters and clinical characteristics of the participants are summarized in Table 1. TABLE 1 Basic Characteristics of the Study Subjects. Male p value a Female p value a Normal Syndrome Metabolic Normal Syndrome Metabolic N = 291 N = 82 N = 463 N = 197 Age (years) 45.48 +- 16.21 51.90 +- 13.84 <.001 40.29 +- 14.07 53.33 +- 11.72 <.001 Smoking, n (%) 69 (23.7) 18 (22.0) .739 3 (0.6) 1 (0.5) .655 Opium addiction, n (%) 64 (22.0) 17 (20.7) .807 21 (4.5%) 15 (7.6%) .111 BMI (kg/m2) 24.99 +- 4.22 29.18 +- 3.86 <.001 26.38 +- 4.86 30.81 +- 5.17 <.001 WC (cm) 88.26 +- 11.94 101.29 +- 9.37 <.001 83.30 +- 11.83 98.16 +- 10.47 <.001 WHR 0.89 +- 0.07 0.96 +- 0.05 <.001 0.82 +- 0.08 0.93 +- 0.08 <.001 SBP (mmHg) 114.98 +- 16.01 127.99 +- 16.92 <.001 108.98 +- 15.56 122.60 +- 17.43 <.001 DBP (mmHg) 74.98 +- 9.39 82.38 +- 11.28 <.001 71.79 +- 10.46 78.11 +- 9.46 <.001 FPG 90.86 +- 24.05 116.20 +- 41.39 <.001 86.67 +- 18.34 120.40 +- 47.38 <.001 TG (mg/dl) 127.03 +- 66.25 221.15 +- 163.26 <.001 100.11 +- 43.90 184.71 +- 78.75 <.001 HDL (mg/dl) 46.43 +- 10.55 37.87 +- 7.27 <.001 52.97 +- 12.41 44.31 +- 9.57 <.001 LDL (mg/dl) 109.31 +- 31.35 101.07 +- 37.57 .077 109.29 +- 35.07 113.65 +- 40.42 .188 Cholestrol (mg/dl) 181.08 +- 37.04 181.66 +- 43.56 .912 182.07 +- 43.53 194.74 +- 45.56 .001 WBC (x103/mL) 6.76 +- 1.68 7.08 +- 1.68 .131 6.30 +- 1.66 6.92 +- 1.52 <.001 Neutrophil (x103/mL) 3.46 +- 1.25 3.72 +- 1.38 .119 3.35 +- 1.21 3.70 +- 1.12 .001 Lymphocyte (x103/mL) 2.48 +- 0.73 2.51 +- 0.73 .793 2.24 +- 0.66 2.46 +- 0.68 <.001 Monocyte (x103/mL) 0.58 +- 0.17 0.61 +- 0.16 .174 0.51 +- 0.14 0.54 +- 0.15 .005 RBC (x106/mL) 5.33 +- 0.54 5.35 +- 0.60 .745 4.78 +- 0.47 4.78 +- 0.49 .951 HB (gr/dl) 15.19 +- 1.26 15.33 +- 1.28 .378 13.24 +- 1.18 13.40 +- 1.38 .177 HCT (%) 46.04 +- 3.51 45.87 +- 4.22 .741 41.08 +- 3.57 41.31 +- 4.23 .516 PLT(x103/mL) 220.67 +- 49.95 215.24 +- 45.32 .376 254.18 +- 62.34 255.26 +- 56.46 .834 MPV (fL) 10.27 +- 0.87 10.22 +- 0.77 .640 10.54 +- 0.94 10.47 +- 0.90 .355 RDW-SD 42.97 +- 3.14 42.76 +- 3.24 .602 42.94 +- 2.98 43.56 +- 3.10 .017 RDW-CV 13.89 +- 1.35 13.94 +- 1.26 .787 14.02 +- 1.34 14.16 +- 1.39 .242 NLR 1.49 +- 0.68 1.59 +- 0.76 .233 1.57 +- 0.62 1.61 +- 0.67 .539 PLR 94.85 +- 31.90 91.59 +- 28.99 .405 120.76 +- 40.62 110.52 +- 37.39 .003 PMR 403.18 +- 128.29 367.38 +- 96.79 .007 529.20 +- 169.01 499.46 +- 162.80 .037 MHR 0.013 +- 0.005 0.0176 +- 0.01 <.001 0.010 +- 0.004 0.013 +- 0.004 <.001 NHR 0.08 +- 0.04 0.10 +- 0.04 <.001 0.07 +- 0.03 0.09 +- 0.03 <.001 Abbreviations: BMI, Body mass index; DBP, Diastolic blood pressure; FPG, Fasting plasma glucose; HB, Haemoglobin; HCT, Haematocrit; HDL, High-density lipoprotein; LDL, Low-density lipoprotein; MHR, Monocyte to HDL ratio; MPV, Mean platelet volume; NHR, Neutrophil to HDL ratio; NLR, Neutrophil to lymphocyte ratio; PLR, Platelet to lymphocyte ratio; PLT, Platelet; PMR, Platelet to monocyte ratio; RBC, Red blood cell; RDW-CV, Red cell distribution width - coefficient of variation; RDW-SD, Red cell distribution width - standard deviation; SBP, Systolic blood pressure; TG, Triglycerides; WBC, White blood count; WC, Waist circumference; WHR, Waist-to-hip ratio. a Independent sample t test was used for quantitive variables, and Pearson chi-square or Fisher's exact test was used for qualitative variables. In both males and females, in the participants with MetS, age, BMI, WC, WHR, systolic blood pressure (SBP), diastolic blood pressure (DBP), FPG, TG, monocyte to HDL ratio (MHR) and Neutrophil to HDL ratio (NHR) were significantly higher compared with the participants without MetS. In females with MetS, WBC, Red Cell distribution width-standard deviation (RDW-SD), Neutrophil, Lymphocyte and Monocyte were significantly higher than the females without MetS (Table 1). HDL and platelet to Monocyte ratio (PMR) were significantly lower in participants with MetS compared with those without MetS in males and females. In females with MetS, the platelet to Lymphocyte ratio (PLR) was significantly lower than those without MetS (Table 1). In males with MetS, smoking, opium addiction, WC, WHR, SBP, DBP, monocyte, RBC, HB, HCT, MHR and NHR were significantly higher than these parameters in females with MetS. In females with MetS, BMI, HDL, LDL, cholesterol, PLT, MPV, PLR and PMR were significantly higher than these parameters in males with MetS (Table 2). TABLE 2 Comparison of characteristics between male and female patients with metabolic syndrome. Syndrome Metabolic p Value a Male Female N = 82 N = 197 Age (years) 51.90 +- 13.84 53.33 +- 11.72 0.381 Smoking, n (%) 18 (22.0) 1 (0.5) <0.001 Opium addiction, n (%) 17 (20.7) 15 (7.6%) 0.002 BMI (kg/m2) 29.18 +- 3.86 30.81 +- 5.17 0.011 WC (cm) 101.29 +- 9.37 98.16 +- 10.47 0.020 WHR 0.96 +- 0.05 0.93 +- 0.08 0.001 SBP (mmHg) 127.99 +- 16.92 122.60 +- 17.43 0.019 DBP (mmHg) 82.38 +- 11.28 78.11 +- 9.46 0.001 FPG 116.20 +- 41.39 120.40 +- 47.38 0.485 TG (mg/dl) 221.15 +- 163.26 184.71 +- 78.75 0.057 HDL (mg/dl) 37.87 +- 7.27 44.31 +- 9.57 <0.001 LDL (mg/dl) 101.07 +- 37.57 113.65 +- 40.42 0.018 Cholestrol (mg/dl) 181.66 +- 43.56 194.74 +- 45.56 0.028 WBC (x103/mL) 7.08 +- 1.68 6.92 +- 1.52 0.434 Neutrophil (x103/mL) 3.72 +- 1.38 3.70 +- 1.12 0.935 Lymphocyte (x103/mL) 2.51 +- 0.73 2.46 +- 0.68 0.640 Monocyte (x103/mL) 0.61 +- 0.16 0.54 +- 0.15 0.001 RBC (x106/mL) 5.35 +- 0.60 4.78 +- 0.49 <0.001 HB (gr/dl) 15.33 +- 1.28 13.40 +- 1.38 <0.001 HCT (%) 45.87 +- 4.22 41.31 +- 4.23 <0.001 PLT(x103/mL) 215.24 +- 45.32 255.26 +- 56.46 <0.001 MPV (fL) 10.22 +- 0.77 10.47 +- 0.90 0.030 RDW-SD 42.76 +- 3.24 43.56 +- 3.10 0.056 RDW-CV 13.94 +- 1.26 14.16 +- 1.39 0.212 NLR 1.59 +- 0.76 1.61 +- 0.67 0.883 PLR 91.59 +- 28.99 110.52 +- 37.39 <0.001 PMR 367.38 +- 96.79 499.46 +- 162.80 <0.001 MHR 0.0176 +- 0.01 0.013 +- 0.004 <0.001 NHR 0.10 +- 0.04 0.09 +- 0.03 0.002 Abbreviations: BMI, Body mass index; DBP, Diastolic blood pressure; FPG, Fasting plasma glucose; HB, Haemoglobin; HCT, Haematocrit; HDL, High-density lipoprotein; LDL, Low-density lipoprotein; MHR, Monocyte to HDL ratio; MPV, Mean platelet volume; NHR, Neutrophil to HDL ratio; NLR, Neutrophil to lymphocyte ratio; PLR, Platelet to lymphocyte ratio; PLT, Platelet; PMR, Platelet to monocyte ratio; RBC, Red blood cell; RDW-CV, Red cell distribution width - coefficient of variation; RDW-SD, Red cell distribution width - standard deviation; SBP, Systolic blood pressure; TG, Triglycerides; WBC, White blood count; WC, Waist circumference; WHR, Waist-to-hip ratio. a The independent sample t test was used for quantitive variables, and the Pearson chi-square test was used for qualitative variables. As shown in Table 3, we have considered the number of metabolic components as a measure to determine the severity of MetS, WBC count, Neutrophil, Lymphocyte, Monocyte, RDW-SD, Red cell distribution width-coefficient of variation (RDW-CV), PLR, MHR and NHR parameters that were significantly correlated with the severity of MetS. The correlation was positive in mentioned parameters except for PLR. WBC was significantly correlated with all metabolic components except age (Table 3). TABLE 3 Results of the correlation analysis between the haematological parameters and the components of the metabolic syndrome. Variables Age (years) b WC (cm) b FPG (mg/dl) b TG (mg/dl) b HDL (mg/dl) b SBP (mmHg) b DBP (mmHg) b Number of components a WBC (x103/mL) -0.001 0.139** 0.095** 0.165** -0.161** 0.080* 0.069* 0.191 Neutrophil (x103/mL) -0.029 0.122** 0.081** 0.095** -0.112** 0.049 0.054 0.164** Lymphocyte (x103/mL) 0.022 0.087** 0.069* 0.188** -0.138** 0.074* 0.047 0.150** Monocyte (x103/mL) 0.059 0.113** 0.045 0.089** -0.139** 0.086** 0.063* 0.089** RBC (x106/mL) -0.027 0.097** -0.021 0.132** -0.128** 0.154** 0.176** -0.029 HB (gr/dl) 0.074* 0.130** -0.001 0.178** -0.142** 0.150** 0.161** -0.026 HCT (%) 0.123** 0.136** -0.044 0.142** -0.053 0.173** 0.188** -0.038 PLT (x103/mL) -0.099** -0.035 -0.026 -0.005 0.117** -0.015 0.025 0.034 MPV (fL) -0.037 -0.031 0.059 -0.085** -0.038 -0.075* -0.092** 0.016 RDW-SD* 0.280** 0.152** -0.046 -0.026 0.088** 0.080** 0.057 0.067* RDW-CV 0.037 0.074* -0.017 -0.025 -0.033 0.060 0.051 0.106** NLR -0.025 0.047 0.030 -0.037 0.002 -0.007 0.020 0.025 PLR -0.077* -0.106** -0.073* -0.158** 0.195** -0.074* -0.022 -0.107** PMR -0.101** -0.126** -0.059 -0.070* 0.204** -0.087** -0.034 -0.054 MHR 0.022 0.231** 0.077* 0.300** -0.649** 0.080* 0.053 0.335** NHR -0.038 0.224** 0.101** 0.268** -0.579** 0.058 0.054 0.371** Abbreviations: DBP, Diastolic blood pressure; FPG, Fasting plasma glucose; HB, Haemoglobin; HCT, Haematocrit; HDL, high-density lipoprotein; MHR, Monocyte to HDL ratio; MPV, Mean platelet volume; NHR, Neutrophil to HDL ratio.NLR, Neutrophil to lymphocyte ratio; PLR, Platelet to lymphocyte ratio; PLT, Platelet; PMR, Platelet to monocyte ratio; RBC, Red blood cell; RDW-CV, Red cell distribution width - coefficient of variation; RDW-SD, Red cell distribution width - standard deviation; SBP, Systolic blood pressure; WBC, White blood count; WC, Waist circumference. Note: *p < .05, **p < .01. a Spearman's correlation coefficient, b Pearson correlation coefficient. Genders had different accuracy of WBC, Neutrophil, Lymphocyte and Monocyte in predicting MetS. The accuracy of WBC was higher for females (AUC = 0.632; p < .001; 95% confidence interval [CI]: 0.594-0.669) than for males (AUC = 0.564; p = .074; 95% confidence interval [CI]: 0.512-0.615) (Table 4). TABLE 4 Areas Under the ROC Curve (AUC), sensitivity and specificity by the optimized cut-off points for WBC, Neutrophil, Lymphocyte and Monocyte in predicting metabolic syndrome. WBC (x103/mL) Neutrophil (x103/mL) Lymphocyte (x103/mL) Monocyte (x103/mL) Male AUC (95% CI) 0.564 (0.512-0.615) 0.566 (0.514-0.617) 0.503 (0.451-0.555) 0.549 (0.497-0.600) Optimal cut-off point 6.34 3.41 2.37 0.43 Sensitivity (%) 68.29 57.32 56.10 93.9 Specificity (%) 46.74 56.70 51.89 17.87 Youden index 0.150 0.140 0.080 0.118 p Value .074 .065 .943 .154 Female AUC (95% CI) 0.632 (0.594-0.669) 0.608 (0.569-645) 0.609 (0.566-0.642) 0.570 (0.532-0.609) Optimal cut-off point 6.15 3.67 2.36 0.44 Sensitivity (%) 70.05 50.25 52.28 77.16 Specificity (%) 52.92 71.0 63.93 36.93 Youden index 0.230 0.212 0.162 0.141 p Value <.001 <.001 <.001 .003 Abbreviation: WBC, White blood count. The cut-off value of WBC was 6.1 (x103/mL), the sensitivity was 70%, the specificity was 52.9% for females, the cut-off value of WBC was 6.3 (x103/mL), the sensitivity was 68.2%, the specificity was 46.7%, for males. Neutrophil for males (AUC = 0.566) and WBC for females (AUC = 0.632) had better accuracy in predicting MetS compared to other parameters (Table 4) . FIGURE 2 Areas Under the ROC curve (AUC) for WBC, Neutrophil, Lymphocyte and Monocyte in predicting metabolic syndrome for males. FIGURE 3 Areas Under the ROC curve (AUC) for WBC, Neutrophil, Lymphocyte and Monocyte in predicting metabolic syndrome for females. 4 DISCUSSION We found that MetS affected the haematological parameters of the patients, including WBC and its subcomponent cell count, RDW, PLR, MHR and NHR. In our study, WBC and its subcomponent cells count had a significant positive correlation with the severity of MetS, especially in females. Our results were in parallel with previous studies, which had reported a significant difference in the WBC and its subcomponent cells, between participants with or without MetS. 11 , 16 , 19 Yang et al. reported that the number of total leukocyte-related parameters were elevated in individuals aged 60 years or above. 20 Ahmadzadeh et al. demonstrated that MetS components were significantly correlated with WBC and its subcomponent cells count. 16 In Hedayati et al. study in western Iran, the means of WBC count in the MetS group were significantly higher than the control group. 21 Consistent with our data in a study by Chen et al., contrary to the platelet-related parameters, the WBC-related parameters had significant changes in patients with MetS. 22 In a study on a total of 100 healthy subjects and 200 patients with MetS, total leukocyte and neutrophil counts were significantly increased in all groups of MetS patients compared to the healthy group. 23 Insulin resistance and chronic inflammation are associated with metabolic syndrome by synthesizing some cytokines leading to an increase in WBC and its subcomponent cells count. 24 Lorenzo et al. observed an association between the increased risk of diabetes and elevated WBC, neutrophil and lymphocyte counts due to insulin resistance/sensitivity mechanism. 25 In addition, the relationship between higher levels of WBC count and higher BMI values has been observed in both sexes. 26 In this study, it was found that in the group of patients with MetS, women had greater BMI, higher cholesterol, PLT and platelet-related ratios, and men had a history of more smoking and opium consumption, higher BP, HB, HCT, RBC, monocytes and monocytes-related ratios. Consistent with our results in another study, it has been determined that the predominant feature of MetS in women was abdominal obesity and impaired lipid profile, and in men, it was high BP and impaired lipid profile. 27 In our study, RDW had a significant positive correlation with the severity of MetS, especially in females; however, this correlation was not observed in mean platelet volume (MPV). So far, minimal studies have been done in this field. Consistent with our data, Farah et al. indicated that both RDW and MPV markers increased as the severity of MetS increased. 28 Abdel-Moneim et al. found higher levels of MPV in MetS patients. 23 Zhao et al. demonstrated that MPV was inversely related to MS in women. 29 In another study, no significant difference in the MPV between those with and without MetS was observed. 16 In our study, MHR and NHR had a significant positive correlation with the severity of MetS. A recent study demonstrated that NHR and Lymphocyte to HDL ratio (LHR) were significantly correlated with the prevalence of MetS; also, the correlation was more profound in females. 30 A recent study showed that both MHR and NHR were significantly increased in patients with nascent MetS. 31 Considering that monocyte is an indicating factor for inflammatory conditions and atherosclerosis, 32 , 33 some studies revealed that the ratio of MHR is a suitable predictor to determine the development and severity of MetS and cardiovascular events. 34 , 35 According to our result, Neutrophil to Lymphocyte ratio (NLR) was not recognized as a MetS predictive factor. Ryder et al. observed no association between NLR and obesity or insulin resistance. 36 Contrary to our results, it was found in two studies that patients with MetS had a higher NLR. 23 , 37 Liu et al. relieved that the risk of MetS increased with increasing NLR, and NLR was mentioned as a factor for predicting the development of MetS. 38 In addition, this ratio has been mentioned as a predictive factor for diabetes in obese individuals. 39 PLR had a significant negative correlation with all metabolic components except DBP in this study. In another study, it was reported that the amount of PLR in patients with MetS was higher than in patients without MetS, and the amount of PLR had a significant positive correlation with C-reactive protein (CRP) levels. 40 In Abdel-Moneim et al. study, the PLR was significantly higher in all patients with MetS than in healthy subjects. 23 The cut-off points for WBC and its subcomponent cell counts are used to determine the potential risk of developing MetS. The cut-off value of WBC was 6.1 (x103/mL) and 6.3 (x103/mL) for females and males, respectively, in our study. Our results are confirmatory of previous study findings. Pei et al. reported a cut-off point of 5.6 (x103/mL) for men and 5.8 (x103/mL) for women, 41 and De Oliveira et al. reported a cut-off point of 7.5 (x103/mL) for men and 5.6 (x103/mL) for women. 42 4.1 Limitations Our study had multiple limitations. Firstly, this is a cross-sectional study, and the analysis of the causative effects was not performed. Secondly, the study sample size was small, and the number of males was smaller than females. Further, our study population included people from southeastern Iran, Kerman; we cannot generalize our results to the whole Iran population. 4.2 Future directions The measurement of haematological parameters is easily available in most parts of the world. However, unfortunately, in public health policies, these parameters do not have a place in the diagnosis and follow-up of patients with MetS, which will cause these patients to be missed and impose a lot of financial and social costs on the global health system. The results of our study can b an incentive to conduct prospective studies that will lead to the inclusion of haematological parameters in the diagnostic criteria of MetS. Measuring these haematological parameters is cost-effective and convenient and facilitates screening patients suspected of MetS and their follow-up. Prospective studies are required to explain the causality effects between MetS and haematological parameters, confirm our data and evaluate the need to change the risk assessment criteria for MetS. 5 CONCLUSION The higher levels of WBC and its subcomponent cell count, RDW, MHR and NHR could redict an increased chance of developing MetS, regardless of gender differences. WBC also correlated with MetS components, such as WC, FPG, TG, HDL, SBP and DBP; these parameters are easy to access in patients. Considering that no study has been done on this topic in the population of Southeast Iran, our findings provide additional evidence for using these markers for the early detection of MetS components, which ultimately improves the existing clinical practice in identifying and following MetS patients. AUTHOR CONTRIBUTIONS Mohammad Javad Najafzadeh: Conceptualization (equal); investigation (equal); writing - original draft (equal); writing - review and editing (equal). Amir Baniasad: Conceptualization (equal); formal analysis (equal); methodology (equal); writing - review and editing (equal). Reza Shahabinejad: Data curation (equal); investigation (equal); software (equal); writing - original draft (equal). Mahdieh Mashrooteh: Formal analysis (equal); investigation (equal); methodology (equal); software (equal). Dr Hamid Najafipour: Conceptualization (equal); investigation (equal); methodology (equal); project administration (equal); writing - review and editing (equal). Mohammad Hossein Gozashti: Conceptualization (equal); investigation (equal); methodology (equal); project administration (equal); writing - original draft (equal); writing - review and editing (equal). FUNDING INFORMATION The Kerman University of Medical Sciences funded this research project (Reg. No. 95000008). CONFLICT OF INTEREST The authors declare that they have no conflict of interest. ETHICAL APPROVAL The study protocol was reviewed and approved by the ethics committee of the Kerman University of Medical Sciences (ethic code: IR.KMU.REC.1395.775). Informed consent was obtained from all participants in the study. ACKNOWLEDGEMENT We appreciate all people who participate in the study. Also, we would like to thank the Kerman University of Medical Sciences for funding us. DATA AVAILABILITY STATEMENT The data supporting this study's findings are available from the corresponding author upon reasonable request. |
PMC10000635 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051015 foods-12-01015 Communication Marine Capture Fisheries from Western Indian Ocean: An Excellent Source of Proteins and Essential Amino Acids Jensen Ida-Johanne Conceptualization Methodology Validation Writing - original draft Writing - review & editing Funding acquisition 12* Bodin Nathalie Conceptualization Methodology Software Validation Writing - review & editing Funding acquisition 34 Govinden Rodney Conceptualization Validation Writing - review & editing Funding acquisition 3 Elvevoll Edel Oddny Methodology Validation Writing - review & editing Funding acquisition 1 Bronze Maria Do Rosario Academic Editor Duarte Noelia Academic Editor Serra Teresa Academic Editor 1 Norwegian College of Fishery Science, Faculty of Biosciences, Fisheries, and Economics, UiT-The Arctic University of Norway, N-9037 Tromso, Norway 2 Department of Biotechnology and Food Science, Norwegian University of Science and Technology, NTNU, 7491 Trondheim, Norway 3 Seychelles Fishing Authority (SFA), Fishing Port, Victoria P.O. Box 449, Mahe, Seychelles 4 Sustainable Ocean Seychelles, BeauBelle, Mahe, Seychelles * Correspondence: [email protected] or [email protected] 27 2 2023 3 2023 12 5 101520 12 2022 15 1 2023 24 1 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). The Republic of Seychelles is located in Western-Central Indian Ocean, and marine capture fisheries play a key role in the country's economic and social life in terms of food security, employment, and cultural identity. The Seychellois are among the highest per capita fish-consuming people in the world, with a high reliance on fish for protein. However, the diet is in transition, moving towards a Western-style diet lower in fish and higher in animal meat and easily available, highly processed foods. The aim of this study was to examine and evaluate the protein content and quality of a wide range of marine species exploited by the Seychelles industrial and artisanal fisheries, as well as to further to assess the contribution of these species to the daily intake recommended by the World Health Organization (WHO). A total of 230 individuals from 33 marine species, including 3 crustaceans, 1 shark, and 29 teleost fish, were collected from the Seychelles waters during 2014-2016. All analyzed species had a high content of high-quality protein, with all indispensable amino acids above the reference value pattern for adults and children. As seafood comprises almost 50% of the consumed animal protein in the Seychelles, it is of particular importance as a source of essential amino acids and associated nutrients, and as such every effort to sustain the consumption of regional seafood should be encouraged. Small Island Developing States (SIDS) pelagic and reef species essential nutrients ocean food security amino acid composition protein French Research Institute for Sustainable DevelopmentEuropean Fisheries Partnership AgreementThe Arctic University of Norway2061344 NTNU Norwegian University of Science and TechnologyThe present work is a contribution to the SEYFISH project ("Nutrients and Contaminants in Seychelles fisheries resources") with the financial support of the French Research Institute for Sustainable Development (IRD) and the European Fisheries Partnership Agreement (EU-FPA), as well as UiT--The Arctic University of Norway, project SECURE, Cristin grant ID 2061344. The publication charges for this article were funded by the NTNU Norwegian University of Science and Technology. pmc1. Introduction The significance of global food and nutrition security is anchored in the United Nations Sustainable Development Goals (SDGs) SDG2 "Zero Hunger" and SDG 3 "Good Health and Well-Being" . It is strongly encouraged that an increased food production should come from well-managed ocean resources. Land-based resources are limited, and agricultural food production is one of the major greenhouse gas (GHG) emitters . Seafood plays an important role in food and nutrition security, particularly in middle-income countries . The nutritional recommendations to eat fish are based on their lipid content and fatty acid composition , although seafood is also an important source of vitamins and minerals and high-quality proteins that are important for human health and disease prevention. Seafood is also recognized as a rich source of taurine , considered to have a positive impact on cardiovascular diseases . Seafood may also be a source of toxic heavy metals such as mercury, arsenic, lead, and cadmium , as well as persistent organic pollutants such as dioxins and dioxin-like polychlorinated biphenyls (PCBs) . However, public agencies in Europe have reviewed available evidence through 2021 and concluded that the possible adverse effects of mercury, dioxin, and dioxin-like PCB exposure are offset by the benefits of seafood consumption on cardio-metabolic diseases in general and that seafood consumption during pregnancy is likely to benefit the neurocognitive development of children . Additionally, a relatively recent review conducted by Hibbeln et al. concluded with moderate and consistent evidence that seafood consumption during pregnancy and childhood had beneficial associations with neurocognitive outcomes. Since 1986, an ongoing research project in the Seychelles (Seychelles Child Development Study, SCDS; accessed on 1 January 2023) has been examining associations between maternal methylmercury exposure and neurodevelopment in children . The Republic of Seychelles, one of the 38 United Nations member states of the Small Island Developing States' group, is located in Western-Central Indian Ocean. It includes a land surface of only 459 km2 divided into 115 tropical islands scattered within an Exclusive Economic Zone (EEZ) of 1.3 million square kilometers . The majority of the population resides on three islands of a large submerged mid-oceanic shelf called the Mahe Plateau. Marine capture fisheries play a key role in the country's economic and social life. In addition to the industrial tuna fisheries being a major pillar of the economy, artisanal fisheries are continuously of great importance to the local population in terms of food security, employment, and cultural identity. Fish is seen not only as a staple food but also as a delicacy in the local Creole cuisine, and the Seychellois are among the highest per capita fish-consuming people in the world, with a high reliance on fish for protein--consuming about 59 kg per year measured as live weight , which is equivalent to 48% of the animal protein consumed . Pregnant women and mothers have been reported to consume as many as 12 meals consisting of fish per week . However, the diet in the Seychelles, as elsewhere, is in transition, moving towards a Western-style diet lower in fish and higher in animal meat and easily available, highly processed foods . This has contributed to the increase in the prevalence of obesity (BMI >= 30 kg/m2) between 1998 and 2004 from 4 to 15% in men and from 23 to 34% in women , and it highlights the importance of fish in the diet. Adequate protein intake is essential for tissue maintenance and growth, with amino acids being important as building blocks of proteins and as intermediates in various metabolic pathways. The nutritional quality of a protein is dependent on the content of indispensable, also called essential, amino acids, i.e., amino acids that are not synthesized in our body to meet the human requirements. The World Health Organization (WHO) recommends a daily dietary intake of protein of 830 mg protein/kg body weight for healthy adults; an additional 1, 9, and 31 g protein/day for pregnant women in the first, second, and third trimester, respectively; and 910 mg/kg body weight for children, in addition to specific recommendations for each of the indispensable amino acids . The protein content of marine capture fisheries can significantly vary between species and even within species depending on habitat, region, and season . Access to local and up-to-date food composition data is therefore essential for dietary counselling, clinical nutrition, and improvements in nutrition security and the development of effective nutrition-related policies . To our understanding, the protein contents and quality of different marine capture species from the Seychelles have not been investigated, nor have any data been published. The objectives of this work were to examine the amino acid composition and to evaluate the protein content and protein quality of a wide range of marine species exploited by Seychelles industrial and artisanal fisheries, as well as further to assess the contribution of these species to the daily intake of proteins and essential amino acids recommended by the WHO. 2. Materials and Methods 2.1. Sample Collection and Preparation A total of 230 individuals from 33 marine species, including 3 crustaceans, 1 shark, and 29 teleost fish, were collected from Seychelles waters during 2014-2016 (Table 1). Nearshore species were caught on the Mahe Plateau, where most of the artisanal fishing grounds are located , and offshore species were caught around the Mahe Plateau within the exclusive economic zone (EEZ) . After their capture, all organisms were measured (cephalothorax length (CL) for crustaceans, lower jaw-fork length (LJFL) for swordfish, and fork length (FL) and total length (TL) for other species) and weighted, and a piece of the edible part was collected from the tail for crustaceans and dorsal muscle for other species before being immediately stored at -80 degC. Samples were then freeze-dried for 72 h and ground to powder before amino acid analyses. 2.2. Amino Acid Composition and Protein Content Amino acid composition was analyzed by dissolving approximately 40 mg of dried samples in 0.7 mL of distilled H2O and 0.5 mL of 20 mM norleucine (internal standard), which was then hydrolyzed as previously described . Following hydrolysis, 100 mL aliquots of the hydrolysates were evaporated under nitrogen gas until complete dryness and re-dissolved to a suitable concentration in a lithium citrate buffer at pH 2.2. All amino acids were chromatographically analyzed using an ion exchange column followed by ninhydrin post column derivatization on a Biochrom 30 amino acid analyzer (Biochrom Co., Cambridge, UK). Amino acid residues were identified using the A9906 physiological amino acid standard (Sigma Chemical Co., St. Louis, MO, USA), as described previously . The concentrations of 20 amino acids (histidine, his; isoleucine, ile; leucine, leu; lysine, lys; methionine, met; phenylalaline, phe; threonine, thr; valine, val; alanine, ala; b-alanine; b-ala; arginine, arg; asparagine, asn; aspartic acid, asp; cysteine, cys; glutamine, gln; glutamic acid, glu; glycine, gly; hydroxyproline, hyp; proline, pro; serine, ser; tyrosine, tyr; taurine, tau) were converted from dry weight to wet weight by using a mean moisture percentage of 72-81%, depending on species, and expressed in mg per 100 g of raw edible portion (noted mg/100 g). Tryptophan is denatured during acid hydrolysis, while glutamine and asparagine deaminate during acid hydrolysis and were therefore included in the category of glutamate and aspartic acid. Protein content (g/100 g) was determined as the sum of the individual amino acid residues (the molecular weight of each amino acid after the subtraction of the molecular weight of H2O), as recommended by the FAO , using norleucine as internal standard. 2.3. Statistical Analyses Statistical analyses were performed using IBM SPSS statistics 27. All samples were measured in duplicate, and the number of individuals analyzed from each marine species is presented in Table 1. 3. Results 3.1. Protein Content The protein content, calculated as the sum of amino acids minus the molecular weight of water, was relatively constant among all fish species , varying between 13 and 17 g/100 g. Crustaceans had a lower protein content of approximately 11-12 g/100 g. The total amount of essential amino acids (EAAs) constituted half of the protein content for all species . 3.2. Distribution of Essential Amino Acids The distribution of EAAs was similar for all fish species , with leucine and the commonly limiting amino acid lysine being the most abundant amino acids (1500-1700 mg/100 g and 1800-2000 mg/100 g, respectively). These two amino acids were also the most abundant in crustaceans, although their contributions were slightly lower than in fish. The concentrations of histidine were highly variable among the 33 studied species (from 250 to 1350 mg/100 g), with the highest relative content being measured in tunas and mackerels. The contents of threonine (776-1093 mg/100 g), valine (867-1091 mg/100 g), methionine (450-697 mg/100 g), isoleucine (799-1071 mg/100 g), and phenylalanine (650-908 mg/100 g) were higher in the fish species compared with the crustaceans (on average 569, 646, 402, 623, and 578 mg/100 g, respectively). 3.3. Taurine Concentration The concentration of taurine considerably varied within and among the studied species . Skipjack tuna and common dolphinfish were lowest in taurine (<20 mg/100 g), while humpback red snapper and peacock hind were highest in taurine (440 mg/100 g). 4. Discussion 4.1. Protein Content The protein content was calculated based on the amount of total amino acids minus the molecular weight of water, as recommended by the FAO . This procedure efficiently hydrolyzes most of the peptide bonds while also reducing some amino acids. Tryptophan is denatured during acid hydrolysis, while glutamine and asparagine deaminate during acid hydrolysis and were therefore included in the category of glutamate and aspartic acid . This may have resulted in a potential underestimation of the actual protein content and a lower protein content compared with that measured with the commonly used Kjeldahl s method . The protein content was similarly high in all marine species, with the exceptions of spanner crab and lobsters that showed slightly lower protein contents. 4.2. Contribution to Daily Recommended Intake The Codex nutritional reference values for protein are based on the best available scientific knowledge of the daily amount needed for good health (830 mg protein/kg body weight for adults and 910 mg kg body weight for children). Based on these reference values and considering a portion size of 150 g for adults and 75 g for children, the contributions of one portion of each capture fishery species to the recommended dietary intake (RDI) for a 65 kg adult person, a 65 kg pregnant woman in the third trimester, and a 10-year-old child (average body weight of 30 kg) were estimated . One portion of swordfish and crustaceans (spiny lobster and spanner crab) would cover 30% of the adult, pregnant woman, and child RDIs. All other species would contribute 40-45% of these RDIs. The FAO and WHO have recommended the dietary intake of each of the indispensable amino acids, based on growth and nitrogen balance. The percentage coverage of each of these amino acids for a 65 kg person by a 150 g portion of different species is illustrated in Figure 6. One portion of crustacean from the Palinuridae (spiny lobster) and Raninidae (spanner crab) families covered 50% of phenylalanine; 60-67% of valine, leucine, isoleucine, and histidine; and 90% of threonine, methionine, and lysine. One portion of fish covered approximately 70% of the daily recommended amount of phenylalanine and around or above 100% of the daily recommended amount of the other indispensable amino acids. 4.3. Protein Quality In addition to being building blocks for protein synthesis, each amino acid has its own metabolic pathway. The 20 proteogenic amino acids are classified as non-essential or essential. The nine essential amino acids (threonine, valine, methionine, isoleucine, leucine, phenylalanine, lysine, histidine, and tryptophan) cannot be synthesized in the human body from naturally occurring precursors at a rate needed to meet the metabolic requirements. In this work, protein quality was determined based on the amount of essential amino acids. All analyzed fish species were high in lysine and threonine, which are strictly indispensable . One portion of crustacean meat or fish filet was found to significantly contribute to the daily requirements of the indispensable amino acids, and one portion of fish filet was found to meet the requirements of threonine, methionine, and lysine. However, it is important to mention that all samples were analyzed raw, and several factors may influence amino acid contents during processing and household preparations such as boiling, baking, frying, and smoking , which may affect the amino acid contribution to the diet. These values thus indicate the amount available in pre-processed food and not the exact actually absorbed amount. The chemical score of amino acids used to assess the amount of limiting amino acids, can be used to determine if a diet meets the required amount of indispensable amino acids. The chemical score equals the ratio between each indispensable amino acid in the food protein and the corresponding amino acid in a reference protein proposed by the FAO/WHO. The protein-digestibility-corrected amino acid score can thereafter be calculated as the amino acid score multiplied by the true digestibility in humans . Proteins of animal source normally have a chemical score of 1.0, while the scores of cereal proteins normally range from 0.4 to 0.6. All species analyzed in this work had a high protein quality, with the contents of all indispensable amino acids above the reference-scoring pattern for adults indicating that the protein quality was superior. As tryptophan was denatured during the acid hydrolysis of the samples, it was not possible to assess whether it is a limiting amino acid. 4.4. Taurine Accumulating evidence supports the idea that an increased dietary intake of taurine, a naturally occurring sulfonic acid, may be beneficial, as it has been documented to attenuate hypertension, suppress atherosclerosis, and exhibit antioxidative and anti-inflammatory properties . Fish is recognized as a rich source of taurine , and urinary taurine may be used as a marker of the level of fish consumption . In this study, taurine content greatly varied not only between species but also within different specimens of a given species. As a free amino acid, taurine is easily lost in handling and preparation, and its content is often found to significantly vary, even within one fillet . A stricter control of all parts of the value chain would be necessary to avoid such variation. The levels of taurine measured in this study were within normal ranges compared to seafood in general, but as this is the first report of the levels of taurine for many of these species, comparison is challenging. The highest content was analyzed in the demersal species, humpback red snapper and tomato hind. 4.5. Regional Food and Nutrition Security Food traditions are important. Food provides nutrients, and changes in lifestyles include nutrition transitions; the decreasing consumption of local foods is often associated with an increase in the consumption of carbohydrate-dense and highly processed foods. Such foods are normally cheaper and excessive in sugar, fat and additives. The high intake of refined food products has led to a worldwide elevated burden of overweight and obesity , and Seychellois are not an exception . Malnutrition, excessive caloric consumption, and coexisting micronutrient deficiencies combined with declining activity levels may imply increases in and the earlier onset of lifestyle diseases, and global food systems may be leading to the poorer health of many . 5. Conclusions This study provides detailed information on the concentrations of essential and non-essential amino acids and the protein content and quality of a wide range of tropical capture fishery species from the Seychelles (Western Indian Ocean) caught in both nearshore and offshore waters. The species' contributions to the recommended daily intake values of indispensable amino acids from the WHO were assessed, and implications for regional food and nutrition security was discussed. The captured fish species analyzed in this work had high contents of high-quality protein, with all indispensable amino acids above the reference value pattern for adults and children. Such species with high protein contents of superior quality are perceived as healthy foods. As fish makes up as much as 48% of the consumed animal protein in the Seychelles, it is of particular importance as a source of essential amino acids and associated nutrients such as fatty acids, taurine, vitamins and minerals. Accordingly, every effort to sustain the consumption of regional fish should be encouraged. Acknowledgments The authors would like to thank all fishermen and crews who assisted with sampling. A special thank goes to the SFA staff (in alphabetic order: Clara Belmont, Dora Lesperance, Kettyna Gabriel, Maria Rose, Natifa Pillay, Rodney Melanie, Rona Arrisol, and Stephanie Hollanda) for their help in processing the samples and to Emmanuel Chassot (IOTC) for assisting with data management. The authors are also grateful to chief engineer Guro K. Edvinsen and senior engineer Hanne K. Maehre for analytical contributions. Author Contributions Conceptualization, I.-J.J., N.B. and E.O.E.; methodology, I.-J.J., N.B. and R.G.; software, I.-J.J. and N.B.; validation, I.-J.J., E.O.E., N.B. and R.G.; writing--original draft preparation, I.-J.J.; writing--review and editing, E.O.E., I.-J.J., N.B. and R.G.; funding acquisition, I.-J.J., E.O.E., N.B. and R.G. All authors have read and agreed to the published version of the manuscript. Data Availability Statement The data are available on request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Fishing locations of the 33 marine species on and around the Mahe Plateau, the Seychelles (Western Indian Ocean). Figure 2 Protein and sum of essential amino acid (threonine, valine, methionine, isoleucine, leucine, phenylalanine, lysine, and histidine; tryptophan is denatured during acid hydrolysis and is thus not included) content (g/100 g) in species caught in the Seychelles waters. Figure 3 Distribution of essential amino acids (mg/100 g) in species caught in the Seychelles waters. Thr, threonine; Val, valine; Met, methionine; Ile, isoleucine; Leu, leucine; Phe, phenylalanine; Lys, lysine; His, histidine (tryptophan is denatured during acid hydrolysis and is thus not included). Figure 4 Taurine concentration (mg/100 g) in species caught in the Seychelles waters. Mean values are represented by an x, median values are represented by -, the boxes cover the interquartile range, and the whiskers represent the minimum and maximum values without outliers. Outliers are plotted as individual points. Figure 5 Percent of recommended dietary intake (RDI) of protein for an adult (65 kg), a pregnant woman in the 3rd trimester (65 kg), or for a 10-year-old child (30 kg), that is covered by one portion of 150 or 75 g, respectively. Recommended dietary intake values from Codex and the WHO are used. Figure 6 Percent of recommended dietary intake of the essential amino acids for a 65 kg person by a portion of 150 g. Recommended dietary allowance values from the WHO are used. Thr, threonine; Val, valine; Met, methionine; Ile, isoleucine; Leu, leucine; Phe, phenylalanine; Lys, lysine; His, histidine; (tryptophan is denatured during acid hydrolysis and is thus not included). foods-12-01015-t001_Table 1 Table 1 Marine species collected from the Seychelles waters, with associated details. Length (presented as mean +- SD) refers to the mean carapace length for crustaceans, the mean lower jaw-fork length for swordfish, and the mean fork length for other teleost fish and for sharks. N = number of individuals. Group Family Scientific Name English Name FAO Code Habitat Fishing Area Fishing Year N Length Weight Crustacean Palinuridae Panulirus longipes Longlegged spiny lobster LOJ reef-associated Nearshore 2014 5 8.2 +- 0.7 0.6 +- 0.1 Panulirus penicillatus Pronghorn spiny lobster NUP reef-associated Nearshore 2014 4 9.3 +- 1.3 0.7 +- 0.2 Raninidae Ranina ranina Spanner crab RAQ reef-associated Nearshore 2014 5 9.7 +- 1.1 0.3 +- 0.1 Shark Carcharhinidae Carcharhinus falciformis Silky shark FAL reef-associated Nearshore 2015 5 79.8 +- 11.4 NA Teleost fish Siganidae Siganus argenteus Streamlined spinefoot IGA reef-associated Nearshore 2015 4 26.0 +- 1.1 0.3 +- 0.0 Scaridae Scarus ghobban Blue-barred parrotfish USY reef-associated Nearshore 2015 4 29.8 +- 3.5 0.5 +- 0.2 Acanthuridae Acanthurus mata Elongate surgeonfish DGW reef-associated Nearshore 2016 4 46.8 +- 1.5 2.1 +- 0.1 Balistidae Canthidermis maculata Rough triggerfish CNT reef-associated Nearshore 2015 5 31.8 +- 3.9 0.7 +- 0.2 Carangidae Gnathanodon speciosus Golden trevally GLT reef-associated Nearshore 2016 2 66.0 +- 19.8 5.8 +- 4.6 Carangoides malabaricus Malabar trevally NGS reef-associated Nearshore 2016 2 67.5 +- 6.4 4.6 +- 0.4 Carangoides fulvoguttatus Yellowspotted trevally NGU reef-associated Nearshore 2016 9 52.4 +- 9.0 2.6 +- 1.2 Elagatis bipinnulata Rainbow runner RRU reef-associated Nearshore 2015 5 75.1 +- 15.5 NA Lethrinidae Lethrinus crocineus Yellowtail emperor ICZ reef-associated Nearshore 2016 5 36.0 +- 5.1 1.0 +- 0.5 Lethrinus microdon Smalltooth emperor LEN reef-associated Nearshore 2016 4 45.8 +- 2.2 1.5 +- 0.2 Lethrinus nebulosus Spangled emperor LHN reef-associated Nearshore 2016 1 41 1.2 Lethrinus variegatus Slender emperor LHV reef-associated Nearshore 2016 10 29.0 +- 1.3 0.4 +- 0.1 Lethrinus mahsena Sky emperor LTQ reef-associated Nearshore 2016 10 33.7 +- 3.2 0.9 +- 0.2 Lutjanidae Aprion virescens Green jobfish AVR reef-associated Nearshore 2015 14 52.2 +- 3.3 2.0 +- 0.4 Etelis coruscans Deepwater longtail red snapper ETC reef-associated Nearshore 2014 10 57.3 +- 11.3 2.8 +- 1.4 Lutjanus bohar Two-spot red snapper LJB reef-associated Nearshore 2015 15 58.6 +- 15.6 4.4 +- 2.7 Lutjanus gibbus Humpback red snapper LJG reef-associated Nearshore 2016 9 36.0 +- 3.6 1.0 +- 0.3 Lutjanus sebae Emperor red snapper LUB reef-associated Nearshore 2014 10 56.8 +- 6.9 4.2 +- 1.5 Serranidae Cephalopholis argus Peacock hind CFF reef-associated Nearshore 2014 8 30.5 +- 3.3 0.5 +- 0.2 Epinephelus fasciatus Blacktip grouper EEA reef-associated Nearshore 2016 3 NA 0.4 +- 0.1 Epinephelus chlorostigma Brownspotted grouper EFH reef-associated Nearshore 2014 9 37.2 +- 2.0 0.7 +- 0.1 Cephalopholis sonnerati Tomato hind EFT reef-associated Nearshore 2014 9 40.2 +- 4.4 1.1 +- 0.4 Epinephelus multinotatus White-blotched grouper EWU reef-associated Nearshore 2014 10 64.0 +- 5.6 4.2 +- 1.1 Sphyraenidae Sphyraena barracuda Great barracuda GBA reef-associated Nearshore 2014 5 106.2 +- 9.0 7.6 +- 2.7 Scombridae Gymnosarda unicolor Dogtooth tuna DOT reef-associated Offshore 2016 1 93 12.2 Rastrelliger kanagurta Indian mackerel RAG pelagic-neritic Offshore 2014 10 25.9 +- 0.5 0.3 +- 0.0 Katsuwonus pelamis Skipjack tuna SKJ pelagic-oceanic Offshore 2014 3 45.4 +- 3.5 2.0 +- 0.4 Coryphaenidae Coryphaena hippurus Common dolphinfish DOL pelagic-neritic Offshore 2014 10 99.1 +- 7.2 6.4 +- 1.7 Xiphiidae Xiphias gladius Swordfish SWO pelagic-oceanic Offshore 2014 20 158.9 +- 32.1 NA Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. UN Transforming Our World: The 2030 Agenda for Sustainable Development UN New York, NY, USA 2015 10.2903/j.efsa.2012.2557 2. Hilborn R. Banobi J. Hall S.J. Pucylowski T. Walsworth T.E. The environmental cost of animal source foods Front. Ecol. Environ. 2018 16 329 335 10.1002/fee.1822 3. Hoegh-Guldberg O. Caldeira K. Chopin T. Gaines S. Haugan P. Hemer M. Howard J. Konar M. Krause-Jensen D. Lindstad E. The ocean as a solution to climate change: Five Opportunities for Action Available online: (accessed on 8 May 2021) 4. FAO The State of World Fisheries and Aquaculture 2022. Towards Blue Transformation FAO Rome, Italy 2022 266p 5. Thomsen S.T. Assuncao R. Afonso C. Boue G. Cardoso C. Cubadda F. Garre A. Kruisselbrink J.W. Mantovani A. Pitter J.G. Human health risk--benefit assessment of fish andother seafood: A scoping review Food Sci. Nutr. 2022 62 7479 7502 10.1080/10408398.2021.1915240 6. Bodin N. Lesperance D. Albert R. Hollanda S. Michaud P. Degroote M. Churlaud C. Bustamante P. Trace elements in oceanic pelagic communitites in the western Indian Ocean Chemosphere 2017 174 354 362 10.1016/j.chemosphere.2017.01.099 28183061 7. Robinson J.P.W. Maire E. Bodin N. Hempson T.N. Graham N.A.J. Wilson S.K. MacNeil M.A. Hicks C.C. Climate-induced increases in micronutrient availability for coral reef fisheries One Earth 2022 5 98 108 10.1016/j.oneear.2021.12.005 35128392 8. Larsen R. Eilertsen K.E. Elvevoll E.O. Health benefits of marine foods and ingredients Biotechnol. Adv. 2011 29 508 518 10.1016/j.biotechadv.2011.05.017 21683131 9. Jensen I.J. Walquist M. Liaset B. Elvevoll E.O. Eilertsen K.E. Dietary intake of cod and scallop reduces atherosclerotic burden in female apolipoprotein E-deficient mice fed a Western-type high fat diet for 13 weeks Nutr. Metab. 2016 13 8 10.1186/s12986-016-0068-z 10. Spitze A.R. Wong D.L. Rogers Q.R. Fascetti A.J. Taurine concentrations in animal feed ingredients; cooking influences taurine content J. Anim. Physiol. Anim. Nutr. 2003 87 251 262 10.1046/j.1439-0396.2003.00434.x 11. Elvevoll E.O. Eilertsen K.E. Brox J. Dragnes B.T. Falkenberg P. Olsen J.O. Kirkhus B. Lamglait A. Osterud B. Seafood diets: Hypolipidemic and antiatherogenic effects of taurine and n-3 fatty acids Atherosclerosis 2008 200 396 402 10.1016/j.atherosclerosis.2007.12.021 18242615 12. Yamori Y. Liu L. Ikeda K. Miura A. Mizushima S. Miki T. Nara Y. Disease W.H.-C. Alimentary Comprarison Study G. Distribution of twenty-four hour urinary taurine excretion and association with ischemic heart disease mortality in 24 populations of 16 countries: Results from the WHO-CARDIAC study Hypertens. Res. 2001 24 453 457 10.1291/hypres.24.453 11510759 13. Sabino M.A. Bodin N. Govinden R. Arrisol R. Churlaud C. Pethybridge H. Bustamante P. The role of tropical small-scale fisheries in trace element delivery for a Small Island Developing State community, the Seychelles Mar. Pollut. Bull. 2022 181 113870 10.1016/j.marpolbul.2022.113870 35835052 14. Sardenne F. Bodin N. Medieu A. Antha M. Arrisol R. Le Grand F. Bideau A. Munaron J.-M. Le Loc'h F. Chassot E. Benefit-risk associated with the consumption of fish bycatch from tropical tuna fisheries Environ. Pollut. 2020 267 115614 10.1016/j.envpol.2020.115614 33254649 15. EFSA Scientific Opinion on Health Benefits of Seafood (Fish and Shellfish) Consumption in Relation to Health Risks Associated with Exposure to Methylmercury EFSA Parma, Italy 2014 16. FAO WHO Report of the Joint FAO/WHO Expert Consultation on the Risks and Benefits of Fish Consumption FAO Rome, Italy 2011 17. VKM Benefit and Risk Assessment of Fish in the Norwegian Diet. Scientific Opinion of the Scientific Steering Committee of the Norwegian Scientific Committee for Food and Environment VKM Oslo, Norway 2022 18. Munschy C. Vigneau E. Bely N. Heas-Moisan K. Olivier N. Pollono C. Hollanda S. Bodin N. Legacy and emerging organic contaminants: Levels and profiles in top predator fish from the western Indian Ocean in relation to their trophic ecology Environ. Res. 2020 188 109761 10.1016/j.envres.2020.109761 32562947 19. Hibbeln J.R. Spiller P. Brenna J.T. Golding J. Holub B.J. Harris W.S. Kris-Etherton P. Lands B. Connor S.L. Myers G. Relationships between seafood consumption during pregnancy and childhood and neurocognitive development: Two systematic reviews Prostaglandins Leukot Essent Fat. Acids 2019 151 14 36 10.1016/j.plefa.2019.10.002 20. Davidson P.W. van Wijngaarden E. Shamlaye C. Strain J.J. Myers G.J. Putting findings from the Seychelles Child Development Study into perspective: The importance of a historical special issue of the Seychelles Medical and Dental Journal Neurotoxicology 2020 76 111 113 10.1016/j.neuro.2019.10.010 31706981 21. Strain J.J. Love T.M. Yeates A.J. Weller D. Mulhern M.S. McSorley E.M. Thurston S.W. Watson G.E. Mruzek D. Broberg K. Associations of prenatal methylmercury exposure and maternal polyunsaturated fatty acid status with neurodevelopmental outcomes at 7 years of age: Results from the Seychelles Child Development Study Nutrition Cohort 2 Am. J. Clin. Nutr. 2021 113 304 313 10.1093/ajcn/nqaa338 33330939 22. Strain J.J. Yeates A.J. van Wijngaarden E. Thurston S.W. Mulhern M.S. McSorley E.M. Watson G.E. Love T.M. Smith T.H. Yost K. Prenatal exposure to methyl mercury from fish consumption and polyunsaturated fatty acids: Associations with child development at 20 mo of age in an observational study in the Republic of Seychelles Am. J. Clin. Nutr. 2015 101 530 537 10.3945/ajcn.114.100503 25733638 23. FAO Fishery and Aquaculture Country Profiles The Republic of Seychelles. Available online: (accessed on 20 October 2022) 24. Seychelles Fisheries Sector Policy and Strategy 2019 Available online: (accessed on 20 October 2022) 25. Strain J.J. Eating fish for two Nutr. Bull. 2014 39 181 186 10.1111/nbu.12088 25132804 26. Cardoso I. Bovet P. Viswanathan B. Luke A. Marques-Vidal P. Nutrition transition in a middle-income country: 22-year trends in the Seychelles Eur. J. Clin. Nutr. 2012 67 135 140 10.1038/ejcn.2012.199 23249880 27. Bovet P. Chiolero A. Shamlaye C. Paccaud F. Prevalence of overweight in the Seychelles: 15 year trends and association with socio-economic status Obes. Rev. 2008 9 511 517 10.1111/j.1467-789X.2008.00513.x 18673305 28. WHO Protein and Amino Acid Requirements in Human Nutrition. Report of a Joint WHO/FAO/UNU Expert Consultation WHO Geneva, Switzerland 2007 29. Lewis J. Codex Nutrient Reference Values FAO Rome, Italy 2019 30. Khalili Tilami S. Sampels S. Nutritional value of fish: Lipids, proteins, vitamins, and minerals Rev. Fish. Sci. Aquac. 2017 26 243 253 10.1080/23308249.2017.1399104 31. Greenfield H. Southate D. Food Composition Data: Production, Management, and Use FAO Rome, Italy 2003 32. Robinson J.P.W. Robinson J. Gerry C. Govinden R. Freshwater C. Graham N.A.J. Diversification insulates fisher catch and revenue in heavily exploited tropical fisheries Sci. Adv. 2020 6 eaaz0587 10.1126/sciadv.aaz0587 32128420 33. Maehre H.K. Dalheim L. Edvinsen G.K. Elvevoll E.O. Jensen I.J. Protein Determination-Method Matters Foods 2018 7 5 10.3390/foods7010005 29301260 34. Moore S. Stein W.H. Chromatographic determination of amino acids by the use of automatic recording equipment Meth. Enzymol. 1963 6 819 831 10.1016/0076-6879(63)06257-1 35. Maehre H.K. Hamre K. Elvevoll E.O. Nutrient evaluation of rotifers and zooplankton: Feed for marine fish larvae Aquacult. Nutr. 2013 19 301 311 10.1111/j.1365-2095.2012.00960.x 36. Maclean W.C. Harnly J.M. Chen J. Chevassus-Agnes S. Gilani G. Livesey G. Mathioudakis B. Munoz De Chavez M. Devasconcellos M.T. Warwick P. Food Energy--Methods of Analysis and Conversion Factors Food and Agriculture Organization of the United Nations Technical Workshop Report FAO Rome, Italy 2003 37. Pickering M. Newton P. Amino acid hydrolysis: Old problems, new solutions LC GC Mag. Sep. Sci. 1990 8 778 781 38. Larsen R. Stormo S.K. Dragnes B.T. Elvevoll E.O. Losses of taurine, creatine, glycine and alanine from cod (Gadus morhua L.) fillet during processing J. Food Comp. Anal. 2007 20 396 402 10.1016/j.jfca.2006.10.001 39. Larsen R. Mierke-Klemeyer S. Maehre H. Elvevoll E.O. Retention of health beneficial components during cold-smoking of African catfish (Ciarias gariepinus) fillets Arch. Fur Lebensm. 2010 61 31 35 10.2376/0003-925X-61-31 40. FAO Protein Quality Evaluation in Human Diets. Report of a Joint FAO/WHO Expert Consultation FAO Rome, Italy 1991 41. Jong C.J. Sandal P. Schaffer S.W. The role of taurine in mitochnodria health: More than just an antioxidant Molecules 2021 26 4913 10.3390/molecules26164913 34443494 42. Schaffer S.W. Ito T. Azyma J. Clinical significance of taurine Amino Acids 2013 46 1 5 10.1007/s00726-013-1632-8 43. Yang J.-Y. Zhang T.-T. Yu Z.-L. Wang C.-C. Zhao Y.-C. Wang Y.-M. Xue C.-H. Taurine alleviates trimethylamine N-oxide-induced atherosclerosis by regulating bile acid metabolism in ApoE-/- mice J Agric. Food Chem. 2022 70 5738 5747 10.1021/acs.jafc.2c01376 35486890 44. Gormley T.R. Neumann T. Fagan J.D. Brunton N.P. Taurine content of raw and processed fish fillets/portions Eur. Food Res. Technol. 2007 225 837 842 10.1007/s00217-006-0489-4 45. Gibson R. Lau C.-H.E. Loo R.L. Ebbels T.M.D. Chekmeneva E. Dyer A.R. Miura K. Ueshima H. Zhao L. Daviglus M.L. The association of fish consumption and its urinary metabolites with cardiovascular risk factors: The International Study of Macro-/Micronutrients and Blood Pressure (INTERMAP) Am. J. Clin. Nutr. 2020 111 280 290 10.1093/ajcn/nqz293 31782492 46. Dragnes B.T. Larsen R. Erntsen M.H. Maehre H. Elvevoll E.O. Impact of processing on the taurine content in processed seafood and their corresponding unprocessed raw materials Int. J. Food Sci. Nutr. 2009 60 143 152 10.1080/09637480701621654 18608559 47. Vasquez F. Vita G. Muller D.B. Food Security for an Aging and Heavier Population Sustainability 2018 10 3683 10.3390/su10103683 48. Nash K.L. MacNeil M.A. Blanchard J.L. Cohen P.J. Farmery A.K. Graham N.A.J. Thorne-Lyman A.L. Watsen R.A. Hicks C.C. Trade and foreign fishing mediate global marine nutrient supply Sustain. Sci. 2022 119 e2120817119 10.1073/pnas.2120817119 49. Fanzo J. Drewnowski A. Blumberg J. Miller G. Kraemer K. Kennedy E. Nutrients, foods, diets, people: Promoting healthy eating Curr. Dev. Nutr. 2020 4 nzaa069 10.1093/cdn/nzaa069 32494761 50. Farmery A.K. Alexander K. Anderson K. Blanchard J.L. Carter C.G. Evans K. Fischer M. Fleming A. Frusher S. Fulton E.A. Food for all: Designing sustainable and secure future seafood systems Rev. Fish Biol. Fish 2022 32 101 121 10.1007/s11160-021-09663-x 34092936 |
PMC10000636 | Introduction Diabetes mellitus (DM) has become an important and exacerbating health epidemic, with severe consequences for both patients and health systems. The National Diabetes Strategy of Austria addresses the lack of high-quality data on DM in Austria and the need for developing a national data network. The aims of our study are to establish a cohort including all adult diabetes patients in a district in western Austria, describe the demographic and clinical characteristics of this cohort, and provide an estimation of diabetes prevalence. Methods We recruited a population-based cohort of adult patients with a diagnosis of DM in cooperation with a network of all caregivers. Data collection was based on a case report form, including patient characteristics, clinical parameters and long-term complications. Results In total, 1845 patients with DM were recruited and analysed. We observed an overall prevalence of 5.3% [95% CI: 5.0%-5.5%]. For the subsequent main analysis, we included 1755 patients with DM after excluding 90 patients with gestational DM. There were significant differences between genders in the distribution of specific clinical parameters, patient characteristics, and the long-term complications diabetic foot, amputation and cardiovascular disease. Conclusion To the best of our knowledge, we established the first diabetes cohort study in Austria. Prevalence and the proportion of specific long-term complications were lower when compared to the international context. We assume that the Diabetes Landeck Cohort has reached a high degree of completeness; however, we were not able to identify independent data sources for a valid check of completeness. The aims of this project and study are to establish a cohort including all adult DM patients in a district in western Austria, describe the demographic and clinical characteristics of these patients, and provide an estimation of diabetes prevalence, both at an overall level and stratified by gender and age groups. diabetes mellitus gender prevalence Tyrolean Health Fund (Tiroler Gesundheitsfond, TGF) source-schema-version-number2.0 cover-dateMarch 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.2.6 mode:remove_FC converted:10.03.2023 Haslwanter V , Rochau U , Siebert U , Schonherr H-R , Oberaigner W . A population-based cohort of adult patients with diabetes mellitus in a Western District of Austria: The Diabetes Landeck cohort. Endocrinol Diab Metab. 2023;6 :e395. doi:10.1002/edm2.395 pmc1 INTRODUCTION Diabetes mellitus (DM) has become a considerable and exacerbating health epidemic that has severe consequences for both the patients and health systems. 1 According to the tenth edition of the International Diabetes Federation (IDF) Atlas, approximately 537 million people aged 20 to 79 years suffered from DM worldwide in 2021, representing a prevalence of approximately 10.5%. 2 The global prevalence is expected to increase in the future, with 783 million people predicted to suffer from DM by 2045, corresponding to a prevalence of around 10.9%. 2 In Europe, one in eleven adults has DM, which means that 61 million people suffer from DM, and the prevalence is estimated to increase from 9.2% in 2021 to 10.4% in 2045. Furthermore, the economic burden associated with DM is substantial: The IDF reports that the share of the global health expenditure for DM ranges from 3.4% in the Middle East and North African region to 19.6% in Europe and to 43% in North America and the Caribbean. 2 Patients with DM suffer from a lifelong burden caused by the treatment and disease complications. One of the main challenges is an increased risk of developing long-term complications such as neuropathy, nephropathy and cardiovascular disease. 3 , 4 In addition, patients with DM have an increased risk of early mortality. Worldwide, about 6.7 million people aged between 20 and 79 years are estimated to die from DM or its complications in 2021. 2 For Austria, the number of patients with DM aged 20-79 years in 2021 was estimated at about 450,000, corresponding to a prevalence of 6.6%. In addition, the number of undiagnosed DM cases was estimated at approximately 150,000. 2 Due to the increasing number of patients with DM and the long-term complications, considerable DM-associated healthcare expenditures are also expected in Austria. The annual costs of DM and its comorbidities in Austria amount to an estimated 3 billion euros. 5 Therefore, the main targets of the diabetes strategy of the Austrian Federal Ministry of Labor, Social Affairs, Health and Consumer Protection from 2017 are to reduce the incidence of DM and to prevent long-term complications. 6 This Austrian strategy document states that there is a lack and necessity of valid data on DM in Austria. Moreover, a report of the Austrian Court of Audition on DM prevention and DM care from 2019 points to the lack of high-quality data. 7 To provide valid data on DM in Austria, a project called the Diabetes Landeck Cohort was initiated, with the aim to set up a population-based cohort of patients with DM in a district in western Austria. This cohort should serve as a basis to collect and analyse valid and comprehensive data on DM. The aims of this project and study are to establish a cohort including all adult DM patients in a district in western Austria, describe the demographic and clinical characteristics of these patients, and provide an estimation of diabetes prevalence, both at an overall level and stratified by gender and age groups. 2 SUBJECTS AND METHODS 2.1 Settings of the study and data collection The recruitment of the cohort titled 'Diabetes Landeck Cohort' started in 2018 and ended in June 2021. Eligible participants were adults (aged >=20 years) with type 1 DM (T1DM) or type 2 DM (T2DM) or other types of diabetes (e.g. gestational, latent autoimmune diabetes in adults) and with the main residence in the district of Landeck in the western part of Austria. The district of Landeck is a well-defined region with clear geographical borders, either high mountains within Austria or national borders towards Italy and Switzerland in the south. The district of Landeck with a population of 35,148 persons aged 20 and older consists of the city of Landeck with a population of 6120 persons aged 20 and older and 30 municipalities. This rural area is rather typical for rural areas in Austria. The study was conducted by the Research Unit of Diabetes Epidemiology of the Department of Public Health, Health Services Research and Health Technology Assessment at UMIT TIROL. Before beginning data collection, a case report form (CRF) was designed by diabetes and public health experts. The CRF includes questions on patient characteristics, and questions on time-varying data based on visits at care units. A network connecting all care units has been established including 22 general practitioners, four diabetes specialists in private practice, the hospital in Zams, the Medical University Innsbruck and retirement homes in the district of Landeck. Data collection was mainly performed by two qualified study nurses who visited the care units. Direct documentation by the treating physician was also technically possible but conducted only in individual cases due to the heavy workload of the physicians caused by the COVID-19 pandemic. All cohort data are stored in the pseudonymized web-based database software ASKIMED. 8 ASKIMED provides the possibility to store visits for each patient at different care units and allows the documentation by different persons and the mapping of the necessary access rights. A pseudonym was generated based on each patient's social security number (with a SHA-2 procedure), which makes it impossible, due to the current computing performance, to identify the specific patient based on the pseudonym. For research purposes, data were transferred to the statistical software Stata (Version 17). 9 In order to address aspects of data privacy, all patients had to sign a written consent form. For data protection reasons, the data of the very few patients who did not agree to sign the consent have not been included into the database for analysis. In addition, every care unit signed a contract stating the rights and obligations between the care units and the research group. The ethics committee at the Medical University of Innsbruck approved the complete project and the project was carried out in accordance with the Helsinki Declaration of 1975, as revised in 2008. A scientific advisory board supervised the study and decided in favour of a minimal data set to build up a population-based cohort, given restrictions by the limited budget and the COVID-19 pandemic. 2.2 Patient characteristics We collected the following patient characteristics: diagnosis, age, migration background, diagnosis site, year of diagnosis, diabetes duration, family history of coronary heart disease and family history of DM, participation in the Austrian disease management program, smoking status, participation in a diabetes education program and sufficient knowledge on diabetes according to the physician. All diagnoses were clinically confirmed according to the criteria of the Austrian Diabetes Society (ADS) based on a fasting glucose or oral glucose tolerance test or haemoglobin A1c (HbA1c). 10 The study protocol did not include patients with prediabetes in the study population. Gestational diabetes (GDM) was documented separately. Diagnosis of GDM was based on the HAPO criteria. 11 Age at the last visit was used for the analysis and was cut into categories, namely 20-49, 50-64, 65-74 and >=75 years, taking into account the definition of geriatrics of the elderly/old person. The migration background was classified based on Schenk's approach and adapted to the Austrian situation. 12 During documentation, participants were asked whether their DM diagnosis had been made in the hospital or at a physician's office (diagnosis site). For most patients, the year of diagnosis was recalled from patients' memory, and we decided to collect the exact year only for patients diagnosed during the past 10 years. The diabetes duration was defined as the difference between the year of DM diagnosis and the year of last contact with a care unit and was analysed in the following groups: 0-4 years, 5-9 years, and 10 years or longer. We documented family history of coronary heart disease and family history of DM. Family included children, parents and/or siblings. Participation in the Austrian Disease Management Program during the study period 13 was also documented. Additionally, the smoking status at the time of diagnosis was recorded retrospectively in the categories 'active smoker' or 'ex-smoker' and 'never smoker'. We assessed whether each patient had participated at least once in a diabetes education program and whether the patient had sufficient knowledge on DM according to the physician. Each patient's life status was verified by a study nurse who inspected hospital/private care records and local newspapers in case there were no up-to-date visit records. Pseudonymization prevented linking records with official mortality data. It was therefore impossible to systematically check the patients' life status. 2.3 Clinical parameters Data collection included the following clinical parameters: body mass index (BMI), HbA1c, low-density lipoprotein (LDL), systolic and diastolic blood pressure, microalbumin, physical activity, eye and foot inspection, hypoglycaemia (requiring external help for recovery), diabetes-specific therapy and lipid therapy. Mean values were computed for BMI, HbA1c, LDL, and systolic and diastolic blood pressure, which were collected at each visit. BMI was calculated according to the formula BMI = weight/height2 (kg/m2), and the classification was based on WHO recommendations. In addition, obesity was defined as a BMI >=30 kg/m2. 14 The mean values of HbA1c were categorized into four groups (<6.5%, 6.5%-7.49%, 7.5%-8.99% and >9%) based on the ADS guidelines. 10 Increased blood pressure was defined as systolic pressure >=140 mmHg or diastolic pressure >=90 mmHg according to WHO guidelines. 15 It should be noted that increased blood pressure is based on blood pressure measurements only. We did not collect information on the diagnosis of hypertension and/or medication. LDL is the primary therapeutic target for lipid control in patients with DM. The LDL classification was based on current recommendations of the ESC/EAS. 16 We did not collect information on the diagnosis of hyperlipidaemia, but we surveyed the proportion of patients with well-controlled LDL levels. We recorded whether microalbumin was determined at least at one visit. Patients were asked if they were physically active (defined as at least moderate activity for less than two and a half hours per week). We also documented if an ophthalmologist had performed at least one eye inspection during visits. The inspections followed the recommendations of the ADS. 10 Furthermore, we recorded foot inspections during the study period. Foot inspection was defined as at least the removal of shoes and socks and the examination of the feet by the diabetes assistant or physician. Furthermore, we noted severe hypoglycaemia (requiring external help for recovery) during the study period and calculated a cumulative number of events. Finally, we recorded information on the diabetes-specific therapy. DM treatment categories were lifestyle adaptation only, metformin only, oral antidiabetic drugs (OADs) without metformin, insulin only, insulin and oral antidiabetic drugs, or another form of therapy. Each form of therapy was documented if it was observed during at least one visit over the study period. 2.4 Long-term complications We documented the following long-term complications: neuropathy, nephropathy, retinopathy, cardiovascular and cerebrovascular disease, diabetic foot ulcers and amputations based on the recommendations of the ADS. 10 Neuropathy is defined as nerve injuries due to DM and is confirmed with a positive monofilament test. Nephropathy requires positive albumin results at two subsequent visits. Retinopathy is diagnosed according to the guidelines provided by the Austrian Ophthalmologist Society. 17 Cardiovascular late complication was defined as myocardial infarction, bypass or percutaneous coronary intervention. Cerebrovascular late complication was defined as minor and major strokes, including transient ischaemic attacks. We used the strict definition of the long-term complication diabetic foot according to the guidelines of the ADS 10 and therefore only collected information on the presence of diabetic foot ulcers. For amputations, we documented any non-traumatic amputation due to diabetic foot ulcers. For all late complications, the year of the first occurrence was recorded if the diagnosis was made within the past 10 years. In all other cases, we only documented the diagnosis of the respective late complication without the year of occurrence. For the analysis we counted every long-term complication, not only long-term complications diagnosed in the study period. 2.5 Statistical analysis Patient characteristics are described as counts and percentages for categorical data. We present all results stratified by gender and age groups (Tables S1-S2 only). Fisher's exact test or the chi-squared test were used to test differences across gender and age groups. Statistical significance was established as p < .05. Cases with GDM that did not result in a life-long type of DM (called 'GDM only') were described in a separate part because they differ in many clinical aspects. The main analysis of demographic and clinical parameters does not include cases with 'GDM only'. In case of missing data, we report data and percentages for non-missing values in a first step and present the number of cases with missing values in a second step; this procedure was adopted for every variable reported in the result tables. For the computation of prevalence figures, we included all patients with DM detected in our study in the region of Landeck. Population data for the region of Landeck per age group were obtained from Statistics Austria. 18 According to this population data, 35,148 individuals aged 20 and older were living in the district of Landeck at the time of the study, with slightly more women (17,839, 50.8%) than men (17,309, 49.2%). The population remained rather stable during the study period. Prevalence was computed for the entire district and for subregions, which are defined by the geographic structure of the following districts: Kaunertal, Stanzertal, Sonnenterasse, Oberes Gericht and Paznauntal. Prevalence and 95% confidence intervals (95% CI) are provided for the overall cohort and also stratified by gender, subregions and age groups. We applied the concept of period prevalence, that is, we included patients with at least one visit at a care unit stay during the study period as a prevalent case. All statistical analyses were performed using Stata Version 17. 9 3 RESULTS 3.1 Study population/diabetes Landeck cohort The recruitment started in 2018 and ended in June 2021. In total, we recruited and analysed 1845 cases with DM including 90 cases with 'GDM only'. For the main analysis, we excluded 'GDM only' cases resulting in a total 1755 cases with DM, 5.7% with T1DM, 92.4% with T2DM and 1.9% with another DM diagnosis. Of the 1755 patients, 25 (1.4%) died during the study and two subjects were lost in the follow-up. At the last visit, 9.6% were aged 20-49 years, 27.2% were 50-64 years old, 26.6% were aged 65-74 years, and 36.6% were 75 years or older. We observed 812 (46.3%) women and 943 (53.7%) men. Patient characteristics of the cohort according to gender are presented in Table 1. We only describe results with significant differences between female and male patients. Age distributions differed between genders. For example, in the age group >=75, the proportion of women with DM was higher than for men. Among patients with DM, women had a longer diabetes duration (55.9% >=10 years vs. 49.2% in men) and more often reported a family history of diabetes (38% in women vs. 31.7% in men). In men, we observed a substantially higher percentage of active smokers (14.1% in men vs. 7.2% in women) and ex-smokers compared to women (53.4% in men vs. 19.4% in women). TABLE 1 Patient characteristics stratified by gender (N = 1755, 'GDM only' excluded) Female Male Total p-value a N % b N % b N % b Age group 20-49 67 8.3 101 10.7 168 9.6 <.001* 50-64 170 20.9 308 32.7 478 27.2 65-74 215 26.5 251 26.6 466 26.6 >=75 360 44.3 283 30.0 643 36.6 Total 812 100.0 943 100.0 1755 100.0 Missing values c 0 0 0 0 0 0 Diagnosis T1DM 40 4.9 59 6.3 99 5.7 .112 T2DM 760 93.7 856 91.3 1616 92.4 Other DM 11 1.4 23 2.5 34 1.9 Total 811 100.0 938 100.0 1749 100.0 Missing values 1 0.1 5 0.5 6 0.3 Diagnosis site Hospital 243 32.1 325 36.8 568 34.6 .048* Private practice 514 67.9 559 63.2 1073 65.4 Total 757 100.0 884 100.0 1641 100.0 Missing values 55 6.8 59 6.3 114 6.5 Diabetes duration 0-4 years 185 22.8 258 27.4 443 25.2 .016* 5-9 years 173 21.3 221 23.4 394 22.5 >=10 years 454 55.9 464 49.2 918 52.3 Total 812 100.0 943 100.0 1755 100.0 Missing values 0 0 0 0 0 0 Family history of diabetes No 482 62.0 611 68.3 1093 65.4 .007* Yes 295 38.0 283 31.7 578 34.6 Total 777 100.0 894 100.0 1671 100.0 Missing values 35 4.3 49 5.2 84 4.8 Family history of coronary heart disease No 621 79.7 727 81.4 1348 80.6 .386 Yes 158 20.3 166 18.6 324 19.4 Total 779 100.0 893 100.0 1672 100.0 Missing values 33 4.1 50 5.3 83 4.7 Participation in disease management program No 726 89.4 843 89.4 1569 89.4 1.000 Yes 86 10.6 100 10.6 186 10.6 Total 812 100.0 943 100.0 1755 100.0 Missing values 0 0 0 0 0 0 Life status Alive 800 98.5 928 98.4 1728 98.5 .416 Deceased 12 1.5 13 1.4 25 1.4 Lost/moved 0 0.0 2 0.2 2 0.1 Total 812 100.0 943 100.0 1755 100.0 Missing values 0 0 0 0 0 0 Migration background No 705 87.3 793 84.3 1498 85.6 .087 Yes 103 12.7 148 15.7 251 14.4 Total 808 100.0 941 100.0 1749 100.0 Missing values 4 0.5 2 0.2 6 0.3 Smoking status Active smoker 55 7.2 125 14.1 180 10.9 .000* Ex-smoker 147 19.4 474 53.4 621 37.7 Never smoker 557 73.4 288 32.5 845 51.3 Total 759 100.0 887 100.0 1646 100.0 Missing values 53 6.5 56 5.9 109 6.2 Participation in education program No 168 21.0 209 22.5 377 21.8 .483 Yes 632 79.0 720 77.5 1352 78.2 Total 800 100.0 929 100.0 1729 100.0 Missing values 12 1.5 14 1.5 26 1.5 Sufficient diabetes knowledge (according to physician) No 115 15.1 109 12.2 224 13.5 .097 Yes 648 84.9 785 87.8 1433 86.5 Total 763 100.0 894 100.0 1657 100.0 Missing values 49 6.0 49 5.2 98 5.6 a Fisher's exact test or chi-squared test (non-missing values only). b All percentages are based on non-missing values (valid percentage). c Missing values are shown for each variable. * Statistical significance, p < .05. Gender differences were found for specific clinical parameters, such as BMI, HbA1c, LDL, determination of microalbumin, DM therapy and physical activity. Table 2 shows the comparison of these characteristics between genders. We observed more overweight subjects (i.e. BMI = 25.0-29.99) in men than women. The proportion of obesity (i.e. BMI>=30) was similar for both genders. Women had a significantly greater proportion of well-controlled HbA1c levels (HbA1c < 6.5) than men. The distribution of LDL showed a shift to higher values in men compared to women. Microalbumin was identified more frequently in men than in women, and more men were physically active (28.5% vs. 21.9%). We also observed differences in DM therapy. More women than men only adapted their lifestyle (22.3% vs. 16.2%). In men, metformin was more frequently used than in women, for example, more men (22.9%) received oral antidiabetic drugs therapy AND insulin than women (18.5%). Further details on clinical parameters are shown in Table 2. TABLE 2 Clinical parameters stratified by gender (N = 1755, 'GDM only' excluded) Female Male Total p-value a N % b N % b N % b BMI <18.5 9 1.1 5 0.5 14 0.8 .015* 18.5-24.99 174 22.0 165 18.0 339 19.8 25.0-29.99 275 34.7 379 41.4 654 38.3 >=30 334 42.2 367 40.1 701 41.0 Total 792 100.0 916 100.0 1708 100.0 Missing values c 20 2.5 27 2.9 47 2.7 HbA1c 0-6.49 320 41.9 325 36.2 645 38.8 .033* 6.5-7.49 254 33.3 303 33.7 557 33.5 7.5-8.99 158 20.7 235 26.2 393 23.7 9-99 31 4.1 35 3.9 66 4.0 Total 763 100.0 898 100.0 1661 100.0 Missing values 49 6.0 45 4.8 94 5.4 LDL <55 77 10.2 110 12.6 187 11.5 .017* 55-69 92 12.2 111 12.7 203 12.5 70-99 202 26.7 273 31.3 475 29.2 >=100 386 51.0 378 43.3 764 46.9 Total 757 100.0 872 100.0 1629 100.0 Missing values 55 6.8 71 7.5 126 7.2 Blood pressure (measured) Within normal range 585 74.0 670 73.0 1255 73.4 .661 Increased (>=140/90) 206 26.0 248 27.0 454 26.6 Total 791 100.0 918 100.0 1709 100.0 Missing values 21 2.6 25 2.7 46 2.6 Microalbumin No 361 47.8 357 40.7 718 43.9 .004* Yes 395 52.2 521 59.3 916 56.1 Total 756 100.0 878 100.0 1634 100.0 Missing values 56 6.9 65 6.9 121 6.9 Diabetes therapy Only metformin 313 38.9 403 43.1 716 41.1 .008* OAD without metformin 51 6.3 50 5.3 101 5.8 Only insulin 106 13.2 112 12.0 218 12.5 OAD + insulin 149 18.5 214 22.9 363 20.9 Another form of therapy 6 0.7 5 0.5 11 0.6 Only lifestyle adaptation 179 22.3 152 16.2 331 19.0 Total 804 100.0 936 100.0 1740 100.0 Missing values 8 1.0 7 0.7 15 0.9 Number of occurrences of hypoglycaemia 0 790 97.5 914 97.0 1704 97.3 .812 1 12 1.5 17 1.8 29 1.7 >=2 8 1.0 11 1.2 19 1.1 Total 810 100.0 942 100.0 1752 100.0 Missing values 2 0.2 1 0.1 3 0.2 Physical activity Inactive 568 78.1 609 71.5 1177 74.5 .003* Active 159 21.9 243 28.5 402 25.5 Total 727 100.0 852 100.0 1579 100.0 Missing values 85 10.5 91 9.7 176 10.0 Eye inspection No 303 37.4 355 37.7 658 37.6 .921 Yes 507 62.6 587 62.3 1094 62.4 Total 810 100.0 942 100.0 1752 100.0 Missing values 2 0.2 1 0.1 3 0.2 Foot inspection No 170 21.0 185 19.6 355 20.3 .512 Yes 640 79.0 757 80.4 1397 79.7 Total 810 100.0 942 100.0 1752 100.0 Missing values 2 0.2 1 0.1 3 0.2 Abbreviation: OAD, oral antidiabetic drug. a Fisher's exact test or chi-squared test (non-missing values only). b All percentages are based on non-missing values (valid percentage). c Missing values are shown for each variable. * Statistical significance, p < .05. 3.1.1 Long-term complications The most frequent long-term complication was cardiovascular disease (N = 332, 19.2%), followed by nephropathy (N = 313, 18.1%) and neuropathy (N = 223, 12.9%). We observed significant differences between women and men for diabetic foot, amputation and cardiovascular disease. More men than women suffered from these three long-term complications . Further details for the total population and genders are presented in Table 3. FIGURE 1 Long-term complications stratified by gender (N = 1728, 'GDM only' and missing values excluded) TABLE 3 Long-term complications stratified by gender (N = 1728, 'GDM only' and missing values excluded) Female Male Total p-value a N % N % N % Nephropathy 155 19.4 158 17.0 313 18.1 .211 Retinopathy 17 2.1 21 2.3 38 2.2 .871 Neuropathy 101 12.6 122 13.1 223 12.9 .774 Diabetic foot 20 2.5 49 5.3 69 4.0 .003* Amputation 5 0.6 29 3.1 34 2.0 <.001* Cardiovascular disease 98 12.3 234 25.2 332 19.2 <.001* Cerebrovascular disease b 70 8.8 70 7.5 140 8.1 .377 a Fisher's exact test or chi-squared test (non-missing values only); total missing for long-term complications: N = 27, 1.5%; women N = 12, 1.5% and men N = 15, 1.6%; multiple responses were possible. b Cerebrovascular late complication was defined as minor and major strokes, including transient ischaemic attacks. * Statistical significance, p < .05. In our study, we observed 90 cases with GDM that did not develop T2DM. A mean value of HbA1c < 6 was documented for 94% of patients with GDM. Only three patients with GDM showed a HbA1c >= 6.5. Lifestyle adaptation was sufficient in 82.5% of patients with GDM, and insulin therapy was necessary for 13.4%. See Tables S1-S2 for patient characteristics and clinical parameters according to the age groups 20-49, 50-64, 65-74 and >=75 years. Briefly, we observed differences in the following patient characteristics: diabetes duration, migration background and smoking status. We also found a difference in the participation in education programmes and sufficient diabetes knowledge, but the percentage was still very high across all age groups. For the clinical parameters stratified by age group, we observed differences in the BMI, LDL, microalbumin, diabetes therapy, physical activity, foot inspection, and for most of the long-term complications. Further details on patient characteristics and clinical parameters are described in Tables S1-S2. 3.2 Prevalence In the age group >=20 with a total population of 35,148, we observed an overall diabetes prevalence of 5.3% (95% CI: 5.0-5.5). Prevalence was slightly higher in men (5.5%, 95% CI: 5.1-5.8) than in women (5.1%, 95% CI: 4.7-5.4, difference not statistically significant). Prevalence differed between subregions: The lowest prevalence was observed in the subregion Sonnenterasse with 3.1% (95% CI: 2.4-3.9) and the highest in the central area Landeck and surroundings with 6.1% (95% CI: 5.7-6.5). The prevalence was 3.4% (95% CI: 2.9-3.9) in the region Paznauntal, 4.1% (95% CI: 3.1-5.5) in Kaunertal, 4.6% (95% CI: 4.1-5.3) in Stanzertal and 6.0% (95% CI: 5.4-6.6) in the region Oberes Gericht. Regarding gender, there were no statistically significant differences in prevalence for the subregions. The results for the subregions, genders and age groups are shown in Table 4. TABLE 4 Prevalence by gender (N = 1845, with all DM, including 'GDM only') Females Males Total Prevalence in % (95% CI) Prevalence in % (95% CI) Prevalence in % (95% CI) Total 5.1 (4.7-5.4) 5.5 (5.1-5.8) 5.3 (5.0-5.5) Regions Landeck/surroundings 5.8 (5.3-6.4) 6.5 (5.9-7.1) 6.1 (5.7-6.5) Oberes Gericht 5.9 (5.2-6.8) 6.0 (5.2-6.9) 6.0 (5.4-6.6) Sonnenterasse 3.6 (2.4-4.9) 2.6 (1.7-3.7) 3.1 (2.4-3.9) Kaunertal 4.6 (3.0-6.7) 3.7 (2.3-5.5) 4.1 (3.1-5.5) Paznauntal 2.8 (2.2-3.6) 3.9 (3.2-4.8) 3.4 (2.9-3.9) Stanzer Tal 4.4 (3.6-5.3) 4.9 (4.1-5.8) 4.6 (4.1-5.3) Age groups 20-49 1.8 (1.5-2.1) 1.1 (0.9-1.4) 1.5 (1.3-1.6) 50-64 3.6 (3.0-4.1) 6.3 (5.6-7.0) 4.9 (4.5-5.4) 65-74 11.0 (9.6-12.6) 14.3 (12.6-16.2) 12.5 (11.4-13.7) >=75 15.4 (13.8-17.0) 18.1 (16.1-20.3) 16.5 (15.2-17.8) Abbreviation: 95% CI, 95% confidence interval. 4 DISCUSSION We established, to the best of our knowledge, the first population-based cohort of patients with DM in Austria following the recommendations of the National Diabetes Strategy. Our cohort should support data-driven healthcare decision-making and should contribute improving outcomes of patients with DM. In total, we analysed 1845 adult cases with DM (including 'GDM only') and observed an overall prevalence of 5.3%, with no statistically significant difference between genders but substantial differences between subregions. In addition, we identified differences between women and men in the following patient characteristics: diabetes duration, diagnosis site, family history of diabetes, smoking status, distribution of age groups, and for specific clinical parameters such as BMI, HbA1c, LDL, determination of microalbumin, DM therapy and physical activity. We also observed significant differences between female and male patients for the long-term complications diabetic foot, amputation and cardiovascular disease. In order to derive unmet needs and healthcare services quality gaps, it is important to set our results into the context of other countries. The prevalence of DM varies strongly across the globe and between European countries. The IDF Diabetes Atlas reports a prevalence ranging from 4.0% in Ireland to 15.9% in Turkey, and reports a prevalence for Austria of 6.6%, which is clearly below the European average of 9.2%. 2 The overall prevalence of 5.3% in the district of Landeck is even lower than the prevalence estimate of the IDF for Austria, which lies above the 95% CI of the district Landeck (95% CI: 5.0%-5.5%). One reason for this difference within Austria could be the so-called east-west gap in Austria. A report by the Federal Ministry of Health showed that mortality due to DM in western Austria is lower than in eastern Austria. 19 There is also an evident east-west variation in cardiovascular mortality among persons over 64. Overweight/obesity, which is a major risk factor for DM, 20 is an even more significant problem in the population over the age of 64 in the east of Austria than in the west. 19 However, these estimates are not standardized, for example, for age, sociodemographic characteristics or proportion living in urban areas. Our prevalence estimation for the district of Landeck is in line with the reported prevalence of 4% in Switzerland, 21 which borders Landeck. Germany also shows a southwest-to-northeast gradient. The regional standardized prevalence was highest in the east, with 12.0% (95% CI: 10.3%-13.7%), and lowest in the south, with 5.8% (95% CI: 4.9%-6.7%). 22 However, we cannot exclude an underestimation of the prevalence in our study because we were unable to validate the completeness of the population-based cohort with independent data sources. When comparing the frequencies with the literature, it should be noted that we included all DM patients treated in a region, not only those treated in a study using only a sample of the target population. Therefore, we predominantly compare our figures with results from population-based diabetes registries. In addition, it is worth noting that our proportion of patients with high blood pressure should not be compared with hypertensive patients of study results from the published literature. The same is true for patients with higher LDL levels versus hyperlipidaemia. In the following, we compare the parameters obesity, glycaemic control, smoking status, foot inspections and participation in a diabetes education program with data from the Scottish diabetes registry, 23 the Swedish registry, 24 and with results from the DAWN2 study. 25 We also compare diabetes therapies with Austrian data 26 and with data from a diabetes surveillance system for Germany at the Robert Koch Institute. 22 In our study, 42.2% of women and 40.1% of men were obese. In Scotland, the proportion of obese patients was at 55%. 23 In Sweden, obesity was also more frequent, with a prevalence of 61% for women and 54% for men, although data from Sweden are only available for patients in the age group 30-60. 24 Overall, the proportion of obese patients in our study is lower compared to global numbers and the Austrian average, 27 which fits the east-west gradient in Austria mentioned above. Concerning glycaemic control, the proportion of patients with HbA1c < 7.5% was 55.4% in Scotland 23 and 72.3% in our study (75.2% in women and 69.9% in men). One reason could be that the population in this western part of Austria is less obese and more physically active. Even overweight patients perform substantial physical activity. In addition, diabetic care is rather well structured. The participation in a diabetes education program was 78.7% in the DAWN2 study, 25 which is nearly identical to our results of 78.2%. The DAWN2 study reported 62.5% foot inspections (vs. 79.7% in the Diabetes Landeck Cohort); however, the DAWN2 data were based on self-assessments. In the Diabetes Landeck Cohort, the proportion of active smokers was 10.9% (two times as many men than women), which is lower compared to 14.3% of the study of Panisch et al. 28 and to 17.7% in Scotland. 23 Our findings on diabetes therapy data do not describe the patients' current therapy (at the last contact) but rather summarize all therapy modalities that were documented during the study period. The proportion of patients with DM who did not need diabetes-specific therapy (lifestyle adaptation only) is relatively high (19%) but is similar to the Swedish registry. 24 The percentage of DM patients treated with OADs corresponds to 72% in the publication of Engler et al., 26 which describes patients in the diabetes register in Tyrol. Among 79-year-old people with T2DM in Germany, 29.6% of women and 37.2% of men received metformin monotherapy in 2010. 22 This percentage is higher in our study (38.9% in women and 42.1% in men). Also, the percentage for OAD+ insulin is higher in our cohort, which included all patients with DM compared to the study across Germany that only assessed patients with T2DM. 22 Comparisons of diabetes-specific therapies with the literature should be carefully interpreted as many studies exclude elderly patients. In contrast, in our cohort, all patients with DM were registered, and the proportion of patients aged >=75 was 36%. When comparing our results with published data, it is essential to consider that the frequency of the respective long-term complications depends on several factors, such as diabetes duration, age, gender and distribution of other risk factors. However, it also depends on the definition (and it is worthwhile to mention that definitions of each late complication differ in some respect), the extent of appropriate screening measures and the diagnostic methods used. Nonetheless, in the following paragraph, we attempt to compare our data on long-term complications to the literature. In Germany, the proportion of patients with DM with documented chronic kidney disease (as an indicator of nephropathy) was 15.1% (women: 14.9%; men: 15.3%) in 2013. 22 For the Diabetes Landeck Cohort, the percentage is higher (18.1%) compared to Germany, but differences in the documentation and diagnosis standards complicate a comparison. Desphande et al. 29 have provided an overview of the prevalence of the most common diabetes complications among individuals with T2M. 29 They reported a frequency of nephropathy of 28%. As the disease DM progresses, the frequency of nephropathy is low in the first 10-15 years after the diagnosis and then increases significantly. In the Diabetes Landeck Cohort, the diabetes duration is less than 10 years in 50% of patients; this may explain the lower frequency of nephropathy, with 19.4% in women and 17.0% in men compared to the data of Desphande et al. 29 Another explanation may be the relatively high proportion of patients with an HbA1c below 7.5%. The frequency of nephropathy correlates strongly with blood glucose control. 30 In 2013, 13.5% of adults with DM had documented diabetic polyneuropathy in Germany (women: 12.7%; men: 14.4%). 22 The proportion of patients with neuropathy is similar to that in our study (women: 12.6%; men: 13.1%). In Germany, 37.1% of adults with T2DM have cardiovascular disease, with a significantly lower prevalence in women (30.6%) than in men (42.8%), but this includes long-term complications and comorbidities. 22 In our study, we observed lower frequencies but also a significant gender difference (women: 12.3%; men: 25.2%) for cardiovascular disease as a long-term complication. Our study and the meta-analysis of Einarson 31 show nearly identical results for cerebrovascular long-term complications. The percentage of retinopathy in our cohort is very small (2.2%), but the prevalence of retinopathy increases progressively in patients with DM with increasing duration of the disease, 32 and we could be confronted with problems in communication diagnoses from ophthalmologists to care givers. In Germany, 5.7% of women and 6.6% of men with DM had documented diabetic foot syndrome in 2013. 22 For the Diabetes Landeck Cohort, the percentage is lower for women (2.5%) and similar for men (5.3%). All patient data registered in this study were pseudonymized according to EU data protection laws. This means that the patient can no longer be identified in the research database, and implausible data can no longer be verified. The process of pseudonymization is based on the Austrian social security number, which means that patients cannot be registered without a social security number; however, the proportion of individuals without a social security number was fairly small in our study (about 0.3% of all patients). 33 One strength of our study is that it covers a well-defined population and region, all care providers are clearly identified, and the area is served almost exclusively by one hospital. We established a network of general practitioners, diabetes specialists in private practices, retirement homes and the hospital in a defined district. Cooperation with all care units was excellent, with very few exceptions. Another strength was the use of a workable documentation system. Furthermore, by limiting the data to a minimum basic dataset, we achieved an acceptable data quality and thus gained a valid representation of the quality of care. Our study design could be a prototype for other countries and can contribute to assess a good overview of the quality of care of patients with DM, given a limited budget. We were also able to identify the problems in the data collection, which is important for future updates of the cohort or planning of diabetes registries in Austria. This study has several limitations. First, most of the study period was dominated by the COVID-19 pandemic. The documentation took place during one of the most difficult periods in the Austrian healthcare system in recent decades, and this could have affected our data quality. The care units could not be visited at the predefined time intervals for a more extended period due to the COVID-19 lockdown restrictions, and caregivers had less time to maintain and update records. This may lead to an underestimation of the true prevalence. Second, some general practitioners retired during the study, which limited cooperation. A third limitation that became evident during the study was the lack of coding DM diagnoses in most practice systems. This means that patients with DM can currently only be identified by a free text search, which is usually very time-consuming and can be associated with measurement bias. 34 This applies particularly to patients with DM who are not detected by the search criteria, which can lead to an underestimation of the true prevalence. It should also be considered that not all essential information is stored in the medical records and therefore was not accessible to the study nurse. Most physicians did not have the time to document or complete data themselves. Therefore, the majority of data documentation was completed by the study nurse who did not have direct contact with the patients. A fifth limitation of our study is that only a minimal data set was registered due to budget restrictions and the current conditions. For example, we were not able to collect data on triglycerides, ischaemic cardiomyopathy or other types of cardiomyopathy, or specific event types of cerebrovascular diseases or vascular events. As our results represent the population structure of the district Landeck and were not standardized to the Austrian population, crude comparisons with Austria should be interpreted with caution, see the discussion of the so-called east-west gap in Austria. Another significant limitation, the need for a uniform definition of long-term complications, applies to all published studies and was also a challenge we faced in our study. Finally, we were not able to verify the completeness of the prevalence estimate with independent data resources and could not systematically survey the patients' life statuses. 5 CONCLUSIONS As explicitly stated in Austria's National Strategy Report on diabetes, there is a lack of high-quality data on DM in the country. 6 There are, to the best of our knowledge, no systematic population-based data on patients with DM in Austria. The Diabetes Landeck Cohort closes this gap for one region in Austria and provides, for the first time in Austria, a nearly complete set of patients with DM living in a well-defined region. Some results, such as the diabetes prevalence or the frequency of some long-term complications, are lower compared to international data. We succeeded in establishing a population-based cohort and related database; however, we were not able to identify independent sources to verify our results. Therefore, for the future, we strongly suggest evaluating both completeness and comparability of data with well-accepted methods. In general, documentation by study nurses who should ideally be located in the care units is recommended to obtain valid data, because many important data are not stored in the practice systems or cannot be accessed in a systematic way. To access patients with specific diagnoses, support for the coding of diagnoses by physicians in private practices should be developed and applied in the practice systems. To make the best use of already existing data in the Austrian healthcare system, we recommend developing and/or optimizing systems to link different databases (e.g. civil registration and death data). The Diabetes Landeck Cohort should allow to evaluate and improve the quality of care of patients with DM in the future. In general, the cohort should be optimized and updated because high-quality data provide an essential basis to optimize the care of patients with DM. The data could also be used to supplement a biobank, for long-term monitoring of diabetes patients, for questions in health services research and healthcare economics, and for the investigation of new electronic communication methods between physicians and patients. Further research is needed, and in a subsequent step, we will extend this study by carefully taking the limitations into account. AUTHOR CONTRIBUTIONS Veronika Haslwanter: Conceptualization (lead); data curation (equal); formal analysis (equal); investigation (equal); methodology (equal); project administration (equal); validation (equal); visualization (lead); writing - original draft (lead). Ursula Rochau: Supervision (supporting); writing - original draft (supporting); writing - review and editing (supporting). Uwe Siebert: Supervision (supporting); writing - original draft (supporting); writing - review and editing (supporting). Hans Schoenherr: Conceptualization (supporting); data curation (equal); methodology (equal). Wilhelm Oberaigner: Conceptualization (supporting); formal analysis (equal); methodology (equal); supervision (lead); writing - review and editing (equal). FUNDING INFORMATION We thank the Tyrolean Health Fund (Tiroler Gesundheitsfond, TGF) for funding this project. CONFLICT OF INTEREST The authors have no competing interests. ETHICS APPROVAL The present study was approved by the Ethics Committee of the Medical University of Innsbruck and was carried out in accordance with the Helsinki Declaration of 1975, as revised in 2008. Supporting information Tables S1-S2 Click here for additional data file. ACKNOWLEDGEMENTS We would like to thank all participating general practitioners, diabetes specialists in private practices and the Hospital Zams for their collaboration. DATA AVAILABILITY STATEMENT The data that support the findings of this study are available from Tiroler Gesundheitsfond. Restrictions apply to the availability of these data, which were used under licence for this study. |
PMC10000637 | Introduction Therapy for hypothyroid obese patients is still under definition since the thyrotropin-stimulating hormone (TSH) level is a less reliable marker of euthyroidism than nonobese patients. Indeed, TSH levels positively correlate with body mass index (BMI), and this increase may be a compensatory mechanism aimed at increasing energy expenditure in obese people. In contrast, the correlation of BMI with thyroid hormone levels is not completely clear, and conflicting results have been obtained by several studies. The L-T4 replacement dose is more variable in obese hypothyroid patients than in nonobese patients, and a recent study indicated that the L-T4 replacement dose is related to lean body mass in obese thyroidectomized patients. We aimed to study the correlations of L-T4-administered dose, thyroid hormone levels and TSH secretion with basal metabolic rate (BMR) and total calculated deiodinase activity (GD) in obese and nonobese athyreotic patients. We also looked for individualized L-T4 replacement dose set points to be used in clinical practice. Methods We studied retrospectively 160 athyreotic patients, 120 nonobese and 40 obese. GD was calculated by SPINA Thyr 4.2, the responsiveness of the hypothalamic/pituitary thyrotrope by Jostel's thyrotropin (TSH) index and BMR by the Mifflin-St. Jeor formula, the interplay of GD and BMR with L-T4, thyroid hormones and TSH index (TSHI) was also evaluated. Results In our study, the L-T4 dose was an independent predictor of GD, and approximately 30% of athyreotic patients under L-T4 therapy had a reduced GD; FT4 levels were higher and negatively modulated by BMR in obese athyreotic patients respect to nonobese, in these patients a T4 to T3 shunt, in terms of TSHI suppression is observed suggesting a defective hypothalamic pituitary T4 to T3 conversion and a resistance to L-T4 replacement therapy. Conclusions L-t4 dose is the most important predictor of GD, BMR modulates T4 levels in obese athyreotic patients that are resistant to L-T4 replacement therapy. Because of controversial issues about the optimal T4 replacement dose in obese hypothyroid subjects and the great importance of thyroid hormones in energy homeostasis, glucose and lipid metabolism, body composition and resting energy expenditure (REE), we compared the correlation between L-T4 administered dose, thyroid hormone levels and TSH secretion with estimated basal metabolic rate (BMR) and total deiodinase activity (GD) in obese and nonobese athyreotic subjects. We aimed to define individualized set points that might provide appropriate therapeutic and biochemical targets to be clinically tested in obese and nonobese patients. deiodination hypothyroidism LT-4 treatment obesity source-schema-version-number2.0 cover-dateMarch 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.2.6 mode:remove_FC converted:10.03.2023 Le Moli R , Malandrino P , Russo M , et al. Levothyroxine therapy, calculated deiodinases activity and basal metabolic rate in obese or nonobese patients after total thyroidectomy for differentiated thyroid cancer, results of a retrospective observational study. Endocrinol Diab Metab. 2023;6 :e406. doi:10.1002/edm2.406 pmc1 INTRODUCTION Levothyroxine (L-T4) therapy has a long record in clinical use, with a defined pharmacological profile and safety in hypothyroidism management. Obesity and thyroid disorders are common among the general population and may be associated with both clinical and molecular aspects. This relationship has become epidemiologically relevant in the context of the significantly increased prevalence of obesity worldwide. However, treatment for obese patients with subclinical or overt hypothyroidism is still under definition regarding both the threshold and modality (liquid L-T4 vs. pills; L-T4 monotherapy vs. liothyronine [L-T3]/L-T4 combinations). The prerequisite for treatment with L-T4 is the presence of hypothyroidism, and the goal is restoration of euthyroidism. Achievement of a thyrotropin-stimulating hormone (TSH) value within the age-adjusted euthyroid range is the accepted therapeutic target, as several studies indicate improvement in symptoms, quality of life and cardiovascular risk. 1 , 2 , 3 , 4 However, among euthyroid subjects, TSH levels usually correlate with body mass index (BMI), being higher in obese than in normal subjects. 5 TSH elevation in obese euthyroid people may be a compensatory mechanism in the pituitary-thyroid axis aimed at increasing energy expenditure. 6 , 7 At variance with TSH, the correlation between BMI and thyroid hormones (T4 and T3) is not clear, as several studies obtained conflicting results. Some studies indicate that BMI is negatively related to FT4 and positively related to FT3. Another study, by contrast, indicated hyperactivation of the pituitary-thyroid axis with increased FT4 levels in obese patients. 8 , 9 Other studies describe a decreased FT4/FT3 ratio in obese patients. 5 , 10 , 11 , 12 This adaptation of thyroid hormone homeostasis in obese subjects has been attributed to leptin and insulin actions. 3 The observation of higher TSH and lower FT4 in obese euthyroid people is in accordance with increased L-thyroxine replacement dose in hypothyroid obese patients. L-T4 replacement therapy is approximately 1.6 mg/kg in hypothyroid patients with any functional thyroid tissue, while in obese patients, the correct T4 replacement dose is more variable. Recently, the American Thyroid Association (ATA) task force identified obesity as a morbid condition implying an increase in the L-T4 replacement dose because of reduced thyroid hormone absorption. 2 This observation is reinforced by the evidence that in obese subjects, acute overload of L-T4 administration takes longer to achieve a plasmatic concentration peak in comparison with nonobese people. 1 However, a recent study indicated that in obese thyroidectomized patients, the L-T4 replacement dose is positively related to lean body mass. Indeed, the ideal body weight (IBW) should be preferred to real body weight (RBW) for L-T4 dose titration because lean body mass results in a better predictor of T4 requirement than fat mass. 6 , 7 Because of these controversial issues about the optimal T4 replacement dose in obese hypothyroid subjects and the great importance of thyroid hormones in energy homeostasis, glucose and lipid metabolism, body composition and resting energy expenditure (REE), 10 , 12 , 13 , 14 we compared the correlation between L-T4-administered dose, thyroid hormone levels and TSH secretion with estimated basal metabolic rate (BMR) and total deiodinase activity (GD) in obese and nonobese athyreotic subjects. Moreover, we aimed to define individualized set points that might provide appropriate therapeutic and biochemical targets to be clinically tested in obese and nonobese patients. 2 PATIENTS AND METHODS 2.1 Patients We retrospectively evaluated 1150 thyroidectomized patients referred to our outpatient thyroid clinic between 2010 and 2015 who were also subjected to 131I ablation because of differentiated thyroid cancer (DTC). In all patients, thyroglobulin levels were between 0.01 and 0.5 ng/ml, and antithyroglobulin antibody (TgAb) was negative. In this cohort, devoid of functional thyroid tissue, all circulating T4 levels originated from levothyroxine replacement therapy. These patients obtain circulating T3 from the conversion of exogenous T4 and represent an ideal model to study peripheral tissue ability to generate biologically active hormones. We excluded from the analysis patients with hypothalamic/pituitary, gastric, intestinal or neurological diseases and pregnant women (n = 72) and those who were taking combined T3/T4 thyroid replacement therapy and/or other drugs interfering with thyroid hormone homeostasis (n = 198). Patients with variations in L-T4 daily dose, body weight and thyroid hormone level fluctuations within 3 months before the start of the study were also excluded (n = 720). Finally, 160 athyreotic patients under L-T4 therapy were included in the analysis . All patients were euthyroid on the basis of their TSH, FT4 and FT3 levels within the normal range. FIGURE 1 Flow chart of study patients selection 2.2 Phenotypic evaluation of the study patients Clinical records included a detailed history, physical examination, standardized questionnaire documenting sex, age, height, weight and BMI. BMI was calculated as weight in kilograms divided by the square of height in metres (kg/m2) and considered a categorical variable according to the World Health Organization (WHO). Obesity was defined as BMI >= 30, which is an adequate indicator of obesity and is associated with increased body fat mass. In our study, 120 patients had a BMI >= 30, while 40 had a BMI < 30. 2.3 Basal metabolic rate (BMR) evaluation We evaluated BMR by the Mifflin-St. Jeor formula (MSTF). The MSTF equation is commonly used in the assessment of basal metabolism and is more particularly used in obese patients. The MSTF was also applied differently to female and male sex as follows: Females = 9.99 x weight (kg) + 6.25 x height (cm) - 4.92 x age (years) - 161. Males = 9.99 x weight (kg) + 6.25 x height (cm) - 4.92 x age (years) + 5. We studied the effect of the L-T4 replacement dose on thyroid hormone homeostasis, estimated BMR and total deiodinase activity (GD) in obese and nonobese patients. Data were collected from patients after thyroidectomy, 131I administration and a persistent euthyroid state under replacement therapy for approximately 3 months with any significant change in L-T4 dose administration, daily caloric intake and body weight. A subgroup of 45 patients maintaining the same replacement dose over the last 6 months was also studied to better evaluate the interplay between the L-T4 administered dose and total GD in the long term. 2.4 Evaluation of stimulated deiodination (GD) GD, which reflects the maximum stimulated activity of deiodination, was calculated by SPINA Thyr 4.2 (Structure Parameter Inference Approach by Johannes W. Dietrich, Lab XU44, Bergmannsheil University Hospitals, Ruhr University of Bochum, D-44789 Bochum, NRW, Germany), which is a mathematical tool for the integrated interpretation of laboratory results. SPINA allows calculation of GD from TSH, FT4 and FT3 serum levels obtained from routine laboratory assays. The method is based upon mathematical/cybernetic modelling of processing structures. 15 In particular, the SPINA algorithm is based on equilibrium analysis of a compartmental nonlinear model: GD = b31Km1+FT41+K30TBGFT3a31FT4, where b31 is the clearance exponent for T3, Km1 is the dissociation constant of type 1 deiodinase, K30 is the dissociation constant of T3 at thyroxine-binding globulin, and a31 is the dilution factor for triiodothyronine. On the basis of several studies, normal values of calculated GD vary between 21 and 26 nmol/s. 15 , 16 Hence, a GD < 21 nmol/s is considered low. 2.5 Responsiveness of the hypothalamic/pituitary thyrotrope We also assessed the responsiveness of the hypothalamic/pituitary thyrotrope by Jostel's thyrotropin (TSH) index: (JTSHI) = ln([TSH]) + b[FT4] and obtained a standardized TSH index (TSHI) = JTSHI - 2.7/0.676 for statistical comparison. 2.6 Laboratory measurements Serum TSH was assessed by an ultrasensitive enhanced chemiluminescence immunoassay (ECLIA) assay. Serum hormones were measured by microparticle enzyme immunoassay (Abbot AxSYM-MEIA) with interassay coefficients of variation of less than 10% over the analytical ranges of 1.7-46.0 pmol/L for FT3, 5.15-77.0 pmol/L for FT4 and 0.03-10.0 mU/L for TSH. The within-run and between-run precisions for the FT3, FT4 and TSH assays showed coefficients of variation <5%. Measurement of antithyroglobulin antibodies (TgAbs) by an automated chemiluminescence assay system (AntiTg, Ready Pack). Thyroglobulin levels were measured with a second-generation chemiluminescent Tg immunoassay (Tg Access; Beckman Coulter) with a functional sensitivity of 0.1 ng/ml. 2.7 Statistical analysis Statistical analysis was performed using the SPSS package (IBM SPSS Statistics for Windows, Version 26.0. IBM Corp). For the descriptive analysis, continuous variables were expressed as the mean +- standard deviation (SD) or median (with its 25th-75th percentile); categorical variables were expressed as numbers and percentages. Univariate analysis of variance (ANOVA) was performed to identify predictive variables significantly associated with the clinical outcome. The shapes of the distribution of each variable were evaluated by visual inspection of the population pyramid charts; for distributions of similar shapes, we reported the medians, and for distributions of different shapes, we reported average ranks. The Mann-Whitney U test was used to analyse the continuous variables without a normal distribution. Categorical variables were analysed by the Chi-square test, if cells with fewer than five expected cell numbers were found, by Fisher's exact test. Complete and partial bivariate analysis was used to evaluate no categorical variables, and Pearson's coefficient was computed. Binary logistic regression analysis was performed for the outcome variables. Covariates were selected on the basis of the results of univariate analysis, and the final model was built using forced entry and a hierarchical method. Linearity of the continuous variables with respect to the logit of the dependent variable was assessed by the Box-Tidwell procedure, and a Bonferroni correction was applied using all terms in the model to assess its statistical significance. Multicollinearity was excluded after checking tolerance and variance inflation factor statistics and the proportion of the variance of each predictor's b value attributed to each eigenvalue. The ability of the model to discriminate between outcome categories was investigated in more detail by elaborating the ROC curve. This analysis was performed for LT4 x week/BMR ratio vs deiodinase activity on the basis of the regression outputs. Youden's best cut-off was also calculated, and the greater values were chosen to balance the better sensitivity and specificity for the studied variable. 3 RESULTS 3.1 Patient evaluation according to GD 3.1.1 Univariate analysis Structure parameter influence assay (SPINA) revealed that GD was reduced in 50/160 (31.2%) of the thyroidectomized patients (Table 1). TABLE 1 Characteristics of the 160 athyreotic patients treated with L-T4 Age (years) 44.6 (13.9) Sex (F/M) 117/43 Weight (kg) 73.1 (18.3) Height (cm) 163.6 (12.2) BMI (kg/Height2) 27.1 (6.1) TSH (mU/L) 1.6 (0.4-2.9) FT4 (pmol/L) 14.1 (11.6-21.9) FT3 (pmol/L) 4.1 (2.1-5.4) FT3/FT4ratio 0.25 (0.05) BMR (Kcal/24 h) 1419.1 (265.5) LT-4 x week/BMRr (mg) 0.6 (0.4-1.4) LT-4 x week (mg) 835.7 (238.7) GD (nmol/s) 24.1 (12.0-40.0) GD < 21 nmol/s (n/%) 50/31.2 TSHI 2.0 (0.0-3.9) Note: Data are expressed as mean and standard deviation or median and range between the brackets. Abbreviations: BMR, basal metabolic rate; GD, calculated deiodinases activity; LT4 x week, LT-4 cumulative weekly dose; LT4 x week/BMR, LT-4 cumulative weekly dose/basal metabolic rate; TSHI, TSH index. Patients were divided into two groups according to normal (>=21 nmol/s) or low (<21 nmol/s) GD. Sex, age, BMI and BMR were not different between the two groups (Table 2). Univariate analysis revealed that FT3 and the FT3/FT4 ratio were significantly reduced in patients with low GD compared to patients with normal GD (p < .004-.0001). However, in low GD TSHI, FT4, LT-4 weekly cumulative dose (LT-4 x week) and the ratio between LT-4 weekly cumulative dose and basal metabolic rate (LT-4 x week/BMR) were significantly increased (p < .0001) (Table 2). TABLE 2 Characteristics of the 160 according to calculated deiodination activity (GD) GD activity <21 nmol/s (n = 50) GD activity >=21 nmol/s (n = 110) p Age (years) 42.9 (17.3) 45.2 (12.1) .6 Sex (F/M) 34/16 83/27 .2 Weight (kg) 73.3 (20.2) 73.1 (17.5) .9 Height (cm) 162.9 (16.9) 163.9 (9.5) .7 BMI (kg/Height2) 26.8 (6.7) 27.1 (5.6) .6 TSH (mU/L) a 0.8 (0.1) 1.0 (0.1) .4 FT4 (pmol/L) a 18.1 (2.0) 14.2 (2.6) .0001 FT3 (pmol/L) a 3.7 (0.1) 4.3 (0.0) .0001 FT3/FT4 ratio a 0.20 (0.05) 0.30 (0.05) .0001 BMR (Kcal/24 h) a 1435.1 (40.3) 1409.8 (24.7) .4 LT-4 x week/BMRr (mg/BMR) 0.6 (0.2) 0.5 (0.1) .01 LT-4 x week (mg) 910.2 (259.2) 802.5 (222.3) .006 GD (nmol/s) 18.3 (0.3) 26.6 (0.3) .0001 TSHI a 2.2 (0.1) 1.7 (0.1) .004 a Data are expressed as median and standard error between the brackets. 3.1.2 Binary logistic regression analysis and ROC curve Variables reaching statistical significance by univariate analysis were then analysed by binary logistic regression analysis models. LT-4 x week/BMR was independently and inversely related to GD [B = -3.88, wald = 7.6, R = 0.021 (0.001-0.329; 95% confidence interval (CI)), p = .006], FT3 levels were directly and independently related to GD [B = 2.81, wald = 25.1, R = 17.4 (5.6-53.4, 95% CI) p = .0001]. In contrast, BMR, BMI, body weight, TSH and FT4 were not independently related to GD. To evaluate the effect of LT4 x week/BMR on GD, we used a classic receiver operating characteristic (ROC) model that was very well validated by the study of area under the curve (AUC) = 0.81 +- 0.073 (0.66-0.95, 95% CI, p = .001). To better define the cut-off of LT-4 dose beyond which GD was reduced, we researched the best cut-off of Youden's statistic (YS). YS = 60 indicates that LT-4 x week/BMR > 0.56 mcg x week/kcal is a good predictor of suppressed GD with sensitivity = 83% and specificity = 77% (e.g. a total of 144 mcg of LT-4 daily dose reduces GD in patients with 1800 kcal/die estimated BMR). 3.2 Patient evaluation according to BMI 3.2.1 Linear regression, complete or partial bivariate analysis with calculation of Pearson coefficient FT3 and FT4 were increased in obese patients compared with nonobese patients (p = .07 and p = .01, respectively), while GD and LT-4 x week/BMR were similar in the two groups (Table 3). TABLE 3 Characteristics of 160 patients according to BMI Non-obese (n = 120) BMI >=30 < 35 (n = 20) BMI >=35 (n = 20) p Sex m/f 28/92 8/12 7/13 Age (years) 43.4 (14.5) 46.7 (11.5) 48.3 (11.1) .1 Weight (kg) 65.5 (12.2) 89.3 (10.8) 101.6 (16.4) .00 Height (cm) 163.1 (12.7) 167.9 (10.5) 160.9 (8.8) .4 BMI (weight/[height]2) 24.3 (3.2) 31.6 (1.3) 39.7 (4.9) .00 TSH (mU/L) 0.8 (0.05) 1.0 (0.2) 1.3 (0.2) .6 TSHI 1.7 (0.8) 2.1 (0.7) 2.2 (0.2) .1 FT4 (pmol/L) 15.6 (2.4) 16.7 (3.2) 17.7 (4.2) .01 FT3 (pmol/L) 3.9 (0.6) 4.1 (0.6) 4.3 (0.5) .07 FT3/FT4 ratio 0.25 (0.04) 0.24 (0.05) 0.24 (0.05) .1 BMR 1336.8 (216.6) 1633.4 (254.8) 1688.1 (258.4) .00 GD (nmol/s) 24.1 (4.8) 23.7 (5.5) 24.4 (6.1) .2 GD < 21 (nmol/s) n/% 36/30 7/35 7/35 .3 Lt4 x week/BMR 0.6 (0.1) 0.6 (0.1) 0.6 (0.1) .3 Lt4 x week (mg) 779.2 (214.1) 1006.9 (270.3) 1002.0 (173.1) .00 Note: Data are expressed as media and SD between the brackets. Partial bivariate analysis revealed that FT4 levels were positively related to BMI and negatively related to BMR after subtraction of the BMI effect: p = .01 and p = .02. Pituitary thyreotropic activity, evaluated by TSHI, was positively related to BMI and LT-4 x week/BMR: R = 0.13, p = .05, R = 0.14, p = .03 and inversely related to GD: p = .0004 . FT4 levels were positively related to TSHI in both obese and nonobese patients. In obese patients, the FT4 to TSHI increment was 3.3 times greater than the increment in nonobese patients: 0.1% versus 0.03%, R 2 = .24, p = .001 versus R 2 = .04, p = .02 (BMI >= 30 vs. BMI < 30) . FIGURE 2 Linear correlation of SPINA GD (nmol/s) with TSHI FIGURE 3 (A and B) Correlation between TSHI and FT4 in obese and non-obese athyreotic patients In obese patients (n = 40), FT3 levels were inversely related to TSHI; a TSHI increment of 1 unit was related to an FT3 decrement of 0.095%: R 2 = .1; p = .045. In contrast, FT3 levels were not related to TSHI variations in nonobese patients: R 2 = .002, p = .58 . FIGURE 4 (A and B) Correlation between TSHI and FT3 in obese and non-obese athyreotic patients These data confirm that the feedback sensitivity of thyroid hormones with the pituitary is significantly different in obese and nonobese patients. 4 DISCUSSION Several lines of evidence indicate that hypothyroid patients under levothyroxine replacement therapy may present impaired T3 production and a reduced T3/T4 ratio. 13 The T3 pool derived from intrathyroidal conversion is absent and fails to maintain normal FT3 levels. As a consequence, their peripheral tissues may be underexposed to circulating T3. Our previous data indicate that 29.6% of levothyroxine-treated athyreotic patients have a reduced FT3/FT4 ratio, and this percentage may progressively increase with increasing replacement levothyroxine dose. 17 These changes may be due to an imbalance between central and peripheral deiodinase activity that may disrupt thyroid hormone homeostasis in this subset of hypothyroid patients. 12 , 13 , 14 In our study, we evaluated the total deiodinase activity (GD) by the SPINA cybernetic model. 15 , 16 We found that our athyreotic patients with impaired GD received a larger dose of LT-4 and had increased FT4 and TSHI levels, while the FT3/FT4 ratio and FT3 levels were reduced (all <0.0001). GD was reduced in 31.2% of study patients, confirming our previous report since GD is well correlated with the FT3/FT4 ratio 17 (Table 2). To better evaluate the interplay between GD, BMR and LT-4 weekly cumulative dose (LT-4 x week), we evaluated the ratio between LT-4 x week and basal metabolic rate (LT-4 x week/BMR) calculated by the formula of Mifflin St.-Jeor. By this tool largely used to evaluate BMR in obese patients, 18 , 19 we demonstrated that total GD activity was independently and inversely related to LT-4 x week/BMR. According to this view, we analysed a subgroup of 45 patients with a stable LT-4 dose, caloric intake and level of thyroid hormones for almost six months, and we found that a LT-4 x week/BMR value of 0.56 mcg x week/Kcal can predict the impairment of GD (<21 nmol/s) with good sensitivity and specificity (p = .01). To our knowledge, this is a new finding with a possible clinical implication in athyreotic patients receiving LT-4 substitutive therapy. Interestingly, estimated BMR, BMI, age and sex were similar between the patients with normal or reduced GD, suggesting that LT-4 dose and FT3 production are the two independent stronger predictors of GD. Cross-sectional and longitudinal studies comparing presurgical levels of L-T4 prove that higher L-T4 doses are associated with the suppression of deiodinase activity. 16 FT4 and FT3 were higher in our obese (BMI >= 30) than in nonobese patients (BMI < 30) (p = .01, p = .07), and TSHI was positively related to BMI and LT-4 x week/BMR and inversely related to GD. However, GD and LT-4 x week/BRM were not different between obese and nonobese patients, suggesting that BMI is not an independent determinant of GD. The pituitary thyrotropic activity, expressed by the relationship between TSHI and thyroid hormone levels, was different between nonobese and obese patients. TSHI suppression was constantly exerted by increasing levels of FT4 in nonobese patients, while this suppression was significantly attenuated at higher levels of FT4 in obese patients, suggesting increased hypothalamic-pituitary resistance in response to increased T4 levels. The increment of FT4 for each unit of TSHI increment was significantly higher in obese patients than in nonobese patients (p = .04) . However, in accordance with the FT4 results, increasing levels of FT3 constantly suppressed TSHI in nonobese patients, while this suppression was increased at increasing levels of FT3 in obese patients . This T4 to T3 shunt, in terms of TSHI suppression observed in obese patients, suggests a defective hypothalamic pituitary T4 to T3 conversion. Moreover, FT4 levels were positively related to BMI as well as to T4 dose but only partially and inversely related to BMR when BMI effect was subtracted (Pearson, p = .01). Considering that FT4 levels in athyreotic patients are entirely dependent on LT-4 adsorbed dose and on the extent of T4 degradation, 17 this finding unravels a role of BMR on the modulation of FT4 bioavailability both in nonobese and in obese patients, those with greater lean body mass that leads to increased BMR. 6 , 7 , 8 Differently than some recent studies, 20 we did not evidence a statistically significant correlation of GD with BMR, however differently than the others studies we evaluated patients athyreotic by total thyroidectomy and 131I ablation, this might contribute to increase the severity of suppression of the feedback loop and the ability to relay type 1 and type 2 allostatic load to T3 production. Moreover, we did not evaluate separately free fat mass and lean body mass. Under normal conditions, thyroid hormones and TSH are inversely correlated, while in patients with resistance to thyroid hormone, higher thyroid hormone levels correspond to high TSH levels due to a possible condition of resistance to FT4, such as in obese patients. 9 , 10 , 21 , 22 One study demonstrated that deiodinase ubiquitination was an important factor in restoring euthyroidism. Indeed, the ubiquitin proteasome system in the hypothalamus of obese mice fails to maintain adequate function. Hence, a defective function of the ubiquitin proteasome system, resulting in deiodinase imbalance, might play a major role in the regulation of the response to thyroid hormones in obese subjects. 9 , 21 , 22 , 23 Thyroid hormone action is modulated by the hypothalamic pituitary thyroid axis, 6 and cell membrane transport, tissue deiodination and degradation and thyroid hormone metabolism in the liver may play an important role. 9 , 22 , 23 , 24 Metabolism of exogenous substrates in the liver occurs by enzymes that either modify and/or conjugate the functional groups to endogenous substrates to increase their solubility to be readily eliminated. Approximately half of obese subjects display several abnormalities in liver enzymatic activity due to steatosis. 25 , 26 , 27 In particular, increasing BMI and thyroid hormone receptor b are inversely correlated with different stages of nonalcoholic fatty liver disease (NAFLD), 28 , 29 which, in turn, is related to decreased multidrug resistance protein (MRP2) activity in the liver. This condition is associated with alterations in the expression and function of enzymes and transporters resulting in an altered glucuronoconjugation of thyroid hormones. 23 However, our study is descriptive and does not allow any direct evaluation of mechanistic insights related to T4 activation, degradation and stability. 5 CONCLUSIONS Approximately one-third of athyreotic patients under LT-4 replacement therapy have reduced GD. GD activity is inversely and independently related to LT-4 dose and FT3 levels. We found that an LT-4 weekly cumulative dose of 0.56 mcg/kcal was an independent predictor of reduced GD, while sex, age, BMI or BMR were not. FT4 levels are higher in athyreotic obese patients, who therefore appear more resistant to LT-4 replacement therapy. Indeed, FT4 is positively related to BMI and inversely related to BMR, which, in turn, negatively modulates the FT4 increment, especially in patients with elevated body lean mass. Other metabolic pathways both centrally and perimetrically might be involved in FT4 and FT3 degradation. AUTHOR CONTRIBUTIONS Pasqualino Malandrino: Data curation (equal); validation (equal). Marco Russo: Data curation (equal); validation (equal). Dario Tumino: Data curation (equal); validation (equal); visualization (equal). Tommaso Piticchio: Data curation (equal); formal analysis (equal); visualization (equal). Adriano Naselli: Data curation (equal); formal analysis (equal); software (lead); supervision (equal); writing - review and editing (equal). Valentina Rapicavoli: Data curation (equal); resources (equal). Antonino Belfiore: Conceptualization (equal); funding acquisition (equal); methodology (equal); project administration (equal); resources (equal); supervision (equal); validation (equal); visualization (equal); writing - review and editing (equal). Francesco Frasca: Conceptualization (equal); funding acquisition (equal); investigation (lead); methodology (equal); project administration (equal); resources (equal); supervision (equal); validation (equal); visualization (equal); writing - review and editing (equal). Rosario Le Moli: Conceptualization (equal); data curation (equal); formal analysis (equal); investigation (equal); methodology (equal); project administration (equal); software (equal); supervision (equal); validation (equal); visualization (equal); writing - original draft (equal); writing - review and editing (equal). CONFLICT OF INTEREST The authors declare no conflict of interest. INSTITUTIONAL REVIEW BOARD STATEMENT The studies involving human participants were reviewed and approved by Ethics Committee Garibaldi Nesima Hospital - Catania. INFORMED CONSENT Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. DATA AVAILABILITY STATEMENT The data presented in this study are available on request from the corresponding author. |
PMC10000638 | Introduction Oxidative stress known as a predictive marker for cardiovascular and metabolic diseases could be measured through pro-oxidant antioxidant balance (PAB). The present study aimed to evaluate PAB and its association with high-sensitivity C-reactive protein (hs-CRP) in the serum of postmenopausal women with diabetes mellitus. Methods In this case-control study, 99 diabetic and 100 healthy postmenopausal women without diabetes mellitus were recruited. Serum PAB values, hs-CRP, lipid profile, insulin, and vitamin D levels were measured. Moreover, insulin resistance (HOMA-IR, HOMA-b and QUICKI), waist circumference (WC), waist-to-hip ratio (WHR), waist-to-height ratio (WHtR), and body mass index (BMI) were calculated. Results Serum PAB, hs-CRP, insulin resistance, HOMA-b, QUICKI, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and triglycerides (TG) levels were significantly higher in the postmenopausal women with diabetes mellitus, while there was no significant difference in the total cholesterol (TC), serum insulin, WC, WHR, WHtR and vitamin D levels between the groups. Pearson correlation coefficient showed that HDL-C and insulin levels were directly correlated with serum PAB. Also, there was a significant direct relationship between LDL-C and insulin levels and hs-CRP. There was no meaningful relationship between serum insulin and vitamin D levels and other assessed parameters. Backward logistic regression showed a positive relationship between diabetes mellitus and serum PAB and an inverse relationship with serum HDL levels. Conclusions Serum PAB, hs-CRP concentration, and lipid profile were significantly different between postmenopausal women with and without diabetes mellitus. These differences may contribute to the development of coronary complications. Serum PAB, hs-CRP concentration, and lipid profile were significantly different between postmenopausal women with and without diabetes mellitus. These differences may contribute to the development of coronary complications. diabetes high-sensitivity C-reactive protein post menopause Pro-oxidant-antioxidant balance vitamin D Mashhad University of Medical Sciences 10.13039/501100004748 981826 source-schema-version-number2.0 cover-dateMarch 2023 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.2.6 mode:remove_FC converted:10.03.2023 Ehteram H , Raji S , Rahmati M , et al. Association between Pro-oxidant-Antioxidant balance and high-sensitivity C-reactive protein in type 2 diabetes mellitus: A Study on Postmenopausal Women. Endocrinol Diab Metab. 2023;6 :e400. doi:10.1002/edm2.400 pmc1 INTRODUCTION An imbalance between the production of oxidants and their scavengers leads to oxidative stress (OS). OS may also stimulate the production of inflammatory factors, such as high-sensitivity C-reactive protein (hs-CRP). Hs-CRP is an inflammatory marker induced via cytokines, especially interleukin-6 (IL-6). OS and hs-CRP are predictive markers of cardiovascular disease (CVD) and metabolic diseases, including type II diabetes. 1 , 2 , 3 Type II diabetes mellitus (T2DM) is a severe, multifactorial and metabolic disease, which affects women more than men in many countries. T2DM increases CVD risk by 2-3 folds, which leads to a higher mortality rate than in non-diabetic people. 4 Moreover, OS increases in menopausal women, which is associated with loss of ovarian follicular function and oestrogen (E2) production because E2 has antioxidant activity. 5 After menopause, the production of antioxidants is reduced, and OS increases. 6 , 7 Thus, menopause may be a risk factor for OS, CVD, osteoporosis, and diabetes. Although OS and inflammation are well established in postmenopausal women, there are limited studies about pro-oxidant-antioxidant balance (PAB), CRP levels and their association with insulin resistance in diabetic postmenopausal women. 1 , 8 , 9 We sought to assess the serum PAB values using a modified PAB assay to measure the pro-oxidant burden and antioxidant capacity. This study also evaluated hs-CRP and whether serum PAB values are associated with hs-CRP in diabetic postmenopausal women. 2 MATERIALS AND METHODS 2.1 Study groups This case-control study was carried out on 99 postmenopausal women who had recently been diagnosed with only diabetes type II and attended the Women's Health Research Center in Gorgan, Iran. The control group included 100 healthy participants age-matched to the patient group recruited between January 2017 and June 2018 for routine check-ups. This group consisted of postmenopausal women with no diabetes. Clinical history and other relevant data were collected from all participants. They were excluded if they had taken vitamin supplements, hormones, anti-inflammatory drugs, and fish oil capsules. Moreover, smokers and pregnant subjects were excluded from the study. Those suffering a myocardial infarction (MI), acute infection or any acute illnesses were excluded. One hundred and 99 subjects met the inclusion/exclusion criteria. They were informed about the study protocol, written consent was obtained from each participant and the research was approved by the Mashhad University of Medical Sciences Ethics Committee (NO: IR.MUMS.REC.1399.533). 2.2 Anthropometric parameters and blood collection After overnight fasting, 5 ml of venous blood was drawn into EDTA and plain tubes, centrifuged at 2500 rpm for 15 min at room temperature, and serum was allocated to several microtubes and stored at -70degC until analysis. Furthermore, body weight, height, waist circumference (WC), and hip circumference (HC) were measured to calculate the waist-to-hip ratio (WHR), waist-to-height (WHtR), and body mass index (BMI) (kg/m2). 2.3 Biochemical analysis processing Fasting glucose and lipid profile indices, including total cholesterol (TC), triglyceride (TG), and HDL-C, were measured by enzymatic methods and commercial kits using the BT-3000 Auto-analyser (Biotechnica). Moreover, LDL-C was indirectly evaluated in participants with the Friedewald formula. The levels of insulin were assessed using commercial kits using a radioimmunoassay from the Immuno Nuclear Corporation (Stillwater). Insulin resistance was calculated using the HOMA equation: HOMA-IR = [Fasting insulin (mIU/ml) fasting glucose (mM/L)]/22.5. Also, homeostasis model assessment of b-cell function (HOMA-b) and quantitative insulin sensitivity check index (QUICKI) were used to assess b-cell function and insulin sensitivity, respectively, as follows: HOMA-b: (fasting plasma insulin [mU/ml] * 20)/(fasting blood glucose [mmol/l] - 3.5) and QUICKI: 1/(log fasting blood glucose [mmol/l] + log fasting plasma insulin [mU/ml]). Furthermore, serum 25-hydroxyvitamin D [25(OH) D] levels were assessed using a commercial ELISA kit (25-Hydroxyvitamin D ELISA kit; Immuno Diagnostic Systems). 2.4 Measurements of hs-CRP The PEG (polyethylene glycol)-enhanced immuno-turbidometry method and commercially available kits on an Alcyon(r) analyser (Abbott) were used to measure hs-CRP levels. 2.5 Assessment of PAB Serum PAB values were measured in all subjects as previously described by Alamdari et al. 10 In the first step, we added horseradish peroxidase enzyme and chloramine-T as oxidizing agents to TMB. Redox index resulted in the combined activity of a colour cation (by oxidants) or reduced to a colourless compound (by antioxidants). In standard solutions, various proportions (0%-100%) of 250 mM hydrogen peroxide (as an oxidizing substance) were mixed with 3 mM uric acid (in 10 mM NaOH) (as an antioxidant). The absorption of 10 ml samples was measured with an enzyme-linked immunosorbent assay (ELISA) reader at 450 nm for the reference, 630 nm, and the values of PAB were expressed in arbitrary (Hamidi. Koliakos [H.K]) units). 2.6 Statistical methods The normality of the data was assessed by the Kolmogorov-Smirnov test. The mean and SD (for normal distribution) and median and interquartile range (IQR) (for non-normal distribution) were used to describe the study variables. The independent student t-test (for variable normality distribution) was used to compare the mean of study variables between case and control groups. A logistic regression method was used to determine the variables related to diabetes, including age, BMI, PAB, systolic blood pressure (SYSp), diastolic blood pressure (DIAp), GLUCOSE (Glc), insulin, InsulinR, TC, LDL, HDL, TG, hs-CRP and vitamin D. Based on the Hosmer-Lemeshow method, simple logistic regression was utilized to determine the relationship between study variables and diabetes. Then, the variables with p < .2 were added to the final model and analysed using multiple logistic regressions. We used SPSS for Windows software (version 18 software package SPSS Inc). A p-value less than .05 was considered statistically significant. 3 RESULTS 3.1 Participants' characteristics and demographic findings All data showed a normal distribution. Demographic data, including age, BMI, SYSp and DIAp, were not significantly different between the two groups. Except for serum TC, insulin, vitamin D, WC, WHR and WHtR, other laboratory findings in diabetic subjects were significantly different from the non-diabetic subjects (p < .05). Table 1 shows the features of the two groups. TABLE 1 Demographic, cardiovascular risk parameters, and lipid profile indices in diabetic patients and healthy controls Variables Diabetic (n = 99) Control (n = 100) p-value Age (y) 65.33 +- 5.34 61.20 +- 5.990 .283 BMI (kg/m2) 26.6 +- 2.0 26.2 +- 2.4 .253 SYSp (mmHg) 13.03 +- 1.16 13.22 +- 1.20 .951 DIAp (mmHg) 7.49 +- 0.66 7.53 +- 0.75 .729 Glc (mmol/L) 198.72 +- 69.78 94.87 +- 5.68 <.001* TC (mg/dl) 152.56 +- 10.77 150.19 +- 10.68 .122 LDL (mg/dl) 144.72 +- 33.27 131.70 +- 31.43 .005* HDL (mg/dl) 46.03 +- 9.03 49.07 +- 10.10 .026* TG (mg/dl) 166.19 +- 37.01 155.59 +- 26.76 .022* Vit D (ng/ml) 19.29 +- 10.58 19.63 +- 8.12 .798 PAB (H.K) 0.40 +- 0.29 0.22 +- 0.13 <.001* hs-CRP (mg/dL) 5.11 +- 6.03 2.96 +- 3.07 .002* Insulin R** (mU/mL x mmol/L) 4.10 +- 4.22 2.63 +- 2.52 .003* (%) 1.09 +- 0.89 1.85 +- 1.38 <.001* QUICKI 0.32 +- 0.03 0.36 +- 0.04 <.001* Insulin (mU/ml) 9.33 +- 6.31 8.45 +- 6.48 .335 WC (cm) 95.49 +- 7.19 94.31 +- 10.3 .352 HC (cm) 102.54 +- 7.18 101.64 +- 5.88 .337 WHR 0.93 +- 0.07 0.93 +- 0.11 .767 WHtR 0.56 +- 0.04 0.55 +- 0.06 .717 Note: Values represent means +- SD. Comparisons were made by using Student's t-test between groups. Abbreviations: BMI: Body mass index; DIAp: Diastolic blood pressure; Glc: Glucose; HC: Hip Circumference; HDL-C: High-density lipoprotein cholesterol; HOMA-b: Homeostasis Model Assessment-b cell; hs-CRP: High-sensitivity C-reactive protein; Insulin R: Insulin Resistance; LDL-C: Low-density lipoprotein cholesterol; PAB: Pro-oxidant-antioxidant balance; QUICKI: Quantitative Insulin Sensitivity Check Index; SYSp: Systolic blood pressure; TC: Total cholesterol; TG: Triglycerides; Vit D: Vitamin D; WC: Waist circumference; WHR: Waist-to-hip ratio; WHtR: Waist-to-height ratio. *Significance was defined as p < .05. ** Insulin resistance was calculated using the HOMA equation (HOMA-IR). 3.2 PAB values, hs-CRP concentration and insulin resistance among postmenopausal women Serum PAB levels in the diabetic subjects were significantly higher than in the control group (p < .001) (Table 1). Also, serum hs-CRP concentrations were statistically different in the two groups (p = .002) (Table 1). Unsurprisingly, in diabetic women, there was a statistically significant difference in insulin resistance, HOMA-b and QUICKI compared to non-diabetic women (all p < .05), whereas no considerable difference was demonstrated between diabetic patients and healthy participants in serum insulin concentrations (p = .335). 3.3 The relationship between serum PAB values, BMI, and hs-CRP concentrations and other laboratory parameters As shown in Table 2, the Pearson correlation coefficient analysis was performed to evaluate the correlation between serum PAB values, BMI, hs-CRP concentrations and other laboratory parameters. Scatter plots graphically showed a strong and positive uncorrected association between serum PAB values and hs-CRP levels (r = .258 and p = .010) . We did not find any significant correlation between PAB values and insulin resistance (r = .095 and p = .347) . Moreover, serum PAB and hs-CRP levels were positively correlated with serum insulin (r = .212, p = .035; r = .211, p = .037), respectively. Among the other study factors, a significant association was observed between serum PAB values and LDL-C levels (r = .209, p = .038) and a negative correlation with HDL-C levels (r = -0.224 and p = .026). Moreover, a comparison of the relationship between BMI and other values showed a significant correlation between BMI and TG levels (r = .207 and p = .042). In addition, we did not find any association between vitamin D levels and other laboratory parameters listed in this study. TABLE 2 Pearson Correlation Coefficient between study variables in the case group Age BMI Insulin InsulinR LDL HDL TG VitD hs-CRP PAB BMI 0.066 1 0.514 Insulin 0.004 0.001 1 0.967 0.999 InsulinR 0.022 -0.02 0.176 1 0.825 0.847 0.082 LDL-C 0.128 0.114 0.217* -0.021 1 0.206 0.261 0.031 0.834 HDL-C 0.083 0.146 -0.055 0.054 0.039 1 0.415 0.148 0.59 0.598 0.703 TG 0.039 0.207* 0.206* -0.003 0.02 0.022 1 0.701 0.042 0.041 0.975 0.847 0.825 VitD 0.073 0.004 -0.134 0.023 -0.131 -0.083 0.04 1 0.472 0.972 0.185 0.821 0.195 0.412 0.693 hs-CRP -0.02 -0.143 0.211* 0.051 0.076 -0.026 0.025 0.005 1 0.848 0.159 0.037 0.618 0.457 0.801 0.805 0.963 PAB 0.072 -0.023 0.212* 0.095 0.209* 0.224* 0.019 0.002 0.258* 1 0.478 0.819 0.035 0.347 0.038 0.026 0.920 0.984 0.010 Glc 0.055 0.105 -0.117 -0.138 0.136 -0.016 0.017 0.001 -0.063 0.034 0.592 0.302 0.25 0.174 0.181 0.876 0.865 0.997 0.536 0.741 Note: Pearson correlation univariate analysis was used to test the relationship between parameters. Abbreviations: BMI, Body mass index; Glc, Glucose; HDL-C, High-density lipoprotein cholesterol; hs-CRP, High-sensitivity C-reactive protein; Insulin R, Insulin Resistance; LDL-C, Low-density lipoprotein cholesterol; PAB, Pro-oxidant-antioxidant balance; TG, Triglycerides; Vit D, Vitamin D. *Significance was defined as p < .05. FIGURE 1 Pro-oxidant-antioxidant balance in case and control groups FIGURE 2 High-sensitivity C-reactive protein in patients and healthy subjects 3.4 Multiple logistic regressions Logistic regression in the backward approach explained that InsulinR (OR: 1.16, p: 0.012), cholesterol (OR: 1.033; p: 0.047) and LDL-C (OR: 1.017; p: 0.002) levels, and PAB values (OR: 174.89; p < .001) had a positive association with diabetes mellitus in patients compared to non-diabetic women (Table 3). Moreover, these results showed that diabetes had an inverse association with HDL-C (OR: -0.932; p < .001). TABLE 3 Association between study variables and diabetes using multiple logistic regressions Variable OR 95% CI p-value Multiple logistic regression (entered approach, pseudo R 2 = .396) InsulinR 1.151 1.019-1.300 .024 Cholesterol 1.032 0.998-1.066 .063 LDL-C 1.015 1.004-1.026 .007 HDL-C 0.935 0.900-0.973 .001 TG 1.010 0.999-1.022 .073 Vit D 0.997 0.962-1.033 .877 PAB 140.451 16.426-1200.97 <.001 hs-CRP 1.064 0.971-1.166 .186 Multiple logistic regression (backward approach, pseudo R 2 = .373) Insulin R 1.165 1.034-1.313 .012 Cholesterol 1.033 1.01-1.067 .047 LDL-C 1.017 1.006-1.028 .002 HDL-C -0.932 0.897-0.968 <.001 PAB 174.893 21.563-1418.518 <.001 Abbreviations: CI, Confidence interval; HDL-C, High-density lipoprotein cholesterol; hs-CRP, High-sensitivity C-reactive protein; Insulin R, Insulin Resistance; LDL-C, Low-density lipoprotein cholesterol; PAB, Pro-oxidant-antioxidant balance; TG, Triglycerides; Vit D, Vitamin D. Significance was defined as p < .05. 4 DISCUSSION To our knowledge, this is the first case-control study to report PAB values and investigate the relationship between hs-CRP levels and PAB values in postmenopausal women with and without diabetes mellitus. The main finding of the present study was the serum PAB and hs-CRP elevation in diabetic postmenopausal women compared to non-diabetic cases. This finding is in accordance with earlier studies demonstrating the presence of systemic inflammation in diabetes. The increased level of OS is significantly associated with metabolic parameters in diabetic patients. 11 , 12 OS can be induced by inflammation 4 , 13 ; for example, higher concentrations of interleukin-6 are an important stimulant for the production of hs-CRP 14 and inflammation can induce the production of free radicals. 15 The present study showed that serum hs-CRP levels were positively associated with serum PAB values in diabetic women. Moreover, earlier reports support the presence of high OS and hs-CRP levels in stroke, cardiovascular and beta-thalassemia patients. 16 , 17 There is strong evidence of the correlation between inflammation and OS because both factors contribute to the pathogenesis of diabetes. 18 Moreover, diabetic postmenopausal women also had higher levels of blood glucose and HOMA-IR index. In correlation with previous studies, dysregulated lipid metabolism in diabetics has been reported, which could be attributed to increased lipolysis due to impaired insulin function in adipose tissue. In addition, the accumulation of free fatty acids in the liver leads to the high hepatic synthesis of TGs and results in hypertriglyceridemia. 11 , 19 In this study, as shown by Barrett-Connor et al., 20 no relationship was observed in total cholesterol between diabetic and non-diabetic subjects. We did not find any significant difference between serum hs-CRP, glucose, TG, LDL-C levels, and BMI. These results were inconsistent with those of Yang et al. 21 The reason may be due to the menopause subjects and the changes in the oestrogen hormone and its function in the liver. Moreover, parallel to our report, earlier reports have suggested that OS plays a major role in developing insulin resistance. 22 , 23 Consistent with many studies, 23 , 24 we can suggest that diabetic women have significantly altered lipid profiles than healthy postmenopausal subjects. Contrary to our work, many studies have reported that increased BMI values were strongly associated with hs-CRP and OS levels. 25 We suggest that independent of BMI, OS may also be an essential determinant of hs-CRP levels in diabetic people. Therefore, the link between OS and hs-CRP levels may involve pathways unrelated to BMI. In line with the study by Goodarzi et al., 7 there was no significant difference in BMI between the two groups. Moreover, consistent with Zaman et al., the patient and control groups were overweight but not obese. 26 Overweight women are not necessarily diabetic, and diabetes mellitus is not the only reason for the BMI increase in overweight type 2 diabetics; other factors may be involved. In addition, in line with our study, many studies have shown that people with diabetes also have a low BMI, and some have a very low BMI. 26 , 27 On the contrary, unlike some studies, 28 our study found that diabetes mellitus in our diabetic patients was not necessarily dependent on insulin. Therefore, it can be concluded that in people with type 2 diabetes, other factors may have a role in the incidence of diabetes. Hence, it can alter insulin levels in people who have diabetes without a statistically considerable difference from healthy subjects. In contrast to previous literature, 29 , 30 our findings demonstrated a positive relationship between serum hs-CRP and insulin levels because inflammatory markers decrease insulin secretion and signalling in peripheral tissues. Moreover, interleukin-6 decreases insulin signalling in the liver. 31 In the present study, we found an irreversible correlation between PAB values and HDL-C levels in line with A. Cagnacci et al. 32 because oxidants can be reduced by the antioxidant enzyme paraoxonase carried by HDL-C lipoproteins. 33 , 34 Moreover, we found a significant relationship between TG levels and BMI. This finding demonstrated that high TG can cause obesity and ultimately increase BMI in diabetic postmenopausal women. Besides, in contrast to the Cardiovascular Health Study and research by Mendall et al., surprisingly, no relationship was found between hs-CRP levels and BMI in women. Due to this controversy with the prior investigation, we think that diabetes in postmenopausal women can cause these outcomes. Our finding was in agreement with that of Kahn et al., 35 , 36 indicating that diabetic postmenopausal women were characterized by insulin resistance. Moreover, it has been noted that insulin has a significantly negative relationship with higher hs-CRP levels and PAB values. However, in Table 3, PAB values showed a positive correlation with LDL-C levels and an irreversible association with HDL-C levels. Therefore, the evidence supporting these results is that HDL cholesterol is the major lipoprotein carrier of antioxidant enzymes, and LDL is the main factor correlated with oxidative markers. Our study had a few limitations. The present work focused only on PAB values. However, several other factors can affect these biochemical parameters in OS, including sex hormones. Another limitation was the small sample size. 5 CONCLUSIONS We found significantly higher PAB values in diabetic postmenopausal women. Moreover, we demonstrated that increased hs-CRP concentrations are strongly associated with PAB values, a reliable OS marker. This finding was independent of BMI and insulin resistance in diabetic postmenopausal women. Measurement of PAB hs-CRP levels and other biochemical parameters may be a valuable marker for OS and inflammation and a helpful diagnostic factor to prevent injury and develop coronary artery disease. Future studies with larger sample sizes on PAB values and hs-CRP may lead to the more practical use of these two markers in clinical diagnosis and follow-up of diseases and better the quality of life for patients. AUTHOR CONTRIBUTIONS Hassan Ehteram: Conceptualization (supporting); writing - review and editing (equal). Sara Raji: Data curation (equal); writing - original draft (equal); writing - review and editing (equal). Mina Rahmati: Data curation (equal); writing - review and editing (equal). Hanieh Teymoori: Data curation (equal); writing - review and editing (equal). Samaneh Safarpour: Data curation (equal); writing - review and editing (equal). Nahid Poursharifi: Data curation (equal); writing - review and editing (equal). Mona Hashem Zadeh: Data curation (equal); writing - original draft (equal); writing - review and editing (equal). Reza Pakzad: Formal analysis (lead); writing - review and editing (equal). Hossein Habibi: Writing - review and editing (equal). Naser Mobarra: Conceptualization (lead); supervision (lead); writing - review and editing (equal). FUNDING INFORMATION This study is funded by Mashhad University of Medical Sciences (Grant No: 981826) CONFLICT OF INTEREST The authors declared no conflicts of interest. ETHICAL APPROVAL The Ethics Committee of Mashhad University of Medical Sciences approved the study (IR.MUMS.REC.1399.533). ACKNOWLEDGEMENTS The authors are particularly grateful to the patients and their family members who volunteered to participate in this study. DATA AVAILABILITY STATEMENT The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. |
PMC10000639 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051103 foods-12-01103 Article Impact of Longkong Pericarp Extract on the Physicochemical Properties of Alginate-Based Edible Nanoparticle Coatings and Quality Maintenance of Shrimp (Penaeus monodon) during Refrigerated Storage Charoenphun Narin Conceptualization Software Validation Formal analysis Data curation Writing - original draft Funding acquisition 1 Rajasekaran Bharathipriya Methodology Software Data curation Writing - original draft Writing - review & editing 2 Palanisamy Suguna Methodology Investigation Writing - original draft 2 Venkatachalam Karthikeyan Conceptualization Validation Formal analysis Resources Data curation Writing - original draft Writing - review & editing Visualization Supervision Project administration Funding acquisition 3* Pan Jinfeng Academic Editor Zhang Yi-Qi Academic Editor 1 Faculty of Science and Arts, Burapha University, Chanthaburi Campus, Chanthaburi 22170, Thailand 2 International Center of Excellence in Seafood Science and Innovation, Faculty of Agro-Industry, Prince of Songkla University, Songkhla 90110, Thailand 3 Faculty of Innovative Agriculture and Fishery Establishment Project, Prince of Songkla University, Surat Thani Campus, Surat Thani 84000, Thailand * Correspondence: [email protected] or [email protected] 05 3 2023 3 2023 12 5 110331 1 2023 27 2 2023 03 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). The objective of this study was to evaluate the impact of varying concentrations of longkong pericarp extract (LPE) on the physicochemical properties of alginate-based edible nanoparticle coatings (NP-ALG) on shrimp. For developing the nanoparticles, the alginate coating emulsion with different LPE concentrations (0.5, 1.0, and 1.5%) was ultrasonicated at 210 W with a frequency of 20 kHz for 10 min and a pulse duration of 1s on and 4 off. After that, the coating emulsion was separated into four treatments (T): T1: Coating solution containing basic ALG composition and without the addition of LPE or ultrasonication treatment; T2: ALG coating solution converted into nano-sized particles with ultrasonication and containing 0.5% LPE; T3: ALG coating solution converted into nano-sized particles with ultrasonication and containing 1.0% LPE; T4: ALG coating solution converted into nano-sized particles with ultrasonication and containing 1.5% LPE. A control (C) was also used, where distilled water was used instead of ALG coating. Before coating the shrimp, all the coating materials were tested for pH, viscosity, turbidity, whiteness index, particle size, and polydispersity index. The control samples had the highest pH and whiteness index and was followed by the lowest viscosity and turbidity (p < 0.05). Among the T1-T4 coating materials, T4 coating had higher turbidity, particle size, polydispersity index, but lower pH, viscosity, and whiteness index (p < 0.05). To study the quality and shelf-life of the shrimp, all coated shrimp samples were refrigerated at 4 degC for a period of 14 days. At 2-day intervals, physiochemical and microbial analyses were performed. The coated shrimp also had a lower increase in pH and weight loss over the storage period (p < 0.05). Coatings containing 1.5% LPE significantly reduced the polyphenol oxidase activity in the shrimp (p > 0.05). The addition of LPE to NP-ALG coatings demonstrated dose-dependent antioxidant activity against protein and lipid oxidation. The highest LPE concentration (1.5%) led to increased total and reactive sulfhydryl content, along with a significant decrease in carbonyl content, peroxide value, thiobarbituric acid reactive substances, p-anisidine, and totox values at the end of the storage period (p < 0.05). Additionally, NP-ALG-LPE coated shrimp samples exhibited an excellent antimicrobial property and significantly inhibited the growth of total viable count, lactic acid bacteria, Enterobacteriaceae, and psychotropic bacteria during storage. These results suggested that NP-ALG-LPE 1.5% coatings effectively maintained the quality as well as extended the shelf-life of shrimp during 14 days of refrigerated storage. Therefore, the use of nanoparticle-based LPE edible coating could be a new and effective way to maintain the quality of shrimp during prolonged storage. edible coating antimicrobial activity antioxidant activity shelf-life extension quality attributes This article received no external funding. pmc1. Introduction Shrimp has become increasingly popular among consumers for its distinctive taste, but it is not just a delicacy. It is also a rich source of protein, healthy PUFA, vitamins, and minerals . The high moisture content and various nutrients in shrimp makes it more vulnerable to changes in physical, biochemical, and microbiological characteristics and such alterations can negatively impact the quality and shelf-life of shrimp, ultimately resulting in reduced market value . Several chemicals including ethylenediaminetetraacetic acid, benzoic acid, polyphosphates, ascorbic acid, and sodium chloride have been tried as preservatives to maintain the quality and extend the shelf-life of shrimp during storage. However, sulfite-based formulation produces allergic reactions, thus synthetic preservatives exert an adverse effect on human health. Moreover, to preserve the quality of shrimp during prolonged storage, various advanced techniques have been studied, such as modified atmospheric packaging, vacuum packaging, high-pressure treatment, the use of plant extract, and applying edible coatings . The edible coating is a promising and reliable method for extending the shelf-life of shrimp among the various techniques studied . It creates a barrier against oxygen, slows down oxidation, prevents microbial contamination, reduces moisture loss, and preserves flavor . Edible coatings also serve as a means of delivering food additives, such as antioxidants and antimicrobial agents. Using natural plant extracts in these coatings can extend the shelf-life of perishable foods, such as fruits, vegetables, and seafood . Alginate (ALG), a polysaccharide derived from brown algae, is widely used as a coating material. It is considered a GRAS (generally recognized as safe) substance and is composed of D-mannuronic acid and L-guluronic acid. Alginate is popular as a coating material because it can form a strong gel and maintain its insolubility when reacting with multivalent cations . Many studies have shown that incorporating preservatives into coatings can help prevent quality changes in perishable foods during prolonged storage . However, consumers often prefer preservatives that come from natural sources to ensure the safety of their food. Phenolic compounds, which are widely found in plants, are one such example of these natural preservatives. Phenolic compounds are well known for their antimicrobial and antioxidant properties, making them a potent alternative to synthetic agents . Longkong (Aglaia dookkoo Griff.) is an economically valuable, non-climacteric tropical fruit belonging to the Meliaceae family, primarily found in southern Thailand . Longkong fruit is composed of three main parts--pericarp, flesh, and seeds--and the pericarp contains a high level of polyphenols. Studies indicate that longkong pericarp extract (LPE) exhibits multiple biological and pharmacological effects, including radical scavenging, germicidal, cytostatic, antimalarial, and depigmentation . In addition, LPE contains a rich source of lansic acid, lansiosides, lansiolic acid, and iso-onoceratriene. These chemical compounds in LPE could control the hormonal imbalance in humans and promote anti-baldness, antipyretic and anti-feeding activities. Nevertheless, ellagic acid and corilagin in the LPE could promote anti-fibrosis and anti-glaucoma effects. Incorporating nanotechnology and natural plant extracts into the coating medium is an effective approach to preserve the stability of bioactive compounds and as well as food products against deterioration from oxidation and microorganism . Additionally, the use of nanotechnology to produce particles of nano-dimension increases the surface area per unit weight, which results in better dispersion of the active substances in the coating medium, thereby enhancing its functionality and bioactivity . Furthermore, the use of nano-sized particles allows for the controlled and gradual release of bioactive substances during prolonged storage . Additionally, nanotechnology can incorporate active substances without altering the sensory characteristics of foods and increase their shelf-life . Although, the antibacterial activity of LPE and alginate coating is well known, their combination, especially in the nanoparticles system, has not been studied. Therefore, the present study utilized this opportunity to examine the properties of LPE-added ALG-based edible nanoparticles coating and to investigate their preventive effect on the quality maintenance of shrimp during 14 days of storage at refrigerated conditions. 2. Materials and Methods 2.1. Raw Materials, Chemicals, and Reagents Longkong fruits fully ripe were harvested from a local garden in Surat Thani province, Thailand. Black tiger shrimp (Penaeus monodon) measuring 6-8 cm in length were purchased from a nearby farm. The shrimp were placed on ice at a ratio of 1:2 (shrimp:ice) and transported to the lab within an hour. Upon arrival, the shrimp were rinsed with cold water and kept on ice until use, not exceeding 5 h. Food-grade sodium alginate (Keltone LV, ISP, San Diego, CA, USA) was used as a biopolymer for coating formulation. The analytical-grade solvents and chemical agents utilized in this study encompassed chloroform, ethanol, methanol, ethyl acetate, sodium hydroxide, Triton X-100, sodium chloride, 1-3,4-dihydroxyphenylalanine (DOPA), Tris, glycine, sodium ammonium sulfate, 5,5'-dithio-bis-(2-nitrobenzoic acid) (DNTB), ethylenediamine tetra acetic acid (EDTA), guanidine chlorate, dipotassium hydrogen phosphate, sulfosalicylic acid, potassium dihydrogen phosphate, hydrochloric acid, potassium iodide, sodium thiosulfate, urea, anhydrous sodium sulfate, trichloroacetic acid (TCA), ascorbic acid, thiobarbituric acid (TBA) (Merck, Darmstadt, Germany), Tween 80 (Labchem, Zelienople, PA, USA), 1,1,3,3-tetramethoxypropane (MDA) (Sigma-Aldrich, St. Louis, MO, USA), and acetic acid (Lab-Scan, Pathum Wan, Bangkok, Thailand). The media used for the microbiological analyses, namely plate count agar, peptone, deMan, Rogosa, Sharpe (MRS) agar, and violet red bile glucose agar (VRBG), were all analytical grade and purchased from Merck, Darmstadt, Germany. 2.2. Preparation of Longkong Pericarp Extract (LPE) Upon the arrival of the longkong fruits, their pericarps were isolated from the flesh and washed using cold water with 2% ascorbic acid. The longkong pericarp extract (LPE) was obtained as guided by Nagarajan et al. with some modifications. First, the pericarps were dried in a hot air oven at 40 degC until a consistent weight was reached, and the dried sample was grounded into fine powder. To prepare the LPE, 5 g of pericarp powder was mixed with 100 mL of absolute ethanol. The mixture was then heated with agitation in a water bath at 40 degC for 4 h. The resulting solution was then placed in the solvent evaporator to remove the solvent thoroughly at 40 degC. The final LPE was freeze-dried, stored in a sealed amber bottle, and kept at -20 degC until required for further experimentation. 2.3. Preparation of Coating Solution The ALG coating solution was prepared by following the method of Sharifimehr et al. with some modifications, dissolving 1% ALG (w/v) and 4% Tween 80 (v/v) in distilled water, and then adding LPE at different concentrations while constantly stirring. The mixture was thoroughly stirred using a magnetic stirrer to obtain a homogenous solution. The particle size of the coating was reduced to the nanoscale using ultrasonication (Hielscher UP200Ht, Hielscher Ultrasonics GmbH, Teltow, Germany) at 210 W, a frequency of 20 kHz for 10 min with a pulse duration of 1 s on and 4 s off. The temperature was maintained at 25 degC during the ultrasonication process. Four treatments (T) were used: T1: Coating solution containing basic ALG composition and without the addition of LPE or ultrasonication treatment; T2: ALG coating solution converted into nano-sized particles with ultrasonication and containing 0.5% LPE; T3: ALG coating solution converted into nano-sized particles with ultrasonication and containing 1.0% LPE; T4: ALG coating solution converted into nano-sized particles with ultrasonication and containing 1.5% LPE. A control (C) was also used, where distilled water was used instead of ALG coating. All the tested coating solutions were measured for physicochemical properties as shown in Section 2.5. 2.4. Physicochemical Analysis of Coating Solutions 2.4.1. pH The pH of the coating solution was measured using a digital pH meter (Mettler-Toledo GmbH, Giessen, Germany). 2.4.2. Viscosity To determine the viscosity of the coating solution, a digital tabletop Brookfield viscometer (Brookfield DVE viscometer, Middleborough, MA, USA) was utilized. First, 150 mL of the coating solution was placed in a beaker measuring 70 mm in diameter and 125 mm in height. A Viscometer equipped with a number 2 spindle, set to run at a speed of 12 rpm, was used to measure viscosity. The outcomes were expressed in centipoise (cP). 2.4.3. Turbidity The turbidity of the coating solution was measured using a turbidimeter (Hanna Instruments, model HI 93703, Woonsocket, RI, USA), and the results were expressed in percentages. 2.4.4. Whiteness Index The HunterLab colorimeter (MiniScan EZ 4000, Hunter Associates, Inc., Reston, VA, USA) was utilized to determine the color of the coating solution following the methodology outlined by Josewin et al. . After calibrating the instrument using a white standard plate (L = 91.83, a = -0.73, b = 1.52), the lightness (L*), redness/greenness (a*), and yellowness/blueness (b*) were measured. The whiteness index (WI) was subsequently calculated using the following formula:(1) WI=100-(100-L*)2+(a*)2+(b*)2 2.4.5. Particle Size and Polydispersity Index (PDI) The particle size and polydispersity index (PDI) were assessed using a method adapted from Venkatachalam . Backscatter detection at a 170deg scattering angle was utilized in the process. The sample was stabilized within the device for 60 s before data collection at 25 degC. The particle size outcomes were reported in nanometers, while the PDI values were expressed as the polydispersity index. 2.5. Coating and Storage of Shrimps The schematic illustration of the ALG coating preparation and coating of the shrimp is shown in Figure 1. The shrimp at a refrigerated temperature (4-8 degC) were coated by immersing them fully in a respective coating solution, as described in Section 2.4, and subsequently, the coated shrimps were placed on a wire rack for 1 min to allow excess coating material to drip off. The drying process of the coating material on the surface of the shrimp was facilitated by exposing coated shrimps to air blown by an electric fan. Next, the shrimps were arranged in polystyrene trays with 12 shrimps per tray, then wrapped with polyolefin film, and kept at 4 degC for 14 days. At every two-day interval during the storage period, samples were chosen randomly for shelf-life analysis as per the protocol described in Section 2.6. 2.6. Shelf-Life Analyses of NP-ALG-LPE Coated Shrimp during Storage 2.6.1. Weight Loss Prior to weighing, the stored shrimps were first dried off from any surface moisture by using a clean paper towel. Then, they weighed on an electronic weighing balance (SECURA124-1CIT, Sartorius, Goettingen, Germany). The following formula was used to determine the weight loss (WL) in the samples :(2) WL (%)=W1-W2W1x100 where W1 and W2 denote the weight of shrimp on the initial and final day of storage, respectively. 2.6.2. pH The pH of the shrimp samples was determined using a digital pH meter (Mettler-Toledo GmbH, Giessen, Germany). To prepare the samples, 5 g of shrimp meat was homogenized in 25 mL of distilled water for 1 min. The homogenate was then filtered through a muslin cloth and the filtrate was collected and measured according to Ebadi et al. . 2.6.3. Extraction of Polyphenol Oxidase The crude polyphenol oxidase (PPO) was extracted in accordance with the method of Basiri et al. . PPO from the cephalothorax was pulverized in the presence of liquid nitrogen and mixed using a pestle and mortar with 0.05 mol/L sodium phosphate buffer (pH 7.2) containing 1.0 mol/L NaCl and 0.2 g/100 mL Triton X-100 (1:3 w/v ratio). The mixture was mixed continuously for 30 min, followed by refrigerated centrifugation for 30 min at 8000x g. The supernatant was subjected to 40% saturation with sodium ammonium sulfate and left to stand at 4 degC for 30 min and centrifuged (12,000x g for 30 min at 4 degC) to collect the pellet. After centrifugation and precipitation with sodium ammonium sulfate, the resulting pellet was dialyzed using buffer overnight with buffer changes (15 volumes with three changes), and the insoluble material was collected by centrifugation (3000x g at 4 degC for 30 min). The resulting supernatant served as the "crude PPO extract." To determine PPO activity, the method described by Simpson et al. was followed using DOPA as the substrate. Specifically, 100 mL of the crude PPO extract was mixed in 1100 mL of the buffer solution containing 400 mL of 0.05 mol/L phosphate buffer (pH 6.0), 600 mL of 15 mmol/L DOPA, and 100 mL of deionized water. The mixture was incubated at 45 degC and the increase in absorbance at 475 nm was monitored using a UV-160 spectrophotometer (Shimadzu, Kyoto, Japan) for 3 min at 30-s intervals to measure the formation of dopachrome. PPO activity was defined as an increase in absorbance of 0.001 at 475 nm and expressed as one unit of PPO activity. The control was run in the same manner, except that deionized water was used instead of the PPO extract. The relative PPO activity was determined as the residual activity compared to the control by using the following formula:(3) Relative activity (%)=BAx100 where A: PPO activity of control; B: PPO activity of the sample. 2.6.4. Total Sulfhydryl Content The shrimp protein isolate was prepared based on the procedure by Liu et al. . The procedure involved dissolving the sample in a buffer (0.086 M Tris, 0.09 M glycine, 4 mM EDTA) at pH 8 and centrifuging at 10,000x g for 15 min to eliminate any insoluble protein. The total sulfhydryl content was then quantified via Ellman's technique by mixing 4.5 mL of the supernatant with 0.5 mL of Ellman's reagent (10 mM DTNB) and the absorbance was read at 412 nm with a spectrophotometer (RF-15001, Shimadzu, Kyoto, Japan). The sulfhydryl content was quantified in mmol sulfhydryl/g protein using a molar extinction coefficient of 13,600 M-1 cm-1. In addition, the protein content of the isolate was determined using the Biuret method . Reactive Sulfhydryl Content The samples were placed in a solubilizing buffer (0.086 M Tris, 0.09 M glycine, 4 mM EDTA, 8 M urea) at pH 8 and then centrifuged at 10,000x g for 15 min to remove any insoluble protein. The reactive sulfhydryl content was determined using the DTNB assay . Specifically, the supernatant (4.5 mL) was mixed with 10 mM DTNB (0.5 mL), then the absorbance was recorded at both 412 nm and 540 nm using a spectrofluorometer (RF-15001, Shimadzu, Kyoto, Japan). The reactive sulfhydryl value was calculated using the following equations:Reactive sulfhydryl content (mmol/g) = 73.53 x (A412 - 1.6934 x A532 + 0.009932)(4) where A412 and A532 were the absorbance 412 nm and 532 nm of the assay solution, respectively. 2.6.5. Carbonyl Content Carbonyl groups were detected by reactivity with 2,4-dinitrophenylhydrazine (DNPH). The method of Parrilla-Taylor et al. was used. The samples were initially homogenized with 1 mL of 5% sulfosalicylic acid and centrifuged at 15,000x g, then the supernatant was discarded and the pellet was mixed with 10 mM DNPH in 2 M HCl and allowed to incubate at 25 degC for 1 h. The protein was precipitated by adding 0.5 mL of 20% TCA and centrifuged for 5 min at 15,000x g. The HCl was used as blanks. The resulting pellets were washed three times with a solution of ethanol and ethyl acetate (1:1, v/v) and resuspended in 6 M guanidine chlorate, followed by incubation at 37 degC for 15 min. The samples were then centrifuged for 5 min at 15,000x g, and the supernatant was collected for spectrophotometric measurement of protein carbonyl content at the maximum absorbance within the range of 360-401 nm. The results were reported as nmol of carbonyl proteins per gram of sample. 2.6.6. Peroxide Value (PV) The method used for lipid extraction in the shrimp was based on the approach described in the study of Bligh and Dyer . A 25 g sample was homogenized with a mixture of chloroform, methanol, and distilled water (50:100:50) at 9500 rpm for 2 min at 4 degC using an IKA labortechnik homogenizer (Model T18, Bangkok, Thailand). Then, 50 mL of chloroform was added and homogenized at 9500 rpm for 1 min and 25 mL of distilled water was added and homogenized again for 30 s. The homogenate was then centrifuged at 3000 rpm at 4 degC for 15 min using a refrigerated centrifuge (Beckman Coulter, Avanti J-E Centrifuge, Fullerton, CA, USA). Finally, the chloroform phase was collected. It was then transferred into a 125 mL Erlenmeyer flask that contained about 2-5 g of anhydrous sodium sulfate, shaken well, and decanted into a round-bottom flask through a Whatman No.4 filter paper. Subsequently, the solvent was evaporated at 25 degC using an EYELA N-100 rotary evaporator (Tokyo, Japan), and the remaining solvent was removed by nitrogen flush. To determine the peroxide value (PV), the Kim et al. method was adopted. The lipid sample (1.0 g) was combined with 25 mL of a solvent mixture (chloroform: acetic acid) at a ratio of 2:3 (v/v) and vigorously shaken, and 1 mL of saturated potassium iodide was introduced. After incubating the mixture in the dark for 5 min, 75 mL of distilled water was added and shaken followed by the addition of 0.5 mL of starch solution (1%, w/v) as an indicator. The mixture was then titrated against 0.01 N sodium thiosulfate solution, and the peroxide value was expressed as milliequivalents per kilogram of lipid. 2.6.7. Thiobarbituric Acid Reactive Substance (TBARS) The TBARS analysis was performed in accordance with the method of Benjakul and Bauer . To begin, 1 g of minced shrimp meat was combined with 9 mL of a 15% TCA solution that contained 0.375% TBA. The resulting mixture was then subjected to heating in boiling water for 10 min, followed by cooling with running water. After this, the mixture was centrifuged for 20 min at 4000x g, the supernatant was collected, and its absorbance was measured at 532 nm using a UV-160 spectrophotometer (RF-15001, Shimadzu, Kyoto, Japan). The TBARS value was then determined by calculating the standard curve of MDA (ranging from 0-2 ppm) and expressed as mg MDA per kilogram of shrimp meat. 2.6.8. Anisidine Value (AnV) The Anisidine value (AnV) was determined using the method developed by Okpala et al. . First, 100 mg of the lipid sample was dissolved in 25 mL of isooctane. Then, 2.5 mL of the resulting solution was mixed with 0.5 mL of 0.5% AnV in acetic acid (w/v) and kept in dark for 10 min. After that, the reaction mixture was read using a UV-Vis spectrophotometer (RF-15001, Shimadzu, Kyoto, Japan) at 350 nm and the following formula was applied to calculate the AnV value in the samples:(5) AnV=25 x ((1.2xA2)-A1W) where A1 and A2 represent the absorbances measured at 350 nm before and after the addition of AnV, respectively. Furthermore, W stands for the weight of the sample (g). 2.6.9. Total Oxidation Value The totox oxidation (TOTOX) value (TV) was determined using the protocol outlined by de Abreu et al. . The TV is determined by adding the peroxide value and Anisidine value (AnV), as follows:(6) TV=2PV+AnV 2.7. Total Volatile Basic Nitrogen (TVB-N) The Conway microdiffusion technique was used to evaluate the TVB-N level in shrimp using 0.1 M KOH. The results were expressed in milligrams of nitrogen per 100 g of shrimp. 2.8. Microbiological Analyses To conduct the microbiological tests, randomly selected shrimp samples were taken from the same tray, and each test was performed in triplicate. For these experiments, 10-g samples were transferred into sterile zipper bags containing 90 mL of peptone water and were homogenized using a stomacher at 250 rpm for 2 min. Subsequently, serial dilutions were made in test tubes containing 0.1% peptone water, and the diluted samples (0.1 mL) were then spread on the surface of dry media for microbial enumeration. Total viable plate counts (TVC) and psychrotrophic bacteria counts were determined by the pour plate method using plate count agar, which was incubated at 37 degC for 48 h and 7 degC for 10 days, respectively . Enterobacteriaceae was evaluated by the spread plate method using violet red bile glucose (VRBG, Merck, Darmstadt, Germany) agar incubated at 37 degC for 24 h. Lactic acid bacteria were enumerated by the spread plate method using MRS agar incubated at 30 degC for 72 h . The results were expressed in log CFU/g. 2.9. Statistical Analysis The statistical analysis in this study was conducted using a completely randomized design, with all experiments performed in triplicate. Mean values were compared using analysis of variance (ANOVA) and Duncan's multiple range test. The data were analyzed using the Statistical Package for Social Sciences (SPSS) version 6 for Windows (SPSS Inc., Chicago, IL, USA). 3. Results and Discussion 3.1. Physicochemical Properties of Different Coating Solutions 3.1.1. pH, Viscosity, and Turbidity The pH values of the tested coating samples are illustrated in Figure 2A. On average, the pH of the coatings varied between 6.1 and 7.1. The pH of the control and T1 coating was found to be in the neutral range (p < 0.05), while the pH of the NP-ALG-LPE coating (T2-T4) was observed to be impacted by the LPE concentration, showing a decrease in pH as the LPE concentration increased from 0.5% to 1.5% (p < 0.05). Of all the samples, T4 had the lowest pH (at 6.2) value, while the control had the highest pH (at 7.1) value. Generally, the pH of longkong pericarp is slightly acidic (at 4.71) in nature , which may have affected the pH of the NP-ALG-LPE coatings. A study by Chen et al. found that the presence of organic acids in longkong pericarp, such as glycolic acid, malic acid, and citric acid, can contribute to the low pH. Fresh longkong pericarp typically has a relatively stable pH due to its high moisture content . However, the extraction process may have concentrated the naturally occurring organic acids in the pericarp, leading to a decrease in the pH of the coatings . In addition, the use of ascorbic acid during extraction could have also lowered the pH of the LPE. A lower pH in the NP-ALG-LPE coatings can be beneficial as it can inhibit the growth of microorganisms. Gokoglu also noted that organic acids can reduce the pH and inhibit the growth of spoilage and pathogenic bacteria. Similarly, Baek et al. found that the pH of the ALG coating decreased with the addition of grapefruit seed extract due to the presence of organic compounds. The viscosity of the different coatings is shown in Figure 2B. In general, T1 had the highest viscosity, while the control had the lowest viscosity (p < 0.05). Among the NP-ALG-LPE coatings (T2-T4), no significant differences were observed, regardless of the LPE concentrations (p > 0.05). This could be due to the reduction of molecular weight of ALG caused by the ultrasonication process, which decreased the thickening properties of the ALG . In general, the viscosity of the coatings is primarily determined by the material used . The turbulence and cavitation effect of ultrasound causes an irreversible ordered-disordered conformation transition of ALG molecules, resulting in reduced particle size . However, a lower viscosity of the coating could provide a thinner film over the food samples. Pilon et al. stated that a thin layer over the product has better barrier properties during prolonged storage. As noted by Rodriguez-Turienzo et al. , the viscosity of a whey protein isolate-based coating was found to be decreased when ultrasonication treatment was applied in the coating emulsion and thus attributed to the formation of nano-sized particles. The turbidity levels of the different coatings are shown in Figure 2C. The NP-ALG-LPE coatings (T2-T4) that were subjected to ultrasonication had lower turbidity compared to T1 (p < 0.05). This can be attributed to the reduction of starch granules in the ALG caused by the ultrasonication process, which leads to a decrease in opacity . The overall transparency of NP-ALG-LPE coatings was improved as a result of ultrasonication. The solubility of ALG increased in samples T2, T3, and T4 due to ultrasonication, while T1 without ultrasonication had more undissolved particles . Jambrak et al. postulated that ultrasonication destroyed the crystalline region of starch granules, resulting in higher solubility of corn starch. Among the NP-ALG-LPE coatings (T2-T4), turbidity increased with augmenting concentrations of LPE from 0.5% to 1.5%, respectively (p < 0.05). Generally, the higher turbidity of the coating negatively impacts the product's appearance, thus reducing marketability . Increasing the coating ingredients promoted more polymers to be present in the coating, giving rise to greater light scattering . Moreover, the higher concentration of LPE (1.5%) in T4 had a larger particle size. Therefore, more light scattering increased the turbidity of the coating. 3.1.2. Whiteness Index, Particle Size, and Polydispersity Index Figure 3A displays the whiteness index (WI) of the tested coating samples. Generally, coatings with higher WI have a more desirable commercial appeal. The T1 coating had a higher WI than the NP-ALG-LPE coatings (T2-T4) with a significant difference (p < 0.05). The decrease in WI of NP-ALG-LPE coatings (T2-T4) was attributed to the yellowish-brownish color of LPE . As the concentration of LPE increased from 0.5-1.5%, the WI significantly decreased (p < 0.05). According to Lichanporn et al. , brown pigments are mainly present in the pericarp area of longkong fruits. Therefore, the higher amount of LPE (1.5%) used in the coating resulted in a reduction of whiteness. Figure 3B illustrates the particle size of various coatings. In general, the NP-ALG-LPE coatings (T2-T4) had smaller particle sizes than T1 with a significant difference (p < 0.05). The shear force created by ultrasonic cavitation caused a reduction in the particle size of polymers in the coatings . Among the NP-ALG-LPE coatings, the particle size increased as the concentration of LPE increased from 0.5% to 1.5% with a significant difference (p < 0.05). T2 had the smallest particle size, followed by T3 and T4 (p < 0.05). The particle size of NP-ALG-LPE coatings ranged from 218 to 260 nm. This is in line with the findings of Baek et al. and Lin et al. , who reported particle sizes of 206 nm and 247 nm for ultrasonicated ALG and chitosan-based coatings, respectively. Figure 3C shows the polydispersity index (PDI) of the tested coating samples. The PDI measures the heterogeneity of a coating based on its particle size . Typically, a lower PDI indicates a more uniform distribution and smaller size of polymers in the coating . In general, NP-ALG-LPE coatings (T2-T4) had a lower PDI compared to T1 with a significant difference (p < 0.05). This difference was likely due to the disruption of structural integrity, induction of dissociation, and degradation of ALG molecules caused by the cavitation effect of ultrasonication , resulting in a reduction in the size of ALG molecules. This implies that ultrasonicated NP-ALG-LPE coatings have a homogenous dispersion of particles in the medium. Among the NP-ALG-LPE coatings, T2 had the lowest PDI. The increase in LPE concentration from 0.5% to 1.5% resulted in larger particle size and uneven distribution, leading to a higher PDI in T4 coating, followed by T3, as compared to T2 (p < 0.05). This is consistent with the particle size results shown in Figure 3B. Thus, the application of ultrasound decreases the particle size and PDI of the coating. 3.2. Shelf-Life Analysis of Refrigerated Shrimp Using Different ALG-Based Coatings 3.2.1. Physicochemical Properties Figure 4A shows the weight loss of control and coated shrimps during storage at 4 degC. Generally, weight loss represents moisture loss of fresh food, impacting both economic benefits and product quality, making it a crucial factor for marketability . Weight loss in all samples rose as storage days progressed (p < 0.05). The probable cause of weight loss in the shrimp sample was due to the denaturation and degradation of protein through autolysis and microbial enzymes, resulting in structural breakdown during storage . Consequently, the water-holding capacity of shrimp muscle significantly decreased with storage time. At the end of storage, the T4-coated sample had the lowest weight loss as compared to the other samples. No significant difference was noted between the control and T1-coated sample until day 6 of storage (p > 0.05). Afterward, the T1-coated sample displayed a noticeable decrease in moisture loss compared to the control (p < 0.05) due to the barrier properties of ALG, which reduced moisture loss during storage . The semi-permeable coating layer acted as a barrier to the flow of O2, CO2, and H2O, thus reducing moisture loss . Moreover, the protein-polysaccharide complex formed between ALG, and the muscle protein of shrimp could enhance the overall water-holding capacity of shrimp during storage . In addition to this, dehydration needs to be considered because it is one of the important quality parameters during refrigerated storage. According to Song et al. , the primary reason for the decrease in dehydration of products coated with ALG is that the gel coating serves as a sacrificial agent. Thus, moisture in the gel evaporates before any substantial desiccation occurs in coated food. Additionally, no significant difference was observed among NP-ALG-LPE coated shrimp samples until the 10th day of storage. However, at the end of storage, the T4-coated sample had the lowest weight loss, followed by T3, when compared to T2 (p < 0.05). This was due to the formation of an additional layer on the shrimps, providing further resistance to mass transfer and slowing the increase in water loss after a certain storage period . As a result, the addition of LPE to the ALG coating on shrimp effectively reduced moisture loss during prolonged storage. pH is recognized to be one of the important indicators for identifying the changes in pH of control and coated shrimps during storage microbial spoilage in seafood or aquatic products . On day 0, the samples' pH was 6.6 to 6.9, consistent with previous reports . The slightly lowered pH of the NP-ALG-LPE-coated samples (T2-T4) was due to the reduced pH of the coatings with the addition of LPE (0.5-1.5%). This was in agreement with the pH of the NP-ALG-LPE coatings . Overall, pH values were gradually increased in all samples over the storage period. At the end of storage, the lowest pH was observed in the T4-coated shrimp, while the control sample had the highest pH (p < 0.05). The rise in pH was associated with the accumulation of alkaline compounds, primarily produced due to microbial activity. Ebadi et al. noted that endogenous enzymes and microorganisms breaking down protein in shrimp produce volatile bases like ammonia and trimethylamine, causing an increase in shrimp muscle pH during storage . T4-coated shrimp had a smaller increase in pH, which matched the controlled growth of the microorganisms. In general, shrimp is considered unacceptable if the pH exceeds 7.6 . The control sample reached this limit on day 6, while coated shrimp remained within acceptable limits throughout storage. Among NP-ALG-LPE-coated shrimp samples (T2-T4), the pH value of the samples was strongly decreased with increasing LPE concentration (p < 0.05). This suggests that LPE may help to slow microbial growth, thus reducing spoilage and decomposition. Shrimp coated with T4, which contains 1.5% LPE, showed lower pH due to the inhibitory effect of LPE against microbial action. This is because LPE contains high levels of polyphenols , and their activity against microbes is dependent on concentration. Polyphenols interact with microbial membranes, altering their permeability and functions and leading to cell death . 3.2.2. Brown Pigment Enzyme Activities Figure 5A displays the levels of PPO activities in both control and coated shrimp samples during storage at 4 degC. PPO is a major contributor to browning in the shrimp samples during prolonged storage. Enzymatic browning, also known as melanosis, affects the quality and acceptance of shrimps. PPO is the main cause of browning, converting colorless quinones into dark pigments . PPO triggers enzymatic browning by catalyzing the reaction between substrates, oxygen, phenolic compounds, and reaction by-products, causing black spot formation . Among the samples, the control exhibited the highest PPO activity. This was attributed to the basic mechanism of melanosis, in which PPO converts colorless monophenols into diphenols, which then react with oxygen to form highly colored quinones, leading to the formation of brown polymers through reaction with amino acids . Shrimp coated with T4 had the lowest PPO activity during storage due to the higher concentration of LPE (1.5%) compared to T3 (1.0%) and T2 (0.5%) (p < 0.05). The inhibition of PPO activity by phenolic compounds can occur through various mechanisms, including direct inhibition of PPO, scavenging of oxygen, and reduction of quinones back to diphenols to prevent melanin formation . Balti et al. found that shrimps coated with microalgal exopolysaccharides enriched with 1.5% red seaweed extract (contains rich sources of polyphenols) reduced PPO activity more effectively than those coated with 0.5% and 1.0% extract during cold storage. Hence, the results suggest that coating shrimp with 1.5% LPE effectively inhibits PPO activity during storage, making it a promising natural inhibitor. 3.2.3. Protein Oxidation Figure 6A,B depicts the total and reactive sulfhydryl groups of the control and coated shrimps during storage at 4 degC. Sulfhydryl is a highly active group found in myofibrillar protein, which possesses weak secondary bonds . Overall, sulfhydryl content decreased in all samples with increasing storage time (p < 0.05). This indicated that prolonged storage of shrimp had a significant effect on protein oxidation. However, the decrease of sulfhydryl groups was found lower in coated samples (T1-T4) when compared to the control (p < 0.05). Morachis-Valdez et al. also reported that carp fillets coated with chitosan showed less reduction in sulfhydryl content than uncoated ones during 5 months of storage at -18 degC . In general, protein oxidation leads to a decrease in sulfhydryl groups, which become disulfides . Additionally, muscle protein denaturation and aggregation are related to disulfide bonds . Among the NP-ALG-LPE coatings (T2-T4), T2 showed the greatest reduction while T4 showed the least loss in the sulfhydryl groups. The loss of sulfhydryl groups affects the structural, functional, and nutritional properties of shrimp muscle protein . This negative effect can be reduced with the use of LPE which showed a superior protective effect and showed greater stability during storage. The carbonyl content in control and coated shrimps during storage at 4 degC is shown in Figure 6C. Carbonyl is a marker of protein oxidation, measured using DNPH (2,4-Dinitrophenylhydrazine) . Protein oxidation decreases shrimp quality and nutrition due to the loss of essential amino acids and reduced digestibility . Carbonyl content increased in all samples during storage (p < 0.05), indicating oxidative damage to amino acid side chains, such as lysine, proline, arginine, and histidine . On day 4, the control sample had higher carbonyl content than all coated samples (T1-T4), which increased over time. Shrimp coated with NP-ALG-LPE (T2-T4) significantly prevented the level of protein oxidation as compared with the ALG coating alone (T1) and control (p < 0.05). However, the addition of LPE in the NP-ALG coating (T2-T4) did not significantly affect the carbonyl content in the samples until day 8. At the end of storage, samples coated with T3 and T4 had lower carbonyl content than T2 (p < 0.05). Secondary products of lipid oxidation, such as aldehydes (e.g., malondialdehyde and 4-hydroxy-2-nonenal) or ketones, can react with amino acid residues through covalent bonds, known as Michael addition reactions, leading to indirect oxidation of protein . Hence, LPE could scavenge this oxidation process in the stored shrimps as it contains an abundant level of polyphenols which act as a primary scavenger for oxidation and oxidation-induced byproducts. 3.2.4. Lipid Oxidation The peroxide value (PV) is a measure of major lipid oxidation products (hydroperoxides) in a sample . The abstraction of hydrogen from fatty acid double bonds produces fatty acid-free radicals, which further react with oxygen to form hydroperoxides . All samples showed a gradual increase in PV level over storage (p < 0.05), indicating oxidation of fatty acids in shrimp muscle, producing hydroperoxides or peroxides . On day 6 of storage, a significant difference was observed between control and coated samples (p < 0.05), with control samples having higher PV, indicating greater lipid oxidation. Coated samples, however, showed lower PV during storage (p < 0.05), attributed to the oxygen barrier capacity of ALG reducing oxygen diffusion and preventing lipid oxidation . Shrimp coated with T4 had the lowest PV compared to control and T1, T2, and T3 throughout storage (p < 0.05). In addition, PV did not exceed the acceptable limit of 18-20 meq/kg . This showed that the antioxidant activity of LPE in preventing lipid oxidation was concentration dependent. Lipid oxidation is generally accelerated during storage due to high levels of polyunsaturated fatty acids in crustacean cell membranes , but phenolic compounds in LPE might scavenge free radicals, reducing lipid oxidation by lowering lipid radicals . Thus, shrimp coated with T4, with the lowest PV, showed the best protection against lipid oxidation. The increase in TBARS (a measure of lipid oxidation) was due to partial oxidation and dehydration of unsaturated fatty acids . All samples showed a rise in TBARS over 14 days of storage (p < 0.05) . Shrimps are susceptible to lipid oxidation, which can occur through autoxidation, photosensitized oxidation, or enzymatic reactions, such as lipoxygenase, peroxidase, and microbial enzymes . In the present study, shrimp coated with 1.5% LPE (T4) showed the lowest TBARS level compared to T1 and control samples at all storage times (p < 0.05), followed by T3 and T2. The decrease in TBARS values was consistent with the decrease in peroxide values (PV), indicating the ability of LPE to scavenge free radicals and prevent the formation of secondary oxidation products. The NP-ALG coating acted as a barrier to oxygen permeation and a carrier for the antioxidants in LPE, thus reducing the production of secondary oxidation products in shrimp . Notably, T4 had the highest amount of LPE (1.5%) in its coating, thus resulting in the lowest TBARS value. Polyphenols, which are abundant in longkong fruit, have a strong reducing capacity and are known to retard and inhibit lipid oxidation . Souza et al. demonstrated that a polyphenol-rich leaf extract from an Amazonian plant act as a powerful antioxidant in human LDL protein by reducing TBARS levels. Furthermore, the TBARS value of all groups was below the acceptable limit (1-2 mg MDA/kg) , consistent with the study by Dehghani et al. . Thus, shrimp coated with T4 showed higher stability against lipid oxidation. The secondary oxidation products of shrimp were measured using AnV, which detects non-volatile oxidation products, such as aldehydes and ketones . AnV reacts with oxidation products to produce a yellow product . The trend of AnV was similar to other lipid oxidation product assays in this study. The control had higher AnV compared to coated samples, with T4 having the lowest AnV . This suggests that LPE in ALG coating reduces secondary oxidation product formation, especially in non-volatile compounds. The results from AnV are consistent with PV and TBARS, confirming that LPE in ALG coating is an effective way to maintain shrimp quality. The totox value measures the total lipid degradation products and indicates the oxidative stage of a product . The totox value combines primary and secondary oxidation products and is commonly used in the food industry. The totox value increased over time for all samples, with the highest value found in the control . A lower totox value indicates better shrimp quality. Among the NP-ALG-LPE-coated shrimps, T4 showed the lowest totox value compared to T2 and T3. Generally, the coating application on the food surface can cause the migration of compounds from the coating into the food . As a result, coating with a higher LPE content (1.5%) resulted in higher levels of migrated substances and a stronger antioxidant effect in the shrimp, providing greater stability against lipid oxidation. 3.2.5. TVB-N The results of TVB-N levels in control and coated shrimp during storage at 4 degC are shown in Figure 8. TVB-N indicates decomposition of protein by microorganisms or endogenous enzymes in foods, particularly seafoods . An analytical technique called TVB-N (Total Volatile Basic Nitrogen) measures the levels of nitrogenous chemicals in shrimp or seafood products, which reveals the level of freshness. TVB-N provides the total base volatile nitrogen content that begins to accumulate in the tissues with degradation during shrimp storage . TVB-N, the combination of ammonia (from amino acid degradation), di-methylamine (generated by self-degrading enzymes), and trimethylamine (result of spoilage bacteria), generally rises with bacterial growth, enzyme degradation, or a combination of both during storage . At day 0, TVB-N levels of all samples were below 10 mg N/100 g, showing freshness of the raw material. The control's TVB-N content rapidly rose with storage time (p < 0.05), reaching 58.4 mg N/100 g on day 14. Shrimp coated with the NP-ALG-LPE coatings (T2-T4) showed a slower increase in TVB content compared to the control and T1-coated samples during storage (p < 0.05). Additionally, samples coated with T4 had lower TVB content compared to those coated with T2 and T3 (p < 0.05). The European Commission considers TVB-N levels of 30-35 mg N/100 g as the upper acceptable limit . The amount of TVB-N increased as a sum of ammonia, dimethylamine, trimethylamine, and volatile amine compounds. A small amount of ammonia is generally found during the first weeks of storage, and the total volatile alkalinity is slowly increased during storage. This may be caused by the amine removal process (deamination) of amino acids . This study found that the NP-ALG-LPE-coated samples (T2-T4) had remained within acceptable limits at the end of storage. The reduced TVB-N content in the NP-ALG-LPE-coated samples was due to either reduced degradation of non-protein nitrogen compounds or slowed bacterial growth, or both. This shows LPE's antibacterial properties. Olatunde et al. also found that coconut husk extract reduced the increase in TVB in Asian sea-bass slices during 12 days of storage at 4 degC. Moreover, the quality and shelf-life of shrimp coated in active edible coatings made of gelatin and orange peel essential oil were determined. The shelf-life of shrimp was evaluated over a 14-day storage period by TVB-N analysis. Compared to the control group, the addition of orange peel essential oil in the edible gelatin coating improved the quality and prolonged the shelf-life of the shrimp. The incorporation of orange peel essential oil helped in preserving the chemical and microbial quality of the shrimp . 3.2.6. Microbial Analysis The initial total viable count (TVC) of the shrimp was around 2.4-2.9 log CFU/g, similar to the result of 2.5-3 log CFU/g reported by Mohebi et al. . During 14 days of storage, a general increase in TVC was observed in all samples (p < 0.05), with the highest increase seen in the control , reaching 23.15 log CFU/g on the 14th day (p < 0.05). Dipping the shrimp in ALG reduced the increasing trend of TVC but adding LPE significantly enhanced the microbial inhibitory effect (p < 0.05). There were no significant differences among NP-ALG-LPE-coated samples (T2-T4) until day 4 of storage (p > 0.05), but the T4-coated sample had the lowest TVC among all tested samples at the end of storage (p < 0.05). This indicates that the inhibitory effect increased with increasing LPE concentration (0.5-1.5%) (p < 0.05). Phytochemical compounds can damage bacterial cells by disrupting cell membranes and precipitating cell protein, causing death . Kim et al. found that shrimp coated with chitosan-alginate containing grape seed extract reduced TVC during 15 days of storage at 4 degC. Liu et al. reported that an alginate-calcium coating with methanol extract from citrus fruit effectively reduced the increasing trend of TVC compared to those coated with 1% chitosan. Lactic acid bacteria (LAB) are facultative anaerobes and form a significant part of the natural microbiota in seafood stored in anaerobic conditions . On the initial day, the count of lactic acid bacteria (LAB) in all samples was 1.0 log CFU/g . During storage, there was an observed increase in LAB count in all samples (p < 0.05). However, samples coated with T1-T4 exhibited a reduced increment compared to the control (p < 0.05). The NP-ALG coating had an inhibitory effect on LAB growth, which was intensified by the addition of LPE at a different concentration. T1, T2, T3, and T4 reduced the LAB count by 2, 2.16, 2.67, and 3.15 log CFU/g compared to the control, respectively. Khaledian et al. found that shrimp coated with a tragacanth gum-based coating containing lime peel extract inhibited LAB growth during 10-day storage at 4 degC. Among the NP-ALG-LPE-coated shrimp samples, T4 had a lower LAB count due to higher LPE concentration (1.5%) and greater antimicrobial effect. This can be attributed to the presence of terpenoids and lansiosides, which are the two major antibacterial substances found in longkong fruit . A smoked eel fillet coated with carboxy methylcellulose containing rosemary extract (200-800 ppm) also showed a decrease in LAB count . Enterobacteriaceae are used as a measure of hygiene . Their presence in seafood, especially in cases of contaminated water or delayed chilling after capture, increases the likelihood of spoilage . The initial count of Enterobacteriaceae in all samples was 0.95 log CFU/g . During storage, the count of this bacteria increased continuously in all samples (p < 0.05). Among the NP-ALG-LPE-coated shrimp samples (T2-T4), T4 had the lowest count of Enterobacteriaceae as compared to the control (p < 0.05). T4 was effective in slowing the growth of Enterobacteriaceae, followed by T3 and T2 (p > 0.05). This suggests that the antimicrobial effect of LPE was dose dependent. Alsaggaf et al. found that shrimp coated with chitosan enriched with pomegranate peel extract reduced the microbial count during 30-day storage at 4 degC, with higher PPE concentration (0.5-2.0%) improving the antimicrobial activity. At the beginning of shrimp storage, the psychrotrophic bacteria count (PBC) of all samples was recorded at 1.85 log CFU/g . Over time, the PBC of the control sample continued to rise (p < 0.05), but the count decreased in shrimp coated with ALG (T1) (p < 0.05). Gram-negative psychrotrophic bacteria are a key cause of spoilage in iced and/or refrigerated seafood . Shrimp coated with NP-ALG-LPE coatings had an even lower level of PBC as compared to T1 and the control (p < 0.05), which confirms the antibacterial activity of LPE. Other studies have also found similar results in shrimp samples treated with green tea extract and fish fillets treated with microalgal exopolysaccharides and red seaweed extract . The use of NP-ALG enriched with LPE as an active edible coating is a promising solution for maintaining the quality of shrimp during refrigerated storage by limiting the growth of multiple bacteria. 4. Conclusions The study investigated the impact of adding LPE (0.5-1.5%) to the ALG coating, using the ultrasonication process to convert the ALG-LPE coating to become nano-sized, and tested various physiochemical properties in the coating emulsion and as well as on the shrimp samples to control the quality loss and prolong the shelf-life for up to 14 days at 4 degC. The results showed that LPE addition to a coating emulsion significantly increased the viscosity, turbidity, particle size, and polydispersity index of the coating material while lowering the pH and whiteness index. The best results were observed when 1.5% LPE was added to the NP-ALG-based coating and tested shrimp samples, which resulted in lower pH, weight loss, and polyphenol oxidase activity, as well as stronger antioxidant effects against protein and lipid oxidation in the shrimp samples. The study also found that the microbial count, including total viable count, lactic acid bacteria, Enterobacteriaceae, and psychrotrophic bacterial count, was lower in the shrimps coated with the ALG-LPE coating. Thus, the results suggest that the addition of 1.5% LPE to the ALG coating can effectively maintain the quality of shrimp for up to 14 days of storage and is a good alternative to synthetic preservatives. In summary, synergistic effect of an alginate-based nanoparticle coating and LPE can serve as promising approach for the quality maintenance of seafood during storage. This finding could be potentially beneficial for food industries to extend the shelf-life of food products without the addition of synthetic preservatives. Further research is anticipated to rationally designed edible coatings with precisely adjusted coating properties for the effective enhancement of food products' quality and storability. Acknowledgments The authors would like to greatly acknowledge the Prince of Songkla University and Food Innovation and Product Development Laboratory (FIPD) for their provided supplies, equipment, and laboratory spaces to complete this research. The authors would also like to express their gratitude to Burapha University Chanthaburi Campus for the additional support to finish this research. Author Contributions Conceptualization, N.C. and K.V.; methodology, B.R. and S.P.; software, N.C. and B.R.; validation, N.C., B.R. and K.V.; formal analysis, N.C. and K.V.; investigation, B.R. and S.P.; resources, K.V.; data curation, N.C., B.R. and K.V.; writing--original draft preparation, N.C., B.R. and S.P.; writing--review and editing, B.R. and K.V.; visualization, K.V.; supervision, K.V.; project administration, K.V.; funding acquisition, N.C. and K.V. All authors have read and agreed to the published version of the manuscript. Data Availability Statement Data is contained within the article. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Schematic illustration of nanoparticle formation using ultrasonication and coating of shrimp using the LPE-added alginate-based nanoparticle edible coating. Figure 2 The pH (A), viscosity (B), and turbidity (C) of the alginate coating and alginate-based nanoparticle coating with different concentration of LPE. C: distilled water; T1: alginate coating; T2, T3, and T4: alginate-based nanoparticle coating with 0.5, 1.0, and 1.5% LPE, respectively. The different alphabets shown on the bar diagram indicate significant differences. Figure 3 Whiteness index (A), particle size (B), polydispersity index (C) of the alginate coating and alginate-based nanoparticle coating with different concentration of LPE. C: distilled water; T1: alginate coating; T2, T3, and T4: alginate-based nanoparticle coating with 0.5, 1.0, and 1.5% LPE, respectively. The different alphabets shown on the bar diagram indicates significant differences. Figure 4 Changes in weight loss (A) and pH (B) of the shrimps coated with different coatings during storage of 14 days at 4 degC. C: shrimp dipped in distilled water as control; T1: shrimp coated with alginate coating; T2, T3, and T4: shrimp coated with 0.5, 1.0, and 1.5% LPE, respectively. Figure 5 Changes in polyphenol oxidase of the shrimps coated with different coatings during storage of 14 days at 4 degC. C: shrimp dipped in distilled water as control; T1: shrimp coated with alginate coating; T2, T3, and T4: shrimp coated with 0.5, 1.0, and 1.5% LPE, respectively. Figure 6 Changes in total sulfhydral content (A), reactive sulfhydral content (B), and carbonyl content (C) of the shrimps coated with different coatings during storage of 14 days at 4 degC. C: shrimp dipped in distilled water as control; T1: shrimp coated with alginate coating; T2, T3, and T4: shrimp coated with 0.5, 1.0, and 1.5% LPE, respectively. Figure 7 Changes in PV (A), TBARS (B), p-anisidine value (C), and totox value (D) of the shrimp coated with different coating during storage of 14 days at 4 degC. C: shrimp dipped in distilled water as control; T1: shrimp coated with alginate coating; T2, T3, and T4: shrimp coated with 0.5, 1.0, and 1.5% LPE, respectively. Figure 8 Changes in TVB-N value of the shrimps coated with different coatings during storage of 14 days at 4 degC. C: shrimp dipped in distilled water as control; T1: shrimp coated with alginate coating; T2, T3, and T4: shrimp coated with 0.5, 1.0, and 1.5% LPE, respectively. Figure 9 Changes in TVC (A), Lactic acid bacteria (B), Enterobacteriacea (C), and Psychrotrophic bacteria (D) of the shrimps coated with different coating during storage of 14 days at 4 degC. C: shrimp dipped in distilled water as control; T1: shrimp coated with alginate coating; T2, T3, and T4: shrimp coated with 0.5, 1.0, and 1.5% LPE, respectively. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Sabu S. Xavier K.M. Sasidharan A. Efficacy of pomegranate phenolic extract and chitosan as an edible coating for shelf life extension of Indian white shrimp during refrigerated storage J. Packag. Technol. Res. 2021 5 59 67 10.1007/s41783-021-00109-0 2. Licciardello F. Kharchoufi S. Muratore G. Restuccia C. Effect of edible coating combined with pomegranate peel extract on the quality maintenance of white shrimps (Parapenaeus longirostris) during refrigerated storage Food Packag. Shelf Life 2018 17 114 119 10.1016/j.fpsl.2018.06.009 3. Ebadi Z. Khodanazary A. Hosseini S.M. Zanguee N. The shelf life extension of refrigerated Nemipterus japonicus fillets by chitosan coating incorporated with propolis extract Int. J. Biol. Macromol. 2019 139 94 102 10.1016/j.ijbiomac.2019.07.204 31374270 4. Nagarajan M. Rajasekaran B. Benjakul S. Venkatachalam K. Influence of chitosan-gelatin edible coating incorporated with longkong pericarp extract on refrigerated black tiger Shrimp (Penaeus monodon) Curr. Res. Food Sci. 2021 4 345 353 10.1016/j.crfs.2021.05.003 34142095 5. Sharifimehr S. Soltanizadeh N. Hossein Goli S.A. Effects of edible coating containing nano-emulsion of Aloe vera and eugenol on the physicochemical properties of shrimp during cold storage J. Sci. Food Agric. 2019 99 3604 3615 10.1002/jsfa.9581 30628095 6. Kim J.-H. Hong W.-S. Oh S.-W. Effect of layer-by-layer antimicrobial edible coating of alginate and chitosan with grapefruit seed extract for shelf-life extension of shrimp (Litopenaeus vannamei) stored at 4 C Int. J. Biol. Macromol. 2018 120 1468 1473 10.1016/j.ijbiomac.2018.09.160 30267814 7. Alexandre S. Vital A.C.P. Mottin C. do Prado R.M. Ornaghi M.G. Ramos T.R. Guerrero A. Pilau E.J. do Prado I.N. Use of alginate edible coating and basil (Ocimum spp.) extracts on beef characteristics during storage J. Food Sci. Technol. 2021 58 3835 3843 10.1007/s13197-020-04844-1 34471307 8. Song Y. Liu L. Shen H. You J. Luo Y. Effect of sodium alginate-based edible coating containing different anti-oxidants on quality and shelf life of refrigerated bream (Megalobrama amblycephala) Food Control 2011 22 608 615 10.1016/j.foodcont.2010.10.012 9. Maqsood S. Benjakul S. Abushelaibi A. Alam A. Phenolic compounds and plant phenolic extracts as natural antioxidants in prevention of lipid oxidation in seafood: A detailed review Compr. Rev. Food Sci. Food Saf. 2014 13 1125 1140 10.1111/1541-4337.12106 10. Venkatachalam K. Influence of prolonged salting on the physicochemical properties of duck egg white Braz. Arch. Biol. Technol. 2018 61 10.1590/1678-4324-2018180134 11. Khaledian S. Basiri S. Shekarforoush S.S. Shelf-life extension of pacific white shrimp using tragacanth gum-based coatings containing Persian lime peel (Citrus latifolia) extract LWT 2021 141 110937 10.1016/j.lwt.2021.110937 12. Xing Y. Li W. Wang Q. Li X. Xu Q. Guo X. Bi X. Liu X. Shui Y. Lin H. Antimicrobial nanoparticles incorporated in edible coatings and films for the preservation of fruits and vegetables Molecules 2019 24 1695 10.3390/molecules24091695 31052263 13. Baek J.H. Lee S.-Y. Oh S.-W. Enhancing safety and quality of shrimp by nanoparticles of sodium alginate-based edible coating containing grapefruit seed extract Int. J. Biol. Macromol. 2021 189 84 90 10.1016/j.ijbiomac.2021.08.118 34419539 14. Pilon L. Spricigo P.C. Miranda M. de Moura M.R. Assis O.B.G. Mattoso L.H.C. Ferreira M.D. Chitosan nanoparticle coatings reduce microbial growth on fresh-cut apples while not affecting quality attributes Int. J. Food Sci. Technol. 2015 50 440 448 10.1111/ijfs.12616 15. Josewin S.W. Ghate V. Kim M.-J. Yuk H.-G. Antibacterial effect of 460 nm light-emitting diode in combination with riboflavin against Listeria monocytogenes on smoked salmon Food Control 2018 84 354 361 10.1016/j.foodcont.2017.08.017 16. Venkatachalam K. The different concentrations of citric acid on inhibition of Longkong pericarp browning during low temperature storage Int. J. Fruit Sci. 2015 15 353 368 10.1080/15538362.2015.1009970 17. Khazaei N. Esmaiili M. Emam-Djomeh Z. Effect of active edible coatings made by basil seed gum and thymol on oil uptake and oxidation in shrimp during deep-fat frying Carbohydr. Polym. 2016 137 249 254 10.1016/j.carbpol.2015.10.084 26686127 18. Basiri S. Shekarforoush S.S. Aminlari M. Akbari S. The effect of pomegranate peel extract (PPE) on the polyphenol oxidase (PPO) and quality of Pacific white shrimp (Litopenaeus vannamei) during refrigerated storage LWT-Food Sci. Technol. 2015 60 1025 1033 10.1016/j.lwt.2014.10.043 19. Simpson B.K. Marshall M.R. Otwell W.S. Phenol oxidase from shrimp (Penaeus setiferus): Purification and some properties J. Agric. Food Chem. 1987 35 918 921 10.1021/jf00078a017 20. Liu G. Xiong Y.L. Contribution of lipid and protein oxidation to rheological differences between chicken white and red muscle myofibrillar proteins J. Agric. Food Chem. 1996 44 779 784 10.1021/jf9506242 21. Ellman G.L. Tissue sulfhydryl groups Arch. Biochem. Biophys. 1959 82 70 77 10.1016/0003-9861(59)90090-6 13650640 22. Robinson H.W. Hogden C.G. The biuret reaction in the determination of serum proteins. 1. A study of the conditions necessary for the production of a stable color which bears a quantitative relationship to the protein concentration J. Biol. Chem. 1940 135 707 725 10.1016/S0021-9258(18)73134-7 23. Xia X. Kong B. Liu Q. Liu J. Physicochemical change and protein oxidation in porcine longissimus dorsi as influenced by different freeze-thaw cycles Meat Sci. 2009 83 239 245 10.1016/j.meatsci.2009.05.003 20416749 24. Parrilla-Taylor D.P. Zenteno-Savin T. Magallon-Barajas F.J. Antioxidant enzyme activity in pacific whiteleg shrimp (Litopenaeus vannamei) in response to infection with white spot syndrome virus Aquaculture 2013 380 41 46 10.1016/j.aquaculture.2012.11.031 25. Bligh E.G. Dyer W.J. A rapid method of total lipid extraction and purification Can. J. Biochem. Physiol. 1959 37 911 917 10.1139/y59-099 13671378 26. Benjakul S. Bauer F. Biochemical and physicochemical changes in catfish (Silurus glanis Linne) muscle as influenced by different freeze-thaw cycles Food Chem. 2001 72 207 217 10.1016/S0308-8146(00)00222-3 27. Okpala C.O.R. Choo W.S. Dykes G.A. Quality and shelf life assessment of Pacific white shrimp (Litopenaeus vannamei) freshly harvested and stored on ice LWT-Food Sci. Technol. 2014 55 110 116 10.1016/j.lwt.2013.07.020 28. de Abreu D.P. Losada P.P. Maroto J. Cruz J. Lipid damage during frozen storage of Atlantic halibut (Hippoglossus hippoglossus) in active packaging film containing antioxidants Food Chem. 2011 126 315 320 10.1016/j.foodchem.2010.10.048 29. Rezaeifar M. Mehdizadeh T. Mojaddar Langroodi A. Rezaei F. Effect of chitosan edible coating enriched with lemon verbena extract and essential oil on the shelf life of vacuum rainbow trout (Oncorhynchus mykiss) J. Food Saf. 2020 40 e12781 10.1111/jfs.12781 30. Chen X. Ren L. Li M. Qian J. Fan J. Du B. Effects of clove essential oil and eugenol on quality and browning control of fresh-cut lettuce Food Chem. 2017 214 432 439 10.1016/j.foodchem.2016.07.101 27507495 31. Lichanporn I. Srilaong V. Wongs-Aree C. Kanlayanarat S. Postharvest physiology and browning of longkong (Aglaia dookkoo Griff.) fruit under ambient conditions Postharvest Biol. Technol. 2009 52 294 299 10.1016/j.postharvbio.2009.01.003 32. Farajzadeh F. Motamedzadegan A. Shahidi S.-A. Hamzeh S. The effect of chitosan-gelatin coating on the quality of shrimp (Litopenaeus vannamei) under refrigerated condition Food Control 2016 67 163 170 10.1016/j.foodcont.2016.02.040 33. Gokoglu N. Novel natural food preservatives and applications in seafood preservation: A review J. Sci. Food Agric. 2019 99 2068 2077 10.1002/jsfa.9416 30318589 34. Bonilla J. Atares L. Vargas M. Chiralt A. Effect of essential oils and homogenization conditions on properties of chitosan-based films Food Hydrocoll. 2012 26 9 16 10.1016/j.foodhyd.2011.03.015 35. Rodriguez-Turienzo L. Cobos A. Diaz O. Effects of edible coatings based on ultrasound-treated whey proteins in quality attributes of frozen Atlantic salmon (Salmo salar) Innov. Food Sci. Emerg. Technol. 2012 14 92 98 10.1016/j.ifset.2011.12.003 36. Cheng J. Cui L. Effects of high-intensity ultrasound on the structural, optical, mechanical and physicochemical properties of pea protein isolate-based edible film Ultrason. Sonochemistry 2021 80 105809 10.1016/j.ultsonch.2021.105809 37. Amiri A. Sharifian P. Soltanizadeh N. Application of ultrasound treatment for improving the physicochemical, functional and rheological properties of myofibrillar proteins Int. J. Biol. Macromol. 2018 111 139 147 10.1016/j.ijbiomac.2017.12.167 29307807 38. Jambrak A.R. Herceg Z. Subaric D. Babic J. Brncic M. Brncic S.R. Bosiljkov T. Cvek D. Tripalo B. Gelo J. Ultrasound effect on physical properties of corn starch Carbohydr. Polym. 2010 79 91 100 10.1016/j.carbpol.2009.07.051 39. Jiang S. Ding J. Andrade J. Rababah T.M. Almajwal A. Abulmeaty M.M. Feng H. Modifying the physicochemical properties of pea protein by pH-shifting and ultrasound combined treatments Ultrason. Sonochemistry 2017 38 835 842 10.1016/j.ultsonch.2017.03.046 40. Salvia-Trujillo L. Rojas-Grau M.A. Soliva-Fortuny R. Martin-Belloso O. Effect of processing parameters on physicochemical characteristics of microfluidized lemongrass essential oil-alginate nanoemulsions Food Hydrocoll. 2013 30 401 407 10.1016/j.foodhyd.2012.07.004 41. Lin L. Gu Y. Cui H. Moringa oil/chitosan nanoparticles embedded gelatin nanofibers for food packaging against Listeria monocytogenes and Staphylococcus aureus on cheese Food Packag. Shelf Life 2019 19 86 93 10.1016/j.fpsl.2018.12.005 42. Acevedo-Fani A. Salvia-Trujillo L. Rojas-Grau M.A. Martin-Belloso O. Edible films from essential-oil-loaded nanoemulsions: Physicochemical characterization and antimicrobial properties Food Hydrocoll. 2015 47 168 177 10.1016/j.foodhyd.2015.01.032 43. Prakash A. Baskaran R. Vadivel V. Citral nanoemulsion incorporated edible coating to extend the shelf life of fresh cut pineapples Lwt 2020 118 108851 10.1016/j.lwt.2019.108851 44. Lu F. Liu D. Ye X. Wei Y. Liu F. Alginate-calcium coating incorporating nisin and EDTA maintains the quality of fresh northern snakehead (Channa argus) fillets stored at 4 C J. Sci. Food Agric. 2009 89 848 854 10.1002/jsfa.3523 45. Huang J. Chen Q. Qiu M. Li S. Chitosan-based edible coatings for quality preservation of postharvest whiteleg shrimp (Litopenaeus vannamei) J. Food Sci. 2012 77 C491 C496 10.1111/j.1750-3841.2012.02651.x 22515241 46. Nirmal N.P. Benjakul S. Retardation of quality changes of Pacific white shrimp by green tea extract treatment and modified atmosphere packaging during refrigerated storage Int. J. Food Microbiol. 2011 149 247 253 10.1016/j.ijfoodmicro.2011.07.002 21798614 47. Shamshad S. Riaz M. Zuberi R. Qadri R. Shelf life of shrimp (Penaeus merguiensis) stored at different temperatures J. Food Sci. 1990 55 1201 1205 10.1111/j.1365-2621.1990.tb03898.x 48. Maqsood S. Benjakul S. Shahidi F. Emerging role of phenolic compounds as natural food additives in fish and fish products Crit. Rev. Food Sci. Nutr. 2013 53 162 179 10.1080/10408398.2010.518775 23072531 49. Bajpai V. Rahman A. Dung N. Huh M. Kang S. In vitro inhibition of food spoilage and foodborne pathogenic bacteria by essential oil and leaf extracts of Magnolia liliflora Desr J. Food Sci. 2008 73 M314 M320 10.1111/j.1750-3841.2008.00841.x 19241564 50. Benjakul S. Visessanguan W. Tanaka M. Properties of phenoloxidase isolated from the cephalothorax of kuruma prawn (Penaeus japonicus) J. Food Biochem. 2005 29 470 485 10.1111/j.1745-4514.2005.00042.x 51. Jang M.S. Sanada A. Ushio H. Tanaka M. Ohshima T. Inhibitory effect of enokitake extract on melanosis of shrimp Fish. Sci. 2003 69 379 384 10.1046/j.1444-2906.2003.00632.x 52. Balti R. Mansour M.B. Zayoud N. Le Balc'h R. Brodu N. Arhaliass A. Masse A. Active exopolysaccharides based edible coatings enriched with red seaweed (Gracilaria gracilis) extract to improve shrimp preservation during refrigerated storage Food Biosci. 2020 34 100522 10.1016/j.fbio.2019.100522 53. Wu G. Farouk M. Clerens S. Rosenvold K. Effect of beef ultimate pH and large structural protein changes with aging on meat tenderness Meat Sci. 2014 98 637 645 10.1016/j.meatsci.2014.06.010 25089788 54. Morachis-Valdez A.G. Gomez-Olivan L.M. Garcia-Argueta I. Hernandez-Navarro M.D. Diaz-Bandera D. Dublan-Garcia O. Effect of chitosan edible coating on the biochemical and physical characteristics of carp fillet (Cyprinus carpio) stored at-18 C Int. J. Food Sci. 2017 2017 2812483 10.1155/2017/2812483 28634584 55. Soyer A. Ozalp B. Dalmis U. Bilgin V. Effects of freezing temperature and duration of frozen storage on lipid and protein oxidation in chicken meat Food Chem. 2010 120 1025 1030 10.1016/j.foodchem.2009.11.042 56. Hematyar N. Rustad T. Sampels S. Kastrup Dalsgaard T. Relationship between lipid and protein oxidation in fish Aquac. Res. 2019 50 1393 1403 10.1111/are.14012 57. Bazargani-Gilani B. Aliakbarlu J. Tajik H. Effect of pomegranate juice dipping and chitosan coating enriched with Zataria multiflora Boiss essential oil on the shelf-life of chicken meat during refrigerated storage Innov. Food Sci. Emerg. Technol. 2015 29 280 287 10.1016/j.ifset.2015.04.007 58. Ansarian E. Aminzare M. Azar H.H. Mehrasbi M.R. Bimakr M. Nanoemulsion-based basil seed gum edible film containing resveratrol and clove essential oil: In vitro antioxidant properties and its effect on oxidative stability and sensory characteristic of camel meat during refrigeration storage Meat Sci. 2022 185 108716 10.1016/j.meatsci.2021.108716 34839195 59. Rysman T. Van Hecke T. Van Poucke C. De Smet S. Van Royen G. Protein oxidation and proteolysis during storage and in vitro digestion of pork and beef patties Food Chem. 2016 209 177 184 10.1016/j.foodchem.2016.04.027 27173550 60. Lorenzo J.M. Gomez M. Shelf life of fresh foal meat under MAP, overwrap and vacuum packaging conditions Meat Sci. 2012 92 610 618 10.1016/j.meatsci.2012.06.008 22749431 61. Rajasekaran B. Singh A. Nagarajan M. Benjakul S. Effect of chitooligosaccharide and a-tocopherol on physical properties and oxidative stability of shrimp oil-in-water emulsion stabilized by bovine serum albumin-chitosan complex Food Control 2022 137 108899 10.1016/j.foodcont.2022.108899 62. Reesha K. Panda S.K. Bindu J. Varghese T. Development and characterization of an LDPE/chitosan composite antimicrobial film for chilled fish storage Int. J. Biol. Macromol. 2015 79 934 942 10.1016/j.ijbiomac.2015.06.016 26092060 63. Medina I. Gallardo J. Gonzalez M.J. Lois S. Hedges N. Effect of molecular structure of phenolic families as hydroxycinnamic acids and catechins on their antioxidant effectiveness in minced fish muscle J. Agric. Food Chem. 2007 55 3889 3895 10.1021/jf063498i 17451250 64. Souza J.N. Silva E.M. Loir A. Rees J.-F. Rogez H. Larondelle Y. Antioxidant capacity of four polyphenol-rich Amazonian plant extracts: A correlation study using chemical and biological in vitro assays Food Chem. 2008 106 331 339 10.1016/j.foodchem.2007.05.011 65. Dehghani S. Hosseini S.V. Regenstein J.M. Edible films and coatings in seafood preservation: A review Food Chem. 2018 240 505 513 10.1016/j.foodchem.2017.07.034 28946304 66. Buamard N. Benjakul S. Ethanolic coconut husk extract: In vitro antioxidative activity and effect on oxidative stability of shrimp oil emulsion Eur. J. Lipid Sci. Technol. 2017 119 1700131 10.1002/ejlt.201700131 67. Xu D. Xue H. Sun L. Wang Y. Retardation of melanosis development and quality degradation of Litopenaeus vannamei with starving treatment during cold storage Food Control 2018 92 412 419 10.1016/j.foodcont.2018.03.039 68. Shahidi F. Wanasundara P. Extraction and analysis of lipids Food Lipids, Chemistry Nutrition and Biotechnology 2nd ed. Akoh C.C. dan Min D.B. Marcel Dekker Inc. New York, NY, USA 2002 133 168 69. Chi S. Zivanovic S. Penfield M. Application of chitosan films enriched with oregano essential oil on bologna-active compounds and sensory attributes Food Sci. Technol. Int. 2006 12 111 117 10.1177/1082013206063845 70. Arfat Y.A. Benjakul S. Vongkamjan K. Sumpavapol P. Yarnpakdee S. Shelf-life extension of refrigerated sea bass slices wrapped with fish protein isolate/fish skin gelatin-ZnO nanocomposite film incorporated with basil leaf essential oil J. Food Sci. Technol. 2015 52 6182 6193 10.1007/s13197-014-1706-y 26396365 71. Castro P. Millan R. Pennedo J.C. Sanjuan E. Santana A. Caballero M.J. Effect of storage conditions on total volatile base nitrogen determinations in fish muscle extracts J. Aquat. Food Pro. Technol. 2012 21 519 523 10.1080/10498850.2011.610917 72. European Commission The European Commission 1995-2000 Available online: (accessed on 31 January 2023) 73. Wu T.H. Bechtel P.J. Ammonia, dimethylamine, trimethylamine, and trimethylamine oxide from raw and processed Fish by-products J. Aquat. Food Prod. Technol. 2008 17 27 38 10.1080/10498850801891140 74. Olatunde O.O. Benjakul S. Vongkamjan K. Combined effect of ethanolic eoconut husk extract and modified atmospheric packaging (MAP) in extending the shelf life of Asian sea bass slices J. Aquat. Food Prod. Technol. 2019 28 689 702 10.1080/10498850.2019.1629138 75. Alparslan Y. Metin C. Yapici H.H. Baygar T. Gunlu A. Baygar T. Combined effect of orange peel essential oil and gelatin coating on the quality and shelf life of shrimps J. Food. Saf. Food Qual. 2017 68 53 80 10.2376/0003-925X-68-69 76. Mohebi E. Shahbazi Y. Application of chitosan and gelatin based active packaging films for peeled shrimp preservation: A novel functional wrapping design LWT-Food Sci. Technol. 2017 76 108 116 10.1016/j.lwt.2016.10.062 77. Liu X. Jia Y. Hu Y. Xia X. Li Y. Zhou J. Liu Y. Effect of Citrus wilsonii Tanaka extract combined with alginate-calcium coating on quality maintenance of white shrimps (Litopenaeus vannamei Boone) Food Control 2016 68 83 91 10.1016/j.foodcont.2016.03.028 78. Ragasa C.Y. Labrador P. Rideout J.A. Antimicrobial terpenoids from Lansium domesticum Philipp. Agric. Sci. 2006 89 101 79. Munir T. Munawar K.S. Mohyuddin A. An overview of the antibacterial implications of Lansium domesticum J. Basic Appl. Sci. 2018 14 206 209 10.6000/1927-5129.2018.14.32 80. Choulitoudi E. Ganiari S. Tsironi T. Ntzimani A. Tsimogiannis D. Taoukis P. Oreopoulou V. Edible coating enriched with rosemary extracts to enhance oxidative and microbial stability of smoked eel fillets Food Packag. Shelf Life 2017 12 107 113 10.1016/j.fpsl.2017.04.009 81. Sallam K.I. Antimicrobial and antioxidant effects of sodium acetate, sodium lactate, and sodium citrate in refrigerated sliced salmon Food Control 2007 18 566 575 10.1016/j.foodcont.2006.02.002 17471315 82. Alsaggaf M.S. Moussa S.H. Tayel A.A. Application of fungal chitosan incorporated with pomegranate peel extract as edible coating for microbiological, chemical and sensorial quality enhancement of Nile tilapia fillets Int. J. Biol. Macromol. 2017 99 499 505 10.1016/j.ijbiomac.2017.03.017 28274866 83. Raeisi M. Tajik H. Aliakbarlu J. Mirhosseini S.H. Hosseini S.M.H. Effect of carboxymethyl cellulose-based coatings incorporated with Zataria multiflora Boiss. essential oil and grape seed extract on the shelf life of rainbow trout fillets LWT-Food Sci. Technol. 2015 64 898 904 10.1016/j.lwt.2015.06.010 84. Tsai G. Su W.-H. Chen H.-C. Pan C.-L. Antimicrobial activity of shrimp chitin and chitosan from different treatments Fish. Sci. 2002 68 170 177 10.1046/j.1444-2906.2002.00404.x |
PMC10000640 | (1) Background: A healthful diet, regular physical activity, and weight management are cornerstones for cancer prevention and control. Yet, adherence is low in cancer survivors and others, calling for innovative solutions. Daughters, dUdes, mothers, and othErs fighting cancer Together (DUET) is a 6-month, online, diet-and-exercise, weight-loss intervention to improve health behaviors and outcomes among cancer survivor-partner dyads. (2) Methods: DUET was tested in 56 dyads (survivors of obesity-related cancers and chosen partners) (n = 112), both with overweight/obesity, sedentary behavior, and suboptimal diets. After baseline assessment, dyads were randomized to DUET intervention or waitlist control arms; data were collected at 6-months and analyzed using chi-square, t-tests, and mixed linear models (a < 0.05). (3) Results: Retention was 89% and 100% in waitlisted and intervention arms, respectively. Dyad weight loss (primary outcome) averaged -1.1 (waitlist) vs. -2.8 kg (intervention) (p = 0.044/time-by-arm interaction p = 0.033). Caloric intake decreased significantly in DUET survivors versus controls (p = 0.027). Evidence of benefit was observed for physical activity and function, blood glucose, and c-reactive protein. Dyadic terms were significant across outcomes, suggesting that the partner-based approach contributed to intervention-associated improvements. (4) Conclusions: DUET represents a pioneering effort in scalable, multi-behavior weight management interventions to promote cancer prevention and control, calling for studies that are larger in size, scope, and duration. neoplasms survivors diet exercise weight reduction programs internet randomized controlled trial American Institute for Cancer Research585363 American Cancer SocietyCRP-19-175-06-COUN National Cancer InstituteP01 CA229997 This research was funded by the American Institute for Cancer Research, 585363; the American Cancer Society CRP-19-175-06-COUN; and the National Cancer Institute, P01 CA229997. pmc1. Introduction Over the past four decades, the prevalence of obesity has tripled , as has the number of cancer survivors . While these trends are not directly related, obesity is an acknowledged risk factor for 13 different cancers . Obesity and weight gain after a cancer diagnosis also are linked to a poorer prognosis . Thus, weight management has been suggested as a means of primary cancer prevention , as well as tertiary prevention of second cancers and other prevalent forms of comorbidity among cancer survivors . Given that social support plays an integral role in achieving the lifestyle changes that underlie weight management , buddy systems have been implemented in weight loss programs to enhance efficacy . With respect to cancer, buddy pairings that unite cancer survivors with individuals in their social networks and which emphasize weight management to prevent or control cancer may be one way to expand reach and fortify intervention uptake, especially those that are minimal touch. To date, there have been two dyadic interventions that have promoted diet, physical activity, and weight loss as a means of cancer prevention and control among both cancer survivors and their family members. The Daughters and Mothers (DAMES) study included 68 breast cancer survivors (mothers) who had overweight or obesity and were insufficiently active, paired with their adult biological daughters who had similar lifestyle behaviors and increased adiposity. Randomized dyads received 12 months of bimonthly print materials that were either team-tailored, individually tailored, or standardized (not tailored at all) . Although baseline-to-follow-up improvements in weight, waist circumference, and accelerometry-measured physical activity were observed across all arms, as hypothesized, the tailored interventions resulted in significantly greater improvements in waist circumference and physical activity than the standardized intervention. Surprisingly, only the individually or personally tailored intervention and not the team-tailored intervention (which provided tailored feedback for both the survivor and daughter simultaneously) resulted in significant reductions in body mass index ([BMI] among cancer survivors but not daughters. While DAMES was deemed feasible in terms of safety, retention (90%), and achievement of the accrual target, recruitment of mother-daughter dyads was difficult (3% enrollment rate [#consented/#contacted]). Subsequently, "Healthy Moves" focused on survivor-spouse dyads (n = 22), which were randomized to intervention arms that targeted the survivor alone (using a modified version of the DAMES individually tailored intervention) or both the survivor and spouse (using a modified version of the DAMES team-tailored intervention, plus nine videoconferencing sessions with a marriage counselor to enhance spousal communication) . Compared to DAMES, Healthy Moves had a higher enrollment rate (12.7%); however, despite intensive efforts to enhance communication among couples assigned to the spouse-survivor arm, improvements in weight and fruit and vegetable consumption (and related physical function outcomes) were similar across arms. Spouses receiving the intervention versus spouses of dyads randomized to the survivor-alone condition (i.e., who did not receive the intervention) experienced significant improvements in lifestyle behaviors and physical function outcomes. Building on this research, the current study, Daughters, dUdes, mothers, and othErs fighting cancer Together (DUET) trial, sought to expand eligibility by encouraging survivors to select any partner they felt could participate with them in the weight loss intervention . Furthermore, instead of relying on a tailored, mailed print intervention that was informed by mailed surveys, DUET incorporated newer technology, i.e., Fitbits(r) and Aria(r) scales, for monitoring and used a more contemporary and scalable web-based platform to deliver the intervention. Herein, we report the main outcomes of the DUET trial that was aimed at promoting weight loss among cancer survivors and their chosen supportive partners. Hypotheses were that dyads assigned to the DUET intervention would lose significantly more weight (primary outcome) at 6-month follow-up than dyads assigned to the waitlisted control; moreover, the intervention also would result in more favorable changes in secondary outcomes, such as other measures of adiposity (e.g., waist circumference), diet quality, physical activity, quality-of-life, and physical performance, as well as related biomarkers (e.g., insulin, glucose, total and high-density lipoprotein (HDL) cholesterol, triglycerides, leptin, adiponectin, and c-reactive protein (CRP). 2. Materials and Methods 2.1. Overview DUET was a single-blinded, 2-arm randomized controlled trial (RCT) that tested a 6-month web-based lifestyle intervention against a waitlist control among 56 dyads. Each dyad comprised a survivor of an obesity-related cancer, and their chosen partner, both of whom had obesity or overweight, were insufficiently active, and consumed suboptimal diets. This trial was approved by the University of Alabama at Birmingham (UAB) Institutional Review Board (300003882) and registered within ClinicalTrials.gov (NCT04132219). Detailed methods for DUET were published upon attainment of the accrual target of 56 dyads which ensued over a 9-month period ranging from October 2020 to July 2021 ; these methods are briefly summarized below. 2.2. Participants: Recruitment, Screening, Consent, and Randomization Study invitations were distributed to adult survivors of localized renal cancer and loco-regional ovarian, colorectal, prostatic, endometrial, and female breast cancers (obesity-related cancers with 5-year survival rates >70%) identified from the UAB cancer registry, and listings of individuals expressing previous interest in lifestyle RCTs. Additionally, the Love Research Army ) (accessed on 2 March 2023 initiated a series of email "blasts" to its members, and a recruitment website was established. Study staff provided telephone follow-up on mailings and contacts by placing up to six calls at various days and times. The study was explained, and interested survivors were screened for eligibility. Inclusion criteria were: (1) BMI >=25 kg/m2; (2) moderate-to-vigorous physical activity (MVPA) < 150 min/week; (3) English speaking and writing; (4) educational attainment >=5th grade; and (5) daily internet use and mobile phone ownership. Exclusion criteria were: (1) adhering to modified diets or enrolled in structured diet or exercise programs; (2) recent physician's advice to limit PA and/or health issues precluding unsupervised PA or weight loss; and (3) residence in an assisted-nursing facility. Once eligibility was established and cancer case status (type and date of diagnosis) was verified by treating physicians of self-referrals, the survivor was asked to identify a partner with whom they interacted in person on at least a biweekly basis. Partners had identical inclusion/exclusion criteria (cancer survivorship was optional). Telephone or Zoom calls were scheduled to review the study and acquire signed consent electronically (Adobe Sign(r), San Jose, CA, USA). Participants completed baseline assessments, and their addresses were used to derive rural-urban commuting area codes (RUCA) as well as to estimate the distance between dyad members using Google Maps ) (accessed on 2 March 2023) since rural-urban status and proximity of dyad pairs could potentially affect access to healthy food procurement and exercise opportunities and support . Dyads were randomly and evenly assigned to the DUET intervention or waitlist control using a permuted block design (block size = 4). 2.3. DUET Intervention The DUET web-based intervention was adapted from two previously established programs: (1) the tailored mail-based, dyadic DAMES intervention, which was expanded to meet the needs of cancer survivors beyond those with just post-menopausal breast cancer and for partners beyond just biological daughters ; and (2) SurvivorSHINE, a web-based diet and exercise program for cancer survivors . Like both of these interventions, DUET was theoretically grounded on Social Cognitive Theory (SCT) and emphasized skills training, modeling, incremental goal setting (with reinforcement), overcoming barriers, and self-monitoring (through the incorporation of new technologies, i.e., Fitbits and Aria Scales) . Concepts from Interdependence Theory and the Theory of Communal Coping also guided the dyads' commitment to relationship quality and the development of mutual goals to promote the adoption and maintenance of health behaviors and the provision and request for social support. Upon randomization, one dyad member was mailed a box of supplies that included two sets of Portion Doctor (r) tableware, two Fitbits (Inspire(r)), two Aria 2(r) digital scales, and two sets of instructions to connect to MyFitnessPal(r) to automate weight and exercise tracking and provide additional reinforcement and support. Fitbit accounts also were linked to the password-protected, interactive DUET website, which formed the central core of the DUET intervention. Here each dyad member received tailored guidance over 24 weeks based on World Cancer Research Fund--American Institute of Cancer Research (WCRF-AICR) guidelines . Thus, each dyad member was encouraged to set incremental goals that would eventually lead over the course of the 6-month intervention to exercising (including aerobic, resistance, flexibility, and balance) at least 150 min a week and adhering to a plant-based diet that included ample amounts of whole grains, vegetables and fruit (V and F), and limited amounts of red and processed meats, sugar, and refined (fast) food, while promoting a loss of roughly 0.5 kg per week. The website was designed with the following key features: (1) My Profile; (2) Topical Content; (3) Tip of the Day; (4) Sessions; (5) Tools; (6) News You Can Use; and (7) Support. Participants initially logged in to "My Profile" to enter age, height, gender, and current data on night-time snacking and intakes of V and F, whole and processed grains, red and processed meats, added sugars, supplement use, and alcohol. Survivors were prompted for data on cancer type, treatment, and coping style (Fighting Spirit or Fatalist) , which were used to provide tailored feedback, e.g., graphical displays with motivational messaging on overcoming treatment-related barriers (such as intolerance of high fiber V and Fs among survivors of colorectal cancer treated with a colostomy or urinary incontinence among survivors of prostate cancer), and calorie budgets to promote a loss of 0.5 kg w-1 . Additionally, discrete tabs were provided to facilely reference topical information on healthy weight, healthy eating, and exercise. Daily tips for weight management, diet, and exercise were continually refreshed over the 6-month intervention as a means to enhance engagement. Furthermore, 24 weekly interactive sessions averaging 15 min in length were created using Articulate Storyline software (Articulate Global, LLC, New York, NY, USA) to guide participants through topics such as portion control, grocery shopping and food preparation, and various forms of exercise (aerobic, resistance, balance, and flexibility). A variety of tools also were provided on goal setting, customized meal plans, recipes, grocery lists, exercise guides, etc., in formats that could be downloaded and printed off. A tab entitled "News You Can Use" provided "take-away" summaries of recently released news stories and research pertaining to diet and exercise for cancer control. Finally, the webpage offered tips, such as active listening, to enhance dyad-based support. To enhance engagement with the website, Short Message System (SMS) text messages were issued thrice weekly. On each Monday of the 24-week intervention, dyads received a "push" message with a direct website link to the newly-released weekly session. On Wednesdays, dyads received a support message to reinforce the weekly content, and on Fridays, a "call-to-action" inquired about progress towards incremental goals. 2.4. Waitlist Control Waitlisted dyads received all DUET resources and programming once 6-month follow-up data were collected. 2.5. Measures Because DUET was implemented during COVID-19, several measures were adapted for remote assessment and were captured via Zoom(r) (San Jose, CA, USA); validation study results were published previously . 2.5.1. Anthropometric Measures (Captured at Baseline and 6 Months) Body Weight (Primary Outcome): The participant stepped on the scale without shoes in light clothing while the process was captured via Zoom (focusing on the scale display to make sure the scale was "zeroed-out" prior to weighing). Weight was measured twice, with both values averaged for analyses. Waist circumference: This was measured with unmarked ribbons to reduce bias ; two sets of ribbons were mailed to each dyad for assessments. Participants bared their midriffs on Zoom and were instructed to place one end of the ribbon on the umbilicus. Partners then wrapped the ribbon evenly around the waist. Participants rotated in front of the camera, and assessors ensured the ribbon was parallel to the floor and snug against the skin. Upon exhale, partners used a felt-tip marker to mark the point of overlap. The process was repeated with the 2nd ribbon. Both ribbons were returned to the study office, where they were measured by staff, and the average was used for analyses. 2.5.2. Dietary Intake (Captured at Baseline and 6 Months) Two 24-h dietary recalls of a non-consecutive weekday and weekend day were conducted via telephone by a registered dietitian using the Automated Self-Administered (ASA-24) dietary assessment tool ) (accessed on 2 March 2023) at baseline and 6-months. Averaged intakes were obtained for calories, and diet quality was assessed using the Healthy Eating Index (HEI)-2015 . 2.5.3. Physical Activity and Sleep (Captured at Baseline and 6 Months) Objective PA data were captured using Actigraphs(r) (Fort Walton, FL, USA) with instructions to wear the device at waist level on a provided belt during waking hours and to move the device to a provided wristband upon retiring to sleep. This procedure was followed for 7 days and accompanied by a written log. Minutes of MVPA were then downloaded and processed with similar methods used previously . The Godin Leisure Time Exercise Questionnaire (GLTEQ) was administered online, giving excellent reliability and validity among cancer survivors . 2.5.4. Physical Performance (Captured at Baseline and 6 Months) Several physical performance measures were adapted for remote delivery and validated . Dyads were mailed soccer cones, measuring tapes, 8' lengths of cord, and stickers before assessments to perform measures. Trained assessors recorded images of testing via Zoom and then replayed them to capture accurate times and observations. Once data were entered into databases, videos were erased. Details of the remote assessment of the 30-s chair stand, 8' get-up-and-go, sit-and-reach, back scratch, 2-min step test, and balance testing are reported by Pekmezi et al. and Hoenemeyer et al. 2.5.5. Circulating Biomarkers (Captured at Baseline and 6 Months) Participants received print and video instruction ) (accessed on 2 March 2023) to self-collect 5 dried blood spots (DBS) on a designated card. These were dried for >4 h at room temperature, then inserted into a foil pouch with desiccant and frozen (0 Fdeg or below) until analyzed. DBS eluents were batch-tested against known standards for insulin, glucose, leptin, adiponectin, high-density lipoprotein (HDL), and total cholesterol, triglycerides, and c-reactive protein (CRP) at the University of Washington as described previously . To assure the validity of DBS assays in the current sample and data presented in this report, assays were performed using traditional multiplex methods on sera collected via phlebotomy from 36 participants at baseline and compared to a matched analysis of assays performed on DBS samples collected at the same time. Coefficients (R2) generated by ordinal logistic regression indicated strong correlations and were as follows: glucose = 0.981; CRP = 0.979; triglycerides = 0.979; total cholesterol = 0.963; leptin = 0.919; HDL = 0.899; adiponectin = 0.799; and insulin = 0.700. Values are expressed in plasma equivalent terms. 2.5.6. Online Surveys Online surveys were administered via REDCap(r) ) (accessed on 2 March 2023) at baseline, 6-months, though demographic information was only collected at baseline. Comorbidity/Symptoms: The Older Americans Resources and Services (OARS) Comorbidity Index (modified version) assessed the number of medical conditions and functional impact by ascertaining either an affirmative or negative response to 21 unique medical conditions and 22 symptoms . Quality of Life (QOL): QOL was measured by the PROMIS global QOL to assess physical, mental, and social health domains, as well as usual activities, fatigue, and pain . Self-Efficacy is a central construct of SCT and is domain specific. Instruments with good internal consistency (a = 0.70-0.95) were selected to assess self-efficacy for dietary weight management and exercise . Social Support: Validated 5-point scales by Sallis et al. assessed social support for exercise and dietary change. Barriers: Thirty-one common barriers were assessed to a diet low in fat and sugar, increased V and Fs and whole grains, and exercise (e.g., "low calorie foods don't taste good," "I don't know how to cook or prepare low calorie foods," "I am too busy to (...follow a low calorie diet...exercise)," "I don't enjoy exercise, etc.) . Demographics: Self-reported data were collected on height, race/ethnicity, age, educational and marital status, current smoking status, income range, and relationship with the dyadic partner (i.e., spouse, child, parent, sibling, friend, or other). Because eHealth literacy also could serve as a potential moderator, three items from the eHEALS eHealth Literacy scale, , i.e., "I know how to use the Internet to answer my health questions; "I know how to use the health information I find on the Internet to help me," and "I have the skills I need to evaluate the health resources I find on the Internet" (5-item Likert scale ranging from strongly agree to disagree) were included in the baseline online survey. In addition, given substantial data showing that depression is a strong moderator of weight loss interventions , the Center for Epidemiologic Studies of Depression (CES-D; Boston short form; 20 items; yes-no format), was administered at baseline . 2.5.7. Safety All participants were encouraged to call a toll-free study number to report any adverse events. In addition, changes in health status were systematically ascertained in both study arms at 3 and 6 months. Events considered permanently disabling, life-threatening, or resulting in overnight hospitalization were deemed "serious," with attribution to the intervention explored further. 2.6. Statistical Considerations While accrual, retention, and safety form the basis of this feasibility trial, between-arm differences (i.e., differences between the intervention group and the waitlist control group) in weight loss (primary outcome) from baseline to 6 months were formally tested. Power calculations were performed using nQuery (version 8.5; GraphPad Software DBA Statistical Solutions, San Diego, CA, USA). These calculations assumed a standard deviation of 4.6 kg for the mean weight loss, as presented in our DUET protocol paper . Assuming a sample size of 25 dyads/arm, a common standard deviation of 4.6 kg, a two-sided two-group t-test, and a significance level of 5%, there was >80% power to detect between-arm differences in weight loss of -3.72 kg or greater. To determine whether important demographic and clinical characteristics of the sample were evenly distributed between the two study arms, the chi-square test (or Fisher's exact test if the assumptions for the chi-square were not valid) for categorical study variables and the two-group t-test for continuous study variables were used. Distributions of continuous study variables were examined using stem-and-leaf, box, and normal probability plots and the Kolmogorov-Smirnov test; variables deviating from normal distribution were log10 transformed prior to analysis. All analyses were performed using SAS software (version 9.4; SAS Institute, Inc., Cary, NC, USA). Arm differences in weight loss were assessed using an intent-to-treat approach. General linear mixed models, in particular, mixed model repeated measures analyses, were used to test for between-arm differences (two study arms), within-arm differences (three time points), and the arm-by-time point interaction simultaneously. These analyses were performed using PROC MIXED of SAS. This method accounts for the repeated measurements as well as the covariance between survivors, partners, or dyad members. A compound symmetry covariance matrix was assumed. This method provides tests of statistical significance (Type 3 tests which produce an F value and a p-value) for the between-arm effect, within-arm effect, and the interaction effect. When any of these effects were statistically significant, the Tukey-Kramer multiple comparisons test (performed using PROC MIXED of SAS) was used to determine which specific pairs of means for that effect were significantly different and also identified the time points at which those differences occurred. Such testing was helpful in comparing the multiple groups of survivors, partners, and dyads over two time points (when most outcomes were assessed) as well as three points (for survey data). Analyses were performed separately for survivors, partners, and combined dyads (thus, three sets of analyses were performed). Post-randomization exclusions (i.e., the three participants who either received gastrointestinal surgery or developed a cancer recurrence within 2 weeks of randomization) were omitted from 6-month analyses. Otherwise, all available data were used, though if a dyad member dropped out, data from that dyad were excluded from the dyadic analysis. Analyses of secondary continuous outcomes were performed using general linear mixed models, as described in the previous paragraph for the primary outcome. These analyses were again followed by the use of the Tukey-Kramer multiple comparisons tests (for post hoc testing). 3. Results 3.1. Study Sample, Retention, and Safety Sample characteristics are shown in Table 1. Overall, participants were diverse in terms of race, age, and geography (Alabama, Illinois, Mississippi, North Carolina, and Tennessee). Most were female, urban dwellers, and non-smokers, and roughly half reported being college graduates and currently employed with annual incomes >USD 50,000. Mean levels of V and F intake and PA were low as compared to the guidelines , while the average BMI was in the obese range, and participants reported an average of three other health conditions in addition to their cancer diagnosis. Survivors tended to be "long-term" (i.e., having diagnoses more than 5 years out), with most reporting early-stage cancers, of which a high proportion were breast cancer. A small number of previous cancer diagnoses (four breast, two gynecologic, and one testicular) were reported among partners. Given the high proportion of breast cancer survivors with spousal partners, survivors were significantly more likely to gender identify as female, while partners were more likely to report as male (p-values < 0.05). There were no statistically significant differences detected between the intervention vs. the waitlist control arms for any of the characteristics collected. The CONSORT diagram shows an enrollment rate of ~5.5% (n = 61/n = 1114). Of the 112 participants enrolled, three exclusions occurred within two weeks after randomization within the waitlisted arm (one survivor developed a cancer recurrence, another received emergency gastrointestinal surgery, and one partner received bariatric surgery), all of which were discontinued from the study and analysis, since all of these conditions affect the primary outcome (weight status). Additionally, three waitlisted partners were lost to follow-up (two of three dropped out when their survivor did so). Thus, the retention rate was 89% in the waitlisted arm and 100% in the intervention arm; the difference was statistically significant (p = 0.027). Adverse events totaled 14 and 16 in waitlist and intervention arms, respectively. All events were non-attributable, and all except four were non-serious (two cancer recurrences, one myocardial infarction, and one acute cholecystitis), with no differences in events noted between the waitlist control and intervention and the intervention arms. 3.2. Changes in Adiposity Significant weight loss occurred in both study arms (Table 2), though the magnitude of weight loss was significantly larger in survivors, partners, and dyads randomized within the DUET intervention arm. Dyads assigned to the DUET intervention lost significantly more weight (an average of 2.8 kg or 3.2% of their body weight) as compared to dyads that were waitlisted (who lost an average of 1.1 kg or 1.2% of their body weight). Findings related to waist circumference paralleled the results for weight loss, but differences between the two study arms did not reach statistical significance. 3.3. Changes in Dietary Intake and Physical Activity Both study arms also significantly reduced their caloric intakes, with reductions being particularly notable among partners and dyads within the DUET intervention arm. However, calorie intake was significantly less among survivors assigned to the intervention arm than those randomized to the waitlist control (Table 2). While values for diet quality increased among intervention participants as compared to decreasing values among controls over the study period, these differences did not achieve statistical significance. Both study arms also showed significant increases in MVPA assessed either via self-report or accelerometry over the study period, though increases among survivors within the DUET study arm were of greater magnitude. That being said, differences between study arms did not reach statistical significance. 3.4. Changes in Physical Performance As shown in Table 3, both study arms experienced significant improvements in several indices of physical performance (i.e., 30-s chair stand, 8' get-up-and-go, sit-and-reach, and 2-min step test) over the study period, with DUET intervention arm Survivors showed improvements of greater magnitude for all four tests and DUET dyads in 3-out-of-4 tests (i.e., all except the 2-min step test). DUET partners also showed notable improvements in the 30-s chair stand. However, in comparing improvements in the two study arms over time, significant differences were only detected for the flexibility measure, i.e., the sit-and-reach among survivors and dyads. 3.5. Changes in Circulating Biomarkers As shown in Table 4, both study arms experienced significant decreases in circulating glucose, and while decreases were particularly noteworthy among DUET-assigned partners and dyads and among waitlisted survivors, these beneficial effects did not differ in statistical significance between study arms. Significant decreases over time also were observed among partners and dyads in both study arms for total cholesterol, as well as HDL cholesterol among all three subgroups (i.e., survivors, partners, and dyads). The effects on HDL cholesterol were particularly notable among dyad members of both study arms, with DUET dyads experiencing significantly greater decreases in HDL cholesterol than waitlisted dyads. Similarly, levels of CRP also decreased significantly over time among survivors and dyads in both study arms, and while these differences were particularly notable among survivors within the DUET intervention arm, no statistically significant differences were noted when the waitlist vs. the intervention arm were compared. Data on the adipokines, leptin, and adiponectin were less consistent, though significantly higher increases in leptin were observed among partners in the DUET intervention than among the waitlist control. No differences were detected for circulating levels of insulin or triglycerides. 3.6. Changes in Patient-Reported Outcomes Significant improvements in physical QOL were observed over time among survivors of both study arms, though differences were not observed among other subgroups and also not for mental QOL. Further, no differences between the DUET intervention arm vs. the waitlist control were detected (Table S1). While social support and self-efficacy for both diet and exercise increased over the 6-month period for intervention participants compared to decreasing levels among controls, these differences did not achieve statistical significance. In contrast, barriers decreased significantly over time in both study groups, with survivors and dyads reporting significantly fewer barriers to pursuing a low-calorie diet and partners and dyads reporting significantly fewer barriers toward exercise. While no statistically significant differences were identified between arms, p-values for time-by-arm interactions approached significance (e.g., p = 0.051). 3.7. Model Dyadic Terms Of note, the models generated for DUET uncovered several significant dyadic terms, suggesting that the relationship established between the survivor and their partner appeared important for influencing effects on body weight (p = 0.009), waist circumference (p = 0.023), diet quality (p = 0.009), objectively-assessed PA (p's < 0.001), sleep efficiency (p < 0.001), most physical performance tests (except the sit-and-reach) (p's < 0.007), HDL cholesterol (p = 0.038), CRP (p = 0.002) and mental health (PROMIS; p < 0.001). Trends also were noted for caloric intake (p = 0.083), diet self-efficacy (p = 0.089), sleep fragmentation index (p = 0.0504), adiponectin (p = 0.095), and total cholesterol (p = 0.076). 4. Discussion The DUET diet and exercise intervention was found to be feasible and resulted in significant weight loss among cancer survivors and their chosen partners. The 3.2% loss in body weight was not only statistically significant as compared to the 1.2% weight loss among controls but also is considered clinically significant and of the magnitude shown to exert favorable effects on glucose control and blood lipids by the American Heart Association, American College of Cardiology and The Obesity Society guidelines panel for the management of obesity and overweight . This intervention is one of a handful of dyadic-based lifestyle interventions among cancer survivors and among the few that promote change in multiple behaviors. Additionally, it is the only one that has employed a web-based platform. Moreover, while evidence is less consistent across both dyad members and as compared to the waitlist control, the DUET intervention also was associated with favorable effects on waist circumference, caloric restriction, self-reported and objective PA, as well as physical performance and blood glucose and CRP. While the relatively modest sample size may have limited power to detect differences in self-efficacy and social support, the intervention appeared to decrease the number of barriers affecting adherence to a calorically-restricted diet or increased PA. Thus, the theoretical concepts of SCT on which the DUET intervention was framed appear supported by these data and should be preserved in future trials. DUET achieved these favorable effects with minimal touch and excellent retention and safety; hence, results support future web-based interventions. Heretofore, variable success has been reported for multi-behavior, web-based interventions among cancer survivors. Bantum et al. evaluated a comprehensive, 6-week, web-based symptom management program (including PA, weight management, and a healthful, plant-based diet) among 352 adult survivors of various cancers in Hawaii. The "Surviving and Thriving" intervention resulted in significant increases in moderate PA as compared to waitlisted controls but did not show concomitant increases in V and F intake (weight status was unreported). Kanera et al. reported similar findings with a 6-month, web-based program ("Cancer Aftercare Guide") among 462 Dutch survivors of mixed cancers and again found significant increases in self-reported moderate PA (+74.7 min/week) in the intervention arm but failed to detect significant increases in V and F intake in controlled analyses. By achieving improvements in calorie control and evidence of improvement in both self-reported and objectively-measured PA, DUET contributes to the unfolding science related to multi-behavior, web-based interventions and also demonstrates an impact on weight status. DUET is the first web-based intervention that reduced obesity and decreased caloric intake among cancer survivors, though, like the other studies also failed to detect significant differences in diet quality (albeit the sample size was six-to-nine-fold smaller and likely precluded our ability to detect between-arm differences). DUET also promoted improvements in levels of CRP and glucose that have been observed in other more intensive weight loss interventions among cancer survivors, though few differences were detected in other biomarkers typically associated with weight loss, i.e., leptin, adiponectin, and insulin . Curiously, some participant groups (e.g., partners assigned to the intervention) experienced increases in leptin rather than decreases and decreases in HDL cholesterol despite increased PA. Seasonal variation in circulating lipids may explain this latter finding since most participants were accrued during the summer when HDL cholesterol peaks and then descends towards its winter time nadir (corresponding to a 6-month follow-up) . Unlike our previous Reach-out to ENhancE Wellness among older cancer survivors (RENEW) RCT that found significant improvements in both physical and mental QOL with a home-based diet and exercise weight management intervention among 641 breast, prostate, and colorectal cancer survivors , the current RCT only detected significant improvements in physical QOL, which it observed in both study arms. A probable explanation for this discrepancy was the smaller sample size, as the DAMES study , which also has a more modest sample size, likewise was unable to detect changes in this outcome. An innovation of DUET was the expansion of dyadic composition beyond the family unit. This expansion increased the intervention reach and also likely enhanced DUET's accrual (56 dyads in 9 months) versus Healthy Moves (22 dyads over 15 months) and DAMES (68 dyads over 2 years). While spouses and children still comprised roughly two-thirds of chosen partners, other family members, friends, and neighbors comprised the remainder. Furthermore, the significance of the dyadic term for most outcomes suggested that the synergy of the dyads was still strong, despite eligibility not being contingent on family relatedness. Yet, the inability to ascertain an eligible and willing partner served as an enrollment barrier. With social isolation and a lack of companionship reported at levels topping 20% in Western countries , and rates of living alone nearing 30% , there is an obvious need to explore effective "match-making" where cancer survivors could be paired with others based on factors that have the potential to create the synergy observed with naturally-occurring partnerships. The DUET trial had several strengths. It formally tested a theoretically-grounded intervention using a randomized controlled design and validated, rigorous measures that were captured by trained staff blinded to study condition. Moreover, the study sample was diverse in geographic range and age, and the proportion of non-Hispanic Whites was representative of the U.S. population (62.5% vs. 59.3%) . In addition, retention was excellent in both the intervention and waitlisted arms over the 6-month study period--exceeding the 70% benchmark that characterizes a tier-1 trial ; however, drop-out was significantly higher in the waitlisted arm. While differential drop-out can result in bias, this threat is considered minimal given the relatively small numbers (n = 6). Like many feasibility studies, DUET had a relatively modest sample size which may have undermined the statistical power to detect differences. Moreover, because of the focus on feasibility, statistical comparisons, of which there were many, were uncontrolled for multiple testing and may have uncovered spurious findings. The trial also did not assess whether weight loss and behavior changes were maintained long-term. Another weakness was the relatively low enrollment rate, i.e., uptake, especially among survivors of cancers other than breast, which may limit the generalizability of findings. However, the preponderance of breast cancer survivor participation in the current study is a phenomenon that has been reported commonly in traditionally delivered clinical interventions , as well as those that are digital . 5. Conclusions DUET represents a pioneering effort in scalable, multi-behavior weight management interventions to promote cancer prevention and control. DUET not only demonstrated feasibility, safety, and excellent retention, it also promoted significant weight loss via caloric restriction and increased PA. Moreover, it exerted favorable effects on physical performance and markers of glucose metabolism and inflammation. Future studies are needed which are larger in size, scope, and duration and which assess longer-term effects of buddy-system interventions--interventions that include family members but that also extend cancer control to friends and neighbors. Acknowledgments The authors thank all of the cancer survivors and their partners who participated in this trial. We also acknowledge the tremendous contributions and dedication of Teri Hoenemeyer. Additionally, we are grateful to our staff and students: J. Ryan Buckman, MPH, Jackie Causey, Grey Freylersythe, Lauren King, MPH, Doctorre McDade, Christopher Reid, Abigail Sims, MPH, and Fariha Tariq, MPH. We also acknowledge the shared resources available at the University of Arizona's Comprehensive Cancer Center (P30 CA023074). Finally, we appreciate the following supportive resources available at UAB: the O'Neal Comprehensive Cancer Center (P30 CA013148), the Cancer Prevention and Control and Cancer Research for Students (CaRES) training programs (T32 CA047888 and R25CA76023), and Metabolism Shared Resource of the Nutrition and Obesity Research Center (P30 DK079626 and P30 DK56336). Supplementary Materials The following supporting information can be downloaded at: Table S1. Self-reported quality-of-life and factors associated with behavior change among participants in wait-listed vs. DUET study arms at baseline, and 6-month follow-up. Click here for additional data file. Author Contributions Conceptualization, D.W.P., R.A.O., T.E.C., L.Q.R., D.F., H.J.B., K.Y.W. and W.D.-W.; methodology, D.W.P., R.A.O., T.E.C., L.Q.R., K.Y.W. and W.D.-W.; software, W.W.C. and D.F.; validation, D.W.P., R.A.O., W.W.C. and W.D.-W.; formal analysis, R.A.O.; investigation, W.W.C., H.K., D.F., K.B.P. and W.D.-W.; resources, W.D.-W.; data curation, W.W.C., H.K. and K.B.P.; writing--original draft preparation, D.W.P. and W.D.-W.; writing--review and editing, D.W.P., R.A.O., T.E.C., L.Q.R., W.W.C., H.K., D.F., K.B.P., H.J.B., K.Y.W. and W.D.-W.; visualization, D.F. and W.D.-W.; supervision, W.W.C. and W.D.-W.; project administration, W.W.C. and W.D.-W.; funding acquisition, W.D.-W. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of the University of Alabama at Birmingham (IRB#: 300003882; date of initial approval: October 28, 2019) and registered within ClinicalTrials.gov (NCT04132219). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are available on request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 DUET CONSORT diagram. cancers-15-01577-t001_Table 1 Table 1 Characteristics of DUET study participants overall and by arm *. Characteristic Overall (n = 112) Waitlist (n = 56) Intervention (n = 56) Cancer Diagnoses among Survivors (n/%)- Breast - Other + 45/56 (80.3%) 11/56 (19.7%) 22/28 (78.6%) 6/28 (21.4%) 23/28 (82.1%) 5/28 (17.9%) Cancer Stage among Survivors (n/%)- I - II - III - Unknown 23/56 (41.1%) 6/56 (10.7%) 1/56 (1.8%) 26/56 (46.4%) 11/28 (39.3%) 3/28 (10.7%) 0/28 (0%) 14/28 (50%) 12/28 (42.9%) 3/28 (10.7%) 1/28 (3.6%) 12/28 (42.9%) Months elapsed since diagnosis: Mean (sd)/range 67.3 (72.0)/0-303 63.1 (62.7)/12-206 71.0 (80.4)/10-303 Partner's relationship to survivor, miles between - Sibling - Child - Friend - Other (parent, son-in-law, neighbor) Miles separating dyad members: Mean (sd)/range 23/56 (41.1%) 7/56 (12.5%) 6/56 (10.7%) 17/56 (30.4%) 3/56 (5.3%) 7.1 (12.4)/0-56 12/28 (42.9%) 5/28 (17.8%) 3/28 (10.7%) 8/28 (28.6%) 0/28 (0%) 5.7 (7.5)/0-31 11/28 (39.3%) 2/28 (7.2%) 3/28 (10.7%) 9/28 (32.1%) 3/28 (10.7%) 8.5 (16.0)/0-56 Female gender (n/%) 86 (76.8%) 42/56 (75.0%) 44/56 (78.6%) Age (years): Mean (sd)/range 58.4 (12.7)/23-81 58.7 (13.0)/24-81 58.1 (12.5)/23-78 Race (n/%)- Non-Hispanic White - Non-Hispanic Black - Other Race or Ethnicity ++ 70 (62.5%) 41 (36.6%) 1 (0.9%) 34 (60.7%) 22 (39.3%) 0 (0%) 36 (64.3%) 19 (33.9%) 1 (1.8%) Educational Status (n/%)- High School Degree or Less - Some College/Junior College/Trade School - College Degree/Post-Grad - Refused/Unknown 16 (14.3%) 34 (30.4%) 59 (53.5%) 2 (1.8%) 8 (14.3%) 14 (25.0%) 32 (57.1%) 2 (3.6%) 8 (14.3%) 20 (35.7%) 28 (50.0%) 0 (0%) eHealth Literacy: Mean (sd)/range P 3.93 (0.77)/1.63-5.0 4.06 (0.78) 1.63-5.0 3.93 (0.77)/1.63-5.0 Income (n/%)- Less than USD50k/year - USD50k/year or more - Refused/Unknown 18 (16.1%) 48 (42.8%) 46 (41.1%) 7 (12.5%) 28 (50.0%) 21 (37.5%) 11 (19.6%) 20 (35.7%) 25 (44.6%) Employment (n/%)- Employed - Retired - Other (student, disabled) 62 (55.4%) 36 (32.1%) 14 (12.5%) 32 (57.1%) 17 (30.4%) 7 (12.5%) 30 (53.6%) 19 (33.9%) 7 (12.5%) Rural Residence (n/%) 9 (8.0%) 6 (10.7%) 3 (5.4%) BMI: Mean (sd)/range 32.4 (2.75) 25-51 33.3 (6.8) 25-51 31.4 (4.9) 25-45 Daily servings of Vand Fs: Mean (sd)/range 1.8 (1.1) 0.2-6.3 1.9 (1.2) 0.2-6.3 1.7 (1.0) 0.2-3.8 Weekly minutes of MVPA: Mean (sd)/range 43.8 (54.7) 0-280 43.7 (59.1) 0-225 43.8 (60.5) 0-280 Current Smoker (n/%) 5 (4.5%) 1 (1.8%) 4 (7.1%) Number of Comorbidities: Mean (sd)/range 3.0 (2.6) 0-11 3.5 (2.7) 0-11 2.6 (2.4) 0-10 Depressive Symptoms (CESD): Mean (sd)/range SS 3.2 (6.3) 0-24 3.9 (6.1) 0-24 3.9 (6.1) 0-24 * No statistically significant differences were observed between the intervention vs. the waitlist control arm. + Other cancer diagnoses among survivors were colorectal (n = 1), gynecologic (n = 2), renal (n = 2), and prostatic (n = 6). ++ Hispanic ethnicity in combination with various racial groups. P Potential range 1-5. SS Potential range 0-24. cancers-15-01577-t002_Table 2 Table 2 Baseline to 6-m differences in adiposity, diet, and physical activity in waitlist vs. DUET intervention study arms. Waitlist Control (WL) DUET Intervention Significance (p-Values) Baseline Mean (SD) 6M Mean (SD) Baseline Mean (SD) 6M Mean (SD) Between Arm Within Arm Time x Arm Weight (kg) (Primary Outcome) Survivors Partners Dyads 88.0 (17.6) 96.7 (22.6) 92.4 (20.6) 86.9 (16.1) 95.3 (22.9) 91.3 (20.3) 86.8 (18.2) 87.6 (14.0) 87.2 (16.1) 83.8 (18.4) * 84.9 (15.6) * 84.4 (16.9) * 0.656 0.069 0.044 0.001 0.001 <0.001 0.090 0.280 0.033 Waist Circumference (cm) Survivors Partners Dyads 107.1 (17.7) 111.3 (15.3) 109.2 (16.7) 104.6 (14.3) 108.4 (16.4) * 106.6 (15.6) * 106.0 (11.7) 106.5 (9.5) 106.2 (10.6) 102.5 (13.4) * 103.3 (9.5) * 102.9 (11.5) * 0.667 0.144 0.128 0.003 <0.001 <0.001 0.579 0.594 0.339 Calorie Intake/day Survivors Partners Dyads 1645.9 (535.0) 1553.5 (483.7) 1596.9 (515.6) 1532.0 (568.9) 1467.2 (501.3) 1497.7 (542.5) 1400.0 (439.0) 1570.2 (498.3) 1485.1 (473.1) 1265.6 (305.1) 1378.5 (408.3) * 1321.0 (360.6) * 0.027 0.716 0.053 0.046 0.012 0.001 0.836 0.286 0.365 Healthy Eating Index Survivors Partners Dyads 51.8 (10.9) 55.5 (10.5) 54.0 (10.6) 50.8 (12.2) 54.6 (12.2) 52.6 (12.5) 53.9 (13.7) 52.2 (12.0) 53.1 (12.8) 58.8 (15.2) 53.9 (13.4) 56.4 (14.4) 0.092 0.416 0.396 0.360 0.874 0.589 0.163 0.587 0.130 MVPA Self-Report Survivors Partners Dyad 51.9 (61.6) 43.3 (61.8) 45.6 (61.4) 57.3 (61.9) 58.1 (75.4) 57.9 (68.6) 48.5 (67.8) 39.1 (52.9) 43.1 (60.4) 103.9(104.7) * 81.6(113.5) 92.7(108.8) * 0.448 0.299 0.099 0.011 0.025 <0.001 0.189 0.356 0.229 MVPA Accelerometry Survivors Partners Dyads 162.6 (172.4) 148.0 (169.0) 157.7 (171.3) 122.2 (190.2) 146.0 (278.4) 139.0 (236.8) + 108.5 (122.6) 146.4 (121.8) 127.5 (122.4) 150.3 (199.0) * 132.0 (271.4) 141.0 (236.4) 0.726 0.655 0.633 0.047 <0.001 <0.001 0.265 0.819 0.600 * Post-hoc analyses show significant improvements from baseline (p < 0.05); + Post-hoc analyses show significant declines from baseline (p < 0.05). cancers-15-01577-t003_Table 3 Table 3 Baseline to 6-m differences in physical performance in waitlist vs. DUET intervention study arms. Waitlist Control (WL) DUET Intervention Significance (p-Values) Baseline Mean (SD) 6M Mean (SD) Baseline Mean (SD) 6M Mean (SD) Between Arm Within Arm Time x Arm 30-sec Chair Stand (reps) Survivors Partners Dyads 10.5 (2.7) 11.5 (3.0) 11.0 (2.9) * 12.2 (3.9) 14.0 (5.8) 13.1 (5.0) 10.3 (3.1) 11.0 (3.4) 10.7 (3.2) 13.1 (3.6) * 13.5 (3.3) * 13.3 (3.4) * 0.645 0.624 0.871 <0.001 <0.001 <0.001 0.301 0.995 0.380 8' Get-up-and-Go (sec) Survivors Partners Dyads 8.1 (2.7) 7.4 (1.5) 7.8 (2.3) 7.6 (2.8) 7.3 (1.8) 7.6 (2.3) 7.7 (1.7) 7.6 (1.6) 7.6 (1.6) 6.9 (1.5) * 7.1 (2.0) 7.0 (1.8) * 0.290 0.980 0.107 0.004 0.084 0.001 0.470 0.304 0.211 Sit-and-Reach (cm) Survivors Partners Dyads -0.2 (4.6) 0.3 (3.6) 0.1 (4.0) -0.3 (4.7) 0.8 (3.6) 0.4 (4.1) -0.8 (3.0) -0.5 (3.8) -0.6 (3.4) 0.6 (2.0) * 0.5 (2.8) 0.6 (2.4) * 0.889 0.583 0.573 0.036 0.025 0.002 0.015 0.520 0.043 Back Scratch (cm) Survivors Partners Dyads -3.7 (3.5) -2.5 (3.4) -3.0 (3.4) -3.7 (3.2) -2.3 (3.3) -2.9 (3.3) -3.3 (3.5) -3.7 (3.6) -3.5 (3.6) -2.9 (3.9) -3.3 (3.6) -3.1 (3.7) 0.627 0.260 0.511 0.187 0.237 0.104 0.217 0.699 0.147 2-min Step Test (reps) Survivors Partners Dyads 81.0 (22.9) 79.7 (20.8) 80.3 (22.0) 80.4 (23.0) 83.0 (22.5) 80.9 (22.6) 79.9 (23.1) 79.1 (24.2) 79.4 (23.4) 84.4 (23.5) * 87.3 (23.6) 85.8 (23.4) 0.806 0.766 0.442 0.396 0.016 0.037 0.258 0.298 0.077 * Post-hoc analyses show significant improvements from baseline (p < 0.05). cancers-15-01577-t004_Table 4 Table 4 Baseline to 6-month differences in circulating biomarkers in waitlisted vs. DUET intervention study arms. Waitlist (WL) Control DUET Intervention Significance (p-Values) Baseline Mean (SD) 6M Mean (SD) Baseline Mean (SD) 6M Mean (SD) Between Arm Within Arm Time x Arm Glucose (mg/dL) Survivors Partners Dyads 87.8 (25.6) 88.2 (17.8) 88.3 (22.6) 62.6 (34.2) * 65.8 (44.3) 61.2 (38.0) 89.0 (23.1) 82.7 (21.6) 85.9 (22.4) 71.4 (23.1) 67.4 (34.5) * 69.4 (29.2) * 0.175 0.599 0.185 0.002 0.002 <0.001 0.188 0.393 0.148 Total Chol. (mg/dL) Survivors Partners Dyads 269.5 (69.6) 280.3 (58.2) 276.9 (64.4) 252.1 (34.0) 238.1 (42.1) 245.6 (39.2) 268.2 (50.7) 269.5 (59.4) 268.8 (54.6) 255.5 (44.8) 249.3 (38.4) 252.3 (41.3) 0.713 0.969 0.868 0.139 0.005 0.002 0.791 0.303 0.377 HDL Chol. (mg/dL) Survivors Partners Dyads 74.2 (11.7) 72.3 (10.8) 73.5 (11.4) 68.1 (14.4) + 67.6 (12.7) 68.0 (13.8) + 68.4 (10.9) 69.0 (11.9) 68.7 (11.3) 63.2 (10.6) 61.2 (10.5) + 62.1 (10.5)+ 0.130 0.119 0.017 <0.001 0.002 <0.001 0.593 0.405 0.977 CRP (mg/dL) Survivors Partners Dyads 2.0 (1.4) 3.2 (2.6) 2.7 (2.1) 1.7 (1.5) 2.7 (2.2) 2.3 (1.9) 4.5 (7.4) 2.7 (4.0) 3.6 (6.0) 3.0 (3.5) 2.3 (4.5) 2.6 (4.0) * 0.182 0.062 0.403 0.041 0.142 0.025 0.824 0.328 0.183 Leptin Survivors Partners Dyads 102.2 (71.6) 113.0 (76.7) 108.1 (75.0) 91.6 (50.2) 116.0 (79.3) 103.7 (68.6) 101.8 (65.1) 74.0 (51.0) 88.2 (59.7) 100.9 (86.5) 83.8 (74.8) 92.1 (80.1) 0.767 0.017 0.161 0.689 0.945 0.705 0.735 0.523 0.448 Adiponectin Survivors Partners Dyads 34.6 (23.1) 26.3 (9.9) 31.1 (18.8) 31.9 (15.2) 24.7 (9.4) 28.4 (13.4) 26.8 (20.8) 25.2 (13.6) 26.0 (17.5) 23.8 (11.1) 23.3 (12.6) 23.5 (11.7) 0.283 0.380 0.113 0.343 0.174 0.043 0.137 0.247 0.408 * Post-hoc analyses show significant improvements from baseline (p < 0.05); + Post-hoc analyses show significant declines from baseline (p < 0.05) Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000641 | Diagnostics (Basel) Diagnostics (Basel) diagnostics Diagnostics 2075-4418 MDPI 10.3390/diagnostics13050897 diagnostics-13-00897 Brief Report Development of a Method for Detection of SARS-CoV-2 Nucleocapsid Antibodies on Dried Blood Spot by DELFIA Immunoassay Damiani Verena Conceptualization Formal analysis Data curation Writing - original draft 12* Pizzinato Erika Methodology 12 Cicalini Ilaria Conceptualization 12 Demattia Gianmaria Methodology 1 Zucchelli Mirco Methodology 12 Natale Luca Methodology 1 Palmarini Claudia Methodology 1 Di Marzio Claudia Methodology 1 Federici Luca Writing - review & editing 12 De Laurenzi Vincenzo Writing - review & editing Funding acquisition 12 Pieragostino Damiana Conceptualization Data curation Writing - review & editing 12 Agarwal Vikas Academic Editor 1 Center for Advanced Studies and Technology (CAST), "G. d'Annunzio" University of Chieti-Pescara, 66100 Chieti, Italy 2 Department of Innovative Technologies in Medicine and Dentistry, "G. d'Annunzio" University of Chieti-Pescara, 66100 Chieti, Italy * Correspondence: [email protected]; Tel.: +39-0871355582 27 2 2023 3 2023 13 5 89701 2 2023 21 2 2023 25 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Antibodies against the SARS-CoV-2 nucleocapsid protein are produced by the immune system in response to SARS-CoV-2 infection, but most available vaccines developed to fight the pandemic spread target the SARS-CoV-2 spike protein. The aim of this study was to improve the detection of antibodies against the SARS-CoV-2 nucleocapsid by providing a simple and robust method applicable to a large population. For this purpose, we developed a DELFIA immunoassay on dried blood spots (DBSs) by converting a commercially available IVD ELISA assay. A total of forty-seven paired plasma and dried blood spots were collected from vaccinated and/or previously SARS-CoV-2-infected subjects. The DBS-DELFIA resulted in a wider dynamic range and higher sensitivity for detecting antibodies against the SARS-CoV-2 nucleocapsid. Moreover, the DBS-DELFIA showed a good total intra-assay coefficient of variability of 14.6%. Finally, a strong correlation was found between SARS-CoV-2 nucleocapsid antibodies detected by the DBS-DELFIA and ELISA immunoassays (r = 0.9). Therefore, the association of dried blood sampling with DELFIA technology may provide an easier, minimally invasive, and accurate measurement of SARS-CoV-2 nucleocapsid antibodies in previously SARS-CoV-2-infected subjects. In conclusion, these results justify further research to develop a certified IVD DBS-DELFIA assay for detecting SARS-CoV-2 nucleocapsid antibodies useful for diagnostics as well as for serosurveillance studies. SARS-CoV-2 nucleocapsid antibodies ELISA DELFIA dried blood spot This research received no external funding. pmc1. Introduction Today, almost three years after the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was first identified in China, the COVID-19 pandemic continues spreading across the world. Indeed, despite a massive vaccination campaign that has been activated against the pandemic worldwide, new highly transmissible SARS-CoV-2 variants are continuously emerging, and it is becoming increasingly clear that they may evade neutralizing antibodies generated by previous infection and/or vaccination and thus contribute to the virus circulation . SARS-CoV-2 is an enveloped, single-stranded, positive-sense RNA virus . The four main structural proteins encoded by the genome include the spike (S), membrane (M), envelope (E), and nucleocapsid (N) proteins . The S protein is a trimeric protein comprising two subunits, namely S1 and S2. The S1 subunit mediates binding to host cells via interactions between its receptor-binding domain (RBD) and the human receptor angiotensin-converting enzyme 2 (ACE2), whereas the S2 subunit is responsible for membrane fusion, which is required for virus entry . The N protein, in which the viral genome is encapsulated, plays a fundamental function in viral RNA transcription, replication, and virion assembly. Although the N protein is highly immunogenic and a major target for antibody response , the S protein was employed to develop vaccines first. The M and E viral structural proteins have not been investigated as vaccine targets due to their inability to induce complete immune protection; indeed, only a significant cellular immune response was elicited, whereas no robust humoral immunity was detected . Currently, the majority of vaccines available on the European market target the spike protein, which is the leading immunogenic protein . However, recently, the N protein has attracted much attention for vaccine development because it is more conservative, more stable, and has fewer mutations than the S protein . In response to SARS-CoV-2 infection or vaccination, most individuals develop antibodies to the N and S proteins within 1 or 2 weeks, and these antibodies can be measured as an indicator of COVID-19 prevalence; moreover, they allow for the monitoring of seroconversion in the population and are essential elements in developing strategies for SARS-CoV-2 infection prevention and control . Plasma and sera isolated from venous blood represent the conventional sample types used for the evaluation of SARS-CoV-2 antibody responses. However, the collection of these samples is invasive and requires trained personnel and equipment for immediate processing. Once collected, plasma and sera must be stored and shipped refrigerated. Therefore, dried blood spot (DBS) testing, already applied in the fields of anti-doping, toxicology, newborn screening, and the diagnosis of infectious diseases, has been validated for the measurement of SARS-CoV-2 IgG antibodies against the N and S proteins . The most commonly used method for serological tests is the enzyme-linked immunosorbent assay (ELISA). During the SARS-CoV-2 pandemic, several ELISA methods were developed to determine SARS-CoV-2 antigens and antibodies, qualitatively and quantitatively, with great sensitivity and specificity . However, a colorimetric ELISA is affected by a narrow linear range for the optical density (OD), which is common to absorbance-based measurements. For this reason, an unknown sample concentration could fall outside the standard curve, introducing the challenge of testing multiple dilutions from the same, potentially limited, sample. The aim of this study was to develop a time-resolved fluorometry-based dissociation-enhanced lanthanide fluorescence immunoassay for detecting nucleocapsid antibodies to SARS-CoV-2 by using dried blood spots (DBS-DELFIA). The newly developed assay was compared to a commercially available, certified in vitro diagnostic (IVD) qualitative ELISA. The DBS-DELFIA test resulted in higher sensitivity and a wider dynamic range compared to the ELISA test. These results justify further research to develop a certified IVD for SARS-CoV-2 IgG anti-N detection by DBS-DELFIA technology. 2. Materials and Methods 2.1. Study Subjects Forty-seven subjects were enrolled in this study. Sex, age, vaccination doses, and SARS-CoV-2 infection history are reported in Table S1. The study was conducted at the Center for Studies and Advanced Technologies (CAST), "G. D'Annunzio" of Chieti-Pescara, Italy, in accordance with the Declaration of Helsinki and the approval no. 16 of 1 July 2021 of the Ethics Committee of "G. D'Annunzio" University of Chieti-Pescara. Written informed consent forms were obtained from all the enrolled subjects. 2.2. Plasma and Dried Blood Spots Collection Whole blood was collected via venipuncture in Vacumed sodium citrate tubes (3.2%, FL MEDICAL s.r.l., Padova, Italy) to prevent coagulation, and processed within 6 h of collection. DBS samples were prepared from venous whole blood by transferring approximately 40 mL of blood to each circle of a filter paper card. Cards were then air-dried for at least two hours, placed into bags with a desiccant dehumidifier, and stored at -20 degC. The remaining whole blood was centrifuged at 3000 rpm for 12 min. Plasma aliquots were taken and transferred into sterile microtubes and stored at -80 degC until analysis. 2.3. SARS-CoV-2 Nucleocapsid IgG ELISA SARS-CoV-2 NP IgG ELISA kit [CE-IVD] (ImmunoDiagnostics, Hong Kong, China) was used following manufacturer's recommendations. Briefly, 50 mL of negative control, 100 mL of the test sample (diluted plasma 1:100), and 100 mL of Assay Buffer (ImmunoDiagnostics) as blank were added onto the antigen-coated plate in duplicate, as the test recommended. Subsequently, the plate was incubated at room temperature (RT) for 1 h. Then, each well was manually washed 3 times with Wash Buffer (ImmunoDiagnostics) included in the kit. Next, 100 mL of Detection Solution (ImmunoDiagnostics) was added to each well and incubated for 1 h at RT. Then, after the wash step, 100 mL of Substrate Solution (ImmunoDiagnostics) was added to each well and incubated for 15 min at RT, protected from light. Finally, we added 100 mL of Stop Solution (ImmunoDiagnostics) to each well, and, after 10 min, absorbance was measured at 450 nm by Victor Nivo microplate reader (PerkinElmer, Turku, Finland). 2.4. Conversion from SARS-CoV-2 Nucleocapsid IgG ELISA to SARS-CoV-2 Nucleocapsid IgG DBS-DELFIA The antigen-coated plate from SARS-CoV-2 NP IgG ELISA kit (ImmunoDiagnostics) was used. DBS disks were punched out into 3.2 mm disks by using the PerkinElmer DBS Puncher, while plasma samples were diluted 1:100 with DELFIA Assay Buffer (PerkinElmer). Next, DBS disks were extracted with 100 mL DELFIA Assay Buffer directly onto the antigen-coated plate, whereas 100 mL of diluted plasma was transferred to the plate, and both DBS disks and plasma samples were incubated for 2 h at RT on a plate shaker set at 300 rpm. Then, after removing DBS disks and plasma samples, each well was manually washed 4 times with DELFIA Wash Solution (PerkinElmer). Next, 100 mL (200 ng/mL) of DELFIA Eu-labeled Anti-human IgG antibody (PerkinElmer) was added to each well and incubated for 1 h at RT on a plate shaker set at 300 rpm. Subsequently, each well was washed 6 times with DELFIA Wash Solution. Finally, 200 mL of DELFIA Enhancement Solution (PerkinElmer) was added, and the plate was read after 10 min of incubation time by Victor Nivo microplate reader using fluorescence (TRF) settings. 2.5. Linearity, Precision Study, and Statistical Analysis To test the linearity of both ELISA and DBS-DELFIA immunoassays, a SARS-CoV-2 anti-N IgG positive sample was diluted sequentially 12 times with a seronegative plasma and tested in duplicate. Dilution percentages are listed in Table S2. Data for linearity and intra-assay precision were collected by the same operator. All statistical analyses were carried out using GraphPad Prism 9.0.2 software (GraphPad Software, La Jolla, CA, USA). 3. Results 3.1. Conversion from ELISA to DBS-DELFIA Immunoassay for SARS-CoV-2 Nucleocapsid Antibody Detection The study population included fifteen subjects who had never tested positive for SARS-CoV-2 infection, thirty-two subjects who reported a positive nasopharyngeal swab (NPS) test, and forty-three individuals who had completed the vaccination schedule. The age of the subjects ranged from 25 to 60 (mean = 35.6), and 70% were female (Table S1). Firstly, we conducted an experiment to set up the method for converting ELISA into DBS-DELFIA by using paired plasma samples and dried blood spots from two subjects, one negative and one positive for SARS-CoV-2 infection. The cut-off value used by the qualitative SARS-CoV-2 nucleocapsid antibody ELISA kit was 0.2 OD. We analyzed, in parallel and in duplicate, the plasma diluted 100 times both by ELISA and by DELFIA, and the paired DBSs by DELFIA following the manufactures' instructions (Supplementary Table S3). The positive subject, with 2.47 +- 0.06 OD, had almost comparable values for plasma DELFIA and DBS-DELFIA (510,238 +- 11,641.8 TRF and 496,748 +- 96,228.75 TRF, respectively). On the other hand, the negative subject showed different values between plasma DELFIA and DBS-DELFIA (24,067.5 +- 678.11 TRF and 93,547.5 +- 12,087.99 TRF). For this reason, we attempted to improve the DBS-DELFIA protocol using the commercially available ELISA-to-DELFIA conversion kit protocol distributed by PerkinElmer, which is fully described in the Section 2. 3.2. DBS-DELFIA Method Linearity and Performance Assessment To prove the linearity of the method for quantifying IgG anti-N, we performed sequential dilutions using two blood samples, one that tested negative (dilution percentage 0, 0.07 OD) and one that tested positive (dilution percentage 100, 3.33 OD) for anti-SARS-CoV-2 nucleocapsid antibodies by the ELISA method. We analyzed 12 paired dilutions by both ELISA and the new DBS-DELFIA method, using different sample matrices (plasma or DBS) (Table S2). The DELFIA method had a wider dynamic range than conventional ELISA . While the linear range of DELFIA covered dilution percentages up to 100 with an R2 equal to 0.97, the ELISA method only showed linearity below the 40-dilution percentage (R2 = 0.97). Next, we examined the cut-off value referred to in the ELISA kit and noted that for 1:100 (0.191 +- 0.03 OD) and 2:100 (0.241 +- 0.03 OD) dilutions, the test resulted in negative and positive results, respectively. However, the paired samples analyzed by DBS-DELFIA measured 30,351 +- 4912 TRF and 53,652 +- 17,674 TRF, showing a higher sensitivity potential. In order to evaluate intra-assay precision, we evaluated the coefficient of variability in percentage (CV%) of both obtained curves. We obtained 7% and 15.2 CV% for ELISA and DBS-DELFIA, respectively. Finally, we calculated the Limit of Detection (LOD) and Limit of Quantification (LOQ) for both methods by using blank sample replicates (n = 8), obtaining 0.09 and 0.18 values, respectively, for the ELISA immunoassay, and 2797 and 3787 values for the DBS-DELFIA one. 3.3. Evaluation of SARS-CoV-2 IgG Anti-N Detection by ELISA vs. DBS-DELFIA Immunoassay All paired plasma samples and dried blood spots were collected and analyzed by both ELISA and DBS-DELFIA immunoassays for the detection of SARS-CoV-2 nucleocapsid antibodies (Table S4). A significant positive correlation (r = 0.9, p < 0.0001) between IgG anti-N measured on DBS and plasma with DELFIA and ELISA immunoassays, respectively, was observed . Fifteen subjects had never tested positive for SARS-CoV-2 infection using NPS. At the same time, thirty-two subjects tested positive at different times after being tested for SARS-CoV-2 nucleocapsid antibodies. All subjects without SARS-CoV-2 infection tested negative on the ELISA test (OD < 0.2). Overall, these subjects presented a mean TRF below 2.0 x 104. Despite significant differences in IgG anti-N levels evaluated by both ELISA and DBS-DELFIA between subjects who were negative or positive for SARS-CoV-2 infection , the receiver operator characteristic (ROC) curve analysis revealed higher sensitivity (87.5%, AUC: 0.91, p < 0.0001) for the DBS-DELFIA assay compared to ELISA (71.88%, AUC: 0.92, p < 0.0001) while the specificity was 100% for both assays . Moreover, the ROC analysis showed that 87.5% sensitivity was achieved above 20,120 TRF; therefore, we set the cut-off value for the DBS-DELFIA assay at 2.0 x 104 . Five samples (PZ 24, 25, 32, 44, and 47) tested negative when analyzed with ELISA but appeared above the assessed cut-off value when analyzed with the DBS-DELFIA method. By calculating the percentage difference between the cut-off value and LOQ for both evaluated methods, we observed that the ELISA test's positivity limit was 2% higher than the LOQ, while the DBS-DELFIA positivity limit was more than 80% higher than its relative LOQ. 4. Discussion With thousands of new cases daily, the ongoing scenario indicates that the SARS-CoV-2 pandemic will continue to evolve . Indeed, as SARS-CoV-2 continues to spread in human populations with fewer susceptible hosts, the risk of selecting more infectious variants or antibody-evasive mutations is expected to increase. Even to avoid undiagnosed cases of SARS-CoV-2 infection in emergencies , viral tests are used for the assessment of current infection with SARS-CoV-2 by testing respiratory tract specimens (throat swabs, sputum, nasopharyngeal swabs, nasal swabs, and bronchoalveolar lavage fluid). There are two main types of viral tests: nucleic acid amplification tests (NAATs, such as reverse transcription polymerase chain reaction) and antigen tests . However, the collection of samples from the respiratory tract is relatively complicated and causes significant discomfort to subjects. Moreover, the costs of NAATs continue to remain high. Therefore, growing interest has been placed in developing serological tests for the detection of anti-SARS-CoV-2 antibodies to help identify people who have been infected with SARS-CoV-2, have recovered from COVID-19, or have been vaccinated . Notably, DBS specimens have been seen to be reliably used as an alternative to serum samples for SARS-CoV-2 antibody measurement, facilitating serosurveillance efforts . We have already validated the GSP(r)(r)/DELFIA(r)(r) Anti-SARS-CoV-2 IgG Kit for measuring anti-S1 antibodies by using DBS, demonstrating the feasibility of such serological tests for high-throughput serosurveillance . However, to ascertain whether a recent SARS-CoV-2 infection occurred among subjects who have previously been vaccinated for the prevention of COVID-19, high-sensitivity immunoassays for SARS-CoV-2 anti-nucleocapsid antibodies are required. Notably, there is evidence of longer durability of anti-spike antibodies after vaccination with the mRNA vaccine in subjects with previous infection, and the risk of new SARS-CoV-2 infection appears to be higher in previously uninfected individuals . This is why knowing the SARS-CoV-2 nucleocapsid antibody profile is extremely important. For this purpose, we established a dissociation-enhanced lanthanide fluorescence (DELFIA) immunoassay for the evaluation of SARS-CoV-2 anti-nucleocapsid IgG antibody status by analyzing dried blood spots (DBS-DELFIA). We converted a commercially available CE-IVD SARS-CoV-2 nucleocapsid protein IgG ELISA kit validated with serum or plasma samples into the DBS-DELFIA. Our results confirm a wider dynamic range and higher sensitivity of the DBS-DELFIA compared to the ELISA immunoassay in recognizing IgG anti-N against SARS-CoV-2, revealing lower amounts of antibodies even when several days have passed since the previous infection. Moreover, the total intra-assay coefficient of variability of the DBS-DELFIA was 14.6%, indicating good assay precision. Finally, a strong correlation between SARS-CoV-2 nucleocapsid antibodies detected by the DBS-DELFIA and ELISA immunoassays was found (r = 0.9). Therefore, the association of dried blood sampling with DELFIA technology, in addition to a simple, non-invasive approach and little time-consuming sample preparation, may provide a more sensitive and accurate measurement of SARS-CoV-2 nucleocapsid antibodies than tests currently available, particularly for low detectable or higher quantizable antibody levels among SARS-CoV-2-infected subjects, with the further potential of being fully automated . 5. Conclusions In summary, we developed a new serological immunoassay specifically for the detection of SARS-CoV-2 nucleocapsid antibodies employing DELFIA technology. The assay showed increased sensitivity and appears to be particularly suitable for high-throughput serosurveillance studies, thus justifying further research aimed at developing a certified IVD DBS-DELFIA assay for detecting SARS-CoV-2 nucleocapsid antibodies. Supplementary Materials The following supporting information can be downloaded at: Table S1: Clinical parameters of the 47 plasma-DBS paired samples collected. Table S2: Linearity assessment of ELISA and DBS-DELFIA methods analyzing SARS-CoV-2 IgG anti-N on sequential dilutions. Table S3: Data of the set-up experiment for ELISA to DELFIA and next to DBS-DELFIA conversion. Table S4: Data of analytical sensitivity of ELISA and DBS-DELFIA immunoassays used to evaluate assay precision. Click here for additional data file. Author Contributions Conceptualization, V.D., I.C. and D.P.; Methodology, V.D. and E.P.; Sample Collection, G.D., M.Z., C.P., C.D.M. and L.N.; Formal analysis, V.D.; Data curation, V.D. and D.P.; Writing--original draft preparation, V.D. and E.P.; Writing--review and editing, D.P., L.F. and V.D.L.; Funding acquisition, V.D.L. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Ethics Committee of "G. d'Annunzio" University, Chieti, Italy (protocol no. 16 of 1 July 2021). Informed Consent Statement Written informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are available in Supplementary Materials. Conflicts of Interest The authors declare no conflict of interest. Figure 1 DBS-DELFIA vs. ELISA dilutions linearity range. (A) DBS-DELFIA dilutions results are expressed as TRF, whereas ELISA results are expressed as OD values. (B) Magnification of rectangular black dashed section highlighting ELISA linearity range. Red dashed line expressed ELISA cut-off value (0.2 OD). Figure 2 Correlation between DBS-DELFIA and plasma ELISA for SARS-CoV-2 nucleocapsid IgG antibody detection. Figure 3 ELISA vs. DBS-DELFIA performance analysis. (A,C) Scatter plots and (B,D) receiver operating characteristic (ROC) curves for SARS-CoV-2 nucleocapsid antibody detection by ELISA and DBS-DELFIA, respectively. IgG anti-N concentration in each subject. The plots show mean +- standard deviation. ROC curves indicate the area under the curve (AUC). **** p < 0.0001. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Dhama K. Nainu F. Frediansyah A. Yatoo M.I. Mohapatra R.K. Chakraborty S. Zhou H. Islam R. Mamada S.S. Kusuma H.I. Global emerging Omicron variant of SARS-CoV-2: Impacts, challenges and strategies J. Infect. Public Health 2023 16 4 14 10.1016/j.jiph.2022.11.024 36446204 2. Guimaraes P.O. Quirk D. Furtado R.H. Maia L.N. Saraiva J.F. Antunes M.O. Kalil Filho R. Junior V.M. Soeiro A.M. Tognon A.P. Emerging Variants of SARS-CoV-2 And Novel Therapeutics Against Coronavirus (COVID-19) N. Engl. J. Med. 2022 385 406 415 10.1056/NEJMoa2101643 3. Akkiz H. The Biological Functions and Clinical Significance of SARS-CoV-2 Variants of Corcern Front. Med. 2022 9 849217 10.3389/fmed.2022.849217 4. Ahmad L. Implication of SARS-CoV-2 Immune Escape Spike Variants on Secondary and Vaccine Breakthrough Infections Front. Immunol. 2021 12 4563 10.3389/fimmu.2021.742167 5. Alexandersen S. Chamings A. Bhatta T.R. SARS-CoV-2 genomic and subgenomic RNAs in diagnostic samples are not an indicator of active replication Nat. Commun. 2020 11 6059 10.1038/s41467-020-19883-7 33247099 6. Abavisani M. Rahimian K. Mahdavi B. Tokhanbigli S. Siasakht M.M. Farhadi A. Kodori M. Mahmanzar M. Meshkat Z. Mutations in SARS-CoV-2 structural proteins: A global analysis Virol. J. 2022 19 220 10.1186/s12985-022-01951-7 36528612 7. Schaub J.M. Chou C.-W. Kuo H.-C. Javanmardi K. Hsieh C.-L. Goldsmith J. DiVenere A.M. Le K.C. Wrapp D. Byrne P.O. Expression and characterization of SARS-CoV-2 spike proteins Nat. Protoc. 2021 16 5339 5356 10.1038/s41596-021-00623-0 34611365 8. Lutomski C.A. El-Baba T.J. Bolla J.R. Robinson C.V. Multiple Roles of SARS-CoV-2 N Protein Facilitated by Proteoform-Specific Interactions with RNA, Host Proteins, and Convalescent Antibodies JACS Au 2021 1 1147 1157 10.1021/jacsau.1c00139 34462738 9. Bai Z. Cao Y. Liu W. Li J. The SARS-CoV-2 Nucleocapsid Protein and Its Role in Viral Structure, Biological Functions, and a Potential Target for Drug or Vaccine Mitigation Viruses 2021 13 1115 10.3390/v13061115 34200602 10. Chen J. Deng Y. Huang B. Han D. Wang W. Huang M. Zhai C. Zhao Z. Yang R. Zhao Y. DNA Vaccines Expressing the Envelope and Membrane Proteins Provide Partial Protection Against SARS-CoV-2 in Mice Front. Immunol. 2022 13 727 10.3389/fimmu.2022.827605 11. Li M. Wang H. Tian L. Pang Z. Yang Q. Huang T. Fan J. Song L. Tong Y. Fan H. COVID-19 vaccine development: Milestones, lessons and prospects Signal Transduct. Target. Ther. 2022 7 146 10.1038/s41392-022-00996-y 35504917 12. Hassine I.H. COVID-19 vaccines and variants of concern: A review Rev. Med. Virol. 2021 32 e2313 10.1002/rmv.2313 34755408 13. Oronsky B. Larson C. Caroen S. Hedjran F. Sanchez A. Prokopenko E. Reid T. Nucleocapsid as a next-generation COVID-19 vaccine candidate Int. J. Infect. Dis. 2022 122 529 530 10.1016/j.ijid.2022.06.046 35788417 14. Wu W. Cheng Y. Zhou H. Sun C. Zhang S. The SARS-CoV-2 nucleocapsid protein: Its role in the viral life cycle, structure and functions, and use as a potential target in the development of vaccines and diagnostics Virol. J. 2023 20 6 10.1186/s12985-023-01968-6 36627683 15. Feng W. Xiang Y. Wu L. Chen Z. Li Q. Chen J. Guo Y. Xia D. Chen N. Zhang L. Nucleocapsid protein of SARS-CoV-2 is a potential target for developing new generation of vaccine J. Clin. Lab. Anal. 2022 36 e24479 10.1002/jcla.24479 35527696 16. Damasceno D.H.P. Amaral A.A. Silva C.A. Simoes e Silva A.C. The impact of Vaccination worldwide on SARS-CoV-2 infection: A Review on Vaccine Mechanisms, Results of Clinical Trials, Vaccinal Coverage and Interactions with Novel Variants Curr. Med. Chem. 2022 29 2673 2690 10.2174/0929867328666210902094254 34473613 17. Nair S. Chen X. Biology of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and the humoral immunoresponse: A systematic review of evidence to support global policy-level actions and research Glob. Health J. 2021 6 38 43 10.1016/j.glohj.2021.11.005 34840854 18. Coltart C.E.M. Wells D. Sutherland E. Fowler A. National cross-sectional survey of 1.14 million NHS staff SARS-CoV-2 serology tests: A comparison of NHS staff with regional community seroconversion rates BMJ Open 2021 11 e049703 10.1136/bmjopen-2021-049703 19. Galla L. Cosma C. Bertan M. Altinier S. Zaninotto M. Basso D. Burlina A. Padoan A. Plebani M. Seroprevalence of SARS-CoV-2 Antibodies in Italy in Newborn Dried Blood Spots Clin. Chem. Lab. Med. 2022 61 311 315 10.1515/cclm-2022-0948 36282964 20. Liu F. Nguyen M. Vijayakumar P. Kaplan A. Meir A. Dai Y. Wang E. Walsh H. Ring A.M. Omer S.B. Newborn Dried Blood Spots for Serologic Surveys of COVID-19 Pediatr. Infect. Dis. J. 2020 39 e454 e456 10.1097/INF.0000000000002918 33105339 21. Havers F.P. Reed C. Lim T. Montgomery J.M. Klena J.D. Hall A.J. Fry A.M. Cannon D.L. Chiang C.-F. Gibbons A. Seroprevalence of Antibodies to SARS-CoV-2 in 10 Sites in the United States, 23 March-12 May 2020 JAMA Intern. Med. 2020 180 1576 10.1001/jamainternmed.2020.4130 22. Styer L.M. Hoen R. Rock J. Yauney E. Nemeth K. Bievenue R. Parker M.M. High-Throughput Multiplex SARS-CoV-2 IgG Microsphere Immunoassay for Dried Blood Spots: A Public Health Strategy for Enhanced Serosurvey Capacity Microbiol. Spectr. 2021 9 e0013421 10.1128/Spectrum.00134-21 34319133 23. Zava T.T. Zava D.T. Validation of dried blood spot sample modifications to two commercially available COVID-19 IgG antibody immunoassays Bioanalysis 2021 13 13 28 10.4155/bio-2020-0289 33319585 24. Damjanovic A. Styer L.M. Nemeth K. Yauney E. Rock J.M. Bievenue R. Hoen R. Ehrbar D. Kay D.M. Caggana M. Utility of Newborn Dried Blood Spots to Ascertain Seroprevalence of SARS-CoV-2 Antibodies Among Individuals Giving Birth in New York State, November 2019 to November 2021 JAMA Netw. Open 2022 5 e2227995 10.1001/jamanetworkopen.2022.27995 35994287 25. Luo J. Brakel A. Krizsan A. Ludwig T. Motzing M. Volke D. Lakowa N. Grunewald T. Lehmann C. Wolf J. Sensitive and Specific Serological ELISA for the Detection of SARS-CoV-2 Infections Virol. J. 2022 19 50 10.1186/s12985-022-01768-4 35305688 26. Jacot D. Moraz M. Coste A.T. Aubry C. Sacks J.A. Greub G. Croxatto A. Evaluation of sixteen ELISA SARS-CoV-2 serological tests J. Clin. Virol. 2021 142 104931 10.1016/j.jcv.2021.104931 34365228 27. Rivera-Olivero I.A. Henriquez-Trujillo A.R. Kyriakidis N.C. Ortiz-Prado E. Laglaguano J.C. Vallejo-Janeta A.P. Lozada T. Garcia-Bereguiain M.A. Diagnostic Performance of Seven Commercial COVID-19 Serology Tests Available in South America Front. Cell. Infect. Microbiol. 2022 12 10.3389/fcimb.2022.787987 28. WHO Coronavirus (COVID-19) Dashboard|WHO Coronavirus (COVID-19) Dashboard with Vaccination Data Available online: (accessed on 19 January 2023) 29. Papi C. Spagni G. Alexandre A. Calabresi P. Della Marca G. Broccolini A. Unprotected stroke management in an undiagnosed case of Severe Acute Respiratory Syndrome Coronavirus 2 infection J. Stroke Cerebrovasc. Dis. 2020 29 104981 10.1016/j.jstrokecerebrovasdis.2020.104981 32807416 30. COVID-19 Testing: What You Need to Know|CDC Available online: (accessed on 19 January 2023) 31. Lisboa Bastos M. Tavaziva G. Abidi S.K. Campbell J.R. Haraoui L.-P. Johnston J.C. Lan Z. Law S. MacLean E. Trajman A. Diagnostic accuracy of serological tests for COVID-19: Systematic review and meta-analysis BMJ 2020 370 m2516 10.1136/bmj.m2516 32611558 32. Zheng X. Duan R.H. Gong F. Wei X. Dong Y. Chen R. Liang M.Y. Tang C. Lu L. Accuracy of serological tests for COVID-19: A systematic review and meta-analysis Front. Public Health 2022 10 923525 10.3389/fpubh.2022.923525 36589993 33. Sims M.D. Podolsky R.H. Childers K.L. Higgins B. Trueman J. Homayouni R. Voss D.R. Berkiw-Scenna N. Keil H. Kennedy R.H. Dried blood spots are a valid alternative to venipuncture for COVID-19 antibody testing J. Immunol. Methods 2023 513 113420 10.1016/j.jim.2022.113420 36596443 34. Toh Z.Q. Higgins R.A. Anderson J. Mazarakis N. Do L.A.H. Rautenbacher K. Ramos P. Dohle K. Tosif S. Crawford N. The use of dried blood spots for the serological evaluation of SARS-CoV-2 antibodies J. Public Health 2021 44 e260 e263 10.1093/pubmed/fdab011 35. Cicalini I. del Boccio P. Zucchelli M. Rossi C. Natale L. de Mattia G. de Bellis D. Damiani V. Tommolini M.L. Pizzinato E. Validation of the GSP(r)/DELFIA(r) Anti-SARS-CoV-2 IgG Kit Using Dried Blood Samples for High-Throughput Serosurveillance and Standardized Quantitative Measurement of Anti-Spike S1 IgG Antibody Responses Post-Vaccination Vaccines 2022 10 514 10.3390/vaccines10040514 35455263 36. Stellini R. Gianello R. Gomarasca W. Durability of anti-spike antibodies after vaccination with mRNA SARS-CoV-2 vaccine is longer in subjects with previous infection: Could the booster dose be delayed? Infection 2022 50 1573 1577 10.1007/s15010-022-01816-9 35391649 37. Hall V. Foulkes S. Insalata F. Kirwan P. Saei A. Atti A. Wellington E. Khawam J. Munro K. Cole M. Protection against SARS-CoV-2 after COVID-19 Vaccination and Previous Infection N. Engl. J. Med. 2022 386 1207 1220 10.1056/NEJMoa2118691 35172051 |
PMC10000642 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050698 cells-12-00698 Review Cell Replacement Therapy for Type 1 Diabetes Patients: Potential Mechanisms Leading to Stem-Cell-Derived Pancreatic b-Cell Loss upon Transplant Shilleh Ali H. Conceptualization Writing - original draft Writing - review & editing 1* Russ Holger A. Conceptualization Writing - original draft Writing - review & editing 123* Kim Sang Woo Academic Editor 1 Barbara Davis Center for Diabetes, School of Medicine, University of Colorado Anschutz Medical Campus, Aurora, CO 80045, USA 2 Department of Pharmacology and Therapeutics, School of Medicine, University of Florida, Gainesville, FL 32610, USA 3 Diabetes Institute, School of Medicine, University of Florida, Gainesville, FL 32610, USA * Correspondence: [email protected] (A.H.S.); [email protected] (H.A.R.) 22 2 2023 3 2023 12 5 69816 11 2022 09 2 2023 20 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Cell replacement therapy using stem-cell-derived insulin-producing b-like cells (sBCs) has been proposed as a practical cure for patients with type one diabetes (T1D). sBCs can correct diabetes in preclinical animal models, demonstrating the promise of this stem cell-based approach. However, in vivo studies have demonstrated that most sBCs, similarly to cadaveric human islets, are lost upon transplantation due to ischemia and other unknown mechanisms. Hence, there is a critical knowledge gap in the current field concerning the fate of sBCs upon engraftment. Here we review, discuss effects, and propose additional potential mechanisms that could contribute toward b-cell loss in vivo. We summarize and highlight some of the literature on phenotypic loss in b-cells under both steady, stressed, and diseased diabetic conditions. Specifically, we focus on b-cell death, dedifferentiation into progenitors, trans-differentiation into other hormone-expressing cells, and/or interconversion into less functional b-cell subtypes as potential mechanisms. While current cell replacement therapy efforts employing sBCs carry great promise as an abundant cell source, addressing the somewhat neglected aspect of b-cell loss in vivo will further accelerate sBC transplantation as a promising therapeutic modality that could significantly enhance the life quality of T1D patients. cell replacement therapy type 1 diabetes stem-cell-derived b-like cells autoimmune diabetes transplantation ischemia transdifferentiation dedifferentiation cell death pancreatic progenitor NIH/NIDDKR01DK12044 R01DK132387 NIDDK/HIRN RRIDSCR_014393 the Culshaw Junior Investigator Award in DiabetesUC24 DK104162 pre-doctoral training grants in Stem Cell Biology5T32AR007411-35 Interdisciplinary Bioengineering Research Training in DiabetesT32 DK 120520 Work in the laboratory of H.A.R. is or was supported by NIH/NIDDK grants R01DK12044, R01DK132387, a new investigator award from NIDDK/HIRN RRID: SCR_014393; UC24 DK104162, the Culshaw Junior Investigator Award in Diabetes, the JDRF, the Children's Diabetes Foundation and the Gate's Grubstake Fund. A.H.S. was supported by pre-doctoral training grants in Stem Cell Biology-5T32AR007411-35 and Interdisciplinary Bioengineering Research Training in Diabetes--T32 DK 120520. pmc1. Introduction The pancreas consists of two main compartments, the exocrine and endocrine tissue, both with distinct functions. Exocrine tissue consists predominantly of acinar cells that release digestive enzymes into the duodenum via a ductal system, making up most of the cell mass found in the organ. The pancreas also contains endocrine cells that are organized together into highly vascularized cell clusters called the islets of Langerhans. Endocrine cells within islets secret hormones that exquisitely regulate and maintain blood sugar levels within a tight physiological range. Representing only about ~1-2% of the organ tissue, the main endocrine cells are insulin-producing b-cells, glucagon-producing a-cells, somatostatin-producing d-cells, polypeptide-producing PP cells, and ghrelin-producing e-cells . Out of all endocrine cells, only b-cells express and secrete insulin in response to elevations in blood glucose levels. b-cell dysfunction or loss is key in contributing toward the development of diabetes, and much research has focused on this fascinating cell type. Diabetes presents as two major subtypes. In both, the inadequate release of insulin results in hyperglycemia that can be life-threatening. The most common diabetes form, type 2 diabetes (T2D), affecting 462 million people globally, is characterized by the insulin resistance of peripheral tissues and (subsequent) b-cell dysfunction, exhaustion, and loss . In type 1 diabetes (T1D), the patient's own insulin-producing b-cells are specifically destroyed through an autoimmune-mediated attack predominantly of T-cells, resulting in insulin deficiency. Type 1 diabetes (T1D) is a chronic condition that affects 1 in 500 Americans by the age of 15 . Current treatment for both T1D and late-stage T2D consists of injecting endogenous insulin. Exogenous insulin replacement therapy falls short of recapitulating the exact physiological function of a b-cell, and patients are susceptible to acute and long-term complications . Hypoglycemic conditions, induced by injecting too much insulin, can result in a life-threatening coma and are a constant risk for patients living with T1D and a practical cure that would alleviate the risks and concerns of current insulin therapy is desperately needed. Therefore, research efforts have focused on promoting cell replacement therapy approaches, such as b-cell proliferation and/or neogenesis and islet transplantation, to identify a practical cure for T1D patients. In this review, we will discuss aspects of cell replacement therapy with a focus on current and potential underappreciated challenges associated with stem-cell-derived b-cell transplantation. 2. Current and Potential Cell Replacement Strategies for T1D Patients Islet Transplantation to Restore b-Cell Mass A proof of principal for a potential practical cure has been shown with the establishment of the Edmond protocol in 2000. In this protocol, isolated allogenic cadaveric islets are infused in the portal vein of long-standing T1D patients that receive non-steroid immunosuppression . Importantly, islet recipients achieve on average ~35 months of insulin independence . Subsequently, islet transplantation was often performed in conjunction with kidney transplantation . However, there are several challenges associated with this procedure that prevent it from becoming widely accessible for patients. A major drawback with islet transplantation is the limited availability of high-purity isolated human cadaveric donor islet material. This is required to restore euglycemia in patients. Typically, each patient receives 10,000 islet equivalents (IEQs) per kilogram of body weight, an amount that usually needs to be extracted from two donor pancreases. In addition, initial clinical trials showed some patients requiring multiple islet infusions throughout the study, further highlighting the need for an abundant source of functional insulin-producing cells. The chronic immune suppression of patients, especially in children and adolescents, is problematic due to long-term complications, including severe and chronic infections and malignancy. In addition, studies have shown that immune suppressive agents impair b-cell function and survival using animal models . Indeed, functional cadaveric islet grafts are frequently lost within 2-5 years due to recurring autoimmunity, side effects of immunosuppressants, and other unknown mechanisms . The lack of sufficient donor islets has prompted the search for alternative and abundant sources of functional b(-like) cells for replacement therapy purposes, and much progress has been made using different approaches. Since porcine insulin has been shown to be physiologically well-matched to humans, the xenotransplantation of porcine islets has been considered an effective strategy to provide adequate amounts of islet material to treat T1D patients. However, immunological responses, such as instant blood-mediated inflammatory reaction (IBMIR) , hyperacute, and cellular rejection, remain major hurdles to overcome and improve porcine islet survival . Therefore, several strategies have been explored to overcome immune complications in this setting. The development of genetically modified pigs lacking the expression of certain surface proteins that play key roles in immune rejection upon porcine islet transplantation demonstrates promising results in improving porcine islet survival through combating IBMIR and hyperacute rejection. Preclinical studies also revealed that the blockade of co-stimulatory cell surface molecules suppresses T cell activation, hampers cellular rejection, and improves islet survival in vivo. With no evidence of porcine endogenous retrovirus (PERV) transmission, clinical studies of porcine islet xenotransplantation in T1D patients showed initial successes; however, most recipients failed to maintain long-term normal glycemic levels. Encapsulating pig islets has been suggested as effective in reducing xenogeneic immune rejection and prolonging graft survival and was tested in two clinical studies in T1D patients but showed only minimal reduction in their daily insulin needs . 3. Alternative Approaches to Increase Functional b-Cell Mass Other strategies have aimed at inducing b-cell replication or neogenesis via the transdifferentiation of pancreatic non-b cells to replenish the b-cell mass. Recently, DYRK1A inhibition in conjugation with other pathway manipulations has been shown to be effective in inducing increased b-cell proliferation . If safe b-cell-specific delivery modalities can be identified, inducing the proliferation of remaining b-cells in diabetic patients might provide a viable therapeutic strategy . Transdifferentiation refers to the change in functional cell phenotype of a differentiated cell into another rather than the differentiation of a less specialized stem cell into a functional cell type. Pancreatic duct ligation (PDL) has been shown to promote b-cell transdifferentiation from ductal, acinar, and alpha cells in mice , although some observations could not be repeated in another study . Similarly, the overexpression of MafA, Pdx1, and Ngn3 can trigger b-cell transdifferentiation predominantly from mouse acinar cells in vivo . Recently, a group induced alpha cell transdifferentiation into b-cells in vivo by infusing adeno-associated viruses carrying Pdx1 and MafA into the pancreatic duct of NOD mice . However, such experimental strategies, while carrying the potential for endogenous b-cell repopulation in T1D patients, are awaiting translation to human systems and/or clinical settings. 4. Stem-Cell-Derived b-Cells as an Abundant Cell Source One attractive approach that has advanced rapidly and shows tremendous potential as an abundant source of functional insulin-producing cells for clinical use is the direct differentiation of human pluripotent stem cells (hPSCs) into stem cell-derived b-like cells (sBCs) . Mouse development studies have identified critical transcription factors and signaling events during pancreas organogenesis, and subsequent work defined the necessary culture conditions to mimic key development stages to direct the differentiation of sBC from pluripotent stem cells . Specifically, several groups focused their efforts on utilizing recombinant proteins and small molecules to generate subsequent cell types resulting in pancreatic cells: definitive endoderm generation , posterior gut specification , formation of pancreatic bipotent progenitors , and endocrine differentiation . Although sBCs generated with early protocols were glucose-responsive, cells still displayed features of immature, fetal-like b-cells, and thus performed poorly in dynamic glucose-stimulated insulin secretion (dGSIS) perifusion assays. Several methods, such as the manipulation of key signaling pathways , media composition and in vitro culture extension , and the use of surface markers and fluorescence-activated cell sorting (FACS) to enrich reaggregated sBCs resulted in a more mature b-cell phenotype that closely resembles primary adult islets . Clear criteria defining a mature, functional b-cell that allows distinction from b-cell surrogates has recently been discussed in detail elsewhere . Interestingly, sBC maturation also seems to be accomplished upon transplantation, which in return restored euglycemia and reversed diabetes in preclinical mouse models . However, the early events taking place during the immediate engraftment of sBC have not been studied in detail. A considerable body of work has shown that the majority of functional b-cell mass is lost from human islets upon engraftment, suggesting that such drastic effects may also apply to sBCs due to unknown underlying cellular and molecular mechanisms . Potential mechanisms that might occur are: (i) cell death, (ii) dedifferentiation, and (iii) transdifferentiation. Recent work identified distinct human b-cell subpopulations in sBC and human islets in vitro and provided the possibility of (iv) b-subtype interconversion upon engraftment as an additional mechanism. Hence, there is a critical knowledge gap in the research field concerning the fate of sBC upon engraftment. Expanding our knowledge of the contributing mechanisms would expedite our progress in promoting the current approaches of delivering sBC as an effective cell replacement therapy for T1D patients. In addition, sBC cell therapy might represent an attractive treatment modality for T2D patients due to the absence of reoccurring autoimmunity if allogeneic rejection can be avoided in a localized manner. In this review, we will discuss findings and potential mechanisms driving b-cell loss upon engraftment , its implications for cell replacement therapy, and strategies to improve our understanding of the events affecting human b-cells upon transplantation. 5. Engrafted b-Cell Loss via Cell Death Classically, the main mechanism associated with islet cell death in vitro and in vivo is apoptosis. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) has shown that a considerable proportion of isolated human islets harvested from donor pancreases are lost in vitro within 5 days of culture due to apoptosis . The high propensity of b-cells to undergo apoptosis within islet preparations might be due to the increased metabolic rate that is not met under cell culture conditions. Later studies determined the caspase cascade as the major intrinsic mediator of apoptosis in cultured islets after being exposed to toxic levels of glucose . Glucotoxicity downregulated BCL-2 (anti-apoptotic protein) expression in isolated islets, which acts as an intrinsic signal to activate caspase-associated apoptosis . Similarly, after 24 h of transplantation, TUNEL staining showed that approximately a quarter of all b-cells in human islets engrafted in the kidney capsule of immunodeficient mice are lost due to apoptosis . Overall, although the exact mechanisms contributing to primary b-cell death upon transplantation are poorly understood, the current literature predominantly attributes the observed loss to ischemia and nutrient deprivation. Mediated by endothelial cells, in situ, pancreatic islets are highly vascularized and are under a continuous supply of oxygen and nutrients, ensuring optimal function. However, this supply is lost during the islet isolation process, which involves the use of digestive enzymes and mechanical force to separate the islets from the native organ . Due to loss of blood flow and imperfect culturing conditions, endothelial cells, which are critical in providing cellular matrix proteins that fine-tune the function of b-cells, eventually die in vitro . Since blood flow is abolished after isolation, islets are under an acute nutrient deficiency and exposed to oxidative stress mediated by hypoxia. Isolated islets depend on passive nutrition diffusion to satisfy the activities of the highly metabolic b-cells. Thus, culture conditions are insufficient in supplying uniform O2 levels to all b-cells, especially cells located at the core of the islet, negatively effecting b-cell survival in vitro, with necrotic cores present, especially in larger islets due to low oxygen accessibility . Similarly transplanted human and rodent islets have been shown to have reduced graft oxygen tension in the initial stages of engraftment and to suffer a drastic loss of b-cells in vivo . Therefore, several in vitro pre-transplant priming methods have been adopted to improve islet survival in transplants, such as oxygenation treatment, culture in hyperoxic conditions, and modulation of seeding density; however, these strategies have failed to be exceedingly successful . Further mechanistic analysis revealed several signaling pathways, such as anaerobic glycolysis and hypoxia-inducible factor (HIF)-related pathways, to be associated with b-cell survival under hypoxic conditions; however, further investigation is required . Finally, other necrotic-regulated mechanisms such as pyroptosis , ferroptosis , and necroptosis have been implicated to contribute toward b-cell death during islet isolation, culture, and transplantation; however, these mechanisms have not been comprehensively elucidated as of yet. Most in vivo sBC studies have focused on the metabolic action and long-term therapeutic capacity of engrafted, surviving cells using preclinical animals starting at 3-4 weeks post-transplantation when grafts are fully vascularized. However, most studies have largely neglected the early phase of sBC transplantation. In a recent study, sBCs constitutively expressing luciferase were transplanted subcutaneously or under the kidney capsule of immunodeficient mice, and total graft mass was quantified using bioluminescence . As expected, on the day of transplantation, robust expression of luciferase was detected; however, 7 days post-transplant, this expression was significantly reduced in both sites, indicating substantial graft loss. In addition, the hPSC cell line employed also contains a GFP reporter driven by the insulin promoter, allowing quantification of sBCs before and after transplant. Flow cytometry analysis revealed that approximately 70% of sBCs were lost, while the total graft was only 50% reduced within the first 7 days of engraftment, indicating a preferential loss of sBCs compared to other cells present. The main drivers of graft loss are considered to be ischemia-induced hypoxia and nutrition deprivation. Amino acid supplementation and adjusting the physiological oxygen levels to 5% in culture improved sBC graft survival significantly. In situ, pancreatic islets are abundantly vascularized with a continuous supply of oxygen and nutrients; therefore, this study further highlights the importance of vascularization to sBC survival and function in vivo. Pepper and colleagues showed the pre-vascularization of the subcutaneous site followed by the transplantation of pancreatic endoderm (PEC) cells improved stem cell-derived b-cell functionality and survival in vivo, providing further evidence for the notion that appropriate vascularization is critical for b-cell survival and function . Several groups focused on engrafting sBCs that incorporate endothelial cells alone or in combination with mesenchymal cells . In a recent elegant study, micro-vessels isolated from adipose tissue have been shown to improve and accelerate the vascularization of sBC grafts, as well as their survival and function in vivo . Finally, using oxygen-generating biomaterials shows promising results to improve islet survival in vivo that could be applied to future sBC engraftments . Altogether, the literature has provided data suggesting hypoxia and nutrient deprivation as two key contributors to sBC graft decline that can be mitigated by providing better engraftment solutions. Understanding what distinguishes sBCs that survive the first week of engraftment from sBCs that are lost during this period could provide additional means to preserve total functional graft mass. 6. De-Differentiation upon Transplantation: Do b-Cells Revert to Progenitor/Precursor Cells? Cell death is the most explored mechanism contributing to immediate b-cell loss upon transplant in vivo, while other potential means resulting in the observed b-cell loss have not been thoroughly investigated. One such mechanism could be b-cell dedifferentiation. Dedifferentiation is loosely defined as the loss of a mature, functional b-cell phenotype and/or the acquisition of progenitor/precursor traits . Dedifferentiation of mouse b-cells has been shown by: upregulation in the expression of progenitor genes (e.g., Foxo1, Neurog3, and Aldh1a3), enrichment of disallowed b-cell makers (e.g., Hk2, Ldha, and Mct1) , mis-localization, loss or reduced expression of key b-cell transcription factors (e.g., Nkx6.1, MafA, and Pdx1), and altered expression of metabolic genes (e.g., Glut2 and Gck) . In vitro and in vivo models of type 2 diabetes (T2D) have suggested oxidative stress, ER stress, and nutritional stress as main contributors to b-cell dedifferentiation. The loss of FOXO1 expression has been suggested as a key trigger of b-cell dedifferentiation. Mice lacking Foxo1 developed hyperglycemia and b-cell dysfunction . A lineage-tracing analysis revealed that Foxo1-deficient b-cells dedifferentiated into a progenitor cell population expressing NEUROG3 (a key early endocrine progenitor maker), as well as the early developmental markers OCT4, NANOG, and L-MYC . Another recent study revealed a decline of FOXO1 expression as well as key b-cell markers NKX6.1 and MAFA in db/db mouse islets and human islets isolated from T2D patients compared to controls . Additional analysis revealed an upregulation of the progenitor marker ALDH1A3 , specifically in b-cells of T2D patients, providing evidence for dedifferentiation in humans similar to mice . These results were further supported by earlier animal studies using Foxo1-/- or db/db deficient mice, in which mouse b-cells similarly showed an upregulation in ALDH1A3 and NEUROG3 with concomitant downregulation of the expression of b-cell markers MAFA and NKX6.1 . SOX9, a transcription factor critical for pancreas development, has also been suggested as a novel regulator of b-cell dedifferentiation into a developmentally earlier, pancreatic progenitor-like cell type. The von Hippel-Lindau/hypoxia-inducible factor (VHL/HIF) has been implicated previously to regulate cellular responses to hypoxia . Hebrok and colleagues have shown that the deletion of Vhlh in mice resulted in glucose intolerance, reduced b-cell mass, and decreased the expression of key b-cell markers (MafA, Pdx1, and insulin) . Further protein analysis revealed the loss of Vhlh in mice triggered a progenitor program that resulted in the emergence of SOX9-expressing cells. Using mouse and rat insulinoma cell lines, hypoxic oxygen levels triggered similar dedifferentiation programs in cultured b-cells by perturbing the expression of key b-cell genes while increasing Sox9 expression . Investigating b-cell dedifferentiation in the human setting has been more challenging. A series of early reports suggested that b-cells dedifferentiate upon culturing in vitro, resulting in the derivation of a proliferative cell population but also a significant reduction in insulin expression . However, the dedifferentiation hypothesis was contested by others as simply being the result of b-cell death and the expansion of pre-existing mesenchymal cells within cell preparations. Using genetic lineage-tracing analysis on isolated primary human islets employing Cre/lox technology that was previously restricted to transgenic animals provided direct evidence for b-cell dedifferentiation as initially postulated . These results were subsequently confirmed by another group . Subsequent work revealed a critical role in the activation of key developmental pathways during the dedifferentiation process, suggesting potential leverage points to prevent b-cell dedifferentiation . Indeed, the inhibition of signaling pathways induces redifferentiation of expanded, dedifferentiated human b-cells marked by increased insulin and key b-cell marker expression . Similar experiments using lineage-traced mouse islets demonstrated mouse b-cell dedifferentiation in vitro but a lack of proliferation, unlike human b-cells, providing an example of distinct species differences . Although it is challenging to prove the occurrence of dedifferentiation of human b-cells in situ, recent studies implicated dedifferentiation events. The increased frequency of a chromogranin A-positive hormone-negative (CPHN) population in T1D and T2D isolated human islets sections and decreased b-cell levels have been reported and could represent dedifferentiation, although the authors suggested emerging regeneration as the sources of CPHN cells . Similarly, higher levels of non-endocrine cells compared to endocrine-expressing cells were detected in sections of human islets transplanted into T1D patients . Taken together, these studies provide strong support for the occurrence of b-cell dedifferentiation in mice and humans and point toward its potential implication in sBC transplantation settings. However, additional studies are needed to provide rigorous evidence for the potential dedifferentiation of human b-cells in different transplantation settings. Taking advantage of genetic lineage tracing or barcoding strategies would likely provide key advances to fill current knowledge gaps. 7. Transdifferentiation upon Transplantation: Do b-Cells Convert into Other Hormone-Expressing Cells? Another potential mode of b-cell loss upon transplantation could be transdifferentiation. Transdifferentiation is referred to as the loss of the b-cell phenotype and acquisition of features of other endocrine hormone-expressing cells. Transdifferentiation can occur via two different mechanisms: a direct shift into displaying characteristics of other hormone-expressing endocrine cells in addition to the loss of b-cell features, or indirectly via dedifferentiating into a precursor/progenitor stage first, followed by the acquisition of other endocrine cell features. All endocrine cells arise from early pancreatic cells marked by PDX1 expression during development . Pancreatic lineages become further specified by specific transcription factors, some of which exhibit antagonistic actions at key lineage decisions by inhibiting each other. PTF1a and NKX6.1 are critical in specifying pancreas progenitors further. PTF1a expression gives rise to exocrine tissue . In contrast, NKX6.1 expression segregates the ductal and endocrine lineages from the exocrine acinar cells during development . Notch induction of bipotent progenitors marked by PDX1 and NKX6.1 expression triggers endocrine differentiation, giving rise to a transient expression NEUROG3 . This is followed by endocrine cell specification, which is also governed by key transcription factors: PAX4 expression gives rise to b-and d-cells , while Arx4 is necessary for a-cell development . Using knock-out approaches, these studies showed that ARX and PAX4 act antagonistically to give rise to their respective lineage endocrine hormone cell types. These differentiated endocrine hormone cells have been viewed as terminally differentiated cells that acquire their specialized function and lose the ability to proliferate and differentiate into other cell types. However, several mouse studies have shown that endocrine cells can transdifferentiate into other cell types under forced genetic and diabetic stress conditions. As a proof of principal, b-cell transdifferentiation into a-cells has been initially explored via altering key transcription factors using mouse models. Overexpression of ARX in all endocrine cells showed a drastic reduction in the levels of d-cells and increased PP cells . Similarly, the overexpression of ARX in all b-cells resulted in transdifferentiation into PP cells. Furthermore, the loss of key b-cell transcription factors also resulted in a-cell transdifferentiation. Supported by lineage-tracing analysis, endocrine precursor and b-cell-specific KO mouse models of Nkx6.1 revealed a significant reduction in the b-cells and a significant increase in all other hormone-expressing cells (a, d, PP, e) . Further studies revealed enrichment in Neurog3 expression in adult islets of cell-type-specific KO Nkx6.1 mice, suggesting that b-cells are potentially dedifferentiating into a precursor cell type before acquiring a non-b-endocrine cell phenotype . Similarly, the deletion of the tinman domain of Nkx2.2 in conjugation with lineage analysis resulted in a-cell transdifferentiation and eventually hyperglycemia in young adult mice (3.5 weeks and older) . Importantly, in the described studies, no significant changes in the total endocrine cell mass or any polyhormonal cells were found, suggesting two things: the transdifferentiation events included a loss of b-cell phenotype with some aspects of dedifferentiation first, followed by acquiring an a-cell type phenotype thereafter. Other than pancreatic developmental transcription factors, a recent study ablated Xbp1 in adult mouse islets, a major regulator of the unfolded protein response (UPR) and b-cell function, which resulted in a-cell transdifferentiation and subsequently in hyperglycemia and diabetes. Interestingly, these studies observed an increase in the expression of the progenitor marker Sox9, suggesting a dedifferentiating phase before the a-cell transition . In sum, mouse studies revealed conditions of b-cell transdifferentiation; however, findings were based on hormone expression analysis but mostly lacked comprehensive functional assays. Only limited observations have been reported documenting human b-cell transdifferentiation in vitro and in vivo. Supported by lineage tracing analysis, the reaggregation of islet cells in vitro resulted in the transition of b-cells into a-cells . Moreover, there are several lines of evidence suggesting the occurrence of transdifferentiation of human diabetic islets. IHC analysis of T2D human isolated islets showed higher levels of bi-hormonal cells expressing glucagon and insulin . This study also described a cell population that expresses the a-cell hormone glucagon and the b-cell marker NKX6.1 but not insulin and could represent a midway transdifferentiation phenotype between a-cells. Another study showed higher levels of polyhormonal cells in lean-isolated T2D primary islets, further supporting the occurrence of transdifferentiation . However, these studies were performed on isolated human islets in vitro, and therefore one potential explanation for the emergence of bi/polyhormonal cells could be attributed to the isolation process and poor culturing conditions. Moreover, currently, it is unknown if transdifferentiation occurs in T1D islets due to the rarity of such samples and the difficulty in capturing islets at different stages of the disease in situ. Currently, the field lacks a comprehensive composition analysis of transplanted pancreatic islets in both human and rodent models. In the setting of sBC grafts, scRNA-seq, protein, and functional analyses showed an improvement in sBC maturation and functionality upon engraftment . However, it appears that the graft displays more a-cells than b-cells compared to sBCs in vitro that possess more insulin-expressing cells . Interestingly, polyhormonal cells generated as an unwanted byproduct during the differentiation of sBCs seem to resolve in vivo into a-like cells . Polyhormonal cells lack NKX6.1 expression; thus, the transition of these cells into a-like cells upon engraftment might resemble aspects of the transdifferentiation observed in the Nkx6.1 KO mice . In addition, enterochromaffin cells have recently been identified as an unwanted off-target differentiation product . While examples exist that convincingly demonstrate transdifferentiation in animal models, the available human data are limited. To corroborate transdifferentiation phenomena in sBC transplantation settings will require careful additional experimentation. Understanding the underlying molecular mechanisms of transdifferention, as well as dedifferentiation, might allow formulating strategies to preserve a pristine b-cell phenotype upon transplantation. 8. b-Cell Subtype Interconversion upon Transplant While in the past b-cells have been commonly viewed as a rather homogenous population, functionally different b-cells were already described decades ago, and the concept of b-cell heterogeneity and subpopulations has recently received increased attention . Indeed, b-cells can be subdivided into distinct subpopulations both in mice and humans. Several molecular markers label mature/immature b-cell subpopulations in mouse islets, such as E-cadherin, FLTP and UCN3 . Recent advances in scRNAseq revealed 3-5 distinct b-cell subpopulations based on differential mRNA transcription in cadaveric human islets . Dr. Grompe and associates identified four b-cell subtypes (b1, 2, 3, and 4) in adult human pancreas marked by the surface protein markers CD9 and ST8SIA1 . A sorting strategy using antibodies against CD9 and ST8SIA1 allowed differential gene expression analysis that revealed distinct transcriptional gene profiles, many associated with b-cell functionality. Glucose-stimulated insulin secretion (GSIS) experiments showed that b1 cells, the most abundant subtype in healthy individuals, are also the most functional subpopulation. Interestingly, the distribution of b-cell subpopulations in cadaveric T2D human islets is skewed towards less functional subtypes. Altogether, this study provides a thorough characterization of human b-subtypes in both healthy and diseased islets; however, the distribution of these subtypes in transplant settings has yet to be discovered. Aside from molecular markers, b-cell subtypes can be categorized based on insulin secretory profiles , functional properties (pacemaker cells such as Hub cells and first responders) , and other phenotypes. Supplemented with human islet data, animal studies have identified a subpopulation of b-cells that acquire a senescence and senescence-like secretory phenotype (SASP) . Senescence is a state in which cells cease to divide but remain metabolically active with an altered phenotype. Some but not all senescent b-cells exhibit SASP by secreting a mixture of chemokines, cytokines, and ECM molecules, among others. The activation of the DNA damage response (DDR) due to cellular stress has been demonstrated to give rise to senescence and SASP in different cell systems . Thompson et al. showed SASP-like b-cells, both mouse and human, exist in higher numbers in T1D compared to healthy pancreatic islets. Other than being growth-arrested, SASP cells exhibit non-cell-autonomous activities by secreting factors that affect the viability and function of neighboring (b-) cells and trigger the chemotaxis of immune cells, which leads to the progression of diabetes in T1D mouse models. Strikingly, using senolytic drugs, the clearance of SASP-like b-cells prevents diabetes in a T1D mouse model and preserves b-cell mass and functionality. Interestingly, a similar senescence/SASP b-subpopulation was also identified and characterized in both human and mouse T2D islets. Similarly, clearance of senescent cells in T2D via senolysis improved glucose levels and b-cell function and identity . Altogether, these findings highlight the consequence of SASP cells on disease development and progression, and it will be vital to determine if senescence and SASP occur in sBCs in vitro and in vivo. Heterogeneity in sBC clusters is less investigated, but multiple scRNA data sets have been published showing distinct subpopulations. Veres et al. performed the first comprehensive scRNA-seq on stem-cell-derived islets at various stages of the differentiation protocol and revealed an endocrine hormone-expressing cell population consisting of b-, polyhormonal cells, an endocrine non-hormone+ population, and a previously unreported enterochromaffin population . However, no detailed analysis focused on the sBC subpopulation was performed. These results were subsequently supported by two other groups identifying similar populations with different cell distributions . Recently, we adopted a more specific approach by performing scRNA-seq analysis on sorted sBCs and identified 7 b-cell subtypes based on an unsupervised cluster analysis. These data revealed a mature b-cell subpopulation marked by the expression of surface ENTPD3 (also known as NTPDase3) protein. The sorting of ENTPD3+ sBC followed by reaggregation and functional evaluation using dynamic GSIS revealed secretion patterns similar to primary human islets, while did not, providing the first evidence for defined heterogeneity within sBCs . Interestingly, these data also revealed other b-subtypes, such as polyhormonal sBC, a proliferative cell sBC population, IGF2+ sBCs, and a distinct CD9-labeled sBC population, suggesting that experiments to further stratify sBCs in vitro and in vivo are warranted. Such studies will cumulatively contribute toward optimizing sBC grafts' functionality in vivo, further promoting cell therapy as the most practical treatment for diabetic patients. 9. Conclusions The lack of donor islets has prompted us and others to generate sBCs as an abundant source of human functional human b-cells. Preclinical studies demonstrated the ability of sBCs to function in vivo and restore euglycemia in diabetic mouse models. However, in vivo studies have also demonstrated that the majority of sBCs, similarly to cadaveric human pancreatic islets, are lost upon transplantation, but the underlying mechanisms remain largely unknown. Currently, there are multiple ongoing clinical trials in T1D patients with some promising results, further promoting stem cell-derived pancreatic progenitors and sBCs as an attractive therapeutic approach for diabetes treatment. Current literature demonstrates that considerable sBC loss occurs upon transplant into preclinical animal models, suggesting this likely happens in humans as well. Understanding the underlying mechanisms resulting in b-cell loss will allow the formulation of strategies to counteract these undesired effects, thereby providing improvements to current cell replacement efforts. Ultimately, improving b-cell survival and function will reduce the graft size needed to restore euglycemia in patients and thus also reduce the cost of this therapy. Currently, b-cell death is largely perceived as the main contributor to b-cell graft mass loss upon transplant. However, we argue that this basic view might be an oversimplified representation of what actually might occur to b-cells upon transplant. Indeed, recent work has highlighted underappreciated aspects of b-cell biology, most notably the plasticity displayed by b-cells under certain experimental and disease conditions, as well as b-cell heterogeneity. We hypothesize here that b-cell mass loss upon engraftment likely constitutes a melange of different phenomena occurring in parallel. It will require focused experimental approaches that allow accurate cell fate tracking in vivo to determine if other mechanisms in addition to b-cell death contribute significantly to graft loss. The notion that sBCs could transdifferentiate into other hormone-expressing cells might be exploited for improved b-cell function. Other islet cell types have been shown to be critical for b-cell function; thus, generating stem-cell-derived islet-like structures upon transplant might be beneficial. However, sBCs dedifferentiate into a less differentiated, potentially proliferative state, which could require additional (long-term) safety considerations. Obtaining detailed molecular and cellular insights into potential mechanisms will allow the development of preventative strategies. We envision such efforts to be similar to current research efforts aimed at reducing the well-recognized problem posed by ischemia-induced graft loss. Indeed, many research groups use innovative strategies, most notably in the bioengineering space, to provide adequate oxygen and nutrient supply to islets and/or b-cells immediately after transplant to counteract ischemia-induced detrimental effects. Another important question relates to the presence and/ or formation of distinct b-cell subpopulations upon transplant. Indeed, we and others have shown that b-cell subpopulations exhibit differential functions. Obviously, providing the most functional sBC to patients in the cell replacement setting is highly desired over subpar phenotypes. However, it is tantalizing to speculate how different b-cell subpopulations might interact with the immune system and if an immune-privileged population can be identified and exploited for replacement therapy. Reduced immunogenicity of a small b-cell subpopulation has indeed been proposed in autoimmune animal models previously. Furthering our understanding of human b-cell biology in the setting of sBC transplantation might also provide interesting insights that could apply to autoimmune mechanisms in patients and provide novel model systems to investigate them. Overall, understanding the exact mechanisms contributing to sBC loss upon transplant will allow the development of informed strategies to prevent sBC loss and might allow the delivery of greater numbers of sBC with superior function. Such second-generation cell replacement therapy approaches would effectively provide distinct advantages, such as the need for reduced graft mass to be transplanted. Overall, advances in stem cell-based cell replacement therapy for diabetic patients have been impressive within the last few years, and we anticipate that progress will further accelerate to allow this curative approach to reach a wide patient population in the near future. Acknowledgments We would like to thank Jessie Barra and Roberto Castro-Gutierrez and other members of the Russ lab for stimulating discussion and comments. Author Contributions Conceptualization, A.H.S. and H.A.R.; writing--original draft preparation, A.H.S. and H.A.R.; writing--review and editing, A.H.S. and H.A.R. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. H.A.R. was a SAB member at Sigilon Therapeutics and Prellis Biologics and was a consultant for Eli Lilly and Minutia. Figure 1 Reported and potential molecular and cellular mechanisms driving human pancreatic b-cell loss upon transplantation. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Ionescu-Tirgoviste C. Gagniuc P.A. Gubceac E. Mardare L. Popescu I. Dima S. Militaru M. A 3D map of the islet routes throughout the healthy human pancreas Sci. Rep. 2015 5 14634 10.1038/srep14634 26417671 2. Khan M.A.B. Hashim M.J. King J.K. Govender R.D. Mustafa H. Al Kaabi J. Epidemiology of Type 2 Diabetes--Global Burden of Disease and Forecasted Trends J. Epidemiol. Glob. Health 2020 10 107 111 10.2991/jegh.k.191028.001 32175717 3. Patterson C. Guariguata L. Dahlquist G. Soltesz G. Ogle G. Silink M. Diabetes in the young--A global view and worldwide estimates of numbers of children with type 1 diabetes Diabetes Res. Clin. Pract. 2014 103 161 175 10.1016/j.diabres.2013.11.005 24331235 4. Patrick A.W. Williams G. Adverse effects of exogenous insulin. Clinical features, management and prevention Drug Saf. 1993 8 427 444 10.2165/00002018-199308060-00004 8329148 5. Carlson M.G. Campbell P.J. Intensive insulin therapy and weight gain in IDDM Diabetes 1993 42 1700 1707 10.2337/diab.42.12.1700 8243815 6. Sjolie A.K. Ocular complications in insulin treated diabetes mellitus. An epidemiological study Acta Ophthalmol. Suppl. 1985 172 1 77 2988264 7. Ryan E.A. Lakey J.R. Rajotte R.V. Korbutt G.S. Kin T. Imes S. Rabinovitch A. Elliott J.F. Bigam D. Kneteman N.M. Clinical outcomes and insulin secretion after islet transplantation with the Edmonton protocol Diabetes 2001 50 710 719 10.2337/diabetes.50.4.710 11289033 8. Shapiro A.M. Lakey J.R. Ryan E.A. Korbutt G.S. Toth E. Warnock G.L. Kneteman N.M. Rajotte R.V. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen N. Engl. J. Med. 2000 343 230 238 10.1056/NEJM200007273430401 10911004 9. Shapiro A.M. Ricordi C. Hering B.J. Auchincloss H. Lindblad R. Robertson R.P. Secchi A. Brendel M.D. Berney T. Brennan D.C. International trial of the Edmonton protocol for islet transplantation N. Engl. J. Med. 2006 355 1318 1330 10.1056/NEJMoa061267 17005949 10. Markmann J.F. Rickels M.R. Eggerman T.L. Bridges N.D. Lafontant D.E. Qidwai J. Foster E. Clarke W.R. Kamoun M. Alejandro R. Phase 3 trial of human islet-after-kidney transplantation in type 1 diabetes Am. J. Transplant. 2021 21 1477 1492 10.1111/ajt.16174 32627352 11. Niclauss N. Bosco D. Morel P. Demuylder-Mischler S. Brault C. Milliat-Guittard L. Colin C. Parnaud G. Muller Y.D. Giovannoni L. Influence of donor age on islet isolation and transplantation outcome Transplantation 2011 91 360 366 10.1097/tp.0b013e31820385e6 21344706 12. Rangel E.B. Tacrolimus in pancreas transplant: A focus on toxicity, diabetogenic effect and drug-drug interactions Expert Opin. Drug Metab. Toxicol. 2014 10 1585 1605 10.1517/17425255.2014.964205 25255841 13. Barton F.B. Rickels M.R. Alejandro R. Hering B.J. Wease S. Naziruddin B. Oberholzer J. Odorico J.S. Garfinkel M.R. Levy M. Improvement in outcomes of clinical islet transplantation: 1999-2010 Diabetes Care 2012 35 1436 1445 10.2337/dc12-0063 22723582 14. Goto M. Tjernberg J. Dufrane D. Elgue G. Brandhorst D. Ekdahl K.N. Brandhorst H. Wennberg L. Kurokawa Y. Satomi S. Dissecting the instant blood-mediated inflammatory reaction in islet xenotransplantation Xenotransplantation 2008 15 225 234 10.1111/j.1399-3089.2008.00482.x 18957045 15. Johansson H. Lukinius A. Moberg L. Lundgren T. Berne C. Foss A. Felldin M. Kallen R. Salmela K. Tibell A. Tissue factor produced by the endocrine cells of the islets of Langerhans is associated with a negative outcome of clinical islet transplantation Diabetes 2005 54 1755 1762 10.2337/diabetes.54.6.1755 15919797 16. Nilsson B. The instant blood-mediated inflammatory reaction in xenogeneic islet transplantation Xenotransplantation 2008 15 96 98 10.1111/j.1399-3089.2008.00460.x 18447877 17. Gill R.G. Wolf L. Daniel D. Coulombe M. CD4+ T cells are both necessary and sufficient for islet xenograft rejection Transplant. Proc. 1994 26 1203 7913259 18. Olack B.J. Jaramillo A. Benshoff N.D. Kaleem Z. Swanson C.J. Lowell J.A. Mohanakumar T. Rejection of porcine islet xenografts mediated by CD4+ T cells activated through the indirect antigen recognition pathway Xenotransplantation 2002 9 393 401 10.1034/j.1399-3089.2002.01070.x 12371935 19. Koulmanda M. Laufer T.M. Auchincloss H. Smith R.N. Prolonged survival of fetal pig islet xenografts in mice lacking the capacity for an indirect response Xenotransplantation 2004 11 525 530 10.1111/j.1399-3089.2004.00174.x 15479462 20. Kirchhof N. Shibata S. Wijkstrom M. Kulick D.M. Salerno C.T. Clemmings S.M. Heremans Y. Galili U. Sutherland D.E. Dalmasso A.P. Reversal of diabetes in non-immunosuppressed rhesus macaques by intraportal porcine islet xenografts precedes acute cellular rejection Xenotransplantation 2004 11 396 407 10.1111/j.1399-3089.2004.00157.x 15303976 21. Lindeborg E. Kumagai-Braesch M. Moller E. Phenotypic and functional characterization of human T cell clones indirectly activated against adult pig islet cells Xenotransplantation 2006 13 41 52 10.1111/j.1399-3089.2005.00257.x 16497211 22. Aghazadeh Y. Nostro M.C. Cell Therapy for Type 1 Diabetes: Current and Future Strategies Curr. Diab. Rep. 2017 17 37 10.1007/s11892-017-0863-6 28432571 23. Sakata N. Sumi S. Yoshimatsu G. Goto M. Egawa S. Unno M. Encapsulated islets transplantation: Past, present and future World J. Gastrointest. Pathophysiol. 2012 3 19 26 10.4291/wjgp.v3.i1.19 22368783 24. Matsumoto S. Tan P. Baker J. Durbin K. Tomiya M. Azuma K. Doi M. Elliott R.B. Clinical porcine islet xenotransplantation under comprehensive regulation Transplant. Proc. 2014 46 1992 1995 10.1016/j.transproceed.2014.06.008 25131091 25. Matsumoto S. Abalovich A. Wechsler C. Wynyard S. Elliott R.B. Clinical Benefit of Islet Xenotransplantation for the Treatment of Type 1 Diabetes EBioMedicine 2016 12 255 262 10.1016/j.ebiom.2016.08.034 27592597 26. Morozov V.A. Wynyard S. Matsumoto S. Abalovich A. Denner J. Elliott R. No PERV transmission during a clinical trial of pig islet cell transplantation Virus Res. 2017 227 34 40 10.1016/j.virusres.2016.08.012 27677465 27. Rachdi L. Kariyawasam D. Aiello V. Herault Y. Janel N. Delabar J.M. Polak M. Scharfmann R. Dyrk1A induces pancreatic b cell mass expansion and improves glucose tolerance Cell Cycle 2014 13 2221 2229 10.4161/cc.29250 24870561 28. Wang P. Alvarez-Perez J.C. Felsenfeld D.P. Liu H. Sivendran S. Bender A. Kumar A. Sanchez R. Scott D.K. Garcia-Ocana A. A high-throughput chemical screen reveals that harmine-mediated inhibition of DYRK1A increases human pancreatic beta cell replication Nat. Med. 2015 21 383 388 10.1038/nm.3820 25751815 29. Dirice E. Walpita D. Vetere A. Meier B.C. Kahraman S. Hu J. Dancik V. Burns S.M. Gilbert T.J. Olson D.E. Inhibition of DYRK1A Stimulates Human b-Cell Proliferation Diabetes 2016 65 1660 1671 10.2337/db15-1127 26953159 30. Shen W. Taylor B. Jin Q. Nguyen-Tran V. Meeusen S. Zhang Y.Q. Kamireddy A. Swafford A. Powers A.F. Walker J. Inhibition of DYRK1A and GSK3B induces human b-cell proliferation Nat. Commun. 2015 6 8372 10.1038/ncomms9372 26496802 31. Wang P. Karakose E. Liu H. Swartz E. Ackeifi C. Zlatanic V. Wilson J. Gonzalez B.J. Bender A. Takane K.K. Combined Inhibition of DYRK1A, SMAD, and Trithorax Pathways Synergizes to Induce Robust Replication in Adult Human Beta Cells Cell Metab. 2019 29 638 652.e635 10.1016/j.cmet.2018.12.005 30581122 32. Ackeifi C. Wang P. Karakose E. Manning Fox J.E. Gonzalez B.J. Liu H. Wilson J. Swartz E. Berrouet C. Li Y. GLP-1 receptor agonists synergize with DYRK1A inhibitors to potentiate functional human b cell regeneration Sci. Transl. Med. 2020 12 eaaw9996 10.1126/scitranslmed.aaw9996 32051230 33. Wang P. Karakose E. Choleva L. Kumar K. DeVita R.J. Garcia-Ocana A. Stewart A.F. Human Beta Cell Regenerative Drug Therapy for Diabetes: Past Achievements and Future Challenges Front. Endocrinol. 2021 12 671946 10.3389/fendo.2021.671946 34. Chung C.H. Hao E. Piran R. Keinan E. Levine F. Pancreatic b-cell neogenesis by direct conversion from mature a-cells Stem Cells 2010 28 1630 1638 10.1002/stem.482 20653050 35. Wang R.N. Kloppel G. Bouwens L. islet-cell differentiation and islet growth in the pancreas of duct-ligated adult rats Diabetologia 1995 38 1405 1411 10.1007/BF00400600 8786013 36. Pan F.C. Bankaitis E.D. Boyer D. Xu X. Van de Casteele M. Magnuson M.A. Heimberg H. Wright C.V. Spatiotemporal patterns of multipotentiality in Ptf1a-expressing cells during pancreas organogenesis and injury-induced facultative restoration Development 2013 140 751 764 10.1242/dev.090159 23325761 37. Xu X. D'Hoker J. Stange G. Bonne S. De Leu N. Xiao X. Van de Casteele M. Mellitzer G. Ling Z. Pipeleers D. Beta cells can be generated from endogenous progenitors in injured adult mouse pancreas Cell 2008 132 197 207 10.1016/j.cell.2007.12.015 18243096 38. Cavelti-Weder C. Shtessel M. Reuss J.E. Jermendy A. Yamada T. Caballero F. Bonner-Weir S. Weir G.C. Pancreatic duct ligation after almost complete b-cell loss: Exocrine regeneration but no evidence of b-cell regeneration Endocrinology 2013 154 4493 4502 10.1210/en.2013-1463 24029238 39. Zhou Q. Brown J. Kanarek A. Rajagopal J. Melton D.A. In vivo reprogramming of adult pancreatic exocrine cells to beta-cells Nature 2008 455 627 632 10.1038/nature07314 18754011 40. Xiao X. Guo P. Shiota C. Zhang T. Coudriet G.M. Fischbach S. Prasadan K. Fusco J. Ramachandran S. Witkowski P. Endogenous Reprogramming of Alpha Cells into Beta Cells, Induced by Viral Gene Therapy, Reverses Autoimmune Diabetes Cell Stem Cell 2018 22 78 90.e4 10.1016/j.stem.2017.11.020 29304344 41. Pagliuca F.W. Millman J.R. Gurtler M. Segel M. Van Dervort A. Ryu J.H. Peterson Q.P. Greiner D. Melton D.A. Generation of functional human pancreatic b cells in vitro Cell 2014 159 428 439 10.1016/j.cell.2014.09.040 25303535 42. Rezania A. Bruin J.E. Arora P. Rubin A. Batushansky I. Asadi A. O'Dwyer S. Quiskamp N. Mojibian M. Albrecht T. Reversal of diabetes with insulin-producing cells derived in vitro from human pluripotent stem cells Nat. Biotechnol. 2014 32 1121 1133 10.1038/nbt.3033 25211370 43. Russ H.A. Parent A.V. Ringler J.J. Hennings T.G. Nair G.G. Shveygert M. Guo T. Puri S. Haataja L. Cirulli V. Controlled induction of human pancreatic progenitors produces functional beta-like cells in vitro EMBO J. 2015 34 1759 1772 10.15252/embj.201591058 25908839 44. Veres A. Faust A.L. Bushnell H.L. Engquist E.N. Kenty J.H. Harb G. Poh Y.C. Sintov E. Gurtler M. Pagliuca F.W. Charting cellular identity during human in vitro b-cell differentiation Nature 2019 569 368 373 10.1038/s41586-019-1168-5 31068696 45. Hebrok M. Hedgehog signaling in pancreas development Mech. Dev. 2003 120 45 57 10.1016/s0925-4773(02)00331-3 12490295 46. Murtaugh L.C. Melton D.A. Genes, signals, and lineages in pancreas development Annu. Rev. Cell Dev. Biol. 2003 19 71 89 10.1146/annurev.cellbio.19.111301.144752 14570564 47. Pan F.C. Wright C. Pancreas organogenesis: From bud to plexus to gland Dev. Dyn. 2011 240 530 565 10.1002/dvdy.22584 21337462 48. Seymour P.A. Sander M. Historical perspective: Beginnings of the beta-cell: Current perspectives in beta-cell development Diabetes 2011 60 364 376 10.2337/db10-1068 21270248 49. D'Amour K.A. Agulnick A.D. Eliazer S. Kelly O.G. Kroon E. Baetge E.E. Efficient differentiation of human embryonic stem cells to definitive endoderm Nat. Biotechnol. 2005 23 1534 1541 10.1038/nbt1163 16258519 50. D'Amour K.A. Bang A.G. Eliazer S. Kelly O.G. Agulnick A.D. Smart N.G. Moorman M.A. Kroon E. Carpenter M.K. Baetge E.E. Production of pancreatic hormone-expressing endocrine cells from human embryonic stem cells Nat. Biotechnol. 2006 24 1392 1401 10.1038/nbt1259 17053790 51. Kroon E. Martinson L.A. Kadoya K. Bang A.G. Kelly O.G. Eliazer S. Young H. Richardson M. Smart N.G. Cunningham J. Pancreatic endoderm derived from human embryonic stem cells generates glucose-responsive insulin-secreting cells in vivo Nat. Biotechnol. 2008 26 443 452 10.1038/nbt1393 18288110 52. Gouon-Evans V. Boussemart L. Gadue P. Nierhoff D. Koehler C.I. Kubo A. Shafritz D.A. Keller G. BMP-4 is required for hepatic specification of mouse embryonic stem cell-derived definitive endoderm Nat. Biotechnol. 2006 24 1402 1411 10.1038/nbt1258 17086172 53. Hebrok M. Kim S.K. Melton D.A. Notochord repression of endodermal Sonic hedgehog permits pancreas development Genes Dev. 1998 12 1705 1713 10.1101/gad.12.11.1705 9620856 54. Nostro M.C. Sarangi F. Yang C. Holland A. Elefanty A.G. Stanley E.G. Greiner D.L. Keller G. Efficient generation of NKX6-1+ pancreatic progenitors from multiple human pluripotent stem cell lines Stem Cell Rep. 2015 4 591 604 10.1016/j.stemcr.2015.02.017 55. Rezania A. Bruin J.E. Riedel M.J. Mojibian M. Asadi A. Xu J. Gauvin R. Narayan K. Karanu F. O'Neil J.J. Maturation of human embryonic stem cell-derived pancreatic progenitors into functional islets capable of treating pre-existing diabetes in mice Diabetes 2012 61 2016 2029 10.2337/db11-1711 22740171 56. Chen S. Borowiak M. Fox J.L. Maehr R. Osafune K. Davidow L. Lam K. Peng L.F. Schreiber S.L. Rubin L.L. A small molecule that directs differentiation of human ESCs into the pancreatic lineage Nat. Chem. Biol. 2009 5 258 265 10.1038/nchembio.154 19287398 57. Nostro M.C. Sarangi F. Ogawa S. Holtzinger A. Corneo B. Li X. Micallef S.J. Park I.H. Basford C. Wheeler M.B. Stage-specific signaling through TGFb family members and WNT regulates patterning and pancreatic specification of human pluripotent stem cells Development 2011 138 861 871 10.1242/dev.055236 21270052 58. Velazco-Cruz L. Goedegebuure M.M. Maxwell K.G. Augsornworawat P. Hogrebe N.J. Millman J.R. SIX2 Regulates Human b Cell Differentiation from Stem Cells and Functional Maturation In Vitro Cell Rep. 2020 31 107687 10.1016/j.celrep.2020.107687 32460030 59. Zhu S. Russ H.A. Wang X. Zhang M. Ma T. Xu T. Tang S. Hebrok M. Ding S. Human pancreatic beta-like cells converted from fibroblasts Nat. Commun. 2016 7 10080 10.1038/ncomms10080 26733021 60. Zhou X. Nair G.G. Russ H.A. Belair C.D. Li M.L. Shveygert M. Hebrok M. Blelloch R. LIN28B Impairs the Transition of hESC-Derived b Cells from the Juvenile to Adult State Stem Cell Rep. 2020 14 9 20 10.1016/j.stemcr.2019.11.009 61. Velazco-Cruz L. Song J. Maxwell K.G. Goedegebuure M.M. Augsornworawat P. Hogrebe N.J. Millman J.R. Acquisition of Dynamic Function in Human Stem Cell-Derived b Cells Stem Cell Rep. 2019 12 351 365 10.1016/j.stemcr.2018.12.012 62. Docherty F.M. Riemondy K.A. Castro-Gutierrez R. Dwulet J.M. Shilleh A.H. Hansen M.S. Williams S.P.M. Armitage L.H. Santostefano K.E. Wallet M.A. ENTPD3 Marks Mature Stem Cell-Derived b-Cells Formed by Self-Aggregation In Vitro Diabetes 2021 70 2554 2567 10.2337/db20-0873 34380694 63. Nair G.G. Liu J.S. Russ H.A. Tran S. Saxton M.S. Chen R. Juang C. Li M.L. Nguyen V.Q. Giacometti S. Recapitulating endocrine cell clustering in culture promotes maturation of human stem-cell-derived beta cells Nat. Cell Biol. 2019 21 263 274 10.1038/s41556-018-0271-4 30710150 64. Molakandov K. Berti D.A. Beck A. Elhanani O. Walker M.D. Soen Y. Yavriyants K. Zimerman M. Volman E. Toledo I. Selection for CD26 Front. Endocrinol. 2021 12 635405 10.3389/fendo.2021.635405 65. Kaestner K.H. Campbell-Thompson M. Dor Y. Gill R.G. Glaser B. Kim S.K. Sander M. Stabler C. Stewart A.F. Powers A.C. What is a b cell?--Chapter I in the Human Islet Research Network (HIRN) review series Mol. Metab. 2021 53 101323 10.1016/j.molmet.2021.101323 34416394 66. Augsornworawat P. Maxwell K.G. Velazco-Cruz L. Millman J.R. Single-Cell Transcriptome Profiling Reveals b Cell Maturation in Stem Cell-Derived Islets after Transplantation Cell Rep. 2020 32 108067 10.1016/j.celrep.2020.108067 32846125 67. Balboa D. Barsby T. Lithovius V. Saarimaki-Vire J. Omar-Hmeadi M. Dyachok O. Montaser H. Lund P.E. Yang M. Ibrahim H. Functional, metabolic and transcriptional maturation of human pancreatic islets derived from stem cells Nat. Biotechnol. 2022 40 1042 1055 10.1038/s41587-022-01219-z 35241836 68. Faleo G. Russ H.A. Wisel S. Parent A.V. Nguyen V. Nair G.G. Freise J.E. Villanueva K.E. Szot G.L. Hebrok M. Mitigating Ischemic Injury of Stem Cell-Derived Insulin-Producing Cells after Transplant Stem Cell Rep. 2017 9 807 819 10.1016/j.stemcr.2017.07.012 69. Paraskevas S. Maysinger D. Wang R. Duguid T.P. Rosenberg L. Cell loss in isolated human islets occurs by apoptosis Pancreas 2000 20 270 276 10.1097/00006676-200004000-00008 10766453 70. McKenzie M.D. Jamieson E. Jansen E.S. Scott C.L. Huang D.C. Bouillet P. Allison J. Kay T.W. Strasser A. Thomas H.E. Glucose induces pancreatic islet cell apoptosis that requires the BH3-only proteins Bim and Puma and multi-BH domain protein Bax Diabetes 2010 59 644 652 10.2337/db09-1151 19959756 71. Federici M. Hribal M. Perego L. Ranalli M. Caradonna Z. Perego C. Usellini L. Nano R. Bonini P. Bertuzzi F. High glucose causes apoptosis in cultured human pancreatic islets of Langerhans: A potential role for regulation of specific Bcl family genes toward an apoptotic cell death program Diabetes 2001 50 1290 1301 10.2337/diabetes.50.6.1290 11375329 72. Costes S. Vandewalle B. Tourrel-Cuzin C. Broca C. Linck N. Bertrand G. Kerr-Conte J. Portha B. Pattou F. Bockaert J. Degradation of cAMP-responsive element-binding protein by the ubiquitin-proteasome pathway contributes to glucotoxicity in beta-cells and human pancreatic islets Diabetes 2009 58 1105 1115 10.2337/db08-0926 19223597 73. McCall M. Pawlick R. Kin T. Shapiro A.M. Anakinra potentiates the protective effects of etanercept in transplantation of marginal mass human islets in immunodeficient mice Am. J. Transplant. 2012 12 322 329 10.1111/j.1600-6143.2011.03796.x 22053751 74. Ricordi C. Lacy P.E. Finke E.H. Olack B.J. Scharp D.W. Automated method for isolation of human pancreatic islets Diabetes 1988 37 413 420 10.2337/diab.37.4.413 3288530 75. Nikolova G. Jabs N. Konstantinova I. Domogatskaya A. Tryggvason K. Sorokin L. Fassler R. Gu G. Gerber H.P. Ferrara N. The vascular basement membrane: A niche for insulin gene expression and Beta cell proliferation Dev. Cell 2006 10 397 405 10.1016/j.devcel.2006.01.015 16516842 76. Linn T. Schmitz J. Hauck-Schmalenberger I. Lai Y. Bretzel R.G. Brandhorst H. Brandhorst D. Ischaemia is linked to inflammation and induction of angiogenesis in pancreatic islets Clin. Exp. Immunol. 2006 144 179 187 10.1111/j.1365-2249.2006.03066.x 16634789 77. Olsson R. Carlsson P.O. Oxygenation of cultured pancreatic islets Adv. Exp. Med. Biol. 2006 578 263 268 10.1007/0-387-29540-2_42 16927703 78. Maechler P. de Andrade P.B. Mitochondrial damages and the regulation of insulin secretion Biochem. Soc. Trans. 2006 34 824 827 10.1042/BST0340824 17052207 79. Carlsson P.O. Liss P. Andersson A. Jansson L. Measurements of oxygen tension in native and transplanted rat pancreatic islets Diabetes 1998 47 1027 1032 10.2337/diabetes.47.7.1027 9648824 80. Carlsson P.O. Palm F. Andersson A. Liss P. Chronically decreased oxygen tension in rat pancreatic islets transplanted under the kidney capsule Transplantation 2000 69 761 766 10.1097/00007890-200003150-00015 10755523 81. Carlsson P.O. Palm F. Andersson A. Liss P. Markedly decreased oxygen tension in transplanted rat pancreatic islets irrespective of the implantation site Diabetes 2001 50 489 495 10.2337/diabetes.50.3.489 11246867 82. Davalli A.M. Ogawa Y. Ricordi C. Scharp D.W. Bonner-Weir S. Weir G.C. A selective decrease in the beta cell mass of human islets transplanted into diabetic nude mice Transplantation 1995 59 817 820 7701574 83. Carlsson P.O. Palm F. Mattsson G. Low revascularization of experimentally transplanted human pancreatic islets J. Clin. Endocrinol. Metab. 2002 87 5418 5423 10.1210/jc.2002-020728 12466329 84. Stokes R.A. Cheng K. Deters N. Lau S.M. Hawthorne W.J. O'Connell P.J. Stolp J. Grey S. Loudovaris T. Kay T.W. Hypoxia-inducible factor-1a (HIF-1a) potentiates b-cell survival after islet transplantation of human and mouse islets Cell Transplant. 2013 22 253 266 10.3727/096368912X647180 22710383 85. Rodriguez-Brotons A. Bietiger W. Peronet C. Magisson J. Sookhareea C. Langlois A. Mura C. Jeandidier N. Pinget M. Sigrist S. Impact of Pancreatic Rat Islet Density on Cell Survival during Hypoxia J. Diabetes Res. 2016 2016 3615286 10.1155/2016/3615286 26824040 86. Komatsu H. Kang D. Medrano L. Barriga A. Mendez D. Rawson J. Omori K. Ferreri K. Tai Y.C. Kandeel F. Isolated human islets require hyperoxia to maintain islet mass, metabolism, and function Biochem. Biophys. Res. Commun. 2016 470 534 538 10.1016/j.bbrc.2016.01.110 26801563 87. Komatsu H. Cook C. Wang C.H. Medrano L. Lin H. Kandeel F. Tai Y.C. Mullen Y. Oxygen environment and islet size are the primary limiting factors of isolated pancreatic islet survival PLoS ONE 2017 12 e0183780 10.1371/journal.pone.0183780 28832685 88. Garcia-Contreras M. Tamayo-Garcia A. Pappan K.L. Michelotti G.A. Stabler C.L. Ricordi C. Buchwald P. Metabolomics Study of the Effects of Inflammation, Hypoxia, and High Glucose on Isolated Human Pancreatic Islets J. Proteome Res. 2017 16 2294 2306 10.1021/acs.jproteome.7b00160 28452488 89. Maillard E. Juszczak M.T. Langlois A. Kleiss C. Sencier M.C. Bietiger W. Sanchez-Dominguez M. Krafft M.P. Johnson P.R. Pinget M. Perfluorocarbon emulsions prevent hypoxia of pancreatic b-cells Cell Transplant. 2012 21 657 669 10.3727/096368911X593136 21944582 90. Cantley J. Grey S.T. Maxwell P.H. Withers D.J. The hypoxia response pathway and b-cell function Diabetes Obes. Metab. 2010 12 (Suppl. S2) 159 167 10.1111/j.1463-1326.2010.01276.x 21029313 91. Cheng K. Ho K. Stokes R. Scott C. Lau S.M. Hawthorne W.J. O'Connell P.J. Loudovaris T. Kay T.W. Kulkarni R.N. Hypoxia-inducible factor-1alpha regulates beta cell function in mouse and human islets J. Clin. Investig. 2010 120 2171 2183 10.1172/JCI35846 20440072 92. Cantley J. Walters S.N. Jung M.H. Weinberg A. Cowley M.J. Whitworth T.P. Kaplan W. Hawthorne W.J. O'Connell P.J. Weir G. A preexistent hypoxic gene signature predicts impaired islet graft function and glucose homeostasis Cell Transplant. 2013 22 2147 2159 10.3727/096368912X658728 23127310 93. Moritz W. Meier F. Stroka D.M. Giuliani M. Kugelmeier P. Nett P.C. Lehmann R. Candinas D. Gassmann M. Weber M. Apoptosis in hypoxic human pancreatic islets correlates with HIF-1alpha expression FASEB J. 2002 16 745 747 10.1096/fj.01-0403fje 11923216 94. Zhou R. Tardivel A. Thorens B. Choi I. Tschopp J. Thioredoxin-interacting protein links oxidative stress to inflammasome activation Nat. Immunol. 2010 11 136 140 10.1038/ni.1831 20023662 95. Wali J.A. Gurzov E.N. Fynch S. Elkerbout L. Kay T.W. Masters S.L. Thomas H.E. Activation of the NLRP3 inflammasome complex is not required for stress-induced death of pancreatic islets PLoS ONE 2014 9 e113128 10.1371/journal.pone.0113128 25405767 96. do Amaral A.S. Pawlick R.L. Rodrigues E. Costal F. Pepper A. Galvao F.H. Correa-Giannella M.L. Shapiro A.M. Glutathione ethyl ester supplementation during pancreatic islet isolation improves viability and transplant outcomes in a murine marginal islet mass model PLoS ONE 2013 8 e55288 10.1371/journal.pone.0055288 23424628 97. Miwa I. Ichimura N. Sugiura M. Hamada Y. Taniguchi S. Inhibition of glucose-induced insulin secretion by 4-hydroxy-2-nonenal and other lipid peroxidation products Endocrinology 2000 141 2767 2772 10.1210/endo.141.8.7614 10919261 98. Carobbio S. Ishihara H. Fernandez-Pascual S. Bartley C. Martin-Del-Rio R. Maechler P. Insulin secretion profiles are modified by overexpression of glutamate dehydrogenase in pancreatic islets Diabetologia 2004 47 266 276 10.1007/s00125-003-1306-2 14689183 99. Li C. Buettger C. Kwagh J. Matter A. Daikhin Y. Nissim I.B. Collins H.W. Yudkoff M. Stanley C.A. Matschinsky F.M. A signaling role of glutamine in insulin secretion J. Biol. Chem. 2004 279 13393 13401 10.1074/jbc.M311502200 14736887 100. Koulajian K. Ivovic A. Ye K. Desai T. Shah A. Fantus I.G. Ran Q. Giacca A. Overexpression of glutathione peroxidase 4 prevents b-cell dysfunction induced by prolonged elevation of lipids in vivo Am. J. Physiol. Endocrinol. Metab. 2013 305 E254 E262 10.1152/ajpendo.00481.2012 23695217 101. Skouta R. Dixon S.J. Wang J. Dunn D.E. Orman M. Shimada K. Rosenberg P.A. Lo D.C. Weinberg J.M. Linkermann A. Ferrostatins inhibit oxidative lipid damage and cell death in diverse disease models J. Am. Chem. Soc. 2014 136 4551 4556 10.1021/ja411006a 24592866 102. Kaczmarek A. Vandenabeele P. Krysko D.V. Necroptosis: The release of damage-associated molecular patterns and its physiological relevance Immunity 2013 38 209 223 10.1016/j.immuni.2013.02.003 23438821 103. Daly K.A. Liu S. Agrawal V. Brown B.N. Johnson S.A. Medberry C.J. Badylak S.F. Damage associated molecular patterns within xenogeneic biologic scaffolds and their effects on host remodeling Biomaterials 2012 33 91 101 10.1016/j.biomaterials.2011.09.040 21967802 104. Paredes-Juarez G.A. Sahasrabudhe N.M. Tjoelker R.S. de Haan B.J. Engelse M.A. de Koning E.J.P. Faas M.M. de Vos P. DAMP production by human islets under low oxygen and nutrients in the presence or absence of an immunoisolating-capsule and necrostatin-1 Sci. Rep. 2015 5 14623 10.1038/srep14623 26419792 105. Land W.G. Agostinis P. Gasser S. Garg A.D. Linkermann A. DAMP-Induced Allograft and Tumor Rejection: The Circle Is Closing Am. J. Transplant. 2016 16 3322 3337 10.1111/ajt.14012 27529775 106. Land W.G. Agostinis P. Gasser S. Garg A.D. Linkermann A. Transplantation and Damage-Associated Molecular Patterns (DAMPs) Am. J. Transplant. 2016 16 3338 3361 10.1111/ajt.13963 27421829 107. Itoh T. Takita M. SoRelle J.A. Shimoda M. Sugimoto K. Chujo D. Qin H. Naziruddin B. Levy M.F. Matsumoto S. Correlation of released HMGB1 levels with the degree of islet damage in mice and humans and with the outcomes of islet transplantation in mice Cell Transplant. 2012 21 1371 1381 10.3727/096368912X640592 22546320 108. Pepper A.R. Gala-Lopez B. Pawlick R. Merani S. Kin T. Shapiro A.M. A prevascularized subcutaneous device-less site for islet and cellular transplantation Nat. Biotechnol. 2015 33 518 523 10.1038/nbt.3211 25893782 109. Vlahos A.E. Cober N. Sefton M.V. Modular tissue engineering for the vascularization of subcutaneously transplanted pancreatic islets Proc. Natl. Acad. Sci. USA 2017 114 9337 9342 10.1073/pnas.1619216114 28814629 110. Song W. Chiu A. Wang L.H. Schwartz R.E. Li B. Bouklas N. Bowers D.T. An D. Cheong S.H. Flanders J.A. Engineering transferrable microvascular meshes for subcutaneous islet transplantation Nat. Commun. 2019 10 4602 10.1038/s41467-019-12373-5 31601796 111. Yu M. Agarwal D. Korutla L. May C.L. Wang W. Griffith N.N. Hering B.J. Kaestner K.H. Velazquez O.C. Markmann J.F. Islet transplantation in the subcutaneous space achieves long-term euglycaemia in preclinical models of type 1 diabetes Nat. Metab. 2020 2 1013 1020 10.1038/s42255-020-0269-7 32895576 112. Butler M.J. Sefton M.V. Cotransplantation of adipose-derived mesenchymal stromal cells and endothelial cells in a modular construct drives vascularization in SCID/bg mice Tissue Eng. Part A 2012 18 1628 1641 10.1089/ten.TEA.2011.0467 22655687 113. Weaver J.D. Headen D.M. Aquart J. Johnson C.T. Shea L.D. Shirwan H. Garcia A.J. Vasculogenic hydrogel enhances islet survival, engraftment, and function in leading extrahepatic sites Sci. Adv. 2017 3 e1700184 10.1126/sciadv.1700184 28630926 114. Aghazadeh Y. Poon F. Sarangi F. Wong F.T.M. Khan S.T. Sun X. Hatkar R. Cox B.J. Nunes S.S. Nostro M.C. Microvessels support engraftment and functionality of human islets and hESC-derived pancreatic progenitors in diabetes models Cell Stem Cell 2021 28 1936 1949.e8 10.1016/j.stem.2021.08.001 34480863 115. Coronel M.M. Geusz R. Stabler C.L. Mitigating hypoxic stress on pancreatic islets via in situ oxygen generating biomaterial Biomaterials 2017 129 139 151 10.1016/j.biomaterials.2017.03.018 28342320 116. Coronel M.M. Liang J.P. Li Y. Stabler C.L. Oxygen generating biomaterial improves the function and efficacy of beta cells within a macroencapsulation device Biomaterials 2019 210 1 11 10.1016/j.biomaterials.2019.04.017 31029812 117. Liang J.P. Accolla R.P. Soundirarajan M. Emerson A. Coronel M.M. Stabler C.L. Engineering a macroporous oxygen-generating scaffold for enhancing islet cell transplantation within an extrahepatic site Acta Biomater. 2021 130 268 280 10.1016/j.actbio.2021.05.028 34087442 118. Santini-Gonzalez J. Simonovich J.A. Castro-Gutierrez R. Gonzalez-Vargas Y. Abuid N.J. Stabler C.L. Russ H.A. Phelps E.A. In vitro generation of peri-islet basement membrane-like structures Biomaterials 2021 273 120808 10.1016/j.biomaterials.2021.120808 33895491 119. Takahashi Y. Sekine K. Kin T. Takebe T. Taniguchi H. Self-Condensation Culture Enables Vascularization of Tissue Fragments for Efficient Therapeutic Transplantation Cell Rep. 2018 23 1620 1629 10.1016/j.celrep.2018.03.123 29742420 120. Holmberg J. Perlmann T. Maintaining differentiated cellular identity Nat. Rev. Genet. 2012 13 429 439 10.1038/nrg3209 22596319 121. Weir G.C. Aguayo-Mazzucato C. Bonner-Weir S. b-cell dedifferentiation in diabetes is important, but what is it? Islets 2013 5 233 237 10.4161/isl.27494 24356710 122. Bensellam M. Laybutt D.R. Jonas J.C. The molecular mechanisms of pancreatic b-cell glucotoxicity: Recent findings and future research directions Mol. Cell Endocrinol. 2012 364 1 27 10.1016/j.mce.2012.08.003 22885162 123. Bensellam M. Jonas J.C. Laybutt D.R. Mechanisms of b-cell dedifferentiation in diabetes: Recent findings and future research directions J. Endocrinol. 2018 236 R109 R143 10.1530/JOE-17-0516 29203573 124. Talchai C. Xuan S. Lin H.V. Sussel L. Accili D. Pancreatic b cell dedifferentiation as a mechanism of diabetic b cell failure Cell 2012 150 1223 1234 10.1016/j.cell.2012.07.029 22980982 125. Kim-Muller J.Y. Zhao S. Srivastava S. Mugabo Y. Noh H.L. Kim Y.R. Madiraju S.R. Ferrante A.W. Skolnik E.Y. Prentki M. Metabolic inflexibility impairs insulin secretion and results in MODY-like diabetes in triple FoxO-deficient mice Cell Metab. 2014 20 593 602 10.1016/j.cmet.2014.08.012 25264246 126. Guo S. Dai C. Guo M. Taylor B. Harmon J.S. Sander M. Robertson R.P. Powers A.C. Stein R. Inactivation of specific b cell transcription factors in type 2 diabetes J. Clin. Investig. 2013 123 3305 3316 10.1172/JCI65390 23863625 127. Kim-Muller J.Y. Fan J. Kim Y.J. Lee S.A. Ishida E. Blaner W.S. Accili D. Aldehyde dehydrogenase 1a3 defines a subset of failing pancreatic b cells in diabetic mice Nat. Commun. 2016 7 12631 10.1038/ncomms12631 27572106 128. Cinti F. Bouchi R. Kim-Muller J.Y. Ohmura Y. Sandoval P.R. Masini M. Marselli L. Suleiman M. Ratner L.E. Marchetti P. Evidence of b-Cell Dedifferentiation in Human Type 2 Diabetes J. Clin. Endocrinol. Metab. 2016 101 1044 1054 10.1210/jc.2015-2860 26713822 129. Ishida E. Kim-Muller J.Y. Accili D. Pair Feeding, but Not Insulin, Phloridzin, or Rosiglitazone Treatment, Curtails Markers of b-Cell Dedifferentiation in Diabetes 2017 66 2092 2101 10.2337/db16-1213 28506962 130. Stancill J.S. Cartailler J.P. Clayton H.W. O'Connor J.T. Dickerson M.T. Dadi P.K. Osipovich A.B. Jacobson D.A. Magnuson M.A. Chronic b-Cell Depolarization Impairs b-Cell Identity by Disrupting a Network of Ca Diabetes 2017 66 2175 2187 10.2337/db16-1355 28550109 131. Kaelin W.G. The von Hippel-Lindau tumour suppressor protein: O2 sensing and cancer Nat. Rev. Cancer 2008 8 865 873 10.1038/nrc2502 18923434 132. Semenza G.L. HIF-1 mediates metabolic responses to intratumoral hypoxia and oncogenic mutations J. Clin. Investig. 2013 123 3664 3671 10.1172/JCI67230 23999440 133. Puri S. Akiyama H. Hebrok M. VHL-mediated disruption of Sox9 activity compromises beta-cell identity and results in diabetes mellitus Genes Dev. 2013 27 2563 2575 10.1101/gad.227785.113 24298056 134. Beattie G.M. Itkin-Ansari P. Cirulli V. Leibowitz G. Lopez A.D. Bossie S. Mally M.I. Levine F. Hayek A. Sustained proliferation of PDX-1+ cells derived from human islets Diabetes 1999 48 1013 1019 10.2337/diabetes.48.5.1013 10331405 135. Beattie G.M. Leibowitz G. Lopez A.D. Levine F. Hayek A. Protection from cell death in cultured human fetal pancreatic cells Cell Transplant. 2000 9 431 438 10.1177/096368970000900314 10972342 136. Beattie G.M. Montgomery A.M. Lopez A.D. Hao E. Perez B. Just M.L. Lakey J.R. Hart M.E. Hayek A. A novel approach to increase human islet cell mass while preserving beta-cell function Diabetes 2002 51 3435 3439 10.2337/diabetes.51.12.3435 12453897 137. Halvorsen T.L. Beattie G.M. Lopez A.D. Hayek A. Levine F. Accelerated telomere shortening and senescence in human pancreatic islet cells stimulated to divide in vitro J. Endocrinol. 2000 166 103 109 10.1677/joe.0.1660103 10856888 138. Gershengorn M.C. Hardikar A.A. Wei C. Geras-Raaka E. Marcus-Samuels B. Raaka B.M. Epithelial-to-mesenchymal transition generates proliferative human islet precursor cells Science 2004 306 2261 2264 10.1126/science.1101968 15564314 139. Russ H.A. Bar Y. Ravassard P. Efrat S. In vitro proliferation of cells derived from adult human beta-cells revealed by cell-lineage tracing Diabetes 2008 57 1575 1583 10.2337/db07-1283 18316362 140. Joglekar M.V. Joglekar V.M. Joglekar S.V. Hardikar A.A. Human fetal pancreatic insulin-producing cells proliferate in vitro J. Endocrinol. 2009 201 27 36 10.1677/JOE-08-0497 19171567 141. Russ H.A. Ravassard P. Kerr-Conte J. Pattou F. Efrat S. Epithelial-mesenchymal transition in cells expanded in vitro from lineage-traced adult human pancreatic beta cells PLoS ONE 2009 4 e6417 10.1371/journal.pone.0006417 19641613 142. Friedman-Mazursky O. Elkon R. Efrat S. Redifferentiation of expanded human islet b cells by inhibition of ARX Sci. Rep. 2016 6 20698 10.1038/srep20698 26856418 143. Russ H.A. Sintov E. Anker-Kitai L. Friedman O. Lenz A. Toren G. Farhy C. Pasmanik-Chor M. Oron-Karni V. Ravassard P. Insulin-producing cells generated from dedifferentiated human pancreatic beta cells expanded in vitro PLoS ONE 2011 6 e25566 10.1371/journal.pone.0025566 21984932 144. Sintov E. Nathan G. Knoller S. Pasmanik-Chor M. Russ H.A. Efrat S. Inhibition of ZEB1 expression induces redifferentiation of adult human b cells expanded in vitro Sci. Rep. 2015 5 13024 10.1038/srep13024 26264186 145. Toren-Haritan G. Efrat S. TGFb Pathway Inhibition Redifferentiates Human Pancreatic Islet b Cells Expanded In Vitro PLoS ONE 2015 10 e0139168 10.1371/journal.pone.0139168 26418361 146. Lenz A. Toren-Haritan G. Efrat S. Redifferentiation of adult human b cells expanded in vitro by inhibition of the WNT pathway PLoS ONE 2014 9 e112914 10.1371/journal.pone.0112914 25393025 147. Bar Y. Russ H.A. Sintov E. Anker-Kitai L. Knoller S. Efrat S. Redifferentiation of expanded human pancreatic b-cell-derived cells by inhibition of the NOTCH pathway J. Biol. Chem. 2012 287 17269 17280 10.1074/jbc.M111.319152 22457355 148. Weinberg N. Ouziel-Yahalom L. Knoller S. Efrat S. Dor Y. Lineage tracing evidence for in vitro dedifferentiation but rare proliferation of mouse pancreatic beta-cells Diabetes 2007 56 1299 1304 10.2337/db06-1654 17303800 149. Md Moin A.S. Dhawan S. Cory M. Butler P.C. Rizza R.A. Butler A.E. Increased Frequency of Hormone Negative and Polyhormonal Endocrine Cells in Lean Individuals With Type 2 Diabetes J. Clin. Endocrinol. Metab. 2016 101 3628 3636 10.1210/jc.2016-2496 27472443 150. Md Moin A.S. Dhawan S. Shieh C. Butler P.C. Cory M. Butler A.E. Increased Hormone-Negative Endocrine Cells in the Pancreas in Type 1 Diabetes J. Clin. Endocrinol. Metab. 2016 101 3487 3496 10.1210/jc.2016-1350 27300574 151. Md Moin A.S. Cory M. Ong A. Choi J. Dhawan S. Butler P.C. Butler A.E. Pancreatic Nonhormone Expressing Endocrine Cells in Children With Type 1 Diabetes J. Endocr. Soc. 2017 1 385 395 10.1210/js.2017-00081 28782056 152. Van Hulle F. De Groot K. Hilbrands R. Van de Velde U. Suenens K. Stange G. De Mesmaeker I. De Paep D.L. Ling Z. Roep B. Function and composition of pancreatic islet cell implants in omentum of type 1 diabetes patients Am. J. Transplant. 2022 22 927 936 10.1111/ajt.16884 34735732 153. Jonsson J. Carlsson L. Edlund T. Edlund H. Insulin-promoter-factor 1 is required for pancreas development in mice Nature 1994 371 606 609 10.1038/371606a0 7935793 154. Offield M.F. Jetton T.L. Labosky P.A. Ray M. Stein R.W. Magnuson M.A. Hogan B.L. Wright C.V. PDX-1 is required for pancreatic outgrowth and differentiation of the rostral duodenum Development 1996 122 983 995 10.1242/dev.122.3.983 8631275 155. Krapp A. Knofler M. Frutiger S. Hughes G.J. Hagenbuchle O. Wellauer P.K. The p48 DNA-binding subunit of transcription factor PTF1 is a new exocrine pancreas-specific basic helix-loop-helix protein EMBO J. 1996 15 4317 4329 8861960 156. Hald J. Sprinkel A.E. Ray M. Serup P. Wright C. Madsen O.D. Generation and characterization of Ptf1a antiserum and localization of Ptf1a in relation to Nkx6.1 and Pdx1 during the earliest stages of mouse pancreas development J. Histochem. Cytochem. 2008 56 587 595 10.1369/jhc.2008.950675 18347078 157. Schisler J.C. Jensen P.B. Taylor D.G. Becker T.C. Knop F.K. Takekawa S. German M. Weir G.C. Lu D. Mirmira R.G. The Nkx6.1 homeodomain transcription factor suppresses glucagon expression and regulates glucose-stimulated insulin secretion in islet beta cells Proc. Natl. Acad. Sci. USA 2005 102 7297 7302 10.1073/pnas.0502168102 15883383 158. Schisler J.C. Fueger P.T. Babu D.A. Hohmeier H.E. Tessem J.S. Lu D. Becker T.C. Naziruddin B. Levy M. Mirmira R.G. Stimulation of human and rat islet beta-cell proliferation with retention of function by the homeodomain transcription factor Nkx6.1 Mol. Cell. Biol. 2008 28 3465 3476 10.1128/MCB.01791-07 18347054 159. Kawaguchi Y. Cooper B. Gannon M. Ray M. MacDonald R.J. Wright C.V. The role of the transcriptional regulator Ptf1a in converting intestinal to pancreatic progenitors Nat. Genet. 2002 32 128 134 10.1038/ng959 12185368 160. Schaffer A.E. Freude K.K. Nelson S.B. Sander M. Nkx6 transcription factors and Ptf1a function as antagonistic lineage determinants in multipotent pancreatic progenitors Dev. Cell 2010 18 1022 1029 10.1016/j.devcel.2010.05.015 20627083 161. Gu G. Dubauskaite J. Melton D.A. Direct evidence for the pancreatic lineage: NGN3+ cells are islet progenitors and are distinct from duct progenitors Development 2002 129 2447 2457 10.1242/dev.129.10.2447 11973276 162. Gradwohl G. Dierich A. LeMeur M. Guillemot F. neurogenin3 is required for the development of the four endocrine cell lineages of the pancreas Proc. Natl. Acad. Sci. USA 2000 97 1607 1611 10.1073/pnas.97.4.1607 10677506 163. Sussel L. Kalamaras J. Hartigan-O'Connor D.J. Meneses J.J. Pedersen R.A. Rubenstein J.L. German M.S. Mice lacking the homeodomain transcription factor Nkx2.2 have diabetes due to arrested differentiation of pancreatic beta cells Development 1998 125 2213 2221 10.1242/dev.125.12.2213 9584121 164. Prado C.L. Pugh-Bernard A.E. Elghazi L. Sosa-Pineda B. Sussel L. Ghrelin cells replace insulin-producing beta cells in two mouse models of pancreas development Proc. Natl. Acad. Sci. USA 2004 101 2924 2929 10.1073/pnas.0308604100 14970313 165. Sosa-Pineda B. Chowdhury K. Torres M. Oliver G. Gruss P. The Pax4 gene is essential for differentiation of insulin-producing beta cells in the mammalian pancreas Nature 1997 386 399 402 10.1038/386399a0 9121556 166. Collombat P. Mansouri A. Hecksher-Sorensen J. Serup P. Krull J. Gradwohl G. Gruss P. Opposing actions of Arx and Pax4 in endocrine pancreas development Genes Dev. 2003 17 2591 2603 10.1101/gad.269003 14561778 167. Collombat P. Hecksher-Sorensen J. Broccoli V. Krull J. Ponte I. Mundiger T. Smith J. Gruss P. Serup P. Mansouri A. The simultaneous loss of Arx and Pax4 genes promotes a somatostatin-producing cell fate specification at the expense of the beta-cell lineages in the mouse endocrine pancreas Development 2005 132 2969 2980 10.1242/dev.01870 15930104 168. Collombat P. Hecksher-Sorensen J. Krull J. Berger J. Riedel D. Herrera P.L. Serup P. Mansouri A. Embryonic endocrine pancreas and mature beta cells acquire alpha and PP cell phenotypes upon Arx misexpression J. Clin. Investig. 2007 117 961 970 10.1172/JCI29115 17404619 169. Schaffer A.E. Taylor B.L. Benthuysen J.R. Liu J. Thorel F. Yuan W. Jiao Y. Kaestner K.H. Herrera P.L. Magnuson M.A. Nkx6.1 controls a gene regulatory network required for establishing and maintaining pancreatic Beta cell identity PLoS Genet. 2013 9 e1003274 10.1371/journal.pgen.1003274 23382704 170. Taylor B.L. Liu F.F. Sander M. Nkx6.1 is essential for maintaining the functional state of pancreatic beta cells Cell Rep. 2013 4 1262 1275 10.1016/j.celrep.2013.08.010 24035389 171. Papizan J.B. Singer R.A. Tschen S.I. Dhawan S. Friel J.M. Hipkens S.B. Magnuson M.A. Bhushan A. Sussel L. Nkx2.2 repressor complex regulates islet b-cell specification and prevents b-to-a-cell reprogramming Genes Dev. 2011 25 2291 2305 10.1101/gad.173039.111 22056672 172. Herrera P.L. Adult glucagon-producing cells differentiate from two independent cell lineages Development 2000 127 2317 2322 10.1242/dev.127.11.2317 10804174 173. Lee K. Chan J.Y. Liang C. Ip C.K. Shi Y.C. Herzog H. Hughes W.E. Bensellam M. Delghingaro-Augusto V. Koina M.E. XBP1 maintains beta cell identity, represses beta-to-alpha cell transdifferentiation and protects against diabetic beta cell failure during metabolic stress in mice Diabetologia 2022 65 984 996 10.1007/s00125-022-05669-7 35316840 174. Spijker H.S. Ravelli R.B. Mommaas-Kienhuis A.M. van Apeldoorn A.A. Engelse M.A. Zaldumbide A. Bonner-Weir S. Rabelink T.J. Hoeben R.C. Clevers H. Conversion of mature human b-cells into glucagon-producing a-cells Diabetes 2013 62 2471 2480 10.2337/db12-1001 23569174 175. Butler A.E. Campbell-Thompson M. Gurlo T. Dawson D.W. Atkinson M. Butler P.C. Marked expansion of exocrine and endocrine pancreas with incretin therapy in humans with increased exocrine pancreas dysplasia and the potential for glucagon-producing neuroendocrine tumors Diabetes 2013 62 2595 2604 10.2337/db12-1686 23524641 176. Spijker H.S. Song H. Ellenbroek J.H. Roefs M.M. Engelse M.A. Bos E. Koster A.J. Rabelink T.J. Hansen B.C. Clark A. Loss of b-Cell Identity Occurs in Type 2 Diabetes and Is Associated With Islet Amyloid Deposits Diabetes 2015 64 2928 2938 10.2337/db14-1752 25918235 177. Pipeleers D. The biosociology of pancreatic B cells Diabetologia 1987 30 277 291 10.1007/BF00299019 3301479 178. Pipeleers D.G. Heterogeneity in pancreatic beta-cell population Diabetes 1992 41 777 781 10.2337/diab.41.7.777 1612191 179. Pipeleers D. Kiekens R. Ling Z. Wilikens A. Schuit F. Physiologic relevance of heterogeneity in the pancreatic beta-cell population Diabetologia 1994 37 (Suppl. S2) S57 S64 10.1007/BF00400827 7821741 180. Bader E. Migliorini A. Gegg M. Moruzzi N. Gerdes J. Roscioni S.S. Bakhti M. Brandl E. Irmler M. Beckers J. Identification of proliferative and mature beta-cells in the islets of Langerhans Nature 2016 535 430 434 10.1038/nature18624 27398620 181. van der Meulen T. Mawla A.M. DiGruccio M.R. Adams M.W. Nies V. Dolleman S. Liu S. Ackermann A.M. Caceres E. Hunter A.E. Virgin Beta Cells Persist throughout Life at a Neogenic Niche within Pancreatic Islets Cell Metab. 2017 25 911 926.e916 10.1016/j.cmet.2017.03.017 28380380 182. Bosco D. Rouiller D.G. Halban P.A. Differential expression of E-cadherin at the surface of rat beta-cells as a marker of functional heterogeneity J. Endocrinol. 2007 194 21 29 10.1677/JOE-06-0169 17592017 183. Segerstolpe A. Palasantza A. Eliasson P. Andersson E.M. Andreasson A.C. Sun X. Picelli S. Sabirsh A. Clausen M. Bjursell M.K. Single-Cell Transcriptome Profiling of Human Pancreatic Islets in Health and Type 2 Diabetes Cell Metab. 2016 24 593 607 10.1016/j.cmet.2016.08.020 27667667 184. Xin Y. Dominguez Gutierrez G. Okamoto H. Kim J. Lee A.H. Adler C. Ni M. Yancopoulos G.D. Murphy A.J. Gromada J. Pseudotime Ordering of Single Human b-Cells Reveals States of Insulin Production and Unfolded Protein Response Diabetes 2018 67 1783 1794 10.2337/db18-0365 29950394 185. Dorrell C. Schug J. Canaday P.S. Russ H.A. Tarlow B.D. Grompe M.T. Horton T. Hebrok M. Streeter P.R. Kaestner K.H. Human islets contain four distinct subtypes of beta cells Nat. Commun. 2016 7 11756 10.1038/ncomms11756 27399229 186. Wojtusciszyn A. Armanet M. Morel P. Berney T. Bosco D. Insulin secretion from human beta cells is heterogeneous and dependent on cell-to-cell contacts Diabetologia 2008 51 1843 1852 10.1007/s00125-008-1103-z 18665347 187. Hiriart M. Ramirez-Medeles M.C. Functional subpopulations of individual pancreatic B-cells in culture Endocrinology 1991 128 3193 3198 10.1210/endo-128-6-3193 2036985 188. Salem V. Silva L.D. Suba K. Georgiadou E. Neda Mousavy Gharavy S. Akhtar N. Martin-Alonso A. Gaboriau D.C.A. Rothery S.M. Stylianides T. Leader b-cells coordinate Ca Nat. Metab. 2019 1 615 629 10.1038/s42255-019-0075-2 32694805 189. Nunemaker C.S. Wasserman D.H. McGuinness O.P. Sweet I.R. Teague J.C. Satin L.S. Insulin secretion in the conscious mouse is biphasic and pulsatile Am. J. Physiol. Endocrinol. Metab. 2006 290 E523 E529 10.1152/ajpendo.00392.2005 16249252 190. Gilon P. Ravier M.A. Jonas J.C. Henquin J.C. Control mechanisms of the oscillations of insulin secretion in vitro and in vivo Diabetes 2002 51 (Suppl. S1) S144 S151 10.2337/diabetes.51.2007.s144 11815474 191. Matveyenko A.V. Liuwantara D. Gurlo T. Kirakossian D. Dalla Man C. Cobelli C. White M.F. Copps K.D. Volpi E. Fujita S. Pulsatile portal vein insulin delivery enhances hepatic insulin action and signaling Diabetes 2012 61 2269 2279 10.2337/db11-1462 22688333 192. Johnston N.R. Mitchell R.K. Haythorne E. Pessoa M.P. Semplici F. Ferrer J. Piemonti L. Marchetti P. Bugliani M. Bosco D. Beta Cell Hubs Dictate Pancreatic Islet Responses to Glucose Cell Metab. 2016 24 389 401 10.1016/j.cmet.2016.06.020 27452146 193. Thompson P.J. Shah A. Ntranos V. Van Gool F. Atkinson M. Bhushan A. Targeted Elimination of Senescent Beta Cells Prevents Type 1 Diabetes Cell Metab. 2019 29 1045 1060.e10 10.1016/j.cmet.2019.01.021 30799288 194. Rodier F. Coppe J.P. Patil C.K. Hoeijmakers W.A. Munoz D.P. Raza S.R. Freund A. Campeau E. Davalos A.R. Campisi J. Persistent DNA damage signalling triggers senescence-associated inflammatory cytokine secretion Nat. Cell Biol. 2009 11 973 979 10.1038/ncb1909 19597488 195. Tchkonia T. Zhu Y. van Deursen J. Campisi J. Kirkland J.L. Cellular senescence and the senescent secretory phenotype: Therapeutic opportunities J. Clin. Investig. 2013 123 966 972 10.1172/JCI64098 23454759 196. Kirkland J.L. Tchkonia T. Cellular Senescence: A Translational Perspective EBioMedicine 2017 21 21 28 10.1016/j.ebiom.2017.04.013 28416161 197. Aguayo-Mazzucato C. Andle J. Lee T.B. Jr. Midha A. Talemal L. Chipashvili V. Hollister-Lock J. van Deursen J. Weir G. Bonner-Weir S. Acceleration of beta Cell Aging Determines Diabetes and Senolysis Improves Disease Outcomes Cell Metab. 2019 30 129 142.e124 10.1016/j.cmet.2019.05.006 31155496 |
PMC10000643 | Adenoid cystic carcinoma (AdCC), a rare heterogenous disease, presents diagnostic, prognostic, and therapeutic challenges. To obtain more knowledge, we conducted a retrospective study on a cohort of 155 patients diagnosed in 2000-2022 with AdCC of the head and neck in Stockholm and investigated several clinical parameters in correlation to treatment and prognosis in the 142/155 patients treated with curative intent. The strongest favourable prognostic factors were early disease stage (stage I and II) as compared to late disease (stage III and IV) and major salivary gland subsite as compared to other subsites, with the best prognosis in the parotid gland, irrespective of the stage of the disease. Notably, in contrast to some studies, a significant correlation to survival was not found for perineural invasion or radical surgery. However, similar to others, we confirmed that other common prognostic factors, e.g., smoking, age, and gender, did not correlate to survival and should not be used for prognostication of AdCC of the head and neck. To conclude, in AdCC early disease stage, major salivary gland subsite and multimodal treatment were the strongest favourable prognostic factors, while this was not the case for age, gender and smoking nor perineural invasion and radical surgery. adenoid cystic carcinoma prognostic factors subsites perineural invasion treatment Swedish Cancer Foundation20 0704 20 0778 21 0292 JCIA Stockholm Cancer Society201092 201242 210292 221293 221082 Swedish Cancer and Allergy Foundation10367 10662 The Stockholm City Council20180037 20200059 20210855 Magnus Bergvalls stiftelse2020-03737 2021-04359 Stiftelsen Sigurd och Elsa Goljes MinneLA2020-0070 LA2021-0005 The Stockholm RegionKarolinska InstitutetThis research was funded by the Swedish Cancer Foundation (grant no. 20 0704; 20 0778 and 21 0292 JCIA), the Stockholm Cancer Society (grant no. 201092; 201242, 210292, 221293, 221082), the Swedish Cancer and Allergy Foundation (grant no. 10367; 10662), the Stockholm City Council (grant no. 20180037; 20200059, 20210855), Karolinska Institutet (Magnus Bergvalls stiftelse (grant no. 2020-03737, 2021-04359), Stiftelsen Sigurd och Elsa Goljes Minne (LA2020-0070; LA2021-0005, The Karolinska Institutet and the Stockholm Region. pmc1. Introduction Adenoid cystic carcinoma (AdCC) is a rare malignancy originating from secretory glands with yet unknown aetiology and accounts for approximately 1% of malignant head and neck tumours . It is mainly found in the major and minor salivary glands, there accounting for approximately 30% of malignant salivary gland tumours . However, AdCC can also arise in secretory tissue in other areas of the head and neck region and, more rarely, also in secretory glands outside this area, e.g., in the oesophagus, breast, lung, prostate, and vulva . AdCC can occur at all ages but is most common later in life (Fifth to Sixth decade) . It frequently presents with unspecific symptoms, slow growth, the perineural invasion, and it is often problematic to diagnose because of the difficulty in defining its clinical differentiation and distinguishing it from benign tumours . Moreover, there is a lack of good prognostic markers , and conventionally used clinical prognostic markers, e.g., gender, age, or smoking status, are less reliable in AdCC . AdCC is usually treated with surgery upfront, often followed by postoperative radiotherapy, but despite an aggressive therapy regimen late (>5 years after primary treatment), local relapse (15-85%) and distant metastases (25-55%) are common . Due to its slow growth, the disease often runs an indolent course, with 5-year survival rates of 80-85%, but contrary to other head and neck cancers, these promising numbers decline when following 15-year survival rates, which are 50-60% and 30-35%, respectively . Prospective clinical trials are uncommon, but many retrospective studies have shown that disease-free survival (DFS) is better when multimodal therapy is given . One report showed that patients treated with surgery as well as radiotherapy had a much better 5-year local control rate than those given only surgery . Another study compared radiotherapy alone with a combination of surgery and radiotherapy; moreover, they also found that local control rates were better with multimodal treatment . However, even if DFS is prolonged with combination therapy upon a 5-year follow-up, its effect on long-term (10-15 years) DFS and overall survival (OS) may differ, which in turn is due to the very long follow-up time needed depending on the slow progression of AdCC . Moreover, there are no standard recommendations on systemic chemotherapy (ChT) in AdCC patients, and specific ChT regimens have not been proven effective in clinical trials . It has, however, been suggested that the lack of effect of cytotoxic agents is related to the very slow-growing biology of this tumour, but interestingly even in the spread and progressive disease, the treatment response to ChT is very limited (objective response < 20%) . Nonetheless, although not effective, ChT is still being used due to the lack of other effective treatment options for spreading disease . Importantly, however, AdCC is a rare and heterogenic disease, and many of the studied cohorts have been small, and moreover, since clinical studies are rare, it has been difficult to draw general conclusions. For this reason, here, we attempted to collect a large cohort of AdCC patients over more than a 20-year period to investigate clinical characteristics and long-time survival in AdCC patients originating from the head and neck region. 2. Materials and Methods 2.1. Patient's Characteristics Patients and tumour samples. Through the Swedish Cancer Registry, between 2000 and 2022, in total, 155 patients were identified as diagnosed with AdCC within the head and neck region (ICD-10: C00.5, C01.9, C04.9, C05.9, C06.9, C07.9, C08.0, C09.9, C11.9, C30.0, and C31.9) in the County of Stockholm/Gotland. Following the identification of these 155 patients in the Swedish Cancer Registry, we could successfully collect and examine the individual charts of all 155 patients at Karolinska University Hospital. Patients' charts (collected prospectively during regular contact with the patients during the follow-up period) were assessed. Their tumours were reclassified according to the 8th Edition UICC TNM classification of Malignant Tumours, and age, gender, smoking, performance status (Eastern Cooperative Oncology Group, ECOG), perineural growth pattern, type of treatment, recurrence, time to recurrence and location of recurrence, as well as survival were recorded. Treatment was classified as surgery, radiotherapy (RT) (ablative doses of minimum 64-68 Gy, including both external beam radiotherapy and brachytherapy), chemoradiotherapy (CRT), or palliative therapy (including RT, CRT, and chemotherapy (ChT), but not at curative doses). Smoking data were obtained whenever noted in the charts and were classified as "never" smokers or "ever" smokers (including recent and previous smokers). All 155 patients were described and characterized according to the criteria above, but 13 were excluded from further survival analyses due to that they did not receive curative treatment or because they were not disease free 6 months after treatment. Furthermore, for the survival analysis, the 142 patients treated with curative intent were grouped into three subgroups. Most patients (118 cases) were included in the group defined as receiving multimodal treatment and receiving both surgery and RT with or without concomitant chemotherapy (CRT). The remaining patients were divided into two minor groups, one group with 14 cases obtaining only oncological treatment with RT or CRT and one group (10 cases) having surgery alone. DFS was followed through the charts of the patients, while OS was also followed through our access to the Swedish Death Registry from the Karolinska University Hospital. It is of note that there was no standard follow-up recommendation for AdCC patients and, therefore, in our cohort, all patients did not have the same follow-up period. Nevertheless, the standard follow-up for head and neck cancer (HNC) patients in our institution is clinical controls every 3rd month for the first 2 years and every 6th month until 5 years after treatment. Most patients treated without surgery undergo a computer tomography scan 3 months post-treatment to ensure that there is no residual tumour; the latter is, however, a practice of the more recent years. Most AdCC patients in this cohort continued with clinical controls 2 times yearly after 5 years due to the fact that late recurrences are common. Moreover, the date of death is automatically provided for all citizens that have deceased in Sweden through the digital medical journal system, providing reliable OS data. The study was conducted according to ethical permissions 99-237, 2005/431-31/4, 2009/1278-31/4, 2012/83-31/2, 2017/1035-31/2, 2019-05211 and 2022-05287-02 from the Ethics Committee at Karolinska Institutet, Stockholm, Sweden, the Stockholm Regional Ethical Review Board, and the Swedish Ethical Review Authority. 2.2. Statistical and Survival Analysis The Kaplan-Meier method was used to estimate overall survival (OS) and disease-free survival (DFS). OS was defined as the time from diagnosis until the date of death of any cause or end of follow-up, whichever comes first. DFS was defined as the time from diagnosis until local or distant recurrence confirmed by radiology or clinical check-up. The chi-square test was used to evaluate differences in categorical data; for continuous variables, an independent two-tailed t-test was used. A test result below 5% was considered statistically significant. R version 3.4.1 was used for data management and analysis. 3. Results 3.1. Patient and Tumour Characteristics In our cohort of 155 AdCC cases, there was a general predominance of female patients (96 females (F) and 59 males (M); F:M ratio 1.6:1) irrespective of tumour site, with the highest female ratio in AdCC of the parotid gland (F:M ratio 2.5:1) and the lowest female ratio in AdCC of the oral cavity (F:M ratio 1.25:1). For details see Table 1. In addition, there was a clear predominance of AdCC localized in the major salivary glands (parotid and submandibular gland), the nasal cavity, the paranasal sinus, and the oral cavity (27.1%, 27.1%, 15.5%, and 17.4%, respectively) (Table 1). More specifically, 55.5% of the cases resided in the major salivary glands (if including the two cases of the sublingual gland). Less common tumour sites accounted for 20 cases (12.9%), and these were located in the oropharynx (7 cases, 4.5%), the lip (5 cases, 3.5%), the sublingual gland (2 cases, 1.3%), the larynx (2 cases, 1.3%), the nasopharynx (2 cases, 1.3%), and finally one case each (1 case, 0.65%) was located in the ear and hypopharynx. A minority of the patients (8/155, 5.2%) presented with spread disease upon diagnosis; furthermore, in one case, the metastasis status was unknown because the patient did not want to undergo further diagnostics due to a high age at diagnosis (90 years old). This patient did not die of AdCC and became 94 years old. Stage IV disease (including stages IVa, IVb, and IVc) dominated (54 cases, 35.3%), followed by stage II (44 cases, 28.8%), stage I (34 cases, 22.2%), and stage III (21 cases, 13.7%) (Table 1). Male patients tended to present with lower stage; 33/59 (55.9%) stage I and II disease, and 26/59 (44.1%) with stage III and IV, as compared to female patients, where the corresponding proportions were 45/96 (46.9%) and 49/96 (51%), respectively (in two female patients no staging could be assessed). Stage IV was predominant in AdCC in the nasal cavity and paranasal sinuses (17/24, 70.8%) in contrast to the submandibular gland, where only five patients presented with stage IV disease (5/42, 11,9%). Perineural tumour growth was predominantly found in 100/155 (64.5%) cases, while 52/155 (33.5%) cases presented without a perineural growth pattern and in three cases, the status was not assessable. Perineural growth was over-represented in cases arising from the major salivary glands 62/84 (73.8%), whereas only 38/69 (55.1%) arising from other sites presented with a perineural growth pattern. In total, 142/155 (91.6%) patients were eligible for treatment with curative intent and were disease-free six months after finishing treatment and included in the further survival analyses. Of the excluded 13 patients, seven had advanced local and/or spread disease at diagnosis, and four were not eligible for intensive treatment due to high biological age and poor performance status. In addition, one patient did not accept treatment, and one showed signs of residual disease after surgical resection and full dose postoperative external beam radiotherapy (not disease-free after a definitive curative treatment regimen). Out of 142 patients treated with curative intent, 140 presented with locoregional diseases, while 2 had both locoregional and distant diseases. The latter two patients both had one single metastasis. One patient had metastasis in the liver (primary AdCC in the parotid gland), while the other one had metastasis in the lung (primary AdCC in the nasopharynx). However, both metastases were stable under the curative treatment of the locoregional disease and could later be radically extirpated (negative surgical margins), but unfortunately, both these patients presented with distant recurrence in the lung within three years after finished treatment. A summary of all the patients treated with curative intent is presented in Table 2. 3.2. Patients Treated with Curative Intent and Their Clinical Characteristics The 142 patients treated with curative intent were for the sake of simplicity divided into three subgroups, and the characteristics of these patients are presented in Table 2. The largest group (118 cases) was defined as receiving multimodal treatment and receiving both surgery and RT with or without concomitant chemotherapy (CRT). The second group (14 cases) obtained only oncological treatment with RT or CRT, and the third group (10 cases) had surgery alone. Of the 118/142 (83.1%) cases treated with multimodal treatment, 89/142 (62.7%) underwent surgery followed by RT, while 29/142 (20.4%) had surgery and postoperative RT (PORT) with concomitant ChT (CRT). Of the 14/142 (9.9%) patients that did not undergo surgery, six received only RT, while eight obtained (CRT). The remaining 10/142 (7%) patients were treated with surgery as a single treatment modality. Below further details per treatment category and patient characteristics are presented. 3.2.1. The Multimodal Treatment Group This group consisted of 118 patients and was heterogenic, with no subsite or disease stage being overrepresented, suggesting that the multimodal treatment modality was not chosen due to the more advanced disease stage or a certain localization. More specifically, of the 118 patients, 27/118, 22.9% (vs. 34/142, 22.2%) cases were stage I, 38/118, 32.2% (vs. 43/142, 30.3%) stage II, 18/118, 15.2% (vs. 20/142, 14.1%) stage III, and 35/118, 29.7% (vs 45/142, 31.7%) stage IV. Any recurrence was observed in 39/118 (33.1%) patients. These recurrences were most frequently deriving from primary AdCCs in the nasal and paranasal sinuses 11/18 (61.1%), followed by those in the oral cavity 8/17 (47.1%), the submandibular gland 12/35 (34.3%), 4/14 (28.6%) in other sites, and finally, they were least commonly found 4/34 (11.8%) in AdCC primaries of the parotid gland. Locoregional recurrence (LRR), occurring in 15 cases (15/118, 12.7%), was most common in the nasal cavity and paranasal sinuses with 8/15 (53.3%) cases, where all but one had initially presented with stage IV disease at diagnosis. Interestingly 11/15 (73.3%) cases with LRR developed even distant metastases. However, only six of these patients died within five years of the diagnosis of LRR. Distant metastases were developed by 36/118 (30.5%) patients, of which 25/36 (69.4%) were diagnosed only with distant metastases and 11/36 (30.6%) with both LRR and distant diseases (see above). Distant metastasis without LRR was most common among primary AdCC located in the oral cavity 5/17 (29.4%), followed by those located in the submandibular gland 10/35 (28.6%), while the distant disease was less common in primary AdCC at other sites. The most frequent site of distant metastases was the lungs (25/36, 69.5%), followed by the liver (4/36, 11.1%), which was also the most common site of secondary metastases after the lungs. Other sites of the distant disease include: the bones, the kidneys and the adrenal glands, the brain, the skin, the lymph nodes, and the thyroid gland. 3.2.2. The RT and CRT Group In this group, the majority 12/14 (85.7%) presented with advanced disease (ten cases with stage IV and two with stage III disease), and 12/14 (85.7%) presented with any late recurrence, of which five cases had LRR, four had both LRR and distant metastasis, whereas two developed distant disease alone. 3.2.3. The Surgery Alone Group In this group, all 10 cases had the early-stage disease (stage I and II), and the surgery was radical in 8 cases, while in the 2 cases where the surgery was not radical, both patients were recommended PORT, but this was not conducted initially. One patient undergoing non-radical surgery later presented with locoregional recurrence and was treated with RT, while the other never encountered a recurrence. Of the eight cases having a radical extirpation of the tumour, one patient still presented with a locoregional recurrence (nasal cavity). 3.3. Survival Analysis Several survival analyses were performed on the 142 patients treated with curative intent, and some of the most notable parameters are presented below. 3.3.1. Long-Term Disease-Free Survival and Overall Survival for Patients Treated with Curative Intent Initial survival analysis for DFS and OS was performed including all 142 patients treated with curative intent, irrespective of treatment modality. For this cohort, the 5-year DFS was 64.9%, the 10-year DFS was 49.6%, and the 15-year DFS was 37.7%, while the 5-year OS was 83.5%, the 10-year OS was 59.4%, and the 15-year OS was 42.5% . 3.3.2. Long-Term Disease-Free Survival Separated for Different Treatment Modalities Patients receiving multimodal treatment had a significantly better DFS compared to patients receiving single treatments, i.e., the 14 patients in the RT and CRT group and the 10 patients receiving surgery alone (log-rank: p < 0.0001). More specifically, the five-year DFS was 70.7% in patients treated with multimodal treatment compared to 37% in patients receiving single modality treatment , and similar data were found for OS . 3.3.3. Long-Term Disease-Free Survival Separated for Treatment with Surgery and RT, and Surgery and CRT To examine whether postoperative CRT improved survival compared to PORT, we compared these groups and found a non-significant tendency of a longer DFS and OS in the PORT group . Notably, both groups were heterogenic; however, although there was no clear disproportion of different AdCC subsites, the stages of the disease proportions were imbalanced. Advanced disease was clearly overrepresented in the group receiving postoperative CRT (Stage IV: 13/29, 44.8%, Stage III: 6/29, 20.7%), whereas the group receiving PORT had more early-stage disease cases (Stage I: 24/89, 27%; Stage II: 31/89, 34.8%). Furthermore, in this cohort, we investigated how radical surgery (negative surgical margin in the pathology report) correlated with DFS (recurrence of any kind) and OS; however, there was no significant difference between the groups (data not shown). 3.3.4. Long-Term Disease-Free Survival in Relation to Gender, Age, and Smoking in AdCC To investigate whether gender, age, and smoking today used as classic prognostic factors in the clinic are prognostic in AdCC, we performed a separate survival analysis for these parameters. For age, we used the median age in our cohort, which was 58.5 years, and we separated the survival analysis for age >58.5 and <58.5 years. For smoking, a dichotomized smoking status (Ever vs. Never) was used. We could not confirm any of these factors to be prognostic for DFS in our cohort of 142 patients receiving treatment with curative intent, and the data are presented for DFS in Figure 4A-C, respectively, and for OS in Supplementary Figure S3. 3.3.5. Long-Term Disease-Free and Overall Survival for Patients Treated with Curative Intent, Independent of Treatment Modality, in Relation to Perineural Growth Status Perineural tumour growth was observed in 95 (66.9%) of all cases undergoing curative treatment and was, in general, more frequent in the major salivary glands and most frequent in tumours from the submandibular gland (31/38; 81.6%), followed by the parotid gland 28/41 (68.3%), the oral cavity 16/25 (64%), the nasal cavity and paranasal sinuses 13/21 (61.9%). Perineural invasion was, however, much less common (7/17, 41.2%) in AdCC arising from other sites. Markedly, perineural growth did not always correlate with a higher disease stage since 55.9% of the cases had perineural invasion in stage I, 74.4% in stage II, 80% in stage III and 62.2% in stage IV. Furthermore, it is of note that there were no significant differences in DFS nor OS when separating the analysis for perineural invasion, dichotomizing positive and negative status, according to the pathological report . 3.3.6. Long-Term Disease-Free and Overall Survival for Patients Treated with Curative Intent, Independent of Treatment Modality, in Relation to Tumour Stage at Diagnosis As mentioned above, of the 142 patients, 34 (23.9%) presented tumours in stage I, 43 (30.3%) in stage II, 20 (14.1%) in stage III, and 45 (31.7%) in stage IV. Patients presented with the lower disease stage, treated with curative intent, had a significantly better DFS and OS, with DFS presented in Figure 6A and OS in Supplementary Figure S4A. After performing the survival analysis for each individual stage, it became clear that stages could be clustered into two groups, with early-stage group clustering stages I and II and advanced-stage group clustering stages III and IV. Stages I and II clustered together had significantly better DFS and OS ; more specifically, five-year DFS and OS were 83.6% and 90.0%, respectively, for the early-stage group, and 41.3% and 75.0%, respectively, for the advanced stage group. In addition, 10-year DFS and OS were 69.2% and 72.3%, respectively, for the early-stage group and 23.9% and 42.7%, respectively, for the advanced-stage group. 3.3.7. Long-Term Disease-Free and Overall Survival in Patients Treated with Curative Intent, Independent of Treatment Modality, in Relation to Most Common Tumour Subsites To investigate whether primary AdCC within different subsites of the head and neck region have a different prognosis or not, we analysed DFS and OS according to the four most frequent subsites, independent of treatment modality. Altogether 125 cases of primary AdCC, all originating from either the parotid gland (41/125, 32.8%), the submandibular gland (38/125, 30.4%), the nasal cavity, the paranasal sinuses (25/125, 20%), and the oral cavity (20/125, 16%), were included in the analyses. Primaries from the nasal cavity and paranasal sinuses (Subsite C) had the least favourable DFS with a 5-year DFS of 42%, followed by the oral cavity at 61.2% (Subsite D), the submandibular grand 68.3% (Subsite B), and the parotid gland 81.1% (Subsite A), and for details of 15-year DFS, see Figure 7A. For OS analysis, the differences between the groups were still significant (log-rank p = 0.0062) but less obvious . The early recurrences of the primaries from the nasal cavity and paranasal sinuses seemed to influence the OS only marginally. The five-year OS for the submandibular gland, the oral cavity and the nasal cavity and paranasal sinuses was very similar (82.4%, 82.5%, and 73.7%, respectively). Primaries from the parotid gland had the best OS (5-year OS of 94.3%), and for details of 15-year OS, see Figure 7B. Since the prognosis of AdCC in the major salivary glands seemed to be more favourable when compared to that in other sites, a further analysis was performed comparing DFS and OS for all major salivary glands (parotid, submandibular, and sublingual gland) to DFS and OS of all other sites . AdCC in major salivary glands had better DFS and OS compared to other sites (log-rank test, DFS: p = 0.00016; OS: p = 0.0061); more explicitly, five-year DFS and OS were 73.5% and 88.7%, respectively in AdCC of the major salivary glands as compared to 53.7% and 76.8%, respectively of AdCC in other sites. Moreover, 10-year DFS and OS for salivary glands were 61.3% and 63.6%, respectively, compared to 33.8% and 54.2%, respectively, for other sites. For further details and 15-year DFS and OS, see Figure 8A,B and Supplementary Table S1. 3.3.8. Multivariable Analysis for Different Prognostic Factors Used in the Survival Analyses for the Entire Cohort, Treated with Curative Intent A multivariable subgroup analysis, including all patients treated with curative intent, irrespective of treatment modality, was performed. Multivariable analyses for DFS and OS were performed for gender, age, smoking status, stage of disease (I + II vs. III + IV), perineural invasion, and dichotomizing surgery and RT/CRT vs. surgery or RT/CRT alone, as well as subsites (Salivary glands vs. Other subsides). For DFS, see Figure 9 below and for OS, see Supplementary Figure S5. (Univariable analysis was very similar and is shown in Supplementary Table S2.) The multivariable analysis confirmed that typical clinical prognostic factors, such as gender, age, smoking status, and perineural growth pattern, were not applicable in AdCC. However, the multivariable analysis supported that the stage of the disease, as well as the subside of the AdCC, could be used as a prognostic factor. In addition, the multivariable analysis supported that combinational treatment of surgery and RT w/o ChT was superior to single treatment modality and resulted in improved survival; for details of DFS, see Figure 9 above and for OS, see Supplementary Figure S5. 4. Discussion In this retrospective study with a follow-up for >20 years, a large cohort of head and neck AdCC patients was examined with regard to clinical characteristics and long-time survival in correlation to gender, age, smoking, perineural invasion, tumour stage, subsite, and treatment modality. We could disclose that both DFS and OS were superior following multimodal treatment modality vs. single treatment modalities. Moreover, patients with tumour stages I and II had a better prognosis than those with stage III and IV disease and those with major salivary gland AdCC had a better outcome than those with AdCC at other subsites, irrespective of disease stage. In contrast, gender, age, smoking status, and perineural growth pattern in the tumour, factors otherwise used for prognostication in many solid cancers , did not affect either DFS or OS. Based on the findings above, we first compared our cohort to those of others and found no major differences in the characteristics of patients included in the study as compared to those included in other reports . More specifically, the overrepresentation of female patients (F:M ratio 1.6:1) and the median age of 58.5 years at diagnosis were not unexpected since AdCC patients often are younger than other head and neck cancer patients and AdCC is more common in females . The latter has been discussed to be linked to biological differences between the sexes and different hormonal levels , but this probably plays a less important role in the AdCC of the head and neck region. Moreover, only around 5% of the patients presented with metastatic disease at diagnosis, suggesting, in line with the present literature, that AdCC is mainly a locoregional disease . Furthermore, similar to other reports, most patients were treated with curative intent, and the majority were treated multimodally with surgery and PORT . We then investigated DFS and OS taking into account the whole cohort and independent of treatment modality; similar to others, we also showed that neither gender, age, or smoking influenced DFS, although some studies have shown that older age is associated with higher disease stage and consequently poorer DFS . OS was, however, worse in older patients, both in our cohort as well as in others, which was not really unexpected since these patients were older . Notably, however, we found that AdCC in the major salivary gland subsite and lower disease stage were prognostic factors, and these data are presented in more detail below upon also discussing treatment modality. Combinational treatment (e.g., surgery and PORT) has been suggested to be important for achieving good control of AdCC in general , and both locoregional and distant relapses were shown to be less common in patients treated with a combination of surgery and PORT . In our cohort, this was not different, and we also reported that multimodal treatment correlated significantly with fewer recurrences of any kind. In line with the current literature, during the present observation period, 53/142 (37.3%) patients presented recurrent disease of any kind after curative treatment . Recurrences were most common in the group receiving only RT or CRT (85.7%) and least common in the group treated with surgery and PORT (33.0%). A large American cohort study compared surgery alone to surgery and PORT and found similar to us that the latter is superior and significantly improves survival . They, however, suggested that the role of PORT needs to be studied further in low-risk patients. Here we showed that surgery alone can be considered in low-stage disease (Stage I-II) when PORT is contraindicated for any reason. Nevertheless, in our cohort, surgery as a single treatment modality was only used in patients with limited disease, but it did accomplish very good local control (one LRR) with no distant disease (10 patients). Induction chemotherapy (ICT) is, to our knowledge, not commonly used in the treatment of AdCC of the head and neck region and was not often reported in the studies we have cited above . In line with the literature, only five patients in this cohort received ICT; in four cases, a combination of Cisplatin and Fluorouracil and, in one case, even Docetaxel was added. In these cases, the primaries originated from different subsites (parotid-, submandibular gland, oropharynx, oral-, and nasal cavity); four had a stage IV disease, and one case was an early-stage disease. Three patients were even operated on for the primary tumour, and they all received CRT after ICT. ICT did not seem to improve survival in our patients since 3/5 patients presented recurrent disease within two years, while one died of secondary cancer; however, one patient with an early-stage AdCC in the parotid gland is still alive after seven years. Interestingly, we could not show that radical surgery (negative surgical margins) could be used as a possible prognostic marker, as suggested in other studies . In fact, in our cohort, we could not show any significant difference in either DFS or OS with regard to if patients underwent radical surgery or not. The latter suggested that PORT successfully treats the possible residual disease after surgery and confirmed its importance for good and long disease control. The fact that patients with tumour stages I and II had longer survival than those with tumour stages III and IV was not unexpected since this is usually the case for most solid tumours, including AdCC . The most important part of staging is the size of the tumour (T-stage) , as AdCC rarely primarily spreads to the lymph nodes . We confirmed that smaller tumours had a better prognosis and that staging should be used as a clinical prognostic factor in AdCC of the head and neck region similar to that for many other tumours as well as AdCC . We also showed that subsites could be used as a prognostic factor since major salivary gland AdCC had a significantly better prognosis than AdCC in other subsites, and notably, especially AdCC within the parotid gland had the best prognosis. This has previously been shown in some cohorts but not in all studies since, in other studies, subsite did not influence the prognosis . In this AdCC cohort, there was a large proportion of primaries within the major salivary glands (57% of patients treated with curative intent), which is not always the case in the literature . Moreover, they also had a better prognosis, which is in line with some reports, but different from some other studies . Additionally, we showed that AdCC arising in the nasal cavity and paranasal sinuses had the least favourable DFS; however, this did not, to the same extent, influence the OS. These patients were more often diagnosed with advanced disease and had the highest rate of locoregional relapses, which can successfully be treated with, e.g., re-irradiation, achieving good control of the disease. Similar data have also been shown in other cohorts . Of note, the frequency of perineural growth in our cohort was in line with that shown previously by others . Moreover, perineural growth was most common in the major salivary glands (especially the submandibular gland), where AdCC had a more favourable outcome compared to other sites in the head and neck. However, the perineural invasion was not linked to a higher disease stage, as one could have expected. Others have shown that perineural growth is associated with positive surgical margins and linked to a higher frequency of LRR or poorer prognosis in general . Some have also shown that perineural invasion, especially of larger nerves, is linked to distant metastases and, thereby, to a less favourable prognosis . In contrast to previous findings, none of the latter was the case in our cohort; on the contrary, we found that both LRR and metastatic disease were overrepresented in the group of patients presenting without perineural invasion. Our findings are, therefore, in contrast with a meta-analysis conducted by Ju et al. that indicated that perineural invasion was strongly associated with poor DFS and OS . The reason for this discrepancy we do not presently know, but also others have shown similar findings . On the other hand, AdCC is a rare disease with many parameters that have not yet been resolved, so despite the fact that both cohorts are fairly large, the differences could be due to chance or other unknown factors. Notably, there are some limitations in this study. Firstly, although we have many patients, the study is still a retrospective one, and not all patients were followed for >15 years. Furthermore, patient treatment varied to some extent, although most patients received multimodal treatments. Finally, the follow-ups of the patients were not identical, which potentially could, in some cases, depending on the follow-up, affect the DFS data; still, this is also the case for many other cohorts reported by others. However, this was not the case for the OS data, which were derived through the Swedish death registry. Other limitations are that we do not include studies of immunohistological or molecular biomarkers in this study, of which some have been described before to be of specific value and where some could be of use for targeted therapies . More specifically, e.g., recurrent rearrangements in MYB or MYBL1 genes, overexpression of MYB-NFIB transcripts, or mutations that activate the NOTCH pathway can be of value for prognostication and/or specific therapies . Moreover, the identification of different immunological markers could also be of value . These are all important issues that require specific thorough investigations that we need to follow up on. We have, however, previously investigated the potential role of both human papillomaviruses (HPV) and polyomaviruses in AdCC and could not show that they were associated with the development of AdCC . Moreover, finding HPV in tumours resembling AdCC implied that the diagnosis was not AdCC . In summary, we confirmed that AdCC is more common in female patients; it appears in adult patients, most often in their fifth decade, but can even appear in younger and elderly aged patients. The strongest prognostic factors are the stage of the disease, as well as the subsite, with the best prognosis of AdCC in the major salivary glands. Furthermore, multimodal treatment was superior to single-treatment modalities. Contrary to what others have shown, we could not find any significant correlation in survival in regard to perineural invasion and radical surgery. Furthermore, we found that other common prognostic factors, such as smoking status, age, and gender, often applicable for other cancer types, should not be used as prognostic factors in the AdCC of the head and neck. This study underlined the need to in the future find new molecular prognostic markers and additional treatment options for this disease and identify patients with poor prognosis upfront since, so far, systemic oncological treatments do not show sufficiently promising results . 5. Conclusions In this retrospective study, we confirmed that AdCC was more common in female patients, often occurring in the fifth decade, although it can appear in younger and older ages. In our cohort, the strongest favourable prognostic factors were early disease stage (stage I and II) and major salivary gland subsite, with the best prognosis in the AdCC of the parotid gland. Furthermore, multimodal treatment was superior to single-treatment modalities. In contrast to some other reports, we did not report any significant correlation between perineural invasion, radical surgery, and survival. However, in addition, similar to other studies, we confirmed that other common prognostic factors, such as smoking status, age, and gender, should not be used as prognostic factors in the AdCC of the head and neck region. Supplementary Materials The following supporting information can be downloaded at: Supplementary Table S1. 1-, 5-, 10-, and 15-year disease free survival (DFS) and overall survival (OS). Supplementary Table S2. Univariable analysis of disease-free survival (DFS) and overall survival (OS) in patients treated with curative intent. Supplementary Figure S1. Overall survival (OS) of patients treated with curative intent separated for treatment modality, multimodal treatment (Yes) and surgery or RT w/o ChT (No). Supplementary Figure S2. Overall survival (OS) of patients treated with curative intent separated for treatment modality, surgery + PORT and surgery + postoperative CRT. Supplementary Figure S3. Overall survival of patients treated with curative intent independent of treatment modality, depending on gender (A), age (B) and smoking (C). Supplementary Figure S4. Overall survival of patients with tumours staged I-IV (A) or with tumours staged I and II vs. III and IV (B). Supplementary Figure S5. Multivariable analysis of overall survival (OS) in patients treated with curative intent. Click here for additional data file. Author Contributions Conceptualization, M.Z. and S.F.; methodology, M.Z. and S.F.; software, A.B., M.Z. and S.F.; validation, M.Z., S.F., T.D. and A.B.; formal analysis, A.B. and M.Z.; investigation, M.Z. and S.F.; resources, S.F.; data curation, M.Z., A.B., A.N., T.D. and S.F.; writing--original draft preparation, M.Z. and T.D.; writing--review and editing, M.Z., A.B., A.N., T.D. and S.F.; visualization, A.B. and M.Z.; supervision, S.F., T.D. and A.N.; project administration, S.F. and T.D.; funding acquisition, S.F. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by ethical permissions 99-237, 2005/431-31/4, 2009/1278-31/4, 2012/83-31/2, 2017/1035-31/2, 2019-05211 and 2022-05287-02 from the Ethics Committee at Karolinska Institutet, Stockholm, Sweden, the Stockholm Regional Ethical Review Board, and the Swedish Ethical Review Authority. Informed Consent Statement Informed consent was obtained from all subjects involved in the study, according to the ethical permissions stated above. Data Availability Statement The data presented in this study are available on request from the corresponding author, but cannot be made publicly available due to Swedish laws on personal confidential information. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Disease-free survival (A) and overall survival (B) of all patients treated with curative intent independent of treatment modality. Figure 2 Disease-free survival of patients treated with curative intent separated for treatment modality, multimodal treatment (Yes) and surgery or RT w/o ChT (No). Figure 3 Disease-free survival of patients treated with curative intent separated for treatment modality, surgery + PORT and surgery + postoperative CRT. Figure 4 Disease-free survival of patients treated with curative intent independent of treatment modality, depending on gender (A), age (B), and smoking (C). Figure 5 Disease-free survival (A) and overall survival (B) of patients with tumours with perineural growth (Yes, Orange) and without perineural growth (No, Blue). Figure 6 Disease-free survival of patients with tumours staged I-IV (A) or with tumours staged I and II vs. III and IV (B). Figure 7 Disease-free survival (A) and overall survival (B) in patients treated with curative intent, independent of treatment modality, in relation to most common tumour subsites. Tumour subsites: --- Parotid Gland; --- Submandibular Gland; --- Nasal Cavity and Paranasal Sinuses; --- Oral Cavity. Figure 8 Disease-free survival (A) and overall survival (B) in patients treated with curative intent, independent of treatment modality, in relation to all tumour subsites, separated for major salivary glands in blue and all other sites together in orange. Figure 9 Multivariable analysis of DFS in patients treated with curative intent. cancers-15-01499-t001_Table 1 Table 1 Patients and tumour characteristics, entire cohort. Total (%) Parotid Gland (%) Submandibular Gland (%) Nasal Cavity & Paranasal Sinuses (%) Oral Cavity (%) Other Sites (%) Number of patients 155 42 42 24 27 20 Gender Female 96 (61.9) 30 (71.4) 26 (61.9) 14 (58.3) 15 (55.6) 11 (55) Male 59 (38) 12 (28.6) 16 (38) 10 (41.7) 12 (44.4) 9 (45) Median Age 60 53.5 62.5 52.5 61 67 [IQR] Stage I 34 (21.9) 11 (26.2) 9 (21.4) 0 6 (22.2) 8 (40) (AJCC 8th Edition) II 44 (28.4) 15 (35.7) 14 (33.3) 4 (16.7) 8 (29.6) 3 (15) III 21 (13.5) 4 (9.5) 12 (28.6) 3 (12.5) 0 2 (10) IV 54 (34.8) 12 (28.6) 5 (11.9) 17 (70.8) 13 (48.1) 7 (35) Unknown 2 (1.3) 0 2 (4.8) 0 0 0 Perineural Growth Yes 100 (64.5) 29 (69) 32 (76.2) 16 (66.7) 16 (59.3) 7 (35) No 52 (33.5) 12 (28.6) 10 (23.8) 8 (33.3) 10 (37) 12 (60) Unknown 3 (1.9) 1 (2.4) 0 0 1 (3.7) 1 (5) Smoking Ever 81 (52.3) 20 (47.6) 24 (57.1) 15 (62.5) 15 (55.6) 7 (35) Never 72 (46.5) 22 (52.4) 16 (38.1) 9 (37.5) 12 (44.4) 13 (65) Unknown 2 (1.3) 0 2 (4.8) 0 0 0 cancers-15-01499-t002_Table 2 Table 2 Patients treated with curative intent and tumour characteristics. Total (%) Parotid Gland (%) Submandibular Gland (%) Nasal Cavity & Paranasal Sinuses (%) Oral Cavity (%) Other Sites (%) Number of patients 142 41 38 21 25 17 Gender Female 88 (62) 30 (73.2) 23 (60.5) 12 (57.1) 13 (52) 10 (58.8) Male 54 (38) 11 (26.8) 15 (39.5) 9 (42.9) 12 (48) 7 (41.2) Median Age [IQR] 58.5 53 59.5 53 61 66 Age Group <58.5 71 (50) 25 (61) 18 (47.4) 14 (66.7) 9 (36) 5 (29.4) >=58.5 71 (50) 16 (39) 20 (52.6) 7 (33.3) 16 (64) 12 (70.6) Stage I 34 (23.9) 11 (26.8) 9 (23.7) 0 (0) 6 (24) 8 (47.1) II 43 (30.3) 15 (36.6) 14 (36.8) 4 (19) 7 (28) 3 (17.6) III 20 (14.1) 4 (9.8) 11 (28.9) 3 (14.3) 0 (0) 2 (11.8) IV 45 (31.7) 11 (26.8) 4 (10.5) 14 (66.7) 12 (48) 4 (23.5) Perineural Growth Yes 95 (66.9) 28 (68.3) 31 (81.6) 13 (61.9) 16 (64) 7 (41.2) No 46 (32.4) 12 (29.3) 7 (18.4) 8 (38.1) 9 (36) 10 (58.8) Unknown 1 (0.7) 1 (2.4) 0 (0) 0 (0) 0 (0) 0 (0) Smoking ever 75 (52.8) 19 (46.3) 22 (57.9) 13 (61.9) 14 (56) 7 (41.2) never 67 (47.2) 22 (53.7) 16 (42.1) 8 (38.1) 11 (44) 10 (58.8) Metastasis at Diagnosis no 140 (98.6) 40 (97.6) 38 (100) 21 (100) 25 (100) 16 (94.1) yes 2 (1.4) 1 (2.4) 0 (0) 0 (0) 0 (0) 1 (5.9) Surgery no 14 (9.9) 5 (12.2) 0 (0) 2 (9.5) 3 (12.0) 4 (23.5) yes 128 (90.1) 36 (87.8) 38 (100) 19 (90.5) 22 (88) 13 (76.5) Radical Surgery no 97 (75.8) 30 (83.3) 28 (73.7) 15 (78.9) 19 (86.4) 5 (38.5) yes 31 (24.2) 6 (16.7) 10 (26.3) 4 (21.1) 3 (13.6) 8 (61.5) RT no 10 (7) 2 (4.9) 3 (7.9) 1 (4.8) 0 (0) 4 (23.5) yes 132 (93.0) 39 (95.1) 35 (92.1) 20 (95.2) 25 (100) 13 (76.5) CRT no 105 (73.9) 34 (82.9) 30 (78.9) 11 (52.4) 18 (72) 12 (70.6) yes 37 (26.1) 7 (17.1) 8 (21.1) 10 (47.6) 7 (28) 5 (29.4) Induction ChT no 137 (96.5) 40 (97.6) 37 (97.4) 20 (95.2) 24 (96) 16 (94.1) yes 5 (3.5) 1 (2.4) 1 (2.6) 1 (4.8) 1 (4) 1 (5.9) Surgery + RT/CRT No 24 (16.9) 7 (17.1) 3 (7.9) 3 (14.3) 3 (12.0) 8 (47.1) Yes 118 (83.1) 34 (82.9) 35 (92.1) 18 (85.7) 22 (88.0) 9 (52.9) Surgery + CRT 29 (24.6) 5 (14.7) 8 (22.9) 8 (44.4) 5 (22.7) 3 (33.3) Surgery + RT 89 (75.4) 29 (85.3) 27 (77.1) 10 (55.6) 17 (77.3) 6 (66.7) Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000644 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051005 foods-12-01005 Review Fermentation for Designing Innovative Plant-Based Meat and Dairy Alternatives Boukid Fatma 1* Hassoun Abdo 23* Zouari Ahmed 4 Tulbek Mehmet Caglar 5 Mefleh Marina 6 Ait-Kaddour Abderrahmane 7 Castellari Massimo 8 Kowalczewski Przemyslaw Lukasz Academic Editor 1 ClonBio Group Ltd., 6 Fitzwilliam Pl, D02 XE61 Dublin, Ireland 2 Univ. Littoral Cote d'Opale, UMRt 1158 BioEcoAgro, USC ANSES, INRAe, Univ. Artois, Univ. Lille, Univ. Picardie Jules Verne, Univ. Liege, Junia, F-62200 Boulogne-sur-Mer, France 3 Sustainable AgriFoodtech Innovation & Research (SAFIR), F-62000 Arras, France 4 LRGP, UMR 7274 CNRS--Universite de Lorraine, 2 Avenue de La Foret de Haye TSA, 40602, F-54518 VANDOEUVRE, France 5 Saskatchewan Food Industry Development Centre, Saskatoon, SK S7M 5V1, Canada 6 Department of Soil, Plant and Food Science (DISSPA), University of Bari Aldo Moro, 70126 Bari, Italy 7 Universite Clermont Auvergne, INRAE, VetAgro Sup, UMRF, F-63370 Lempdes, France 8 Institute of Agriculture and Food Research and Technology (IRTA), Food Industry Area, Finca Camps i Armet s/n, 17121 Monells, Spain * Correspondence: [email protected] (F.B.); [email protected] (A.H.) 27 2 2023 3 2023 12 5 100506 2 2023 13 2 2023 17 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Fermentation was traditionally used all over the world, having the preservation of plant and animal foods as a primary role. Owing to the rise of dairy and meat alternatives, fermentation is booming as an effective technology to improve the sensory, nutritional, and functional profiles of the new generation of plant-based products. This article intends to review the market landscape of fermented plant-based products with a focus on dairy and meat alternatives. Fermentation contributes to improving the organoleptic properties and nutritional profile of dairy and meat alternatives. Precision fermentation provides more opportunities for plant-based meat and dairy manufacturers to deliver a meat/dairy-like experience. Seizing the opportunities that the progress of digitalization is offering would boost the production of high-value ingredients such as enzymes, fats, proteins, and vitamins. Innovative technologies such as 3D printing could be an effective post-processing solution following fermentation in order to mimic the structure and texture of conventional products. plant proteins precision fermentation food innovation safety health and nutrition digitalization This research received no external funding. pmc1. Introduction Since the Neolithic times, fermentation has been a natural process passed down through generations to produce several types of foods and beverages. Fermentation was primarily used for food preservation and shelf-life extension. Nowadays, there are a large variety of fermented animal and plant products made using a wide range of raw materials, microorganisms, and manufacturing techniques . Fermented meat products such as salami, ham, and sausages have traditionally been produced all around the world and currently occupy a special position in the gastro-economic trade of meat products . Fermented dairy products including cheese and yoghurt are staples known to contain potentially probiotic microorganisms such as lactic acid bacteria . Fermentation has also been applied to plant-based sources such as coffee, bread, chocolate, wine, and olives in order to improve their nutritional value, aroma and taste, texture, and stability . Overall, fermented foods are a crucial part of the human diet owing to their health benefits and particular flavor, aroma, and texture . As a part of the shift towards plant-based diets and alternative proteins, dairy and meat alternatives have become increasingly popular around the world. Cereals, legumes, oil seeds, nuts, and vegetables are the main sources used to make alternative products . These plant-based sources have lower nutritional quality than animal proteins and might induce undesirable flavors. Furthermore, plant proteins have different compositions and structures than animal proteins resulting in different functional features, including solubility, gelling, emulsifying, and foaming . Thus, one-to-one replacement of animal proteins with plant proteins to achieve a similar texture and mouthfeel to that of traditional products could be challenging. Among the strategies to recreate a meat/dairy-like experience, manufacturers rely on the use of additives such as fats, starches, flavorings, colorings, and stabilizers . Several plant-based products are perceived to be less nutritious (i.e., high in sugar, salt, and fat and low in protein contents) compared to animal products. Furthermore, the excessive use of additives is not appreciated by many consumers seeking natural and clean-labelled ingredients. Alternatively, manufacturers rely on processing for protein functionalization, for example, of textured vegetable proteins, to create a meat-like fibrous structure . However, most of the resulting products could be classified as ultra-processed foods, i.e., the group 4 of NOVA classification (a classification system that groups all foods according to the nature, extent, and purposes of the industrial processes they undergo) . The healthiness of ultra-processed foods is a controversial debate because of potential health issues such as cancer, obesity, and cardiovascular diseases . Therefore, fermentation has been playing a pivotal role in recent years to make plant-based alternatives to meat/dairy with minimal additives/processing . There are three types of fermentation used in plant-based products. Traditional fermentation intends to produce traditional foods and beverages. Various microorganisms (mainly bacteria and yeasts) are involved in the fermentation of plant-based raw materials . These microorganisms may be indigenously present on the substrate or added as a starter culture, or they may be present in the ingredient(s) . Depending on the raw material properties, the used microorganism, process conditions and substrate composition, fermentation can induce several changes in the organoleptic and nutritional properties of fermented plant-based ingredients/foods. Several pieces of recent research have underlined the potential of fermentation to improve the sensory and nutritional quality of plant-based fermented products . Biomass fermentation aims to obtain single cell proteins (SCP). Recently, precision fermentation emerged as a novel targeted technology for food applications with the aim of providing high-value compounds . This novel technology leverages metabolic engineering tools to serve as a factory of ingredients such as protein, pigments, vitamins, and fats to upgrade the quality of plant-based alternatives . Furthermore, fermentation is important for sustainable food solutions and can provide a positive impact on the sustainability index for the food industry . Compared to conventional protein sources, biomass fermentation for protein production can rapidly produce a very high protein content among other nutrients. Furthermore, food by-products and wastes can be used as substrates to be transformed in high-value food and feed products . This is an environmentally friendly strategy that could encourage a more energy-efficient and sustainable economy. Within this framework, this review aims to provide a better understanding of the opportunities and challenges related to the use of fermentation for developing plant-based products. Firstly, the market landscape and segmentation of fermented plant-based products is addressed with a focus on dairy and meat alternatives. Secondly, the impact of fermentation on the sensory and nutritional characteristics of plant-based dairy and meat products is discussed. Finally, a section is dedicated to capture the role of digitalization in facing the current and forthcoming challenges of fermented plant-based foods. 2. Market Landscape of Conventional and Innovative Plant-Based Foods Made Using Fermentation The global market of plant-based products made using fermentation was valued at USD 329.29 million in 2021 and is expected to reach USD 422.26 million by 2026, with a compound annual growth rate (CAGR) of 5.0% . Figure 2 illustrates the evolution of the number of the products launched in the global market during the last two decades (2002-2022). From 2002 to 2012, the number of fermented plant-based products increased, but at a low speed. During the period 2013-2022, this market witnessed an exponential growth and reached a peak in 2021 owing to the expansion of fermented plant-based meat and dairy alternatives , and it is expected to keep growing in the forthcoming years. The outbreak of COVID-19 contributed to fueling consumer interest in fermented plant-based products, owing to their health benefits such as boosting immune system performance and ameliorating gut health and inflammatory responses . The global fermented plant-based market is highly fragmented, including conventional (e.g., bakery) and emerging (i.e., alternative to meat and dairy) products. Overall, bakery is the largest fragment, followed by dairy alternatives, sauces and seasonings, meat alternatives, and vegetables, as well as ready-to-eat meals (Table 1). Europe holds the largest share of the market of fermented plant-based products, with a 73% market share, followed by Asia-Pacific, North America, Latin America, the Middle East, and Africa (Table 2). Europe generated the highest revenue of USD 96.99 million in 2020, which is projected to reach USD 129.47 million by 2026 . In 2020, consumer interest in vegan (1.9%), vegetarian (3.1%), and flexitarian (57.1%) foods increased across Europe . Plant-based dairy and meat alternatives produced by fermentation account for 26.30% and 8.56%, respectively, of the total fermented plant-based foods sold on the market (Table 1). There is now a wide range of commercial plant-based dairy products. Even though plant-based "yoghurt" is still at an early stage, it has registered the highest demand in the market of fermented plant-based foods . Manufacturers have focused on diversifying their product portfolio through broadening their range of flavors, while using health-beneficial probiotic bacteria . The global plant-based "yoghurt" market was estimated at USD 2.02 billion in 2020, and it is expected to achieve significant growth in the coming years . Europe dominates ~50% of this market share . Regarding the fermented "cheese" market, the global market was valued at USD 2.70 million in 2019 and is expected to reach USD 4.58 billion by 2025, at a CAGR of 8.91% . Fermented plant-based "cheese" is available in different types (e.g., camembert, Roquefort, and feta), forms (shreds, blocks, and slices), and textures (hard, soft, creamy, and spreadable) to fit different uses . This industry is increasingly relying on fermentation technologies to produce high-quality products while reducing the use of starches and vegetable fats . Plant-based fermented dairy products are generally made by fermentation of aqueous extracts obtained from different raw plant-based materials . Based on Mintel GNPD, the most used fermented ingredients derive from coconut (n = 575), soy (n = 570), oat (n = 143), rice (n = 73), wheat (n = 55), nut (n = 42), bean (n = 27), and almond (n = 27). In most cases, the used ferments are not reported on the products' labels, apart from lactic acid bacteria (n = 409) and vegan bacteria culture (n = 10). Lactic acid fermentation of plant-based products is commonly applied for improving product palatability and nutritional and health-promoting quality . Products such as tempeh (fermented soybean) and fermented tofu were not included because they were not originally designed as meat alternatives . Recently, the use of precision fermentation is emerging to unlock the potential of alternative products through providing targeted ingredients with particular features to deliver a unique dairy/meat-like experience. According to GFI database accessed on 20 September 2022), 151 startups use fermentation in developing plant-based alternatives, where precision fermentation (n = 68, 45%) attracts the highest interest, followed by fermentation to obtain protein-rich biomass (n = 60, 40%) and classic fermentation (n = 23, 15%). These startups produce fermented ingredients (n = 60) as well as dairy (n = 59), meat (n = 138), seafood (n = 32), and egg (n = 13) alternatives. Most of these products are not available on supermarket shelves and are still under development. 3. Impact of Fermentation on the Quality of Plant Based-Dairy Alternatives 3.1. Plant-Based Beverages (Milk Alternatives) Plant-based beverages are water-soluble extracts of legumes, oil seeds, nuts, cereals, or pseudocereals, and have a similar appearance and consistency to milk . To date, there is no consensus definition and/or classification of plant-based beverages/drinks, and these are also referred to as alternatives/substitutes to milk/dairy or plant-based milks or dairy. Categorizing plant-based beverages as cow's milk substitutes is still debatable. Raw milk has been defined by the Commission Regulation (of the EU) No 605/2010 as milk secreted from the mammary glands such as animals and humans. Thus, plant-based beverages do not meet the definition of milk. Furthermore, several studies reported that plant-based sources are not nutritionally equivalent to milk . Therefore, plant-based beverages can be considered as a separate category offering fermented/non-fermented options to consumers. The main plant sources used for making plant-based beverages are soy, almond, coconut, oat, and rice . Due to the increased demand for this category, there is room for new sources such as lentils, peanuts, quinoa, lupin, and peas . The common process for manufacturing non-fermented beverages includes wet milling, filtration, formulation (the addition of ingredients), sterilization, homogenization, emulsification, and storage. This process can be adjusted by adding/replacing some operations to fit the specific features of the raw materials and avoid the deterioration of product quality . The process of fermented plant-based beverages follows the same steps of the non-fermented variants, but with the addition of two supplementary steps, conditioning (to reach the optimal temperature for the growth of the microorganisms) and fermentation (under specific conditions suitable for the used microorganism(s)) . Plant proteins has been shown to be efficient carriers of probiotics . Most plant-based beverages are fermented using lactic acid bacteria, mainly Lactobacillus (Lactobacillus spp., L. casei, L. helveticus, L. fermentum, L. reuteri, L. acidophilus, L. rhamnosus and L. johnsonii), Bifidobacterium (Bifidobacterium animalis ssp. Lactis), Streptococcus (Streptococcus thermophilus), and Enterococcus (Enterococcus faecium) . Plant-based substrates favor high viability of fermenting microorganisms, and consequently result in probiotic dairy-free products (>106 CFU/mL of lactic acid bacteria cocci) . The duration of fermentation of plant-based beverages is typically 12-24 h, depending on the raw plant-based material properties, the used microorganism(s) and the final product features . Excessive fermentation time (>24 h) may result in the formation of undesirable compounds that can negatively impact the nutritional and organoleptic qualities . The use of fermentation enhanced the nutritional value and palatability of plant-based beverages. Pectinases released during lactic fermentation by Lactobacillus and Streptococcus improved the content of proteins, the amino acid profile, and protein digestibility . It also contributed to the formation of peptides with bioactive activities, e.g., ACE-inhibitory, antioxidative, and anti-microbial activities . Starters such as Lactobacillus and Bifidobacteria enhanced minerals content and availability as well as antioxidant properties in beverages made with soy, chickpea, and red bean . In addition, fermentation promoted an increase in organic acid and short-chain fatty acids concentrations, which in turn enhanced the absorption and solubility of minerals (calcium, iron, and zinc) and vitamins . Lactic bacteria could synthesize vitamins that are naturally lacking or absent in plant sources, such as vitamin B and K . Additionally, the activity of b-glucosidase increased the isoflavone content, which contributed to enhancing the digestion of plant-based beverages . Lactic bacteria were found to be efficient in mitigating antinutrients such as stachyose, raffinose, phytate, oligosaccharides, tannins and protease inhibitors, which in turn increased the bioavailability of minerals . Similarly, raffinose was reduced by 60% in fermented moringa leaves and beetroot extract with Lactobacillus plantarum and Enterococcus hirae . Furthermore, fermentation increased antibacterial activity against Bacillus cereus, Escherichia coli, Listeria monocytogenes and Staphylococcus aureus . Radical scavenging activity and phenolic content as well as minerals (calcium an iron) improved after fermentation. C. vulgaris and soy extract fermented using Lactobacillus fermentum and Lactobacillus rhamnosus resulted in improved polyphenol content and dietary antioxidant capacity compared to fermented soy extract . Fermentation mitigated plant protein allergens resulted in the loss of the IgE-binding ability of epitopes . For instance, conglycinin (7S) and glycinin (11S) were found to be drastically reduced in soymilk fermented by Enterococcus faecalis VB43 . In addition, Lactiplantibacillus plantarum subsp. plantarum reduced allergens (b-conglycinin and glycinin) in chickpea-based beverages . The beany flavor related to n-hexanal and n-hexanol, acetate, isovalerate, and 2-methylbutyrate was dramatically reduced by Bacillus subtilis, Lactobacillus species and edible fungi (e.g., Naematelia aurantialba, Lycoperdon pyriforme, Phellinus igniarius and Agrocybe cylindracea) . This resulted in reduced bitterness in fermented soymilk and legumes beverages . On the other hand, fermentation favored the formation of desirable volatile flavors such as acetoin, diacetyl (2,3-butanedione), and acetaldehyde, depending on the used microorganism(s) . For instance, new and pleasant aromatic notes were perceived after fermenting lupin, such as cheesy aromas (using Lb. reuteri), fatty aromas (using Lb. brevis, Lb. delbrueckii), and roasted aromas (using Lb. amilolyticus, Lb. helveticus) . For texture, no thickening effect is required in milk-like beverages, and thus fermentation is primarily used to improve taste and flavor . Therefore, the key selection criterion of plant materials is their solubility, in order to avoid powdery mouthfeel . The formation of antimicrobial compounds against e.g., Bacillus cereus, Staphylococcus aureus, and Pseudomonas aeruginosa can contribute to extending shelf-life . 3.2. Spoonable Yoghurt-Like Products Plant-based yoghurt-like products are generally made by fermenting aqueous extracts obtained from oat, pea, cashew, almond, coconut, and soy . Unlike fermented beverages, the major challenge of plant-based "yoghurt" alternatives is recreating the viscous texture of dairy yoghurt. For this reason, commercial non-fermented plant-based "yoghurt" alternatives rely chiefly on the use of thickening agents (e.g., natural gums, proteins, starches, pectin, and agar) to reach the desired consistency and ensure product stability . However, adding additives might negatively impact the consumers' acceptance, as they might prefer clean-labelled products. In fermented plant-based yoghurt, S. thermophilus and L. delbrueckii subsp. Bulgaricus strains are the most-used starters, with additional optional species for enhancing the nutritional and/or organoleptic qualities. S. thermophilus and L. delbrueckii subsp. Bulgaricus were reported to efficiently change oat protein concentrates' structure to favor their aggregation, which resulted in increased consistency . Fermented oat-based yoghurt products are appreciated for their desirable flavor and texture . It was reported that the use of lactic bacteria producers of exopolysaccharide (e.g., Weissella confusa) might improve viscosity and mouthfeel, reaching similar features to those of conventional dairy . Weissella confusa increased the viscosity and the water-holding capacity of quinoa-based yoghurt . It also improved protein digestibility and resulted in high final viable cell counts (>109 CFU/mL) . Fermenting sprouted tiger nut tubers with Lactobacillus bulgaricus and Streptococcus thermophilus provided probiotic products with increased protein and amino acids contents, improved sensory attributes, and reduced levels of anti-nutritional compounds . 3.3. Plant-Based "Cheese" Alternatives Plant-based "cheese" analogues can be subdivided into two categories: those made using fermentation and those not. Non-fermented cheeses are the most available in the market, and they are made using vegetable oils/fats (e.g., sunflower, coconut or palm oils) and polysaccharides (e.g., starches, gums, and fibers) as the main ingredients . For vegetable oil-based "cheese" alternatives, the selection of the type of fat is crucial for determining the quality of the end product. For instance, fresh and ripened Edam-type "cheeses" prepared using palm oil had a similar fat content and texture to their dairy counterparts , while those made of sunflower oil showed high spreadability and low firmness . Overall, the use of fat enables manufacturers to imitate the texture and meltability of dairy cheese, but not the stretchiness and flow . Polysaccharides were used in the making of soft "cheeses" such as Mozzarella, and resulted in a low fat content, soft texture, and some semblance of stretching . In dairy products, stretchability is related to the weakening of non-covalent casein-casein interactions, which is not possible using plant-based proteins . This explains why most fermented commercial "cheese" alternatives have a low protein content, and manufacturers rely chiefly on fats and polysaccharides . A few studies have investigated other matrices, with protein content around 20%. A prototype of cheddar "cheese" made by including 30% of zein showed similar softness, stretchability, and meltability to the dairy type . Soy and cashew-based "cheese" alternatives also showed high protein and low fat contents, while being appreciated for their color and flavor . The use of plant-based fermentation in "cheese" is not yet widespread on the market . The impact of fermentation on plant-based dairy is summarized in Table 3. Plant-based extracts from cashew, soy, and nuts were fermented using lactic bacteria and then used for making "cheese" alternatives . A soy "cheese" spread made using lactic bacteria and the addition of glucono-d-lactone showed improved texture . The use of lactic bacteria and/or Geotrichum candidum improved the sensory properties of soy-based "cheese" in its fresh state and after ripening (10 degC for 28 days) . It was also reported that this "cheese" had better features than that made only using lactic bacteria . A prolonged fermentation (7 days) of soy protein isolates with L. helveticus strains enabled the formation of cheese flavors (3-methylbutanal, 2-methylbutanal and benzaldehyde). This result might be an opportunity to modulate fermentation time to obtain natural flavoring compounds . More matrices are being explored in the literature. Different pea protein isolates and olive oil emulsions were fermented with a commercial bacterial inoculum (VegaTM) . The authors found that optimal fermentation-induced pea protein gels can be produced with 10% protein content and 10% olive oil levels without compromising gel hardness. A peanut extract-based spread was prepared using probiotic microorganism Lactobacillus rhamnosus NCDC18 . The product showed acceptable appearance, yet no sensory characterization was performed. Fermenting flaxseed oil cake using a combination of lactic acid bacteria, Penicillium camemberti, and Geotrichum candidum enabled the manufacture of a camembert-like "cheese" with improved oil oxidative stability . Fermented rice milk was prepared using lactic acid bacteria and various coagulation agents (gelatin, xanthan gum, or agar). Although gelatin treatment enabled the best sensory scores, its animal origin can limit its use in vegan alternatives . Fermented cashew using quinoa starter inoculum (dominated by Pediococcus and Weissella) was used to make "cheese" alternatives. The nutritional quality was marginally changed, while allergenicity associated with cashew was drastically reduced. Moreover, a high viable bacterial count was recorded (108-109 CFU/g) . foods-12-01005-t003_Table 3 Table 3 Impact of fermentation on the quality of plant-based products. Substrate Starters Effects References Broad bean and chickpea beverages Streptococcus thermophilus, Lactobacillus delbrueckii subsp. Bulgaricus, and a mixture of Lactobacillus casei and XPL-1, which is a mixed culture containing Lactococcus lactis subsp. cremoris, Lactococcus lactis subsp. lactis, Leuconostoc species, Lactococcus lactis subsp. lactis biovar. Diacetylactis, and a Streptococcus thermophilus strain Improvement in antioxidants (AOX) content and viscosity Red bean beverage Streptococcus thermophilus TISTR 894 (ST), Lactobacillus plantarum 299 V, and Lactobacillus casei 388, as a single or a mixed culture fermentation Improvement in AOX content Bean (Phaseolus vulgaris) beverage 10 lactobacillus strains Decrease in saturated fat and increase in unsaturated fat. Barley:finger millet: moth bean Lactobacilli acidophilus and a probiotic bacterium Increase in polyphenol content Chickpea beverage Streptococcus thermophilus (ST), a co-culture of ST with Lactococcus lactis and a co-culture of ST with Lactobacillus plantarum Decrease in saturated fat, phytic acids and increase in minerals Soymilk Lactobacillus casei PLA5 Increase in b-glucosidase, minerals and AOX activity and decrease in polyphenols content Bean (Phaseolus vulgaris) Streptococcus thermophilus + Lactobacillus Bulgaricus subs Debulgaricus, Lactobacillus acidophilus La-5 + Bifidobacterium animalis Bb-12 + Streptococcus thermophilus, Lactobacillus rhamnosus yoba + Streptococcus thermophilus and Fiti, Lactobacillus rhamnosus GR1 + Streptococcus thermophilus Increase in B vitamins and decrease in verbascose, stachyose and raffinose Soymilk Lactobacillus rhamnosus and Lactobacillus casei Increase in b-glucosidase activity and aglycones Black soybean beverage Lactiplantibacillus plantarum WGK 4, Streptococcus thermophilus Dad 11, and Lactiplantibacillus plantarum Dad 13 Increase in AOX activity and aglycone content Soymilk Enterococcus faecalis VB43 Reduction in the immunoreactivity of soybean allergens Moringa leaves and beetroot extract drink Lactobacillus plantarum and Enterococcus hirae Reduction in reffinose by 60%, increase in antibacterial activity against pathogenes and improvement of radical scavenging activity and phenolic content as well as minerals C. vulgaris and soy extract Lactobacillus fermentum and Lactobacillus rhamnosus Increase in polyphenol content and dietary antioxidant capacity Chickpea beverage Lactiplantibacillus plantarum subsp. plantarum Reduction in the immunoreactivity of chickpeas proteins Soymilk Lycoperdon pyriforme Decrease in the green off-flavor Pea protein isolate drink Lactobacillus Plantarum Reduction in the off-flavor VOC (aldehydes and ketones) Soybean beverage Naematelia aurantialba Increase in AOX activity, nutrient content and decrease in the oddly flavored VOC Oat-based "yoghurt" S. thermophilus and L. delbrueckii subsp. Bulgaricus Improvement in the texture and flavor Quinoa-based "yoghurt" Weissella confusa Improvement in the viscosity Soy-based "cheese" Lactic bacteria and/or Geotrichum candidum Improvement in the sensorial properties Pea protein isolate "cheese" Lactobacillus plantarum, perolens, fermentum, casei, Leuconostoc mesenteroids subsp. Cremoris and Pedicoccus pentasaceus Increase in cheesy aroma, acid and salty and reduction in the immunoreactivity of allergenic proteins Flaxseed oil-based "cheese" Penicillium camemberti and Geotrichum candidum Production of camembert-like cheese with good oil oxidative stability Cashew-based "cheese" Quinoa starter inoculum (dominated by Pediococcus and Weissella) Reduction in allergenicity associated with cashew and increase in the viable bacterial count 4. Impact of Fermentation on the Quality of Plant-Based Meat In conventional fermented meat products, lactic bacteria (Staphylococcus carnosus and Staphylococcus xylosus) can accelerate the degradation of proteins and fats to produce flavor compounds, enhance palatability, develop compact meat quality, and extend shelf-life (by inhibiting the growth of food-borne pathogenic bacteria such as Escherichia coli and Enterobacteriaceae) . The most used lactic bacteria were Lactobacillus spp., such as L. plantarum, L. sake, L. paracasei, and L. fermentum . Lactic acid bacteria were also used for the fermentation of plant-based protein ingredients to produce plant-based meat alternatives. These bacteria (Lactobacillus plantarum P1, Lactobacillus brevis R, Lactobacillus acidophilus 336, and Lactobacillus acidophilus 308) improved water/oil holding capacities and reduced protein oxidation of soy press cake . The inclusion of fermented soy products (10%) in meat alternative formulations improved texture (by increasing juiciness) and flavor (by reducing bitterness and balancing taste) . Edible fungi species (Lentinus edodes, Coprinus comatus and Pleurotus ostreatus) were used as ingredients in making fermented sausages. Among the different species, extruded Coprinus comatus and soybean protein showed improved functionality and fibrous structure. The resulting meat alternatives had desirable physicochemical and textural properties, taste, and flavor. Regarding the aroma profile, the curing and fermenting process contributed to the increased volatile compounds' contents, while fermented sausages without curing showed undesired flavors . Textured vegetable proteins (made by extruding soy protein, corn starch and wheat gluten) were fermented using B. subtilis, which resulted in improving the chewiness, hardness, integrity index, and layered structure . Biomass fermentation was used in producing QuornTM brand products accessed on 20 September 2022). The process relied on a continuous fermentation of glucose from roasted barley malt and nitrogen from ammonia by an edible fungi (Fusarium venenatum) . Mycoprotein exhibited organoleptic properties resembling meat, but with longer shelf-life and lower fat content . Furthermore, mycoprotein contains all essential amino acids, and has a protein digestibility corrected amino acid score (PDCAAS) of 0.99 . Yeast biomass could be incorporated into alternative meat formulations to improve their flavor . Particularly, yeast extracts from Saccharomyces cerevisiae have been widely used as flavoring agents in many meat products . They are also commonly used in plant-based meat products to impart meat flavor and umami taste. Commercial yeast extract products include Marmite(r) and Vegemite(r), which are by-products of fermentation. Torula yeast (Candida utilis), obtained using a continuous fermentation process, can be also added as a flavoring agent owing to its natural smoky umami flavor. Precision fermentation enabled the production of targeted ingredients that could benefit the mimicking of meat products. Soy leghemoglobin has been produced by an engineered yeast Pichia pastoris to give the flavor and color of animal meat to plant-based burgers (Impossible Foods) . Since it is a genetically edited ingredient, there were considerable concerns over its safety . A recent metanalysis of the literature showed that foods containing recombinant soy leghemoglobin are unlikely to present an unacceptable risk of allergenicity or toxicity to consumers . HemamiTM (Motif FoodWorks, Boston, MA, USA) is another yeast-derived heme protein. This product delivers an umami flavor and meaty aroma which likely can improve consumers' sensory perception of plant-based products. Precision fermentation is also used to make fats with similar molecular structures as their animal-derived counterparts . Melt & Marble has developed a precision fermentation-derived beef fat alternative via yeasts. Significant amounts of vitamin B12 were produced using co-fermentation of Propionibacterium freudenreichii and Lactobacillus brevis in wheat bran . This vitamin can be used to fortify plant-based meat products since it is naturally absent in plant-based sources and exclusively found in animal products. Precision fermentation-derived enzymes could be used as post-processors to address the functional limitations of plant proteins. Established enzymes manufacturers can lead this sector through using their expertise to develop a custom portfolio of enzymes able to overcome plant protein challenges. Computational biology "omics" and process engineering would contribute to screening and identifying the potential existing variants and designing new variants with new functionalities. 5. Role of Digitalization in the Innovation of Fermented Plant-Based Dairy and Meat Alternative Products As the plant-based market expands, so does progress in innovation and technological advances in many food-related sectors. Recent technological innovations have been driven by the emergence of the fourth industrial revolution (Industry 4.0) and its advanced technologies such as artificial intelligence (AI), big data, and the Internet of Things (IoT), among others . Many publications have shown that digital technologies and other Industry 4.0 innovations could provide tremendous opportunities to improve food quality and traceability, and boost food sustainability . Attempts to create innovative solutions for healthier diets with alternative proteins have been accelerated with the advent of digital technologies and other advanced related innovations. It was reported that the application of digital technologies, such as AI, smart sensors, robotics, and augmented reality, during fermentation can improve the monitoring and performance of the process . For example, throughout the fermentation process of rice wine, multiple parameters such as temperature, humidity, percentage of sugar and alcohol, and acidity can be measured using IoT, allowing manufacturers to virtually monitor the whole fermentation process online . Incorporation of Industry 4.0 technologies into fermentation facilities, "Fermentation 4.0" has recently been discussed, highlighting its potential to solve relevant problems such as the implementation of complex culture conditions . The implementation of such advanced technologies to achieve automatic detection and control of beer fermentation was recently reviewed and thoroughly discussed . These technologies will enable a better understanding of the fermentation process (classic/biomass) and thus an advanced control of the process to reach desired texture/viscosity. Sophisticated and automated processes, such as automated computer vision would help in modulating process conditions to maximize the production and the quality. Monitoring the process will offer a better understanding of the synergy among the bacteria/fungi used and how they interact within different media (different raw materials unlike milk). Three-dimensional (3D) cameras and hyperspectral imaging (HSI) would enable real-time monitoring, and thus the process can be adjusted and optimized in real-time during processing . For example, in a recent study, HSI was used to predict and quantify the total acid content and reducing sugar content of fermented grains . The results showed that HSI can be used to monitor the fermented grains' process in a rapid and non-destructive manner compared to traditional methods. Fermentation contributes to improving flavor through the formation of volatile compounds, thus advanced analytical tools for their identification/quantification are deemed crucial to modulate the process/ingredients/microorganisms to reach desirable sensory properties such as aroma and flavor. To this end, omics technologies and bioinformatics tools, in addition to AI and big data, are being increasingly investigated . Omics techniques applied during the production of vegetable-fermented foods and beverages could help us to determine and quantify microbial composition, understand the metabolic and functional properties of the microbial communities, detect changes associated with their development, and identify the metabolites that they produce . High-throughput analytical techniques, such as advanced mass spectrometry, chromatography, and nuclear magnetic resonance spectroscopy, are being used to select optimum fermentation conditions, measuring all the enzymes and metabolites produced by the microbes . The application of digital technologies and other advanced innovations has been demonstrably efficient in upgrading the quality and safety of fermented foods and beverages . Nevertheless, similar applications could be expected in the near future in the sector of fermented plant-based dairy and meat alternative products. Recent advances in Industry 4.0 technologies have enabled vast progress in precision fermentation due to recent advances in AI, bioinformatics, and systems and computational biology . Precision fermentation scales up the production of plant-based ingredients that will require advanced technologies for monitoring and optimizing the process. Such advances could offer quality standardization by detecting any potential anomalies in the fermentation process (e.g., mutation), and stable productivity, and thus could be more cost-effective food production methods. Advances in precision fermentation are expected to be key elements in the future to target taste and texture, enhance the shelf-life of plant-based fermented food products, and to mimic their animal counterparts . Additive manufacturing (or 3D printing) is among the emerging technological alternatives to traditional food production methods. 3D food printing and its derivatives (i.e., 4D, 5D, and 6D printing) have experienced a rapid evolution over the last few years, revolutionizing many aspects of the food industry . According to our search inquiry on the Scopus database, there has been a significant increase in the number of publications and citations reporting on the application of 3D printing in the food sector over the last decade . 3D food printing integrates digital gastronomy with additive manufacturing technology . Applications of 3D printing in the food sector are gaining increasing relevance due to their high potential in the production of personalized food, the reduction of food waste, and time and energy savings among other benefits . The combination of 3D food printing with AI can increase exploration of novel protein sources from plants, insects, fungi, and algae . AI and machine learning can be applied to perform a thorough analysis of ingredients that can be used in the plant-based industry to produce alternative products with similar molecular structure, taste, and texture to products of animal origin . A wide variety of plant-based materials have been proven to be suitable for 3D food printing . For example, a recent study investigated the possibility of producing hybrid meat analogues prepared by printing pea protein isolate and chicken mince as plant and animal protein ingredients at different ratios . Although there is a transition toward more plant-based diets being more sustainable and environmentally friendly, the question of the ability of such a diet to fulfill nutritional requirements such as the composition of essential amino acids in proteins is still under debate. In a recent study, Conzuelo and co-authors developed a digital tool that can provide a combination of protein ingredients (e.g., soy products, microalgae, and press cakes) used in the formulation of nutritious dairy/meat analogues and snacks with a high-to-excellent protein quality . However, the organoleptic attributes and technological performance of the protein combinations were not addressed in this study. The use of 3D printing as a post-processing technique of fermented plant-based products might boost their quality to a higher level in terms of nutrition, taste, aspect, and color, as well as product stability. It was reported that 3D printing of processed dairy cheese enabled higher dimensional stability, color, casein retention rate, and a final porosity compared to traditional methods . The technique could be applied to make plant-based fermented "cheese" alternatives to guarantee a better structure, color, and taste. For instance, making plant-based hard "cheese" is still a challenging task, and thus the use of 3D printing might give the possibility of creating "cheese" alternatives with complex geometries. Similarly, extrusion-based 3D printing of meat alternatives can make 3D structures similar to animal products. Bacterial species are now being mixed with various bioinks to produce functional complex materials using 3D printing . These systems could use fermentation substrates to make probiotic products with tailored features to facilitate personalized nutrition . Further optimization of this technique would enable the production of novel 3D-printed fermented meat and dairy alternative products with improved nutritional and functional properties. Thus, combining 3D-food printing technology and biotechnology approaches would help us to achieve products with equivalent taste, nutrition, structure and flavor to those of animal products . As 3D printing is well adapted to create new foods and textures, it is expected that this revolutionary technology will be the next big thing in the next decades. Finding a solution to overcome the limitations of 3D printing technology, such as consumer acceptance, insufficient institutionalization, and lack of standardized food material, would boost its use . To summarize, the application of digital technologies in fermented plant-based dairy and meat alternative products is currently limited, but further innovations are expected to accelerate the development of these products. 6. Conclusions Fermentation is attracting plenty of attention as a green sustainable solution to design plant-based alternatives with similar features to those of traditional foods. The key advantages of traditional and biomass fermentation are the familiarity and versatility of microorganisms. These microorganisms are also of natural origin and sustainably processed, providing minimally processed ingredients/biomass/products with enhanced organoleptic properties and health benefits. The addition of probiotics to the manufacture of plant-based dairy alternatives ensured improved nutritional and organoleptic values and an improved shelf-life of the products (beverages, yoghurt, and cheese). More research is required to select adequate starter suitable for fermenting plant-based material to offer nutritionally balanced products with desirable taste and flavor. Precision fermentation could deliver targeted compounds that might upgrade the quality of plant-based foods to match that of animal products. However, its main challenges include consumer perception of genetically engineered products, scalability, and ethical and regulatory concerns. Consumer understanding of precision fermentation is still lacking, and thus raising awareness about innovative food technologies is required to avoid the gap between consumers and the science behind their foods. Close collaboration between different actors in the food value chain is needed to overcome other obstacles. In the future, qualitative and quantitative sensorial studies are required for a better understanding of consumer acceptance/rejection of precision fermentation. The digestibility of plant-based fermented meat and dairy products and its effects on the composition of gut microbiota are scarcely investigated. These studies are of high relevance for understanding the impact of fermentation on human health. More research is also required to harness the opportunities offered by digitalization and other Industry 4.0 technologies for healthier and more sustainable food. Author Contributions Conceptualization, F.B.; writing--original draft preparation, F.B. and A.H.; writing--review and editing, F.B., A.H., A.Z., M.C.T., M.M., A.A.-K. and M.C. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 The use of fermentation to make innovative plant-based products. Figure 2 Evolution of launches of fermented plant-based foods in the global market. A search was conducted on Mintel GNPD (Global New Products Database) on 15 September 2022. Date: from 1 January 2002, to 15 September 2022. The inclusion criteria of the search were as follows: the Super-Category matches Food and Drink, the Claims match one or more of [Vegan/No Animal Ingredients; Plant Based], and Ingredient Search matches Fermented as one of the Ingredients. A total of 4379 products were retrieved. Figure 3 Number of publications and citations/year on 3D food printing over the last decade. The search was done on 1 November 2022, using the following search query criteria on the Scopus database: TITLE (3D food printing) OR (Additive food manufacturing). foods-12-01005-t001_Table 1 Table 1 Fermented plant-based foods sold in the global market. Fermented Plant-Based Foods and Drinks Number of Products and Percentage 1 Foods Bakery 1371 (31.2%) Dairy alternatives 1152 (26.3%) * "Yoghurt" 420 (9.6%) * Hard "cheese" and semi-hard "cheese" 210 (4.8%) * Soft "cheese" and semi-soft "cheese" 188 (4.3%) * Processed "cheese" 141 (3.3%) * Ice cream 100 (2.3%) * Drinks 50 (1.1%) * Fresh cheese and cream cheese 35 (1%) * Margarine 8 (0.18%) Sauces and seasonings 467 (10.2%) Meat alternatives 378 (8.6%) Ready-to-eat meals 312 (7.1%) Vegetables 93 (2.1%) Drinks Nutritional drinks and other beverages 260 (5.9%) Carbonated soft drinks 109 (2.5%) Alcoholic beverages 40 (0.9%) Juice drinks 39 (0.9%) Sports and energy drinks 8 (0.2%) Ready-to-drink beverages 4 (0.1%) 1 Total of fermented plant-based products = 4379. A search was conducted on Mintel GNPD (Global New Products Database) on 15 September 2022. Date: from 1 January 1996, to 15 September 2022. The inclusion criteria of the search were as follows: the Super-Category matches Food and Drink, the Claims match one or more of [Vegan/No Animal Ingredients; Plant Based], and Ingredient Search matches Fermented as one of the Ingredients. A total of 4379 products were retrieved. foods-12-01005-t002_Table 2 Table 2 Landscape of the global market of fermented plant-based foods. Region Number (% of Total) 1 Top 10 Brands Europe 3128 (71%) Fentimans: Alpro; Tesco Finest; M & S Food; M & S The Bakery; Tesco; Asda Extra Special; BFree; Marks & Spencer; Sojasun Asia Pacific 729 (17%) Maggi; Pascual; East Bali Cashews; Javara; Lo Bros.; Prima Ham Try Veggie; Remedy Kombucha, Coles; Elle & Vire; Fentimans North America 299 (7%) Genuine Health; Field Roast Chao; Nuts for Cheese Naked & Saucy; BFree; Field Roast Chao Vegan Creamery; Health-Ade Pop; Booch; Brami; Hu Latin America 154 (4%) Nomoo; Ile de France; Milkaut; Liane; Mun; Augusta; Emporium Vida; Neptune; Nogurt; Soignon Middle East & Africa 69 (2%) Woolworths Food; Soignon; Carrefour; Fry's Special Vegetarian; Kefir Life; Moya; Vigo Kombucha; Woolworths; Fynbos Fine Foods; Herman Brot 1 Total of fermented plant-based products = 4379. A search was conducted on Mintel GNPD (Global New Products Database) on 15 September 2022. Date: from 1 January 1996, to 15 September 2022. The inclusion criteria of the search were as follows: the Super-Category matches Food and Drink, the Claims match one or more of [Vegan/No Animal Ingredients; Plant Based], and Ingredient Search matches Fermented as one of the Ingredients. A total of 4379 products were retrieved. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Owusu-Kwarteng J. Agyei D. Akabanda F. Atuna R.A. Amagloh F.K. Plant-Based Alkaline Fermented Foods as Sustainable Sources of Nutrients and Health-Promoting Bioactive Compounds Front. Sustain. Food Syst. 2022 6 197 10.3389/fsufs.2022.885328 2. Chakanya C. Arnaud E. Muchenje V. Hoffman L.C. Fermented Meat Sausages from Game and Venison: What Are the Opportunities and Limitations? J. Sci. Food Agric. 2020 100 5023 5031 10.1002/jsfa.9053 29635774 3. Bis-Souza C.V. Barba F.J. Lorenzo J.M. Penna A.L.B. Barretto A.C.S. New Strategies for the Development of Innovative Fermented Meat Products: A Review Regarding the Incorporation of Probiotics and Dietary Fibers Food Rev. Int. 2019 35 467 484 10.1080/87559129.2019.1584816 4. Dimidi E. Cox S.R. Rossi M. Whelan K. Fermented Foods: Definitions and Characteristics, Impact on the Gut Microbiota and Effects on Gastrointestinal Health and Disease Nutrients 2019 11 1806 10.3390/nu11081806 31387262 5. Mefleh M. Darwish A.M.G. Mudgil P. Maqsood S. Boukid F. Traditional Fermented Dairy Products in Southern Mediterranean Countries: From Tradition to Innovation Fermentation 2022 8 743 10.3390/fermentation8120743 6. Tangyu M. Muller J. Bolten C.J. Wittmann C. Fermentation of Plant-Based Milk Alternatives for Improved Flavour and Nutritional Value Appl. Microbiol. Biotechnol. 2019 103 9263 9275 10.1007/s00253-019-10175-9 31686143 7. Jakubczyk K. Kaldunska J. Kochman J. Janda K. Chemical Profile and Antioxidant Activity of the Kombucha Beverage Derived from White, Green, Black and Red Tea Antioxidants 2020 9 447 10.3390/antiox9050447 32455926 8. Sanlier N. Gokcen B.B. Sezgin A.C. Health Benefits of Fermented Foods Crit. Rev. Food Sci. Nutr. 2019 59 506 527 10.1080/10408398.2017.1383355 28945458 9. Boukid F. Rosell C.M. Rosene S. Bover-Cid S. Castellari M. Non-Animal Proteins as Cutting-Edge Ingredients to Reformulate Animal-Free Foodstuffs: Present Status and Future Perspectives Crit. Rev. Food Sci. Nutr. 2022 62 6390 6420 10.1080/10408398.2021.1901649 33775185 10. Boukid F. The Realm of Plant Proteins with Focus on Their Application in Developing New Bakery Products Advances in Food and Nutrition Research Zhou W. Gao J. Academic Press Cambridge, MA, USA 2021 Volume 99 in press 11. Boukid F. Castellari M. How Can Processing Technologies Boost the Application of Faba Bean (Vicia faba L.) Proteins in Food Production? eFood 2022 3 e18 10.1002/efd2.18 12. Boukid F. Lamri M. Dar B.N. Garron M. Castellari M. Vegan Alternatives to Processed Cheese and Yogurt Launched in the European Market during 2020: A Nutritional Challenge? Foods 2021 10 2782 10.3390/foods10112782 34829064 13. Boukid F. Castellari M. Veggie Burgers in the EU Market: A Nutritional Challenge? Eur. Food Res. Technol. 2021 247 2445 2453 10.1007/s00217-021-03808-9 34230809 14. Baune M.-C. Terjung N. Tulbek M.C. Boukid F. Textured Vegetable Proteins (TVP): Future Foods Standing on Their Merits as Meat Alternatives Futur. Foods 2022 6 100181 10.1016/j.fufo.2022.100181 15. Marron-Ponce J.A. Flores M. Cediel G. Monteiro C.A. Batis C. Associations between Consumption of Ultra-Processed Foods and Intake of Nutrients Related to Chronic Non-Communicable Diseases in Mexico J. Acad. Nutr. Diet. 2019 119 1852 1865 10.1016/j.jand.2019.04.020 31262695 16. Satija A. Bhupathiraju S.N. Spiegelman D. Chiuve S.E. Manson J.A.E. Willett W. Rexrode K.M. Rimm E.B. Hu F.B. Healthful and Unhealthful Plant-Based Diets and the Risk of Coronary Heart Disease in U.S. Adults J. Am. Coll. Cardiol. 2017 70 411 422 10.1016/j.jacc.2017.05.047 28728684 17. Julia C. Martinez L. Alles B. Touvier M. Hercberg S. Mejean C. Kesse-Guyot E. Contribution of Ultra-Processed Foods in the Diet of Adults from the French NutriNet-Sante Study Public Health Nutr. 2018 21 27 37 10.1017/S1368980017001367 28703085 18. Melini F. Melini V. Luziatelli F. Ficca A.G. Ruzzi M. Health-Promoting Components in Fermented Foods: An up-to-Date Systematic Review Nutrients 2019 11 1189 10.3390/nu11051189 31137859 19. Mefleh M. Faccia M. Natrella G. De Angelis D. Pasqualone A. Caponio F. Summo C. Development and Chemical-Sensory Characterization of Chickpeas-Based Beverages Fermented with Selected Starters Foods 2022 11 3578 10.3390/foods11223578 36429170 20. Ben-Harb S. Saint-Eve A. Panouille M. Souchon I. Bonnarme P. Dugat-Bony E. Irlinger F. Design of Microbial Consortia for the Fermentation of Pea-Protein-Enriched Emulsions Int. J. Food Microbiol. 2019 293 124 136 10.1016/j.ijfoodmicro.2019.01.012 30690292 21. Nakamura M. Kitamura Y. Kokawa M. Development of a Cheese-Type Food Using Rice Milk Food Sci. Technol. Res. 2016 22 605 609 10.3136/fstr.22.605 22. Chai K.F. Ng K.R. Samarasiri M. Chen W.N. Precision Fermentation to Advance Fungal Food Fermentations Curr. Opin. Food Sci. 2022 47 100881 10.1016/j.cofs.2022.100881 23. Williams R.A. Opportunities and Challenges for the Introduction of New Food Proteins Annu. Rev. Food Sci. Technol. 2021 12 75 91 10.1146/annurev-food-061220-012838 33400558 24. Terefe N.S. Recent Developments in Fermentation Technology: Toward the next Revolution in Food Production Food Engineering Innovations across the Food Supply Chain Elsevier Amsterdam, The Netherlands 2022 89 106 10.1016/B978-0-12-821292-9.00026-1 25. Cichonska P. Ziarno M. Legumes and Legume-Based Beverages Fermented with Lactic Acid Bacteria as a Potential Carrier of Probiotics and Prebiotics Microorganisms 2022 10 91 10.3390/microorganisms10010091 26. Finnigan T. Needham L. Abbott C. Mycoprotein: A Healthy New Protein With a Low Environmental Impact Sustainable Protein Sources Academic Press Cambridge, MA, USA 2016 305 325 10.1016/B978-0-12-802778-3.00019-6 27. Boukid F. Plant-Based Meat Analogues: From Niche to Mainstream Eur. Food Res. Technol. 2021 247 297 308 10.1007/s00217-020-03630-9 28. Boukid F. Pera J. Parlade J. Castellari M. Fungal Bioconversion of Brewery By-Products: Assessment of Fatty Acids and Sterols Profiles Qual. Assur. Saf. Crop. Foods 2022 14 202 211 10.15586/qas.v14i4.1198 29. Souza Filho P.F. Zamani A. Taherzadeh M.J. Edible Protein Production by Filamentous Fungi Using Starch Plant Wastewater Waste Biomass Valorization 2019 10 2487 2496 10.1007/s12649-018-0265-2 30. Sar T. Larsson K. Fristedt R. Undeland I. Taherzadeh M.J. Demo-Scale Production of Protein-Rich Fungal Biomass from Potato Protein Liquor for Use as Innovative Food and Feed Products Food Biosci. 2022 47 101637 10.1016/j.fbio.2022.101637 31. Chakraborty A. Bhowal J. Bioconversion of Jackfruit Seed Waste to Fungal Biomass Protein by Submerged Fermentation Appl. Biochem. Biotechnol. 2022 1 14 10.1007/s12010-022-04063-8 32. Research and Markets Fermented Plant-Based Alternatives Market--A Global and Regional Analysis: Focus on Applications, Products, Patent Analysis, and Country Analysis-Analysis and Forecast, 2019-2026 Available online: (accessed on 25 September 2022) 33. Mintel Emerging Trends in the Plant-Based Industry|Mintel.Com Available online: (accessed on 24 September 2022) 34. Fonseca S.C. Rivas I. Romaguera D. Quijal M. Czarlewski W. Vidal A. Fonseca J.A. Ballester J. Anto J.M. Basagana X. Association between Consumption of Fermented Vegetables and COVID-19 Mortality at a Country Level in Europe medRxiv 2020 10.1101/2020.07.06.20147025 35. Shahbazi R. Sharifzad F. Bagheri R. Alsadi N. Yasavoli-Sharahi H. Matar C. Anti-Inflammatory and Immunomodulatory Properties of Fermented Plant Foods Nutrients 2021 13 1516 10.3390/nu13051516 33946303 36. BIS Research Fermented Plant Based Alternatives Market Analysis and Forecast Upto 2026|BIS Research Available online: (accessed on 25 September 2022) 37. Statista Europe: Share of People Following a Vegetarian Diet, by Country|Statista Available online: (accessed on 24 September 2022) 38. Reports and Insights Plant-Based Yogurt Market Size, Share Report and Global Industry Analysis 2022-2028 Available online: (accessed on 24 September 2022) 39. Montemurro M. Pontonio E. Coda R. Rizzello C.G. Plant-Based Alternatives to Yogurt: State-of-the-Art and Perspectives of New Biotechnological Challenges Foods 2021 10 316 10.3390/foods10020316 33546307 40. Canon F. Maillard M.B. Famelart M.H. Thierry A. Gagnaire V. Mixed Dairy and Plant-Based Yogurt Alternatives: Improving Their Physical and Sensorial Properties through Formulation and Lactic Acid Bacteria Cocultures Curr. Res. Food Sci. 2022 5 665 676 10.1016/j.crfs.2022.03.011 35434647 41. GlobeNewsWire Vegan Cheese Market Growth Available online: (accessed on 17 October 2021) 42. Pua A. Tang V.C.Y. Goh R.M.V. Sun J. Lassabliere B. Liu S.Q. Ingredients, Processing, and Fermentation: Addressing the Organoleptic Boundaries of Plant-Based Dairy Analogues Foods 2022 11 875 10.3390/foods11060875 35327297 43. Boeck T. Sahin A.W. Zannini E. Arendt E.K. Nutritional Properties and Health Aspects of Pulses and Their Use in Plant-Based Yogurt Alternatives Compr. Rev. Food Sci. Food Saf. 2021 20 3858 3880 10.1111/1541-4337.12778 34125502 44. Demir H. Simsek M. Yildirim G. Effect of Oat Milk Pasteurization Type on the Characteristics of Yogurt LWT 2021 135 110271 10.1016/j.lwt.2020.110271 45. Mefleh M. Pasqualone A. Caponio F. Faccia M. Legumes as Basic Ingredients in the Production of Dairy-Free Cheese Alternatives: A Review J. Sci. Food Agric. 2022 102 8 18 10.1002/jsfa.11502 34453343 46. Penna Franca P.A. Duque-Estrada P. da Fonseca e Sa B.F. van der Goot A.J. Pierucci A.P.T.R. Meat Substitutes-Past, Present, and Future of Products Available in Brazil: Changes in the Nutritional Profile Futur. Foods 2022 5 100133 10.1016/j.fufo.2022.100133 47. Rasika D.M. Vidanarachchi J.K. Rocha R.S. Balthazar C.F. Cruz A.G. Sant'Ana A.S. Ranadheera C.S. Plant-Based Milk Substitutes as Emerging Probiotic Carriers Curr. Opin. Food Sci. 2021 38 8 20 10.1016/j.cofs.2020.10.025 48. Zakidou P. Paraskevopoulou A. Aqueous Sesame Seed Extracts: Study of Their Foaming Potential for the Preparation of Cappuccino-Type Coffee Beverages LWT-Food Sci. Technol. 2021 135 110258 10.1016/j.lwt.2020.110258 49. Chalupa-Krebzdak S. Long C.J. Bohrer B.M. Nutrient Density and Nutritional Value of Milk and Plant-Based Milk Alternatives Int. Dairy J. 2018 87 84 92 10.1016/j.idairyj.2018.07.018 50. Makinen O.E. Wanhalinna V. Zannini E. Arendt E.K. Foods for Special Dietary Needs: Non-Dairy Plant-Based Milk Substitutes and Fermented Dairy-Type Products Crit. Rev. Food Sci. Nutr. 2016 56 339 349 10.1080/10408398.2012.761950 25575046 51. Smith N.W. Dave A.C. Hill J.P. McNabb W.C. Nutritional Assessment of Plant-Based Beverages in Comparison to Bovine Milk Front. Nutr. 2022 9 1785 10.3389/fnut.2022.957486 36003838 52. Chavan M. Gat Y. Harmalkar M. Waghmare R. Development of Non-Dairy Fermented Probiotic Drink Based on Germinated and Ungerminated Cereals and Legume LWT-Food Sci. Technol. 2018 91 339 344 10.1016/j.lwt.2018.01.070 53. Amini R.K. Islam M.Z. Kitamura Y. Kokawa M. Utilization of Fermented Rice Milk as a Novel Coagulant for Development of Paneer (Soft Cheese) Foods 2019 8 339 10.3390/foods8080339 31408978 54. Canaviri Paz P. Janny R.J. Hakansson A. Safeguarding of Quinoa Beverage Production by Fermentation with Lactobacillus Plantarum DSM 9843 Int. J. Food Microbiol. 2020 324 108630 10.1016/j.ijfoodmicro.2020.108630 32305832 55. Verni M. Demarinis C. Rizzello C.G. Baruzzi F. Design and Characterization of a Novel Fermented Beverage from Lentil Grains Foods 2020 9 893 10.3390/foods9070893 32645950 56. Lopes M. Pierrepont C. Duarte C.M. Filipe A. Medronho B. Sousa I. Legume Beverages from Chickpea and Lupin, as New Milk Alternatives Foods 2020 9 1458 10.3390/foods9101458 33066342 57. Valero-Cases E. Cerda-Bernad D. Pastor J.J. Frutos M.J. Non-Dairy Fermented Beverages as Potential Carriers to Ensure Probiotics, Prebiotics, and Bioactive Compounds Arrival to the Gut and Their Health Benefits Nutrients 2020 12 1666 10.3390/nu12061666 32503276 58. Karovicova J. Kohajdova Z. Minarovicova L. Laukova M. Greifova M. Greif G. Hojerova J. Utilisation of Quinoa for Development of Fermented Beverages Potravin. Slovak J. Food Sci. 2020 14 465 472 10.5219/1323 59. Singh B.P. Bhushan B. Vij S. Antioxidative, ACE Inhibitory and Antibacterial Activities of Soy Milk Fermented by Indigenous Strains of Lactobacilli Legum. Sci. 2020 2 e54 10.1002/leg3.54 60. Demarinis C. Verni M. Pinto L. Rizzello C.G. Baruzzi F. Use of Selected Lactic Acid Bacteria for the Fermentation of Legume-Based Water Extracts Foods 2022 11 3346 10.3390/foods11213346 36359959 61. De Mendonca Brandao H.C.A.D.N.T. De Mendonca Brandao W.A.P.L.N.T. De Mendonca S.N.T.G. Felsner M.L. Probiotic Fermented Rice Extract Beverage: An Alternative Food for Lactose Intolerants and People Allergic to Bovine Milk and Soy Protein Braz. J. Food Technol. 2021 24 e2020119 10.1590/1981-6723.11920 62. Nawaz M.A. Tan M. Oiseth S. Buckow R. An Emerging Segment of Functional Legume-Based Beverages: A Review Food Rev. Int. 2020 38 1064 1102 10.1080/87559129.2020.1762641 63. Sethi S. Tyagi S.K. Anurag R.K. Plant-Based Milk Alternatives an Emerging Segment of Functional Beverages: A Review J. Food Sci. Technol. 2016 53 3408 3423 10.1007/s13197-016-2328-3 27777447 64. Garcia Arteaga V. Leffler S. Muranyi I. Eisner P. Schweiggert-Weisz U. Sensory Profile, Functional Properties and Molecular Weight Distribution of Fermented Pea Protein Isolate Curr. Res. Food Sci. 2021 4 1 10 10.1016/j.crfs.2020.12.001 33385169 65. Silva A.R.A. Silva M.M.N. Ribeiro B.D. Health Issues and Technological Aspects of Plant-Based Alternative Milk Food Res. Int. 2020 131 108972 10.1016/j.foodres.2019.108972 32247441 66. Undhad T. Hati S. Makwana S. Significance of Storage Study on ACE Inhibitory, Antioxidative, Antimicrobial Activities, and Biotransformation of Isoflavones of Functional Fermented Soy-Based Beverage J. Food Process. Preserv. 2021 45 e15062 10.1111/jfpp.15062 67. Vasilean I. Aprodu I. Garnai M. Munteanu V. Patrascu L. Preliminary Investigations into the Use of Amylases and Lactic Acid Bacteria to Obtain Fermented Vegetable Products Foods 2021 10 1530 10.3390/foods10071530 34359401 68. Naprasert J. Suttisansanee U. Kemsawasd V. Single And Mixed Lactic Acid Bacteria Culture Fermentation In Red Bean Milk For Development Of A Functional Beverage Malays. Appl. Biol. 2019 48 139 145 69. Aydar E.F. Tutuncu S. Ozcelik B. Plant-Based Milk Substitutes: Bioactive Compounds, Conventional and Novel Processes, Bioavailability Studies, and Health Effects J. Funct. Foods 2020 70 103975 10.1016/j.jff.2020.103975 70. Dubey M.R. Patel V.P. Probiotics: A Promising Tool for Calcium Absorption Open Nutr. J. 2018 12 59 69 10.2174/1874288201812010059 71. Kumari A. Angmo K. Monika S. Bhalla T.C. Functional and Technological Application of Probiotic L. casei PLA5 in Fermented Soymilk Int. Food Res. J. 2018 25 2164 2172 72. Ziarno M. Brys J. Parzyszek M. Veber A. Effect of Lactic Acid Bacteria on the Lipid Profile of Bean-Based Plant Substitute of Fermented Milk Microorganisms 2020 8 1348 10.3390/microorganisms8091348 32899320 73. Anino C. Onyango A. Imathiu S. Maina J. Effect of Lactic Acid Bacteria Starter Cultures on Vitamin and Oligosaccharide Composition of Milk Extracted from Three Common Bean (Phaselous vulgaris L.) Varieties J. Food Res. 2019 8 103 10.5539/jfr.v8n3p103 74. Leksono B.Y. Cahyanto M.N. Rahayu E.S. Yanti R. Utami T. Enhancement of Antioxidant Activities in Black Soy Milk through Isoflavone Aglycone Production during Indigenous Lactic Acid Bacteria Fermentation Fermentation 2022 8 326 10.3390/fermentation8070326 75. Hati S. Vij S. Singh B.P. Mandal S. b-Glucosidase Activity and Bioconversion of Isoflavones during Fermentation of Soymilk J. Sci. Food Agric. 2015 95 216 220 10.1002/jsfa.6743 24838442 76. Castro-Alba V. Lazarte C.E. Perez-Rea D. Carlsson N.G. Almgren A. Bergenstahl B. Granfeldt Y. Fermentation of Pseudocereals Quinoa, Canihua, and Amaranth to Improve Mineral Accessibility through Degradation of Phytate J. Sci. Food Agric. 2019 99 5239 5248 10.1002/jsfa.9793 31062366 77. Vanajakshi V. Vijayendra S.V.N. Varadaraj M.C. Venkateswaran G. Agrawal R. Optimization of a Probiotic Beverage Based on Moringa Leaves and Beetroot LWT-Food Sci. Technol. 2015 63 1268 1273 10.1016/j.lwt.2015.04.023 78. Csatlos N.-I. Simon E. Teleky B.-E. Szabo K. Diaconeasa Z.M. Vodnar D.-C. Ciont C. Pop O.-L. Development of a Fermented Beverage with Chlorella Vulgaris Powder on Soybean-Based Fermented Beverage Biomolecules 2023 13 245 10.3390/biom13020245 36830613 79. El Mecherfi K.E. Todorov S.D. De Albuquerque M.A.C. Denery-Papini S. Lupi R. Haertle T. De Melo Franco B.D.G. Larre C. Allergenicity of Fermented Foods: Emphasis on Seeds Protein-Based Products Foods 2020 9 792 10.3390/foods9060792 32560210 80. Biscola V. de Olmos A.R. Choiset Y. Rabesona H. Garro M.S. Mozzi F. Chobert J.M. Drouet M. Haertle T. Franco B.D.G.M. Soymilk Fermentation by Enterococcus Faecalis VB43 Leads to Reduction in the Immunoreactivity of Allergenic Proteins b-Conglycinin (7S) and Glycinin (11S) Benef. Microbes 2017 8 635 643 10.3920/BM2016.0171 28726509 81. Skrzypczak K. Jablonska-Rys E. Gustaw K. Slawinska A. Wasko A. Radzki W. Michalak-Majewska M. Gustaw W. Reinforcement of the Antioxidative Properties of Chickpea Beverages through Fermentation Carried out by Probiotic Strain Lactobacillus Plantarum 299v J. Pure Appl. Microbiol. 2019 13 1 12 10.22207/JPAM.13.1.01 82. Meinlschmidt P. Schweiggert-Weisz U. Eisner P. Soy Protein Hydrolysates Fermentation: Effect of Debittering and Degradation of Major Soy Allergens LWT-Food Sci. Technol. 2016 71 202 212 10.1016/j.lwt.2016.03.026 83. Zhu Y.Y. Thakur K. Feng J.Y. Cai J.S. Zhang J.G. Hu F. Wei Z.J. B-Vitamin Enriched Fermented Soymilk: A Novel Strategy for Soy-Based Functional Foods Development Trends Food Sci. Technol. 2020 105 43 55 10.1016/j.tifs.2020.08.019 84. Nedele A.K. Gross S. Rigling M. Zhang Y. Reduction of Green Off-Flavor Compounds: Comparison of Key Odorants during Fermentation of Soy Drink with Lycoperdon Pyriforme Food Chem. 2021 334 127591 10.1016/j.foodchem.2020.127591 32721838 85. Emkani M. Oliete B. Saurel R. Effect of Lactic Acid Fermentation on Legume Protein Properties, a Review Fermentation 2022 8 244 10.3390/fermentation8060244 86. Sun T. Jiang H. Yang K. Li X. Wang S. Yao H. Wang R. Li S. Gu Y. Lei P. Nutritional Function and Flavor Evaluation of a New Soybean Beverage Based on Naematelia Aurantialba Fermentation Foods 2022 11 272 10.3390/foods11030272 35159425 87. Shi Y. Singh A. Kitts D.D. Pratap-Singh A. Lactic Acid Fermentation: A Novel Approach to Eliminate Unpleasant Aroma in Pea Protein Isolates LWT 2021 150 111927 10.1016/j.lwt.2021.111927 88. Zheng Y. Fei Y. Yang Y. Jin Z. Yu B. Li L. A Potential Flavor Culture: Lactobacillus Harbinensis M1 Improves the Organoleptic Quality of Fermented Soymilk by High Production of 2,3-Butanedione and Acetoin Food Microbiol. 2020 91 103540 10.1016/j.fm.2020.103540 32539956 89. Schlegel K. Leidigkeit A. Eisner P. Schweiggert-Weisz U. Technofunctional and Sensory Properties of Fermented Lupin Protein Isolates Foods 2019 8 678 10.3390/foods8120678 31847102 90. Ahsan S. Khaliq A. Chughtai M.F.J. Nadeem M. Din A.A. Hlebova M. Rebezov M. Khayrullin M. Mikolaychik I. Morozova L. Functional Exploration of Bioactive Moieties of Fermented and Non-Fermented Soy Milk with Reference to Nutritional Attributesax J. Microbiol. Biotechnol. Food Sci. 2020 10 145 149 10.15414/jmbfs.2020.10.1.145-149 91. Wang S. Chelikani V. Serventi L. Evaluation of Chickpea as Alternative to Soy in Plant-Based Beverages, Fresh and Fermented LWT 2018 97 570 572 10.1016/j.lwt.2018.07.067 92. Aduol K.O. Onyango A.N. Imathiu S.M. Proximate, Microbial and Sensory Characteristics of Cowpea Milk Fermented with Probiotic Starter Cultures Eur. J. Agric. Food Sci. 2020 2 1 7 10.24018/ejfood.2020.2.4.65 93. Bruckner-Guhmann M. Vasil'eva E. Culetu A. Duta D. Sozer N. Drusch S. Oat Protein Concentrate as Alternative Ingredient for Non-Dairy Yoghurt-Type Product J. Sci. Food Agric. 2019 99 5852 5857 10.1002/jsfa.9858 31206178 94. Bruckner-Guhmann M. Benthin A. Drusch S. Enrichment of Yoghurt with Oat Protein Fractions: Structure Formation, Textural Properties and Sensory Evaluation Food Hydrocoll. 2019 86 146 153 10.1016/j.foodhyd.2018.03.019 95. Laaksonen O. Kahala M. Marsol-Vall A. Blasco L. Jarvenpaa E. Rosenvald S. Virtanen M. Tarvainen M. Yang B. Impact of Lactic Acid Fermentation on Sensory and Chemical Quality of Dairy Analogues Prepared from Lupine (Lupinus Angustifolius L.) Seeds Food Chem. 2021 346 128852 10.1016/j.foodchem.2020.128852 33476950 96. Klost M. Gimenez-Ribes G. Drusch S. Enzymatic Hydrolysis of Pea Protein: Interactions and Protein Fractions Involved in Fermentation Induced Gels and Their Influence on Rheological Properties Food Hydrocoll. 2020 105 105793 10.1016/j.foodhyd.2020.105793 97. Grasso N. Alonso-Miravalles L. O'Mahony J.A. Composition, Physicochemical and Sensorial Properties of Commercial Plant-Based Yogurts Foods 2020 9 252 10.3390/foods9030252 32110978 98. Shin J.S. Kim B.H. Kim H.S. Baik M.Y. Optimization of Pea Protein and Citrus Fiber Contents for Plant Based Stirred Soymilk Yogurt Using Response Surface Methodology Food Sci. Biotechnol. 2022 31 1691 1701 10.1007/s10068-022-01180-2 36313001 99. Kim H.J. Han M.J. The Fermentation Characteristics of Soy Yogurt with Different Content of D-Allulose and Sucrose Fermented by Lactic Acid Bacteria from Kimchi Food Sci. Biotechnol. 2019 28 1155 1161 10.1007/s10068-019-00560-5 31275715 100. Bruckner-Guhmann M. Banovic M. Drusch S. Towards an Increased Plant Protein Intake: Rheological Properties, Sensory Perception and Consumer Acceptability of Lactic Acid Fermented, Oat-Based Gels Food Hydrocoll. 2019 96 201 208 10.1016/j.foodhyd.2019.05.016 101. Zannini E. Jeske S. Lynch K. Arendt E.K. Development of Novel Quinoa-Based Yoghurt Fermented with Dextran Producer Weissella Cibaria MG1 Int. J. Food Microbiol. 2018 268 19 26 10.1016/j.ijfoodmicro.2018.01.001 29316448 102. Lorusso A. Coda R. Montemurro M. Rizzello C.G. Use of Selected Lactic Acid Bacteria and Quinoa Flour for Manufacturing Novel Yogurt-like Beverages Foods 2018 7 51 10.3390/foods7040051 29614769 103. Ogundipe O.O. Fasogbon B.M. Ogundipe F.O. Oredope O. Amaezenanbu R.U. Nutritional Composition of Non-Dairy Yogurt from Sprouted Tigernut Tubers J. Food Process. Preserv. 2021 45 e15884 10.1111/jfpp.15884 104. Myhan R. Jelinski T. Bialobrzewski I. Sadowska J. Jachimczyk E. The Effect of Milk Fat Substitution on the Rheological Properties of Edam-Type Cheese Eur. Food Res. Technol. 2020 246 2443 2450 10.1007/s00217-020-03587-9 105. Butt N.A. Ali T.M. Hasnain A. Development of Rice Starch-Based Casein and Fat Mimetics and Its Application in Imitation Mozzarella Cheese J. Food Process. Preserv. 2020 44 e14928 10.1111/jfpp.14928 106. Kamath R. Basak S. Gokhale J. Recent Trends in the Development of Healthy and Functional Cheese Analogues--A Review LWT 2022 155 112991 10.1016/j.lwt.2021.112991 107. Shabani J. Sarfarazi M. Mirzaei H. Jafari S.M. Influence of the Sunflower Oil Content, Cooking Temperature and Cooking Time on the Physical and Sensory Properties of Spreadable Cheese Analogues Based on UF White-Brined Cheese Int. J. Dairy Technol. 2016 69 576 584 10.1111/1471-0307.12305 108. Mattice K.D. Marangoni A.G. Evaluating the Use of Zein in Structuring Plant-Based Products Curr. Res. Food Sci. 2020 3 59 66 10.1016/j.crfs.2020.03.004 32914121 109. Karimi R. Azizi M.H. Ghasemlou M. Vaziri M. Application of Inulin in Cheese as Prebiotic, Fat Replacer and Texturizer: A Review Carbohydr. Polym. 2015 119 85 100 10.1016/j.carbpol.2014.11.029 25563948 110. Harper A.R. Dobson R.C.J. Morris V.K. Moggre G.J. Fermentation of Plant-Based Dairy Alternatives by Lactic Acid Bacteria Microb. Biotechnol. 2022 15 1404 1421 10.1111/1751-7915.14008 35393728 111. Lopusiewcz L. Smietana N. Lichwiarska E. Mazurkiewicz-Zapalowicz K. Gefrom A. Drozlowska E. The Biotransformation of Lupine Seeds by Lactic Acid Bacteria and Penicillium Camemberti into a Plant-Based Camembert Alternative, and Its Physicochemical Changes during 7 Weeks of Ripening Fermentation 2022 8 447 10.3390/fermentation8090447 112. Oyeyinka A.T. Odukoya J.O. Adebayo Y.S. Nutritional Composition and Consumer Acceptability of Cheese Analog from Soy and Cashew Nut Milk J. Food Process. Preserv. 2019 43 e14285 10.1111/jfpp.14285 113. Masia C. Jensen P.E. Petersen I.L. Buldo P. Design of a Functional Pea Protein Matrix for Fermented Plant-Based Cheese Foods 2022 11 178 10.3390/foods11020178 35053910 114. Jeewanthi R.K.C. Paik H.D. Modifications of Nutritional, Structural, and Sensory Characteristics of Non-Dairy Soy Cheese Analogs to Improve Their Quality Attributes J. Food Sci. Technol. 2018 55 4384 4394 10.1007/s13197-018-3408-3 30333634 115. Li Q. Xia Y. Zhou L. Xie J. Evaluation of the Rheological, Textural, Microstructural and Sensory Properties of Soy Cheese Spreads Food Bioprod. Process. 2013 91 429 439 10.1016/j.fbp.2013.03.001 116. Li Y. Zhang X. Yang J.J. Ma X.Y. Jia X.D. Du P. Li A.L. Influence of the Addition of Geotrichum Candidum on the Microbial, Chemical, Textural, and Sensory Features of Soft Soy Cheese J. Food Process. Preserv. 2020 44 e14823 10.1111/jfpp.14823 117. Shirotani N. Bygvraa Hougaard A. Lametsch R. Agerlin Petersen M. Rattray F.P. Ipsen R. Proteolytic Activity of Selected Commercial Lactobacillus Helveticus Strains on Soy Protein Isolates Food Chem. 2021 340 128152 10.1016/j.foodchem.2020.128152 33032150 118. Sharma P. Sharma D. Amin A. Development of a Functional Fermented Peanut-Based Cheese Analog Using Probiotic Bacteria Biotechnologia 2018 99 435 441 10.5114/bta.2018.79973 119. Lopusiewicz L. Drozlowska E. Tarnowiecka-Kuca A. Bartkowiak A. Mazurkiewicz-Zapalowicz K. Salachna P. Biotransformation of Flaxseed Oil Cake into Bioactive Camembert-Analogue Using Lactic Acid Bacteria, Penicillium Camemberti and Geotrichum Candidum Microorganisms 2020 8 1266 10.3390/microorganisms8091266 32825460 120. Chen J.M. Al K.F. Craven L.J. Seney S. Coons M. McCormick H. Reid G. O'connor C. Burton J.P. Nutritional, Microbial, and Allergenic Changes during the Fermentation of Cashew 'Cheese' Product Using a Quinoa-Based Rejuvelac Starter Culture Nutrients 2020 12 648 10.3390/nu12030648 32121191 121. Wang Z. Wang Z. Ji L. Zhang J. Zhao Z. Zhang R. Bai T. Hou B. Zhang Y. Liu D. A Review: Microbial Diversity and Function of Fermented Meat Products in China Front. Microbiol. 2021 12 1322 10.3389/fmicb.2021.645435 122. Wang Y. Han J. Wang D. Gao F. Zhang K. Tian J. Jin Y. Research Update on the Impact of Lactic Acid Bacteria on the Substance Metabolism, Flavor, and Quality Characteristics of Fermented Meat Products Foods 2022 11 2090 10.3390/foods11142090 35885333 123. Razavizadeh S. Alencikiene G. Vaiciulyte-Funk L. Ertbjerg P. Salaseviciene A. Utilization of Fermented and Enzymatically Hydrolyzed Soy Press Cake as Ingredient for Meat Analogues LWT 2022 165 113736 10.1016/j.lwt.2022.113736 124. Yuan X. Zhu X. Sun R. Jiang W. Zhang D. Liu H. Sun B. Sensory Attributes and Characterization of Aroma Profiles of Fermented Sausages Based on Fibrous-like Meat Substitute from Soybean Protein and Coprinus Comatus Food Chem. 2022 373 131537 10.1016/j.foodchem.2021.131537 34776312 125. Maung T.T. Gu B.Y. Kim M.H. Ryu G.H. Fermentation of Texturized Vegetable Proteins Extruded at Different Moisture Contents: Effect on Physicochemical, Structural, and Microbial Properties Food Sci. Biotechnol. 2020 29 897 907 10.1007/s10068-020-00737-3 32582452 126. Hyde K.D. Xu J. Rapior S. Jeewon R. Lumyong S. Niego A.G.T. Abeywickrama P.D. Aluthmuhandiram J.V.S. Brahamanage R.S. Brooks S. The Amazing Potential of Fungi: 50 Ways We Can Exploit Fungi Industrially Fungal Divers. 2019 97 1 136 10.1007/s13225-019-00430-9 127. Hashempour-Baltork F. Khosravi-Darani K. Hosseini H. Farshi P. Reihani S.F.S. Mycoproteins as Safe Meat Substitutes J. Clean. Prod. 2020 253 119958 10.1016/j.jclepro.2020.119958 128. Derbyshire E. Food-Based Dietary Guidelines and Protein Quality Definitions--Time to Move Forward and Encompass Mycoprotein? Foods 2022 11 647 10.3390/foods11050647 35267280 129. Kyriakopoulou K. Keppler J.K. van der Goot A.J. Functionality of Ingredients and Additives in Plant-Based Meat Analogues Foods 2021 10 600 10.3390/foods10030600 33809143 130. Campagnol P.C.B. dos Santos B.A. Wagner R. Terra N.N. Pollonio M.A.R. The Effect of Yeast Extract Addition on Quality of Fermented Sausages at Low NaCl Content Meat Sci. 2011 87 290 298 10.1016/j.meatsci.2010.11.005 21123002 131. Ismail I. Hwang Y.H. Joo S.T. Meat Analog as Future Food: A Review J. Anim. Sci. Technol. 2020 62 111 120 10.5187/jast.2020.62.2.111 32292920 132. Fraser R.Z. Shitut M. Agrawal P. Mendes O. Klapholz S. Safety Evaluation of Soy Leghemoglobin Protein Preparation Derived From Pichia Pastoris, Intended for Use as a Flavor Catalyst in Plant-Based Meat Int. J. Toxicol. 2018 37 241 262 10.1177/1091581818766318 29642729 133. Jin Y. He X. Andoh-Kumi K. Fraser R.Z. Lu M. Goodman R.E. Evaluating Potential Risks of Food Allergy and Toxicity of Soy Leghemoglobin Expressed in Pichia Pastoris Mol. Nutr. Food Res. 2018 62 1700297 10.1002/mnfr.201700297 28921896 134. Teng T.S. Chin Y.L. Chai K.F. Chen W.N. Fermentation for Future Food Systems: Precision Fermentation Can Complement the Scope and Applications of Traditional Fermentation EMBO Rep. 2021 22 e52680 10.15252/embr.202152680 33908143 135. Xie C. Coda R. Chamlagain B. Varmanen P. Piironen V. Katina K. Co-Fermentation of Propionibacterium Freudenreichiiand Lactobacillus Brevisin Wheat Bran for in Situproduction of Vitamin B12 Front. Microbiol. 2019 10 1541 10.3389/fmicb.2019.01541 31333632 136. Xie C. Coda R. Chamlagain B. Edelmann M. Deptula P. Varmanen P. Piironen V. Katina K. In Situ Fortification of Vitamin B12 in Wheat Flour and Wheat Bran by Fermentation with Propionibacterium Freudenreichii J. Cereal Sci. 2018 81 133 139 10.1016/j.jcs.2018.05.002 137. Hassoun A. Ait-kaddour A. Abu-mahfouz A.M. Rathod N.B. Bader F. Barba F.J. Cropotova J. Galanakis C.M. Jambrak A.R. Lorenzo M. The Fourth Industrial Revolution in the Food Industry--Part I: Industry 4.0 Technologies Crit. Rev. Food Sci. Nutr. 2022 1 17 10.1080/10408398.2022.2034735 138. Hassoun A. Cropotova J. Trif M. Rusu A.V. Bobis O. Nayik G.A. Jagdale Y.D. Saeed F. Afzaal M. Mostashari P. Consumer Acceptance of New Food Trends Resulting from the Fourth Industrial Revolution Technologies: A Narrative Review of Literature and Future Perspectives Front. Nutr. 2022 9 972154 10.3389/fnut.2022.972154 36034919 139. Hassoun A. Prieto M.A. Carpena M. Bouzembrak Y. Marvin H.J.P. Pallar N. Barba F.J. Punia S. Chaudhary V. Ibrahim S. Exploring the Role of Green and Industry 4.0 Technologies in Achieving Sustainable Development Goals in Food Sectors Food Res. Int. 2022 162 112068 10.1016/j.foodres.2022.112068 36461323 140. Hassoun A. Jagtap S. Garcia-Garcia G. Trollman H. Pateiro M. Lorenzo J.M. Trif M. Rusu A.V. Aadil R.M. Simat V. Food Quality 4.0: From Traditional Approaches to Digitalized Automated Analysis J. Food Eng. 2023 337 111216 10.1016/j.jfoodeng.2022.111216 141. Hassoun A. Alhaj Abdullah N. Ait-Kaddour A. Ghellam M. Besir A. Zannou O. Onal B. Aadil R.M. Lorenzo J.M. Mousavi Khaneghah A. Food Traceability 4.0 as Part of the Fourth Industrial Revolution: Key Enabling Technologies Crit. Rev. Food Sci. Nutr. 2022 10.1080/10408398.2022.2110033 35950635 142. Viejo C.G. Fuentes S. Editorial: Special Issue "Implementation of Digital Technologies on Beverage Fermentation" Fermentation 2022 8 127 10.3390/fermentation8030127 143. Vosahlik J. Hart J. Measurability of Quality in Fermentation Process of Rice Wine by IoT in the Field of Industry 4.0 Agron. Res. 2021 19 1318 1324 144. Alarcon C. Shene C. Fermentation 4.0, a Case Study on Computer Vision, Soft Sensor, Connectivity, and Control Applied to the Fermentation of a Thraustochytrid Comput. Ind. 2021 128 103431 10.1016/j.compind.2021.103431 145. Violino S. Figorilli S. Costa C. Pallottino F. Internet of Beer: A Review on Smart Technologies from Mash to Pint Foods 2020 9 950 10.3390/foods9070950 32709156 146. Ozdogan G. Lin X. Sun D.W. Rapid and Noninvasive Sensory Analyses of Food Products by Hyperspectral Imaging: Recent Application Developments Trends Food Sci. Technol. 2021 111 151 165 10.1016/j.tifs.2021.02.044 147. Caporaso N. Whitworth M.B. Fowler M.S. Fisk I.D. Hyperspectral Imaging for Non-Destructive Prediction of Fermentation Index, Polyphenol Content and Antioxidant Activity in Single Cocoa Beans Food Chem. 2018 258 343 351 10.1016/j.foodchem.2018.03.039 29655743 148. Jiang X. Tian J. Huang H. Hu X. Han L. Huang D. Luo H. Nondestructive Visualization and Quantification of Total Acid and Reducing Sugar Contents in Fermented Grains by Combining Spectral and Color Data through Hyperspectral Imaging Food Chem. 2022 386 132779 10.1016/j.foodchem.2022.132779 35349904 149. Rizo J. Guillen D. Farres A. Diaz-Ruiz G. Sanchez S. Wacher C. Rodriguez-Sanoja R. Omics in Traditional Vegetable Fermented Foods and Beverages Crit. Rev. Food Sci. Nutr. 2020 60 791 809 10.1080/10408398.2018.1551189 30582346 150. McClements D.J. Future Foods: How Modern Science Is Transforming the Way We Eat Springer Basel, Switzerland 2019 9783030129958 151. Gonzalez Viejo C. Harris N.M. Fuentes S. Quality Traits of Sourdough Bread Obtained by Novel Digital Technologies and Machine Learning Modelling Fermentation 2022 8 516 10.3390/fermentation8100516 152. Viejo C.G. Fuentes S. Low-Cost Methods to Assess Beer Quality Using Artificial Intelligence Involving Robotics, an Electronic Nose, and Machine Learning Fermentation 2020 6 104 10.3390/fermentation6040104 153. Hassoun A. Bekhit A.E. Jambrak A.R. Regenstein J.M. Chemat F. Morton J.D. Gudjonsdottir M. Carpena M. Prieto A. Varela P. The Fourth Industrial Revolution in the Food Industry--Part II: Emerging Food Trends Trends Crit. Rev. Food Sci. Nutr. 2022 1 31 10.1080/10408398.2022.2106472 35930319 154. Tyndall S.M. Maloney G.R. Cole M.B. Hazell N.G. Augustin M.A. Critical Food and Nutrition Science Challenges for Plant-Based Meat Alternative Products Crit. Rev. Food Sci. Nutr. 2022 1 16 10.1080/10408398.2022.2107994 35972071 155. Singh S. Yap W.S. Ge X.Y. Min V.L.X. Choudhury D. Cultured Meat Production Fuelled by Fermentation Trends Food Sci. Technol. 2022 120 48 58 10.1016/j.tifs.2021.12.028 156. Viejo C.G. Fuentes S. Hernandez-Brenes C. Smart Detection of Faults in Beers Using Near-Infrared Spectroscopy, a Low-Cost Electronic Nose and Artificial Intelligence Fermentation 2021 7 117 10.3390/fermentation7030117 157. Landerneau S. Lemarie L. Marquette C. Petiot E. Green 3D Bioprinting of Plant Cells: A New Scope for 3D Bioprinting Bioprinting 2022 27 e00216 10.1016/j.bprint.2022.e00216 158. Ghazal A.F. Zhang M. Mujumdar A.S. Ghamry M. Progress in 4D/5D/6D Printing of Foods: Applications and R & D Opportunities Crit. Rev. Food Sci. Nutr. 2022 1 24 10.1080/10408398.2022.2045896 159. Liu X. Le C. Yu J. Zhao L. Wang K. Tao Y. Renard C.M.G.C. Hu Z. Trends and Challenges on Fruit and Vegetable Processing: Insights into Sustainable, Traceable, Precise, Healthy, Intelligent, Personalized and Local Innovative Food Products Trends Food Sci. Technol. 2022 125 12 25 10.1016/j.tifs.2022.04.016 160. Teng X. Zhang M. Mujumdar A.S. 4D Printing: Recent Advances and Proposals in the Food Sector Trends Food Sci. Technol. 2021 110 349 363 10.1016/j.tifs.2021.01.076 161. Soni R. Ponappa K. Tandon P. A Review on Customized Food Fabrication Process Using Food Layered Manufacturing LWT 2022 161 113411 10.1016/j.lwt.2022.113411 162. Escalante-Aburto A. Santiago G.T. Alvarez M.M. Chuck-Hernandez C. Advances and Prospective Applications of 3D Food Printing for Health Improvement and Personalized Nutrition Compr. Rev. Food Sci. Food Saf. 2021 20 5722 5741 10.1111/1541-4337.12849 34643023 163. Le-Bail A. Maniglia B.C. Le-Bail P. Recent Advances and Future Perspective in Additive Manufacturing of Foods Based on 3D Printing Curr. Opin. Food Sci. 2020 35 54 64 10.1016/j.cofs.2020.01.009 164. Tomasevic I. Putnik P. Valjak F. Pavlic B. Sojic B. Bebek Markovinovic A. Bursac Kovacevic D. 3D Printing as Novel Tool for Fruit-Based Functional Food Production Curr. Opin. Food Sci. 2021 41 138 145 10.1016/j.cofs.2021.03.015 165. Bedoya M.G. Montoya D.R. Tabilo-Munizaga G. Perez-Won M. Lemus-Mondaca R. Promising Perspectives on Novel Protein Food Sources Combining Artificial Intelligence and 3D Food Printing for Food Industry Trends Food Sci. Technol. 2022 128 38 52 10.1016/j.tifs.2022.05.013 166. Ping Y. Feng X. Blank I. Liu Y. Strategies to Improve Meat-like Properties of Meat Analogs Meeting Consumers' Expectations Biomaterials 2022 287 121648 10.1016/j.biomaterials.2022.121648 35780575 167. Marinova D. Bogueva D. Alternative Proteins Food in a Planetary Emergency Springer Nature Singapore 2022 121 151 168. Wang M. Li D. Zang Z. Sun X. Tan H. Si X. Tian J. Teng W. Wang J. Liang Q. 3D Food Printing: Applications of Plant-Based Materials in Extrusion-Based Food Printing Crit. Rev. Food Sci. Nutr. 2022 62 7184 7198 10.1080/10408398.2021.1911929 33856247 169. Wang T. Kaur L. Furuhata Y. Aoyama H. Singh J. 3D Printing of Textured Soft Hybrid Meat Analogues Foods 2022 11 478 10.3390/foods11030478 35159628 170. Conzuelo Z.R. Robyr R. Kopf-bolanz K.A. Optimization of Protein Quality of Plant-Based Foods Through Digitalized Product Development Front. Nutr. 2022 9 902565 10.3389/fnut.2022.902565 35619962 171. Deng Y. Cheng K. Shao X. Dong X. Jiang H. Xiao G. Effects of Drying Method on the Stability and Quality of Post-Processing of 3D-Printed Processed Cheese Dry. Technol. 2022 1 8 10.1080/07373937.2022.2126494 172. Kyle S. 3D Printing of Bacteria: The Next Frontier in Biofabrication Trends Biotechnol. 2018 36 340 341 10.1016/j.tibtech.2018.01.010 29402473 173. Kewuyemi Y.O. Kesa H. Adebo O.A. Trends in Functional Food Development with Three-Dimensional (3D) Food Printing Technology: Prospects for Value-Added Traditionally Processed Food Products Crit. Rev. Food Sci. Nutr. 2021 62 7866 7904 10.1080/10408398.2021.1920569 33970701 174. Cabrera-Barjas G. Banerjee A. Valdes O. Moncada M. Sirajunnisa A.R. Surendhiran D. Ramakrishnan G. Rani N.S. Hamidi M. Kozani P.S. Food Biotechnology: Innovations and Challenges Future Foods: Global Trends, Opportunities, and Sustainability Challenges Bhat R. Academic Press Cambridge, MA, USA 2022 697 719 175. Lee J. A 3D Food Printing Process for the New Normal Era: A Review Processes 2021 9 1495 10.3390/pr9091495 |
PMC10000645 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12050972 foods-12-00972 Article Thiamine and Biotin: Relevance in the Production of Volatile and Non-Volatile Compounds during Saccharomyces cerevisiae Alcoholic Fermentation in Synthetic Grape Must Evers Marie Sarah Conceptualization Methodology Formal analysis Investigation Data curation Writing - original draft Writing - review & editing Visualization 12 Roullier-Gall Chloe Conceptualization Methodology Formal analysis Resources Data curation Writing - review & editing Visualization Supervision Project administration Funding acquisition 1 Morge Christophe Resources Writing - review & editing Project administration Funding acquisition 2 Sparrow Celine Resources Writing - review & editing Project administration Funding acquisition 2 Gobert Antoine Conceptualization Resources Writing - review & editing Supervision 2 Vichi Stefania Formal analysis Writing - review & editing 3 Alexandre Herve Conceptualization Methodology Resources Writing - review & editing Visualization Supervision Project administration Funding acquisition 1* Corona Onofrio Academic Editor 1 Institut Universitaire de la Vigne et du Vin Jules Guyot, UMR PAM, Universite de Bourgogne, 2 Rue Claude Ladrey, 21000 Dijon, France 2 Sofralab SAS, 79 Avenue Alfred Anatole Thevenet, 51530 Magenta, France 3 Food Science and Gastronomy Department, University of Barcelona, Nutrition, INSA (Institut de Recerca en Nutricio I Seguretat Alimentaria), 08921 Santa Coloma de Gramenet, Spain * Correspondence: [email protected] 24 2 2023 3 2023 12 5 97213 1 2023 17 2 2023 21 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Vitamins are major cofactors to numerous key metabolic pathways in enological yeasts, and both thiamine and biotin, notably, are believed to be essential to yeast fermentation and growth, respectively. In order to further assess and clarify their role in winemaking, and in the resulting wine, alcoholic fermentations of a commercial Saccharomyces cerevisiae active dried yeast were conducted in synthetic media containing various concentrations of both vitamins. Growth and fermentation kinetics were monitored and proved the essential character of biotin in yeast growth, and of thiamine in fermentation. The synthetic wine volatile compounds were quantified, and notable influences of both vitamins appeared, through a striking positive effect of thiamine on the production of higher alcohols, and of biotin on fatty acids. Beyond the evidence of this influence on fermentations and on the production of volatiles, this work proves, for the first time, the impact held by vitamins on wine yeasts' exometabolome, investigated through an untargeted metabolomic analysis. This highlighted chemical differences in the composition of synthetic wines through a notably marked influence of thiamine on 46 named S. cerevisiae metabolic pathways, and especially in amino acid-associated metabolic pathways. This provides, overall, the first evidence of the impact held by both vitamins on the wine. thiamine biotin vitamins Saccharomyces cerevisiae alcoholic fermentation wine volatiles metabolomics Association Nationale Recherche Technologie (ARNT)2019/0747 This research was funded by the Association Nationale Recherche Technologie (ARNT), grant number 2019/0747. pmc1. Introduction Wine is a highly complex and versatile drink, presenting a large variety of chemical compounds that derive from numerous sources during vinification and play significant roles in the sensory profiles and perception of the final product. As such, over 8000 volatile compounds have been estimated to be found in wines, a significant number of those resulting from fermentation processes . It has been shown that those generated by yeast metabolism during the alcoholic fermentation are affected by the type of fermenting yeast, as well as the fermentative conditions . As such, nutrient availability in the yeast cultivation medium stands as a significant factor regarding the production of volatile compounds during fermentation , susceptible of highly modifying the final aromatic profile in the resulting wine. Vitamins, consequently, being involved in numerous yeast metabolic pathways, such as those of fatty acids, amino acids, or carbohydrates' metabolisms, as well as those of sulfur and nitrogen , stand as highly significant nutritional compounds. Thus, detrimental effects resulting from their deficiencies in the yeast have been reported, such as impairments of the growth and fermentation kinetics and the production or accumulation of possibly detrimental metabolites . Amongst all water-soluble vitamins, biotin and thiamin have been distinguished as particularly significant regarding yeast metabolism: biotin, is, indeed, essential to yeast growth , while thiamine has been recognized as essential to fermentation processes . However, such investigations remain ancient, and there is no updated knowledge regarding an overall significance of the vitaminic nutrition of yeasts on the production of volatiles during alcoholic fermentation. The aim, here, is therefore to investigate the overall impact of both biotin and thiamin on the course of alcoholic fermentation, assessing their significance regarding kinetics and the production of volatile compounds in a synthetic grape must matrix. 2. Materials and Methods 2.1. Chemicals, Reagents, and Materials Cultivation medium: Glucose, fructose, malic acid, citric acid, potassium dihydrogen phosphate, potassium sulfate, magnesium sulfate heptahydrate, sodium chloride, ammonium chloride, tyrosine, tryptophane, isoleucine, aspartic acid, glutamic acid, arginine, leucine, threonine, glycine, glutamine, alanine, valine, methionine, phenylalanine, serine, histidine, lysine, cysteine, proline, manganese (II) sulfate monohydrate, copper (II) sulfate, cobalt (II) chloride hexahydrate, boric acid, ammonium molybdate, sodium hydroxide, myo-inositol, pantothenic acid hemi-calcium salt, thiamine hydrochloride, nicotinic acid, pyridoxine, biotin, ergosterol, oleic acid, and Tween 80 were purchased from Sigma (Merck, Darmstadt, Germany). Calcium chloride dihydrate and sodium bicarbonate were purchased from Honeywell (Charlotte, NC, USA). Zinc sulfate heptahydrate was purchased from Prolabo (VWR, Avantor, Radnor, PA, USA). Potassium iodide was purchased from Merck (Darmstadt, Germany). Flow cytometry: Phosphate buffer saline (PBS) was purchased from Sigma (Merck, Darmstadt, Germany). The 5-6-carboxyfluorescein diacetate (cFDA) was purchased from Molecular Probes (Thermo Fisher Scientific, Waltham, MA, USA). Headspace solid-phase microextraction coupled to gas chromatography (HS-SPME-GC/MS): Ethanol 96%, 1-propanol, 3-methylbutyl acetate, 3-methylbutanol, ethyl octanoate, ethyl decanoate, 1-hexanol, 2-phenylethyl alcohol, and octanoic acid were purchased from Sigma-Aldrich (St Louis, MO, USA). SPME fiber divinylbenzene/carboxen/polydimethylsiloxane 50/30 mm, 1 cm-long (DVB/CAR/PDMS), were from Supelco (Bellefonte, PA, USA). 2.2. Cultivation Medium A synthetic must MS300 was used for the fermentations, adapted from Bely and colleagues . The base medium, adjusted at pH 3.3, thus contained the following components (expressed per liter): glucose 100 g, fructose 100 g, DL malic acid 6 g, citric acid 6 g; mineral salts: KH2PO4 750 mg, KH2SO4 500 mg, MgSO4*7H2O 250 mg, CaCl2*2H2O 155 mg, NaCl 200 mg; nitrogen compounds: NH4Cl 460 mg, L-tyrosine 18 mg, L-tryptophane 179 mg, L-isoleucine 33 mg, L-aspartic acid 45 mg, L-glutamic acid 120 mg, L-arginine 374 mg, L-leucine 48 mg, L-threonine 76 mg, L-glycine 18 mg, L-glutamine 505 mg, L-alanine 145 mg, L-valine 45 mg, L-methionine 31 mg, L-phenylalanine 38 mg, L-serine 79 mg, L-histidine 33 mg, L-lysine 17 mg, L-cysteine 13 mg; trace elements: MnSO4*H2O 4 mg, ZnSO4*7H2O 4 mg, CuSO4*5H2O 1 mg, KI 1 mg, CoCl2*6H2O 0.4 mg, H3BO3 1 mg, (NH4)6Mo7O24 1 mg; anaerobic growth factors: ergosterol 3 mg, oleic acid 1 mg, Tween 80 100 mg; vitamins: myo-inositol 20 mg, pantothenic acid hemi-calcium salt 1.5 mg, nicotinic acid 2 mg, pyridoxine 0.25 mg. In order to evaluate the impact of both thiamine and biotin on the production of volatile compounds, this base medium was declined under nine different variations, relying on three different concentration modalities for each of the vitamins: a total absence of the given vitamin, a concentration fit to trigger deficiencies in yeasts , and the original concentration of the MS300 as described by Bely et al. . The nine different cultivation medium variations are described in Table 1. 2.3. Yeast and Inoculation Procedure A commercial strain of Saccharomyces cerevisiae "Selectys(r) La Marquise" (Sofralab, Magenta, France) was selected for fermentation during this experiment. The active dried yeast (ADY) was rehydrated according to the manufacturer's specifications, in 10 mL of mineral water, and left to rest for half an hour at 28 degC. The density of viable cells was then determined by flow cytometry on a BD AccuriTM C6 Plus (Becton, Dickinson and Company, Franklin Lakes, NJ, USA). The fluorophore used to detect the viable cells was 5-6-carboxyfluorescein diacetate (cFDA) dissolved in acetone at a final concentration of 1500 mM. Then, 3 mL of this solution was added to 100 mL of the yeast suspension diluted in PBS for the measurement. This yeast suspension was then taken and used to inoculate 200 mL of the selected synthetic must medium at an initial concentration of 106 viable cells/mL. 2.4. Fermentation Conditions and Monitoring All fermentations were carried out in MS300 synthetic must medium, according to the variations described in Section 2.2, previously sterilized by vacuum filtration on a 0.22 mm pored membrane-based apparatus (Millipore Steritop, Merck, Darmstadt, Germany). Erlenmeyer flasks of a 250 mL total volume, containing 200 mL of the sterile cultivation medium and covered with sterile cotton wool to avoid any cross-contaminations, were used for the conduction of the fermentations, and set at 28 degC without shaking. A volume of 500 mL of all samples was collected every 3 h during the first two days, except for the initial first 12 h (lag phase), during which no sampling was performed. Samplings were then collected at 24 h intervals after the completion of the exponential phase. All cultures were conducted in triplicate. The viable cell concentration was determined similarly to the procedure used for inoculation, thus allowing for the monitoring of biomass growth. Fermentation progress was monitored according to the procedure described by Seguinot and colleagues , by weighing of the Erlenmeyer flasks, with the weight loss being assimilated to the amount of CO2 released over the course of fermentation. As such, the maximal rate of CO2 production, or the maximal fermentation rate, was calculated by the derivation of CO2 production over time . 2.5. Analytical Methods Endpoint fermentation samples were centrifuged at 13,000 rpm for 3 min, and the collected supernatants were stored at -20 degC until analysis. 2.5.1. Enzymatic Determination of Central Carbon Metabolites Acetic acid, acetaldehyde, D-lactic acid, and glycerol were enzymatically quantified using a Y15 enzymatic autoanalyzer (Biosystems, Barcelona, Spain), calibrated and configured for the associated enzymatic kits (references 12930, 12820, 12812, 12801; Biosystems, Barcelona, Spain). Ethanol was determined by Fourier Transform Infrared (FTIR) spectroscopy with an OenoFoss type 4101 apparatus (FOSS Electric, Hillerod, Denmark). 2.5.2. HS-SPME-GC/MS Analysis of Volatile Compounds Extraction of volatiles was carried out by means of a Combi-pal autosampler (CTC Analytics, Zwingen, Switzerland). A volume of 2 mL of sample was placed into a 10 mL vial that was then fitted with a polytetrafluoroethylene (PTFE)/silicone septum and maintained under continuous agitation (250 rpm) at 40 degC during 10 min for sample conditioning. Then, a DVB/CAR/PDMS fiber was exposed to the sample headspace at the same temperature and stirring during 30 min and immediately desorbed in the gas chromatograph injector. Volatile compounds were analyzed by gas chromatography coupled to quadrupolar mass selective spectrometry using an Agilent 5973 Network detector (Agilent Technologies, Palo Alto, CA, USA). Analytes were separated on a Supelcowax-10 (Supelco), 60 m x 0.25 mm I.D., with a 0.25 mm film thickness. Column temperature was held at 40 degC for 10 min, increased to 150 degC at 3 degC/min, then to 250 degC at 15 degC/min, holding for 5 min. The injector temperature was 260 degC and the time of desorption of the fiber into the injection port was fixed at 5 min. Helium was the carrier gas, at a flow rate of 1.5 mL/min. The temperature of the ion source was 230 degC and of the transfer line was 280 degC. Electron impact mass spectra were recorded at 70 eV ionization energy, 5.1 scan/s. GC-MS analysis was performed in the complete scanning mode (SCAN), in the 35-300 u mass range. Identification of compounds was carried out by comparison of their mass spectra and retention times with those of standard compounds or with those available in the mass spectrum library Wiley 6 and in the literature, respectively. Response factors of volatile compounds were calculated using a calibration curve, obtained by analyzing a hydroalcoholic solution (ethanol 10%, v/v) with different concentrations of reference compounds. 2.5.3. UHPLC-Q-ToF-MS Non-Targeted Analysis UHPLC-Q-ToF-MS non-targeted analyses of the yeast metabolome were performed using an ultra-high-pressure liquid chromatography (UHPLC) Dionex Ultimate 3000 apparatus (Thermo Fisher Scientific, Waltham, MA, USA) coupled to a MaXis plus MQ ESI-Q-ToF mass spectrometer (MS) (Bruker, Bremen, Germany). Non-polar compounds were eluted in reverse phase on an Acquity BEH C18 1.7% m, 100 x 2.1 mm column (Waters, Guyancourt, France), using mobile phases 0.1% MS grade formic acid in ultrapure water (Milli-Q, Merck, Darmstadt, Germany) (eluent A) and 0.1% formic acid in 95% MS grade acetonitrile (Biosolve, Dieuze, France) (eluent B). The elution was performed as a gradient at a flow rate of 400 mL/min at 40 degC, with its initial 5% eluent B phase being held for 1.10 min, before increasing to reach 95% eluent B at 6.40 min until the end of the elution after a total 10 min analytical run. The nebulizer pressure was set at 2 bar, the nitrogen dry gas flowing through the apparatus at 10 L/min, while ionization was performed as electrospray in either positive or negative mode. Ion transfer parameters were set at 500 V as endplate offset, and 3500 or 4500 V as capillary voltage for negative and positive ionization mode, respectively. The mass spectrometer system was calibrated prior to analyses using sodium formate in enhanced quadratic mode (errors < 0.5 ppm). Mass ranges stood between 100 and 1000 m/z, while the stability of the UHPLC-Q-TOF-MS-MS system was ensured through the analysis of quality controls, prepared as a mix of all the analyzed samples, at the beginning, the end, and every ten samples during the analysis. All the samples were randomly analyzed, and the injection of the sodium formate calibrant at the beginning of each run allowed recalibrations of the spectrum. Software Bruker Compass MetaboScape (v. 8.0.1, Bruker, Mannheim, Germany) was used for pre-treatment analysis comprising: the internal mass recalibration of the spectrum, extraction of molecular features (m/z couple, retention time), and alignment and annotation using SmartFormula (isotopic profile). Feature extraction has been realized for an absolute intensity higher than 1000 and the presence of the feature in more than 20% of the total number of samples. Recursive extraction has been used to find features without an intensity threshold if it presents in more than 20% of samples. SmartFormula annotation has been realized with the criteria: mass error < 10 ppm and mSigma < 20. Isolated significant features were assigned possible annotations using the MassTRIX application , and its associated databases, such as KEGG , as accessed on 11 November 2022. 2.6. Statistical Analyses Data analyses were performed with a statistical treatment and graphically (PCA biplots, chromatograms, boxplots) represented using the R software (version 4.0.3) (R foundation, Vienna, Austria). Nonparametric Scheirer-Ray-Hare (chosen as a two-way ANOVA equivalent) and Dunn (relying on a Bonferroni correction) tests from the rcompanion and rstatix package were used for statistical comparison of the growth and fermentation kinetics data, as well as of the samples resulting from the quantification of volatiles and carbon metabolites in wines. A p-value < 0.05 was used as the acceptance criterion for rejecting the null hypothesis. PCA and one-way ANOVA were used for statistical analysis of the metabolomics datasets, at confidence levels of a = 5% and a = 1%, respectively. Heatmaps in the volatilome and exometabolome investigations were drawn for biotin and thiamine on all their respective significantly affected compounds or features. 3. Results and Discussion 3.1. Impact on the Growth and Fermentation Kinetics Both thiamine and biotin were evaluated for their ability to support the growth of S. cerevisiae in MS300 synthetic must. Significant differences in growth kinetics were observed due to the initial concentration of biotin present in the cultivation medium. Three growth parameters were assessed during the experiment: the maximal specific growth rate (mmax) attained during the exponential phase, the generation time (G), the maximal population reached during cultivation (ymax), the duration of the alcoholic fermentation process (tAF), and the maximal production rate of CO2 during fermentation (rCO2max). The ability of yeasts to complete the fermentation processes, with complete consumption of sugars in the medium, was also evaluated. A strong dose-related effect of biotin on growth appeared throughout yeast monitoring, with significant increases (p < 0.05) arising for all growth parameters . Notable rises in the mmax and ymax values, as well as a significant drop in the generation time G values, can be found for initial biotin concentrations as little as 0.5 mg/L in the growth medium, suggesting that low doses of biotin might be sufficient in ensuring the coverage of the yeast requirements in the vitamin. Such an influence stands as both in regard to the growth celerity, as well as its intensity, allowing for the achievement of higher yeast populations when biotin is sufficiently provided in the growth medium. A saturation effect appeared here, since no further significant increase (p > 0.05) in the yeast maximal population can be found between cells grown in 0.5 mg/L or 3 mg/L of biotin . However, no significant effect of biotin on the fermentation rates could be found , suggesting that the vitamin solely affects growth, and that its deficiencies do not impair the fermentative capacity of the yeast, although in higher populations. While concurrent conclusions of this matter were drawn on anaerobically grown S. cerevisiae yeasts in previous investigations , other works rather concluded in, opposingly, a significant influence held by the vitamin on fermentation rates . Although the yeasts used in the assays led by Bohlscheid and colleagues were pre-starved in a biotin-deficient medium before inoculation, which might have contributed to inducing the notable differences in behaviors observed here , those used by Ough and Kunkee were rehydrated active dried yeasts, in a similar fashion to the yeasts used in the present work . The essential character of biotin appears to be, however, a strain-dependent character in S. cerevisiae, since all strains do not possess the capacity for biotin de novo biosynthesis . It is not impossible, as such, that the difference in fermentative behaviors that appears here might rather reflect such a disparity in regard to biotin biosynthesis capacity. There is, however, an overall consensus of past and present findings regarding the essential role assumed by biotin in yeast growth and viability, its significance being, as such, asserted. In contrast to biotin, thiamine appeared to exert an effect solely on fermentation kinetics and had no impact on the considered growth parameters . Interestingly, the significance of thiamine in regard to fermentation does not appear to be dose-related in a linear fashion, since distinctive differences in the fermentation rates only appeared between the yeasts cultivated in 50 mg/L and 250 mg/L of thiamine, the total absence of the vitamin rather standing as an intermediate, and highly versatile condition. While rising concentrations of the initial thiamine in the medium allowed for dose-dependent decreases in the duration of the fermentation process, this dose-dependent trend did not appear in the fermentation rates, since yeasts fermented in 50 mg/L appeared to display the lowest fermentation rates at any given biotin level. The intermediate fermentation rates resulting from the complete absence of thiamine in the cultivation medium might result from the establishment, by the yeast, of a survival strategy to resist the inadequate growth conditions. It has indeed been shown that, when grown in the complete absence of thiamine, S. cerevisiae displays thiamine synthesis-related proteins Thi4 and Thi5 as amongst its most abundant proteins , both acting as suicide enzymes to maintain thiamine levels . Interestingly, THI4 expression has been positively correlated to the S. cerevisiae fermentation rate , as well as negatively correlated to the thiamine concentration in the medium . Overall, the overexpression of THI4 has been proven to improve glucose fermentation and the resulting ethanol production ; as such, it could be envisioned that, in the absence of thiamine, a metabolic balancing system is established within the yeast as a response to the deficiency, allowing it to maintain the sufficient fermentative capacity that has been observed here. Surprisingly, however, these findings regarding the role exerted by thiamine on both S. cerevisiae growth kinetics do not agree with previous conclusions found on the matter by Labuschagne and colleagues; as such, notable decreases in S. cerevisiae EC1118 growth rates have been noted in thiamine limitation conditions, with significant effects observed for doses lower than 125 mg/L , while thiamine, whether absent or as low as 50 mg/L, did not induce any growth impairment here. Although these findings regarding the impact held by thiamine on the yeast growth do not concur, the research conducted by Labuschagne and colleagues does agree with the present results, that conclude in the essential character of thiamine for alcoholic fermentation . A difference in fermentative behavior in regard to thiamine does, however, appear since increases in thiamine have led to rather proportional increases in the yeast fermentation capacity in the work conducted by Labuschagne , rather than inducing a maintained fermentation state in complete absence of thiamine, as observed here. Such diverging findings might be an expression of individual differences between the different strains used, or might reflect the differences in physiological states of the yeasts at inoculation; as such, the fermentations conducted by Labuschagne and colleagues relied on thiamine-starved, pre-cultivated yeasts, rather than directly rehydrated commercial ADY that were used in the present experiment, and specifically chosen to replicate common yeasting conditions that may be found in winemaking contexts. Such differences in the yeasts' physiological states might have impacted their behavior in regard to thiamine and led to the observed differences. Plus, the constant agitation applied on the fermentation broth during the assays led by Labuschagne and colleagues, opposingly to the static mode selected here, might have notably impacted thiamine assimilation by the fermenting yeasts, as it does other nutrients . All in all, those results highlight the essential character held by biotin on yeast growth, allowing for it to reach sufficient population and rates for proper course of the biological processes. Similarly, thiamine's significance in regard to fermentation has been further assessed here, although a complete thiamine deprivation has been associated with maintained fermentation rates, possibly through the establishment of a survival strategy by the yeast, or the possibility that S. cerevisiae might have had some leftover inner thiamine resources that have allowed it to cope with thiamine deprivation. 3.2. Impact on Central Carbon Metabolites In order to further assess the impact held by both vitamins on the yeast metabolism, acetaldehyde, acetic acid, D-lactic acid, and glycerol were enzymatically quantified on the endpoint fermentation supernatants for all wines, and ethanol was determined through FT-IR spectroscopy (Table 4 and Table 5). Consistent with the evidence that all fermentations managed to achieve a complete consumption of sugars, ethanol did not display any significant dependence (p > 0.05) in regard to the initial concentrations of the vitamins (Table 4 and Table 5). Interestingly, amongst all four investigated compounds, solely acetaldehyde and ethanol did not display any significant variation (p > 0.05) regarding the initial thiamine concentration in the medium (Table 5). Such an effect on acetaldehyde does contradict previous knowledge, that established thiamine deficiencies to be associated with higher contents of sulfur-binding compounds, including acetaldehyde , as well as the strong dependency manifested by pyruvate decarboxylase (Pdc) towards thiamine for the formation of acetaldehyde from pyruvate. However, some other findings, although recognizing the clear influence of thiamine deficiencies on sulfur-binding ketoacids, did not, opposingly, find any impact of it on acetaldehyde itself . Here, the absence of any effect of thiamine on acetaldehyde contents in wines clearly appeared. Biotin, however, appeared to have a significant effect on acetaldehyde accumulation (p < 0.05) (Table 4). Such a rise in acetaldehyde contents might indirectly result from other influences held by the vitamin on carbon metabolism, although its exact cause remains obscure. In a surprising fashion, acetic acid displayed significant (p < 0.05) and opposed responses to both vitamins (Table 4 and Table 5); as such, while it appears to increase concurrently with the initial thiamine concentration in the medium, its contents seem to drop when increasing the initial biotin. Interestingly, however, the statistical interaction between both vitamins, by itself, does not hold any significant influence (p > 0.05) on the acetic acid contents formed during alcoholic fermentation. The positive effect held by the initial thiamine on acetic acid here appears consistent with previous conclusions, that found higher concentrations of acetic acid to be associated with higher thiamine . Although the negative effect held by the initial biotin on acetic acid contents appears to be more surprising, evidence show that there is a lesser incorporation of acetic acid into fatty acids in biotin-deficient yeast cells , suggesting that, in the absence of the vitamin, there might be an accumulation of acetic acid in the medium, leading to the higher contents that were observed here. While biotin did not have any significant influence (p > 0.05) on the D-lactic acid contents of the produced wines, a clear effect of thiamine (p < 0.05) on those contents appeared (Table 4 and Table 5). Interestingly, D-lactic acid concentrations seemed to be decreasing against the initial thiamine contents. This negative effect most likely results from increases in lactic acid production favored by thiamine depletions due to the defective thiamine pyrophosphate (TPP)-dependent pyruvate decarboxylase (Pdc) activity , that therefore redirects pyruvate towards lactic acid synthesis. Finally, glycerol contents in the final wines displayed a significant dependency (p < 0.05) towards both the initial biotin and thiamine concentrations in the growth medium (Table 4 and Table 5); as such, the amounts of glycerol produced noticeably increased when increasing both vitamins. This appears consistent with previous findings, that concluded towards drops in the glycerol contents when fermenting in thiamine-deficient media , while the effect of biotin on glycerol production has not been assessed yet. The statistical interaction between both vitamins, however, did not display any significant influence (p > 0.05) on glycerol production. All in all, the consequent involvement of thiamine in the yeast central carbon metabolism (CCM) is potently highlighted here, as a reflection of its cofactor role towards several CCM-associated enzymes . While the effect held by biotin on these metabolites appeared slightly less consequent, its noticeable influence on both acetaldehyde and acetic acid emerged here, overall concluding towards the clear existence of an effect of both vitamins on the yeast metabolism during alcoholic fermentation. 3.3. Relevance for the Production of Volatile Compounds during Alcoholic Fermentation The volatile profiles of wines resulting from the alcoholic fermentation of S. cerevisiae on the medium containing different concentrations of both thiamine and biotin were analyzed by HS-SPME-GC/MS, with the aim of assessing any potential effect held by those vitamins on the wine volatilome. Significant differences between the various vitamin doses have been investigated through Scheirer-Ray-Hare nonparametric tests, and hierarchical clustering of the compounds that displayed a significant reliance for either biotin or thiamine was performed using the average concentrations for those compounds at each considered vitamin level . Amongst the 31 compounds that were analyzed, 13 displayed a significant influence (p < 0.05) of the initial biotin composition, 11 were significantly correlated (p < 0.05) to the initial thiamine concentration, and only 1, methyl heptenone, was significantly impacted by the interaction between both vitamins, but interestingly, not by thiamine or biotin considered individually (Table 5 and Table 6). It is also relevant to note that only isovaleric acid was impacted by both initial thiamine and biotin individually. 3.3.1. Biotin Influence on Wine Volatile Compounds Out of the 31 analyzed, 13 volatile compounds were significantly influenced by the initial biotin concentration in the synthetic must medium , offering the first evidence for the relevance of biotin in the production of wine volatiles, on a substantial scale of analyzed compounds. Fatty Acids, Fatty Alcohols, and Fatty Acid Ethyl Esters (FAEEs) A clear effect held by the vitamin on fatty acid metabolism appeared here, since 7 of the 13 fatty acids, fatty alcohols, and fatty acid ethyl esters (FAEEs) displayed significant (p > 0.05) dose-related variations in their contents . Notable increases in all the fatty acid-related compounds appeared when increasing the initial biotin concentration in the medium, therefore highlighting a strong stimulatory effect of the vitamin on those compounds. Interestingly, it must be noted that such increases do appear for all derived compounds of a given fatty acid, with its associated alcohol and ethyl esters being heightened in a similar fashion, suggesting that the stimulatory influence of the vitamin on those compounds appears at the previous stages of the fatty acid synthesis pathway. This effect is to be explained by the essential role played by biotin in fatty acid synthesis and elongation, with the vitamin acting as a cofactor to Acc1 and Hfa1 in the conversion of acetyl-CoA towards malonyl-CoA, which stands as the first step of the fatty acid synthesis pathway, before malonyl-CoA enters the FAS system . Ethyl nonanoate, however, did not present any significant effect of the initial biotin on its concentration (0.39 +- 0.05 mg/L, 0.48 +- 0.17 mg/L and 0.52 +- 0.16 mg/L at 0, 0.5, and 3 mg/L of biotin, respectively), although its precursor, nonanol, did indeed show similar increases to the other impacted compounds, although to an apparently lesser degree . Such a difference might be solely resulting from the decreased transfer of ethyl esters to the medium when increasing the fatty acid carbon chain length . However, the other longer-chained fatty acids and associated alcohols and FAEEs, such as decanoic acid (C10), decanol, ethyl decanoate, and ethyl laurate (C12), displayed no significant differences in the concentrations in regard to the initial biotin contents of the medium. Three of the impacted compounds, ethyl hexanoate, ethyl octanoate, and octanoic acid, do, interestingly, display odor activity values (OAV) above 1, detected above their olfactory thresholds of 14 mg/L, 5 mg/L, and 5 mg/L, respectively . This suggests that the initial biotin in the medium might lead to notable changes in the olfactory profiles of wines. No definite conclusion, however, can be drawn on the matter without any sensory analysis to definitely assess the impact held by the vitamin on wine aromatic profiles; this, however, strikingly proves the relevance of sufficient biotin in grape must for the production of fatty acids and FAEEs. Higher Alcohols, Higher Aldehydes, Fusel Acids, and Associated Esters (Ehrlich Pathway) In addition to the clear impact held by the initial biotin on the production of fatty acids and derivatives that has been observed here, a lesser influence of the vitamin on the production of higher alcohols and derivatives has appeared, since 4 of the 10 higher alcohols, higher aldehydes, and associated esters, that can be linked to the Ehrlich pathway, display significant differences in their contents in regard to the initial biotin contents of the synthetic must . Interestingly, two notable groups of affected compounds seemed to materialize in regard to their behaviors in the face of the initial biotin concentration: while propyl acetate and isovaleric acid increased in contents when increasing biotin, isoamyl alcohol and isoamyl acetate appeared to decrease against the vitamin . Surprisingly, no consensus appears to exist in regard to the higher levels of isoamyl alcohol and its corresponding acetate observed here; as such, albeit Bohlscheid and colleagues reported decreases of the isoamyl alcohol contents in case of biotin deprivation , a complete absence of the effect of the vitamin on the production of this compound had been previously reported by Gutierrez and associates . With the yeast strains used in these various studies being different, it is not impossible that this disagreement in behavior could rather be a strain-dependent character, or reflect their individual tolerances towards low biotin availabilities. This could also result from disparities in the yeasts' physiological states, since, notably, the yeasts used by Bohlscheid and colleagues were pre-starved in a biotin-deficient medium before inoculation . The increase in the other compounds when increasing biotin might rather be related to their amino acid origin. Reductions in the cellular amino acid contents of yeasts have been found when grown in biotin-deficient medium , which might justify the limited formation of their associated higher alcohols, aldehydes, and esters in case of biotin deficiency. Although the exact metabolic impact of biotin in the production of these compounds is not clear, such a behavior can be overall associated with the role held by the vitamin as a cofactor to several carboxylases, including pyruvate and acetyl-CoA carboxylases as well as in the Ehrlich pathway decarboxylation step of a-keto acids towards their associated aldehydes . Amongst these impacted contents, both isovaleric acid and isoamyl alcohol were found in concentrations above their odor detection thresholds of 0.7 and 7 mg/L, respectively , further supporting the idea that biotin might play a role in the sensory properties of final wines. Interestingly, the ratio between the total ethyl esters (calculated as sum of the ethyl esters quantified here) and total acetate esters (calculated as sum of the acetate esters quantified here) appears to be significantly (p < 0.05) influenced by the initial biotin in the medium, increasing alongside the vitamin. Surprisingly, though, the total esters (calculated as sum of the esters quantified here) do not present any significant (p > 0.05) dependency regarding biotin, suggesting that the vitamin indeed modulates the ester profile of the wines, orienting it preferentially towards ethyl esters rather than acetate esters, and offering the first evidence of such an effect of the vitamin on the ester formation in wine. Central Carbon Metabolism-Derived Compounds Notable increases in the diethyl succinate concentrations were found when increasing the initial biotin contents in the synthetic must medium , rising from 0.35 +- 0.05 to 0.45 +- 0.03 mg/L when increasing biotin from 0 to 3 mg/L. On the other hand, significant decreases in the concentration of benzaldehyde formed during fermentation appeared when raising the initial biotin doses , dropping from 52.50 +- 16.62 mg/L to 29.62 +- 4.91 mg/L when reaching 3 mg/L of the initial vitamin in the synthetic must. All in all, such results further suggest the influence that biotin might have on the yeast central carbon metabolism during alcoholic fermentation. 3.3.2. Thiamine Influence on Wine Volatile Compounds Out of the 31 analyzed, 11 volatile compounds were significantly influenced by the initial thiamine concentration in the synthetic must medium . Fatty Acids, Fatty Alcohols, and Fatty Acid Ethyl Esters (FAEEs) Interestingly, amongst those affected compounds, a noticeable impact of the initial thiamine on fatty acids and derivatives appeared here: 5 of the 13 fatty acids, fatty alcohols, and FAEEs indeed displayed significant (p < 0.05) dose-related variations in their contents . Surprisingly, only the longer medium-chain fatty acids (MCFAs) that were analyzed here exhibited a significant dependence towards the initial thiamine in the medium, since solely decanoic acid (C10), decanol, ethyl decanoate, and ethyl laurate (C12) were affected here. Those compounds were associated with increases alongside the initial thiamine concentration in the medium , and the results appear consistent with previous findings by Labuschagne during the alcoholic fermentation of S. cerevisiae . On the other hand, butanol and ethoxy propanol were associated with notable decreases in concentrations when increasing the initial thiamine, in accordance with previous findings by Labuschagne . Interestingly, out of those five significantly impacted compounds, solely ethoxy propanol was present in concentrations above its 100 mg/L detection threshold , suggesting that the initial thiamine might contribute to modulating the sensory profiles of wines. Higher Alcohols, Higher Aldehydes, Fusel Acids, and Associated Esters (Ehrlich Pathway) Besides the influence held by the vitamin on fatty acid metabolism, an impact of the initial thiamine concentration in the synthetic must medium on higher alcohols, higher aldehydes, and their derivative esters appeared here , since 5 of those compounds out of their total 10 displayed significant (p < 0.05) variations in regard to thiamine contents. As such, a clear increase in the final concentrations of phenyl ethyl alcohol, phenyl ethyl acetate, and isovaleric acid appeared when raising thiamine doses . Such a phenomenon appears consistent with the thiamine-dependency exhibited by the Tkl1/2 transketolase of the pentose phosphate pathway (PPP), as well as the decarboxylation step (Aro10, Pdc1/2/5) of the Ehrlich pathway . The Ehrlich pathway precursor phenylpyruvate is, indeed, a product of the shikimate pathway, itself deriving from the PPP through phosphoenolpyruvate, and further converted to phenyl ethanol. It is, however, most surprising that the other higher alcohols do not display a similar behavior and dependency in regard to thiamine, since isoamyl alcohol and isoamyl acetate did not express any significant effect of the initial vitamin on their contents , consistent with the absence of any influence of the initial thiamine on isoamyl alcohol, as found by Labuschagne using S. cerevisiae . Amongst those compounds, both isovaleric acid and phenyl ethanol were found in concentrations above their olfactory detection thresholds of, respectively, 7 and 14 mg/L . It also appears relevant to note that the ratio between superior alcohols and the total esters (calculated as the sum of the individual esters quantified here) was significantly (p < 0.05) impacted by the initial thiamine in the medium, and diminished against thiamine doses, suggesting that, although the vitamin stands as a cofactor in the initial steps of the Ehrlich pathway, it overall favors the production of esters in the final wine. Significant (p < 0.05) increases in the total esters were indeed found in regard to the initial thiamine, rising from 42.5 +- 0.9 to 71.6 +- 1.8 mg/L when increasing thiamine from 0 to 250 mg/L. This first evidence of an influence held by the vitamin on ester production in wines suggests that sufficient thiamine doses, therefore, might contribute to expanding the ester profiles of wines, and, as such, possibly take part in the definition of the aromatic signature of the wine, since most esters are associated with fruity, pleasant notes . Although the individual esters reported here were mostly under their detection thresholds, and therefore individually might not play any role in the sensory profiles of wines, their synergistic, combined effect might prove significant . Central Carbon Metabolism-Derived Compounds In addition to what was observed in volatiles from the fatty acid metabolism and the Ehrlich pathway, thiamine appeared to have a significant influence on only one carbon metabolism-derived compound out of the four quantified here . As such, ethyl benzaldehyde rose from 2.89 +- 0.35 mg/L to 3.49 +- 0.33 mg/L when similarly increasing thiamine from 0 to 250 mg/L, which might be originating from the dependency held by benzaldehyde lyase (BAL) towards thiamine pyrophosphate for the synthesis of acetoin from benzaldehyde and acetaldehyde , although benzaldehyde itself, interestingly, did not present any significant differences in regard to initial thiamine. This might, however, be rather dependent of acetaldehyde contents, since those appear to have decreased in an opposite fashion to ethyl benzaldehyde. It could be envisioned, as such, that since the conversion of benzaldehyde and acetaldehyde towards acetoin cannot be as efficiently performed in case of thiamine limitation, the esterification of benzaldehyde towards ethyl benzaldehyde could preferentially occur. 3.4. Biotin and Thiamine Impact on the Wine Metabolome An untargeted UPLC-qTOF-MS analysis of the wines collected at the endpoint of the alcoholic fermentation was performed in order to investigate the influence held by both biotin and thiamine on the S. cerevisiae metabolome. The retrieved features were filtered to only retain those existing in at least two of the three biological replicates of all modalities, to ensure significance in their presence in the final wines. Subsequently to those treatments, 3870 features have been detected in both positive and negative ionization modes and analyzed using both PCA at a significance level of a = 0.05 and one-way ANOVA at a significance level of a = 0.01. In order to investigate possible differences between the metabolomes of S. cerevisiae when grown in different initial concentrations of either biotin or thiamine, a principal component analysis (PCA) was performed on the intensities obtained for all 3870 extracted features, subsequently leading to planar representations in which components PC1, PC2, and PC3 accounted for, respectively, 20.4%, 13.5%, and 5.6% of the observed variations . No clear discrimination between groups appeared here, and especially no differentiation could be found between the intermediary and lowest vitamin doses. As such, for all further metabolomic investigations, the wines obtained from S. cerevisiae grown in 3 mg/L of biotin ("high biotin") were separated from those grown in both 0.5 and 0 mg/L of biotin ("low biotin"), for which data were taken into account together, as one unique condition. Similarly, wine resulting from alcoholic fermentation in 250 mg/L of thiamine ("high thiamine") was considered separately from those obtained in both 50 and 0 mg/L of thiamine ("low thiamine"). 3.4.1. Biotin Influence on the Wine Metabolome ANOVA performed on the extracted features allowed for the isolation of 208 compounds that presented significant differences (p < 0.01) between both high and low biotin conditions, therefore accounting for less than 10% of the extracted features. Putative annotations were assigned to all significant extracted features according to the KEGG, Metlin, and Oligonet online databases and tools . Consequently, out of those 208 notable features, 164 were assigned a possible chemical formula, and therefore annotated at level 4 ; as such, those annotated features were the sole ones to be considered in all further investigations of the impact of biotin on the S. cerevisiae metabolome. Amongst those, a remarkably higher proportion of features appeared to be significantly more intense in high biotin than in low biotin, since 133 of those appeared significant (p < 0.01) in wines obtained from 3 mg/L of biotin, against 31 resulting from those associated with the lower doses of the vitamin. Hierarchical clustering of those features led to a clear distinction between wines obtained from higher and lower initial contents of the vitamin, although two low biotin samples appeared less properly discriminated from the high biotin ones , which strongly highlights the impact of biotin on the yeast metabolism. To further assess the nature of this influence, Van Krevelen diagrams were plotted on the features' O/C and H/C ratios for both biotin conditions, and the elemental compositions of them were investigated . Notable differences in the chemical composition of wines appeared between the features associated with each biotin condition; as such, while high biotin-resulting wines displayed a dominance of CHON-based features, the predominating ones in low biotin wines were the phosphorus-containing features (CHOP, CHONP, and CHONPS). It is, however, relevant to note that those predominant features are, in either biotin condition, found in the lipid (O/C < 0.6 and H/C > 1.3) and polyphenol (O/C < 1.2 and H/C < 1.3) regions of the Van Krevelen diagram, consistent with their predicted chemical families . In addition, an interesting influence exerted by biotin on the proportion in the CHONS features appeared here, since they dropped from 11.3% in high biotin to 6.5% in low biotin. Subsequently to the features' nature assumption, hypothetical annotations were assigned using the MassTRIX application (Helmholtz Zentrum Munchen, Munchen, Germany), and its associated databases, such as KEGG , as accessed on 11 November 2022. A limited number of the specific biomarkers were successfully annotated in the databases, since less than 10% of those were matched, reflecting the high complexity and current low understanding of the wine composition . Those annotations have allowed to identify the metabolic pathways associated with the changes in the exometabolome changes observed as a result of the initial biotin in the medium . Unsurprisingly, a higher number of pathways were impacted by a high initial biotin, although the number of associated biomarkers to each of those pathways remains limited, as a consequence of the low number of annotated features overall. No pathway appeared here, as such, to be notably more affected than others; however, it appears relevant to note that the impacted metabolic pathways here are reliant on major pathways, such as central carbon metabolism, amino acid metabolism, and lipid metabolism. While its influence on the CCM might be a reflection of the biotin-dependent enzymes Acc1/Hfa1 and Pyc1/2 , its influence on amino acids appears less clear, although they might be an indirect result of these biotin-dependent reactions, and notably the Pyc1/2 conversion of pyruvate towards oxaloacetate . The absence of any more significance in the effect of high biotin on the lipid metabolism appeared, however, highly surprising, notably in regard to the essential role held by the vitamin in the first steps of the synthesis of fatty acids . A similar impact on carbon metabolism was found in the low biotin wines, through the fructose and mannose metabolism. Surprisingly, low biotin was also found to impact riboflavin metabolism. All in all, the untargeted approach allows to conclude, for the first time, on the actual influence of the initial biotin on the yeast exometabolome during alcoholic fermentation. 3.4.2. Impact of Thiamine on the Wine Metabolome In order to determine the extent of the influence of thiamine on the yeast exometabolome, an ANOVA was performed on the 3870 extracted features, and resulted in the isolation of 515 that were associated with significant differences (p < 0.01) in regard to the initial concentration of the vitamin, therefore accounting for nearly 15% of the extracted features. However, only 378 of those specific features were assigned possible chemical formulas through putative annotations in online databases, in a similar fashion to what was performed on biotin-associated features, and these were the sole ones to be considered for their significant influence in the S. cerevisiae thiamine-dependent metabolome. It is also relevant to note that the number of features associated with thiamine here was significantly higher than those associated with biotin, since it more than doubled its amount, also suggesting a greater influence of thiamine on the yeast metabolism. Similar to what was observed in the assessment of the influence of biotin on the yeast exometabolome, a high proportion of those features were significantly more intense in high thiamine, amounting to 378, while only 69 features appeared significantly more intense (p < 0.01) in low thiamine wines. Such a differentiation also appeared quite clearly through hierarchical clustering of these 515 total biomarkers, since it resulted in a distinct discrimination between high and low thiamine wines , strongly testifying on the influence exerted by the vitamin on the wine exometabolome. An investigation of the chemical composition of the thiamine-affected features was led, in order to find the precise nature of the effect exercised by the vitamin on the yeast metabolism during fermentation. As such, Van Krevelen diagrams were drawn based on the O/C and H/C atomic ratios of the specific features for both the low and high thiamine conditions, and their elemental composition was investigated in an effort to define their chemical identities. Surprisingly, as opposed to what was observed for biotin, there was no clear difference between the elemental composition proportions of the high and low thiamine features , although there was a slight increase in the percentage of CHO markers in high biotin, amounting to 11.7% of the features, against 7.2% in low thiamine, as well as a slight increase in phosphorus-containing ones in low thiamine, reaching 39.1% of the features, against 33.0% in high thiamine. Interestingly, however, these CHO features in high thiamine conditions appeared to be amongst the most intense ones, found mostly where the amino sugars (O/C > 0.6 and H/C > 1.5) and carbohydrates (O/C > 0.8, 1.6 < H/C < 2.7) were expected on the Van Krevelen diagrams. Plus, it appears relevant to note that the strongly represented CHON biomarkers, not significantly varying in proportion between both thiamine conditions, also appeared to display similar behaviors in regard to their intensities and nature, whether obtained from low or high thiamine wines. They are, indeed, associated with a number of high intensities, as observed on Van Krevelen diagrams, and both were found in the area in which polyphenols were expected (O/C < 1.2 and H/C < 1.3), consistent with the high proportion of polyphenol compounds that were predicted . Thiamine did not appear, consequently, to interfere in any way with the frequency of formation of neither polyphenols nor lipids; however, the high thiamine exometabolome did display increases in the proportion of carbohydrates and amino sugars formed during alcoholic fermentation, while low thiamine appeared to induce noticeable increases in the proportion of synthetized proteins. A limited number of biomarkers for both thiamine conditions was annotated at level 3 in online databases, since less than 5% of those were matched with an identification. However, this further annotation process has led to the identification of the pathways in which those biomarkers are involved, and therefore, associated with the exometabolome changes that were observed here . In contrast to the low number of pathways affected by the initial biotin, there was, here, a thorough influence of high thiamine on numerous metabolic pathways, since 46 named pathways were matched through KEGG for its associated biomarkers, while a far lesser number of metabolic pathways were influenced by low thiamine. Interestingly, high thiamine seems to exert a strong influence on both amino acid and carbon metabolisms, since they displayed the highest number of total annotated biomarkers (53 and 51, respectively), distributed in numerous, more specified pathways. Thiamine, as such, appears to be a relevant actor in amino acid synthesis, matching with 10 biomarkers, as well as being strongly involved in phenylalanine, valine, leucine, isoleucine, tyrosine, and lysine metabolisms, more specifically. The overall effect of the vitamin on amino acid metabolism might reflect, overall, its role as a cofactor involved in the Ehrlich pathway, with TPP being essential to the decarboxylation step; however, it might also reflect, indirectly, the relevance of pyridoxine in the transamination of the Ehrlich pathway as well, which can proceed to the conversion of an amino acid towards its associated a-ketoacid, and inversely . Since thiamine biosynthesis relies on pyridoxine in case of thiamine deficiencies, in a mechanism that takes priority over all other pyridoxine-dependent reactions , it is not impossible that this effect on amino acid synthesis also reflects the indirect influence of pyridoxine. Plus, such an influence of thiamine on both phenylalanine and tyrosine, more specifically, appears consistent with the role held by the vitamin in the pentose phosphate pathway through TPP-dependent transketolases Tkl1/2 , from which prephenate is derived, itself a precursor in the biosynthesis of both aromatic amino acids . Similarly, the differences observed regarding both valine and leucine metabolism might reflect the thiamine-dependent character of acetolactate synthase Ilv2 , which is involved in the synthesis of both branched-chain amino acids . All in all, such an influence held by thiamine on the amino acid pathways in the S. cerevisiae exometabolome appears consistent with our previous observations regarding the wine volatile compounds, which highlighted a similar effect. The strong influence exerted by high thiamine on carbon metabolism, similarly, affects a large number of pathways involved in the CCM, and highlights, notably, a remarkable effect on keto acids, such as 2-oxocarboxylic acids, C5-branched dibasic acids, and pyruvate, which might reflect the TPP-dependency of the Ehrlich decarboxylation reaction of a-keto acids towards their associated aldehydes , as well as the conversion of pyruvate towards S-acetolactate through TPP-dependent Ilv2 . The observed influence of thiamine on both the propanoate metabolism and the TCA cycle biomarkers, similarly, might be a reflection of the cofactor role played by TPP in the conversion of a-ketoglutarate to succinyl-CoA by Kgd1 , which is further derived towards propanoyl-CoA . This leads to, overall, more evidence of the involvement of thiamine in the CCM, consistent with previous conclusions evaluating its relevance in the volatile compounds that were produced, as well as remarkable compounds, such as acetic acid or glycerol. In addition, high thiamine appears to affect other notable metabolic pathways, such as those of lipids, as well as having a striking involvement in the synthesis of other cofactors and vitamins, and notably in the metabolism of vitamin B3 through its apparent connection with the nicotinate and nicotinamide metabolism, strongly suggesting the possibility for synergistic effects of vitamins in the yeast metabolism. Interestingly, however, low thiamine seemed to have a similar effect on S. cerevisiae metabolic pathways to that of low biotin, likewise impacting riboflavin metabolism in a fashion that still remains unclear. Thiamine, overall, appeared here as a striking actor in the yeast metabolism, intervening in numerous metabolic pathways, and proving its strong influence on the wine metabolome during alcoholic fermentation. This investigation, on a larger scale, is the first evidence of the impact held by both biotin and thiamine on the wine exometabolome during the alcoholic fermentation by S. cerevisiae. 4. Conclusions The effect of both thiamine and biotin on yeasts in an enological context was investigated, from both a kinetic and metabolic standpoint, detailing their influence on the yeast growth and fermentation rates, as well as on the volatile profiles of the final wines. This study has, therefore, provided novel information in regard to the significance of both vitamins in enology, thoroughly assessing the essential character of biotin regarding growth, and the high relevance of thiamine in the course of fermentation processes. Both vitamins have been proven to have a notable influence on the metabolic course of the fermenting yeast, significantly impacting nearly half of the quantified volatile compounds and being further highlighted in an untargeted LC-MS investigation on the wine non-volatile metabolome. A significant role of both thiamine and biotin was found on volatile compounds resulting from the fatty acids and the higher alcohol metabolisms, with the clear role of biotin in the synthesis of fatty acids being highlighted here, and a noticeable influence of thiamine on the central carbon metabolism appearing as well. A striking influence of thiamine on numerous metabolic pathways was also proven here, with its evident role in the wine metabolome being shown here on a large untargeted scale, concomitant with the first evidence of the impact of biotin on the yeast exometabolome. All in all, this exploration of the influence held by both thiamine and biotin on wine composition provides leads for further exploration of the role of vitamins in an enological context, suggesting a possibility for an impact of the wine sensory profiles by their concentrations in the grape musts. As such, varying biotin and thiamine contents of musts may affect wine production, and notably, since thiamine contents of musts were shown to be correlated with their vineyard sites and geographical origin , which might factor in inducing differences in the final wines. This, in addition, highlights the evidence of an influence of the yeasts' physiological state on the management and response to deficiencies, through the appearance of notable differences between the present results, and those resulting from the inoculation of pre-starved yeasts in the medium, as observed in other previous studies. Such an effect, similarly, offers grounds for further understanding of these physiological states when inoculating wine yeasts, evidently a source for relevant oenological applications. Acknowledgments The authors would like to most sincerely thank Remy Romanet for performing the metabolomic analyses as part of the DiVVA platform. Supplementary Materials The following supporting information can be downloaded at: Table S1: Impact of both biotin and thiamine on Saccharomyces cerevisiae growth and fermentation kinetics in synthetic wine. Table S2: Saccharomyces cerevisiae-impacted metabolic pathways in high and low biotin conditions. Table S3: Saccharomyces cerevisiae-impacted metabolic pathways in high and low thiamine conditions. Figure S1: Metabolic impact of biotin (a) and thiamine (b) on S. cerevisiae. Click here for additional data file. Author Contributions Conceptualization, M.S.E., H.A., C.R.-G. and A.G.; methodology, M.S.E., H.A. and C.R.-G.; formal analysis, M.S.E., C.R.-G. and S.V.; investigation, M.S.E.; resources, H.A., C.R.-G., C.M., C.S. and A.G.; data curation, M.S.E. and C.R.-G.; writing--original draft preparation, M.S.E.; writing--review and editing, M.S.E., H.A., C.R.-G., C.M., C.S., A.G. and S.V.; visualization, M.S.E., H.A. and C.R.-G.; supervision, H.A., C.R.-G. and A.G.; project administration, H.A., C.R.-G., C.M. and C.S.; funding acquisition, H.A., C.R.-G., C.M. and C.S. All authors have read and agreed to the published version of the manuscript. Data Availability Statement Data are contained within the article and the Supplementary Materials. Conflicts of Interest Authors Marie Sarah Evers, Chloe Roullier-Gall, Stefania Vichi and Herve Alexandre declare no conflict of interest. Authors Christophe Morge, Celine Sparrow and Antoine Gobert provided the yeast strain, did not take part in discussing the results and have no influence on the correctness of the results. Figure 1 Growth and fermentation profiles of S. cerevisiae under different biotin and thiamine concentrations. Bluer and darker blue shades indicate higher initial thiamine in the cultivation medium, while pinker and darker pink shades indicate higher initial biotin in the cultivation medium. Figure 2 Hierarchical clustering of the biotin (left) and thiamine (right) significantly impacted volatile compounds in the final wines: effect of both vitamins on their profiles. (A-C): Biotin-impacted volatile compounds ((A): Fatty acid metabolites, (B): Ehrlich pathway metabolites, (C): Central carbon metabolites). (D-F): Biotin-impacted volatile compounds ((D): Fatty acid metabolites, (E): Ehrlich pathway metabolites, (F): Central carbon metabolites). Figure 3 Intensity and distribution of the wine extracted metabolomic features after alcoholic fermentation in varying biotin and thiamine concentrations. (A) Principal component analysis individual plot according to the initial biotin content in the growth medium. (B) Hierarchical clustered heatmap of the one-way ANOVA significantly different (p < 0.01) level 4 annotated features between high and low biotin fermentation conditions. (C) Principal component analysis individual plot according to the initial thiamine content in the growth medium. (D) Hierarchical clustered heatmap of the one-way ANOVA significantly different (p < 0.01) level 4 annotated features between high and low thiamine fermentation conditions. Figure 4 Metabolomic foot-printing of S. cerevisiae grown in high or low biotin conditions based on LC-MS-MS untargeted analysis data. Van Krevelen diagrams based on the H/C and O/C atomic ratios for the significantly more intense features (extracted through ANOVA; p < 0.01), histogram proportion of their elemental compositions, and pie chart distribution of the predicted families for these biomarkers in high biotin conditions (left, 133 markers) and low biotin conditions (right, 31 markers). The point size indicates the relative intensity of the extracted masses. Figure 5 Metabolomic foot-printing of S. cerevisiae grown in high or low thiamine conditions based on LC-MS-MS untargeted analysis data. Van Krevelen diagrams based on the H/C and O/C atomic ratios for the significantly more intense features (extracted through ANOVA; p < 0.01), histogram proportion of their elemental compositions, and pie chart distribution of the predicted families for these biomarkers in high thiamine conditions (left, 378 markers) and low thiamine conditions (right, 79 markers). The point size indicates the relative intensity of the extracted masses. foods-12-00972-t001_Table 1 Table 1 Thiamine and biotin contents of the modified MS300 synthetic musts. Medium Thiamine (mg/L) Biotin (mg/L) T0 0 0 0 0.5 B1+ 250 0 B8+ 0 3 - 50 0.5 B1+ 0.5 + 50 3 T+ 250 3 foods-12-00972-t002_Table 2 Table 2 Initial biotin impact on the growth and fermentation kinetics. Parameter Initial Biotin Concentration (mg/L) 0 0.5 3 mmax (h-1) 1 0.12 +- 0.00 a 0.15 +- 0.01 b 0.16 +- 0.00 b G (h) 2 5.7 +- 0.2 a 4.7 +- 0.2 ab 4.6 +- 0.2 b ymax (cells/mL) 3 8.0 x 107 +- 1.7 x 107 a 2.3 x 108 +- 4.6 x 107 b 1.8 x 108 +- 6.5 x 107 b tAF (h) 4 138 +- 11 a 117 +- 11 a 126 +- 11 a rCO2max (gCO2/200 mL/h) 5 0.9 +- 0.4 a 1.1 +- 0.8 a 0.8 +- 0.2 a Note: different letters in the same line indicate significant differences between conditions (Scheirer-Ray-Hare and Dunn tests, p < 0.05). 1 mmax: maximal specific growth rate; 2 G: generation time; 3 ymax: maximal population; 4 tAF: duration of the alcoholic fermentation; 5 rCO2max: maximal CO2 production rate. foods-12-00972-t003_Table 3 Table 3 Initial thiamine impact on the growth and fermentation kinetics. Parameter Initial Thiamine Concentration (mg/L) 0 50 250 mmax (h-1) 1 0.16 +- 0.07 a 0.15 +- 0.06 a 0.13 +- 0.07 a G (h) 2 4.8 +- 0.1 a 4.9 +- 0.1 a 5.3 +- 0.1 a ymax (cells/mL) 3 1.5 x 108 +- 6.0 x 107 a 1.4 x 108 +- 8.3 x 107 a 1.1 x 108 +- 3.5 x 107 a tAF (h) 4 136 +- 22 a 134 +- 17 a 112 +- 19 b rCO2max (gCO2/200 mL/h) 5 1.1 +- 0.9 ab 0.7 +- 0.1 a 0.9 +- 0.2 b Note: different letters in the same line indicate significant differences between conditions (Scheirer-Ray-Hare and Dunn tests, p < 0.05). 1 mmax: maximal specific growth rate; 2 G: generation time; 3 ymax: maximal population; 4 tAF: duration of the alcoholic fermentation; 5 rCO2max: maximal CO2 production rate. foods-12-00972-t004_Table 4 Table 4 Central carbon metabolites in the final wines produced in varying initial biotin concentrations. Compound Initial Biotin Concentration (mg/L) 0 0.5 3 Acetaldehyde (mg/L) 108.9 +- 32.8 a 144.1 +- 19.8 a 162.4 +- 6.5 b Acetic acid (g/L) 1.4 +- 0.2 a 1.3 +- 0.4 ab 1.0 +- 0.2 b D-lactic acid (g/L) 0.2 +- 0.0 a 0.2 +- 0.1a 0.2 +- 0.0 a Glycerol (g/L) 8.0 +- 0.6 a 7.0 +- 0.8 b 6.5 +- 0.3 b Ethanol (% v/v) 9.6 +- 0.1a 9.6 +- 0.1a 9.4 +- 0.3 a Note: different letters in the same line indicate significant differences between conditions (Scheirer-Ray-Hare and Dunn tests, p < 0.05). foods-12-00972-t005_Table 5 Table 5 Central carbon metabolites in the final wines produced in varying initial thiamine concentrations. Compound Initial Thiamine Concentration (mg/L) 0 50 250 Acetaldehyde (mg/L) 126.7 +- 37.8 a 135.8 +- 35.8 a 151.7 +- 7.1a Acetic acid (g/L) 1.1 +- 0.4 a 1.2 +- 0.2 ab 1.4 +- 0.2 b D-lactic acid (g/L) 0.3 +- 0.0 a 0.2 +- 0.0 b 0.2 +- 0.0 b Glycerol (g/L) 6.8 +- 0.6 a 7.0 +- 0.8 a 7.8 +- 1.06 b Ethanol (% v/v) 9.5 +- 0.2 a 9.6 +- 0.1 a 9.5 +- 0.3 a Note: different letters in the same line indicate significant differences between conditions (Scheirer-Ray-Hare and Dunn tests, p < 0.05). foods-12-00972-t006_Table 6 Table 6 Biotin-dependent volatile contents in the final wines. Compound (mg/L) Initial Biotin Concentration (mg/L) 0 0.5 3 Benzaldehyde 52.57 +- 16.76 a 36 +- 10.69 ab 29.56 +- 4.93 b Diethyl succinate 0.34 +- 0.05 a 0.47 +- 0.1 ab 0.44 +- 0.05 b Ethyl dec-9-enoate 0.69 +- 0.16 a 1.77 +- 3.06 ab 1.19 +- 0.4 b Ethyl hexanoate 38.29 +- 4.54 a 41.67 +- 7.95 a 52.56 +- 4.33 b Ethyl octanoate 51.43 +- 4.76 a 54.22 +- 6.7 a 71.56 +- 9.03 b 1-Heptanol 9.14 +- 1.57 a 14.22 +- 4.29 b 14.22 +- 2.44 b Isoamyl acetate 83 +- 18.15 a 81.44 +- 28.85 ab 59.78 +- 13.46 b Isoamyl alcohol 75,173.43 +- 8173.63 a 67,535.22 +- 19,964.24 a 49,623.22 +- 4767.42 b Isovaleric acid 2787.86 +- 156.47 a 3099.56 +- 401.25 ab 3427.22 +- 327 b 1-Nonanol 8.14 +- 1.46 a 11.33 +- 4.06 b 10.44 +- 1.59 b Octanoic acid 6334 +- 695.2 a 6620.89 +- 1678.83 a 8288.22 +- 730.84 b 1-Octanol 37.71 +- 5.59 ab 35.78 +- 11.02 a 48.56 +- 8.9 b Propyl acetate 2.26 +- 0.36 a 3.3 +- 0.81ab 4.06 +- 0.94 b Note: different letters in the same line indicate significant differences between conditions (Scheirer-Ray-Hare and Dunn tests, p < 0.05). foods-12-00972-t007_Table 7 Table 7 Thiamine-dependent volatile contents in the final wines. Compound (mg/L) Initial Thiamine Concentration (mg/L) 0 50 250 1-Butanol 255.5 +- 60.25 a 193 +- 89 ab 152.67 +- 69.36 b Decanoic acid 1400.25 +- 235.59 a 1463.75 +- 198.13 a 1966.78 +- 343.87 b 1-Decanol 15.59 +- 3 a 17.71 +- 1.72 a 23.29 +- 4.47 b 3-Ethoxy-1-propanol 2002.12 +- 328.39 a 1375.25 +- 200.96 b 1027 +- 263.77 b Ethyl benzaldehyde 2.89 +- 0.35 a 2.74 +- 0.45 a 3.49 +- 0.33 b Ethyl decanoate 19.62 +- 4.72 a 23.75 +- 4.3 ab 37.33 +- 11.79 b Ethyl laurate 2.67 +- 0.93 a 3.36 +- 0.7 a 5.1 +- 1.28 b Isovaleric acid 2944.5 +- 235.19 a 2973.38 +- 282.2 a 3434.78 +- 455.67 b Phenyl ethyl acetate 2.67 +- 0.67 a 3.6 +- 0.61 a 8.1 +- 3.15 b Phenyl ethyl alcohol 20,728.25 +- 4410.14 a 21,323.5 +- 4424.84 a 35,286 +- 9176.91 b Note: different letters in the same line indicate significant differences between conditions (Scheirer-Ray-Hare and Dunn tests, p < 0.05). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Francis I.L. Newton J.L. Determining Wine Aroma from Compositional Data Aust. J. Grape Wine Res. 2005 11 114 126 10.1111/j.1755-0238.2005.tb00283.x 2. Rapp A. Mandery H. Wine Aroma Experientia 1986 42 873 884 10.1007/BF01941764 3. Chen L. Darriet P. Strategies for the Identification and Sensory Evaluation of Volatile Constituents in Wine Compr. Rev. Food Sci. Food Saf. 2021 20 4549 4583 10.1111/1541-4337.12810 34370385 4. Torrens J. Urpi P. Riu-Aumatell M. Vichi S. Lopez-Tamames E. Buxaderas S. Different Commercial Yeast Strains Affecting the Volatile and Sensory Profile of Cava Base Wine Int. J. Food Microbiol. 2008 124 48 57 10.1016/j.ijfoodmicro.2008.02.023 18423920 5. Estevez P. Gil M.L. Falque E. Effects of Seven Yeast Strains on the Volatile Composition of Palomino Wines Int. J. Food Sci. Technol. 2004 39 61 69 10.1046/j.0950-5423.2003.00755.x 6. Patel S. Shibamoto T. Effect of 20 Different Yeast Strains on the Production of Volatile Components in Symphony Wine J. Food Compos. Anal. 2003 16 469 476 10.1016/S0889-1575(03)00021-8 7. Vilanova M. Ugliano M. Varela C. Siebert T. Pretorius I.S. Henschke P.A. Assimilable Nitrogen Utilisation and Production of Volatile and Non-Volatile Compounds in Chemically Defined Medium by Saccharomyces cerevisiae Wine Yeasts Appl. Microbiol. Biotechnol. 2007 77 145 157 10.1007/s00253-007-1145-z 17846763 8. Jimenez-Lorenzo R. Bloem A. Farines V. Sablayrolles J.M. Camarasa C. How to Modulate the Formation of Negative Volatile Sulfur Compounds during Wine Fermentation? FEMS Yeast Res. 2021 21 1 15 10.1093/femsyr/foab038 9. Seguinot P. Ortiz-Julien A. Camarasa C. Impact of Nutrient Availability on the Fermentation and Production of Aroma Compounds Under Sequential Inoculation With M. Pulcherrima and S. Cerevisiae Front. Microbiol. 2020 11 1 19 10.3389/fmicb.2020.00305 32082274 10. Saerens S.M.G. Delvaux F. Verstrepen K.J. Van Dijck P. Thevelein J.M. Delvaux F.R. Parameters Affecting Ethyl Ester Production by Saccharomyces cerevisiae during Fermentation Appl. Environ. Microbiol. 2008 74 454 461 10.1128/AEM.01616-07 17993562 11. Gobert A. Tourdot-Marechal R. Morge C. Sparrow C. Liu Y. Quintanilla-Casas B. Vichi S. Alexandre H. Non-Saccharomyces Yeasts Nitrogen Source Preferences: Impact on Sequential Fermentation and Wine Volatile Compounds Profile Front. Microbiol. 2017 8 2175 10.3389/fmicb.2017.02175 29163451 12. Evers M.S. Roullier-Gall C. Morge C. Sparrow C. Gobert A. Alexandre H. Vitamins in Wine: Which, What for, and How Much? Compr. Rev. Food Sci. Food Saf. 2021 20 2991 3035 10.1111/1541-4337.12743 33884746 13. Ough C.S. Davenport M.A. Joseph K. Effects of Certain Vitamins on Growth and Fermentation Rate of Several Commercial Active Dry Wine Yeasts Am. J. Enol. Vitic. 1989 40 208 213 10.5344/ajev.1989.40.3.208 14. Xing H. Impact of Thiamine and Pyridoxine on Alcoholic Fermentations of Synthetic Grape Juice Washington State University Pullman, WA, USA 2007 15. Bataillon M. Rico A. Sablayrolles J.-M. Salmon J.M. Barre P. Early Thiamin Assimilation by Yeasts under Enological Conditions: Impact on Alcoholic Fermentation Kinetics J. Ferment. Bioeng. 1996 82 145 150 10.1016/0922-338X(96)85037-9 16. Ferreira D. Galeote V. Sanchez I. Legras J.L. Ortiz-Julien A. Dequin S. Yeast Multistress Resistance and Lag-Phase Characterisation during Wine Fermentation FEMS Yeast Res. 2017 17 1 11 10.1093/femsyr/fox051 17. Hosono K. Aida K. Uemura T. Effect of Pantothenic Acid on Respiratory Activity of Aerobically Grown Saccharomyces cerevisiae J. Gen. Appl. Microbiol. 1972 18 189 199 10.2323/jgam.18.189 18. Eschenbruch R. Sulfite and Sulfide Formation during Winemaking--A Review Am. J. Enol. Vitic. 1974 25 157 161 10.5344/ajev.1974.25.3.157 19. Wainwright T. Production of H2S by Yeasts: Role of Nutrients J. Appl. Microbiol. 1971 34 161 171 10.1111/j.1365-2672.1971.tb02275.x 20. Eglinton J. Henschke P. Can the Addition of Vitamins during Fermentation Be Justified? Aust. Grapegrow. Winemak. 1993 352 47 21. Hannemann W. Ausscheidung von Essigsaure Durch Garende Hefen und Die Reinigung und Charakterisierung Einer NADP-Spezifischen Aldehyd-Dehydrogenase Aus Saccharomyces cerevisiae Johannes Gutenberg-Universitat Mainz Mainz, Germany 1985 22. Ough C.S. Kunkee R.E. Effects of Acid Additions to Grape Juice on Fermentation Rates and Wine Qualities Am. J. Enol. Vitic. 1967 18 11 17 23. Bely M. Sablayrolles J.-M. Barre P. Description of Alcoholic Fermentation Kinetics: Its Variability and Significance Am. J. Enol. Vitic. 1990 41 319 324 10.5344/ajev.1990.41.4.319 24. Shinohara T. Oshida A. Yanagida F. Goto S. Vitamin Requirements of Wine Yeasts and Influence of Growth Temperature J. Inst. Enol. Vitic. 1996 31 1 8 25. Jones R.P. Pamment N. Greenfield P.F. Alcohol Fermentation by Yeasts--The Effect of Environmental and Other Variables Process Biochem. 1981 16 42 49 26. Wagele B. Witting M. Schmitt-Kopplin P. Suhre K. Masstrix Reloaded: Combined Analysis and Visualization of Tran-Scriptome and Metabolome Data PLoS ONE 2012 7 e39860 10.1371/journal.pone.0039860 22815716 27. Kanehisa M. Goto S. KEGG: Kyoto Encyclopedia of Genes and Genomes Nucleic Acids Res. 2000 28 27 30 10.1093/nar/28.1.27 10592173 28. Mangiafico S. Rcompanion: Functions to Support Extension Education Program Evaluation. R Package Version 2.4.18 2022 Available online: (accessed on 11 November 2022) 29. Kassambara A. Rstatix: Pipe-Friendly Framework for Basic Statistical Tests. R Package Version 0.7.0 2021 Available online: (accessed on 11 November 2022) 30. Oura E. Suomalainen H. Biotin and the Metabolism of Baker's Yeast J. Inst. Brew. 1978 84 283 287 10.1002/j.2050-0416.1978.tb03889.x 31. Bohlscheid J.C. Fellman J.K. Wang X.D. Ansen D. Edwards C.G. The Influence of Nitrogen and Biotin Interactions on the Performance of Saccharomyces in Alcoholic Fermentations J. Appl. Microbiol. 2007 102 390 400 10.1111/j.1365-2672.2006.03180.x 17241344 32. Castor J.G.B. Archer T.E. Amino Acids in Must and Wines, Proline, Serine and Threonine Am. J. Enol. Vitic. 1956 7 19 25 33. Muller E.H. Richards E.J. Norbeck J. Byrne K.L. Karlsson K.A. Pretorius G.H.J. Meacock P.A. Blomberg A. Hohmann S. Thiamine Repression and Pyruvate Decarboxylase Autoregulation Independently Control the Expression of the Saccharomyces cerevisiae PDC5 Gene FEBS Lett. 1999 449 245 250 10.1016/S0014-5793(99)00449-4 10338141 34. Chatterjee A. Abeydeera N.D. Bale S. Pai P.J. Dorrestein P.C. Russell D.H. Ealick S.E. Begley T.P. Saccharomyces cerevisiae THI4p Is a Suicide Thiamine Thiazole Synthase Nature 2011 478 542 546 10.1038/nature10503 22031445 35. Brion C. Ambroset C. Sanchez I. Legras J.L. Blondin B. Differential Adaptation to Multi-Stressed Conditions of Wine Fermentation Revealed by Variations in Yeast Regulatory Networks BMC Genom. 2013 14 681 10.1186/1471-2164-14-681 36. Li M. Petteys B.J. McClure J.M. Valsakumar V. Bekiranov S. Frank E.L. Smith J.S. Thiamine Biosynthesis in Saccharomyces cerevisiae Is Regulated by the NAD+-Dependent Histone Deacetylase Hst1 Mol. Cell. Biol. 2010 30 3329 3341 10.1128/MCB.01590-09 20439498 37. Shi X. Zou Y. Chen Y. Ying H. Overexpression of THI4 and HAP4 Improves Glucose Metabolism and Ethanol Production in Saccharomyces cerevisiae Front. Microbiol. 2018 9 1 10 10.3389/fmicb.2018.01444 29403456 38. Labuschagne P.W.J. Impact of Exogenous Thiamine on Kluyveromyces Marxianus under Oenological Conditions Master's Thesis Stellenbosch University Stellenbosch, South Africa 2020 39. Casalta E. Aguera E. Picou C. Rodriguez-Bencomo J.J. Salmon J.M. Sablayrolles J.M. A Comparison of Laboratory and Pilot-Scale Fermentations in Winemaking Conditions Appl. Microbiol. Biotechnol. 2010 87 1665 1673 10.1007/s00253-010-2652-x 20461506 40. Rollero S. Roberts S. Bauer F.F. Divol B. Agitation Impacts Fermentation Performance as Well as Carbon and Nitrogen Metabolism in Saccharomyces cerevisiae under Winemaking Conditions Aust. J. Grape Wine Res. 2018 24 360 367 10.1111/ajgw.12338 41. Jackson R. Wine Science: Principles, Practice, Perception Academic Press Cambridge, MA, USA 2008 10.1016/B978-0-12-373646-8.X5001-X 42. Perli T. Wronska A.K. Ortiz-Merino R.A. Pronk J.T. Daran J.M. Vitamin Requirements and Biosynthesis in Saccharomyces cerevisiae Yeast 2020 37 283 304 10.1002/yea.3461 31972058 43. Ribereau-Gayon P. Dubourdieu D. Doneche B.B. Lonvaud A.A. Towey J. Handbook of Enology, Volume 1: The Microbiology of Wine and Vinifications Wiley New York, NY, USA 2021 9781119584698 44. Delfini C. Castino M. Ciolfi G. Thiamine Addition to Musts in Order to Reduce Ketoacids Formation and Increase the SO2 Efficacy in Wines Riv. Di Vitic. Enol. 1980 33 572 589 45. Mizunaga T. Kuraishi H. Aida K.O. Levels of Biotin Enzymes during Unbalanced Growth and Death of Biotin-Deficient Yeast Cells J. Gen. Appl. Microbiol. 1973 19 1 9 10.2323/jgam.19.1 46. Mendes-Ferreira A. Mendes-Faia A. The Role of Yeasts and Lactic Acid Bacteria on the Metabolism of Organic Acids during Winemaking Foods 2019 9 1231 10.3390/foods9091231 32899297 47. Radler F. Schtitz H. Glycerol Production of Various Strains of Saccharomyces Am. J. Enol. Vitic. 1982 33 36 40 10.5344/ajev.1982.33.1.36 48. Hiltunen J.K. Chen Z. Haapalainen A.M. Wierenga R.K. Kastaniotis A.J. Mitochondrial Fatty Acid Synthesis--An Adopted Set of Enzymes Making a Pathway of Major Importance for the Cellular Metabolism Prog. Lipid Res. 2010 49 27 45 10.1016/j.plipres.2009.08.001 19686777 49. Hoja U. Marthol S. Hofmann J. Stegner S. Schulz R. Meier S. Greiner E. Schweizer E. HFA1 Encoding an Organelle-Specific Acetyl-CoA Carboxylase Controls Mitochondrial Fatty Acid Synthesis in Saccharomyces cerevisiae J. Biol. Chem. 2004 279 21779 21786 10.1074/jbc.M401071200 14761959 50. Wakil S.J. Titchener E.B. Gibson D.M. Evidence for the Participation of Biotin in the Enzymic Synthesis of Fatty Acids BBA--Biochim. Biophys. Acta 1958 29 225 226 10.1016/0006-3002(58)90177-X 51. Nykanen L. Nykanen I. Production of Esters By Different Yeast Strains in Sugar Fermentations J. Inst. Brew. 1977 83 30 31 10.1002/j.2050-0416.1975.tb03787.x 52. Miller G.H. Whisky Science: Appendix 1, Threshold Data Whisky Science Springer International Publishing Cham, Switzerland 2019 53. Gutierrez L.E. Effect of Some Vitamins and Micronutrient Deficiencies on the Production of Higher Alcohols by Saccharomyces cerevisiae Sci. Agric. 1993 50 484 489 10.1590/S0103-90161993000300024 54. Moat A.G. Ahmad F. Alexander J.K. Barnes I.J. Alteration in the Amino Acid Content of Yeast during Growth under Various Nutritional Conditions J. Bacteriol. 1969 98 573 578 10.1128/jb.98.2.573-578.1969 5784213 55. Morris C.P. Lim F. Wallace J.C. Yeast Pyruvate Carboxylase: Gene Isolation Biochem. Biophys Res. Commun. 1987 145 390 396 10.1016/0006-291X(87)91334-9 3036126 56. Labuschagne P.W.J. Divol B. Thiamine: A Key Nutrient for Yeasts during Wine Alcoholic Fermentation Appl. Microbiol. Biotechnol. 2021 105 953 973 10.1007/s00253-020-11080-2 33404836 57. Schymanski E.L. Jeon J. Gulde R. Fenner K. Ruff M. Singer H.P. Hollender J. Identifying Small Molecules via High Resolution Mass Spectrometry: Communicating Confidence Environ. Sci. Technol. 2014 48 2097 2098 10.1021/es5002105 24476540 58. Rivas-Ubach A. Liu Y. Bianchi T.S. Tolic N. Jansson C. Pasa-Tolic L. Moving beyond the van Krevelen Diagram: A New Stoichiometric Approach for Compound Classification in Organisms Anal. Chem. 2018 90 6152 6160 10.1021/acs.analchem.8b00529 29671593 59. Roullier-Gall C. David V. Hemmler D. Schmitt-Kopplin P. Alexandre H. Exploring Yeast Interactions through Metabolic Profiling Sci. Rep. 2020 10 1 10 10.1038/s41598-020-63182-6 31913322 60. Bordet F. Roullier-Gall C. Ballester J. Vichi S. Quintanilla-Casas B. Gougeon R.D. Julien-Ortiz A. Kopplin P.S. Alexandre H. Different Wines from Different Yeasts? "Saccharomyces cerevisiae Intraspecies Differentiation by Metabolomic Signature and Sensory Patterns in Wine" Microorganisms 2021 9 2327 10.3390/microorganisms9112327 34835452 61. Hazelwood L.A. Daran J.M. van Maris A.J.A. Pronk J.T. Dickinson J.R. The Ehrlich Pathway for Fusel Alcohol Production: A Century of Research on Saccharomyces cerevisiae Metabolism Appl. Environ. Microbiol. 2008 74 2259 2266 10.1128/AEM.02625-07 18281432 62. Masi A. Mach R.L. Mach-Aigner A.R. The Pentose Phosphate Pathway in Industrially Relevant Fungi: Crucial Insights for Bioprocessing Appl. Microbiol. Biotechnol. 2021 105 4017 4031 10.1007/s00253-021-11314-x 33950280 63. Xiao W. Rank G.H. Branched Chain Amino Acid Regulation of the ILV2 Locus in Saccharomyces cerevisiae Genome 1990 33 596 603 10.1139/g90-088 2227406 64. Wiese J. Propionate Metabolism in Yeast and Plants Ph.D. Thesis Heinrich Heine Universitat Dusseldorf Dusseldorf, Germany 2014 65. Reiter T. Montpetit R. Byer S. Frias I. Leon E. Viano R. Mcloughlin M. Halligan T. Hernandez D. Figueroa-Balderas R. Transcriptomics Provides a Genetic Signature of Vineyard Site and Offers Insight into Vintage-Independent Inoculated Fermentation Outcomes mSystems 2021 6 e00033-21 10.1128/mSystems.00033-21 33850038 |
PMC10000646 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050701 healthcare-11-00701 Brief Report Overlooking of Individuals with Cardiometabolic Risk by Evaluation of Obesity Using Waist Circumference and Body Mass Index in Middle-Aged Japanese Women Wakabayashi Ichiro Tian Yaohua Academic Editor Ma Jixuan Academic Editor Department of Environmental and Preventive Medicine, School of Medicine, Hyogo Medical University, Nishinomiya 663-8501, Hyogo, Japan; [email protected]; Tel.: +81-798-45-6561; Fax: +81-798-45-6563 27 2 2023 3 2023 11 5 70106 2 2023 24 2 2023 24 2 2023 (c) 2023 by the author. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Waist circumference is often used for the diagnosis of visceral obesity and metabolic syndrome. In Japan, obesity in women is defined by the government as a waist circumference of >=90 cm and/or BMI of >=25 kg/m2. However, there has been a controversy for almost two decades as to whether waist circumference and its above-optimal cutoff are appropriate for the diagnosis of obesity in health checkups. Instead of waist circumference, the waist-to-height ratio has been recommended for the diagnosis of visceral obesity. In this study, the relationships between the waist-to-height ratio and cardiometabolic risk factors, including diabetes, hypertension and dyslipidemia, were investigated in middle-aged Japanese women (35~60 years) who were diagnosed as not having obesity according to the above Japanese criteria of obesity. The percentage of subjects showing normal waist circumference and normal BMI was 78.2%, and about one-fifth of those subjects (16.6% of the overall subjects) showed a high waist-to-height ratio. In subjects with normal waist circumference and normal BMI, odds ratios of high vs. not high waist-to-height ratio for diabetes, hypertension and dyslipidemia were significantly higher than the reference level. A considerable proportion of women who have a high cardiometabolic risk might be overlooked at annual lifestyle health checkups in Japan. body mass index cardiometabolic risk obesity waist circumference waist-to-height ratio Japan Society for the Promotion of Science21H03386 The present work was funded by the Grant-in-Aid for Scientific Research (No. 21H03386) from the Japan Society for the Promotion of Science. pmc1. Introduction Waist circumference is used for diagnosing metabolic syndrome in the criteria by the International Diabetes Federation . In Japan, individuals aged 40~74 years are legally obliged to receive annual lifestyle health checkups for the prevention of cardiovascular diseases. In health checkups, obesity is a required item for diagnosing metabolic syndrome and is defined by the Government of Japan as a waist circumference of >=90 cm and/or BMI of >=25 kg/m2 in women . However, there is criticism about the large optimal cutoff of waist circumference in the criteria . In fact, in our previous study using ROC analysis for relationships between obesity indices and diabetes, the age-dependent optimal cutoffs of waist circumference for Japanese women and men aged 35~70 years were 80~82 cm and 84~85 cm, respectively . In Chinese adults, the optimal cutoffs have been reported to be 75 cm in women and 85 cm in men . Thus, there is a possibility that a considerable proportion of women with high cardiometabolic risk are overlooked in health checkups using the above criteria of metabolic syndrome in Japan. A previous meta-analysis study showed that the waist-to-height ratio is better than BMI and waist circumference for discriminating cardiometabolic risk , whereas it was shown in another meta-analysis study that there was no evidence supporting higher suitability of waist-to-height ratio than BMI and waist circumference . It is not known what proportion of individuals receiving health checkups are overlooked when the above criteria of obesity using waist circumference and BMI are used for the evaluation of cardiometabolic risk in Japanese women. In this concise study, the relationships between waist-to-height ratio and cardiometabolic risk factors, including diabetes, hypertension and dyslipidemia, were therefore investigated in middle-aged women who were diagnosed as not having obesity according to the above Japanese criteria of metabolic syndrome using waist circumference and BMI. 2. Methods 2.1. Subjects The subjects were 18,574 Japanese women aged 35~60 years who underwent annual health checkups at their workplaces. This study was approved by the Hyogo College of Medicine Ethics Committee (No. 3003 in 2020). Histories of alcohol consumption, cigarette smoking, regular exercise, medication therapy and illness were surveyed by questionnaires. In the self-written questionnaire paper, participants were first asked "Are you a habitual cigarette smoker?" Cigarette smokers were defined as participants who had smoked for 6 months or longer and had smoked for the past month or longer. Then the participants who had been smokers were further asked "What is your average cigarette consumption per day?" The response categories for this question were "20 or less cigarettes per day", "21 or more and 40 or less cigarettes per day" and "41 or more cigarettes per day". Because there were no subjects with very heavy cigarette consumption (41 or more cigarettes per day), the subjects were divided into three groups in this study: nonsmokers, light smokers (20 or less cigarettes per day) and heavy smokers (21 or more cigarettes per). The average alcohol consumption of each subject per week was also reported on questionnaires. Frequency of habitual alcohol drinking was asked in the questionnaire as "How frequently do you drink alcohol?" The frequency of weekly alcohol drinking was categorized as "every day" (regular drinkers), "sometimes" (occasional drinkers) and "never" (nondrinkers). Subjects with a habit of regular exercise were defined as those doing exercise almost every day for 30 min or longer per day. Thus, the subjects were divided into two groups: those with and those without a habit of regular exercise. A quantitative evaluation of regular exercise was not performed in this study. 2.2. Measurements Height, body weight and waist circumference at the navel level (according to the definition of the Japanese Committee for the Diagnostic Criteria of Metabolic Syndrome ) were measured. Large waist circumference and high BMI were defined as >=90 cm and >=25 kg/m2, respectively. High waist-to-height ratios were defined as >=0.51 (35~49 years of age), 0.52 (50~59 years of age) and 0.54 (60 years of age) . Blood pressure was measured by trained nurses, who were part of the local health-checkup company, with a mercury sphygmomanometer once on the day of the health checkup after each subject had rested quietly in a sitting position. Korotkoff phase V was used to define diastolic pressure. Fasted blood was collected from each subject in the morning. Serum triglycerides, HDL cholesterol and LDL cholesterol were measured by enzymatic methods using commercial kits, pureauto S TG-N, cholestest N-HDL and cholestest LDL (Sekisui Medical Co., Ltd., Tokyo, Japan), respectively. Hemoglobin A1c was measured by the NGSP (National Glycohemoglobin Standardization Program)-approved technique using the latex cohesion method with a commercial kit (Determiner HbA1c, Kyowa Medex, Tokyo, Japan). Since the standards of hemoglobin A1c used for measurement are different in the NGSP method and JDS (the Japan Diabetes Society) method, hemoglobin A1c values were calibrated by using a formula proposed by JDS : hemoglobin A1c (NGSP) (%) = 1.02 x hemoglobin A1c (JDS) (%) + 0.25%. The criterion for each cardiometabolic risk factor was defined as follows: high blood pressure, systolic blood pressure >= 140 mmHg and/or diastolic blood pressure >= 90 mmHg ; low HDL cholesterol, HDL cholesterol < 50 mg/dL; high triglycerides, triglycerides >= 150 mg/dL ; diabetes, hemoglobin A1c >= 6.5% . Subjects receiving drug therapy for hypertension, dyslipidemia and diabetes were also included in the above definitions of hypertension, dyslipidemia and diabetes, respectively. 2.3. Statistical Analyses Statistical analyses were performed using a computer software program (IBM SPSS Statistics for Windows, Version 25.0. IBM Corp, Armonk, NY, USA). In logistic regression analysis, crude and adjusted odds ratios of subjects with a high waist-to-height ratio vs. those without a high waist-to-height ratio were calculated for diabetes, hypertension or dyslipidemia. In multivariable logistic regression analysis, age, BMI, and/or habits of smoking, alcohol drinking and regular exercise were used as other explanatory variables. Categorical variables were compared by using Pearson's chi-square test. Continuous variables were compared between the two groups with and without a high waist-to-height ratio by Student's unpaired t-test in univariable analysis and analysis of covariance (ANCOVA) followed by Student's t-test with Bonferroni's multiplicity correction in multivariable analysis. Age, habits of smoking, alcohol drinking and regular exercise, and current history of medication therapy for hypertension, dyslipidemia or diabetes were adjusted in ANCOVA. Probability (p) values less than 0.05 were defined as significant. 3. Results Table 1 shows characteristics of the overall subjects and subjects not showing large waist circumferences and high BMI. The percentages of subjects showing large waist circumference, high BMI and high waist-to-height ratio were 11.6%, 20.3% and 37.1%, respectively. The percentage of subjects showing normal waist circumference and a high waist-to-height ratio was 25.4%, while no subjects showed large waist circumference and normal waist-to-height ratio. The percentage of subjects showing normal waist circumference and normal BMI was 78.2%, and about one-fifth of those subjects (16.6% of the overall subjects) showed a high waist-to-height ratio. The mean with standard deviation and the range of waist circumferences (cm) of the subjects not showing high BMI and large waist circumference but showing high waist-to-height ratio was 83.2 +- 3.2 (71.0~89.5). In overall subjects and subjects showing normal waist circumference and normal BMI, logistic regression analysis was performed using the waist-to-height ratio (subjects with vs. subjects without high waist-to-height ratio) as an explanatory variable and each of the cardiometabolic risk factors (diabetes, hypertension and dyslipidemia) as an independent variable (Table 2). In subjects with normal waist circumference and normal BMI, crude and adjusted odds ratios of high vs. not high waist-to-height ratio for diabetes and dyslipidemia were all significantly higher than the reference level and tended to be decreased when age and BMI were adjusted. The odds ratios for diabetes and dyslipidemia remained significant (odds ratios: 1.79 [1.28~2.51] for diabetes and 1.54 [1.40~1.70) for dyslipidemia) when age, BMI and lifestyles (histories of smoking, alcohol drinking and regular exercise) were adjusted. Crude and lifestyle-adjusted odds ratios for hypertension were significantly higher than the reference level, while the odds ratios were not significantly different from the reference level when BMI was adjusted. Tendencies that are similar to the tendencies of the results in analyses using subjects who showed normal waist circumference and normal BMI were found in overall subjects, although the odds ratios tended to be higher in overall subjects than in subjects showing normal waist circumference and normal BMI. Table 3 shows the results of comparisons by univariable analysis of cardiometabolic risk factors between the subjects without and with a high waist-to-height ratio in the subject group not showing large waist circumference and high BMI. Prevalences of diabetes, hypertension and dyslipidemia and percentages of subjects receiving medication therapy for diabetes, hypertension and dyslipidemia were significantly higher in the subjects with a high waist-to-height ratio than in the subjects without a high waist-to-height ratio. Hemoglobin A1c, systolic and diastolic blood pressure, log-transformed triglycerides and LDL cholesterol were significantly higher and HDL cholesterol was significantly lower in the subjects with a high waist-to-height ratio than in the subjects without a high waist-to-height ratio. These differences in cardiometabolic risk factors between the subject groups with and without high waist-to-height ratio were also found in multivariable analysis with adjustment for age, habits of smoking, alcohol drinking and regular exercise, and current history of medication therapy for diabetes, hypertension or dyslipidemia (Table 4). 4. Discussion About three-fourths of the overall subjects had normal waist circumference and normal BMI levels. However, about one-fifth of those subjects had high waist-to-height ratios and showed significantly higher risks for diabetes, hypertension and dyslipidemia than those not having high waist-to-height ratios in univariable analysis and lifestyle-adjusted multivariable analysis. The odds ratios for diabetes and dyslipidemia remained significant even after adjustment for BMI. Interestingly, a high waist-to-height ratio was not associated with hypertension when BMI was adjusted in the multivariable analysis. This finding is reasonable since blood pressure, which strongly depends on cardiac output, is directly influenced by general obesity but not by visceral obesity. The mean levels of the variables of cardiometabolic risk factors were also significantly different between the subject groups with and without a high waist-to-height ratio. The above findings suggest that in the group of subjects showing normal waist circumference and normal BMI levels, subjects with a high waist-to-height ratio have a higher cardiometabolic risk than subjects without a high waist-to-height ratio. This indicates that a considerable proportion (16.6% in this study) of women who have a high cardiometabolic risk were overlooked at annual lifestyle health checkups in Japan. No subjects showed a large waist circumference and normal waist-to-height ratio; in other words, all subjects with large waist circumferences showed a high waist-to-height ratio. Therefore, waist-to-height ratio, instead of waist circumference, is better for diagnosing obesity more correctly in health checkups for the purpose of early prevention of cardiovascular diseases in middle-aged women. The above criteria of obesity using waist circumference and BMI have been used since 2008 when a new law about lifestyle health checkups was established in Japan. However, there has been a controversy for almost two decades as to whether waist circumference and its above-optimal cutoff for women are appropriate for the diagnosis of obesity in health checkups . In the present study, using the cutoff of waist circumference for women (90 cm or larger) in health checkups, the proportion of subjects diagnosed as having a large waist circumference was 11.6%, which was much lower than the proportions of subjects diagnosed as having high BMI (20.3%) and high waist-to-height ratio (37.1%). Thus, it is strongly suspected that a considerable number of individuals with visceral obesity are overlooked when the current Japanese criteria for obesity for women are used. The waist-to-height ratio is a convenient variable because the appropriate waist circumference is estimated by an easy calculation using an approximately optimal cutoff value of 0.5, i.e., the upper limit of the appropriate waist circumference (cm) is approximately half of the height (cm). A waist-to-height ratio of 0.5 or higher has been demonstrated to be a more favorable criterion for the identification of higher metabolic individuals than the criteria of other anthropometric indices including BMI, waist circumference and waist-to-hip ratio in Asians . Age-dependent cut-off values obtained in ROC analysis with diabetes as an outcome were used in this study. Using these criteria, 37.1% of overall subjects showed a high waist-to-height ratio. Further studies are needed to determine practically useful optimal cut-off values of waist-to-height ratio for predicting the risk of future cardiovascular events. 5. Conclusions A considerable proportion (16.6% in this study) of women who have a high cardiometabolic risk are thought to be overlooked at annual lifestyle health checkups in Japan. Waist-to-height ratio, instead of waist circumference, is recommended to be used for the diagnosis of visceral obesity in health checkups. Institutional Review Board Statement This study was approved by the Hyogo College of Medicine Ethics Committee (No. 3003 in 2020). Informed Consent Statement The database used in this study was supplied from a local health checkup system without individual identification, and no informed consent was obtained from each subject. Data Availability Statement The data can be obtained by reasonable request from the corresponding author on reasonable requests. Conflicts of Interest The author declares no conflict of interest. healthcare-11-00701-t001_Table 1 Table 1 Characteristics of the overall subjects and subjects not showing large waist circumference and high BMI. Overall Without Large WC and High BMI Number 18,574 14,529 Age (years) 47 9 +- 7.0 47.6 +- 7.0 Height (cm) 156.4 +- 5.6 156.6 +- 5.5 Weight (kg) 55.0 +- 9.3 51.5 +- 5.9 BMI (kg/m2) 22.5 +- 3.6 21.0 +- 2.1 WC (cm) 78.1 +- 9.7 74.6 +- 6.8 WHtR 0.500 +- 0.064 0.477 +- 0.046 High BMI (%) 20.3 0 Large WC (%) 11.6 0 High WHtR 37.1 21.3 Therapy for diabetes (%) 1.4 0.7 Therapy for hypertension (%) 8.4 5.8 Therapy for dyslipidemia (%) 4.5 3.2 Diabetes (%) 2.6 1.4 Hypertension (%) 20.2 14.2 Dyslipidemia (%) 36.2 23.2 Drinkers (%) Occasional 30.0 30.5 Regular 9.4 10.1 Smokers (%) Light 17.5 17.9 Heavy 0.8 0.8 Regular exercise (%) 5.3 5.4 Hemoglobin A1c (%) 5.38 +- 0.52 5.32 +- 0.42 Systolic BP (mmHg) 121.9 +- 17.0 119.0 +- 15.9 Diastolic BP (mmHg) 72.3 +- 11.2 70.4 +- 10.5 Triglycerides (mg/dL) 78 (55, 115) 72 (52, 103) HDL cholesterol (mg/dL) 65.3 +- 14.9 67.4 +- 14.8 LDL cholesterol (mg/dL) 114.0 +- 30.0 110.6 +- 29.1 Shown are numbers, proportions, means with standard deviations and medians with interquartile ranges in parentheses of the variables. WC, waist circumference; WHtR, waist-to-height ratio; BP, blood pressure. healthcare-11-00701-t002_Table 2 Table 2 Odds ratios for diabetes, hypertension and dyslipidemia of subjects with high waist-to-height ratio vs. subjects without high waist-to-height ratio in the overall subjects and subjects not showing large waist circumference and high BMI. Overall (n = 18574) Without Large WC and High BMI (n = 14529) Crude OR Diabetes 5.56 (4.49~6.87) * 3.11 (2.35~4.14) * Hypertension 3.46 (3.21~3.72) * 2.05 (1.85~2.27) * Dyslipidemia 3.49 (3.27~3.71) * 2.63 (2.43~2.86) * Age-adjusted OR Diabetes 4.82 (3.89~5.96) * 2.38 (1.78~3.17) * Hypertension 3.04 (2.82~3.28) * 1.62 (1.45~1.80) * Dyslipidemia 3.18 (2.98~3.39) * 2.27 (2.09~2.47) * Age and lifestyle-adjusted OR Diabetes 4.81 (3.88~5.96) * 2.42 (1.81~3.23) * Hypertension 3.06 (2.83~3.30) * 1.62 (1.46~1.80) * Dyslipidemia 3.19 (2.99~3.41) * 2.30 (2.11~2.51) * Age, lifestyle and BMI-adjusted OR Diabetes 1.97 (1.52~2.56) * 1.79 (1.28~2.51) * Hypertension 1.02 (0.92~1.13) 1.03 (0.91~1.16) Dyslipidemia 1.64 (1.50~1.79) * 1.54 (1.40~1.70) * Shown are odds ratios with 95% confidence intervals in parentheses. Crude and adjusted odds ratios of subjects with high waist-to-height ratio vs. those without high waist-to-height ratio were calculated for diabetes, hypertension or dyslipidemia. In multivariable logistic regression analysis, age, BMI, and/or lifestyle (habits of smoking, alcohol drinking and regular exercise) were used as other explanatory variables. OR, odds ratio; WC, waist circumference. Asterisks denote significantly high odds ratios compared with the reference level of 1.00 (*, p < 0.01). healthcare-11-00701-t003_Table 3 Table 3 Comparisons of cardiometabolic risk factors between subjects without and with a high waist-to-height ratio in the subject group not showing large waist circumference and high BMI in univariable analysis. Without High WHtR (n = 11,441) With High WHtR (n = 3088) WHtR 0.460 (0.459~0.460) 0.539 (0.538~0.540) * Diabetes (%) 0.94 2.88 * Hypertension (%) 12.1 22.1 * Dyslipidemia (%) 25.0 46.8 * Therapy for diabetes (%) 0.52 1.55 * Therapy for hypertension (%) 12.1 22.1 * Therapy for dyslipidemia (%) 2.67 5.08 * Hemoglobin A1c (%) 5.289 (5.282~5.296) 5.429 (5.409~5.448) * Systolic blood pressure (mmHg) 117.8 (117.5~118.1) 123.6 (123.1~124.2) * Diastolic blood pressure (mmHg) 69.6 (69.5~69.8) 73.1 (72.7~73.4) * Log[triglycerides (mg/dL)] 1.849 (1.845~1.853) 1.969 (1.961~1.977) * LDL cholesterol (mg/dL) 108.0 (107.5~108.5) 120.3 (119.2~121.4) * HDL cholesterol (mg/dL) 68.5 (68.3~68.8) 63.0 (62.5~63.5) * Shown are percentages and means with 95% confidence intervals of the variables. Triglycerides did not show a normal distribution and were used for analysis after logarithmic transformation with a base of 10. WHtR, waist-to-height ratio. Asterisks denote significant differences from subjects without a high waist-to-height ratio (*, p < 0.01). healthcare-11-00701-t004_Table 4 Table 4 Comparisons of cardiometabolic risk factors between subjects without and with a high waist-to-height ratio in the subject group not showing large waist circumference and high BMI in multivariable analysis. Without High WHtR (n = 11,441) With High WHtR (n = 3088) Hemoglobin A1c (%) 5.302 (5.295~5.309) 5.382 (5.369~5.395) * Systolic blood pressure (mmHg) 118.3 (118.0~118.5) 121.9 (121.4~122.4) * Diastolic blood pressure (mmHg) 69.9 (69.7~70.1) 72.1 (71.7~72.4) * Log[triglycerides (mg/dL)] 1.853 (1.849~1.857) 1.953 (1.946~1.961) * LDL cholesterol (mg/dL) 108.7 (108.2~109.2) 117.6 (116.6~118.5) * HDL cholesterol (mg/dL) 68.6 (68.3~68.8) 62.8 (62.3~63.3) * Shown are means with 95% confidence intervals of the variables. Triglycerides did not show a normal distribution and were used for analysis after logarithmic transformation with a base of 10. Age, habits of smoking, alcohol drinking and regular exercise, and current history of medication therapy for hypertension, dyslipidemia or diabetes were adjusted in ANCOVA. WHtR, waist-to-height ratio. Asterisks denote significant differences from subjects without a high waist-to-height ratio (*, p < 0.01). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Alberti K.G. Zimmet P. Shaw J. IDF Epidemiology Task Force Consensus Group The metabolic syndrome--A new worldwide definition Lancet 2005 366 1059 1062 10.1016/S0140-6736(05)67402-8 16182882 2. Matsuzawa Y. Metabolic syndrome--definition and diagnostic criteria in Japan J. Atheroscler. Thromb. 2005 12 301 10.5551/jat.12.301 16394611 3. Yamagishi K. Iso H. The criteria for metabolic syndrome and the national health screening and education system in Japan Epidemiol. Health 2017 39 e2017003 10.4178/epih.e2017003 28092931 4. Miyatake N. Wada J. Matsumoto S. Nishikawa H. Makino H. Numata T. Re-evaluation of waist circumference in metabolic syndrome: A comparison between Japanese men and women Acta Med. Okayama 2007 61 167 169 17593953 5. Nakamura K. Nanri H. Hara M. Higaki Y. Imaizumi T. Taguchi N. Sakamoto T. Horita M. Shinchi K. Tanaka K. Optimal cutoff values of waist circumference and the discriminatory performance of other anthropometric indices to detect the clustering of cardiovascular risk factors for metabolic syndrome in Japanese men and women Environ. Health Prev. Med. 2011 16 52 60 10.1007/s12199-010-0165-y 21432217 6. Wakabayashi I. Daimon T. Receiver-operated characteristics (ROCs) of the relationships between obesity indices and multiple risk factors (MRFs) for atherosclerosis at different ages in men and women Arch Gerontol. Geriatr. 2012 55 96 100 10.1016/j.archger.2011.06.028 21871669 7. Zeng Q. He Y. Dong S. Zhao X. Chen Z. Song Z. Chang G. Yang F. Wang Y. Optimal cut-off values of BMI, waist circumference and waist:height ratio for defining obesity in Chinese adults Br. J. Nutr. 2014 112 1735 1744 10.1017/S0007114514002657 25300318 8. Ashwell M. Gunn P. Gibson S. Waist-to-height ratio is a better screening tool than waist circumference and BMI for adult cardiometabolic risk factors: Systematic review and meta-analysis Obes. Rev. 2012 13 275 286 10.1111/j.1467-789X.2011.00952.x 22106927 9. Sommer I. Teufer B. Szelag M. Nussbaumer-Streit B. Titscher V. Klerings I. Gartlehner G. The performance of anthropometric tools to determine obesity: A systematic review and meta-analysis Sci. Rep. 2020 10 12699 10.1038/s41598-020-69498-7 32728050 10. Japanese Society of Internal Medicine Metabolic Syndrome-Definition and Diagnostic Criteria in Japan J. Jpn. Soc. Int. Med. 2005 94 794 809 (In Japanese) 11. Kashiwagi A. Kasuga M. Araki E. Oka Y. Hanafusa T. Ito H. Tominaga M. Oikawa S. Noda M. Kawamura T. International clinical harmonization of glycated hemoglobin in Japan: From Japan Diabetes Society to National Glycohemoglobin Standardization Program values J. Diabetes Investig. 2012 3 39 40 10.1111/j.2040-1124.2012.00207.x 24843544 12. Flack J.M. Adekola B. Blood pressure and the new ACC/AHA hypertension guidelines Trends Cardiovasc. Med. 2020 30 160 164 10.1016/j.tcm.2019.05.003 31521481 13. Alberti K.G. Eckel R.H. Grundy S.M. Zimmet P.Z. Cleeman J.I. Donato K.A. Fruchart J.C. James W.P. Loria C.M. Smith S.C. Jr. Harmonizing the metabolic syndrome: A joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity Circulation 2009 120 1640 1645 19805654 14. Teramoto T. Sasaki J. Ishibashi S. Birou S. Daida H. Dohi S. Egusa G. Hiro T. Hirobe K. Iida M. Japan Atherosclerosis Society. Executive summary of the Japan Atherosclerosis Society (JAS) guidelines for the diagnosis and prevention of atherosclerotic cardiovascular diseases in Japan-2012 version J. Atheroscler. Thromb. 2013 20 517 523 10.5551/jat.15792 23665881 15. American Diabetes Association Diagnosis and classification of diabetes mellitus Diabetes Care 2010 33 (Suppl. 1) S62 S69 10.2337/dc10-S062 20042775 16. Ho S.Y. Lam T.H. Janus E.D. Hong Kong Cardiovascular Risk Factor Prevalence Study Steering Committee Waist to stature ratio is more strongly associated with cardiovascular risk factors than other simple anthropometric indices Ann. Epidemiol. 2003 13 683 691 10.1016/S1047-2797(03)00067-X 14599732 17. Lyu L.C. Hsu C.Y. Yeh C.Y. Lee M.S. Huang S.H. Chen C.L. A case-control study of the association of diet and obesity with gout in Taiwan Am. J. Clin. Nutr. 2003 78 690 701 10.1093/ajcn/78.4.690 14522726 18. Hsieh S.D. Muto T. The superiority of waist-to-height ratio as an anthropometric index to evaluate clustering of coronary risk factors among non-obese men and women Prev. Med. 2005 40 216 220 10.1016/j.ypmed.2004.05.025 15533532 19. Hsieh S.D. Muto T. Metabolic syndrome in Japanese men and women with special reference to the anthropometric criteria for the assessment of obesity: Proposal to use the waist-to-height ratio Prev. Med. 2006 42 135 139 10.1016/j.ypmed.2005.08.007 16376978 |
PMC10000647 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051089 foods-12-01089 Article Sensory Profile and Consumer Liking of Sustainable Salamis Differing in Wild Boar Meat and Seasoning Ingredients Addition Freschi Pierangelo Conceptualization Validation Resources Funding acquisition Braghieri Ada Conceptualization Methodology Validation Formal analysis Writing - original draft * Pacelli Corrado Validation Writing - review & editing Visualization Langella Emilia Investigation Riviezzi Amelia Maria Software Paolino Rosanna Data curation Cosentino Carlo Validation Supervision Project administration Hormann-Wallner Marlies Academic Editor Laureati Monica Academic Editor School of Agricultural, Forestry, Food and Environmental Sciences (SAFE), University of Basilicata, 85100 Potenza, Italy * Correspondence: [email protected]; Tel.: +39-0971-205044 03 3 2023 3 2023 12 5 108903 2 2023 22 2 2023 01 3 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). The production of game meat is a proven way of promoting sustainable food, which is also consistent with the proper management of the expansion of the wild boar population in Italy. In the present study, we investigated consumer response to sensory attributes and consumer preference for ten types of "cacciatore" salamis prepared with different mixtures of wild boar/pork (30/50 or 50/50) and spice ingredients. PCA analysis showed a clear characterization of the salamis based on the first component with the hot pepper powder and fennel types differing from the others. For the second component, salamis without flavorings could be discriminated by those flavored with aromatized garlic wine or with black pepper only. The main findings of the hedonic test revealed that products with hot pepper and fennel seeds received the highest ratings, as well as satisfactory acceptance in the consumer test sensory analysis for eight out of ten products. The panelists and consumers' ratings were influenced by the flavors used, but not by the ratio of wild boar to pork. This gives us the opportunity to produce more cost-effective and environmentally friendly products, as doughs with a high proportion of wild boar meat can be used without affecting product preference. cacciatore salami wild boar meat sensory properties consumer liking Department of Ambiente Energia e Territorio--Basilicata Region. Prog. Epos P.O. FESR19AE.2015//D.01297 This research was funded by the Department of Ambiente Energia e Territorio--Basilicata Region. Prog. Epos P.O. FESR [Grant number 19AE.2015//D.01297]. pmc1. Introduction The increase in harvested wild boar from less than 50,000 in the mid-1980s to more than 300,000 in recent times indicates the recent and significant demographic spread of wild boar throughout Italy ; on a global scale, the rapid and extensive global spread has made this species one of the 100 worst invasive animals and pests . Damage currently attributed to wild boar includes: agricultural crops, rangelands, commercial woodlands; predation and disease transmission to livestock such as African swine fever and pseudorabies ; environmental damage to native plant and wildlife species; wetlands and water quality in surface waters; green spaces (e.g., parks, suburban/urban landscapes); threats to human health , including infectious diseases caused by bacterial pathogens (e.g., Yersinia enterocolitica, Brucella suis, Salmonella spp., Leptospira spp., and Escherichia coli), endoparassitosis (e.g., Trichinella spp.), and collisions between wild boars and vehicles resulting in property damage, human injuries, and death . The application of culling plans to limit the damage of this species is leading to widespread availability of wild boar meat in many developed countries , resulting in an increased supply of processed products. Although a market for processed game meat from hunting has been established in France and Spain , game meat in Italy is primarily consumed by hunters and their families, and the lack of a structured food supply chain limits the distribution of these products to a few regions in central northern Italy, mainly Umbria and Tuscany . On the other hand, consumer interest in game meat, which is considered more "natural" and with excellent sensory and nutritional properties, is growing. Wild game meat is considered more "natural" as wild animals are not exposed to the stress associated with industrial breeding and, when properly hunted, do not have the stress of transport to the slaughterhouse . In addition, the use of game in meat products would also be more sustainable than the use of pure pork. The use of a fine grain of the dough and of flavorings could mitigate the organoleptic differences in a very heterogeneous raw material, mainly due to the different age and quality of the carcasses of the hunted animals. Furthermore, the guidelines recently issued in Italy for the hygiene of wild boar meat will certainly allay the food safety concerns of unsettled consumers, while also creating supply chains for wild boar meat in the context of its wide availability and increasing consumer demand . In light of this last point, and given the importance of consumer attitudes and purchasing behavior for various types of food preparation and for wild boar meat in particular , in the present study we investigated consumer response to sensory attributes and consumer preference for ten types of "cacciatore" salamis, prepared with various blends of wild boar, pork meat and spice ingredients. 2. Materials and Methods The study was carried out in the Basilicata region, where 7,225 wild boars were culled in the 2017 hunting season and which had a population of about 89,000 wild boars at last census. This region has a long tradition of producing cured meat as testified by Marco Gavio Apicius, between 25 BC and AD 37, who wrote down the recipe for the oldest lucanian salami in the second book of De re coquinaria, which became known as "lucanica". In this study we utilized a "Cacciatore" type salami, one of the most widespread types of salami in Italy, made from pork and various adipose components and spices, which originated in northern Italy during the Longobard invasions. 2.1. Raw Material The processing of wild boar meat began in the first ten days of November with the culling of four adult wild boars in the Appennino Lucano-Val d'Agri-Lagonegrese National Park: two males weighing 60 and 65 kg and two females weighing 50, and 53 kg. The animals were culled by selection hunters using the waiting method. This hunting strategy is distinguished by remarkably low ante mortem stress for the animals, as the hunters remain stationary in a predetermined location and wait for the animal to approach them . The carcasses were then sent to the cutting laboratory after a veterinary inspection and trichinoscopic examination. The pork cuts from a commercial hybrid line were supplied by the same plant where the raw meat was processed, stuffed, and then seasoned. Our research does not fall within Directive 63/210 of the European Parliament and of the Council on the protection of animals used for experimental purposes (transposed into Italian law by Legislative Decree 26/2014) and, thus, it does not require any authorization from the national competent authorities. The protocol code of the certification of our Ethics Committee is OpBA 05_2023_UNIBAS. 2.2. Product Processing Sausage making was performed according to a randomized block design with ten batches and three replicates. After rapid cooling at 0/2 degC, the wild boar meat obtained from the whole carcasses was put into the mincer and cut into pieces of the typical size (3-4 mm) of a "Cacciatore" salami. The pork cuts used (bacon and shoulder) were minced separately, in the same way as the wild boar meat. Ten different mixtures were made from the three sources of minced meat (wild boar meat, pork belly, and p rk shoulder), which differed in the ratio of wild boar to pork and in the seasoning ingredients added (i.e., garlic-flavored "Aglianico del Vulture" red wine, hot pepper, fennel seeds, black pepper, and wine), as depicted in Table 1. The garlic-flavored wine was obtained by steeping three garlic cloves, cut lengthwise, for 24 h per liter of wine. All the spices used in the experiment were commercial products. The same amounts of NaCl (23 g/kg), NaNO2 (50 p.p.m.), fructose (3 g/kg), and lactic ferments (1 mL/kg) were added to the ten batches. Lactic ferments consisted of a mixture of Lactobacillus sakei, Staphylococcus carnosus, and Staphylococcus xylosus (Sacco Clerici Comp.). Each dough was immediately filled into natural casings with a diameter of about 4 cm. Hand tied every 10 cm, the casings were hung in pairs on aluminum rods . The salamis were then weighed and placed in a drying cell at 22 degC and 85% relative humidity. The weight loss percentage ranged from 38.27 (30-CF) to 45.96% (50-CF) (Table A1). Thirty-six hours after filling, the salamis were sprayed with a suspension of Penicillium spores, which resulted in the formation of white molds on the outside of the casing. The pH was measured inside the sausage on days 2, 4, 6 and at the end of ripening after 30 days (Table A2) using a portable pH meter HI931410 (Hanna Instruments, Woonsocket, RI) and a combined glass electrode. The apparatus was calibrated with 4.01 and 7.01 buffer solutions, according to the manufacturer's methodology. The chemical, physical, and sensory properties of the cured meat were evaluated at the end of the curing phase (30 days at 16 degC and 65-70% RH). 2.3. Proximate Physical and Chemical Analyses The sausages were weighed immediately after their preparation and successively after 10, 20, and 30 days, at the end of the maturing process. At the end of the curing phase (30 d), color was measured on three 0.4 mm slices of each batch using a Spectrophotometer CM-2600d (Minolta Co., Osaka, Japan), using the illuminant A and 10deg observer in L*, a*, b* color space . The chemical analyses were carried out with NIRS (Near Infrared Reflectance Spectroscopy, Foodscan Foss, Hillerod, Denmark) at the end of ripening on each sausage sample (Table A3). 2.4. Sensory Analises 2.4.1. Panel Selection and Training Fifteen subjects were recruited among regular eaters of sausages (defined as consuming the product at least once a week). Ten panelists were selected (four males and six females, between 29 and 61 yr. of age) in accordance with ISO standards . For this purpose, the four basic tastes were used . For this purpose, sucrose (Carlo Erba, Milan, Italy), sodium chloride (Carlo Erba, Milan, Italy), citric acid (Carlo Erba, Milan, Italy) and quinine hydrochloride (Sigma-Aldrich, St Louis, MO, USA) at three levels each were used . The panelists were informed about the taste of each basic concentration. Then, a 10 mL quantity of high and low concentration for each taste solution was served blind. The panelists rinsed their mouths with filtered, de-ionized water between tests. De-ionized water was also used to prepare two blanks. Totaling ten samples (taste solutions and blanks) were presented in random order. The panelists had to identify the intensity (low and high) of each taste solution. The inability to recognize eight out of the 10 taste solutions was used as cutoff point for selection purposes . Afterwards, panelists were trained for the scale use . 2.4.2. Quantitative Descriptive Sensory Analysis A quantitative descriptive analysis (QDA) method was used to assess the sensory profile of the sausages. During preliminary sessions, the panelists were asked to taste some slices of the samples and, on the basis of the available literature , they were encouraged by the panel leader to describe their taste, odor, flavor, appearance and texture and to develop and agree on a consensus list of 22 attributes and their definitions (Table 2). Standard reference products specific to each identified attribute were administered, with two points of the scale anchored to the reference material. In particular, assessors were repeatedly exposed to the reference samples (three times), indicating the corresponding intensity levels. Subsequently, panelists re-assessed the two levels of intensity of each attribute in blind conditions. Tests were performed in a controlled sensory analysis laboratory , equipped with individual booths, under red lighting to mask color differences in the samples, except during the evaluation of appearance, carried out in white fluorescent lighting conditions. For each sample, two 0.4 mm slices (one for appearance, and one for odor/flavor and texture) were obtained using a commercial slicing machine, and immediately served to the panelists at room temperature (20-23 degC). Each sample, coded with three-digit randomized numbers, was served in random order and evaluated in triplicate. For each daily session, five samples were presented. Assessors had to drink a sip of still water at the beginning of the sensory evaluation and to eat unsalted crackers between samples to try to cancel the sensations caused by the previous sample. Attributes were rated on the basis of 100 mm unstructured lines with anchor points at each end (0 = absent, 100 = very strong). 2.4.3. Consumer Testing Seventy-eight consumers (average age 31 yr.s; 48% men and 52% women) participated in the test. Consumers were recruited among regular eaters of sausages (i.e., consuming the product at least once a week). Under white fluorescent lighting, each consumer evaluated three 0.4 cm thick slices of each sausage in the same controlled sensory analysis laboratory described for QDA. The samples were presented in a random order. Consumers had to drink a sip of still water at the beginning of the sensory evaluation and to eat unsalted crackers between samples to try to cancel the sensations caused by the previous sample. For each product, they expressed an overall liking and a liking for appearance, odor, flavor and texture. Consumers rated their liking on a 9-point hedonic scale, with "extremely unpleasant" (1) at the left end and "extremely pleasant" (9) at the right end . The panelists and all consumers were informed about publication of the study. 2.5. Statistical Analysis One-way analysis of variance was used to test the effect of the product affecting color parameters and chemical composition. To identify differences between products the value of least significant difference (LSD) was calculated. Data on batch weights and pH were analyzed using a mixed procedure with product (ten levels) as non-repeated factor and ripening time (four levels) as repeated factor. Sensory profile data were subjected to ANOVA with product (ten levels = two pork/wild boar ratios x five flavor combinations), assessor (ten levels), replicate (three levels) and their first order interactions as factors. Data on consumer test were analyzed by ANOVA with product (ten levels), gender (two levels), age (three levels = 18-39, 40-59, over 60 yr.) and their first order interactions as factors. Principal component analysis (PCA) was performed on the sensory profile and on pH and color parameters to study the relationship between sensory attributes and these parameters. Data were analyzed by R software . 3. Results 3.1. Color Characteristics All the color parameters, depicted in Table 3, were significantly affected by product. Lightness (p < 0.0001) and redness (p < 0.0001) was higher in 30 and 30-CF products, while 50-PGW salami showed lower L* value and 30-PW, 30-P and 50-PW had lower a* index. As for b* parameter (p = 0.025), 30-CF had the higher value and 50-PW the lower one. 3.2. Quantitative Descriptive Sensory Analysis The ANOVA showed that there were no significant interactions between product x replication or product x assessor, thus indicating that both the training program and the reference frame used were appropriate to reach high reliability of the panel, as the products were consistently evaluated both in different replications and by different assessors. Product significantly affected the perception of almost all sensory attributes, determining an appreciable differentiation in the different salamis (Table A4). In fact, as for appearance parameters, 50-CF sample was perceived with the highest meat color intensity (73.57 +- 2.69, p < 0.0001) and the lowest brightness (28.80 +- 2.66, p < 0.0001), while 30 salami showed the opposite intensities for the parameters considered (35.10 +- 2.69 and 51.37 +- 2.66, p < 0.0001, for meat color and brightness intensities, respectively). These two products differed markedly also for fennel and hot pepper odor, with the highest intensities in the 50-CF salami (63.13 +- 2.34 and 48.53 +- 2.02, p < 0.0001, for fennel and hot pepper odor, respectively) and the lowest in the 30 sample (3.73 +- 2.34 and 3.13 +- 2.02, p < 0.0001, for fennel and hot pepper odor, respectively). As for flavor attributes, the 30-PGW salami tended to show the highest overall intensity (44.17 +- 3.15, p = 0.068), especially when compared to the 30 salami (22.97 +- 5.28, p = 0.068). Fennel and hot pepper flavor perception confirmed what was found for odor, as the 50-CF salami was perceived with the highest intensities (64.47 +- 3.14 and 48.9 +- 2.60, p < 0.001, for fennel and hot pepper flavor, respectively) and the 30 sample with the lowest fennel flavor (4.90 +- 3.14, p < 0.0001) while the 50 sample showed the lowest hot pepper flavor (5.37 +- 2.60, p < 0.0001). This latest product was also perceived with the highest wild flavor intensity (23.57 +- 2.68, p = 0.001); on the contrary, the lowest wild flavor intensity was perceived in the 30-PW salami (7.83 +- 2.76, p = 0.001). Again, the 50-CF product had the highest intensity for spiced flavor (65.70 +- 2.97, p < 0.001) while the 50 salami showed the lowest intensity (11.07 +- 2.97, p < 0.0001). As for the texture profile, 50-CF sample showed the highest intensities for tenderness, cohesiveness and chewiness (57.90 +- 3.05, p = 0.023, 69.67 +- 2.27 and 67.63 +- 2.70, p < 0.0001, respectively) while 30 salami had the lowest values for these parameters (40.53 +- 3.05, p = 0.023, 47.57 +- 2.27 and 44.73 +- 2.70, p < 0.0001, respectively). The PCA bi-plot of the Cacciatore type salami sensory profile, salami pH and color parameters provided a multivariate graphical representation of the product space showing the relationship between the sensory attributes and the other variables. The first two principal components of PCA explained 65.04% of the variance in the data (46.69% for PC1 and 21.35% for PC2, respectively). In particular, pH (0.58), tenderness (0.62), cohesiveness (0.67), chewiness (0.62), hot pepper flavor (0.90) and odor (0.90), fennel flavor (0.90) and odor (0.94), spiced (0.80), meat color (0.67) and color uniformity (0.91) showed positive correlations with PC1, whereas color parameters, such as L* (-0.73), a* (-0.62), b* (-0.49), brightness (-0.64), wine odor (-0.43), garlic flavor (-0.34), and black pepper flavor (-0.28) were negatively correlated with this axis. On the contrary, PC2 showed positive correlations with wine odor (0.69), and wild flavor (0.89) and negative correlations with spiced (-0.27), chewiness (-0.40), cohesiveness (-0.41), and tenderness (-0.55). This allows a marked characterization of the salamis based on the first component. Products with added chili powder and fennel seeds (30-CF and 50-CF) were differentiated by the other products and located on the right side. On the second component, salamis without flavorings (codes 30 and 50) could be discriminated by those flavored with aromatized garlic wine (30-PGW and 50-PW) or with black pepper only (30-P and 50-P). 3.3. Consumer Testing The results of hedonic testing on 78 consumers are depicted in Table 4. Almost all the products were rated at scores above the neutral point (5 = neither pleasant nor unpleasant), showing that the tested products were perceived as being characterized by a good eating quality. Gender significantly affected all the consumer liking parameters considered, with men attributing the higher scores to the product. Looking at the age groups, appearance (p = 0.045), flavor (p = 0.002), and texture (p = 0.041) of the product were significantly rated best by the youngest consumers. Product factor significantly affected all the liking parameters. In particular, the sausages from the two batches with chili powder and fennel seeds (30-CF and 50-CF) received the highest scores for overall liking, appearance and odor (p = 0.001), and flavor (p = 0.031) and texture (p = 0.040). In contrast, the sausages that only contained black pepper (30-P and 50-P) were rated worst (Table 5). 4. Discussions Significant differences were found between products for color parameters. This result may be ascribed to the different formulations used in the manufacturing of the ten salamis. In particular, higher percentages of pork meat used in 30 and 30CF salamis may have produced their higher lightness, while the lower value of L* index in 50-PGW product may be due to higher percentage of wild boar meat together with garlic/red wine addition. As for processed meat, other authors found darker color (lower L* values) in wild boar hams compared with Yorkshire hams. Marchiori and de Felicio reported a darker coloration in the wild boar meat compared with pork. Game meat, in fact, has a typical dark red color due to a higher concentration of myoglobin as a result of the intense physical activity of wild animals . The QDA method allowed us to significantly characterize the different products. In agreement with Brankovic Lazic et al. , the ingredients used might have significantly affected sensory properties of cacciatore salami. In fact, the 50-CF sample, added with larger quantities of hot pepper powder and fennel seeds, stood out among other products and in particular from sample 30, for highest meat color intensity, fennel and hot pepper odor and flavor intensities, and for textural properties. Probably, hot pepper powder and wild boar meat produced higher meat color intensities. The sensory characterization of the products may be highlighted even more by PCA graphical representation, where 30-CF and 50-CF salamis stand out from the others on the first component. In addition, similar to stretched curd , pH was positively related to some textural attributes, such as cohesiveness, tenderness, and chewiness, whereas a negative correlation was found between pH and oiliness. Despite the fact that we are comparing two extremely different products in the raw material and in the transformation process, in dry-cured fermented sausages textural properties have been mainly related to pH, as stated by Gimeno et al. , explaining the variability of texture among different brands of Chorizo de Pamplona. In fact, pH evolution during the ripening process strongly affects the changes in textural attributes and if the pH falls below its isoelectric point, a firmer product is obtained . As for color parameters (L*, a*, b*), they were negatively related with brightness on the second component. In the consumer test, the products with chili powder and fennel seeds achieved higher liking scores compared with black pepper salami. On this regard, it is worth noting that in Basilicata, the region where the experiment was conducted, many typical cured meats are made with these flavors, so the consumers' familiarity with these products may have influenced their evaluations. The addition of wild boar meat did not affect consumer liking. Products significantly differed for all the liking parameters. As reported in previous studies on sausages and on hams , taste is the most important factor affecting purchasing and consumption of dry-cured products, followed by appearance and texture. A significant gender effect for all the liking parameters was observed, contrastingly with what was reported by Razmaite et al. for traditional sausages in Lithuania. Although we did not evaluate the effect of information on consumer liking for wild boar products, other studies showed a mostly positive attitude toward wild game products compared to women. In addition, age affected appearance, flavor, and texture liking, with higher scores from the younger consumers. Again, Razmaite et al. did not find any effect of age on sausages liking, except for the innovations with reduced salt content that were most favorably accepted by the middle-age generation. In addition, Czarniecka-Skubina et al. found that consumers aged 30-40 yr. with higher education and income readily accepted game meat and agreed to eat it in the future. On the contrary, other authors reported that young people have a more negative attitude towards wild game meat. The positive response to mixed salami could be an opportunity to also drastically reduce the carbon footprint and land use in the production of a widely used and appreciated product such as salami. In our case with a 50% mixed salami, we could halve the carbon footprint and land use values, which for pork are equivalent to 5 kgCO2-eq kg-1 meat and 40-75 m2 yr. kg-1 protein, respectively . 5. Conclusions Despite the increasing occurrence of wild boar, comparatively few studies deal with quality aspects of meat products of this species. The results of our study, indicating the predominantly satisfactory results of both panel and consumer tests, contribute to a better use of the resource dry-cured wild boar meat. Furthermore, our results show that consumer ratings were influenced by the flavors used, but not by the ratio of wild boar to pork. This opens up the possibility of producing more cost-effective products, as doughs with a high proportion of wild boar meat, in our case 50%, can be used without affecting the acceptability of the product. The development of a wild game meat supply chain could be an effective strategy for the supply of a sustainable alternative to production on intensive livestock farms, for the development of rural territories, and for controlling the growth of wild animal populations. The next studies will look at the variability associated with the effects of hunting methods and field dressing procedures as factors influencing the microbiological quality of the meat and consequently the sensory characteristics of the products made from it. In addition, it would be interesting to investigate the effect of information on the use of wild boar meat in the manufacturing of sausages on consumer liking. Acknowledgments We are grateful to Carlo Gilio and Sandrino Caffaro, from the "Dipartimento dell'Ambiente del Territorio e dell'Energia" of the Basilicata Region, for their help in finding the wild boar carcasses and in taking care of all the preparatory health conditions for processing the meat. Author Contributions Conceptualization, P.F. and A.B.; methodology, A.B.; software, A.M.R.; validation, P.F., A.B. and C.C.; formal analysis, A.B.; investigation, E.L.; resources, P.F.; data curation, R.P.; writing--original draft preparation, A.B.; writing--review and editing C.P.; visualization, C.P.; supervision, C.C.; project administration, C.C.; funding acquisition, P.F. All authors have read and agreed to the published version of the manuscript. Data Availability Statement Data are available on request. Conflicts of Interest The authors declare no conflict of interest. Appendix A foods-12-01089-t0A1_Table A1 Table A1 Average values of the batch weights (kg) of sausages during drying and curing, and total weight loss (%) (mean +- SE). Batch Days of Ripening Total Weight Loss, 0-30 d 0 10 20 30 30 2.70 +- 0.10 1.66 +- 0.16 1.58 +- 0.13 1.49 +- 0.11 44.81 30-PGW 2.71 +- 0.21 1.74 +- 0.12 1.65 +- 0.11 1.57 +- 0.10 42.07 30-PW 2.75 +- 0.40 1.72 +- 0.29 1.64 +- 0.27 1.57 +- 0.25 42.91 30-P 2.71 +- 0.41 1.78 +- 0.24 1.71 +- 0.20 1.64 +- 0.16 39.48 30-CF 2.59 +- 0.49 1.75 +- 0.40 1.67 +- 0.38 1.60 +- 0.36 38.22 50 2.70 +- 0.36 1.69 +- 0.26 1.60 +- 0.24 1.51 +- 0.21 44.07 50-PGW 2.62 +- 0.24 1.70 +- 0.30 1.58 +- 0.23 1.45 +- 0.24 44.66 50-PW 2.84 +- 0.11 1.73 +- 0.07 1.64 +- 0.06 1.54 +- 0.05 45.77 50-P 2.73 +- 0.04 1.76 +- 0.05 1.67 +- 0.03 1.57 +- 0.01 42.49 50-CF 2.72 +- 0.18 1.61 +- 0.11 1.54 +- 0.11 1.47 +- 0.11 45.96 p 0.056 0.184 0.078 0.265 foods-12-01089-t0A2_Table A2 Table A2 pH values on day 2, 4, 6 and 30 of ripening (mean +- SE) (1). Batch Days of Ripening 2 4 6 30 30 5.79 +- 0.08 5.76 +- 0.12 5.87 a +- 0.09 5.63 a +- 0.04 30-PGW 5.66 +- 0.05 5.46 +- 0.05 5.51 +- 0.03 5.96 +- 0.05 30-PW 5.63 +- 0.87 5.42 +- 0.08 5.54 +- 0.12 5.92 +- 0.09 30-P 5.57 +- 0.11 5.58 +- 0.1 5.58 +- 0.04 6.17 +- 0.87 30-CF 5.68 +- 0.04 5.48 +- 0.08 5.69 +- 0.06 6.19 b +- 0.07 50 5.71 +- 0.03 5.64 +- 0.05 5.77 +- 0.09 5.85 +- 0.08 50-PGW 5.61 +- 0.06 5.40 +- 0.05 5.24 b +- 0.13 5.81 +- 0.12 50-PW 5.60 +- 0.05 5.44 +- 0.05 5.58 +- 0.04 5.75 +- 0.05 50-P 5.56 +- 0.07 5.41 +- 0.03 5.53 +- 0.12 5.94 +- 0.10 50-CF 5.74 +- 0.07 5.52 +- 0.1 5.81 +- 0.08 5.74 +- 0.08 p 0.078 0.227 0.033 0.041 (1) a, b = p < 0.05. foods-12-01089-t0A3_Table A3 Table A3 Proximate chemical composition of the products at the end of ripening (mean +- SE) (1). Batch Dry Matter Collagen Connective Total Protein Fat NaCl Ash 30 67.94 a +- 1.97 2.47 +- 0.07 8.71 a +- 0.25 28.34 a +- 0.82 30.61 a +- 0.89 3.79 +- 0.11 8.03 +- 0.23 30-PGW 66.03 +- 1.91 2.62 +- 0.08 9.23 +- 0.27 30.03 +- 0.87 32.42 +- 0.94 4.02 +- 0.12 8.51 +- 0.25 30-PW 66.31 +- 1.92 2.6 +- 0.08 9.15 +- 0.27 29.78 +- 0.86 32.15 +- 0.93 3.98 +- 0.12 8.44 +- 0.24 30-P 64.86 +- 1.88 2.71 +- 0.08 9.55 b +- 0.28 31.06 +- 0.90 33.54 b +- 0.97 4.15 +- 0.12 8.08 +- 0.26 30-CF 64.78 b +- 1.88 2.71 +- 0.08 9.57 b +- 0.28 31.14 b +- 0.90 33.62 b +- 0.97 4.16 +- 0.12 8.82 +- 0.26 50 66.69 +- 1.93 2.57 +- 0.07 9.05 +- 0.26 29.44 +- 0.85 31.79 +- 0.92 3.94 +- 0.11 8.34 +- 0.24 50-PGW 66.89 +- 1.94 2.55 +- 0.07 9.00 +- 0.26 29.27 +- 0.85 31.61 +- 0.92 3.91 +- 0.11 8.29 +- 0.24 50-PW 67.25 +- 1.95 2.52 +- 0.07 8.90 +- 0.26 28.95 +- 0.84 31.26 +- 0.91 3.87 +- 0.11 8.20 +- 0.24 50-P 66.17 +- 1.92 2.61 +- 0.08 9.19 +- 0.27 29.90 +- 0.87 32.29 +- 0.94 4.00 +- 0.12 8.47 +- 0.25 50-CF 67.31 +- 1.95 2.52 +- 0.07 8.88 +- 0.26 28.90 +- 0.84 31.20 +- 0.90 3.86 +- 0.11 8.19 +- 0.24 p 0.044 0.349 0.023 0.028 0.029 0.088 0.213 (1) a, b = p < 0.05. foods-12-01089-t0A4_Table A4 Table A4 Sensory profile, as assessed by a 10-member trained panel, of Cacciatore salami (mean +- SE) (1) Attribute Batch p 30 30-PGW 30-PW 30-P 30-CF 50 50-PGW 50-PW 50-P 50-CF Appearance Color uniformity 29.70 +- 2.76 B 29.80 +- 2.76 B 37.22 +- 2.83 33.40 +- 2.76 39.40 +- 2.76 a 33.37 +- 2.76 B 39.07 +- 2.76 a 33.83 +- 7.76 B 34.57 +- 2.76 47.50 +- 2.76 Aa 0.001 Meat color 35.10 +- 2.69 Ba 38.53 +- 2.76 B 47.83 +- 2.76 A 37.47 +- 2.69 B 41.00+- 2.69 B 46.77 +- 2.69 A 61.33 +- 2.6 A 56.93 +- 2.69 A 51.27 +- 2.69 A 73.57 +- 2.69 A <0.0001 Fat color 54.37 +- 3.37 53.97 +- 3.37 52.54 +- 3.47 60.30 +- 3.37 Aa 47.30 +- 3.37 B 54.47 +- 3.37 55.17 +- 3.37 48.77 +- 3.37 b 48.47 +- 3.37 b 54.30 +- 3.37 0.227 Brightness 51.37 +- 2.66 A 49.47 +- 2.66 45.71 +- 2.73 Aa 47.43 +- 2.66 A 45.83 +- 2.66 47.40 +- 2.66 33.87 +- 2.66 B 39.37 +- 2.66 bB 36.43 +- 2.66 Bb 28.80 +- 2.66 B <0.0001 Fat percentage 58.63 +- 2.25 Aa 57.97 +- 2.25 A 55.97 +- 2.31 58.67 +- 2.25 A 53.23 +- 2.25 52.70 +- 2.25 49.10 +- 2.25 Bb 56.43 +- 2.25 56.53 +- 2.25 44.67 +- 2.25 B 0.002 Fat diameter 44.27 +- 2.87 A 50.03 +- 2.87 Aa 36.54 +- 2.94 B 38.00 +- 2.87 B 37.00 +- 2.87 B 44.07 +- 2.87 A 49.20 +- 2.87 A 45.73 +- 2.87 38.0 +- 2.87 b 44.37 +- 2.87 B 0.002 Exudate 39.40 +- 3.46 34.00 +- 3.46 B 48.62 +- 3.55 Aa 38.30 +- 3.46 39.77 +- 3.46 37.80 +- 3.46 38.96 +- 3.46 45.03 +- 3.46 A 36.83 +- 3.46 ab 34.23 +- 3.46 B 0.102 Odor Overall odor 22.97 +- 5.28 B 33.20 +- 5.28 B 38.24 +- 5.43 41.47 +- 5.28 A 39.17 +- 5.28 Aa 27.97 +- 5.28 b 31.07 +- 5.28 34.03 +- 5.28 25.86 +- 5.28 B 31.77 +- 5.28 B 0.258 Fennel 3.73 +- 2.34 B 10.30 +- 2.34 A 14.94 +- 2.40 a 13.40 +- 2.34 B 49.67 +- 2.34 C 6.17 +- 2.34 B 9.60 +- 2.34 6.73 +- 2.34 b 5.63 +- 2.34 63.13 +- 2.34 <0.0001 Wine 20.33 +- 2.82 18.27 +- 2.82 7.68+-2.90 B 15.43 +- 2.82 12.70 +- 2.82 14.77 +- 2.82 21.43 +- 2.82 A 13.20 +- 2.82 B 13.20 +- 2.82 B 12.47 +- 2.82 B 0.033 Garlic 14.97 +- 2.75 14.87 +- 2.75 10.89 +- 2.82 B 10.17 +- 2.75 Bb 8.63 +- 2.75 B 12.97 +- 2.75 18.37 +- 2.75 Aa 17.73 +- 2.75 A 12.20 +- 2.75 16.03 +- 2.75 A 0.195 Red hot pepp. 3.13 +- 2.02 B 3.73 +-2.02 B 7.13+-2.08 7.80 +- 2.02 B 47.30 +- 2.02 A 2.73 +- 2.02 B 6.50 +- 2.02 B 5.10 +- 2.02 B 4.43 +- 2.02 B 48.53 +- 2.02 A <0.0001 Taste Bitter 20.90 +- 3.43 B 19.00 +- 3.43 20.61 +- 3.53 23.90 +- 3.43 16.80 +- 3.43 B 21.87 +- 3.43 20.27 +- 3.43 30.00+- 3.43 A 27.20 +- 3.43 A 22.57 +- 3.43 0.249 Flavor Overall flavor 29.77 +- 3.15 Bb 44.17 +- 3.15 A 40.42 +- 3.24 39.70 +- 3.15 A 36.37 +- 3.15 32.47 +- 3.15 36.13 +- 3.15 38.63 +- 3.15 Aa 35.40 +- 3.15 33.17 +- 3.15 a 0.068 Fennel 4.90 +- 3.14 Bb 11.40 +- 3.14 B 12.41 +- 3.22 B 15.57 +- 3.14 Ba 63.20 +- 3.14 A 5.43 +- 3.14 B 9.43 +- 3.14 B 8.23 +- 3.14 B 9.40 +- 3.14 B 64.47 +- 3.14 A <0.0001 Red hot pepp. 6.23 +- 2.60 B 9.27 +- 2.60 8.89 +- 2.60 9.10 +- 2.60 B 46.50 +- 2.60 A 5.37 +- 2.60 B 9.00 +- 2.60 6.67 +- 2.60 B 8.57 +- 2.60 B 48.90 +- 2.60 A <0.0001 Black pepper 14.13 +- 3.62 Ca 41.43 +-3.62 72.99 +- 3.62 A 69.33 +- 3.62 A 19.77+- 3.62 a 4.27 +- 3.62 Bb 29.17 +- 3.62 42.03 +- 3.62 C 65.43 +- 3.62 A 15.30 +- 3.62 a <0.0001 Wild 22.33 +- 2.68 A 13.63 +- 2.68 7.83 +- 2.76 B 10.90 +- 2.68 10.70 +- 2.68 23.57 +- 2.68 Aa 19.03 +- 2.68 A 14.97 +- 2.68 b 14.53 +- 2.68 b 20.50 +- 2.68 A 0.001 Spiced 23.57 +- 2.97 24.87 +- 2.97 D 41.54 +- 3.05 40.00 +- 2.97 63.53 +- 2.97 A 11.07 +- 2.97 Bb 20.57 +- 2.97 Da 21.50 +-2.97 35.47 +- 2.97 C 65.70 +- 2.97 A <0.0001 Garlic 9.53 +- 2.77 B 15.50 +- 2.77 9.66 +- 2.84 B 10.13+- 2.77 6.83 +- 2.77 Bb 10.53 +- 2.77 20.06 +- 2.77 A 15.93 +- 2.77 a 14.83 +- 2.77 10.20 +- 2.77 0.030 Wine 14.83 +- 3.00 18.73 +- 3.00 15.06 +- 3.07 21.10 +- 3.00 a 12.33+- 3.00 13.90 +- 3.00 22.90 +- 3.00 A 18.13 +- 3.00 17.03 +- 3.00 10.97 +- 3.00 Bb 0.116 Texture Tenderness 40.53 +- 3.05 Bb 50.73 +- 3.05 53.21 +- 3.13 A 55.30 +- 3.05 A 52.33 +- 3.05 A 52.63 +- 3.05 A 52.93 +- 3.05 A 53.97 +- 3.05 A 51.57 +- 3.05 a 57.90 +- 3.05 A 0.023 Cohesiveness 47.57 +- 2.27 B 56.53 +- 2.27 C 59.77 +-2.34 C 61.97 +- 2.27 b 57.93 +- 2.27 C 60.33 +- 2.27 C 61.27 +- 2.27 65.07 +- 2.27 A 61.70 +- 2.27 b 69.67 +- 2.27 Aa <0.0001 Chewiness 44.73 +- 2.70 B 57.83 +- 2.70 B 58.56 +- 2.77 B 61.10 +- 2.70 B 55.43 +- 2.70 B 57.80 +- 2.70 B 61.37 +- 2.70 B 62.23 +- 2.70 B 58.13+- 2.70 B 67.63 +- 2.70 B <0.0001 Oiliness 44.60 +- 5.24 b 44.60 +- 5.24 B 46.91 +- 5.38 B 51.87 +- 5.24 b 49.43 +- 5.24 b 53.20 +- 5.24 67.03 +- 5.24 Aa 56.93 +- 5.24 49.50 +- 5.24 48.97 +- 5.24 0.170 (1) a, b, = p < 0.05; A, B, C, D = p < 0.01. Figure A1 Salami in the last seasoning period. Figure 1 Principal component analysis (PCA). foods-12-01089-t001_Table 1 Table 1 Ingredients utilized in the ten batches. Batch Proportions of Wild Boar Meat/Pork Shoulder/Pork Belly Ground Black Pepper Black Pepper Grains Hot Pepper Powder Fennel Seeds Wine Garlic Flavored Wine g/kg ml/kg 30 30/40/30 - - - - - - 30-PGW 1 4 - - - 30 30-PW 1 4 - - 30 - 30-P 1 4 - - - - 30-CF - - 5 2 - - 50 50/20/30 - - - - - - 50-PGW 1 4 - - - 30 50-PW 1 4 - - 30 - 50-P 1 4 - - - - 50-CF - - 5 2 - - foods-12-01089-t002_Table 2 Table 2 List of attributes and reference frame used by 10-member panel for Cacciatore salami sensory profiling. Attributes Definition Intensity Low (<20) High (>80) Appearance Color uniformity Presence of a darker external halo in the slice due to an anomalous drying process Two-month-seasoned sausage Bresaola Meat color Intensity of the characteristic red color of the lean of the sausage Red orange = 2.5 YR 1 Dark red = 10 RP 1 Fat color Intensity of fat color White = 10 Y 1 Pink = 10 R 1 Brightness Intensity of the characteristic red color (dark-light) of the cured sausage White Black Fat percentage Percentage of fat on the slice surface Bresaola Hungarian salami Fat diameter Mincing type of fat in the slice Hungarian salami Soppressata salami Exudate Amount of liquid fat on the surface Seasoned sausage Cacciatore salami Odor Overall odor Level of overall odor before eating the sample Fifteen-day-seasoned sausage Napoli salami Fennel Odor associated with fennel seed Cacciatore salami Lucanian sausage Wine Odor associated with red wine Water Red wine Flavor Overall flavor Level of overall flavor Fifteen-day-seasoned sausage Napoli salami Fennel Flavor associated with fennel seed Low High Black pepper Flavor associated with the presence of sweet pepper powder Lucanian sausage Napoli salami Hot pepper Flavor associated with hot pepper Seasoned sausage Lucanian sausage with hot pepper Spiced Wine Flavor associated with mixed aromatic spices Flavor associated with red wine Seasoned sausage Water Hungarian salami Red wine Wild Characteristic odor of seasoned wild boar meat Seasoned sausage Wild boar sausage Garlic Flavor associated with garlic Low High Texture Tenderness Effort required to bite thorough lean and to make the sample ready to be swallowed Two-month-seasoned cubed sausage Cubed Hungarian salami Cohesiveness Mechanical textural attribute relating to the degree to which sausage can be deformed before it breaks Cubed cooked ham Dry sausage Chewiness Number of chews until reaching a state ready for swallowing Cubed cooked ham Dry cured ham Oiliness Perception of the amount of fat released by the product during mastication Cubed dry cured ham Cubed Pancetta 1 Color definitions as in Munsell Book of Color (X Rite color. Europe GmbH). foods-12-01089-t003_Table 3 Table 3 Color parameters of the ten products (mean +- SE) (1). Batch L* a* b* 30 42.69 +- 1.23 aA 28.03 +- 0.76 aA 31.48 +- 2.36 a 30-PGW 39.34 +- 1.23 bA 25.73 +- 0.76 B 29.59 +- 2.36 A 30-PW 37.87 +- 1.23 aAB 23.64 +- 0.76 B 29.10 +- 2.36 30-P 35.70 +- 1.23 aB 23.31 +- 0.76 B 29.14 +- 2.36 A 30-CF 41.71 +- 1.23 bA 30.7 +- 0.76 bA 39.7 +- 2.36 bB 50 36.47 +- 1.23 B 25.28 +- 0.76 B 34.06 +- 2.36 50-PGW 34.56 +- 1.23 aB 26.07 +- 0.76 bB 30.20 +- 2.36 50-PW 36.80 +- 1.23 B 23.53 +- 0.76 B 27.12 +- 2.36 50-P 35.19 +- 1.23 B 27.66 +- 0.76 B 32.99 +- 2.36 50-CF 37.38 +- 1.23 B 27.32 +- 0.76 B 33.25 +- 2.36 p <0.0001 <0.0001 0.025 (1) A, B = p < 0.01; a, b: p < 0.05. foods-12-01089-t004_Table 4 Table 4 Hedonic test of Cacciatore salami: effect of gender and age (mean +- SE) (1). Liking Gender p Class of Age (1) p F M I II III Overall 6.11 +- 0.14 A 6.59 +- 0.11 B 0.002 6.46 +- 0.06 6.24 +- 0.17 6.37 +- 0.26 0.484 Appearance 6.12 +- 0.13 A 6.46 +- 0.11 B 0.005 6.54 +- 0.06 6.13 +- 0.17 6.19 +- 0.26 0.045 Odor 6.08 +- 0.13 a 6.27 +- 0.11 b 0.037 6.36 +- 0.06 6.13 +- 0.16 5.94 +- 0.25 0.146 Flavor 6.01 +- 0.15 a 6.30 +- 0.13 b 0.023 6.34 +- 0.07 a 5.93 +- 0.18 b 6.18 +- 0.29 0.002 Texture 6.03 +- 0.13 A 6.56 +- 0.11 B <0.0001 6.51 +- 0.06 a 6.07 +- 0.17 b 6.31 +- 0.26 0.041 (1) A, B = p < 0.01; a, b p < 0.05. I = 18-39 yr.; II = 50-59 yr.; III = >60 yr. foods-12-01089-t005_Table 5 Table 5 Hedonic test of Cacciatore salami: effect of products (mean +- SE) (1). Batch Liking Overall Liking Appearance Odor Flavor Texture 30 6.69 +- 0.36 a 6.62 +- 0.35 A 6.48 +- 0.34 a 5.93 +- 0.39 b 6.21 +- 0.35 Bb 30-PGW 6.32 +- 0.36 6.11 +- 0.35 b 6.36 +- 0.34 a 6.07 +- 0.39 b 6.25 +- 0.35 30-PW 6.47 +- 0.36 6.45 +- 0.35 a 6.21 +- 0.34 b 6.25 +- 0.39 6.24 +- 0.35 30-P 5.65 +- 0.36 Bb 5.84 +- 0.35 B 5.45 +- 0.34 B 5.54 +- 0.39 B 5.81 +- 0.35 Bb 30-CF 7.35 +- 0.36 A 7.25 +- 0.35 Aa 7.27 +- 0.34 Aa 7.11 +- 0.39 Aa 7.18 +- 0.35 Aa 50 6.24 +- 0.36 6.74 +- 0.35 A 6.12 +- 0.34 b 6.22 +- 0.39 6.36 +- 0.35 50-PGW 6.02 +- 0.36 Bb 5.78 +- 0.35 B 5.54 +- 0.34 B 5.85 +- 0.39 B 6.10 +- 0.35 B 50-PW 6.69 +- 0.36 a 6.60 +- 0.35 A 6.33 +- 0.34 a 6.53 +- 0.39 6.60 +- 0.35 50-P 5.73 +- 0.36 B 5.57 +- 0.35 B 5.54 +- 0.34 B 5.61 +- 0.39 B 5.82 +- 0.35 B 50-CF 7.07 +- 0.36 Aa 6.98 +- 0.35 A 6.95 +- 0.34 A 7.03 +- 0.39 A 7.04 +- 0.35 Aa p 0.001 0.001 0.001 0.031 0.040 (1) A, B = p < 0.01; a, b p < 0.05. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Roila R. Branciari R. Primavilla S. Miraglia D. Vercillo F. Ranucci D. Microbial, physicochemical and sensory characteristics of salami produced from wild boar (Sus scrofa) Slovak J. Food Sci. 2021 15 475 483 10.5219/1551 36838542 2. Fulgione D. Buglione M. The Boar War: Five Hot Factors Unleashing Boar Expansion and Related Emergency Land 2022 11 887 10.3390/land11060887 3. Penrith M.L. Current status of African swine fever CABI Agric. Biosci. 2020 1 11 10.1186/s43170-020-00011-w 4. Corn J.L. Stallknecht D.E. Mechlin N.M. Luttrell M.P. Fischer J.R. Persistence of Pseudorabies Virus in Feral Swine Populations J. Wildl. Dis. 2004 40 307 310 10.7589/0090-3558-40.2.307 15362832 5. Frolich K. Thiede S. Kozikowski T. Jakob W. A review of mutual transmission of important infectious diseases between livestock and wildlife in Europe Ann. N. Y. Acad. Sci. 2002 969 4 13 10.1111/j.1749-6632.2002.tb04343.x 12381556 6. Lombardini M. Meriggi A. Fozzi A. Factors influencing wild boar damage to agricultural crops in Sardinia (Italy) Curr. Zool. 2017 63 507 514 10.1093/cz/zow099 29492010 7. Pacini M.I. Bonelli F. Briganti A. Citi S. Papini R.A. Sgorbini M. Wildlife ungulate rescue and emergency services in the Pisa area (Tuscany, Italy): Evaluation of a 9-years period (2010-2018) Front. Vet. Sci. 2020 7 626 10.3389/fvets.2020.00626 33033722 8. Paulsen P. Vali S. Bauer F. Quality traits of wild boar mould-ripened salami manufactured with different selections of meat and fat tissue, and with and without bacterial starter cultures Meat Sci. 2011 89 486 490 10.1016/j.meatsci.2011.05.014 21664055 9. Russo C. Balloni S. Altomonte I. Martini M. Nuvoloni R. Cecchi F. Pedonese F. Salari F. Sant'ana Da Silva A.M. Torracca B. Fatty acid and microbiological profile of the meat (longissimus dorsi muscle) of wild boar (Sus scropha scropha) hunted in Tuscany Ital. J. Anim. Sci. 2016 16 1 8 10.1080/1828051X.2016.1261006 10. Marescotti M.E. Caputo V. Demartini E. Gaviglio A. Discovering market segments for hunted wild game meat Meat Sci. 2019 149 163 176 10.1016/j.meatsci.2018.11.019 30557774 11. Ramanzin M. Amici A. Casoli C. Esposito L. Lupi P. Marsico G. Mattiello S. Olivieri O. Ponzetta M.P. Russo C. Meat from wild ungulates: Ensuring quality and hygiene of an increasing resource Ital. J. Anim. Sci. 2010 9 e61 10.4081/ijas.2010.e61 12. Fantechi T. Fantini C. Scozzafava G. Casini L. Consumer preferences for wild game meat: Evidence from a hybrid choice model on wild boar meat in Italy Agr. Foods Econ. 2022 10 23 10.1186/s40100-022-00231-w 13. Czarniecka-Skubina E. Stasiak D.M. Latoch A. Owczarek T. Hamulka J. Consumers' Perception and Preference for the Consumption of Wild Game Meat among Adults in Poland Foods 2022 11 830 10.3390/foods11060830 35327252 14. Presidenza del Consiglio dei Ministri Conferenza Permanente Per i Rapporti Tra lo Stato, le Regioni e le Province Autonome di Trento e Bolzano 2021 Available online: (accessed on 4 October 2022) 15. Adduci F. Elshafie H.S. Labella C. Musto M. Freschi P. Paolino R. Ragni M. Cosentino C. Abatement of the clostridial load in the teats of lactating cows with lysozyme derived from donkey milk J. Dairy Sci. 2019 102 6750 6755 10.3168/jds.2019-16311 31155256 16. Cosentino C. Faraone D. Paolino R. Freschi P. Musto M. Short communication: Sensory profile and acceptability of a cow milk cheese manufactured by adding jenny milk J. Dairy Sci. 2016 99 228 233 10.3168/jds.2015-10107 26506544 17. Cosentino C. Paolino R. Rubino M. Freschi P. Effect of the Addition of Donkey Milk on the Acceptability of Caciotta Cow Cheese Animals 2022 12 1444 10.3390/ani12111444 35681908 18. Paolino R. Freschi P. Claps S. Tarricone S. Sepe L. Cosentino C. Effect of donkey milk addition on the acceptability of Caprino, a typical goat cheese from Basilicata region, Italy Int. J. Dairy Technol. 2022 75 460 462 10.1111/1471-0307.12837 19. Petrescu D.C. Vermeir I. Petrescu-Mag R.M. Consumer Understanding of Food Quality, Healthiness, and Environmental Impact: A Cross-National Perspective Int. J. Environ. Res. Public Health 2020 17 169 10.3390/ijerph17010169 20. Hoffman L.C. Muller M. Schutte D.W. Calitz F.J. Crafford K. Consumer expectations, perceptions and purchasing of South African game meat S. Afr. J. Wildl. 2005 35 33 42 21. Paleari M.A. Moretti V.M. Beretta G. Mentasti T. Bersani C. Cured products from different animal species Meat Sci. 2003 63 4854 4889 10.1016/S0309-1740(02)00108-0 22. Schimpl A. Bauer F. Paulsen P. Quality aspects of a spreadable raw sausage product manufactured from wild boar meat Arch. Lebensmittelhyg. 2010 61 153 159 23. Zochowska-Kujawska J. Lachowicz K. Sobczak M. Gajowiecki L. Kotowicz M. Zych A. Medrala D. Effects of massaging on hardness, rheological properties, and structure of four wild boar muscles of different fibre type content and age Meat Sci. 2007 75 595 602 10.1016/j.meatsci.2006.09.018 22064023 24. Regione Basilicata Braia: Tavolo Emergenza Cinghiali, le Azioni del Dipartimento Available online: (accessed on 6 October 2022) 25. Commission Internationale de l'Eclairage Colorimetry 2nd ed. Pubblication CIE 15.2 Commission Internationale de l'Eclairage Vienna, Austria 1986 26. ISO Sensory Analysis--General Guidelines for the Selection, Training and Monitoring of Selected Assessors and Expert Sensory Assessors ISO/TC 34/SC 12 Technical Committee Geneva, Switzerland 2012 1 28 27. Braghieri A. Piazzolla N. Carlucci A. Bragaglio A. Napolitano F. Sensory properties, consumer liking and choice determinants of Lucanian dry cured sausages Meat Sci. 2016 111 122 129 10.1016/j.meatsci.2015.09.003 26398006 28. Stone H. Sidel J.L. Sensory Evaluation Practices 3rd ed. Elsevier Academic Press London, UK 2004 21 67 29. Murray J.M. Delahunty C.M. Baxter I.A. Descriptive sensory analysis: Past, present and future Food Res. Int. 2001 34 461 471 10.1016/S0963-9969(01)00070-9 30. Ruiz J. Garcia C. Muriel E. Andres A.I. Ventanas J. Influence of sensory characteristics on the acceptability of dry cured ham Meat Sci. 2002 61 347 354 10.1016/S0309-1740(01)00204-2 22061062 31. Ruiz J.N. Villanueva N.D.M. Favaro-Trindade C.S. Contreras-Castillo C.J. Physicochemical, microbiological and sensory assessments of Italian salami sausages with probiotic potential Sci. Agric. 2014 71 204 211 10.1590/S0103-90162014000300005 32. ISO/TC 34/SC 12 ISO Sensory Analysis--General Guidance for the Design of Test Technical Committee Geneva, Switzerland 2007 1 16 33. Markey O. Souroullas K. Fagan C.C. Kliem K.E. Vasilopoulou D. Jackson K.G. Humphries D.I. Grandison A.S. Givens D.I. Lovegrove J.A. Consumer acceptance of dairy products with a saturated fatty acid reduced, monounsaturated fatty acid enriched content J. Dairy Sci. 2017 100 7953 7966 10.3168/jds.2016-12057 28803023 34. Peryam D.R. Pilgrim F.J. Hedonic scale method of measuring food preferences Food Technol. 1957 11 9 14 35. R Version 3.6.1 Available online: (accessed on 12 November 2022) 36. Townsend W.E. Brown W.L. McCampbell H.C. Davis C.E. Chemical, physical, curing and sensory properties of hams from Yorkshire, crossbred and wild pigs J. Anim. Sci. 1979 49 1219 1226 10.2527/jas1979.4951219x 37. Marchiori A.F. de Felicio P.E. Quality of wild boar meat and commercial pork Sci. Agr. 2003 60 1 5 10.1590/S0103-90162003000100001 38. Hedrick H.B. Aberle E.D. Forrest J.C. Judge M. Merkel R.A. Properties of fresh meat Principles of Meat Science 3rd ed. Kendall Hunt Publishing Company Dubuque, Iowa 1994 123 131 39. Brankovic Lazic I. Jovanovic J. Simunovic S. Raseta M. Trbovic D. Baltic T. Ciric J. Evaluation of sensory and chemical parameters of fermented sausages Meat Technol. 2019 60 84 90 10.18485/meattech.2019.60.2.2 40. Braghieri A. Piazzolla N. Romaniello A. Paladino F. Ricciardi A. Napolitano F. Effect of adjuncts on sensory properties and consumer liking of Scamorza cheese J. Dairy Sci. 2015 98 1479 1491 10.3168/jds.2014-8555 25465632 41. Gimeno O. Ansorena D. Astiasaran I. Bello J. Characterization of chorizo de Pamplona: Instrumental measurements of colour and texture Food Chem. 2000 69 195 200 10.1016/S0308-8146(99)00239-3 42. Gonzalez-Fernandez C. Santos E.M. Rovira J. Jaime I. The effect of sugar concentration and starter culture on instrumental and sensory textural properties of chorizo-Spanish dry-cured sausage Meat Sci. 2006 74 467 475 10.1016/j.meatsci.2006.04.019 22063051 43. Saccani G. Fornelli G. Zanardi E. Characterization of textural properties and changes of myofibrillar and sarcoplasmic proteins in salame felino during ripening Int. J. Food Prop. 2013 16 1460 1471 10.1080/10942912.2011.595027 44. Napolitano F. Girolami A. Faraone D. Chaudry M.M.A. Braghieri A. Appearance, consumer liking and preferences of Lucanian 'Soppressata' salami Meat. Sci. 2020 167 108159 10.1016/j.meatsci.2020.108159 32388189 45. Morales R. Guerrero L. Claret A. Guardia M.D. Gou P. Beliefs and attitudes of butchers and consumers towards dry-cured ham Meat Sci. 2008 80 1005 1016 10.1016/j.meatsci.2008.04.015 22063829 46. Razmaite V. Sveistien R. Jatkauskiene V. Siukscius A. Pork Consumption Frequencies, Attitudes and Sensory Acceptance of Traditional Products in Lithuania Foods 2022 11 3292 10.3390/foods11203292 47. Niewiadomska K. Kosicka-Gebska M. Gebski J. Gutkowska K. Jezewska-Zychowicz M. Sulek M. Game Meat consumption-conscious choice or just a game? Foods 2020 9 1357 10.3390/foods9101357 32987949 48. Xie X. Huang L. Li J.J. Zhu H. Generational differences in perceptions of food health/risk and attitudes toward organic food and game meat: The case of the COVID-19 crisis in China Int. J. Environ. Res. Public Health 2020 17 3148 10.3390/ijerph17093148 32366016 49. Nikolovski G. Stojanovska M. Miovska M. The attitude towards hunting of the local population from two settlements in Suburb of Skopje, Macedonia South-East Eur. For. SEEFOR 2011 2 23 32 10.15177/seefor.11-03 50. Nijdam D. Rood T. Westhoek H. The price of protein: Review of land use and carbon footprints from life cycle assesment of animal foods products and their substitutes Food Policy 2012 37 760 770 10.1016/j.foodpol.2012.08.002 |
PMC10000648 | Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells12050678 cells-12-00678 Review An Updated Review of Hypertrophic Scarring Mony Manjula P. 1+ Harmon Kelly A. 1+ Hess Ryan 1 Dorafshar Amir H. 1 Shafikhani Sasha H. 23* Goldust Mohamad Academic Editor 1 Department of Surgery, Division of Plastic & Reconstructive Surgery, Rush University Medical Center, Chicago, IL 60612, USA 2 Department of Medicine, Division of Hematology and Oncology and Cell Therapy, Rush University Medical Center, Chicago, IL 60612, USA 3 Cancer Center, Rush University Medical Center, Chicago, IL 60612, USA * Correspondence: [email protected] + These authors contributed equally to this work. 21 2 2023 3 2023 12 5 67820 12 2022 01 2 2023 08 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Hypertrophic scarring (HTS) is an aberrant form of wound healing that is associated with excessive deposition of extracellular matrix and connective tissue at the site of injury. In this review article, we provide an overview of normal (acute) wound healing phases (hemostasis, inflammation, proliferation, and remodeling). We next discuss the dysregulated and/or impaired mechanisms in wound healing phases that are associated with HTS development. We next discuss the animal models of HTS and their limitations, and review the current and emerging treatments of HTS. normal (acute) wound healing hypertrophic scar keloids animal models treatments National Institutes of Health (NIH)R01AI150668 RO1DK107713 This work was supported by the National Institutes of Health (NIH) grants R01AI150668, and RO1DK107713 (all to S.H.S). pmc1. Introduction Wound healing is a complex physiologic process in which the body attempts to replace destroyed and damaged tissue with newly generated tissue and restore the skin's barrier functions. It is an overlapping and sequential process of hemostasis, inflammation, proliferation, and remodeling that involves communication between many different cell types . When this process does not occur in a sequential and finite manner, aberrant wound healing with hypertrophic scarring (HTS) or keloids may occur. These fibroproliferative disorders can be appreciated as elevated scars above the skin level with abundant deposition of extracellular matrix (ECM) components, especially collagen . Although HTS and keloids are often used interchangeably, they are not the same. In HTS, excess scarring is limited to the original site of injury, whereas in keloids, scarring can extend beyond the original wound and is often regarded as a form of benign skin tumor . Scarring is a major clinical problem, affecting some 100 million patients in the developed world alone . The reported prevalence of hypertrophic scarring ranges from 32 to 72% . Hypertrophic scars are particularly prevalent among adult burn patients, with those with darker skin, younger age, female sex, burns greater than 20% of total body surface area (TBSA), and burns on the neck and upper limbs experiencing the highest risk . Following burn injury, nearly 75% of patients develop neuropathic pain . Factors such as scar height, pigmentation, vascularity, and hyperplasia have been associated with increased levels of pain . In one study, nearly 60% of patients who underwent bilateral reduction mammoplasty or median sternotomy incision developed HTS postoperatively, with an increased risk in those who were young . Keloids have been reported in all ethnic groups, but they are significantly more common in individuals of African, Asian, and to a lesser degree, Hispanic descent, with the incidence ranging from 0.09% amongst the European white population, to 16% in the black population in Africa . Severe HTS may result in scar contractures which can be significantly disfiguring and disabling and may lead to loss of mobility and affect patients' ability to carry out routine daily activities . In patients with severe burns, HTS is associated with decreased quality of life and delayed reintegration into society, in part due to the effect on self-esteem and the resultant desire to hide the scarring . Globally, the wound care cost is estimated to be nearly $20.8 billion annually, with $4 billion per year associated with HTS treatment in the United States alone . Hypertrophic wound care remains one of the largest markets without definitive drug therapy. The global hypertrophic and keloid scar treatment market size is expected to reach $37.9 billion US dollars by 2026 with a compound annual growth rate (CAGR) of 9.9% . Hypertrophic scars typically occur in the second to third decade of life and present 1-2 months following injury . The scar experiences a rapid growth phase for the first 6 months, followed by regression . HTS arises as increased induration and dyspigmentation limited to the site of initial injury in areas of high tension, such as the shoulders, neck, prosternum, knees, and ankles . Diagnosis of hypertrophic scarring is made clinically. Scoring systems such as the Vancouver Scar Scale, Seattle Scar Scale, Hamilton Scar Scale, and Patient and Observer Scar Assessment Scale may be used to assess the degree of hypertrophy . These scales are based on clinical parameters such as lesion thickness, color, pliability, pain, and itching; however, the resulting scar scores are variable, as they are based on subjective clinical assessment. Combining the scar scales with more objective data such as high-resolution ultrasound scanning may be beneficial . In this review article, we provide an overview of normal (also known as, acute) wound healing phases; namely, hemostasis, inflammation, proliferation, and remodeling. We next provide an updated review of the dysregulated and/or impaired mechanisms of HTS associated with each phase of wound healing. We then discuss the animal models of HTS and their limitations, and review the current and emerging treatments of HTS. 2. Overview of Normal Wound Healing To gain better understanding of the pathophysiology underlying HTS, it is essential to appreciate the processes underlying normal (acute) wound healing in the acute setting. Normal wound healing occurs in four overlapping and complex phases; namely, hemostasis, inflammation, proliferation, and remodeling . 2.1. Phase 1: Hemostasis Hemostasis begins immediately after injury and could last for several hours. As an immediate response to limit blood loss after injury, the blood vessels' smooth muscle contracts via vasoconstrictors, such as endothelin, released by the damaged endothelial cells . This is followed by blood clot formation, which occurs in two steps: primary hemostasis and secondary hemostasis. During primary hemostasis, rearrangement and transformation of the actin cytoskeleton occur in platelets, allowing a change in their morphology from disk-shaped to fried egg-shaped cells. This, in turn, causes platelets to interact with each other and the surrounding extracellular matrix (ECM) through activated integrins, allowing for the development of the platelet plug . During secondary hemostasis, thrombin becomes activated via the intrinsic and extrinsic coagulation pathways . Activated thrombin cleaves soluble fibrinogen into fibrin and cross-links them to form fibrin mesh, which is incorporated into the fibrin clot at the site of injury to form a thrombus, which enmeshes aggregated platelets and leukocytes into a stronger structure known as the platelet plug . The platelet plug serves three important functions during wound healing: to prevent blood loss after injury, to serve as a source of chemokines and growth factors needed to initiate the inflammatory phase, and to function as a provisional scaffold for inflammatory leukocytes migration into damaged tissue . 2.2. Phase 2: Inflammation Following injury, the inflammatory phase begins within minutes, peaks in 2-3 days, and can last 1-2 weeks, depending on the extent of the injury . The primary functions of the inflammation phase during wound healing are to protect wounds against invading pathogens and to jumpstart the subsequent inflammatory and non-inflammatory responses needed for proper healing . The inflammatory phase can be divided into early and late phases. During the early phase of inflammation, endothelial cells increase the expression of adhesion molecules, resulting in the recruitment and extravasation of inflammatory cells, such as neutrophils, monocyte, lymphocytes, and mast cells . Leukocytes recruitment is mediated by pro-inflammatory cytokines, such as interleukin-1 (IL-1), IL-6, and tumor necrosis factor alpha (TNF-a), which are released from degranulating platelets, keratinocytes, endothelial cells, and tissue-resident macrophages . Upon arrival into the wound, monocytes differentiate into pro-inflammatory M1 macrophages, which function to further amplify inflammatory responses by producing more pro-inflammatory cytokines, and assist neutrophils in destroying invading pathogens . During the late phase of inflammation, the macrophages polarize into the anti-inflammatory M2 phenotype, which play pivotal roles in the resolution of the inflammatory responses and in the initiation of the proliferation phase through the production of a spectrum of angiogenic and growth factor mediators, such as vascular endothelial growth factor (VEGF), PDGF, and FGF2 . 2.3. Phase 3: Proliferation The proliferative phase, also known as the new tissue regeneration phase, begins approximately 3 days after injury and lasts for about 2-3 weeks. The main events during the proliferative phase are provisional matrix replacement with granulation tissue, angiogenesis, and re-epithelization . Initially, fibroblasts migrate to the site of injury in response to mediators released by platelets and macrophages, such as PDGF, transforming growth factor beta (TGF-b), and connective tissue growth factor (CTGF) . To replace the provisional matrix with granulation tissue, fibroblasts release extracellular matrix (ECM) components (primarily type III collagen, fibronectin, proteoglycans, and hyaluronic acid) . Granulation tissue is composed of ECM components, fibroblasts, proliferating blood vessels, macrophages, and lymphocytes, and it is an important indicator of wound healing progression . During re-epithelization, M2 macrophages and keratinocytes produce and release EGF and TGF-b, which in turn induce proliferation and cell migration in epithelial cells bordering the wound edges to re-establish the epidermis integrity at the wound site . Stem cells from hair follicles and sebaceous glands differentiate into keratinocytes to aid in this process . Angiogenesis involves the creation of new vasculature that is 3 to 10 times denser than what is found in normal tissue . It is critical in facilitating the transport of immune cells, oxygen, and nutrients for the cells participating in healing . Angiogenesis is triggered by local hypoxia and several soluble factors, including VEGF (most prominent factor), PDGF, fibroblast growth factor-basic (bFGF), the serine protease thrombin, and members of the TGF-b family . Following the completion of wound healing, most of the newly formed capillaries will regress . 2.4. Phase 4: Remodeling The remodeling phase begins 2-3 weeks following injury and can last up to a year or even longer . Matrix maturation and tissue remodeling depend on the balance between the degradation of extracellular matrix (ECM) components in granulation tissue and their replacement by connective tissue components, namely collagen I. Early in the remodeling phase, ECM components (e.g., collagen III, fibronectin, and hyaluronic acid) are degraded by matrix metalloproteinases (MMPs) . Because of the destructive nature of the MMPs, they are tightly regulated by the tissue inhibitors of metalloproteinases (TIMPs) . Moreover, fibroblasts differentiate into myofibroblasts which produce thick bundles of collagen I to replace most of the collagen III . Over time, collagen I fiber bundles increase in diameter, resulting in increased wound tensile strength; however, the healed tissue never fully regains the properties of uninjured skin, resulting in a mostly acellular and avascular scar . Scar tissue contains collagen bundles that are smaller and more disorganized and, thus, prone to dehiscence . Over time, wound contraction occurs as the result of myofibroblasts bringing the wound edges together with the contractile function of their actin filaments . 3. Hypertrophic Scarring Associated with Wound Healing Phases A large body of evidence suggests that excessive inflammation generates pro-fibrotic molecules, which in turn activate fibroblasts, resulting in HTS . In addition, excessive angiogenesis and prolonged re-epithelialization can extend the release of pro-fibrotic growth factors . In the last few years, many biomolecules have been implicated in HTS; however, their exact mechanisms have yet to be fully elucidated, in part due to the complexity and overlapping nature of wound-healing processes. Here, we will examine each phase of wound healing with respect to HTS formation. 3.1. Phase 1: Hemostasis The fibrin provisional matrix deposited during hemostasis has been implicated in the activation of myofibroblasts and the formation of HTS . In fact, high-density fibrin clot deposition during the early phase of healing may predict the formation of HTS , as calculated by a multiscale mathematical model ; however, more studies related to fibrin content and rate of fibrinolysis in experimental models are required to validate the role of the fibrin provisional matrix in the formation of HTS. In addition, during hemostasis, platelets release a multitude of pro-fibrotic growth factors such as PDGF, VEGF, TGF-b1, and CTGF, which have been linked to the formation of HTS . Interestingly, platelet-rich plasma (PRP) obtained from platelets of the peripheral blood is considered to be a therapeutic option for HTS, as it reduces the expression of pro-fibrotic molecules such as TGF-b1 and CTGF . These reports suggest that while naive platelets may be anti-fibrotic in nature when activated excessively, they can contribute to HTS development. Clearly, more studies are needed to evaluate the role of naive versus activated platelets in HTS. 3.2. Phase 2: Inflammation Excessive inflammation is the best elucidated pathophysiological reason for HTS formation . As such, many of the accepted therapeutics target inflammation . Excessive infection and tissue necrosis in severe burn wounds cause increased pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs), toll-like receptor (TLR) signaling, and infiltration of inflammatory cells to the wound site . Surprisingly, studies of HTS have found chemokine expression to be variable. In a study using the rabbit ear as a model for HTS, the expression of chemokines such as Chemokine (C-C motif) ligand 3 (CCL3), CCL7, and CCL13 maintained increased expression for 21 to 35 days, while CCL2, CCL4, CCL5, and chemokine (C-X3-C motif) ligand 1 (CX3CL1) were maintained at high levels for 21 to 56 days . Another study reported that the expression of CCL3, CXCL1, CXCL2, CXCR2, C3, and Interleukin 10 (IL-10) was reduced in human HTS, 52 weeks following surgery . In another study, SDF1/CXCR4 signaling was found to be increased in human HTS tissue . The underlying reasons for this variability remain unknown and require future investigation. Inflammatory cells release various factors such as interleukins, interferon, and growth factors . Increased expression of pro-inflammatory and pro-fibrotic growth factors activate fibroblasts and are thus implicated in HTS formation . Interestingly, in a study of HTS tissue at 3 h following surgery, the expression of certain pro-inflammatory factors such as IL-6, IL-8, and CCL2 was found to be reduced during the early phase of healing . Intriguingly, inadequate pro-inflammatory responses have also been reported in hypofibrotic diabetic wounds early after injury, rendering them vulnerable to infection and impaired healing . This delay in inflammatory responses during the acute phase of healing early after injury and its role in the formation of HTS should be further investigated. IL-6 is a major cytokine that influences the middle and late phases of healing, as it is involved in shifting inflammation from acute to chronic by enhancing monocyte recruitment, M2 macrophage polarization, and ECM deposition . IL-6 is highly expressed in HTS and is considered to be a therapeutic target for the treatment of HTS . The IL-6/STAT3 (Signal transducer and activator of transcription 3) pathway activates many of the genes required for ECM production and fibroblast proliferation, leading to HTS . Other than the IL-6 and inflammatory chemokines, other inflammatory cytokines that are highly expressed in HTS include IL-1b, IL-4, IL-8, IL-17, IL-13, and IL-22 . Some of these cytokines, such as IL-4 and IL-13, have been under investigation as therapeutic targets for HTS . The expression of IL-10, an anti-inflammatory cytokine and promising therapeutic molecule, has been found to be low in patients with hypertrophic scarring compared to those with non-hypertrophic scarring . Some studies have suggested that IL-10 directly influences fibroblasts by activating the STAT3 or AKT signaling pathways . It has also been reported that IL-10 reduces scar formation by regulating the TLR4/NF-kB pathway in dermal fibroblasts . However, further investigation is required to elucidate the role of IL-10 in preventing HTS development. Additionally, the expression of other cytokines, such as IL-24, IL-36, IL-37, IL-1RA, and TNF-a, has been found to be low in HTS . TNF-stimulated gene 6 (TSG-6) has been found to suppress scarring by downregulating the IRE1a/TRAF2/NF-kB signaling pathway . Moreover, alteration in the fatty acid metabolism influences inflammation and can result in excessive scarring . In a recent study, the expression of sterol regulatory element-binding protein-1 (SREBP1) and fatty acid synthase (FASN) was shown to be reduced at mRNA and protein levels in pathological HTS and in HTS-derived fibroblasts . In another study, the expression of fatty acid desaturase 1 and 2 (FAD1 and FAD2)--key enzymes in the polyunsaturated fatty acids (PUFAs) metabolism with demonstrated anti-inflammatory function --were lower in keloids and keloid-derived fibroblasts . However, the mechanism of altered lipid profile in HTS has not been explored. It is possible that alterations in lipid metabolism might influence HTS through changes in the inflammatory pathways, given that fatty acids play an important role in regulating inflammation . 3.3. Phase 3: Proliferation Events of the proliferative phase, such as angiogenesis and ECM deposition, are highly active in HTS, whereas re-epithelialization is prolonged in HTS as keratinocytes remain continually activated . Consequently, the granulation tissue becomes denser during HTS formation than in normal scarring . In HTS, cells such as keratinocytes, endothelial cells, and fibroblasts release many pro-fibrotic growth factors such as TGF-b, PDGF, VEGF, and CTGF . This pro-fibrotic environment, in turn, induces fibroblasts to produce more ECM proteins such as collagen, fibronectin, laminin, periostin, fibrillin, and tenascin; however, the expression of certain ECM proteins such as hyaluronic acid, dermatopontin, and decorin are found to be altered or reduced . Fibroblasts of the deep dermis are responsible for the production of additional factors such as osteopontin, angiotensin-II, and peroxisome proliferator-activated receptor (PPAR)-a and contribute to scarring more than fibroblasts of the superficial dermis . Recently, it has been revealed that fibroblasts in the upper dermis also contribute to scarring by producing IL-11, which in turn activates myofibroblasts . TGF-b plays an important role in the formation of HTS, and the TGF-b/SMAD (Suppressor of Mothers against Decapentaplegic) signaling pathway is considered to be a potential therapeutic target of HTS . Molecules such as SMAD interacting protein and bacterial PAMPs such as lipopolysaccharides (LPS) may induce HTS by enhancing the TGF-b1/SMAD signaling pathway . Endothelial cells isolated from porcine burn wounds show that endothelial dysfunction and altered expression of angiogenic genes such as endothelin-1, angiopoietin-1, angiopoietin-2, and angiogenin may result in HTS . In turn, angiogenesis is stimulated by the microvesicles released from the myofibroblast . Factors released from these vesicles may result in HTS, as many of them are pro-fibrotic in nature . During hypertrophic scarring, keratinocytes remain in their activated state for a prolonged duration of time . Dysregulation in the Notch signaling of keratinocytes may also contribute to HTS formation . Notch 1 signaling and intracellular domains such as Jagged1 and Hes1 are highly expressed in the epidermis of hypertrophic scar patients . This leads to the enhanced expression of pro-fibrotic factors, such as TGF b1, TGF b2, CTGF, IGF-1, VEGF, and EGF . In addition, epithelial-mesenchymal transition increases ECM deposition and has been shown to contribute to HTS formation . Moreover, keratinocytes produce HMGB1, which activates fibroblasts, resulting in HTS formation . However, certain factors released from keratinocyte-like pigment epithelium-derived growth factor (PEDF) are associated with reduced angiogenesis and HTS formation . Interestingly, among different growth factors, FGF-2 has an anti-scarring effect since it up-regulates the expression of MMP-1 and hepatocyte growth factor (HGF), although further investigations are required to clarify its therapeutic potential . 3.4. Phase 4: Remodeling In HTS, the balance of ECM synthesis and remodeling is dysregulated . Both fibroblasts and myofibroblasts continue to deposit collagen III and collagen I in HTS . The persistence of myofibroblasts due to defects in apoptosis results in the deposition of excessive fibrous collagen I and scarring . The presence of nodules containing myofibroblasts is a peculiar feature of HTS . Mechanical stretch and TGF-b can stimulate the differentiation of fibroblasts to myofibroblasts, contributing to HTS formation . Metalloproteinases (MMPs), such as MMP1 and MMP7, are downregulated during HTS formation, resulting in reduced degradation of ECM components such as collagen I, collagen III, and fibronectin . Administration of MMP1 has been shown to improve scarring . The tissue inhibitors of metalloproteinases (TIMPs), such as TIMP1 and TIMP2, reduce the action of MMPs during HTS development . In contrast, expression of MMP2, MMP9, and MMP13 are shown to be increased in HTS . This upregulation may be a compensatory response to elevated levels of ECM in HTS, but it remains unclear and requires future investigation. In HTS, reduced expression of matrix remodeling proteins results in the disorganization of ECM components . Treatment with decorin, a matricellular protein involved in collagen fiber organization, has been shown to reduce HTS formation . In addition, targeting the lysil hydroxylase enzyme, involved in the formation of pyridinoline cross-links, reduces the activity of fibroblast proliferation by regulating TGF-b1 . 4. Animal Models of Hypertrophic Scarring While two-dimensional and three-dimensional cell culture-based in vitro models can be useful for investigating the mechanism of fibroblast in producing excessive ECM and potential therapeutic molecules, the absence of immune and vascular components in these models limits the physiological relevance of the findings emerging from these studies with respect to the mechanisms underlying HTS formation in tissue . In the past several decades, many attempts have been made to develop animal models of HTS in different species. Despite these attempts, there are no animal models that can fully recapitulate HTS in humans. Descriptions of each animal model for HTS, as well as their advantages and disadvantages, are summarized in Table 1. The rabbit ear model has been widely used to study HTS formation despite the involvement of chondrocytes during the healing, where skin and perichondrial layers are removed from the ventral side of the rabbit ear to generate an HTS-like condition . The advantages of this model include the simplicity of the procedure, ease of handling of the animal, and ability to create multiple wounds; however, the ventral side of the ear is difficult to handle because of its low thickness, and precaution needs to be taken to avoid damage of the underlying cartilage during the procedure . To reduce damage to the rabbit ear cartilage during the procedure, cryosurgery has been attempted to remove the perichondrial layer . In another rabbit ear model for HTS, thermal burn injury has been attempted to create a more elevated scar within a shorter duration which better mimics an HTS condition in humans . However, thermal injury has to be precisely controlled to avoid variability in scarring . Injecting anhydrous alcohol into the subcutaneous and superficial fascia regions of the dorsal skin of a rabbit has been used to model HTS ; however, this model appears to be more appropriate for skin fibrosis than the HTS due to the absence of a healing response. Deep burn injury to the dorsal side of porcine skin creates raised scar tissue and has been used in some studies as a model for HTS . Although there are structural similarities between human and pig skin, the high costs associated with the production of this animal model and the difficulty in handling it have lessened its popularity for HTS studies. Several groups have also attempted to develop rodent (mouse and rat) models for HTS . These murine models are inexpensive to produce and easy to handle, but wound healing patterns in rodents differ from that of humans due to the rapid contraction of the panniculus carnosus muscles . To mitigate the effect of rapid wound contraction in rodents, splinting excision wounds have been attempted . For example, splinted full-thickness skin wounds in rodents recapitulate mechanical tension in the wound bed, and the lack of neo-epithelium in this model amplifies myofibroblast function, culminating in hypertrophic features, which are similar to HTS in humans . Similarly, mechanical pressure applied to a wound by a biomechanical loading device also produces HTS-like features in mice . C-X-C motif chemokine receptor 3 (CXCR3)-deficient mice develop thick keratinized scars and have been used in some studies to model HTS, but deficient dermal maturation with poor collagen content has been observed . Hence, the role of CXCR-3 and its effect on matrix development require further investigation. By resecting the abdominal wall muscle on the ventral side of mice that produces contractile forces, another murine wound model for scarring has been created, but it is not comparable with the healing mechanism underlying HTS . Some attempts have been made to develop a xenograft model of HTS by grafting tissue from human HTS onto nude mice . These mice displayed scar thickness and collagen bundle orientation and morphology resembling human HTS . However, a lack of an immune response and difficulty in maintaining nude mice may obstruct the study of therapeutic molecules in this model. Developing an ideal animal model for HTS is exceptionally challenging, as the scar endotype is difficult to control in experimental settings . The aforementioned animal models all fall short; therefore, developing an ideal animal model is essential to support studies related to the formation of and therapy for HTS. 5. Conventional and Emerging Treatments for Hypertrophic Scarring 5.1. Conventional Therapies Treatments of hypertrophic scars often focus on correction of factors that are associated with pathological scar development as described above. These include wound stabilization, minimizing mechanical irritation, balancing wound healing phases, attenuating pro-fibrotic mechanisms, inducing anti-fibrotic mechanisms, and promoting the remodeling of collagenous scar components. Published guidelines on the treatment of hypertrophic scars and keloids include many different modalities without one single, widely accepted protocol . Several treatments and techniques have been shown to prevent the development of hypertrophic scar development. (These conventional treatments have been summarized in Table 2). Reduction in tension on the dermal layer when closing wounds is effective and can be achieved with fascial and subcutaneous tensile reduction sutures in wounds of adequate depth . Additionally, dermal closure using sutures arranged in a zig-zag pattern or using z-plasties should be performed whenever possible . Closure with 3-0 VLoc 90 barbed suture (VLoc, Covidien, North Haven, CT, USA) compared to interrupted suture with 4-0 nylon produced significant improvements in the Vancouver scar scale (VSS) and patient and observer scar assessment scale (POSAS) scores in patients undergoing anterolateral thigh flap procedures with identical methods of deep closure between groups . Following the closure of initial wounds, several therapies can also be applied early in the healing process. Similarly to the aforementioned suturing techniques, wound stabilization using paper tape or silicone sheets can also prevent the dermal inflammation that contributes to hypertrophic scar and keloid formation . Wound compression using pressure garment therapy at 15-40 mmHg has been shown to improve outcomes . Regarding ideal pressure, one review of pressure garment therapy for the treatment of burn wounds found that the application of pressure at 17-24 mmHg resulted in improved scar height, softness, and cosmetic appearance compared to a pressure below 5 mmHg . Cohesive silicone sheets that added pressure to the wound also outperformed silicone gel sheets in improving scar assessment scale scores . Intermittent application of pressure through regular massage therapy has not been shown to improve outcomes, suggesting that constant pressure must be applied . Topical agents applied to heal wounds have also been shown to reduce hypertrophic scar formation, including flavonoids and silicone cream . The local injection of Botulinum toxin-A postoperatively has also been shown to significantly improve scar assessment scale scores compared to controls . In a recent study of optimal dosing of Botulinum toxin-A, postoperative injections of 8 units showed significantly improved Stony Brook Scar Evaluation Scale (SBSES) scores compared to the injections of 4 units . The culture of human fibroblasts with Botulinum toxin-A resulted in decreased proliferation, migration, and secretion of pro-fibrotic factors, while JNK phosphorylation levels were increased, providing evidence for possible mechanisms of this benefit . Scar revision is the simplest method of treating pre-existing HTS and encompasses procedures aimed at excisional debulking of hypertrophic scar tissue ( and Table 2). Closure during these procedures is specifically directed at providing favorable cosmetic results and should employ the methods described above for prophylaxis against scar recurrence. To be effective, scar revisions should be performed over 1 year from the original injury to give adequate time for the scar to mature , as immature scars are prone to hypertrophic healing and give poor results after scar revision . However, excision may not be necessary, as more conservative measures have proven to be effective. For example, in one study, mechanical disruption of existing hypertrophic scars using microneedle roller therapy improved scar pigmentation to resemble surrounding tissue more closely, and significantly improved both the mean patient satisfaction scale (PSS) and observer satisfaction scale (OSS) between preoperative and postoperative sampling . Another study found that microneedle therapy improved modified Vancouver scar scale (mVSS) scores significantly more than carbon dioxide (CO2) laser therapy for hypertrophic scars . This benefit may be explained by microneedle therapy disrupting existing collagen and stimulating the release of MMP-9 . Pharmacologic agents have also been used frequently in the treatment of hypertrophic scars, with common agents including corticosteroids, chemotherapeutic agents, and Botulinum toxin-A. Corticosteroids provide benefits through their potent anti-inflammatory effects and are believed to induce local vasoconstriction when applied to hypertrophic scars and keloids. Tapes and plasters containing corticosteroids effectively treat hypertrophic scars and keloids when applied to these lesions and should be positioned to avoid contact with surrounding tissue . The most common use of corticosteroids in the treatment of hypertrophic scars and keloids by far is the intralesional injection of triamcinolone (TAC). A recent literature review and meta-analysis of this therapy found that compared to 5-FU and verapamil, TAC alone improved scar vascularity . However, TAC therapy also had higher rates of skin atrophy and telangiectasias, especially at the commonly used dose of 40 mg/mL . Significant differences in favor of other agents were found for scar height (5-FU, TAC + 5-FU), scar pliability (TAC + 5-FU, Botulinum toxin-A), scar pigmentation (TAC + 5-FU), VSS score (TAC + 5-FU, TAC + platelet rich plasma), and POSAS score (bleomycin) when compared against TAC alone . A study of TAC vs. TAC + 5-FU found significant differences favoring TAC + 5-FU in mean reduction in scar height, overall POSAS score, and the overall rate of efficacy. Rates of telangiectasias (commonly known as "spider veins"), skin atrophy, hypopigmentation, and recurrence were significantly higher in the group receiving TAC, while the rates of ulceration were significantly higher in the group receiving TAC + 5-FU . A literature review and meta-analysis of intralesional Botulinum toxin-A injection found significantly improved visual analog scale (VAS) scores compared to intralesional corticosteroid and placebo injection . In a split-scar study of patients with existing hypertrophic scars, injection of Botulinum toxin-A was found to significantly improve mean VSS score post-treatment as compared to the placebo control . The energy-based therapy is well established as a treatment modality for hypertrophic scars and keloids, with its use dating back to the 1980s . Lasers are the mainstay of energy-based treatments, with a multitude of different laser devices utilizing different wavelengths for specific targets . Laser therapy is often used in the treatment of formed hypertrophic scars but can also be used preventatively in the early postoperative period. In a split-scar study of patients undergoing total knee arthroplasties, scar treatment with a 595 nm pulsed-dye laser was associated with significantly improved overall VSS scores compared to an untreated scar . The guidelines for the use of energy-based treatment for acne scars have included specific recommendations for use with hypertrophic acne scars and keloids. In patients with active acne, a 1064 nm ND:YAG laser is preferred, and pulsed-dye vascular lasers are the laser treatment of choice for hypertrophic acne scars. Pulsed-dye lasers (PDL) may also be used to assist with the delivery of 5-FU and/or TAC. Non-laser devices, including Tixel (Novoxel, Ltd., Berlin, Germany) and EnerJet (PerfAction Technologies Ltd., Rehovot, Israel), were also recommended for the treatment of hypertrophic acne scars . Similar guidelines for traumatic scars recommend non-ablative fractional laser (NAFL) for hypertrophic scars, except in the presence of significant thickness and textural irregularity, where ablative fractional laser (AFL) therapy is preferred . In a study comparing no laser treatment, CO2 laser treatment alone, and intense pulsed light (IPL) + CO2 laser, both treatment groups had statistically significant improvements in POSAS score and Manchester scar scale (MSS) score compared to the placebo, without significant difference between the treatment groups. The only significant difference between treatment groups was in favor of the combination therapy for scar color and texture, indicating that CO2 alone is sufficient and IPL can be used for an additional benefit for these specific factors . Regarding protocols for CO2 laser, a study of varying densities for fractional CO2 laser treatment found that high (25.6%) density significantly improved VAS and POSAS scores compared to low (7.4%) and medium (12.6%) densities in treating mature hypertrophic burn scars . A split-scar study of low-energy CO2 fractional laser treatment showed significantly improved POSAS scores for all elements except for patient-scored irregularity compared to the control for pediatric patients with early-stage hypertrophic burn scars . A study of CO2, PDL, and CO2 + PDL for the treatment of hypertrophic burn scars found significant improvements in posttreatment POSAS for all treatment groups. Focused analyses found that scar height was improved by PDL or CO2 + PDL for scars <0.3 cm, and a significant reduction in scar height was achieved by CO2 + PDL only for scars older than 9 months. Although the guidelines for hypertrophic acne scars include the use of laser-assisted delivery of corticosteroids, a study of fractional ER:YAG laser alone or in combination with topical clobetasol found no significant benefit from the addition of steroids, with both treatment groups achieving significant posttreatment improvements in scar thickness and POSAS scores . Recently, studies have compared IPL to non-laser therapies. Significant differences in scar pliability, hyperpigmentation, and median VAS favored IPL vs. silicone sheet, but significant differences in VAS and histopathological characteristics favored cryotherapy vs. IPL . 5.2. Emerging Treatments Given the prevalence of hypertrophic scarring, new treatments are continually developed. Intralesional TAC, for example, was found to improve scar height, pliability, and pigmentation when combined with 5-FU and reduced the number of treatment sessions and remission time when combined with 1550 nm erbium glass fractional laser treatment (Table 3) . While Botox A with TAC showed no difference in scar appearance, it significantly reduced pain and pruritis . Scars treated with RFA plus verapamil and 5-FU experienced the fastest scar volume reduction with relief of symptoms and hyperemia compared to either agent alone . Additionally, the combination of intense pulse light (IPL) and CO2 laser significantly improved scar color and texture . The combination of lasers with 5-FU and/or TAC delivered intralesionally or via laser assistance has thus been recommended for the treatment of hypertrophic acne scars . The role of angiotensin II in scar activity has recently been examined . Human dermal fibroblasts treated with losartan, an angiotensin II type 1 receptor antagonist, displayed decreased contractile activity, fibroblast migration, gene expression of TGF-b1, type 1 collagen, and MCP-1, while reducing monocyte migration and adhesion . In rat models, the consumption of losartan showed decreased cross-sectional area and elevation index in scars, with decreased a-SMA+ and CD68+ during immunostaining . Another in vivo model demonstrated a reduced incidence of hypertrophic scarring with decreased inflammation, collagen and fibroblast cellularity, vascularization, and myofibroblast activity with the topical administration of oxandrolone and hyaluronic acid gel . Clinically, the administration of dipeptidyl peptidase-4 inhibitors was shown to reduce the risk of hypertrophic scarring and keloid onset by less than half in patients who underwent sternotomy, while 1,4-diaminobutane (1,4 DAB) in breast reduction patients resulted in significantly greater scar satisfaction and less scar hardness measured by Rex Durometer . Autologous fat grafting also presents as a novel therapy to improve the function and appearance of scars. While the underlying mechanism is unknown, exposure to adipocytes decreased the expression of the myofibroblast marker a-SMA and ECM components . The reprogramming of myofibroblasts was found to be triggered by BMP-4 (bone morphogenetic protein 4) and activation of PPARg (peroxisome proliferator-activated receptor gamma) signaling, which initiated tissue remodeling . As is the case in many other fields of medicine, stem cells are also a promising therapeutic target for HTS. Mesenchymal stem cells (MSC) isolated from the mouse whisker hair follicle outer root sheath were applied to an in vivo full-thickness wound model . A quantitative evaluation revealed reduced inflammation, cellularity, and collagen filaments, as well as thinner dermal and epidermal layers in the MSC-treated wounds, indicating a reduction in hypertrophic scars. Another study examined the effect of combined treatment with a non-ablative laser and human stem cell-conditioned medium on burn-induced hypertrophic scar formation . The treatment group was found to have reduced erythema, trans-epidermal water loss, and scar thickness. Platelet-rich plasma (PRP) has also been identified as a promising therapy for scarring. In one study, primary dermal fibroblasts isolated from hypertrophic scars were cultured in a medium supplemented with 5% PRP or platelet-poor plasma (PPP) . The PRP group was found to have reduced expression of TGF-b1 and connective tissue growth factor (CTGF) mRNA. Other studies have examined combination treatments with both PRP and ablative fractional CO2 lasers and have found the combination to be more beneficial than either treatment alone . In addition, identifying the molecular targets for potential treatments is an ongoing source of investigation. Co-cultures of anti-inflammatory cluster of differentiation 206 (CD206)+ macrophages and fibroblasts showed decreased expression of fibrotic factors, such as type 1 and 2 collagen, alpha-smooth muscle actin, connective tissue growth factor, and TGF-b, with upregulation of MMP-1. IL-6 was also found to be increased in the medium, with an increase in anti-fibrotic gene expression when IL-6 was added to fibroblasts. Cytotherapy with cultured CD206+ macrophages or a direct administration of recombinant human IL-6 has been shown to dampen the expression of pro-fibrotic mediators (e.g., COL1A1 *, COL2A1 *, a-SMA *, CTGF *, and TGF-b1) in fibroblast in cell culture studies . In vitro studies of fibroblasts have revealed that IFN-g inhibits collagen synthesis . IFN-g knockout mice were found to have reduced wound closure, lower wound breaking strength, and dampened expression of collagen type 1A (COL1A1) and collagen type 3 A1 (COL3A1) mRNA, but a greater expression of MMP-2 (gelatinase) mRNA . The study concluded IFN-g may be involved in both the proliferation and maturation stages of wound healing and, therefore, may be a target for potential treatments. 6. Conclusions As this review illustrates, there has been significant knowledge gained in the field of hypertrophic scarring. A pro-fibrotic environment results in excessive collagen deposition and, therefore, hypertrophic scar formation. In this review article, and for the first time, we highlighted the defective and impaired mechanisms underlying HTS that are associated with each phase of wound healing (hemostasis, inflammation, proliferation, and remodeling). This was an attempt to demonstrate the multifaceted nature of the phase-specific dysregulations and impaired mechanisms that underlie HTS development. We further discussed the current animal models and their limitations in order to highlight the need for better animal models that can more closely reproduce the human condition with respect to HTS development. We also reviewed the current and emerging therapies, which further demonstrate the inadequacy of therapies to address HTS. There is still much to be discovered in regard to the underlying mechanisms contributing to HTS development. A better understanding of the impaired mechanisms underlying HTS would surely lead to the development of more effective targeted therapies to treat this debilitating and costly pathological condition. Acknowledgments We would like to acknowledge Okensama La-Anyane and Ayaan Ahmed for their insightful comments regarding this manuscript. Author Contributions Conceptualization, M.P.M., K.A.H., A.H.D. and S.H.S.; methodology, M.P.M., K.A.H., A.H.D. and S.H.S.; investigation, M.P.M. and K.A.H.; resources, M.P.M. and K.A.H.; writing--original draft preparation, M.P.M., K.A.H. and R.H.; writing--review and editing, M.P.M., K.A.H., A.H.D. and S.H.S.; funding acquisition, S.H.S. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 The phases of acute wound healing, including hemostasis (I), inflammation (II), proliferation (III), and remodeling (IV). Hemostasis begins soon after wounding with vasoconstriction and blood clot formation. This is followed by the infiltration of inflammatory cells. Then, re-epithelialization occurs with collagen deposition and angiogenesis during the proliferation phase. Finally, the remodeling phase occurs with collagen remodeling and maturation, wound contraction, and scar tissue formation. Figure 2 Phases of normal wound healing versus aberrant wound healing with the formation of hypertrophic scars. Events such as higher fibrin clot deposition, infiltration of higher number of inflammatory cells, prolonged re-epithelialization, and excess angiogenesis can result in excessive and improper collagen deposition and, therefore, formation of hypertrophic scars. Figure 3 The events and molecules associated with hypertrophic scar (HTS) formation. The pro-fibrotic molecules generated from the high-density blood clot induce excessive inflammation, angiogenesis, and prolonged re-epithelization. The resultant excessive production of extracellular matrix and fibroblast differentiation and improper matrix remodeling then causes formation of hypertrophic scarring. Molecules with increased expression are denoted with (+), whereas those with decreased expression are denoted with (-). The players included in this figure are discussed in the text. cells-12-00678-t001_Table 1 Table 1 Animal models for hypertrophic scarring (HTS). Model Scar Location Advantages Disadvantages Rabbit ear HTS * model Ventral side of rabbit ear Simple, reliable model Ease of handling Possibility of creating multiple wounds Skin of the ventral side is too difficult to handle because of low thickness Involvement of cells other than skin cells during healing, such as chondrocytes Risk of damaging the underlying cartilage Modified rabbit ear HTS model--use of cryosurgery Ventral side of rabbit ear Low risk of damaging the cartilage Skin of the ventral side is too difficult to handle because of low thickness Involvement of cells other than skin cells during healing, such as chondrocyte Modified Rabbit ear HTS model--application of thermal injury Ventral side of rabbit ear Elevated scar within short duration compared to the typical rabbit ear HTS model Skin of the ventral side is too difficult to handle because of low thickness Involvement of cells other than skin cells during healing, such as chondrocytes Uncontrolled thermal injury can cause variability in scarring effect HTS model on rabbit by injecting anhydrous alcohol Dorsal skin HTS-like appearance comparable to the rabbit ear HTS model Low cost Ease of handling Absence of healing response Burn hypertrophic model on porcine skin Dorsal Skin Elevated scar comparable to human scar High cost Difficult to handle HTS model by splinting of rat wound Dorsal skin HTS-like features by reducing the formation of neo-epithelium Low cost Ease of handling Splinting may create a higher and more persistent tensional state Scar on CXCR3 * deficient mouse Dorsal skin Simple, reliable model Ease of handling The model requires further validation HTS model produced by grafting human xenografts on nude mice Dorsal skin Establishment of human scar on an animal model Difficulty in maintaining nude mice Absence of immune response in mice HTS model by resecting abdominal wall muscle on mice Ventral skin, abdominal region Simple and reliable method Ease of handling Not comparable with general scar development after burn injury or trauma * Abbreviations: HTS (hypertrophic scar); CXCR3 (C-X-C motif chemokine receptor 3). cells-12-00678-t002_Table 2 Table 2 Conventional treatments for hypertrophic scarring. Treatment Mechanism Tensile reduction suture closure Reduces tension on the dermal layer when closing wound Paper tape, silicone sheets Hydration, increased temperature, prevent dermal inflammation Wound compression Reduces capillary perfusion, accelerated collagen maturation Laser Destroys microvascularization, resulting in hypoperfusion and hypoxia Silicone cream Hydration of the stratum corneum and cytokine-mediated signaling from keratinocytes to dermal fibroblasts Flavonoids Anti-inflammatory, antioxidant, anti-bacterial Botulinum toxin-A Decreases proliferation, migration, and secretion of pro-fibrotic factors from fibroblasts Scar excision Removal of affected tissue Microneedle Disruption of existing collagen, stimulation of MMP-9 * release Corticosteroids Anti-inflammatory, local vasoconstriction Botulinum toxin-A Decreases proliferation, migration, and secretion of pro-fibrotic factors from fibroblasts * Abbreviations: MMP-9 (metalloproteinase 9). cells-12-00678-t003_Table 3 Table 3 Emerging therapeutics for hypertrophic scarring. Treatment Proposed Mechanism Corticosteroids + 5-Fluorouracil Anti-inflammatory, local vasoconstriction, inhibit fibroblasts proliferation, decrease collagen synthesis Laser + Verapamil + 5-Fluorouracil or corticosteroids Destroy microvascularization resulting in hypoperfusion and hypoxia, inhibit fibroblast proliferation, decrease collagen synthesis, anti-inflammatory CO2 * Laser + Intense Pulse Light Destroy microvascularization resulting in hypoperfusion and hypoxia, promote new dermal collagen formation and rapid differentiation of keratinocytes Losartan Fibroblasts with decreased contractile activity, migration, and adhesion Oxandolone + hyaluronic acid gel Decrease inflammation, collagen and fibroblast cellularity, vascularization, and myofibroblast activity Dipeptidyl peptidase-4 inhibitors Attenuate collagen synthesis and deposition 1,4-Diaminobutane Inhibits collagen cross-linking Autologous fat grafting Decreases the expression of the myofibroblast marker a-SMA * and ECM * components Stem cells Reduce inflammation, cellularity, and collagen filaments Platelet-rich plasma Reduces expression of TGF-b1 * and CTGF mRNA CD206 * + Macrophages and Fibroblasts Increase MMP-1 * and decrease expression of pro-fibrotic factors, COL1A1 *, COL2A1 *, a-SMA *, CTGF *, and TGF-b1 * IL-6 * Increases expression of anti-fibrotic genes IFN-g * Increases expression of COL1A1 * and COL3A1 * mRNA and decreases expression of MMP-2 * (gelatinase) * Abbreviations: CO2 (carbon dioxide); CD206 (cluster of differentiation 206); IL-6 (interleukin 6); IFN-g (Interferon gamma); a-SMA (alpha smooth muscle actin); ECM (extracellular matrix); TGF-b1 (transforming growth factor beta-1); COL1A1 (collagen type I alpha 1 chain); COL2A1 (collagen type II alpha 1 chain); COL3A1 (collagen type III alpha 1 chain), CTGF (connective tissue growth factor); TGF-b1 (transforming growth factor beta 1); MMP-2 (matrix metalloproteinase 2). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Gurtner G.C. Werner S. Barrandon Y. Longaker M.T. Wound repair and regeneration Nature 2008 453 314 321 10.1038/nature07039 18480812 2. Diegelmann R.F. Evans M.C. Wound healing: An overview of acute, fibrotic and delayed healing Front. Biosci. 2004 9 283 289 10.2741/1184 14766366 3. Ogawa R. Akita S. Akaishi S. Aramaki-Hattori N. Dohi T. Hayashi T. Kishi K. Kono T. Matsumura H. Muneuchi G. Diagnosis and treatment of keloids and hypertrophic scars--Japan scar workshop consensus document 2018 Burn. Trauma 2019 7 39 10.1186/s41038-019-0175-y 31890718 4. Limandjaja G.C. Niessen F.B. Scheper R.J. Gibbs S. Hypertrophic scars and keloids: Overview of the evidence and practical guide for differentiating between these abnormal scars Exp. Dermatol. 2021 30 146 161 10.1111/exd.14121 32479693 5. Marshall C.D. Hu M.S. Leavitt T. Barnes L.A. Lorenz H.P. Longaker M.T. Cutaneous scarring: Basic science, current treatments, and future directions Adv. Wound Care 2018 7 29 45 10.1089/wound.2016.0696 29392092 6. Lawrence J.W. Mason S.T. Schomer K. Klein M.B. Epidemiology and impact of scarring after burn injury: A systematic review of the literature J. Burn. Care Res. 2012 33 136 146 10.1097/BCR.0b013e3182374452 22138807 7. Gauglitz G.G. Korting H.C. Pavicic T. Ruzicka T. Jeschke M.G. Hypertrophic scarring and keloids: Pathomechanisms and current and emerging treatment strategies Mol. Med. 2011 17 113 125 10.2119/molmed.2009.00153 20927486 8. Thompson C.M. Hocking A.M. Honari S. Muffley L.A. Ga M. Gibran N.S. Genetic risk factors for hypertrophic scar development J. Burn. Care Res. 2013 34 477 482 10.1097/BCR.0b013e3182a2aa41 23966119 9. Shu F. Liu H. Lou X. Zhou Z. Zhao Z. Liu Y. Bai X. Luo P. Zheng Y. Xiao S. Analysis of the predictors of hypertrophic scarring pain and neuropathic pain after burn Burns 2022 48 1425 1434 10.1016/j.burns.2021.08.007 34670714 10. Mahdavian Delavary B. van der Veer W.M. Ferreira J.A. Niessen F.B. Formation of hypertrophic scars: Evolution and susceptibility J. Plast. Surg. Hand Surg. 2012 46 95 101 10.3109/2000656X.2012.669184 22471257 11. Brissett A.E. Sherris D.A. Scar contractures, hypertrophic scars, and keloids Facial Plast. Surg. 2001 17 263 272 10.1055/s-2001-18827 11735059 12. Huang C. Wu Z. Du Y. Ogawa R. The epidemiology of keloids Textbook on Scar Management Springer Cham, Switzerland 2020 29 35 13. Kouotou E.A. Nansseu J.R. Omona Guissana E. Mendouga Menye C.R. Akpadjan F. Tounkara T.M. Zoung-Kanyi Bissek A.C. Ndjitoyap Ndam E.C. Epidemiology and clinical features of keloids in Black Africans: A nested case-control study from Yaounde, Cameroon Int. J. Derm. 2019 58 1135 1140 10.1111/ijd.14610 31429935 14. Kwan P. Hori K. Ding J. Tredget E.E. Scar and contracture: Biological principles Hand Clin 2009 25 511 528 10.1016/j.hcl.2009.06.007 19801124 15. Hawkins H. Herndon D. Total Burn Care Saunders Elsevier Philadelphia, PA, USA American Medical Association Chicago, IL, USA 2007 608 619 16. Wurzer P. Forbes A.A. Hundeshagen G. Andersen C.R. Epperson K.M. Meyer W.J. Kamolz L.P. Branski L.K. Suman O.E. Herndon D.N. Two-year follow-up of outcomes related to scarring and distress in children with severe burns Disabil. Rehabil. 2017 39 1639 1643 10.1080/09638288.2016.1209579 27685196 17. Aarabi S. Longaker M.T. Gurtner G.C. Hypertrophic scar formation following burns and trauma: New approaches to treatment PLoS Med. 2007 4 e234 10.1371/journal.pmed.0040234 17803351 18. Atiyeh B. Ibrahim A. Nonsurgical management of hypertrophic scars: Evidence-based therapies, standard practices, and emerging methods: An update Aesthetic Plast. Surg. 2020 44 1345 1347 10.1007/s00266-020-01766-3 32766896 19. Brusselaers N. Pirayesh A. Hoeksema H. Verbelen J. Blot S. Monstrey S. Burn scar assessment: A systematic review of different scar scales J. Surg. Res. 2010 164 e115 e123 10.1016/j.jss.2010.05.056 20828761 20. Elrefaie A.M. Salem R.M. Faheem M.H. High-resolution ultrasound for keloids and hypertrophic scar assessment Lasers Med. Sci. 2020 35 379 385 10.1007/s10103-019-02830-4 31240510 21. Rodrigues M. Kosaric N. Bonham C.A. Gurtner G.C. Wound Healing: A Cellular Perspective Physiol. Rev. 2019 99 665 706 10.1152/physrev.00067.2017 30475656 22. Gale A.J. Continuing education course #2: Current understanding of hemostasis Toxicol. Pathol. 2011 39 273 280 10.1177/0192623310389474 21119054 23. Palta S. Saroa R. Palta A. Overview of the coagulation system Indian J. Anaesth. 2014 58 515 10.4103/0019-5049.144643 25535411 24. Martin P. Wound healing--aiming for perfect skin regeneration Science 1997 276 75 81 10.1126/science.276.5309.75 9082989 25. Singer A.J. Clark R.A. Cutaneous wound healing N. Engl. J. Med. 1999 341 738 746 10.1056/NEJM199909023411006 10471461 26. Ellis S. Lin E.J. Tartar D. Immunology of wound healing Curr. Dermatol. Rep. 2018 7 350 358 10.1007/s13671-018-0234-9 30524911 27. Turabelidze A. Dipietro L.A. Inflammation and wound healing Endod. Top. 2011 24 26 38 10.1111/etp.12012 28. Kolaczkowska E. Kubes P. Neutrophil recruitment and function in health and inflammation Nat. Rev. Immunol. 2013 13 159 175 10.1038/nri3399 23435331 29. Mittal M. Nepal S. Tsukasaki Y. Hecquet C.M. Soni D. Rehman J. Tiruppathi C. Malik A.B. Neutrophil activation of endothelial cell-expressed TRPM2 mediates transendothelial neutrophil migration and vascular injury Circ. Res. 2017 121 1081 1091 10.1161/CIRCRESAHA.117.311747 28790198 30. Bettinger D.A. Pellicane J.V. Tarry W.C. Yager D.R. Diegelmann R.F. Lee R. Cohen I. DeMaria E.J. The role of inflammatory cytokines in wound healing: Accelerated healing in endotoxin-resistant mice J. Trauma 1994 36 810 813 810-813; discussion 813 10.1097/00005373-199406000-00010 8015002 31. Nosenko M. Ambaryan S. Drutskaya M. Proinflammatory cytokines and skin wound healing in mice Mol. Biol. 2019 53 653 664 10.1134/S0026893319050121 32. Kim S.Y. Nair M.G. Macrophages in wound healing: Activation and plasticity Immunol. Cell Biol. 2019 97 258 267 10.1111/imcb.12236 30746824 33. Kloc M. Ghobrial R.M. Wosik J. Lewicka A. Lewicki S. Kubiak J.Z. Macrophage functions in wound healing J. Tissue Eng. Regen. Med. 2019 13 99 109 10.1002/term.2772 30445662 34. Eming S.A. Martin P. Tomic-Canic M. Wound repair and regeneration: Mechanisms, signaling, and translation Sci. Transl. Med. 2014 6 265sr266 10.1126/scitranslmed.3009337 35. Kotwal G.J. Chien S. Macrophage differentiation in normal and accelerated wound healing Macrophages 2017 62 353 364 36. Demidova-Rice T.N. Hamblin M.R. Herman I.M. Acute and impaired wound healing: Pathophysiology and current methods for drug delivery, part 1: Normal and chronic wounds: Biology, causes, and approaches to care Adv. Ski. Wound Care 2012 25 304 10.1097/01.ASW.0000416006.55218.d0 37. Duncan M.R. Frazier K.S. Abramson S. Williams S. Klapper H. Huang X. Grotendorst G.R. Connective tissue growth factor mediates transforming growth factor beta-induced collagen synthesis: Down-regulation by cAMP FASEB J. 1999 13 1774 1786 10.1096/fasebj.13.13.1774 10506580 38. Barker T.H. Engler A.J. The provisional matrix: Setting the stage for tissue repair outcomes Matrix Biol. 2017 60 1 4 10.1016/j.matbio.2017.04.003 28527902 39. Broughton G. Janis J.E. Attinger C.E. The basic science of wound healing Plast Reconstr Surg 2006 117 12S 34S 10.1097/01.prs.0000225430.42531.c2 16799372 40. Hakkinen L. Larjava H. Koivisto L. Granulation tissue formation and remodeling Endod. Top. 2011 24 94 129 10.1111/etp.12008 41. Rousselle P. Braye F. Dayan G. Re-epithelialization of adult skin wounds: Cellular mechanisms and therapeutic strategies Adv. Drug Deliv. Rev. 2019 146 344 365 10.1016/j.addr.2018.06.019 29981800 42. Pastar I. Stojadinovic O. Yin N.C. Ramirez H. Nusbaum A.G. Sawaya A. Patel S.B. Khalid L. Isseroff R.R. Tomic-Canic M. Epithelialization in Wound Healing: A Comprehensive Review Adv. Wound Care 2014 3 445 464 10.1089/wound.2013.0473 43. DiPietro L.A. Angiogenesis and wound repair: When enough is enough J. Leukoc. Biol. 2016 100 979 984 10.1189/jlb.4MR0316-102R 27406995 44. Grotendorst G. Grotendorst C. Gilman T. Production of growth factors (PDGF & TGF-beta) at the site of tissue repair Prog. Clin. Biol. Res. 1988 266 47 54 3164110 45. Uutela M. Wirzenius M. Paavonen K. Rajantie I. He Y. Karpanen T. Lohela M. Wiig H. Salven P. Pajusola K. PDGF-D induces macrophage recruitment, increased interstitial pressure, and blood vessel maturation during angiogenesis Blood 2004 104 3198 3204 10.1182/blood-2004-04-1485 15271796 46. Nissen N.N. Polverini P.J. Gamelli R.L. DiPietro L.A. Basic fibroblast growth factor mediates angiogenic activity in early surgical wounds Surgery 1996 119 457 465 10.1016/S0039-6060(96)80148-6 8644013 47. Pardali E. Goumans M.-J. ten Dijke P. Signaling by members of the TGF-b family in vascular morphogenesis and disease Trends Cell Biol. 2010 20 556 567 10.1016/j.tcb.2010.06.006 20656490 48. Sorg H. Tilkorn D.J. Hager S. Hauser J. Mirastschijski U. Skin wound healing: An update on the current knowledge and concepts Eur. Surg. Res. 2017 58 81 94 10.1159/000454919 27974711 49. Diller R.B. Tabor A.J. The role of the extracellular matrix (ECM) in wound healing: A review Biomimetics 2022 7 87 10.3390/biomimetics7030087 35892357 50. Xue M. Jackson C.J. Extracellular matrix reorganization during wound healing and its impact on abnormal scarring Adv. Wound Care 2015 4 119 136 10.1089/wound.2013.0485 25785236 51. Reinke J.M. Sorg H. Wound repair and regeneration Eur. Surg. Res. 2012 49 35 43 10.1159/000339613 22797712 52. Shanmugam V.K. Fernandez S.J. Evans K.K. McNish S. Banerjee A.N. Couch K.S. Mete M. Shara N. Postoperative wound dehiscence: Predictors and associations Wound Repair Regen. 2015 23 184 190 10.1111/wrr.12268 25683272 53. Spiliotis J. Tsiveriotis K. Datsis A.D. Vaxevanidou A. Zacharis G. Giafis K. Kekelos S. Rogdakis A. Wound dehiscence: Is still a problem in the 21th century: A retrospective study World J. Emerg. Surg. 2009 4 1 5 10.1186/1749-7922-4-12 19144131 54. Li B. Wang J.H.-C. Fibroblasts and myofibroblasts in wound healing: Force generation and measurement J. Tissue Viability 2011 20 108 120 10.1016/j.jtv.2009.11.004 19995679 55. Wang Z.C. Zhao W.Y. Cao Y. Liu Y.Q. Sun Q. Shi P. Cai J.Q. Shen X.Z. Tan W.Q. The Roles of Inflammation in Keloid and Hypertrophic Scars Front. Immunol. 2020 11 603187 10.3389/fimmu.2020.603187 33343575 56. Bellemare J. Roberge C.J. Bergeron D. Lopez-Valle C.A. Roy M. Moulin V.J. Epidermis promotes dermal fibrosis: Role in the pathogenesis of hypertrophic scars J. Pathol. A J. Pathol. Soc. Great Br. Irel. 2005 206 1 8 10.1002/path.1737 15772942 57. Molina E.A. Hartmann B. Oliver M.A. Kirkpatrick L.D. Keyloun J.W. Moffatt L.T. Shupp J.W. Travis T.E. Carney B.C. Angiogenic gene characterization and vessel permeability of dermal microvascular endothelial cells isolated from burn hypertrophic scar Sci. Rep. 2022 12 12222 10.1038/s41598-022-16376-z 35851095 58. Fuchs P. Calitz C. Pavlovic N. Binet F. Solbak S.M.O. Danielson U.H. Kreuger J. Heindryckx F. Gerwins P. Fibrin fragment E potentiates TGF-b-induced myofibroblast activation and recruitment Cell. Signal. 2020 72 109661 10.1016/j.cellsig.2020.109661 32334027 59. Wang Y. Guerrero-Juarez C.F. Qiu Y. Du H. Chen W. Figueroa S. Plikus M.V. Nie Q. A multiscale hybrid mathematical model of epidermal-dermal interactions during skin wound healing Exp. Derm. 2019 28 493 502 10.1111/exd.13909 30801791 60. Werner S. Grose R. Regulation of wound healing by growth factors and cytokines Physiol. Rev. 2003 83 835 870 10.1152/physrev.2003.83.3.835 12843410 61. Nam S.M. Kim Y.B. The effects of platelet-rich plasma on hypertrophic scars fibroblasts Int. Wound J. 2018 15 547 554 10.1111/iwj.12896 29781178 62. Agrawal A. Ding J. Agrawal B. Kwan P.O. Tredget E.E. Stimulation of toll-like receptor pathways by burn eschar tissue as a possible mechanism for hypertrophic scarring Wound Repair Regen. 2021 29 810 819 10.1111/wrr.12940 34043867 63. Wang J. Hori K. Ding J. Huang Y. Kwan P. Ladak A. Tredget E.E. Toll-like receptors expressed by dermal fibroblasts contribute to hypertrophic scarring J. Cell Physiol. 2011 226 1265 1273 10.1002/jcp.22454 20945369 64. Williams H. Suda S. Dervish S. Yap Y.T. Holland A.J. Medbury H.J. Monocyte M1/M2 profile is altered in paediatric burn patients with hypertrophic scarring Wound Repair Regen. 2021 29 996 1005 10.1111/wrr.12960 34272902 65. Chen B. Li H. Xia W. The role of Th1/Th2 cell chemokine expression in hypertrophic scar Int. Wound J. 2020 17 197 205 10.1111/iwj.13257 31691483 66. van den Broek L.J. van der Veer W.M. de Jong E.H. Gibbs S. Niessen F.B. Suppressed inflammatory gene expression during human hypertrophic scar compared to normotrophic scar formation Exp. Derm. 2015 24 623 629 10.1111/exd.12739 25939875 67. Ding J. Hori K. Zhang R. Marcoux Y. Honardoust D. Shankowsky H.A. Tredget E.E. Stromal cell-derived factor 1 (SDF-1) and its receptor CXCR4 in the formation of postburn hypertrophic scar (HTS) Wound Repair Regen. 2011 19 568 578 10.1111/j.1524-475X.2011.00724.x 22092795 68. Henry G. Garner W.L. Inflammatory mediators in wound healing Surg. Clin. N. Am. 2003 83 483 507 10.1016/S0039-6109(02)00200-1 12822721 69. Zhang D. Li B. Zhao M. Therapeutic Strategies by Regulating Interleukin Family to Suppress Inflammation in Hypertrophic Scar and Keloid Front. Pharm. 2021 12 667763 10.3389/fphar.2021.667763 70. Butzelaar L. Schooneman D.P. Soykan E.A. Talhout W. Ulrich M.M. van den Broek L.J. Gibbs S. Beelen R.H. Mink van der Molen A.B. Niessen F.B. Inhibited early immunologic response is associated with hypertrophic scarring Exp. Derm. 2016 25 797 804 10.1111/exd.13100 27249786 71. Roy R. Zayas J. Mohamed M.F. Aboonabi A. Delgado K. Wallace J. Bayat M. Kuzel T.M. Reiser J. Shafikhani S.H. IL-10 Dysregulation Underlies Chemokine Insufficiency, Delayed Macrophage Response, and Impaired Healing in Diabetic Wounds J. Investig. Dermatol. 2022 142 692 704.e614 10.1016/j.jid.2021.08.428 34517005 72. Roy R. Zayas J. Singh S.K. Delgado K. Wood S.J. Mohamed M.F. Frausto D.M. Estupinian R. Giurini E.F. Kuzel T.M. Overriding impaired FPR chemotaxis signaling in diabetic neutrophil stimulates infection control in murine diabetic wound eLife 2022 11 e72071 10.7554/eLife.72071 35112667 73. Wood S. Jayaraman V. Huelsmann E.J. Bonish B. Burgad D. Sivaramakrishnan G. Qin S. Dipietro L.A. Zloza A. Zhang C. Pro-inflammatory chemokine CCL2 (MCP-1) promotes healing in diabetic wounds by restoring the macrophage response PLoS ONE 2014 9 e91574 10.1371/journal.pone.0091574 24618995 74. Goldufsky J. Wood S.J. Jayaraman V. Majdobeh O. Chen L. Qin S. Zhang C. DiPietro L.A. Shafikhani S.H. Pseudomonas aeruginosa uses T3SS to inhibit diabetic wound healing Wound Repair Regen. 2015 23 557 564 10.1111/wrr.12310 25912785 75. Kaplanski G. Marin V. Montero-Julian F. Mantovani A. Farnarier C. IL-6: A regulator of the transition from neutrophil to monocyte recruitment during inflammation Trends Immunol. 2003 24 25 29 10.1016/S1471-4906(02)00013-3 12495721 76. Fernando M.R. Reyes J.L. Iannuzzi J. Leung G. McKay D.M. The pro-inflammatory cytokine, interleukin-6, enhances the polarization of alternatively activated macrophages PLoS ONE 2014 9 e94188 10.1371/journal.pone.0094188 24736635 77. Juhl P. Bondesen S. Hawkins C.L. Karsdal M.A. Bay-Jensen A.C. Davies M.J. Siebuhr A.S. Dermal fibroblasts have different extracellular matrix profiles induced by TGF-b, PDGF and IL-6 in a model for skin fibrosis Sci. Rep. 2020 10 17300 10.1038/s41598-020-74179-6 33057073 78. Xue H. McCauley R.L. Zhang W. Martini D.K. Altered interleukin-6 expression in fibroblasts from hypertrophic burn scars J. Burn. Care Rehabil. 2000 21 142 146 10.1097/00004630-200021020-00010 10752747 79. Ray S. Ju X. Sun H. Finnerty C.C. Herndon D.N. Brasier A.R. The IL-6 trans-signaling-STAT3 pathway mediates ECM and cellular proliferation in fibroblasts from hypertrophic scar J. Investig. Dermatol. 2013 133 1212 1220 10.1038/jid.2012.499 23303450 80. Kwan P.O. Ding J. Tredget E.E. Serum Decorin, Interleukin-1b, and Transforming Growth Factor-b Predict Hypertrophic Scarring Postburn J. Burn Care Res. 2016 37 356 366 10.1097/BCR.0000000000000271 27828834 81. Zhang J. Qiao Q. Liu M. He T. Shi J. Bai X. Zhang Y. Li Y. Cai W. Han S. IL-17 Promotes Scar Formation by Inducing Macrophage Infiltration Am. J. Pathol. 2018 188 1693 1702 10.1016/j.ajpath.2018.04.005 29753790 82. Nguyen J.K. Austin E. Huang A. Mamalis A. Jagdeo J. The IL-4/IL-13 axis in skin fibrosis and scarring: Mechanistic concepts and therapeutic targets Arch. Dermatol. Res. 2020 312 81 92 10.1007/s00403-019-01972-3 31493000 83. Shi J. Li J. Guan H. Cai W. Bai X. Fang X. Hu X. Wang Y. Wang H. Zheng Z. Anti-fibrotic actions of interleukin-10 against hypertrophic scarring by activation of PI3K/AKT and STAT3 signaling pathways in scar-forming fibroblasts PLoS ONE 2014 9 e98228 10.1371/journal.pone.0098228 24878845 84. Shi J. Shi S. Xie W. Zhao M. Li Y. Zhang J. Li N. Bai X. Cai W. Hu X. IL-10 alleviates lipopolysaccharide-induced skin scarring via IL-10R/STAT3 axis regulating TLR4/NF-kB pathway in dermal fibroblasts J. Cell Mol. Med. 2021 25 1554 1567 10.1111/jcmm.16250 33410606 85. Ma L. Hua L. Yu W. Ke L. Li L. TSG-6 inhibits hypertrophic scar fibroblast proliferation by regulating IRE1a/TRAF2/NF-kB signalling Int. Wound J. 2022 10.1111/iwj.13950 36056472 86. Yang J.-X. Li S.-Y. Chen M.-L. He L.-R. The role of altered fatty acid in pathological scars and their dermal fibroblasts Chin. J. Traumatol. 2022 25 218 223 10.1016/j.cjtee.2022.03.006 35478090 87. Yang J. Chen M. He L. To Explore Ideas From the Altered Metabolites: The Metabolomics of Pathological Scar J. Craniofacial Surg. 2022 33 1619 1625 10.1097/SCS.0000000000008470 35045014 88. Gromovsky A.D. Schugar R.C. Brown A.L. Helsley R.N. Burrows A.C. Ferguson D. Zhang R. Sansbury B.E. Lee R.G. Morton R.E. D-5 fatty acid desaturase FADS1 impacts metabolic disease by balancing proinflammatory and proresolving lipid mediators Arterioscler. Thromb. Vasc. Biol. 2018 38 218 231 10.1161/ATVBAHA.117.309660 29074585 89. Fritsche K.L. The science of fatty acids and inflammation Adv. Nutr. 2015 6 293S 301S 10.3945/an.114.006940 25979502 90. Vinolo M.A. Rodrigues H.G. Nachbar R.T. Curi R. Regulation of inflammation by short chain fatty acids Nutrients 2011 3 858 876 10.3390/nu3100858 22254083 91. van der Veer W.M. Niessen F.B. Ferreira J.A. Zwiers P.J. de Jong E.H. Middelkoop E. Molema G. Time course of the angiogenic response during normotrophic and hypertrophic scar formation in humans Wound Repair Regen. 2011 19 292 301 10.1111/j.1524-475X.2011.00692.x 21539647 92. Machesney M. Tidman N. Waseem A. Kirby L. Leigh I. Activated keratinocytes in the epidermis of hypertrophic scars Am J. Pathol. 1998 152 1133 1141 9588880 93. Ma L. Gan C. Huang Y. Wang Y. Luo G. Wu J. Comparative proteomic analysis of extracellular matrix proteins secreted by hypertrophic scar with normal skin fibroblasts Burn. Trauma 2014 2 76 83 10.4103/2321-3868.130191 94. Lian N. Li T. Growth factor pathways in hypertrophic scars: Molecular pathogenesis and therapeutic implications Biomed Pharm. 2016 84 42 50 10.1016/j.biopha.2016.09.010 95. Eremenko E. Ding J. Kwan P. Tredget E.E. The Biology of Extracellular Matrix Proteins in Hypertrophic Scarring Adv. Wound Care 2022 11 234 254 10.1089/wound.2020.1257 96. Varkey M. Ding J. Tredget E.E. Differential collagen-glycosaminoglycan matrix remodeling by superficial and deep dermal fibroblasts: Potential therapeutic targets for hypertrophic scar Biomaterials 2011 32 7581 7591 10.1016/j.biomaterials.2011.06.070 21802722 97. Huang X. Gu S. Liu C. Zhang L. Zhang Z. Zhao Y. Khoong Y. Li H. Gao Y. Liu Y. CD39+ fibroblasts enhance myofibroblast activation by promoting IL-11 secretion in hypertrophic scars J. Investig. Dermatol. 2022 142 1065 1076.e1019 10.1016/j.jid.2021.07.181 34537192 98. Zhang T. Wang X.F. Wang Z.C. Lou D. Fang Q.Q. Hu Y.Y. Zhao W.Y. Zhang L.Y. Wu L.H. Tan W.Q. Current potential therapeutic strategies targeting the TGF-b/Smad signaling pathway to attenuate keloid and hypertrophic scar formation Biomed. Pharm. 2020 129 110287 10.1016/j.biopha.2020.110287 32540643 99. Fang X. Hu X. Zheng Z. Tao K. Wang H. Guan H. Shi J. Ji P. Cai W. Bai X. Smad interacting protein 1 influences transforming growth factor-b J. Mol. Histol. 2019 50 503 514 10.1007/s10735-019-09844-w 31595443 100. Li X.P. Liu P. Li Y.F. Zhang G.L. Zeng D.S. Liu D.L. LPS induces activation of the TLR4 pathway in fibroblasts and promotes skin scar formation through collagen I and TGF-b in skin lesions Int. J. Clin. Exp. Pathol. 2019 12 2121 2129 31934034 101. Ogawa R. Akaishi S. Endothelial dysfunction may play a key role in keloid and hypertrophic scar pathogenesis-keloids and hypertrophic scars may be vascular disorders Med. Hypotheses 2016 96 51 60 10.1016/j.mehy.2016.09.024 27959277 102. Laberge A. Merjaneh M. Arif S. Larochelle S. Moulin V.J. Shedding of proangiogenic microvesicles from hypertrophic scar myofibroblasts Exp. Derm. 2021 30 112 120 10.1111/exd.14178 32813921 103. Wilgus T.A. Vascular endothelial growth factor and cutaneous scarring Adv. Wound Care 2019 8 671 678 10.1089/wound.2018.0796 104. Li B. Gao C. Diao J.S. Wang D.L. Chu F.F. Li Y. Wang G. Guo S.Z. Xia W. Aberrant Notch signalling contributes to hypertrophic scar formation by modulating the phenotype of keratinocytes Exp Derm. 2016 25 137 142 10.1111/exd.12897 26566963 105. Yuan F.L. Sun Z.L. Feng Y. Liu S.Y. Du Y. Yu S. Yang M.L. Lv G.Z. Epithelial-mesenchymal transition in the formation of hypertrophic scars and keloids J. Cell Physiol. 2019 234 21662 21669 10.1002/jcp.28830 31106425 106. Zhao J. Yu J. Xu Y. Chen L. Zhou F. Zhai Q. Wu J. Shu B. Qi S. Epidermal HMGB1 activates dermal fibroblasts and causes hypertrophic scar formation in reduced hydration J. Investig. Dermatol. 2018 138 2322 2332 10.1016/j.jid.2018.04.036 29787749 107. Ma D. Chen L. Shi J. Zhao Y. Vasani S. Chen K. Romana-Souza B. Henkin J. DiPietro L.A. Pigment epithelium-derived factor attenuates angiogenesis and collagen deposition in hypertrophic scars Wound Repair Regen. 2020 28 684 695 10.1111/wrr.12828 32585760 108. Eto H. Suga H. Aoi N. Kato H. Kuno S. Tabata Y. Yoshimura K. Therapeutic potential of fibroblast growth factor-2 for hypertrophic scars: Upregulation of MMP-1 and HGF expression Lab. Investig. 2012 92 214 223 10.1038/labinvest.2011.127 21946856 109. Spyrou G.E. Naylor I.L. The effect of basic fibroblast growth factor on scarring Br. J. Plast. Surg. 2002 55 275 282 10.1054/bjps.2002.3831 12160530 110. Ulrich D. Ulrich F. Unglaub F. Piatkowski A. Pallua N. Matrix metalloproteinases and tissue inhibitors of metalloproteinases in patients with different types of scars and keloids J. Plast. Reconstr. Aesthetic Surg. 2010 63 1015 1021 10.1016/j.bjps.2009.04.021 111. Oliveira G.V. Hawkins H.K. Chinkes D. Burke A. Tavares A.L. Ramos-e-Silva M. Albrecht T.B. Kitten G.T. Herndon D.N. Hypertrophic versus non hypertrophic scars compared by immunohistochemistry and laser confocal microscopy: Type I and III collagens Int. Wound J. 2009 6 445 452 10.1111/j.1742-481X.2009.00638.x 20051096 112. Moulin V. Larochelle S. Langlois C. Thibault I. Lopez-Valle C.A. Roy M. Normal skin wound and hypertrophic scar myofibroblasts have differential responses to apoptotic inductors J. Cell Physiol. 2004 198 350 358 10.1002/jcp.10415 14755540 113. Sidgwick G.P. Bayat A. Extracellular matrix molecules implicated in hypertrophic and keloid scarring J. Eur. Acad Derm. Venereol. 2012 26 141 152 10.1111/j.1468-3083.2011.04200.x 21838832 114. Gabbiani G. The myofibroblast in wound healing and fibrocontractive diseases J. Pathol. 2003 200 500 503 10.1002/path.1427 12845617 115. Junker J.P. Kratz C. Tollback A. Kratz G. Mechanical tension stimulates the transdifferentiation of fibroblasts into myofibroblasts in human burn scars Burns 2008 34 942 946 10.1016/j.burns.2008.01.010 18472340 116. Darby I.A. Laverdet B. Bonte F. Desmouliere A. Fibroblasts and myofibroblasts in wound healing Clin. Cosmet. Investig. Dermatol. 2014 7 301 25395868 117. Lee D.E. Trowbridge R.M. Ayoub N.T. Agrawal D.K. High-mobility Group Box Protein-1, Matrix Metalloproteinases, and Vitamin D in Keloids and Hypertrophic Scars Plast. Reconstr. Surg. Glob. Open 2015 3 e425 10.1097/GOX.0000000000000391 26180726 118. Travis T.E. Ghassemi P. Prindeze N.J. Moffatt L.T. Carney B.C. Alkhalil A. Ramella-Roman J.C. Shupp J.W. Matrix Metalloproteinases Are Differentially Regulated and Responsive to Compression Therapy in a Red Duroc Model of Hypertrophic Scar Eplasty 2018 18 e1 29375731 119. Keskin E.S. Keskin E.R. Ozturk M.B. Cakan D. The Effect of MMP-1 on Wound Healing and Scar Formation Aesthetic Plast Surg. 2021 45 2973 2979 10.1007/s00266-021-02369-2 34075460 120. McArdle C. Abbah S.A. Bhowmick S. Collin E. Pandit A. Localized temporal co-delivery of interleukin 10 and decorin genes using amediated by collagen-based biphasic scaffold modulates the expression of TGF-b1/b2 in a rabbit ear hypertrophic scarring model Biomater. Sci. 2021 9 3136 3149 10.1039/D0BM01928C 33725045 121. Ying M. Chen Y. Yuan B. Knockdown of lysyl oxidase like 1 inhibits the proliferation and pro-fibrotic effects of transforming growth factor-b1-induced hypertrophic scar fibroblasts Can. J. Physiol. Pharm. 2021 99 1272 1279 10.1139/cjpp-2021-0242 122. Raktoe R.S. Rietveld M.H. Out-Luiting J.J. Kruithof-de Julio M. van Zuijlen P.P. van Doorn R. El Ghalbzouri A. The effect of TGFbRI inhibition on fibroblast heterogeneity in hypertrophic scar 2D in vitro models Burns 2021 47 1563 1575 10.1016/j.burns.2021.01.004 33558094 123. Moulin V.J. Three-dimensional model of hypertrophic scar using a tissue-engineering approach Myofibroblasts Springer Berlin/Heidelberg, Germany 2021 419 434 124. Sun Q. Yu E.T. Zhou Y. Tong S. Sun X. Li K.Z. Lv M.Z. Guo S. The Effects of Timing of Postoperative Radiotherapy on Hypertrophic Scar in a Rabbit Model Med. Sci. Monit. Int. Med. J. Exp. Clin. Res. 2020 26 e921263 10.12659/MSM.921263 125. Huang J. Chen J. Wo Y. Wang X. Zhang Y. Chen X. Zhang Z. Biskup E. CO2 Fractional Laser Combined with 5-Fluorouracil Ethosomal Gel Treatment of Hypertrophic Scar Macro-, Microscopic, and Molecular Mechanism of Action in a Rabbit Animal Model Rejuvenation Res. 2021 24 131 138 10.1089/rej.2019.2204 32578501 126. Tunca M. Gamsizkan M. Yurekli A. Alp Goksel B. Fuat Cicek A. Caliskan E. Cryosurgery to remove perichondrium for the rabbit ear hypertrophic scar model: A simplified method Acta Dermatovenerol. Alp. Pannonica Adriat. 2019 28 57 59 10.15570/actaapa.2019.15 31233167 127. Friedrich E.E. Niknam-Bienia S. Xie P. Jia S.X. Hong S.J. Mustoe T.A. Galiano R.D. Thermal injury model in the rabbit ear with quantifiable burn progression and hypertrophic scar Wound Repair Regen. 2017 25 327 337 10.1111/wrr.12518 28370931 128. Zu W. Jiang B. Liu H. Establishment of a long-term hypertrophic scar model by injection of anhydrous alcohol: A rabbit model Int. J. Exp. Pathol. 2021 102 105 112 10.1111/iep.12389 33710702 129. Alrobaiea S.M. Ding J. Ma Z. Tredget E.E. A novel nude mouse model of hypertrophic scarring using scratched full thickness human skin grafts Adv. Wound Care 2016 5 299 313 10.1089/wound.2015.0670 27366591 130. Davidson J.M. Yu F. Opalenik S.R. Splinting Strategies to Overcome Confounding Wound Contraction in Experimental Animal Models Adv. Wound Care 2013 2 142 148 10.1089/wound.2012.0424 24527337 131. Son D.O. Hinz B. A Rodent Model of Hypertrophic Scarring: Splinting of Rat Wounds Methods Mol. Biol. 2021 2299 405 417 10.1007/978-1-0716-1382-5_27 34028757 132. Naldaiz-Gastesi N. Bahri O.A. Lopez de Munain A. McCullagh K.J.A. Izeta A. The panniculus carnosus muscle: An evolutionary enigma at the intersection of distinct research fields J. Anat. 2018 233 275 288 10.1111/joa.12840 29893024 133. Aarabi S. Bhatt K.A. Shi Y. Paterno J. Chang E.I. Loh S.A. Holmes J.W. Longaker M.T. Yee H. Gurtner G.C. Mechanical load initiates hypertrophic scar formation through decreased cellular apoptosis FASEB J. 2007 21 3250 3261 10.1096/fj.07-8218com 17504973 134. Sylakowski K. Hwang M.P. Justin A. Whaley D. Wang Y. Wells A. The matricellular protein decorin delivered intradermally with coacervate improves wound resolution in the CXCR3-deficient mouse model of hypertrophic scarring Wound Repair Regen. 2022 30 436 447 10.1111/wrr.13017 35470921 135. Yates C.C. Krishna P. Whaley D. Bodnar R. Turner T. Wells A. Lack of CXC chemokine receptor 3 signaling leads to hypertrophic and hypercellular scarring Am. J. Pathol. 2010 176 1743 1755 10.2353/ajpath.2010.090564 20203286 136. Yates C.C. Whaley D. Kulasekeran P. Hancock W.W. Lu B. Bodnar R. Newsome J. Hebda P.A. Wells A. Delayed and deficient dermal maturation in mice lacking the CXCR3 ELR-negative CXC chemokine receptor Am. J. Pathol. 2007 171 484 495 10.2353/ajpath.2007.061092 17600132 137. Jimi S. Saparov A. Koizumi S. Miyazaki M. Takagi S. A novel mouse wound model for scar tissue formation in abdominal muscle wall J. Vet. Med. Sci. 2021 83 1933 1942 10.1292/jvms.21-0464 34719609 138. Li Y. Shan X. Mao Q. Xiang R. Cai Z. Botulinum toxin type a intralesional monotherapy for treating human hypertrophic scar in a dose-dependent manner: In an animal model J. Plast. Reconstr. Aesthetic Surg. 2021 74 3186 3195 10.1016/j.bjps.2021.03.062 34158272 139. Li Z. Li S. Li K. Jiang X. Zhang J. Liu H. A highly simulated scar model developed by grafting human thin split-thickness skin on back of nude mouse: The remodeling process, histological characteristics of scars Biochem. Biophys Res. Commun. 2020 526 744 750 10.1016/j.bbrc.2020.03.140 32265030 140. Ud-Din S. Bayat A. Classification of distinct endotypes in human skin scarring: SCAR--A novel perspective on dermal fibrosis Adv. Wound Care 2022 11 109 120 10.1089/wound.2020.1364 33677998 141. Nast A. Eming S. Fluhr J. Fritz K. Gauglitz G. Hohenleutner S. Panizzon R.G. Sebastian G. Sporbeck B. Koller J. German S2k guidelines for the therapy of pathological scars (hypertrophic scars and keloids) J. Dtsch Derm. Ges. 2012 10 747 762 10.1111/j.1610-0387.2012.08012.x 142. Ogawa R. Akaishi S. Huang C. Dohi T. Aoki M. Omori Y. Koike S. Kobe K. Akimoto M. Hyakusoku H. Clinical applications of basic research that shows reducing skin tension could prevent and treat abnormal scarring: The importance of fascial/subcutaneous tensile reduction sutures and flap surgery for keloid and hypertrophic scar reconstruction J. Nippon Med. Sch. 2011 78 68 76 10.1272/jnms.78.68 21551963 143. Ogawa R. Akaishi S. Kuribayashi S. Miyashita T. Keloids and Hypertrophic Scars Can Now Be Cured Completely: Recent Progress in Our Understanding of the Pathogenesis of Keloids and Hypertrophic Scars and the Most Promising Current Therapeutic Strategy J. Nippon Med. Sch. 2016 83 46 53 10.1272/jnms.83.46 27180789 144. Chen H.H. Huang T. Hsu Y. Chen H. Prospective clinical trial comparing barbed dermal suture and interrupted suture closure of the anterolateral thigh flap donor site in a Taiwanese population based on the Vancouver scar scale and the patient and observer scar assessment scale J. Plast. Reconstr. Aesthet. Surg. 2021 74 3196 3211 10.1016/j.bjps.2021.06.015 34247963 145. Ogawa R. Keloid and Hypertrophic Scars Are the Result of Chronic Inflammation in the Reticular Dermis Int. J. Mol. Sci. 2017 18 606 10.3390/ijms18030606 28287424 146. Macintyre L. Baird M. Pressure garments for use in the treatment of hypertrophic scars--a review of the problems associated with their use Burns 2006 32 10 15 10.1016/j.burns.2004.06.018 16413399 147. Engrav L.H. Heimbach D.M. Rivara F.P. Moore M.L. Wang J. Carrougher G.J. Costa B. Numhom S. Calderon J. Gibran N.S. 12-Year within-wound study of the effectiveness of custom pressure garment therapy Burns 2010 36 975 983 10.1016/j.burns.2010.04.014 20537469 148. Moortgat P. Meirte J. Maertens K. Lafaire C. De Cuyper L. Anthonissen M. Can a Cohesive Silicone Bandage Outperform an Adhesive Silicone Gel Sheet in the Treatment of Scars? A Randomized Comparative Trial Plast. Reconstr. Surg. 2019 143 902 911 10.1097/PRS.0000000000005369 30601234 149. Rogers N.T. Fancourt D. Cultural Engagement Is a Risk-Reducing Factor for Frailty Incidence and Progression J. Gerontol. Ser. B Psychol. Sci. Soc. Sci. 2020 75 571 576 10.1093/geronb/gbz004 30624696 150. Lee H.J. Jang Y.J. Recent Understandings of Biology, Prophylaxis and Treatment Strategies for Hypertrophic Scars and Keloids Int. J. Mol. Sci. 2018 19 711 10.3390/ijms19030711 29498630 151. Kalasho B.D. Kikuchi R. Zoumalan C.I. Silicone-Based Scar Cream for Post Upper Eyelid Blepharoplasty-associated Cicatricial and Hypertrophic Scarring J. Drugs Dermatol. 2019 18 440 446 31141848 152. Kim S.H. Lee S.J. Lee J.W. Jeong H.S. Suh I.S. Clinical trial to evaluate the efficacy of botulinum toxin type A injection for reducing scars in patients with forehead laceration: A double-blinded, randomized controlled study Medicine 2019 98 e16952 10.1097/MD.0000000000016952 31441893 153. Huang R.L. Ho C.K. Tremp M. Xie Y. Li Q. Zan T. Early Postoperative Application of Botulinum Toxin Type A Prevents Hypertrophic Scarring after Epicanthoplasty: A Split-Face, Double-Blind, Randomized Trial Plast. Reconstr. Surg. 2019 144 835 844 10.1097/PRS.0000000000006069 31568286 154. Abedini R. Mehdizade Rayeni N. Haddady Abianeh S. Rahmati J. Teymourpour A. Nasimi M. Botulinum Toxin Type A Injection for Mammoplasty and Abdominoplasty Scar Management: A Split-Scar Double-Blinded Randomized Controlled Study Aesthetic Plast. Surg. 2020 44 2270 2276 10.1007/s00266-020-01916-7 32813130 155. Chen Z. Pang R. Wei Z. Zhang H. Liu W. Li G. The effect of botulinum toxin injection dose on the appearance of surgical scar Sci. Rep. 2021 11 13670 10.1038/s41598-021-93203-x 34211099 156. Park G.S. An M.K. Yoon J.H. Park S.S. Koh S.H. Mauro T.M. Cho E.B. Park E.J. Kim K.H. Kim K.J. Botulinum toxin type A suppresses pro-fibrotic effects via the JNK signaling pathway in hypertrophic scar fibroblasts Arch. Dermatol. Res. 2019 311 807 814 10.1007/s00403-019-01975-0 31501922 157. Shin J.U. Park J. Lee J.H. Lee K.H. Kim Y.O. Yun C.O. Lee W.J. Extramarginal excision is preferable for hypertrophic scars Int. J. Derm. 2014 53 1138 1144 10.1111/ijd.12481 24898827 158. Schweinfurth J.M. Fedok F. Avoiding pitfalls and unfavorable outcomes in scar revision Facial Plast. Surg. 2001 17 273 278 10.1055/s-2001-18822 11735060 159. Vijaya Lakshmi Y. Swetha Reddy L. Naga Neelima Devi K. Phani Kumar K. Guru Karthik G. Srinivas Chakravarthy P. Nageswar Rao K. Evaluation of Microneedling Therapy in Management of Facial Scars J. Craniofacial Surg. 2020 31 e214 e217 10.1097/SCS.0000000000006145 160. Esmat S. Shokeir H.A. Samy N.A. Mahmoud S.B. Sayed S.S.E.D. Shaker E. Hilal R.F. Automated Microneedling Versus Fractional CO2 Laser in Treatment of Traumatic Scars: A Clinical and Histochemical Study Dermatol. Surg. 2021 47 1480 1485 10.1097/DSS.0000000000003227 34468410 161. El Mofty M. Esmat S. Hunter N. Mashaly H.M. Dorgham D. Shaker O. Ibrahim S. Effect of different types of therapeutic trauma on vitiligo lesions Dermatol. Ther. 2017 30 e12447 10.1111/dth.12447 27998025 162. Wong T.-S. Li J.Z.-H. Chen S. Chan J.Y.-W. Gao W. The efficacy of triamcinolone acetonide in keloid treatment: A systematic review and meta-analysis Front. Med. 2016 3 71 10.3389/fmed.2016.00071 163. Zhuang Z. Li Y. Wei X. The safety and efficacy of intralesional triamcinolone acetonide for keloids and hypertrophic scars: A systematic review and meta-analysis Burns 2021 47 987 998 10.1016/j.burns.2021.02.013 33814214 164. Khalid F.A. Mehrose M.Y. Saleem M. Yousaf M.A. Mujahid A.M. Rehman S.U. Ahmad S. Tarar M.N. Comparison of efficacy and safety of intralesional triamcinolone and combination of triamcinolone with 5-fluorouracil in the treatment of keloids and hypertrophic scars: Randomised control trial Burns 2019 45 69 75 10.1016/j.burns.2018.08.011 30340861 165. Bi M. Sun P. Li D. Dong Z. Chen Z. Intralesional Injection of Botulinum Toxin Type A Compared with Intralesional Injection of Corticosteroid for the Treatment of Hypertrophic Scar and Keloid: A Systematic Review and Meta-Analysis Med. Sci. Monit. 2019 25 2950 2958 10.12659/MSM.916305 31006769 166. Elshahed A.R. Elmanzalawy K.S. Shehata H. ElSaie M.L. Effect of botulinum toxin type A for treating hypertrophic scars: A split-scar, double-blind randomized controlled trial J. Cosmet. Derm. 2020 19 2252 2258 10.1111/jocd.13627 32668499 167. Henderson D.L. Cromwell T.A. Mes L.G. Argon and carbon dioxide laser treatment of hypertrophic and keloid scars Lasers Surg. Med. 1984 3 271 277 10.1002/lsm.1900030402 6427540 168. Nouri K. Rivas M.P. Stevens M. Ballard C.J. Singer L. Ma F. Vejjabhinanta V. Elsaie M.L. Elgart G.W. Comparison of the effectiveness of the pulsed dye laser 585 nm versus 595 nm in the treatment of new surgical scars Lasers Med. Sci. 2009 24 801 810 10.1007/s10103-009-0698-8 19572180 169. Seago M. Shumaker P.R. Spring L.K. Alam M. Al-Niaimi F. Rox Anderson R. Artzi O. Bayat A. Cassuto D. Chan H.H. Laser Treatment of Traumatic Scars and Contractures: 2020 International Consensus Recommendations Lasers Surg. Med. 2020 52 96 116 10.1002/lsm.23201 31820478 170. Tawfic S. Sayed S. Nada A. Manaa D. Shalaby S. Low-Density Fractional Laser in the Treatment of Hypertrophic Postburn Scars: A Randomized Clinical Trial Dermatol. Surg. 2020 46 e38 e44 10.1097/DSS.0000000000002293 31851017 171. Won T. Ma Q. Chen Z. Gao Z. Wu X. Zhang R. The efficacy and safety of low-energy carbon dioxide fractional laser use in the treatment of early-stage pediatric hypertrophic scars: A prospective, randomized, split-scar study Lasers Surg. Med. 2022 54 230 236 10.1002/lsm.23459 34487566 172. Manuskiatti W. Kaewkes A. Yan C. Ng J.N. Glahn J.Z. Wanitphakdeedecha R. Hypertrophic Scar Outcomes in Fractional Laser Monotherapy Versus Fractional Laser-Assisted Topical Corticosteroid Delivery: A Randomized Clinical Trial Acta Derm. Venereol. 2021 101 adv00416 10.2340/00015555-3781 33686446 173. Babu P. Meethale Thiruvoth F. Chittoria R.K. Intense pulsed light vs silicone gel sheet in the management of hypertrophic scars: An interventional comparative trial in the Indian population J. Cosmet. Laser Ther. 2019 21 234 237 10.1080/14764172.2018.1525741 30300020 174. Salem S.A.M. Abdel Hameed S.M. Mostafa A.E. Intense pulsed light versus cryotherapy in the treatment of hypertrophic scars: A clinical and histopathological study J. Cosmet. Derm. 2021 20 2775 2784 10.1111/jocd.13971 175. Salameh F. Shumaker P.R. Goodman G.J. Spring L.K. Seago M. Alam M. Al-Niaimi F. Cassuto D. Chan H.H. Dierickx C. Energy-based devices for the treatment of Acne Scars: 2022 International consensus recommendations Lasers Surg. Med. 2022 54 10 26 10.1002/lsm.23484 34719045 176. Daoud A.A. Gianatasio C. Rudnick A. Michael M. Waibel J. Efficacy of Combined Intense Pulsed Light (IPL) With Fractional CO Lasers Surg. Med. 2019 51 678 685 10.1002/lsm.23092 31090087 177. Waibel J.S. Wulkan A.J. Rudnick A. Daoud A. Treatment of hypertrophic scars using laser-assisted corticosteroid versus laser-assisted 5-fluorouracil delivery Dermatol. Surg. 2019 45 423 430 10.1097/DSS.0000000000001678 30365461 178. Shin J. Cho J.T. Park S.I. Jung S.N. Combination therapy using non-ablative fractional laser and intralesional triamcinolone injection for hypertrophic scars and keloids treatment Int. Wound J. 2019 16 1450 1456 10.1111/iwj.13213 31475470 179. Rasaii S. Sohrabian N. Gianfaldoni S. Hadibarhaghtalab M. Pazyar N. Bakhshaeekia A. Lotti T. Ramirez-Pacheco L.A. Lange C.S. Matta J. Intralesional triamcinolone alone or in combination with botulinium toxin A is ineffective for the treatment of formed keloid scar: A double blind controlled pilot study Dermatol. Ther. 2019 32 e12781 10.1111/dth.12781 30422367 180. Teplyi V. Grebchenko K. THE USAGE OF RADIOFREQUENCY ABLATION FOR TREATMENT OF KELOIDS AND HYPERTROPHIC SCARS Probl. Radiac Med. Radiobiol. 2019 24 561 573 10.33145/2304-8336-2019-24-561-573 31841496 181. Murphy A. LeVatte T. Boudreau C. Midgen C. Gratzer P. Marshall J. Bezuhly M. Angiotensin II Type I Receptor Blockade Is Associated with Decreased Cutaneous Scar Formation in a Rat Model Plast. Reconstr. Surg. 2019 144 803e 813e 10.1097/PRS.0000000000006173 182. Sobec R.L. Fodor L. Bodog F. Topical Oxandrolone Reduces Ear Hypertrophic Scar Formation in Rabbits Plast. Reconstr. Surg. 2019 143 481 487 10.1097/PRS.0000000000005275 30688891 183. Wang P. Gu L. Yang X. An Y. Bi H. Zhao Z. Wang Q. Du S. Qin Z. Dipeptidyl Peptidase-4 Inhibitor Reduces the Risk of Developing Hypertrophic Scars and Keloids following Median Sternotomy in Diabetic Patients: A Nationwide Retrospective Cohort Study Using the National Database of Health Insurance Claims of Japan Plast. Reconstr. Surg. 2021 147 1067e 1068e 10.1097/PRS.0000000000007935 184. Dolynchuk K.N. Tredget E.E. A Preliminary Report of the Biochemical and Clinical Effects of 1,4-Diaminobutane on Prevention of Human Hypertrophic Scars Plast. Reconstr. Surg. 2020 145 76e 84e 10.1097/PRS.0000000000006413 31881614 185. Hoerst K. van den Broek L. Sachse C. Klein O. von Fritschen U. Gibbs S. Hedtrich S. Regenerative potential of adipocytes in hypertrophic scars is mediated by myofibroblast reprogramming J. Mol. Med. 2019 97 761 775 10.1007/s00109-019-01772-2 30891616 186. Li H. Ziemer M. Stojanovic I. Saksida T. Maksimovic-Ivanic D. Mijatovic S. Djmura G. Gajic D. Koprivica I. Krajnovic T. Mesenchymal Stem Cells From Mouse Hair Follicles Reduce Hypertrophic Scarring in a Murine Wound Healing Model Stem Cell Rev. Rep. 2022 18 2028 2044 10.1007/s12015-021-10288-7 35080748 187. Joo S.Y. Cho Y.S. Lee S.Y. Seo C.H. Regenerative effect of combined laser and human stem cell-conditioned medium therapy on hypertrophic burn scar Burns 2022 in press 10.1016/j.burns.2022.06.007 188. Dai Z. Lou X. Shen T. Sun Y. Xiao Y. Zheng X. Wang X. Peng Y. Guo Y. Guo Y. Combination of ablative fractional carbon dioxide laser and platelet-rich plasma treatment to improve hypertrophic scars: A retrospective clinical observational study Burn. Trauma 2021 9 tkab016 10.1093/burnst/tkab016 34337088 189. Ebrahimi Z. Alimohamadi Y. Janani M. Hejazi P. Kamali M. Goodarzi A. Platelet-rich plasma in the treatment of scars, to suggest or not to suggest? A systematic review and meta-analysis J. Tissue Eng. Regen. Med. 2022 16 875 899 10.1002/term.3338 35795892 190. Kurachi I. Kurita E. Takushima A. Suga H. Human CD206+ Macrophages Show Antifibrotic Effects on Human Fibroblasts through an IL-6-Dependent Mechanism In Vitro Plast Reconstr Surg 2021 147 231e 239e 10.1097/PRS.0000000000007563 33370071 191. Kanno E. Tanno H. Masaki A. Sasaki A. Sato N. Goto M. Shisai M. Yamaguchi K. Takagi N. Shoji M. Defect of interferon g leads to impaired wound healing through prolonged neutrophilic inflammatory response and enhanced MMP-2 activation Int. J. Mol. Sci. 2019 20 5657 10.3390/ijms20225657 31726690 |
PMC10000649 | This study aimed to add real-world evidence to the literature regarding the effectiveness and safety of durvalumab consolidation (DC) after concurrent chemoradiotherapy (CCRT) in the treatment of unresectable stage III non-small cell lung cancer (NSCLC). Using a hospital-based NSCLC patient registry and propensity score matching in a 2:1 ratio, we conducted a retrospective cohort study of patients with unresectable stage III NSCLC who completed CCRT with and without DC. The co-primary endpoints were 2-year progression-free survival and overall survival. For the safety evaluation, we evaluated the risk of any adverse events requiring systemic antibiotics or steroids. Of 386 eligible patients, 222 patients--including 74 in the DC group--were included in the analysis after propensity score matching. Compared with CCRT alone, CCRT with DC was associated with increased progression-free survival (median: 13.3 vs. 7.6 months, hazard ratio[HR]: 0.63, 95% confidence interval[CI]: 0.42-0.96) and overall survival (HR: 0.47, 95% CI: 0.27-0.82) without an increased risk of adverse events requiring systemic antibiotics or steroids. While there were differences in patient characteristics between the present real-world study and the pivotal randomized controlled trial, we demonstrated significant survival benefits and tolerable safety with DC after the completion of CCRT. real-world study concurrent chemoradiotherapy durvalumab non-small cell lung cancer Korean government (MSIT)NRF-2019M3E5D1A02067951 Chonnam National University Hwasun Hospital Institute for Biomedical ScienceHCRI 21018 Chonnam National University Hospital Biomedical Research InstituteBCRI202109-85 This research was supported by the Bio & Medical Technology Development Program of the National Research Foundation (NRF), funded by the Korean government (MSIT) (NRF-2019M3E5D1A02067951 to Cheol-Kyu Park); a grant from the Chonnam National University Hwasun Hospital Institute for Biomedical Science (HCRI 21018 to Cheol-Kyu Park); and a grant from the Chonnam National University Hospital Biomedical Research Institute (BCRI202109-85 to Nakyung Jeon). pmc1. Introduction Stage III non-small cell lung cancer (NSCLC) accounts for 18% of all clinical stages at the initial diagnosis of NSCLC . The presentation of stage III NSCLC is heterogeneous, and the tumor extent at diagnosis varies from resectable to unresectable, depending on the size and spread of the tumor . Patients with unresectable stage III NSCLC with good performance statuses usually receive concurrent chemoradiotherapy (CCRT) as a standard of care among possible strategies, such as single or combination surgery, radiation therapy (RT), or chemotherapy . The median overall survival (OS) of these patients was estimated as 12 to 24 months despite the standard therapy . Durvalumab is an anti-programmed cell death-ligand 1 (PD-L1) antibody that inhibits the interaction of PD-L1 and PD-1 in the tumor tissue . The PACIFIC trial, a randomized, placebo-controlled, phase 3 trial, demonstrated the survival benefits of durvalumab consolidation (DC) therapy in patients with unresectable stage III NSCLC without progression following CCRT . DC after CCRT improved 5-year OS dramatically compared with placebo (median: 47.5 vs. 29.1 months, hazard ratio [HR]: 0.72, 95% confidence interval [CI]: 0.59-0.89) in the PACIFIC trial. In 2018, the US Food and Drug Administration approved durvalumab for patients with unresectable stage III NSCLC without disease progression following platinum-based CCRT. Practice guidelines recommend DC for up to 12 months . In general, traditional clinical trials, including the PACIFIC trial, have strict participant eligibility criteria and are performed under near-ideal experimental conditions wherein patients are highly controlled, compliant, and adherent . In fact, the median 5-year OS of the control group included in the PACIFIC trial was 29.1 months, meaning that the control group was representative of stage III NSCLC patients with relatively favorable OS. Durvalumab has been reimbursed by the Korean National Health Insurance System (the Health Insurance Review and Assessment, HIRA) since 1 April 2020. To be eligible for reimbursement, patients are required to have PD-L1 expression > 1% and to be within 42 days of CCRT completion, in accordance with the eligibility criteria for the PACIFIC trial. However, some NSCLC patients choose DC as an alternative treatment, even if they are not eligible for reimbursement under the Korean National Health Insurance System. To advance the existing real-world evidence, we investigated the effectiveness and safety of DC with 2-year follow-up data of patients with characteristics deviating from those required by the PACIFIC trial. 2. Materials and Methods 2.1. Data Source We conducted a retrospective cohort study using lung cancer patient registry data collected by lung cancer specialists at Chonnam National University Hwasun Hospital since 2011. The registry contains personal details (e.g., patient identifiers, date of birth, sex, and type of health insurance), medical history related to lung cancer (e.g., family and personal history of cancer, smoking status, comorbidities), lung cancer characteristics (e.g., stage at registry enrollment, performance status), pulmonary comorbidities, pulmonary function, epidermal growth factor receptor (EGFR) mutation, anaplastic lymphoma kinase (ALK) rearrangement, and PD-L1 tumor proportional score (TPS). Additionally, longitudinal treatment information is available in the registry through institutional electronic health records, allowing this study to capture the type, dose, and date of administration of treatment at the patient level. At the time of analysis after data collection, all patient information was anonymized. 2.2. Study Cohort Patients with unresectable locally advanced stage III NSCLC who completed CCRT between December 2014 and December 2020 were included in the analysis. Clinical stages were defined by the Union for International Cancer Control (UICC) TNM classification: the seventh edition was applied to patients included until December 2015, and the eighth edition was applied to patients included since January 2016. At the time of the analysis, all the stages of enrolled patients were defined by the eighth edition. The clinical stages of patients whose index date was before January 2016 were changed according to the eighth edition. We excluded patients who had documented disease progression during CCRT. CCRT completion was defined as receiving at least two cycles of platinum-based chemotherapy concurrently with RT at a total dose ranging from 54-66 Gy. The concurrent chemotherapeutic regimens were weekly paclitaxel (45 mg/m2) plus either cisplatin (20 mg/m2) or carboplatin (AUC 2). Given the dose intensity of the weekly regimen compared with the 3-week interval regimen, patients who received four or more cycles of weekly paclitaxel plus cisplatin or carboplatin were included. Follow-up chest computed tomography was performed 4-8 weeks after CCRT completion and repeated every 8-12 weeks thereafter. Clinical responses to treatment were defined according to the Response Evaluation Criteria in Solid Tumors (RECIST), version 1.1 . For patients eligible for DC after CCRT, durvalumab (10 mg/kg) was administered via intravenous infusion every 2 weeks and continued for up to 12 months until the occurrence of confirmed progression, death, initiation of alternative cancer therapy, an intolerable adverse event, or other reasons resulting in discontinuation of durvalumab. We compared the DC group with a group of patients who did not receive any adjuvant treatment after CCRT completion (CCRT alone) as historical controls. The study design is illustrated in Figure 1. The index date for the DC group was the date of durvalumab initiation after or on the last date of CCRT. For the CCRT-alone group, via propensity score (PS) matching, we risk-set sampled two historical controls for each patient who received DC. To ascertain the index date of the historical controls in the CCRT-alone group, we estimated the gap between the initiation of durvalumab and the completion of CCRT by subtracting the last date of CCRT from the first date of durvalumab administration for each DC patient and added the gap to the last date of CCRT of the matched historical controls. The matched pairs were excluded altogether if any patients in either the DC or CCRT-alone group had any of the following conditions between registry enrollment and cohort entry: initiation of durvalumab more than 3 months after CCRT completion, surgery for NSCLC during or after CCRT, or confirmed NSCLC progression. 2.3. Outcome Measures The co-primary outcomes were progression-free survival (PFS) and OS at 1 and 2 years, defined as the durations (measured in months) from cohort entry to disease progression and to death from any cause, respectively. Patients were followed for up to 2 years. The patients who were lost to follow-up or did not progress to a study outcome through the study period were censored at the end of the study period, which was on 27 January 2022. Both disease progression and death were ascertained primarily via manual investigation of electronic health records. If necessary, we obtained information through a national death registration system to confirm survival statuses and dates of death. The secondary outcome was treatment-related adverse events associated with durvalumab, defined as any event requiring systemic antibiotics or steroids within a year after the index date. Given that the assessed safety outcome was a composite outcome of two domains (antibiotics and steroids), we further examined the safety outcome by type of drug administered (antibiotics or steroids), route of administration (intravenous or oral only), and duration of antibiotics (>=5 days or >=10 days) and steroids (>=14 days or >=28 days). 2.4. Statistical Analysis The following patient characteristics were assessed during and after CCRT (before the index date): age, sex, smoking history (never, current, previous), body mass index (<18.5, 18.5-25, >=25, or unknown), Eastern Cooperative Oncology Group (ECOG) performance status (0, 1, 2, or unknown), history of chronic obstructive pulmonary disease, history of interstitial lung disease, tumor histologic type (non-squamous or squamous), NSCLC disease stage (IIIA, IIIB, or IIIC), EGFR mutation (wild type, mutant, or unknown), ALK rearrangement (negative, positive, or unknown), PD-L1 expression (>=1%, <1%, or unknown), chemotherapy regimen (cisplatin or carboplatin), chemotherapy cycles completed (three, four, five, or six), RT fraction during CCRT, RT dose during CCRT, history of radiation pneumonitis. We also assessed anemia (hemoglobin level < 12 g/dL), thrombocytopenia (platelet count < 130 x 103 mL), decreased liver function (aspartate aminotransferase level > 38 U/L or alanine aminotransferase level > 42 U/L), and kidney function (estimated glomerular filtration rate < 60 or 60-90 mL/min/1.73 m2) at the hospital visit for the last CCRT. All the characteristics were used to estimate the PSs for the DC and CCRT-alone groups, except for that reflecting PD-L1 expression. A sensitivity analysis that included PD-L1 expression status in the PS estimation was conducted. We used PS matching to account for potential differences in baseline risk between the DC and CCRT-alone groups and compared patient characteristics between the groups before and after PS matching to demonstrate the success of baseline risk balancing via t-test or chi-square analysis. The Cox proportional hazards analysis included the DC indicator variable (i.e., DC group or CCRT-alone group) and 21 variables used in a multivariable logistic regression analysis to estimate PSs. Kaplan-Meier curves were plotted according to the primary outcomes. HRs comparing the incidences of progression and death in the DC group vs. the CCRT-alone group were estimated by fitting Cox proportional hazards regression models. We analyzed 37 subgroups based on patient characteristics, including age, sex, and other clinical factors related to lung cancer prognosis. For the safety analysis, multivariable logistic regression models were used to estimate the association between DC and the risk of any treatment-related adverse events requiring systemic antibiotic or steroid use. All analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC, USA). 3. Results 3.1. Patients' Characteristics Of 620 patients in the registry, 386 were included in the PS matching after applying the inclusion and exclusion criteria. A total of 222 patients (CCRT alone: 148, DC: 74) were finally selected after the PS matching . The number of days between CCRT completion and DC initiation ranged from 0-66 days (median: 28 days), while 11 patients (14.9%) in the DC group exceeded 42 days. The calendar year distributions of index dates by comparison groups are shown in Figure 3. Among 74 patients in the DC group, 36 patients (48.6%)--including seven durvalumab users before its regulatory approval (December 2018)--initiated DC before durvalumab became eligible for reimbursement by the Korean National Health Insurance System on 1 April 2020. All the remaining patients presumably received reimbursements for DC because they had the records of durvalumab initiation within 42 days after CCRT completion and positive PD-L1 expression, except for one patient who did not have a documented PD-L1 test result. Baseline characteristics were compared before and after PS matching (Table 1). Before matching, compared with the CCRT-alone group, more patients in the DC group had chronic obstructive pulmonary disease, had stage IIIA or IIIC, had positive results for PD-L1 expression (PD-L1 TPS >= 1%), received carboplatin rather than cisplatin in their paclitaxel-based chemotherapy regimens, and developed radiation pneumonitis. The DC group also had better kidney function, as indicated by estimated glomerular filtration rates. Patients in the DC group received more chemotherapy cycles and RT at a higher fraction than patients in the CCRT-alone group. The intergroup differences became statistically non-significant after PS matching, except for the difference in PD-L1 expression status. This was expected because PD-L1 expression status was purposely not included in the logistic regression model and PS estimation. In a sensitivity analysis where PD-L1 expression status was included in the PS estimation, 180 patients (CCRT-alone group: 120 patients; DC group: 60 patients) remained in the analysis. After PS matching, there were no differences in baseline characteristics between the groups (Table 1). The overall median follow-up was 18.4 months, that is, 561 days (range: 0 to 730 days in the CCRT-alone group and 8 to 730 days in the DC group). None of the patients were receiving durvalumab at the data cutoff point. 3.2. Survival Outcomes and Post-Progression Treatment As of the data cutoff on 27 January 2022, among all included patients, 162 patients had experienced disease progression (73.0%, 119 in the CCRT-alone group and 43 in the DC group), and 127 patients had died. The types and sites of progression and post-progression treatment of the 162 patients are described in Supplementary Table S1. When we limited the maximum length of follow-up to 2 years (corresponding with the primary outcome definitions), 145 patients experienced disease progression (65.32%, 106 in the CCRT-alone group and 39 in the DC group), and 95 patients died (42.79%, 76 in the CCRT-alone group and 19 in the DC group). The median PFS was 8.8 months (95% CI: 7.8-10.0). The DC group had a significantly longer median PFS (13.3 vs. 7.6 months, HR: 0.63, 95% CI: 0.42-0.96) and a higher 2-year PFS rate (47.3% vs. 28.4%) than the CCRT-alone group . Among all covariates included in the multivariable Cox regression model, stage IIIB (HR: 1.64, 95% CI: 1.01-2.65) and stage IIIC (HR: 1.98, 95% CI: 1.01-3.89) were significantly associated with PFS (Supplementary Table S2). The median OS duration and 2-year OS rate were 24.3 months and 48.6%, respectively, in the CCRT-alone group. In the DC group, the median OS was not reached, and the corresponding 2-year OS rate was 74.4%. The findings demonstrated that DC provided OS benefits for NSCLC patients after CCRT completion (adjusted HR; 0.47, 95% CI: 0.27-0.82) . Poor ECOG performance status scores (scores > 1, HR: 3.73, 95% CI: 1.23-11.32) and stage IIIB (HR: 2.15, 95% CI: 1.17-4.00) were associated with an increased risk of death (Supplementary Table S2). A sensitivity analysis was performed wherein PD-L1 expression status was well balanced between the comparison groups. While the point estimates of HRs for PFS and OS were similar to the results of the main analyses (Table 2), the sample size was underpowered for significance due to the loss of patients during PS matching . 3.3. Safety Outcomes There were 114 patients (51.4%) who developed treatment-related adverse events, defined by the necessitated use of systemic antibiotics or steroids during follow-up (Table 3). Overall, DC did not increase the risk of treatment-related adverse events requiring antibiotics or steroids. Instead, DC therapy was associated with a decreased risk of any events necessitating systemic antibiotic or steroid administration (adjusted HR: 0.472, 95% CI: 0.242-0.921) or the administration of antibiotics alone (adjusted HR: 0.436, 95% CI: 0.220-0.865). 3.4. Subgroup Analysis Overall, the survival benefits of DC were consistently observed across subgroups, especially when a subgroup was well-powered. Several subgroup analyses yielded statistically significant intergroup differences in PFS but not in OS, including body mass index < 18.5, pneumonitis not requiring steroid treatment, and unknown EGFR mutation or ALK rearrangement . 4. Discussion In this real-world data analysis, we evaluated PFS and OS outcomes among patients who received DC for stage III NSCLC after CCRT. Consistent with the pivotal phase 3 trial (the PACIFIC trial) , DC after CCRT was well tolerated and effective in real-world patients with unresectable stage III NSCLC. Given that real-world data analyses generally have less-restrictive eligibility criteria than clinical trials, the present study can help improve our understanding of the patient populations that could benefit from receiving DC after CCRT. Included patients in the present study were all Asians (Koreans) with longer durations from CCRT to DC, which differed from the PACIFIC study design . The median age of the patients at the start of the follow-up was 66 years old, which is younger than the patients' ages attributed to several real-world studies (67-72 years old) but older than those of the seminal randomized controlled trial (RCT; 64 years old) . This study also included patients with ECOG performance status scores of 2, whereas the PACIFIC trial was limited to patients with ECOG performance status scores < 2. Nevertheless, favorable PFS and OS results were observed regardless of these discrepancies. A meta-analysis of 16 real-world studies (RWSs) designed to evaluate the effectiveness and safety of DC also revealed great differences in patient characteristics and treatment strategies between RWSs and the PACIFIC trial . Despite such differences, the meta-analysis demonstrated the safety and effectiveness of durvalumab in different clinical settings. The present study confirms the survival benefits associated with DC in an NSCLC-representative population in South Korea. The median duration of follow-up was 18.4 months, with a maximum follow-up time of 24 months. Follow-up was limited to 24 months to balance concurrent and historical data in terms of the length of the follow-up . In the real world, patients whose data are being analyzed concurrently with their follow-up can only be followed for a limited time, even though they might survive longer. Allowing longer follow-ups for only one group can yield biased results . However, it is unknown whether it truly leads to biased estimates of survival benefits because our study was not designed to evaluate the existence of depletion of susceptibility. An analysis of PACIFIC trial data that summarized interim findings (with approximately 2 years of follow-up data) yielded HR estimates of 0.51 for PFS (95% CI, 0.41-0.63) and of 0.68 for OS (95% CI 0.47-0.997) . The present study yielded HR estimates for PFS and OS within the CIs from the PACIFIC trial. Of note, "estimate agreement"--defined by real-world HR estimates that fall within the 95% CIs of the corresponding RCT estimates--is a metric used to assess agreement between RCT and RWS findings. One unique aspect of this study was that the median PFS of the overall population and treatment groups were relatively short compared to PFS durations reported previously. This could be explained by including patients with ECOG performance status scores of 2 or a higher percentage of stage IIIC patients in our study. The median PFS was 8.8 months, with an intergroup difference in median PFS of 5.7 months (DC vs. CCRT alone: 13.3 vs. 7.6 months). The intergroup difference in median PFS was 11 months in the PACIFIC trial, with the group-specific median PFS durations of 17.2 and 5.6 months for DC and CCRT alone , respectively. A recent Chinese RWS determined an intergroup difference in median PFS of 8.6 months (17.5 and 8.9 months for DC and CCRT alone, respectively), which was slightly longer than what we found . Nevertheless, the 2-year OS of patients with stage III NSCLC treated with DC was comparable between this RWS and the PACIFIC trial. This implies that stage III NSCLC patients in regular practice have a relatively poor prognosis with CCRT alone and that prognosis can improve (in terms of survival) with the use of DC after CCRT. Subgroup analysis may help identify the specifications of patients who would have been excluded from the PACIFIC trial but could benefit from DC. However, this study was not designed to identify such subgroup populations, and few trials are powered to detect treatment effects in subgroups. In the sensitivity analysis, wherein we created matched comparison groups by baseline characteristics, including PD-L1 expression (<1%, >=1%, or unknown), the estimated HRs for PFS and OS aligned with the results obtained from the main analysis. However, the sample size for the sensitivity analysis was underpowered for significance . The subgroup analysis of the PACIFIC trial showed the possibility of PD-L1 TPS as a biomarker for DC, and evidence of survival benefits according to PD-L1 expression in the target population has been of great interest in literature . Similarly, the PACIFIC trial and several RWSs have shown that the presence of oncogenic driver mutations, including mutations of EGFR and ALK, tend to negatively affect the prognosis of patients with DC . Even considering the lack of power in previous studies and the effect of CCRT on the tumor microenvironment , the application of DC is expected to be limited in oncogene-addicted NSCLC due to the unfavorable efficacy of single-agent immune checkpoint inhibitors and the harmful toxicity of post-progression sequential treatment with EGFR tyrosine kinase inhibitors . In the present study, there was no variation in the survival benefit of DC according to PD-L1 expression and driver mutations. Notably, the molecular testing results were unknown for most patients (PD-L1: 33.7%, EGFR: 62.0%, ALK: 65.5%). In addition to PD-L1 and driver oncogenes, circulating tumor DNA (ctDNA) has been proposed as another candidate biomarker for DC, which could predict recurrence (minimal/molecular residual disease) following curative-intent treatment and the response of immune checkpoint inhibitor consolidation . Therefore, large-scale prospective trials with patient selection based on biomarker analysis are warranted to consolidate the evidence regarding the role of additional therapy after definitive treatment for locally advanced NSCLC. 5. Conclusions The findings of the present study suggest that DC's efficacy (demonstrated in its pivotal phase 3 trial) is evident in real-world clinical practice, making DC feasible as the global standard of care for patients with unresectable stage III NSCLC. Continuing to generate real-world evidence is necessary with longer follow-up of more patients, with the potential for expanding the indications of DC or making regulatory decisions toward meaningful use of this effective and innovative therapy. Supplementary Materials The following supporting information can be downloaded at: Figure S1: Sensitivity analysis: (a) PFS and (b) OS. PFS: progression-free survival, OS: overall survival, CCRT: concurrent chemoradiotherapy, CI: confidence interval, NR: not reached, HR: hazard ratio; Figure S2: Subgroup analysis: (a) PFS and (b) OS. PFS: progression-free survival, OS: overall survival, BMI: body mass index, ECOG: Eastern Cooperative Oncology Group, COPD: chronic obstructive pulmonary disease, PD-L1: programmed death-ligand 1, TPS: tumor proportional score, EGFR: epidermal growth factor receptor, ALK: anaplastic lymphoma kinase; Table S1: Post-treatment progression pattern and subsequent treatment in patients with CCRT alone and durvalumab consolidation (DC). Values are presented as numbers (%). CCRT: concurrent chemoradiotherapy, PD: progressive disease, OP: operation, RT: radiotherapy, CRT: chemoradiotherapy, TKI: tyrosine kinase inhibitor, ICI: immune checkpoint inhibitor, BSC: best supportive care; Table S2: Multivariable Cox regression analysis for PFS and OS. PFS: progression-free survival, OS: overall survival, HR: hazard ratio, CI: confidence interval, DC, durvalumab consolidation, BMI: body mass index, eGFR: estimated glomerular filtration rate, ECOG: Eastern Cooperative Oncology Group, PS: performance status, COPD: chronic obstructive pulmonary disease, RT, radiotherapy. Click here for additional data file. Author Contributions Conceptualization, C.-K.P., N.J. and I.-J.O.; methodology, C.-K.P., N.J. and I.-J.O.; software, C.-K.P., N.J. and I.-J.O.; validation, C.-K.P., N.J., H.-K.P., H.-J.O., Y.-C.K., H.-L.J., Y.-H.K., S.-J.A. and I.-J.O.; formal analysis, C.-K.P., N.J., H.-L.J. and I.-J.O.; investigation, C.-K.P., H.-K.P., H.-J.O., N.J. and I.-J.O.; resources, C.-K.P., N.J., H.-K.P., H.-J.O., Y.-C.K., Y.-H.K., S.-J.A. and I.-J.O.; data curation, C.-K.P., N.J., H.-L.J. and I.-J.O.; writing--original draft preparation, C.-K.P., N.J. and I.-J.O.; writing--review and editing, C.-K.P., N.J., H.-K.P., H.-J.O., Y.-C.K., H.-L.J., Y.-H.K., S.-J.A. and I.-J.O.; visualization, C.-K.P., N.J. and I.-J.O.; supervision, C.-K.P., N.J., Y.-C.K., H.-L.J., Y.-H.K., S.-J.A. and I.-J.O.; project administration, C.-K.P. and I.-J.O.; funding acquisition, C.-K.P. and N.J. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and Good Clinical Practice guidelines. This study was approved by the Institutional Review Board of Chonnam National University Hwasun Hospital (IRB No. CNUHH-2022-025). Informed Consent Statement Patient consent was waived because of the retrospective nature of the study and because the analysis used anonymous clinical data. Data Availability Statement The data presented in this study are available on request from the corresponding author. The data are not publicly available due to institutional data-sharing restrictions. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Study design. CCRT: concurrent chemoradiotherapy; wCCRT: weekly CCRT regimen. Figure 2 Patient selection flow chart. NSCLC: non-small cell lung cancer; CCRT: concurrent chemoradiotherapy; wCCRT: RT: radiation therapy; PS: propensity score. Figure 3 Calendar year distribution of index dates. CCRT: concurrent chemoradiotherapy; DC: durvalumab consolidation. The red line indicates 1 April 2020, when durvalumab began to be reimbursed by Korean National Health Insurance System. Figure 4 Survival analysis results: (a) PFS and (b) OS. CCRT: concurrent chemoradiotherapy; PFS: progression-free survival; CI: confidence interval; NR: not reached; HR: hazard ratio; OS: overall survival. cancers-15-01606-t001_Table 1 Table 1 Baseline characteristics of patients with concurrent chemoradiotherapy alone and durvalumab consolidation before and after propensity score matching. Before After Primary Analysis Sensitivity Analysis Characteristics CCRT Alone (n = 294) DC (n = 91) p-Value CCRT Alone (n = 148) DC (n = 74) p-Value CCRT Alone (n = 120) DC (n = 60) p-Value Age, years Mean (SD) 67.4 (8.5) 66.3 (8.2) 0.247 66.6 (8.4) 65.9 (8.5) 0.589 66.2 (8.0) 66.0 (7.8) 0.912 Median (range) 67.1 (34-84) 66.1 (45-81) 0.949 66.1 (34-81) 66.1 (45-78) 0.834 66.0 (40-81) 66.1 (49-78) 0.926 Sex Female, n (%) 20 (6.8%) 7 (7.7%) 0.772 11 (7.4%) 5 (6.8%) 0.854 9 (7.5%) 4 (6.7%) 0.839 Male, n (%) 274 (93.2%) 84 (92.3%) 137 (92.6%) 69 (93.2%) 111 (92.5%) 56 (93.3%) Smoking Never smoker, n (%) 40 (13.6%) 11 (12.1%) 0.697 22 (14.9) 11 (14.8) 0.916 18 (15.0%) 8 (13.3%) 0.725 Current smoker, n (%) 99 (33.7%) 35 (38.5%) 50 (33.8) 27 (36.5) 35 (29.2%) 21 (35%) Ex-smoker, n (%) 155 (52.7%) 45 (49.5%) 76 (51.4) 36 (48.7) 67 (55.8) 31 (51.7%) BMI, kg/m2 Mean (SD) 24.1 (9.0) 22.9 (3.1) 0.216 23.3 (3.4) 22.8 (3.1) 0.327 23.4 (3.4) 23.1 (2.9) 0.587 BMI < 18.5, n (%) 12 (4.1%) 8 (8.8%) 0.346 8 (5.4%) 5 (6.8%) 0.880 6 (5.0%) 4 (6.7%) 0.883 18.5 <= BMI < 25, n (%) 182 (61.9%) 56 (61.5%) 90 (60.8%) 47 (63.5%) 74 (61.7%) 35 (58.3%) 25 <= BMI, n (%) 86 (29.3%) 22 (24.2%) 41 (27.7%) 17 (23.0%) 33 (27.5%) 16 (26.7%) Unknown BMI, n (%) 14 (4.8%) 5 (5.5%) 9 (6.1%) 5 (6.8%) 7 (5.8%) 5 (8.3%) ECOG performance status 0, n (%) 122 (41.5 %) 39 (42.9%) 0.760 68 (46.0%) 35 (47.3%) 0.742 58 (48.3%) 25 (41.7%) 0.855 1, n (%) 152 (51.7%) 44 (48.4%) 66 (44.6%) 31 (41.9%) 52 (43.3%) 30 (50.0%) 2, n (%) 7 (2.4%) 4 (4.4%) 4 (2.7%) 4 (5.4%) 4 (3.3%) 2 (3.3%) Unknown, n (%) 13 (4.4%) 4 (4.4%) 10 (6.8%) 4 (5.4%) 6 (5.0%) 3 (5.0%) Comorbidity COPD, n (%) 166 (56.4%) 67 (73.6%) 0.003 100 (67.6%) 53 (71.6%) 0.538 83 (69.2%) 41 (68.3%) 0.909 ILD, n (%) 13 (4.4%) 8 (8.8%) 0.109 6 (4.0%) 4 (5.4%) 0.647 5 (4.2%) 3 (5.0%) 0.798 Histologic type Non-squamous, n (%) 105 (35.7%) 30 (33.0%) 0.631 57 (38.5%) 25 (33.8%) 0.491 49 (40.8%) 22 (36.7%) 0.590 Squamous, n (%) 189 (64.3%) 61 (67.0%) 91 (61.5%) 49 (66.2%) 71 (59.2%) 38 (63.3%) Stage (TNM 8th) IIIA, n (%) 146 (49.7%) 47 (51.6%) 0.001 75 (50.7%) 37 (50.0%) 0.267 61 (50.8%) 31 (51.7%) 0.969 IIIB, n (%) 136 (45.3%) 31 (34.1%) 61 (41.2%) 26 (35.1%) 48 (40.0%) 23 (38.3%) IIIC, n (%) 12 (4.1%) 13 (14.3%) 12 (8.1%) 11 (14.9%) 11 (9.2%) 6 (10.0%) EGFR mutation Wild type, n (%) 101 (34.4%) 25 (27.5%) 0.415 45 (30.4%) 22 (29.7%) 0.640 39 (32.5%) 20 (33.3%) 0.551 Mutant, n (%) 12 (4.1%) 3 (3.3%) 8 (5.4%) 2 (2.7%) 6 (5.0%) 1 (1.7%) Unknown, n (%) 181 (61.6%) 63 (70.2%) 95 (64.2%) 50 (67.6%) 75 (62.5%) 39 (65.0%) ALK rearrangement Negative, n (%) 82 (27.9%) 25 (27.5%) 0.992 40 (27.0%) 22 (29.7%) 0.891 39 (32.5) 19 (31.7) 0.996 Positive, n (%) 9 (3.1%) 3 (3.3%) 5 (3.4%) 2 (2.7%) 4 (3.3) 2 (3.3) Unknown, n (%) 203 (69.1%) 63 (69.2%) 103 (69.6%) 50 (67.6%) 77 (64.2) 39 (65.0) PD-L1 immunohistochemistry (SP263) TPS <1%, n (%) 82 (27.9%) 14 (14.9%) <0.0001 41 (27.7%) 10 (13.5%) <0.0001 27 (22.5%) 10 (16.7%) 0.284 TPS >=1%, n (%) 112 (29.0%) 65 (69.2%) 45 (30.4%) 51 (68.9%) 59 (49.2%) 37 (61.7%) Unknown, n (%) 175 (45.3%) 15 (16.0%) 62 (41.9%) 13 (17.6%) 34 (28.3%) 13 (21.7%) Chemotherapy regimen Pac-Cis, n (%) 275 (93.5%) 71 (78.0%) <0.0001 131 (88.5) 65 (87.8) 0.883 109 (90.8%) 52 (13.3%) 0.3911 Pac-Car, n (%) 19 (6.5%) 20 (22.0%) 17 (11.5%) 9 (12.2) 11 (9.2%) 8 (87.0%) Chemotherapy cycle Mean (SD) 5.6 (0.63) 5.9 (0.41) <0.0001 5.8 (0.48) 5.9 (0.43) 0.681 5.9 (0.42) 5.8 (0.46) 0.714 3, n (%) 1 (0.34%) 0 (0.0%) 0.001 0 (0.0%) 0 (0.0%) 0.878 0 (0.0%) 0 (0.0%) 0.931 4, n (%) 20 (6.8%) 2 (2.2%) 6 (4.1%) 2 (2.7%) 3 (2.5%) 2 (3.3%) 5, n (%) 80 (27.2%) 9 (9.9%) 14 (9.5%) 7 (9.5%) 11 (9.2%) 6 (10.0%) 6, n (%) 193 (65.6%) 80 (87.9%) 128 (86.5%) 65 (87.8%) 106 (88.3%) 52 (86.7%) RT fraction, mean (SD) 28.7 (2.18) 29.7 (1.10) <0.0001 29.7 (1.18) 29.7 (1.09) 0.805 29.7 (1.15) 29.7 (1.20) 0.787 RT dose in Gy, mean (SD) 61.3 (2.85) 61.6 (2.63) 0.457 61.8 (2.90) 61.8 (2.78) 0.923 62.0 (2.97) 62.0 (2.82) 1.000 Radiation pneumonitis RP without treatment, n (%) 168 (57.1%) 65 (71.4%) 0.033 95 (64.2%) 49 (66.2%) 0.823 84 (70.0%) 41 (68.3%) 0.479 RP with treatment, n (%) 49 (16.7%) 13 (14.3%) 24 (16.2%) 13 (17.6%) 13 (10.8%) 10 (16.7%) No RP, n (%) 77 (26.2%) 13 (14.3%) 29 (19.6%) 12 (16.2%) 23 (19.2%) 9 (15.0%) Anemia No, n (%) 156 (52.0%) 39 (42.9%) 0.126 66 (44.6%) 32 (43.2%) 0.848 56 (46.7%) 29 (48.3%) 0.833 Yes, n (%) 141 (48.0%) 52 (57.1%) 82 (55.4%) 42 (56.8%) 64 (53.3%) 31 (52.7%) Thrombocytopenia No, n (%) 270 (91.8%) 87 (95.6%) 0.227 141 (95.3%) 70 (94.6%) 0.827 113 (94.2%) 58 (96.7%) 0.468 Yes, n (%) 24 (8.2%) 4 (4.4%) 7 (4.7%) 4 (5.4%) 7 (5.8%) 2 (3.3%) Liver Failure No, n (%) 264 (89.8%) 85 (93.4%) 0.301 136 (91.9%) 68 (91.9%) 1.000 109 (90.8%) 54 (90.0%) 0.857 Yes, n (%) 30 (10.2%) 6 (6.6%) 12 (8.1%) 6 (8.1%) 11 (9.2%) 6 (10.0%) Kidney function by eGFR eGFR >=90, n (%) 111 (37.7%) 49 (53.9%) 0.015 63 (42.6%) 38 (51.4%) 0.460 54 (45.0%) 29 (48.3%) 0.684 60 <= eGFR < 90, n (%) 148 (50.3%) 37 (40.7%) 74 (50.0%) 31 (41.9%) 59 (49.2%) 26 (43.3%) 0 <= eGFR < 60, n (%) 35 (11.9%) 5 (6.5%) 11 (7.4%) 5 (6.8%) 7 (5.83%) 5 (8.3%) SD: standard deviation; BMI; body mass index; CCRT: concurrent chemoradiotherapy; ECOG: Eastern Cooperative Oncology Group; COPD: chronic obstructive pulmonary disease; ILD: interstitial lung disease; EGFR: epidermal growth factor receptor; ALK: anaplastic lymphoma kinase; PD-L1: programmed cell death ligand-1; TPS: tumor proportional score; Pac-Cis: paclitaxel plus cisplatin; Pac-Car: paclitaxel plus carboplatin; RT: radiotherapy; RP: radiation pneumonitis; eGFR: estimated glomerular filtration rate. cancers-15-01606-t002_Table 2 Table 2 Survival analysis results. Main Analysis (N = 222) Sensitivity Analysis (N = 180) No. of survival at 2 years/total patients in a group (%) PFS DC: 35/74 (47.3%) DC: 28/60 (46.7%) CCRT: 42/148 (28.4%) CCRT: 38/120 (31.7%) OS DC: 55/74 (74.4%) DC: 42/60 (70.0%) CCRT: 72/148 (48.6%) CCRT: 69/120 (57.5%) Median survival time in months (95% CI) PFS DC: 13.5 (6.9-NR) DC: 12.7 (8.7-NR) CCRT: 7.8 (6.7-9.5) CCRT: 8.9 (6.9-13.6) OS DC: NR (23.5-NR) DC: NR (23.4-NR) CCRT: 24.3 (17.5-NR) CCRT: NR (22.7-NR) Hazard Ratio (95% CI) PFS 0.630 (0.416-0.957) 0.647 (0.405-1.013) OS 0.469 (0.270-0.815) 0.477 (0.306-1.037) PFS: progression-free survival; OS: overall survival; CI: confidence interval; DC: durvalumab consolidation; CCRT: concurrent chemoradiotherapy; NR: not reached. cancers-15-01606-t003_Table 3 Table 3 Multivariable analysis of durvalumab consolidation for risk of antibiotics or steroid use after completion of concurrent chemoradioatherapy. Outcome Type Number of Events Odds Ratio (95% Confidence Interval) CCRT (n = 148) DC (n = 74) Unadjusted Adjusted Systemic use of antibiotics or steroid 80 (54.1%) 34 (46.0%) 0.723 (0.413-1.265) 0.472 (0.242-0.921) * Systemic use of antibiotics, regardless of duration 70 (47.3%) 27 (36.5%) 0.640 (0.361-1.135) 0.436 (0.220-0.865) * >=5 days 67 (45.3%) 27 (36.5%) 0.695 (0.391-1.232) 0.511 (0.261-1.002) >=10 days 51 (34.5%) 23 (31.1%) 0.858 (0.472-1.559) 0.681 (0.346-1.340) IV antibiotics, regardless of duration 39 (26.4%) 8 (10.8%) 0.339 (0.149-0.769) 0.205 (0.080-0.525) Systemic use of steroids, regardless of duration 42 (27.7%) 23 (31.8%) 1.177 (0.640-2.166) 0.800 (0.391-1.639) >=14 days 34 (23.0%) 17 (23.0%) 1.000 (0.515-1.941) 0.738 (0.339-1.610) >=28 days 28 (18.9%) 16 (21.6%) 1.182 (0.593-2.536) 0.952 (0.423-2.141) IV steroids, regardless of duration 23 (15.5%) 8 (10.8%) 0.659 (0.279-1.554) 0.367 (0.132-1.020) * p < 0.05. CCRT: concurrent chemoradiotherapy; DC: durvalumab consolidation; IV: intravenous. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
PMC10000650 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12050990 foods-12-00990 Article Distinct Bacterial Communities in Sao Jorge Cheese with Protected Designation of Origin (PDO) Coelho Marcia C. Formal analysis Investigation Writing - original draft 1 Malcata Francisco Xavier Methodology Writing - review & editing Supervision 23 Silva Celia C. G. Methodology Formal analysis Writing - review & editing Supervision 14* Alvarenga Nuno Academic Editor Cordoba Maria G. Academic Editor Martins Antonio Pedro Louro Academic Editor Dias Joao Academic Editor 1 School of Agrarian and Environmental Sciences, University of the Azores, 9700-042 Angra do Heroismo, Portugal 2 LEPABE--Laboratory for Process Engineering, Environment, Biotechnology and Energy, Department of Chemical Engineering, Faculty of Engineering, University of Porto, 4200-465 Oporto, Portugal 3 ALiCE--Associate Laboratory in Chemical Engineering, Faculty of Engineering, University of Porto, Rua Dr. Roberto Frias, 4200-465 Oporto, Portugal 4 Institute of Agricultural and Environmental Research and Technology (IITAA), University of the Azores, 9700-042 Angra do Heroismo, Portugal * Correspondence: [email protected] 26 2 2023 3 2023 12 5 99030 1 2023 22 2 2023 24 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Sao Jorge cheese is an iconic product of the Azores, produced from raw cow's milk and natural whey starter (NWS). Although it is produced according to Protected Designation of Origin (PDO) specifications, the granting of the PDO label depends crucially on sensory evaluation by trained tasters. The aim of this work was to characterize the bacterial diversity of this cheese using next-generation sequencing (NGS) and to identify the specific microbiota that contributes most to its uniqueness as a PDO by distinguishing the bacterial communities of PDO and non-PDO cheeses. The NWS and curd microbiota was dominated by Streptococcus and Lactococcus, whereas Lactobacillus and Leuconostoc were also present in the core microbiota of the cheese along with these genera. Significant differences (p < 0.05) in bacterial community composition were found between PDO cheese and non-certified cheese; Leuconostoc was found to play the chief role in this regard. Certified cheeses were richer in Leuconostoc, Lactobacillus and Enterococcus, but had fewer Streptococcus (p < 0.05). A negative correlation was found between contaminating bacteria, e.g., Staphylococcus and Acinetobacter, and the development of PDO-associated bacteria such as Leuconostoc, Lactobacillus and Enterococcus. A reduction in contaminating bacteria was found to be crucial for the development of a bacterial community rich in Leuconostoc and Lactobacillus, thus justifying the PDO seal of quality. This study has helped to clearly distinguish between cheeses with and without PDO based on the composition of the bacterial community. The characterization of the NWS and the cheese microbiota can contribute to a better understanding of the microbial dynamics of this traditional PDO cheese and can help producers interested in maintaining the identity and quality of Sao Jorge PDO cheese. cheese microbiota lactic acid bacteria Leuconostoc fermented foods metagenomic analysis bacterial diversity high throughput sequencing Portuguese Foundation for Science and Technology (FCT)UID/CVT/0153/2016 LA/P/0045/2020 national funds through FCT/MCTES (PIDDAC)UIDB/00511/2020 FCTSFRH/BD/110227/2015 This work was financially supported by the Portuguese Foundation for Science and Technology (FCT), Project UID/CVT/0153/2016; LA/P/0045/2020 (ALiCE), and Base and Programmatic UIDB/00511/2020 (LEPABE), funded by national funds through FCT/MCTES (PIDDAC). M.C. Coelho also acknowledges financial support from FCT, grant SFRH/BD/110227/2015. pmc1. Introduction The microbiota of raw milk cheeses is quite complex and includes many non-starter lactic acid bacteria (LAB) strains originally derived from the milk itself or introduced by the manufacturing environment; these bacteria are important for the ripening of the cheese and the development of the expected flavor . Interest in the functional and structural diversity of the microbiota in raw milk cheeses has increased because these cheeses have a more intense and unique flavor compared to cheeses produced from pasteurized milk (Montel et al., 2014). Several studies have attempted to describe the microbiota of traditional cheeses and the distinct stages of cheese manufacture and ripening . Culture-dependent methods have been the preferred choice, but they are labor-intensive and inherently biased . Therefore, culture-independent techniques and next-generation sequencing (NGS) technology have played a key role in recent studies on microbial communities in traditional cheeses . In addition, NGS methods can reveal the existence of subdominant populations within the cheese microbiota that are difficult to detect using culture-dependent methods. These populations may be responsible (at least in part) for the differentiating flavors of raw milk cheeses. The interactions of the subdominant (or rare) microbiota with the dominant microbiota also likely play an important role in the development of the key flavor and aroma notes of these cheeses . Sao Jorge cheese is a very popular Portuguese cheese produced from raw cow's milk on the island of Sao Jorge in the Azores. It bears major economic and social importance on the island. This cheese exhibits a yellowish, hard or semi-hard paste and a crumbly texture. It is produced from raw cow's milk to which a natural whey starter (NWS) is added, obtained from the whey of the previous day's cheese-making. The NWS is added in a ratio of about 1.5-2/1000 (1.5/1000, for NWS acidity 50-65 degD; 2/1000, for NWS acidity 40-50 degD). Acidification takes place at 30 degC and is followed by cooking the curd at 35-36 degC, draining the whey, shaping the curd, and salting and pressing the curd. This is followed by the ripening process, which lasts at least 3-4 months. By the end of the ripening, Sao Jorge cheese has small, irregular eyes, and its flavor is characterized by strong, clean and slightly spicy notes that become more intense as it matures. Therefore, the indigenous microbiota of the raw milk and the starter culture (NWS) is important for the subsequent ripening process, as both actively contribute to the characteristic aroma and spicy flavor of the final product. However, the variability of the final product--expected in view of its being manufactured from raw milk, is often sufficient to compromise the PDO seal. In order to obtain this seal, it is not enough for the cheese to be produced by a certified cheese maker according to PDO specifications; each batch of cheese is indeed also subjected to sensory testing by a trained panel from an independent certifying body (Confraria do Queijo de Sao Jorge). As a result, a large percentage of cheeses (40-60%) produced according to PDO specifications will be eventually denied PDO status, yet they will still be suitable for selling at lower prices. Although a few studies have attempted to identify and characterize the dominant bacteria in Sao Jorge cheese , they all relied on culture-dependent methods, unable to fully decipher the diversity associated with this type of dairy product. In addition, the growth media used in culture-dependent methods are not truly selective for differentiation among bacterial communities . Therefore, complementary, more in-depth studies are needed to fully elucidate the role of the entire microbiota of this cheese. To achieve this goal, a detailed identification and characterization of the microbiota using culture-independent methods appears important for the eventual selection of tailor-made starter cultures (SLAB) and/or auxiliary cultures specifically designed to control the fermentation of this cheese, thus helping reduce variability, achieve the best sensory characteristics during ripening, and guarantee a higher percentage of PDO cheeses. The aim of this work was accordingly to apply culture-independent and NGS methods to characterize the bacterial communities in milk, NWS, curd, and final cheese under a concerted program to shed light on the dynamics of the microbiota and to rationalize the failure to receive the PDO seal on microbiological grounds. 2. Materials and Methods 2.1. Sampling Collection Samples of milk, NWS, curd and cheese (9 months ripening) from the traditional production of Sao Jorge cheese were taken aseptically on four different occasions at "UNIQUEIJO: Union of Agricultural Cooperatives" on the island of Sao Jorge (the main producer of PDO cheese). One milk sample and one sample of the NWS were taken from the respective tank in each sampling period. In addition, two samples of the curd were taken each time from different vats, so a total of 8 samples were taken. The milk samples (ca. 100 mL), the NWS (ca. 100 mL) and the cheese curd (ca. 250 g) were kept in sterile individual bottles, refrigerated (4 degC) during transport and stored at -20 degC until DNA extraction. From each of the four production dates, several batches of cheeses (from different vats) matured for 9 months were subjected to sensory analysis by the Sao Jorge Private Control Body and Cheese Certification. After classification, 8 cheeses granted PDO status and 8 cheeses without PDO certification were collected for analysis. The cheese samples (ca. 500 g) were vacuum packed and kept refrigerated (4 degC) until DNA extraction. 2.2. Sample Preparation and DNA Extraction Bacterial cells in milk and NWS samples were concentrated by centrifugation (10 mL) at 7000x g for 10 min (Beckman J2-HS centrifuge). The supernatant was discarded, and the pellet was washed twice with TE buffer (Tris-EDTA: 2M Tris HCl + 0.5M EDTA, pH 8.0); the pellet was then resuspended in 1 mL of TE buffer before DNA extraction. For the preparation of cheese and curd aliquots, 5 g of the sample was weighed and 45 mL of 2% sodium citrate buffer was added, followed by homogenization in a stomacher (400 Circulator, Seward Medical, London, UK) for 2 min at 230 rpm. Total genomic DNA was extracted using the UltraClean(r) extraction kit Microbial DNA Isolation Kit (MoBio, Carlsbad, CA, USA). The quantity and quality of extracted DNA were evaluated by measuring absorbance at 260 and 280 nm (LVis Plate, Fluorstar Omega, BMG Labtech). The quality of the extracted DNA was confirmed via 1.5% agarose (w/v) gel electrophoresis. Only two milk samples yielded good-quality DNA after extraction. Therefore, a total of 30 samples, including raw milk (n = 2), NWS (n = 4), curd (n = 8), PDO cheese (n = 8), and non-PDO cheese (n = 8) were analyzed. 2.3. High-Throughput Sequencing The samples were prepared for Illumina Sequencing by 16S rRNA gene amplification of the bacterial community. The DNA was amplified for the hypervariable V3-V4 region with specific primers and further reamplified in a limited-cycle PCR reaction to add sequencing adapters and dual indexes. PCR reactions were first performed for each sample using the KAPA HiFi HotStart PCR Kit according to the manufacturer's recommendations: 0.3 mM of each PCR primer: forward primer Bakt_341F 5'-CCTACGGGNGGCWGCAG-3' and reverse primer Bakt_805R 5'-GACTACHVGGGTATCTAATCC-3' (Herlemann et al., 2011, Klindworth et al., 2013), and 12.5 ng of template DNA was collected accordingly in a total volume of 25 mL. PCR conditions included denaturation at 95 degC for 3 min, followed by 25 cycles of 98 degC for 20 s, 55 degC for 30 s, and 72 degC for 30 s and a final extension at 72 degC for 5 min. In the second PCR reactions, indexes and sequencing adapters were added to both ends of the amplified target region according to the manufacturer's recommendations (Illumina, 2013). Negative PCR controls were used for all amplification procedures. PCR products were then purified in one step, normalized using the SequalPrep 96-well plate kit (ThermoFisher Scientific, Waltham, MA, USA) (Comeau et al., 2017), pooled, and pair-end sequenced in the Illumina MiSeq(r) sequencer using V3 chemistry, according to the manufacturer's instructions (Illumina, San Diego, CA, USA) at Genoinseq (Cantanhede, Portugal). 2.4. Bioinformatics Sequence data were processed at Genoinseq (Cantanhede, Portugal). Raw reads were extracted from an Illumina MiSeq(r) System in the fastq format and quality-filtered using PRINSEQ v. 0.20.4 to remove sequencing adapters and reads with fewer than 150 bases and trim bases with an average quality lower than Q25 in a 5-base window. The forward and reverse reads were merged by overlapping paired-end reads with AdapterRemoval v. 2.1.5 using default parameters. The QIIME package v. 1.8.0 was used for operational taxonomic unit (OTU) generation, taxonomic identification, sample diversity and richness index calculation. Sample IDs were assigned to the merged reads and converted to the fasta format. Chimeric merged reads were detected and removed using UCHIME against the Greengenes database v. 13.8 (DeSantis et al., 2006). OTUs were selected at a 97% similarity threshold using the open reference strategy. Merged reads were pre-filtered by removing sequences with a similarity below 60% against Greengenes database v. 13.8, and the remaining merged reads were then clustered at 97% similarity against the same database. The merged reads that failed clustering in the previous step were de novo clustered into OTUs at 97% similarity. OTUs with fewer than two reads were removed from the OTU table. A representative sequence of each OTU was then selected for taxonomy assignment. 2.5. Statistical Analysis Alpha diversity indices Chao1, dominance, equitability, goods coverage, observed species, Shannon and Simpson were calculated to reflect the diversity and richness of bacterial communities in the different samples. Chao1 rarefaction curves were also calculated. The OTU profiles of each sample were normalized (total sum normalization, TSS) and compared with the Bray-Curtis distance metric. The calculated Bray-Curtis distances were later used to sort the OTU profiles using principal coordinate analysis (PCoA). A Pearson correlation network was constructed based on the relative number of readings assigned to each genus in cheeses with and without PDO status. The underlying relationships between the genera observed in the cheeses were also analyzed using Spearman correlation. All analyses were performed using the program Calypso, v. 8.84 accessed on 29 January 2023). To determine which genera provided significant discrimination between cheeses with and without PDO status, a stepwise discriminant analysis was performed using Wilks' lambda. The assumptions of normality and homogeneity of the variance-covariance matrices of each group were tested using the Kolmogorov-Smirnov test and Box's M test, respectively. To evaluate possible differences between cheeses with and without PDO status for the main taxonomic genera, the nonparametric Wilcoxon-Mann-Whitney test (for a = 0.05) was also applied. Statistical tests were performed using IBM SPSS Statistics (v. 25, IBM Corporation, New York, NY, USA). 3. Results Based on 97% similarity, a total of 1612 operational taxonomic units (OTUs) were identified (out of a total of 2,039,272 sequence reads), of which, 1580 OTUs were identified in the 30 analyzed samples of milk, NWS, curd, and cheese with and without PDO status (PDO cheese and non-PDO cheese, respectively). Only 32 OTUs (representing 0.01% of the total number of reads) could not be identified. The average value of sequence frequency per sample was 67,976 reads/sample from a minimum of 48,194 reads/sample (PDO cheese) to a maximum of 89,059 reads/sample (non-PDO cheese). Although only two milk samples produced DNA for NGS, they were included in the results to understand the microbial dynamics from milk to curd. 3.1. Alpha Diversity The rarefaction curve showed a tendency to flatten, indicating that bacterial communities were adequately covered in all samples analyzed. This finding was confirmed by the estimated coverage index of the samples (Good's coverage), above 99% in all samples, indicating a good description of microbial diversity (Table 1). The richness and diversity of the bacterial community were assessed for the samples of raw milk, NWS, curd and ripened cheese with or without PDO status, and the assignment was performed using different alpha diversity indices (OTU, Chao1, dominance, equitability, Shannon index and Simpson index) as shown in Table 1. The Chao analysis, which estimates species richness, showed good richness in the samples. There were no significant differences in the Chao1 index (p > 0.05), in contrast to the other diversity indices (p < 0.05) between the species richness of milk and that of NWS, curd and ripened cheese (PDO and non-PDO). The dominance index showed a significantly higher value (p < 0.05) for NWS and curd than for milk and cheese (Table 1). This value indicates a several-fold lower diversity in NWS and curd, which is confirmed by the significantly lower values (p < 0.05) in these samples when considering the number of observed different OTUs, equitability, Shannon index and Simpson index. Conversely, greater species diversity was observed in milk, as indicated by the higher number of different OTUs (p < 0.05) and Shannon index (p < 0.05) compared to all other samples. In contrast, an increase in species diversity, reflected in Shannon and Simpson indices, was observed from NWS to cheese (p < 0.05). There were no significant differences (p > 0.05) in species diversity between cheeses with and without PDO, although cheese with PDO resulted in slightly higher Shannon and Simpson index values (Table 1). 3.2. Taxonomic Composition of Bacterial Communities The relative abundance of sequences identified at the family and genus level is shown in Figure 1. The major families found in milk were Pseudomonadaceae (14-52%), Moraxellaceae (21-4%), Enterobacteriaceae (13-27%) and Streptococcaceae (11-12%). In the M1 sample, the dominant genus was Pseudomonas, whereas the genus Acinetobacter was found in a greater proportion in the M2 sample. Although the milk had a lower abundance of bacteria of the genus Streptococcus, the M2 sample exhibited a greater abundance of this genus than the M1 milk sample . The dominant family in NWS was Streptococcaceae, with a relative abundance exceeding 99%. The Streptococcaceae family also dominated in curd (91 to 99%), although bacteria from the Enterobacteriaceae (1 to 7%), Moraxellaceae (<2%) and Staphylococcaceae (<1%) families were also detected in some samples . Although communities from the Listeriaceae family were detected in milk (<0.3%), it should be noted that no OTUs from this family were found in NWS, curd or cheese samples. At the genus level, the bacterial population in NWS was dominated by the genus Streptococcus (69-92%), followed by the genera Lactococcus (8-31%) and Lactobacillus (0.004-0.9%). NWS samples W2 and W3 were characterized by a higher percentage of Lactococcus ; these samples were also characterized by the presence of bacteria belonging to the genus Acetobacter (0.2-0.3%). The genus Streptococcus was also dominant (65-91%) in the cheese curd , followed by the genus Lactococcus (6-42%). The genera Acinetobacter (0.08-1.1%), Serratia (0-3.3%) and Macrococcus (0.005-1.4%) were detected in curd to a much lesser extent. Regarding the microbiota in aged cheeses (9 months), several differences can be observed between samples of non-PDO (nPDO) and PDO cheeses . The Streptococcaceae family was dominant in non-PDO cheese samples, with the exception of sample 7 (nPDO7). However, there was a marked decrease in the Streptococcaceae family from curd (>99%) to cheese (29-73%). The Lactobacillaceae family was second-most abundant in non-PDO cheeses (12-34%), except in sample 7 (64%). In all non-PDO cheeses, the relative abundance of the Leuconostocaceae family was less than 5% . Bacteria of the families Staphylococcaceae (2-3%, in samples 1 and 2) and Enterococcaceae (0.2-1%) were still detected in some non-PDO cheeses. In contrast, the predominant families in PDO cheeses were: Lactobacillaceae (26-45%), Streptococcaceae (24-36%), Leuconostocaceae (13-31%) and Enterococcaceae (1.5-3%). Thus, according to the sensory evaluation by the trained tasters, there was a clear difference between the cheeses that obtained PDO status and those that did not. The most striking difference concerned the relative abundance of the Leuconostocaceae family, which was higher in all PDO cheeses . In addition, the Lactobacillaceae family was more represented in the PDO cheeses, whereas the Streptococcaceae family dominated in the non-PDO cheeses. At the genus level, Streptococcus (14-58%), Lactobacillus (11-50%) and Lactococcus (9-39%) were the dominant genera of non-PDO (nPDO) cheeses. Among the subdominant microbiota, the following genera were detected: Leuconostoc (0.23-4.1%), Enterococcus (0.22-1.2%), Staphylococcus (0.01-1.75%), Pediococcus (0-1.75%), Macrococcus (0-1.6%), Acinetobacter (0-0.6%), Weissella (0-0.5%), Citrobacter (0-0.4%), Chryseobacterium (0-0.16%), Delftia (0-0.12%) and Enhydrobacter (0-0.11%). In PDO cheeses, the diversity of dominant genera increased, with the genus Lactobacillus standing out. In these cheeses, the dominant genera were Lactobacillus (25-55%), Streptococcus (9-27%), Leuconostoc (8-28%), Lactococcus (8-26%) and Enterococcus (1.5-3.3%), whereas the subdominant genera were Weissella (0.01-2.6%), Macrococcus (0.04-0.75%), Pediococcus (0.02-0.64%), Staphylococcus (0.04-0.56%), Chryseobacterium (0.02-0.25%), Vibrio (0-0.23%), Delftia (0.04-0.17%) and Acinetobacter (0.01-0.16%). Although the genus Lactobacillus was recently reclassified into 25 genera , the name of this genus is retained in this study to denote all organisms classified by 2020. 3.3. Beta Diversity of Bacterial Communities The bacterial communities in the cheese, curd, NWS and milk used in cheese production differ significantly from each other, as shown by the principal coordinate analysis . The first two PCoA axes accounted for 94% of the total variability, with PCoA1 and PCoA2 describing 63% and 31% of the variability, respectively. The first axis (PCoA1) refers to the differentiation of the NWS, curd and cheese populations. PCoA2 differentiates the bacterial community in milk. At both levels (family and genus), there was a high degree of dissimilarity between the bacterial community in the milk and the remaining samples. On the other hand, no differences were found between the NWS and curd samples, as they were grouped together. Some degree of dissimilarity was also found between the bacterial communities of the non-PDO and PDO cheeses, especially at the family level . At the genus level, one sample of cheese without PDO status (sample 7) was included in the PDO group . Cluster analysis confirmed the differentiation observed between the samples at the genus level . A clear separation of the milk cluster--with the highest degree of dissimilarity--from the other clusters was evident. A cluster of NWS and curd samples showed a high degree of similarity and was dominated by the genus Streptococcus. The cluster for non-PDO cheese included six of the eight non-PDO cheese samples and shared the high relative abundance of Streptococcus with the cluster for NWS and curd. On the other hand, the genera Lactobacillus, Leuconostoc and Enterococcus were positively differentiated in the cluster for PDO cheeses. Two samples of non-PDO cheeses (samples 6 and 7) were also included in this cluster. Although they were included in the same cluster as the PDO cheeses, these samples differed in the low abundance of OTUs of the genus Leuconostoc . 3.4. Distinction of Bacterial Communities in PDO and Non-PDO Cheeses PCoA based on Bray-Curtis distance matrix on cheese samples was performed to visualize the differences in the bacterial community between the non-PDO and PDO cheeses. As shown in Figure 4a, the bacterial communities in the PDO cheeses were closer, and more similar to each other than in the non-PDO cheeses. The results of the PCoA analysis were consistent with the network for the bacterial communities of the Sao Jorge cheeses . The network with the interactions of OTUs identified at the genus level unfolded the difference between the PDO cheeses (blue circles) and the non-PDO cheeses (red circles). The genera Leuconostoc, Enterococcus, Lactobacillus, Weissella and Vibrio were associated with PDO cheeses, whereas Streptococcus, Lactococcus, Staphylococcus, Citrobacter, Serratia, Enhydrobacter and Acinetobacter were associated with non-PDO cheeses. To determine which genus best characterizes PDO cheese, a stepwise discriminant analysis was performed that identified the genus Leuconostoc as the variable that significantly differentiates PDO cheese (p < 0.05). These results were confirmed by a nonparametric analysis of the OTUs assigned to the dominant genera in these cheeses . Compared to the non-PDO cheeses, the cheeses bearing the PDO label had a higher proportion of OTUs of the genera Lactobacillus (p < 0.05), Leuconostoc (p < 0.001), and Enterococcus (p < 0.01). In contrast, the PDO cheeses had lower Streptococcus OTUs (p < 0.05) than non-PDO cheeses. The pattern of co-occurrence and exclusion of OTUs in the cheese samples is shown in Figure 6. A strong negative correlation is observed between Streptococcus and Lactobacillus, suggesting that a reduction in Streptococcus dominance is necessary for the development of Lactobacillus during cheese ripening. Negative correlations are also observed between Streptococcus and Leuconostoc and between Streptococcus and Enterococcus, although to a lesser extent. Conversely, the genera Staphylococcus and Acinetobacter exhibited a strong positive correlation with Streptococcus. 4. Discussion Despite the small sample size, the results of a-diversity in milk are consistent with other studies that have found higher species diversity in raw milk compared to cheeses produced from it . The high level of species diversity in milk decreases significantly when moving to NWS and curd. These samples have high dominance values associated with low equitability and lower Shannon and Simpson indices, indicating low diversity in bacterial community composition with dominant populations. NWSs were generally characterized by a relatively simple microbiota. This LAB community is generally thermophilic and well adapted to the particular physicochemical conditions of NWSs . The decrease in biodiversity observed during the transition from milk to curd is expected because the lactic acid production of LAB from NWS lowers the pH, which contributes to cheese curd formation and inhibits pathogen growth from raw milk . However, the biological richness of raw milk is of great importance as it can provide a desirable microbiota associated with specific characteristics of raw milk cheeses . Similar results were obtained with Poro cheese, an artisanal Mexican cheese also produced from raw cow's milk and inoculated with fermented NWS from the previous batch . During the production of this cheese, the bacterial diversity in the milk was high and decreased significantly in the NWS and curd, although it increased again during cheese ripening . Concerning the taxonomic composition of bacterial communities, the results were similar to those reported for milk and curd in the production of traditional Italian cheeses . Other studies provided similar results to our work, with the phylum Proteobacteria predominant in milk, followed by Firmicutes and Bacteroidetes . In contrast, Quigley et al. reported that Firmicutes accounted for ca. 80% of the bacterial community in raw milk in Ireland. The presence of Proteobacteria may unfold hygiene problems in milk, as this phylum includes a wide range of Gram-negative pathogenic bacteria . It should be noted that milk samples were collected from the cold storage tank, knowing that during storage, populations of psychotropic bacteria dominate, which have been reported to contribute to the spoilage of dairy products . The present study also confirms the previous data of Kongo et al. , according to which Enterobacteriaceae were detected in the milk used for the production of Sao Jorge cheese. The presence of high numbers of these bacteria is generally considered an indicator of poor hygiene, and if pathogenic species are also present, this can pose a health risk; it also has a negative effect on the sensory quality of the finished cheese . In contrast, the presence of the genera Lactococcus, Lactobacillus, Leuconostoc and Enterococcus in the milk samples, albeit at relatively low levels, may be critical to the development of desired flavor characteristics during cheese ripening . These bacteria exhibit significant lipolytic and proteolytic activities, so they strongly influence the quality of cheese produced from raw milk . As for the NWS, all samples had a bacterial community dominated by Streptococcaceae, which accounted for 99.1% to 99.9% of the total population. Thus, there was a significant change in the bacterial community during the transition from milk to NWS. This change was predictable since NWS was mainly associated with backslopping, and this method tends to favor the bacterial community best adapted to the fermentation of milk . These results are also in agreement with those of Fontina PDO cheese, where a low correlation was found between the microbiota of raw milk and curd, which was influenced by the composition of the NWS added as a starter culture . The microbial composition at the genus level of the NWS, which was dominated exclusively by Streptococcus and Lactococcus, was similar to starter cultures traditionally used in the production of aged cheese . The bacteria of these genera are known to play a crucial role in acidifying milk at the beginning of cheese making. However, the less frequent presence of the genus Lactobacillus distinguishes this NWS from the one used in the manufacture of other artisanal cheeses . In these cheeses, the NWS showed a microbiota dominated by the genera Lactobacillus and Streptococcus, as was also the case in Silter PDO cheese . This difference is probably due to the heat treatment applied in the production of these cheeses (39-54 degC), i.e., higher temperatures than those commonly used for Sao Jorge cheese (35-36 degC). According to some authors , an increase in temperature during heat treatment leads to a decrease in Lactococcus spp. and an increase in Lactobacillus spp. Acetobacter was also detected in NWS, but is not common in this habitat, although it has been described in some traditional cheeses (Jin et al., 2018). In addition, the presence of Enterococcus in NWS has been reported by some authors (Giannino et al., 2009, Silvetti et al., 2017). However, this genus was essentially not detected in the NWS samples tested. Similar results were obtained in starter cultures used in the production of Italian and Mexican cheeses . Although the genera Leuconostoc and Enterococcus were not detected in the NWS, they were present in lower proportions in the curd, likely imported from the milk. Should they find the right conditions in the cheese ecosystem, such LAB genera would become dominant in the cheese microbiota. The dominant microbiota in the NWS (Streptococcus and Lactococcus) was also found in the curd, whereas the dominant genera in the milk, namely, Pseudomonas and Acinetobacter, underwent a substantial reduction in the curd. These results are comparable to those reported in previous studies on different artisanal cheeses (Quigley et al., 2013, Aldrete-Tapia et al., 2014, De Pasquale et al., 2014). It is known that the changes in the food environment during the fermentation phase exert some selection pressure on the microbial populations present in raw milk, which ultimately favors the growth of LAB. As mentioned earlier, several studies have been published on the microbiota of Sao Jorge cheese . However, all of these studies resorted to cultivation-dependent methods and did not attempt to distinguish between cheeses with and without PDO status. Although cultivation-dependent methods are essential for isolating microorganisms characteristic of cheese, they may underestimate some microbial communities--particularly species that are less well-adapted to growth under conditions commonly used for isolation in the laboratory. With the recent development of new sequencing techniques, it has become possible to assess the composition of bacterial communities in these ecosystems without the bias that their isolation represents. This is, in fact, the first study to apply these methods to gain a better understanding of the microbial community of Sao Jorge cheese. However, this technique is limited to the identification of bacterial communities at the genus level. In addition, NSG methodologies may also introduce some bias due to the methods used in sampling, DNA extraction, PCR amplification, and sequencing (reviewed by Hugerth and Andersson ). According to our results, the ripening of Sao Jorge cheese is dominated by the genera Lactobacillus, Streptococcus, Leuconostoc, and Lactococcus. In general, these dominant genera are similar to those previously found in ripened cheeses produced from raw milk . Previous studies on the microbiota of Sao Jorge cheese also refer to Lactobacillus as the dominant genus at the end of ripening ; however, the genus Enterococcus accounted for 62% of isolates in the curd and 30-37% in the cheese. Given the high selection of Enterococcus by culture media commonly used for bacterial isolation , it is possible that the dominance of Enterococcus reported for this and other traditional cheeses was overestimated. In studies using culture-independent methods, this genus was not found expressively in the microbiota of ripening cheeses . Among the lactobacilli, two species were described as dominant in an earlier study on Sao Jorge cheese: Lacticaseibacillus paracasei and Lacticaseibacillus rhamnosus . In this study, Lactococcus lactis was also identified as dominant in the curd, but no Streptococcus spp. were isolated from the Sao Jorge cheese, despite the dominance of this genus observed in the present study. As for the subdominant microbiota, the genera Weissella, Macrococcus and Pediococcus should be highlighted as potential contributors to cheese texture and flavor . The presence of the genus Vibrio has also been described in Herve PDO cheese, suggesting that these bacteria may play an important role in the ripening process . However, due to their low abundance and sporadic occurrence, this genus is not expected to have a positive impact on the flavor of Sao Jorge cheese. To assess which genera best distinguished PDO cheeses, a discriminant analysis pointed to the genus Leuconostoc (p < 0.05). This result is not surprising since some species of the genus Leuconostoc are well-adapted to the cheese environment and may play an important role in flavor development during ripening . Therefore, our results support a clear distinction between PDO and non-PDO cheeses in terms of the bacterial community. It should be noted that this classification depends solely on the evaluation of a group of tasting experts who grant (or do not) the PDO label based on the sensory characteristics of the cheese. Because this cheese is produced from raw milk without the addition of a well-defined starter culture, it is subject to wide batch-to-batch variations that often disqualify it for PDO status. When samples were taken for this work, the rejection of PDO status was over 50% of batches. Therefore, it seems crucial to know what is expected in terms of the microbiota of said PDO cheese in order to improve the sensory quality of the final product, which could eventually allow a higher percentage of PDO approval. As Leuconostoc has been found to be essential for the differentiation of PDO cheese, it is important to determine the factors that allow the development of these bacteria in the cheese during ripening. The differentiation resulting from the development of Leuconostoc may result from the environment created by the particular microbial ecology of each vat. The presence of a specific microbial community can favor the development of beneficial bacteria for the flavor development of the cheese, which guarantees the awarding of PDO status. It should also be noted that the genera characteristic of Sao Jorge cheese, such as Leuconostoc and Lactobacillus, showed negative correlations with bacteria considered contaminants, e.g., Staphylococcus and the proteobacteria Acinetobacter, Serratia, Klebsiella, Erwinia, Citrobacter, Enhydrobacter and Bacillus. Similar results were reported by Zheng et al. , who observed a negative correlation between Lactobacillus and Lactococcus, and Acinetobacter and Staphylococcus in Kazak artisan cheese. The pattern of co-occurrence and exclusion suggests that good milk quality, low levels of contaminating bacteria and good equipment hygiene may control the dominance of Streptococcus during cheese ripening. Such control would allow the growth of Leuconostoc spp. as well as Lactobacillus spp. and Enterococcus spp., thus ensuring the proper development of the intended characteristic flavors in Sao Jorge PDO cheese. Thus, our results indicate that LAB populations, especially of Leuconostoc and Lactobacillus, dominate the microbiota of Sao Jorge PDO cheese and limit the development of spoilage bacteria, as in other cheeses . Recently, Lactobacillus and Lactococcus were also shown to positively correlate with cheese quality in traditional Chinese cheeses . 5. Conclusions The unique characteristics of Sao Jorge PDO cheese are related to the microbiota present in its ingredients (milk and NWS), which in turn are controlled by the production process and the ripening period. Milk stored in tanks and used for cheese production is dominated by Gram-negative bacteria of the genera Pseudomonas and Acinetobacter, whereas Lactococcus and Streptococcus were detected in lower numbers. On the other hand, the microbial composition of NWS was dominated by Streptococcus, followed by Lactococcus, which should play a positive role in curd acidification. These genera were retained in the curd, with a decrease in Streptococcus and an increase in Lactococcus. However, during ripening, a decrease in Streptococcus and an increase in Lactobacillus and Leuconostoc communities were observed. Thus, the microbiota of Sao Jorge cheese was dominated by the genera Lactobacillus, Streptococcus, Leuconostoc and Lactococcus. This work contributed to clearly distinguishing between PDO and non-PDO cheeses in terms of bacterial community composition. PDO status is assigned using empirical methods based on sensory analysis by a tasting panel. PDO cheeses have been found to own a distinctive bacterial community in which the genus Leuconostoc is a distinguishing feature. Leuconostoc bacteria are associated with the development of flavor during the ripening process, so they should play a major role in the final sensory characteristics of Sao Jorge PDO cheese. In addition to the genus Leuconostoc, PDO cheeses were characterized by a higher occurrence of the genera Lactobacillus and Enterococcus and a lower occurrence of Streptococcus compared to non-PDO cheeses. The pattern of co-occurrence and exclusion of OTUs in cheese samples suggests that the presence of contaminating bacteria does not favor the development of bacteria associated with PDO status. Therefore, good milk quality appears to be essential for the development of a community rich in the genera Leuconostoc and Lactobacillus characteristic of Sao Jorge PDO cheese. The results of this study will allow a better understanding of the bacterial community of Sao Jorge cheese and its use to distinguish between non-PDO and PDO cheeses by applying culture-independent techniques. This information is important for developing strategies to increase the percentage of cheeses that can obtain the PDO label, which will ultimately have a positive impact on the economic aspects of Sao Jorge cheese production. Supplementary Materials The following supporting information can be downloaded at: Figure S1: Rarefaction curves of the variable region of 16S rRNA sequences from samples obtained during Sao Jorge cheese manufacture. (A) Milk (samples S16 and S28); (B) whey used as starter culture in the manufacture of Sao Jorge cheese (samples S24-27); (C) curd (samples S17-S23, S30); (D) non-PDO cheese (samples S1-S6, S12, S29). (E) PDO cheeses (samples S7-S11, S13-S15). Click here for additional data file. Author Contributions M.C.C.: experimentation, data analysis and writing--original draft. F.X.M.: methodology, supervision and writing--review and editing. C.C.G.S.: methodology, data analysis, supervision and writing--review and editing. All authors have read and agreed to the published version of the manuscript. Data Availability Statement The data presented in this study are openly available in the Sequence Read Archive database (NCBI) under BioProject PRJNA908105. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Relative abundance (%) of sequences identified at the family (a) and genus (b) level in milk (M), whey (W), curd (C), non-PDO cheeses (nPDO) and PDO cheeses (PDO). Only taxa contributing more than 0.1% of the total abundance in at least one sample are shown. Figure 2 Principal coordinate analysis (PCoA) based on the Bray-Curtis distance matrix of operational taxonomic units (OTUs) identified at the family (a) and genus (b) level of milk, NWS (whey), and cheese samples (PDO and nPDO cheeses). The ellipses were drawn by hand to help visualizing the different sample types. Figure 3 Dendrogram and heat map representing the distribution (%) of bacterial genera in samples of milk, whey (NWS), curd, non-PDO cheeses (nPDO) and PDO cheeses (PDO). Only OTUs that occur with an abundance of more than 0.1% in at least one sample were included. The grouping of the samples was obtained by hierarchical clustering. The color code, from blue to red, indicates the Bray-Curtis distance metric, where the color blue represents maximum dissimilarity and red shows maximum similarity. In the upper part, the color code indicates the relative abundance of OTUs in the sample, ranging from white (low abundance) to black (high abundance). Figure 4 (a) Principal coordinate analysis (PCoA) based on the Bray-Curtis distance matrix of OTUs identified at the genus level of PDO and non-PDO cheese samples. The ellipses were drawn by hand to help visualizing the different cheese types. (b) Correlation network of co-occurrence patterns of OTUs classified at the genus level between non-PDO cheeses (red circles) and PDO cheeses (blue circles). The size of each node indicates the richness of OTUs belonging to each taxonomic group. Lines connecting two nodes represent significant positive correlations (p < 0.05). Figure 5 Comparison of OTUs identified at the genus level between samples of non-PDO cheese (nPDO) and PDO cheese. The genera compared were: (a) Streptococcus, (b) Lactobacillus, (c) Lactococcus, (d) Leuconostoc, (e) Enterococcus, (f) Staphylococcus, (g) Pediococcus, (h) Macrococcus, (i) Weissela and (j) Serratia. * p < 0.05, ** p < 0.01, *** p < 0.001. Figure 6 Co-occurrence and co-exclusion relationships between genera observed in cheeses, based on Spearman's correlation. Only OTUs with a frequency greater than 0.5% in at least one sample are shown. The color of the scale bar indicates the type of correlation, with +1 indicating a positive correlation (dark red) and -1 indicating a negative correlation (dark blue). foods-12-00990-t001_Table 1 Table 1 Alpha diversity indices observed in samples of raw milk, NWS, curd, non-PDO cheese (nPDO) and PDO cheese (PDO). Values of the mean and standard deviation (SD) are indicated. Samples N Chao1 SD Dominance * SD Equitability * SD Good's Coverage SD N. OTUs Observed * SD Shannon Index * SD Simpson Index * SD Milk 2 446 130 0.0935a 0.0697 0.5625a 0.0596 0.9987 0.0004 402a 115 4.860a 0.750 0.907a 0.070 NWS 4 302 45 0.3874b 0.0788 0.2467b 0.0441 0.9988 0.0002 139b 11 1.759b 0.340 0.613b 0.079 Curd 8 344 83 0.3588b 0.0910 0.2609b 0.0621 0.9986 0.0003 206b.c 43 2.002b 0.491 0.641b 0.091 nPDO 8 420 117 0.2000a 0.0267 0.3750c 0.0227 0.9986 0.0003 250c 58 2.975c 0.273 0.800a 0.027 PDO 8 401 71 0.1659a 0.0388 0.4084c 0.0235 0.9985 0.0003 232b.c 42 3.200c 0.199 0.834a 0.039 * Different letters within a column represent significant differences (p < 0.05). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Settanni L. Moschetti G. Non-starter lactic acid bacteria used to improve cheese quality and provide health benefits Food Microbiol. 2010 27 691 697 10.1016/j.fm.2010.05.023 20630311 2. Coelho M.C. Malcata F.X. Silva C.C. Lactic acid bacteria in raw-milk cheeses: From starter cultures to probiotic functions Foods 2022 11 2276 10.3390/foods11152276 35954043 3. Calasso M. Ercolini D. Mancini L. Stellato G. Minervini F. Di Cagno R. De Angelis M. Gobbetti M. Relationships among house, rind and core microbiotas during manufacture of traditional Italian cheeses at the same dairy plant Food Microbiol. 2016 54 115 126 10.1016/j.fm.2015.10.008 4. Londono-Zapata A.F. Durango-Zuleta M.M. Sepulveda-Valencia J.U. Herrera C.X.M. Characterization of lactic acid bacterial communities associated with a traditional Colombian cheese: Double cream cheese LWT-Food Sci. Technol. 2017 82 39 48 10.1016/j.lwt.2017.03.058 5. Motato K.E. Milani C. Ventura M. Valencia F.E. Ruas-Madiedo P. Delgado S. Bacterial diversity of the Colombian fermented milk "Suero Costeno" assessed by culturing and high-throughput sequencing and DGGE analysis of 16S rRNA gene amplicons Food Microbiol. 2017 68 129 136 10.1016/j.fm.2017.07.011 28800820 6. Silvetti T. Capra E. Morandi S. Cremonesi P. Decimo M. Gavazzi F. Giannico R. De Noni I. Brasca M. Microbial population profile during ripening of Protected Designation of Origin (PDO) Silter cheese, produced with and without autochthonous starter culture LWT 2017 84 821 831 10.1016/j.lwt.2017.06.022 7. Choi J. In Lee S. Rackerby B. Frojen R. Goddik L. Ha S.-D. Park S.H. Assessment of overall microbial community shift during Cheddar cheese production from raw milk to aging Appl. Microbiol. Biotechnol. 2020 104 6249 6260 10.1007/s00253-020-10651-7 32451588 8. Terzic-Vidojevic A. Veljovic K. Tolinacki M. Zivkovic M. Lukic J. Lozo J. Fira D. Jovcic B. Strahinic I. Begovic J. Diversity of non-starter lactic acid bacteria in autochthonous dairy products from Western Balkan Countries-technological and probiotic properties Food Res. Int. 2020 136 109494 10.1016/j.foodres.2020.109494 32846575 9. Yeluri Jonnala B.R. McSweeney P.L. Sheehan J.J. Cotter P.D. Sequencing of the cheese microbiome and its relevance to industry Front. Microbiol. 2018 9 1020 10.3389/fmicb.2018.01020 29875744 10. Riquelme C. Camara S. Maria de Lurdes N. Vinuesa P. da Silva C.C.G. Malcata F.X. Rego O.A. Characterization of the bacterial biodiversity in Pico cheese (an artisanal Azorean food) Int. J. Food Microbiol. 2015 192 86 94 10.1016/j.ijfoodmicro.2014.09.031 25440551 11. Dugat-Bony E. Garnier L. Denonfoux J. Ferreira S. Sarthou A.-S. Bonnarme P. Irlinger F. Highlighting the microbial diversity of 12 French cheese varieties Int. J. Food Microbiol. 2016 238 265 273 10.1016/j.ijfoodmicro.2016.09.026 27710867 12. Porcellato D. Skeie S.B. Bacterial dynamics and functional analysis of microbial metagenomes during ripening of Dutch-type cheese Int. Dairy J. 2016 61 182 188 10.1016/j.idairyj.2016.05.005 13. Ceugniez A. Taminiau B. Coucheney F. Jacques P. Delcenserie V. Daube G. Drider D. Use of a metagenetic approach to monitor the bacterial microbiota of "Tomme d'Orchies" cheese during the ripening process Int. J. Food Microbiol. 2017 247 65 69 10.1016/j.ijfoodmicro.2016.10.034 27817942 14. Afshari R. Pillidge C.J. Dias D.A. Osborn A.M. Gill H. Cheesomics: The future pathway to understanding cheese flavour and quality Crit. Rev. Food Sci. Nutr. 2020 60 33 47 10.1080/10408398.2018.1512471 30285475 15. Rocha R. Vaz Velho M. Santos J. Fernandes P. Serra da estrela pdo cheese microbiome as revealed by next generation sequencing Microorganisms 2021 9 2007 10.3390/microorganisms9102007 34683326 16. Alessandria V. Ferrocino I. De Filippis F. Fontana M. Rantsiou K. Ercolini D. Cocolin L. Microbiota of an Italian Grana-like cheese during manufacture and ripening, unraveled by 16S rRNA-based approaches Appl. Environ. Microbiol. 2016 82 3988 3995 10.1128/AEM.00999-16 27107125 17. Dalmasso A. Civera T. Pattono D. Cardazzo B. Bottero M.T. Characterization of microbiota in Plaisentif cheese by high-throughput sequencing LWT-Food Sci. Technol. 2016 69 490 496 10.1016/j.lwt.2016.02.004 18. Neviani E. Bottari B. Lazzi C. Gatti M. New developments in the study of the microbiota of raw-milk, long-ripened cheeses by molecular methods: The case of Grana Padano and Parmigiano Reggiano Front. Microbiol. 2013 4 36 10.3389/fmicb.2013.00036 23450500 19. Kongo J.M. Ho A.J. Malcata F.X. Wiedmann M. Characterization of dominant lactic acid bacteria isolated from Sao Jorge cheese, using biochemical and ribotyping methods J. Appl. Microbiol. 2007 103 1838 1844 10.1111/j.1365-2672.2007.03423.x 17953594 20. Cogan T. Beresford T. Steele J. Broadbent J. Shah N. Ustunol Z. Invited review: Advances in starter cultures and cultured foods J. Dairy Sci. 2007 90 4005 4021 10.3168/jds.2006-765 17699017 21. Kongo J.M. Gomes A.P. Malcata F.X. Monitoring and identification of bacteria associated with safety concerns in the manufacture of Sao Jorge, a Portuguese traditional cheese from raw cow's milk J. Food Prot. 2008 71 986 992 10.4315/0362-028X-71.5.986 18522034 22. Kongo J.M. Gomes A.M. Malcata F.X. McSweeney P. Microbiological, biochemical and compositional changes during ripening of Sao Jorge-a raw milk cheese from the Azores (Portugal) Food Chem. 2009 112 131 138 10.1016/j.foodchem.2008.05.067 23. Domingos-Lopes M. Stanton C. Ross P. Dapkevicius M. Silva C. Genetic diversity, safety and technological characterization of lactic acid bacteria isolated from artisanal Pico cheese Food Microbiol. 2017 63 178 190 10.1016/j.fm.2016.11.014 28040167 24. Schmieder R. Edwards R. Quality control and preprocessing of metagenomic datasets Bioinformatics 2011 27 863 864 10.1093/bioinformatics/btr026 21278185 25. Schubert M. Lindgreen S. Orlando L. AdapterRemoval v2: Rapid adapter trimming, identification, and read merging BMC Res. Notes 2016 9 1 7 10.1186/s13104-016-1900-2 26725043 26. Caporaso J.G. Kuczynski J. Stombaugh J. Bittinger K. Bushman F.D. Costello E.K. Fierer N. Pena A.G. Goodrich J.K. Gordon J.I. QIIME allows analysis of high-throughput community sequencing data Nat. Methods 2010 7 335 336 10.1038/nmeth.f.303 20383131 27. Edgar R.C. Haas B.J. Clemente J.C. Quince C. Knight R. UCHIME improves sensitivity and speed of chimera detection Bioinformatics 2011 27 2194 2200 10.1093/bioinformatics/btr381 21700674 28. Zheng J. Wittouck S. Salvetti E. Franz C.M. Harris H. Mattarelli P. O'Toole P.W. Pot B. Vandamme P. Walter J. A taxonomic note on the genus Lactobacillus: Description of 23 novel genera, emended description of the genus Lactobacillus Beijerinck 1901, and union of Lactobacillaceae and Leuconostocaceae Int. J. Syst. Evol. Microbiol. 2020 70 2782 2858 10.1099/ijsem.0.004107 32293557 29. Aldrete-Tapia A. Escobar-Ramirez M.C. Tamplin M.L. Hernandez-Iturriaga M. High-throughput sequencing of microbial communities in Poro cheese, an artisanal Mexican cheese Food Microbiol. 2014 44 136 141 10.1016/j.fm.2014.05.022 25084655 30. Ramezani M. Hosseini S. Ferrocino I. Amoozegar M. Cocolin L. Molecular investigation of bacterial communities during the manufacturing and ripening of semi-hard Iranian Liqvan cheese Food Microbiol. 2017 66 64 71 10.1016/j.fm.2017.03.019 28576374 31. De Filippis F. La Storia A. Stellato G. Gatti M. Ercolini D. A selected core microbiome drives the early stages of three popular Italian cheese manufactures PLoS ONE 2014 9 e89680 10.1371/journal.pone.0089680 24586960 32. Gobbetti M. Di Cagno R. Calasso M. Neviani E. Fox P.F. De Angelis M. Drivers that establish and assembly the lactic acid bacteria biota in cheeses Trends Food Sci. Technol. 2018 78 244 254 10.1016/j.tifs.2018.06.010 33. Canon F. Nidelet T. Guedon E. Thierry A. Gagnaire V. Understanding the mechanisms of positive microbial interactions that benefit lactic acid bacteria co-cultures Front. Microbiol. 2020 11 2088 10.3389/fmicb.2020.02088 33013761 34. Calasso M. Minervini F. De Filippis F. Ercolini D. De Angelis M. Gobbetti M. Attenuated Lactococcus lactis and surface bacteria as tools for conditioning the microbiota and driving the ripening of semisoft Caciotta cheese Appl. Environ. Microbiol. 2020 86 e02165-19 10.1128/AEM.02165-19 31862717 35. Quigley L. O'Sullivan O. Stanton C. Beresford T.P. Ross R.P. Fitzgerald G.F. Cotter P.D. The complex microbiota of raw milk FEMS Microbiol. Rev. 2013 37 664 698 10.1111/1574-6976.12030 23808865 36. Rizzatti G. Lopetuso L. Gibiino G. Binda C. Gasbarrini A. Proteobacteria: A common factor in human diseases BioMed Res. Int. 2017 2017 9351507 10.1155/2017/9351507 29230419 37. Parente E. Ricciardi A. Zotta T. The microbiota of dairy milk: A review Int. Dairy J. 2020 107 104714 10.1016/j.idairyj.2020.104714 38. Hantsis-Zacharov E. Halpern M. Culturable psychrotrophic bacterial communities in raw milk and their proteolytic and lipolytic traits Appl. Environ. Microbiol. 2007 73 7162 7168 10.1128/AEM.00866-07 17890340 39. Yann D. Pauline G. Usefulness of natural starters in food industry: The example of cheeses and bread Food Nutr. Sci. 2014 5 1679 1691 10.4236/fns.2014.517181 40. Dolci P. De Filippis F. La Storia A. Ercolini D. Cocolin L. rRNA-based monitoring of the microbiota involved in Fontina PDO cheese production in relation to different stages of cow lactation Int. J. Food Microbiol. 2014 185 127 135 10.1016/j.ijfoodmicro.2014.05.021 24960294 41. Irlinger F. Layec S. Helinck S. Dugat-Bony E. Cheese rind microbial communities: Diversity, composition and origin FEMS Microbiol. Lett. 2015 362 1 11 10.1093/femsle/fnu015 25670699 42. Rossetti L. Fornasari M.E. Gatti M. Lazzi C. Neviani E. Giraffa G. Grana Padano cheese whey starters: Microbial composition and strain distribution Int. J. Food Microbiol. 2008 127 168 171 10.1016/j.ijfoodmicro.2008.06.005 18620769 43. Morandi S. Battelli G. Silvetti T. Goss A. Cologna N. Brasca M. How the biodiversity loss in natural whey culture is affecting ripened cheese quality? The case of Trentingrana cheese LWT 2019 115 108480 10.1016/j.lwt.2019.108480 44. Adamberg K. Kask S. Laht T.-M. Paalme T. The effect of temperature and pH on the growth of lactic acid bacteria: A pH-auxostat study Int. J. Food Microbiol. 2003 85 171 183 10.1016/S0168-1605(02)00537-8 12810281 45. Dias S. Oliveira M. Semedo-Lemsaddek T. Bernardo F. Probiotic potential of autochthone microbiota from Sao Jorge and Parmigiano-Reggiano cheeses Food Nutr. Sci. 2014 5 1793 46. Hugerth L.W. Andersson A.F. Analysing microbial community composition through amplicon sequencing: From sampling to hypothesis testing Front. Microbiol. 2017 8 1561 10.3389/fmicb.2017.01561 28928718 47. Biolcati F. Ferrocino I. Bottero M.T. Dalmasso A. The bacterial and fungal microbiota of "Robiola di Roccaverano" protected designation of origin raw milk cheese Front. Microbiol. 2022 12 776862 10.3389/fmicb.2021.776862 35173686 48. Jingkai J. Jianming Z. Zhenmin L. Huaxi Y. Dynamic changes of microbiota and texture properties during the ripening of traditionally prepared cheese of China Arch. Microbiol. 2020 202 2059 2069 10.1007/s00203-020-01921-z 32488559 49. Nero L.A. Andretta M. Almeida T.T. Ferreira L.R. Camargo A.C. Yamatogi R.S. Carvalho A.F. Call D.R. Lactic microbiota of the minas artisanal cheese produced in the serro region, Minas Gerais, Brazil LWT 2021 148 111698 10.1016/j.lwt.2021.111698 50. Papadakis P. Konteles S. Batrinou A. Ouzounis S. Tsironi T. Halvatsiotis P. Tsakali E. Van Impe J.F. Vougiouklaki D. Strati I.F. Characterization of bacterial microbiota of PDO Feta cheese by 16S metagenomic analysis Microorganisms 2021 9 2377 10.3390/microorganisms9112377 34835502 51. Perin L.M. Sardaro M.L.S. Nero L.A. Neviani E. Gatti M. Bacterial ecology of artisanal Minas cheeses assessed by culture-dependent and-independent methods Food Microbiol. 2017 65 160 169 10.1016/j.fm.2017.02.005 28399998 52. Fuka M.M. Wallisch S. Engel M. Welzl G. Havranek J. Schloter M. Dynamics of bacterial communities during the ripening process of different Croatian cheese types derived from raw ewe's milk cheeses PLoS ONE 2013 8 e80734 10.1371/journal.pone.0080734 24278315 53. Delcenserie V. Taminiau B. Delhalle L. Nezer C. Doyen P. Crevecoeur S. Roussey D. Korsak N. Daube G. Microbiota characterization of a Belgian protected designation of origin cheese, Herve cheese, using metagenomic analysis J. Dairy Sci. 2014 97 6046 6056 10.3168/jds.2014-8225 25064656 54. Hemme D. Foucaud-Scheunemann C. Leuconostoc, characteristics, use in dairy technology and prospects in functional foods Int. Dairy J. 2004 14 467 494 10.1016/j.idairyj.2003.10.005 55. Pedersen T.B. Ristagno D. McSweeney P.L.H. Vogensen F.K. Ardo Y. Potential impact on cheese flavour of heterofermentative bacteria from starter cultures Int. Dairy J. 2013 33 112 119 10.1016/j.idairyj.2013.03.003 56. Zheng X. Xu X. Ma Y. Zhu L. Xiao J. Deng L. Shi X. Wang B. Diversity and potential function of bacterial communities during milk fermentation of Kazak artisanal cheese Process Biochem. 2021 111 191 200 10.1016/j.procbio.2021.11.005 57. Zhang K. Jia M. Guo Z. Li Y. Li B. Li X. Evaluation of bacterial diversity of traditional cheese in Tarbagatay Prefecture, China, and its correlation with cheese quality Food Sci. Nutr. 2021 9 3155 3164 10.1002/fsn3.2275 34136180 |
PMC10000651 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051047 foods-12-01047 Article Microbial Load of Fresh Blueberries Harvested by Different Methods Wang Peien Formal analysis Data curation Writing - original draft Writing - review & editing 1 Hur Minji Formal analysis Data curation Writing - original draft 1 Cai Yixin Data curation 2 Takeda Fumiomi Conceptualization Writing - review & editing Funding acquisition 3 DeVetter Lisa Conceptualization Data curation Writing - review & editing Funding acquisition 2 Chen Jinru Conceptualization Methodology Formal analysis Writing - review & editing Supervision Funding acquisition 1* Osaili Tareq Academic Editor 1 Department of Food Science and Technology, The University of Georgia, Griffin, GA 30223, USA 2 Department of Horticulture and Northwestern Washington Research and Extension Center, Washington State University, Mount Vernon, WA 98273, USA 3 Appalachian Fruit Research Station, Agricultural Research Service, US Department of Agriculture, Kearneysville, WV 25430, USA * Correspondence: [email protected]; Tel.: +1-770-412-4738 01 3 2023 3 2023 12 5 104706 12 2022 17 1 2023 26 1 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Currently, more and more growers are transitioning to the use of over-the-row machine harvesters for harvesting fresh market blueberries. This study assessed the microbial load of fresh blueberries harvested by different methods. Samples (n = 336) of 'Draper' and 'Liberty' northern highbush blueberries, which were harvested using a conventional over-the-row machine harvester, a modified machine harvester prototype, ungloved but sanitized hands, and hands wearing sterile gloves were collected from a blueberry farm near Lynden, WA, in the Pacific Northwest at 9 am, 12 noon, and 3 pm on four different harvest days during the 2019 harvest season. Eight replicates of each sample were collected at each sampling point and evaluated for the populations of total aerobes (TA), total yeasts and molds (YM), and total coliforms (TC), as well as for the incidence of fecal coliforms and enterococci. The harvest method was a significant factor (p < 0.05) influencing the TA and TC counts, the harvest time was a significant factor influencing the YM counts, while the blueberry cultivar was an insignificant (p > 0.05) factor for all three indicator microorganisms. These results suggest that effective harvester cleaning methods should be developed to prevent fresh blueberry contamination by microorganisms. This research will likely benefit blueberry and other fresh fruit producers. blueberry blueberry machine harvester food safety handpicking Washington State Department of AgricultureK2538 The project was sponsored by a Specialty Crop Block Grant from the Washington State Department of Agriculture awarded to Lisa Wasko DeVetter (K2538). pmc1. Introduction Blueberry (Vaccinium spp.) consumption in the United States has steadily increased over the past few decades, primarily due to increased consumer awareness of the health benefits associated with blueberry consumption. A large quantity of blueberries is produced on a global scale to meet this increasing consumer demand . The United States is currently the largest blueberry-producing country, with a production share of approximately 56% between 2009 and 2011 . Within the United States, Washington has become the largest blueberry-producing area in the world, with ca. 75,000 kg of utilized production in 2020, and about half of its utilized production goes to the fresh market . Blueberries for the fresh market are primarily harvested by hand during the early harvest season . Handpicking favors high-quality fruit with a firmer texture and longer postharvest shelf life . However, a lack of worker availability and increasing labor costs are major constraints for berry growers that rely on hand labor. Thus, growers are increasingly transitioning to the use of over-the row (OTR) machine harvesters to harvest blueberries for the fresh market . During the harvesting process of a top-loading machine, blueberries are separated from bushes by the vibration of plastic beater bars mounted on rotary shaking drums. The separated fruit fall onto catching plates and adjacent hard surfaces at the bottom of the harvester . The fruit roll onto horizontal conveyors that transport the fruit to the rear of the harvester, where they fall into small buckets that elevate the fruit to the top platform of the harvester. Once at the top platform, debris and leaves, clustered berries, as well as small, immature berries are separated as the blueberries fall through a fast-moving upward stream of air. The air-sorted blueberries are dropped onto a horizontal manual grading/inspection belt to remove some crushed and diseased fruit, before dropping into the harvest lugs. Although there are several advantages associated with OTR machine harvesting, machine-harvested blueberries have internal bruise damage . The hard surfaces of machine harvesters, such as the fruit-shaking beater rods and plastic-catching plates, create significant damage to the harvested berries, including increased bruising and reduced firmness compared to hand-harvested fruit . As a result, berries harvested by mechanical harvesters have a shorter shelf life than hand-harvested blueberries . To reduce the blueberry quality losses caused by mechanical harvesting, several attempts have been made to modify standard OTR harvesters by replacing the hard surfaces with softer materials . An OTR machine prototype with modified fruit catching plates and intermittent soft surfaces suspended above the hard surfaces was used to harvest blueberries . Brown et al. and Sargent et al. reported that the modification of the hard catch plates on a blueberry machine harvester with a soft cushioning material significantly reduced the physical impacts to the harvested berries. However, it is currently unknown whether such modifications would adversely affect the microbial safety of harvested fruit. The objective of this study was to compare the microbial load of fresh blueberries harvested using a standard OTR machine harvester, a prototype modified machine harvester with softer catching surfaces, ungloved but cleaned and sanitized hands, and hands wearing sterile gloves. 2. Materials and Methods 2.1. Sample Collection Samples of 'Liberty' and 'Draper' northern highbush (Vaccinium corymbosum) blueberries were collected at two different geographic locations within a commercial blueberry farm located near Lynden, WA, in the Pacific Northwest region of the United States (latitude 48deg58''38.35, longitude-122deg21''27.07 [whatismygps.com, accessed on 10 January 2020], altitude 40 m [how-far.net]) in July and September of 2019. The blueberry samples were collected at three different times (9 am, 12 noon, and 3 pm) on four different harvesting days separated by two months. The average temperatures at 9 am, 12 noon, and 3 pm on the four harvest days were 18.33 degC +- 0.48, 21.67 degC +- 2.59, and 23.89 degC +- 3.60, respectively, and the relative humidity values were 89.65% +- 6.10, 72.78% +- 17.30, and 62.40% +- 22.59, respectively. At each sampling point, eight replicates of the blueberry samples were collected using each of the four harvesting methods: ungloved but cleaned and sanitized hands, standard OTR machine harvester, and modified OTR machine harvester prototype, while the samples collected by hands wearing sterile nitrile gloves (Fisher Scientific, Pittsburg, PA, USA) served as controls. The workers on both machine harvesters also wore sterile nitrile gloves. Both types of machine harvesters were manufactured by the Oxbo International Corporation (Lynden, WA, USA). The catch surfaces and catch plates of the conventional OTR harvester were made of proprietary hard polycarbonate plastic. As for the modified OTR prototype, both hard catch surfaces and the center of each catch plate were replaced with a type of soft, food-grade elastomeric polymer within a slender polycarbonate frame left on each plate . The handpicked blueberry samples were randomly collected from different locations on multiple blueberry bushes, while the berries harvested using the conventional and modified OTR machines were randomly collected from stackable berry lugs, with one sample being collected per lug. The collected blueberries were placed in sterile Whirl-Pak bags (Nasco, Fort Atkinson, WI, USA) and kept in a cooler (Rubbermaid; Newell Brands Inc, Atlanta, GA, USA) with ice packs (VWR International LLC, Radnor, PA, USA) in the field and during transportation to a laboratory in Lynden, WA, USA, for the sample processing. 2.2. Sample Processing and Transportation The collected blueberry samples were stored in the cooler no longer than 1 h after each harvest time point, and they were processed immediately upon arrival at the laboratory. Each blueberry sample (25 g) taken from fruit harvested in the field was placed in a fresh, sterile Whirl-Pak bag and homogenized (Stomacher 80, Seward Ltd., West Sussex, UK) in 50 mL of sterile 0.2 M phosphate-buffered saline (PBS, pH 7.4) for 1 min at normal speed. The resulting homogenate, in a volume of 7.5 mL, was transferred to a 15 mL conical centrifuge tube (Fisher Scientific) containing 2.5 mL of 60% glycerol to make the final glycerol concentration of 15% (v/v). Each sample was vigorously mixed and then stored at -20 degC before being transported overnight by aircraft to Atlanta, GA, followed by ground transportation to our laboratory in Griffin, GA, in an insulated polystyrene foam container (Polar Tech 266C Thermo Chill Insulated Carton with Foam Shipper, 19'' x 12'' x 16'', Genoa, IL, USA) with 5 lb. of dry ice (Ice Company LLC, Lynden, WA, USA). The samples were analyzed immediately upon arrival at the laboratory on the University of Georgia Griffin Campus. 2.3. Microbiological Analyses The frozen blueberry homogenates described above were thawed upon arrival at the laboratory. The individual blueberry homogenates (100 mL) were inoculated in duplicate on four different microbiological media, including tryptic soy agar (TSA), potato dextrose agar (PDA) acidified with 10% tartaric acid to pH 3.5, MacConkey agar (MAC), and enterococcus agar (EA). All the microbial media used in the study were purchased from Becton Dickinson Inc. (Franklin Lake, NJ, USA). Serial dilutions were made in sterile 0.1 M PBS when necessary, before inoculation onto the various microbiological media. The incubation condition for the TA, TC, and enterococci was at 37 degC for 24 to 48 h, fecal coliforms at 44.5 degC for 24 h, and YM at 25 degC for 48 to 72 h. The colonies were enumerated after the incubation, and the results were presented as the log colony-forming units per gram of fresh blueberry sample (log CFU/g). The detection limit of the plate count assay was 1.12 log CFU/g. 2.4. Statistical Analysis A split-plot design analysis of variance (ANOVA) was used in the current study. Each sampling location with one blueberry cultivar planted was considered to be the main plot experimental unit, while the cultivar itself was the main plot factor. The sampling time and harvesting method were two factors nested within each level of the main plot, forming subplot factors. The random error term was reflected by the effect of the repeated visits, as well as by the interaction effect between the visits and different cultivars. All the microbiological data were fitted into a general linear model, and Fisher's least significant difference test was used to separate the means (SAS, version 9.4, Statistical Analysis System Institute, Inc., Cary, NC, USA). Differences were considered significant when the p values were smaller than 0.05. The numbers of positive samples for the fecal coliforms and enterococci were recorded to calculate the respective incidence of these two indicators in the total number of samples collected within the same factor. 3. Results According to the results of the type III test, the harvesting method was a significant (p < 0.05) factor influencing the counts of TA and TC, the sampling time was a significant factor influencing only the counts of YM on the sampled blueberries, while the blueberry cultivar was an insignificant (p > 0.05) factor for the counts of all three indicator microorganisms (Table 1). No significant interactions between or among the independent variables were observed. The samples of the two blueberry cultivars had similar (p > 0.05) mean TA, YM, and TC counts (Table 2). In general, the average TA and TC counts on the sampled blueberries were low, with the mean TA count being ca. 2 log CFU/g and the mean TC count being less than 0.5 log CFU/g. In comparison, the mean YM counts were relatively higher, being in the 3 log CFU/g range in the berries from both cultivars. On average, the berries collected at the three different time intervals had similar mean TA counts (Table 2). Similar mean YM counts were found for the 9 am and 12 noon samples, and the counts for the samples collected at these two intervals were significantly higher (p < 0.05) than the same count for the 3 pm samples. The 3 pm samples also had significantly lower mean TC counts compared to the 9 am samples. The blueberries harvested using the standard OTR machine harvester and modified OTR machine harvester prototype had significantly higher (p < 0.05) mean TA counts than those harvested using hands with sterile gloves and ungloved but cleaned and sanitized hands (Table 2). The berry samples collected using the modified OTR prototype also had significantly higher TC counts than those harvested using the other three different methods. However, the levels of the mean YM counts for all the collected berry samples were not different (p > 0.05). When the microbial counts for the berries harvested using different methods were separated by the time of the sample collection, the samples harvested using the two types of machines had significantly higher (p < 0.05) TA counts than those harvested using the other two methods in the 9 am and 3 pm samples (Table 3). For the noon samples, however, the samples collected using bare hands had similar TA counts to the samples collected using the other three methods. Furthermore, the samples collected using the conventional machine harvester and both bare and gloved hands had similar (p > 0.05) mean TC counts at the 9 am sampling points, and these counts were significantly lower than the mean TC count recovered from the samples harvested using the modified machine harvester prototype at the same time interval. For the noon samples, however, the mean TC count for the fruit harvested using the modified OTR machine harvester prototype was only significantly higher than for the two groups of handpicked fruits. The results in Table 4 demonstrate that the samples of both blueberry cultivars collected using the two machine harvesters had significantly higher (p < 0.05) mean TA counts than those harvested using bare and gloved hands, an observation similar to the results of the overall statistical analysis in Table 2. The 'Liberty' berries harvested using the conventional machine harvester and using bare and gloved hands had similar (p > 0.05) mean TC counts, which were significantly lower than the TC count for the samples harvested using the modified machine harvester prototype. No enterococci were detected in any of the collected samples, whereas seven fecal coliforms were recovered from the berry samples harvested using the modified OTR machine harvester prototype. The percentage of fecal coliform positive samples was 2.1%. Three out of the seven isolates were from the 9 am samples, and the other four isolates were from the noon samples. 4. Discussion 4.1. The Effect of Cultivars The 'Draper' blueberry is a mid-season cultivar known for its firm texture, whereas the 'Liberty' is a late-season cultivar with outstanding flavor and good shelf life . Different cultivars of fresh produce may have different intrinsic factors, such as the pH, water content, surface morphology, etc., which may affect the level and diversity of the microorganisms associated with fresh produce . However, the results of the current study revealed no difference (p > 0.05) in the microbial counts of the samples of the two cultivars (Table 2). One possible interpretation of this observation is that the overall phytochemical profiles of these two cultivars might be quite similar and had an indistinguishable influence on the microbial counts for the collected blueberry samples. Fresh produce grown in the field are susceptible to contamination by microorganisms from irrigation water, soil, manure fertilizers, insects, domestic and wild animals, and produce handlers . Extrinsic factors such as climate conditions, geographic locations, and agricultural practices (i.e., irrigation, spraying, fertilization, etc.) may influence the introduction of foodborne pathogens into fresh produce . For example, when irrigation water was inoculated with Salmonella, the furrow-irrigated cantaloupes (Cucumis melo subsp. melo) had a microbial load that was two to four times higher than the drip-irrigated ones . Given the fact that the two blueberry cultivars sampled in the current study were grown in adjacent areas on one farm, under similar climate conditions and management they could be loaded with similar levels of microorganisms sourced from the environment. 4.2. YM Counts in Sampled Blueberries Fresh fruit, such as blueberries, are rich in sugar and nutrients, have a high moisture content, and are ideal media for microorganisms to thrive on . The pH of the 'Draper' and 'Liberty' blueberries was estimated to be 3.4 and 3.5, respectively . A relatively low pH is favorable for the growth of acid-tolerant spoilage microorganisms such as yeasts, molds, and other fungi, which may cause blueberry rot and deterioration . The mean count of YM in the samples collected in the current study was ca. 3.80 log CFU/g (Table 2). Slightly higher levels of YM counts were reported in several other studies in the United States. For example, a study performed in Georgia recovered 4.49 log CFU/g of YM cells in blueberry samples collected in the dumping area of packing lines . The different observations between the two studies might be caused by multiple factors, such as the cultivar and climate differences, sample locations, and harvesting methods used. As stated previously, the Washington samples were collected from 'Draper' and 'Liberty' northern highbush blueberry plants, while the Georgia samples were collected from southern highbush (V. corymbosum unknown cvs.) and 'Rabbiteye' (V. virgatum unknown cvs.) blueberry plants. In addition to the cultivar and species variation, the blueberry samples from Georgia were collected at fresh blueberry packing facilities , while the ones from Washington were sampled in the field. Microorganisms can be introduced into blueberries through contaminated harvesting tools, containers, harvesters, gondolas, forklifts, or during the packing process . Furthermore, the mean YM counts for the Washington samples presented in Table 2 were the mean populations of both the hand-harvested samples in the field. The Georgia samples were, nevertheless, all collected by machine harvesters--no handpicked fruit were included. This may be another contributing factor to the observed differences. Other than these factors, the hot and humid weather condition in the southeast and the possible waiting time between harvest and packing during the busy harvesting season could also make the microbial load of harvest blueberries relatively higher . 4.3. The Effect of Harvesting Time The 9 am samples in the present study had significantly higher (p < 0.05) mean YM and TC counts than the 3 pm samples (Table 2). This could be explained by the temperature drops overnight and the water condensation on the surface of the fruit in the early morning hours of the day. These conditions may favor the survival and persistence of some microorganisms. Harvested fruit in berry lugs usually stay at the edge of the field for a period of time after harvest before being transported to refrigerated facilities. During this time, uncovered fruit may be exposed to solar radiation, which could be lethal to some microorganisms on the surface of the fruit . These factors could explain why the YM and TC counts for the samples collected in the afternoon were relatively lower than those for the samples collected in the morning. 4.4. The Effect of Harvesting Methods 4.4.1. Machine Harvesting The samples of blueberries harvested using the conventional machine harvester in the current study had mean TA, YM, and TC counts of 2.30, 3.87, and 0.31 log CFU/g, respectively, and the counts for the samples of blueberries harvested using the modified machine harvester prototype were not significantly different (p > 0.05) from these values, except for the mean TC counts (Table 2). The samples of machine-harvested blueberries collected from the fields in other states in the United States seem to have relatively higher microbial counts according to the results of several previous studies. Popa et al. reported mean TA, yeast, mold, and TC counts of 4.03, 4.32, 4.57, and 1.12 log CFU/g, respectively, for machine-harvested blueberries from fields in Michigan. The TA, yeast, and mold counts for the blueberries collected from lowbush blueberry (V. angustifolium) fields in Maine were reported to be within the range of 3.22 to 3.51, 2.27 to 2.72, and 2.46 to 3.35 log CFU/g, respectively . The sampled Michigan blueberries were from 'Bluecrop' northern highbush blueberry, while the Maine blueberries were harvested from unnamed, 'wild' lowbush blueberry stands. These crops were grown under different environmental conditions and under different management compared to the fruit sampled in the present study. The production and postharvest handling practices likely varied as well, especially for the lowbush blueberry. The differences in the microbial counts for the fresh berries observed in the previous studies could be attributed to these factors. Bruising is one of the biggest challenges associated with mechanically harvested blueberries. Studies have shown that the modified harvester prototype with soft catching surfaces usually maintains better firmness of harvested fruit compared to standard harvesters with hard surfaces, which ultimately enhances the shelf life . It is encouraging to see that the TA counts for the samples harvested using the two different kinds of machines were not significantly different (p > 0.05; Table 2). However, the samples harvested using the modified harvester prototype did have significantly higher (p < 0.05) TC counts and a relatively more frequent presence of fecal coliforms. A higher level of coliforms can be an indication of unsanitary conditions, potentially related to contamination by polluted water, uncleaned surfaces, and employees with poor personal hygiene. Thus, preventative measures should be encouraged when the modified machine harvester is used to harvest fresh market fruit. 4.4.2. Hand vs. Machine Harvesting Regardless of the type of harvester, the microbial loads of the blueberries harvested using the machine harvesters were significantly higher (p < 0.05) than those of the blueberries harvested by hand (Table 2, Table 3 and Table 4). The most probable explanation is that machine harvesters may cause more internal bruises and external injuries to blueberries during harvesting, which accelerate the penetration and invasion of microorganisms into the fruit. A previous study observed higher incidences of plant diseases in machine-harvested blueberries during postharvest storage . Blueberries can also be contaminated by microorganisms on the harvester surfaces if it is not appropriately cleaned or sanitized. The machine harvesters used in this study were cleaned and sanitized once a day in the late afternoon with water emitted from a high-pressure sprayer (Lisa Wasko DeVetter, personal communication). In comparison, the hand pickers were required to wash their hands whenever they started or returned to harvesting activities. The blueberry contact surfaces on the machine harvesters were thus not frequently cleaned nor sanitized, and this is a probable source of microbial contamination for the harvested blueberries. Although the hand-harvested blueberries had relatively lower microbial loads, harvesting blueberries by hand may also project the risk of introducing human pathogens via the direct contact of the berries with contaminated hands. For example, Staphylococcus aureus is a typical type of pathogen that is usually carried by food handlers . In the current study, fecal coliforms were found in some collected samples. Harvesting using gloved hands served as a control in this study. The other method of hand harvesting was done by following good agricultural, as well as appropriate hand washing, practices; therefore, the fruit harvested using this method had a similar microbial quality as the control. On machine harvesters, graders may also come into contact with fruit as they move along the inspection belt before falling into a lug. Graders and anyone who comes into direct contact with fruit pose the same risks as hand harvesters and should likewise follow good agricultural and handwashing practices. These results reemphasized the importance of having handwashing facilities in blueberry fields during the harvest seasons. 5. Conclusions This study found that fruit harvested using the modified machine harvester prototype had similar (p > 0.05) TA and YM counts to fruit harvested using the conventional harvester, although the TC counts were higher for the fruit harvested using the modified harvester. However, the berry samples harvested using the two types of machine harvesters had significantly higher (p < 0.05) microbial loads than the handpicked samples, which emphasizes the importance of the routine cleaning and sanitation of machine harvesters. This research will benefit fresh berry and, perhaps, other fresh fruit producers. Acknowledgments The authors wish to sincerely thank Kathryn Vanweerdhuizen, Brian Foote, and Scott Korthuis from Oxbo International Inc. and the blueberry growers in Washington State for participating in the project. The authors would also like to acknowledge Weifan Wu and Seulgi Lee from the University of Georgia for their assistance during the sample processing. Author Contributions Conceptualization, J.C., L.D. and F.T.; methodology, J.C., P.W. and M.H.; formal analysis, P.W. and M.H.; data curation, P.W., M.H. and Y.C.; original draft preparation, P.W. and M.H.; writing--review and editing, J.C., L.D. and F.T.; supervision, J.C.; project administration, J.C.; funding acquisition, L.D., J.C. and F.T. All authors have read and agreed to the published version of the manuscript. Data Availability Statement Data will be provided upon reasonable request. Conflicts of Interest The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. foods-12-01047-t001_Table 1 Table 1 Results of the type III tests for fixed effects by the statistical model of blueberry harvesting (a = 0.05). Effect DF Type III SS Mean Square F Value Pr > F TA Cultivar 1 0.79 0.79 1.28 0.2589 Sampling time 2 1.29 0.64 1.04 0.3558 Harvest method 3 33.93 11.31 18.21 <0.0001 Cultivar*time 2 0.86 0.43 0.69 0.5007 Cultivar*method 3 0.57 0.19 0.31 0.8216 Time*method 12 7.30 0.61 0.98 0.4679 YM Cultivar 1 0.02 0.02 0.16 0.6914 Sampling time 2 2.34 1.17 7.81 0.0005 Harvest method 3 0.37 0.12 0.83 0.4775 Cultivar*time 2 0.58 0.29 1.94 0.1457 Cultivar*method 3 0.55 0.18 1.22 0.3012 Time*method 12 1.65 0.14 0.92 0.5301 TC Cultivar 1 0.68 0.68 1.06 0.3038 Sampling time 2 3.47 1.74 2.70 0.0690 Harvest method 3 8.63 2.88 4.47 0.0043 Cultivar*time 2 0.70 0.35 0.55 0.5800 Cultivar*method 3 2.90 0.97 1.50 0.2137 Time*method 12 11.00 0.92 1.42 0.1541 TA: total aerobes; YM: total yeasts and molds; TC: total coliforms. DF: Degree of freedom. Pr > F: p value, which reflects the significance of the effect. A value smaller than 0.05 is a significant effect; *: Interaction between the two variables. foods-12-01047-t002_Table 2 Table 2 Overall mean microbial load on the sampled fresh blueberries of various cultivars harvested using various methods at different sampling time points. Total Aerobes Total Yeast and Molds Total Coliforms log CFU/g Cultivar Draper (n = 192) 2.03 +- 1.02 A 3.87 +- 0.47 A 0.40 +- 0.90 A Liberty (n = 144) 2.10 +- 0.55 A 3.85 +- 0.26 A 0.36 +- 0.73 A Sampling time 9 am (n = 120) 2.10 +- 0.83 A 3.96 +- 0.36 A 0.46 +- 0.85 A 12 pm (n = 120) 2.12 +- 0.85 A 3.86 +- 0.43 B 0.41 +- 0.82 AB 3 pm (n = 96) 1.95 +- 0.87 A 3.73 +- 0.35 B 0.25 +- 0.78 B Harvesting method Modified OTR prototype machine harvester (n = 88) 2.49 +- 0.48 A 3.90 +- 0.39 A 0.68 +- 0.93 A OTR machine harvester (n = 72) 2.30 +- 0.52 A 3.87 +- 0.31 A 0.31 +- 0.75 B Ungloved sanitized hands (n = 88) 1.81 +- 1.00 B 3.80 +- 0.37 A 0.29 +- 0.81 B Hands with sterile gloves (n = 88) 1.70 +- 0.94 B 3.86 +- 0.48 A 0.24 +- 0.73 B Means of each experiment variable followed by different letters in the same column are significantly different (p < 0.05). foods-12-01047-t003_Table 3 Table 3 Microbial load on sampled fresh blueberries harvested using different methods at individual sampling time points. Modified OTR Machine Harvester Prototype (n = 88) Standard OTR Machine Harvester (n = 72) Ungloved Sanitized Hands (n = 88) Hands with Sterile Gloves (n = 88) log CFU/g Total aerobes 9 am (n = 120) 2.65 +- 0.50 a 2.37 +- 0.56 a 1.66 +- 0.87 b 1.80 +- 0.83 b 12 pm (n = 120) 2.43 +- 0.34 a 2.29 +- 0.30 a 2.04 +- 1.08 ab 1.74 +- 1.03 c 3 pm (n = 96) 2.34 +- 0.57 a 2.24 +- 0.64 a 1.71 +- 0.98 b 1.51 +- 0.92 b Total yeasts and molds 9 am (n = 120) 4.03 +- 0.34 A 3.98 +- 0.33 A 3.88 +- 0.30 A 3.98 +- 0.46 A 12 pm (n = 120) 3.89 +- 0.47 AB 3.90 +- 0.29 A 3.79 +- 0.43 AB 3.88 +- 0.47 AB 3 pm (n = 96) 3.76 +- 0.32 B 3.77 +- 0.27 A 3.69 +- 0.34 B 3.69 +- 0.45 B Total coliforms 9 am (n = 120) 0.91 +- 1.04 a 0.33 +- 0.66 b 0.31 +- 0.74 b 0.27 +- 0.78 b 12 pm (n = 120) 0.76 +- 0.89 a 0.44 +- 0.43 ab 0.25 +- 0.96 b 0.13 +- 0.69 b 3 pm (n = 96) 0.26 +- 0.73 a 0.41 +- 1.02 a 0.14 +- 0.64 a 0.14 +- 0.66 a According to the results in Table 1, the harvest method is a significant factor influencing only the total aerobic and coliform counts, and the sampling time is a significant factor influencing only the yeast and mold counts. Values of the variables followed by different uppercase letters in each column are significantly different (p < 0.05). Values followed by different lowercase letters in each row are significantly different (p < 0.05). foods-12-01047-t004_Table 4 Table 4 Microbial loads on sampled fresh blueberries of different cultivars as affected by the harvesting methods. Modified OTR Prototype Machine Harvester (n = 88) Standard OTR Machine Harvester (n = 72) Ungloved Sanitized Hands (n = 88) Hands with Sterile Gloves (n = 88) log CFU/g Total aerobes Draper (n = 192) 2.50 +- 0.53 a 2.25 +- 0.60 a 1.74 +- 1.28 b 1.64 +- 1.16 b Liberty (n = 144) 2.47 +- 0.42 a 2.41 +- 0.31 a 1.89 +- 0.45 b 1.76 +- 0.57 b Total coliforms Draper (n = 192) 0.54 +- 0.98 a 0.33 +- 0.84 a 0.35 +- 0.88 a 0.39 +- 0.87 a Liberty (n = 144) 0.85 +- 0.84 a 0.20 +- 0.55 b 0.26 +- 0.72 b 0.07 +- 0.43 b According to the results in Table 1, the harvesting method is a significant factor influencing only the total aerobic and coliform counts. Values in the same row followed by the same letters are significantly different (p < 0.05). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Kalt W. Cassidy A. Howard L.R. Krikorian R. Stull A.J. Tremblay F. Zamora-Ros R. Recent research on the health benefits of blueberries and their anthocyanins Adv. Nutr. 2020 11 224 236 10.1093/advances/nmz065 31329250 2. Szajdek A. Borowska E.J. Bioactive compounds and health-promoting properties of berry fruit: A review Plant Foods Hum. Nutr. 2008 63 147 156 10.1007/s11130-008-0097-5 18931913 3. Food and Agriculture Organization of the United Nations Production: Crops: Blueberry Available online: (accessed on 18 July 2022) 4. National Agricultural Statistics Service (NASS) United States Department of Agriculture (USDA) Noncitrus Fruit and Nuts 2020 Summary Available online: (accessed on 18 July 2022) 5. DeVetter L.W. Yang W.Q. Takeda F. Korthuis S. Li C. Modified over-the-row machine harvesters to improve northern highbush blueberry fresh fruit quality Agriculture 2019 9 13 10.3390/agriculture9010013 6. Takeda F. Yang W.Q. Li C. Freivalds A. Sung K. Xu R. Hu B. Williamson J. Sargent S. Applying new technologies to transform blueberry harvesting Agronomy 2017 7 33 10.3390/agronomy7020033 7. Li C. Yu P. Takeda F. Krewer G. A miniature instrumented sphere to understand impacts created by mechanical blueberry harvesters HortTechnology 2013 23 425 429 10.21273/HORTTECH.23.4.425 8. Brown G.K. Schulte N.I. Timm E.J. Beaudry R.M. Peterson D.L. Hancock J.F. Takeda F. Estimates of mechanization effects on fresh blueberry quality Appl. Eng. Agric. 1996 12 21 26 10.13031/2013.25435 9. Sargent S.A. Takeda F. Williamson J.G. Berry A.D. Harvest of southern highbush blueberry with a modified, over-the-row mechanical harvester: Use of soft-catch surfaces to minimize impact bruising Agronomy 2021 11 1412 10.3390/agronomy11071412 10. Takeda F. Krewer G. Andrews E.L. Mullinix B. Peterson D.L. Assessment of the V45 blueberry harvester on rabbiteye blueberry and southern highbush blueberry pruned to V-shaped canopy HortTechnology 2008 18 130 138 10.21273/HORTTECH.18.1.130 11. Peterson D.L. Wolford S.D. Timm E.J. Takeda F. Fresh market quality blueberry harvester Trans. ASABE 1997 40 535 540 10.13031/2013.21298 12. Gallardo K. Lu L. Zilberman D. Jung A.R. Adoption of mechanization solutions for harvesting fresh market blueberries Proceedings of the Agricultural and Applied Economics Association Conferences Annual Meeting Atlanta, GA, USA 21-23 July 2019 Available online: (accessed on 18 July 2022) 13. Cai Y. Takeda F. Foote B. DeVetter L.W. Effects of machine-harvest interval on fruit quality of fresh market northern highbush blueberry Horticulturae 2021 7 245 10.3390/horticulturae7080245 14. Sargent S.A. Takeda F. Williamson J.G. Berry A.D. Harvest of southern highbush blueberry with a modified, over-the-row mechanical harvester: Use of handheld shakers and soft catch surfaces Agriculture 2020 10 4 10.3390/agriculture10010004 15. Hancock J.F. Blueberry Plant Denominated 'Liberty' U.S. Patent Application 10/350,343 2004 16. Hancock J.F. Blueberry Plant Denominated 'Draper' U.S. Patent Application 10/350,344 2004 17. Mostafidi M. Sanjabi M.R. Shirkhan F. Zahedi M.T. A review of recent trends in the development of the microbial safety of fruits and vegetables Trends Food Sci.Technol. 2020 103 321 332 10.1016/j.tifs.2020.07.009 18. Di Vittori L. Mazzoni L. Battino M. Mezzetti B. Pre-harvest factors influencing the quality of berries Sci. Hortic. 2018 233 310 322 10.1016/j.scienta.2018.01.058 19. Alegbeleye O.O. Singleton I. Sant'Ana A.S. Sources and contamination routes of microbial pathogens to fresh produce during field cultivation: A review Food Microbiol. 2018 73 177 208 10.1016/j.fm.2018.01.003 29526204 20. Rodrigues C. da Silva A.L.B.R. Dunn L.L. Factors impacting the prevalence of foodborne pathogens in agricultural water sources in the southeastern United States Water 2019 12 51 10.3390/w12010051 21. Reque P.M. Steffens R.S. Silva A.M.D. Jablonski A. Flores S.H. Rios A.D.O. Jong E.V.D. Characterization of blueberry fruits (Vaccinium spp.) and derived products Food Sci. Technol. 2014 34 773 779 10.1590/1678-457X.6470 22. Gunduz K. Serce S. Hancock J.F. Variation among highbush and rabbiteye cultivars of blueberry for fruit quality and phytochemical characteristics J. Food Compos. Anal. 2015 38 69 79 10.1016/j.jfca.2014.09.007 23. Perez-Lavalle L. Carrasco E. Valero A. Strategies for microbial decontamination of fresh blueberries and derived products Foods 2020 9 1558 10.3390/foods9111558 33126448 24. Hintz T. Matthews K.K. Di R. The use of plant antimicrobial compounds for food preservation BioMed Res. Int. 2015 2015 246264 10.1155/2015/246264 26539472 25. Quansah J.K. Gazula H. Holland R. Scherm H. Li C. Takeda F. Chen J. Microbial quality of blueberries for the fresh market Food Control 2019 100 92 96 10.1016/j.foodcont.2018.12.034 26. Machado-Moreira B. Richards K. Brennan F. Abram F. Burgess C.M. Microbial contamination of fresh produce: What, where, and how? Compr. Rev. Food Sci. Food Saf. 2019 18 1727 1750 10.1111/1541-4337.12487 33336968 27. Brandl M.T. Mandrell R.E. Fitness of Salmonella enterica serovar Thompson in the cilantro phyllosphere Appl. Environ. Microbiol. 2002 68 3614 3621 10.1128/AEM.68.7.3614-3621.2002 12089050 28. Lopez-Galvez F. Gil M.I. Allende A. Impact of relative humidity, inoculum carrier and size, and native microbiota on Salmonella ser. Typhimurium survival in baby lettuce Food Microbiol. 2018 70 155 161 10.1016/j.fm.2017.09.014 29173622 29. Dreux N. Albagnac C. Federighi M. Carlin F. Morris C.E. Nguyen C. Viable but non-culturable Listeria monocytogenes on parsley leaves and absence of recovery to a culturable state J. Appl. Microbiol. 2007 103 1272 1281 10.1111/j.1365-2672.2007.03351.x 17897231 30. Ongeng D. Geeraerd A.H. Springael D. Ryckeboer J. Muyanja C. Mauriello G. Fate of Escherichia coli O157: H7 and Salmonella enterica in the manure-amended soil-plant ecosystem of fresh vegetable crops: A review Crit. Rev. Microbiol. 2015 41 273 294 10.3109/1040841X.2013.829415 24083946 31. Popa I. Hanson E.J. Todd E.C. Schilder A.C. Ryser E.T. Efficacy of chlorine dioxide gas sachets for enhancing the microbiological quality and safety of blueberries J. Food Prot. 2007 70 2084 2088 10.4315/0362-028X-70.9.2084 17900086 32. Crowe K.M. Bushway A.A. Bushway R.J. Effect of alternative postharvest treatment on the microbial quality of lowbush blueberries Small Fruit Rev. 2005 4 29 39 10.1300/J301v04n03_03 33. Mehra L.K. MacLean D.D. Savelle A.T. Scherm H. Postharvest disease development on southern highbush blueberry fruit in relation to berry flesh type and harvest method Plant Dis. 2013 97 213 221 10.1094/PDIS-03-12-0307-RE 30722315 34. Cappai F. Benevenuto J. Ferrao L.F.V. Munoz P. Molecular and genetic bases of fruit firmness variation in blueberry--A review Agronomy 2018 8 174 10.3390/agronomy8090174 35. Becker R.E.N. Wardenburg B.J. Staphylococcus aureus and the skin: A longstanding and complex interaction Skinmed 2015 13 111 120 26137737 |
PMC10000652 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12051093 foods-12-01093 Article Farmer vs. Industrial Practices: Impact of Variety, Cropping System and Process on the Quality of Durum Wheat Grains and Final Products Samson Marie-Francoise Methodology Validation Formal analysis Investigation Data curation Writing - original draft Writing - review & editing 1 Boury-Esnault Anais Formal analysis Data curation Writing - original draft 1 Menguy Ewen Formal analysis Resources 2 Avit Valentin Data curation Writing - original draft 2 Canaguier Elodie Formal analysis Resources 1 Bernazeau Bruno Methodology Resources 2 Lavene Patrice Resources 2 Chiffoleau Yuna Methodology Visualization Funding acquisition 3 Akermann Gregori Investigation 3 Moinet Kristel Conceptualization Methodology Resources Supervision Project administration Funding acquisition 4 Desclaux Dominique Conceptualization Methodology Validation Formal analysis Investigation Resources Data curation Writing - original draft Writing - review & editing Visualization Supervision Project administration Funding acquisition 2* Sissons Mike Academic Editor 1 IATE, Univ Montpellier, INRAE, Institut Agro, 34060 Montpellier, France 2 INRAE, UE DiaScope, UE 0398, 34130 Mauguio, France 3 INNOVATION, Univ Montpellier, CIRAD, INRAE, Institut Agro, 34060 Montpellier, France 4 Biocivam 11, 11800 Trebes, France * Correspondence: [email protected] 03 3 2023 3 2023 12 5 109317 1 2023 24 2 2023 27 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). The consumption of artisanal and organic pasta made on-farm from ancient varieties is increasing in France. Some people, namely, those suffering from digestive disorders following the consumption of industrial pasta, consider these artisanal pasta to be more digestible. Most of them have linked these digestive disorders to the ingestion of gluten. We analyzed in this study the impact of industrial and artisanal practices on the protein quality of durum wheat products. The varieties recommended by the industry (IND) were compared to those used by farmers (FAR): the FAR being on average much richer in protein. However, the solubility of these proteins analyzed by Size Exclusion-High Performance Liquid Chromatography (SE-HPLC) and their in vitro proteolysis by digestive enzymes vary little between the two groups of varieties, while differences between varieties in each group are observable. The location of grain production and the tested cropping systems (zero vs. low input) have a low impact on protein quality. Yet, more contrasting modalities should be studied to validate this point. The type of production process (artisanal vs. industrial) is, among those studied, the factor having the greatest impact on protein compositionPasta produced by the artisanal method contains a higher sodium dodecyl sulfate (SDS)-soluble protein fraction and are more in-vitro proteolyzed. Whether these criteria are indicative of what happens during a consumer's digestion remains to be determined. It also remains to be assessed which key stages of the process have the greatest influence on protein quality. participatory research ancient variety gluten farming system pasta making process protein digestibility semolina Fondation de France00081238 FEADER and Region OccitanieFEADER 1.2.2020 This research was funded by Fondation de France, (project : Gluten : Mythe ou Realite?") grant number 00081238 and by FEADER and Region Occitanie (project Activa-Ble), FEADER 1.2.2020. pmc1. Introduction Durum wheat, Triticum turgidum L. ssp. Durum, is one of the few cereals intended exclusively for human consumption. It is generally transformed into semolina and then into pasta or couscous. Twenty years of participatory research carried out in the south of France with partners from farming, artisanal and industrial sectors have highlighted differences in practices between these different operators . These differences are to be found not only in the choice of durum wheat varieties and the agronomic practices, but also in pasta processing. Farmers and craftsmen use mainly ancient (heritage) durum wheat varieties or those resulting from participatory breeding programs and grow them under organic conditions. They transform their harvest into semolina on a stone-mill and produce pasta under slow conditions by avoiding high-temperature drying. On the other hand, industrialists use mainly modern (elite) durum wheat varieties, grown in a conventional mode, ground on roller-mills and transformed into pasta under controlled conditions with high or very high temperature drying step . Durum wheat grains contain about 14% protein . Wheat proteins are classified according to their solubility : albumins (soluble in water), globulins (soluble in saline solutions), gliadins (soluble in dilute alcohols) and glutenins (soluble in dilute acids and bases). Gliadins and glutenins are storage proteins. Gliadins are monomeric proteins, grouped into four classes: o-, a-, g-gliadins. Glutenins appear in the form of polymers composed of two types of subunits: low-molecular-weight glutenin subunits (LMW-GS) and high-molecular-weight glutenin subunits (HMW-GS). During the transformation processes, under the effect of hydration and energy supply, these polymers rearrange, disulfide bonds disrupt and new bonds are created . Gliadins interact with this glutenin skeleton by non-covalent bonds and together form gluten. Gliadins contribute to the viscosity and extensibility of gluten while glutenins are the major element of elasticity . There is a current craze for local varieties or so-called "old" or heritage varieties. One of the reasons for this enthusiasm is the association made between local, regional products and attributes such as fresh, tasty, nutritional value, healthy and safe . The selection of durum wheat, during the second half of the 20th century, resulted in an increase in yield and resistance to diseases, a decrease in protein content and a change in protein composition with the introduction of glutenin alleles favorable to pasta quality (i.e., HMW 7 + 8. LMW-2) but also by a reduction in genetic diversity . The proportions of the different protein classes have also changed; some authors indicate that modern genotypes have a higher glutenin to gliadin ratio (Glu/Glia) than that of old varieties and that the expression of type B LMW subunits is twice as high in the former. At the same time, they noted no significant differences in the expression of a-, g-gliadins between the two groups, but they mentioned a significant decrease in o-gliadins. In another article, De Santis et al. reported a content of unextractable polymeric protein (UPP) twice as high for the modern variety Saragolla as for the old variety Cappelli and also a lower expression of o-gliadins. Several studies comparing modern wheats and ancient wheats have attempted to link the harmful effects of gluten to the year of creation and the change in protein composition. It usually appears that old varieties are no less toxic than recent varieties . However, the study by Ianiro et al. showed that patients with NCGS (non-coeliac gluten sensitivity) expressed fewer gastrointestinal and extra-intestinal symptoms after ingesting pasta made with an "old" variety of durum wheat (cv.Senatore Cappelli, date of release = 1923) compared to ingesting pasta from a mixture of recent varieties. Some people suffering from digestive disorders and declaring themselves NCGS (non-coeliac gluten sensitive), questioned during a specific survey or during personal discussions, declare that they can consume wheat-based farm products without experiencing the inconveniences they usually feel when eating industrial semolina or pasta. From an agronomic point of view, the effects of nitrogen fertilization on the storage protein composition of cereals are well documented. A meta-analysis carried out over the period 1960-2019 highlighted an increase in nitrogen fertilization from 9.8 to 93.8 kg N. ha-1. y-1. At the same time, this increase in fertilization resulted in a higher gluten content, with more gliadins. Over the period, the authors estimated the increase in the intake of gliadins in the ration, going from 2.4 to 3.8 kg.y-1, representing a jump of +58%. According to these same authors, this increase would be positively correlated with that of celiac disease. The increase in nitrogen fertilization resulted in an increase in the proportion of monomeric gliadins and, on the other hand, decreased the proportion of large glutenin polymers . This result is supported by the data of Wan et al. who indicated that increasing the level of nitrogen fertilization led to an increase in o-gliadin expression with increased proportions of o5 gliadin. With regard to the processes, numerous works concerning their impact on the components of the finished product, in particular, proteins, have been carried out. The industrial process for making pasta differs from traditional methods by extensive refining of the semolina, control over its hydration, by controlling temperatures and pressures throughout the process and finally by the characteristics of drying (short duration and high or even ultra-high temperature). A high protein content (12 to 15% depending on the country) is required in the industry to produce semolina and give pasta good resistance to cooking, the absence of stickiness and optimal firmness . When the protein content is lower, it can be compensated by strong gluten . The improvement in the final quality of pasta has also been achieved through technological progress with the introduction of drying at high, or even very high temperature . Drying at high temperature and especially very high temperature has made it possible to significantly improve pasta quality, in particular, when using medium-quality wheat . At very high temperatures, the gluten network formed by gliadins and glutenins stiffens under the effect of heat via disulfide bonds, which results, on the biochemical level, in protein denaturation, reinforcement of the gluten around starch granules and a decrease in protein solubility in many solvents such as acetic acid . Protein digestibility is estimated to be over 90% in rich countries . Food processes involving altering or modifying the supramolecular and molecular structures make them more accessible, or not accessible, to digestive enzymes . Gluten proteins may be resistant to enzymatic hydrolysis due to the presence of proline-rich regions, and the process, depending on its intensity, may reduce digestibility , similarly to the presence of other compounds such as fiber . There are few studies on the digestibility of proteins as a function of parameters such as the date of registration of the variety or the cropping system. Concerning soft wheat, Gulati et al. note an increase in digestibility linked to the decrease in the protein content of "modern" wheat. Ma and Baik , after analyzing different species of wheat (soft wheat, durum wheat, einkorn, spelt, emmer), show that proteins that do not constitute gluten (albumins, globulins) affect protein digestibility. This article is the result of a participatory research work involving farmers, actors in agricultural advice and the artisanal and industrial sectors, doctors, researchers and trainers from various disciplines, consumers and people diagnosed with NCGS. Their collaboration aimed to understand the combined impact of varieties, cropping systems, and processes on the quantity, quality and in vitro digestibility of gluten proteins in based-durum wheat products. They compared the protein quality of farmhouse and industrial semolina and pasta. 2. Materials and Methods 2.1. Grain Samples Twenty varieties of durum wheat (Triticum t. durum) were grown and harvested at the DiaScope experimental station (INRAE, Mauguio, France 43deg7'14'' N, 35deg9'03'' E) in 2015-2016 (H2016 = harvest in 2016), 2016-2017 (H2017) and 2018-2019 (H2019) (No trial was conducted in 2017-2018). Among these varieties, eight categorized as IND have been chosen from the list of varieties recommended by the SIFPAF (Union of Industrial Pasta Manufacturers of France) and the CFSI (French Committee for Industrial Semolina); six varieties categorized as FAR are used by farmers and dedicated to the artisanal sectors of the Occitanie region; six other varieties, identified as DIV, are currently underutilized or no-longer-used genetic resources (Table 1). The agronomic device was a randomized split plot with 3 replications implemented only under zero input (ZI) conditions in H2016 (previous crop: chickpea) and H2019 (previous alfalfa) and in two cropping systems in H2017: ZI (previous and antecedent Sainfoin, nitrogen residue: 80 units, no fertilizer, herbicide or pesticide input) and LI (low inputs) (previous chickpea, nitrogen residue: 80 units, supply of 150 units of Nitrogen + application of 2 herbicides). In H2019, varieties were also grown at the Purpan Engineering School (Toulouse, France 1deg24'06'' E, 43deg35'43'' N) under zero input conditions. At harvest, the grains were cleaned to remove all foreign elements and stored in a cold room. 2.2. Criteria Acquired on the Whole Grain The thousand-kernel weight (TKW), the specific weight (SW) and the total protein content were measured on each sample of whole grain, by near-infrared spectroscopy (NIRS) using the calibrations developed by Foss (2017) and integrated in the Infratec Nova spectrometer (Foss, Hillerod, Denmark) located at INRAE-Mauguio. The principle behind NIRS is that specific organic molecules absorb specific wavelengths of near infrared light energy. The absorptions are directly correlated with the concentration of the organic molecules in the sample. 2.3. Milling Wheat Grains and Making Pasta Three processes were carried out to produce semolina and pasta: "Artisanal 1", "Artisanal 2" and "Industrial". Artisanal 1 process: ten varieties, harvested in 2016, were selected among the 20 varieties listed in Table 1 to make pasta on a farm in Bezouce (30320, France) by a farmer (Francois C.) who is used to producing durum wheat, grinding it, and making pasta on his own farm using an artisanal process. For the "Artisanal 1" process, the farmer has ground 5 kg of grain of each variety with his Tyrol-type stone-mill (SM 1). Then, he made tagliatelle on his P3-type pasta press (La Monferrina, Moncalieri, Italy), which has a mixing capacity of 3 kg and is capable of producing 8 to 10 kg.h-1. The pasta was then air dried (Table 2). Artisanal 2 process: based on 2017 harvest, only two varieties of durum wheat belonging to "IND" group (cv. Anvergur, cv. Claudio) and two others belonging to "FAR" group (cv. LA1823 and cv. Bidi17) were used. The semolina was produced on a Tyrol-type stone-mill (SM 2) at the Lycee Agricole d'Auzeville (Castanet-Tolosan, France). Subsequently, part of the semolina was transformed into artisanal pasta at the Lycee Agricole d'Auzeville on a pasta press Mac 30 (Italpast, Parma, Italy) equipped with a tagliatelle bronze die. The pasta was then dried on racks in a ventilated chamber at 37 degC. Industrial process: the other part of the semolina (SM 2) was used to produce "industrial"-type pasta at JRU IATE (INRAE Montpellier, France) using an experimental mini-press (Sercom, Montpellier, France) from 700 g of semolina hydrated at 47% (dry basis). After hydration, the semolina was mixed for 20 min to form agglomerates which were then extruded under vacuum through a pappardelle die (width 33 mm, thickness approx. 0.6 mm), coated with Teflon and maintained at 40 degC. Pasta was then dried at very high temperature (90 degC) in a drying chamber regulated by temperature and humidity (CS-40, CTS, Auriol, France) (Table 2). 2.4. Preparation of Samples and Physico-Chemical Analysis The grains and dry pasta were ground with a Cyclotec 1093 laboratory mill (Foss Tecator, Nanterre, France) for biochemical analysis. Pasta were cooked to their optimal cooking time (OCT) according to the AACC 66-50.01 standard , then drained and frozen before being freeze-dried and ground with an IKA A10 basic grinder (IKA, Staufen, Germany). The moisture content of ground products and semolina was determined according to the AACC 44-15.02 method , and their total protein content was determined according to the AACC 46-12.01 method with 5.7 as nitrogen-to-protein conversion factor. 2.5. Protein Extraction Procedures Proteins were extracted from lab-milled and stone-milled grains according to Morel et al. with some modifications. The sample (160 mg) was suspended in 20 mL of 0.1 M sodium phosphate buffer (pH 6.9) containing 1% (w/v) sodium dodecyl sulfate (SDS) and shaken for 80 min at 60 degC. After centrifugation (37,000x g, 30 min, 20 degC), the supernatant was stored (-20 degC) until the Size Exclusion-High Performance Liquid Chromatography (SE-HPLC) analysis and the pellet was re-suspended in 5 mL of SDS-phosphate buffer before being sonicated for 3 min at 7.5 watts and centrifuged under the same conditions. The new supernatant was kept until analysis. For heat-treated products (dry and OCT-cooked pasta), the first extraction step was the same as before. However, the pellet was re-suspended in sodium phosphate-SDS buffer containing 20 mM of dithioerythritol (DTE) and the content of the tube purged with argon. The suspension was then stirred for 1 h at 60 degC before being sonicated for 3 min at 7.5 watts and then centrifuged. The supernatant was kept until the SE-HPLC analysis. 2.6. Measurement of Protein Size Distribution by SE-HPLC The proteins obtained after the different extraction steps were separated by SE-HPLC using a TSKgel G4000 SWXL column (7.8 mm i.d. x 30 cm, TOSO BIOSCIENCE GmbH, Griesheim, Germany) according to Dachkevitch and Autran on an Alliance system (Waters, Saint Quentin en Yvelines, France). Proteins were eluted at room temperature with 0.1 M sodium phosphate buffer (pH 6.9) containing 0.1% (w/v) SDS at a flow rate of 0.7 mL. min-1, and the absorbance was recorded at 214 nm. The first chromatogram corresponds to the SDS-soluble proteins and can be divided into five fractions, from F1 to F5, according to Morel et al. . The F1 and F2 fractions include the large and small glutenin aggregates, respectively; F3 contains o-gliadins and high-molecular-weight albumins such as b-amylase; F4 includes a-, g-gliadins; and F5 contains albumins-globulins. The total area under the second chromatogram of the non-thermally treated products was calculated and named Fi, for insoluble fraction. The total protein content of the samples was estimated from the sum of the total areas of the two chromatograms corresponding to the same sample. Fractions F1 to F5 and Fi were expressed as a percentage of this sum. The proportion of unextractable polymeric proteins (UPP%) was calculated as follows: UPP% = (Fi x 100)/(F1 + F2 + Fi). In the case of heat-treated products, the first chromatogram was not cut into 5 fractions; only the total area was retained to obtain SDS-soluble proteins expressed in % of the total proteins. The proportion of DTE-soluble proteins (i.e., after DTE reduction and sonication) was calculated from the total area of the second chromatogram and also expressed in % of the total proteins. The non-extracted protein fraction was calculated by subtracting the sum of the SDS-soluble and DTE-soluble protein contents from the Kjeldahl total protein content. When the fraction of non-extracted protein was negative, due to high recovery, this value was forced to 0, and the sum of the SDS soluble and DTE extracted protein fractions was corrected to reach 100%. The gliadin to glutenin ratio was calculated as follow: Glia/glu = (F4/(F1 + F2 + Fi)). 2.7. In Vitro Protein Digestibility The digestibility of OCT cooked pasta samples was evaluated by measuring the rate of proteolysis in vitro using the Protein Digestibility Assay kit (Neogen, Auchincruive, UK), with a few modifications to the standard procedure. For each pasta sample, proteolysis was carried out on two 250 mg samples, according to manufacturer specifications and trypsin/chymotrypsin digestion was conducted for 3 h. Digestion was stopped by immersing the tubes in boiling water. After cooling, the tubes were centrifuged at 4696x g for 15 min at 15 degC. After centrifugation, the supernatants were set aside and the pellets frozen. Protein digestibility was then estimated by determining the amount of nitrogen remaining in the pellets from two extractions of the same sample using the Kjeldahl method. The extent of proteolysis after 1 h of peptic digestion followed by 3 h of tryptin/chymotryptin digestion was expressed as the percentage of the initial protein content of the sample that remained after digestion. 2.8. Statistical Processing ANOVA, Multivariate analyzes of variance (MANOVA) and Pillai trace tests with a 95% confidence interval were carried out in order to assess the significance of the impacts of the factors studied on all the variables measured, and rank the influence of these factors. The significance of the differences between modalities was tested by SNK (Student-Newman-Keuls) tests within a 95% and 99% confidence interval. Correlation analyses allowed performing existing links between the different variables. Both SAS Software (SAS 9.3 TS-Version windows 6.2.9200. SAS Institute Inc. SAS and all other SAS Institute Inc., Cary, NC, USA) and R packages were used to perform these analyses: readxl, xlsx, FactoMineR, Ggally, agricolae, tidyverse and rlist R Studio (version 4.0.2). 3. Results 3.1. Whole Grain Quality Criteria Three main criteria--thousand-kernel weight (TKW), specific weight (SW) and protein content (PROT)--usually considered in the case of durum wheat commercial contracts have been estimated on whole grains by NIRS. These estimates have been made on the whole grain samples of 4, 18 or 20 varieties according to the year. Five environments were considered: 3 years of harvest (2016, 2017, 2019) and, for some years, 2 farming systems (ZI and LI) or 2 different locations (Mauguio and Purpan). A classic ANOVA comparing the impact of the varieties and the impact of the environments on these data made it possible to define the contribution of each of the factors to the variability of the criteria (Table 3). As expected, TKW and SW were mainly heritable criteria (the variety effect explained 67% and 68% of the variability, respectively) while the environment played a less role (20 to 29% of the total variability), whereas the protein content was equally dependent on the variety (42%) and the environment (48%). 3.1.1. Thousand-Kernel Weight Across all years, the TKW of FAR varieties was significantly higher than that of IND and DIV varieties (Table 3). For instance, in H2016, the TKW varied from 37.9 g (cv. RG218) to 51.9 g (cv. RG1110). The varieties used by the farmers showed, on average, higher TKW (49.6 g) than those recommended by the industrial sector (40.8 g) and, the others underutilized genetic resources (42.6 g). Only four varieties were tested in 2019 (H2019) in two different locations (Mauguio and Purpan); their TKW varied, on average, from 47.5 g (cv. Anvergur) to 60.2 g (cv. LA1823). The TKW was higher in Purpan (mean = 56.6 g) than in Mauguio (mean = 51.4 g). When selecting the data for these 4 varieties from the others years, we noted also a higher and significant TKW for the two varieties used by the farmers ("FAR") than for the two used by the industrial sector ("IND"), regardless of the year and cropping conditions. 3.1.2. Specific Weight The specific weight is an essential data factor to be considered in case of commercial contracts. It varied on average, from 80.7 to 86.7 kg.hL-1 (Table 3). Only in 2017 was the usual threshold of 76 kg.hL-1 not reached for one variety (cv. Miradoux). Whatever the year of production and the cropping system or location, the specific weight of the varieties used by the farmers (overall mean FAR = 85.2 kg.hL-1) was always greater than that of the varieties recommended and used by the industrial sector (overall mean IND = 82.2 kg.hL-1). However, no significant difference was found with varieties classified in "DIV" (overall mean DIV = 82.6 kg.hL-1). The low input cropping system implemented in 2017 led to a higher specific weight (83.5 kg.hL-1) than the zero-input cropping system (80.7 kg.hL-1). In H2019, for the four studied varieties, the average specific weight varied between 84.4 kg.hL-1 (cv. Anvergur) and 87.0 kg.hL-1 (cv. LA1823). The specific weight was, on average, higher at Mauguio (86.7 kg.hL-1) compared to Purpan (84.7 kg.hL-1). 3.1.3. Protein Content The protein content of the grains harvested in 2016 varied from 7.5% db (cv. Karur) to 10.6% db (cv. RG218). The varieties used by farmers and artisans (FAR) contained significantly more proteins (mean = 9.7% db) than the varieties recommended by the industrial sector (IND) (8.6% db) (Table 3). On the 2017 harvest, protein levels were higher (14.1% db for FAR and 12.7% db for IND) than in H2016 despite the higher grain yield. There was no noticeable effect of cropping system (ZI or LI) on protein content. Concerning the grains harvested in 2019, at the Mauguio site, the protein level was lower than in H2017 (11.3% db vs. 13.6% db, respectively), when the same four varieties were considered. On the Purpan site, the protein content was slightly lower (10.2% db) than at Mauguio. Similar observations were noted in each environment: the varieties used by the farmers presented a higher protein content (general mean FAR = 12.5% db) than those used by the industrial sector (general mean IND = 11.0% db). This has to be linked with the level of the grain yield, which was significantly lower for FAR (2.4 t.ha-1) than for IND (3.2 t.ha-1) . 3.2. Whole Grain Protein Composition Samples of harvested grains (H2016, H2017, H2019) were ground in a laboratory mill, and the protein analysis was performed using the SE-HPLC method. Six protein fractions (F1 to F5 and Fi) were obtained after SE-HPLC separation (Table S1). The analysis of variance of these fractions allowed us to identify the main factors responsible of the variability of each of them (Table 4). Except for fraction F1, all the others fractions varied mainly according to the year. Then, the factor "variety" also contributed in large part to the variability of all the fractions but to varying degrees. When considering only the year 2017, the main part of the variability of the criteria was due to the varieties (Table 4). Variety largely explained the variance of the protein fractions (Table 4). The Cropping system effect and the Variety x Cropping system interaction help to explain F2 and Fi fractions. The ANOVA on fractions F1, F3 and F5 showed higher statistical residuals (around 20% of the explained variance, compared to less than 10% for the other fractions), meaning that others unexplained factors may have an impact on these fractions. As expected, the major part of the grain proteins was a-, g-gliadins (F4 fraction), followed by small glutenin aggregates proteins (F2) and albumins and globulins (F5) . 3.2.1. Year Effect Except for fraction F1, significant differences appeared for the others fractions between the 2 years: H2016 and H2017 . A significant increase was noted in the insoluble fraction (Fi) (from 6% in 2016 to 11% in 2017) and in F4 (from 29% to 36%). This compensated for a decrease in F2 (from 23% to 18%) and F5 (from 23% to 18%). It seems that the higher protein content measured in 2017 (+2.6%) was due to gliadins and unextractable proteins. To assess more deeply the year effect, Table 5 includes only the data concerning the zero-input cropping system (present the 3 years) and the data concerning the four varieties common to the 3 years. The year effect was very high for all the parameters and higher than the variety effect except for the F1 fraction. 3.2.2. Cropping System Effect When comparing the two cropping systems (zero input, ZI; low input, LI) implemented only in 2017 in Mauguio (Table 6), the Fi and F5 fractions were significantly higher in zero input (11.1% and 18.3%) than in low input (9.6% and 17.8%), respectively, while the percentages of large and small glutenin aggregates (F1 and F2) were significantly lower in zero input (6.9% and 18.5%) than in low input (7.3% and 19.3%). The others fractions (F3 and F4) were not significantly different. The contribution of the cropping system effect to the protein composition is mentioned Table 6. The cropping system had a significant effect on total protein content, F2, Fi and UPP and to a lesser extent on F1 and F5. The fractions F3 and F4 and the ratio gliadin to glutenin (Glia/Glu) were not impacted by the considered cropping system. 3.2.3. Location Effect Only the data of H2019 are available to assess the impact of the location (Mauguio vs. Purpan) on the protein quality parameters. Data are presented Table 6. The location had an important impact on protein content, then on F1, F3, F4 and F5 fractions. No effect was observed on F2, Fi and UPP. In these two locations, durum wheat was grown under zero-input condition. 3.2.4. Variety Effect Concerning the varieties, the major coefficients of variation were obtained for Fi and F1 fractions . When zooming on each of these protein fractions measured in 2016 , the percentage of Fi varied from 3.3% (cv. Oued Zenati) to 8.5% (cv. Claudio). Compared to the average of the 19 varieties, with a highest Fi content Claudio had, conversely, a lower proportion of soluble glutenin aggregates (F1 and F2, 6.2 and 20.6%, respectively) and of o-gliadins and high-molecular-weight albumins (F3). Other varieties, such as cv. Oued Zenati, were characterized by a higher proportion of F1 and F2 (6.8 and 23.3%) and a lower proportion of Fi (3.3%). The Vivit variety was distinguished by the lowest proportion of F4 (27.0%) and a significant amount of albumins and globulins (F5) (24.0%). 3.2.5. Comparison of Farmer Varieties (FAR) with Industry Recommended Varieties (IND) The comparison of the means of protein fractions measured on the varieties used by the farmers (FAR) and those recommended by the industry (IND) showed significant differences, except for F1 (Table 7). The higher protein level noted on the FARs seemed essentially due to an increase in the F2 and F4 fractions, whereas the fraction of insoluble proteins Fi was slightly higher for the IND varieties (Table 7). The same trend, i.e., significant higher F4 fractions and lower F5 and Fi fractions in FAR varieties compared to IND varieties, was systematically observed regardless of the environment . 3.3. Protein Composition of Stone-Milled Grains (Semolina) Measurements on semolina were carried out for four varieties harvested in 2017: Anvergur, Claudio (IND) and Bidi17, LA1823 (FAR). This semolina was produced at Lycee Agricole d'Auzeville on a Tyrol stone-mill (SM 2). The total protein content of stone-milled grains samples assessed by the Kjeldahl method was similar to those measured on whole grain at harvest by NIRS (except for cv. Anvergur and Claudio, which have a slightly higher--2 points--NIRS protein content) (Table S2). The protein composition of semolina obtained from stone-milled grains is presented in Table S2. The results are very similar to those measured on grains ground with a laboratory (lab-milled) . The F1, F2 and F4 fractions are very similar between the two measurements (lab-milled grains and stone-milled grains), they are correlated at 81%, 92% and 84%, respectively. However, the F3 fraction appears less well correlated (6%), and the correlations of the F5 and Fi fractions are lower (59% and 49%). When considering the whole data set, the correlation is high (r2 = 0.99). The contribution of each factor to the variability of each parameter analyzed in stone-milled semolina is mentioned in Table 8. Compared with the percentage obtained on lab-milled grains (Table 3), we noted a similar impact of each factor for the protein content, a greater impact here of the factor "variety" except for F3, F4 and F5, a lesser importance of the environment, except for F5, and a greater impact of the interaction Variety x Cropping system, notably for F3, F4 and F5. Regarding the type of variety (Table S2), for all the fractions, except F4, significant differences exist between varieties used by the farmers (FAR) and those recommended by the industry (IND). The latter have fractions F3, F5 and Fi higher than FAR varieties, and less medium and small glutenin aggregates (F1, F2). 3.4. Protein Composition of Dry Pasta The semolina (stone-milled grains) of the 4 varieties were used to make pasta, according to two types of process: "Industrial" and "Artisanal 2". The "Industrial" process differed from the "Artisanal 2" by 2 points: the extrusion was under vacuum, temperature controlled, and the drying of the pasta was performed at very high temperature (90 degC instead of 37 degC for the artisanal process) (Table 2). The protein composition of dry pasta was analyzed by SE-HPLC (Table S3). Because the extrusion and drying processes induced a decrease in solubility of some proteins, DTE was used to recover them, and the Fi fraction was renamed "DTE-soluble" fraction. The contribution of the varieties, cropping systems and process to the variability of the protein fractions in dry pasta is mentioned Table 9. For fractions F1 to F5, the process alone (Artisanal 2 vs. Industrial) explained between 88 and 98% of the variability. For fraction F5 (albumin and globulin), the interaction Variety x cropping system was significant. The unextractable proteins were depending on the process, on the variety and on their interactions. Even if the varieties explained very few of the variability of the protein fractions, when grouping on one hand the two varieties used by the farmers (FAR) (cv. LA1823, cv. BIDI17) and on the other hand, the two others recommended by the industry (IND) (cv. Anvergur and cv. Claudio), we noted significant differences mainly for the fractions of proteins soluble only in the DTE (34.6% for FAR and 37% for IND), and for the unextractable proteins (4% for FAR and 1.5% for IND). The impact of the process was major on the proportion of the protein fractions in the dry pasta. Whatever the cropping system (zero input, ZI, or low input, LI), the "Industrial" process led to a large increase in the DTE-soluble fraction (DTE-soluble represented 53.9% of the total protein in "Industrial" process vs. 19.3% in the "Artisanal 2" process) and to a decrease mostly in the F2 (26% of reduction) and F4 fractions . Under the zero input condition, the two varieties Bidi17 and LA1823, used mainly by the artisanal sector, showed a greater difference than the two other varieties (cv. Anvergur and cv. Claudio) for the F2 and DTE-soluble fractions. In low-input (LI) conditions, the difference was not significant for DTE-soluble fraction (Table S3). 3.5. Comparison between Lab-Milled Grains, Stone-Milled Grains (Semolina) and Dry Pasta (Artisanal 2 and Industrial Process) under the Cropping System Zero Input (ZI) The comparison between the distribution of the fractions measured in the semolina (stone-milled grains) and in the dry pasta produced according to the two processing modes (Artisanal 2 and Industrial) highlighted the impact of the mixing, extrusion and drying stages . The main differences were the important decrease in the F1 fraction, significantly reduced by the IND pasta process and the large increase in the DTE-soluble fraction. The other fractions varied little between artisanal dry pasta and semolina. When adding lab-milled grains to the comparison, only F1, F2 and F4 fractions were well correlated with those of stone-milled grains (Table 10). Surprisingly, F3, F5 and Fi (or DTE-soluble) were not correlated. No correlation was found between lab-milled grains and dry pasta fractions. 3.6. Protein Composition of Cooked Pasta Protein fractions were also analyzed by SE-HPLC on two different sets of pasta cooked at OCT. The first set of pasta was obtained from ten varieties of durum wheat chosen among the 20 harvested at the Mauguio site, in 2016 (Table 1), milled and processed by a farmer (Table 2 = Artisanal 1). The second set was composed of pasta produced with four varieties of durum wheat (Anvergur, Claudio, Bidi 17, LA1823), processed by three methods: "Industrial", "Artisanal 1" or "Artisanal 2" (Table 2). The results of the dry pasta of this second set were presented in the previous section. Because cooked pastas are heat-treated products, the first chromatogram obtained from each sample was not cut into five fractions; only the total area was retained to obtain SDS-soluble proteins expressed in % of the total proteins. The proportion of DTE-soluble proteins (i.e., after DTE reduction and sonication) was obtained from the total area of the second chromatogram and also expressed in % of the total proteins. The difference between the protein contents calculated from the total areas and the protein content measured by Kjeldahl allowed us to deduce the unextractable part of proteins (unextractable). It should be noticed that, with varieties processed according to the "Artisanal 1" process, no unextractable fraction was recovered (Tables S4 and S5). 3.6.1. First Set: Proteins Composition of the Cooked Artisanal Pasta Made from 10 Varieties of Durum Wheat The 10 samples of cooked pasta, made from the grains of the varieties harvested in 2016, showed significant differences in the fraction of soluble proteins in the SDS: from 34.3% (cv. Karur) to 48% (cv. Pescadou) and conversely for DTE-soluble proteins . There was no correlation between protein content and the others parameters (Tables S4 and S5). Although there was, in this set, an imbalance between the number of varieties recommended by the industry (eight varieties) and those used by the farmers (two varieties), we did not observe differences between these two types on the protein fractions. However, the difference regarding the protein content remained (Table S4). 3.6.2. Second Set: Proteins Composition of the Cooked Pasta Coming from Four Varieties of Durum Wheat Processed with Two Methods: "Industrial" and "Artisanal", and In Vitro Proteolysis Data Four varieties, among which were cv. Anvergur and cv. Claudio, recommended by the industry, and cv. LA1823 and cv. Bidi17, used by the farmers, were harvested in 2017, in the Mauguio site, then stone-milled and then processed according to two types of methods: "Industrial" and "Artisanal 2". The "Industrial" process differed from the "Artisanal 2" process in two ways: the extrusion that was under vacuum and temperature controlled, and the drying of pasta that was made at a very high temperature (90 degC instead of 37 degC or ambient temperature for the artisanal process). In this results section, pasta were produced according to two artisanal processes: "Artisanal 1" on wheat grains harvested in 2016 (see results above), and "Artisanal 2" on wheat grains harvested in 2017 (Table 2). In order to compare the effect of process, we considered, in Table 11, only the data coming from "Artisanal 2" and "Industrial" processes. The ANOVA highlighted the importance of the factor "Process" on proteins soluble in SDS and DTE (Table 11) and the factor "variety" had an impact on the amount of unextractable proteins and total protein content measured on cooked pasta. The four varieties exhibited a total protein content ranging from 12 (cv. Anvergur) to 15% db (cv. Bidi 17). The quantity of proteins solubilized with DTE ranged in average from 61.5 (cv. Claudio) to 68.7% (cv. LA1823), and those that were unextractable from 6.9% to 14.8% (Table S5). The cropping system, zero input or low input, did not have an impact on the considered parameters except on total protein content. The protein content was significantly and negatively correlated with DTE-soluble protein content (r2 = -0.54) at 0.0001% and with in vitro-digested proteins (r2 = -0.63). When also considering the data coming from the "Artisanal 1" process, we confirmed the importance of the process on the protein solubility. Both "Artisanal" processes led to pasta containing the most easily soluble proteins (33.3% of SDS-soluble, against 19.9% for industrial pasta) . Consequently, "Industrial" pasta presented a high level of highly polymerized proteins, soluble only with DTE. No significant difference was observed for the unextractable proteins. An in vitro test was carried out to measure the protein digestibility of OCT cooked pasta. It was performed with gastric and intestinal enzymes (pepsin, chymotrypsin and trypsin). The amount of digested proteins was measured at the end of the digestion step (= 100 - % of proteins remaining in the solid portion of the sample). The quantity of digested proteins was significantly higher in artisanal pasta (36.2% on average) than in industrial ones (28.6%) (Table S5). 4. Discussion Survey data and one-on-one interviews highlighted that some consumers, claiming to suffer from gastric disorders, consume pasta again when it comes from farmers or artisanal short chains. Such products are most often elaborated from local or old varieties, grown organically. The grains are then processed on the farm, ground with stone-mills and used to make pasta, produced under "slow" conditions, avoiding, for instance, drying at high temperatures, as is performed in industry. Starting from these observations, the objectives of this study were to evaluate the importance of various factors: variety, cropping system, process method on a major component of cereals, and proteins, in particular, those of gluten, identified in the literature as possible NCGS/NCWS (non-coeliac gluten sensitivity/non-coeliac wheat sensitivity) triggers. To date, no study has made it possible to link with certainty any constituent of wheat to NCGS/NCWS (i.e., gluten, ATIs, FODMAPs, wheat germ agglutinin) . This study was not intended to establish this link. The differences between artisanal and industrial type products were therefore approached from the angle of the protein composition of the grains and semolina, the degree of aggregation/reticulation of the proteins in the pasta following the process and their susceptibility to in-vitro proteolysis. 4.1. Evolution of Quality Parameters from Grain to Pasta Wheat proteins undergo structural changes during the pasta making process due to changes in hydration during mixing and drying, the input of mechanical energy during mixing and extrusion, and the intensity of heat treatment during drying. During the mixing phase, the semolina particles are hydrated, and the mobility of the compounds, including that of the proteins, is favored. It is only then, during extrusion, under the effect of mechanical stress, that glutenins and gliadins interact and form inter-molecular disulfide bonds that constitute the network of gluten . Regarding the 10 varieties for which we have data on lab-milled grains and cooked pasta, the protein content of ground grains is, as expected, predictive of the protein content of cooked pasta (correlation coefficient r2 = 0.88). The F5 fraction measured on ground grains also seems to be a fairly good predictor of the amount of soluble protein in SDS (r2 = 0.63). When considering the four varieties (Anvergur, Claudio, LA1823, Bidi17) for which we have data on ground grains, semolina, dry pasta and cooked pasta, we observe that the protein content of cooked pasta is very correlated to the total protein content present in grain (r2 = 0.92) or in dry pasta (r2 = 0.87). The total protein content of ground grains is also significantly and negatively correlated with the F3 (r2 = -0.78) and F5 (r2 = -0.86) fractions, and it is positively correlated with fraction F4 (r2 = 0.85) analyzed on ground grains. Between the lab-milled grains and the stone-milled grains (semolina), the F1, F2 and F4 fractions are strongly correlated (r2 = +0.89 to 0.93). Due to thermal treatment, the fractions analyzed on dry pasta are not correlated with those measured on lab-milled grains. The amount of SDS-soluble protein of cooked pasta is strongly and positively correlated (r2 = +0.89 to 0.95) with the F1 to F4 fractions measured on dry pasta. Thus, it is also strongly correlated with the amount of SDS-soluble protein (sum of fractions F1 to F5) measured on dry pasta (+0.93) and inversely correlated with the DTE-soluble protein content of dry pasta (-0.92). The other parameters measured on cooked pasta, such as DTE-soluble proteins, unextractable proteins, and protein digested in vitro, show no significant correlation with other parameters. In the artisanal process, with drying at a low temperature (37 degC), a slight decrease in the solubility of proteins is observed on the dry pasta, in comparison with the semolina, in particular on the F1 fraction, concomitant with an increase in the insoluble fraction. This loss of protein solubility between semolina and dry pasta reflects a change in protein behavior despite a "slow" processing mode. The quantity of SDS-soluble proteins thus goes from 89.2% in the semolina to 80.1% in the dry artisanal pasta. This decrease is significant and consistent with the conclusion of Aktan & Khan and De Zorzi et al. , who showed that drying induces a decrease in protein solubility. However, it is not in agreement with the work of Wagner et al. , who showed no significant difference in SDS-soluble proteins between semolina and fresh pasta (produced with the same equipment and under the same extrusion conditions as our industrial-type pasta). This result is to be compared with that of Martin et al. , who showed an insolubilization of glutenins and to a lesser extent of gliadins during resting phases experienced at 20, 30 or 40 degC and applied before drying. These authors indicate that there are no further differences in protein solubility between the pasta that have undergone a resting period and those that have not, as soon as they are all dried at a high temperature. 4.2. Differences between Varieties Used by Farmers and Those Recommended by Industrialists The French pasta companies annually update a list of around 10 varieties chosen from among the 568 varieties of durum wheat currently registered in the European catalogue. In 2010, they created a network of field trials called "Observatory" to "assess the agronomic aptitudes impacting the physical traits and technological qualities of the main cultivated varieties and those potentially in the making. Pasta and semolina manufacturers thus intend to transcribe precisely, in particular with breeders, the needs of processing industries to all actors in the durum wheat sector". Among the technological criteria, they are mainly interested in semolina yield, semolina and pasta aspect, and cooking quality. Farmers who mill their own harvest and turn it into pasta, use a small number of varieties of durum wheat. They are looking for locally adapted varieties or heritage varieties that give a particular flavor to the pasta. They often considered their varieties as "old" even the cv. LA1823, which comes from a participatory plant breeding project and was released in 2006! It was therefore difficult to obtain a sample of varieties representative of the diversity cultivated for each sector and which is balanced for statistical analyses. The varieties recommended by the French industry of semolina and pasta and the varieties used by peasant pasta makers differ significantly on several quality criteria measured on their grains such as TKW, SW, and protein content. Concerning TKW and SW, our data show that they are higher for varieties used by smallholder processors compared to the recent varieties used by industry. This could be explained by a shorter filling period for modern varieties. Thus, Nazco et al. reported a 12% reduction in the duration of filling of modern varieties compared to landraces, which would explain their lower weight. If we consider the protein content, the results of the first set of data, coming from the analysis of the grains of 20 varieties and the cooked pasta made from 10 varieties chosen among the 20 ("Artisanal 1"), show a higher protein content for the varieties used by peasant pasta makers. The yield/protein dilution curve confirms that this protein level is linked to the lower yields obtained for these varieties. This can be explained by the role of breeding which reduced grain protein concentration due to improved yields, but without impacting pasta cooking quality . We confirmed these results thanks to the second set of data, comprising four varieties for which we have data on each of the following products: lab-milled grains, stone-milled grains (semolina), dry pasta and cooked pasta. Whatever the product, its protein content is always significantly higher for the varieties used by peasant pasta makers. The proportion of the different protein classes is different between varieties recommended by the industry and the farmers. We observe that the increase in the protein content in the varieties used by farmers is expressed mainly by an increase in the F1, F2 and F4 fractions, to the detriment of the F5 fraction. Concerning the F4 fraction (a-, g-gliadins), the superiority measured in ground grains of the varieties processed by the farmers is consistent with the literature synthesis of Suter and Bekes , stating that "Because of the trend of decreases in overall protein content [...] older varieties are higher in gliadin content". It must be added that De Santis et al. or Pronin et al. noted no significant differences in the expression of a-, g-gliadins (F4) between modern and old genotypes but mentioned a significant decrease in o-gliadins (F3) in the modern genotypes. Breeding programs have aimed to improve technological quality by selecting certain glutenin alleles and increasing glutenins quantity . The new varieties are characterized by a better expression of LMW-GS, allowing us to obtain a lower Glia/Glu ratio and therefore to increase the strength of the gluten . According to Mefleh et al. , the genotypic variation in grain protein percentage among old varieties was more strongly associated with glutenin than with gliadin content. We do not note any difference concerning the ratio Glia/Glu between varieties used by farmers and those used by the industry probably due to too few varieties studied in both categories. In another article, De Santis et al. reported a content of unextractable polymeric protein (UPP) twice as high for the modern variety Saragolla as for the old variety Cappelli. Our data also show an increase in the proportion of UPP in the grains and semolina of the varieties recommended by industry, a proportion related, according to Sissons et al. , to the strength of gluten. This is correlated with the level of DTE-soluble proteins in dry pasta, which is also higher in the varieties recommended by the industry. However, after cooking, we do not see anymore any differences for this parameter. The four varieties tested in 2017 do not differ in the Glu/Glia ratio between modern and old genotypes. On the contrary, in 2016, the Glu/Glia ratio was indeed higher for the modern genotypes on milled grain. Several studies comparing modern wheats and ancient wheats have attempted to link the harmful effects of gluten to the year of creation and the change in protein composition. Some authors, such as Branlard and coworkers , have suggested that modern breeding would have led to varieties with more glutenin polymers or larger polymers, stronger glutens and therefore potentially more difficult to digest. It appears that old varieties are not all less toxic than recent varieties . However, the study by Ianiro et al. shows that patients with NCGS showed fewer gastrointestinal and extra-intestinal symptoms after ingesting pasta made with an "old" variety of durum wheat (Senatore Cappelli, date of release = 1923) compared to ingesting pasta from a mixture of recent varieties. This is consistent with the behavior observed among customers of the farmer pasta makers involved in our study. It would seem that the division of IND vs. FAR, in our study, is of little relevance to explain the great variability observed between the varieties on the quality of the proteins and their proteolysis in vitro. Indeed, concerning the unextractable proteins in cooked pasta, the cv. Claudio (IND) and Bidi 17 (FAR) have the highest percentage, whereas cv. Anvergur (IND) and LA1823 (FAR) have the lowest. Moreover, the results of protein solubility, obtained on pasta produced using a single process, do not make it possible to oppose the varieties used by farmers/artisans (FAR) and those recommended by the industry (IND). 4.3. Genotypic Variation for Quality Parameters Our results on protein solubility ( DTE-soluble proteins) of cooked pasta produced by the same farmer (same milling and processing conditions, set 1) show a different aggregation state depending on the variety. Indeed, there are more SDS-soluble proteins from cooked pasta made with cv. Bidi17, Pescadou and Miradoux. This difference can, perhaps, be explained by the glutenin composition. Most of the varieties with high SDS-soluble protein content have the HMW-GS type 7 + 8 encoded at the Glu-B1 locus (Pescadou, Miradoux, Claudio), while those with a lower proportion of SDS-soluble proteins have the 6 + 8 type (Joyau, Karur). In between are varieties with the HMW-GS 20x + 20y type (Bidi17, Qualidou, LA1823) or 13 + 16 type (Fabulis, Anvergur). This could allow us to link a weak and less cohesive gluten to more SDS-soluble proteins. However, this would only be a partial explanation, as in durum wheat, the work of Carrillo and colleagues has shown that it is most often the HMW-GS 20x + 20y type varieties that have a weaker gluten, and conversely, those of 7 + 8 or 6 + 8 types have medium to high gluten (see ). Moreover, HMW-GS composition alone is not sufficient to predict the gluten strength, as it is accepted that group B LMW-GS remain the key factor determining factor of gluten quality in durum wheat. Determination of the allelic composition of group B LMW-GS could help to better explain protein solubility in relation to gluten strength. 4.4. Effect of Environment on Protein Content and Composition The comparison of the 3 years of culture shows that the biotic and abiotic environmental factors play a determining role in the protein content and the composition of the grains of durum wheat, compared to the variety. The environmental variables considered in our study were: year, location and cropping system. For harvested grain, protein content was highest in 2019, followed by 2017 and 2016. The interaction between year and variety was significant for this parameter. Protein composition measured on lab-milled grains was also greatly affected by the year, with the exception of the F1 fraction (large glutenin aggregates), which is mainly variety dependent. The ratio Glia/Glu and the F4 fraction are also affected by the interaction year x var. These results are in line with those of Zivancev et al. , who observed that the year of production has no significant effect on the percentage ranges for glutenins. However, they had a considerable impact on the percentage ranges for gliadins. Climatic conditions (heat stress and drought) and cropping system are known to affect protein synthesis leading to variation in strength of gluten and affecting the final quality of semolina and flour . Thus, Labuschagne et al. showed that in response to stress conditions (high temperature and drought) SDS-extractable HMW-GS and gliadins increase, while SDS-unextractable HMW-GS decrease. DuPont et al. show that under a high temperature regimen (37/28 degC) and with the addition of NPK, the total protein content had increased, mainly due to an increase in o-gliadin and HMW-GS. When comparing the two cropping systems (zero input, ZI; low input, LI) implemented only in 2017, significant differences were noted on total protein content, F2, Fi and UPP, and to a lesser extent on F1 and F5 measured on lab-milled grains. The lower total protein content obtained with ZI results in a lower level of glutenins. These cropping systems were little contrasted: in both cases, the previous crop was a legume, the only difference was a nitrogen supply of 150 u and the use of a herbicide for the LI. However, these results are partially in line with those of Park et al. , who concluded that N fertilizer increased flour protein, relative concentration of glutenin and relative amount of high-molecular-weight glutenin subunits (HMW-GS/LMW-GS). Trials on more contrasting cropping systems over several years would make it possible to more accurately assess the impact of the cropping system. We only have measurements on whole and ground grains to assess the effect of the location of grain production (Mauguio and Purpan in 2019). This factor has a significant effect on the total quantity of proteins, and to a lesser extent on some fractions. Numerous authors showed that the environmental factor was the main source of variation in protein content for modern varieties. Criteria such as rainfall, drought, sowing date and nitrogen fertilization can cause significant differences in grain protein content . 4.5. Impact of Transformation Process on Protein Profile of Pasta After discussing the role of variety and the role of the environment, we are interested here in the impact of the type of processing on the protein status. Among the studied factors (variety, year, location, cropping system, and processing), our results show that the processing is the main factor affecting the protein fractions and in particular the protein solubility. The comparison of the types of processes implemented in this study, two different "Artisanal" types and one "Industrial" type, shows that the latter reduces the solubility of proteins in the SDS more strongly than the "Artisanal" types and that this difference remains, despite the cooking of the pasta. This confirms the results of Wagner et al. , indicating that the protein network formed during extrusion is largely insolubilized during drying and the greatest loss of protein solubility therefore occurs during the drying step. Our data, on dry pasta, show that the industrial process implemented, with the application of a very high temperature during drying, leads to a loss of protein solubility in the SDS. This data are in line with those obtained by Wagner et al. , Petitot et al. or Bruneel et al. . The heat treatment during cooking further reduces this solubility to 19.9%, close to the values of Wagner et al. , Petitot et al. and Brunel et al. , which range from 13 to 18%. Thus, the aggregation of gluten proteins during the cooking of pasta reinforces the density of the protein network by the formation of an intermolecular covalent cross-linking inducing an additional loss of solubility of the proteins in the SDS. With low temperature drying, the amount of SDS-soluble proteins in cooked artisanal pasta is higher than in industrial pasta, despite the cooking step. This difference in SDS solubility seems to be correlated with the degree of in vitro proteolysis. This rate is higher for artisanal pasta than for industrial pasta. These results agree with those of Petitot et al. , who show that the digestion of proteins is increased for pasta dried at a lower temperature compared to that of cooked pasta dried at a very high temperature. 5. Conclusions The role of varieties, and in particular the difference between the so-called modern varieties and the so-called old varieties, is much debated regarding the quality of gluten. By comparing, in average, the varieties recommended by the industry (IND) and those used by farmers who process pasta (FAR), we observed that only the total amount of protein varies: the FAR being, on average, much richer in protein. However, the solubility of these proteins, their proteolysis in vitro by digestive enzymes and the fractions analyzed by SE-HPLC vary little between the two groups of varieties, while differences between varieties inside each group are greatly observable. The impact of the location of production and the tested cropping systems on protein quality was low. Yet, more contrasting modalities should be studied before confirming or invalidating this point. The pasta production process, approached here by artisanal-type production and industrial-type production, is, among those studied, the factor having the greatest impact on protein quality. Pasta produced by the artisanal method contains a higher SDS-soluble protein fraction and is more easily proteolyzed. Whether these criteria are indicative of what happens during a consumer's digestion remains to be determined. It also remains to be assessed which key stages of the process have the greatest influence on protein quality. Acknowledgments The authors acknowledge the DiaScope (Mauguio) INRAE team for the grain production and NIRS analysis. The authors want to thank Lycee Agricole d'Auzeville and, mainly, Arnaud Marcodini for the milling and "Artisanal 2" pasta processing. They also want to thank warmly Francois Caizergues for the milling and "Artisanal 1" pasta processing. Supplementary Materials The following supporting information can be downloaded at: Table S1. Protein composition measured on lab-milled grain using SE-HPLC analysis: protein fractions from F1 to F5 (as a percentage of the total protein content), Fi (SDS-unextractable polymeric proteins as a percentage of the total protein content), UPP% (100 x Fi/(F1 + F2 + Fi)) and the gliadin/glutenin ratio (F4/F1 + F2 + Fi) and total protein content (% db). Values are the mean of replicates; Table S2. Protein composition measured on stone-milled grain (semolina) using SE-HPLC analysis: protein fractions from F1 to F5 (as a percentage of the total protein content), Fi (SDS-unextractable polymeric proteins as a percentage of the total protein content), UPP% (100 x Fi/(F1 + F2 + Fi)) and the gliadin/glutenin ratio (Glia/Glu = (F4/(F1 + F2 + Fi)). Total protein content measured by Kjeldahl method (% db). Values are the mean of replicates; Table S3. Protein composition measured on dry pasta using SE-HPLC analysis: protein fractions from F1 to F5 (as a percentage of the total protein content), DTE-soluble (DTE-soluble proteins as a percentage of the total protein content) and unextractable (unextractable proteins remaining after two extraction steps as a percentage of the total protein content); Table S4. Total protein content (% db) and protein composition measured on OCT cooked pasta using SE-HPLC analysis: SDS-soluble (SDS-soluble protein as a percentage of the total protein content), DTE-soluble (DTE-soluble proteins as a percentage of the total protein content). Values are mean of replicates; Table S5. Total protein content measured by Kjeldahl method (% db). Protein composition measured on OCT cooked pasta using SE-HPLC analysis: SDS-soluble (SDS-soluble protein as a percentage of the total protein content), DTE-soluble (DTE-soluble proteins as a percentage of the total protein content) and unextractable (unextractable proteins remaining after two extraction steps as a percentage of the total protein content). In vitro-digested protein (% of the total protein content). Not determined: nd. Values are mean of duplicates. Click here for additional data file. Author Contributions Conceptualization, D.D., K.M. and M.-F.S.; methodology, G.A., Y.C., D.D., K.M. and M.-F.S.; grain production, B.B., K.M., P.L. and D.D.; pasta production E.C., A.B.-E. and E.M., biochemical analysis, E.C., A.B.-E. and M.-F.S.; statistical analysis, V.A., M.-F.S. and D.D.; writing--original draft preparation, V.A., M.-F.S., A.B.-E. and D.D.; review and editing, M.-F.S. and D.D.; project administration, K.M. and D.D.; funding acquisition, K.M. and D.D. All authors have read and agreed to the published version of the manuscript. Data Availability Statement In the frame of open science encouraged by INRAE, the raw data will be available as soon as possible on Conflicts of Interest The authors declare no conflict of interest. Figure 1 Yield (t.ha-1) vs. Protein content (% db NIRS) in the 5 environments. FAR: varieties used by the farmer' pasta makers; IND: varieties recommended by the industry; DIV: underutilized genetic resources. Figure 2 Comparison of 2 years (H2016, left; H2017, right), same cropping system (ZI) and same location (Mauguio). Proportion of the different protein fractions (F1 to Fi) in lab-milled grains. The size of the small outer circles is proportional to the varietal coefficient of variation of each fraction. Figure 3 Proportion of Fi (left) and F4 (right) fractions measured on lab-milled grains of 20 varieties, cultivated under zero input and harvested in 2016 in Mauguio. Figure 4 Comparison of the protein fractions measured on ground grains, cultivated in zero (ZI)- or low (LI)-input conditions and harvested in 2016 (H2016) or 2017 (H2017) in Mauguio. For each fraction, the means and error bars (SD) of the two groups of varieties are compared: IND = varieties recommended by the industry; FAR = varieties used by the farmers making pasta. Figure 5 Comparison of the protein fractions measured on lab-milled grains and on stone-milled grains (or semolina). Means and error bars (SD) of 4 varieties: Anvergur, Claudio (IND) and Bidi17, LA1823 (FAR) harvested in 2017 (H2017) and grown under ZI conditions. Figure 6 Differences between protein fractions (F1 to F5, DTE-soluble and unextractable) between dry pasta made according "Artisanal 2" and "Industrial" processes with stone-milled grains grown under ZI cropping system. When the bar is directed upwards, the effect is greater with the IND process; when it is directed downwards, the effect is greater with the ART_2 process. * Difference significant at 5%. Figure 7 Comparison of the relative proportions of the different protein fractions between stone-milled grains (semolina) grown in 2017 under ZI cropping system and 2 different processes: "Artisanal 2" (ART_2) and "Industrial" (IND). Figure 8 Protein composition measured on freeze-dried cooked pasta from set 1 (H2016 Artisanal 1 process) using SE-HPLC analysis: SDS-soluble and DTE-soluble proteins (as a percentage of the total protein content) according to the cultivar. Figure 9 Protein composition measured on freeze-dried OCT cooked pasta from Set 2, using SE-HPLC analysis: SDS-soluble and DTE-soluble proteins (as a percentage of the total protein content), total protein content (% db) and proportion of in vitro-digested proteins (% of the total protein content) according to process and cropping system. Means and error bars (SD). Bar values marked with the same letter are not significantly different based on Student-Newman-Keuls (SNK) test performed at a = 0.05. foods-12-01093-t001_Table 1 Table 1 Varieties of wheat and number of replicates analyzed by type of product (whole grain, lab-milled grain, stone-milled grain or lab-milled grain, dry and cooked pasta) by variety and over the 3 years of harvest (H2016, H2017, H2019). Variety Group H2016 H2017 H2019 Whole Grain Lab-Milled Grain Cooked Pasta Whole Grain Lab-Milled Grain Stone-Milled Grain Dry Pasta Cooked Pasta Whole Grain Lab-Milled Grain Anvegur IND 2 3 2 2 4 4 8 8 5 4 Claudio IND 2 3 2 2 4 4 8 8 5 4 Fabulis IND 2 3 2 2 4 - - - - - Joyau IND 2 3 2 2 4 - - - - - Karur IND 2 3 2 2 4 - - - - - Miradoux IND 2 3 2 2 4 - - - - - Pescadou IND 2 3 2 2 4 - - - - - Qualidou IND 2 3 2 2 4 - - - - - Bidi 17 FAR 2 3 2 2 4 4 8 8 5 5 Bidi 17_2 * FAR 2 3 - 2 4 - - - - - LA1823 FAR 2 3 2 2 4 4 8 8 5 4 LA1823_2 * FAR - 3 - 2 - - - - - - Oued Zenati FAR - 3 - 2 4 - - - - - RG218 FAR 2 3 - 2 4 - - - - - Clovis DIV 2 3 - 2 4 - - - - - Desire DIV 2 3 - 2 4 - - - - - Meliani 1B DIV 2 3 - 2 4 - - - - - Vivit DIV 2 3 - 2 4 - - - - - RG1110 DIV - 3 - 2 4 - - - - - RG69649 DIV 2 3 - 2 4 - - - - - IND: varieties recommended by industry, FAR: varieties used by farmers, DIV: underutilized genetic resources, -: not available. * sources of the seeds are different. foods-12-01093-t002_Table 2 Table 2 Pasta making conditions used in "Artisanal 1", "Artisanal 2" and, "Industrial" processes. Process Year of Harvest Milling Type of Pasta Press (Manufacturer, Location) Quantity of Semolina (kg) Hydration (% db) Vacuum during Extrusion Extrusion Temperature (degC) Drying Step Artisanal 1 2016 SM 1 P3 (La Montferrina, Montcalieri, Italy) 3 ~50 no Not controlled Air drying Artisanal 2 2017 SM 2 Mac 30 (Italpast, Parma, Italy) 3 ~50 no Not controlled 37 degC 48 h Industrial 2017 SM 2 Experimental mini-press (Sercom, Montpellier, France) 0.7 47 Yes 40 degC 90 degC 5 h SM: stone milling. foods-12-01093-t003_Table 3 Table 3 Average of the thousand-kernel weight (TKW), specific weight (SW) and protein content (PROT) measured on durum wheat whole grains by NIRS method, according to the year, the cropping system and the location. Varieties used by farmers (FAR), varieties recommended by industry (IND), other varieties (DIV). Contribution (%) of each factor (variety and environment) to the variability of the criteria TKW, SW and PROT. Traits Year of Harvest H2016 H2017 H2019 Cropping System Zero Input Zero Input Low input Zero Input Zero Input Location Mauguio Mauguio Mauguio Mauguio Purpan Number of Studied Varieties 18 20 20 4 4 TKW (g) Mean Min Max 42.6 37.9 51.9 39.8 33.8 48.5 48.2 40.8 57.1 51.4 44.5 57.5 56.6 52.1 64.2 FAR 49.6 a SS 44.5 a 54.5 a 55.2 a 60.4 a IND 40.8 c 38.0 b 44.7 c 47.5 b 52.7 b DIV 42.6 b 38.6 b 47.5 b - - mean of cv. Bidi17 and cv. LA1823 (FAR) 49.4 a 45.7 a 52.7 a 55.2 a 60.4 a mean of cv. Anvergur and cv. Claudio (IND) 42.8 b 42.6 b 48.4 b 47.5 b 52.7 b SW (kg.hL-1) Mean Min Max 82.5 80.1 85.7 80.7 74.0 86.0 83.5 79.6 86.0 86.7 85.3 87.6 84.7 83.1 86.0 FAR 85.4 a 83.4 a 84.6 a 87.1 a 85.3 a IND 81.7 c 78.8 c 82.3 b 86.4 b 84.2 b DIV 83.0 b 81.0 b 84.1 a - - mean of cv. Bidi17 and cv. LA1823 (FAR) 85.4 a 85.0 a 85.2 a 87.1 a 85.3 a mean of cv. Anvergur and cv. Claudio (IND) 82.8 b 83.7 b 84.6 b 86.4 b 84.2 b PROT (% db) Mean Min Max 9.0 7.5 10.6 13.6 11.2 17.5 13.6 12.1 16.3 11.3 10.7 12.1 10.2 9.1 10.9 FAR 9.7 a 13.6 b 14.6 a 11.9 a 10.8 a IND 8.6 c 12.7 c 12.7 b 10.7 b 9.7 b DIV 9.3 b 14.6 a 14.0 a - - mean of cv. Bidi17 and cv. LA1823 (FAR) 9.4 a 13.2 a 13.8 a 11.9 a 10.8 a mean of cv. Anvergur and cv. Claudio (IND) 8.4 b 11.5 b 13.2 a 10.7 b 9.7 b SS Factor/SS total TKW SW PROT Variety 67% 68% 42% Environment 29% 20% 48% Error 4% 12% 10% SS In a same sub-column, numbers followed by the same letter are not significantly different based on Student-Newman-Keuls (SNK) test performed at a = 0.05. -: not available. foods-12-01093-t004_Table 4 Table 4 Contribution of each factor (Variety, Type, Year, Cropping system, and Location) to the variability of parameters concerning protein composition determined by SE-HPLC. Variance percentages coming from ANOVAs on lab-milled grains from the 2017 harvest (Sum of squares of each factor/total sum of squares) for each of the protein fraction. SE-HPLC Protein Composition Total Protein F1 + F2 F3 F4 F5 Fi UPP Glia/Glu All years Variety (n = 19) 37% **,++ 17% ** 10% ** 10% ** 4% * 14% ** 15% ** 26% ** 15% ** Type (n = 3) 2% ns 5% ** 2% * 3% ** 1% * 8% ** 7% ** 6% ** 12% ** Year (n = 3) 1% ns 68% ** 67% ** 80% ** 82% ** 62% ** 64% ** 49% ** 59% ** Cropping system (n = 2) 2% * 2% ** <1% ns <1% * <1% * 3% ** 3% ** <1% * 1% ** Location (n = 2) <1% ns <1% ns <1% ns <1% * <1% * <1% * <1% * <1% ns 4% ** Var x CS 11% ns 1% ns 4% ns 3% ** 2% ns 5% ** 3% ** 6% ** 3% ** Var x Location <1% ns <1% ns <1% ns <1% ns <1% ns 1% * 1% * <1% ns <1% * Residual 46% 6% 18% 4% 10% 7% 5% 11% 4% 2017 Variety 67% ** 72% ** 68% ** 75% ** 60% ** 53% ** 55% ** 78% ** 83% ** Cropping system 4% * 13% ** 1% 2% * 4% * 17% ** 19% ** 1% * 5% ** Variety x cropping system 10% 12% ** 7% 14% * 15% 24% ** 22% ** 14% ** 10% ** Residual 19% 3% 24% 9% 21% 6% 4% 7% 2% + F1 = large glutenin aggregates; F2 = medium and small glutenin aggregates; F3 = o-gliadins and high-molecular-weight albumins; F4 = - and g-gliadins; F5 = albumins-globulins, Fi = insoluble fraction; UPP: unextractable polymeric proteins; Glia/Glu: gliadin to glutenin ratio. ++ The contribution is calculated as the sum of squares of the factor/total sum of squares of the model and expressed in %. The model is as follows: Param = Type of variety + Variety effect + Year effect + Cropping system effect + Location effect + (Var x CS) + (Var x location)+ residual. ** Significant at 0.1%, * significant at 5%; ns: not significant. foods-12-01093-t005_Table 5 Table 5 Means of the protein quality parameters measured on ground grains (lab-milled grains) of cvs Anvergur, Claudio, Bidi17 and LA1823 grown under zero-input conditions and harvested in Mauguio in 2016, 2017 and 2019. Harvested Year Total Protein (% db) SE-HPLC Protein Fraction (%) UPP (%) Glia/Glu F1 F2 F3 F4 F5 Fi 2016 8.1 c SS 6.9 a 22.7 a 12.4 a 29.5 c 22.3 a 6.1 c 17.1 c 0.83 c 2017 10.2 b 6.6 a 18.1 c 9.2 c 37.0 a 17.8 c 11.3 a 31.6 a 1.03 a 2019 10.8 a 6.8 a 20.0 b 10.0 b 33.7 b 19.7 b 9.6 b 26.4 b 0.92 b Sum of squares = SS effect/SS model Variety 30% ** 71% ** 33% ** 1% 6% ** 6% 35% ** 37% ** 13% ** Year 65% ** 6% 64% ** 79% ** 90% ** 83% ** 62% ** 61% ** 78% ** Var x year 4% ** 6% 1% 2% 3% * 5% 1% 1% 6% * Residuals 1% 17% 2% 18% 1% 6% 2% 1% 3% SS Data followed by the same letter are not significantly different at 5% level. ** Significant at 0.1%, * significant at 5%. foods-12-01093-t006_Table 6 Table 6 Means of the protein quality parameters measured on ground grains (lab-milled grains) grown under zero-input conditions (ZI) and low input conditions (LI) in Mauguio in 2017, and Sum of squares (SS Effect/SS Model) for the variety and location effects. Mean of protein quality criteria measured on ground grains harvested in 2019 in 2 locations (Mauguio and Purpan), and Sum of squares (SS Effect/SS Model) for the variety effect and location effect. Total Protein (% db) SE-HPLC Protein Fraction (%) UPP % Glia/Glu F1 F2 F3 F4 F5 Fi 2017 LI 11.3 a SS 7.3 a 19.3 a 9.4 a 36.6 a 17.8 b 9.6 b 26.6 b 1.01 a ZI 10.8 b 6.9 b 18.5 b 9.2 a 36.1 a 18.3 a 11.1 a 30.3 a 0.99 a Sum of squares Variety ** ** ** ** ** ** ** ** ** Cropping system ** * ** ns ns * ** ** ns 2019 Mauguio 10.8 a SS 6.8 b 20.0 a 10.0 b 33.7 a 19.7 b 9.6 a 26.4 a 0.92 a Purpan 8.8 b 7.3 a 20.6 a 10.3 a 32.6 b 20.7 a 8.5 a 23.3 a 0.89 b Sum of squares Variety * ** ** * * ns * ** ** Location ** * ns * * * ns ns * SS Data followed by the same letter are not significantly different at 5% level. ** Significant at 0.1%, * significant at 5%; ns, not significant. foods-12-01093-t007_Table 7 Table 7 Means of protein content and protein fractions measured on lab-milled grains (H2016, H2017 and H2019) in 3 groups of varieties. FAR: varieties used by the farmer pasta makers; IND: varieties recommended by the industry; DIV: underutilized genetic resources. SE-HPLC Protein Fraction (%) UPP (%) Glia/Glu Total Protein (% db) F1 F2 F3 F4 F5 Fi FAR 7.4 a SS 21.3 a 10.4 b 34.1 a 19.7 b 7.1 c 19.8 c 0.95 a 10.5 a IND 7.1 a 20.5 b 10.4 b 32.7 c 20.2 a 9.1 a 24.8 a 0.89 b 9.1 c DIV 7.1 a 20.0 c 10.9 a 33.5 b 20.1 a 8.4 b 23.4 b 0.95 a 10.2 b SS Data followed by the same letter are not significantly different at 5% level. foods-12-01093-t008_Table 8 Table 8 Contribution of each factor (Sum of squares of the different factors (Variety, Cropping system)/total sum of squares) to the variability of the protein fractions in semolina (stone-milled grains). The model of analysis of variance was: Param = Variety + Cropping system + (variety x cropping system) + residuals, with Param = the protein fraction, UPP or the protein content. % SS Factor/Total SS SE-HPLC Protein Composition Total Protein F1 F2 F3 F4 F5 Fi UPP Glia/Glu Variety 93% ** 84% ** 46% * 49% * 22% ** 68% * 83% ** 56% * 86% ** Cropping system <0.3% 0.3% <1% 8% * 48% ** <0.1% <0.1% 3% 1.5% ** Var x CS 4% * 0.7% 30% * 30% * 28% ** 13% 4.5% 28% * 12.5% ** Residuals 3% 15% 23% 13% 2% 19% 12.5% 13% 0% ** Significant at 0.01%, * significant at 5%. foods-12-01093-t009_Table 9 Table 9 Contribution of each factor to the variability of the protein fractions in dry pasta, calculated by the ratio: Sum of squares of the factor/total sum of squares of the model. SE-HPLC Protein Composition F1 F2 F3 F4 F5 DTE-Soluble Unextr. Variety (n = 4) 1% 0.6% 1% 1% 2% 1% 29% * Cropping system (CS) (n = 2) <<1% 0.4% 2% 1% 2% * <<1% 6% Process (n = 2) 94% * 98% * 92% * 96% * 88% * 98% * 49% * Residual of the model (including the interactions) 4% 1% 5% 2% 8% (Var x CS = 6% *) 1% 16% (Var x Process = 7% *) * Significant at 5%. foods-12-01093-t010_Table 10 Table 10 Correlations between the protein fractions analyzed on ground grains and those analyzed on semolina and dry pasta. Lab-Milled Grains SE-HPLC Protein Composition F1 F2 F3 F4 F5 Fi or DTE-Soluble with Stone-milled grains 0.89 ** 0.93 ** 0.15 ns 0.89 ** 0.66 ns 0.55 ns with Dry Pasta 0.06 ns 0.06 ns 0.05 ns 0.11 ns 0.23 ns 0.05 ns ** Significant at 0.1%; ns: not significant. foods-12-01093-t011_Table 11 Table 11 Contribution of each factor to the variability of the parameters: percentage of proteins soluble in the SDS, proteins soluble in the DTE, unextractable proteins, and total protein content. %SS Factor/SS Tot SE-HPLC Protein Composition Total Protein SDS-Soluble DTE-Soluble Unextractable Proteins Variety 4% * 16% * 34% * 56% ** Cropping system 1% 1% 4.5% 4% * Process 84% ** 53% ** 0.5% 12% ** Residual of the model (including the significant interactions) 11% (var x process = 3% * var x CS = 4% *) 30% (var x process = 5.5% * var x CS = 11% *) 61% (var x process = 17.5% *) 28% (var x process = 5% * var x CS = 14% ** process x CS = 2% *) ** Significant at 0.1%; * significant at 5%; ns, not significant. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Desclaux D. Samson M.-F. Chiffoleau Y. Gluten Mythe ou Realite? A la Recherche des Personnes Hypersensibles. The Conversation (Ed. Francaise) 2018 Available online: (accessed on 6 December 2022) 2. Chiffoleau Y. Echchatbi A. Rod J. Gey L. Akermann G. Desclaux D. Jard G. Kessari M. Moinet K. Peres J. When social innovation re-orientates technological innovations in agri-food systems: The case of local chains around wheat Innovations 2021 64 41 63 10.3917/inno.pr2.0095 3. France Agrimer 2021. Qualite des Bles Durs--Recolte 2021 Available online: https//www.franceagrimer.fr/Actualite/Filieres/Cereales/2021/Qualite-des-bles-durs-francais-recolte-2021 (accessed on 6 December 2022) 4. Osborne T.B. The Vegetable Proteins Longmans, Green and Co. London, UK 1924 154p 5. Shewry P.R. Popineau Y. Lafiandra D. Belton P. Wheat glutenin subunits and dough elasticity: Findings of the EUROWHEAT project Trends Food Sci. Technol. 2000 11 433 441 10.1016/S0924-2244(01)00035-8 6. Shewry P.R. Tatham A.S. Forde J. Kreis M. Miflin B.J. The classification and nomenclature of wheat gluten proteins: A reassessment J. Cereal Sci. 1986 4 97 106 10.1016/S0733-5210(86)80012-1 7. Aprile M.C. Caputo V. Nayga R.M. Jr. Consumers' preferences and attitudes toward local food products J. Food Prod. Mark. 2015 22 19 42 10.1080/10454446.2014.949990 8. Subira J. Pena R.J. Alvaro F. Ammar K. Ramdani A. Royo C. Breeding progress in the pasta-making quality of durum wheat cultivars released in Italy and Spain during the 20th Century Crop Pasture Sci. 2014 65 16 26 10.1071/CP13238 9. Ruiz M. Giraldo P. The influence of allelic variability of prolamins on gluten quality in durum wheat: An overview J. Cereal Sci. 2021 101 103304 10.1016/j.jcs.2021.103304 10. Fu Y.-B. Somers D.J. Genome-wide reduction of genetic diversity in wheat breeding Crop Sci. 2009 49 161 168 10.2135/cropsci2008.03.0125 11. De Santis M.A. Giuliani M.M. Giuzio L. De Vita P. Lovegrove A. Shewry P.R. Flagella Z. Differences in gluten protein composition between old and modern durum wheat genotypes in relation to 20th century breeding in Italy Eur. J. Agron. 2017 87 19 29 10.1016/j.eja.2017.04.003 28769550 12. Pronin D. Borner A. Weber H. Scherf K.A. Wheat (Triticum aestivum L.) breeding from 1891 to 2010 contributed to increasing yield and glutenin contents but decreasing protein and gliadin contents J. Agric. Food Chem. 2020 68 13247 13256 10.1021/acs.jafc.0c02815 32648759 13. De Santis M.A. Cunsolo V. Giuliani M.M. Di Francesco A. Saletti R. Foti S. Flagella Z. Gluten proteome comparison among durum wheat genotypes with different release date J. Cereal Sci. 2020 96 103092 10.1016/j.jcs.2020.103092 14. Prandi B. Tedeschi T. Folloni S. Galaverna G. Sforza S. Peptides from gluten digestion: A comparison between old and modern wheat varieties Food Res. Int. 2017 91 92 102 10.1016/j.foodres.2016.11.034 28290332 15. Ficco D.B.M. Prandi B. Amaretti A. Anfelli I. Leonardi A. Raimondi S. Pecchioni N. De Vita P. Faccini A. Sforza S. Comparison of gluten peptides and potential prebiotic carbohydrates in old and modern Triticum turgidum ssp genotypes. Food Res. Int. 2018 120 568 576 10.1016/j.foodres.2018.11.007 31000273 16. Pronin D. Borner A. Scherf K.A. Old and modern wheat (Triticum aestivum L.) cultivars and their potential to elicit celiac disease Food Chem. 2020 339 127952 10.1016/j.foodchem.2020.127952 33152854 17. Ianiro G. Rizzatti G. Napoli M. Matteo M.V. Rinninella E. Mora V. Fanali C. Leonetti A. Benedettelli S. Mele M.C. A durum wheat variety-based product is effective in reducing symptoms in patients with non-celiac gluten sensitivity: A double-blind randomized cross-over trial Nutrients 2019 11 712 10.3390/nu11040712 30934747 18. Akermann G. Coeurquetin P. Consommer sans gluten: Trajectoires individuelles d'eviction Faire Sans, Faire Avec Moins: Les Nouveaux Horizons de L'innovation Goulet F. Vinck D. Presses des Mines, Collection Sciences sociales Paris, France 2022 53 68 19. Penuelas J. Gargallo-Garriga A. Janssens I. Ciais P. Obersteiner M. Klem K. Urban O. Zhu Y.-G. Sardans J. Could global intensification of nitrogen fertilisation increase immunogenic proteins and favour the spread of coeliac pathology? Foods 2020 9 1602 10.3390/foods9111602 33158083 20. Godfrey D. Hawkesford M.J. Powers S.J. Millar S. Shewry P.R. Effects of crop nutrition on wheat grain composition and end use quality J. Agric. Food Chem. 2010 58 3012 3021 10.1021/jf9040645 20131902 21. Wan Y. Gritsch C.S. Hawkesford M.J. Shewry P.R. Effects of nitrogen nutrition on the synthesis and deposition of the o-gliadins of wheat Ann. Bot. 2013 113 607 615 10.1093/aob/mct291 24344140 22. Dexter J.E. Matsuo R.R. Influence of protein content on some durum wheat quality parameters Can. J. Plant Sci. 1977 57 717 727 10.4141/cjps77-105 23. Sissons M. Cutillo S. Marcotuli I. Gadaleta A. Impact of durum wheat protein content on spaghetti in vitro starch digestion and technological properties J. Cereal Sci. 2020 98 103156 10.1016/j.jcs.2020.103156 24. Troccoli A. Borrelli G.M. De Vita P. Fares C. Di Fonzo N. Mini review: Durum wheat quality: A multidisciplinary concept J. Cereal Sci. 2000 32 99 113 10.1006/jcrs.2000.0322 25. Sissons M.J. Ames N.P. Hare R.A. Clarke J.M. Relationship between glutenin subunit composition and gluten strength measurements in durum wheat J. Sci. Food Agric. 2005 85 2445 2452 10.1002/jsfa.2272 26. Sissons M. Role of durum wheat composition on the quality of pasta and bread Food 2008 2 75 90 27. Pollini C.M. Panto F. Nespoli A. Sissons M. Abecassis J. Manufacture of pasta products Durum Wheat: Chemistry and Technology 2nd ed. Sissons M. Abecassis J. Marchylo B. Carcea M. AACC International Inc. Saint Paul, MN, USA 2012 436 476 28. Malcolmson L.J. Matsuo R.R. Balshaw R. Textural optimization of spaghetti using response surface methodology: Effects of drying temperature and durum protein level Cereal Chem. 1993 70 417 423 29. Zweifel C. Handschin S. Escher F. Conde-Petit B. Influence of high-temperature drying on structural and textural properties of durum wheat pasta Cereal Chem. 2003 80 159 167 10.1094/CCHEM.2003.80.2.159 30. Aktan B. Khan K. Effect of high-temperature drying of pasta on quality parameters and on solubility, gel electro-phoresis, and reversed-phase high-performance liquid chromatography of protein components Cereal Chem. 1992 69 288 295 31. Gilani G.S. Xiao C.W. Cockell K.A. Impact of antinutritional factors in food proteins on the digestibility of protein and the bioavailability of amino acids and on protein quality Br. J. Nutr. 2012 108 S315 S332 10.1017/S0007114512002371 23107545 32. Joye I. Protein digestibility of cereal products Foods 2019 8 199 10.3390/foods8060199 31181787 33. Smith F. Pan X. Bellido V. Toole G.A. Gates F.K. Wickham M.S.J. Shewry P.R. Bakalis S. Padfield P. Mills E.N.C. Digestibility of gluten proteins is reduced by baking and enhanced by starch digestion Mol. Nutr. Food Res. 2015 59 2034 2043 10.1002/mnfr.201500262 26202208 34. Wu T. Taylor C. Nebl T. Ng K. Bennett L.E. Effects of chemical composition and baking on in vitro digestibility of proteins in breads made from selected gluten-containing and gluten-free flours Food Chem. 2017 233 514 524 10.1016/j.foodchem.2017.04.158 28530606 35. Gouseti O. Lovegrove A. Kosik O. Fryer P.J. Mills E.C. Gates F.K. Tucker G. Latty C. Shewry P.R. Bakalis S. Exploring the role of cereal dietary fiber in digestion J. Agric. Food Chem. 2019 67 8419 8424 10.1021/acs.jafc.9b01847 31267740 36. Gulati P. Brahma S. Graybosch R.A. Chen Y. Rose D.J. In vitro digestibility of proteins from historical and modern wheat cultivars J. Sci. Food Agric. 2020 100 2579 2584 10.1002/jsfa.10283 31975391 37. Ma F. Baik B. Influences of grain and protein characteristics on in vitro protein digestibility of modern and ancient wheat species J. Sci. Food Agric. 2021 101 4578 4584 10.1002/jsfa.11100 33474737 38. Cereals & Grains Association Pasta and Noodle Cooking Quality-Firmness AACC Approved Methods of Analysis 11th ed. Method 66-50.01 Cereals & Grains Association St. Paul, MN, USA 1999 10.1094/AACCIntMethod-66-50.01 39. Cereals & Grains Association Moisture-Air oven methods AACC Approved Methods of Analysis 11th ed. Method 44-15.02 Cereals & Grains Association St. Paul, MN, USA 1999 10.1094/AACCIntMethod-44-15.02 40. Cereals & Grains Association Crude Protein--Kjeldahl Method, Boric Acid Modification AACC Approved Methods of Analysis 11th ed. Method 46-12.01 Cereals & Grains Association St. Paul, MN, USA 1999 10.1094/AACCIntMethod-46-12.01 41. Morel M.-H. Dehlon P. Autran J.C. Leygue J.P. Bar-L'Helgouac'H C. Effects of temperature, sonication time, and power settings on size distribution and extractability of total wheat flour proteins as determined by size-exclusion high-performance liquid chromatography Cereal Chem. 2000 77 685 691 10.1094/CCHEM.2000.77.5.685 42. Dachkevitch T. Autran J.C. Prediction of baking quality of bread wheats in breeding programs by size-exclusion high-performance liquid chromatography Cereal Chem. 1989 66 448 456 43. Kucek L.K. Veenstra L.D. Amnuaycheewa P. Sorrells M.E. A grounded guide to gluten: How modern genotypes and processing impact wheat sensitivity Compr. Rev. Food Sci. Food Saf. 2015 14 285 302 10.1111/1541-4337.12129 33401796 44. Brouns F. van Rooy G. Shewry P. Rustgi S. Jonkers D. adverse reactions to wheat or wheat components Compr. Rev. Food Sci. Food Saf. 2019 18 1437 1452 10.1111/1541-4337.12475 33336916 45. Scherf K.A. Immunoreactive cereal proteins in wheat allergy, non-celiac gluten/wheat sensitivity (NCGS) and celiac disease Curr. Opin. Food Sci. 2019 25 35 41 10.1016/j.cofs.2019.02.003 46. Geisslitz S. Weegels P. Shewry P. Zevallos V. Masci S. Sorrells M. Gregorini A. Colomba M. Jonkers D. Huang X. Wheat amylase/trypsin inhibitors (ATIs): Occurrence, function and health aspects Eur. J. Nutr. 2022 61 2873 2880 10.1007/s00394-022-02841-y 35235033 47. Icard-Verniere C. Feillet P. Effects of mixing conditions on pasta dough development and biochemical changes Cereal Chem. 1999 76 558 565 10.1094/CCHEM.1999.76.4.558 48. Petitot M. Abecassis J. Micard V. Structuring of pasta components during processing: Impact on starch and protein digestibility and allergenicity Trends Food Sci. Technol. 2009 20 521 532 10.1016/j.tifs.2009.06.005 49. De Zorzi M. Curioni A. Simonato B. Giannattasio M. Pasini G. Effect of pasta drying temperature on gastrointestinal digestibility and allergenicity of durum wheat proteins Food Chem. 2007 104 353 363 10.1016/j.foodchem.2006.11.057 50. Wagner M. Morel M.-H. Bonicel J. Cuq B. Mechanisms of heat-mediated aggregation of wheat gluten protein upon pasta processing J. Agric. Food Chem. 2011 59 3146 3154 10.1021/jf104341w 21370874 51. Martin C. Morel M.-H. Reau A. Cuq B. Kinetics of gluten protein-insolubilisation during pasta processing: Decoupling between temperature-dependent effects J. Cereal Sci. 2019 88 103 109 10.1016/j.jcs.2019.05.014 52. Desclaux D. Chiffoleau Y. Nolot J.M. Pluralite des agricultures biologiques: Enjeux pour la construction des marches, le choix des varietes et les schemas d'amelioration des plantes Innov. Agron. 2009 4 297 306 53. Nazco R. Villegas D. Ammar K. Pena R.J. Moragues M. Royo C. Can Mediterranean durum wheat landraces contribute to improved grain quality attributes in modern cultivars? Euphytica 2011 185 1 17 10.1007/s10681-011-0588-6 54. Motzo R. Fois S. Giunta F. Relationship between grain yield and quality of durum wheats from different eras of breeding Euphytica 2004 140 147 154 10.1007/s10681-004-2034-5 55. Xynias I.N. Mylonas I. Korpetis E.G. Ninou E. Tsaballa A. Avdikos I.D. Mavromatis A.G. Durum wheat breeding in the mediterranean region: Current status and future prospects Agronomy 2020 10 432 10.3390/agronomy10030432 56. Suter D.A.I. Bekes F. Who is to blame for the increasing prevalence of dietary sensitivity to wheat? Cereal Res. Commun. 2021 49 1 19 10.1007/s42976-020-00114-0 57. De Vita P. Nicosia O.L.D. Nigro F. Platani C. Riefolo C. Di Fonzo N. Cattivelli L. Breeding progress in morpho-physiological, agronomical and qualitative traits of durum wheat cultivars released in Italy during the 20th century Eur. J. Agron. 2007 26 39 53 10.1016/j.eja.2006.08.009 58. Branlard G. Faye A. Rhazi L. Tahir A. Lesage V. Aussenac T. Genetic and environmental factors associated to glutenin polymer characteristics of wheat Foods 2020 9 683 10.3390/foods9050683 32466243 59. Mefleh M. Conte P. Fadda C. Giunta F. Piga A. Hassoun G. Motzo R. From ancient to old and modern durum wheat varieties: Interaction among cultivar traits, management, and technological quality J. Sci. Food Agric. 2018 99 2059 2067 10.1002/jsfa.9388 30267406 60. Mefleh M. Conte P. Fadda C. Giunta F. Motzo R. From seed to bread: Variation in quality in a set of old durum wheat cultivars J. Sci. Food Agric. 2019 100 4066 4074 10.1002/jsfa.9745 30977135 61. Branlard G. Faye A. Meleard B. Le Gouis J. Oury F.-X. Ravel C. Quelle hypothese approfondir pour apprehender la sensibilite non celiaque au gluten: 1. L'abondance du gluten et la diversite des prolamines Ind. Cereales 2019 211 34 37 62. Branlard G. Rhazi L. Meleard B. Le Gouis J. Ravel C. Quelle hypothese approfondir pour apprehender la sensibilite non celiaque au gluten: 2. Les polymeres des glutenines Ind. Cereales 2019 212 26 29 63. Carrillo J.M. Vazquez J.F. Orkellana J. Relationship between gluten strength and glutenin proteins in durum wheat cultivars Plant Breed. 1990 104 325 333 10.1111/j.1439-0523.1990.tb00443.x 64. Magallanes-Lopez A.M. Ammar K. Morales-Dorantes A. Gonzalez-Santoyo H. Crossa J. Guzman C. Grain quality traits of commercial durum wheat varieties and their relationships with drought stress and glutenins composition J. Cereal Sci. 2017 75 1 9 10.1016/j.jcs.2017.03.005 65. Zivancev D. Torbica A.M. Tomic J.M. Hajnal E.P.J. Belovic M. Mastilovic J.S. Kevresan Z.S. Effect of Climate Change on Wheat Quality and HMW Glutenin Subunit Composition in the Pannonian Plain Cereal Chem. 2016 93 90 99 10.1094/CCHEM-05-15-0101-R 66. Labuschagne M. Guzman C. Phakela K. Wentzel B. van Biljon A. Solvent Retention Capacity and Gluten Protein Composition of Durum Wheat Flour as Influenced by Drought and Heat Stress Plants 2021 10 1000 10.3390/plants10051000 34067794 67. DuPont F. Hurkman W. Vensel W. Chan R. Lopez R. Tanaka C. Altenbach S. Differential accumulation of sulfur-rich and sulfur-poor wheat flour proteins is affected by temperature and mineral nutrition during grain development J. Cereal Sci. 2006 44 101 112 10.1016/j.jcs.2006.04.003 68. Park H. Clay D.E. Hall R.G. Rohila J.S. Kharel T.P. Clay S.A. Lee S. Winter wheat quality responses to water, environment, and nitrogen fertilization Commun. Soil Sci. Plant Anal. 2014 45 1894 1905 10.1080/00103624.2014.909833 69. Ames N.P. Clarke J.M. Marchylo B.A. Dexter J.E. Woods S.M. Effect of Environment and Genotype on Durum Wheat Gluten Strength and Pasta Viscoelasticity Cereal Chem. 1999 76 582 586 10.1094/CCHEM.1999.76.4.582 70. Rharrabti Y. Royo C. Villegas D. Aparicio N. del Moral L.G. Durum wheat quality in Mediterranean environments: I. Quality expression under different zones, latitudes and water regimes across Spain Field Crops Res. 2003 80 123 131 10.1016/S0378-4290(02)00176-4 71. Rharrabti Y. Villegas D. Royo C. Martos-Nunez V. del Moral L.G. Durum wheat quality in Mediterranean environments: II. Influence of climatic variables and relationships between quality parameters Field Crops Res. 2003 80 133 140 10.1016/S0378-4290(02)00177-6 72. Giunta F. Pruneddu G. Motzo R. Grain yield and grain protein of old and modern durum wheat cultivars grown under different cropping systems Field Crops Res. 2018 230 107 120 10.1016/j.fcr.2018.10.012 73. Petitot M. Brossard C. Barron C. Larre C. Morel M.-H. Micard V. Modification of pasta structure induced by high drying temperatures. Effects on the in vitro digestibility of protein and starch fractions and the potential allergenicity of protein hydrolysates Food Chem. 2009 116 401 412 10.1016/j.foodchem.2009.01.001 74. Bruneel C. Pareyt B. Brijs K. Delcour J.A. The impact of the protein network on the pasting and cooking properties of dry pasta products Food Chem. 2010 120 371 378 10.1016/j.foodchem.2009.09.069 |
PMC10000653 | Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods12050980 foods-12-00980 Article Effect of Catechin on the Formation of Polycyclic Aromatic Hydrocarbons in Camellia oleifera Oil during Thermal Processing Pei Wenjun Formal analysis Investigation Data curation Writing - original draft 1 Wang Jiaqi Validation Project administration 2 Zhang Lu Validation Data curation 1 Guo Yiwen Writing - review & editing 1 Cao Minjie Writing - review & editing 1 Liu Ruijie Conceptualization Supervision Funding acquisition 1* Chang Ming Supervision Funding acquisition 1 Wang Xingguo Writing - review & editing Supervision 1 Artieda Diana Ansorena Academic Editor De Leonardis Antonella Academic Editor 1 National Key Laboratory, School of Food Science and Technology, Jiangnan University, 1800 Lihu Road, Wuxi 214122, China 2 China National Center for Food Safety Risk Assessment, Southeast of the Intersection of Huangmuchang Road and Dongbai Street, Chaoyang, Beijing 100022, China * Correspondence: [email protected]; Tel.: +86-0510-85876799 25 2 2023 3 2023 12 5 98010 12 2022 30 1 2023 20 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Polycyclic aromatic hydrocarbons (PAHs) in oil are affected by many factors, including temperature, time, and PAHs precursors. Phenolic compounds, as beneficial endogenous components of oil, are often associated with the inhibition of PAHs. However, studies have found that the presence of phenols may lead to increased levels of PAHs. Therefore, this study took Camellia oleifera (C. oleifera) oil as the research object, in order to study the effect of catechin in the formation of PAHs under different heating conditions. The results showed that PAH4 were generated rapidly during the lipid oxidation induction period. When the addition of catechin was >0.02%, more free radicals were quenched than generated, thus inhibiting the generation of PAH4. ESR, FT-IR, and other technologies were employed to prove that when the catechin addition was <0.02%, more free radicals were produced than quenched, causing lipid damage and increasing PAHs intermediates. Moreover, the catechin itself would break and polymerize to form aromatic ring compounds, ultimately leading to the conclusion that phenolic compounds in oil may be involved in the formation of PAHs. This provides suggestions for the flexible processing of phenol-rich oil to balance the retention of beneficial substances, and for the safe control of hazardous substances in real-life applications. C. oleifera oil polycyclic aromatic hydrocarbon phenolic compounds lipid oxidation "National key research and development program" of the Ministry of Science and Technology2020YFC1606805 This work was supported by the "National key research and development program" of the Ministry of Science and Technology, PRC (2020YFC1606805). pmc1. Introduction PAHs are a class of organic compounds that contain multiple (>=2) aromatic rings commonly found in various natural or artificial media we are exposed to daily, such as water, soil, air, and food . Several studies have shown that long-term exposure to PAHs can lead to many adverse health outcomes, including abnormal lung function, chronic obstructive pulmonary disease, and various cancers . As a result of their lipophilic and hydrophobic properties, PAHs are more likely to accumulate in the food chain , and are usually produced during high-temperature processes, such as grilling, frying, smoking, baking, etc. . Among them, oil is the primary source of dietary exposure to PAHs. Even so, the exact mechanism of PAHs formation in oil has not been well verified; however, some researchers believe that PAHs formation is strongly related to fats, and speculate that fatty acids generate many free radicals during their oxidation to hydroperoxides, which form PAHs through reactions such as intramolecular addition or small molecule polymerization . Oil is an essential part of the human diet, and C. oleifera oil is a vegetable oil rich in beneficial ingredients and is widely used by consumers. Considering the susceptibility of edible oil to PAHs and their risk to human health during daily high-temperature use, it is crucial to investigate the causes of PAHs generation in C. oleifera oil, in order to control PAHs contamination. Previously, numerous studies have shown that phenolic compounds can inhibit the formation of PAHs due to their antioxidant properties ; however, it was found that the contents of phenolic compounds, phytosterols, and another lipid companions in C. oleifera oil extracted by the hot pressing process increased with an increase in temperature, while the contents of Benzo(a)pyrene (BaP) also increased . There are also studies found where the contents of PAH8 [sum of benzo(a)anthracene, chrysene, benzo(b)fluoranthene, benzo(k)fluoranthene, benzo(a)pyrene, dibenzo(a,h)anthracene, benzo(g,h,i)perylene, and indeno(1,2,3-c,d)pyrene] in chargrilled chicken wings marinated with black tea high in phenolic contents were much higher than that in control groups . Min et al., as well as others, found higher levels of PAH8 in their heated meat model system than in the control group when EGCG [(-)-epigallocatechin gallate] was used as an antioxidant at concentrations of 50 mg/kg and 100 mg/kg . We also found in this study that the rate of increase of PAHs in crude camellia oil (CCO) was faster than that in refined camellia oil (RCO) under the same heat treatment conditions. Therefore, we speculated whether some substances in the CCO caused an increase in PAHs after heat treatment. Based on this, we found that many studies pointed out the presence of the precursors of PAHs in food, such as amino acids, lipids, polyphenols, cellulose, etc. These precursors may undergo condensation polymerization to form PAHs under high-temperature conditions . Therefore, the aim of this study was to elucidate the effect of phenols present in C. oleifera oil, which may be PAHs precursors, on the formation of PAHs and the mechanism of their formation. In order to detect accurately the PAHs formed after heating, changes in PAH4 [sum of benzo(a)anthracene, benzo(b)fluoranthene, chrysene, benzo(a)pyrene] were monitored in this study, which was stated to be a more suitable approach to quantify the contamination level of PAHs in oil . The findings of this study will provide valuable and innovative reference data, as well as scientific basis for the control of PAHs in edible oils. 2. Materials and Methods 2.1. Materials The CCO and RCO were obtained from Zhejiang Jiusheng Oil Seed Oil Co., Ltd. in Jiande City, Zhejiang Province. The standards of catechin were purchased from Sigma-Aldrich (St. Louis, MO, USA). Reagents for all experiments were obtained from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). 2.2. Preparation of C. oleifera Oil with Different Catechin Contents and Heating Conditions Preparation of oil samples containing different catechin concentrations: The catechin standard was accurately weighed, first dissolved in ethanol, then added corresponding volumes of ethanol solution to C. oleifera oil, and then dried under nitrogen at 35 degC. The final mass concentrations of catechin were 50, 100, 200, and 400 mg/kg, respectively, and stored at -4 degC. A blank group without catechin was set as the control. Preparation of heated oil samples for CCO and RCO: 500 mL crude oil and 500 mL refined oil were heated continuously at 180 +- 5 degC for 6 h in a thermostatic oil bath (Gongyi City Yuhua Instrument Co., Ltd., Zhengzhou, China). An aliquot of 60 mL oil was collected every 1 h during the 6 h continuous heating process and the collected samples were stored at -4 degC until further analysis. Preparation of Rancimat pro-oxidation oil samples with catechin addition: The above catechin-added samples were heated with Rancimat to promote oxidation. The induction Period (IP) was measured with a Rancimat892 apparatus from Metrohm Nordic ApS (Glostrup, Denmark), at 140 degC, 160 degC, 180 degC, 200 degC, and 20 L/h airflows. 2.3. Analysis of PAHs PAHs were detected by high performance liquid chromatography-fluorescence detection (HPLC-FLD) according to the previous study . The correlation coefficient was greater than 0.999. The limits of detection (LOD) of Chr, BaA, BbF, and BaP in C. oleifera oil were 0.25, 0.29, 0.023, and 0.061 mg/kg. Meanwhile, the corresponding limits of quantitation (LOQ) were 0.85, 0.98, 0.076, and 0.21 mg/kg, respectively. The recovery rates of PAH4 ranged from 70.2% to 110%. Calculation of LOD and LOQ: An aliquot of 0.5 g oil sample was weighed and PAHs mixed standard solution was added quantitatively to achieve concentrations of 2 mg/kg, 10 mg/kg, and 20 mg/kg, respectively, for 6 repetitions. The detection limit of the object to be measured was set according to international regulations . The concentration under S/N = 3 is defined as LOD, while the concentration under S/N = 10 is defined as the limits of quantitation LOQ. 2.4. Determination of Physicochemical Parameters, Fatty Acid Composition, Minor Constituents Content, and Heating Products The physicochemical parameters (acid value and peroxide value) of this study were determined according to the AOCS official method . Fatty acid composition was analyzed by Gas Chromatography (GC) according to the method reported previously . Minor constituents contents (tocopherols, phytosterols, and total phenols) and PAHs intermediates were analyzed according to the previous study , and some modifications were made on this basis. Details are provided in the Supplementary Materials to the manuscript. 2.5. Electron Spin Resonance (ESR) Spectroscopy ESR spectra were measured using a Bruker spectrometer (BrukerGmbH, Karlsruhe, Germany). Samples with and without catechin were used to detect free radical contents after heating; accurately weighed 0.3% a-phenyl-n-tert-butyl nitrogen oxide (PBN), dissolved in the oil sample, mix thoroughly, and heat. Next, 100 mL of the heated oil sample was placed in a nuclear magnetic tube with an inner diameter of 4 mm. Then, it was placed in the resonant cavity of the measurement instrument. The ESR detection parameters refer to Chen . 2.6. Fourier Transform Infrared Spectroscopy (FT-IR) Analysis The FT-IR spectrum of oil used a Fourier transform infrared spectrometer (Vertex 70, Bruker Banner Lane, Coventry, Germany). The main operating parameters of the instrument were as follows: measuring spectrum range of 4000~400 cm-1, scanning times of 64, resolution of 4 cm-1, and 2 times gain. Data were collected in log (1/R) format. After the collection program was programmed, the infrared spectrum of the sample was directly collected. Each sample was scanned three times, and its average spectrum was calculated and stored until analysis. 2.7. Identification of Catechin and Its Structures by High Performance Liquid Chromatography Ultraviolet Spectrometry (HPLC-UV) and Ultra Performance Liquid Chromatography Coupled with Quadrupole Time-of-Flight Tandem Mass Spetrometry (UPLC-Q-TOF/MS) HPLC-UV was used to detect catechin. Briefly, the sample was dissolved with methanol, then the catechin in the sample were extracted by shock, followed by ultrasound for 20 min, and then placed in a -18 degC refrigerator for 10 min. After centrifugation, the supernatant was taken. Liquid phase and mass spectrometry detection conditions were referenced in previous studies . 2.8. Statistical Analysis All experiments were performed in triplicate and results were expressed as means +- standard errors. Structures were produced with the use of ChemDraw. Statistical significance was analyzed by one-way ANOVA, using the SPSS 22.0 package (IBM, New York, NY, USA). The results were regarded as statistically significant when p < 0.05. Curve fits were performed using the Origin 8.0 (Electronic Arts Inc., Redwood City, CA, USA). 3. Results and Discussion 3.1. Changes in Physicochemical Properties, Lipid Companion, Fatty Acid Composition, and PAH4 of Crude C. oleifera Oil and Refined C. oleifera Oil during Heating Combined with previous studies, we found that the rate of increase of PAH4 in CCO and RCO differed under the same heating conditions. Figure 1A shows the change in PAH4 contents of CCO and RCO after heating at 180 degC for 6 h. It can be seen that the 21.88 mg/kg PAH4 contents in RCO are higher than that in CCO (9.88 mg/kg) without heating. The contents of PAH4 in both samples increased after heating, which may be due to the fact that high temperature is conducive to the formation of PAHs, through the massive cleavage of organic compounds . After heating for 6 h, PAH4 in CCO and RCO both increased significantly (p < 0.05), results which were 82.61% and 36.01% higher than their initial PAH4 contents. It can be seen that the rate of increase of PAH4 in CCO was faster than that of RCO. This is a curious phenomenon, so this study performed a series of index measurements; thus, the main differences between CCO and RCO were identified and the reasons for this phenomenon were investigated. The changes in fatty acids in C. oleifera oil under prolonged heating were firstly examined. As shown in Table 1, the main saturated fatty acids of CCO and RCO were palmitic acid and stearic acid, and the unsaturated fatty acids were oleic and linoleic acid. With the extension of heating time, the unsaturated fatty acids in CCO and RCO both decreased obviously, and palmitic acid contents gradually increased. Acid value (AV) and peroxide value (POV) are the most basic and important parameters to evaluate oil quality . The changes in AV and POV of RCO and CCO during the heating process are shown in Figure 1B,C. They both underwent a continuous increase in AV and POV. It is clear, however, that the AV and POV growth rates in CCO are slower than those in RCO, which might be due to the fact that CCO contained more lipid companions. Tocopherols, phytosterols, and polyphenols are natural nutrients in C. oleifera oil that affect the quality of the oil . Changes in the contents of tocopherols, phytosterols, and total phenols during heating are shown in Table 1, from which it can be seen that the contents of tocopherols, phytosterols, and total phenols in unheated CCO were higher than those of the lipid companion in RCO. During the heating process, the contents of tocopherols, phytosterols, and total phenols in both oils underwent a significant reduction (p < 0.05). Tocopherols, phytosterols, and total phenols in CCO decreased by 82.5%, 22.5%, and 59.7%, respectively, and the RCO phytosterols and total phenols decreased by 10.7% and 42%, after 6 h heating, respectively. It can be seen that the loss of lipid companions in CCO was more serious throughout the heating process; as a result, it was speculated that some reactions of lipid companions play a role in the heating process of CCO, resulting in the differences between CCO and RCO. In summary, it is worth exploring whether the growth of PAHs in CCO was the result of more abundant lipid companions. Through a series of experiments, combined with a literature review, we hypothesized that the increased production of PAHs in C. oleifera oil may be related to phenolic compounds. Therefore, we conducted correlation analysis of heating time, fatty acid composition, lipid companion, and PAH4 in CCO (Table 2), and found that the formation of PAH4 was negatively correlated with the reduction of total phenol, which suggested that the depleted phenolic compounds might be involved in the formation of PAH4. Therefore, we chose refined oil that had most of its lipid companions removed as the research object. We first studied the effect of phenolic compounds on the formation of PAHs in refined oil under high temperature, and then explored the mechanism of its influence on the formation of PAHs. 3.2. Effects of Catechin on the Formation of PAH4 in C. oleifera Oil under Different Heating Temperatures, Heating Times, and Catechin Additions According to previous research , several phenolic compounds present in C. oleifera oil were selected for separate addition experiments, and the change in PAH4 contents in the oil was examined . This phenomenon made us think about why catechin caused an increase of PAHs in the heating process of oil. Therefore, we chose catechin as the research object, and explored its mechanism in the subsequent study. As shown in Figure 2A, at 0.01% catechin addition and 60 min heating time, the contents of BaP, BbF, BaA, and Chr increased with an increase in temperature. At 160 degC, they reached maximum values of 2.87, 2.94, 4.22, and 14.12 mg/kg, respectively, and then began to decrease. Therefore, 160degC was chosen as the experimental temperature for subsequent experiments. Additionally, the PAH4 contents showed the same trend with an extension in heating time when the catechin addition was 0.01% and the heating temperature was 160 degC . After heating for 20 min, the PAH4 contents increased significantly compared with the control group (p < 0.01). Similar findings were found in the study of Li et al. . In this study, the changes in PAH4 in C. oleifera oil during high-temperature oxidation were monitored under the rancimat pro-oxidation model, which is based on the principle of passing purified air into a sample that has reached a set heating temperature; the sudden increase in the conductivity of deionized water due to the release of volatile acids during oxidation is indicated as the end of the induction period, which can be defined as a measure of the antioxidant properties of the oil . The IPs of C. oleifera oil used in this study were 110 min, 20 min, 10 min, and 6 min at 140 degC, 160 degC, 180 degC, and 200 degC, respectively. When the temperature of the system was higher than 160 degC, the contents of PAH4 increased first, and then decreased with an increase in temperature and time. This may have been due to the rapid decomposition and oxidation of oil in the induction period, leading to the rapid formation of PAHs in this system; meanwhile, in the later oxidation period, under the strong high-temperature reaction, PAHs may react with other components in the oil to produce other substances, resulting in a reduction in the PAHs already generated . Therefore, the PAH4 in samples at 180 degC and 200 degC were lower than that at 160 degC when heated for 60 min. Furthermore, the PAH4 contents continued to increase in the first 20 min of the induction period when heated at 160 degC, and then during the period of 20-60 min, the oil was completely oxidized, and the contents of the PAH4 generated in the early stage gradually decreased under the severe reaction conditions, along with other components. Based on the above research, it could be concluded that temperature was an important factor in the formation of PAHs, which was also confirmed in a previous study . As a result, the temperature was relatively lower at 140 degC compared to 160 degC, and the rate of increase of PAH4 contents was lower than that for PAH4 in the sample at 160 degC. Combined with the above experimental results, we chose 160 degC and 20 min as our experimental conditions. Considering the results of both heating temperature and heating time, the formation of PAHs was influenced by both factors. Furthermore, the addition of catechin may have also had an effect on PAH4 production. For the purpose of finding the most effective concentration for PAH4 production, different additions of catechin (0.005%, 0.01%, 0.02%, 0.04%) were added. The samples were heated at 160 degC for 20 min, and the results are shown in Figure 2C. It was found that catechin had a significant effect (p < 0.005) on the increase in the formation of each PAH at additions of 0.005% and 0.01%. Most of the PAHs decreased slightly at a 0.02% concentration of catechin, with a significant decrease observed at a 0.04% concentration. Therefore, it was speculated that some cleavage and polymerization reactions of catechin at certain concentrations during the oxidation of oil at high temperatures may have led to an increase in PAH4 contents in oil. The generation of PAHs in oil at high temperatures is mainly due to the cleavage and oxidation of triglycerides in oil, resulting in the production of cyclized alkane compounds; these, in turn, lead to the generation of a large number of aldehydes, ketones, and hydrocarbons; at the same time, this generates a large number of free radicals, eventually leading to a significant increase in PAHs . However, it is very difficult to study the formation process of PAHs via intermediates because some intermediates formed by PAHs are very active and the process is transient. Only some relatively stable intermediates can be used to speculate on the formation process of PAHs. Therefore, in order to investigate the main mechanism of the increase in PAHs contents in C. oleifera oil caused by catechin, the roles of catechin addition on fatty acid composition, PAHs intermediates, free radical intensity, and quantity of C. oleifera oil at high temperature were further determined. 3.3. Effect of Catechin on Fatty Acid Composition during Heat Treatment In order to explore the effect of catechin on the fatty acid stability of C. oleifera oil under high-temperature conditions, the changes in fatty acid composition under different conditions of catechin addition were studied. As shown in Table 3, after heating at different temperatures for 60 min, the contents of oleic acid and linoleic acid decreased significantly with an increase in temperature, and the contents of linoleic acid decreased. As the heating time increased from 0 to 60 min, samples with 0.01 % catechin at 160 degC showed a significant increase in palmitic and stearic acids from the 20th min and a significant decrease in oleic and linoleic acids from the 15th min. This result was in keeping with the results of Bansal et al. . Under high temperature, unsaturated fatty acids in C. oleifera oil are very unstable and highly susceptible to oxidation and degradation; thus, some unsaturated fatty acids will generate saturated fatty acids and lead to an increase in palmitic and stearic acid contents. When the samples without catechin were compared with those containing catechin, significant changes in fatty acids began to occur after heating at 160 degC for 20 min (p < 0.05). When catechin was added at 0.02% and 0.04%, the contents of oleic acid and linoleic acid were higher than those of the control group, which indicated that catechin could inhibit the degradation of unsaturated fatty acids at these concentrations. When catechin was added at 0.01%, the contents of oleic acid were significantly lower than those of the control group, and the contents of linoleic acid were lower than that of the control group when catechin was added at 0.005 %. These results were similar to those obtained by Yuan . This indicated that catechin may affect fatty acid decomposition and oxidation after heat decomposition when the addition amount of catechin was 0.005 or 0.01%. 3.4. Effect of Catechin on PAHs Intermediates during Heating The research found that the degradation of oil in the process of the high-temperature thermal reaction itself produced free radicals, oleic acid, and other unsaturated fatty acids that are oxidized to generate hydroperoxides. Hydroperoxides are extremely unstable, and will be further cleaved to produce small molecules of substances, generating hydroxyl radicals and alkoxy radicals, alkoxy radicals further break down, thereby generating aldehydes, olefins, and other compounds . Aldehydes are important substrates for some hazard factors such as heterocyclic amines, polycyclic aromatic hydrocarbons, etc. Liu Li found that when aldehydes were added to oil alone, a significant increase in PAHs occurred in the system . Studies also have shown that the addition of nutrients, such as tea polyphenols to oil, can intensify the production of aldehydes in oil at high temperatures . Therefore, it was presumed that the addition of catechin to C. oleifera oil could promote the production of aldehydes and eventually lead to an increase in PAH4 contents. Figure 3A displays the graph of the relative content changes of aldehydes (RCA) generated in C. oleifera oil after heating at different temperatures (140 degC, 160 degC, 180 degC, and 200 degC) for 60 min. A total of 13 aldehydes were detected in this experiment, including nonanal, octanal, 2-undecenal, etc. It can be seen that high-temperature treatment led to a large number of aldehydes generated, and the overall contents levels also differed; the RCA continued to increase with increasing temperature during the first 20 min induction period of oil oxidation, from 0% to 31.92, 49.1, 54.31, and 69.12%, and the increasing trend was similar to that of PAH4. When in the oxidation period of oil, the primary oxidation products generated in the induction period were rapidly degraded to generate secondary oxidation products, and volatile compounds such as aldehydes were produced in large quantities in this phase, so that RCA continued to increase; however, the PAHs generated in the induction phase were rapidly depleted in this phase, and the rate of PAHs consumption in this phase was faster than the rate of PAHs production from aldehydes. Thus, this may have led to a continuous increase in aldehydes, but a decrease in PAH4. Figure 3B shows the changes in RCA produced by the heat treatment of C. oleifera oil with different additions of catechin. It can be seen that the RCA were higher in the samples with 0.005% and 0.01% catechin (59.94% and 55.01%), which were significantly higher than that of the control group (p < 0.05). This was probably attributed to the addition of 0.005% and 0.01% catechin that caused the oxidative degradation of fatty acids, which led to an increase in RCA. When the samples with 0.02% and 0.04% of catechin were added, the RCA were 38.17% and 39.86%, respectively, which were lower than the control group; this may have been due to the fact that the high concentration of catechin captured the aldehydes in the system, and thus reduced the contents of aldehydes in the samples. 3.5. Electron Spin Resonance Studies on the Mechanism of PAH Formation ESR is a modern analytical method for the direct and effective detection of compounds containing unpaired electrons . The free radicals generated during the oxidation of oil can be trapped, which is also particularly valuable for studying the mechanism of free radical generation with PAHs. Figure 4A shows the total number of free radicals at different temperatures (140, 160, 180, and 200 degC) with different heating times. It can be seen from the figure that the higher the temperature, the higher the total number of spins at the same heating time; the total number of radicals increased to 2.94, 3.79, 5.21, and 7.12, respectively, at 10 min, and the number of radicals also showed a trend of increasing and then decreasing with time. Based on the above results, it can be seen that during the induction period of oil oxidation, the free radicals were more active and the total number of spins increased, while no free radicals were detected after heating at 160 degC, 180 degC, and 200 degC for about 20 min. This may have been due to the fast bimolecular reaction rate between the newly generated lipoxy radicals and radical adducts, leading to rapid quenching of radical adducts in the system. According to the preliminary literature review and experiments, we hypothesized that the increase in PAHs due to catechin addition was related to the free radicals generated by catechin. Figure 4B shows the free radical intensity of oil with different additions of catechin after heating at 160 degC for 20 min; the higher the peak, the more free radicals in the analytes. Therefore, it can be seen from the graph that the intensities of free radicals when catechin was added at 0.005% and 0.01% were higher than the control group, and with an increase in catechin addition, catechin showed an inhibitory effect on free radicals. According to previous studies, phenolic-rich components have a better inhibitory effect on the formation of PAHs . However, in this study, it was found that at certain concentrations, the addition of catechin not only increased the free radical contents in the samples, but also led to an increase in PAHs. Wang et al. found higher levels of both PAH8 and free radicals after margination of chicken wings with phenolic-rich beer than the control, and the free radical contents were positively correlated with the PAH8 contents . Similar results were found in a study by Jongberg et al. . This suggested that phenolic compounds may produce reactive oxygen species, such as hydrogen peroxide and hydroxyl radicals, through reductive coupling with oxygen at high temperatures, and caused oxidative damage to lipids as well as their own oxidation to phenoxy radicals, thus leading to a system more favorable to form PAHs. From the results of free radical detection combined with PAH4 generation, the generation of PAH4 at high temperature in the more active stage of free radicals was positively correlated with free radicals, and the intensity of free radicals in the catechin-added samples also showed a correlation with the contents of PAH4. 3.6. FT-IR Spectroscopy Analysis In order to more clearly investigate the role of catechin on the formation of PAHs during the heat treatment of C. oleifera oil, this study also used triglycerides as the object to study the liposome system and verified these results using FT-IR. The FT-IR spectra indicated specific functional groups and their infrared absorption wavelengths. Clearly visible variations based on the position and intensity of some bands were used to explain the structural changes of the triglycerides. As shown in Figure 5A, the spectrum of triglyceride containing four major peaks near 3007, 1742, 1236, and 721 cm-1 showed the stretching vibrational positions of the unsaturated CH=CH, C=O stretching, C-O stretching, and CH=CH bending out of the plane, respectively; the two strong peaks near 2852 and 2921 cm-1 corresponded to the asymmetric stretching and symmetric stretching of -CH2, respectively; 1462 cm-1 corresponded to -CH2 bending; 1376 cm-1 corresponded to -CH3 bending; and 1119 and 1097 cm-1 both corresponded to C-O stretching . In the figure, the higher the transmittance of the peak, the higher the contents of the specific functional groups. It was observed that the 2921 and 2852 cm-1 positions and transmittances of the samples, with and without catechin, were relatively stable after heat treatment because these two positions corresponded to the methyl and methylene groups, respectively, which are the most stable segments of triglycerides. It was also shown in several spectral bands studied that the transmittance of the peaks at 3007 and 721 cm-1 increased with heating time, and the spectra of the catechin-added samples showed higher transmittance than those of the non-catechin-added samples . These results confirmed the degradation and oxidation of the lipids occurred during heating, more hydrolysis and oxidation reactions occurred in the functional groups of the oil samples, and the added catechin at high temperature that may be coupled with oxygen reduction to produce reactive oxygen species, such as hydrogen peroxide and hydroxyl radicals, causing oxidative damage to lipids. This result was consistent with the result of Section 3.3 in this study. It is worth mentioning that the transmittances and peak positions at 1742 and 1236 cm-1 in the spectra of the oil samples after heat treatment also changed, with the peak at 1742 cm-1 reaching the lowest transmittance after 40 min of heating with the addition of catechin, followed by the samples at 10 min and 20 min, and the transmittance of the two samples were relatively close . This indicates that the amount of stretching C=O vibrations of their ester-carbonyl chromophores increased slightly over time. This may have resulted from the breakdown of the acylglycerol ester bond and the oxidation of the C=C group to the C=O bond, causing the peak position shifted to 1744 cm-1. This change may be related to the overlap of the functional group of the aldehyde or other secondary oxidation products with the ester carbonyl functional group of the triglyceride. In the induction period of the first 20 min, aldehydes and other oxidation products were gradually generated, while some of them are transformed into other substances. In the oxidation period after 20 min, aldehydes and other oxidation products were generated in large quantities, so the transmittance continued to decrease. This result was consistent with the result of Section 3.4 in this study. Similar results were observed for the 1236 cm-1 bands, where the lowest transmittance was observed after the addition of catechin for 40 min, followed by the samples at 10 min and 20 min, and the peak positions were shifted to 1235 cm-1 . However, these changes were due to the breakdown of C=C bonds and the formation of C-H units. Although the spectra in this experiment could not directly show the formation of PAHs, the FT-IR spectral results were consistent with the above chemical analysis results of the oxidative degradation of fatty acids, free radical changes, formation of PAHs intermediates, and degradation and oxidation of catechin, which further confirmed the influence of catechin on PAH formation. 3.7. Analysis of Catechin in Thermal Processing Figure 6A shows the changes in catechin contents during the heating of C. oleifera oil containing 0.01% catechin at different temperatures (140 degC, 160 degC, 180 degC, and 200 degC) for 60 min. As seen in the figure, the contents of catechin decreased significantly (p < 0.05) with time at all temperatures. At the high temperatures of 180 degC and 200 degC, the catechin contents decreased rapidly when heated for about 15 min, and completely degraded by 20 min, which may have been due to the volatilization of catechin and the loss of thermal reactions dominated in the high-temperature system, resulting in faster consumption of catechin. Similar results emerged in Liu's study of TBHQ loss patterns in heating systems . It was also found that the separately added catechin produced a new substance, named catechin 1, during the heating. Figure 6B shows the liquid chromatograms of 0.01% catechin when heated at 160 degC for 0, 5, 10, 20, 40, and 60 min. The retention time of catechin was 13.92 min, and the retention time of catechin 1 was 14.31 min. It can be seen that the peak represented by the catechin gradually decreased in the first 20 min, and a new peak appeared behind catechin at about 20 min. At 40 min, both catechin and catechin 1 were degraded and absent, thus, it was hypothesized that catechin 1 was the oxidation product of catechin in this system. The catechin 1 was then isolated and detected using UPLC-QTOF-MS . According to the mass-to-charge ratio (m/z) and the previously reported tea polyphenols under the action of free radicals, the hydrogen of the phenolic hydroxyl group of the A ring itself is converted into phenoxy free radicals, and then further converted into quinones, so it was speculated that catechin 1 was a quinone converted from catechin after oxidation . When all the catechin was degraded, the quinones disappeared quickly. The overall results showed that throughout the thermal reaction, catechin participated in some chemical reactions to degrade themselves; moreover, by comparing the amount of PAH4 production and performing correlation analysis, it was found that the PAH4 contents continued to increase during the phase of gradual degradation of catechin, and the catechin contents showed a significant negative correlation with PAH4 production when the catechin addition was <0.02%. During the induction period of lipid oxidation, PAH4 were generated rapidly. Figure 7 shows the speculative path of PAH4 formation in C. oleifera oil. When catechin addition was <0.02%, more free radicals were produced than quenched, resulting in free radicals that attacked triglyceride and caused the oxidation and decomposition of fatty acids. Meanwhile, the hydrogenoxides produced may also generate cyclic compounds, such as cyclohexene through intramolecular cyclization, which eventually generated benzene rings. Molecular growth was then achieved through HACA (hydrogen atoms removed-acetylene molecules added) mechanism to form PAHs . This led to an increase in PAHs intermediates and PAH4. Moreover, unstable groups such as C-O, C-OH, and C-C in the benzene ring side chain of catechin itself would break at high temperature, and the C=C bond on the benzene ring skeleton would break and polymerize to form naphthalene, anthracene, and other dense aromatic ring compounds , which eventually led to an increase in PAH4 in the system. 4. Conclusions In summary, this study suggested that oxidative decomposition of catechin at high temperature may be involved in the formation of PAHs. The results showed that the rapid growth of PAH4 in crude oil was associated with phenolic compounds compared to refined oil. Therefore, catechin was selected as the subject after the experiment and the effect of catechin on PAH4 formation in C. oleifera oil under high temperature was investigated. It was found that the formation of PAHs was inhibited when the catechin addition was >0.02%, while it was involved in the formation of PAHs when the addition of catechin was <0.02%. We further found the oxidative cleavage reaction of catechin at high-temperature, which generated free radicals attacking triglycerides and promoted the formation of PAHs intermediates. As well as the cyclization reaction of catechin itself after cleavage, it led to the increase of PAH4 content in the system. Therefore, in order to control the formation of PAHs in oil rich in phenolic compounds, prolonged use of oil at high temperature (>140 degC) is not recommended. When the oil uses phenol as an antioxidant, the phenol, with its better heat resistance, should be selected in the process of high-temperature use, and the appropriate addition amount should be screened out. Acknowledgments The raw material of this work is provided by Zhejiang Jiusheng Oil Tea Technology Co., Ltd. (Industrial Zone of Datong Town, Hangzhou 311614, China). Supplementary Materials The following supporting information can be downloaded at: The supplementary experimental method is the specific details of Section 2.4 related experiments. Click here for additional data file. Author Contributions Conceptualization, R.L.; Validation, J.W. and L.Z.; Formal analysis, W.P.; Investigation, W.P.; Data curation, W.P. and L.Z.; Writing--original draft, W.P.; Writing--review & editing, Y.G., M.C. (Minjie Cao), W.P. and X.W.; Supervision, R.L., M.C. (Ming Chang) and X.W.; Project administration, J.W.; Funding acquisition, R.L. and M.C. (Ming Chang). All authors have read and agreed to the published version of the manuscript. Data Availability Statement Data is contained within the article or Supplementary Materials. Conflicts of Interest The authors declare no conflict of interest. Abbreviations AOCS: American Oil Chemists Society; AV: acid value; BaA: benzo(a)anthracene; BaP: benzo(a)pyrene; BbF: benzo(b)fluoranthene; CCO: crude C. oleifera oil; RCO: refined C. oleifera oil; Chr: chrysene; C. oleifera: camellia oleifera; HACA: hydrogen atoms removed-acetylene molecules added; IP: induction period; LOD, limit of detection; LOQ, limit of quantification; PAHs: polycyclic aromatic hydrocarbons; PAH4: sum of benzo(a)anthracene, benzo(b)fluoranthene, chrysene, benzo(a)pyrene; PAH8: sum of benzo(a)anthracene, chrysene, benzo(b)fluoranthene, ben-zo(k)fluoranthene, benzo(a)pyrene, dibenzo(a,h)anthracene, benzo(g,h,i)perylene, and indeno(1,2,3-c,d)pyrene; POV: peroxide value; FT-IR: Fourier Transform infrared spectroscopy; HPLC-FLD: high performance liquid chromatography-fluorescence detection. Figure 1 Changes of PAH4 in CCO and RCO heating for 6 h (A); The AV and POV changes of CCO and RCO during heating for 6 h (B,C); Effects of different addition of catechin, naringin, chlorogenic acid, and epigallocatechin on the formation of PAH4 in C. oleifera oil after heated at 180 degC (D). Bars with different letters show significant differences (p < 0.05). Figure 2 The PAH4 contents in C. oleifera oil at different heating temperatures under 0.01% catechin addition and 60 min heating time (A); The PAH4 contents in C. oleifera oil at different heating times under 0.01% catechin addition and 160 degC heated (B); and the PAH4 contents in C. oleifera oil at different catechin addition, heated at 160 degC for 60 min (C). Bars with different letters show significant differences (p < 0.05). Figure 3 Relative contents of aldehydes in C. oleifera oil under different heating temperatures and times (A); Relative contents of aldehydes after adding different catechin addition and heated at 160 degC for 20 min (B). Bars with different letters show significant differences (p < 0.05). Figure 4 Total amounts of spin under different heating temperatures and times (A); ESR spectrum of different catechin addition at 160 degC for 20 min (B). Figure 5 FT-IR spectra of triolein added with catechin heated for different times at 160 degC (A); Several characteristic peaks of (A): 3007 cm-1 (B) (a), 721 cm-1 (B) (b), 1742 cm-1 (B) (c), 1236 cm-1 (B) (d). Figure 6 Contents of catechin at different temperatures with heating time (A); Chromatograms of catechin and catechin 1 at 160 degC for different heating times (B); Mass spectrum of catechin 1 and speculated fragment structure (C). Figure 7 Speculative path of PAH4 formation in C. oleifera oil. foods-12-00980-t001_Table 1 Table 1 Fatty acid composition and lipid companion contents of crude C. Oleifera oil and refined C. Oleifera oil during heating. Heating Time (h) Fatty Acid (g/100 g) CCO RCO C16:0 C18:0 C18:1 C18:2 C16:0 C18:0 C18:1 C18:2 0 8.11 +- 0.08 d 2.36 +- 0.08 a 80.15 +- 0.11 a 9.44 +- 0.06 a 9.82 +- 0.08 c 2.56 +- 0.06 a 78.18 +- 0.07 a 8.81 +- 0.11 a 1 8.29 +- 0.08 c 2.45 +- 0.04 a 79.37 +- 0.09 b 9.31 +- 0.08 ab 9.89 +- 0.08 bc 2.59 +- 0.04 a 78.07 +- 0.05 ab 8.71 +- 0.06 ab 2 8.42 +- 0.06 bc 2.57 +- 0.18 a 78.84 +- 0.13 c 9.16 +- 0.06 bc 10.02 +- 0.06 abc 2.67 +- 0.18 a 77.79 +- 0.06 bc 8.56 +- 0.06 bc 3 8.47 +- 0.04 abc 2.37 +- 0.08 a 78.58 +- 0.06 d 9.13 +- 0.06 c 10.04 +- 0.14 ab 2.70 +- 0.11 a 77.73 +- 0.13 bc 8.48 +- 0.13 bc 4 8.45 +- 0.11 abc 2.41 +- 0.06 a 78.42 +- 0.06 de 9.02 +- 0.05 cd 10.10 +- 0.02 ab 2.74 +- 0.12 a 77.67 +- 0.13 c 8.42 +- 0.09 c 5 8.51 +- 0.14 ab 2.36 +- 0.08 a 78.31 +- 0.11 ef 8.88 +- 0.08 d 10.11 +- 0.08 ab 2.71 +- 0.13 a 77.56 +- 0.18 c 8.38 +- 0.05 c 6 8.61 +- 0.09 a 2.53 +- 0.06 a 78.19 +- 0.08 f 8.60 +- 0.11 e 10.14 +- 0.078 a 2.78 +- 0.15 a 77.44 +- 0.28 c 8.35 +- 0.11 c Heating Time (h) Total Phenols Contents (mg/kg) Phytosterol Contents (mg/g) Tocopherol Contents (mg/kg) CCO RCO CCO RCO CCO RCO 0 36.92 +- 2.70 a 17.20 +- 1.98 a 3.87 +- 0.11 a 3.28 +- 0.16 a 249.09 +- 5.35 a ND 1 27.65 +- 2.65 b 14.42 +- 1.57 ab 3.70 +- 0.28 a 3.16 +- 0.29 a 165.56 +- 4.01 b ND 2 21.03 +- 2.76 c 13.21 +- 2.24 ab 3.51 +- 0.14 ab 3.11 +- 0.19 a 139.24 +- 3.27 c ND 3 20.12 +- 2.65 cd 12.13 +- 1.98 ab 3.37 +- 0.22 b 3.05 +- 0.14 a 121.77 +- 5.55 d ND 4 18.06 +- 2.09 d 11.10 +- 2.60 b 3.20 +- 0.78 bc 3.00 +- 0.56 a 80.52 +- 2.26 e ND 5 15.42 +- 1.82 e 10.16 +- 2.76 b 3.04 +- 0.16 c 2.94 +- 0.98 a 53.44 +- 4.94 f ND 6 14.89 +- 1.34 e 9.97 +- 2.84 b 3.00 +- 0.27 c 2.93 +- 0.41 a 43.58 +- 4.02 f ND Decreasing percentage (%) 59.7 42.0 22.5 10.7 82.5 0 Note: ND--not detected. The value carrying different letters are significantly different (p < 0.05) from control and each other when comparing all pairs of columns. Results presented are the means of three values followed by their standard deviation. n = 3. Decreasing percentage (%) = (C0 - C6)/C0. C0: The contents of lipid companion when heating time was 0 h; C6: The contents of lipid companion when heating time was 6 h. foods-12-00980-t002_Table 2 Table 2 Pearson correlation coefficients of heating time, fatty acid composition, lipid companion, and PAH4 in crude C. Oleifera oil. PAH4 Heating Time C16:0 C18:0 C18:1 C18:2 Total Phenols Phytosterol Tocopherol PAH4 1 - - - - - - - - Heating time 0.964 ** 1 - - - - - - - C16:0 0.974 ** 0.943 * 1 - - - - - - C18:0 0.534 0.457 0.699 1 - - - - - C18:1 -0.992 ** -0.930 * -0.979 ** -0.606 1 - - - - C18:2 -0.827 -0.930 * -0.866 -0.512 0.780 1 - - - Total phenols -0.982 ** -0.912 * -0.983 ** -0.661 0.997 ** 0.771 1 - - Phytosterol -0.983 ** -0.993 ** -0.953 * -0.466 0.959 ** 0.885 * 0.942 * 1 - Tocopherol -0.998 ** -0.960 ** -0.977 ** -0.547 0.989 ** 0.834 0.980 ** 0.976 ** 1 Note: * p < 0.05; ** p < 0.01. foods-12-00980-t003_Table 3 Table 3 Fatty acid composition of C. oleifera oil added with catechin at different heating temperatures, heating times, and different addition of catechin. T (degC) t (min) Catechin Addition (%) Fatty Acid (g/100 g) C16:0 C18:0 C18:1 C18:2 0 60 0.010 8.11 +- 0.08 fg 2.36 +- 0.08 e 80.15 +- 0.11 a 9.44 +- 0.06 a 140 60 0.010 9.11 +- 0.08 b 2.53 +- 0.06 cde 78.77 +- 0.05 b 7.94 +- 0.29 c 160 60 0.010 9.27 +- 0.06 ab 2.77 +- 0.04 ab 78.25 +- 0.16 e 6.32 +- 0.27 d 180 60 0.010 9.30 +- 0.03 a 2.87 +- 0.01 a 78.20 +- 0.09 e 6.31 +- 0.23 d 200 60 0.010 9.35 +- 0.06 a 2.91 +- 0.04 a 78.15 +- 0.11 e 6.24 +- 0.21 d 160 0 0.010 8.11 +- 0.08 fg 2.36 +- 0.08 e 80.15 +- 0.11 a 9.41 +- 0.03 ab 160 5 0.010 8.17 +- 0.06 efg 2.40 +- 0.04 de 78.78 +- 0.04 b 9.36 +- 0.06 ab 160 10 0.010 8.23 +- 0.06 defg 2.41 +- 0.02 de 78.61 +- 0.10 bcd 9.28 +- 0.09 ab 160 15 0.010 8.28 +- 0.09 def 2.44 +- 0.02 de 78.58 +- 0.13 bcd 9.19 +- 0.06 ab 160 20 0.010 8.32 +- 0.06 de 2.57 +- 0.17 cd 78.39 +- 0.19 cde 8.94 +- 0.18 b 160 40 0.010 8.87 +- 0.13 c 2.63 +- 0.05 bc 78.35 +- 0.13 de 7.76 +- 0.37 c 160 60 0.010 9.27 +- 0.06 ab 2.77 +- 0.04 ab 78.25 +- 0.16 e 6.32 +- 0.27 d 160 20 0.000 8.09 +- 0.08 g 2.41 +- 0.04 de 78.72 +- 0.09 b 8.60 +- 0.23 a 160 20 0.005 8.39 +- 0.06 d 2.49 +- 0.04 cde 78.37 +- 0.09 de 8.31 +- 0.08 ab 160 20 0.010 8.32 +- 0.06 de 2.57 +- 0.18 cd 78.39 +- 0.19 cde 8.94 +- 0.19 b 160 20 0.020 8.22 +- 0.04 defg 2.37 +- 0.08 e 78.67 +- 0.04 bc 9.30 +- 0.02 ab 160 20 0.040 8.30 +- 0.12 de 2.41 +- 0.06 de 78.81 +- 0.09 b 9.37+-0.16 ab Note: The value carrying different letters are significantly different (p < 0.05) from control and each other when comparing all pairs of columns. Results presented are the means of three values followed by their standard deviation (n = 3). Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Lopez-Jimenez F.J. Ballesteros-Gomez A. Rubio S. Determination of polycyclic aromatic hydrocarbons (PAH4) in food by vesicular supramolecular solvent-based microextraction and LC-fluorescence detection Food Chem. 2014 143 341 347 10.1016/j.foodchem.2013.07.136 24054249 2. Wenzl T. Simon R. Anklam E. Kleiner J. Analytical methods for polycyclic aromatic hydrocarbons (PAHs) in food and the environment needed for new food legislation in the European Union TrAC Trends Anal. Chem. 2006 25 716 725 10.1016/j.trac.2006.05.010 3. Shadi A. Mazandarani M. Nikpour Y. Concentrations of Polycyclic Aromatic Hydrocarbons (PAHS) in Sediments of Khowre-Musa System (Persian Gulf) World 2012 4 83 86 4. Tfouni S.A. Souza N.G. Neto M.B. Loredo I.S. Leme F.M. Furlani R.P. Polycyclic aromatic hydrocarbons (PAHs) in sugarcane juice Food Chem. 2009 116 391 394 10.1016/j.foodchem.2009.02.040 5. Wongmaneepratip W. Vangnai K. Effects of oil types and pH on carcinogenic polycyclic aromatic hydrocarbons (PAHs) in grilled chicken Food Control 2017 79 119 125 10.1016/j.foodcont.2017.03.029 6. Garcia-Lomillo J. Viegas O. Gonzalez-SanJose M.L. Ferreira I.M. Influence of red wine pomace seasoning and high-oxygen atmosphere storage on carcinogens formation in barbecued beef patties Meat Sci. 2017 125 10 15 10.1016/j.meatsci.2016.11.009 27871036 7. Li Z. Impact on Trace Nutrients and Antioxidant Activity of Camellia Seed Oil in Processing Master's Thesis Jiangnan University Wuxi, China 2015 8. Wang C. Xie Y. Qi J. Yu Y. Bai Y. Dai C. Li C. Xu X. Zhou G. Effect of Tea Marinades on the formation of polycyclic aromatic hydrocarbons in charcoal-grilled chicken wings Food Control 2017 93 325 333 10.1016/j.foodcont.2017.12.010 9. Min S. Patra J.K. Shin H.-S. Factors influencing inhibition of eight polycyclic aromatic hydrocarbons in heated meat model system Food Chem. 2018 239 993 1000 10.1016/j.foodchem.2017.07.020 28873662 10. Britt P.F. Buchanan A.C. Owens C.V. Jr. Skeen J.T. Does glucose enhance the formation of nitrogen containing polycyclic aromatic compounds and polycyclic aromatic hydrocarbons in the pyrolysis of proline? Fuel 2004 83 1417 1432 10.1016/j.fuel.2004.02.009 11. Cheng W. Liu G. Wang X. Liu X. Liu B. Formation of Benzo(a)pyrene in Sesame Seeds During the Roasting Process for Production of Sesame Seed Oil J. Am. Oil Chem. Soc. 2015 92 1725 1733 10.1007/s11746-015-2734-0 12. Shi L.-K. Zhang D.-D. Liu Y.-L. Incidence and survey of polycyclic aromatic hydrocarbons in edible vegetable oils in China Food Control 2016 62 165 170 10.1016/j.foodcont.2015.10.037 13. Zhang Y. Zhai X. Gao L. Jin J. Zhong Q. Sun C. Yan L. Liu R. Akoh C.C. Jin Q. Quality of Wood-Pressed Rapeseed Oil J. Am. Oil Chem. Soc. 2017 94 767 777 10.1007/s11746-017-2986-y 14. Zhao J. Li X. Lu X. Zhang Q. Zhang W. Determination of four polycyclic aromatic hydrocarbons in edible oil by PLC with frozen degreasing J. Beijing Univ. Agric. 2022 37 113 116 15. Liu R. Liu R. Shi L. Zhang Z. Zhang T. Lu M. Chang M. Jin Q. Wang X. Effect of refining process on physicochemical parameters, chemical compositions and in vitro antioxidant activities of rice bran oil LWT 2019 109 26 32 10.1016/j.lwt.2019.03.096 16. Gao P. Cao Y. Liu R. Jin Q. Wang X. Phytochemical Content, Minor-Constituent Compositions, and Antioxidant Capacity of Screw-Pressed Walnut Oil Obtained from Roasted Kernels Eur. J. Lipid Sci. Technol. 2019 121 1800292 10.1002/ejlt.201800292 17. Xu L. Mei X. Chang J. Wu G. Zhang H. Jin Q. Wang X. Comparative characterization of key odorants of French fries and oils at the break-in, optimum, and degrading frying stages Food Chem. 2022 368 130581 10.1016/j.foodchem.2021.130581 34399180 18. Chen H. Free Radical Mechanism Study of Polar TAG Polymers Formed During Oil Thermal Oxidation Ph.D. Thesis Jiangnan University Wuxi, China 2019 19. Hu C. Kitts D.D. Evaluation of antioxidant activity of epigallocatechin gallate in biphasic model systems in vitro Mol. Cell. Biochem. 2001 218 147 155 10.1023/A:1007220928446 11330830 20. Zhang Y. Chen X. Zhang Y. Analytical chemistry, formation, mitigation, and risk assessment of polycyclic aromatic hydrocarbons: From food processing to in vivo metabolic transformation Compr. Rev. Food Sci. Food Saf. 2021 20 1422 1456 10.1111/1541-4337.12705 33506545 21. Koh E. Surh J. Food types and frying frequency affect the lipid oxidation of deep frying oil for the preparation of school meals in Korea Food Chem. 2015 174 467 472 10.1016/j.foodchem.2014.11.087 25529707 22. Aladedunye F.A. Curbing thermo-oxidative degradation of frying oils: Current knowledge and challenges Eur. J. Lipid Sci. Technol. 2015 117 1867 1881 10.1002/ejlt.201500047 23. Esposto S. Taticchi A. Di Maio I. Urbani S. Veneziani G. Selvaggini R. Sordini B. Servili M. Effect of an olive phenolic extract on the quality of vegetable oils during frying Food Chem. 2015 176 184 192 10.1016/j.foodchem.2014.12.036 25624222 24. Li T. Change of Bitter Peptide Content of Cold-Pressed Flaxseed Oil during Accelerated Oxidation Master's Thesis Jiangnan University Wuxi, China 2019 25. Garcia-Moreno P.J. Perez-Galvez R. Guadix A. Guadix E.M. Influence of the parameters of the Rancimat test on the determination of the oxidative stability index of cod liver oil LWT 2012 51 303 308 10.1016/j.lwt.2012.11.002 26. Liu J. Shen M. Liu X. Liang L. Wu Y. Zhang J. Xu X. Liu G. The loss and fate of BaA, Chr, BbF, and BaP (PAH4) tracked by stable isotope during frying Food Chem. 2021 374 131769 10.1016/j.foodchem.2021.131769 34920410 27. Kamal N.H.A. Selamat J. Sanny M. Simultaneous formation of polycyclic aromatic hydrocarbons (PAHs) and heterocyclic aromatic amines (HCAs) in gas-grilled beef satay at different temperatures Food Addit. Contam. Part A 2018 35 848 869 10.1080/19440049.2018.1425553 28. Guillen M.D. Goicoechea E. Palencia G. Cosmes N. Evidence of the Formation of Light Polycyclic Aromatic Hydrocarbons during the Oxidation of Edible Oils in Closed Containers at Room Temperature J. Agric. Food Chem. 2008 56 2028 2033 10.1021/jf072974h 18318493 29. Bansal G. Zhou W. Barlow P.J. Lo H.-L. Neo F.-L. Performance of palm olein in repeated deep frying and controlled heating processes Food Chem. 2010 121 338 347 10.1016/j.foodchem.2009.12.034 30. Yuan F. Study on the effect of Tea Polyphenol on the Oxidative Stability of Palm Oil and Instant Noodles during Frying of Instant Noodles Master's Thesis Henan University of Technology Zhengzhou, China 2021 31. Kandhro A. Sherazi S. Mahesar S. Bhanger M. Talpur M.Y. Rauf A. GC-MS quantification of fatty acid profile including trans FA in the locally manufactured margarines of Pakistan Food Chem. 2008 109 207 211 10.1016/j.foodchem.2007.12.029 26054282 32. Liu L. Effects of Heat Treatment on the Formation Regularity of Benzo(a)pyrene in Edible Oil Master's Thesis Yangzhou University Yangzhou, China 2017 33. Liu S. Regulation Mechanism of Tea Polyphenols on Heterocyclic Aromatic Amines Formation in Grilled Mutton Patties Master's Thesis Bohai University Jinzhou, China 2021 34. Zhao B. Application of Electron Spin Resonance (ESR) Technology in Biology and Medicine Chin. J. Magn. Reson. 2010 27 51 67 35. Viegas O. Yebra-Pimentel I. Martinez-Carballo E. Simal-Gandara J. Ferreira I.M.P.L.V.O. Effect of Beer Marinades on Formation of Polycyclic Aromatic Hydrocarbons in Charcoal-Grilled Pork J. Agric. Food Chem. 2014 62 2638 2643 10.1021/jf404966w 24605876 36. Wang C. Xie Y. Wang H. Bai Y. Dai C. Li C. Xu X. Zhou G. The influence of natural antioxidants on polycyclic aromatic hydrocarbon formation in charcoal-grilled chicken wings Food Control 2018 98 34 41 10.1016/j.foodcont.2018.11.012 37. Jongberg S. Torngren M.A. Gunvig A. Skibsted L.H. Lund M.N. Effect of green tea or rosemary extract on protein oxidation in Bologna type sausages prepared from oxidatively stressed pork Meat Sci. 2013 93 538 546 10.1016/j.meatsci.2012.11.005 23273462 38. Oussama A. Elabadi F. Platikanov S. Kzaiber F. Tauler R. Detection of Olive Oil Adulteration Using FT-IR Spectroscopy and PLS with Variable Importance of Projection (VIP) Scores J. Am. Oil Chem. Soc. 2012 89 1807 1812 10.1007/s11746-012-2091-1 39. Liu C. Migration and Transformation Rules of TertiaryButyl Hydroquinone (TBHQ) and Its Antioxidation Mechanism under Thermal Treatment Master's Thesis Henan University of Technology Zhengzhou, China 2013 40. Sang S. Tian S. Wang H. Stark R.E. Rosen R.T. Yang C.S. Ho C.-T. Chemical studies of the antioxidant mechanism of tea catechins: Radical reaction products of epicatechin with peroxyl radicals Bioorg. Med. Chem. 2003 11 3371 3378 10.1016/S0968-0896(03)00367-5 12878131 41. Chen B.H. Chen Y.C. Formation of Polycyclic Aromatic Hydrocarbons in the Smoke from Heated Model Lipids and Food Lipids J. Agric. Food Chem. 2001 49 5238 5243 10.1021/jf0106906 11714310 42. Peng Y. Wu S. The structural and thermal characteristics of wheat straw hemicellulose J. Anal. Appl. Pyrolysis 2010 88 134 139 10.1016/j.jaap.2010.03.006 |
PMC10000654 | Healthcare (Basel) Healthcare (Basel) healthcare Healthcare 2227-9032 MDPI 10.3390/healthcare11050672 healthcare-11-00672 Article Comparing Full and Pre-Term Neonates' Heart Rate Variability in Rest Condition and during Spontaneous Interactions with Their Parents at Home Kokkinaki Theano Conceptualization Methodology Writing - original draft Writing - review & editing Supervision 1* Markodimitraki Maria Formal analysis Investigation 2 Giannakakis Giorgos Methodology Formal analysis Investigation Writing - original draft 3 Anastasiou Ioannis Formal analysis Investigation 4 Hatzidaki Eleftheria Conceptualization Methodology Writing - original draft 5 Mohamed Abdel-Latif Academic Editor 1 Child Development and Education Unit, Laboratory of Applied Psychology, Department of Psychology, University of Crete, 74150 Rethymnon, Greece 2 Department of Preschool Education, University of Crete, 74150 Rethymnon, Greece 3 Institute of Computer Science, Foundation for Research and Technology, 70013 Heraklion, Greece 4 Cardiology Department, University Hospital of Heraklion, University of Crete, 71500 Heraklion, Greece 5 Department of Neonatology/Neonatal Intensive Care Unit, University Hospital of Heraklion, School of Medicine, University of Crete, 71500 Heraklion, Greece * Correspondence: [email protected]; Tel.: +30-28310-77536 24 2 2023 3 2023 11 5 67221 1 2023 18 2 2023 22 2 2023 (c) 2023 by the authors. 2023 Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ). Background: Preterm neonates show decreased HRV compared to those at full-term. We compared HRV metrics between preterm and full-term neonates in transfer periods from neonate rest state to neonate-parent interaction, and vice versa. Methods: Short-term recordings of the HRV parameters (time and frequency-domain indices and non-linear measurements) of 28 premature healthy neonates were compared with the metrics of 18 full-term neonates. HRV recordings were performed at home at term-equivalent age and HRV metrics were compared between the following transfer periods: from first rest state of the neonate (TI1) to a period in which the neonate interacted with the first parent (TI2), from TI2 to a second neonate rest state (TI3), and from TI3 to a period of neonate interaction with the second parent (TI4). Results: For the whole HRV recording period, PNN50, NN50 and HF (%) was lower for preterm neonates compared to full-terms. These findings support the reduced parasympathetic activity of preterm compared to full-term neonates. The results of comparisons between transfer period simply a common coactivation of SNS and PNS systems for both full and pre-term neonates. Conclusions: Spontaneous interaction with the parent may reinforce both full and pre-term neonates' ANS maturation. heart rate variability (HRV) time-domain indices frequency-domain indices non-linear measurements very low-frequency band (VLF) PNN50 preterm neonates full-term neonates spontaneous neonate-parent interaction University of Crete10792-668/08.02.2021 This work is funded by the Special Account for Research Funds of University of Crete, Grant Number: 10792-668/08.02.2021. The ACP was funded by the Special Account for Research Funds of University of Crete. pmc1. Introduction Heart rate variability (HRV) constitutes a non-invasive biomarker and refers to the physiological fluctuations and time intervals between spontaneous and successive heartbeats . HRV is widely used to efficiently assess the regulatory activity of the autonomic nervous system (ANS) by its sympathetic and parasympathetic components and makes it possible to evaluate the balance between these two branches within the ANS . HRV analysis has diagnostic and prognostic potential for detecting and monitoring dysregulation due to disease in neonates, and gives paramount hints about the newborns' wellbeing and socioemotional and cognitive development . The ANS undergoes significant maturation between 38-weeks' gestation . The sympathetic system develops early in pregnancy, while parasympathetic control emerges later in the perinatal period . Gestational (GA) and postmenstrual age (PMA) have the largest influence on HRV . The ANS of preterm infants is underdeveloped and the multiple control loops responsible for homeostasis may not yet work synergistically . Prematurity delays maturation of HRV (Fyfe et al., 2014) and preterm birth has been associated with decreased HRV (Aye et al., 2018). Lower HRV values indicate abnormal adaptation with impaired function of the ANS and vulnerability to stress, while an increase in HRV represents physical and mental adaptability along with efficient autonomic mechanisms . Furthermore, environmental challenges in the postnatal days play a crucial role in the development of the parasympathetic system and the maturational course of sympathetic regulation may be altered by physiological challenges in the NICU . Preterm infants in the Neonatal Intensive Care Unit (NICU) experience chronic exposure to stressors . When the underdevelopment of the ANS of preterm infants and NICU stressors are combined, ANS maturation of preterm neonates may be further delayed and impaired, with consequences on their overall development that persist later in life . Interventions are needed to reduce the adverse environmental impacts on ANS development to mitigate exposure to stressors in NICUs and to enhance maturation of the ANS of preterm neonates . 1.1. HRV Variations between Pre-Term Neonates/Infants Evidence based on measures of neonates, mainly in the NICU or in a laboratory setting, shows that HRV of preterm infants is less complex and slower compared to full-term neonates at the same postmenstrual age . At birth and within the first weeks of life, preterm infants display lower scores in certain time-and frequency-domain parameters of HRV: in mean RR, lower values of root mean square of the difference between adjacent NN intervals (RMSSD), standard deviation of the NN intervals (SDNN), total power (TP) and very low (VLF) frequency power. The most significant differences have been found in the high frequency power parameter (HF), which increases with gestational age. Preterm neonates have higher or lower low frequency power values (LF) compared to full-terms . As for the relative changes (%), the power in the HF and LF spectrum revealed the most marked increase with gestational age . Evidence on the LF/HF ratio is contradictory. There is a negative correlation with gestational age at birth and the LF/HF ratio was higher in preterm infants , or the LF/HF ratio does not differ significantly between preterm and full-term infants . Furthermore, decreased complexity of HRV dynamics in preterm compared to full-term infants is evidenced by non-linear indices, as this has been shown by only three relevant studies. Compared to full-terms, preterm infants have more linear and less chaotic patterns, smaller values of sample entropy, higher values of a1 but no variations in a2 .HRV variations between preterm and full-term infants are evident right after birth; they remain at preterm theoretical term age and they persist even beyond term-equivalent age . Methodological variations and the lack of consensus in neonatal HRV analysis makes synthesis and comparisons between investigations very difficult, if not impossible . However, the above review provides evidence that, generally, premature infants show decreased HRV compared to full-terms according to differences in time and frequency domain parameters and in non-linear indices. These variations imply that early in life, compared to the HRV of term counterparts, HRV of preterm infants is characterized by a reduction in sympathetic, and even more markedly, parasympathetic activities, and a relative sympathovagal imbalance, which results in the impaired function of ANS . 1.2. HRV of Pre-Term Infants in Different Contexts and Conditions and in Interactions with Their Parents Evidence of HRV variations in preterm and full-term infants in different conditions/contexts, involving mostly mother-infant sensory stimulation, is rarely investigated in the naturalistic environment, therefore studies are limited and the evidence is contradictory. In particular, in preterm infants(aged 33 weeks), the LF/HF ratio was similar during caregiving epochs and sleep epochs, though the LF/HF ratio increased during periods of caregiving for massage-treated male infants. This suggests an increase in sympathetic response during a physiologically demanding time period . In the first 4post-term months, HF of preterm infants was higher during pre-feeding, decreased during feeding and returned to the pre-feeding level during post-feeding, though LF did not show a similarly consistent pattern. This shows an adaptive response to stimulation that requires increased attention, or metabolic output . Furthermore, transfer of the preterm infant (34 weeks) from the open-crib to Kangaroo Care (KC, sensory stimulation from being in skin-to-skin contact with the mother's chest) decreased the values of LF and HF and, conversely, the LF/HF ratio was higher in KC. Overall, KC produced changes in HRV that indicated a decrease in stress . In the course of maternal natural breathing and physical contact, the LF power in 3-5-month-old full-term infants did not differ according to the period (pre-rest, respiration, post-rest). No correlation was found between the mothers' LF power and 3-5-month-old infants' LF power during the paced breathing period. Young infants showed a delayed increase in the LF components after termination of maternal-paced breathing, possibly due to their immaturity . It is difficult to integrate these results due to the methodological heterogeneity of relevant studies. Despite that, this literature provides evidence of HRV variations in pre-term and full-term neonates according to various contexts of sensory stimulation. 1.3. Research Questions In this study, we addressed the following research questions: Do HRV parameters (time-domain, frequency-domain indices and non-linear measurements) of preterm and full-term infants vary in the total duration of three transfer periods from rest state to spontaneous interaction with the parent, and vice versa? Do HRV parameters of preterm and full-term infants vary between three transfer periods, that is between: (a) rest state 1 (TI1) and spontaneous interaction between neonate-parent 1 (mother or father) (TI2), (b) spontaneous interaction between neonate-parent 1 (TI2) and rest state 2 (TI3), and (c) rest state 2 (TI3) and spontaneous interaction between neonate-parent 2 (mother or father) (TI4)? 2. Material and Methods 2.1. Participants One hundred and two mothers, fathers and neonates participated in the study in two groups. The first group included 18 parents and their infants born at full-term >= 37 weeks gestational age (GA) with no medical complications. The second group consisted of 28 parents and their preterm infants. Ninety three percent of preterm infants included in this study were moderate-to-late pre-terms (32-36 weeks GA) and only 7% of them were healthy pre-terms with a GA of 31 weeks. Exclusion criteria included: perinatal asphyxia, neurological pathologies, malformation syndromes and major malformations, sensory deficits, metabolic genetic disease, or CNS infection. Six mothers in the full-term group and three mothers in the preterm group were not included in the final sample due to: delayed answer to the researcher for participation in the study, neonates' hospitalization, or time constraints. No differences were observed between participating and non-participating families on family demographic, or infant medical status. Demographics and infant medical status of the two groups are reported in Table 1 and Table 2. The data show no group differences in maternal and paternal education years. Mothers and fathers of premature neonates were slightly older than parents of full-terms. All families were middle-class , both parents were older than 20 years of age, they did not suffer from a psychiatric illness, and they did not have issues with drug or substance abuse; mothers were married to the child's father and in all families at least one parent was employed. 2.2. Procedure After ethical approval (see in notes), parents were approached shortly after birth at the Neonatal Intensive Care Unit (NICU) of the Neonatology Clinic of the General University Hospital of Crete (Greece) (preterm) and at private Maternity/Gynecological Clinic Mitera of Heraklion (full-term). Firstly, the medical staff of the above clinics asked the parent's consent to provide their communication information to the members of the research team. After parental consent, a neonatologist and a psychologist (both members of the research team) informed the parents about the aim and the procedure of the study. Parents who accepted to participate were asked to sign the consent form. In the course of the same meeting, parents were asked to answer questions regarding family sociodemographic characteristics and the neonate's birth characteristics. Then, the first visit to the family's home for the video-recording was scheduled at a time when both parents would be available and when the neonate was expected to be alert. The video-recording was performed within the first four to five weeks after birth at term-equivalent age for both groups, that is, for preterm neonates at mean PMA 39.57 weeks (SD = 2.41, min-max = 37-45 weeks) and for full-term neonates at mean PMA 42.55 (SD = 1.75, min-max = 39-46 weeks). Newborns with a post-conceptual age of more than 38 weeks are relatively mature in terms of sympathovagal balance (Javorka et al., 2017).The whole recording lasted 30 min and it was segmented in three time intervals (TI) as follows TI1: resting state 1 (no neonate-parent interaction, HRV measurement of the infant in a supine position) (7 min), TI 2: interaction of the neonate with the first parent (8 min), TI3: resting state 2 (no neonate-parent interaction, HRV measurement of the infant in a supine position) (7 min), TI4: interaction of the neonate with the second parent (8 min). For the interaction, the only instruction given to the parents was: "Play as you normally do with your young baby". The recording took place in a room and at a position chosen by the parents prohibiting any third-party intervention. If the neonate became distressed, or either the parents, or the researcher considered that the visit should be postponed for some reason, it was rescheduled as soon as possible thereafter. 2.3. Heart Rate Variability Analysis 2.3.1. Heart Rate Variability Data Collection Neonate HRV measurements were carried out through SEER 1000, ECG Recorder, and General Electric (Version 1.0, 2067634-077 Revision F). The device was used by a trained operator under the direct supervision of a licensed healthcare practitioner. The device is suitable for use for pediatric patients, including those patients weighing less than 10 kg. For the data collection, the device was connected via Bluetooth to an Android mobile smartphone. Recording and HRV measurement was stopped if there was excessive restlessness or crying. 2.3.2. Heart Rate Variability Data Processing Once the recording was completed, an ECG analysis software (General Electric, Athens, Greece)package was used for data collection. The ECG preprocessing and the HRV parameters extraction analysis was performed using custom scripts written on the MATLAB r2018b platform. During the preprocessing phase, the ECG signal was detrended by subtracting time series polynomial fit or order 60. The R components of the QRS complex were detected and the RR Intervals (RRI) were calculated. The ectopic heartbeats (irregular heartbeats deviated from normal) were also detected and excluded by adopting the HRV signal approach (percentage change of 70% over the averaged previous 5 heartbeats). The whole preprocessing procedure is described in . 2.3.3. Heart Rate Variability Analysis Short-term recordings of HRV parameters of premature and full-term neonates were performed . Calculated HRV features were based on time-domain indices (quantification of the amount of HRV observed during monitoring periods), frequency-domain values (calculation of the absolute or relative amount of signal energy within component bands) and non-linear measurements (quantification of the unpredictability and complexity of a series of interbit intervals) (Table 3). 2.4. Statistical Analysis Data were tested for their normality using the Kolmogorov-Smirnov test. Firstly, they were analyzed, controlling for differences in their HRV parameters between the two groups (full-term and preterm infants) for the whole recording time and for each time interval separately, using the independent samples t-test or Mann-Whitney test. Secondly, the effect of parent interaction was investigated, controlling for pairwise differences between two conditions (no interaction, parent interaction) within each group (full-term and preterm infants) using Pairwise t-test or Wilcoxon signed-rank test. The statistical significance level was set to a = 0.05. All statistical analyses were performed using custom scripts in the MATLAB R2018b platform environment. It should be noted that we compared the NN50 between the two groups only in the total duration of the recording. The NN50 was excluded from the analysis between rest states 1 and 2 and interaction with the first/second parent as these had different durations. Only the pNN50 was utilized in these cases as it is not affected by the recording duration. 3. Results 3.1. Comparing HRV Parameters between Pre-Term Neonates in the Whole Recording Duration The comparison of HRV parameters between preterm infants for the whole recording duration showed that RMSSD (U = 344, z = 2.060, p = 0.039), the pNN50 (U = 344, z = 2.059, p = 0.037), the HF (%) (U = 341, z = 1.992, p = 0.046) and the VLF (%) (t(44) = 0.424, p = 0.046) of preterm neonates was statistically significantly reduced in relation to the full-terms (Table 4). 3.2. Comparing HRV Parameters of Pre-Term Neonates between Resting Condition 1 and Interaction with the First Parent The comparison of HRV parameters of full-term infants between resting condition 1 and interaction with the first parent (Table 5) shows that HRm(t(17) = -3.61, p = 0.002) and total power (z = -2.11, p = 0.035) was increased, while the VLF peak (z = 2.22, p = 0.026), DFA a (t(17) = 4.07, p = 0.001), DFA a1(t(17) = 2.44, p = 0.030), DFA a2 (t(17) = 3.48, p = 0.004) were significantly reduced. The comparison of HRV parameters of preterm infants between resting condition 1 and interaction with the first parent (Table 6) shows that the preterm HRm(t(27) = -3.45, p = 0.002), total power (t(27) = -2.64, p = 0.014) and VLF (%) (t(27) = -4.87, p < 0.001) was increased, while the DFA a2 (t(27) = 2.49, p = 0.021)was reduced. 3.3. Comparing HRV Parameters of Pre-Term Infants between Interaction with the First Parent and Resting Condition 2 For the interaction between the first parent and resting condition 2, preterm HRm(t(27) = 2.78, p = 0.010), total power (z =2.02, p = 0.043), and VLF (%) (t(27) = 3.35, p = 0.002) decreased, while LF (%) (t(27) = -2.366, p = 0.025) increased (Table 7). For the interaction between the first parent and resting condition 2, full-term infants' VLF and DFAa increased (Table 8). 3.4. Comparing HRV Parameters between Pre-Term Infants between Resting Condition 2 and Interaction with the Second Parent For the interaction between rest condition 2 and the second parent, VLF % increased (Table 9 and Table 10). However, LF% decreased only for full-terms and VLF peak decreased only for pre-terms. The significant HRV behavior parameters in the investigated interaction patterns (resting condition 1 (no interaction), interaction with the 1st parent, resting condition 2 (no interaction), interaction with the 2nd parent) are depicted in Figure 2. 4. Discussion We aimed to compare HRV parameters between full-term and preterm neonates, and between transfer periods from rest state to spontaneous interaction of neonates with their parents at home, and vice versa. A comparison of HRV parameters between preterm infants in the four time intervals, in total, showed that PNN50, NN50and HF (%) of preterm infants was significantly decreased compared to full-terms. This is consistent with findings showing that preterm infants score lower in time-domain parameters compared to full term infants, and with evidence showing that increasing prematurity has been associated with lower HF . The pNN50 is closely correlated with PNS activity and the HF band reflects parasympathetic activity . Reduced pNN50 and HF(%) of premature infants compared to full-terms may be attributed to the early disruption of autonomic development, which causes immaturity of ANS . The sympathetic system shows steady development throughout the fetal period and develops earlier than the parasympathetic system. The latter begins to develop during the first trimester and development continues until birth but it undergoes accelerated maturation at 25-32 weeks' gestation. The normal steep increase in vagal tone (which reflects the parasympathetic division activation) occurs around 37-38 weeks at a time when premature newborns may already have been in an ex utero environment. In infants born prematurely, the normal third trimester increase in parasympathetic tone may be dampened in the ex utero environment, compared to that of the inutero third trimester fetus . Stressful environmental stimuli in the NICU (e.g., invasive procedures, mechanical ventilation, loud noise, and bright lights) may have impeded normal maturation of the ANS . Deficits in HRV parameters in the preterm population may persist after birth up to term-equivalent age . We indicated that between TI1 and TI2, certain common HRV parameters changed in the same direction between pre-terms, while others varied. In particular, HR and total power increased and a2 decreased for both groups, while DFA, DFA a1 and VLF peak decreased for full terms, and VLF (%) increased for pre-terms. HR increase indicates a rise in SNS activity . Total power, the sum of the energy of VLF, LF, and HF bands for short-terms recordings, represents the overall variability . HF are expressions of PNS activation, while LF contains contributions of both the SNS and PNS influences . Thus, a total power increase implies coactivation of the SNS and PNS systems. Non-linear indices reflect the unpredictability of a time series, which results from the complexity of the mechanisms that regulate HRV . These measurements quantify the properties of heart rate dynamics, such as response patterns and self-correlation, which are caused by complex interplays between vagal and sympathetic regulations . In this context, a2 characterizes ultraslow changes in the heart rate (below the frequency of sympathetic tone) and reflects the regulatory mechanisms that limit fluctuation of the beat cycle . It is noted that a decrease in DFAa reflects an adverse adaptive situation not related to "slow recovery" processes or vagal activity (URL: mathnet.ru/php/archive.phtml?wshow=paper&jrnid=ivp&paperid=200&option_lang=eng accessed on 17 February 2023) Taken together, for both full and pre-term infants, the transfer period from TI1 to TI2 is associated with an increase in overall variability and coactivation of the SNS and PNS systems, along with a decrease in the regularity of heartbeats. An interesting finding of this study is that between TI1-TI2 and TI2-TI3, the HR of pre-terms changed but in opposite directions. In particular, in the transfer period, TI1-TI2, preterm HR increased, while in TI2-TI3, preterm HR decreased. Thus, between transfer periods TI1-TI2-TI3, we indicated a fluctuating SNS activation of preterm infants. Between TI3-TI4, LF (%) decreased only for full-terms, implying decreases in SNS and PNS influences. Given that an infant's state influences arousal, attention and affect , these patterns of SNS and PNS activation in TI3-TI4 may be attributed to fatigue of young infants after 22-min transfer periods [7 min (rest state 1) + 8 min (interaction) + 7 min (rest state 2)]. 5. Conclusions In accordance with the previous literature, a comparison of HRV parameters across the four time intervals showed lower scores in certain time-domain parameters (PNN50, NN50) and in HF (%), a frequency-domain parameter of preterm infants compared to full-term infants. These findings support the reduced parasympathetic activity of preterm compared to full-term neonates. Furthermore, HRV metric changes across the transfer periods from rest conditions to spontaneous neonate-parent interaction, and vice versa, imply a common coactivation of the SNS and PNS systems for both full and pre-term neonates (TI1-TI2), a fluctuating SNS activation for pre-terms (TI1-TI2 and TI2-TI3), and decreases in SNS and PNS activation for full-terms (TI3-TI4). 5.1. Limitations of This Study To deepen our understanding of HRV variations between preterm and full-term infants, larger samples are needed for the measurement of both long-term HRV metrics. An investigation into the correlation of HRV parameters with maternal lifestyle and delivery mode is needed .Due to the small sample size of preterm infants, we were not powered to detect differences between subgroups according to gestational age and we did not control variations in HRV parameters between preterm infants according to parent gender. This is important because mothers and fathers vary in the interactive patterns with their infants and between full-term and preterm infants . 5.2. Implications for Practice The findings of this study highlight the utility of HRV in neonatology and the importance of introducing the HRV in as many NICUs as possible in order to improve neonatal care . In order to enhance family-centered and family-integrated developmental care practices in the NICU, high priority should be given to facilitate and reinforce the parent-preterm infant physical and emotional closeness and parental involvement in the infant's care. In this context, the concept of parents as "partners in care" rather than "visitors" should be further supported. This will have long-term positive implications for infant development, benefits for parental mental health and for the development of parent-infant bonding, along with implications for the wellbeing of health professionals. Furthermore, it is vital to increase the awareness of healthcare specialists about the critical need to enable parents' access to the NICU and an active engagement of parents in the primary care of hospitalized newborns . Acknowledgments We are deeply indebted to the neonates and their families for offering their time, cooperation and patience to participate in this study. Author Contributions Conceptualization: T.K. and E.H.; Methodology: T.K., G.G. and E.H.; Formal analysis and investigation: M.M., G.G. and I.A.; Writing--original draft preparation: T.K., G.G. and E.H.; Writing--review and editing: T.K., G.G. and E.H.; Supervision: T.K. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement All parents of neonates gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Research Ethics Committee of the University of Crete (Project identification code: 46/15.04.2021), the University Hospital of Heraklion (Project identification code:471/14/09.06.2021) and the Maternity/Gynecological Clinic Mitera of Heraklion (Crete) (Project identification code: 491/21.1.2021). Informed Consent Statement Informed consent was obtained from all individual participants included in the study. Data Availability Statement The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Box-plots of the HRV parameters RMSSD, VLF, LF, HF for the whole recording duration. Asterisk denotes statistically significant difference at 0.05 level. Figure 2 HRV parameters (HRm, total power, DFA a2) behavior in the investigated interaction patterns (resting condition 1 (no interaction), interaction with the parent, resting condition 2 (no interaction), interaction with the parent). healthcare-11-00672-t001_Table 1 Table 1 Neonate medical variables of the two groups of the sample. Neonate Medical Variables Preterm Neonates (N = 28) Full-Term Neonates (N = 18) M SD range M SD range GA * (weeks) 34.03 1.52 31-36 38.61 1.09 37-40 PMA ** (weeks) at video-recording 39.57 2.41 37-45 42.55 1.75 39-46 Birth weight *** 2200 450.68 1520-3240 3310 208.12 2820-3750 Birth height 45.16 2.47 41-50 51.05 1.06 49-53 Male/Female ratio 19/9 10/8 Notes: * GA: Gestational age; ** PMA: Postmenstrual age; *** 23 neonates(82.1% of pre-terms) had birth weight < 2500 g and 8 neonates (28.5%) <2000 g. healthcare-11-00672-t002_Table 2 Table 2 Family demographic of the two groups of the sample. Parental Characteristics Families of Preterm Neonates (N = 28) Families of Full-Term Neonates (N = 18) M SD range M SD range Maternal age (years) 35.71 5.28 25-49 32.88 5.00 24-42 Maternal education (years) 15.28 2.50 6-18 15.88 2.32 12-18 Paternal age (years) 41.64 5.77 31-56 37.05 5.99 29-47 Paternal education (years) 14.78 2.79 6-18 15.00 2.49 12-18 healthcare-11-00672-t003_Table 3 Table 3 HRV parameters measured in pre-term infants. Parameter Definition Unit Time-domain HRm Mean heart rate bpm HRstd Standard deviation of instantaneous heart rate values bpm HRV triangular index Integral of the density of the RR interval histogram divided by its height - SDNN Standard deviation of NN intervals s rMSSD Root mean square of consecutive RR interval differences s NN50 Number of adjacent NN intervals that differ from each other by more than 50 ms - pNN50 Percentage of successive NN intervals that differ by more than 50 ms % Frequency-domain VLF_peak Peak frequency of the very low-frequency band Hz LF_peak Peak frequency of the low-frequency band Hz HF_peak Peak frequency of the high-frequency band Hz VLF (%) Normalized VLF power - LF (%) Normalized LF power - HF (%) Normalized HF power - LF/HF Ratio of LF-to-HF power - Total power Total power of the ECG spectrogram Hz Non-linear DFA a Detrended fluctuation analysis coefficient - DFA a1 Detrended fluctuation analysis, which describes short-term fluctuations - DFA a2 Detrended fluctuation analysis, which describes long-term fluctuations - healthcare-11-00672-t004_Table 4 Table 4 Comparison of HRV parameters between preterm infants for the whole recording duration. HRV Features Full-Term Neonates Preterm Neonates p-Value Difference HRm 157.5 162.9 0.134 ns SDNN 0.042 0.035 0.156 ns HR_std 18.9 16.2 0.359 ns RMSSD 0.037 0.022 0.039 | NN50 271.1 99.0 0.039 | pNN50 6.4 2.3 0.037 | HRV_Tri 8.6 8.0 0.465 ns VLF_peak 0.020 0.019 0.674 ns LF_peak 0.06 0.06 0.767 ns HF_peak 0.18 0.18 0.743 ns Total power 367.6 233.1 0.209 ns VLF (%) 0.359 0.448 0.046 | LF (%) 0.43 0.40 0.290 ns HF (%) 0.19 0.13 0.046 | LF/HF 3.59 4.87 0.180 ns DFA a 1.03 1.04 0.923 ns DFA a1 1.10 1.10 0.989 ns DFA a2 0.96 0.95 0.869 ns Note: Bold type denotes a statistically significant difference between groups. The differentiations are depicted in Figure 1. Ns means 'non-significant' and arrows show the direction of variation for a specific parameter. healthcare-11-00672-t005_Table 5 Table 5 Comparison of full-term infants' HRV parameters between resting condition 1 and interaction with the 1st parent. HRV Feature Resting Condition 1 Interaction between Full-Term Neonate-1st Parent p-Value Difference HRm 154.8 160.2 0.002 | SDNN 0.033 0.038 0.344 ns HR_std 14.9 16.8 0.247 ns RMSSD 0.029 0.039 0.112 ns pNN50 4.6 7.8 0.085 ns HRV_Tri 6.7 6.9 0.654 ns VLF_peak 0.037 0.035 0.026 | LF_peak 0.07 0.06 0.144 ns HF_peak 0.20 0.19 0.794 ns Total power 175.0 322.4 0.035 | VLF (%) 0.033 0.064 0.171 ns LF (%) 0.65 0.61 0.165 ns HF (%) 0.28 0.28 0.947 ns LF/HF 3.47 3.23 0.665 ns DFA a 1.02 0.95 0.059 ns DFA a1 1.01 0.96 0.405 ns DFA a2 1.03 0.91 0.004 | Note: Bold type denotes a statistically significant difference between groups. Ns means 'non-significant' and arrows show the direction of variation for a specific parameter. healthcare-11-00672-t006_Table 6 Table 6 Comparison of preterm infants' HRV parameters between resting condition 1 and interaction with the 1st parent interaction and interaction with the 1st parent. HRV Feature Resting Condition 1 Interaction between Preterm Neonate-1st Parent p-Value Difference HRm 160.6 166.9 0.002 | SDNN 0.030 0.032 0.511 ns HR_std 14.1 16.1 0.125 ns RMSSD 0.018 0.025 0.064 ns pNN50 1.7 3.2 0.059 ns HRV_Tri 6.6 6.4 0.650 ns VLF_peak 0.036 0.036 0.289 ns LF_peak 0.07 0.06 0.168 ns HF_peak 0.18 0.19 0.367 ns Total power 78.1 136.7 0.014 | VLF (%) 0.043 0.102 0.000 | LF (%) 0.68 0.63 0.122 ns HF (%) 0.25 0.23 0.653 ns LF/HF 4.13 4.01 0.856 ns DFA a 1.11 1.06 0.117 ns DFA a1 1.10 1.11 0.860 ns DFA a2 1.07 0.98 0.021 | Note: Bold type denotes a statistically significant difference between groups. Ns means 'non-significant' and arrows show the direction of variation for a specific parameter. healthcare-11-00672-t007_Table 7 Table 7 Comparison of preterm infants' HRV parameters between interaction with the 1st parent and resting condition 2. HRV Feature Interaction between Preterm Infant-1st Parent Resting Condition 2 p-Value Difference HRm 166.9 161.7 0.010 | SDNN 0.032 0.032 0.983 ns HR_std 16.1 14.8 0.361 ns RMSSD 0.025 0.019 0.255 ns pNN50 3.2 1.9 0.211 ns HRV_Tri 6.4 6.4 0.958 ns VLF_peak 0.036 0.036 0.112 ns LF_peak 0.06 0.07 0.062 ns HF_peak 0.19 0.21 0.608 ns Total power 136.7 96.9 0.043 | VLF (%) 0.102 0.046 0.002 | LF (%) 0.63 0.70 0.025 | HF (%) 0.23 0.23 0.848 ns LF/HF 4.01 4.30 0.647 ns DFA a 1.06 1.09 0.183 ns DFA a1 1.11 1.23 0.121 ns DFA a2 0.98 0.99 0.833 ns Note: Bold type denotes a statistically significant difference between groups. Ns means 'non-significant' and arrows show the direction of variation for a specific parameter. healthcare-11-00672-t008_Table 8 Table 8 Comparison of full-term infants' HRV parameters between interaction with the 1st parent and resting condition 2. HRV Feature Interaction between Full-Term Neonate-1st parent Resting Condition 2 p-Value Difference HRm 160.2 156.8 0.142 ns SDNN 0.038 0.039 0.988 ns HR_std 16.8 16.4 0.948 ns RMSSD 0.039 0.031 0.097 ns pNN50 7.8 6.5 0.193 ns HRV_Tri 6.9 7.2 0.462 ns VLF_peak 0.035 0.037 0.022 | LF_peak 0.06 0.06 0.580 ns HF_peak 0.19 0.18 0.862 ns Total power 322.4 231.8 0.145 ns VLF (%) 0.064 0.039 0.256 ns LF (%) 0.61 0.67 0.145 ns HF (%) 0.28 0.25 0.239 ns LF/HF 3.23 4.20 0.248 ns DFA a 0.99 1.11 0.019 | DFA a1 1.02 1.22 0.060 ns DFA a2 0.91 1.00 0.118 ns Note: Bold type denotes a statistically significant difference between groups. Ns means 'non-significant' and arrows show the direction of variation for a specific parameter. healthcare-11-00672-t009_Table 9 Table 9 Comparison of full-term infants' HRV parameters between resting condition 2 and interaction with the 2nd parent. HRV Feature Resting Condition 2 Interaction between Full-Term Neonate-2nd Parent p-Value Difference HRm 156.8 157.1 0.882 ns SDNN 0.039 0.041 0.487 ns HR_std 16.4 18.7 0.078 ns RMSSD 0.031 0.038 0.147 ns pNN50 6.5 6.2 0.492 ns HRV_Tri 7.2 6.8 0.366 ns VLF_peak 0.037 0.034 0.126 ns LF_peak 0.06 0.06 0.673 ns HF_peak 0.18 0.18 0.812 ns Total power 231.8 330.9 0.170 ns VLF (%) 0.039 0.119 0.013 | LF (%) 0.67 0.59 0.045 | HF (%) 0.25 0.25 0.994 ns LF/HF 4.20 3.25 0.293 ns DFA a 1.11 1.06 0.223 ns DFA a1 1.22 1.15 0.451 ns DFA a2 1.00 0.98 0.548 ns Note: Bold type denotes a statistically significant difference between groups. Ns means 'non-significant' and arrows show the direction of variation for a specific parameter. healthcare-11-00672-t010_Table 10 Table 10 Comparison of preterm infants' HRV parameters between resting condition 2 and interaction with the 2nd parent. HRV Feature Resting Condition 2 Interaction between Preterm Neonate-2nd Parent p-Value Difference HRm 161.7 162.0 0.812 ns SDNN 0.032 0.031 0.699 ns HR_std 14.8 13.8 0.394 ns RMSSD 0.019 0.019 0.964 ns NN50 20.1 22.6 0.833 ns pNN50 1.9 1.9 0.737 ns HRV_Tri 6.4 6.8 0.261 ns VLF_peak 0.036 0.034 0.050 | LF_peak 0.07 0.06 0.125 ns HF_peak 0.21 0.18 0.463 ns Total power 96.9 132.4 0.116 ns VLF (%) 0.046 0.107 0.008 | LF (%) 0.70 0.66 0.130 ns HF (%) 0.23 0.20 0.394 ns LF/HF 4.30 5.07 0.384 ns DFA a 1.09 1.06 0.308 ns DFA a1 1.23 1.13 0.091 ns DFA a2 0.99 0.97 0.536 ns Note: Bold type denotes a statistically significant difference between groups. Ns means 'non-significant' and arrows show the direction of variation for a specific parameter. Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References 1. Cerritelli F. Frasch M.G. Antonelli M.C. Viglione C. Vecchi S. Chiera M. Manzotti A. A Review on the Vagus Nerve and Autonomic Nervous System During Fetal Development: Searching for Critical Windows Front. Neurosci. 2021 15 721605 10.3389/fnins.2021.721605 34616274 2. Fyfe K.L. Yiallourou S.R. Wong F.Y. Odoi A. Walker A.M. Horne R.S. The Effect of Gestational Age at Birth on Post-Term Maturation of Heart Rate Variability Sleep 2015 38 1635 1644 10.5665/sleep.5064 25902805 3. Latremouille S. Lam J. Shalish W. Sant'Anna G. Neonatal heart rate variability: A contemporary scoping review of analysis methods and clinical applications BMJ Open 2021 11 e055209 10.1136/bmjopen-2021-055209 34933863 4. Longin E. Gerstner T. Schaible T. Lenz T. Konig S. Maturation of the autonomic nervous system: Differences in heart rate variability in premature vs. term infants J. Perinat. Med. 2006 34 303 308 10.1515/jpm.2006.058 16856820 5. Aye C.Y.L. Lewandowski A.J. Oster J. Upton R. Davis E. Kenworthy Y. Boardman H. Yu G.Z. Siepmann T. Adwani S. Neonatal autonomic function after pregnancy complications and early cardiovascular development Pediatr. Res. 2018 84 85 91 10.1038/s41390-018-0021-0 29795212 6. Joshi R. Kommers D. Guo C. Bikker J.-W. Feijs L. van Pul C. Andriessen P. Statistical Modeling of Heart Rate Variability to Unravel the Factors Affecting Autonomic Regulation in Preterm Infants Sci. Rep. 2019 9 7691 10.1038/s41598-019-44209-z 31118460 7. Della Longa L. Dragovic D. Farroni T. In Touch with the Heartbeat: Newborns' Cardiac Sensitivity to Affective and Non-Affective Touch Int. J. Environ. Res. Public Health 2021 18 2212 10.3390/ijerph18052212 33668108 8. Helander E. Khodor N. Kallonen A. Varri A. Patural H. Carrault G. Pladys P. Comparison of linear and non-linear heart rate variability indices between preterm infants at their theoretical term age and full term newborns Proceedings of the EMBEC & NBC 2017: Joint Conference of the European Medical and Biological Engineering Conference (EMBEC) and the Nordic-Baltic Conference on Biomedical Engineering and Medical Physics (NBC) Tampere, Finland 11-15 June 2017 Eskola H. Vaisanen O. Viik J. Hyttinen J. Springer Singapore 2017 Volume 65 10.1007/978-981-10-5122-7_39 9. Suga A. Uraguchi M. Tange A. Ishikawa H. Ohira H. Cardiac interaction between mother and infant: Enhancement of heart rate variability Sci. Rep. 2019 9 20019 10.1038/s41598-019-56204-5 31882635 10. McCain G.C. Ludington-Hoe S.M. Swinth J.Y. Hadeed A.J. Heart Rate Variability Responses of a Preterm Infant to Kangaroo Care J. Obstet. Gynecol. Neonatal Nurs. 2005 34 689 694 10.1177/0884217505281857 11. Filho L.F.M.D.S. de Oliveira J.C.M. Ribeiro M.K.A. Moura M.C. Fernandes N.D. de Sousa R.D. Pedrino G.R. Rebelo A.C.S. Evaluation of the autonomic nervous system by analysis of heart rate variability in the preterm infants BMC Cardiovasc. Disord. 2019 19 198 10.1186/s12872-019-1166-4 31420010 12. Lucchini M. Fifer W.P. Sahni R. Signorini M.G. Novel heart rate parameters for the assessment of autonomic nervous system function in premature infants Physiol. Meas. 2016 37 1436 1446 10.1088/0967-3334/37/9/1436 27480495 13. Smith S.L. Haley S. Slater H. Moyer-Mileur L.J. Heart rate variability during caregiving and sleep after massage therapy in preterm infants Early Hum. Dev. 2013 89 525 529 10.1016/j.earlhumdev.2013.01.004 23361061 14. Chiera M. Cerritelli F. Casini A. Barsotti N. Boschiero D. Cavigioli F. Corti C.G. Manzotti A. Heart Rate Variability in the Perinatal Period: A Critical and Conceptual Review Front. Neurosci. 2020 14 561186 10.3389/fnins.2020.561186 33071738 15. Selig F.A. Tonolli E.R. Da Silva V.C.M. De Godoy M.F. Heart rate variability in preterm and term neonates Arq. Bras. Cardiol. 2011 96 443 449 10.1590/S0066-782X2011005000059 21584479 16. Cardoso S. Silva M.J. Guimaraes H. Autonomic nervous system in newborns: A review based on heart rate variability Child's Nerv. Syst. 2017 33 1053 1063 10.1007/s00381-017-3436-8 28501900 17. Landrot I.D.R. Roche F. Pichot V. Teyssier G. Gaspoz J.-M. Barthelemy J.-C. Patural H. Autonomic nervous system activity in premature and full-term infants from theoretical term to 7 years Auton. Neurosci. 2007 136 105 109 10.1016/j.autneu.2007.04.008 17556047 18. Mulkey S.B. Kota S. Swisher C.B. Hitchings L. Metzler M. Wang Y. Maxwell G.L. Baker R. du Plessis A.J. Govindan R. Autonomic nervous system depression at term in neurologically normal premature infants Early Hum. Dev. 2018 123 11 16 10.1016/j.earlhumdev.2018.07.003 30025221 19. Brown L. Heart Rate Variability in Premature Infants During Feeding Biol. Res. Nurs. 2007 8 283 293 10.1177/1099800406298542 17456589 20. Hollingshead A.B. Four Factor Index of Social Status Department of Sociology, Yale University New Haven, CT, USA 1975 unpublished working paper 21. Giannakakis G. Tsiknakis M. Vorgia P. Focal epileptic seizures anticipation based on patterns of heart rate variability parameters Comput. Methods Programs Biomed. 2019 178 123 133 10.1016/j.cmpb.2019.05.032 31416541 22. Shaffer F. Ginsberg J.P. An Overview of Heart Rate Variability Metrics and Norms Front. Public Health 2017 5 258 10.3389/fpubh.2017.00258 29034226 23. Giannakakis G. Marias K. Tsiknakis M. A stress recognition system using HRV parameters and machine learning techniques Proceedings of the 2019 IEEE 8th International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW) Cambridge, UK 3-6 September 2019 269 272 10.1109/aciiw.2019.8925142 24. Mulkey S.B. du Plessis A.J. Autonomic nervous system development and its impact on neuropsychiatric outcome Pediatr. Res. 2018 85 120 126 10.1038/s41390-018-0155-0 30166644 25. DiPietro J.A. Hodgson D.M. Costigan K.A. Hilton S.C. Johnson T.R.B. Fetal Neurobehavioral Development Child Dev. 1996 67 2553 2567 10.2307/1131640 9022256 26. Heart rate variability: Standards of measurement, physiological interpretation, and clinical use. Task Force of the European Society of Cardiology and the North American Society of Pacing Electrophysiology Circulation 1996 93 1043 1065 10.1161/01.CIR.93.5.1043 8598068 27. Patural H. Pichot V. Flori S. Giraud A. Franco P. Pladys P. Beuchee A. Roche F. Barthelemy J.-C. Autonomic maturation from birth to 2 years: Normative values Heliyon 2019 5 e01300 10.1016/j.heliyon.2019.e01300 30899829 28. Suzuki M. Hiroshi T. Aoyama T. Tanaka M. Ishii H. Kisohara M. Iizuka N. Murohara T. Hayano J. Nonlinear Measures of Heart Rate Variability and Mortality Risk in Hemodialysis Patients Clin. J. Am. Soc. Nephrol. 2012 7 1454 1460 10.2215/CJN.09430911 22723446 29. Field T. Fathers' interactions with their high-risk infants Infant Ment. Health J. 1981 2 249 256 10.1002/1097-0355(198124)2:4<249::AID-IMHJ2280020407>3.0.CO;2-I 30. Dietz P. Watson E.D. Sattler M.C. Ruf W. Titze S. Van Poppel M. The influence of physical activity during pregnancy on maternal, fetal or infant heart rate variability: A systematic review BMC Pregnancy Childbirth 2016 16 326 10.1186/s12884-016-1121-7 27784276 31. Thiriez G. Bouhaddi M. Mourot L. Nobili F. Fortrat J.-O. Menget A. Franco P. Regnard J. Heart rate variability in preterm infants and maternal smoking during pregnancy Clin. Auton. Res. 2009 19 149 156 10.1007/s10286-009-0003-8 19255805 32. Kozar M. Tonhajzerova I. Mestanik M. Matasova K. Zibolen M. Calkovska A. Javorka K. Heart rate variability in healthy term newborns is related to delivery mode: A prospective observational study BMC Pregnancy Childbirth 2018 18 264 10.1186/s12884-018-1900-4 29945544 33. Yogman M.W. Games fathers and mothers play with their infants Infant Ment. Health J. 1981 2 241 248 10.1002/1097-0355(198124)2:4<241::AID-IMHJ2280020406>3.0.CO;2-8 34. Levy-Shiff R. Mogilner M.B. Mothers' and fathers' interactions with their preterm infants during the initial period at home J. Reprod. Infant Psychol. 1989 7 25 37 10.1080/02646838908403568 35. Cena L. Biban P. Janos J. Lavelli M. Langfus J. Tsai A. Youngstrom E.A. Stefana A. The Collateral Impact of COVID-19 Emergency on Neonatal Intensive Care Units and Family-Centered Care: Challenges and Opportunities Front. Psychol. 2021 12 630594 10.3389/fpsyg.2021.630594 33716895 36. Griffin T. A Family-Centered "Visitation" Policy in the Neonatal Intensive Care Unit That Welcomes Parents as Partners J. Perinat. Neonatal Nurs. 2013 27 160 165 10.1097/JPN.0b013e3182907f26 23618937 37. O'Brien K. Bracht M. Macdonell K. McBride T. Robson K. O'Leary L. Christie K. Galarza M. Dicky T. Levin A. A pilot cohort analytic study of Family Integrated Care in a Canadian neonatal intensive care unit BMC Pregnancy Childbirth 2013 13 (Suppl. 1) S12 10.1186/1471-2393-13-S1-S12 23445639 |
Subsets and Splits