text
stringlengths 66
377k
|
|---|
Evaluation of Aspergillus IgG, IgM antibody for diagnosing in chronic pulmonary aspergillosis
Chronic pulmonary aspergillosis (CPA) is a slowly progressing pulmonary fungal infectious disease caused by Aspergillus. Aspergillus IgG, IgM are now considered to be valuable biomarkers in CPA diagnosing.Our research attempts to evaluate the effectiveness of Aspergillus IgG, IgM in diagnosing CPA. In our study, CPA patients were younger than the patients who suffered other pulmonary disease. The most common underlying disease in CPA patients was pulmonary tuberculosis. And the most common clinical symptom was hemoptysis. The comparison among the groups indicated statistical significance with regard to Aspergillus IgG and IgM between the CPA and other pulmonary disease groups (P < .01). The Aspergillus-specific IgG and IgM in infectious group exhibited higher levels than those in colonization group (P < .01). The area under the receiver operating characteristic curve of Aspergillus IgG was 0.762 (95% confidence interval: 0.664-0.860) (P < .01).Aspergillus-specific IgG offers great diagnostic value with regard to CPA.Abbreviations: AUC = area under the curve, BALF = bronchoalveolar lavage fluid, COPD = chronic obstructive pulmonary disease, CPA = chronic pulmonary aspergillosis, ELISA = enzyme-linked immunosorbent assay, NTM = nontuberculous mycobacteria, ROC = receiver operating characteristic.
# Introduction
Chronic pulmonary aspergillosis (CPA) is a slowly progressing pulmonary fungal infectious disease caused by Aspergillus. Approximately 3,000,000 people suffer from CPA worldwide, and the 5-year survival rate is estimated at 50%. [bib_ref] Prognostic factors for survival of 194 patients with low infiltrate leukemia, Maddox [/bib_ref] The onset of CPA is insidious, and the diagnosis is difficult. However, the early-stage diagnosis of CPA and the intervention therapy can significantly reduce the fatality of CPA. Therefore, a new effective serological marker is considered important for the confirmation of early-stage intervention. In 2016, the European Respiratory Society and the European Society of Clinical Microbiology and Infectious Diseases issued the first CPA guideline, which confirmed the value and significance of the use of Aspergillusspecific IgG for CPA diagnosis. [bib_ref] Chronic pulmonary aspergillosis: rationale and clinical guidelines for diagnosis and management, Denning [/bib_ref] Currently, the standardized assay for Aspergillus-specific IgG exhibits certain limitations. The majority of the clinical data is derived from European countries, and there have been no relevant data reported in China to date. There were few researches on Aspergillus-specific IgM and had not enough evidence to prove the diagnostic value. In the present study, the diagnostic value for CPA patients was confirmed via the measurement of the Aspergillus-specific IgG, IgM in CPA patients, which provided the basis for the development of CPA investigation and Aspergillus-specific IgG, IgM tests in China.
# Materials and methods
## Study objects
The patients selected were those who visited the outpatient department and/or were hospitalized in the Beijing Chao-yang Hospital, Capital Medical University, due to cough, hemoptysis. The patients were recruited from May 2016 to April 2018. The patient inclusion criteria were as follows: Patients with lesions on chest computed tomography, such as sarcoidosis, cavity, fibrosis, and aspergilloma; patients with Aspergillus detection in lower respiratory tract samples; patients with suspicion of CPA as demonstrated by clinical diagnosis. The exclusion criteria were the following: Patients with severe immune deficiency, namely agranulocytosis, receptor receiving allogeneic hematopoietic stem cell transplantation, long-term administration of large dose of steroid hormone, administration of T cells immunosuppressor in the past 90 days, severe hereditary immunodeficiency disease, and acquired immune deficiency syndrome. All the patients provided the complete clinical information required for the conduct of the study.
A total of 144 blood samples were collected for the detection of Aspergillus-specific IgG, IgM. In the cases selected into the study, there were 79 male and 65 female with an average age of (56.74 ± 15.67). In these cases, there were 22 cases with tuberculosis, 20 cases with bronchiectasis, 19 cases with chronic obstructive pulmonary disease (COPD), 7 cases with pulmonary sarcoidosis, 6 cases with pulmonary fibrosis, 4 cases with pneumoconiosis, 3 cases with lung tumor, 1 case with pulmonary vasculitis, and 1 case with nontuberculous mycobacterium pulmonary disease (6 cases coexist 2 kinds of diseases).
# Methods
This was a prospective study and was approved by the Ethics Committee of Beijing Chao-yang Hospital, Capital Medical University. All the enrolled patients had signed the informed consent. The enrolled patients finished all the relevant tests according to the routine diagnosis of the disease, such as serum G test, serum and bronchoalveolar lavage fluid (BALF) GM test, serum specific IgE and Aspergillus fumigatus specific IgE, and pulmonary high-resolution computed tomography. The clinical information from the enrolled patients was collected, including laboratory examinations, imaging, microbiological examination, and therapeutic treatment. Aspergillus-specific IgG, IgM test were conducted following the collection of the blood samples. According to the diagnostic criteria, the cases were divided into 4 groups: Group 1a (proven CPA), Group 1b (possible CPA), Group 2 (Aspergillus colonization), Group 3 (other pulmonary disease).
2.2.1. Diagnostic criteria. Diagnosis of CPA [bib_ref] Chronic pulmonary aspergillosis: rationale and clinical guidelines for diagnosis and management, Denning [/bib_ref] : proven CPA (meeting 1 of requirements as follows):
(1) microscopic examination of sterile specimens: the specimens were collected by needle aspiration and/or biopsy. Histopathology, cytopathology, and/or direct microscopic examination indicated Aspergillus fungi infection, associated with corresponding tissue damage; (2) sterile specimen culture: Aspergillus was cultured from the samples derived from pulmonary lesions as demonstrated by clinical sampling and imaging using sterile operation (not including BALF). possible CPA (meeting (1) to [bib_ref] Diagnostic significance of Aspergillus species isolated from respiratory samples in an adult..., Tashiro [/bib_ref] Aspergillus colonization was established by positive Aspergillus detection from sputum culture for 2 or more times (≥2 times in the absence of compliance with the diagnostic criteria for CPA, allergic bronchopulmonary aspergillosis, invasive pulmonary aspergillosis and/or other diseases). And the pulmonary diseases absorbed without antifilamentous fungi agents.
## Collection of aspergillus-specific antibody samples.
Following hospitalization, 5 mL of blood was obtained, and centrifuged to collect the supernatant for the Aspergillus-specific IgG, IgM measurement. This was carried out by enzyme-linked immunosorbent assay (ELISA). The experiment was conducted according to the instructions for A fumigate IgG, IgM antibody quantitative detection kit (Dynamiker, China, LOT No. 160801). The Aspergillus-specific IgG, IgM critical value that was lower than 50 AU/mL (<50 AU/mL) was considered negative, whereas a value of higher than 60 AU/mL (>60 AU/mL) was considered positive.
2.2.3. Statistical analysis. SPSS19.0 (SPSS Inc, Chicago, IL) was used to analyze the data. The measurement data were expressed as mean ± standard deviation or median (interquartile range), and analyzed by t test or Kruskal-Wallis test/1-way analysis of variance. The enumeration data were expressed as percentage, and analyzed by the chi-square test. P value of less than .05 (P < .05) was considered statistically significant.
# Results
## Characteristics of the study population
One hundred forty four cases were included. Seventy cases had CPA (16 cases had proven CPA), 28 cases found filamentous fungi and had no evidence to diagnosis mycotic infection. And there were 46 control cases (had other respiratory disease). The clinical characteristics of the different groups are shown in [fig_ref] Table 1: Clinical characteristics of different groups [/fig_ref].
## Diagnostic capability of aspergillus-specific antibody testing based on pulmonary aspergillosis
The results of the serum G, GM, and BALF GM testing among the 3 groups exhibited no statistical significance. The differences in Aspergillus-specific IgG, IgM were significant (both P < .05) . The comparison among the groups indicated statistical significance with regard to Aspergillus IgG, IgM between Group 1 and other groups (both P < .01).
## Identification capability of aspergillus-specific antibody on aspergillus infection and colonization
The comparison between Group 1 and Group 2 indicated that the variables Aspergillus-specific IgG, IgM in the infection group exhibited higher levels than those in the colonization group (P < .01). However, no statistical significance was noted in the serum G test, GM test, BALF GM detection (P > .05) . There was no statistical significance between Group 2 and control group (Group 3).
## Diagnostic capability evaluation of aspergillusspecific antibody for the detection of disease
Receiver operating characteristic (ROC) curve was applied to evaluate the accuracy of Aspergillus IgG and IgM with regard to was statistically significant (P < .01). The sensitivity and specificity of Aspergillus-specific IgM were 58.82% and 68.00%, respectively. The ROC curve was shown in [fig_ref] Figure 1: The ROC curve of aspergillus IgG and aspergillus IgM [/fig_ref].
# Discussion
The rate of Aspergillus infection in fungal infections has been shown to increase daily. Certain studies that were conducted in Japan indicated that the detection rate of deep fungi gradually increased, and that Aspergillus had been one of the main fungi responsible for respiratory tract infection, as demonstrated by The comparison of laboratory results of different groups. specimen culture. [bib_ref] Epidemiology of visceral mycoses in autopsy cases in Japan: comparison of the..., Kume [/bib_ref] [bib_ref] Diagnostic significance of Aspergillus species isolated from respiratory samples in an adult..., Tashiro [/bib_ref] Currently, a paucity of investigations has been conducted with regard to CPA in China, compared with Japan. Compared with the patients suffered from other pulmonary disease, the CPA patients had no differences in sex ratio. CPA patients were younger than the patients who suffered other pulmonary disease. The most common underlying disease in CPA patients was pulmonary tuberculosis. In Aspergillus colonization patients, the most common underlying disease was COPD, bronchiectasis and pulmonary fibrosis. These results were relevant to the high incidence of tuberculosis in China. In CPA patients, the most common clinical symptom was hemoptysis. Pyrexia and dyspnea were not familiar. The inflammation markers in CPA patients were similar to others. Lower respiratory tract specimens could seldom find Aspergillus in CPA patients. In our research, only 4 of 16 CPA patients had positive microbiological results. The most abnormal radiological changes in CPA patients were pulmonary aspergilloma and pulmonary cavitation.
The accurate diagnosis of CPA is challenging. Thus, the identification of a serologic marker of CPA is imperative for the timely treatment of Aspergillus infection. The serum GM test had low sensitivity to CPA diagnosis in the previous study. [bib_ref] Comparison of Aspergillus galactomannan antigen testing with a new cut-off index and..., Kitasato [/bib_ref] In our study, the serum GM test is difficult to distinguish CPA from other pulmonary diseases, and there were no statistical significant differences between the groups. There were numerical differences in BALF GM between CPA group and other pulmonary disease group. The process of bronchoalveolar lavage is hard to standardize. BALF is very hard to obtain from critically ill patients, aged patients, or those with other bronchoscope contraindications. Thus, its application is limited. Only 31 BALF specimens were obtained in this study, so the diagnostic value of BALF GM still need to be verified with more larger sample-size studies.
In clinical practice, physicians frequently diagnose CPA based on clinical manifestations, although the diagnosis for CPA requires pathological specimen verification. Previous studies demonstrated that Aspergillus-specific IgG could be helpful to diagnose CPA. Dumollard et al tested 436 serum samples in a prospective multicentre study. A total of 3 different kits were used to test CPA that had high sensitivity and specificity. The AUC of the ROC curves were higher than 0.9. [bib_ref] Prospective evaluation of a new Aspergillus IgG enzyme immunoassay kit for diagnosis..., Dumollard [/bib_ref] Fujiuchi et al tested Aspergillus-specific IgG in 269 serum samples from the National Hospital Organization Asahikawa Medical Center. The analysis indicated that 96 samples were in compliance with the CPA diagnostic criteria, whereas the antibody levels were significantly higher than those in the control group. The AUC of the ROC curve was 0.94 (test kit was from Phadia, Uppsala, Sweden). [bib_ref] Evaluation of a quantitative serological assay for diagnosing chronic pulmonary Aspergillosis, Fujiuchi [/bib_ref] In the present study, 16 CPA patients were confirmed cases, whereas the remaining 54 were clinically diagnosed cases. The Aspergillus-specific IgG levels were significantly higher than those determined in the other pulmonary disease groups. The AUC of the ROC curve of the Aspergillus-specific IgG was 0.762, whereas the sensitivity and specificity were 70.00% and 82.80%, respectively. Aspergillus-specific IgG provided optimal guidance for the clinical diagnosis of CPA. This study was the first that was conducted on Aspergillus-specific IgG testing of Chinese CPA patients. The data indicated the diagnostic value of Aspergillus-specific IgG on CPA among Chinese subjects, which further provided valuable insight for the diagnostic criteria of Aspergillus-specific IgG for the Chinese participants. In addition, the studies provided valuable information for the design of further CPA epidemiological studies in China.
The increase in the Aspergillus-specific IgG antibody is very common in CPA. [bib_ref] Chronic cavitary and fibrosing pulmonary and pleural aspergillosis: case series, proposed nomenclature..., Denning [/bib_ref] [bib_ref] Clinical characteristics and treatment outcomes of chronic pulmonary aspergillosis, Jhun [/bib_ref] The reports that have examined the diagnostic value of specific IgM are limited. Certain studies indicated that the increase in the Aspergillus-specific IgM antibody was noted in over 50% of the CPA cases, [bib_ref] Use of recombinant mitogillin for improved serodiagnosis of Aspergillus fumigatus-associated diseases, Weig [/bib_ref] whereas specific CPA cases that did not reveal a significant increase in Aspergillus IgG, exhibited a specific IgM increase. The possible reason was attributed to the combination of IgM with different Aspergillus antigens as opposed to IgG. [bib_ref] Follow-up of anti-Aspergillus IgG and IgA antibodies in bone marrow transplanted patients..., Centeno-Lima [/bib_ref] [bib_ref] Class-specific antibody determination against Aspergillus fumigatus by means of the enzyme-linked immunosorbent..., Kauffman [/bib_ref] Aspergillus-specific IgM exhibited supplementary function for the diagnosis of CPA. Until now, we do not clarify whether the duration and rise time is different between Aspergillus IgG and IgM. We want to determine that Aspergillus IgM has early diagnosis value which Aspergillus IgG did not have. Therefore, more researches with high quality, large samples, and adequate follow up are required for further verification. Although the Aspergillus-specific IgM exhibited statistical significance among the CPA, colonization and other pulmonary disease group, the sensitivity and specificity of this marker were limited. The positive likelihood ratio was relatively low. Furthermore, the combination of the Aspergillus IgG and IgM tests exhibited no advantages with regard to the diagnosis of CPA compared with the single use of the Aspergillusspecific IgG test. This might be related to the low sensitivity and specificity of IgM for the diagnosis of CPA. The diagnostic value of the Aspergillus-specific IgM for CPA was limited.
Several clinical challenges remain to be clarified with regard to the differentiation between the Aspergillus colonization and infection. [bib_ref] Prospective evaluation of a new Aspergillus IgG enzyme immunoassay kit for diagnosis..., Dumollard [/bib_ref] Aspergillus colonization has frequently been considered a precondition and/or primary stage of infection. In certain studies, these subjects were investigated in 1 group. [bib_ref] Evaluation of the Aspergillus western blot IgG kit for diagnosis of chronic..., Oliva [/bib_ref] In the present study, the Aspergillus-specific IgG levels in CPA patients were significantly higher compared with those noted in the Aspergillus colonization patients. The latter patients exhibited similar Aspergillus-specific IgG levels with those corresponding to patients with other pulmonary diseases. Future studies should include a higher number of Aspergillus colonization patients.
The present study highlights the application of Aspergillusspecific IgG in the monitoring of CPA and/or the evaluation of the therapeutic response during follow-up. This is due to the collection of the serum samples for Aspergillus-specific antibody testing compared with BALF samples that are considerably difficult to obtain. In the present study, 2 cases were followed up and it was found that Aspergillus-specific IgG did not decrease within the 5-month of therapy, whereas it was significantly Moreover, the levels of Aspergillus-specific IgG during therapy will be monitored during the follow up of all the enrolled patients for a long time in order to observe the changes during the disease progression. We hope to use the Aspergillus-specific IgG markers in the clinical therapy of CPA. However, the majority of the kits are used to detect pulmonary aspergillosis that is caused by A fumigatus. With regard to the non-A fumigatus CPA patients, negative results of Aspergillus specific IgG testing may be observed. In the present study, 6 cases were diagnosed as CPA and exhibited Aspergillus flavus and/or Aspergillus niger that were cultured from lower respiratory tract samples. All 6 patients had low Aspergillus-specific IgG levels (<80 AU/m). A total of 4 out of 6 patients exhibited negative results of the Aspergillus IgG test (<50 AU/mL). Thus, the Aspergillus-specific IgG test exhibited certain limitations. Currently, no studies have been reported that have used the A flavus/Aspergillus nidulans specific IgG ELISA kit. [bib_ref] Antibody testing in aspergillosisquo vadis?, Page [/bib_ref] The data of the present study indicated that Aspergillus-specific IgG, IgM could distinguish CPA from other pulmonary diseases effectively, and offers great diagnostic value with regard to CPA. Furthermore, the diagnostic value of Aspergillus-specific IgG, IgM on the identification of infection and colonization proved very invaluable. And the meaning of the long-term follow-up of CPA requires verification by large sample-size studies.
[fig] 1, 3: -b-D-glucan (n = 93), pg/mL Serum GM (n = 93) BALF GM (n = 31) Aspergillus IgG (n = 144), AU/mL Aspergillus IgM (n = 144), AU/mL [/fig]
[fig] Figure 1: The ROC curve of aspergillus IgG and aspergillus IgM. [/fig]
[table] Table 1: Clinical characteristics of different groups. COPD = chronic obstructive pulmonary disease, ESR = erythrocyte sedimentation rate, NTM = nontuberculous mycobacteria. [/table]
|
Case Report: Trans-axillary Artery Access for Rescue Stent Implantation in an Infant With Retrograde Non-passable Aortic Coarctation
Primary surgical repair remains the traditional treatment for patients with critical duct-dependent coarctation of the aorta (CoA). Initial surgical repair might not be possible or associated with higher risks if additional comorbidities arise in small infants and neonates. Balloon angioplasty (BA) has been described as a rescue strategy for these children. We describe the feasibility of a palliative BA and rescue stent implantation via an alternative antegrade right-axillary artery approach in an initially inoperable infant with pneumonia and respiratory failure and severe CoA, where the stenosis was not passable by traditional retrograde femoral access. This case adds new aspects to the therapy of critical CoA: Stent implantation provides a bridge to surgery in critically ill infants and does not preclude successful surgical repair. Further, if the classic retrograde approach is not possible, the right axillary artery access should be considered as an alternative to pass the stenosis.
# Introduction
Coarctation of the aorta (CoA) accounts for 5-8% of all congenital heart defects (CHD). Surgical repair remains the traditional treatment of critical duct-dependent CoA in newborns and small infants. However, choice of treatment depends on a number of factors including age, weight and clinical condition accounting for the overall risk of the surgical procedure. Balloon angioplasty (BA) and stent implantation is associated with less acute complications compared to surgery, but they are more likely to require planned reintervention.
The axillary artery (AA) access has gained more interest as an alternative vascular access in children especially for duct stenting and CoA. However, it still represents an often-unused access route.
In this report, we present a case of a critically ill infant who underwent palliative stenting of severe CoA by AA access, where retrograde passage of the stenotic segment was not possible.
## Case presentation
A 3-month old infant (4.6 kg) was admitted to an external hospital with Respiratory Syncytial Virus (RSV) bronchiolitis and secondary Haemophilus influenzae Type B pneumonia with respiratory failure needing invasive ventilation. Ventilation quickly became challenging, requiring escalation to high frequency oscillation (HFO) ventilation. The pulmonary situation did not improve and the child developed additional cardiac decompensation requiring high inotrope support and two short periods of cardio pulmonary resuscitation. The patient was transferred to our pediatric cardiac intensive care unit (ICU). On initial examination the patient presented in a significantly reduced clinical condition, with prolonged recapillarisation time of 5 s, skin mottling, cold peripheries and no palpable femoral pulses, an active precordium with a 3/6 systolic murmur. X-ray showed diffuse bi-pulmonal infiltrations, a left pleural effusion and atelectasis of the left upper lobe.
Transthoracic echocardiography revealed severe CoA with minimal antegrade blood flow and an associated systolic peak pressure gradient of 65 mmHg in the presence of a severely impaired left ventricular (LV) function (FS 23%). The arterial duct was almost closed. The LV was hypertrophic with a mild mitral valve regurgitation. The right ventricular function was also impaired and pulmonary artery hypertension present (Pulmonary acceleration time: 25 ms and right ventricular pressure >60 mmHg). Considering the clinical condition with severe biventricular congestive heart failure and respiratory failure, we opted against a primary surgical repair in favor of an emergency interventional transcatheter rescue strategy.
The right femoral artery was first accessed for the intervention. However, a guide wire could not be advanced across the CoA due to the severity of the stenosis. The wire passed through the slightly open arterial duct into the pulmonary artery (PA). Severe pulmonary hypertension was diagnosed with a peak PA systolic pressure of 106 mmHg and mean pressure of 79 mmHg.
To access the stenosis a 4F-sheath (Terumo; Somerset, New Jersey, US) was inserted ultrasound guided in the right AA. Antegrade passage of the nearly closed aortic isthmus was achieved with a 0.021-inch Teflon-coated wire of a 2.7 F Progreat microcatheter (Terumo; Somerset, New Jersey, US). The invasively measured pressure values in the transverse aortic arch proximal to the stenosis (114/40/71 mmHg) and in the descending aorta (DAO) distal to the stenosis (47/40/43 mmHg) corresponded to an invasively measured gradient of 67 mmHg, respectively.
The Progreat catheter was snared by a 5 mm Amplatzer Goose Neck Snare (Medtronic; Dublin, Ireland) from the DAO to build a stable loop. Via the Progreat microcatheter the wire was exchanged to a 0.014-inch Ironman wire (Abbot; Chicago, Illinois, US), and snared again in the DAO. The Progreat catheter was removed and a 2.5 × 20 mm Ryujin Plus balloon (Terumo; Somerset, New Jersey, US) was advanced and positioned over the CoA. After confirming the correct position, the stenosis was pre-dilated with 12.5 atm. After the pre-dilatation a premounted 5.0 x 13 mm Pro kinetic Energy Stent (Biotronic, Berlin, Germany) was advanced. Angiography revealed extreme CoA and the pre-mounted stent completely occluded the lumen with angiographically no visible forward flow to the DAO. The stent was expanded to 5.5 mm with a balloon pressure of 14 atm. The angiography showed the desired result with a harmonically modeled stent and a mild residual gradient of 17 mmHg. After the intervention, the axillary sheath was removed and the AA carefully compressed. The patient received a single intravenous dose of 100 U heparin after placement of the femoral sheath, followed by a continuous heparin infusion of 10 U/kg/h.
After the procedure, the patient remained on the cardiac ICU and stabilized over the next days. Pneumonia and pulmonary edema improved and extubation was achieved 4 days after the intervention. Despite clinical improvement, echocardiography still showed a peak gradient over the stent of 64 mmHgso that surgical therapy was still indicated. The patient underwent surgical stent removal and extended resection with end-to-end anastomosis and duct ligation one week after the intervention. The stent was easily removed during surgery. The patient showed an uncomplicated postoperative course and was extubated a few hours after the surgery. She was discharged 2 weeks after the intervention in a good clinical condition with no residual non-invasive blood pressure gradient between right arm (82/53 mmHg) and leg (81/59 mmHg). Echocardiography at discharge showed a mildly accelerated flow of 2.1 m/s over the isthmus region, LV function was still impaired with a FS of 18% and LV hypertrophy. Due to the reduced LV function and arterial hypertension therapy with Enalapril was started.
Follow-up visits were performed 2 and 4 weeks after discharge showing no clinical signs of heart failure and a peak Dopplergradient of 16 mmHg over the isthmus region. Six months and 1.5 years after the intervention, ventricular function (FS > 28%) and LV hypertrophy have normalized, the aortic arch shows a great surgical result and the patient presents in a good clinical condition.
# Discussion
We describe successful antegrade BA and stent implantation via the AA for severe CoA in a critically ill infant with LV heart failure, pulmonary artery congestion and respiratory failure as a bridge to surgical repair.
BA or stent implantation for duct-dependent native CoA has its limitations with a higher rate of reinterventions, aneurysm formation and especially in neonates trauma or thrombosis of the accessed artery (6, 7). Primary surgical therapy should therefore be preferred if possible. However, in critically ill newborns and infants with additional comorbidities an interventional approach has evolved as a life-saving treatment option, bridging to corrective surgery. The interventional approach, especially with stent implantation, remains reserved for a selected group of patients and is mainly performed by conventional retrograde femoral access. In our case, under emergency conditions, we were forced to choose an alternative access route because the stenosis was so severe that it could not be passed by a traditional retrograde approach. Schranz and Michel-Behnke described a detailed description of the AA access for cardiac interventions in newborns and promote it as an attractive alternative approach especially compared to carotid and femoral access (3). Other authors describe the AA access for BA of the aortic valve or CoA in neonates as a safe but underrepresented access (4, 5, 10). The AA access can be even easier in patients with CoA and LV failure presenting with decreased perfusion of the femoral artery. Therefore, AA remains easier to palpate and access, while femoral pulses and perfusion are weakened. One could argue that the carotid artery access offers the same advantages, which has been used successfully by others . However, vascular complications including occlusion, thromboembolism or severe hematoma, which may occur especially in small children, are not as relevant using the AA compared to the carotid artery. Especially as it is not an end artery and arm perfusion remains guaranteed during the procedure. In our case, heparin was given after insertion of the sheath in the femoral artery as per our protocol. Therefore, we also decided to use the AA access, since puncture of the carotid artery under heparinization has higher risks. AA access has been described previously for stent implantation for CoA e.g., in a case series of neonates with stable hemodynamics and preserved ventricular function. In contrast to our patient, the inventions were electively planned with a surgical exposure of the AA punctured under direct vision. This approach requires surgical repair of the AA at the end of the procedure. Esmaeili et al. recently pushed boundaries using the AA access for a planned stent implantation in a neonate with 1.2 kg, severe CoA and pulmonary hypertension (14).
CoA-stenting as a rescue maneuver is less invasive compared to surgical therapy. This allows the patient additional time to recover clinically as well as from LV dysfunction before undergoing surgical repair. Concerns that the stent may complicate surgical therapy could not been confirmed in our case, as it was easy to removed after 1 week.
Despite our case clearly describes a severe congenital CoA complicated by pulmonary infection, we would like to point out the interesting correlation between Haemophilus influenzae infections and aortic aneurysm or aortitis, which have been reported previously.
This report is limited to a single case, however adds new aspects to the therapy of critically ill children with severe CoA. Firstly, rescue stent implantation for patients with severe compromised hemodynamics provides a successful bridge to surgery. Secondly, the right AA access is an attractive alternative for patients with severe narrowing of the isthmus where traditional retrograde passage cannot be achieved. The percutaneous access of the AA can safely be performed ultrasound guided by an experienced interventionist and does not require surgical cut down.
# Conclusion
Percoutaneous AA access for antegrade BA and stenting represents an attractive alternative approach in critically ill infants with severe CoA who require emergency intervention and conventional retrograde passage is not possible. Rescue stent implantation does not preclude successful surgical repair if removal of the stent follows timely.
# Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
# Ethics statement
Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.
# Author contributions
SD performed the intervention. JM, SE, and SD were involved in the clinical decision making and care plan of the child. AP performed the surgery. JM and SD collected the clinical and scientific findings and wrote the manuscript. All authors discussed the results and contributed to the final manuscript. |
Rice In Vivo RNA Structurome Reveals RNA Secondary Structure Conservation and Divergence in Plants
Supplemental Figure 1 Preliminary experiment for optimizing concentration of DMStreatment. The fragment from 1 251 nt to 1 334 nt of 18S rRNA was analyzed. Lanes 1-5: 0, 0.75, 1.5, 2.5, 4% (v/v) DMS concentration, respectively, for 15 min at 28°C; lanes 6 to 9: A/C/G/U sequencing lanes. The red and blue stars show the obviously modified A and C in concentration course, respectively.
in Arabidopsis. P values between two groups were calculated by KS test.
(C) DFT of average DMS reactivity of selected regions in mRNAs with high and low mRNA GC Two individual examples for two kinds of relationship between sequence identity and structure similarity. High sequence identity and high structure similarity (group-1) (A) and low sequence identity and low structure similarity (group-4) (B) exhibit two kinds of relationship between sequence identity and structure similarity. The results in rice are coloured with red and the ones in Arabidopsis with blue. The reactivity is compared based on the sequence alignment of selected regions. Best matched regions are coloured in a light purple background, the mismatches on A and C in a green background, and the mismatches on G and U in a yellow background. Arc diagrams were shown to compare in vivo RNA secondary structure in the selected regions.
[formula] A A A C C UG G G A U G CG C A A G G A A G C UG A C G AG CG G G A G G C C C U C A C G G G C CG C A C C G C UG G C CG A C C C UG A U C U U C UG UG A A G G G U U C G AG U UG G AG C A C G C C UG U C G A A A C C CG U A A G G C G C A A G G A A G C UG A U U G G C G G G --U C C U C G C G G G G C A C CG C C G A C C G A C C UUG A U C U U C UG A G A A G G G U U C G AG UG UG AG C A U G C C UG U C G A U750 G G C G G A U C C U C G C G G G U G C A C C G C C G A C C G A C C U U G A U C U U C U G A G A A G G U U G A G G C U G C U G A C A U G U G U G C G A G UC A A C GA G U A A A C C C G U A A G G C G C A A G G A A G C U G A G A G C A U G C C U G U C G G G A C G U G G U U U G 700 800 3' 5' 750 C U G A C A U G C G U G C G A G UC G A C G G G U C U G A A A C C U G G G A U G C G C A A G G A A G C U G A G A G C A C G C C U G U C G G G A C U U G G C G G G A C C U C G G G G G C A C C G C C G C C G A C C U U G A U C U U C U G G A A G G U U G A G G U G C U C G A GG UC C AG A A U A A G G GCG AG A C CG U C A CG CU CG AC AC C AG G GU CG UC A A C U C CG C CG UC AC CG C C AC CG UCGG CG ACG C CG AG CCG C UCG C CG UC U A CG CCG G U U C A A A A U G ACGG AG AAG A CG UG ACG C UG GAG A CG GA CG UUG UG A CG GCG A A G GU A A U G G G GA CG U U A A A G GA U C A A G AG C CA U UG A U A G U U U A C A A G A LOC_Os09g07350.1 AT4G12730.1 C A A A G A A C CU U U A C A U C A U U U C UG UG A A G G GA A U C A A G G GA CG C CU U A A C AG G CU U C CU U C UG C UUG UG U UG G G GA C A UG GU U A UG GC U A C UG UG A A A A A C U A A G A A C CU U U A C A U C A U C U C UG U U A A A G GA A U C A A A G GU CG UC UC A A U C G GU U A C CU U C [/formula]
[fig] Supplemental, Figure 2: High agreement between Structure-seq and conventional gel-based RNA secondary structure probing. (A) Left: gel-based probing. Right: the comparison of Structure-seq (purple bars) and gel-based probing (orange line) from 1 363 nt to 1 430 nt of 18S rRNA. (B) Left: gel-based probing. Right: the comparison of Structure-seq (purple bars) and gel-based probing (orange line) from 156 nt to 253 nt of 25S rRNA. Lane 1 and 2, without and with 2.5% DMS treatment; lane 3 to 6, A/C/G/U sequencing lanes. The positions of each well-visible A and C are coloured by red and blue stars, respectively. PCC and P values were shown in the corresponding region. [/fig]
[fig] Supplemental, Figure 3 Supplemental, Figure 4 Supplemental, Figure 5: High correlation of PPV with sensitivity and F1 score. (A)-(B) Correlation of PPV with sensitivity (A) and F1 score (B) in rice. Both P values are less than 2.2E-16. (C)-(D) Correlation of PPV with sensitivity (C) and F1 score (D) in Arabidopsis. Both P values are less than 2.2E-16. A ...... ...... Global pattern of mRNA structurome in rice. (A) Average DMS reactivity in selected regions of mRNAs with more than one stop per nucleotide.23 607 mRNAs filtered by length were used for this analysis. Wilcoxon rank-sum test (when alternative hypothesis was "less") was used for P value calculation between 5' UTR and CDS, CDS and 3' UTR. P values were shown in the corresponding region.(B) Average DMS reactivity around the alternative splice sites between 88 724 spliced and 5 862 unspliced events in 5' splice site and 88 724 spliced and 6 035 unspliced events in 3' splice site. (C)-(E) Discrete Fourier transform (DFT) of average DMS reactivity of selected regions in all the mRNAs (C), mRNAs with high and low mRNA GC content (D), mRNAs with high and low translation efficiency (E). (F) Effect of the number of m 6 A peaks on Gini index in 3' UTR. (Yellow) 7 308 transcripts without m 6 A peaks; (orange) 2 064 transcripts with single m 6 A peak; (red) 2 475 transcripts with multiple m 6 A peaks. P values were calculated by Wilcoxon rank-sum test. (G) DMS reactivity pattern around the conserved motif (UGUAMM, M = A or C) identified in m 6 A-seq (Li et al., 2014). 4 221 motifs with m 6 A and 4 194 motifs without m 6 A were analyzed. P values were calculated by Kolmogorov-Smirnov (KS) test. Effect of GC content on the global pattern of mRNA structurome in Arabidopsis. (A)-(B) Average GC content (A) and average DMS reactivity (B) of transcripts with the highest 15% and lowest 15% mRNA GC content with more than one stop per nucleotide (11 207 in total) [/fig]
[fig] Supplemental, Figure 6: Comparison of 25S rRNA fragment in rice and Arabidopsis. The fragments are from 695 nt to 800 nt in rice and from 703 nt to 805 nt in Arabidopsis.(A) Gel-based probing in rice and Arabidopsis. Lane 1 and 2, without and with 2.5% DMS treatment; lane 3 to 6, A/C/G/U sequencing lanes. The positions of each well-visible A and C are coloured by red and blue, respectively.(B) Comparison of gel reactivity (top) and relative DMS reactivity (bottom) in rice andArabidopsis. In the sequence alignment, mismatches are highlighted in a green background, best matches are highlighted in a purple background, and the region with highly conserved in vivo RNA secondary structure is marked in white. Pairwise alignments were performed in EMBOSS Needle. PCC and P value were shown in the corresponding region.(C) Phylogenetic structure of the fragments coloured by DMS reactivity. The orange line represents the highly conserved structure probing region identified in gel-based probing and Structure-seq analysis. The probed sites using gel-based probing (SupplementalFigure 6Aand 6B top) and Structure-seq (SupplementalFigure 6Bbottom) show very similar DMS modification patterns between rice and Arabidopsis. The pattern of DMS reactivity is consistent with the phylogeny-derived structures (SupplementalFigure 6C). [/fig]
|
Radiologic and Pathologic Findings of a Follicular Variant of Papillary Thyroid Cancer with Extensive Stromal Fat: A Case Report
Thyroid cancer may have small adipose structures detected by microscopy. However, there are no reports of thyroid cancer with gross fat evaluated by radiological methods. We reported a case of a 58-year-old woman with a fat containing thyroid mass. The mass was hyperechoic and ovoid in shape with a smooth margin on ultrasonography. On computed tomography, the mass had markedly low attenuation suggestive of fat, and fine reticular and thick septa-like structures. The patient underwent a right lobectomy. The mass was finally diagnosed as a follicular variant of papillary thyroid cancer with massive stromal fat.
# Introduction
Fat-containing lesions in the thyroid are very rare [bib_ref] Lipomatous lesions of the thyroid gland: a review, Schröder [/bib_ref]. Thyroid lesions containing gross fat include heterotopic fat nests, thyrolipoma, teratoma, liposarcoma, and thyroid cancers of the papillary or follicular type [bib_ref] Metastatic liposarcoma of the thyroid gland, Azar [/bib_ref] [bib_ref] Thyroid papillary carcinoma lipomatous type: report of two cases, Bisi [/bib_ref] [bib_ref] Thyroid neoplasms containing mature fat: a report of two cases and review..., Derienzo [/bib_ref] [bib_ref] Papillary carcinoma of the thyroid gland with lipomatous stroma: report of a..., Vestfrid [/bib_ref] [bib_ref] An unusual mature thyroid teratoma on CT and 99Tcm scintigraphy imaging in..., Zhang [/bib_ref]. Most of these cases are only described upon obtaining pathological data after surgery; radiologic methods are of limited use [bib_ref] Study of intrathyroid fat-containing lesions using CT imaging with literature review, Kim [/bib_ref]. With the approval of the Institutional Review Board of our hospital, we reported the case of an intrathyroid nodule composed mainly of gross fat detected by computed Radiologic and Pathologic Findings of a Follicular Variant of Papillary Thyroid Cancer with Extensive Stromal Fat: A Case Report tomography (CT) and confirmed by lobectomy as a follicular variant of papillary thyroid cancer (FVPTC) with mature fat.
## Case report
A 58-year-old woman presented with an incidentally detected mass in the thyroid on routine ultrasonography (US) at a primary clinic. The patient had a 10-year history of diabetes mellitus and hypertension, and was currently taking medication. She had no medical history of thyroid disease. Physical examination revealed a firm mass of approximately 3 cm in size in the right anterior neck. Laboratory tests revealed that serum anti-microsomal antibody, anti-thyroglobulin antibody, thyroglobulin, calcitonin, T3, free T4, and thyroid-stimulating hormone levels were within normal ranges.
Ultrasonography revealed an ovoid lesion with a smooth margin of approximately 2.9 x 2 x 1.3 cm in size. The nodule was generally hyperechoic, with the peripheral portion more hyperechoic than the central area. Posterior to the mass, curtain-like hyperechoic shadowing was observed. The perithyroidal fat, muscle, and vertebral bodies were [fig_ref] Figure 1: 58-year-old woman diagnosed with follicular variant of papillary thyroid cancer with mature... [/fig_ref]. Doppler US did not show any vascular signals in the nodule. Based on these findings, a diagnosis of an indeterminate nodule was made. A board-certified radiologist performed US-guided fine needle aspiration using a freehand technique with a 23-gauge needle and a 5-mL disposable plastic syringe. Cytologic smears were made on glass slides and immediately fixed in 95% alcohol for both Papanicolaou staining and May-Grunwald-Giemsa staining. Cytology revealed atypia of undetermined significance or a follicular lesion of The thyroid CT protocol at our hospital includes precontrast, early, and delayed phases. Pre-contrast images were acquired from the mandibular angle to the thoracic inlet. Early phase images (40 seconds) were scanned from the hyoid bone to the thoracic inlet to cover the entire thyroid and level III-VI lymph nodes. Delayed scans (90 seconds) covered the area from the skull base to the upper mediastinum. All scans were acquired with a slice thickness of 2.5 mm and reconstruction increment of 2.5 mm. Precontrast CT showed markedly low attenuation (mean CT number, -80 Hounsfield units [HU]), in a well-defined mass with fine reticular and thick septa-like structures in the parenchyma of the right mid-to-upper portion of the thyroid [fig_ref] Figure 1: 58-year-old woman diagnosed with follicular variant of papillary thyroid cancer with mature... [/fig_ref]. To measure the enhancement pattern of the tumor, 25 mm 2 of the region of interest was placed in the area of the fatty mass, including the reticular and thick septa-like soft tissue lesions. The tumor was enhanced in the early phase (mean CT number, 16 HU) and was washed out in the delayed phase (mean CT number, -13 HU) [fig_ref] Figure 1: 58-year-old woman diagnosed with follicular variant of papillary thyroid cancer with mature... [/fig_ref]. The tumor was ovoid in the longitudinal direction, had a lobular appearance in the axial plane, and was completely encapsulated in the thyroid parenchyma, with no evidence of extrathyroidal extension. No pathological lymph nodes were identified. Based on these findings, we suspected a fat-containing, neoplastic lesion of the thyroid gland, such as thyrolipoma, teratoma, lipoma, lowgrade liposarcoma, or a follicular tumor with stromal fat. Considering the results from cytopathology and the gene mutation analysis, we could not rule out the possibility of a malignant lesion. Therefore, we performed a right lobectomy.
The gross specimen was fixed with formalin solution. It showed a yellowish, homogeneous, and well-demarcated nodular lesion measuring 2.5 x 2.2 x 1.5 cm at the upperto-mid pole. The mass was 1.5 cm away from the isthmic resection margin.
Microscopic examination revealed a well-circumscribed, thin, fibrous, and capsulated nodule with mature fat. There were atypical follicular cells with nuclear enlargement, nuclear overlapping, chromatin clearing, and nuclear grooves without papillary structures in the nodule [fig_ref] Figure 1: 58-year-old woman diagnosed with follicular variant of papillary thyroid cancer with mature... [/fig_ref]. On immunohistochemistry, the tumor cells stained positive for HBME-1 (DAKO, Glostrup, Denmark, 1:100) and galectin-3 (Novocastra, Newcastle Upon Tyne, UK, 1:200). Finally, we tested for the V600E and K601E mutations in exon 15 of the BRAF gene and codons 12, 13, and 61 of the KRAS and NRAS genes. We reconfirmed positive NRAS 61 mutation, negative BRAF and negative KRAS mutation. The final pathologic diagnosis was follicular variant papillary thyroid carcinoma with mature fat.
# Discussion
In our case, the tumor was a yellowish, homogeneous, and well-demarcated nodular lesion on gross examination, with negative value of attenuation on CT, suggesting fat. However, it was difficult to recognize the tumor as a fat containing mass on US examination. The hyperechoic nature of the mass and the curtain-like hyperechoic shadowing posterior to the mass may suggest rich fatty component. We suspected that the hyperechoic shadowing might have been due to a reverberating artifact caused by innumerable fat globules and septa of the tumor, similar to the artifact frequently observed along the thick abdominal walls that consist of several layers of fat globules surrounded by connective tissues, skin, and fascia.
Computed tomography and gross specimen findings of our case were very similar to those of thyrolipoma, a benign fat-containing tumor [bib_ref] Thyrolipoma and thyrolipomatosis: 5 case reports and historical review of the literature, Ge [/bib_ref]. Thyrolipoma is considered a variant of follicular adenoma, which is characterized by a well-circumscribed and encapsulated nodule resulting from the proliferation of thyroid follicles admixed with mature fat [bib_ref] Thyrolipoma and thyrolipomatosis: 5 case reports and historical review of the literature, Ge [/bib_ref] [bib_ref] Adenolipomas of the head and neck: analysis of 6 cases, Daboin [/bib_ref]. This case could have been misdiagnosed as a follicular neoplasm or thyrolipoma, as the tumor cells did not show a papillary growth pattern upon microscopic examination and the gross features were very similar to thyrolipoma. Two reports provided the CT and magnetic resonance imaging findings of thyrolipoma [bib_ref] Case 53: adenolipoma of the thyroid gland, Borges [/bib_ref]. The mass predominantly had fat attenuation and distinct margins in the thyroid gland. However, the authors only performed precontrast CT on the lesion. Therefore, it is uncertain whether thyrolipoma and thyroid cancer with massive stromal fat can be differentiated by CT with contrast enhancement. In our case, the tumor was enhanced in the early phase and slightly washed out in the delayed phase.
Fat-containing thyroid cancer of the papillary or follicular type, have only been reported from pathological studies [bib_ref] Thyroid papillary carcinoma lipomatous type: report of two cases, Bisi [/bib_ref] [bib_ref] Thyroid neoplasms containing mature fat: a report of two cases and review..., Derienzo [/bib_ref] [bib_ref] Papillary carcinoma of the thyroid gland with lipomatous stroma: report of a..., Vestfrid [/bib_ref]. Fat tissue was observed only by microscopy, so it was assumed that small amounts of fat tissue could not be detected by CT. We also assumed that the detectability of the fat component is variable, as it depends on the total number of fat cells in the mass. Mature teratoma can be Korean J Radiol 16(6), Nov/Dec 2015 kjronline.org recognized if the fat-containing mass contains calcification, bone or cartilage, and cystic components. However, this is not possible, as radiological images do not always exhibit all these attributes [bib_ref] An unusual mature thyroid teratoma on CT and 99Tcm scintigraphy imaging in..., Zhang [/bib_ref]. CT images of liposarcoma (3) in the thyroid gland are very similar to those in our case, and differentiating between teratoma, thyrolipoma, liposarcoma, and papillary thyroid cancer with massive fat may be impossible using radiology alone.
The origins of the fat tissue are unknown. Some investigators have insisted that fat cells are derived from displaced remnants of embryonic structures in the thyroid. In amyloid goiter, Schröder and Böcker (1) postulated that adipose tissue may be derived from metaplasia of stromal fibroblasts due to either impaired circulation and diffusion or tissue hypoxia caused by amyloid deposition. DeRienzo and Truong [bib_ref] Thyroid neoplasms containing mature fat: a report of two cases and review..., Derienzo [/bib_ref] suggested that the fat cells in thyrolipoma and follicular carcinoma might be neoplastic tissue because 1) fat did not exist in the normal thyroid, 2) fat was a major component of the tumor, 3) there was no evidence of degenerative changes such as dystrophic calcification, cystic or hemorrhagic degeneration, or amyloid deposition, 4) fat was widely distributed in the tumor, and 5) there were similar fat-containing tumors in other organs such as thymolipoma, parathyroid adenoma, and fat-containing fibroadenoma of the breast. Our case showed similar microscopic features to theirs in agreement with the suggestion that the fat cells might be neoplastic cells themselves.
In summary, we identified FVPTC with stromal fat using US and CT. The diagnosis was confirmed by histopathology, immunochemical staining, and gene analysis. These findings included a hyperechoic mass with a smooth margin with a reverberating artifact and no detectable Doppler signal on US. CT revealed a well-encapsulated fat-attenuation mass with enhancing septa-like structures. FVPTC should be included in the differential diagnosis of fat-containing focal lesions in the thyroid, in addition to thyrolipoma, teratoma, liposarcoma, and small focal fat nests near the capsule. Moreover, fine needle aspiration cytology with gene analysis may be helpful in determining the nature of fatty tumors before surgery.
[fig] Figure 1: 58-year-old woman diagnosed with follicular variant of papillary thyroid cancer with mature fat. A. Longitudinal image of right thyroid using gray-scale ultrasonography shows hyperechoic, ovoid mass with smooth margin. Peripheral portion of mass is more echogenic than central area. Curtain-like hyperechoic shadowing was observed posterior to mass. Perithyroidal fat, muscle, and vertebral bodies were therefore not clearly visualized. B-D. Pre-contrast computed tomography (CT) shows fatty (mean CT number, -80 Hounsfield units, HU), well-defined mass in parenchyma of right mid-to-upper portion of thyroid (B). Mass has several fine reticular and thick septa-like soft tissue lesions, which show contrast enhancement. Post-contrast CT shows higher enhancement in early phase (C, at 40 seconds; mean CT number, 16 HU) of tumor than in delayed phase (D, 90 seconds; mean CT number, -13 HU). Tumor is lobular in appearance in axial plane (B-D), ovoid in longitudinal direction, and completely located in thyroid parenchyma without evidence of extrathyroidal extension. E. High-power magnification (1:400, hematoxylin and eosin staining) reveals tumor cells with enlarged nuclei, chromatin clearing, and nuclear grooves. Tumor cells are surrounded by mature adipose tissue. with negative for BRAF mutations and positive for an NRAS 61 mutation. [/fig]
|
Repeated cerebellar infarction in the affected nondominant vertebral artery distribution with reversible vertebral artery occlusion elicited by head tilt: illustrative case
BACKGROUND Bow hunter's syndrome or stroke (BHS) is characterized by rotational vertebrobasilar insufficiency elicited by rotation of the neck. It is caused by dynamic and reversible occlusion of the vertebral artery (VA). Reversible symptoms of rotational vertebrobasilar insufficiency are described as bow hunter's syndrome, although brain infarction is rarely reported as bow hunter's stroke.OBSERVATIONS A 70-year-old man experienced repeated cerebellar infarctions three times in the posterior inferior cerebellar artery (PICA) distribution of the nondominant right VA connecting the basilar artery. The onset of symptoms indicating cerebellar infarcts and the patient's head position changes were unrelated. Dynamic digital angiography (DA) revealed that the nondominant right VA was occluded by an osteophyte from the C4 vertebral body, and the right PICA branches were shown to be passing through the distal right VA from the left VA. These findings were observed when the patient's head was tilted to the right. An arterio-arterial embolic mechanism was suggested as the cause of repeated cerebellar infarctions.LESSONS Transient nondominant VA occlusion has been rarely reported as a cause of BHS when the head is tilted. To confirm the diagnosis of BHS, additional head tilt is recommended when performing dynamic DA in patients with a cervical osteophyte. https://thejns.org/doi/abs/10.3171/CASE2061 KEYWORDS repeated cerebellar infarction; bow hunter's syndrome; stroke; cervical osteophyte; transient vertebral artery occlusion; head tilt ABBREVIATIONS 3D-CTA = three-dimensional computed tomography angiography; BHS = bow hunter's syndrome or stroke; DA = digital angiography; DSA = digital subtraction angiography; DWI = diffusion-weighted imaging; MRA = magnetic resonance angiography; MRI = magnetic resonance imaging; PICA = posterior inferior cerebellar artery; VA = vertebral artery.
Bow hunter's syndrome or stroke (BHS) is characterized by rotational vertebrobasilar insufficiency. This condition is elicited by neck rotation in most cases and by neck extension or flexion in some cases. BHS is caused by dynamic and reversible occlusion of the vertebral artery (VA) and a lack of collateral blood supply to the brainstem. [bib_ref] Surgical treatment of vertebral artery insufficiency caused by cervical spondylosis, Bakay [/bib_ref] [bib_ref] Surgical treatment of vertebral artery insufficiency caused by cervical spondylosis, Nagashima [/bib_ref] [bib_ref] Mechanical occlusion of the vertebral artery. A new clinical concept, Husni [/bib_ref] [bib_ref] Bow hunter's syndrome revisited: 2 new cases and literature review of 124..., Jost [/bib_ref] Although the term "bow hunter's stroke" was coined by Sorensen for a patient with suspected brainstem infarction, 5 reversible symptoms of rotational vertebrobasilar insufficiency are usually described as bow hunter's syndrome. [bib_ref] Bow hunter's syndrome revisited: 2 new cases and literature review of 124..., Jost [/bib_ref] BHS associated with recurrent cerebellar infarction and transient occlusion of the nondominant VA induced by head tilt has not been reported in the literature. In this study, we report a case involving an older man who experienced recurrent cerebellar infarctions three times in the last 2 years after undergoing a cervical left laminoplasty procedure that had been performed 5 years earlier. The infarctions occurred in the posterior inferior cerebellar artery (PICA) distribution of the affected nondominant VA, with no infarction occurring in any other region. Transient occlusion of the right VA by an osteophyte at the level of C4 was demonstrated during dynamic digital angiography (DA) when the patient's head was tilted to the right. No symptoms or additional infarction occurred after surgery with posterior fixation of the cervical spine (C3-C6).
## Illustrative case
A 70-year-old man underwent left laminoplasty at C3-C6 in a different hospital 5 years ago for ossification of the posterior longitudinal ligament. He had a prominent medical history of cerebellar infarction in the lateral region of the right PICA distribution that was detected 2 years ago. For this he had started taking oral clopidogrel (75 mg/day). He was admitted to our hospital with complaints of the sudden onset of dizziness, nausea, and vomiting. Diffusion-weighted imaging (DWI) revealed an additional cerebellar infarction in the medial region of the right PICA distribution, and magnetic resonance angiography (MRA) revealed that the right VA was occluded [fig_ref] FIG. 1: MRI and MRA on the patient's first admission to our hospital [/fig_ref] , B, D, and E). Three days later, repeated MRA showed recanalization of the nondominant right VA connecting the basilar artery [fig_ref] FIG. 1: MRI and MRA on the patient's first admission to our hospital [/fig_ref]. Echocardiography and 24-hour Holter monitor electrocardiography results were negative, and the etiology of his recurrent cerebellar infarction could not be elucidated. He was discharged from the hospital, and the following medications were prescribed without anticoagulants: aspirin (100 mg/day) and clopidogrel (75 mg/day).
Two months later, the man returned to our hospital with symptoms identical to those reported during the previous admission. DWI and MRA showed a new small cerebellar infarction in the right side of the inferior cerebellar vermis, with no infarction in any other region, as well as another right VA occlusion [fig_ref] FIG. 2: MRI and MRA on the patient's second admission to our hospital [/fig_ref] , B, D, and E). The right VA appeared patent on MRA on the 12th day after the second admission to our hospital [fig_ref] FIG. 2: MRI and MRA on the patient's second admission to our hospital [/fig_ref] and F). Three-dimensional computed tomography angiography (3D-CTA) showed that the right VA was laterally compressed by an osteophyte from the C4 vertebral body with no stenosis [fig_ref] FIG. 3: 3D-CTA [/fig_ref]. Nineteen days later, during the man's second admission, he complained of symptoms similar to those in his first admission, including dizziness, nausea, and vomiting. MRA revealed an identical right VA occlusion with no new infarction on DWI. During subsequent dynamic digital subtraction angiography (DSA) and DA procedures with the patient's head in the neutral position, the nondominant right VA and the vermian and hemispheric branches of the right PICA were injected without stenosis [fig_ref] FIG. 4: Dynamic bilateral vertebral arteriography [/fig_ref]. No severe stenosis or occlusion was revealed in the right VA when the neck was rotated to the right [fig_ref] FIG. 4: Dynamic bilateral vertebral arteriography [/fig_ref] or left [fig_ref] FIG. 4: Dynamic bilateral vertebral arteriography [/fig_ref] or when the neck was flexed or extended. However, the right VA was occluded by an osteophyte at the C4 level when his head was tilted to the right, which elicited no complaints of dizziness, nausea, or vomiting [fig_ref] FIG. 4: Dynamic bilateral vertebral arteriography [/fig_ref]. A similar occlusion was not induced by head tilt to the left [fig_ref] FIG. 4: Dynamic bilateral vertebral arteriography [/fig_ref]. No stenosis or occlusion was demonstrated in the dominant left VA on dynamic DA. The distal right VA was demonstrated in a retrograde fashion by injection of the left VA when the patient's head was tilted to the right [fig_ref] FIG. 4: Dynamic bilateral vertebral arteriography [/fig_ref] , and both the vermian and hemispheric branches of the right PICA were subsequently opacified [fig_ref] FIG. 4: Dynamic bilateral vertebral arteriography [/fig_ref] and K). The man was asymptomatic while his head was tilted to the right for more than 1 minute.
The patient denied that changes in the head or neck position caused the symptoms or the stroke. An arterio-arterial embolism from the transiently occluded right VA to the right PICA distribution was suggested as a cause of the repeated cerebellar infarction in this patient. He underwent posterior cervical fixation at C3-C6 at Niigata University Medical and Dental Hospital. After the posterior cervical fixation, both VAs were shown to operate continuously without stenosis from the origin to the basilar artery on 3D-CTA . He had no symptoms, and no additional cerebellar infarction developed for 1 year 6 months after the surgery.
# Discussion
In this patient, repeated cerebellar infarction was restricted to the PICA distribution with reversible occlusion of the ipsilateral nondominant right VA and was demonstrated three times on MRI and MRA in the past 2 years after a C3-C6 left laminoplasty that had been performed 5 years earlier. 3D-CTA showed an osteophyte of C4 compressing the right VA laterally without stenosis. On dynamic DA, occlusion of the right VA was demonstrated when the patient tilted his head to the right, although no severe stenosis or occlusion was observed in other head and neck positions. No stenosis or occlusion was shown in the dominant left VA. In this patient, without demonstration of the right VA occlusion elicited by head tilt to the right, diagnosis of BHS could not be made, even after dynamic DA was performed. In elderly male patients (most ≥40 years), reversible symptomatic VA stenosis/ occlusion caused by an osteophyte of the degenerative cervical vertebral body is one of the common characteristics in BHS when changing head and neck positions. [bib_ref] Surgical treatment of vertebral artery insufficiency caused by cervical spondylosis, Bakay [/bib_ref] [bib_ref] Surgical treatment of vertebral artery insufficiency caused by cervical spondylosis, Nagashima [/bib_ref] [bib_ref] Mechanical occlusion of the vertebral artery. A new clinical concept, Husni [/bib_ref] [bib_ref] Bow hunter's syndrome revisited: 2 new cases and literature review of 124..., Jost [/bib_ref] [bib_ref] Bow hunter's stroke, Sorensen [/bib_ref] [bib_ref] Rotational vertebral artery occlusion: mechanisms and long-term outcome, Choi [/bib_ref] In our patient, left laminoplasty of C3-C6 might have predisposed him to contralateral cervical osteophyte formation or growth at the C4 vertebral body.
BHS is the descriptive term for rotational vertebrobasilar insufficiency. This condition is elicited by neck rotation in most cases and by neck extension or, rarely, by neck flexion in some patients. BHS is caused by dynamic and reversible occlusion of the VA and a lack of collateral blood supply to the brainstem, either because of a hypoplastic or an absent contralateral VA or because of a deficient circle of Willis. [bib_ref] Bow hunter's syndrome revisited: 2 new cases and literature review of 124..., Jost [/bib_ref] Both VAs have adequate flow, and the patient remains asymptomatic when one VA is positionally occluded. [bib_ref] Positional occlusion of the vertebral artery: a rare cause of embolic stroke, Grossmann [/bib_ref] In 1978, Sorensen was the first to use the term "bow hunter's stroke" in a case of suspected brainstem infarction due to a VA injury, 5 although brain infarction is rare in BHS. [bib_ref] Positional occlusion of the vertebral artery: a rare cause of embolic stroke, Grossmann [/bib_ref] [bib_ref] Bow hunter's stroke due to prominent degenerative spinal disorder, Andereggen [/bib_ref] [bib_ref] Wake-up stroke in a young woman with rotational vertebral artery occlusion due..., Okawa [/bib_ref] [bib_ref] Positional occlusion of vertebral artery due to cervical spondylosis as rare cause..., Nishikawa [/bib_ref] The suspected cause of cerebral or cerebellar infarction in BHS is hemodynamic impairment [bib_ref] Surgical treatment of vertebral artery insufficiency caused by cervical spondylosis, Bakay [/bib_ref] [bib_ref] Surgical treatment of vertebral artery insufficiency caused by cervical spondylosis, Nagashima [/bib_ref] [bib_ref] Mechanical occlusion of the vertebral artery. A new clinical concept, Husni [/bib_ref] [bib_ref] Bow hunter's syndrome revisited: 2 new cases and literature review of 124..., Jost [/bib_ref] [bib_ref] Bow hunter's stroke due to prominent degenerative spinal disorder, Andereggen [/bib_ref] or arterio-arterial embolism from the affected VA. [bib_ref] Positional occlusion of the vertebral artery: a rare cause of embolic stroke, Grossmann [/bib_ref] [bib_ref] Bow hunter's stroke due to prominent degenerative spinal disorder, Andereggen [/bib_ref] [bib_ref] Wake-up stroke in a young woman with rotational vertebral artery occlusion due..., Okawa [/bib_ref] [bib_ref] Positional occlusion of vertebral artery due to cervical spondylosis as rare cause..., Nishikawa [/bib_ref] In our patient, the hemodynamic mechanism for repeated cerebellar infarction in the right PICA distribution was less likely given that he was asymptomatic when the right VA was occluded at the C4 level. Moreover, the right intracranial VA and branches of the right PICA were demonstrated during dynamic DA of the left VA when the patient's head was tilted to the right. However, repeated cerebellar infarcts developed in the right PICA distribution in the 2 years after the C3-C6 left laminoplasty procedure that had been performed 5 years earlier; nonetheless, he denied that these events occurred during neck positional changes. Arterio-arterial embolism from the transiently occluded right VA was a likely mechanism for cerebellar infarction in this patient, although the precise mechanism remains unexplained for repeated infarction restricted to the ipsilateral PICA distribution.
BHS is caused by an atypical pattern of transient occlusion of a nondominant VA terminating in the PICA. [bib_ref] Rotational vertebral artery occlusion: mechanisms and long-term outcome, Choi [/bib_ref] [bib_ref] Bow hunter's stroke caused by a nondominant vertebral artery occlusion: case report, Matsuyama [/bib_ref] [bib_ref] A case of bow hunter's stroke caused by non-dominant vertebral artery, Yeh [/bib_ref] [bib_ref] Rotational vertebral artery syndrome due to compression of nondominant vertebral artery terminating..., Noh [/bib_ref] In addition, BHS that results from nondominant VA compression, similar to that in our patient, does not terminate in the PICA and has recently been reported in 20 cases. 14 Among these cases, BHS caused by thromboembolism without VA dissection was extremely rare in patients with degenerative cervical spondylosis. 9,10,14 Such a condition was reported in two patients with wake-up stroke in the PICA distribution: ipsilateral in one patient and contralateral to the affected nondominant right VA in the other. 9,10 The affected nondominant right VA was stenotic at C4-C5 or C5-C6 in the neutral position but completely occluded with head rotation to the right. 9,10 However, in our patient, repeated cerebellar infarction was not a wake-up stroke, and the nondominant right VA was not stenotic in the neutral head position [fig_ref] FIG. 4: Dynamic bilateral vertebral arteriography [/fig_ref] , although complete occlusion of the right VA was induced by head tilt to the right [fig_ref] FIG. 4: Dynamic bilateral vertebral arteriography [/fig_ref]. BHS induced by head tilt is an extremely rare condition. [bib_ref] Rotational vertebral artery occlusion: mechanisms and long-term outcome, Choi [/bib_ref] [bib_ref] Dominant vertebral artery occlusion during ipsilateral head tilt, Choi [/bib_ref] Treatment for BHS includes conservative treatment and surgery. Conservative treatment includes the avoidance of head rotation, cervical collars, and/or antiplatelet or anticoagulation therapy. [bib_ref] Bow hunter's syndrome revisited: 2 new cases and literature review of 124..., Jost [/bib_ref] [bib_ref] Rotational vertebral artery occlusion: mechanisms and long-term outcome, Choi [/bib_ref] [bib_ref] Rotational vertebral artery occlusion: a mechanism of vertebrobasilar insufficiency, Kuether [/bib_ref] Conservative management with antiplatelet agents may be considered the first line of treatment for BHS. [bib_ref] Rotational vertebral artery occlusion: mechanisms and long-term outcome, Choi [/bib_ref] Dual-antiplatelet therapy with aspirin and clopidogrel was used in our patient. The combination of these antiplatelet agents synergistically inhibits platelet aggregation. [bib_ref] Clopidogrel with aspirin in acute minor stroke or transient ischemic attack, Wang [/bib_ref] In comparison with aspirin monotherapy, such dual-antiplatelet therapy reduces the rate of stroke occurrence from 11.7% to 8.2% in patients who have transient ischemic attack or minor stroke. [bib_ref] Clopidogrel with aspirin in acute minor stroke or transient ischemic attack, Wang [/bib_ref] However, repeated cerebellar infarction occurred in our patient while on the therapy. Some possible reasons for these repeated occurrences include the limited contribution of these antiplatelet agents because the affected nondominant VA showed no stenosis in the neutral head position and/or antiplatelet resistance to one or both antiplatelet agents. [bib_ref] Antiplatelet resistance in stroke, Topçuoglu [/bib_ref] Nearly 50% of patients who receive conservative treatment are vulnerable to infarcts or have residual neurological deficits. [bib_ref] Rotational vertebral artery occlusion: a mechanism of vertebrobasilar insufficiency, Kuether [/bib_ref] Surgical intervention is proposed if medical treatment fails. 14 All previously reported patients with BHS caused by nondominant VA compression required surgical management. 14 Therefore, surgical treatment was considered in the current patient. An anterior approach is often used to remove osteophytes at C4-C6. [bib_ref] Surgical treatment of vertebral artery insufficiency caused by cervical spondylosis, Bakay [/bib_ref] [bib_ref] Surgical treatment of vertebral artery insufficiency caused by cervical spondylosis, Nagashima [/bib_ref] [bib_ref] Bow hunter's syndrome revisited: 2 new cases and literature review of 124..., Jost [/bib_ref] [bib_ref] Wake-up stroke in a young woman with rotational vertebral artery occlusion due..., Okawa [/bib_ref] [bib_ref] Positional occlusion of vertebral artery due to cervical spondylosis as rare cause..., Nishikawa [/bib_ref] [bib_ref] Bow hunter's syndrome by nondominant vertebral artery compression: a case report, literature..., Iida [/bib_ref] [bib_ref] Rotational vertebral artery occlusion-series of 9 cases, Lu [/bib_ref] [bib_ref] Rotational vertebrobasilar ischemia: hemodynamic assessment and surgical treatment, Vilela [/bib_ref] [bib_ref] Diagnosis and management of bow hunter's syndrome: 15-year experience at Barrow Neurological..., Zaidi [/bib_ref] PICA-to-PICA bypass surgery has also been performed in a patient. [bib_ref] Posterior inferior cerebellar artery to posterior inferior cerebellar artery in situ bypass..., Kan [/bib_ref] Recently, successful endovascular treatment with angioplasty and stenting for contralateral VA stenosis was reported as an attempt to augment collateral flow. [bib_ref] Endovascular treatment for bow hunter's syndrome: case report, Sugiu [/bib_ref] In a patient with transient occlusions of the VA by an osteophyte from a cervical vertebral body while changing neck positions and associated with repeated embolic infarction in the posterior circulation, a coil embolization procedure was performed at the distal portion of the affected VA to prevent further embolic infarction in the posterior circulation. [bib_ref] A subtype of bow hunter's syndrome requiring specific method for detection: a..., Mori [/bib_ref] The current patient underwent C3-C6 left laminoplasty for ossification of the posterior longitudinal ligament 5 years ago, and posterior fusion of C3-C6 was subsequently selected. After surgery, he experienced no additional symptoms or further cerebellar infarctions.
## Observations
Repeated cerebellar infarction developed in the right PICA distribution in our patient. No VA stenosis was observed on 3D-CTA and DSA in the neutral head position, although DA revealed that the nondominant right VA was occluded by an osteophyte from the C4 vertebral body and that the right PICA branches passed through the distal right VA from the left VA without producing any symptoms when the patient's head was tilted to the right. Transient nondominant VA occlusion has been rarely reported as a cause of BHS when the head is tilted.
## Lessons
In patients who show an association between repeated cerebellar infarction of unknown cause and degenerative cervical spondylosis, the occurrence of BHS should be considered even in the following instances: VAs are not stenotic in the neutral head position, a nondominant VA connects to the opposite VA and the basilar artery, and no characteristic symptoms are induced by head and neck positional changes. To gain further information, additional head tilt is recommended during a dynamic DA study for patients with suspected BHS.
[fig] FIG. 1: MRI and MRA on the patient's first admission to our hospital (A, B, D, and E) and follow-up MRA at his first admission (C and F). A: DWI showing a high signal intensity lesion in the medial right PICA distribution. Brain (B) and cervical (E) MRA on admission. The right VA is not demonstrated. D: FLAIR image showing a previous right cerebellar infarction as a small hypointensity lesion in the lateral right PICA distribution. Brain (C) and cervical (F) MRA 3 days later. The nondominant right VA is observed from the origin to the junction of the left VA and basilar artery. FLAIR = fluid-attenuated inversion recovery. [/fig]
[fig] FIG. 2: MRI and MRA on the patient's second admission to our hospital (A, B, D, and E) and follow-up MRA at his second admission (C and F). A: DWI showing a small, high signal intensity lesion in the right side of the inferior cerebellar vermis. Brain (B) and cervical (E) MRA on admission. The right VA is not observed as it was on his previous admission (seeFig. 1B and E). D: FLAIR image showing past cerebellar infarcts in the right PICA distribution as hypointense lesions. Brain (C) and cervical (F) MRA 12 days later. The right VA appears again as previously demonstrated (seeFig. 1Cand F). [/fig]
[fig] FIG. 3: 3D-CTA. A: Posterior view with volume-rendering method showing C3-C6 left laminoplasty performed at a previous clinic. B: Coronal reformatted image. An osteophyte of C4 (red arrow) compresses the right VA laterally without stenosis. [/fig]
[fig] FIG. 4: Dynamic bilateral vertebral arteriography. A-C: DSA of the right VA at the neutral position of the head and neck. The hemispheric (arrows) and vermian (arrowheads) branches of the right PICA are demonstrated by injection of the right VA. A: Anterior-posterior view. B: Lateral view. C: Lateral view of 3D-rotation angiography with the volume-rendering method. D-H: Dynamic DA of the right VA. No stenosis is seen in the right VA at the neutral position (D), rotation of the neck to the right (E) and to the left (F), and head tilt to the left (H). Occlusion of the right VA is demonstrated at C4 level during head tilt to the right (G). I-K: Dynamic DSA of the left VA during head tilt to the right. I and J: Anterior-posterior view. K: Lateral view. The distal portion of the right VA is opacified in a retrograde fashion from the left VA (I), and both PICA hemispheric (arrows) and vermian (arrowheads) branches are subsequently shown (J and K). [/fig]
|
Effect of Daily Ingestion of Quercetin-Rich Onion Powder for 12 Weeks on Visceral Fat: A Randomised, Double-Blind, Placebo-Controlled, Parallel-Group Study
Quercetin, which is frequently found in vegetables such as onion, is widely found to have biological activities such as visceral fat reduction. Therefore, we performed this randomised double-blind placebo-controlled parallel-group study and analysed the effects of daily intake of quercetin-rich onion on visceral fat for 12 weeks. Seventy healthy Japanese subjects whose body mass index (BMI) was ≥23 and <30 were recruited and randomly assigned to either the quercetin-rich onion group or placebo group. The subjects ingested 9 g of onion powder per day for 12 weeks. We conducted medical interviews, hematological and biological tests; measured body composition and vital signs; and analysed the Food Frequency Questionnaire weeks 0, 4, 8, and 12. Abdominal fat area was measured using computed tomography scanning at weeks 0 and 12. No significant differences in visceral fat area (VFA) were observed between the two groups. However, in subjects whose high-density lipoprotein cholesterol was lower, VFA was significantly lower in the quercetin-rich onion group. In addition, alanine aminotransferase was significantly lower in the quercetin-rich onion group than in the placebo group. Thus, the results suggest that quercetin-rich onion may be beneficial for preventing obesity and improving liver function.Nutrients 2020, 12, 91 2 of 12 including prevention of metabolic syndrome for fresh vegetables, research on and development of various vegetables containing highly functional components is expected.Quercetin is a flavonoid that is abundant in vegetables such as onions and in fruit, tea, and wine[3][4][5]. It exhibits antioxidant[6,7]and antihypertensive effects[8,9]. In a placebo-controlled double-blind parallel-group comparison study for subjects with body mass index (BMI) of ≥25 kg/m 2 , daily intake of tea containing 110 mg quercetin glycoside for 12 weeks significantly decreased visceral fat area (VFA) at weeks 8 and 12 after intake compared to that in the placebo food intake group [10]. The mechanism underlying suppression of fat accumulation by quercetin was indicated to involve suppression of the expression of peroxisome proliferator-activated receptor γ which is related to fat accumulation, and sterol regulatory element-binding protein 1 and fatty acid synthase which are related to fatty acid synthesis, and to increase the expression of cAMP which is related to lipolysis[11][12][13].Onion is the one of the most popular vegetables in the world, and a major source of quercetin (28.4-48.6 mg/100 g)[14]. In the current clinical trial, we used the two quercetin-rich onion cultivars 'Quergold' and 'Sarasara-gold' to adjust the amount of quercetin contained in the test food. These onions were developed via selective breeding to develop onions containing increased amounts of quercetin. One 'Quergold' onion bulb (~140 g) contains 109 mg quercetin aglycone[15], and one 'Sarasara-gold' onion bulb (~200 g) contains 200 mg quercetin glycoside. The above findings suggest that the intake of 'Quergold' and 'Sarasara-gold' onions could provide sufficient quercetin in the range of reasonable daily volume of onion for visceral fat reduction.Therefore, to evaluate the effects of intake of quercetin-rich onion on visceral fat, we conducted a randomised double-blind placebo-controlled parallel-group study involving healthy subjects with BMI ≥23 kg/m 2 and <30 kg/m 2 , this range was defined as normal-high obesity and class I obesity.
# Introduction
According to the 2016 National Health and Nutrition Survey Report of the Ministry of Health, Labour and Welfare, Japan, the proportion of the ≥20-year-old population suspected to have visceral fat obesity is 29.9% for men and 14.4% for women . Visceral fat has higher fat synthesis and degradation activity than subcutaneous fat, and during fasting, visceral fat supplies free fatty acids and glycerol, which are formed on degradation of triglycerides (TGs), to the liver via the portal vein. It is thought that this excessive inflow of free fatty acids and glycerol into the liver induces insulin resistance, which in turn leads to abnormal glucose metabolism, abnormal lipid metabolism and hypertension. Thus, since visceral fat accumulation is related to lifestyle-related diseases, reduction of visceral fat has become an important aspect of prevention of such diseases. In 2015, the system of 'Foods with Function Claims' was established in Japan. Because this system permits health claims Interviews were conducted by a research doctor and a nurse to obtain medical information, as well as for checking vital signs, body composition measurement and blood sampling, during the second (week 0), third (week 4), fourth (week 8) and fifth (week 12) visits. VFA, subcutaneous abdominal fat area (SFA) and total abdominal fat area (TFA) were analysed at the second and fifth visits using CT scans. In addition, the subjects filled the Food Frequency Questionnaire Based on Food Groups (FFQg; Kenpakusha, Tokyo, Japan) from the second to fifth visits. The subjects ingested 9 g of powder (quercetin-rich onion or quercetin-free onion powder) daily for 12 weeks, at any time and in association with their favourite cooking method. Subjects daily recorded whether they ingested the test food or not, in a diary. During the entire course of this study, the subjects were asked to maintain their daily activities, including eating and exercise habits, to avoid any supplements and health foods and to restrict the consumption of onion and quercetin-rich foods. The subjects used a diary to record their daily activities, which was reviewed by a medical doctor or a nurse during each visit.
The primary outcome was the change in VFA. The secondary outcomes were TFA, SFA, body weight (BW), BMI, body fat rate (BFR), abdominal circumference, hospital blood pressure (BP), home BP and thiobarbituric acid reactive substances (TBARS). In addition, safety outcomes changed in CBCs, liver function, renal function, lipid profiles and blood glucose profiles.
## Study subjects
One-hundred fifty-eight volunteers were screened on their first visit; all the volunteers provided written informed consent to participate in this study. After screening, 70 healthy Japanese subjects whose BMI was ≥23 kg/m 2 and <30 kg/m 2 were enrolled. Inclusion and exclusion criteria are summarized in. The methods used for assignments and blinding are briefly described: Okamoto Plant Breeding Co., Ltd., which provided the test foods, assigned the food identification numbers, which were randomly assigned to either quercetin-rich onion powder or placebo powder, and were then printed on the food packaging. Food identification number information was strictly sealed and then transferred Nutrients 2020, 12, 91 4 of 12 to a third-party data center independent of the study (Media Educational Center, Hokkaido Institute of Information Technology, Ebetsu, Hokkaido, Japan). The eligible subjects were randomly assigned to the quercetin-rich onion or placebo onion groups and stratified by sex, age, and BMI during the first visit. The assignments were computer generated and based on stratified block randomisation at a third-party data center; the block size was 32 (gender (male/female), age (30s, 40s, 50s and 60s), and BMI (<24, <26, <28 and ≥28.1 kg/m 2 )). The research collaborators at Hokkaido Information University, including medical doctors, nurses, statistical analysts, and clinical research coordinators, were blinded to the food identification numbers and the assignment information during the trial period. This information was disclosed only after all the analytical data were collected and the subjects to be included in the efficacy analysis and the method to be used for statistical analyses were finalized.
## Preparation of the test food
The quercetin-rich onion cultivars 'Quergold' and 'Sarasara-gold' were cultivated in Hokkaido, Japan. The quercetin-rich onion was supplied from Okamoto Plant Breeding Co., Ltd. (Hokkaido, Japan). The manufacturing process for the onion powder was as follows: peeled onions were soaked in hypochlorous acid solution for 20 min, thoroughly rinsed with water, cut into 2-mm wide pieces, dried at 45 - C for 30 h and then at 45 - C for 4 h, sterilized at 60 - C for 120 min and finally powderized. The quercetin-rich onion powder contained 'Quergold' and 'Sarasara-gold' at a ratio of 4:6. White onions, cultivated in U.S., did not have detectable levels of quercetin, and were used for the placebo powder, and the placebo powder used commercial products. The powders were analysed at Japan Food Research Laboratories (Hokkaido, Japan), the amount of quercetin was analysed at Food Research Institute, National Agriculture and Food Research Organization, and analytical results for the nutrient composition of the quercetin-rich onion and placebo powders are summarized in. Both powders were identical in appearance. The intake amount of quercetin in this trial was taken as a quantity for which an effect can be expected. This intake amount was decided on the basis of a previous study, which reported that 72 mg of quercetin (aglycone equivalent) leads to fat reduction. Surveys in Hokkaido reported the daily intake of quercetin to be approximately 16 mg (ranging from 0.5 to 56.8 mg/day). Therefore, the daily intake was set at approximately 90 g containing approximately 60 mg of quercetin in terms of onion consumption in this trial. Since both the onion and quercetin quantities used in this trial are within the range of daily intake, it can be said that the intake in this study is a safe amount for consumption.
## Physical, hematological and biological assessments
Blood was collected from subjects after a 12-h fast and used for the following TBARS, hematological examinations: white blood cell (WBC), red blood cell (RBC), hemoglobin (Hb) and platelet (Plt) counts and hematocrit (Ht). Biological examinations were as follows: liver function (aspartate aminotransferase (AST), alanine aminotransferase (ALT), γ-glutamyltranspeptidase (γ-GTP), alanine phosphatase (ALP), and lactate dehydrogenase (LDH)); renal function (blood urea nitrogen (BUN), creatinine (CRE), and uric acid (UA)); lipid profiles (total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), Nutrients 2020, 12, 91 5 of 12 high-density lipoprotein cholesterol (HDL-C) and TG); and blood glucose profiles (fasting plasma glucose (FPG), hemoglobin A1c (HbA1c) and insulin).
The hematological tests were performed at Sapporo Clinical Laboratory, Inc. (Hokkaido, Japan). TBARS were measured using TBARS Assay Kit (Cayman Chemical, Ann Arbor, MI, USA) at Hokkaido Information University. Body composition (BW, BFR and BMI) were measured using Body Composition Analyser DC-320 (Tanita Corp, Tokyo, Japan) at Hokkaido Information University.
Hospital BP was measured by a nurse with the Automatic Blood Pressure Monitor HEM-7080IC (Omron Healthcare Co., Ltd., Kyoto, Japan), using the upper arm region of the non-dominant arm after >10-min rest. Three sequential measurements were performed, and the median of the measurements was taken at each evaluation point. Home BP was measured by the subjects with the HEM-7080IC instrument, using the upper arm region of the non-dominant arm. The subjects measured their BP daily for 1 week before visits 2-5, within 1 h after waking up (morning BP) and before going to bed (evening BP). Three sequential measurements were obtained, and the median of the measurements was taken each day. The average BP during 3 d before each evaluation point was evaluated.
## Measurement of abdominal fat area
For measurement of abdominal fat area, the subjects underwent computed tomography (CT) scanning, which was performed using a GEMINI GXL 16 system (PHILIPS, Amsterdam, Netherlands). The total fat area was calculated as the sum of the SFA and VFA. The imaging settings were as follows: tube voltage, 120 kVp; mA value, 150 mA; window level, 35; and window width, 350. A single slice on the L4 level was selected. At the time of measurement, the subjects lay on their back, with both hands up and breathing held at maximum exhalation. Abdominal body fat areas (SFA and VFA) were calculated on the basis of an abdominal CT scan image by using visceral fat measurement software (Fat Checker, J-MAC SYSTEM, INC., Sapporo, Japan).
## Ffqg
The FFQg is a semi-quantitative dietary assessment used to estimate nutrient intake on the basis of the subject's regular diet. This questionnaire included 29 food groups and 10 types of cooking methods. For each question, the subjects reported the weekly amount and frequency of food intake for the past month at each visit. From the report, regular and nutrient intakes (calorie, protein, lipid, carbohydrate, dietary fibre and sodium chloride equivalent) were estimated.
## Safety assessment
We considered an adverse event to be any undesirable or unintentional sign (including abnormal fluctuations in laboratory values), symptoms or illness that occurred during the food intake period. A side effect was defined as adverse events for which a causal relationship with the test food could not be completely denied ('probably related to intake of test food' or 'related to intake of test food'). In this study, events that occurred between the start date of intake and the end date of intake were recorded. We further assessed the incidence of unfavourable symptoms and findings and abnormal changes in laboratory variables as adverse events. The severity of adverse events and their relationship with the test food were classified according to the protocol criteria set by the investigator. Laboratory variables related to safety outcome were assessed according to the guideline on the side effect criteria, as defined by the Japanese Society of Chemotherapy, and the excluded anomaly levels, as defined by the investigator. All adverse events were reported as follows: symptoms, occurrence date and severity (mild/moderate/severe); relation to test food (not relevant/probably not relevant/probably relevant/relevant/not assessable); and continuation or discontinuation, treatment, and outcome (day).
## Ethics
This study was conducted according to the guidelines laid down in the Declaration of Helsinki (revised by the Fortaleza General Meeting of the World Medical Association), and all procedures Nutrients 2020, 12, 91 6 of 12 involving human subjects were approved by the ethics committee of Hokkaido Information University (Ebetsu, Hokkaido, Japan; approved on May 25, 2018; approval number: 2018-05). Written informed consent was obtained from all subjects. This study is also in compliance with the ethical guidelines on medical research in humans as per the Ministry of Education, Culture, Sports, Science and Technology and the Ministry of Health, Labour and Welfare. This trial was registered at University Hospital Medical Information Network-Clinical Trials Registry (UMIN-CTR) (www.umin.ac.jp/ctr/index.htm; registered on 17 July 2018; registration number: UMIN000033410).
# Statistical analysis
Student's t-tests were used to analyse the primary, secondary, FFQg and safety outcomes by comparing changes in values at baseline to those at week 4, week 8 and/or week 12 between the two subject groups. For subject characteristics, the Fisher's exact probability test was used for sex and the Mann-Whitney U-test was used for the intake rate; Student's t-tests were used for the other subject characteristics. All statistical analyses were performed using SPSS version 25.0 (IBM Japan, Ltd., Tokyo, Japan), and p < 0.05 was considered statistically significant.
## Sample size
The sample size was statistically determined to obtain a power of 80% with a two-sided significance level of 5%. The data from a previous clinical trial on the effect of quercetin on VFAindicated that to demonstrate a difference in VFA at week 12 (postulated to have an intergroup difference of 10 cm 2 with a standard deviation of 13 cm 2 ), a sample size of 60 (30 in each group) was required. Assuming a 15% loss in the follow-up rate, 70 subjects (35 in each group) were enrolled.
# Results
## Flow chart for subject involvement in trial and subject characteristics
A flow chart of the subject's involvement throughout the trial schedule is presented in. Subjects who provided informed consent (n = 158) were assessed for eligibility, and from these, a total of 70 subjects were enrolled in this study. All enrolled subjects were randomised to one of the two intervention groups (placebo group, n = 35; quercetin-rich onion group, n = 35). Three subjects dropped out for personal reasons before the trial started, one subject withdrew because of elevated TC at week 0 (before test food intake), two subjects withdrew because they started undergoing medical treatment, and three subjects dropped out due to personal reasons. Finally, 61 subjects completed this trial: 31 in the quercetin-rich onion group and 30 in the placebo group.
Sixty-seven subjects, excluding the three subjects who dropped out before the start of the trial, were included in the safety analysis. We excluded seven subjects from the efficacy analysis: abnormal value baseline (n = 2), missing data of primary outcome (n = 1), menopause symptoms (n = 1), changes in activity due to injury (n = 1) and diagnosis of obesity (n = 2). Therefore, we included 27 subjects in the quercetin-rich onion group and 27 in the placebo group. In addition, two subjects (quercetin-rich onion group) were only excluded from efficiency analysis of body composition, BP, and the oxidative marker at all points because they were absent at visit 2 (baseline), and one subject (quercetin-rich onion group) was only excluded from efficiency analysis of body composition, BP, and the oxidative marker at week 4 because of absence at visit 3 (week 4).
Data for sex ratio, mean age, body weight, BFR, BMI, abdominal circumference and intake rate for each group are showed in. These characteristics did not significantly differ between the two groups. These results proved the appropriate allocation of subjects between quercetin-rich onion and placebo groups. marker at all points because they were absent at visit 2 (baseline), and one subject (quercetin-rich onion group) was only excluded from efficiency analysis of body composition, BP, and the oxidative marker at week 4 because of absence at visit 3 (week 4).
Data for sex ratio, mean age, body weight, BFR, BMI, abdominal circumference and intake rate for each group are showed in. These characteristics did not significantly differ between the two groups. These results proved the appropriate allocation of subjects between quercetin-rich onion and placebo groups.
## Efficacy of quercetin-rich onion in case of vfa
The effects of quercetin-rich onion on VFA are summarized in. There were no significant differences between the quercetin-rich onion and placebo groups in terms of changes in VFA in all the subjects analysed. Next, we performed exploratory subgroup analysis for HDL-C. In subjects whose HDL-C was lower to normal (40-74 mg/dL; 74 mg/dL was the 75th percentile of baseline HDL-C-values), VFA was significantly lower in the quercetin-rich onion group than in the placebo group (p = 0.046).
## Efficacy of quercetin-rich onion in case of tfa, sfa and body composition
The effects of quercetin-rich onion on TFA, SFA and body composition (BW, BFR, BMI and abdominal circumference) are summarized in. There were no significant differences between the quercetin-rich onion and placebo groups in terms of changes in these parameters.
## Efficacy of quercetin-rich onion in case of bp and the oxidative marker
To determine the effects of quercetin-rich onion on BP, changes in hospital systolic BP (SBP), hospital diastolic BP (DBP), home SBP and DBP (morning and evening) and the oxidative marker that is TBARS, were evaluated. There were no statistically significant differences between the quercetin-rich onion and placebo groups for these parameters.
## Assessment of dietary nutrients in case of liver marker
As exploratory analysis, we evaluated the effects of quercetin-rich onion on liver marker, changes in AST, ALT, γ-GTP, ALP and LDH. ALT was significantly lower in the quercetin-rich onion group than in the placebo group (p = 0.035).
## Assessment of dietary nutrients of subjects during the study
For assessment of dietary nutrients, the subjects filled the FFQg. Protein intake at week 4 and 8 (p = 0.008, 0.041), lipid intake at week 8 (p = 0.044) and dietary fibre intake at week 8 (p = 0.011) for the quercetin-rich onion group were significantly lower than those for the placebo group. However, there were no statistically significant differences in the intake of calories, carbohydrates, and sodium chloride equivalent between the quercetin-rich onion and placebo groups. Therefore, we consider that this difference did not affect our results. In addition,shows that the test foods contained different amounts of sodium (placebo onion powder: 12.2 mg, quercetin-rich onion powder: 1.8 mg). However, we consider that this difference did not affect our results, as it is a minimal amount as compared to the subjects' average daily intake of sodium, which is approximately 3,000 mg.
# Safety analysis
For safety analysis of quercetin-rich onion, complete blood counts (WBC count, RBC count, Hb, Ht and Plt count), renal function (BUN, CRE and UA), lipid profile (TC, LDL-C, HDL-C and TG) and blood glucose profile (FPG, HbA1c and homeostasis model assessment of insulin resistance [HOMA-IR]) were evaluated. Minimal changeswere observed in each group. During monitoring of adverse effects (variation in clinical data and clinical observations), four subjects (placebo: n = 3; quercetin-rich onion group: n = 1) briefly experienced mild gastrointestinal symptoms after test food intake. For the other adverse events, all symptoms were mild, and the subjects recovered in a few days; the principal investigator judged that they were not related to test food intake.
# Discussion
In this clinical trial, we assessed the effect of daily quercetin-rich onion intake for 12 weeks on visceral fat. No significant differences were observed between the quercetin-rich onion and placebo groups in efficacy analysis. However, in subjects whose HDL-C was ≥40 mg/dL and <74 mg/dL, VFA was significantly lower in the quercetin-rich onion group than in the placebo group. In addition, the levels of ALT, a liver marker, were significantly lower in the quercetin-rich onion group than in the placebo group.
VFA, the primary endpoint, did not significantly differ between the quercetin-rich onion and placebo groups. However, exploratory subgroup analysis based on HDL showed that VFA was significantly improved in quercetin-rich onion group compared to the placebo group subjects with HDL-C ≥ 40 mg/dL and <74 mg/dL. A meta-analysis involving 18 studies in Asia and Australia, including 222,975 subjects, demonstrated that obesity indicators in relation to BMI and waist circumference had a stronger relationship with HDL-C and TG than with TC and/or LDL-C. These findings suggested that the effect of quercetin-rich onion on visceral fat was strong in subjects with low HDL-C.
The quercetin-rich onion and placebo groups did not show significant differences with respect to the second outcome, including TFA, SFA, BW, BMI, BFR and abdominal circumference. A previous clinical trial suggested that dietary restriction for 14 d did not reduce SFA; however, VFA was found to decrease. In addition, in vivo experiments in the same study showed that the expression of genes related to fat metabolism (beta3-adrenergic receptor, hormone sensitive lipase, peroxisome proliferator-activated receptor-gamma, and uncoupling protein-2) changed in the VFA because of dietary restrictions but did not change in subcutaneous fat. These findings suggested that VFA was affected earlier than SFA in our trial. Moreover, the placebo group showed an increase in BFR from baseline to 12 weeks after the intake, but the quercetin-rich onion group showed suppression of this increase.
BP, including hospital BP and home BP, did not improve on intake of quercetin-rich onion. A meta-analysis study reported that an intake amount of ≥500 mg quercetin was required to obtain an effect on BP. Previous clinical trial with hypertensive and obese subjects reported that BP improved on daily intake of onion epidermis extract for 6 weeks, amounting to 162 mg quercetin per day. The levels of TBARS, the oxidation marker, did not improve on intake of quercetin-rich onion in our clinical trial. Quercetin exhibits antioxidant activity by regulating gene expression related to the production of reactive oxygen species. Consumption of 1 g quercetin daily for 2 weeks significantly reduced TBARS levels after exercise in a crossover comparative study. Thus, it is necessary to reconsider the intake amount of quercetin used for analysing its BP-improving and antioxidant effects.
The levels of ALT were significantly lower in the quercetin-rich onion group than in the placebo group. There are several studies on the effects of quercetin on liver function in in vivo experiments. For example, quercetin was found to inhibit liver fat accumulation in Western-diet model mice. There are two types of non-alcoholic fatty liver disease (NAFLD): simple fatty liver (non-alcoholic fatty liver [NAFL]) with mild symptoms and non-alcoholic steatohepatitis (NASH) with severe symptoms; it is important to prevent progression before the onset of NASH. In this trial, although the effect of daily intake of quercetin-rich onion on VFA was limited, it can be expected to cause improvement in NAFL. In addition, VFA tended to decrease among subjects consuming the quercetin-rich onion powder whose ALT values were on the higher end of the physiological range (placebo: 1.6 ± 8.2 cm 2 , quercetin-rich onion: −6.4 ± 12.2 cm 2 , p = 0.060). However, this study did not include sufficient evaluation items for liver function and NAFL. Further studies are required to examine such items to verify the effect of quercetin-rich onion on liver function.
In this clinical trial, no severe side effects or severe adverse events were observed during physical and blood examinations or reported in the medical interviews. These results confirm the safety of daily intake of quercetin-rich onion for 12 weeks.
A limitation of this study is that we showed the effect of quercetin-rich onion on VFA with limited subjects. Therefore, modification of inclusion criteria, exclusion criteria and number of subjects included in the trial is required to further clarify the effects of quercetin-rich onion on VFA. In our clinical trial, ALT levels improved on consumption of quercetin-rich onion; however, the items for evaluation of liver function were limited. Therefore, we need to reconsider the variables used to prove the effect of quercetin-rich onion on liver functions. Another limitation of this study was the ingestion period; the effects of long-term ingestion need to be evaluated using other evaluation outcomes.
# Conclusions
In this 12-week randomised double-blind placebo-controlled parallel-group comparative study, we assessed the effect of daily intake of quercetin-rich onion for 12 weeks on visceral fat. No significant differences were observed between the quercetin-rich onion and placebo groups in efficacy analysis. However, in the case of subjects whose HDL-C levels were low, the VFA was significantly lower in the quercetin-rich onion group. In addition, ALT was significantly lower in the quercetin-rich onion group than in the placebo group. The findings indicate that consumption of quercetin-rich onion may be beneficial for preventing obesity and improving liver function.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6643/12/1/91/s1,: Dietary nutrients for the placebo and quercetin-rich onion groups,: Complete blood counts, renal function, lipid profile and blood glucose profile,: Adverse events. |
A semi-automated technique for labeling and counting of apoptosing retinal cells
Background: Retinal ganglion cell (RGC) loss is one of the earliest and most important cellular changes in glaucoma. The DARC (Detection of Apoptosing Retinal Cells) technology enables in vivo real-time non-invasive imaging of single apoptosing retinal cells in animal models of glaucoma and Alzheimer's disease. To date, apoptosing RGCs imaged using DARC have been counted manually. This is time-consuming, labour-intensive, vulnerable to bias, and has considerable inter-and intra-operator variability.Results: A semi-automated algorithm was developed which enabled automated identification of apoptosing RGCs labeled with fluorescent Annexin-5 on DARC images. Automated analysis included a pre-processing stage involving local-luminance and local-contrast "gain control", a "blob analysis" step to differentiate between cells, vessels and noise, and a method to exclude non-cell structures using specific combined 'size' and 'aspect' ratio criteria. Apoptosing retinal cells were counted by 3 masked operators, generating 'Gold-standard' mean manual cell counts, and were also counted using the newly developed automated algorithm. Comparison between automated cell counts and the mean manual cell counts on 66 DARC images showed significant correlation between the two methods (Pearson's correlation coefficient 0.978 (p < 0.001), R Squared = 0.956. The Intraclass correlation coefficient was 0.986 (95% CI 0.977-0.991, p < 0.001), and Cronbach's alpha measure of consistency = 0.986, confirming excellent correlation and consistency. No significant difference (p = 0.922, 95% CI: −5.53 to 6.10) was detected between the cell counts of the two methods.Conclusions:The novel automated algorithm enabled accurate quantification of apoptosing RGCs that is highly comparable to manual counting, and appears to minimise operator-bias, whilst being both fast and reproducible. This may prove to be a valuable method of quantifying apoptosing retinal cells, with particular relevance to translation in the clinic, where a Phase I clinical trial of DARC in glaucoma patients is due to start shortly.
# Background
Glaucoma is a chronic degenerative optic neuropathy that results in irreversible loss of retinal ganglion cells (RGC; the neurons that relay information from the retina to the cortex). RGC loss, coupled with degeneration of the RGC axons, results in optic disc "cupping" and a progressive visual field loss that is characteristic of glaucoma [bib_ref] Quantitative assessment of structural damage in eyes with localized visual field abnormalities, Bagga [/bib_ref]. In glaucoma, most RGC loss occurs through the process of apoptosis (programmed cell death) [bib_ref] Retinal ganglion cell death in experimental glaucoma and after axotomy occurs by..., Quigley [/bib_ref]. Apoptosis has a central role in several other neurodegenerative diseases [bib_ref] Glaucoma and Alzheimer's disease in the elderly, Bizrah [/bib_ref] [bib_ref] Apoptosis and caspases in neurodegenerative diseases, Friedlander [/bib_ref] [bib_ref] Apoptosis and human neurodegenerative diseases, Berry [/bib_ref] , as well as glaucoma, with evidence that the targeting of pro-apoptotic activity may be neuroprotective against Neurodegeneration [bib_ref] Glaucoma and Alzheimer's disease in the elderly, Bizrah [/bib_ref] [bib_ref] Apoptosis and caspases in neurodegenerative diseases, Friedlander [/bib_ref] [bib_ref] Apoptosis and human neurodegenerative diseases, Berry [/bib_ref] [bib_ref] Neuroprotection in glaucoma -is there a future role?, Baltmr [/bib_ref] [bib_ref] Assessment of neuroprotection in the retina with DARC, Guo [/bib_ref] [bib_ref] Ocular neuroprotection by siRNA targeting caspase-2, Ahmed [/bib_ref] [bib_ref] Prospects for antiapoptotic drug therapy of neurodegenerative diseases, Waldmeier [/bib_ref] [bib_ref] Targeting programmed cell death in neurodegenerative diseases, Vila [/bib_ref].
Glaucoma is often diagnosed late in the course of the disease using the gold standard method of perimetry, since visual field defects are not detected until up to 40% of RGCs have been lost [bib_ref] Retinal ganglion cell atrophy correlated with automated perimetry in human eyes with..., Quigley [/bib_ref]. However, since timely intervention can halt (but not reverse) glaucomatous progression, much recent research has focused on identifying early diagnostic markers of glaucoma. RGC apoptosis has been shown to be one of the initial pathological processes in glaucoma [bib_ref] Real-time imaging of single nerve cell apoptosis in retinal neurodegeneration, Cordeiro [/bib_ref] [bib_ref] Detection of early neuron degeneration and accompanying microglial responses in the retina..., Naskar [/bib_ref] , and its detection could facilitate early diagnosis and management of this condition. One of the first events in apoptosis is externalisation of phosphatidylserine (a membrane phospholipid) from the inner to the outer leaflet of the cell membrane. Annexin V is a protein with a high affinity to exposed phosphatidylserine [bib_ref] In vivo detection and imaging of phosphatidylserine expression during programmed cell death, Blankenberg [/bib_ref]. Imaging of radiolabeled Annexin V therefore enables detection of apoptotic cells. Clinical studies have utilized Technetium-99 m radiolabeled Annexin V for the non-invasive detection and serial imaging of apoptosis in various clinical settings, such as acute myocardial ischemia [bib_ref] Non-invasive in vivo imaging of myocardial apoptosis and necrosis, Flotats [/bib_ref] , cardiac allograft rejection, breast cancer [bib_ref] In vivo and in vitro measurement of apoptosis in breast cancer cells..., Yang [/bib_ref] and anti-cancer treatment induced apoptosis [bib_ref] In vivo imaging of apoptosis by 99mTc-Annexin V scintigraphy: visual analysis in..., Kartachova [/bib_ref] [bib_ref] In vivo imaging of radiation-induced apoptosis in follicular lymphoma patients, Haas [/bib_ref].
Recently, our laboratory has developed a technique by which Annexin V is labeled with a fluorescent marker, which is subsequently intravitreally administered [bib_ref] Real-time imaging of single nerve cell apoptosis in retinal neurodegeneration, Cordeiro [/bib_ref]. A 488 nm wavelength argon laser is used to excite the administered annexin V-bound fluorophore, and a photodetector system with a 521-nm cut-off filter enables detection of the fluorescence light emission. The fluorescent retinas are imaged with Confocal laser scanning ophthalmoscopy. This novel technology has enabled the non-invasive in vivo realtime visualisation of single retinal cells undergoing apoptosis, and has been given the acronym DARC (Detection of Apoptosing Retinal Cells). [bib_ref] Real-time imaging of single nerve cell apoptosis in retinal neurodegeneration, Cordeiro [/bib_ref] DARC has been used in animal models of glaucoma [bib_ref] Imaging apoptosis in the eye, Cordeiro [/bib_ref] and Alzheimer's disease [bib_ref] Imaging multiple phases of neurodegeneration: a novel approach to assessing cell death..., Cordeiro [/bib_ref] , highlighting the role of apoptosis in the early stages of both diseases. It has also been studied in the evaluation of neuroprotective strategies in animal models of glaucoma, such as glutamate modulation [bib_ref] Assessment of neuroprotective effects of glutamate modulation on glaucoma-related retinal ganglion cell..., Guo [/bib_ref] , amyloid-beta targeting therapyand topical Coenzyme Q10 [bib_ref] Assessment of neuroprotection in the retina with DARC, Guo [/bib_ref].
To date, quantitative assessment of RGC apoptosis has been a manual process. The number of apoptosing RGC's is counted by one or more persons using software such as Ima-geJ®. Such manual assessment procedures have several disadvantages related to the precision and accuracy of cell counts. In terms of precision, manual quantification involves subjective judgment increasing operator-dependency -especially when images are of low qualitypotentially leading to substantial intra-and inter-operator variability. In terms of accuracy, if the operator is not blinded then this technique is potentially vulnerable to bias. Furthermore manual quantification is time-consuming and labourintensiveespecially if more than one individual is needed to maximise precision and accuracyrendering the analysis of a large number of images challenging.
In this study, a semi-automated technique has been developed for the quantification of apoptosing retinal cells on DARC images. A total of 66 DARC images were analysed by a novel automated algorithm and by 3 human operators. The total cell counts of the automated algorithm were compared to the mean cell counts of three human operators. The automated algorithm was found to minimise operator-dependency while providing fast, accurate, and reproducible cell-counts.
# Methods
## Darc images
DARC images were randomly selected from a database of approximately 3000 DARC images of rat eyes, which had either undergone surgically-induced intraocular pressure (IOP) elevation or had been exposed to neurotoxic substances or various treatments, at different time points. Images were captured as described in previous publications [bib_ref] Real-time imaging of single nerve cell apoptosis in retinal neurodegeneration, Cordeiro [/bib_ref] [bib_ref] Imaging multiple phases of neurodegeneration: a novel approach to assessing cell death..., Cordeiro [/bib_ref] [bib_ref] Imaging apoptosis in the eye, Cordeiro [/bib_ref] and operators were blinded to the type of insult which the eyes had undergone. The quality of images spanned a wide range in order to investigate the robustness of the technique. [fig_ref] Figure 1: Images A, B & C are examples of DARC images before undergoing... [/fig_ref] below represents examples of the variation in quality of the DARC images. Note that apoptosing retinal cells, imaged using a confocal laser scanning ophthalmoscope, appear as 'white spots' on the retina as previously described [bib_ref] Real-time imaging of single nerve cell apoptosis in retinal neurodegeneration, Cordeiro [/bib_ref] [bib_ref] Imaging apoptosis in the eye, Cordeiro [/bib_ref].
Cropping and re-sizing of DARC images pre-analysis DARC images were cropped to remove descriptive text at the bottom and to eliminate peripheral noise. They were then re-sized to 600 pixels square using the bilinear interpolation algorithm built into the "image resize" function in Adobe Photoshop (Adobe Inc). This was done purely in order to reduce image-processing time and we see no systematic influence of this level of down-sampling on processing of a random sample of the images tested.
# Manual analysis
Manual image analysis was performed by three blinded operators using ImageJ® (National Institutes of Mental Health, USA). The ImageJ 'multi-point' tool was used to label each structure in the image classed as an apoptosing cell. As each cell is labeled it is assigned a unique number enabling manual quantification of the total number of visible single apoptosing retinal cells an example of a manually labeled DARC image is shown.
# Automated analysis
The Matlab® (Mathworks Ltd) programming environment was used to develop a program for labeling and counting apoptosing retinal cells in DARC images. The stages of the semi-automated analysis performed by the program are described below. Of note, it is possible to automate the cropping and re-sizing of images by adding these functions to the Matlab script. This will enable the image analysis to be fully automated.
## Stage 1: pre-processing
A single DARC image can contain wide fluctuations in mean luminance and contrast levels within a given imageregion, which can interfere with subsequent thresholding and spatial filtering. To counteract this local luminance and contrast, structure was "flattened" within each image. Specifically, the mean and standard deviation of the grey levels in the locale of a given pixel are computed and used to effectively convert the pixel grey-level into a local z-score. To compute statistics within a locale we used convolution with Gaussian spatial filters, i.e. the local mean luminance of a pixel and its locale is simply a Gaussian blurred version of the original:
[formula] μ ¼ G s ⊗Ið1Þ [/formula]
where I is the source image and G s is a two-dimensional Gaussian filter (standard deviation, s). Similarly, the Gaussian-weighted standard deviation can be computed as follows:
[formula] σ ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi G s ⊗I 2 −μ 2 qð2Þ [/formula]
so that the final pre-processed image is then:
[formula] Z ¼ G s −μ σð3Þ [/formula]
The resultant Z is then processed with a conventional Laplacian-of-Gaussian (Δ 2 G t ) spatial-frequency band-pass filter (with standard deviation, u) to highlight high-energy isotropic image-structure. The operation of such a filter on DARC images that have and have not been pre-processed is illustrated in [fig_ref] Figure 2: The effects of pre-processing illustrated on a DARC image [/fig_ref] -D below. [fig_ref] Figure 2: The effects of pre-processing illustrated on a DARC image [/fig_ref] is the original image and 2B the result of filtering it with the Δ 2 G t filter. Note the weak (low-contrast) filter-responses in the lower left portion of the image in 2B. [fig_ref] Figure 2: The effects of pre-processing illustrated on a DARC image [/fig_ref] shows the pre-processed version of [fig_ref] Figure 2: The effects of pre-processing illustrated on a DARC image [/fig_ref] (generated with Eqs 1-3); note the uniformity of luminance and contrast structure therein. [fig_ref] Figure 2: The effects of pre-processing illustrated on a DARC image [/fig_ref] is a Laplacian-filtered version of the pre-processed image [fig_ref] Figure 2: The effects of pre-processing illustrated on a DARC image [/fig_ref]. Note that the filter response is now much more spatially uniform than in 2B. The candidate vasculature and cell-structure is now visible across the whole image, and will remain so after global thresholding used to isolate discrete image structure. The parameters used to pre-process the 600 pixel square source images were: s = 64 pixels, u = 1.5 pixels.
## Stage 2: cell identification
To identify image structure as cells we first apply imagethresholding to the filtered images; this simply sets all grey levels falling too near to the mean grey level of the whole image, to zero. The threshold (T) was fixed at 1.8 × the standard deviation of the image grey-level, which generally gives good subjective delineation of cell and vessel structure in the image. We then employed "blob-analysis" (using the regionprops routine in MatLab®) on the isolated regions that resulted from thresholding. This yields various features of each blob including the length along major (L maj ) and minor (L min ) axes length, its area (A) and the location of its centroid ([C x ,C y ]). We next perform categorisation of image structure based on these estimates. In [fig_ref] Figure 3: An illustration of the DARC image shown inFigure 2dafter undergoing thresholding and... [/fig_ref] , blobs have been categorized as cells (red), vessels (green) or noise (blue), based on the following criteria:
[formula] For noise : A < A min ; for vessels : L maj =L min À Á > Aspect min ; [/formula]
and all other blobs are classed as cells.
Pilot studies were performed to maximize agreement between the automated and manual cell counts (n.b. the inclusion of this stage is why we refer to the technique as 'semi-automated' rather than fully automated). Setting A min (minimum area -in square pixels-for a blob to be a candidate cell) to 9.0 and Aspect min (the minimum aspect ratio for a blob to be a candidate blood vessel) to 3.0 yielded total cell counts which best corresponded with mean manual cell count of three inexperienced and masked operators, and was therefore chosen and fixed for automated quantification. This is an important step as altering these parameters results in different classification of blobs. This is particularly true for the A min parameter, as this determines the minimum cut-off size for a blob to be classified as a cell rather than noise. The pilot studies enabled the five Matlab algorithm script parameters (s, u, T, A min and Aspect min ) to be fixed at the point of image analysis, enabling fully automated analysis by a single operator.
## Study protocol
For the purpose of this study, 66 post-insult images were picked randomly from this database with two exclusion criteria: the presence of "white" vessels (thought to be arising from Annexin 5 binding to the vascular endothelium) and insufficient image-quality to support manual cell identification. These images were analyzed using both manual and automated techniques, and this sample-size selected to reflect limits on the operator time available for manual counting. The study protocol is summarized in the flow chart [fig_ref] Figure 4: Flowchart summarizing the protocol followed for the manual and automated analysis of... [/fig_ref].
As the automated algorithm parameters were fixed, only one operator was needed to perform the automated image analysis.
# Statistical methods
Pearson's R, Intraclass correlation coefficient (ICC) and Cronbach's Alpha Reliability Coefficient were used to test the correlation, consistency and reliability between manual and automated cell counts. We used Bland-Altman plots to assess the level of agreement between the gold standard (mean manual) cell count and the automated cell count. The paired samples t-test was used to test for a statistically significant difference between manual and automated cell counts.
# Results
## Duration of image analysis
Manual labeling of the cells on an image by a single operator to obtain a total cell count took an average of 3 min ± 2 min (Mean ± 1.96 Standard Deviation). In contrast, generating a labeled image and a total cell count with the automated algorithm took an average of 9 sec ± 2 sec. As all the Matlab script parameters were fixed, the script was only run once on each image. below illustrates a DARC image before and after undergoing manual and automated labeling: Example of a DARC image before and after undergoing manual and automated labeling. Image A represents a cropped DARC image before undergoing labeling. Image B represents the same DARC image after undergoing manual labeling using ImageJ® 'multi-point selections' tool, which marks and numbers each selected spot on the image. Image C represents the same DARC image after undergoing automated cell labeling using the novel Matlab® script. In Image C, structures identified as 'cells' have been labeled in green, whilst 'non-cellular' structures have been labeled in red. , manual labeling using the ImageJ® 'multi-point selections' tool enables the marking and numbering of each spot on the DARC image (Image Ib). Image Ic represents the same DARC image after undergoing automated cell labeling using the novel Matlab® script. In Image Ic, structures identified by novel Matlab® script as 'cells' have been labeled in green, whilst 'non-cellular' structures have been labeled in red. The script automatically calculates the total number of spots identified as cells.
## Examples of automated labeling
## As shown in
## Mean manual cell counts vs automated cell counts analysis
Pearson's correlation coefficient for the mean manual cell counts and the automated cell counts was 0.978, p < 0.001 (two-tailed significance). The R squared, as illustrated in [fig_ref] Figure 6: Correlation between the Mean Manual Cell Counts and the Automated Cell Counts... [/fig_ref] , was 0.956. The Intraclass correlation coefficient was 0.986 (95% CI 0.977-0.991, p < 0.001). Cronbach's alpha measure of consistency was 0.986. These results indicate a highly significant correlation and consistency between the mean manual cell counts and the automated cell counts.
In 36 (54.5%) of 66 DARC images, the automated cell count was higher than the mean manual cell count. The mean manual cell count for the 66 DARC images was 125.7 cells, whereas the mean automated cell count was 126.0 cells. The mean automated cell count was therefore 0.23% higher than the mean manual cell count. There was no significant difference between the mean manual and automated cell counts (p = 0.922, 95% CI −5.53 to 6.10).
A Bland-Altman 'percent difference' plot was constructed as recommended for method comparison studies in which agreement is to be assessed for a wide measurements range [bib_ref] Measuring agreement in method comparison studies, Bland [/bib_ref] [bib_ref] Application of the Bland-Altman plot for interpretation of method-comparison studies: a critical..., Dewitte [/bib_ref]. As shown in [fig_ref] Figure 7: Bland-Altman Plot of Percent Difference of Automated Cell Counts from Mean Manual... [/fig_ref] , there was strong agreement between the two methods, with 64 (97%) of 66 images cell counts lying within the 95% limits of agreement. The two images lying beyond 1.96 SD from the mean (normally referred to as 95% limits of agreement) are discussed in the next section. There was a tendency for the automated algorithm to underestimate the cell count in DARC images with high cell numbers (>200 cells). As shown in [fig_ref] Figure 7: Bland-Altman Plot of Percent Difference of Automated Cell Counts from Mean Manual... [/fig_ref] , the ratio of the difference of the automated cell counts from the mean manual cell counts was within 1.96 standard deviations of the mean difference for all >200 cell counts. This indicates that the extent of undercounting was minimal. A larger sample size of DARC images with >200 cell counts is needed to assess for a statistically significant difference in automated and mean manual cell counts.
Cell count differences beyond the 95% limits of agreement represents a DARC image in which the automated cell count was higher than the mean manual cell count. The image contained non-cellular fluorescent structure (pink arrow) which represents injection artifact, as well as a dark blob (blue arrow) which represents either a bubble (resulting from intravitreal injection) or a haemorrhage. The apoptosing cells in the image exhibited poor fluorescence, making manual cell identification challenging. This is reflected by the large inter-operator variation: The difference between the manual cell counts of operator 1 and operator 2, operator 1 and operator 3, and operator 2 and operator 3 were 32 cells, 41 cells and 9 cells respectively. The mean manual cell count was 29 cells (highest manual cell count = 53 cells), whereas the automated cell count was 73 cells. The higher automated cell count may be due to higher sensitivity of the automated method. On the other hand, the 'granular' nature of the retinal background may have resulted in false positive detection of cells.
In [fig_ref] Figure 9: The second DARC image with cell count differences beyond the 95% limits... [/fig_ref] , the cells were very poorly fluorescent, making accurate labeling by an operator difficult. The cell counts for operators 1, 2 and 3 were 21 cells, 0 cells and 7 cells respectively. Hence, while the cell count of operator 1 (=21 cells) was close to the automated cell count (=19 cells), operator 2 did not judge any of the structures to be fluorescent labeled cells. Arguably, such an image with poorly fluorescent cells should not be used to judge the extent of apoptosis, as the manual analysis results are variable and contentious. The higher cell count acquired by the automated technique may be due to higher sensitivity in detecting poorly fluorescent cells, or due to detection of structures which in reality would be not judged as cells because they are not strongly fluorescent. In the presence of such wide variation in manual labeling results, confirmation of the true presence of fluorescent cells is only possible with histological analysis.
Undercounted DARC images [fig_ref] Figure 1: Images A, B & C are examples of DARC images before undergoing... [/fig_ref] shows examples of a DARC image with >200 apoptosing RGCs in which cells were undercounted by the automated algorithm.
Green labeled spots on [fig_ref] Figure 1: Images A, B & C are examples of DARC images before undergoing... [/fig_ref] represent spots which were labeled and counted as 'cells' by the algorithm. Pink spots represent spots labeled as non-cellular structure and therefore not counted as cells by the algorithm. The white circle on the image shows examples of noise correctly identified as such by the algorithm The first DARC image with cell count differences beyond the 95% limits of agreement. (labeled in pink). On the other hand, the yellow arrows shows spots which should be labeled as cells, but the algorithm in this case has labeled as non-cellular structure (labeled in pink). This is due to the small size and low luminance of these spots. Another example is shown in [fig_ref] Figure 1: Images A, B & C are examples of DARC images before undergoing... [/fig_ref]. [fig_ref] Figure 1: Images A, B & C are examples of DARC images before undergoing... [/fig_ref] demonstrates how difficult it can be to distinguish background noise from apoptosing retinal cells (see in particular inside the white dashed circle). The yellow arrows point towards examples of pink spots which were likely to be labeled as cells by the operators. Once again, the small size and low fluorescence of these spots prevented labeling by the algorithm, but also served to prevent mislabeling of noise as cells. Another reason why RGC spots were undercounted by the algorithm was due to the shape of the spots, as shown in [fig_ref] Figure 1: Images A, B & C are examples of DARC images before undergoing... [/fig_ref].
The yellow circle contains spots labeled as noncellular structure (in pink) by the algorithm, which should have been labeled as RGC spots (in green). This is due the elongated non-circular shape of the spots on the image (resulting from image aberration), which prevents them being labeled as cells by the algorithm (see 'Methods' section).
## Analysis of individual manual operator cell counts vs automated cell counts
As shown in , there is a highly significant Pearson correlation (p < 0.001) between the manual cell counts measured by all three operators, as well as between the automated cell count and each operator's manual cell count.
Within all three manual cell counts and the automated cell counts, the Intraclass correlation coefficient is 0.994 (p < 0.001, 95% CI 0.991 -0.996). Cronbach's alpha measure of consistency was 0.986 for operator 1 and automated cell counts, 0.983 for operator 2 and automated cell counts, 0.980 for operator 3 and automated cell counts and 0.986 for the mean manual (all three operators) cell count and automated cell count. The automated cell count was within 1.96 standard deviations from the manual cell counts of operators 1, 2 and 3 in 61 (92.4%), 62 (93.9%) and 63 (95.5%) images analysed respectively. Overall, there was no significant difference between the three operators' cell counts (ANOVA, p = 0.319). [fig_ref] Table 2: Bland Altman test of agreement results between automated cell counts and each... [/fig_ref] below illustrates the strength of agreement between the automated and the manual counts, as well as the inter-operator agreement.
The inter-operator 95% limits of agreement were wider than those between the mean manual and automated cell counts, indicating wider inter-operator variability. The 66 DARC images contained an average of 126 cells each. Applying the average discrepancy (bias) of 5.6% between methods, this would result in automated cell count difference of 7 cells, which is not clinically important.
# Discussion
Cell counting has numerous applications in the field of biological imaging [bib_ref] Rapid automated cell quantification on HIV microfluidic devices, Alyassin [/bib_ref] [bib_ref] Enumeration of leukocyte infiltration in solid tumors by confocal laser scanning microscopy, Biggerstaff [/bib_ref] [bib_ref] A novel approach to automated cell counting for studying human corneal epithelial..., Bandekar [/bib_ref]. Although manual counting by an experienced cell biologist remains the gold standard, this process is time-consuming, monotonous, nonreproducible and subject to bias. The procedure proposed here counts cells in DARC images of variable quality to a level of confidence that is comparable to the gold-standard manual method. This technique has the advantages of being fast, accurate, reproducible and non-labour intensive. Fixing the algorithm parameters before image analysis enabled a non-biased objective quantification of cells that minimises cell count variability arising from inter-observer variability.
Various methods have been developed for automated retinal image analysis [bib_ref] Automated feature extraction in color retinal images by a model based approach, Li [/bib_ref] [bib_ref] Model-Based Illumination Correction in Retinal Images, Grisan [/bib_ref] [bib_ref] Retinal image analysis based on mixture models to detect hard exudates, Sanchez [/bib_ref] [bib_ref] Improved automatic detection and segmentation of cell nuclei in histopathology images, Al-Kofahi [/bib_ref]. Fluorescence images present specific challenges for the development of automated methods of cell counting, particularly the problem of background noise being mislabeled as cells [bib_ref] Automated neurite labeling and analysis in fluorescence microscopy images, Xiong [/bib_ref]. Distinguishing fluorescent particles from background noise and mild non-specific staining is therefore a crucial step in the development of algorithms enabling automated labeling and counting of fluorescent cells [bib_ref] Enumeration of leukocyte infiltration in solid tumors by confocal laser scanning microscopy, Biggerstaff [/bib_ref]. Increasing the image-thresholding level (without preprocessing) minimizes the impact of noise on cell-counts but results in more fluorescent cells being missed. The pre-processing stage of our algorithm minimises the impact of noise on local image statistics (such as local mean luminance and contrast) allowing us to use lower thresholds and so detect more cells without mislabeling noise.
Fluorescent cells may present as circular regions containing relatively uniform luminance structure, or may be more non-uniform in terms of shape and luminance [bib_ref] Automated neurite labeling and analysis in fluorescence microscopy images, Xiong [/bib_ref]. Non-uniform cell shape is a common problem in 2D histological sections of 3D specimens, in which cells may be partially present or damaged due to the sectioning process [bib_ref] Improved automatic detection and segmentation of cell nuclei in histopathology images, Al-Kofahi [/bib_ref]. Uneven luminance commonly occurs due to uneven fluorescent staining [bib_ref] Computerized image analysis as a tool to quantify infiltrating leukocytes: a comparison..., Johansson [/bib_ref] [bib_ref] Automated tool for the detection of cell nuclei in digital microscopic images:..., Byun [/bib_ref] , and the image acquisition process [bib_ref] Illumination correction of retinal images using Laplace interpolation, Leahy [/bib_ref]. The latter may also result in local contrast variability, also impeding the accuracy of automated analysis [bib_ref] Luminosity and contrast normalization in retinal images, Foracchia [/bib_ref]. In the context of fluorescence image analysis, this limits the utility of automated cell enumeration algorithms relying on cell-shape and luminance [bib_ref] Endothelial cell density in donor corneas: a comparison of automatic software programs..., Hirneiss [/bib_ref] [bib_ref] Arraycount, an algorithm for automatic cell counting in microwell arrays, Kachouie [/bib_ref]. To surpass these challenges, Byun et al. [bib_ref] Automated tool for the detection of cell nuclei in digital microscopic images:..., Byun [/bib_ref] used Laplacian-of-Gaussian filtering followed by searching for local maxima using cell size and distance between cells for the detection of cell nuclei in immunofluorescent retinal images acquired by confocal microscopy. In comparison to manual counting, their automated technique counted outer nuclei layer (ONL) nuclei with an average error of 3.67% (0-6.07%) and inner nuclear layer (INL) nuclei with an average error of 8.55% (0-13.76%). Accuracy of the technique was compromised in the INL due to variability in nuclei size and shape [bib_ref] Automated tool for the detection of cell nuclei in digital microscopic images:..., Byun [/bib_ref]. Large variability in cell size may indeed limit the accuracy of automated cell enumeration. Our algorithm utilizes a minimum cell size parameter rather than the mean or median cell size for categorization of cells after image pre-processing and thresholding. This has the advantage of maximizing detection of various size cells, (see [fig_ref] Figure 1: Images A, B & C are examples of DARC images before undergoing... [/fig_ref] in 'Methods' as an example) yet minimizing detection of noise and any other smaller background structures. This may be problematic in images containing small cells similar in size to background noise, which is why pre-processing is a crucial step for minimizing error in such images. It is possible to add a 'maximum' cell size cut-off to our algorithm, but this was not required for DARC images. Even in normal 'non-fluorescein' images, the presence of noise, fluctuating luminance and non-regular cell structure is a recognized barrier to automated retinal image analysis [bib_ref] Automated feature extraction in color retinal images by a model based approach, Li [/bib_ref] [bib_ref] Retinal image analysis based on mixture models to detect hard exudates, Sanchez [/bib_ref] [bib_ref] Retinal Images: Noise Segmentation, Akram [/bib_ref] [bib_ref] Observation Model Based Retinal Fundus Image Normalization and Enhancement, Yao [/bib_ref]. The algorithm presented here utilized image pre-processing, thresholding and blob analysis to enable detection of non-uniform and irregular fluorescent apoptosing retinal cells from noise, and other non-cellular structures (such as parts of blood vessels). We suggest that our algorithm may be more widely applicable to cell labeling problems in both retinal and other biological images with poor image quality and various shaped structures (e.g. elongated structures such blood vessels or nerves), but this is yet to be tested.
There are no studies we can find which have developed automated techniques for labeling and counting of single apoptosing retinal cells. This limits the comparability of our automated cell detection method to other methods. have utilized a cell penetrating fluorescent peptide probe (TcapQ) in an in vivo rat model of glaucoma to image single apoptosing RGCs by ex vivo fluorescence imaging [bib_ref] Single-cell imaging of retinal ganglion cell apoptosis with a cell-penetrating, activatable peptide..., Barnett [/bib_ref]. Counting of the apoptosing retinal cells was computer-assisted; the authors state that quantification of RGCs was performed by Scion image analysis software (Scion Corp), and that an experienced observer (who was blinded to the procedure) performed the counting process. The quantification of RGCs was therefore operator-dependent and not comparable to our automated algorithm. More recently, Qiu X et al. used a confocal scanning laser ophthalmoscope (CSLO) to enable in vivo fluorescence imaging of activated apoptosing RGCs displaying TcapQ probe activation [bib_ref] Single-cell resolution imaging of retinal ganglion cell apoptosis in vivo using a..., Qiu [/bib_ref]. Strong fluorescent cell-specific signals were observed with in vivo imaging in the RGC layer of eyes of living rats pre-treated with NMDA followed by TcapQ488. Image analysis was performed manually; cell signals were counted by a human operator using ImageJ software. The authors performed automated cell counting in a 'subset' of animals using "Find Maxima" in Ima-geJ to confirm manual counting. Noise tolerance level was pre-set, while edge and center (optic disc) maxima were excluded from the analysis field. Once again, an accurate and efficient automated method of cell quantification would be of great use in such studies. The evolving ability to image single apoptosing retinal cells in vivo and the potential of this technology to be used in humans in the future highlights the need for an accurate method of quantifying apoptosing RGCs that is not operator-dependent.
A weakness of the algorithm is that the automated cell counts tended to be lower than the mean manual cell counts for DARC images with RGC counts of >200 cells. Although these cell counts were within 1.96 SD from the mean difference as shown on [fig_ref] Figure 7: Bland-Altman Plot of Percent Difference of Automated Cell Counts from Mean Manual... [/fig_ref]. The two principal factors for RGC spots being mislabeled as noncellular structures were 1) Elongated non-circular RGC spots (due to image aberration), and 2) small and low luminance spots. For the former, the algorithm could be equipped with a function in which the operator adjusts the minimum aspect ratio for DARC images in which image acquisition has resulted in RGC spots appearing elongated. This has not been tested in this study. As for small and low luminance spots, reducing the cell size cut-off or lowering the luminance threshold may result in more noise being mislabeled as cells. Furthermore, pink spots which have been labeled as cells by operators in [fig_ref] Figure 1: Images A, B & C are examples of DARC images before undergoing... [/fig_ref] are not clear-cut apoptosing RGC spots, and may be argued to be noise rather than apoptosing cells. It is important to note that overall, the average automated cell count discrepancy was 5.6% higher than the mean manual cell count. The pattern of lower total cell counts obtained by the automated algorithm in images with >200 cells may be due to inadequately sized sample (14 out of 66 DARC images contained >200 cells as per mean manual count). A future comparative study of DARC images with >200 cells will shed more light on this. As DARC is a fairly new technology and still experimental, it is still not established whether such small low luminance spots are cellular or non-cellular. Arguably, only clear-cut RGC spots should be labeled and counted by manual or automated methods to minimize bias. As DARC imaging improves, visualization of small apoptosing RGC will become easier. Furthermore, if this technique succeeds in humans (Human clinical trials due to start soon), apoptosing RGC's should be larger and easier to identify.
A further weakness of our study is our assumption of the three operators' mean cell count as a gold-standard apoptosing cell count. In reality, even an experienced operator cannot be assumed to be able to label and count apoptosing retinal cells in DARC images with 100% accuracy, and this method is subjective. The operator needs to be able to distinguish positive-labeled cells, which may be difficult due to the small size of apoptosing retinal cells, the presence of non-specific staining, and the 'granular' nature of the retinal background especially apparent in poor quality images. To eliminate any subjective bias in the automated method, a pilot study was performed to determine and preset the optimum minimum cell size cut-off which could be applied to DARC images of variable quality. Furthermore, our comparison of total cell counts may not be the sharpest instrument for looking at relative strengths and weaknesses of operators and algorithms. It is possible to use a more "multi-local" analysis, looking at differences in correspondence of assigned labels within a locale to provide a more detailed comparison of manual and automated analysis techniques, and this is an approach we are currently evaluating.
# Conclusion
The novel Matlab software script described in this study enables fast, reproducible and non-operator dependent semi-automated labeling and counting of apoptosing retinal cells. The automated cell counts have significant correlation and consistency with the gold-standard mean manual cell counts, with no significant difference being detected. The method utilises fixed parameters, thus enabling analysis by relatively inexperienced operators. If image cropping and/or re-sizing is needed, it can be incorporated into the Matlab algorithm to make the image analysis process fully automated. This automated technique may prove to be a valuable method of quantifying apoptosing retinal cells, with particular relevance to translation in the clinic, where a Phase I clinical trial of DARC in glaucoma patients is due to start shortly.
## Availability of supporting data
The cell count results of the operators and the automated algorithm are available in the LabArchives repository, [Dataset DOI:10.6070/H4HM56D2 and 'https://mynotebook. labarchives.com/share/Bizrah/MjAuOHwzNzM4Ny8xNi9Ucm VlTm9kZS8zODcyMTExMDMyfDUyLjg'].
[fig] Figure 1: Images A, B & C are examples of DARC images before undergoing manual or automated cell labeling. [/fig]
[fig] Figure 2: The effects of pre-processing illustrated on a DARC image. (A) A raw DARC image. (B) The same image filtered with a Δ 2 G t filter (C) A pre-processed version of (A) (corrected for local variation in luminance structure). (D) A Laplacian filtered version of (C). Compare (D) to (B) and note presence of additional image structure in (D). [/fig]
[fig] Figure 4: Flowchart summarizing the protocol followed for the manual and automated analysis of the DARC images. [/fig]
[fig] Figure 3: An illustration of the DARC image shown inFigure 2dafter undergoing thresholding and a novel 'blob analysis' stage to classify blobs as cells (red), vessels (green) or noise (blue). [/fig]
[fig] Figure 6: Correlation between the Mean Manual Cell Counts and the Automated Cell Counts of the 66 DARC images. The continuous line is the best-fit line, and the adjacent dotted lines represent the 95% confidence intervals. [/fig]
[fig] Figure 7: Bland-Altman Plot of Percent Difference of Automated Cell Counts from Mean Manual Cell Counts. [/fig]
[fig] Figure 9: The second DARC image with cell count differences beyond the 95% limits of agreement. [/fig]
[table] Table 2: Bland Altman test of agreement results between automated cell counts and each individual operator's cell counts, as well as inter-operator agreement [/table]
|
Epigenetic Modifications in Pediatric Acute Lymphoblastic Leukemia
Aberrant epigenetic modifications are well-recognized drivers for oncogenesis. Pediatric acute lymphoblastic leukemia (ALL) is no exception and serves as a model toward the significant impact these heritable alterations can have in leukemogenesis. In this brief review, we will focus on the main aspects of epigenetics, which control leukemogenesis in pediatric ALL, mainly DNA methylation, histone modification, and microRNA alterations. As we continue to gain better understanding of the driving mechanisms for pediatric ALL at both diagnosis and relapse, therapeutic interventions directed toward these pathways and mechanisms can be harnessed and introduced into clinical trials for pediatric ALL.
# Introduction
Epigenetics is the study of biochemical modifications of chromatin [bib_ref] Epigenetics of acute lymphocytic leukemia, Garcia-Manero [/bib_ref] and have been implicated in the pathogenesis of cancer [bib_ref] Cancer epigenomics: beyond genomics, Sandoval [/bib_ref]. Epigenetic modifications to DNA are not secondary to changes to the nucleotide sequence itself but rather heritable changes affecting the activity of genes and their cellular expression. Examples include DNA methylation, histone modification, and alterations in non-coding microRNAs (miRNAs). Each of these mechanisms can alter how genes are expressed or silenced without modifying the DNA sequence. If these epigenetic modifications lead to silencing of tumor suppressor genes or activation of oncogenes then it is easy to conceptualize how leukemogenesis can occur.
Unlike chromosomal translocations or gene mutations, which are permanent, hypermethylation of gene promoters is a reversible event that could be targeted with therapeutic agents designed to alter aberrant epigenetic events. Incorporating epigenetic modifying agents into the treatment of pediatric ALL is an exciting approach that theoretically could have a significant impact in the treatment of this disease. This would be particularly true for relapse ALL, which is highly hypermethylated [bib_ref] Epigenetic reprogramming reverses the relapse-specific gene expression signature and restores chemosensitivity in..., Bhatla [/bib_ref] [bib_ref] DNA methylation patterns at relapse in adult acute lymphocytic leukemia, Garcia-Manero [/bib_ref] , and accounts for more deaths than any other pediatric disease and remains the fifth most common pediatric cancer overall [bib_ref] Childhood acute lymphoblastic leukaemia and relapse, Gaynon [/bib_ref].
In this brief review, we will focus on the three main areas of epigenetics, which have been implicated in the leukemogenesis of pediatric ALL; DNA hypermethylation, histone modification, and microRNA alterations. As we continue to gain better understanding of the driving mechanisms for pediatric ALL at both diagnosis and relapse, therapeutic interventions directed toward these pathways and mechanisms can be harnessed and introduced into clinical trials.
## Dna hypermethylation
Gains of DNA methylation tend to occur in the gene promoter region and are one of the most studied epigenetic abnormalities in oncogenesis [bib_ref] Methylation matters, Costello [/bib_ref]. The methylation occurs at cytosine (C) bases located 5 to guanosine (G) in a CpG dinucleotide and often in regions rich in repetitive CpGs known as CpG islands. The methyl groups are transferred to the CpG dinucleotide via DNA methyltransferases (Dnmt1, Dnmt3a, and Dnmt3b) and serve to transcriptionally silence genes downstream of the methylated promoter. When aberrant methylation occurs in a cancer cell, it typically results in hypermethylation of tumor suppressor genes. This can lead to disruption of key molecular pathways such as apoptosis, DNA repair pathways, cell cycle checkpoints, and cell differentiation as well as result in activation of metastasis/invasion pathways, drug resistance, and proliferation signal transduction [bib_ref] The role of DNA hypermethylation in the pathogenesis and prognosis of acute..., Roman-Gomez [/bib_ref].
Various groups have used DNA methylation studies to investigate the underlying epigenetic mechanisms in childhood leukemia. In a large cohort of 137 B-lineage and 30 T-lineage pediatric ALL cases, distinct DNA methylation signatures with significant concordant correlation of gene expression were found to be characteristic of various cytogenetic sub-types [bib_ref] Integrated genetic and epigenetic analysis of childhood acute lymphoblastic leukemia, Figueroa [/bib_ref]. In fact, a core set of epigenetically deregulated genes, common to all cases, was identified; suggesting their central role in leukemia initiation and maintenance. Likewise, DNA methylation interrogation of 69 pediatric B-ALL and 42 non-leukemic control samples revealed 325 genes hypermethylated and down regulated, and 45 genes hypomethylated and up-regulated across all the samples, irrespective of subtype [bib_ref] Epigenetic deregulation in pediatric acute lymphoblastic leukemia, Chatterton [/bib_ref]. Furthermore, gene ontology analysis of these epigenetically deregulated genes highlighted the role of genes involved in cell signaling, cellular development, cell survival, and apoptosis. Another study investigating 764 cases of newly diagnosed ALL and 27 cases of relapse, identified 9406 predominantly hypermethylated CpG sites, independent of cytogenetic background, with each cytogenetic subtype displaying a unique set of hyper-and hypomethylated sites. These differentially hypermethylated CpG sites were enriched for genes in the transcriptional regulatory network such as NANOG, OCT4, SOX2, and REST. These genes are known to be regulated by a polycomb group of proteins and have been identified as targets for hypermethylation in solid tumors [bib_ref] Mapping and analysis of chromatin state dynamics in nine human cell types, Ernst [/bib_ref] , leukemia [bib_ref] Prognostic DNA methylation patterns in cytogenetically normal acute myeloid leukemia are predefined..., Deneberg [/bib_ref] , and lymphoma [bib_ref] Array-based DNA methylation analysis in classical Hodgkin lymphoma reveals new insights into..., Ammerpohl [/bib_ref].
MLL-rearranged infant leukemia is one specific ALL subtype that has been shown to exhibit distinct promoter hypermethylation [bib_ref] Hypermethylation of specific microRNA genes in MLL-rearranged infant acute lymphoblastic leukemia: major..., Stumpel [/bib_ref] [bib_ref] Promoter hypermethylation in MLL-r infant acute lymphoblastic leukemia: biology and therapeutic targeting, Schafer [/bib_ref] [bib_ref] Suppression of the let-7b microRNA pathway by DNA hypermethylation in infant acute..., Nishi [/bib_ref] [bib_ref] Specific promoter methylation identifies different subgroups of MLLrearranged infant acute lymphoblastic leukemia,..., Stumpel [/bib_ref]. Stumpel and colleagues identified a distinct DNA methylation pattern dependent on the presence and type of MLL-fusion partner in a cohort of 57 newly diagnosed infant ALL patients [bib_ref] Specific promoter methylation identifies different subgroups of MLLrearranged infant acute lymphoblastic leukemia,..., Stumpel [/bib_ref]. In addition, the degree of hypermethylation appeared to correlate with a higher risk of relapse among infants carrying t (4;11) or t translocations. In another study of 5 MLL-rearranged infant ALL samples, genes known to be involved in oncogenesis and tumor progression (DAPK1, CCR6, HRK, LIFR, and FHIT ) were differentially methylated suggesting a role in the leukemogenesis of MLL-rearranged ALL [bib_ref] Promoter hypermethylation in MLL-r infant acute lymphoblastic leukemia: biology and therapeutic targeting, Schafer [/bib_ref]. As well, four of five genes that were hypermethylated and silenced were able to be re-expressed in vitro when exposed to DNMTi and regain their functional roles, thus pointing to the clinical potential epigenetic therapy may have in the treatment of infant leukemia.
Relapsed ALL is a highly aggressive disease marked predominantly by drug resistance [bib_ref] In vitro cellular drug resistance in children with relapsed/refractory acute lymphoblastic leukemia, Klumper [/bib_ref]. Efforts are currently being undertaken to identify the role of epigenetic mechanisms in driving relapse and chemoresistance [bib_ref] Integrated genomic analysis of relapsed childhood acute lymphoblastic leukemia reveals therapeutic strategies, Hogan [/bib_ref]. Genome-wide DNA methylation profiling performed on 33 matched relapse-diagnosis pairs demonstrated that the relapsed genome was distinctly more hypermethylated compared to matched samples at diagnosis [bib_ref] Integrated genomic analysis of relapsed childhood acute lymphoblastic leukemia reveals therapeutic strategies, Hogan [/bib_ref]. In this study, 1147 CpG sites corresponding to 905 genes were differentially hypermethylated at relapse. About a third of these genes exhibited concordant down-regulation of mRNA expression. Many of the known regulators of the Wnt pathway were hypermethylated and down regulated at relapse, including inhibitors of the β-catenin/TCF/LEF activity, as well as APC, WT1, cadherins (CDH1, CDH11), and SOX genes (SOX2, SOX8, SOX11, SOX21). Interestingly, PTPRO, a negative feedback inhibitor of the Wnt pathway that binds to Wnt and blocks its association with other receptors [bib_ref] Identification of ptpro as a novel target gene of Wnt signaling and..., Kim [/bib_ref] , was also hypermethylated and down regulated. This suggests that the Wnt pathway is over-activated at relapse and that aberrant DNA methylation may play a significant role in the activation of this pathway in relapsed ALL (3). Re-expression of these hypermethylated and down regulated genes was observed when leukemia cell lines were treated with decitabine. As well, enhanced chemosensitivity was observed when ALL cell lines and primary patient ALL samples were pretreated with decitabine followed by conventional cytotoxic chemotherapy (4).
In summary, DNA hypermethylation appears to play a significant role in the leukemogenesis of ALL and may be an important contributor toward relapse. As more studies interrogate the specific genes and or pathways influenced by hypermethylation in pediatric ALL, we will gain further insight toward strategies to therapeutically target these aberrant epigenetic changes and hopefully begin to make a greater impact in the treatment of this disease.
## Histone modifications
Histones are small basic proteins involved in the spatial organization of DNA within the nucleus. The chromatin environment influences the "on-off " transcriptional states of a gene depending on the post-translational modifications of the histone proteins [bib_ref] Translating the histone code, Jenuwein [/bib_ref]. Numerous covalent histone tail modifications, the most prominent being methylation and acetylation, can directly affect gene transcription [bib_ref] Combinatorial patterns of histone acetylations and methylations in the human genome, Wang [/bib_ref]. These modifications are highly specific for the particular amino acid position on the N-terminal tails of the histones. For example, H3K4me3, H3K9 acetylation, H3K14 acetylation, and H3K79me2 are associated with open chromatin structures and linked with transcriptional activation, while H3K9me3 and H3K27me3 are associated with closed chromatin, and hence transcriptional repression. These histone marks are regulated by the balance between competing enzymes such as the histone lysine methyltransferases (HKMTs) and histone demethylases (HKDMs), and the histone acetyltransferases (HATs) and histone deacetylases (HDACs) [bib_ref] Regulation of chromatin structure by site-specific histone H3 methyltransferases, Rea [/bib_ref]. Moreover, multiple histone modifications can be associated with critical regulatory elements of transcription such as enhancers, which can determine cell fate and differentiation [bib_ref] Combinatorial patterns of histone acetylations and methylations in the human genome, Wang [/bib_ref] [bib_ref] A unique chromatin signature uncovers early developmental enhancers in humans, Rada-Iglesias [/bib_ref].
Mutations in epigenetic modifying genes are common in hematologic malignancies, including ALL [bib_ref] Myeloid disorders working group of the International Cancer Genome, clinical and biological..., Papaemmanuil [/bib_ref] [bib_ref] mutations prevail in relapse cases of high hyperdiploid childhood acute lymphoblastic leukemia, Inthal [/bib_ref] [bib_ref] NSD2 links dimethylation of histone H3 at lysine 36 to oncogenic programming, Kuo [/bib_ref] [bib_ref] Global chromatin profiling reveals NSD2 mutations in pediatric acute lymphoblastic leukemia, Jaffe [/bib_ref] [bib_ref] CREBBP mutations in relapsed acute lymphoblastic leukaemia, Mullighan [/bib_ref] [bib_ref] Multiple mechanisms deregulate EZH2 and histone H3 lysine 27 epigenetic changes in..., Khan [/bib_ref]. These mutations can result in a gain or loss of function of key genes known to regulate histone marks. Jaffe and colleagues, in pediatric ALL cell lines, have used global chromatin profiling, a tandem mass spectrometry strategy, to measure levels of histone modifications on bulk chromatin [bib_ref] Global chromatin profiling reveals NSD2 mutations in pediatric acute lymphoblastic leukemia, Jaffe [/bib_ref]. In this work, a novel cluster of cell lines with a specific epigenetic signature was identified, characterized by increased dimethylation of histone H3 at lysine 36 (H3K36me2) and decreased unmodified H3K36. Approximately half of the cell lines in this cluster harbored the t (4;14) translocation, which is known to induce overexpression of NSD2 [bib_ref] Regulation of chromatin structure by site-specific histone H3 methyltransferases, Rea [/bib_ref] [bib_ref] Detection of t(4;14)(p16.3;q32) chromosomal translocation in multiple myeloma by reverse transcription-polymerase chain..., Malgeri [/bib_ref] [bib_ref] The t(4;14) translocation in myeloma dysregulates both FGFR3 and a novel gene,..., Chesi [/bib_ref]. NSD2 is a member of the HKMTs that catalyze the conversion of unmodified H3K36 to mono-and dimethylated forms [bib_ref] NSD2 links dimethylation of histone H3 at lysine 36 to oncogenic programming, Kuo [/bib_ref]. Upon targeted sequencing in an extensive patient sample set, NSD2 mutations were found to be enriched in ETV6-RUNX1 and TCF3-PBX1 sub-types of pediatric B-ALL, while no mutations were identified in 30 adult ALL samples. These were gainof-function mutations and their overexpression led to a global increase in H3K36me2, with concomitant decrease in H3K27me3. Similar results were reported by others [bib_ref] Point mutation E1099K in MMSET/NSD2 enhances its methyltranferase activity and leads to..., Oyer [/bib_ref] , showing these mutations affect expression of a number of genes involved in normal lymphoid development.
Accumulating evidence suggests that histone modification is an important aspect of MLL-fusion mediated transformation and leukemogenesis [bib_ref] H3K79 methylation profiles define murine and human MLL-AF4 leukemias, Krivtsov [/bib_ref] [bib_ref] Targeting epigenetic programs in MLL-rearranged leukemias, Bernt [/bib_ref]. It has been shown that wild type MLL SET domain is a methyltransferase, modifying histone H3 on lysine 4 (H3K4), and positively regulating gene expression of multiple Hox genes [bib_ref] SET domain methyltransferase activity to Hox gene promoters, Milne [/bib_ref]. In addition, MLL mediated transcriptional regulation involves recruitment of HAT, such as CBP [bib_ref] MLL and CREB bind cooperatively to the nuclear coactivator CREB-binding protein, Ernst [/bib_ref] and MOF [bib_ref] Physical association and coordinate function of the H3 K4 methyltransferase MLL1 and..., Dou [/bib_ref]. Furthermore, DOT1L, a histone methyltransferase that methylates lysine 79 on histone H3 (H3K79), has been associated with multiple MLL-fusion partners such as AF9, AF10, AF17, and ENL [bib_ref] hDOT1L links histone methylation to leukemogenesis, Okada [/bib_ref] [bib_ref] A role for the MLL fusion partner ENL in transcriptional elongation and..., Mueller [/bib_ref] [bib_ref] Dot1a-AF9 complex mediates histone H3 Lys-79 hypermethylation and repression of ENaCα in..., Zhang [/bib_ref] , and has emerged as an attractive therapeutic target [bib_ref] Targeting epigenetic programs in MLL-rearranged leukemias, Bernt [/bib_ref]. Several groups have used small molecule inhibitors to demonstrate the feasibility of pharmacological inhibition of DOT1L enzymatic activity in preclinical models of MLL-rearranged leukemia [bib_ref] Selective killing of mixed lineage leukemia cells by a potent small-molecule DOT1L..., Daigle [/bib_ref] [bib_ref] Abrogation of MLL-AF10 and CALM-AF10-mediated transformation through genetic inactivation or pharmacological inhibition..., Chen [/bib_ref] [bib_ref] Potent inhibition of DOT1L as treatment of MLL-fusion leukemia, Daigle [/bib_ref] and are now under clinical investigation in a phase I study for adults with advanced hematologic malignancies, including acute leukemia with rearrangement of the MLL gene (NCT01684150). One inhibitor in particular, EPZ-5676, has shown potent activity in its ability to selectively inhibit the DOT1L histone methyltransferase, resulting in cell death of acute leukemia cell lines harboring MLL translocations as well as complete tumor regression in a rat xenograft model of MLL-rearranged leukemia following continuous iv infusion of EPZ-5676 [bib_ref] Potent inhibition of DOT1L as treatment of MLL-fusion leukemia, Daigle [/bib_ref].
In order to identify novel mutations in relapsed ALL, Mullighan and colleagues performed targeted resequencing of 300 genes in 23 matched relapse-diagnosis B-ALL pairs [bib_ref] CREBBP mutations in relapsed acute lymphoblastic leukaemia, Mullighan [/bib_ref]. The authors identified novel mutations in CREBBP, a gene encoding the transcriptional coactivator CREB binding protein with HAT activity. The overall frequencies of these sequence and/or deletional mutations were 18.3% in relapse cases [bib_ref] CREBBP mutations in relapsed acute lymphoblastic leukaemia, Mullighan [/bib_ref]. However, particularly high incidences of somatic CREBBP alterations (63%) were found in the high hyperdiploidy relapse cases. Of note, the majority of these mutations occurred in the HAT domain [bib_ref] mutations prevail in relapse cases of high hyperdiploid childhood acute lymphoblastic leukemia, Inthal [/bib_ref]. Although less common, mutations in other important epigenetic regulators were also seen such as NCoR1 (Nuclear corepressor complex), EP300 (a paralog of CREBBP), EZH2 (histone methyltransferase gene), and CTCF (zinc finger protein involved in histone modifications) [bib_ref] CREBBP mutations in relapsed acute lymphoblastic leukaemia, Mullighan [/bib_ref]. Additionally, transcriptome sequencing has identified relapse-specific mutations in CBX3 (encoding heterochromatin protein), PRMT2 (gene encoding protein arginine methyltransferase 2), and MIER3 (involved in chromatin binding); providing further evidence of aberrant epigenetic mechanisms that play a role at relapse [bib_ref] Relapse-specific mutations in NT5C2 in childhood acute lymphoblastic leukemia, Meyer [/bib_ref].
Epigenetic alterations are not only restricted to B-ALL, but are a notable feature of T-ALL, particularly the aggressive subtype early T-cell precursor (ETP) ALL. Whole genome sequencing of 12 cases of ETP ALL identified mutations in genes encoding components of the polycomb repressor complex 2 (PRC2), including deletions and sequence mutations of EZH2, SUZ12, and EED [bib_ref] The genetic basis of early T-cell precursor acute lymphoblastic leukaemia, Zhang [/bib_ref]. Loss of function mutations and deletions of EZH2 and SUZ12 genes have also been found in T-ALL, where authors implicate the tumor suppressor role of the PRC2 complex [bib_ref] Genetic inactivation of the polycomb repressive complex 2 in T cell acute..., Ntziachristos [/bib_ref].
In addition to the discovery of somatic mutations in epigenetic machinery in ALL, mRNA expression of HDACs has been shown to be dysregulated. Higher mRNA expression of HDAC7 and HDAC9 in a study of 94 childhood ALL cases was shown to correlate with poor prognosis [bib_ref] Differential expression of HDAC3, HDAC7 and HDAC9 is associated with prognosis and..., Moreno [/bib_ref]. Similarly, another group identified the correlation of HDAC4 overexpression with prednisone poor response, T-ALL phenotype, and a high initial WBC [bib_ref] The expression of histone deacetylase 4 is associated with prednisone poor-response in..., Gruhn [/bib_ref]. Given the compelling evidence of HDAC's involvement in tumor development and progression, inhibitors of HDACs have emerged as an attractive therapeutic option in hematologic malignancies [bib_ref] Histone deacetylase inhibitors in cancer therapy, Lane [/bib_ref]. Through a connectivity map search (52) for agents, which could potentially reverse the characteristic gene expression signature specific for relapse ALL [bib_ref] Biologic pathways associated with relapse in childhood acute lymphoblastic leukemia: a Children's..., Bhojwani [/bib_ref] and potentially endow chemosensitivity, vorinostat (HDACi) was identified as the most promising candidate (4). In fact, vorinostat not only modulated the gene expression signature characteristic of relapse in ALL cell lines and patient samples, but showed a synergistic effect when given sequentially with chemotherapy (4). The fact that vorinostat showed significant alteration of gene expression correlating with histone modifications, indicates that the perturbation of histone marks may have a key role in aberrant gene regulation at relapse. Bachmann and colleagues have reported glucocorticoid resistance associated with epigenetic silencing of the BIM gene in pediatric ALL and showed synergistic effect of vorinostat with dexamethasone in both in vitro and in vivo models [bib_ref] Epigenetic silencing of BIM in glucocorticoid poor-responsive pediatric acute lymphoblastic leukemia, and..., Bachmann [/bib_ref]. The potential importance of these changes is highlighted by the promising activity of several other drugs from the same class that target epigenetic alterations [bib_ref] The epigenomics of cancer, Jones [/bib_ref].
In summary, similar to the influence DNA hypermethylation has in pediatric ALL leukemogenesis, maintenance, and relapse, aberrant epigenetic changes involving histones have been associated with disease progression and relapse in ALL. With growing experience using HDACi in hematologic malignancies, including pediatric trials (NCT01483690, NCT01321346), the impact of these agents will become clearer as well as their role in future relapse and upfront ALL studies.
## Microrna alterations
MicroRNAs are a class of small endogenous single stranded non-coding ribonucleic acids (RNA) composed of roughly 22 nucleotides that are primarily involved in post-transcriptional gene regulation. miRNAs play a critical regulatory role in targeting mRNAs for cleavage or translational repression, with greater than 1,000 miRNAs currently identified in the human genome. MicroRNA genes are preferentially localized to CpG islands, which leads to the plausible mechanism that they can be controlled through aberrant epigenetic regulation (e.g., hypermethylation, histone modification) [bib_ref] Methylation of human microRNA genes in normal and neoplastic cells, Weber [/bib_ref].
Altered expression of miRNAs has been implicated in leukemogenesis and appears to have the ability to influence critical growth regulatory pathways in ALL [bib_ref] Differential microRNA expression in childhood B-cell precursor acute lymphoblastic leukemia, Ju [/bib_ref] [bib_ref] Identification of new microRNA genes and aberrant microRNA profiles in childhood acute..., Schotte [/bib_ref] [bib_ref] MicroRNA expression signatures accurately discriminate acute lymphoblastic leukemia from acute myeloid leukemia, Mi [/bib_ref] [bib_ref] miRNA expression profiles in chronic lymphocytic and acute lymphocytic leukemia, Zanette [/bib_ref]. An example of the functional impact miRNA can have in B-cell ALL was reported with the restoration of miR-196b expression, which led to significant down-regulation of c-myc and its effector genes fhTERT, Bcl-2, and AATF, suggesting a tumor suppressor function role for miR-196b [bib_ref] Potential tumor suppressive function of miR-196b in B-cell lineage acute lymphoblastic leukemia, Bhatia [/bib_ref]. Some specific miRNAs that have been implicated in pediatric ALL include miRNA (miR) miR-34, miR-128, miR-142, and miR-181, all reported to be over expressed [bib_ref] Differential microRNA expression in childhood B-cell precursor acute lymphoblastic leukemia, Ju [/bib_ref] [bib_ref] Differential miRNA expression in childhood acute lymphoblastic leukemia and association with clinical..., De Oliveira [/bib_ref] [bib_ref] MicroRNA patterns associated with clinical prognostic parameters and CNS relapse prediction in..., Zhang [/bib_ref] and miR-100 and miR-196b, both under expressed [bib_ref] Identification of new microRNA genes and aberrant microRNA profiles in childhood acute..., Schotte [/bib_ref] [bib_ref] Differential miRNA expression in childhood acute lymphoblastic leukemia and association with clinical..., De Oliveira [/bib_ref]. Schotte and colleagues investigated 397 miRNAs using qRT-PCR in 81 pediatric ALL cases in comparison to 17 normal CD34 + stem cell controls [bib_ref] MicroRNA characterize genetic diversity and drug resistance in pediatric acute lymphoblastic leukemia, Schotte [/bib_ref]. Unique miRNA signatures were identified for various ALL sub-types including ETV6-RUNX1, MLL-rearranged, T-ALL, hyperdiploidy, and E2A-PBX1. Overall, expression of miR-143 and miR-140 were found to be 70-and 140-fold lower in the B-ALL samples compared to controls (p FDR = 0.0007 and p FDR = 0.001, respectively). Hyperdiploid samples showed a clustering of high expression of miR-98, miR-222, miR-223, and miR-511 and the ETV6-RUNX1 cases had a 5-to 1700-fold increase expression in miR-99a, miR-100, miR-125b, and miR-383 compared to controls (p FDR < 0.001). Together these findings lend support for epigenetic alterations involving miRNAs in the leukemogenesis of some of the more common variants of pediatric ALL.
Aberrant miRNA expression has been implicated in leukemia drug resistance and lower event-free survival (EFS). Schotte and colleagues identified a lower expression of miR-454 (1.9fold lower) in leukemia blasts with l-asparaginase resistance (p FDR = 0.017) and patient samples resistant to vincristine and www.frontiersin.org daunorubicin were found to have over expression of miR-99a, miR-100, and miR-125b (14-to 25-fold) (p FDR ≤ 0.002 and p FDR < 0.05, respectively) [bib_ref] MicroRNA characterize genetic diversity and drug resistance in pediatric acute lymphoblastic leukemia, Schotte [/bib_ref]. In terms of EFS, six miR-NAs (miR-33, -215, -369-5p, -496, -518d, and -599) were associated with worse survival (HR 1.3-1.52, 95% CI 1.01-2.04; 0.003 ≤ p ≤ 0.046) and another eight (miR-10a, -134, -214, -484, -572, -580, -624, and -627) with greater EFS (HR 0.59-0.82, 95% CI 0.41-0.99, 0.004 ≤ p ≤ 0.045) [bib_ref] MicroRNA characterize genetic diversity and drug resistance in pediatric acute lymphoblastic leukemia, Schotte [/bib_ref]. The authors concluded that the miRNAs associated with a more favorable outcome likely had tumor suppressor activity through their signaling of apoptosis (miR-10a), inhibition of proliferation (miR-10a and miR-214), and oncogene SOX2 down-regulation (miR-134).
In a report of 18 matched-pair diagnosis and relapse (n = 8) or diagnosis and remission (n = 10) pediatric ALL samples, data was summarized for the most differentially expressed miRNAs [bib_ref] A set of miRNAs that involve in the pathways of drug resistance..., Han [/bib_ref]. Down-regulation of miR-23a and miR-223 was observed at time of relapse compared to remission whereas miR130b, -181, and -708 were over expressed at relapse. Specifically, the expression of miR-708 was greater in relapse samples and lower in remission samples when compared to diagnosis whereas miR-223 was up-regulated in remission samples compared to diagnosis and confirmed with qRT-PCR. These two miRNAs at diagnosis along with miR-27a were shown to correlate significantly with 3-year relapse-free survival (p = 0.0483, 0.0079, and 0.0024, respectively) and thus could potentially be used as prognostic biomarkers for newly diagnosed patients. The functional impact these miRNAs had on gene expression was described as well with targets identified for BMI1, transcription factor necessary for hematopoietic stem cell and leukemia stem cell self-renewal, in miR-27a and miR-128b as well as E2F1, master cell cycle regulator, a target of miR-223. The variations in miRNA expression that exist between diagnostic, remission, and relapse samples identified by Han and colleagues suggest that critical epigenetic mechanisms exist through these non-coding miRNAs that may assist in driving leukemogenesis and disease recurrence.
In an analysis of 353 diagnostic bone marrow samples from patients with ALL (<15 years of age, n = 179), 65% had at least one of 13 previously identified miRNAs hypermethylated [bib_ref] Epigenetic regulation of miRNA genes in acute leukemia, Agirre [/bib_ref]. These 13 miRNAs were found to be regulated by methylation and histone modification and associated with a closed chromatin conformation of 11 CpG islands close to where the 13 miRNAs resided. The hypermethylation was associated with miRNA under expression but could be reversed with decitabine.
In summary, aberrant miRNA expression, particularly secondary to methylation, is a common finding in ALL. These data support that epigenetic modifications of specific miRNAs are associated with chemotherapy resistance and clinical outcomes. As these modifications can be secondary to DNA hypermethylation [bib_ref] MicroRNA characterize genetic diversity and drug resistance in pediatric acute lymphoblastic leukemia, Schotte [/bib_ref] [bib_ref] microRNA-199a-3p, DNMT3A, and aberrant DNA methylation in testicular cancer, Chen [/bib_ref] [bib_ref] Epigenetic silencing of miRNA-9 is associated with HES1 oncogenic activity and poor..., Fiaschetti [/bib_ref] [bib_ref] Molecular mechanisms of regulation and action of microRNA-199a in testicular germ cell..., Gu [/bib_ref] [bib_ref] Epigenetic regulation of microRNAs in acute lymphoblastic leukemia, Roman-Gomez [/bib_ref] , exposure to agents such as DNMTi could reverse the aberrant expression, normalize miRNA levels, and ultimately lead to improved clinical outcomes.
## Clinical trials investigating epigenetic modifying therapies in pediatric all
The majority of clinical experience using epigenetic modifying agents in the treatment of acute leukemia has been in adults [bib_ref] A phase II trial of decitabine and vorinostat in combination with chemotherapy..., Burke [/bib_ref] [bib_ref] Phase I study of 5-aza-2'-deoxycitidine, alone or in combination with hyper-CVAD, in..., Garcia-Manero [/bib_ref]. The Children's Oncology Group (COG) piloted a phase I study investigating decitabine (10 mg/m 2 /day × 5 days/week × 2 weeks) in children with relapsed/refractory acute leukemia that closed prematurely due to low patient accrual (NCT00042796, unpublished). No maximum tolerated dose (MTD) was identified and 5/15 patients reported grade 3/4 cytopenias (anemia, thrombocytopenia, and leukopenia) that were possibly related to the study drug.
Similar to the DNMTi, HDACi (e.g., vorinostat, panobinostat) have been studied in the treatment of acute leukemia, primarily as single agents and almost exclusively in adults [bib_ref] A phase I study of intravenous LBH589, a novel cinnamic hydroxamic acid..., Giles [/bib_ref] [bib_ref] Phase 1 study of the histone deacetylase inhibitor vorinostat (suberoylanilide hydroxamic acid..., Garcia-Manero [/bib_ref]. The COG completed a phase I study investigating vorinostat in combination with 13 cis-retinoic acid in children with refractory/recurrent solid tumors and vorinostat alone for patients with refractory leukemia [bib_ref] Pediatric phase I trial and pharmacokinetic study of vorinostat: a Children's Oncology..., Fouladi [/bib_ref]. Six patients with refractory leukemia were enrolled with 2 DLTs reported at the solid tumor MTD (230 mg/m 2 /day) including an elevated AST (n = 1), hyperbilirubinemia (n = 1), elevated GGT (n = 1), and hypokalemia (n = 1). As the solid tumor MTD for vorinostat did not appear tolerable for patients with hematologic malignancies, there was no further dose finding attempt in this study. Currently, there is a phase I study of panobinostat in children with refractory hematologic malignancies open through the therapeutic advances in childhood leukemia and lymphoma (TACL) Consortium (NCT01321346).
The first study incorporating a DNMTi and HDACi followed by chemotherapy for children and adults with relapsed/refractory ALL was recently completed [bib_ref] A phase II trial of decitabine and vorinostat in combination with chemotherapy..., Burke [/bib_ref]. In this phase II trial, decitabine (15 mg/m 2 /day) and vorinostat (230 mg/m 2 divided BID) were given over four consecutive days prior to re-induction chemotherapy (vincristine, prednisone, PEG-asparaginase, doxorubicin) (NCT00882206) [bib_ref] A phase II trial of decitabine and vorinostat in combination with chemotherapy..., Burke [/bib_ref]. Thirteen eligible patients enrolled with a median age of 16 (range, 3-54) years. There was a single toxic death occurring on study attributed to the chemotherapy regimen, which included a grade five hemorrhage/bleeding (n = 1). A second patient experiencing grade five hypoxia/acute respiratory distress died on day 4 of study attributed to disease progression (n = 1). There were an additional 14 grade 3/4 serious adverse events, which were at least possibly attributed to decitabine or vorinostat, the most common being fever with neutropenia (n = 2) and infection (blood) with neutropenia (n = 5). Results of the eight patients evaluable for response, identified a CR rate of 50% (n = 4/8) (95% CI 15.7-84.3%) and an overall response rate (CR + PR) of 75% (n = 6/8) (95% CI 34.9-96.8%). As well, minimal residual disease (MRD) negativity by flow cytometry was observed in 4/8 patients (50%, CI: 15.7-84.3%). Five of the eight patients who completed the study proceeded to allogeneic hematopoietic cell transplantation (four in second CR and one in third CR). Three patients succumbed to transplant related deaths without evidence of leukemia while the remaining two patients remain alive with no evidence of disease. Based on the results of this study, a pediatric trial for relapse/refractory ALL combining decitabine and vorinostat with re-induction chemotherapy is currently open through the TACL Consortium (NCT01483690; R21CA161688-01).
## Summary
Underlying epigenetic alterations in pediatric ALL are common events, which appear to be more common at relapse than Frontiers in Pediatrics | Pediatric Oncology diagnosis. Thus children with relapse ALL may be an ideal population for clinical trials incorporating epigenetic modifying agents aimed at reversing these aberrant signatures. Whether such trials will lead to improved clinical outcomes has yet to be determined but early findings in studies incorporating these agents have been encouraging.
In conclusion, leukemogenesis of pediatric ALL is heavily influenced by epigenetics, particularly DNA hypermethylation, histone modification, and alterations in miRNA expression. Epigenetic modifying agents such as DNMTi and HDACi as well as newer therapies (e.g., histone methyltransferase inhibitors) are now being incorporated into early phase clinical trials for relapse leukemia. As more trials for children with relapse ALL, incorporating epigenetic therapies into standard and/or novel salvage regimens, are developed and completed, we will have a better understanding as to which patients might benefit the most using this approach and ultimately where these agents may be best served in treating pediatric ALL.
May 2014 | Volume 2 | Article 42 | 1
May 2014 | Volume 2 | Article 42 | 2
May 2014 | Volume 2 | Article 42 | 3
May 2014 | Volume 2 | Article 42 | 4
May 2014 | Volume 2 | Article 42 | 5
May 2014 | Volume 2 | Article 42 | 7
Conflict of Interest Statement:The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.Received |
Financial Literacy and Physician Wellness: Can a Financial Curriculum Improve an Obstetrician/Gynecologist Resident and Fellow's Well-Being?
[bib_ref] Intervention for physician burnout: a systematic review, Wiederhold [/bib_ref] [bib_ref] Implementation of a comprehensive curriculum in personal finance for medical fellows, Bar-Or [/bib_ref] [bib_ref] A longitudinal study of financial difficulties and mental health in a national..., Richardson [/bib_ref] [bib_ref] Prevalence of depression and depressive symptoms among resident physicians: a systematic review..., Mata [/bib_ref] [bib_ref] Money matters: a resident curriculum for financial management, Mizell [/bib_ref] [bib_ref] The relationship of financial stress to overall stress and satisfaction, Bailey [/bib_ref] [bib_ref] Understanding the impact of employer-provided financial education on workplace satisfaction, Tahira [/bib_ref]
# Materials and methods
This was a prospective cohort study that measured obstetrician/gynecologist (OB/GYN) resident and fellow's sense of well-being and financial stress through validated surveys before and after implementation of a personal finance curriculum. All 46 OB/GYN residents and fellows at a single institution and currently enrolled in the graduate medical education (GME) program were invited to participate in the personal finance curriculum.
Our intervention consisted of a five-part personal finance curriculum (45 minutes each) that was performed by the primary author throughout the 2019 to 2020 academic year. The primary author has spent the last 5 years in both a formal and informal capacity for advising physicians on their finances. The covered topics were chosen based on the primary authors experience, as well as what trainees have previously expressed interest in learning. [bib_ref] Personal finance education for residents: a qualitative study of resident perspectives, Shappell [/bib_ref] These included the following: (1) personal finance education and its relationship to well-being, (2) an overview of financial terms and principles, (3) budgeting, (4) debt planning, and (5) investing and giving (►Fig. 1). The first three topics were given in a live round-table discussion format. A formal presentation was used as the background with free-flowing discussion throughout. The final two presentations were moved to a video-conferencing application (due to the novel coronavirus disease 2019 [COVID-19] pandemic) and involved a lecture format followed by question and answer (Q&A) session. Following each lecture, the presentation was provided in a PowerPoint PDF document with a modified transcript of what was discussed. This was available to all residents and fellows to educate those who were unable to participate in the live sessions and review at their own discretion.
All OB/GYN trainees who were able and willing to participate in the study were asked to take a pre-and postintervention survey. The primary outcomes were trainee financial health and stress as measured by both the previously validated the 9-Item Expanded Well-Being Index (E-WBI) and the 13question Financial Stress Scale-College Version (FSS-CV) survey. [bib_ref] Ability of a 9-item well-being index to identify distress and stratify quality..., Dyrbye [/bib_ref] [bib_ref] The development, evaluation, and validation of a financial stress scale for undergraduate..., Northern [/bib_ref] Two primary outcomes were included as well-being can be multifactorial and not directly related to financial stress. [bib_ref] Ability of a 9-item well-being index to identify distress and stratify quality..., Dyrbye [/bib_ref] Secondary outcomes included resident and fellow attitudes and comfort levels on the role of financial education in a medical curriculum. These attitudes were recorded based on a 7-point Likert's scale. Questions were chosen based on previously described physician attitudes toward financial education. [bib_ref] Implementation of a comprehensive curriculum in personal finance for medical fellows, Bar-Or [/bib_ref] [bib_ref] Money matters: a resident curriculum for financial management, Mizell [/bib_ref] [bib_ref] Personal finance education for residents: a qualitative study of resident perspectives, Shappell [/bib_ref] Those who declined to take the surveys were excluded from analysis but were still be able to actively participate in all facets of the personal finance curriculum.
Our sample size was limited by the set number of OB/GYN trainees, 46 total. However, with a mean FSS-CV score of 22.54 AE 7.39, we calculated the ability to detect a 20% improvement in FSS-CV score with 80% power and 0.05 α if 43 residents/fellows participate. Given this would involve a difficult to achieve 95% participation rate, we acknowledged our limitations beforehand in failing to reject the null hypothesis in the event; no statistical significance was achieved.
All data were captured directly into the study database developed with the Research Electronic Data Capture (RED-Cap). 14 Exposure groups were compared using Chi-square, Fisher's exact, and Wilcoxon's tests as appropriate. The statistical analysis was performed using SAS 9.4 (Cary, NC) with significance level at 0.05. Institutional review board review was provided and obtained given all residents could opt out of study participation by not completing the surveys.
# Results
Out of the eligible resident and fellows, 35 agreed to participate in the study by taking the initial survey. The representation included a majority of those who identify as predominantly female which is representative of the GME program's demographics. They also were well distributed between lower level residents, upper level residents, and fellows (►Table 1). Of those who agreed to participate in the There was a significant improvement in the median expanded well-being index score after exposure to the personal finance curriculum (2 vs. 1, p 0.05). There was no significant difference in the median financial stress scale survey before and after the personal finance curriculum (22 vs. 20, p ¼ 0.06; ►Table 3). Residents and fellows expressed a better understanding of financial planning topics (4 vs. 5, p ¼ 0.007) and the financial industry (3 vs. 5, p 0.001). After completing the curriculum, there was a significantly stronger agreement that financial literacy improves their sense of well-being (4 vs. 5, p ¼ 0.018; ►Table 4).
# Discussion
In this prospective study, OB/GYN residents and fellows who participated in a personal finance curriculum showed improved well-being as scored by the E-WBI. In addition, the trainees agreed that financial education improved their sense of well-being.
Financial literacy programs are a common occurrence across multiple groups and industries, including those within health care. [bib_ref] Understanding the impact of employer-provided financial education on workplace satisfaction, Tahira [/bib_ref] [bib_ref] Financial literacy around the world: an overview, Lusardi [/bib_ref] [bib_ref] An elective course in personal finance for health care professionals, Chui [/bib_ref] [bib_ref] Developing a financial literacy program with survivors of intimate partner violence: the..., Stylianou [/bib_ref] Although measured outcomes and goals of programs differ, common findings include the desire for financial literacy and an appreciation for the obtained knowledge. Programs that have specifically assessed physician attitudes toward a personal finance curriculum have all consistently showed residents' and fellows' desire to improve their financial literacy. [bib_ref] Implementation of a comprehensive curriculum in personal finance for medical fellows, Bar-Or [/bib_ref] [bib_ref] Money matters: a resident curriculum for financial management, Mizell [/bib_ref] [bib_ref] Personal finance education for residents: a qualitative study of resident perspectives, Shappell [/bib_ref] [bib_ref] Financial education for radiology residents: significant improvement in measured financial literacy after..., Boehnke [/bib_ref] As studies have shown the appetite for financial literacy is apparent, ours is the first to directly measure its association with physician well-being and financial stress.
Physician well-being is a complex concept that involves multiple inputs, including work-life balance, quality of life, resilience, mindfulness, coping strategies, and mood. [bib_ref] Assessment of physician wellbeing, part two: beyond burnout, Lall [/bib_ref] Financial well-being is not included in this list; however, financial stress and strain play a direct roll on our selfefficacy beliefs and mental health. [bib_ref] Health effects of indebtedness: a systematic review, Turunen [/bib_ref] With improved financial literacy, it is logical that financial stress will decrease 9 and in turn improve most of the components of physician wellbeing. Burnout and work-related stress have been described as a "mismatch between the perception of what is expected and the reality of what is possible." [bib_ref] Burnout or mismatch between expectations and reality?, Hajela [/bib_ref] We would concur that all stress can be decreased when expectations and reality become more in line. When a resident or fellow goes through a personal finance curriculum, they are exposed to the reality of the financial world. This in turn creates more realistic financial expectations and in turn decreases financial stress. Although we did not show a significant difference in the course decreasing financial stress, this was likely due to being underpowered and not due to the lack of association.
# Limitations
Although our study is the first to demonstrate an association with financial education and physician well-being, it is not without limitations. The first involves the unmasking of the participants. The residents and fellows in our GME program were notified beforehand that they were participating in a study assessing the potential benefits of a personal finance curriculum. This knowledge could have led to the Hawthorne effect. However, the participants were not aware of the specific outcomes being measured or the scoring system, limiting result bias. In addition, those who saw no benefit to the course may have declined to participate in the postintervention survey biasing our intervention results away from the null hypothesis. Another notable limitation relates to the generalizability. This education initiative was given at a single site, to a single specialty, and specific to residents and fellows. Both medical students and faculty have a different set of financial needs and our personal finance curriculum may not show the same association in these groups. Plus pay differs substantially between subspecialties and geographic regions which could have a bearing on how applicable the curriculum may be to all GME programs.In addition, the external validity of this study is not apparent. This course consisted of five parts and given by one individual. Our study is unable to distinguish if the association between the personal finance curriculum and improved well-being is due to the instructor and specific course design or becoming more financial literate in general. A final limitation is that surveys were given during an unprecedented time for medical education. The distribution to program structures and resident and fellow schedules due to the COVID-19 pandemic cannot be quantified. It is difficult to evaluate if the improvement in scores was due to the personal finance curriculum or other circumstantial changes. Despite the fact that stress and anxiety levels of residents/fellows were increasing during the COVID-19 pandemic, 23 our trainees showed an improvement in the well-being index. This may further substantiate our findings that financial literacy improves one's sense of well-being. We also witnessed that residents or fellows do not need to be exposed to a whole personal financial curriculum to see benefit. The median attendance for residents and fellows were 2.8 lectures, just over 50% of the entire curricula. Even a small amount of an exposure to monetary topics can improve one's financial literacy and well-being.
Given the overarching limitations of our study, further studies must be done. It should be evaluated at other GME sites and in different specialty programs which we are evaluating following this pilot. In addition, separate studies should be done to assess the effect on medical students. Furthermore, other curricula need to be evaluated to discover if improvement is due to a specific course structure or if it is the financial education itself that is associated with an improved sense of well-being.
# Conclusion
In conclusion, we demonstrated that a personal financial course is associated with an improvement in a resident and fellow's sense of well-being. Incorporation and further study of a personal finance component into the wellness curriculum should be considered by all OB/GYN programs.
## Conflict of interest
The primary author is a registered investment advisor with Navigo Wealth Management.
[fig] Figure 1: The list of covered topics that topics were chosen based on the primary authors experience, as well as what trainees have previously expressed interest in learning. study by taking the first survey, 21 of 35 (60%) took part in the follow-up survey. The mean number of individual course lectures attended during a live session was around 3 (►Table 2). [/fig]
[table] Table 1: Personal characteristics of the OB/GYN residents and fellows [/table]
[table] Table 2: OB/GYN resident/fellow exposure and participation in the personal finance curriculum [/table]
[table] Table 3: Expanded well-being index and financial stress scores before and after exposure to the personal finance curriculum [/table]
[table] Table 4: Trainee attitudes on financial education [/table]
|
Acute ischaemic stroke interventions: large vessel occlusion and beyond
Care for acute ischaemic stroke is one of the most rapidly evolving fields due to the robust outcomes achieved by mechanical thrombectomy. Large vessel occlusion (LVO) accounts for up to 38% of acute ischaemic stroke and comes with devastating outcomes for patients, families and society in the pre-intervention era. A paradigm shift and a breakthrough brought mechanical thrombectomy back into the spotlight for acute ischaemic stroke; this was because five randomised controlled trials from several countries concluded that mechanical thrombectomy for acute stroke offered overwhelming benefits. This review article will present a comprehensive overview of LVO management, techniques and devices used, and the future of stroke therapy. In addition, we review our institution experience of mechanical thrombectomy for posterior and distal circulation occlusion.
# Introduction
Care for acute ischaemic stroke is one of the most rapidly evolving fields due to the robust outcomes achieved by mechanical thrombectomy. Until the mid-1990s, little to no progress had been made in the treatment of stroke. Intravenous tissue plasminogen activator (tPA) was found to have some (but limited) benefit. [bib_ref] Blazing the frontiers of stroke therapy, Hopkins [/bib_ref] However, due to the narrow window of treatment (that was initially 3 hours then extended to 4.5 hours) where only 3.4%-5.2% of patients who had a stroke receive tPA, 2 the low rate of vessel recanalisation (13%-50%) [bib_ref] Site of arterial occlusion identified by transcranial Doppler predicts the response to..., Saqqur [/bib_ref] [bib_ref] The benefits of intravenous thrombolysis relate to the site of baseline arterial..., Silva [/bib_ref] [bib_ref] Systemic thrombolysis in patients with acute ischemic stroke and internal carotid artery..., Paciaroni [/bib_ref] [bib_ref] Thrombolytic therapy in acute occlusion of the intracranial internal carotid artery bifurcation, Jansen [/bib_ref] and the low rate of favourable outcomes (12.9%-30%), extensive efforts have been placed to achieve outcomes similar or equivalent to the cardiac field. In 2013, three randomised controlled trials (RCTs) failed to show real benefit for catheter-based stroke interventions for large vessel occlusions (LVOs). [bib_ref] Endovascular therapy after intravenous t-PA versus t-PA alone for stroke, Broderick [/bib_ref] [bib_ref] Endovascular treatment for acute ischemic stroke, Ciccone [/bib_ref] [bib_ref] A trial of imaging selection and endovascular treatment for ischemic stroke, Kidwell [/bib_ref] Two years later, a paradigm shift and a breakthrough brought back mechanical thrombectomy for acute ischaemic stroke into the spotlights; this was because five RCTs from several countries concluded that mechanical thrombectomy (MT) for acute stroke offered overwhelming benefits. [bib_ref] A randomized trial of intraarterial treatment for acute ischemic stroke, Berkhemer [/bib_ref] [bib_ref] Endovascular therapy for ischemic stroke with perfusion-imaging selection, Campbell [/bib_ref] [bib_ref] Randomized assessment of rapid endovascular treatment of ischemic stroke, Goyal [/bib_ref] [bib_ref] Thrombectomy within 8 hours after symptom onset in ischemic stroke, Jovin [/bib_ref] [bib_ref] Stent-retriever thrombectomy after intravenous t-PA vs. t-PA alone in stroke, Saver [/bib_ref] LVO accounts for up to 38% of acute ischaemic stroke and came with devastating outcomes for patients, families and society in the pre-intervention era. Efficacy of intervention in preselected patients presenting within 24 hours has been an established treatment paradigm. EpIdEmIology And nATurAl hIsTory of lVo Stroke is the second leading cause of death worldwide, with a 16 million incidence.The prevalence of stroke among the US population increases with age starting with 2.7% among people 20 years of age, 6% over 60 years and reaching 13% for people above 80 years. Each year, there are ~800 000 new or recurrent cases of stroke. With all the advancements, 17.5% still die, making it the fifth leading cause of death in the USA.Patients who survive stroke may end up with long-term disability, requiring rehabilitation with an associated annual cost of more than US$40 billion in the USA alone.In addition, the disability burden attributed to stroke continues to grow, with an estimated increase to 68 million disability-adjusted life years in 2020. The incidence of LVO, as defined in the introduction, compromises 24% to 38% of acute ischaemic stroke. The proportion increases to 46% on including A2 and P2 segments. [bib_ref] Significance of large vessel intracranial occlusion causing acute ischemic stroke and TIA, Smith [/bib_ref] The estimated 10 000 annual thrombectomies performed in recent years remains well below LVO incidence [bib_ref] A population-based incidence of acute large vessel occlusions and thrombectomy eligible patients..., Rai [/bib_ref] and suggests a potentially significant future increase in the utilisation of endovascular stroke therapies. [bib_ref] Epidemiology, natural history, and clinical presentation of large vessel ischemic stroke, Rennert [/bib_ref] Twothirds of LVO occur in the anterior circulation, mainly in the Internal Carotid Artery (ICA) and Middle Cerebral Artery, and the remaining occur in the posterior circulation with equivalent distribution among the Vertebral Artery, Basilar Arterty and Posterior Cerebral Artery. [bib_ref] Significance of large vessel intracranial occlusion causing acute ischemic stroke and TIA, Smith [/bib_ref] Tandem occlusions can occur in less than 10%. Open access A meta-analysis reported a double proportion of poor outcome (modified Rankin Scale (mRS) 3-6) and mortality in patients with LVO as compared with patients without LVO in the pre-endovascular era (64 vs 24%, p<0.0001, 26.2 vs 1.3%, p<0.0001, respectively). [bib_ref] Ischemic strokes due to largevessel occlusions contribute disproportionately to stroke-related dependence and..., Malhotra [/bib_ref] Intravenous tPA has two main limitations: it has a narrow treatment window of up to 4.5 hours since 2008 and high pharmacological resistance for more proximal occlusions (4%-8% for ICA vs 31%-44% distal recanalisation). prE-InTErVEnTIon workflow The terms 'time is brain' and 'picture to puncture' reflect the importance of early reperfusion for clinical outcomes established by several studies. [bib_ref] Picture to puncture": a novel time metric to enhance outcomes in patients..., Nogueira [/bib_ref] [bib_ref] Time to angiographic reperfusion and clinical outcome after acute ischaemic stroke: an..., Khatri [/bib_ref] [bib_ref] Analysis of workflow and time to treatment and the effects on outcome..., Goyal [/bib_ref] [bib_ref] Analysis of workflow and time to treatment on thrombectomy outcome in the..., Menon [/bib_ref] Every 30 min delay in MT decreases favourable outcomes by 11%, [bib_ref] Predictors of functional dependence despite successful revascularization in large-vessel occlusion strokes, Shi [/bib_ref] and every 15 min acceleration at initiating tPA is associated with 4% greater odds of walking independently at discharge. [bib_ref] Estimated impact of emergency medical service triage of stroke patients on comprehensive..., Katz [/bib_ref] Tremendous effort has been made by individual institutions and the healthcare system overall to overcome barriers to reduce the time for stroke treatment. [bib_ref] Factors influencing in-hospital delay in treatment with intravenous thrombolysis, Mikulík [/bib_ref] [bib_ref] Metrics for measuring quality of care in comprehensive stroke centers: detailed followup..., Leifer [/bib_ref] [bib_ref] Reducing doorto-puncture times for intra-arterial stroke therapy: a pilot quality improvement project, Mehta [/bib_ref] Several modifiable factors have been improved and refined (such as stroke assessment tools, pre-hospital notification, communication between emergency personnel and stroke specialists, 'stroke telemedicine', mobile stroke units and direct transfer to comprehensive stroke centres), resulting in improved stroke care timing. Still, a small proportion of thrombectomy eligible patients, 27%, receive a MT. [bib_ref] A population-based incidence of acute large vessel occlusions and thrombectomy eligible patients..., Rai [/bib_ref] Emergency medical personnel triage suspected patients who had a stroke based on the clinical examination alone. An efficient and rapid assessment has a profound influence on stroke outcomes. Multiple stroke scales have been validated for clinical use in the stroke field and have shown to speed up the triage process. The National Institutes of Health Stroke Scale (NIHSS) is complicated and time intensive (as time is often limited in field assessment); because of this, multiple alternative stroke scales exist specifically for quick field triage such as the Cincinnati Prehospital Stroke Severity Scale, the Los Angeles Motor Scale and the Rapid Arterial Occlusion Evaluation. However, a recent meta-analysis reported that the scales have a low predictive value for the presence of LVO (35% to 50%). [bib_ref] Accuracy of prediction instruments for diagnosing large vessel occlusion in individuals with..., Smith [/bib_ref] Another factor that reduces the time for treatment duration is a prehospital notification. The American Heart Association (AHA)/American Stroke Association (ASA) recommends pre-hospital notification of potential patients who had a stroke to the destination medical centre. [bib_ref] guidelines for the early management of patients with acute ischemic stroke: a..., Powers [/bib_ref] A recent analysis showed that the prehospital notification procedure improves stroke outcome by shortening onset to needle time. [bib_ref] Prehospital notification procedure improves stroke outcome by shortening onset to needle time..., Zhang [/bib_ref] The introduction of mobile stroke units with imaging capabilities, teleconsults with stroke specialist to assist with triaging, and potential to initiate intravenous thrombolysis tPA en route to an advanced stroke centre have been an extra step towards more prompt care and have also been shown to improve care. [bib_ref] Stroke team remote evaluation using a digital observation camera in Arizona: the..., Demaerschalk [/bib_ref] [bib_ref] Efficacy of site-independent telemedicine in the stroke Doc trial: a randomised, blinded,..., Meyer [/bib_ref] [bib_ref] The cost-effectiveness of telestroke in the treatment of acute ischemic stroke, Nelson [/bib_ref] [bib_ref] Cost-Effectiveness of hub-and-spoke TeleStroke networks for the management of acute ischemic stroke..., Switzer [/bib_ref] [bib_ref] Implementing a mobile stroke unit program in the United States: why, how,..., Rajan [/bib_ref] [bib_ref] Prehospital computed tomography angiography in acute stroke management, Kettner [/bib_ref] [bib_ref] Mobile stroke unit reduces time to image acquisition and reporting, Nyberg [/bib_ref] [bib_ref] Establishing the first mobile stroke unit in the United States, Parker [/bib_ref] Finally, there has been a stance to directly transfer suspected patients with LVO to comprehensive stroke centres and bypass non-comprehensive centres to improve treatment time for patients. [bib_ref] Emerging technologies in optimizing pre-intervention workflow for acute stroke, Yaeger [/bib_ref] In a study, the median hospital-to-hospital distance was 14.7 miles, and median transfer time was 104 min. [bib_ref] Transfer delay is a major factor limiting the use of intra-arterial treatment..., Prabhakaran [/bib_ref] Even hospital workflow has been adjusted for better outcomes. The change from the traditional hierarchical linear method of patient care to a single-call stroke code activation where all essential personnel are notified concurrently has been shown to reduce time to treatment in acute stroke care. The Society of NeuroInterventional Surgery recommends stroke code activation, [bib_ref] Initial hospital management of patients with emergent large vessel occlusion (ELVO): report..., Mctaggart [/bib_ref] and similarly, the AHA urges for protocols that limit treatment delays. [bib_ref] guidelines for the early management of patients with acute ischemic stroke: a..., Powers [/bib_ref] mEChAnICAl ThrombECTomy Over the last decade, the natural history of ischaemic stroke has been dramatically improved following the five pivotal trials that set the ground for MT as the standard of care. The first attempt was in 2013, when three RCTs (Interventional Management of Stroke III (IMS III), Mechanical Retrieval and Recanalization of Stroke Clots Using Embolectomy (MR Rescue) and A Randomized Controlled Trial on Intra-arterial vs Intravenous Thrombolysis in Acute Ischemic Stroke (SYNTHESIS)) failed to exhibit the superiority of mechanical thrombectomy compared with standard medical treatment. [bib_ref] Endovascular therapy after intravenous t-PA versus t-PA alone for stroke, Broderick [/bib_ref] [bib_ref] Endovascular treatment for acute ischemic stroke, Ciccone [/bib_ref] [bib_ref] A trial of imaging selection and endovascular treatment for ischemic stroke, Kidwell [/bib_ref] The main limitation present within these three RCTs was the use of old thrombectomy devices and the lack of uniform protocol for confirming LVO on CT angiography, which introduced bias against endovascular therapy. [bib_ref] Emerging technologies in optimizing pre-intervention workflow for acute stroke, Yaeger [/bib_ref] In IMS III, 20% of patients randomised to the endovascular arm did not have a LVO. As a result, the 2013 AHA/ ASA guidelines for the early management of patients with acute ischaemic stroke advised that the "ability to improve patient outcomes has not yet been established" for thrombectomy devices. [bib_ref] Guidelines for the early management of patients with acute ischemic stroke: a..., Jauch [/bib_ref] Open access meta-analysis of the five trials by HERMES collaborators reported that MT for anterior circulation groups significantly reduced overall 90-day disability, with a number needed to treat of 2.6 to reduce the mRS in one patient by at least one point. [bib_ref] Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data..., Goyal [/bib_ref] More recently, the Clinical Mismatch in the Triage of Wake Up and Late Presenting Strokes Undergoing Neurointervention With Trevo (DAWN) and A multicentre RCT of endovascular therapy following imaging evaluation for ischaemic stroke (DEFUSE 3) trials showed functional benefit following MT in patients presenting after 6 hours from insult. Eligible patients for MT had a mismatch on perfusion imaging. The DAWN trial showed a benefit from 6 to 24 hours, and the DEFUSE 3 showed benefit from 6 to 16 hours. [bib_ref] Thrombectomy 6 to 24 hours after stroke with a mismatch between deficit..., Nogueira [/bib_ref] 20 DEFUSE 3 found that MT along with standard medical therapy resulted in a higher percentage of functionally independent patients than standard medical treatment alone (45% vs 17%). [bib_ref] Thrombectomy for stroke at 6 to 16 hours with selection by perfusion..., Albers [/bib_ref] The CT Perfusion to Predict Response to Recanalization in Ischemic Stroke Project (CRISP) trial results are consistent with DEFUSE 3 and DAWN. Perfusion scans were used to identify patients with perfusion mismatch, and results showed that a higher rate of improvement on the NIHSS of >8 points was observed in subjects with a significant mismatch (83% vs 44%). 54 Importantly, this favourable rate of improvement remained regardless of whether patients were treated within or beyond 6 hours. Based on those five RCTs the AHA/ASA published updated guidelines in 2018, recommending that endovascular thrombectomy be considered for LVOs of the posterior circulation up to 6 hours from symptom onset and that intervention is recommended/can be considered up to 16 and 24 hours, respectively, in select patients with anterior circulation LVOs based on perfusion imaging. Also, the 2018 guidelines stress the need for perfusion scan as triaging imaging for patients presenting between 6 and 24 hours from symptom onset. [bib_ref] guidelines for the early management of patients with acute ischemic stroke: a..., Powers [/bib_ref] The posterior circulation is still a grey zone, where perfusion is limited in stratifying patients that may be good candidates for mechanical intervention. Consequently, physician experience and judgement play a paramount role. [bib_ref] Overview of mechanical thrombectomy techniques, Munich [/bib_ref] We reviewed our cohort of mechanical thrombectomy for posterior circulation. Out of 453 thrombectomy procedures, 45 were located in the posterior circulation and were mainly a basilar artery occlusion (87%). A solumbra technique, which is the simultaneous use of a stent retriever and aspiration catheter, was used in more than half of the procedures, and an optimal recanalisation rate (TICI >2 b) was achieved in 69%. Revascularisation was achieved using a single pass in 44%. When compared with the anterior circulation, posterior thrombectomy procedures were associated with a significant higher rate of extravasation (9% vs 4%, p<0.01) and post-procedural symptomatic Intracerebral Hemorrhage (ICH) (13% vs 4 %, p<0.01). Mortality occurred in 38%, and only 20% achieved functional independence at 90 days. While the functional outcomes are lower in posterior circulation compared with the anterior circulation, the comparison should be made to tPA alone for acute basilar occlusion since the natural history of posterior circulation occlusion is worse compared with anterior circulation. [bib_ref] Overview of mechanical thrombectomy techniques, Munich [/bib_ref] When interpreted in this context, mechanical thrombectomy for posterior circulation results in superior outcomes.
We also reviewed our case series of distal thrombectomy procedures. The M2 was involved in 89%, and ~80% of thrombectomy procedures were performed using the Solumbra technique. Peri-procedural complications including distal emboli and subarachnoid hemorrhage (SAH)/ICH occurred at a rate of 7% and 8.5%, respectively. Compared with proximal occlusions, distal emboli occurred at a significantly higher incidence in the distal circulation occlusions (7% vs 2%, p=0.01). Effective recanalisation rates occurred in 89%, with no difference compared with proximal circulation. Favourable outcomes occurred in 64.7%, and mortality in 8%.
## Mt techniques retrievable stents
The improvements introduced to the first-generation Merci device (Concentric Medical, CA, USA) resulted in improved recanalisation rates by the second-generation devices, Solitaire (Medtronic, MN, USA) and Trevo (Stryker, MI, USA). Second-generation devices showed threefold improved recanalisation rates compared with the first-generation. [bib_ref] The 2012 Feinberg Lecture: treatment swift and treatment sure, Saver [/bib_ref] More than 80% of enrolled patients in the five RCTs published in 2015 underwent thrombectomy using second-generation stent retrievers contrast to the Merci device used in earlier RCT resulted in superior outcomes in the endovascular arm. [bib_ref] A randomized trial of intraarterial treatment for acute ischemic stroke, Berkhemer [/bib_ref] [bib_ref] Endovascular therapy for ischemic stroke with perfusion-imaging selection, Campbell [/bib_ref] [bib_ref] Randomized assessment of rapid endovascular treatment of ischemic stroke, Goyal [/bib_ref] [bib_ref] Thrombectomy within 8 hours after symptom onset in ischemic stroke, Jovin [/bib_ref] [bib_ref] Stent-retriever thrombectomy after intravenous t-PA vs. t-PA alone in stroke, Saver [/bib_ref] The safe and effective results achieved by stent retrievers led to off-label use and a push in the industry for refinements in succeeding designs. Effective recanalisation rates (58%-88%), 12-14 high rates of functional independence at 3 months (53%-71% using stent retriever vs 19%-40% using tPA alone), along with the low incidence of complications were achieved in the five RCTs. [bib_ref] A randomized trial of intraarterial treatment for acute ischemic stroke, Berkhemer [/bib_ref] [bib_ref] Endovascular therapy for ischemic stroke with perfusion-imaging selection, Campbell [/bib_ref] [bib_ref] Randomized assessment of rapid endovascular treatment of ischemic stroke, Goyal [/bib_ref] [bib_ref] Thrombectomy within 8 hours after symptom onset in ischemic stroke, Jovin [/bib_ref] [bib_ref] Stent-retriever thrombectomy after intravenous t-PA vs. t-PA alone in stroke, Saver [/bib_ref] Such favourable results encouraged the development of novel retrievable stent designs. Both the Penumbra 3D revascularisation device (Penumbra, CA, USA) and the EmboTrap retrievable stent (Cerenovus/ Johnson & Johnson, NJ, USA) achieved very promising effective revascularisation rates, 84% and 92.5%, respectively. The off-label use of mechanical thrombectomy encompasses posterior and distal circulation. Effective recanalisation rates following posterior mechanical thrombectomy using stent retriever were reported in 90% and good functional outcomes at 3 months were achieved by 46.9%. [bib_ref] Endovascular thrombectomy for acute basilar artery occlusion: a multicenter retrospective observational study, Kang [/bib_ref] Similarly, optimal results have been achieved with the use of stent retrievers for distal circulation occlusions, with successful reperfusion achieved in 85%, favourable outcomes in 60% and symptomatic ICH at 2%. [bib_ref] Mechanical thrombectomy for isolated M2 occlusions: a post hoc analysis of the..., Coutinho [/bib_ref] Aspiration devices Due to the limitations of early aspiration devices and the use of stent-retriever devices in Mr Clean, Extend IA, Escape and Swift Prime, most succeeding trials have used stent-retriever devices as the modality for thrombectomy. However, recent data suggest that aspiration provides an efficient and effective means of recanalisation in patients with LVOs. [bib_ref] Effect of endovascular contact aspiration vs stent retriever on revascularization in patients..., Lapergue [/bib_ref] [bib_ref] Contact aspiration versus stentretriever thrombectomy for distal middle cerebral artery occlusions in..., Phan [/bib_ref] [bib_ref] Initial clinical experience with the adapt technique: a direct aspiration first pass..., Turk [/bib_ref] [bib_ref] ADAPT FAST study: a direct aspiration first pass technique for acute stroke..., Turk [/bib_ref] The Contact Aspiration vs Stent Retriever for Successful Revascularization study assessed outcomes in thrombectomy procedures using contact aspiration compared with stent retriever. The study concluded that outcomes were not different among both modalities; recanalisation rate (85.4%), functional outcomes (45.3%) and morbidity. [bib_ref] Effect of endovascular contact aspiration vs stent retriever on revascularization in patients..., Lapergue [/bib_ref] Promising results pushed the envelope and paved the way for aspiration first pass technique (ADAPT). Efficacy of ADAPT has been endeavoured by the recently completed Comparison of Direct Aspiration vs Stent Retriever as a First Approach (COMPASS) trial. The investigators found no difference in functional outcomes, revascularisation rates (92% aspiration, 89% retrievable stents) and morbidity.Aspiration catheters may be valuable in distal occlusion because there is no need for stent deployment, and it may be more gentle on the vessels causing fewer manipulations.
## Combination therapy
Recently, thrombectomy techniques have been modified where both stent retriever and direct aspiration are used concurrently (the 'Solumbra technique') in an attempt to improve recanalisation rates. In a series from six highvolume centres, Humphries et al reported favourable outcomes following the use of the Solumbra technique for mechanical thrombectomy. The authors reported achieving 88% TICI 2b/3 recanalisation and 44% favourable mRS outcomes at 90 days. [bib_ref] Distal aspiration with retrievable stent assisted thrombectomy for the treatment of acute..., Humphries [/bib_ref] fuTurE of sTrokE ThErApy Technological innovations have resulted in a significant shift in patient care, two of which are worth mentioning; The volumetric impedance phase shift spectroscopy device (Cerebrotech, CA, USA) has demonstrated high sensitivity and specificity in detecting LVOs in patients who had a stroke. It is a non-invasive device that is placed on the patient's head that detects changes in bioimpedance in the brain caused by several pathologies, including ischaemia. Another non-invasive device is the SONAS device (BURL Concepts, CA, USA) that uses transcranial ultrasound and microbubble intravenous contrast to identify potential LVOs.
Perhaps more notably, catheter advancements are occurring at a rapid pace aiming to improve manoeuvrability, efficacy and safety of the devices. The Lazarus Effect Cover (Medtronic, MN, USA) is designed to provide a protective sheath around retrievable stents to protect against distal embolisation. Once the stent engages with the clot, it is re-sheathed to prevent clot fragmentation and dislodgment. [bib_ref] Stent retriever thrombectomy with the cover accessory device versus proximal protection with..., Mokin [/bib_ref] A small case series of 20 patients reported high rate of effective recanalisation and no evidence of distal emboli. [bib_ref] Abstract WMP2: a new device to prevent embolization in new territories during..., Piotin [/bib_ref] Another innovative concept is the R4Q aspiration catheter (MIVI Neuroscience, MN, USA). The distal one-quarter of the catheter functions as an extension of the guide catheter, allowing the full guide catheter to be used to deliver suction. [bib_ref] Novel aspiration catheter design for acute stroke thrombectomy, Long [/bib_ref] The future of stroke will be stem cell therapy to stimulate prompt neurological recovery.
# Conclusion
We are in an era where the second leading cause of death in the world that carries a high burden has become a treatable disease. Huge efforts on all levels have been placed and are still laid to improve time for stroke treatment; beginning with field triage to clot retrieval and post-procedural care and rehabilitation.
Twitter Pascal Jabbour @PascalJabbourMD funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests PJ is a consultant for Medtronic and Micro Vention. ST is a consultant for Stryker. patient consent for publication Not required. provenance and peer review Commissioned; internally peer reviewed.
open access This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http:// creativecommons. org/ licenses/ by-nc/ 4. 0/. orCId id Pascal Jabbour http:// orcid. org/ 0000-0002-8965-2413
[fig] Contributors: All authors reviewed and agreed on the final version of the manuscript. Conception or design of the work: AS, BH, SR, JHW, PJ. Drafting the work: AS, BH, SR, DW, TDA, MD, JD, DJ-MM. Revising the work for valuable intellectual content: MRG, HZ, VR, DMH, RHR. Final approval of the version: ST, PJ. [/fig]
|
Development of new microalgae-based sourdough “crostini”: functional effects of Arthrospira platensis (spirulina) addition
The aim of this work was to evaluate the influence of Arthrospira platensis F&M-C256 (spirulina) incorporation on the nutritional and functional properties of "crostini", a leavened bakery product largely consumed in italy and europe. Sourdough was used as leavening and fermentation agent and three concentrations of A. platensis F&M-C256 were tested: 2%, 6% and 10% (w/w). Despite a lower volume increase compared to the control, the A. platensis F&M-C256 "crostini" doughs reached a technological appropriate volume after fermentation. At the end of fermentation, no significant differences in microorganisms concentrations were observed. A. platensis F&M-C256 "crostini" showed higher protein content compared to the control. considering the european commission Regulation on nutritional claims, "crostini" incorporated with 6% and 10% biomass can be claimed to be a "source of protein". Six and ten percent A. platensis "crostini" also presented significantly higher antioxidant capacity and phenolics. A significantly lower value of in vitro dry matter and protein digestibility between A. platensis F&M-C256 "crostini" and the control was found. The overall acceptability decreased with increasing A. platensis F&M-C256 addition. The combination of spirulina biomass addition and the sourdough technology led to the development of a novel microalgae-based bakery product with nutritional and functional features. . Biochemical composition of A. platensis F&M-C256 biomass used in the experiments. Data are expressed as % (w/w) of algal powder. Results are expressed as average ± standard deviation (n = 3).
## Dough analyses.
Microbiological analyses. Ten grams of doughs at the end of fermentation time were transferred into 90 mL of sterile physiological solution and homogenized for 2 min in a Stomacher Lab Blender 400 (Seward Ltd, Worthing, West Sussex, UK). Afterwards, decimal dilutions were performed and 100 µL of appropriate dilutions were plated onto MR3i medium for enumeration of LAB and onto MYPG agar containing sodium propionate (2 g/L) for enumeration of S. cerevisiae. LAB colonies were counted after incubation for 48-72 h at 30 °C under anaerobic conditions and yeast colonies after incubation for 48 h at 30 °C under aerobic conditions. Plate counts were performed in duplicate.
Physico-chemical analyses. The pH and the total titratable acidity (TTA) of the doughs were determined at the beginning and at the end of the fermentation time. The pH values were measured by a pH-meter (Metrohm Italiana Srl, Varese, Italy).
TTA was determined on doughs and A. platensis F&M-C256 powder by using 10 g of samples, homogenized with 90 mL of distilled water for 3 min and expressed as the amount (mL) of 0.1 N NaOH needed to reach a pH value of 8.5.
The increase in dough volume was assessed by placing 100 g of each dough in a graduated cylinder (1 L). After 2 h of fermentation at 30 °C the volume of the doughs (in mL) was recorded. The volume increase was calculated using the following formula: (∆V/V0) x 100, where ∆V was the difference between the volume after the 2 h of fermentation and the initial volume (V0) [bib_ref] Microbiological and technological characterization of sourdoughs destined for bread-making with barley flour, Zannini [/bib_ref].
"crostini" analyses. HPLC analysis of organic acids. "Crostini" and A. platensis F&M-C256 samples were crushed to powder, diluted ten times and filtered by Amicon ® Ultra-4 Centrifugal Filters (3.000 Da NMWL) (Merck Millipore) before the injection. Lactic and acetic acid were determined by High-Performance Liquid Chromatography (HPLC) analysis (Varian Inc, Palo Alto, CA, USA).
Free amino acid determination. For the free amino acid (FAA) quantification, water/salt soluble extracts (WSE) of "crostini" were prepared as follows. Five grams of the sample were diluted with 20 mL of 50 mM Tris-HCl (pH 8.8), held at 4 °C for 1 h, vortexed at 15-min intervals, and centrifuged at 20,000 × g for 20 min [bib_ref] Selected lactic acid bacteria synthesize antioxidant peptides during sourdough fermentation of cereal..., Coda [/bib_ref]. The supernatants were used for the analysis. Before injection on HPLC for FAA quantification, the sample extracts were prepared according to . Separation was obtained with a Kinetex 5 µm C18 100 Å column (150 × 4.60 mm; Phenomenex, Bologna, Italy) connected to fluorimetric detector (Jasco Europe, Lecco, Italy) with wavelengths set at 293 nm (Ex) and 492 nm (Em) under the conditions reported by Tuberoso et al. [bib_ref] Determination of dansylated amino acids and biogenic amines in Cannonau and Vermentino..., Tuberoso [/bib_ref]. The quantitative analysis was performed using calibration graphs constructed according to the internal standard method.
Proximate chemical composition and colour analyses. The "crostini" biochemical composition was determined as in Abiusi et al. [bib_ref] Growth, photosynthetic efficiency, and biochemical composition of Tetraselmis suecica F&M-M33 grown with..., Abiusi [/bib_ref]. Moisture and ash were analysed following ISTISAN protocols. The total energetic value was calculated following the conversion factors designated in the annex XIV of Regulation (EU) N° 1169/2011 49 .
The colour of "crostini" samples was measured instrumentally using a Minolta CR-400 (Japan) colorimeter with standard illuminant D65 and a visual angle of 2°. The results were expressed in terms of L*, lightness (values increase from 0 to 100%); a*, redness to greenness (+60 to −60 positive to negative values, respectively); b*, yellowness to blueness (+60 to −60 positive to negative values, respectively), according to the CIEL*a*b* system. Chroma, C* ab (saturation), and hue angle, h°a b , were also calculated, as defined by: C* ab = [(a* 2 + b* 2 )] 1/2 ; h°a b = arctan (b*/a*). The measurements were conducted under the same light conditions, using a white standard (L* = 94.61, a* = −0.53, b* = 3.62), under artificial fluorescent light at room temperature, replicated twelve times for each formulation sample (one measurement per "crostino").
Phycocyanin, phenolics and antioxidant capacity determination. Phycocyanin content in A. platensis F&M-C256 lyophilised biomass and in powdered "crostini" was determined according to Herrera et al. [bib_ref] Recovery of c-phycocyanin from the cyanobacterium Spirulina maxima, Herrera [/bib_ref]. "Crostini" recipes (%, w/w). Control (without A. platensis F&M-C256 incorporation) (C) and "crostini" enriched with 2% (Ap 2), 6% (Ap 6) and 10% (Ap 10) (w/w) A. platensis F&M-C256 biomass.
extracted suspending freeze dried A. platensis F&M-C256 powder or "crostini" powders in a 1% (w/v) CaCl 2 solution at pH 6.8 (typically 1 g powder in 20 mL of extraction solution). The extracted phycocyanin was spectrophotometrically quantified and purity was assessed at 620 nm and 280 nm using a UV-Vis spectrophotometric reader (Cary 60 UV-Vis, Agilent Technologies, California, USA). The total phenolic content assay was carried out according to Rajauria et al. [bib_ref] antioxidant and free radical-scavenging capacity of brown seaweed Himanthalia elongata from western..., Rajauria [/bib_ref] using the Folin Ciocalteu assay. Samples of 0.1 g of lyophilised "crostini" were dissolved in 6 mL of deionised water and sonicated for 30 min at the maximum power (frequency 20 kHz, power 130 W) maintaining the temperature below 30 °C by immerging the sample flask in an ice bath (MicrosonTM XL2000, Misonix Inc., Farmingdale, New York, USA). To 100 µL aliquots of each sample, 2 mL of 2% sodium carbonate (Sigma-Aldrich) in water were added. After 2 min, 100 μL of 50% Folin Ciocalteu reagent (Sigma-Aldrich) were added. The reaction mixture was incubated in darkness at 25 °C for 30 min. The absorbance of each sample was measured at 720 nm using a UV-Vis spectrophotometric reader (Cary 60 UV-Vis, Agilent Technologies, California, USA). Results were expressed in gallic acid equivalents (mg GAE g −1 ) through a calibration curve with gallic acid (0 to 150 μg mL −1 , R 2 = 0.9907) (Sigma-Aldrich).
To evaluate the radical scavenging capacity of the "crostini" samples, the 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical scavenging assay was carried out according to . Briefly, the assay was performed by reaction of 1 mL of DPPH radical solution (165 μM, in methanol, Sigma-Aldrich) with 1 mL of sample (0.2 g of lyophilised "crostini" extracted for 30 min in 5 mL of a 1:5 methanol/:water solution). The reaction mixtures were incubated in darkness at 30 °C for 30 min and the absorbances were measured at 517 nm using a UV-Vis spectrophotometric reader (Cary 60 UV-Vis, Agilent Technologies, California, USA). The antioxidant capacity of the samples was expressed as Radical Scavenging Activity (% RSA). Samples RSA capacity was also converted in µg of Vitamin C Equivalent Antioxidant Capacity (VCEAC) per milligram of sample through a calibration curve (ascorbic acid: 0 to 10 µg mL −1 , R 2 = 0.9918).
In vitro digestibility tests. The in vitro digestibility was evaluated by the method of Boisen & Fernández 52 , modified by Niccolai et al. [bib_ref] In vitro toxicity of microalgal and cyanobacterial strains of interest as food..., Niccolai [/bib_ref] , on lyophilised "crostini". The method reproduces the chemical-enzymatic digestion (by gastric and pancreatic juices) that occurs in the proximal tract of the monogastric digestive system. One-gram samples (particle size ≤ 1 mm) were weighed and transferred to 250 mL conical flasks. The analysis comprised two steps of enzymatic digestion: the first performed with porcine pepsin (Applichem, Germany) and the second with porcine pancreatin (Applichem) [bib_ref] Microalgae biomass as an alternative ingredient in cookies: sensory, physical and chemical..., Batista [/bib_ref]. The dry matter (IVD) and protein (IVPD) in vitro digestibility (%) of A. platensis F&M-C256 biomass and of "crostini" samples were calculated from the difference between the initial biomass and the undigested biomass (after correction for the blank assay) divided by the initial biomass and multiplied by 100. Casein (Sigma Aldrich Corp., St. Louis, USA) was utilised as the reference material with 100% digestibility.
Sensory analysis. An untrained panel of 35 people, 17 males and 18 females, aged from 18 to 60, evaluated the "crostini" in terms of colour, smell, taste, texture, global appreciation (6 levels from "very pleasant" to "very unpleasant"). The buying intention was also assessed, from "would certainly buy" to "certainly would not buy" (5 levels). The assays were conducted in a standardized sensory analysis room, according to the standard EN ISO 8589 (2010).
# Statistical analysis.
Statistical analysis of the experimental data was performed using STATISTICA from StatSoft (version 8.0) and using Statgraphics Centurion XV from StatPoint Technologies Inc., through variance analysis (one-way ANOVA), by the Scheffé test -Post Hoc Comparison (significance level of 95%, p < 0.05) and through ANOVA followed by multiple range tests to determine the least significant differences (LSD) (significance level of 95%, p < 0.05). All analyses were conducted at least in triplicate and presented as average ± standard deviation.
## Results and discussion fermentation parameters of dough and chemical analysis of "crostini" and a. platensis f&m-c256.
Two common parameters used to characterize the sourdough, Total Titratable Acidity (TTA) and organic acids concentrations, were determined also on A. platensis F&M-C256 biomass to evaluate its influence on sourdough characteristics. No data concerning TTA in microalgae biomasses are available in the literature, our results showed a TTA value of 15 mL and lactic acid and acetic acid concentrations of 0.84 g/100 g and 0.42 g/100 g, respectively. Consequently, microalgae addition affected the acidification parameters of doughs and "crostini" as discussed below. shows both the values of the principal technological parameters and the cell densities of doughs at the end of the fermentation, and the organic acids and free amino acid (FAA) content in the "crostini".
The control displayed the lowest ∆TTA, 0.70 ± 0.00 mL, and the highest ∆pH, 0.36 ± 0.05 at the end of the fermentation process; the opposite trend was observed for Ap 10, that was characterized by the highest ∆TTA and the lowest ∆pH among the samples. In addition to the acidifying effect, A. platensis F&M-C256 acted also as buffering agent, probably due to residual bicarbonates of the growth medium, hence to a higher ∆TTA did not correspond a greater pH decrease. All the samples at least doubled their volume at the end of fermentation time, although the volume increase of Ap 10 was lower compared to the control, indicating that a high percentage of A. platensis F&M-C256 negatively affects the gluten network development of the dough. Microorganism concentrations were not affected by A. platensis F&M-C256 supplementation, and showed final values of about 8 log CFU/g for L. farciminis and 7 log CFU/g for S. cerevisiae consistent with those reported for stable sourdoughs 2 . Organic acids concentrations, determined in the "crostini" after cooking, were affected by their content in A. platensis F&M-C256 powder rather than by different LAB activity. The amount of lactic acid did not show significant differences among the samples, although a higher lactic acid content was found with the increase of www.nature.com/scientificreports www.nature.com/scientificreports/ A. platensis F&M-C256 concentrations. For acetic acid concentration, a statistically significant (p < 0.05) increase was observed for higher A. platensis F&M-C256 incorporations (Ap 6 and Ap 10).
FAA content increased with the increasing percentage of A. platensis F&M-C256. The addition of 6% or more of microalgal biomass in the recipe significantly (p < 0.05) increased the amino acid content of "crostini". presents the proximate chemical composition of "crostini" incorporated with A. platensis F&M-C256 biomass. "Crostini" samples presented 7.4 to 9.3% of moisture, which is in the range of allowed moisture values for this type of dried bakery foods. For Ap 2 "crostini" there is no significant difference in comparison to the control. At or above 6% A. platensis F&M-C256 biomass addition there is a significant increase in "crostini" moisture content which might be related to a significantly (p < 0.05) higher water absorption capacity of A. platensis F&M-C256 biomass in comparison to wheat flour (approximately + 150%). A significant increase (p < 0.05) in ash content was observed for Ap 6 and Ap 10 "crostini" compared to the control (+50% and + 61%, respectively). The higher ash content could be due to the CO [bib_ref] Lactic acid bacterium and yeast microbiotas of 19 sourdoughs used for traditional/typical..., Minervini [/bib_ref] 2− and HCO 3 − residual in the cyanobacterial biomass. As expected, a higher amount of proteins corresponds to a higher amount of A. platensis F&M-C256 biomass in the product. A 44% and 69% increase of protein content between A. platensis "crostini" (Ap 6 and Ap 10, respectively) and the control was found. These values were comparable with those found in bread enriched with spirulina [bib_ref] Nutritional and physicochemical characteristics of bread enriched with microalgae Spirulina platensis, Ak [/bib_ref] [bib_ref] Vitamins, minerals, protein digestibility and antioxidant activity of bread enriched with spirulina..., Saharan [/bib_ref] and were coherent with theoretical calculation, as also reported by Batista et al. [bib_ref] Microalgae biomass as an alternative ingredient in cookies: sensory, physical and chemical..., Batista [/bib_ref] and Bolanho et al. [bib_ref] Antioxidant and nutritional potential of cookies enriched with Spirulina platensis and sources..., Bolanho [/bib_ref] on cookies with 2 and 5% A. platensis biomass. As expected, in terms of lipids content, no significant variation (p > 0.05) between A. platensis F&M-C256 "crostini" (16.5-18.2%) and the control (16.6%) was found. Carbohydrate content of A. platensis F&M-C256 "crostini", in agreement with the cyanobacterial biomass composition, ranged from 57.8 to 59.7% and was not significantly different from the control. Therefore, the newly developed "spirulina crostini" could be regarded as a very interesting protein-fortified bakery product, reaching values as high as 14-17% protein. In fact, for Ap 6 and Ap 10 "crostini", the protein content corresponds to 12.7 and 14.9%, respectively, of the total energy content of the sample. Therefore, these products can be claimed as "source of protein" considering the European Commission Regulation on nutritional claims which states that "a claim that a food is a source of protein may only be made where at least 12% of the energy value of the food is provided by protein".
In terms of energetic content, the values ranged from 429 to 450 kcal/100 g.
## "crostini" bioactive compounds and antioxidant capacity. the presence of bioactive compounds in
Arthrospira platensis biomass is associated with its antioxidant potential, immunomodulatory, anti-inflammatory and other activities [bib_ref] The antioxidant, immunomodulatory, and anti-inflammatory activities of Spirulina: an overview, Wu [/bib_ref]. Among the main bioactives, A. platensis contains phycocyanin, a blue proteic pigment composed of an open chain tetrapyrrole chromophore (phycocyanobilin) covalently attached to an apoprotein, that presents nutraceutical properties and health benefits, principally attributed to its antioxidant activity [bib_ref] Medical application of Spirulina platensis derived C-phycocyanin. Evid.-Based Compl, Liu [/bib_ref] [bib_ref] C-phycocyanin: a potent peroxyl radical scavenger in vivo and in vitro, Bhat [/bib_ref]. [fig_ref] Figure 1: Linear regression between phycocyanin content [/fig_ref] shows the linear regression between phycocyanin content of "crostini" after cooking and level of incorporation of A. platensis F&M-C256 biomass. The control "crostini" did not contain phycocyanin. As expected, an increase in the A. platensis F&M-C256 biomass content within the "crostini" corresponded to an . Acidification parameters (pH and Total Titratable Acidity, TTA), volume increase (%, ∆V/V0), microorganism concentrations (LAB and yeast) of the doughs at the end of fermentation and organic acids and free amino acid (FAA) content of "crostini". (C = control without A. platensis F&M-C256 incorporation; Ap 2, Ap 6 and Ap 10 indicate doughs and "crostini" enriched with 2, 6 and 10% (w/w) A. platensis F&M-C256 biomass; ∆: difference between the final and the initial value). Results are expressed as average ± standard deviation (n = 3). Different letters in the same column correspond to significant differences (p < 0.05). . Proximate chemical composition (g/100 g) of control (without A. platensis F&M-C256 incorporation) (C) and of "crostini" enriched with 2% (Ap 2), 6% (Ap 6) and 10% (Ap 10) (w/w) A. platensis F&M-C256 biomass. Results are expressed as average ± standard deviation (n = 9). Different letters in the same column correspond to significant differences (p < 0.05).
## Moisture
increase in phycocyanin content. Ap 2, Ap 6 and Ap 10 "crostini" contained 0.1%, 0.5% and 0.9% phycocyanin, respectively. Thus, surpisingly, about all the pigment from the microalgal biomass (which contains 8.2% phycocyanin w/w; results not shown) was still present in the cooked "crostini". Aliquots of A. platensis F&M-C256 biomass (both lyophilised and moistened with 30% water), were heated to simulate the "crostini" baking conditions. The test highlighted an important phycocyanin loss that increased with temperature and exposure time (results not shown). Since phycocyanin reduction during "crostini" baking was much lower compared to that of biomass alone, we hypothesize that some dough components might act as protecting agents against phycocyanin degradation. Further investigation to fully clarify this protective action of dough is necessary. Phenolic compounds, synthesized as secondary metabolites by plants, but also by cyanobacteria, are considered one of the most important classes of natural antioxidants [bib_ref] Production of phenolic compounds from Spirulina maxima microalgae and its protective effects, El-Baky [/bib_ref]. Phenolics, that include simple phenols, flavonoids, phenylpropanoids, tannins, lignins, phenolic acids, and their derivatives 59 , are receiving increasing interest from food manufacturers and consumers for their health benefits [bib_ref] Phenolic content and antioxidant capacity in algal food products, Machu [/bib_ref]. [fig_ref] Figure 2: Figure 2 [/fig_ref] shows the total phenolic content of "crostini". A value of 2.3 mg GAE g −1 was found in the control "crostini", very likely supplied by the extra-virgin olive oil used in the dough preparation [bib_ref] A nutrigenomics approach for the study of anti-aging interventions: olive oil phenols..., Luceri [/bib_ref]. The addition of A. platensis F&M-C256 biomass, with a total phenolic content of 10 mg GAE g −1 ± 0.1, resulted in a significant (p < 0.05) supplementation of phenolic compounds to the "crostini". Ap 10 "crostini" presented the highest phenolic content (3.4 mg GAE g −1 ). Other studies reported a higher total phenolic content of A. platensis biomass 62 , as well as a correlation with its antioxidant activity 20 . Bolanho et al. [bib_ref] Antioxidant and nutritional potential of cookies enriched with Spirulina platensis and sources..., Bolanho [/bib_ref] reported a total phenolic content for A. platensis biomass of 12 mg GAE g −1 , similar to that found in the present study. They also showed an increase of about 65% (from 1.4 to 2.3 mg GAE g −1 ) in total phenolic content in 5% A. platensis cookies when compared to the control, which is higher than the increase found in this study for Ap 6 "crostini" (26%).
The radical-scavenging capacity of A. platensis F&M-C256-enriched "crostini" was tested by the DPPH method. The incorporation of A. platensis F&M-C256 biomass led to a significant (p < 0.05) increase in the Total phenolic content (expressed as gallic acid equivalents mg g −1 dry weight) (white columns) and radical-scavenging capacity (%) (grey columns) of control (without A. platensis F&M-C256 incorporation) (C) and of "crostini" enriched with 2% (Ap2), 6% (Ap6) and 10% (Ap10) (w/w) A. platensis F&M-C256 biomass. Results are expressed as average ± standard deviation (n = 9). Different letters correspond to significant differences (p < 0.05). www.nature.com/scientificreports www.nature.com/scientificreports/ radical-scavenging capacity for 6 and 10% A. platensis F&M-C256 "crostini" compared to the control [fig_ref] Figure 2: Figure 2 [/fig_ref]. Overall, DPPH radical scavenging capacity ranged from 57% to 61%, corresponding to a VCEAC of 0.60 to 0.64 µg per mg of "crostini" (from Ap 2 to Ap 10 "crostini", respectively) [fig_ref] Figure 2: Figure 2 [/fig_ref]. The control showed a radical scavenging capacity of about 55%, indicating that ingredients in the sourdough preparation already confer a strong antioxidant capacity, that is only slightly increased by A. platensis F&M-C256 biomass addition. The results of this work are in agreement with the study conducted by Siriwardhana et al. [bib_ref] Antioxidant activity of Hizikia fusiformis on reactive oxygen species scavenging and lipid..., Siriwardhana [/bib_ref] which reported a high correlation between DPPH radical scavenging activities and total phenolic content. Other authors [bib_ref] Functional characters evaluation of biscuits sublimated with pure phycocyanin isolated from Spirulina..., El Baky [/bib_ref] [bib_ref] Optimization of a process for high fibre and high protein biscuit, Singh [/bib_ref] have studied the antioxidant capacity of A. platensis enriched bakery products (principally cookies), attributing this property to the phycobiliproteins provided by this microorganism. El Baky et al. [bib_ref] Functional characters evaluation of biscuits sublimated with pure phycocyanin isolated from Spirulina..., El Baky [/bib_ref] observed increasing antioxidant activity for biscuits containing A. platensis biomass (from 0.3 to 0.9%). Singh et al. [bib_ref] Optimization of a process for high fibre and high protein biscuit, Singh [/bib_ref] found a linear positive correlation between A. platensis concentration (from 1.6 to 8.4%) in biscuits and antioxidant activity. Our results are in agreement with the findings of El Baky et al. [bib_ref] Functional characters evaluation of biscuits sublimated with pure phycocyanin isolated from Spirulina..., El Baky [/bib_ref] and Singh et al. [bib_ref] Optimization of a process for high fibre and high protein biscuit, Singh [/bib_ref] , considering that the antioxidant capacity increases with the increasing levels of A. platensis F&M-C256 in "crostini".
The extra-virgin olive oil used in the dough preparation presumably contributed to the high antioxidant capacity of control "crostini". The addition of A. platensis F&M-C256 biomass, replacing wheat flour in the doughs, weakly increased the radical-scavenging capacity, with a more moderate effect than observed for phenols. After baking, all A. platensis F&M-C256 "crostini" still presented a high content of phycocyanin, possibly responsible, together with phenolics, for the observed antioxidant capacity.
"crostini" in vitro digestibility. The in vitro digestibility analysis reproduces the chemical-enzymatic attack that occurs in the proximal tract of the monogastric digestive system through a double enzymatic incubation with pepsin and pancreatin [bib_ref] Prediction of the total tract digestibility of energy in feedstuffs and pig..., Boisen [/bib_ref]. Most of the literature on algae digestibility focuses on seaweeds [bib_ref] Seaweed proteins: biochemical, nutritional aspects and potential uses, Fleurence [/bib_ref] [bib_ref] Edible Azorean macroalgae as source of rich nutrients with impact on human..., Paiva [/bib_ref] [bib_ref] Nutritional quality of some wild and cultivated seaweeds: Nutrient composition, total phenolic..., Tibbetts [/bib_ref] and only few studies deal with microalgae [bib_ref] In vitro digestibility of different commercial edible algae products, Machů [/bib_ref] [bib_ref] In vitro digestion of microalgal biomass from freshwater species isolated in Alberta,..., Tibbetts [/bib_ref]. This is the first work, to our knowledge, on in vitro digestibility of cyanobacteria-enriched "crostini".
The in vitro dry matter (IVD) and protein digestibility (IVPD) results are presented in . A. platensis F&M-C256 "crostini" showed lower IVD values than the control, but still above 85%. The reason could be that A. platensis F&M-C256 biomass is not completely digestible (78% ± 2.6). The dry matter digestibility of Ap 2 and Ap 6 "crostini" (88.4% and 86.8%, respectively) was slightly lower compared to the values found for 2 and 6% A. platensis F&M-C256-based cookies (94.6% and 92.1%, respectively) by Batista et al. [bib_ref] Microalgae biomass as an alternative ingredient in cookies: sensory, physical and chemical..., Batista [/bib_ref] , but still within the IVD range of commercial wheat crackers (78.3-93%).
IVPD values reflect those of dry matter digestibility . Protein digestibility is an important factor to estimate the availability of protein for intestinal absorption, which influences the efficiency of protein utilization. Many authors reported very diverse protein digestibility values (from 70 to 84%) for different species of Arthrospira 73-75 . In our study, an IVPD value of 81% ± 3.0 for A. platensis F&M-C256 biomass was found. A. . In vitro dry matter digestibility (IVD, %) (white columns) and in vitro protein digestibility (IVPD, %) (grey columns) of control (without A. platensis F&M-C256 incorporation) (C) and of "crostini" enriched with 2% (Ap2), 6% (Ap6) and 10% (Ap10) (w/w) A. platensis F&M-C256 biomass. Results are expressed as average ± standard deviation (n = 9). Different letters correspond to significant differences (p < 0.05). www.nature.com/scientificreports www.nature.com/scientificreports/ platensis F&M-C256 "crostini" showed significantly (p < 0.05) lower IVPD values than the control . This is in accordance with De Marco et al. [bib_ref] Effects of spirulina biomass on the technological and nutritional quality of bread..., Marco [/bib_ref] , which evaluated the protein digestibility of A. platensis enriched dried pasta (from 5 to 20%). The IVPD reduction at the increasing percentage of A. platensis F&M-C256 biomass addition could be due to the content of algae minor compounds (i.e. phenols or polysaccharides) that may react with gluten, forming insoluble complexes and inhibiting the activity of pepsin and pancreatin enzymes [bib_ref] Food phenolics: Sources, chemistry, effects, applications, Shahidi [/bib_ref]. However, it should be noted that the control "crostini" has 9.9% total protein , with 71% IVPD, which should result in approximately 7 g digestible protein/100 g. On the other hand, the Ap 10 "crostini" has a much higher protein content (16.7%, , so even with a lower IVPD (59%), there is still a higher digestible protein content, of approximately 10 g/100 g. "crostini" colour and appearance. A. platensis F&M-C256 "crostini" presented unusual and innovative tonalities. The different samples showed very distinct colours, as evident by visual observation [fig_ref] Figure 4: Control "crostini" [/fig_ref] and through the instrumental colour parameters analysis represented in .
Regarding luminosity (L*), there was a drastic significant (p < 0.05) decrease from the light control sample (69%) to the very dark A. platensis F&M-C256 "crostini" samples (36 to 27%). These results are in agreement with previous studies, namely A. platensis F&M-C256 cookies 20 and pasta [bib_ref] Incorporation of Chlorella vulgaris and Spirulina maxima biomass in pasta products. Part..., Fradique [/bib_ref]. In terms of chromaticity (C*), which results from the combination of a* and b* coordinates, it is clear that b* (blue/yellow intensity) parameter is dominant over a* (red/ green intensity). In fact, a* values are very close to zero (≤1), which means that green/red colour intensity is very low. On the other hand, the control sample has a b* value of 17, i.e. on the yellow range. By increasing A. platensis F&M-C256 concentration, a significant decrease in b* is observed, reaching values close to 1 for the Ap 10 "crostini". Batista et al. [bib_ref] Microalgae biomass as an alternative ingredient in cookies: sensory, physical and chemical..., Batista [/bib_ref] observed a similar behaviour for A. platensis F&M-C256 cookies, although with much higher a* and b* values and constant hue angle for cookies with 2% and 6% A. platensis F&M-C256 biomass incorporation (h = 117°). In the present study, the hue angle (h°) significantly (p < 0.05) decreases with A. platensis F&M-C256 concentration. The control sample presents an h° of 94°, which is very close to a pure yellow tonality (90°) while Ap 2 and Ap 6 "crostini" show lower values around 84-87° and Ap 10 "crostini" goes down to 77°. From these results, it seems that A. platensis F&M-C256 incorporation is not contributing to a vivid coloration but mainly to a darkening effect. Probably A. platensis F&M-C256 pigments, namely green chlorophylls, are highly degraded during the "crostini" high temperature baking and toasting procedures. As will be detailed below (section 3.5) these colour characteristics of the A. platensis "crostini" are not well appreciated by the sensory panelists.
The "crostini" characteristic dimensions were also measured (using a digital caliper) and it was observed that the surface area of the "crostini" decreased linearly (R 2 = 0.96) when increasing A. platensis F&M-C256 content (results not shown). In fact, the control "crostini" had a surface area of 10.0 ± 1.6 cm 2 , while, Ap 2, Ap 6 and Ap 10 "crostini" showed surface areas of 8.0 ± 1.4 cm 2 , 5.4 ± 1.1 cm 2 and 3.9 ± 0.7 cm 2 . This is in agreement with the lower volume increase observed for A. platensis F&M-C256 doughs . Graça et al. [bib_ref] Impact of Chlorella vulgaris on the rheology of wheat flour dough and..., Graça [/bib_ref] have also reported lower dough volume expansion and lower surface areas on bread enriched with Chlorella vulgaris. . Instrumental color parameters of control (without A. platensis F&M-C256 incorporation) (C) and of "crostini" enriched with 2% (Ap 2), 6% (Ap 6) and 10% (Ap 10) (w/w) A. platensis F&M-C256 biomass. Results are expressed as average ± standard deviation (n = 12). Different letters in the same column correspond to significant differences (p < 0.05).
## Figure 5.
Responses of the sensory analysis panel tasters (n = 35) of control (without A. platensis F&M-C256 incorporation) (C) and of "crostini" enriched with 2% (Ap 2), 6% (Ap 6) and 10% (Ap 10) (w/w) A. platensis F&M-C256 biomass. 0 -"very unpleasant"; 1 -"unpleasant"; 2 -"slightly unpleasant"; 3 -"slightly pleasant"; 4 -"pleasant"; 5 -"very pleasant". Different letters correspond to significant differences (p < 0.05). www.nature.com/scientificreports www.nature.com/scientificreports/ "crostini" sensory evaluation. Sensory analysis assays were carried out with A. platensis F&M-C256 "crostini", at 2, 6 and 10% incorporation level. represents the average scores of the sensorial parameters as evaluated by the panel. From the graph it is clear that the less appreciated samples were Ap 6 and Ap 10 "crostini" (average score 2.5 for global appreciation, "slightly unpleasant") The global sensory appreciation of Ap 2 "crostini" was considered "pleasant" (average score 3.9), similar to the control sample (average score 3.7). In fact, no significant differences (p > 0.05) were found between Ap 2 and control "crostini", regarding taste, smell or texture attributes, while colour was significantly (p < 0.05) less appreciated (average scores 2.6 and 4.0, respectively). Many authors reported the results of sensory analyses of microalgae-based food products such as bread [bib_ref] Nutritional and physicochemical characteristics of bread enriched with microalgae Spirulina platensis, Ak [/bib_ref] [bib_ref] Vitamins, minerals, protein digestibility and antioxidant activity of bread enriched with spirulina..., Saharan [/bib_ref] [bib_ref] Spirulina platensis effect as protein supplement on rheological properties of dough and..., Constantinescu [/bib_ref] [bib_ref] Evaluation of nutritional and sensory properties of bread enriched with Spirulina, Achour [/bib_ref] , pancakes [bib_ref] Nutritional and organoleptic attributes of novel pancake fortified with spirulina (Arthrospira platensis), Kumar [/bib_ref] , croissants 82 , cookies 20,54,64,65 and pasta [bib_ref] Isochrysis galbana and Diacronema vlkianum biomass incorporation in pasta products as PUFA's..., Fradique [/bib_ref] [bib_ref] Blue-green algae (Arthrospira platensis) as an ingredient in pasta: free radical scavenging..., Zouari [/bib_ref] finding that these products reported different appreciation in terms of sensorial acceptance. Bolanho et al. [bib_ref] Antioxidant and nutritional potential of cookies enriched with Spirulina platensis and sources..., Bolanho [/bib_ref] found that the addition of A. platensis biomass (2 and 5%) in the cookies decreased the sensorial acceptance when compared to the control cookie. Singh et al. [bib_ref] Optimization of a process for high fibre and high protein biscuit, Singh [/bib_ref] also reported that the addition of A. platensis (>7% incorporation level) to cookies prepared from sorghum and whole wheat flour negatively affected the textural and sensory attributes of flavour. On the contrary, El-Baky et al. [bib_ref] Functional characters evaluation of biscuits sublimated with pure phycocyanin isolated from Spirulina..., El Baky [/bib_ref] found that cookies supplemented with different levels of A. platensis biomass (0.3, 0.6 and 0.9% incorporation level) were significantly acceptable for colour, odour, flavour, texture and global appreciation. presents the answers given by the panelists in relation to the buying intention. Around 60% of the tasters "would probably buy" and 22% "would certainly buy" the Ap 2 "crostini", the most appreciated product. While, Ap 6 and Ap 10 "crostini" presented a high variability in terms of buying intention. Around 35% of panel tasters would probably buy the Ap 10 "crostini" and less than 15% gave the same response for Ap 6 "crostini". Maybe Ap 6 "crostini" did not show any distinctive character neither in terms of strong taste nor in terms of delicate taste, as confirmed by insecure tasters ("not sure", 42%). Around 25% "certainly would not buy" the Ap 6 and Ap 10 "crostini", principally due to taste and texture .
# Conclusion
The combination of Arthrospira platensis (spirulina) biomass addition and the sourdough technology led to the development of a novel microalgae-based bakery product. The addition of Arthrospira platensis F&M-C256 biomass as ingredient resulted in sourdough "crostini" with an innovative green colour. Increasing microalgae content from 2% to 10% (w/w) resulted in a significant (p < 0.05) increase in the "crostini" protein, phycocyanin and total phenolic content. Considering the European Commission Regulation on nutritional claims, "crostini" incorporated with 6% and 10% A. platensis F&M-C256 can be claimed as "source of protein". A lower value of in vitro dry matter and protein digestibility between A. platensis F&M-C256 "crostini" and the control was found, although IVD values are always above 85%. Two percent A. platensis F&M-C256 "crostini" presented the highest sensory scores and resulted the most appreciated product, even more than the control without cyanobacterial biomass. To improve the acceptance by consumers of 6% and 10% A. platensis F&M-C256 "crostini", educational marketing strategies and formulation enhancements should be considered.
This study suggests that A. platensis F&M-C256-based sourdough "crostini" represent original functional food products, with antioxidant properties (also due to the extra virgin olive oil utilization) and containing alternative source of digestible proteins. Moreover, the use of sourdough fermentation in producing A. platensis F&M-C256 "crostini" can provide an additional health value considering the enhancing of nutritional properties associated to this biotechnology process. This new product could become widely consumed by particular categories of people such as sportsmen, vegetarians, vegans and the elderly, but also by consumers interested in a product with an innovative taste and colour.
[fig] Scientific: RepoRtS | (2019) 9:19433 | https://doi.org/10.1038/s41598-019-55840-1 [/fig]
[fig] Figure 1: Linear regression between phycocyanin content (%, w/w) of control (without A. platensis F&M-C256 incorporation) (0%) and of "crostini" enriched with 2%, 6% and 10% (w/w) A. platensis F&M-C256 biomass. Results are expressed as average ± standard deviation (n = 9). [/fig]
[fig] Figure 2: Figure 2. Total phenolic content (expressed as gallic acid equivalents mg g −1 dry weight) (white columns) and radical-scavenging capacity (%) (grey columns) of control (without A. platensis F&M-C256 incorporation) (C) and of "crostini" enriched with 2% (Ap2), 6% (Ap6) and 10% (Ap10) (w/w) A. platensis F&M-C256 biomass. Results are expressed as average ± standard deviation (n = 9). Different letters correspond to significant differences (p < 0.05). [/fig]
[fig] Figure 4: Control "crostini" (without A. platensis F&M-C256 incorporation) (C) and "crostini" enriched with 2% (Ap 2), 6% (Ap 6) and 10% (Ap 10) (w/w) A. platensis F&M-C256 biomass.Scientific RepoRtS | (2019) 9:19433 | https://doi.org/10.1038/s41598-019-55840-1 [/fig]
|
LazySampling and LinearSampling: fast stochastic sampling of RNA secondary structure with applications to SARS-CoV-2
Many RNAs fold into multiple structures at equilibrium, and there is a need to sample these structures according to their probabilities in the ensemble. The conventional sampling algorithm suffers from two limitations: (i) the sampling phase is slow due to many repeated calculations; and (ii) the endto-end runtime scales cubically with the sequence length. These issues make it difficult to be applied to long RNAs, such as the full genomes of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). To address these problems, we devise a new sampling algorithm, LazySampling, which eliminates redundant work via on-demand caching. Based on LazySampling, we further derive LinearSampling, an end-to-end linear time sampling algorithm. Benchmarking on nine diverse RNA families, the sampled structures from LinearSampling correlate better with the well-established secondary structures than Vienna RNAsubopt and RNAplfold. More importantly, LinearSampling is orders of magnitude faster than standard tools, being 428× faster (72 s versus 8.6 h) than RNAsubopt on the full genome of SARS-CoV-2 (29 903 nt). The resulting sample landscape correlates well with the experimentally guided secondary structure models, and is closer to the alternative conformations revealed by experimentally driven analysis. Finally, LinearSampling finds 23 regions of 15 nt with high accessibilities in the SARS-CoV-2 genome, which are potential targets for COVID-19 diagnostics and therapeutics.
# Introduction
RNAs are involved in many cellular processes, including expressing genes, guiding RNA modification (1), catalyzing reactions [bib_ref] The chemical repertoire of natural ribozymes, Doudna [/bib_ref] and regulating diseases [bib_ref] Long noncoding RNAs: past, present, and future, Kung [/bib_ref]. Knowing the spatial structure of RNAs is one of the keys to better understand these biological processes and further harness RNAs for diagnostics and therapeutics, but determining the 3D structures is hard and expensive. Alternatively, RNA secondary structure is helpful for revealing the structurefunction relationship, and can be used for inferring the 3D structure [bib_ref] Automated RNA tertiary structure prediction from secondary structure and low-resolution restraints, Seetin [/bib_ref]. Therefore, being able to rapidly and accurately predict RNA secondary structures is desired.
Commonly, the minimum free energy (MFE) structure is predicted [bib_ref] Optimal computer folding of large RNA sequences using thermodynamics and auxiliary information, Zuker [/bib_ref] , but these methods do not capture the fact that many RNAs fold into multiple structures at equilibrium [bib_ref] Rich RNA structure landscapes revealed by mutate-and-map analysis, Cordero [/bib_ref] [bib_ref] A decade of riboswitches, Serganov [/bib_ref]. To address this issue, Ding and Lawrence [bib_ref] A statistical sampling algorithm for RNA secondary structure prediction, Ding [/bib_ref] pioneered stochastic sampling, which samples secondary structures according to their probabilities in the ensemble (see [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref]. These samples are not only informative for describing or probing the ensemble, but are also useful for downstream applications. First, we can model the multiple conformations based on the sampled structures, and further cluster them into representative structures [bib_ref] RNA secondary structure prediction by centroids in a Boltzmann weighted ensemble, Ding [/bib_ref]. For example, several studies used sampled structures, together with SHAPE mapping data, to model and cluster the conformations of the frameshifting region in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [bib_ref] Comprehensive in vivo secondary structure of the SARS-CoV-2 genome reveals novel regulatory..., Huston [/bib_ref] [bib_ref] Secondary structural ensembles of the SARS-CoV-2 RNA genome in infected cells, Lan [/bib_ref]. Also, we can use a set of samples to estimate the end-to-end distance of an RNA [bib_ref] ) mRNAs and lncRNAs intrinsically form secondary structures with short end-to-end distances, Lai [/bib_ref]. In addition, we often want to predict the probability that a region is completely unpaired, known as the accessibility of that region, which plays an important role in small interfering RNA (siRNA) drug design [bib_ref] The efficacy of small interfering RNAs targeted to the type 1 insulin-like..., Bohula [/bib_ref] [bib_ref] The impact of target site accessibility on the design of effective siRNAs, Tafer [/bib_ref] [bib_ref] Efficient siRNA selection using hybridization thermodynamics, Lu [/bib_ref]. Accessibility can be easily estimated from the fraction of sampled structures for which the region is completely unpaired (see [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref]. However, widely used as it is, the standard sampling algorithm suffers from two main limitations. First, the sampling phase is slow due to many repeated and redundant calculations, wasting a substantial amount of time especially for large sample sizes. Note that our notion of 'redundant' is unrelated to the one in 'non-redundant sampling', which is a variant to output a collection of unique structures, while standard sampling can sample the same structure more than once. Second, its end-to-end runtime scales cubically with sequence length as it has to calculate the partition function using the standard McCaskill algorithm [bib_ref] The equilibrium partition function and base pair probabilities for RNA secondary structure, Mccaskill [/bib_ref] before sampling. Therefore, both the sampling and partition function phases scale superlinearly with the sequence length. This makes it difficult to be applied to long sequences, such as the full-genome of SARS-CoV-2.
To alleviate these issues, we present three innovations, each built upon the previous one, that together linearize both the sampling and partition function phases to achieve end-to-end linear runtime. The first idea, Full-Saving Sampling, eliminates all redundant calculations in the sampling phase by saving all computations during the partitionfunction phase. The second idea, LazySampling, is an even smarter version that only saves computations that are needed during the sampling phase on the fly, greatly reducing the amount of runtime and memory used for saving. These two ideas are based on the observation that node visits are highly unbalanced: most nodes are never visited while a small fraction of 'important' nodes are visited repeatedly. Finally, we further improve LazySampling by replacing its O(n 3 )-time partition function calculation by our recently proposed O(n)-time approximate algorithm, LinearPartition [bib_ref] LinearPartition: linear-time approximation of RNA folding partition function and base-pairing probabilities, Zhang [/bib_ref]. This combination gives rise to LinearSampling, an end-to-end linear time sampling algorithm that is orders of magnitude faster than the standard one (see for speedup details).
More importantly, in order to fight current and future pandemics, it is of great value to study the SARS-CoV-2 genome structure and find the regions with high accessibilities, which can be potentially used for coronavirus disease (COVID) diagnostics and therapeutics. However, the conventional tools, including the traditional sampling tools, and other methods to estimate accessibility, e.g. constrained partition function (forcing each region of interest to be fully unpaired, and computing the fraction of the resulting constrained partition function over the global one) and direct computation [bib_ref] Thermodynamics of RNA-RNA binding, Mückstein [/bib_ref] [bib_ref] RNA accessibility in cubic time, Bernhart [/bib_ref] , all run in at least O(n 3 ) time and are too slow on such long sequences with the consideration of global, long-range base pairs. In contrast, Lin-earSampling can easily scale up to the whole SARS-CoV-2 genomes without local window constraints, and sample 10 000 structures in only 72 s, achieving 428× speedup (72 s versus 8.6 h) compared with RNAsubopt, a widely used sampling tool in the ViennaRNA package [bib_ref] ViennaRNA Package 2.0. Algorithm, Lorenz [/bib_ref]. We confirm that LinearSampling-derived accessibilities correlate well with the experimentally guided structures [bib_ref] Comprehensive in vivo secondary structure of the SARS-CoV-2 genome reveals novel regulatory..., Huston [/bib_ref] , resulting in 23 regions of 15 nt with high accessibilities, which are potential targets of diagnostics and drug design.
In addition, the lazy saving strategy can also be applied to other stochastic sampling tasks, such as the Gibbs centroid sampling [bib_ref] The Gibbs centroid sampler, Thompson [/bib_ref] and structural alignment sampling [bib_ref] Stochastic sampling of the RNA structural alignment space, Harmanci [/bib_ref]. Also, LinearSampling can load partition function (hypergraph nodes) dumped from LinearTurboFold (28), a tool for simultaneous folding and alignment of RNA homologs, which integrates homologous information into the sampled structures to identify conserved and accessible regions.
# Materials and methods
To facilitate our discussion on sampling, we first formulate the search space of RNA folding using the framework of (directed) hypergraphs [bib_ref] Directed hypergraphs and applications, Gallo [/bib_ref] [bib_ref] Computation of biopolymers: a general approach to different problems, Finkelstein [/bib_ref] which have been used for both the closely related problem of context-free parsing [bib_ref] Better k-best parsing, Huang [/bib_ref] and RNA folding itself [bib_ref] Computation of biopolymers: a general approach to different problems, Finkelstein [/bib_ref] [bib_ref] A combinatorial framework for designing (pseudoknotted) RNA algorithms, Ponty [/bib_ref]. This formulation makes it possible to present the standard sampling algorithm as top-down stochastic backtracing that is mirror-symmetric to the bottom-up partition function computation. Then we present our two novel sampling algorithms which eliminate redundant computations by saving and on-demand caching, respectively. Finally, we present our LinearSampling algorithm which is the first sampling algorithm to run in end-to-end linear time.
## Hypergraph framework
For RNA sequence x = x 1 ... x n , we formalize its folding space as a hypergraph H = V, E . Each node v is a subsequence (i.e. span) x i. j = x i ...x j , and each hyperedge e is a pair node(e), subs(e) representing a decomposition of node(e) into a list of children nodes subs(e), e.g. x i, j , [x i,k , x k+1, j ] divides one span into two smaller ones. For each node v, we define INEDGES(v) = {e | node(e) = v} to be its set of incoming hyperedges, i.e. all decompositions of v.
Hyperedges with one and two child(ren) nodes are called unary and binary hyperedges, respectively. In order to recursively assemble substructures to form the global structure, each hyperedge e = v, subs is associated with a combination function f (e) : S |subs(e)| → S that assembles substructures from subs into a structure for v (here S is the set of dot-bracket strings that represent RNA secondary structure). Each hyperedge e is associated with an (extra) energy term w(e) ∈ R.
We take the classical Nussinov algorithm (8) as a concrete example, which scores secondary structures by counting the number of base pairs. The nodes include singleton spans x i,i that are terminals, and decomposable spans x i, j (1 ≤ i < j ≤ n) that can be decomposed in three ways depending on base x j :
- x j is unpaired, which corresponds to a unary hyperedge
[formula] x i, j , [x i, j −1 ] [/formula]
with the combination function f 1 (a) = "a . " and energy term 0:
[formula] x i, j x i, j −1 . i j - x j pairs with x i (if allowed), denoted by a unary hyper- edge x i, j , [x i +1, j −1 ] [/formula]
with combination function f 1 (a) = "(a)" and energy term −1:
[formula] x i, j ( x i +1, j −1 ) i j - x j pairs with some x k (i < k < j) [/formula]
, dividing x i, j into two smaller spans x i,k−1 and x k+1, j −1 :
[formula] x i, j x i,k−1 ( x k+1, j −1 ) i k j [/formula]
These bifurcations correspond to a set of binary hyperedges with combination function f 2 (a, b) = "a(b)" and energy term −1: [fig_ref] Figure 2: Top [/fig_ref] for an illustration. This framework can easily extend to other folding algorithms such as Zuker [bib_ref] Optimal computer folding of large RNA sequences using thermodynamics and auxiliary information, Zuker [/bib_ref] and LinearFold [bib_ref] LinearFold: linear-time approximate RNA folding by 5 -to-3 dynamic programming and beam..., Huang [/bib_ref] , where nodes are 'labeled spans' such as C i, j for substructures over x i, j with (x i , x j ) paired, M i, j for multiloops over x i, j , etc.
[formula] BINARY(x i, j ) = i < k < j (x k ,x j ) pair { x i, j , [x i,k−1 , x k+1, j −1 ] } See [/formula]
## Classical sampling algorithm under a hypergraph framework
All sampling algorithms consist of the (bottom-up) partition function calculation phase and the (top-down) stochastic backtracing phase which is mirror-symmetric to the former.
The partition function phase. In this bottom-up phase, we first calculate the local partition function Z(v) of each node v (see Supplementary [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref] complexity remains O(n 3 ), but in practice the overhead for saving (line 8) is quite costly and it may run out of memory (see [fig_ref] Figure 4: Runtime and memory comparisons against sequence length on the RNAcentral dataset [/fig_ref].
Classical 'Non-Saving' Sampling. In the sampling phase, the Non-Saving Sampling algorithm (see Supplementary [fig_ref] Figure 2: Top [/fig_ref] recursively and stochastically backtraces from the goal node, in the exact opposite direction to the bottomup partition function phase (see [fig_ref] Figure 2: Top [/fig_ref]. When visiting a node v, it tries to sample a hyperedge e from v's incoming hyperedges INEDGES(v) according to the probability Z(e)/Z(v). This is done by first generating a random number p between 0 and Z(v), and then gradually recovering each incoming hyperedge e, accumulating its Z(e) to a running sum s, until s exceeds p, at which point that current hyperedge e is chosen as sampled. Note that this algorithm in general does not need to recover all incoming hyperedges of v (see 'partial recovery of hyperedges' in Our presentation resembles the original Ding and Lawrence [bib_ref] A statistical sampling algorithm for RNA secondary structure prediction, Ding [/bib_ref] algorithm, but is simpler and cleaner thanks to the hypergraph framework and the mirror symmetry between the bottom-up and top-down phases. In contrast, Ding and Lawrence's recurrences for the two phases are different in nature (see [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref] of their paper), which results in unnecessarily complicated implementations [see Vienna RNAsubopt for an example; RNAstructure's (34) sampling is much closer to our non-saving version, except for being non-recursive]. Ponty [bib_ref] Efficient sampling of RNA secondary structures from the Boltzmann ensemble of low-energy, Ponty [/bib_ref] also exploits this symmetry both in theory and in simulations, but his analysis is for the special case under the simplified Nussinov model. Our work is on the full Turner model [bib_ref] Incorporating chemical modification constraints into a dynamic programming algorithm for prediction of..., Mathews [/bib_ref] , being the first to formulate . Non-Saving Sampling needs to repeatedly recover hyperedges in each sample (most of the node visits are non-unique; see ␣ k ), whereas Full-Saving Sampling does not need to recover any hyperedge thanks to storing all of them during the bottom-up phase, but rather oversaves (only less than half of the nodes are visited during the first five samples; see  k ; in fact, six nodes {x 2,5 , x 3,4 , x 3,5 , x 4,5 , x 5,5 , x 4,4 } are never visited even when k → ∞ as they are not reachable from the goal node x 1,5 ). LazySampling is a smart hybrid between the two, using on-demand caching, thus saving only a small fraction of hyperedges. general sampling (Nussinov, Zuker, LinearFold, etc.) under a unified framework.
## Idea 1: full-saving sampling
One observation is that Non-Saving Sampling wastes time on recovering hyperedges during the sampling phase. First, due to the symmetry, all hyperedges recovered in the sampling phase have already been traversed during the inside phase. To make things worse, a great number of hyperedges are recovered multiple times across different samples because, whenever a node is (re-)visited, its hyperedges need to be re-recovered. This situation worsens with the sample size k. More formally, we define the 'unique visit ratio' among k samples,
[formula] α k = # of unique nodes visited in k samples #of all node visits in k samples(1) [/formula]
which we will see in [fig_ref] Figure 5: In practice, most node visits are repeated, and only a small portion... [/fig_ref] to be extremely small, quickly approaching 0% as k increases, meaning most node visits are repeated ones. Actually, all hyperedges could have been saved during the inside phase, thus alleviating the need to recover hyperedges during sampling. Therefore, we devise a new algorithm, Full-Saving Sampling (see Supplementary Figures S2 and S3 for pseudocode). For each node v, for each of its incoming hyperedges e, we save Z(e), which is e's contribution to the local partition function Z(v), once and for all. Then the sampling phase is easier, only requiring sampling a hyperedge e according to its relative contribution (or 'weight') to v, i.e. Z(e)/Z(v) (line 2 in Supplementary [fig_ref] Figure 3: Illustrations of the three sampling algorithms [/fig_ref]. See Supplementary Sec. 1.2 for the detailed complexity analysis of generating each sample.
## Idea 2: lazysampling = lazy-saving sampling
Though Full-Saving Sampling avoids all re-calculations, it costs too much more space [O(n 3 ) versus O(n 2 )] and significantly more time in practice for saving the whole hypergraph. Actually, the vast majority of nodes are never visited during the sampling phase even for a large sample size. To quantify this, we define the 'visited ratio' to be:
[formula] β k = # of unique nodes visited in k samples # of all nodes in hypergraph(2) [/formula]
Our experiments in [fig_ref] Figure 5: In practice, most node visits are repeated, and only a small portion... [/fig_ref] show that only <0.5% of all nodes in the hypergraph are ever visited for 20 000 samples of a 3048 nt sequence (URS0000D5C703 9606) using Vienna RNAsubopt, so most of the saving is indeed wasted. In other words, node visits are greatly unbalanced, i.e. a small portion of nodes are repeatedly visited, while most nodes are barely visited. Based on this observation, we devise an even smarter algorithm, LazySampling, which is a hybrid between Non-Saving and Full-Saving Samplings (see Supplementary [fig_ref] Figure 4: Runtime and memory comparisons against sequence length on the RNAcentral dataset [/fig_ref]. By 'lazy' we mean only recovering and saving a node v's hyperedges when needed, i.e. the first time v is visited during the sampling phase. In this way, each hyperedge is recovered at most once, and most are not recovered at all. This version balances between space and time, and is the fastest among the three versions in most settings in practice. The complexity analysis of LazySampling is also a hybrid between the Non-and Full-Saving versions, combined together using the ␣ k and  k ratios. We note that the sampling time of LazySampling consists of two parts: (i) the hyperedge recovery (and saving) work and (ii) the sampling . Time complexities of four sampling algorithms (n is the sequence length, k is sample size, ␣ k and  k are the 'unique visit' (Eq. 1) and 'visited' (Eq. 2) node ratios, respectively, and b is beam size)
[formula] Classical Non-Saving Sampling O(n 3 ) O(n 2 ) O(kn 2 ) O(knlog n) O(n) Idea 1: Full-Saving Sampling O(n 3 ) O(n 3 ) O(knlog n) O(kn) O(n) Idea 2: LazySampling O(n 3 ) O(n 2 ) O(␣ k kn 2 + knlog n) O(␣ k knlog n + kn) O( k n 3 ) Idea 3: LinearSampling O(nb 2 ) O(nb) O(␣ k knb + knlog b) O(␣ k knlog b + kn) O( k nb 2 ) [/formula]
The runtime of LazySampling is a hybrid between those of the Non-and Full-Saving ones, i.e. T S lazy = α k T S non (hyperedge recovery) + T S full (backtracing) (Eq. 3). LinearSampling extends LazySampling by replacing the exact partition function with LinearPartition, and achieves end-to-end linear runtime. Note that the sampled derivations are mostly balanced as the depth of derivation scales O(log n) in practice (see Supplementary , therefore the complexity of the average case is close to that of the best case. (backtracing) work after relevant hyperedges are recovered. Part (i) resembles Non-Saving Sampling, but with a ratio of ␣ k , because only a tiny fraction of node visits are unique, and hyperedge recovery is only performed at the first visit to each node. Part (ii) is identical to Full-Saving Sampling (in both cases, all needed hyperedges are already saved). Therefore, we have the following relationships among the time complexities for the sampling phase of these three versions:
[formula] T S lazy (n, k) = α k T S non (n, k) hyperedge recovery + T S full (n, k) backtracing(3) [/formula]
This holds for both the worst-and best-case scenarios in Table 1. The space complexity is easier: LazySampling saves only a fraction ( k ) of all nodes in the hypergraph, thus O( k n 3 ). See for summary.
## Idea 3: linearsampling = lazysampling + linearpartition
LazySampling is the most efficient among all three sampling algorithms above, but the end-to-end runtime still scales O(n 3 ) due to the partition function computation. To address this problem, we use our recently proposed lineartime approximate algorithm, LinearPartition [bib_ref] LinearPartition: linear-time approximation of RNA folding partition function and base-pairing probabilities, Zhang [/bib_ref] , to replace the conventional cubic-time algorithm. It can be used with any of the three sampling algorithms and, in particular, we name the combination, 'LinearPartition + LazySampling', the LinearSampling algorithm as it is the fastest among all combinations and achieves end-to-end linearity.
Supplementary [fig_ref] Figure 5: In practice, most node visits are repeated, and only a small portion... [/fig_ref] describes a simplified pseudocode using the Nussinov-Jacobson energy model. In-spired by LinearPartition, we employ a beam search to prune out nodes with the small partition function (line 11) during the inside phase. So at each position j, only the top b promising nodes 'survive' [i.e. O(nb) nodes survive in total]. Here the beam size b is a user-specified hyperparameter, and the default b = 100 is found to be more accurate for structure prediction than an exact search [bib_ref] LinearPartition: linear-time approximation of RNA folding partition function and base-pairing probabilities, Zhang [/bib_ref]. See Supplementary Sec. 1.3 for the detailed complexity analysis of LinearSampling.
# Results
## Efficiency and scalability
We benchmark the runtime and memory usage on 36 sequences sampled from RNAcentral (37), ranging from 100 to 27 985 nt, which we refer to as the RNAcentral dataset; see [fig_ref] Table 2: Accessibility defect comparisons between in silico predictions [/fig_ref] for statistics. We use a Linux machine (CentOS 7.7.1908) with 2.30 GHz Intel Xeon E5-2695 v3 CPU and 755 GB memory, and gcc 4.8.5.
Comparing Non-Saving, Full-Saving and LazySampling. [fig_ref] Figure 4: Runtime and memory comparisons against sequence length on the RNAcentral dataset [/fig_ref] compares the three sampling strategies under the exact partition function calculation. Non-Saving Sampling and LazySampling have identical partition function runtime, which are both faster than Full-Saving Sampling, benefiting from saving no hyperedges. However, for the sampling phase, Non-Saving has to repeatedly recover hyperedges, resulting in a runtime increase in samplingonly time. LazySampling avoids this cost and reduces the runtime to less than half of Non-Saving's, leading Note that  k is higher in LinearSampling than in RNAsubopt, mainly because the denominator scales quadratically with sequence length n in RNAsubopt, while it scales linearly in LinearSampling; see for the analysis of partition function space complexity (which is related to the number of all nodes in the hypergraph) of different systems. Here n = 3048 nt. Supplementary [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref] demonstrates the trend with sequence length, and Supplementary [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref] shows that most of the visits are concentrated on a few nodes.
to a similar performance as Full-Saving at sampling-only time. Regarding end-to-end runtime, which combines the partition function and the sampling phases, LazySampling is the fastest and Full-Saving is the slowest. For memory usage, Full-Saving uses much more memory, while the other two are close. Supplementary shows the comparisons against sample size k, which confirms that LazySampling is the best, especially when k is large.
We also illustrate why LazySampling is a better strategy. [fig_ref] Figure 5: In practice, most node visits are repeated, and only a small portion... [/fig_ref] shows the unique visit ratio, ␣ k , is <5% when k >1000 for both RNAsubopt and LinearSampling, confirming that LazySampling is able to avoid a large number of recalculations during the sampling phase. On the other hand, the visited ratio  k [fig_ref] Figure 5: In practice, most node visits are repeated, and only a small portion... [/fig_ref] is always smaller than 0.5% and 3% for RNAsubopt and LinearSampling, respectively, and grows slower and slower as the sample size increases, suggesting that saving all hyperedges (i.e. Full-Saving) is not ideal. Supplementary [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref] further confirms that both ␣ k and  k do not increase with sequence length; therefore, this analysis applies to both short and long sequences.
Comparing LinearSampling with Vienna RNAsubopt global and local modes. We compare the efficiency and scalability between LinearSampling and Vienna RNAsubopt (version 2.4.14) in . RNAsubopt (-p mode) implemented the conventional sampling algorithm based on the exact partition function, and it enables local sampling (--maxBPspan mode) by limiting the sequence distance between paired nucleotides. confirms that Lin-earSampling scales end-to-end linearly against sequence length n, and is much faster than RNAsubopt, which has an empirical runtime of O(n 2.8 ). For 50 000 samples, LinearSampling is 208× faster (2 m versus 7 h) than RNAsubopt on the longest sequence of the RNAcentral dataset (URS0001BDA28C 9606, 27 985 nt), and 251× faster on the SARS-CoV-2 sequence (29 903 nt). Regarding sampling-only runtime, LinearSampling is >3× faster. confirms that the memory usage of LinearSampling is also linear, but RNAsubopt requires O(n 2 ) memory. LinearSampling uses 1 GB for SARS-CoV-2, while RNAsubopt needs 20 GB.
It is surprising that RNAsubopt local mode has an empirical complexity of O(n 3.3 ) for end-to-end runtime, and is even slower than its global mode. For a 9211 nt sequence (URS00009BB84A 10090), RNAsubopt local mode, using a span of 70, takes 83.8 min, and its global mode takes 18.7 min, while LinearSampling only takes 34.9 s. Memory-wise, RNAsubopt local mode uses as much as its global mode.
We also observe that RNAsubopt turns to overflow on long sequences, making it less reliable. For example, RNAsubopt overflows on four sequences in the RNAcentral dataset, shown in with open triangles whose endto-end and sampling-only runtime drop unreasonably. resulting in an abnormal exit, with only a few structures generated. Though a self-adapted partition function scale (named pf scale) is used in RNAsubopt, overflow is still unavoidable for some long sequences. In contrast, following LinearPartition, LinearSampling uses log-space for the partition function, which does not have overflow issues.
LinearSampling benefits from both LinearPartition and LazySampling. investigates the runtime of Non-Saving Sampling and LazySampling under the linear partition function, and compares them with LinearPartition with sequence length up to ∼30 000 nt. We observe that the partition function phase is fast and no longer the bottleneck, while the classical Non-Saving Sampling is 5-6× slower than LinearPartition, and takes the majority of endto-end runtime. LazySampling enjoys ∼3.5× speedup from the Non-Saving Sampling and substantially reduces the sampling runtime to roughly the same as LinearPartition's runtime.
We further compare LazySampling with Non-Saving Sampling and RNAsubopt under the exact partition function. RNAsubopt uses the non-saving strategy, but adopts an optimization called the 'Boustrophedon' method [bib_ref] Efficient sampling of RNA secondary structures from the Boltzmann ensemble of low-energy, Ponty [/bib_ref] , which makes it slightly faster than our Non-Saving Sampling. However, it is clear that LazySampling's speedup is more significant.
## Correlation with well-established structures across diverse families
We use the ArchiveII dataset [bib_ref] Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure, Mathews [/bib_ref] [bib_ref] Exact calculation of loop formation probability identifies folding motifs in RNA secondary..., Sloma [/bib_ref] to investigate the correlation between sampled structures and the ground truth structures. This dataset contains diverse families of RNA sequences and their well-established secondary structures, and each family has a reliable source. We follow the preprocessing steps of a previous study [bib_ref] LinearPartition: linear-time approximation of RNA folding partition function and base-pairing probabilities, Zhang [/bib_ref] , and obtain a subset of 2859 sequences distributed in nine families (see .
We investigate the sampled structure's correlation with the ground truth using 'ensemble defect' [bib_ref] Nucleic acid sequence design via efficient ensemble defect optimization, Zadeh [/bib_ref] , the expected number of incorrectly predicted nucleotides over the ensem- ble. It is defined as:
[formula] (S, y * ) = 1 |S| y∈S d(y, y * ) = |y * | − 2 (i, j )∈pairs(y * ) p i, j (S) − j ∈unpaired(y * ) q j (S) [/formula]
where y * is the ground truth structure, and d(y, y * ) is the distance between y and y * , defined as the number of incorrectly predicted nucleotides in y. p i, j (S) is the probability of nucleotide i pairing with nucleotide j in sample S, which can be easily calculated as the number of (i, j) pairs divided by sample size. Similarly, q j (S) is the probability of j being unpaired in the sample S, i.e. q j (S) = 1 − p i, j (S). For this database of non-coding RNAs with well-defined structures, we expect that a lower ensemble defect indicates better structure prediction accuracy. [fig_ref] Figure 8: acc [/fig_ref] shows the ensemble defect differences between LinearSampling and RNAsubopt on each family and overall. Note that the lower ensemble defect suggests better correlation with the ground truth structures. For short families, the differences between LinearSampling and RNAsubopt are either 0 or close to 0, indicating that the sampling qualities of the two systems are similar on these families. However, on the longer families (i.e. 16S and 23S rRNAs), LinearSampling has lower ensemble defects, showing that it performs better on longer sequences. The only family for which LinearSampling performs worse is tmRNA. We also present the results of RNAsubopt local mode, with base pair length limitations of 70 and 150. RNAsubopt local mode does not have a default span size; we choose 70 following the default setting in RNAplfold [bib_ref] Local RNA base pairing probabilities in large sequences, Bernhart [/bib_ref] , and 150 since it is the largest default limit in the local folding literature and software. It is obvious that the local sampling has a much higher (worse) ensemble defect on 23S rRNA, which is probably caused by not predicting the known base pairs beyond the maximum span limit.
An important application of the sampling algorithm is to calculate a region's accessibility. Therefore, we follow Ding and Lawrence [bib_ref] A statistical sampling algorithm for RNA secondary structure prediction, Ding [/bib_ref] , calculating accessibilities of window size 4 from structures generated by LinearSampling and RNAsubopt, as well as directly from RNAplfold, which implemented a dynamic programming algorithm for calculating accessibility in cubic time [bib_ref] RNA accessibility in cubic time, Bernhart [/bib_ref] , and compare them based on the ground truth structures. We denote the measurement of accessibility defect as D, which evaluates the averaged incorrect predictions of the accessibility to the ground truth given a window size. For sampling-based methods, D(S, y * ) is generated from the samples S and is defined as: for length and identity information of each family. [fig_ref] Figure 8: acc [/fig_ref] compares the accessibility defect of Lin-earSampling with RNAsubopt, RNAplfold and their local modes on the ArchiveII dataset. We observe that Lin-earSampling outperforms (or is as good as) all others on seven out of nine families, and is the best overall. Notably, both RNAsubopt's and RNAplfold's local modes are worse than their global modes, with only one exception on Group I Intron, indicating that the local modes are less accurate.
It is worth noting that the lower ensemble and accessibility defects of LinearSampling are inherited from LinearPartition, which was shown to be better correlated with the ground truth structures [bib_ref] LinearPartition: linear-time approximation of RNA folding partition function and base-pairing probabilities, Zhang [/bib_ref] by pruning out structures with low probabilities. To confirm this, we investigate the rootmean-square deviation (RMSD) between the base pairing probability matrices p(S), which is derived from the sample set S, and p , which is generated by Vienna RNAfold or LinearPartition. The results are shown in Supplementary [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref] , suggesting that structures generated by Lin-earSampling strictly match with the ensemble distribution of LinearPartition. Since LinearPartition deviates from the exact partition function of RNAfold with a small margin, the folding landscape structures of LinearSampling have different sampling probabilities compared with RNAsubopt. In most cases, structures with lower folding free energies may have higher probabilities in LinearSampling than in RNAsubopt, while low-probability structures in RNAsubopt may have 0 or close to 0 probabilities in LinearSampling.
## Applications to sars-cov-2
The COVID-19 pandemic swept the world in 2020-2022, and is likely to threaten global health for a long time. Therefore, it is valuable to study the structure of the SARS-CoV-2 genome, which helps us better understand biological processes such as virus replication and translation. Now with the significant speedup, LinearSampling is able to quickly sample structures for the whole genome of SARS-CoV-2, providing an efficient tool for SARS-CoV-2 structural studies. In this section, we run LinearSampling, as well as the conventional tools, on NC 0405512.2, the reference sequence of SARS-CoV-2 (42), investigate the multiple conformations and accessible regions of great interest and show that LinearSampling's structures correlate better with the . The correlation between sampled structures and DRACOinferred structures of the SARS-CoV-2 3 -UTR. LinearSampling and RNAsubopt are represented in solid bars and striped bars, respectively; the two DRACO-inferred conformations are in blue (conformation A) and red (conformation B), respectively. The normalized structural distance of sampled structures against DRACO-inferred conformations are grouped into 10 bins, where the distance of 0 means a perfect match of the whole structure, and the distance of 1 means that all positions in the region are incorrectly predicted. The numbers of sampled structure in each bin are illustrated. The majority of samples from LinearSampling are distributed in bins with smaller structural distance compared with RNAsubopt for both conformations, suggesting that LinearSampling has a better ability to model alternative conformations. Here the sample size is 10 000. experimental models. Then we report the LinearSampling's findings on new targets with high accessibility, which can be potentially used for COVID-19 diagnostics and drug design.
Alternative conformations and base pairs. Experimental studies on SARS-CoV-2 have inferred that multiple conformations exist. For example, DRACO analysis (43) suggests two alternative structures in the 3 -untranslated region (UTR) of SARS-CoV-2 based on the clustering of DMS-MaPseq data. To see if LinearSampling's structures are closer to this experimental analysis, we calculate the normalized structural distance (i.e. the number of incorrectly predicted nucleotides normalized by sequence length) between different algorithms and the two conformations, where the lower (normalized) structural distance indicates better correlation with the DRACO results. shows the number of sampled structures derived from LinearSam-pling and RNAsubopt against normalized structural distance, grouped into 10 bins from the lowest (best) to the highest (worst) of the distances. It is clear that the distribution of distance between LinearSampling and conformation A of the DRACO analysis (solid blue bars) shifts to the direction of lower structural distance compared with RNAsubopt's (striped blue bars). Specifically, for conformation A, there are 7023 among 10 000 of LinearSampling's structures in the bin of [bib_ref] Computation of biopolymers: a general approach to different problems, Finkelstein [/bib_ref] , suggesting that 70% of Lin-earSampling's structures are similar to conformation A. In contrast, the majority of RNAsubopt's sample are in the bins that have structural distance >50%, suggesting that the structures of RNAsubopt are similar to neither conformation A nor conformation B. Therefore, LinearSampling's sample correlates better with the DRACO analysis.
Besides the DRACO analysis, COMRADES analysis provides experimental data for short-and long-range basepairing interactions [bib_ref] The short-and long-range RNA-RNA interactome of SARS-CoV-2, Ziv [/bib_ref] , which can be used to verify the quality of the folding landscape samples regarding the base pairs in different conformations of SARS-CoV-2 genomes. We observed that LinearSampling's oracle structure, compared with that of RNAsubopt, has more base pairs matched with COMRADES analysis with varying sample sizes (Supplementary [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref]. Regarding the full sample landscapes, LinearSampling has a higher ratio of matched base pairs to predicted ones, and its total number of unmatched base pairs is substantially smaller than RNAsubopt against different sample sizes (Supplementary [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref] and B), indicating that LinearSampling has a higher hit rate on the COMRADES-inferred base pairs. The COMRADES data also provides the chimeric reads of each interaction region, where a larger number of chimeric reads indicates high credibility of interaction, and we confirm that LinearSampling has higher average chimeric reads per predicted base pairs for all tested sample sizes (Supplementary [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref]. On the other hand, RNAsubopt has higher coverage of COMRADES base pairs, but costs much longer time (Supplementary [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref]. The difference in sample landscape between LinearSampling and RNAsubopt is illustrated in a Venn diagram (Supplementary [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref]. LinearSampling tends to sample less diverse base pairs, which leads to the benefit of obtaining a higher ratio of matched base pairs and the cost of missing some base pairs shown in the COMRADES analysis; while RNAsubopt tends to sample more diverse base pairs, which has higher coverage but also results in more base pairs that are not supported by experimental data.
Accessibility. We investigate if the accessibilities predicted by LinearSampling match better with the experimentally guided structures, especially on the regions of interest, e.g. the 5 -UTR which has conserved structures and plays a critical role in viral genome replication [bib_ref] Coronavirus cis-acting RNA elements, Madhugiri [/bib_ref]. A number of studies report structural probing data for SARS-CoV-2 genomes [bib_ref] Comprehensive in vivo secondary structure of the SARS-CoV-2 genome reveals novel regulatory..., Huston [/bib_ref] [bib_ref] Secondary structural ensembles of the SARS-CoV-2 RNA genome in infected cells, Lan [/bib_ref] [bib_ref] Genome-wide mapping of SARS-CoV-2 RNA structures identifies therapeutically-relevant elements, Manfredonia [/bib_ref] [bib_ref] Genome-scale deconvolution of RNA structure ensembles, Morandi [/bib_ref] [bib_ref] In vivo structural characterization of the SARS-CoV-2 RNA genome identifies host proteins..., Sun [/bib_ref] [bib_ref] Genomic RNA elements drive phase separation of the SARS-CoV-2 nucleocapsid, Iserman [/bib_ref] , including the selective 2hydroxyl acylation analyzed by primer extension (SHAPE) and dimethyl sulfate (DMS) data. These well-accepted experimental techniques for RNA secondary structure probing have been shown to be able to improve the accuracy of structure prediction [bib_ref] Selective 2'-hydroxyl acylation analyzed by primer extension (SHAPE): quantitative RNA structure analysis..., Wilkinson [/bib_ref] [bib_ref] Accurate SHAPE-directed RNA structure determination, Deigan [/bib_ref] [bib_ref] Quantitative dimethyl sulfate mapping for automated RNA secondary structure inference, Cordero [/bib_ref]. Therefore, we utilize two SHAPE-directed structures [bib_ref] Comprehensive in vivo secondary structure of the SARS-CoV-2 genome reveals novel regulatory..., Huston [/bib_ref] [bib_ref] Genome-wide mapping of SARS-CoV-2 RNA structures identifies therapeutically-relevant elements, Manfredonia [/bib_ref] to evaluate different systems, i.e. LinearSampling, RNAsubopt and RNAplfold. [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref] compares the accessibilities derived from the three systems with the experimentally guided structures based on SHAPE reactivities [bib_ref] Comprehensive in vivo secondary structure of the SARS-CoV-2 genome reveals novel regulatory..., Huston [/bib_ref]. Following the previous study [bib_ref] A statistical sampling algorithm for RNA secondary structure prediction, Ding [/bib_ref] , the accessibilities of window size 4 are visualized in the top sub-figure, and LinearSampling clearly correlates better with the SHAPE-directed structure. For example, RNAsubopt and RNAplfold (span = 150) overestimate the accessibilities around the double-stranded region , and RNAplfold (both span = 150 and span = 800) overestimates in the region ; in contrast, LinearSampling's predictions are close to 0. Also, LinearSampling correctly captures the accessible region around 50, with a high predicted accessibility of nearly 1. We further extend the window size to cover a wider range (from 1 to 11), and visualize the computational results versus the SHAPE-directed structure model in the second row (LinearSampling), the third row (RNAsubopt) and the bottom two rows (RNAplfold). For instance, the black circle at position 52 and window size 5 (indicated by a green arrow) represents a highly accessible region [52, 56] predicted by LinearSampling, which is surrounded by a box, indicating that the prediction is supported by the wetlab experiment. In RNAsubopt and RNAplfold predictions, the same positions are in purple or orange, indicating that they have lower accessibility and are less correlated with the SHAPE reactivities. In addition, LinearSampling predicts the region around position 279, indicated by blue arrows, as an inaccessible region, which is supported by the SHAPE-directed structures. The main differences between the three systems are highlighted in gray shades. In general, LinearSampling's result correlates better with the experimentally guided models.
To further quantify the difference, we calculate the accessibility defects of three important and well-studied regions in SARS-CoV-2, i.e. the 5 -UTR, the frameshifting element (FSE) and the 3 -UTR, shown in [fig_ref] Table 2: Accessibility defect comparisons between in silico predictions [/fig_ref]. LinearSampling has lower (better) accessibility defects on all these three regions, except for the FSE region using the reference structure from Manfredonia et al., on which LinearSampling is the second best (behind RNAplfold, span = 150) and has only 0.0025 difference from the best one.
New findings on accessible targets. Next, we aim to computationally obtain accessible regions as potential targets for diagnostics and drug discovery. A previous study [bib_ref] RNA genome conservation and secondary structure in SARS-CoV-2 and SARS-related viruses: a..., Rangan [/bib_ref] locates unstructured regions of SARS-CoV-2 by scanning the reference sequence with windows of 120 nt, sliding by 40 nt and then calculating base-pairing probabilities using CONTRAfold (52) for these fragments. In total, 75 accessible regions with ≥15 nt are claimed, where each base has the average unpaired probability of at least 0.6. However, this method has two flaws: (i) it is not correct to evaluate accessibility based on unpaired probabilities due to their mutual dependency; and (ii) it neglects long-range base pairs and has to approximate the accessibilities based on local structures, which is less accurate (see [fig_ref] Figure 8: acc [/fig_ref]. Instead, we measure the accessibilities based on samples generated by LinearSampling, setting the window size to be 15 following Rangan et al. [bib_ref] RNA genome conservation and secondary structure in SARS-CoV-2 and SARS-related viruses: a..., Rangan [/bib_ref]. We only show the fragments whose [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref]. The accessibilities derived from LinearSampling correlate better with the unpaired region in the SHAPE-directed structure of the 5 -UTR of SARS-CoV-2 [bib_ref] Comprehensive in vivo secondary structure of the SARS-CoV-2 genome reveals novel regulatory..., Huston [/bib_ref] [bib_ref] Genome-wide mapping of SARS-CoV-2 RNA structures identifies therapeutically-relevant elements, Manfredonia [/bib_ref]. Note that the full sequence was used for the accessibility calculation, but only the 5 -UTR is shown. First row: accessibilities of window size 4 derived from SHAPE-directed structures (blue plus symbol and purple cross symbol; agreed on >90% positions), RNAsubopt (green line), RNAplfold local mode (blue lines) and LinearSampling (red line). Following the previous study [bib_ref] A statistical sampling algorithm for RNA secondary structure prediction, Ding [/bib_ref] , the accessibility of position i stands for the region [i, i + 3]. Second row: accessibilities predicted by LinearSampling with window sizes from 1 to 11. Each prediction is presented with a filled circle, where the color correlates with its accessibility value. The SHAPE-directed structure [bib_ref] Comprehensive in vivo secondary structure of the SARS-CoV-2 genome reveals novel regulatory..., Huston [/bib_ref] is shown in dot-bracket format along the x-axis, and its accessible regions are annotated in boxes. Third, fourth and fifth rows: accessibilities predicted by RNAsubopt and RNAplfold (with a base pairing limit of 150 and 800), respectively. Note that the first row is a special case (window size 4) of the bottom four rows. The main differences are highlighted in gray shading, where LinearSampling is the most accurate in all cases. For example, the green arrows point to window size 5 at position 52 (representing the region [52, 56]), showing the case of a 5 nt accessible region in which LinearSampling predicts the highest accessibility, correlating better with the experimentally guided models. In addition, the blue arrows, pointing to the region around position 279, illustrate an example where an inaccessible region is accurately predicted by LinearSampling but not by other tools. We chose span = 150 and span = 800 for RNAplfold, the former being the largest default limit of local folding in most literature and software, and the latter following a previous study [bib_ref] Efficient siRNA selection using hybridization thermodynamics, Lu [/bib_ref]. . Some of the regions are overlapped, resulting in a total of nine separate accessible regions, which are illustrated in [fig_ref] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure [/fig_ref]. Of the nine regions, two are in ORF1ab, one in ORF3a, one in the M gene, three in the N gene and two in the S (spike) gene whose proteins can recognize and bind with its receptor [bib_ref] Structural and functional properties of SARS-CoV-2 spike protein: potential antivirus drug development..., Huang [/bib_ref].
The constant emergence of mutations may lead to changes in new virus variants. Therefore, it is better to target both conserved and accessible regions. LinearSampling is able to integrate conservation information for sampled structures by loading the partition function dumped from LinearTurboFold [bib_ref] LinearTurboFold: linear-time global prediction of conserved structures for RNA homologs with applications..., Li [/bib_ref] , which simultaneously folds and aligns homologous sequences and can identify conserved regions. This feature of LinearSampling allows mutationinsensitive diagnostics and drug design. The homologyassisted results have been previously reported [bib_ref] LinearTurboFold: linear-time global prediction of conserved structures for RNA homologs with applications..., Li [/bib_ref].
# Discussion
We focus on simplifying and accelerating the stochastic sampling algorithm for a given RNA sequence. Algorithmically, we present a hypergraph framework under which the classical sampling algorithm can be greatly simplified. We further elaborate this sampling framework to formal- ize the classical Non-Saving Sampling, and devise two new sampling algorithms, the Full-Saving Sampling that saves all hyperedges a priori and avoids re-computing in sampling phase, and LazySampling which eliminates redundant work and avoids unnecessary hyperedges saving via on-demand caching. We show that the LazySampling algorithm, i.e. exact partition function followed by a Lazy-Saving sampling, is the fastest among the three sampling strategies. Then we present LinearSampling, which combines LazySampling and LinearPartition.
LinearSampling is the first sampling algorithm to run in linear time without imposing constraints on the base pair distance, and is orders of magnitude faster than Vienna RNAsubopt. We conclude that
- LinearSampling runs linearly both in end-to-end and sampling-only time, and can scale up to long RNA sequence without any overflow issue; - its sampled structures correlate better with the ground truth structures on a diverse database, and with the experimentally guided structure of SARS-CoV-2; - it can be applied to SARS-CoV-2 to discover regions with high accessibilities, which are potential targets for diagnostics and drug design.
## Data availability
Our web server is available at http://linearfold.org/sampling.
[fig] Figure 1: Overview and applications of stochastic sampling of RNA secondary structure. The FMN riboswitch is used as an example, and their centroid structures in the absence of the ligand are shown(7). [/fig]
[fig] Figure 2: Top: the hypergraph framework and the bottom-up partition function calculation. All nodes and hyperedges are shown. Bottom: two sampling instances given the partition function above. The sampling is just top-down stochastic backtracing, i.e. at each visited node, we randomly choose one of its incoming hyperedges, traverse that hyperedge down and recursively visit its children nodes. [/fig]
[fig] Figure 3: Illustrations of the three sampling algorithms: classical Non-Saving Sampling, our Full-Saving Sampling and our LazySampling. For each sample, the newly recovered hyperedges are shown with shadow, and the bold hyperedges are the chosen derivation in stochastic backtracing (pointing downwards) [/fig]
[fig] Figure 4: Runtime and memory comparisons against sequence length on the RNAcentral dataset (cut at 4000 nt) under exact partition function. The sample size is 50 000. (A-C) Partition function, sampling-only and end-to-end runtime. (D) Memory usage. [/fig]
[fig] Figure 5: In practice, most node visits are repeated, and only a small portion of all nodes are visited. (A) Unique visit ratio ␣ k . (B) visited ratio  k . [/fig]
[fig] Figure 6, Figure 7: Runtime and memory usage between RNAsubopt, its local mode (span = 70) and LinearSampling on the RNAcentral dataset and SARS-CoV-2 (sample size 50 000). (A) End-to-end runtime. (B) Sampling-only runtime. (C) Memory usage. The large filled triangles and circles around n = 30 000 represent SARS-CoV-2 sequences, and the open triangles are the sequences that RNAsubopt overflows. Note that the sampling-only time of LinearSampling is already the majority of its end-to-end time on SARS-CoV-2 (64.6 out of 123.4 s). See Supplementary Figure S7 for the comparison within 10 000 nt. LazySampling can reduce sampling phase runtime with both linear (A) and exact (B) partition function. (A) Linear partition function time (using LinearPartition), Non-Saving Sampling and LazySampling against sequence length. Here LP = LinearPartition. (B) The sampling time of RNAsubopt, Non-Saving Sampling and LazySampling under the exact partition function. The runtime of RNAsubopt from Non-Saving Sampling is reduced by 'Boustrophedon' optimization. See Supplementary Figure S8 for the comparison within 10 000 nt, and Supplementary Figure S15 for more runtime analysis regarding the 'Boustrophedon' method. [/fig]
[fig] Figure 8: acc({y * }, i ) − acc(S, i ) where acc(S, i) is the accessibility of region [i, i + 3]: acc(S, i ) = 1 |S| y∈S 1[y i,i +3 = " .... "] LinearSampling is better correlated with the ground truth structure. (A) The ensemble defect difference of each family and overall (averaged by families) of the ArchiveII dataset, comparing LinearSampling, RNAsubopt and its local modes (span = 70 and 150); RNAsubopt is set to be the baseline. (B) The accessibility defect difference among LinearSampling, RNAsubopt, RNAplfold and their local modes, using RNAplfold as the baseline. The sample size is 50 000. The families are ordered by their average sequence lengths, from the shortest to the longest; see [/fig]
[table] Table 2: Accessibility defect comparisons between in silico predictions (LinearSampling, RNAsubopt and RNAplfold) and experimentally guided models [Huston et al. (14) and Manfredonia et al. (41)] on UTRs and FSE regions The 5 -UTR, FSE and 3 -UTR are [1, 450], [13470, 13545] and [29550, 29870], respectively (45). The sample size is 10 000. accessibilities are >0.5, i.e. they are more likely to be open than closed. We list all 23 regions found by LinearSampling in [/table]
|
Nurses’ Occupational and Medical Risks Factors of Leaving the Profession in Nursing Homes
This study aimed to evaluate the association between intention to leave work, and working conditions and health status among female care-staff in nursing homes. A multicenter cross-sectional study included female care-staff in 105 nursing homes for the elderly. We used validated questionnaires to assess occupational, psychosocial and medical data in a multicenter transverse study. Univariate analysis on chi 2 test was performed with stratification according to job (nurse, nursing assistant), and variables found to be significant on each dimension were included on multivariate models. 1428 nursing assistants and 342 registered nurses were included. 391 nursing assistants and 85 registered nurses intended to leave their work with the elderly. The registered nurses' intention to leave was associated with deteriorated care-team or residents relations, and with perceived elevated hardship due to the proximity of residents' death. The nursing assistants' intention to leave was associated with deteriorated management relation, with job insecurity and elevated hardship due to the residents' intellectual deterioration. Impaired physical or psychological health status also correlated with this intention. Policy to reduce voluntary turnover of care-staff in nursing homes for the elderly could be based on multifactorial management, acting on work organization and reducing psychosocial stress.
# Introduction
High rates of staff turnover in nursing homes are not a recent phenomenon, especially among those who work most closely with the residents: registered nurses and nursing assistants. Nursing homes for the elderly differ in functioning from hospitals in that they have lower level of medical presence. They are residential establishments caring for increasingly elderly, dependent persons with multiple comorbidities. Care is given by nursing assistants with various kinds of initial training and by registered nurses. Registered nurses usually ensure technical care and coordinate the work of the nursing assistants. As well as catering and accompaniment, nursing assistants are in charge of hygiene, comfort and preventive and curative care and are under the supervision of a nurse. About 17.5% of new nurses leave their first job within one year of starting their jobs [bib_ref] What does nurse turnover rate mean and what is the rate?, Kovner [/bib_ref]. The proportion of nurses considering or intending to leave the profession varied from 4% to 54% [bib_ref] Nurses' intention to leave the profession: Integrative review, Flinkman [/bib_ref]. The cost of turnover from the organizational perspective includes replacement costs (including training), lost productivity, lost quality and lowered morale [bib_ref] An Exploration of Job, Organizational, and Environmental Factors Associated with High and..., Brannon [/bib_ref] [bib_ref] The costs of turnover in nursing homes, Mukamel [/bib_ref]. A previous study argued that these costs may well be reflected in the quality of care that residents receive [bib_ref] An Exploration of Job, Organizational, and Environmental Factors Associated with High and..., Brannon [/bib_ref] [bib_ref] Turnover among nursing home staff. A review, Cohen-Mansfield [/bib_ref]. Three broad categories of factors that influence turnover have been identified: general economic, work-related and personal factors [bib_ref] An Exploration of Job, Organizational, and Environmental Factors Associated with High and..., Brannon [/bib_ref]. Registered nurses and nursing assistants in nursing homes are subject to strong mental and physical demands (e.g., lifting and carrying, work schedule), and frequently describe their working environment as distressing [bib_ref] Occupational risk factors for upper-limb and neck musculoskeletal disorder among health-care staff..., Pelissier [/bib_ref]. The physical and mental deterioration of elderly patients and the proximity to death were more often perceived as a source of hardship by registered nurses and nursing assistants. Registered nurses and nursing assistants are more often confronted by aggressive patients than in hospitals or facilities for the disabled [bib_ref] Design and Operation of the 2010 National Survey of Residential Care Facilities, Moss [/bib_ref] [bib_ref] Risk factors for work-related stress and subjective hardship in health-care staff in..., Pélissier [/bib_ref]. Registered nurses who had experienced burnout or stress were more likely to intend to leave the profession [bib_ref] Nurses' intention to leave the profession: Integrative review, Flinkman [/bib_ref]. have focused on the relationship between musculoskeletal disorders and intent to leave the profession in staff of nursing-homes for the elderly [bib_ref] Predictors of leaving nursing care: A longitudinal study among Swedish nursing personnel, Fochsen [/bib_ref]. Several studies underlined an association between impaired mental well-being in nursing home care workers and psychosocial factors in health caregivers, but there has been less research on the association between impaired mental well-being in nursing home care workers and health caregivers' intention to leave has hardly been studied [bib_ref] Impaired mental well-being and psychosocial risk: A cross-sectional study in female nursing..., Pélissier [/bib_ref] [bib_ref] Stress, depression, and intention to leave among nurses in different medical units:..., Chiang [/bib_ref] [bib_ref] Linking hospital workers' organisational work environment to depressive symptoms: A mediating effect..., Jolivet [/bib_ref] [bib_ref] Reviewing the effort-reward imbalance model: Drawing up the balance of 45 empirical..., Van Vegchel [/bib_ref]. Understanding why health care workers abandon their current employer and/or their job in the nursing profession is important in retention of health providers. The objective of the present study was to analyze occupational and medical factors associated with the intention to leave work with the elderly, comparing between nursing assistants and registered nurses, considering the differences in their tasks.
# Materials and methods
## Design
The study was designed as a cross-sectional questionnaire survey design.
## Participants
The target population of the survey was female employees working in with elderly patients in nursing homes in the Rhône-Alpes Region of France. The occupational physicians of the Region were asked to participate in the survey by the Regional Department of Businesses, Competition, Consumption, Work and Employment (DIRECCTE), a state business consultancy. Volunteer occupational physicians filled out a working conditions assessment questionnaire, and asked all employees meeting the inclusion criteria in the nursing homes which they oversaw to take part. Only employees who had been working with the elderly for at least 6 months on at least a half-time basis were included. The occupational physician handed over the self-administered questionnaire to subjects included in the study and collected data on working conditions and health status in nursing staff. The number of subjects who refused to participate was anonymously collected by the occupational physicians.
## Data collection
Between October 2009 and September 2010, data were collected by self-administered questionnaires during periodic follow-up medical examinations or during company visits. These self-administered questionnaires comprised personal characteristics (age, gender, family status, number of children), health characteristics and work-related characteristics and used four questionnaires measuring Job related hardship, musculoskeletal complaints, impaired mental well-being, and psychosocial demands [bib_ref] The measurement of effort-reward imbalance at work: European comparisons, Siegrist [/bib_ref] [bib_ref] Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms, Kuorinka [/bib_ref] [bib_ref] Validation of the General Health Questionnaire (GHQ-12) adapted to a work-related context, Lesage [/bib_ref].
## Measures
Job related hardship: The questionnaire included visual analog hardship scales (1 = no hardship, to 10 = great hardship) related to: premises (disrepair, stairs, clutter), lifting and carrying the elderly, patients' mental and physical deterioration, and proximity of death [bib_ref] The measurement of effort-reward imbalance at work: European comparisons, Siegrist [/bib_ref]. The choice of these particular categories was based on the literature. Hardship scores were categorized as slight (≤3), moderate (4-7) or severe hardship (≥8).
Musculoskeletal complaints: Musculoskeletal complaints were assessed on the Nordic Musculoskeletal Questionnaire, comprising multiple choice questions for each body part (such as upper limbs, lower limbs, neck and back): "During the last 12 months, have you had trouble (such as ache, pain, discomfort)?". The Nordic Musculoskeletal Questionnaire is a standardized questionnaire with reliability and validity moderate to high, which makes is a good instrument for epidemiological survey into musculoskeletal disorders and complaints [bib_ref] Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms, Kuorinka [/bib_ref].
Impaired mental well-being: Developed as a screening tool to detect the workers likely to have or to risk of developing psychiatry disorders, the General Health Questionnaire (GHQ) is a measure of the common mental health problems, domains of depression, anxiety, somatic symptoms and social withdrawal. It's the 12-item version (GHQ-12) was used. Respondents had to indicate how frequently they had recently experienced the various symptoms listed. Each item was rated on a 4-point scale, with weights from 0 to 3 according to Likert-like type scale (0 = less than usual, 1 = no more than usual, 2 = rather more than usual, and 3 = much more than usual) [bib_ref] Validation of the General Health Questionnaire (GHQ-12) adapted to a work-related context, Lesage [/bib_ref]. Scores higher than 12 were considered as indicating impaired mental wellbeing [bib_ref] Stress at work and mental health status among female hospital workers, Estryn-Behar [/bib_ref] [bib_ref] Working conditions in call-centers, the impact on employee health: A transversal study...., Charbotel [/bib_ref].
Psychosocial demands: Psychosocial demands at work were assessed on the effort-reward imbalance (ERI) model using the Siegrist questionnaire. In this model, chronic work-related stress is identified as non-reciprocity or imbalance between high efforts spent and low rewards received. Conversely, positive emotions evoked by appropriate social rewards promote well-being, health and survival [bib_ref] The measurement of effort-reward imbalance at work: European comparisons, Siegrist [/bib_ref]. This questionnaire has three scales: two measuring the extrinsic components of 'effort' (6 items) and 'reward' (11 items covering earnings, esteem and job security) and one measuring the intrinsic component of 'over-commitment' (six items) [bib_ref] The measurement of effort-reward imbalance at work: European comparisons, Siegrist [/bib_ref]. Effort was measured either by six items on the demanding aspect of the work environment (three measuring quantitative load, one qualitative load, one increase in total load over time, and one physical load), rated as 1 = does not apply, 2 = does apply but subject does not consider herself distressed, 3 = does apply and subject considers herself somewhat distressed, 4 = does apply and subject considers herself distressed, or 5 = does apply and subject considers herself very distressed. A sum score of these ratings was totalled, as documented in several studies [bib_ref] The measurement of effort-reward imbalance at work: European comparisons, Siegrist [/bib_ref]. According to the ERI model, extrinsic and intrinsic effort scores are directly proportional to effort, whereas the rewards score is inversely proportional to reward. ERI was measured by calculating the ratio between the extrinsic effort index (E) and the inverse reward index (R): E/(R*c), with c as a correction factor (c: 6/11); ERI >1 indicates a critical condition of high-cost/low-gain, or effort-reward imbalance [bib_ref] The measurement of effort-reward imbalance at work: European comparisons, Siegrist [/bib_ref]. To obtain a more precise view of which rewards were related to which indicators of employee health, three occupational rewards (earnings, esteem and job security) were assessed separately. All three reward scores were dichotomized separately, using the most adverse tercile to indicate low reward: the higher the reward score, the lower the reward [bib_ref] The measurement of effort-reward imbalance at work: European comparisons, Siegrist [/bib_ref] [bib_ref] Psychosocial factors at work and self reported health: Comparative results of cross..., Niedhammer [/bib_ref] [bib_ref] Déséquilibre efforts/récompense, Langevin [/bib_ref].
# Ethical considerations
Approval by the French Ministry of Health Research (Comité Consultatif pour le Traitement de l'Information en Matière de Recherche dans le Domaine de la Santé, 09.320) was obtained before starting the study. The participants were given an information leaflet explaining the study objectives and participation was voluntary.
# Data analysis
The continuous quantitative variable of seniority in the establishment was transformed into an ordinal qualitative variable for statistical purposes.
A descriptive step characterized the employee population according to personal characteristics, working conditions, including occupational psychosocial factors, and mental health status. Distinguishing between registered nurses and nursing assistants, associations were sought between the intent to leave work with the elderly and personal factor, occupational factors, psychosocial factors and medical factors. Frequencies were compared on chi-squared tests, with chi-squared trend tests depending on the results of cross-analysis. Ratios of event probabilities per case of intent to leave were studied. As the prevalence of each event was high, odds ratios would not provide a good estimate of relative risk, and a log-linked binomial model was applied, using the PROC GENMOD procedure in the SAS statistical package (version 9.3) with the DIST = BINOMIAL and LINK = LOG options.
Univariate and multivariate analyses were performed stratified on job category (nurse/nursing assistant). The binary response of each case of intent to leave was modeled in two steps: Firstly, all independent variables underwent univariate analysis; secondly, multivariate analyses were performed for each dimension (sociodemographic, occupational, psychosocial, and personal health); variables with p-value ≤ 0.1 were included in a multivariate model on a descending procedure, and variables with p-value < 0.05 were kept in the model. The psychosocial and medical dimension models were adjusted on age. For the psychosocial dimension, the global ERI variable is classically used, but in the present study we sought to investigate the questionnaire sub-dimensions in greater detail, studying the four variables of extrinsic effort and the three types of reward (esteem, earnings, and job security) in the multivariate models. From these four multivariate models, multivariate regression was performed on a descending procedure, with a significance threshold of 5%. In case of non-convergence of PROC GENMOD because the maximum likelihood estimate (MLE) lay on the boundary of the parameter space, the SAS COPY macro was used, which provides a good approximation to the exact maximum likelihood estimates, as well as yielding good estimates of the true population parameters [bib_ref] Estimation of prevalence ratos when PROC GENMOD does not converge, Deddens [/bib_ref].
# Results
## Socio-occupational and medical data
1770 women (19.6% registered nurses, 80.4% nursing assistants) working in direct contact with the elderly in 105 nursing homes were included. Twenty-nine registered nurses and nine nursing assistants refused to participate, leading to a global participation rate of 98%. Four hundred and seventy-six of the 1808 women (26.3%), wished to leave their work with the elderly: 391 nursing assistants (26.8%) and 85 registered nurses (24.2%).
## Factors associated with intention to leave work on univariate analysis
Occupational factors: In both nursing assistants and registered nurses, severe hardship related to residents' physical deterioration, to proximity to death, or to lack of equipment was significantly associated with intention to leave. In nursing assistants, weak attachment to the elderly residents, understaffing, and severe hardship related to lifting and carrying residents, to premises, or to residents' intellectual deterioration, were significantly associated with intention to leave [fig_ref] Table 1: Sociodemographic, occupational, psychosocial and medical factors associated with registered nurses' and nursing... [/fig_ref]. In both nursing assistants and registered nurses, work contract, seniority in work with elderly, working time, night work and type of working hours were not associated with intention to leave. Psychosocial factors: In both nursing assistants and registered nurses, intention to leave was significantly associated with impaired relations with residents, poor reward in terms of earnings, job security, and esteem, over-commitment, and ERI.
In nursing assistants, intention to leave was significantly associated with physical, or verbal aggression, and impaired relations with management, or families.
In registered nurses, intention to leave was significantly associated with impaired relations with the care-team [fig_ref] Table 1: Sociodemographic, occupational, psychosocial and medical factors associated with registered nurses' and nursing... [/fig_ref].
Med Psychosocial factors: In both nursing assistants and registered nurses, intention to leave was significantly associated with impaired relations with residents, poor reward in terms of earnings, job security, and esteem, over-commitment, and ERI.
In nursing assistants, intention to leave was significantly associated with physical, or verbal aggression, and impaired relations with management, or families.
In registered nurses, intention to leave was significantly associated with impaired relations with the care-team [fig_ref] Table 1: Sociodemographic, occupational, psychosocial and medical factors associated with registered nurses' and nursing... [/fig_ref].
ical factors: For both nursing assistants and registered nurses, psychological distress and spinal, or lower-limb complaints were significant. Upper-limb musculoskeletal disorder was significantly associated with intention to leave in nursing assistants [fig_ref] Table 1: Sociodemographic, occupational, psychosocial and medical factors associated with registered nurses' and nursing... [/fig_ref].
## Factors associated with intention to leave work on multivariate analysis
Occupational factors: Severe hardship related to proximity to death was associated with intention to leave in registered nurses, as was severe hardship related to residents' intellectual deterioration and weak attachment to residents in nursing assistants [fig_ref] Table 2: Significant associations between nursing assistants' intention to leave work with the elderly... [/fig_ref]. In this global multivariate analysis, hardship related to inadequate equipment no longer featured as a significant factor [fig_ref] Table 4: Significant associations between registered nurses' and nursing assistants' intention to leave work... [/fig_ref].
Psychosocial factors: Impaired relations with management in nursing assistants, impaired relations with residents or with care team in registered nurses, remained associated with intention to leave. In nursing assistants, high levels of effort and poor reward in terms of job security were associated with intention to leave. Poor esteem from colleagues no longer featured as a significant factor on global multivariate analysis [fig_ref] Table 2: Significant associations between nursing assistants' intention to leave work with the elderly... [/fig_ref].
Medical factors: Lower limb complaints and spinal complaints remained associated with intention to leave in nursing assistants but not registered nurses on global multivariate analysis [fig_ref] Table 4: Significant associations between registered nurses' and nursing assistants' intention to leave work... [/fig_ref].
# Discussion
In the present study, care-workers' intention to leave was significantly related to certain psychosocial factors and impaired health status. This study of 1770 female registered nurses and nursing assistants in 105 nursing homes for the elderly in the Rhône-Alpes Region of France had a 98% response rate. This study has the advantage of having exploited for the administration of the questionnaires the periodic visits conducted by the occupational physicians. This method usually allows a high participation and minimizes the possibility that the answers are influenced by other workers, because the time in the waiting room is very limited [bib_ref] Social psychiatry in the waiting room. What a physician can learn about..., Magnavita [/bib_ref].
Validated questionnaires assessed psychosocial demand, musculoskeletal complaints and psychological distress. The interest of the study lay in correlating intention to leave work with the elderly and working conditions (psychosocial and organizational demand) while taking into account the registered nurses' and nursing assistants' psychological and physical health status. This cross-sectional study does not allow us to know the temporal relation between the different phenomena that influence the intention to leave. However, previous studies have made it clear that the various factors, violence and stress, are cyclically related to each other [bib_ref] The exploding spark. Workplace violence in an infectious disease hospital-A longitudinal study, Magnavita [/bib_ref] [bib_ref] Workplace violence and occupational stress in health care workers: A chicken and..., Magnavita [/bib_ref] [bib_ref] The determinants and consequences of adult nursing staff turnover: A systematic review..., Halter [/bib_ref].
Four hundred and seventy-six (26.3%) of respondents wished to leave their work with the elderly, and nursing assistants (26.8%) more frequently than registered nurses (24.2%). This finding agrees with Fochsen et al., who reported that "being an assistant nurse was associated with leaving nursing care" [bib_ref] Predictors of leaving nursing care: A longitudinal study among Swedish nursing personnel, Fochsen [/bib_ref]. The NEXT-Study (Nurses' early exit study) was established to investigate the degree, reasons, circumstances, and consequences of premature departure from the nursing profession in Europe. The proportion of nurses (24.9%) who had frequently considered leaving their profession in nursing homes was higher in our study than in the French data of the longitudinal NEXT-Study (16.5%), in which nurses were working in residential care for the elderly, hospital structures or home-care [bib_ref] Effort-Reward Imbalance among Nurses in Stable Countries and in Countries in Transition, Hasselhorn [/bib_ref]. The negative relation-ship between age and turnover intent has been widely reported in previous studies [bib_ref] Stress, depression, and intention to leave among nurses in different medical units:..., Chiang [/bib_ref] [bib_ref] The relationship between job satisfaction and intention to leave current employment among..., Ramoo [/bib_ref]. It has been argued that young nurses who belong to generation Y (born after 1980) hold different attitudes and values towards their work, which may influence their retention. According to Lavoie-Tremblay et al., the generation Y nurses reported that recognition was a key motivator. Their needs are stability, flexible work schedules and shifts, recognition, opportunities for professional development, and adequate supervision [bib_ref] Addressing the turnover issue among new nurses from a generational viewpoint, Lavoie-Tremblay [/bib_ref].
The study highlighted a significant association between hardship in relation to proximity to death and intention to leave work with the elderly. Previous study showed that nursing home staff who had had palliative care training were less liable to report severe hardship in relation to proximity to death, suggesting that palliative care training has positive impact on perceived hardship in relation to proximity to death [bib_ref] Pénibilité ressentie et accès aux formations chez le personnel féminin travaillant auprès..., Pélissier [/bib_ref]. Better training in the management of dementia patients and in palliative care might reduce the intention to leave work with the elderly.
A number of researchers have reported an association between workload, stress and turnover intention [bib_ref] Job Perceptions and Intent to Leave Among Direct Care Workers: Evidence from..., Brannon [/bib_ref] [bib_ref] The Influence of Work-Related Fatigue, Work Conditions, and Personal Characteristics on Intent..., Liu [/bib_ref]. In this cross-sectional study, both workplace violence and bad relationships (which often fall into verbal violence) are significantly associated with the intention to leave work. The effect was higher in younger workers; this confirms what observed in previous Italian cohorts of nurse and nursing students [bib_ref] Workplace violence against nursing students and nurses. An Italian experience, Magnavita [/bib_ref]. Previous studies suggested that the psychologically strenuous and stressful nature of the work causes nurses to consider leaving the nursing profession [bib_ref] Average Hospital Length of Stay, Nurses' Work Demands, and Their Health and..., Cho [/bib_ref] [bib_ref] Psychosocial work environment and intention to leave the nursing profession: Results from..., Li [/bib_ref] [bib_ref] Factors associated with the intention to leave among newly graduated nurses in..., Tominaga [/bib_ref]. In the present study, intention to leave was significantly related to certain psychosocial factors. First effort/reward imbalance was significantly related to intention to leave in nursing assistants. The ERI model was used to analyze intention to leave the nursing profession in nurses (n = 21,299) in seven European countries, in the NEXT-Study: ERI was associated with the frequency of considering leaving the nursing profession [bib_ref] Effort-Reward Imbalance among Nurses in Stable Countries and in Countries in Transition, Hasselhorn [/bib_ref]. Secondly, overcommitment was not significantly associated with intention to leave in the present study, in agreement with previous findings [bib_ref] Effort-Reward Imbalance among Nurses in Stable Countries and in Countries in Transition, Hasselhorn [/bib_ref] [bib_ref] Factors associated with the intention to leave among newly graduated nurses in..., Tominaga [/bib_ref]. According to Zeytinoglu et al., nurses with higher levels of career commitment showed statistically significantly lower levels of propensity to leave nursing. Wang et al. found that a higher level of occupational commitment was related to stronger intention to stay [bib_ref] Job satisfaction, occupational commitment and intent to stay among Chinese nurses: A..., Wang [/bib_ref]. Low affective commitment to the profession was associated with greater intention to leave the profession [bib_ref] Nurses' intention to leave the profession: Integrative review, Flinkman [/bib_ref]. Moreover, Takase et al. found that nursing staff commitment was directly related to reduce intention to quit the profession [bib_ref] Nurses' leaving intentions: Antecedents and mediating factors, Takase [/bib_ref]. The present study highlighted weak attachment to residents on the part of nurses and nursing assistants as associated with intention to leave. Then high levels of effort were an important work environment factor for intention to leave, consistent with previous findings [bib_ref] Effort-Reward Imbalance among Nurses in Stable Countries and in Countries in Transition, Hasselhorn [/bib_ref] [bib_ref] The relationship between job satisfaction and intention to leave current employment among..., Ramoo [/bib_ref] [bib_ref] Addressing the turnover issue among new nurses from a generational viewpoint, Lavoie-Tremblay [/bib_ref] [bib_ref] Pénibilité ressentie et accès aux formations chez le personnel féminin travaillant auprès..., Pélissier [/bib_ref] [bib_ref] Job Perceptions and Intent to Leave Among Direct Care Workers: Evidence from..., Brannon [/bib_ref] [bib_ref] The Influence of Work-Related Fatigue, Work Conditions, and Personal Characteristics on Intent..., Liu [/bib_ref] [bib_ref] Workplace violence against nursing students and nurses. An Italian experience, Magnavita [/bib_ref] [bib_ref] Average Hospital Length of Stay, Nurses' Work Demands, and Their Health and..., Cho [/bib_ref] [bib_ref] Psychosocial work environment and intention to leave the nursing profession: Results from..., Li [/bib_ref] [bib_ref] Factors associated with the intention to leave among newly graduated nurses in..., Tominaga [/bib_ref]. Finally, for nursing assistants, poor reward in terms of job security, but in terms of not earnings, was significantly associated with intention to leave on multivariate analysis. Pay was found not a significant predictor for Jordanian nurses to leave their organization [bib_ref] Predicting nurses' turnover intentions by demographic characteristics, perception of health, quality of..., Al-Hussami [/bib_ref]. According to Lavoie-Tremblay et al., the reasons most often given for quitting the profession by nurses were difficult working conditions and a lack of job stability [bib_ref] Pénibilité ressentie et accès aux formations chez le personnel féminin travaillant auprès..., Pélissier [/bib_ref]. Improving the psychosocial work environment and specifically occupational rewards may be helpful in retaining caregivers.
Poor relations with management for nursing assistants and poor relations with the care-team and patients for nurses were significantly associated with intention to leave. Nursing assistants seemed to exchange little with nurses according to an ergonomic study of geriatric care-staff. In France, the working week was reduced to 35 h, which may have had the effect of reducing the time available for verbal exchanges within the team. Moreover, increasing administrative work for nurses and the increased work-load reported may have reduced the time available for verbal exchanges between patients and nurses, impairing the nurse-patient relationship. Cowden et al., in a systematic review of the literature, found a positive relationship between transformational leadership, supportive work environments and staff nurses intentions to stay [bib_ref] Leadership practices and staff nurses' intent to stay: A systematic review, Cowden [/bib_ref]. Creating a supportive work environment, with supportive leadership gives nurses the confidence to stay [bib_ref] Job stress and intent to stay at work among registered female nurses..., Kaewboonchoo [/bib_ref] [bib_ref] Differences of hospitals' organisational climates and nurses' intent to stay: Nurses' perspectives, Mrayyan [/bib_ref]. Healthy and supportive work relationships have been shown to be related to greater intention to stay in work for nurses [bib_ref] Nurse intention to remain employed: Understanding and strengthening determinants, Tourangeau [/bib_ref] [bib_ref] Determinants of hospital nurse intention to remain employed: Broadening our understanding, Tourangeau [/bib_ref].
The present study revealed a significant association between intention to leave and impaired health status. First, lower-limb and spinal musculoskeletal complaints were significantly associated with intention to leave for nursing assistants. Fochsen and Sjögren showed that musculoskeletal problems of the neck/shoulder and knees were associated with leaving nursing care [bib_ref] Predictors of leaving nursing care: A longitudinal study among Swedish nursing personnel, Fochsen [/bib_ref] [bib_ref] Reasons for leaving nursing care and improvements needed for considering a return:..., Sjögren [/bib_ref]. Secondly, psychological distress was significantly associated with intention to leave for both nursing assistants and nurses, in line with the longitudinal study of 6441 caregivers by Kivimaki et al. [bib_ref] Team climate, intention to leave and turnover among hospital employees: Prospective cohort..., Kivimäki [/bib_ref]. found that emotional labor has an indirect effect on intention to leave via burnout nurses experience [bib_ref] Do perceived high performance work systems influence the relationship between emotional labour,..., Bartram [/bib_ref].
# Conclusions
Care-staff in nursing homes for the elderly are exposed to elevated psychological and physical demands in caring for dependent elderly persons often with comorbidities. The present study found that some of the demands experienced by nurses and nursing assistants underlie an intention to leave work with the elderly. The results highlight the importance of perceived working conditions in this intention to leave. Policy to reduce voluntary turnover in nursing homes for the elderly could be based on a multifactorial approach involving work organization, reduced psychosocial demand and access to certain training modules such as in palliative care and the management of patients in dementia. Reorganization to reduce nurses' administrative workload by delegating tasks to administration workers should allow nurses more time for exchanges with both patients and nursing assistants. Having more time to devote to relational aspects would enhance verbal exchange between residents and caregivers, which in turn should improve job satisfaction and quality of care. Greater involvement in drawing up personalized care projects and reinforcing and developing caregiver autonomy in performing everyday care tasks should improve the caregivers' image of the job. Better recognition of the work accomplished could be achieved by increasing the time available for exchanges between colleagues and with management, and by improving job security.
[fig] Author: Contributions: Conceptualization, B.C. and L.F.; Data curation, C.P., E.F. and B.C.; Formal analysis, C.P. and E.F.; Funding acquisition, B.C.; Investigation, C.P. and B.C.; Methodology, C.P., B.C., J.B.F., E.F. and L.F.; Project administration, B.C. and J.B.F.; Supervision, B.C. and L.F.; Validation, C.P., B.C., E.F., J.B.F. and L.F.; Visualization, C.P., J.B.F., E.F. and L.F.; Writing-original draft, C.P. Funding: This research was funded by [Department of Businesses, Competition, Consumption, Work and Employment (DIRECCTE)]. [/fig]
[table] Table 1: Sociodemographic, occupational, psychosocial and medical factors associated with registered nurses' and nursing assistants' intention to leave work with the elderly, on univariate analysis.Table 1. Cont. [/table]
[table] Table 2: Significant associations between nursing assistants' intention to leave work with the elderly and sociodemographic, occupational, psychosocial and medical factors on intermediate multivariate analysis. [/table]
[table] Table 3: Significant associations between registered nurses' intention to leave work with the elderly and sociodemographic, occupational, psychosocial and medical factors on intermediate multivariate analysis. [/table]
[table] Table 4: Significant associations between registered nurses' and nursing assistants' intention to leave work with the elderly and sociodemographic, occupational, psychosocial and medical factors on global multivariate analysis. [/table]
|
Heterogeneity of asthma and COPD overlap
# Introduction
Asthma-COPD overlap (ACO) is a disorder in which features of asthma and COPD are mixed. The definition, clinical features, and course of ACO are still controversial. [bib_ref] Phenotype of asthma-chronic obstructive pulmonary disease overlap syndrome, Rhee [/bib_ref] [bib_ref] Overlap between asthma and COPD: where the two diseases converge, Kim [/bib_ref] Although the Global Initiative for Asthma (GINA) and Global Initiative for Chronic Obstructive Lung Disease (GOLD) outlined 11 items to allow ACO diagnosis, there is no consensus on diagnostic criteria of ACO. [bib_ref] Asthma-chronic obstructive pulmonary disease overlap syndrome (ACOS): current literature review, Papaiwannou [/bib_ref] [bib_ref] The asthma-COPD overlap syndrome, Postma [/bib_ref] [bib_ref] Do we really need asthma-chronic obstructive pulmonary disease overlap syndrome?, Cazzola [/bib_ref] ACO may be simply defined as COPD with airway hyper-responsiveness or as asthma with persistent airflow limitation. [bib_ref] Asthma-COPD overlap syndrome (ACOS) vs 'pure' COPD: a distinct phenotype?, Caillaud [/bib_ref] The prevalence of ACO in those with COPD and asthma is 12%-55% 5,7,8 and 16%-61%, [bib_ref] Do we really need asthma-chronic obstructive pulmonary disease overlap syndrome?, Cazzola [/bib_ref] [bib_ref] Asthma-chronic obstructive pulmonary disease overlap syndrome in the urban Chinese population: prevalence..., Ding [/bib_ref] respectively. In some studies, ACO patients displayed more severe symptoms, poor quality of life (QoL), and frequent exacerbations compared to COPD patients. [bib_ref] The clinical features of the overlap between COPD and asthma, Hardin [/bib_ref] [bib_ref] Characterisation of the overlap COPD-asthma phenotype. Focus on physical activity and health..., Miravitlles [/bib_ref] [bib_ref] Long-term prognosis of asthma, chronic obstructive pulmonary disease, and asthma-chronic obstructive pulmonary..., Lange [/bib_ref] Contrarily, a study reported no difference in severity or mortality between ACO and COPD alone, [bib_ref] Asthma-COPD overlap syndrome (ACOS) vs 'pure' COPD: a distinct phenotype?, Caillaud [/bib_ref] with another study describing better survival in ACO than in COPD. [bib_ref] Defining the asthma-COPD overlap syndrome in a COPD cohort, Cosio [/bib_ref] Surveys in the USA, China, and South Korea reported higher absence rates, medical resource utilizations, and medical costs for ACO patients compared to COPD patients. [bib_ref] Asthma-chronic obstructive pulmonary disease overlap syndrome in the urban Chinese population: prevalence..., Ding [/bib_ref] [bib_ref] Medical utilization and cost in patients with overlap syndrome of chronic obstructive..., Rhee [/bib_ref] [bib_ref] Economic burden in direct costs of concomitant chronic obstructive pulmonary disease and..., Blanchette [/bib_ref] A recent prospective population-based analysis reported the heterogeneity of ACO concerning the onset of asthma. [bib_ref] Long-term prognosis of asthma, chronic obstructive pulmonary disease, and asthma-chronic obstructive pulmonary..., Lange [/bib_ref] The reason for the differing prevalence and clinical outcomes of ACO in different studies could reflect the different definitions of ACO as well as heterogeneity in ACO. However, in order to overcome the disease, it is necessary to confirm that ACO is a disease with various characteristics. It is important to recognize that there are also various subtypes within ACO and that they may show different patterns in their frequency of exacerbations, medical use, and cost.
Therefore, the aims of this study were to investigate the heterogeneity of ACO in subjects with mild-to-moderate airflow limitation and to identify the diverse epidemiological characteristics and subtypes of ACO associated with exacerbations and health care use and cost through the national survey and insurance claims data.
# Methods study design
In the Korean National Health and Nutrition Examination Survey (KNHANES)conducted between 2007 and 2012, subjects aged $40 years with forced expiratory volume in 1 second (FEV 1 )/forced vital capacity (FVC) ,0.7 and FEV 1 $50% predicted were included. Self-reported wheezing was indicated as presence (W+) or absence (W−). An S+ subject was a current smoker or an exsmoker who had smoked $100 cigarettes. An S− subject had never smoked or had smoked ,100 cigarettes in his/her lifetime, which is also used in other surveys. [bib_ref] Cigarette smoking among U.S. Adults by state and region: estimates from the..., Shopland [/bib_ref] [bib_ref] Annual report to the nation on the status of cancer, 1973-1996, with..., Wingo [/bib_ref] The subjects were divided into the following four groups: W−S−, W−S+, W+S−, and W+S+. W+S− was regarded as asthma-predominant ACO, and W+S+ was regarded as COPD-predominant ACO. KNHANES data were analyzed according to the groups. These data were linked to the National Health Insurance (NHI) data, and the use and cost of health care services from January 2007 to December 2012 for each group were analyzed. Exacerbations were defined as deterioration in respiratory symptoms requiring systemic corticosteroids and/or antibiotics for hospitalization or emergency room (ER) visits, as in previous studies. [bib_ref] Medical utilization and cost in patients with overlap syndrome of chronic obstructive..., Rhee [/bib_ref] [bib_ref] Impacts of coexisting bronchial asthma on severe exacerbations in mild-to-moderate COPD: results..., Lee [/bib_ref] [bib_ref] The health care burden of high grade chronic obstructive pulmonary disease in..., Kim [/bib_ref] [bib_ref] Association between chronic obstructive pulmonary disease and gastroesophageal reflux disease: a national..., Kim [/bib_ref] [bib_ref] Health care use and economic burden of patients with diagnosed chronic obstructive..., Kim [/bib_ref] Medical costs and days of health care use not related to the respiratory diseases were excluded from the analysis. Detailed methods are described in previous studies [bib_ref] Medical utilization and cost in patients with overlap syndrome of chronic obstructive..., Rhee [/bib_ref] [bib_ref] Impacts of coexisting bronchial asthma on severe exacerbations in mild-to-moderate COPD: results..., Lee [/bib_ref] [bib_ref] The health care burden of high grade chronic obstructive pulmonary disease in..., Kim [/bib_ref] [bib_ref] Association between chronic obstructive pulmonary disease and gastroesophageal reflux disease: a national..., Kim [/bib_ref] [bib_ref] Health care use and economic burden of patients with diagnosed chronic obstructive..., Kim [/bib_ref] [bib_ref] Identification of subtypes in subjects with mild-to-moderate airflow limitation and its clinical..., Lee [/bib_ref] and are presented in [fig_ref] Figure 1: Study flow diagram [/fig_ref].
Among the variables, household income was divided into quartiles: upper, middle upper, middle lower, and lower. The level of education was based on the graduation level (elementary school, middle school, high school, and college). Marital status was divided into single, married, divorced, and bereaved. Physical activity was divided into three levels by the questionnaire. Physical activity 1 was defined as highly intensive activity that increases respiration or heart rate to a high level, physical activity 2 is moderate activity that leads to a slight increase in respiration or heart rate, and physical activity 3 is low-intensity activity. Self-rated health was also scored by the questionnaire and classed as excellent, good, fair, poor, and very poor. EuroQoL five dimensions questionnaire (EQ-5D) index was used to measure health-related QoL and included mobility, self-care, usual activities, pain/ discomfort, and anxiety/depression.Each item of EQ-5D was graded in three levels; the higher the grade, the lower the QoL. Our study was approved by the National Evidence Based Coordinating Agency Ethics Committee. Due to the retrospective nature of this study, the Ethics Committee waived the requirement for informed consent from subjects. All data accessed were de-identified.
## Statistical analyses
All data are expressed as mean ± SD or n (%). Analysis of variance or Kruskal-Wallis test was used to compare continuous variables among the four groups, and Student's t-test or Mann-Whitney test was used to compare asthmapredominant and COPD-predominant ACO. Categorical variables were analyzed by chi-square test or Fisher's exact test. Multiple logistic regression analysis was used to determine the effect of groups on the deterioration that required hospitalization or emergency room visits, and multiple linear regression analysis was used to determine the effect of groups on medical costs. A four-step model was applied to adjust confounding factors. No factor was adjusted in step 1. Age, sex, and body mass index (BMI) were adjusted in step 2. Age, sex, BMI, and household income were adjusted in step 3. Age, sex, BMI, household income, and FEV 1 were adjusted in step 4. Odds ratio (OR) and 95% confidence interval (CI) or beta coefficients are presented. P-value ,0.05 was considered statistically significant. All statistical analyses were performed using SAS Version 9.2 (SAS Institute Inc., Cary, NC, USA).
# Results
Of the 2,269 subjects with FEV 1 /FVC ,0.7 and FEV 1 .50% predicted, 410 (18%) subjects with W+ belonged to ACO. Among these 410 ACO subjects, 154 never smokers were classified as W+/S− (asthma-predominant ACO) and 256 smokers were classified as W+/S+ (COPD-predominant ACO). Of those without wheezing (W−), 681 were never smokers (W−/S−) and 1,178 were smokers (W−/S+). Although mean age was not different, age distribution was significantly different among the four groups (P=0.001).
Subjects aged 40-50 years were more frequent in the W+ groups and subjects aged 50-70 years were more frequent in the S+ groups. Men were more prevalent in both S+ groups than in both S− groups and in the COPD-predominant ACO group than in the asthma-predominant ACO group (both P,0.001). BMI and waist circumference were the highest in the COPD-predominant ACO group (both P,0.001). Household income and education level were the highest in the W−/S− group among the four groups (P,0.001) and were lower in the asthma-predominant ACO group than in the COPD-predominant ACO group (P,0.001). There were more married subjects in the two S+ groups and more bereaved subjects in the other S− groups (P,0.001). FEV 1 , FVC, and FEV 1 /FVC were lower in the W+ group than in the W− group (P,0.001). Among the W+ groups, FEV 1 /FVC was lower in the COPD-predominant ACO group than in the asthma-predominant ACO group (P=0.002), while FEV 1 and FVC were not different. Based on the responses to the questionnaires, more subjects were diagnosed with COPD in the COPD-predominant ACO group and more subjects were diagnosed with asthma in the asthma-predominant ACO group compared to the other groups (P,0.001). Based on the NHI data, osteoarthritis, osteoporosis, and depression were more common in the S− groups (P=0.012, ,0.001, and ,0.001, respectively) and osteoporosis was most prevalent in the asthma-predominant ACO group (P,0.001) [fig_ref] Table 1: epidemiological characteristics of the four groups according to the presence or absence... [/fig_ref].
Physical activity 1 (high intensity) was more frequent in the S+ groups among the four groups (P,0.001) and in the COPD-predominant ACO group compared to the asthma-predominant ACO group (P=0.026). Regarding the self-rated health levels, both W− groups more often rated their health as excellent or good compared to the W+ groups (P,0.001), with health rated as poorer more often in the asthma-predominant ACO group than in the COPDpredominant ACO group (P=0.001). Both W+ groups showed poorer health-related QoL in all EQ-5D items compared to W− groups (P,0.001), and usual activities and pain/ discomfort scores were lower in the asthma-predominant ACO group than in the COPD-predominant ACO group (P=0.024) [fig_ref] Table 2: Physical activity and health-related quality of life of the four groups with... [/fig_ref].
NHI data revealed that insurance type was not different among the four groups. Hospitalization rate was higher in both W+ groups (P,0.001), but there was no difference between both ACO groups. Mean total medical costs did not differ appreciably among the four groups. Outpatient medical expenses were highest in the COPD-predominant ACO group (P,0.001). This tendency was similar in the number of days of health care use [fig_ref] Table 3: health care utilization and insurance-related data among the four groups Notes [/fig_ref]. ACO groups were prescribed more respiratory medicine compared to both W− groups (P,0.001). Within the ACO groups, the prescription pattern was slightly different according to the smoking history. The asthma-predominant ACO group was prescribed more inhaled corticosteroids (ICS)/ long-acting beta-agonists (LABA). Conversely, the COPDpredominant ACO group was prescribed more long-acting muscarinic antagonists (LAMA) [fig_ref] Table 4: Prescription patterns of respiratory medication among the four groups [/fig_ref].
Multiple logistic regression analysis showed that, when the OR of the W−/S− group was set to 1, the COPD-predominant ACO group was associated with increased OR of any exacerbation (age, sex, BMI, household income, and FEV 1 % predicted -adjusted OR [aOR], 1.79; 95% CI, 1.12-2.85). However, when OR of the asthma-predominant ACO group was set to 1, the COPD-predominant ACO group did not show significant results [fig_ref] Table 5: associations of each group with exacerbations [/fig_ref]. When the OR of W−/S+ group was set to 1, the COPD-predominant ACO group was associated with increased odds of any exacerbation (OR 2.11, 95% CI 1.43-3.10) [fig_ref] Table 6: associations of the four subgroups or each variable with any exacerbation requiring... [/fig_ref].
Compared to the W−/S− group (OR, 1), outpatient and total medical costs were significantly associated with the COPD-predominant ACO group in all adjusted models [fig_ref] Table 7: associations of the four groups with medical cost [/fig_ref]. Compared to the asthma-predominant ACO group, only outpatient medical costs were associated with the COPDpredominant ACO group in all adjusted models [fig_ref] Table 8: association of wheezing subtypes with medical cost Notes [/fig_ref].
# Discussion
We identified the heterogeneity of ACO according to smoking status and found that COPD-predominant ACO might have larger clinical impact, such as lower lung function and higher outpatient medical cost than asthma-predominant ACO. Since our study subjects were selected randomly from a
## 1256
Kim et al nationwide survey, and mostly presented only mild-tomoderate airflow limitation, although their FEV 1 /FVC was ,0.7. Therefore, we defined airway hyper-responsiveness as the presence of self-reported wheezing instead of the bronchial provocation test and/or bronchodilator responsiveness. Nonetheless, the clinical characteristics were well characterized when divided into the four groups (W−/S−, W−/S+, W+/S−, and W+/S+) according to smoking and wheezing status. In this study, ACO (W+) groups had a younger mean age, lower self-reported health and QoL levels, and higher health care utilization than W− groups. S+ groups showed more frequent married status and more intensive physical activity and less common osteoarthritis, osteoporosis, and depression than S− groups, which was mainly due to differences in sex between the two groups. Almost all (1,363/1,434, 95%) of the S+ groups were males, and the majority of the S− groups were females (568/835, 68%). When comparing both ACO groups, males predominated, physical activity was more intensive, and FEV 1 /FVC was lower in the COPD-predominant ACO group than in the asthma-predominant ACO group. Socioeconomic status parameters, such as household income and education levels, self-rated health, and health-related QoL levels, were lower in the asthma-predominant ACO group than in the COPD-predominant ACO group. The asthma-predominant ACO group showed higher prevalence of osteoporosis compared to the COPD-predominant ACO group. This may be due to the higher proportion of women in the asthma-predominant ACO group. There was no difference between the asthma-predominant ACO and the COPD-predominant ACO in the medical utilization, except outpatient costs, which were higher in the COPD-predominant ACO group. Multiple logistic regression analysis showed that the COPD-predominant ACO group was associated with increased odds of exacerbations compared to both groups without wheezing (W−/S− and W−/ S+) but not the asthma-predominant ACO group. Subjects with ACO (W+) were more likely to experience exacerbations and to have lower health-related QoL compared to those without ACO (W−). Both W+ groups also showed higher medical costs and health care utilization than W− groups. Similarly, most previous comparative studies of ACO and COPD showed that ACO had more frequent exacerbations, symptoms, and comorbidities and lower QoL. [bib_ref] Asthma-COPD overlap 2015: now we are six, Gibson [/bib_ref] [bib_ref] The clinical and genetic features of COPD-asthma overlap syndrome, Hardin [/bib_ref] [bib_ref] The asthma-chronic obstructive pulmonary disease overlap syndrome (ACOS): opportunities and challenges, Barrecheguren [/bib_ref] [bib_ref] Characteristics and self-rated health of overlap syndrome, Chung [/bib_ref] In the EPI-SCAN study of a Spanish cohort, more frequent respiratory symptoms and exacerbations and lower FEV 1 , FVC, QoL, and physical activities were observed in ACO patients than in COPD patients. [bib_ref] Characterisation of the overlap COPD-asthma phenotype. Focus on physical activity and health..., Miravitlles [/bib_ref] The PLATINO study of Latin American countries also showed that ACO patients had more symptoms, lower pulmonary function, higher risk of exacerbations and hospitalizations, and worse health status than those with COPD. [bib_ref] Increased risk of exacerbation and hospitalization in subjects with an overlap phenotype:..., Menezes [/bib_ref] These results were similarly observed in the COPDGene Study conducted in the USA and South Korea. [bib_ref] The clinical features of the overlap between COPD and asthma, Hardin [/bib_ref] [bib_ref] Medical utilization and cost in patients with overlap syndrome of chronic obstructive..., Rhee [/bib_ref] In a meta-analysis of ACO-related studies, no differences in sex, smoking amounts, lung function, and 6-minute walk distance were found between ACO and COPD patients, but ACO patients were younger and had higher BMI, greater use of medical resources, and lower health-related QoL. [bib_ref] Asthma and COPD overlap syndrome (ACOS): a systematic review and meta analysis, Alshabanat [/bib_ref] In terms of medical use and cost, Medicare beneficiary data revealed that patients with diagnostic codes of both COPD and asthma had higher medical use and expenses than those with codes of COPD alone. 14 In Korea, medical utilization and cost were reported to be significantly higher in ACO than in COPD or asthma. [bib_ref] Medical utilization and cost in patients with overlap syndrome of chronic obstructive..., Rhee [/bib_ref] [bib_ref] Socioeconomic impact of asthma, chronic obstructive pulmonary disease and asthma-COPD overlap syndrome, Kim [/bib_ref] Systematic reviews and meta-analyses have also revealed more medical use including emergency department visits and hospitalization for ACO patients. [bib_ref] Asthma and COPD overlap syndrome (ACOS): a systematic review and meta analysis, Alshabanat [/bib_ref] [bib_ref] Clinical characteristics of the asthma-COPD overlap syndrome -a systematic review, Nielsen [/bib_ref] In contrast to these previous studies, Cosio et al 12 reported that ACO patients had better survival than COPD patients. In addition, a recent study reported that ACO with early onset of asthma has a better prognosis than COPD. However, ACO with late onset of asthma had poorer prognosis compared to COPD. [bib_ref] Long-term prognosis of asthma, chronic obstructive pulmonary disease, and asthma-chronic obstructive pulmonary..., Lange [/bib_ref] The collective data indicate that there may be heterogeneous groups within ACO patients and the prognosis can be determined accordingly. The aforementioned study showed that the outcome of ACO patients varies according to the time of asthma diagnosis, and our study demonstrated the heterogeneity of ACO according to smoking history. Although the COPD-predominant ACO group had higher household income, education levels, and health-related QoL than the asthma-predominant ACO group, COPD-predominant ACO was associated with higher outpatient medical cost compared to the asthma-predominant ACO group. Multivariate logistic regression analysis adjusted for age, sex, BMI, household income, and FEV 1 showed that the COPD-predominant ACO was associated with a higher rate of hospitalization than the asthma-predominant ACO group, although not statistically significant.
There are some limitations to this study. First, we used arbitrary definitions for the four groups, which did not fully meet the common diagnostic criteria. Because we used KNHANES and NHI data, bronchial provocation test or bronchodilator responsiveness and other respiratory imaging studies were not included. Therefore, the distinction between asthma and ACO proposed in this study were not accurate. It is also difficult to reflect the asthma control status, severity, or medication compliance. For example, asthma symptoms or control can vary depending on how steady the asthma medication is used and it may misclassified to a group of wheezing (−) or (+). This inaccurate assignment is likely to have weakened the statistical power of this study. Second, pulmonary function was measured only once at a specific time point and it could be affected by condition or short-term disease such as upper respiratory infection of patients. This may also be an obstacle to the precise distinction between W+ and W−. However, many previous population-based studies have used "wheezing" as a representative symptom of asthma. [bib_ref] Identification of subtypes in subjects with mild-to-moderate airflow limitation and its clinical..., Lee [/bib_ref] [bib_ref] Socioeconomic impact of asthma, chronic obstructive pulmonary disease and asthma-COPD overlap syndrome, Kim [/bib_ref] [bib_ref] Prevalence and risk factors of asthma and wheezing among us adults: an..., Arif [/bib_ref] [bib_ref] Lung function growth and its relation to airway hyperresponsiveness and recent wheeze...., Xuan [/bib_ref] To overcome these limitations, we used NHI data in conjunction with KNHANES data. NHI data reflect the actual illness and prescription, which can complement the aforementioned drawbacks. As confirmed in the study International Journal of COPD 2018:13 submit your manuscript | www.dovepress.com Dovepress Dovepress 1259 heterogeneity of asthma and COPD results, this classification seemed to represent two diseases to some extent, because the prescription patterns according to this classification were quite characteristic. In addition, it was a nationwide survey that had the advantage of being able to represent the whole population of a country and that included important variables associated with disease, such as socioeconomic status and QoL. Also, this study benefited from obtaining relatively accurate data on medical use and expenses by using NHI claim data. Second, although statistical analysis was performed by adjusting for sex through multiple logistic regression, sex difference associated with smoking history might be too large to overcome. Therefore, some of our results may be due to sex differences between groups and not differences between ACO subtypes.
Particularly, within ACO groups, two subtypes with distinct characteristics were found and they also showed marked differences in terms of exacerbation frequency and medical costs. While asthma-predominant ACO individuals displayed poorer socioeconomic status and QoL compared to the COPD-predominant ACO group, the COPD-predominant ACO group showed higher medical use and expenditure than the asthma-predominant ACO group. In this study, we found that ACO has more exacerbations than COPD with mild airway obstruction and that it can show various phenotypes with different prognoses. Therefore, in order to manage ACO patients well, it is necessary to recognize the characteristics well and appropriately allocate limited medical resources.
[fig] Figure 1: Study flow diagram. Abbreviations: FeV 1 , forced expiratory volume in 1 second; FVC, forced vital capacity; Knhanes, Korea national health and nutrition examination survey; nhI, national health Insurance. International Journal of COPD 2018:13 submit your manuscript | www.dovepress. [/fig]
[table] Table 1: epidemiological characteristics of the four groups according to the presence or absence of wheezing or smoking history Notes: Data are presented as mean ± sD or number (%). Comorbidities based on nhI data. *Comparison between s− and s+ groups within the W+ groups. **Comparison among the four groups. Abbreviations: BMI, body mass index; FeV 1 , forced expiratory volume in 1 second; FVC, forced vital capacity; nhI, national health Insurance; s, smoking; W, wheezing.International Journal of COPD 2018:13 submit your manuscript | www.dovepress.com [/table]
[table] Table 2: Physical activity and health-related quality of life of the four groups with mild-to-moderate airflow limitation Notes: Data are presented as mean ± sD or n (%). *Comparison between s− and s+ groups within the W+ groups. **Comparison among the four groups. "-" indicates no data. Abbreviations: EQ-5D, EuroQoL five dimensions questionnaire; EQ-VAS, EQ-5D visual analogue scale; S, smoking; W, wheezing. [/table]
[table] Table 3: health care utilization and insurance-related data among the four groups Notes: Data are presented as mean ± sD or number (%). *Comparison between s− and s+ groups within the W+ groups. **Comparison among the four groups. "-" indicates no data. Abbreviations: er, emergency room; ICU, intensive care unit; s, smoking; UsD, Us dollars; W, wheezing. [/table]
[table] Table 4: Prescription patterns of respiratory medication among the four groups [/table]
[table] Table 5: associations of each group with exacerbations [/table]
[table] Table 6: associations of the four subgroups or each variable with any exacerbation requiring hospitalization or emergency room visit [/table]
[table] Table 7: associations of the four groups with medical cost [/table]
[table] Table 8: association of wheezing subtypes with medical cost Notes: Model 1: unadjusted. Model 2: adjusted for age, sex, and BMI. Model 3: adjusted for age, sex, BMI, and household income. Model 4: adjusted for age, sex, BMI, household income, and forced expiratory volume in 1 second. Abbreviations: BMI, body mass index; s, smoking; W, wheezing. [/table]
|
Cyclophosphamide responsive primary angiitis of the CNS in a 61-year-old female
Dear Sir -A 61-year-old female with no significant past medical history presented with flu like symptoms including fever, chills and headaches after returning from a trip to Hawaii. First saw her primary care physician who advised conservative management. A couple of days later, she presented with worsening of her headaches and was admitted for meningitis work up. The neurologic exam was unremarkable. A lumbar puncture revealed a pleocytosis with 206 WBC and an elevated total protein of 75 mg/dl. PCR was negative for West Nile, HSV and enterovirus. Malaria smear, leptospira, and crypto-coccal antigen were also negative. A brain MRI was unremarkable except some slight periventricular T2 changes near the anterior horn of the lateral ventricle. She was eventually discharged with presumptive diagnosis of aseptic meningitis. However, she failed to improve with persistent nausea and unsteadiness of gait and vision gradually became blurry. A repeat brain MRI showed multifocal abnormal T2 prolongation in periventricular regions (Figure 1A, B) with contrast enhancement on T1-weighted images (Figure 1C). In addition, MRI of cervical, thoracic and lumbar spine showed significant T2 signal changes in the cervical and thoracic cord with enhancement, both patchy and confluent. She underwent a brain biopsy and the preliminary pathology was concerning for lymphoma or vasculitis. Intravenous methylprednisolone was initiated afterwards and the patient improved immediately.The pathology revealed white matter with perivascular cellular infiltrates. Lumens were nearly obliterated by reactive endothelial cells and, percolating through the blood vessel walls, were populations of predominantly small mature appearing lymphocytes (Figure 2A) accompanied by reactive microglia or macrophages. The inflammatory infiltrates showed no significant cytologic atypia. The blood vessel walls, while perturbed by the passing lymphocytes and microglia, did not show fibrinoid necrosis. No granulomas were identified. Surrounding parenchyma was hypercellular as a result of infiltration Figure 1. Axial Brain MRI sequences prior to treatment. A: T2 weighted image, B: FLAIR showing abnormal signal prolongation in periventricular areas, C: post contrast T1 image showing mild diffuse enhancement.
Dear Sir -A 61-year-old female with no significant past medical history presented with flu like symptoms including fever, chills and headaches after returning from a trip to Hawaii. First saw her primary care physician who advised conservative management. A couple of days later, she presented with worsening of her headaches and was admitted for meningitis work up. The neurologic exam was unremarkable. A lumbar puncture revealed a pleocytosis with 206 WBC and an elevated total protein of 75 mg/dl. PCR was negative for West Nile, HSV and enterovirus. Malaria smear, leptospira, and crypto-coccal antigen were also negative. A brain MRI was unremarkable except some slight periventricular T2 changes near the anterior horn of the lateral ventricle. She was eventually discharged with presumptive diagnosis of aseptic meningitis. However, she failed to improve with persistent nausea and unsteadiness of gait and vision gradually became blurry. A repeat brain MRI showed multifocal abnormal T2 prolongation in periventricular regions [fig_ref] Figure 1: Axial Brain MRI sequences prior to treatment [/fig_ref] , B) with contrast enhancement on T1-weighted images [fig_ref] Figure 1: Axial Brain MRI sequences prior to treatment [/fig_ref]. In addition, MRI of cervical, thoracic and lumbar spine showed significant T2 signal changes in the cervical and thoracic cord with enhancement, both patchy and confluent. She underwent a brain biopsy and the preliminary pathology was concerning for lymphoma or vasculitis. Intravenous methylprednisolone was initiated afterwards and the patient improved immediately.
The pathology revealed white matter with perivascular cellular infiltrates. Lumens were nearly obliterated by reactive endothelial cells and, percolating through the blood vessel walls, were populations of predominantly small mature appearing lymphocytes [fig_ref] Figure 2: A [/fig_ref] accompanied by reactive microglia or macrophages. The inflammatory infiltrates showed no significant cytologic atypia. The blood vessel walls, while perturbed by the passing lymphocytes and microglia, did not show fibrinoid necrosis. No granulomas were identified. Surrounding parenchyma was hypercellular as a result of infiltration The immunohistochemical staining for CD3 revealed that the angiocentric infiltrates were predominantly T-lymphocytes [fig_ref] Figure 2: A [/fig_ref]. Only rare CD20 positive B lymphocytes were present. These changes are most consistent with lymphocytic vasculitis.
She was started on pulse intravenous cyclophosphamide (cytoxan). At 1 year followup, she is feeling quite well, almost completely back to her normal self. A follow-up brain MRI demonstrated a significant resolution of the previously noted lesions [fig_ref] Figure 3: Axial Brain MRI sequences after treatment [/fig_ref].
Vasculitides, in general, are very broad spectrum and can affect any organs including brain either isolated or in the context of a systemic process such as lupus erythemato-sus or polyarteritis nodosa (PAN). However, primary CNS vasculitis or primary angitis of central nervous system (PACNS) is a rare disease characterized by inflammation and necrosis of walls of medium and small sized vessels in the CNS only [bib_ref] The role of imaging in the diagnosis of central nervous system vasculitis, Gomes [/bib_ref] [bib_ref] Diagnosis and treatment of cerebral vasculitis, Berlit [/bib_ref]. The clinical manifestations can be heterogeneous including headache, stroke, myelopathy and encephalopathy. Imaging and other para-clinical findings are nonspecific. Cerebrospinal fluid (CSF) pleocytosis and protein elevation are commonly seen. However, pathology is the gold standard diagnostic test. Spinal cord involvement is rarely seen in PACNS and most of the time is associated with brain lesions [bib_ref] Primary central nervous system vasculitis: analysis of 101 patients, Salvarani [/bib_ref] [bib_ref] Primary angiitis of the CNS with pure spinal cord involvement: a case..., Goertz [/bib_ref]. Goertz et al. [bib_ref] Primary angiitis of the CNS with pure spinal cord involvement: a case..., Goertz [/bib_ref] reported a patient with an isolated homogeneously enhancing lesion in the spinal cord which Clinical Neuropathology, Vol. 32 -No. 1/2013 -Letters to the editor turned out to be PACNS [bib_ref] Primary angiitis of the CNS with pure spinal cord involvement: a case..., Goertz [/bib_ref]. The pattern of spinal cord involvement in our case was multifocal with both confluent and patchy enhancement which makes it different from previously reported cases.
Current treatment recommendations are steroid therapy with intravenous pulse cyclophosphamide (cytoxan). Other alternative treatment options include: rituximab, azathioprine, methotrexate and mycophenolate mophetil [bib_ref] The role of imaging in the diagnosis of central nervous system vasculitis, Gomes [/bib_ref] [bib_ref] Diagnosis and treatment of cerebral vasculitis, Berlit [/bib_ref] [bib_ref] Primary central nervous system vasculitis: analysis of 101 patients, Salvarani [/bib_ref] [bib_ref] Primary angiitis of the CNS with pure spinal cord involvement: a case..., Goertz [/bib_ref] [bib_ref] Primary angiitis of the CNS -an enigma that needs world-wide efforts to..., Berlit [/bib_ref] [bib_ref] Primary angiitis of the central nervous system, Néel [/bib_ref]. The number of cases of PACNS reported in the literature thus far is about 700 [bib_ref] Diagnosis and treatment of cerebral vasculitis, Berlit [/bib_ref]. Given its multifarious clinical presentations, PACNS usually mimics other pathologies such as tumors. In our case, the initial impression was aseptic meningitis but later on the imaging findings and the frozen section report raised the suspicion of lymphoma. However the permanent pathology ascertained the diagnosis of CNS vasculitis and as there was no evidence of other organ involvement, PACNS is our final diagnosis. The patient initially responded to steroids quickly; however her gait and lower limb ataxia gradually worsened over a period of 3 weeks. At that time a pulse cyclophosphamide regimen was recommended. Pulse cyclophosphamide therapy, in addition to steroids, has been used for treatment of PACNS, however it is important to rule infectious causes of vasculitis before starting any immunosuppressive treatment [bib_ref] Diagnosis and treatment of cerebral vasculitis, Berlit [/bib_ref] [bib_ref] Cyclophosphamide treatment of primary angiitis of the central nervous system in children:..., Bitter [/bib_ref]. In our case, the extensive work up did not reveal any evidence of infectious etiologies. The relapse rate in PACNS is estimated to be around 25%, so it is important to keep in mind that a regular follow up is essential for a proper care in PACNS patients [bib_ref] Diagnosis and treatment of cerebral vasculitis, Berlit [/bib_ref]. We conclude that PACNS is a relatively rare pathology that is a great masquerader and creates diagnostic and therapeutic challenges. Pulse cyclophosphamide therapy has been recommended and should be considered early as a therapeutic option.
[fig] Figure 1: Axial Brain MRI sequences prior to treatment. A: T2 weighted image, B: FLAIR showing abnormal signal prolongation in periventricular areas, C: post contrast T1 image showing mild diffuse enhancement. Clinical Neuropathology, Vol. 32 -No. 1/2013 -Letter to the editor by scattered T-lymphocytes and reactive microglia accompanied by some fibrous gliosis. [/fig]
[fig] Figure 3: Axial Brain MRI sequences after treatment. A: T2 weighted image, B: FLAIR, C: post contrast T1 image showing resolution of previously noted lesions. [/fig]
[fig] Figure 2: A: Obliterated vascular lumen by reactive endothelial cells and infiltration of mature lymphocytes. B: Immunohistochemical staining for CD3 showing infiltration of T-lymphocytes. [/fig]
|
Tobacco smoking and risks of >470 diseases in China: a prospective cohort study
Background-Tobacco smoking is estimated to account for >1 million annual deaths in China, and the epidemic continues to increase in men. Large nationwide prospective studies linked to different health records can help assess periodically disease burden attributed to smoking. We examined associations of smoking with an extensive range of disease incidence and mortality in China.Methods-The prospective China Kadoorie Biobank recruited >512,000 adults aged 30-79 years from 10 diverse areas during 2004-2008, and recorded detailed smoking information. During 12year follow-up, 1,137,603 ICD-10 coded hospitalisation events and 48,760 deaths were recorded, involving 476 and 85 distinct conditions, each with ≥100 incident cases and deaths, respectively. Cox regression yielded adjusted hazard ratios (HRs) associating smoking with disease outcomes, adjusting for multiple-testing.Findings-At baseline 67·7% of men and 3·2% of women (overall 29·4%) ever smoked regularly. Compared with never-smokers, ever-smokers had significantly higher risks for 9 of 18 ICD-10 disease chapters examined at ages 35-84 years. For individual conditions, smokers had significantly higher risks of 56 diseases (men 50; women 24) and 22 causes of death (men 17; women 9). Among men, ever-smokers had HR of 1·09 (95% CI 1·08-1·11) for any disease incidence, and also significantly more episodes and longer duration of hospitalisation, particularly those due to cancer and respiratory diseases. For overall mortality, the HRs were greater in urban Chan et al. than in rural men (1·50 [1·43-1·58] vs. 1·25 [1·20-1·30]). Among urban men who began smoking at age <18 years, the HRs were 2·06 (1·89-2·24) for overall mortality and 1·32 (1·27-1·37) for any disease incidence. In this population, 20% of male (urban 24%; rural 16%) and 3% of female deaths were attributed to ever-regular smoking.Interpretation-Among Chinese adults, smoking was associated with higher risks of morbidity and mortality from a wide range of diseases. Among men, the future smoking-attributed disease burden will increase further, highlighting a pressing need for reducing consumption, through widespread cessation and uptake prevention.MethodsStudy populationDetails of the study design and baseline characteristics of CKB have been described elsewhere. 20,21 Briefly, 512,726 adults aged 30-79 years were recruited via multi-stage cluster sampling in 2004-2008 from 10 (four urban, five rural, and one semi-rural) Chan et al.
# Introduction
Worldwide smoking accounted for about 100 million deaths in the 20 th century and is projected to cause one billion deaths this century, mainly in low-and middle-income countries (LMICs), including China.China now consumes about 40% of the world's tobacco, almost exclusively by men.In China the main increase in consumption of manufactured cigarette took place after 1980 and continued until the 2010s, 2 many decades behind that in high-income Western countries. [bib_ref] The causes of cancer: quantitative estimates of avoidable risks of cancer in..., Doll [/bib_ref] Previous nationwide cohort studies established decades apart in China have reliably demonstrated the increasing proportion of adult mortality attributed to smoking in men. [bib_ref] Contrasting male and female trends in tobacco-attributed mortality in China: evidence from..., Chen [/bib_ref] However, the morbidity burden attributed to smoking from a much broader range of diseases has not been properly studied in China (and most other LMICs).
There is compelling evidence that smoking causes multiple diseases, chiefly cancer, cardiovascular disease (CVD) and chronic respiratory diseases.Nevertheless, more recent large prospective studies in high-income Western countries have also identified new diseases associated with smoking, including chronic kidney disease, influenza, and mental disorders. [bib_ref] Evidence of a Causal Relationship Between Smoking Tobacco and Schizophrenia Spectrum Disorders, Scott [/bib_ref] [bib_ref] Cigarette smoking and chronic kidney disease in the general population: a systematic..., Xia [/bib_ref] [bib_ref] Cigarette smoking and the occurrence of influenza -Systematic review, Lawrence [/bib_ref] [bib_ref] Cigarette smoking and the risk of systemic lupus erythematosus, overall and by..., Barbhaiya [/bib_ref] In most LMICs, the available prospective evidence on smoking hazards was chiefly confined to mortality. [bib_ref] Contrasting male and female trends in tobacco-attributed mortality in China: evidence from..., Chen [/bib_ref] [bib_ref] Tobacco Smoking and Mortality in Asia: A Pooled Meta-analysis, Yang [/bib_ref] [bib_ref] Quantifying the association of low-intensity and late initiation of tobacco smoking with..., Yang [/bib_ref] [bib_ref] Association of childhood smoking and adult mortality: prospective study of 120 000..., Thomson [/bib_ref] In China, although many studies have attempted to assess the effects of smoking on morbidity outcomes, they were constrained by the use of non-prospective study designs, small number of outcomes involved (mainly site-specific cancer, chronic obstructive pulmonary disease and CVD), restriction to specific urban cities, and by lack of objective validation of self-reported smoking status. [bib_ref] Smoking and chronic obstructive pulmonary disease in Chinese population: a meta-analysis, Wang [/bib_ref] [bib_ref] Epidemiological characteristics of 561 cases of intracerebral hemorrhage in Chengdu, Yu [/bib_ref] [bib_ref] Risk factors for gastric cancer: a large-scale, population-based case-control study, Zhang [/bib_ref] [bib_ref] Trends in smoking prevalence and implication for chronic diseases in China: serial..., Wang [/bib_ref] [bib_ref] Relation of active, passive, and quitting smoking with incident type 2 diabetes:..., Pan [/bib_ref] Reliable assessment of smoking attributed morbidity, over and above mortality, burden in different populations is needed to inform effective tobacco control nationally and globally.
To help fill the evidence gap, we undertook detailed analyses of smoking and risks of hospitalisations and death in the prospective China Kadoorie Biobank (CKB) of >512,000 adults, [bib_ref] Cohort profile: the Kadoorie Study of Chronic Disease in China (KSCDC), Chen [/bib_ref] [bib_ref] China Kadoorie Biobank of 0.5 million people: survey methods, baseline characteristics and..., Chen [/bib_ref] involving 476 distinct diseases and 85 specific causes of death across all organ systems. areas across China. These were selected through China's nationally representative Disease Surveillance Point System 22 to cover a diverse range of geographical areas, socioeconomic development, risk exposures and disease patterns. For each study site, about 100-150 administrative units (i.e. rural villages or urban street committees) were selected, and all eligible residents aged 35-74 years (n=1,801,167) were identified through local residential records and were invited to participate, of whom 499,439 (28%) participated, plus 13,287 individuals just outside the targeted age range (n=9817 at n=3470 at 75-79 years), resulting in 512,726 participants at baseline age range of 30-79 years.
In local study assessment clinics, trained health workers undertook physical measurements (e.g. height, weight, and blood pressure) and a laptop-based questionnaire interview covering socio-demographic status, lifestyle, environmental factors, female reproductive factors and medical history.
Ethical approvals were obtained by the Ethical Review Committee of the China National Center for Disease Control and Prevention and the Oxford Tropical Research Ethics Committee, University of Oxford prior to commencement of the field work. All participants provided written informed consent.
## Assessment of smoking
The questionnaire on current and past smoking behaviours included age first began to smoke regularly, frequency, amount and type of tobacco smoked, degree of inhalation, and for ex-smokers, age last stopped and main reasons for cessation (due to illnesses or other reasons). To validate self-reported smoking status (along with exposure to household air pollution), exhaled carbon monoxide was also measured using MicroCO meters (Carefusion, San Diego, CA, USA). [bib_ref] Exhaled carbon monoxide and its associations with smoking, indoor household air pollution..., Zhang [/bib_ref] In the present study, participants who had never smoked or smoked <100 cigarettes in their lifetime were classified as "never-regular smokers" and those who smoked ≥1cigarette (or ≥1 gram tobacco) daily for at least 6 months were classified as "ever-regular smokers" (or "ever-smokers" for simplicity). Among participants who had stopped for ≥6 months, those who had stopped smoking due to illnesses were grouped with baseline current smokers as "regular smokers" as they would have significantly elevated disease risk even after stopping, while those who had stopped voluntarily for other reasons (e.g. finance) constituted a separate category to assess the effects of cessation (before ill health). 6
## Follow-up for mortality and morbidity
Participants were followed up for death and any episodes of hospitalisation through electronic linkage via unique personal identification number to established mortality and morbidity (for cancer, stroke, ischaemic heart disease [IHD] and diabetes) registries and to national health insurance systems. All the reported disease events were coded following the International Classification of Diseases, 10 th Revision (ICD-10) by trained medical professionals blinded to baseline information. By 1 Jan 2018, 49,459 (9·6%) participants had died and 5302 (1·0%) were lost to follow-up. Participants were censored upon death, loss to follow-up, or 1 Jan 2018, whichever came first.
## Outcome measures
To enable a "phenome-wide" investigation, all disease events coded up to the first three characters of ICD-10 codes (i.e. the "disease category" component) were reviewed, and, where appropriate, combined (based on knowledge about the disease characteristics), to produce a list of distinct diseases. Several ICD-10 chapters considered irrelevant to the study population (e.g. perinatal-origin diseases [chapter XVI] and congenital conditions [XVII]) were excluded. Specific outcomes (or small number of specific disease groups) with at least 100 incident events (i.e. for those with multiple hospitalised events of same nature, only the first one was considered) or deaths (80 for sex-specific analyses) recorded during follow-up were analysed separately to capture a wide spectrum of conditions while ensuring reasonable statistical precision. Under each ICD-10 chapter, outcomes with <100 (<80 for sex-specific analyses) events were combined as "other disease" of the individual chapter to enable exploratory analysis that may shed light on rarer outcomes of the same disease system. Since the classification for the "other disease" endpoints were based on event number, the ICD-10 criteria used varies slightly in the combined and sex-specific analyses.
For overall mortality and major diseases constituting the top five causes of disability adjusted life years in China, [bib_ref] Global burden of 87 risk factors in 204 countries and territories, 1990-2019:..., Murray [/bib_ref] namely IHD (ICD-10 I20-I25), intracerebral haemorrhage (ICH; I61), ischaemic stroke (IS; I63), COPD (J41-J44), and lung cancer (C34), further indepth analyses by detailed smoking characteristics were also undertaken. The total number of hospitalisation episodes (i.e. including both first and subsequent events) and duration spent in hospitals (i.e. bed-days) attributed to all-cause, CVD, respiratory disease, cancer, and other causes were also examined as indicators of use of health service.
# Statistical analysis
We used Cox regression to estimate hazard ratios (HRs) comparing disease incidence or mortality risks at age-at-risk of 35-84 years in ever-(overall and by certain smoking characteristics) versus never-regular smokers, both overall and separately by sex (given the substantial sex-difference in smoking prevalence [bib_ref] Contrasting male and female trends in tobacco-attributed mortality in China: evidence from..., Chen [/bib_ref]. Ten participants died or were lost to follow-up before reaching 35 years of age, so a total of 512,712 participants were included in the analysis. All analyses were stratified by age-at-risk (5-year groups), study areas (10 sites), and, where appropriate, sex and adjusted for education (no formal school; primary school; middle or high school; college or university) and alcohol drinking (never, occasional, or ever regular). Among men, analyses were also conducted separately in urban (four) and rural/semi-rural (six) areas. All-cause mortality is a competing risk for disease events and cause-specific mortality. Hence, to facilitate observational analyses assessing aetiological questions, [bib_ref] Competing Risk Regression Models for Epidemiologic Data, Lau [/bib_ref] participants were censored at death from any cause to estimate cause-specific HRs, which compared event rates in participants who were alive and free of the event of interest.
For smoking exposures involving more than two categories, group-specific CIs of HRs were calculated using the variance of the log hazard in each category, including the reference group. This enables direct comparison of HRs across any two categories of exposure, instead of just between a fixed reference group and other exposure categories. [bib_ref] Improved estimates of floating absolute risk, Plummer [/bib_ref] The proportional hazards assumption for ever-regular smoking was investigated by plotting and testing correlations between transformed follow-up time and the scaled Schoenfeld residuals [bib_ref] Proportional hazards tests and diagnostics based on weighted residuals, Grambsch [/bib_ref] for incidence and mortality of the top-five major diseases, among men and women.
For disease outcomes showing significant associations with smoking, sensitivity analyses were undertaken with additional adjustment. For top-five major diseases, we further excluded participants with prior medical history of relevant conditions at baseline (e.g. excluding those with prior CVD in analyses of IHD/stroke). Moreover, we also conducted separate analyses among current smokers and ex-smokers who had stopped due to illnesses, who together formed the "regular smoker" category.
To assess the cumulative burden of smoking, the total number of hospitalisations and days in hospital were estimated for ever-versus never-regular smokers using the mean cumulative count. This estimates the total number of events occurring in a population and does not assume independence between hospitalisations and all-cause mortality. [bib_ref] Estimating the Burden of Recurrent Events in the Presence of Competing Risks:..., Dong [/bib_ref] [bib_ref] Nonparametric analysis of recurrent events and death, Ghosh [/bib_ref] [bib_ref] Marginal analysis of recurrent events and a terminating event, Cook [/bib_ref] Overall survival of regular versus never-regular smokers was also examined using Kaplan-Meier curves by sex. Adjusted incidence rates for individual diseases were calculated within age (<65, ≥65 years) and sex strata as
[formula] HR i × overall rate ∑ + n i × HR i ∑ + n i [/formula]
where HR i and n i are hazard ratios and number of events for ever-regular smoker and never-smokers. Overall incidence rates were calculated as weighted means of the ageand sex-specific rates, and total incidence rates are the sum of disease-specific incidence rates. The fraction of all deaths that is attributed to smoking in the study population (the population-attributed fraction [PAF]) was estimated by P(HR -1)/HR, where P is the prevalence of ever regular smoking among those dying during follow-up and HR is the associated hazard ratio for all-cause mortality.
Statistical significance (at the 5% level) was evaluated using both conventional and false discovery rate (FDR) adjusted p-values applied within ICD-10 chapters (and separately in the combined and sex-specific analyses), [bib_ref] Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple..., Benjamini [/bib_ref] the latter of which controls the expected proportion of false positives among all significant associations. [bib_ref] Using phenome-wide association studies to examine the effect of environmental exposures on..., Braun [/bib_ref] [bib_ref] False discovery rate control is a recommended alternative to Bonferroni-type adjustments in..., Glickman [/bib_ref] Unless otherwise specified, we only highlight associations that were statistically significant after FDR adjustment. All analyses were conducted using R software version 3·6·2.
## Role of the funding source
The study funders had no role in study design, data collection, analysis, interpretation, or writing of the report. KHC, NW, LL, and ZC had access to all data and had final responsibility for the decision to submit for publication.
# Results
Of the 512,716 participants included, 41·0% (n=210,201) were men, 66·3% (n=339,794) resided in rural areas, and the mean (SD) age was 52·0 (10·7) years at baseline. Overall, 29·4% (n= 150,801) smoked regularly, much higher in men (67·7%; n=142,205) than women (2·8%; n=8596). Both regular and ex-smokers tended to live in rural areas, to be less educated, more likely to consume alcohol and to have slightly lower BMI and higher prevalence of prior chronic diseases (eTable 1). Compared with female smokers, male smokers were more likely to start at a younger age, to be heavy smokers, and to habitually inhale tobacco smoke into the lungs. Among men, similar differences were also seen between younger and older smokers, with younger smokers more likely to start at age <20 and to smoke persistently manufactured cigarettes.
During 5·5 million person-year (median 12 years) of follow-up, 49,452 (9·6%) participants died and 285,888 (55·8%) were ever hospitalised at age-at-risk of 35-84 years. The "phenome-wide" investigation across 18 ICD-10 chapters included a total of 1,137,603 incident disease events from 476 different conditions, and 48,760 deaths from 85 causes . Of the 18 ICD-10 disease chapters examined, ever-smoking was associated with significantly elevated risks in nine chapters (and no significant associations for the rest), including infectious and parasitic diseases (HR=1·07, 95%CI 1·03-1·11), neoplasms (1·34, 1·30-1·38), endocrine, nutritional and metabolic diseases (1·05, 1·02-1·09), circulatory diseases (1·10, 1·08-1·12), respiratory diseases (1·18, 1·16-1·21), digestive diseases (1·03, 1·01-1·06) and diseases of skin and subcutaneous tissue (1·14, 1·06-1·22) (eFigure 1). For the corresponding mortality analyses, the HRs were generally greater than the morbidity findings (eFigure 2). Although the HRs tended to be somewhat larger in men, the patterns of associations were broadly similar in men and women. Overall, male ever-smokers had 9% (1·09, 1·08-1·11) excess risk of morbidity from any diseases (eFigure 1), higher in urban than rural areas (1·17 [1·15-1·20] vs 1·05 [1·03-1·06]) (eTables 2 and 3). Among women ever-smokers, the HR for any morbidity was 1·04 (1·01-1·06) (eFigure 1).
Across the 476 distinct conditions investigated, ever-smoking was associated with statistically significant higher risks of 87 diseases (or aggregates of similar diseases), with 56 associations remaining significant after FDR adjustment, including 10 different types of circulatory diseases, 14 respiratory diseases, 14 cancers, and 5 digestive diseases . In men, 50 different diseases were significantly positively associated (FDR-adjusted) with smoking (eTable 4), while among women, there were 24 such diseases (eTable 5).
In contrast, seven conditions showed significant inverse associations with smoking in sexcombined analyses (eight in men, none in women). No significant association was found with other conditions. [fig_ref] Figure 1: Adjusted HRs for cause-specific disease incidence significantly associated with everregular smoking [/fig_ref] shows the adjusted HRs for specific conditions showing FDR-significant positive associations with ever-smoking in sex-combined or sex-specific analyses. Overall in the combined analyses, the adjusted HRs ranged from 1·06 (1·02-1·09) for diabetes mellitus (ICD-10: E10-E14) to 3·16 (1·98-5·05) for larynx cancer (C32) [fig_ref] Figure 1: Adjusted HRs for cause-specific disease incidence significantly associated with everregular smoking [/fig_ref]. The HRs were generally higher in men than in women, with a few exceptions including larynx cancer For the five leading diseases, male smokers who had started smoking at younger ages or smoked more cigarettes (or equivalents) per day had consistently higher HRs for both morbidity and mortality in a dose-response manner (eTable 7) and also separately in urban and rural men . Moreover, the HRs were typically higher in urban than in rural men and for mortality than morbidity (eTable 8). Similar, albeit less extreme, findings were also evidence in women (eTable 7). Sensitivity analyses with exclusion of individuals with a prior history of specific diseases yielded similar results (eTables 9 and 10).
For overall mortality, the adjusted HRs were 1·33 (1·29-1·37) in male ever-smokers, higher in urban than in rural areas (1·50 [1·42-1·58] versus 1·25 [1·20-1·30]) (eTables 2, 3, 11). Among women it was 1·44 (1·36-1·52) (eTable 12), again higher in urban than rural areas (1·53 [1·41-1·66] versus 1·38 [1·29-1·48]). These HRs were not materially altered after exclusion of participants with prior cancer at baseline (eTables 2, 3, 11, 12).
Moreover, regular-smokers who started smoking regularly at younger age had higher risks for major disease morbidity and mortality, particularly in urban areas [fig_ref] Figure 2: Adjusted HRs for risks of selected major disease incidence and mortality in... [/fig_ref]. Among urban men who began smoking at <18 years, the HRs were 2·06 (1·89-2·24) for overall mortality and 1·32 (1·27-1·37) for any disease incidence [fig_ref] Figure 2: Adjusted HRs for risks of selected major disease incidence and mortality in... [/fig_ref] , increasing to 2·35 (2·07-2·67) and 1·40 (1·30-1·50), respectively, among those who began at <15 years. Compared to never-smokers, ever-smokers in either sex had worse survival starting at age of 55 years, with smokers reaching 50% survival ~3.5 years earlier (eFigure 23). After accounting for excess mortality (i.e. competing risk), ever-smokers had significantly higher total expected hospitalisations and (to a lesser extent) longer stay in hospital compared with never-smokers (eFigure 24), especially in men and for those attributed to cancer and respiratory disease (eFigure 25). The differences broadly increased along with age-at-risk, with apparent divergence beginning at around 55 to 60 years of age.
Among male ex-smokers who had quitted due to illness, there were significant excess risks of overall mortality (1·61, 1·55-1·66) and morbidity (1·25, 1·22-1·27) (eTable 11), which, although decreasing gradually, persisted beyond 15 years after quitting at baseline (eFigure 26). In contrast, those who had stopped voluntarily (i.e. before developing major diseases) only had small excess risks for overall mortality (1·06, 1·01-1·11) and morbidity (1·05, 1·03-1·08), with the risks approaching those among never smokers after about 5-10 years of quitting. Similarly contrasting differences between the two groups of ex-smokers were also observed for the top-five disease incidence and mortality (eTables [bib_ref] Tobacco Smoking and Mortality in Asia: A Pooled Meta-analysis, Yang [/bib_ref] [bib_ref] Quantifying the association of low-intensity and late initiation of tobacco smoking with..., Yang [/bib_ref].
When considering all FDR-adjusted significant positive (n=56) and inverse (n=7) associations simultaneously, smoking was associated with an additional 13 (men 20; women 10) events for each one event prevented, corresponding to a net absolute excess of 1,217 (men 1,473; women 1,046) events per 100,000 person-years [fig_ref] Figure 3: Incidence and mortality rates from all diseases found to have a FDR-adjusted... [/fig_ref]. For mortality, smoking was associated with a net excess of 230 (men 283; women 194) deaths per 100,000 person-years. Overall in this population, 19·6% of deaths in men (24·3% urban, 16·2% rural) and 2·8% of deaths in women at ages 35-84 years could be attributed to smoking if all the FDR-adjusted significant associations were causal.
# Discussion
This study provided the first comprehensive assessment of the long-term health effects of tobacco smoking on a wide range of diseases in adult men and women in China. Overall, smoking was significantly associated with higher risks of 22 causes of death and 56 individual diseases across all major organ systems, as well as more episodes and longer durations of hospitalisation. The associations were stronger in urban than in rural areas, and in those who started at younger age and smoked larger amount of tobacco. Although the relative risks associated with smoking were still relatively modest for most diseases, among urban men who started smoking before 18 years of age, the HR for overall mortality already approached those observed in high-income Western populations where most smokers started during young adulthood. Furthermore, we also showed that stopping smoking before the onset of major illness is remarkably beneficial.
Decades of epidemiological studies in many Western countries such as the UK and USA have demonstrated substantial hazards of tobacco smoking.diseases (and nine causes of injuries) were likely associated with smoking (see eTable 13). [bib_ref] Spatial, temporal, and demographic patterns in prevalence of smoking tobacco use and..., Reitsma [/bib_ref] For many other diseases (e.g. ICH, heart failure, influenza, or gastrointestinal conditions), the existing evidence is less conclusive, coming mainly from case-control or small cohort studies. [bib_ref] Cigarette smoking and the occurrence of influenza -Systematic review, Lawrence [/bib_ref] [bib_ref] Tobacco Use: A Major Risk Factor of Intracerebral Hemorrhage, Cho [/bib_ref] [bib_ref] Influence of Smoking Status on Risk of Incident Heart Failure: A Systematic..., Lee [/bib_ref] In China and other LMICs, most existing prospective evidence was mainly on mortality from several major diseases. [bib_ref] Contrasting male and female trends in tobacco-attributed mortality in China: evidence from..., Chen [/bib_ref] [bib_ref] Tobacco Smoking and Mortality in Asia: A Pooled Meta-analysis, Yang [/bib_ref] [bib_ref] Quantifying the association of low-intensity and late initiation of tobacco smoking with..., Yang [/bib_ref] [bib_ref] Association of childhood smoking and adult mortality: prospective study of 120 000..., Thomson [/bib_ref] [bib_ref] Cohort profile: The Chennai Prospective Study of Mortality among 500 000 adults..., Gajalakshmi [/bib_ref] In this large prospective study, we found smoking to be associated with increased mortality from 22 causes and increased morbidity from 56 conditions, including many that are still largely under-studied in LMICs (e.g. ICH, heart failure, aortic aneurysm, pulmonary embolism, peptic ulcer).Most of these associations overlap with those reported consistently in previous cohort studies of Western populations.
Long-term prospective studies in Western countries have also demonstrated clearly the long delay between widespread uptake of cigarette smoking in young adult population, first in men then women, and rising morbidity and mortality risks subsequently. [bib_ref] 50-year trends in smoking-related mortality in the United States, Thun [/bib_ref] In these Western populations where the smoking epidemic has matured, most adult smokers started at a young age, smoked a large amount, and persistently consumed manufactured cigarette (as opposed to traditional tobacco products) for decades, which are the hallmark of "high risk smoking patterns". Therefore, contemporary studies in those populations could reliably capture the "full effects" of prolonged cigarette smoking. [bib_ref] The 21st century hazards of smoking and benefits of stopping: a prospective..., Pirie [/bib_ref] In contrast, as the widespread uptake and peak of cigarette smoking in China was much more recent, adult smokers in CKB who were born mainly before 1970s had a later starting age, consumed fewer cigarettes daily, and used more traditional tobacco products (which are less harmful than cigarettes) compared to Western smokers [bib_ref] Contrasting male and female trends in tobacco-attributed mortality in China: evidence from..., Chen [/bib_ref] [bib_ref] Tobacco Smoking and Mortality in Asia: A Pooled Meta-analysis, Yang [/bib_ref] or younger smokers (i.e. those born after 1970s) in China. These differences in smoking patterns likely explain the generally weaker associations in CKB than those reported in Western studies. [bib_ref] The 21st century hazards of smoking and benefits of stopping: a prospective..., Pirie [/bib_ref] [bib_ref] Smoking and mortality--beyond established causes, Carter [/bib_ref] Our findings are, however, highly consistent with previous studies in China [bib_ref] Contrasting male and female trends in tobacco-attributed mortality in China: evidence from..., Chen [/bib_ref] [bib_ref] Tobacco Smoking and Mortality in Asia: A Pooled Meta-analysis, Yang [/bib_ref] [bib_ref] Quantifying the association of low-intensity and late initiation of tobacco smoking with..., Yang [/bib_ref] and other LMICs where the widespread use of cigarette is relatively recent. [bib_ref] Tobacco Smoking and Mortality in Asia: A Pooled Meta-analysis, Yang [/bib_ref] [bib_ref] Quantifying the association of low-intensity and late initiation of tobacco smoking with..., Yang [/bib_ref] Similarly, the greater excess risks for most diseases in urban than rural men smokers likely reflect more advanced smoking epidemic in urban than rural China, as cigarettes had been less available and affordable in rural areas until recent decades.These may also partly explain the lack of apparent associations in the present study for certain conditions known to be linked, albeit modestly, to smoking (e.g. prostate cancer, colorectal cancer, dementia). Likewise, the elevated background disease risk for many diseases, especially in rural areas, due to greater exposure to traditional risk factors (e.g. household air pollution, chronic infection) may also play a role. For certain diseases (e.g. ICH) that are much less common in Western populations, previous studies have reported mixed findings. [bib_ref] Tobacco Use: A Major Risk Factor of Intracerebral Hemorrhage, Cho [/bib_ref] With a much larger number of well-characterised ICH cases (>10,000; >80% confirmed by imaging), we found significantly elevated risks of ICH mortality and incidence, particularly among urban men.
Among women, despite the low smoking prevalence (<3%) and intensity among smokers, we found significant excess risks of 20 distinct conditions. With a few exceptions (e.g. lung cancer, COPD), the HRs for most disease-specific morbidities, but not necessarily mortalities, were more modest in female than male smokers. The lack of apparent sex difference in HRs for lung cancer and COPD, corroborates to previous findings, [bib_ref] Smoking and mortality--beyond established causes, Carter [/bib_ref] [bib_ref] Women and COPD: do we need more evidence?, Gut-Gobert [/bib_ref] suggesting that women's respiratory system appears more vulnerable to the harm of smoking. This may partly reflect the potential protective role of oestrogen in these diseases and the anti-oestrogenic effects of tobacco smoke 45 as well as genetic predisposition of early onset COPD and lung cancer in women. [bib_ref] Women and COPD: do we need more evidence?, Gut-Gobert [/bib_ref] For breast cancer we did not find any significant associations, consistent with previous Chinese studies, [bib_ref] Active and passive smoking with breast cancer risk for Chinese females: a..., Chen [/bib_ref] but not with those in Western populations, which generally showed positive association. [bib_ref] The 21st century hazards of smoking and benefits of stopping: a prospective..., Pirie [/bib_ref] However, the recorded case numbers among smokers in this and other Chinese studies were very small.
Previous Western studies have also reported lower risks of certain diseases among smokers, particularly Parkinson's disease and endometrium cancer.For Parkinson's disease, we have also found a significant 27% lower risk among smokers. For the other conditions, we found no clear associations with smoking, but the case numbers were small (e.g. only 15 corpus uteri cancer cases in female smokers). For six other conditions, however, we found lower risks among smokers, but there was very limited previous evidence. Among these, varicose veins, inguinal hernia, gonarthrosis, and other arthrosis are all strongly linked to family history that cannot be controlled for in CKB; [bib_ref] Clinical and Genetic Determinants of Varicose Veins, Fukaya [/bib_ref] [bib_ref] Smoking is a risk factor for recurrence of groin hernia, Sorensen [/bib_ref] [bib_ref] Tobacco use is not associated with groin hernia repair, a population-based study, Hemberg [/bib_ref] and all except inguinal hernia have been associated with higher BMI, 47-51 while smoking is known to have a weight reduction effect. Further adjustment for BMI, occupation and self-rated health, however, did not change the associations materially, suggesting no strong residual confounding by these factors. Bronchiectasis is associated with severe lung infections, asthma and cystic fibrosis but not known to be linked to smoking.Reverse causation bias can arise from the inclusion of diseased individuals who were less likely to start or more likely to quit early in life. Similar bias may also apply to conjunctiva disorders, which tend to develop during early adulthood with recurrent irritating symptoms or even vision impairment. Sensitivity analyses to exclude the first five years of follow-up, however, did not alter the results materially (eTable 14). Nonetheless, the inverse associations could also represent chance findings, given the large number of tests conducted. With the lack of prior evidence, our findings could only be considered as hypothesis-generating, which require further verifications in other studies.
Our study has several strengths, including large sample size, reliable assessment of smoking exposure, completeness of follow-up, and the broad range of diseases included. However, it also has limitations. First, for rare conditions (e.g. leukaemia, peripheral artery disease) the case numbers remain small, especially for women. Second, we only assessed prior medical history of ~20 major diseases, so we were unable to account fully for reverse causality for many other outcomes. Third, the use of FDR adjustment may have obscured any true but modest associations for certain conditions (e.g. hyperthyroidism) that have been associated with smoking. Fourth, although the baseline smoking habits had been objectively validated, 23 subsequent changes in smoking patterns (e.g. quitting) could yield underestimated HRs. In the 2013-14 resurvey of a random subset of ~25,000 surviving participants, we found that about one-fifth of baseline smokers stopped subsequently. Despite this, we could not directly assess the likely resulting underestimation of risks associated with changes in smoking patterns. Fifth, the record linkage may miss milder (e.g. influenza) or underdiagnosed conditions (e.g. COPD, dementia), resulting in underestimation of the disease burden associated with smoking. Sixth, CKB was not nationally representative, and the study participation was voluntary, so healthy volunteer bias is inevitable. However, the large sample size, diverse areas covered, heterogeneity of exposure, and highly consistent smoking patterns with those reported in other representative surveys in China 6,18 means that our HRs for smoking could still be largely generalisable to the Chinese population. [bib_ref] Enhancing the feasibility of large cohort studies, Manolio [/bib_ref] [bib_ref] Why representativeness should be avoided, Rothman [/bib_ref] Finally, unlike for mortality, we were not able to estimate the total disease morbidity burden attributed to smoking in China, due to the lack of nationally representative disease incidence data and regional variations in accessing health service.
In recent decades various estimates have been made about the smoking-attributed mortality burden in China. We showed that smoking now accounted for 20% and 3% of male and female deaths respectively at age 35-84 years. This would translate into >1 million total deaths per year in China, in line with our previous projection, 6 but much lower than that estimated in the 2019 GBD study (2·1 million men and 0·3 million women). [bib_ref] Spatial, temporal, and demographic patterns in prevalence of smoking tobacco use and..., Reitsma [/bib_ref] The GBD estimates would suggest that about half of all adult male deaths in China (~4·5 million in 2019) could be attributed to smoking. The GBD report used relative risk estimates for 36 causes of death (as opposed to all-cause mortality in this study) derived from pooled analyses of mainly Western cohort and case-control studies. [bib_ref] Spatial, temporal, and demographic patterns in prevalence of smoking tobacco use and..., Reitsma [/bib_ref] Although various adjustments were made, [bib_ref] Spatial, temporal, and demographic patterns in prevalence of smoking tobacco use and..., Reitsma [/bib_ref] it did not appear to fully account for the delayed effects and large urban-rural differences of recent cigarette uptake in China. Nevertheless, our study provided reliable evidence that if the current trends in smoking persist, the future smoking attributed disease burden in Chinese men is likely to increase markedly, whereas that in women it will likely remain low and continue to decline. [bib_ref] Contrasting male and female trends in tobacco-attributed mortality in China: evidence from..., Chen [/bib_ref] In summary, the present study demonstrated substantial hazards of smoking from a wide range of conditions among Chinese men and women. Given the delayed effects, the future risk per individual smoker and overall disease burden attributed to smoking among adult Chinese men will be much greater, especially in those who were born in the 1970s-1980s, who have reached adulthood when nationwide cigarette consumption was high. As demonstrated in the present and many previous studies, [bib_ref] Contrasting male and female trends in tobacco-attributed mortality in China: evidence from..., Chen [/bib_ref] [bib_ref] The 21st century hazards of smoking and benefits of stopping: a prospective..., Pirie [/bib_ref] stopping smoking before the onset of major illness is remarkably beneficial, and widespread smoking cessation, facilitated through increased tobacco tax, effective package warnings, and cessation clinics and helplines,offer China one of the most effective strategies to control the rising burden of chronic diseases over the next few decades.
# Supplementary material
Refer to Web version on PubMed Central for supplementary material.
## Research in context
## Evidence before this study
We searched the PubMed database for relevant prospective cohort studies in China published in English before 1 Jan 2022, using the terms "smoking OR tobacco" AND "China" AND "mortality OR death OR incidence OR risk". The majority of the previous studies were conducted before 2000 and focused only on mortality from several major diseases such as cancer, cardiovascular disease, and chronic respiratory diseases. A few large prospective cohort studies (each with >100,000 participants) in China, including our earlier reports based on the China Kadoorie Biobank also demonstrated significant associations of smoking with incidence of several major diseases including site-specific cancer, and diabetes. Additionally, a recent pooled analysis of six prospective cohorts with >330,000 primarily Chinese adults found significantly higher risk of nasopharyngeal cancer associated with smoking. However, there is no report in China assessing systematically the associations of smoking with a broader range of fatal and non-fatal diseases. Outside of China, only a few large prospective studies, such as the UK Million Women Study (1·3 million women) and pooled data from several US cohort studies (involving 0·5-1 million adults), have examined the associations of smoking with risks of mortality, but not incidence, from a wide range of disease outcomes simultaneously. Overall, these large UK and US studies reported significant associations of smoking with 23 and 35 causes of deaths, respectively.
## Added value of this study
To our knowledge, this is the first large nationwide prospective cohort study in China to systematically examine the associations of smoking with diseases affecting all body systems. After accounting for multiple-testing, smoking was significantly associated with higher disease incidence from 56 (man 50, women 24) specific diseases (10 cardiovascular, 14 respiratory, 14 cancer, five digestive, 13 others), and with higher risks of death from 22 (men 17, women 9) specific causes. This study provides important new evidence on the health effects of smoking beyond mortality from major diseases in China. It also provides additional evidence demonstrating greater relative risk of morbidity and mortality among urban male smokers, especially among those who started young. Furthermore, for the first time, we estimated the excess total and cause-specific (i.e. cardiovascular, respiratory and cancer) hospitalisation episodes and duration of hospitalisation associated with smoking in China. Apart from hazards of smoking, this study also demonstrated that stopping smoking before the onset of major illness is remarkably beneficial.
## Implications of all the available evidence
Decades of epidemiological research have produced strong evidence on the adverse health effects of smoking on an extensive range of major diseases. In China and many other low-and middle-income countries, the future disease morbidity and mortality burden attributed to smoking will likely increase. Large prospective cohort studies with reliable electronic linkage not only to mortality but also to hospital records in diverse populations could help to assess and monitor the evolution of the tobacco epidemic by providing population-specific estimates on the current and future disease burden attributed to smoking. Such evidence is crucial to inform appropriate policy action in tobacco control nationally and globally. Hazard ratios (HRs) were stratified by age-at-risk (5-year groups), sex, and study area and were adjusted for education and alcohol drinking. All analyse were restricted to age-at-risk range of 35-84 years. The solid boxes represent HRs, with the size inversely proportional to the variance of the logarithm of the HR, and the horizontal lines represent 95% confidence intervals. The individual diseases listed included all that showed FDR-adjusted significant associations with smoking, in overall or sex-specific analyses. The solid black and grey boxes indicate FDR-adjusted significant and non-significant associations, respectively. Number of morbidity and mortality events by ICD-10 chapter and their overall associations with ever-regular smoking, for men and women combined
[fig] Figure 1: Adjusted HRs for cause-specific disease incidence significantly associated with everregular smoking [/fig]
[fig] Figure 2: Adjusted HRs for risks of selected major disease incidence and mortality in urban and rural men who started smoking before 20 years old Health. Author manuscript; available in PMC 2022 December 14. [/fig]
[fig] Figure 3: Incidence and mortality rates from all diseases found to have a FDR-adjusted significant association with ever-regular smokingThe bar diagrams indicate the overall absolute morbidity and mortality rates per 100,000 person-years at age-at-risk 35-84 years in never-versus ever-regular smokers, overall and in men and women separately. Within each bar diagram, separate rates were also shown for circulatory (black), respiratory (grey) and other diseases (white). The morbidity analyses included 56 diseases showing positive associations with smoking and seven showing inverse associations. The mortality analyses were based on the 22 causes of death showing significant positive associations with smoking. Health. Author manuscript; available in PMC 2022 December 14. [/fig]
|
Percutaneous closure of residual shunting in a patient with a fenestrated atrial septal defect occluder
Rationale: Fenestrated atrial septal defect (ASD) occlusion has been performed in patients complicated with severe pulmonary hypertension (PH). Nevertheless, the persistent interatrial residual shunting in the fenestration might increase the risk of paradoxical embolism. Percutaneous closure of fenestrated ASD occluder (ASO) has not yet been reported.Patient concerns: A 26-year-old patient with a 25-mm ASD and severe PH underwent ASD closure using a Memory ASO with a waist of 32 and 6-mm custom-made fenestration. Echocardiography revealed the fenestration remained 6 mm and the pulmonary artery pressure decreased to the normal range at the 6-month follow-up.Diagnoses: Persistent interatrial residual shunting in ASO.Interventions: Percutaneous closure of residual interatrial shunting was performed using a waist of 8-mm ASO under guidance of TEE.Outcomes: The fenestration was successfully closed. Neither thromboembolism nor infection events were noted during the 12-month follow-up after the procedure.Lessons: This case illustrates that percutaneous closure of the residual shunting in fenestrated ASO was feasible and safe for shortand long-term.Abbreviations: ASD = atrial septal defect, ASO = atrial septal defect occluder, PH = pulmonary hypertension.
# Introduction
Secundum-type atrial septal defect (ASD) is one of the most common types of congenital heart diseases in adulthood. Compared with the conventional surgical closure, percutaneous closure of ASD provides an effective, minimally invasive, and safe alternative for patients with ASD. [bib_ref] Secundum atrial septal defect: nonoperative closure during cardiac catheterization, King [/bib_ref] Patients with ASDs and pulmonary hypertension (PH) might require atrial communication as decompression in the event of PH crisis or left ventricular diastolic dysfunction. [bib_ref] Shunt reduction with a fenestrated Amplatzer device, Kretschmar [/bib_ref] [bib_ref] Left ventricular conditioning in the elderly patient to prevent congestive heart failure..., Schubert [/bib_ref] Partial occlusion with fenestrated ASD occluder (ASO) has been suggested to be beneficial in patients with severe PH, and it might be superior than complete ASD closure. [bib_ref] Shunt reduction with a fenestrated Amplatzer device, Kretschmar [/bib_ref] [bib_ref] Fenestrated occluders for treatment of ASD in elderly patients with pulmonary hypertension..., Bruch [/bib_ref] However, the fenestration might remain patent in the long-term follow-up. [bib_ref] Fenestrated transcatheter ASD closure in adults with diastolic dysfunction and/or pulmonary hypertension:..., Abdelkarim [/bib_ref] The persistent residual shunting may increase the risk of paradoxical embolism. Herein, we report a case of a patient with successful deployment of the second ASO for closure of the residual shunting in the fenestrated ASO implanted previously.
## Case report
A 26-year-old female, recently diagnosed with secundum-type ASD accompanied by severe PH, was referred to our hospital for therapy. She presented a 1-year history of symptoms, including exertional chest distress and dyspnea. The second heart sound was prominent, and a grade 2/6 systolic murmur was heard at the left midsternal border during physical examination. Her functional class was assessed to be NYHA II. Any marked abnormality was not detected in the laboratory tests. Transthoracic echocardiography (TTE) revealed the presence of a moderately large secundum-type ASD, sized 25 mm. Both aortic and posterior left atrium rims were small (2 mm), while the other rims were sufficient. Also, a bidirectional interatrial shunt was observed, predominantly from left to right. In addition, the right chambers and pulmonary artery were dilated. The diameters of the right atrium and right ventricle end-diastolic were 45 and 51 mm, respectively. The right and left ventricular functions were normal. Tricuspid regurgitation was moderate-to-severe. The pulmonary artery systolic pressure (PASP) gradient was 110 mm Hg as estimated by the Doppler method. Diagnostic catheterization revealed a significantly elevated mean pulmonary artery pressure (mPAP) of 55 mm Hg, and pulmonary vascular resistance index (PVRi) of 4 Wood/U/m 2 .
The patient provided written consent and then underwent sufficient preoperative preparation, including the administration of Remodulin for reducing the pulmonary arterial pressure and oxygen inhalation. Considering the severe PH, Memory ASO (Shape Memory Alloy Co., Ltd, Shanghai, China) with a waist of 32 and 6-mm custom-made fenestration [fig_ref] Figure 1: Actual Memory ASD fenestrated device and 3-dimensional transesophageal echocardiography [/fig_ref] was subsequently inserted by standard ASD closure technique under the guidance of intraoperative transesophageal echocardiography (TEE). Consequently, the interarterial shunt was successfully diminished. A mild bidirectional shunt through the fenestration could be detected by TEE; the right and left ventricular functions remained stable. After the procedure, the patient was treated with Remodulin and aspirin, and then, discharged on day 3, as the recovery was uneventful. However, Aspirin was prescribed for 3 months.
The patient's symptoms improved significantly as noted at the 3-month follow-up. TTE evaluation demonstrated excellent ASO position and a residual left-to-right shunt across the fenestration. PASP decreased to 71 mm Hg. Tricuspid regurgitation improved from mild-to-moderate degree. The enlarged right atrium and ventricle also decreased in size (from 45 to 40 mm and from 51 to 44 mm, respectively).
Half a year after the procedure, TTE revealed normal size of the right heart chambers. A further decrease in PASP was also observed, up to 34 mm Hg. The residual shunt did not exert a significant influence according to hemodynamics. Considering the possibility of paradoxical embolism and patient's concern, complete closure of residual atrial shunting was performed under general anesthesia and guidance of TEE, which estimated the diameter of fenestration in the device as 6 mm. The guidewire moved from inferior vena cava to right atrium, traversing the ASO fenestration, and lodged into the left atrium. Venous access was obtained with 8F sheath via the wire. After the withdrawal of the wire, a standard ASO (Yachuang Biotech Co., Ltd, Xuzhou, China) with 8 mm waist was delivered from the sheath. The left and right discs were successively deployed, and the small ASO overlapped the previously large discs on both sides [fig_ref] Figure 2: Color compared transesophageal echocardiography [/fig_ref]. The position relationships among the sheath, fenestration, and the center hub of the device could be visually displayed by 3dimensional (3D) TEE [fig_ref] Figure 1: Actual Memory ASD fenestrated device and 3-dimensional transesophageal echocardiography [/fig_ref]. The patient was treated by Aspirin after the second procedure. Pre-discharge 24-hour TTE revealed the stationary position of the 2 devices. No interatrial residual shunt was observed. In addition, no thrombi or vegetation was detected on the devices.
Aspirin was prescribed for 3 months after discharge. No thromboembolism or infection events occurred during the 12-month follow-up after the second closure.
The patient consented to publish this case report, which was also approved by the Institutional Review Board of Shanghai Jiao Tong University Affiliated Sixth People's Hospital, Shanghai, China.
# Discussion
Complete closure of the ASD might aggravate the right heart overload in patients with significant PH. A patent foramen ovale (PFO) allowing a pop-off to the left heart may ameliorate the effect of increased overload on the right ventricular function. [bib_ref] Factors that influence the outcome of primary pulmonary hypertension, Rozkovec [/bib_ref] Similarly, the fenestration in ASO could allow a blood overflow in the right-to-left direction as a reduction of right heart overextension in such patients. Therefore, the transcatheterfenestrated ASD closure is considered to be advantageous to patients with severe PH. In order to maintain the interatrial communication patent, fenestrated ASOs were manufactured in various ways, [bib_ref] Shunt reduction with a fenestrated Amplatzer device, Kretschmar [/bib_ref] [bib_ref] Fenestrated occluders for treatment of ASD in elderly patients with pulmonary hypertension..., Bruch [/bib_ref] [bib_ref] Fenestrated transcatheter ASD closure in adults with diastolic dysfunction and/or pulmonary hypertension:..., Abdelkarim [/bib_ref] [bib_ref] Fate of a modified fenestration of atrial septal occluder device after transcatheter..., Schneider [/bib_ref] [bib_ref] Modified technique of stent fenestration of the atrial septum, Stümper [/bib_ref] [bib_ref] Self-fabricated fenestrated Amplatzer occluders for transcatheter closure of atrial septal defect in..., Peters [/bib_ref] which have been shown to be safe and associated with improvements in functional status and pulmonary arterial pressure. [bib_ref] Fenestrated transcatheter ASD closure in adults with diastolic dysfunction and/or pulmonary hypertension:..., Abdelkarim [/bib_ref] Despite that a few ASO fenestrations may close spontaneously after implantation during follow-up, most of the interatrial communications remained open over a prolonged duration in previous cases. [bib_ref] Shunt reduction with a fenestrated Amplatzer device, Kretschmar [/bib_ref] [bib_ref] Fenestrated occluders for treatment of ASD in elderly patients with pulmonary hypertension..., Bruch [/bib_ref] [bib_ref] Fate of a modified fenestration of atrial septal occluder device after transcatheter..., Schneider [/bib_ref] [bib_ref] Modified technique of stent fenestration of the atrial septum, Stümper [/bib_ref] [bib_ref] Self-fabricated fenestrated Amplatzer occluders for transcatheter closure of atrial septal defect in..., Peters [/bib_ref] [bib_ref] Atrial septal defect closure with Occlutech( ASD fenestrated device in a child..., Gonzalez-Barlatay [/bib_ref] Nevertheless, the longstanding fenestration may not be necessary later when the severe PH and right heart overload is improved. Moreover, as the residual shunting across the fenestration may cause a paradoxical embolism. Thus, complete closure of the ASO fenestration should be beneficial to the patients, especially to those with a high risk of peripheral venous thrombosis, such as pregnant women. Venous thrombosis can be secondary to a prothrombotic state in the hemostatic system associated with pregnancy. Compared with the nonpregnant state, the risk of venous thromboembolism was 5-fold higher in pregnant cohort,which significantly increased the risk of paradoxical embolism through atrial communication. A paradoxical embolism can lead to stroke, which is a catastrophic event in young patients. Considering that the current patient was of childbearing age, facing an increased risk of paradoxical embolism if pregnant, the fenestration was completely closed in the absence of the need for decompression in the right heart. To the best of our knowledge, this is the first report describing the percutaneous closure of the residual shunting in the fenestrated ASO. The complete closure was performed 6 months after the fenestrated ASO implantation. Both Doppler-estimated pulmonary arterial pressure and size of the right heart chambers were within the normal range according to TTE. Nevertheless, the improvement of PH was observed during >6-month followup in a majority of the previous cases. [bib_ref] Fenestrated occluders for treatment of ASD in elderly patients with pulmonary hypertension..., Bruch [/bib_ref] [bib_ref] Fenestrated transcatheter ASD closure in adults with diastolic dysfunction and/or pulmonary hypertension:..., Abdelkarim [/bib_ref] [bib_ref] Fate of a modified fenestration of atrial septal occluder device after transcatheter..., Schneider [/bib_ref] Thus, we speculated that the selection of complete closure time might be varied according to the follow-up results.
Different from the atrial septum tissue, the fenestrated device is stiff and thick. The reduced operation space due to the large device in the right atrial as well as the small fenestration would increase the difficulty while inserting the guidewire and sheath from the inferior vena cava to the fenestration. The wire was inserted across the fenestration after several attempts. As a result, the second occluder was implanted successfully under TEE guidance instead of radiation exposure. Two-dimensional (2D) TEE combined with real-time 3D TEE imaging can demonstrate the structures of fenestration and occluder or sheath in an enhanced manner.
Although the size of the second ASO was much smaller than the first, the second prescription of postoperative oral Aspirin was similar to that of the first ASD closure. This phenomenon might be attributed to the difficulty of the tissue encroachment to the second ASO, as the small ASO directly contacts with the previous device instead of the cardiac structures. Some authors proposed that incomplete endothelialization was a mechanism of endocarditis after ASO implantation. [bib_ref] Infective endocarditis after device closure of atrial septal defects: case report and..., Amedro [/bib_ref] However, statistical data on endothelization of 2 devices in such an overlapping state, as well as the effects on platelet aggregation are yet lacking. In this case, the patient did not display any events during the 12-month follow-up after the second operation; no thromboembolism or infections occurred. Furthermore, the safety of this procedure needs to be further validated in an additional number of cases and studies.
# Conclusion
This case report illustrated that minimal invasive percutaneous complete closure of the fenestration in ASO under TEE guidance rather than radiation exposure is feasible and can serve as a therapeutic option in patients with severe PH, who were treated with fenestrated ASOs. Complete closure of the residual atrial shunting in the fenestrated ASO should have the advantage of reversing the heart hemodynamics to the normal condition and avoiding potential paradoxical embolism.
[fig] Figure 1: Actual Memory ASD fenestrated device and 3-dimensional transesophageal echocardiography (TEE) imaging: (A) Fenestration next to the center hub(arrows), (B) Position relationship among center hub fenestration and the sheath (arrows). C = center hub, F = fenestration, S = sheath. [/fig]
[fig] Figure 2: Color compared transesophageal echocardiography (TEE) guidance showing the procedural steps of the fenestration closure: (A) Sheath across the fenestration (arrows), (B) Deployment of the left disc (arrow), (C) Deployment of the right disc (arrow). LA = left atrium, RA = right atrium, SVC = superior vena cava. [/fig]
|
CNN-Peaks: ChIP-Seq peak detection pipeline using convolutional neural networks that imitate human visual inspection
chip-seq is one of the core experimental resources available to understand genome-wide epigenetic interactions and identify the functional elements associated with diseases. the analysis of chip-seq data is important but poses a difficult computational challenge, due to the presence of irregular noise and bias on various levels. Although many peak-calling methods have been developed, the current computational tools still require, in some cases, human manual inspection using data visualization. However, the huge volumes of chip-seq data make it almost impossible for human researchers to manually uncover all the peaks. Recently developed convolutional neural networks (cnn), which are capable of achieving human-like classification accuracy, can be applied to this challenging problem. In this study, we design a novel supervised learning approach for identifying chip-seq peaks using cnns, and integrate it into a software pipeline called cnn-peaks. We use data labeled by human researchers who annotate the presence or absence of peaks in some genomic segments, as training data for our model. the trained model is then applied to predict peaks in previously unseen genomic segments from multiple chip-seq datasets including benchmark datasets commonly used for validation of peak calling methods. We observe a performance superior to that of previous methods. open Scientific RepoRtS | (2020) 10:7933 | https://doi.org/10.1038/s41598-020-64655-4www.nature.com/scientificreports www.nature.com/scientificreports/ additional information as input data (e.g. mappability scores that the fraction of a region that overlaps at least one uniquely mappable read in the genome 8 ).In 9 a neural network model was proposed to denoise ChIP-seq data and increase the performance of calling the peaks. An ensemble approach was also proposed 10 to improve the accuracy of calling ChIP-seq peaks. The latter exploits the output from multiple existing peak-calling software tools to eliminate outliers among the multiple peak calling decisions. These approaches have increased the sensitivity of calling true peaks, but they still suffer from high false-positive rates 8 . In particular, human cancer cell lines are too complicated for one to understand human malignancy using ChIP-seq, since the data patterns vary substantially depending on patients' primary cancer tissues[11][12][13]. False-positive rates of calling ChIP-seq peaks in human cancer cell lines are worse than in other datasets 12 .To resolve these issues, human experts are used to label true peaks using visualization tools such as UCSC genome browser, and the Integrative Genomics Viewer (IGV) 14,15 . False-positive peaks can also be corrected by professional researchers 16 . However, it would be extremely inefficient for human scientists to find all of the peaks in the whole genome for large volumes of ChIP-seq data. Hocking et al.8, have proposed a supervised learning approach based on grid search to optimize a parameter (e.g. a cut-off value) that users need to set up when running peak-calling tools such as MACS2. They use data of which part of the peaks have been labeled by human experts, learn parameter values, and apply the optimized parameters for the rest of the dataset. Unfortunatly, this requires completing labeling task for each individual ChIP-seq dataset and each peak-calling tool.Recently, supervised machine learning methods based on deep neural networks, such as convolutional neural networks (CNN), have been successfully applied to epigenetics, regulatory genomics, and system biology 17,18 . In this article, we develop a novel peak calling software pipeline based on CNN, named CNN-Peaks, that feeds on data partially labeled by human researchers. We identify local peak-calling threshold values, which might differ from one genomic segment to another, and use them to reduce false-positive peak calls, and thus resolve local bias caused by structural variations in, e.g., the human cancer cell lines. Note that some of the peaks in genomic regions that have frequent copy number variations are expected to show higher mapping depth than the others. As training data for CNN-Peaks, we use data labeled by human experts as well as read count information from the preprocessing steps of raw read mapping. Our software tool learns a model for determining proper peak-detection cut-off values in specific genomic regions by taking read mapping patterns in their neighbor regions into account. Our inspiration to use CNNs is the fact that they have been successfully used to solve problems in image processing and natural language translation using data patterns that, like in our case, have features with local dependencies 19,20 ). We also use integrated annotation information (RefSeq) available in NCBI (https://www.ncbi.nlm.nih.gov/) as training input data for building our predictive CNN. The RefSeq data include genomic locations of genes, transcripts and protein encodings that can assist to determine the existence of peaks for human visual inspection.Our CNN-Peaks software package is composed of three main modules: one for preprocessing original input data (labeled data, genome annotation, and read count information) and feed them into our CNN architecture; another for learning a model using the training data; and a final one for predicting peaks for unknown data. We evaluate the performance of CNN-Peaks using some labeled data reserved as test data. We also test CNN-Peaks on ChIP-seq benchmark datasets that are commonly used for evaluation, and compare its performance against other major peak calling tools. We also use our CNN-Peaks tool to analyze various real datasets for histone modification, such as H3K27ac, and for transcription factor binding, such as GATAD2, in various human cell lines.Users can install the CNN-Peaks pipeline using a docker image and run the package using their own data with our trained model for predicting narrow histone modifications and transcription factor binding sites in humans. Experienced users can also build a new predictive model trained on their own labeled data using CNN-Peaks. Our software package includes a desktop application to help experts create labeled data easily by labeling peaks in genomic regions randomly selected from a given ChIP-seq read mapping data (Supplementary text S3). Our pipeline can be also applied to other types of high-throughput sequencing data, such as DNase-seq and ATAC-seq. Our package is available through the Github repository http://github.com/odb9402/CNNPeaks.Scientific RepoRtS |(2020) 10:7933 | https://doi.MACS2 and HOMER data (see Supplementary text 8). SICER seems to always call peaks wider than any other tool. We use the matchPWM function in the "Biostrings" R package to find binding sites of BRCA1, CHD2, and CTCF 41 . This experiment shows a biological relationship between peak significant scores and actual functional genomics.Figure 8shows that the scoring system of CNN-Peaks, which uses both the sigmoid activation from the deep learning model and a Poisson distribution, is stable and appropriate.Figure 8also shows that our CNN-Peaks results are strongly related to binding motifs, and are almost similar to, or better than, the other three peak calling tools.Figures 7 and 8demonstrate that our trained model that was learned using labeled data is quite robust and generalizes well. Indeed, only 12% of total labels came from transcription factor ChIP-seq, and the rest from histone modification ChIP-seq and ATAC-seq.Scientific RepoRtS | (2020) 10:7933 | https://doi.
In the postgenomic era, understanding epigenetic regulation is one of the most important challenges in the biological and medical sciences. To elucidate pathological mechanisms, and pinpoint target genes for developing therapeutics, it is important to assess abnormalities in the genome-wide interactions of proteins and genomic elements that cause gene misregulation. In this regard, chromatin immunoprecipitation followed by sequencing (ChIP-seq) is a technique widely used to identify genomic binding sites for epigenetic regulators, including histones, transcription factors, and DNA/RNA binding proteins 1,2 . ChIP-seq enables the discovery of the interactions between protein complexes and DNA regulatory elements, and their gene regulatory networks. ChIP-seq data analysis has shown how histone modifications and nucleic acid interacting proteins modulate the critical factors of gene regulation, and cell lineage determination and maintenance.
One of the major computational challenges in analyzing ChIP-seq data is to identify peaks in genomic areas where aligned reads are enriched when sequencing reads are mapped to a given reference genome. This task is a challenge for several reasons. Sequencing errors and local bias caused by structural variations complicate solving the peak-calling problem. The diverse patterns of data, which are caused by biological variability, a variety of domains, experiment environment, and sequencing coverage, make the problem even harder. Sensitive and reliable computational methods to determine peaks from noisy backgrounds are especially important for medical studies, where patient samples can be limited, resulting in sub-optimal data quantity and quality.
Several software tools for calling the peaks in ChIP-seq data have been developed based on various probabilistic and unsupervised learning methods, such as MACS2, HOMER, SICER, and SPP. Some of these tools show the high sensitivity of calling the peaks, but suffer from high false-positive error rates. Other tools may needSchool of Computer Science and Engineering, Pusan National University, Busan, 46241, South Korea.Department of Genetics, Stanford University, Stanford, 94305, USA.School of Medicine, Kyung Hee University, Seoul, 02447, South Korea.Department of Computer Science, Boston College, Chestnut Hill, Philadelphia, MA, 02467, USA. ✉ e-mail: [email protected]
# Materials and methods
Data description. We obtained multiple ChIP-seq read mapping datasets in BAM format from the ENCODE data portalfor creating labeled data, and validating CNN-Peaks. This data includes several ChIP-seq datasets for examining histone modification, such as histones H3K36me3, H3K4me3, H3K27me3, H2AFZ, and H3K9ac, and transcription factor binding, such as transcription factors GATAD2, POLR2A, SMARCE1, and MAX in cancer cell domains K562, A549, HepG2, HEK293, and GM12878. We also downloaded ChIP-seq data from leukemia cell line K562 and labeled its background binding of genomic regions to then train our CNN model. From the ChIP-seq read mapping data in BAM format, some genomic segments were randomly selected, and the locations of peaks were labeled using our visualization tool with the BAM alignment 8 data asillustrates.
In addition to the labeled data and read mapping BAM files, we used a curated non-redundant collection of genomic, transcript, and protein sequence records for the human reference sequence (RefSeq) in NCBI, as an additional input vector for our CNN model. The RefSeq data include protein-coding locations and pseudogenes. When researchers manually examine ChIP-seq data to determine the peaks, the genome annotation information, such as transcripts and their corresponding protein records, are commonly viewed together with the given read mapping data. This information often gives a good guess of where peaks might be. To mimic human inspection, we considered this and built our model so that it uses these different annotated types as input. We added this annotation information as a binary vector to represent the presence and absence of transcripts and proteins in each genomic position. www.nature.com/scientificreports www.nature.com/scientificreports/ Data preprocessing. A preprocessing module is included in our CNN-Peaks pipeline. This module converts all input data, including read mapping BAM files, labeled data, and genome annotation information, to vectors with the right shape to be fed into our CNN architecture. We use bins of a fixed size (12,000 bins by default) to normalize the different window sizes of labeled segments into the same size. If a window is smaller than the target window size, the user of CNN-Peaks should either label an additional region to append to this window using our visual inspection tool, or should adjust the parameter of the target window size, or should simply eliminate the window (Supplementary text S1).
Our preprocessing module also includes functions to smooth read count patterns to reduce noises in read alignment data. This smoothing step also helps make our CNN model close to human visual inspection. Without the smoothing and correction step, (raw) depth patterns are very noisy (Supplementary text S1). We use max-pooling and Gaussian filter convolution operations for smoothing. Gaussian filters are commonly used for denoising and smoothing in various domains, such as image processing. The equations of the convolution and max-pooling operations (Eqs. (1) and (2) respectively) are as follows (note "*" indicates a convolution operation. The strides for convolution and max-pooling are 1).
[formula] ∑ * = =− − ⋅ = ≥ − ≤ ≤ + v v n w i n w max _pool n v ( ) where f or all 2 2 (2) k i v,w k Above, ∈ + M w Z , [/formula]
are the filter size of each operation, v, u are input vectors, and v i is the i th component of v. Let X be a vector of read mapping counts, filter a gaussian filter, and X smoothing a smoothed vector of X, then. We built three types of modules with a slimmer structure than the original Inception module asillustrates. The structure is composed of several hidden layers, such as pooling and convolutional layers, as well as the Inception modules, asshows. The Inception-style modules are known to lead to inefficacies in terms of prediction accuracy if they are placed at the beginning of the CNN architecture. To avoid this potential problem, we put traditional convolutional filters, and a max-pooling layer, at the beginning of our CNN-Peaks architecture. The details of our CNN architecture are described as follows. First, two input vectors for annotation information (RefSeq) and read mapping patterns are transformed through the convolution layers into vectors. Then the max-pooling layer reduces the dimension of these vectors. Then there are seven Inception modules. Each Inception module has a layer that concatenates a list of outputs from the filters. Two of these modules, called A-modules, use three types of filters: convolution, max, and average pooling. They provide information about peak patterns and the scale of strong peak signals to our CNN model. However, since max pooling and average pooling have the same number of features as the previous layer, this increases the vector size exponentially as a function of the depth of pooling layers. To resolve this issue, we use the other three Inception modules, called B-modules, which are composed of convolution filters of three types www.nature.com/scientificreports www.nature.com/scientificreports/ without an average pooling. Finally, we use two Inception modules, called C-modules, that have four types of convolution filters with a wider size and longer stride than the filters used in A-and B-modules. At the end of the Inception modules, an average pooling layer followed by fully connected layers is added to reduce dimensions.
[formula] = … X m ax pool max pool L [ _ (1), , _ ( )],(3)pool X w X w , , X X filter X filter L [( )(1), , ( )( )](4) [/formula]
In addition, we use a residual structure between the Inception layers. This helps avoiding vanishing gradient problems. We also use batch normalization as regularization to avoid overfitting 27 while training. output layer of cnn architecture. To determine the presence or absence of peaks in each individual genomic position, the output layer of the CNN architecture needs a number of neurons that are equal to the number of genome bases. This large number of neurons in the output layer usually causes a significant degradation of learning performance. In order to reduce the number of neurons, we designed our CNN model to learn optimal threshold values for genomic segments based on read mapping patterns in a selected window, rather than computing the p-value or the likelihood of the presence of a peak signal in each individual genomic position. This substantially reduces the number of neurons required in the output layer of our CNN model, and prevents performance degradation. Since the output vector size becomes smaller than the input vector, we add an additional operation to expand the output vector size to be identical to the input vector, so that we can predict the presence or absence of peaks in each individual position (See the purple box named as "Expand" in. These expanding vectors are implemented using the "broadcasting vector" standard in Tensorflow and Numpy, which allows operations between vectors of different sizes. The peak calling process of our CNN-Peaks is summarized in.
## Loss function.
Determining the presence or absence of a peak signal is a binary classification problem. We use cross-entropy as a loss function for learning our model. Most methods for classification problems require balancing the trade-off between sensitivity and specificity in performance. Likewise, we need to be careful not to favor only one of these. In peak calling problems with ChIP-seq data, peaks are relatively rare compared to the whole genome size. If a certain method tends to call 'no-peak' (i.e. it has low false-positive error rate), it will show high accuracy, although it misses important peaks. Therefore, we use a weighted cross-entropy as our loss, in Eq. (5). www.nature.com/scientificreports www.nature.com/scientificreports/
[formula] θ Λ Λ Λ = + − − θ θ X y h X w y h X ( , , ) log( ( , )) (1 ) log ( 1 ( , )),(5)i i i i i , , [/formula]
where X is the input read counts vector, Λ the annotation vector, θ the set of parameters in our model, y i the i th element of the input for the labeled data, w a weight for the importance of false-negative calls relative to false-positives calls in the valuation, and
[formula] Λ θ h X ( , ) i, [/formula]
is the i th element in the output predicted for given input X, Λ and model parameter θ. The weight w is determined by a ratio between negative regions (no peaks) and positive regions (peaks) for given data.
In addition, we apply the Top-K method for the loss function. In the Top-K method, sensitivity is regarded as more important than specificity for a high value of K, while specificity is more important than sensitivity for a low value of K. To achieve a balance between sensitivity and specificity, we set L = K/2, where L is the output vector size. Our final loss function is (6)
[formula] ∑ θ κ κ θ Λ Λ = κ = X X ( , , , ) 1 ( , , ),(6)where θ Λ X ( , , ) n [ ] [/formula]
is the n th largest individual weighted cross-entropy loss among all l i . The loss function TopK was optimized using the Adam optimizer that uses backpropagation to adjust model the parameters θ 31 .
Scoring peaks using sigmoid activations. An important task of any peak-calling algorithm is to give a significance score to each peak 32 . We measure the significance of peaks by calculating a p-value for each genomic location under the Poisson distribution (note that the counts for each position in the genome-wide tag data for ChIP experiments is known to follow a Poisson distribution)and by combining these p-values with the sigmoid values from the output layer in the CNN-Peaks architecture. The sigmoid function in the output layer of our CNN-Peaks model generates values that can be interpreted as the probability of the presence of a peak. The score value of each peak called by CNN-Peaks is determined by the product of its sigmoid activation value and −log 10 of the p-value of the Poisson distribution for that peak (Supplementary text S5). This score can help users assess the significance of a particular peak.
# Results
Data preparation and creation of labeled data. We downloaded 16 ChIP-seq datasets and one ATACseq dataset in BAM format, which were mapped to the human reference sequence. Professional experts examined 3,294 genomic segments that were randomly selected . The datasets were assigned to three experts, each dataset being labeled exclusively by one expert. We expect that there is little experts' bias in the labeling process.
Part of these data were labeled and used as training data for our model (note that to put CNN-Peaks through a tough test, our test data does not come from , but rather from completely different datasets). Our data includes ChIP-seq data for identifying histone modification sites, such as H3K4me3 and H3K27ac, which are known as strong cancer biomarkers. On average, about 66% of the peaks labeled by the experts are near the transcription start sites (TSS) of RefSeq genes, as illustrated in the 4 th column of . This fraction varies from 33% to 88% depending on experiments. We expect that ChIP-seq for H3K9ac histone modification target has a large number of peaks nearby promoters whereas H3K9me3 has much fewer peaks nearby TSS. The labels of 'peak' or 'no-peak' were applied while visualizing the raw read alignments in BAM format, together with the RefSeq annotation data obtained from the UCSC genome browser. The labeled data were easily created using our own graphical interface program included in the CNN-Peaks software package. www.nature.com/scientificreports www.nature.com/scientificreports/ cnn-peaks pipeline. We developed the CNN-Peaks pipeline as described in. The CNN-Peaks pipeline takes as training input: labeled data in text format; read alignment data in BAM format; and RefSeq annotation information. The input data is converted to vectors of the right shapes for our CNN architecture. CNN-Peaks trains a predictive model based on the data in these vectors. The model is then applied to unlabeled data in the prediction stage. Our trained model can be used to call peaks for narrow histone modifications and transcription factor binding sites in other human ChIP-seq data (i.e. other than the one in , part of which was used for training) without additional labeling and training. In addition, peaks in ATAC-seq can also be detected using CNN-Peaks.
Unlike most other peak calling methods, CNN-Peaks needs no additional control ChIP-seq data, usually used as a background signal to reduce false-positive errors. In other words, users can determine the peaks of their target ChIP-seq data with no control data. This is critical when peaks are being called for high-throughput sequencing (HTS) data, such as ATAC-seq and DNase-seq, to capture open chromatin regions, because for the HTS data it is almost impossible to make the control dataset (see the information of ATAC-seq in https://informatics.fas. harvard.edu/atac-seq-guidelines.html).
All modules in CNN-Peaks can be run using a single command. Its output, the peaks called out by CNN-Peaks, is provided in BED format. All of our source code is available in a GitHub repository (http://github. com/odb9402/CNNPeaks). The CNN-Peaks pipeline can be easily installed with Docker, which avoids users having to manually install all prerequisite programs and set up complicated system environments. The graphical www.nature.com/scientificreports www.nature.com/scientificreports/ interface program for labeling peaks in a given ChIP-seq BAM file is also included in CNN-Peaks. Experienced users can use CNN-Peaks to build a new predictive model, learned from their own labeled data. optimization of hyperparameters in the cnn architecture. There are several hyper-parameters in our package, as listed in (Supplementary text S2). We manually fine-tuned its various hyperparameters, such as the learning rate and convolution filter feature numbers, although we observe that CNN-peak's performance is not very sensitive to their values. Our CNN architecture uses the Adam Optimizer in TensorFlow-GPU 1.8.0. We trained a predictive model using our CNN-Peaks package with a single Nvidia Quadro P400 GPU within an hour. We used a random subset of all 3,294 genomic segments (see Section 3.1) to train the CNN-peaks model. evaluation using labeled data. Our labeled data in were randomly sampled (90%) to build our training set. ChIP-seq data for peaks labeled in (i) the H3K4me3 histone modification in the K562 human leukemia cell line, and (ii) the H3K27ac histone modification in the GM12878 cell line, were not used for training the predictive model (note that professional experts marked 156 labels for the H3K4me3 data in K562 and 150 for H3K27ac in GM12878). To evaluate our CNN-Peaks prediction model, we used (i) and (ii) as test datasets, comparing prediction results using CNN-Peaks with the labels in (i) and (ii). We counted false-positive and false-negative errors, and measured sensitivity and specificity. To account for both sensitivity and specificity, we also calculated the F1 score for performance evaluation.
We compared our CNN-Peaks with widly-used peak calling tools, including MACS2, HOMER, and SICER. We used default parameters when running these software tools. There are peak calling results available at the ENCODE data portal for our test datasets (i) and (ii), so we included these results in our performance comparison as well. Note that the peak calling results from ENCODE were processed using an ENCODE ChIP-seq pipeline that used MACS2, as well as post-processing steps to reduce false-positive errors.shows the performance comparison for both test datasets. For the GM12878 cell line data (See, the sensitivity of CNN-Peaks dropped by about 10% compared to the other tools. However, CNN-Peaks improved specificity drastically to almost 97%, while the specificity of the other methods was lower than 76%. As a result, our CNN-Peaks increased the F1 score by at least 8% compared to existing tools. Unlike the GM12878 cell line that is known to have a relatively normal karyotype, the genome of the K562 cell line has more complex karyotypes due to frequent abnormal structural variations. For this reason, calling peaks in the K562 cell line data is more challenging than in the GM12878 data. While the sensitivity of CNN-Peaks dropped by almost 10% for the K562 ChIP-seq data, the specificity increased by 24% compared to the average of the other tools. (see. In addition to the sensitivity and specificity, we measured the portion of peaks overlapped by the CNN-Peaks and the other three peak callers on the H3K4me3 histone modification ChIP-seq data in the K562 cell line. We observed that 97%, 94%, and 96% of peaks called by the CNN-Peaks overlapped with MACS2, SICER, and HOMER (see Supplementary text S7). Most peaks called by the other peak callers were captured by CNN-Peaks, while many of the false-positive peaks generated by all the three peak callers were filtered out by CNN-Peaks.
We also used the peaks overlapped by all combinations of MACS2, SICER, and HOMER as new predictors, and compared the predictors with the CNN-Peaks. As a result, the intersection of all the three callers showed the highest F1-score (0.7941) for test dataset (i) among all the combinations, but its performance is still not better than that of CNN-Peaks (F1-score 0.8515). In addition, for test dataset (ii), the intersection also showed the highest F1-score (0.8462), but its performance is still not better than CNN-Peaks (F1-score 0.9397). The intersection of peaks called by all the three peak callers may filter out some false-positive peaks, but still contain more false-positive results than our CNN-Peaks. www.nature.com/scientificreports www.nature.com/scientificreports/ Measuring relative distances between narrow histone modification peaks and transcription start sites. Histone modification H3K4me3 and H3K9ac are commonly used as epigenetic markers. The histone modification markers are highly enriched near transcription start sites (TSS), so the relative distances between histone modification peaks and TSS are used for evaluating ChIP-seq peak callers. To validate the performance of CNN-Peaks for calling narrow histone modification peaks in the H3K4me3 data of the HepG2 cell line, and H3K9ac of GM12878, we measured the relative distances of our peak intervals and TSS using a similarity metric suggested in.shows the ideal shape for the empirical distribution of relative distances between highly correlated intervals.,C describe the empirical distributions of relative distances between TSS and peaks called by CNN-Peaks, MACS2, SCIER, and HOMER for the H3K4me3 and H3K9ac histone modification data. The plots for CNN-Peaks are more similar to the plots in the ideal case, shown in, than the plots of other peak calling tools in both datasets. In particular, CNN-Peaks' plots exhibit an accentuated y-value for small relative distances (x-value) that then drops. While other tools exhibit a similar behavior, they do not do so as markedly. This indicates that the peaks determined by our CNN-Peaks are more related to TSS than the results of other major tools.
peak calling for curated chip-seq benchmark data. In addition to histone modification ChIP-seq data, we validate our CNN-Peaks pipeline with other benchmark data that are popularly used for validating software tools of transcription factor binding ChIP-seq analysis. Unlike the data labeled from labeled by our experts for this study, this manually curated benchmark data have three classes of labels instead of just the classes 'peak' and 'no-peak' . To be specific, they have an additional 'ambiguous' class to mark genomic segments that could not be labeled as either 'peak' or 'no-peak' . In our experiments, we take the 'ambiguous' class as 'no-peak' , which follows the standard of performance measurement for peak calling tools using ChIP-seq benchmark datasets.
We applied our CNN-Peaks, trained with some of our own labeled data from , to predict the benchmark data whose peaks were determined by manual curation in. We compare our results with the ones predicted by three peak calling programs, including MACS2, HOMER, and SICER. We measure the sensitivity, specificity, and F1 scores using the results predicted by each method and the peaks of the benchmark data in 16 as ground truth. We ran these three peak calling tools with their default parameters. The whole alignments of all of the ChIP-seq data relative to the reference sequence were fed as their input. While the sensitivity of CNN-Peaks is almost similarly, or slightly smaller, than MACS2, which showed the best sensitivity among the three existing tools, CNN-Peaks' specificity is the best among the three existing tools, in both the ChIP-seq experiment data for NRSF and SRF targets. Our CNN-Peaks pipeline has substantially better F1 scores than the other tools.
Validation of peak scoring using known binding motifs. Comparing the peaks with statistically significant scores (see Section 2.6) for transcription factor ChIP-seq to known binding motifs is another way to evaluate the performance of ChIP-seq peak calling methods. We matched peaks determined by four methods, including our CNN-Peaks, to the known binding motifs for several cell lines from ENCODE, including BRCA1 in GM12878, CHD2 in K562, and CTCF in HepG2. In this analysis, we determine the genomic position with the highest score as the center of each peak region, and examine only motif matches close to the center of the peak (i.e. within 100 bases around the center of each peak). The goal of this approach is to avoid the analysis to be affected by bias introduced because of the different peak lengths. Users can refine the peak regions determined by CNN-Peaks for the transcription binding ChIP-seq data using a script included in our CNN-Peaks package. We note that for the transcription factor binding ChIP-seq data, CNN-peaks consistently calls peaks wider than . Peak calling performance evaluation using ChIP-seq benchmark data manually curated for SRF target with GM12878 cell line (top) and NRSF target with K562 cell line (bottom). The benchmark dataset includes 6 different types of datasets for the SRF target and NRSF target respectively. These datasets were generated using different methods as described in. Each color represents a different dataset.
## Disscussion
As sequencing has become readily available for biological and medical studies, GWAS studies have identified many DNA bases and structural variations linked not only to genetic diseases but also to common diseases. While GWAS suggests what DNA abnormalities are linked to diseases, such studies cannot characterize the functional aspects of such variations. ChIP-seq has been widely used to understand the mechanisms of genome-wide gene regulation involved in developmental biology and various diseases, including cancer. Many computational approaches have been developed and applied to analyze the ChIP-seq data. However, most of them exhibit a high false-positive ratio of peak signals, and can misinterpret data, especially for histone modification ChIP-seq data.
To resolve this issue, we designed a new approach for calling the peaks in ChIP-seq data based on a convolutional neural network architecture that mimics human visual inspection. Our tool uses genome annotation information, labeled data for peaks that are inspected by human researchers, as well as read mapping data in BAM format for ChIP-seq data. Our ChIP-seq data were generated from GM12878 and K565 human cell lines. Our CNN prediction method, which comes integrated with some useful data preprocessing steps, is available as a software package called CNN-Peaks. This package also includes a graphical visualization interface for users to easily set labels 'peak' or 'no-peak' in genomic segments using read mapping BAM files.
We tested our predictive model and reported test errors using both our own labeled data, and benchmark data that were manually curated by other groups, and that are popularly used to evaluate computational methods for calling peaks. Our CNN-Peaks showed a quite conservative peak calling rate and improved specificity compared to other methods, while not dropping sensitivity much.first proposed the idea of using visual inspection for calling peaks, and showed that manual peak calling could improve the performance of existing peak calling tools. However, they only focused on tuning parameters of existing peak callers using labeled data, while we redefined this idea as a signal processing problem, and designed a new approach based on the convolutional neural network model, which has been successfully used in various signal processing problems. Two important details contribute to its improved filtering out of false-positive peaks: its use of inception modules and its use of additional information, such as RefSeq annotation from experts' visual. These factors increase the specificity and F1 scores of CNN-Peaks' prediction comparing to other existing peak calling tools, even for ChIP-seq data that were generated in the K562 cancer cell line, which is known to be extremely complex due to frequent abnormal structural variations. In addition, we investigated the biological relationship between peak calling results and transcription start site (TSS) for histone ChIP-seq, and with known binding motifs for transcription factor ChIP-seq. Our experiments demonstrate that the peaks of histone modification ChIP-seq data determined by CNN-Peaks are highly correlated with TSS, and that the peaks of transcription factor ChIP-seq are strongly related to known binding motifs.
There is still some room to improve the performance of our pipeline. CNN-peaks can be improved by collecting more labeled data are collected by more experts (see Supplementary text 6). Non-experts' opinions can also be used to improve its performance, since their labeling results should be mostly consistent with that of the experts (see Supplementary text 9). We can further include other types of information that professional experts use for inspecting the peaks in ChIP-seq data via our visualization tool. The more thoroughly curated data are available for various types of ChIP-seq datasets, the more we expect that our performance will improve. In addition to calling peaks, we believe that our CNN-Peaks pipeline could be extended to determine if ChIP-seq datasets are problematic or valid. One straighfoward way of doing so is to look at the confidence scores output by CNN-Peaks and determine a dataset to be problematic if most confidence scores are low. This approach is not fail proof, as one could imagine a problematic dataset with very clear (but wrong) peaks. One way to tackle this, would be to ask experts not only to label peaks, but also to label the correctness of different regions in the training data. A CNN could then be trained to minimized a combined loss for (a) predicting the occurences of peaks, and (b) determining the correctness of different regions in the input data. CNN-Peaks can be a useful toolset for the quality control and accurate analyses of high throughput sequencing data in the epigenetic research community. |
Pathogenesis-related protein 1 suppresses oomycete pathogen by targeting against AMPK kinase complex
h i g h l i g h t sUpon infecting by P. infestans, StPR1 expression was induced in host and secretory StPR1 proteins translocated into pathogen. The translocated PR1 proteins target AMPK complex in P. infestans, and impaired the AMPK activation to downstream targets. StPR1 prevented ROS homeostasis and inhibited the expression of RxLR effector-encoding genes in P. infestans.g r a p h i c a l a b s t r a c tThe potato-Phytophthora infestans system was used as a model to investigate how PR1 proteins protect plants against oomycete pathogens. Results showed that upon pathogen infection, PR1 proteins were induced, and secreted PR1 translocated into pathogen cells to target AMPK kinase complex, thus inhibiting vegetative growth and pathogenicity. Question marks and dotted lines represent scientific problems which need more detailed proofs. NPR1, nonexpressor of pathogenesis-related genes 1; TGA, TGACGbinding factor; SNF1, Sucrose non-fermenting 1; RxLR, the amino acid (Arg-x-Leu-Arg) of conserved motif in effector, where x is any amino acid.a b s t r a c tIntroduction: During the arms race between plants and pathogens, pathogenesis-related proteins (PR) in host plants play a crucial role in disease resistance, especially PR1. PR1 constitute a secretory peptide family, and their role in plant defense has been widely demonstrated in both hosts and in vitro.Keywords:Pathogenesis-related protein 1 AMPK kinase complex Phytophthora infestans Potato late blight Host-pathogen interaction Cross-kingdom translocation However, the mechanisms by which they control host-pathogen interactions and the nature of their targets within the pathogen remain poorly understood. Objectives: The present study was aimed to investigate the anti-oomycete activity of secretory PR1 proteins and elaborate their underlying mechanisms. Methods: This study was conducted in the potato-Phytophthora infestans pathosystem. After being induced by the pathogen infection, the cross-kingdom translocation of secretory PR1 was demonstrated by histochemical assays and western blot, and their targets in P. infestans were identified by yeast-twohybrid assays, bimolecular fluorescence complementation assays, and co-immunoprecipitation assay. Results: The results showed that the expression of secretory PR1-encoding genes was induced during pathogen infection, and the host could deliver PR1 into P. infestans to inhibit its vegetative growth and pathogenicity. The translocated secretory PR1 targeted the subunits of the AMPK kinase complex in P. infestans, thus affecting the AMPK-driven phosphorylation of downstream target proteins, preventing ROS homeostasis, and down-regulating the expression of RxLR effectors. Conclusion: The results provide novel insights into the molecular function of PR1 in protecting plants against pathogen infection, and uncover a potential target for preventing pre-and post-harvest late blight.
## G r a p h i c a l a b s t r a c t
The potato-Phytophthora infestans system was used as a model to investigate how PR1 proteins protect plants against oomycete pathogens. Results showed that upon pathogen infection, PR1 proteins were induced, and secreted PR1 translocated into pathogen cells to target AMPK kinase complex, thus inhibiting vegetative growth and pathogenicity. Question marks and dotted lines represent scientific problems which need more detailed proofs. NPR1, nonexpressor of pathogenesis-related genes 1; TGA, TGACGbinding factor; SNF1, Sucrose non-fermenting 1; RxLR, the amino acid (Arg-x-Leu-Arg) of conserved motif in effector, where x is any amino acid.
# Introduction
Potato late blight is a devastating disease caused by the oomycete Phytophthora infestans, resulting in stem death and tuber rot, and causing severe economic losses reaching $12 billion annually [bib_ref] Applied biotechnology to combat late blight in potato caused by Phytophthora infestans, Haverkort [/bib_ref]. During plant-pathogen interactions, the RxLR (amino acid of conserved motif Arg-x-Leu-Arg) effectors from P. infestans and the resistance-related proteins (R) from Solanum tuberosum are the two key factors determining whether the disease will prevail or not. Besides R proteins, host plants also secrete cross-kingdom proteins that target pathogens to inhibit their growth, development, metabolism, and pathogenicity, including the pathogenesis-related proteins (PR). PR proteins are antimicrobial, low molecular weight, secretory or vacuolartargeted proteins encoded by host plants, and they are induced by pathological conditions during plant-pathogen interactions [bib_ref] NPR1, all things considered, Dong [/bib_ref]. Inducible PR proteins were first identified in the 1970 s in Nicotiana tabacum infected with tobacco mosaic virus and they were later reported in different plant species that had previously been infected by pathogenic oomycetes, fungi, bacteria, and viruses [bib_ref] Samsun' and 'Samsun NN'. II. Changes in protein constitution after infection with..., Van Loon [/bib_ref]. Among them, PR1 is an inherent component of innate immunity in plants, and has been used as markers for enhanced disease resistance for a long time. Its biochemical activity and mode of action remain elusive, being the only PR without functional annotations [bib_ref] Significance of inducible defense-related proteins in infected plants, Van Loon [/bib_ref]. The PR1 family is highly conserved among plants, and homologs have also been found in fungi, insects, and vertebrates, including humans. The cysteinerich secretory protein (CRISP), antigen 5, and PR1 proteins form a superfamily of secreted proteins named CAP together. In plants, PR1 proteins have unique resistance to oomycetes. The anti-oomycete activity of PR1 proteins from tomato and tobacco was directly demonstrated to inhibit zoospore germination of P. infestans in vitro and restricted P. infestans colonization on tomato surfaces in vivo [bib_ref] Pathogenesis-Related PR-1 proteins are antifungal (isolation and characterization of three 14-kilodalton proteins..., Niderman [/bib_ref]. Furthermore, the heterologous expression of the PR1 proteins P14c (from Solanum lycopersicum) and PR-1a (from N. tabacum) inhibited the growth of the oomycete pathogen P. brassicae at concentrations of 20 lg/mL [bib_ref] The sterol-binding activity of Pathogenesis-Related Protein 1 reveals the mode of action..., Gamir [/bib_ref] and restricted zoospore germination of P. infestans [bib_ref] Expression and purification of tobacco PR-1a protein tor function analysis, Li [/bib_ref]. However, the antifungal mechanism of PR1 proteins was a mystery until a CAP protein of yeast was found to bind sterols [bib_ref] Pathogen-Related Yeast (PRY) proteins and members of the CAP superfamily are secreted..., Choudhary [/bib_ref]. Since then, the inhibitory effect of PR1 on pathogens was demonstrated to sequester sterol from pathogens [bib_ref] The sterol-binding activity of Pathogenesis-Related Protein 1 reveals the mode of action..., Gamir [/bib_ref]. Sterol-auxotroph pathogens, such as oomycete Phytophthora, are particularly sensitive to PR1; and when the sterol biosynthesis was restricted, the sterolprototroph fungal pathogens became highly sensitive to PR1, indicating a critical relationship between sterol binding and anti-fungal property, and suggesting some other mechanisms may contribute to the anti-oomycete activity [bib_ref] The sterol-binding activity of Pathogenesis-Related Protein 1 reveals the mode of action..., Gamir [/bib_ref].
Besides sterol sequestration, PR1 inhibited programmed cell death at the pathogen infection sites [bib_ref] Plant and animal PR1 family members inhibit programmed cell death and suppress..., Lincoln [/bib_ref] , and induced the expression of host defense genes by releasing a defense signaling peptide called CAPE1 (for CAP-derived peptide1) [bib_ref] PR1-mediated defence via C-terminal peptide release is targeted by a fungal pathogen..., Sung [/bib_ref]. CAPE1 originates from the last 11 amino acids at C-terminus derived from the tomato PR1 protein P14c, and contributes to broad-spectrum antibacterial activity against Pseudomonas syringae DC3000 and larvae of Spodopteralitura by inducing the expression of defenserelated genes [bib_ref] Quantitative peptidomics study reveals that a wound-induced peptide from PR-1 regulates immune..., Chen [/bib_ref]. The importance of PR1 proteins in plant defense is further evidenced by an increasing number of studies showing that they are targeted by effector proteins from multiple fungal pathogens [bib_ref] Plant and animal PR1 family members inhibit programmed cell death and suppress..., Lincoln [/bib_ref] , such as the TaPR1-5 that is targeted by the ToxA effector from wheat pathogens Parastagonospora nodorum and Pyrenophora tritici-repentis [bib_ref] Wheat PR-1 proteins are targeted by necrotrophic pathogen effector proteins, Breen [/bib_ref] [bib_ref] A dimeric PR-1-type pathogenesis-related protein interacts with ToxA and potentially mediates ToxA-induced..., Lu [/bib_ref] , the PR1a and PR1b that are targeted by barley powdery mildew effector CSEPP0055 [bib_ref] Wheat PR-1 proteins are targeted by necrotrophic pathogen effector proteins, Breen [/bib_ref] [bib_ref] Interaction of barley powdery mildew effector candidate CSEP0055 with the defence protein..., Zhang [/bib_ref] , and the PR1 in Arabidopsis thaliana that is targeted by the virulence factor SsCP1 (cerato-platanin protein) from Sclerotinia sclerotiorum [bib_ref] A cerato-platanin protein SsCP1 targets plant PR1 and contributes to virulence of..., Yang [/bib_ref].
Although sterol sequestration and releasing host defense signaling peptides are two important factors that contributing to the antifungal activity of PR1 proteins, our knowledge regarding the molecular function of PR1 proteins is still incomplete, especially in pathogen cells. Can host secretory PR1 proteins translocate into pathogen cells, thus inhibiting the vegetative growth and pathogenicity by targeting key intracellular proteins or signaling pathways? Given the sterol-defect and high sensitivity of P. infestans to PR1 proteins, the S. tuberosum-P. infestans pathosystem was used to investigate whether host secretory PR1 proteins mediate cross-kingdom anti-oomycete activity. Additionally, we aimed to identify the molecular targets of PR1 proteins in pathogens. Our research provides new insights into the anti-oomycete mechanisms of PR1 proteins. Moreover, the functional analysis of PR1 proteins provides a theoretical basis for breeding resistant varieties, and the identification of their molecular targets in pathogens supply potential target resources for biocontrol against the potato late blight.
# Materials and methods
## Strains and plant materials
The international aggressive oomycete pathogen P. infestans strain T30-4 was used for the pathogenicity and sensitivity tests, and for generating mutants. It was cultured on rye agar medium supplemented with 2% (w/v) sucrose (RSA) or in a liquid rye sucrose broth at 20°C. Agrobacterium GV3101 was used for the transient expression and genetic transformations, and was grown in Luria-Bertani (LB) agar/liquid medium. S. cerevisiae strain Y2HGold was used for yeast transformation. For S. tuberosum transformations, the susceptible potato variety 'Desiree' and the resistant variety 'E3 0 were used. N. benthamiana was used for transient expression and protein interaction assays. All plants were grown at 25°C with a 16-h photoperiod and 55% relative humidity.
## Vector construction
The RNAi and overexpression plasmids were constructed using the Gateway cloning system as previously described by Xiong et al. [bib_ref] Brassinosteriod Insensitive 2 (BIN2) acts as a downstream effector of the Target..., Xiong [/bib_ref] (presented in supplemental materials and methods in detail). For the yeast-two-hybrid (Y2H) assays, all StPR1 and mutations without signal peptide were cloned into pGADT7, and the subunits of PiAMPK were cloned into pGBKT7 through in-fusion HD Cloning kit (Cat No. 639648; Takara Bio USA, Inc). Meanwhile, the StPR1 genes were cloned into pGBKT7 and PiAMPK was cloned into pGADT7 vectors as a swap. For bimolecular fluorescence complementation (BiFC) assays, all the StPR1 and mutations were cloned into pAB862, while the subunits of PiAMPK were cloned into pAB855. pColdTF (Cat No. 3365; TaKaRa) was used for prokaryotic expression. All constructs were sequenced at TSINGKE (Chongqing, China). All primers used for the construction of plasmids are listed in .
## Y2h assays
To screen for the StPR1 targets, key kinases involved in the TOR signaling pathway in P. infestans were selected as potential targets. Y2H assays were performed using the Matchmaker TM Gal4 Two-Hybrid System 3 (Clontech, Palo Alto, CA, USA), as described in the manual. The plasmid pGBKT7-target gene with the empty vector pGADT7 was co-transformed into S. cerevisiae to exclude the false positives caused by autoactivation. The same procedure was performed in the pGADT7-target gene with pGBKT7. Yeast transformants were screened on SD/-Leu/-Trp and further confirmed on SD-Leu/-Trp/-His/-Ade medium with 20 lg/mL X-a-gal and 50 ng/mL AbA.
## Bifc and co-localization assays
BiFC was performed using the protocols described previously [bib_ref] Visualization of protein interactions in living plant cells using bimolecular fluorescence complementation, Walter [/bib_ref]. Target genes were cloned into binary vectors pAB855 tagged with cYFP and pAB862 tagged with nYFP. These vectors were transiently co-expressed in 4-week-old N. benthamiana, and fluorescence was observed 3 days after agroinfiltration. As controls, only the YFP halves as well as single proteins fused to cYFP and nYFP alone were co-transformed into N. benthamiana. The agroinfiltration buffer consisted of 10 mM MES, 10 mM MgCl 2 , 200 lM AS.
Agrobacterium strain with final OD 600 values of 0.1 in agroinfiltration buffer was used. For YFP fluorescence images, cells were analyzed using an excitation wavelength of 514 nm and an emission wavelength of 527 nm with a Leica confocal fluorescence microscope (Leica TCS SP8, Germany). For the co-localization assay, p35S-StPR1 GFP -t35S and pHam34-PiSNF1 RFP -tHam34 fragments were cloned into modified multigene assembly vectors p35Smulti-8GWN through isocaudarner AsisI/AscI and PacI/AscI enzyme digestions, respectively. Co-localization assays were performed in N. benthamiana as described above.
## Genetic transformation of potato lines and p. infestans
For S. tuberosum transformations, the susceptible potato variety 'Desiree' was used to generate StPR1 overexpression lines, and the resistant variety 'E3 0 was used to generate RNAi transgenic lines. The transformation of potato was mediated by A. tumefaciens GV3103, and the detailed procedure was performed as described previously by Millam. A. tumefaciens-mediated transformation was used to acquire P. infestans mutants according to the procedures described by Wu et al. [bib_ref] Establishment of a simple and efficient Agrobacterium-mediated transformation system for Phytophthora palmivora, Wu [/bib_ref] with some modifications. Detailed information is provided in supplemental materials and methods.
## Pathogenicity assays
Pathogenicity assays were conducted according to the method of Whisson et al. [bib_ref] A translocation signal for delivery of oomycete effector proteins into host plant..., Whisson [/bib_ref] with some modifications. The sporangia (1 Â 10 7 sporangia/mL) of P. infestans were harvested from 15day-old cultures by washing with sterile water and were prepared as inoculum. Then, 10 lL of the inoculum were dropped onto the back of detached leaves and tubers of wild-type or transgenic lines. Owing to the inhibited sporangia formation of P. infestans mutants, oomycete disks (7-mm-diameter) were used as the inoculum on detached leaves of 'Desiree'. All leaves and tubers were incubated at 20°C with a 12-h/12-h light/dark cycle, and the relative humidity was greater than 90%. The disease index (DI) of the leaves was assessed at 5 days post-inoculation (dpi), and the size of necrotic patches on tubers was measured at 10 dpi. The disease grade and DI were determined as described previously by Wu et al.. The pathogenicity tests were repeated three times, and each replication contained at least 10 leaves or tubers.
## Prokaryotic expression and purification
The vector used for constructing prokaryotic expression plasmid was pColdTF (Cat No. 3365; TaKaRa), which was engineered to enhance the production of disulphide-bonded proteins, thus improving the intracellular yield, purity and solubility of the recombinant proteins. PR1 genes lacking N-terminal signal peptides were cloned into the pColdTF through in-fusion HD Cloning kit (Cat No. 639648; Takara Bio USA, Inc), and recombinant vectors were transformed into Escherichia coli BL21 competent cells. The protein expression was induced by 0.5 mM IPTG at 16°C for 12 h. For removing proteins improperly folded, only soluble proteins of interest were collected and separated from the highaffinity Ni-charged resin (Qiagen, Valencia, CA, USA), and were visualized by coomassie-stained SDS-PAGE. The elution was dialyzed overnight to appropriate concentrations for further analysis. The expression of empty vector and GUS were used as a control.
## Cross-kingdom translocation of stpr1 and its inhibition on p. infestans in vitro
To test whether the secretory StPR1 proteins could translocate from potato to P. infestans, heterogenous StPR1 and its mutations (StPR1 DSP , StPR1 DSPDTM , and StPR1 DSPDGH ) tagged with GUS were purified and co-incubated with mycelia and sporangia of P. infestans in vitro. The mycelia that co-incubated with GUS only were used as control. The mycelia and sporangia of P. infestans were washed three times and then used for GUS histochemical staining after co-incubation with purified StPR1 and its mutations for 12 h. To test the cross-kingdom translocation of StPR1 in vivo, the intact StPR1 overexpressing potato lines (35S::StPR1 GUS ) were infected with a mixture of mycelia and sporangia for four days. The plants used for the inoculation were from the same line, and GUS histo-chemical assays were conducted every day. Those incubated with 35S::GUS were set as a control. At 2 dpi, the mycelia in/on the plants were also isolated for GUS histochemical assays. To measure the influence of StPR1 on sporangia germination, the 15-day-old sporangia (1 Â 10 7 sporangia/mL) were co-incubated with heterogenous StPR1 at 20°C, and the sporangia germination was calculated at 24 h post-inoculation. Each experiment was repeated three times.
## Measurement of h 2 o 2 and determination of enzyme activity
After being cultured in a liquid medium for 15 days, sporangia and mycelium mixtures of WT and mutants were harvested to measure H 2 O 2 content and the activity of superoxide dismutase (SOD), peroxidase (POD), and catalase (CAT). H 2 O 2 content, and SOD, POD, CAT activities were determined according to procedures specified by the kit manufacturers (Cat No. TO1076, TE0720, TE0423, TE0742; Leagene). Each sample was tested in biological triplicate.
## Co-immunoprecipitation assay and western blot analysis
To demonstrate the interaction of PiSNF1 and StPR1, total protein was extracted from S. cerevisiae expressing PiSNF1 Myc and/or StPR1 HA according to the instructions of Yeast Total Protein Extraction kit (Cat No. KGP650; KeyGEN, Nanjing, China). The total protein was incubated with anti-Myc and protein A/G sepharose (Abcam, ab193262) for 3 h at 4°C. The immunoprecipitation complexes were washed three times in phosphate-buffered saline as described previously by Cai et al. [bib_ref] Plants send small RNAs in extracellular vesicles to fungal pathogen to silence..., Cai [/bib_ref]. The HA-or Myc-tagged proteins were detected using western blot with anti-HA (Cell Signaling Technology, 3724S) and anti-Myc antibodies (Cell Signaling Technology, 9402S), respectively.
For western blot, total protein was extracted using lysis buffer (25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA, and 0.5% NP-40) plus 2% (w/v) polyvinylpolypyrrolidone, 1 mM dithiothreitol, and a protease inhibitor cocktail. Primary antibodies were diluted in 5% BSA as follows: 1:1000 for anti-pAMPK Thr172 and anti-APMKa (Cell Signaling Technology, 2535 and 2532, respectively), 1:2000 for anti-yeast beta Actin (Engibody Biotechnology, AT0014), anti-His (Solarbio, M1001020), anti-plant Actin (Abbkine, A01050), and anti-Gus (Agrisera, AS16 3689).
## In vitro and in vivo phosphorylation assay
LKB1 is the upstream kinase of AMPK. For in vitro phosphorylation assay, PiSNF1 His was incubated with active PiLKB1 in a kinase buffer (25 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 50 mM NaCl, 1 mM DTT, 50 mM ATP and 0.5 mCi of [32P]-ATP). StPR1 proteins (StPR1 DSP , StPR1 DSPDTM , and StPR1 DSPDGH ) were added into the system to check the influence of StPR1 on PiSNF1 activity. The system without the introduction of PR1 was used as a control. The reaction mixtures were incubated at 30°C for 30 min, and then for western blot to analyze the phosphorylation level of AMPKa. For in vivo phosphorylation assay, total protein was extracted from the wild-type P. infestans and mutants using lysis buffer, and then for western blot and ELISA analysis. According to the manual protocol, PathScan Ò Phospho-AMPKa (Thr172) Sandwich ELISA kit (CST, 7955S) was used to detect the phosphorylation level of AMPK; PathScan Ò Total Acetyl-CoA Carboxylase (ACC) Sandwich ELISA kit (CST, #7996C) and PathScan Ò Phospho-ACC (Ser79) Sandwich ELISA kit (CST, #7986C) were used to detect the phosphorylation level of ACC.
# Qrt-pcr analysis
Total RNA was extracted using the EZNA Ò Total RNA kit I (Omega Bio-tek, Inc. Norcross, GA, R6834-01). The first strand of cDNA was synthesized according to the protocol of the Prime-Script TM RT reagent kit with gDNA Eraser (TaKaRa, Otsu, Japan). StEF1a was used as a reference gene for the normalization of S. tuberosum, and PiActin was used as a reference gene for the normalization of P. infestans. The reaction systems had a volume of 20 lL and were assessed using a CFX96 Real-Time System (Bio-Rad, USA) with TB Green Ò Premix Ex Taq TM II (Tli RNaseH Plus). The gene-specific primers designed for the qRT-PCR are listed in .
## Bioinformatics and statistical analysis
The structures and topology of PR1 were predicted using Protter (http://molbiol-tools.ca/Protein_secondary_structure.htm).
The conserved domains of proteins were analyzed online using Pfam (http://pfam.xfam.org/). Multiple nucleic acid and protein sequences were aligned using ClustalX v2.0. The phylogenic tree was constructed using MEGA6.0 with a neighbor-joining method, and there were 1000 bootstrap replicates. Statistical analyses were performed using IBM SPSS v25.0. The differences between control and treatment were determined by Dunnett's t-test, and the differences among independent groups were determined by One-way ANOVA. The significance levels were *P 0.05, **P 0.01, and ***P 0.001.
# Results
## Pr1 proteins are widely distributed in s. tuberosum
PR1 is a multigene (10 genes) family with StPR1.1-1.9 clustered on chromosome 1, and StPR1a on chromosome 9 [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref]. The gene information is provided in Supplementary Table 2. StPR1.1-1.4, StPR1.8-1.9, and StPR1a are predicted to be secretory proteins with a signal peptide (SP), while StPR1.5-1.7 are non-secretory proteins [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref]. The nucleotide and protein sequence identity between StPR1.6 and StPR1.7, and that between StPR1.8 and StPR1.9 were over 99%, while the identity between StPR1a and other StPR1 proteins was less than 50% [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref]. Sequence alignment showed that the conserved GHYTQVVW motif (GH), four highly conserved a-helices (aI-IV) and b-strands (bA-D) were evident in different PR1 proteins. It is worth noting that the aI amino acid sequence, with a conserved motif of LxxHNxARxxVGV (TM, type-specific motif), was evident in secretory but not in nonsecretory PR1 proteins [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref]. Additionally, all PR1 proteins contain six conserved cysteine residues to form disulfide bridges and show a high level of sequence conservation throughout different plant species [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref]. During P. infestans infection, except for PR1.4 and PR1.6/1.7 genes in the sensitive variety 'Desiree', the expression of all PR1 genes was up-regulated (Supplementary [fig_ref] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo [/fig_ref]. Moreover, the transcript levels for all PR1 genes were higher in resistant variety 'E3 0 than in 'Desiree' (Supplementary [fig_ref] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo [/fig_ref]. Among the examined genes, StPR1.2, StPR1.3, and StPR1.8 were up-regulated in both 'Desiree' and 'E3 0 during P. infestans infection, suggesting their role in protecting plants against P. infestans [fig_ref] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo [/fig_ref]. Hence, these genes were selected for further functional characterization.
## Stpr1 is involved in plant defense in vivo and shows antioomycete activity in vitro
To investigate PR1 function in host defense against P. infestans in vivo, transgenic lines overexpressing StPR1.2, StPR1.3 and StPR1.8 as well as RNA interference lines (RNAi) were generated (Supplementary [fig_ref] Figure 3: The GH and TM motifs are important for the anti-oomycete activity of... [/fig_ref]. No differences were found in terms of plant growth among wild-type (WT), control plants (CK), and transgenic lines [fig_ref] Figure 4: Cross-kingdom translocation of StPR1 from host to Phytophthora infestans in vivo and... [/fig_ref]. In overexpressed transgenic plants 35S::StPR1.2 GUS , 35S::StPR1.3 GUS and 35S::StPR1.8 GUS , the disease resistance to P. infestans was enhanced, with the disease index (DI) being lower than 7 in all cases, while in WT and control (35S::GUS) plants, the disease symptoms were severe, with the DI reaching 80.4 and 80.8, respectively [fig_ref] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo [/fig_ref]. To explore the functional differences between secretory and non-secretory PR1 proteins in disease resistance against P. infestans, either the expression of all PR1 genes (RNAi-PR1) or only that of all secretory PR1-encoding genes were knocked down (RNAi-SPR1). Pathogenicity tests showed that both RNAi-PR1 and RNAi-SPR1 transgenic lines were more susceptible to P. infestans infection compared with the WT, as the DI was 97.7 and 72.8 in the RNAi-PR1 and RNAi-SPR1, respectively; and only 5.7 in the WT [fig_ref] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo [/fig_ref]. The disease symptoms and DI were more severe in RNAi-PR1 than those in RNAi-SPR1, suggesting that although both types of PR1 proteins are important, the secretory PR1 proteins may play a more crucial role in disease resistance. Furthermore, StPR1.2, StPR1.3 and StPR1.8 tagged with RFP (OE-StPR1) were heterologously expressed under the Ham34 promoter without signal peptide in P. infestans (Supplementary [fig_ref] Figure 5: PiSNF1 is the target of StPR1 [/fig_ref]. Fluorescence analysis showed that all PR1 proteins were expressed successfully in P. infestans, and the colony growth and sporulation were highly inhibited in PR1 overexpression mutants [fig_ref] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo [/fig_ref] , F and [fig_ref] Figure 5: PiSNF1 is the target of StPR1 [/fig_ref]. Moreover, the virulence of each mutant on leaves was significantly decreased, as the DIs of OE-StPR1.2, OE-StPR1.3 and OE-StPR1.8 were 11.2, 2.5 and 2.1, respectively; whereas the DI of wild-type P. infestans (WT) reached almost 100 [fig_ref] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo [/fig_ref]. Additionally, the infection of PR1 overexpression mutants on tubers was also lower than WT [fig_ref] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo [/fig_ref].
To test the role of SP, TM and GH motifs in StPR1, five types of StPR1 mutations were generated in E. coli or N. benthamiana [fig_ref] Figure 3: The GH and TM motifs are important for the anti-oomycete activity of... [/fig_ref]. In vitro tests showed that 1 mM StPR1.2, StPR1.2 DSP , StPR1.2 DSPDGH and StPR1.2 DSPDTM could inhibit sporangiospore germination significantly; however, the germination rates of StPR1.2 DSPDGH and StPR1.2 DSPDTM were higher than those tuberosum. The secretory PR1 proteins have four highly conserved a-helices (aI-IV) and b-strands (bA-D) to form a a-b-a sandwich structure, while the non-secretory PR1 proteins were absent from the amino acid sequence of aI. All the PR1 proteins have a conserved ''GHYTQVVW" motif, and six conserved cysteine residues (Cys) to form three disulfide bonds. The red blank indicates a specific sequence between secretory and non-secretory PR1 proteins, which was abbreviated as TM (type-specific motif). The words in red background represent completely uniform sequences. of StPR1.2 and StPR1.2 DSP [fig_ref] Figure 3: The GH and TM motifs are important for the anti-oomycete activity of... [/fig_ref]. A similar inhibition pattern was also observed in StPR1.3, StPR1.8, and their mutants [fig_ref] Figure 6: AMPK kinase complex in Phytophthora infestans is the target of StPR1 [/fig_ref]. Further pathogenicity assays for P. infestans were conducted in N. benthamiana leaves that transiently expressed StPR1 and StPR1 mutations tagged with GFP. The abundance of target proteins was measured according to the GFP fluorescence intensity [fig_ref] Figure 7: StPR1 did not change the levels of pAMPKa, while affected the AMPK... [/fig_ref]. To precisely measure the severity of necrosis, three necrosis grades for the treated N. benthamiana leaves were considered as defined by Li et al. [bib_ref] Negative regulators of plant immunity derived from cinnamyl alcohol dehydrogenases are targeted..., Li [/bib_ref] [fig_ref] Figure 3: The GH and TM motifs are important for the anti-oomycete activity of... [/fig_ref]. Data showed that compared with StPR1, the necrosis was higher in StPR1 DSP , StPR1 DTM , StPR1 DSPDTM and StPR1 DGH at 3 dpi (P 0.001), but lower compared with control plants (P 0.001) [fig_ref] Figure 3: The GH and TM motifs are important for the anti-oomycete activity of... [/fig_ref]. The results indicate that both TM and GH are critically important for the function of StPR1 proteins in conferring disease resistance upon late blight infection.
## Secretory verification of stpr1
Co-incubation analysis showed that, in accordance with the secreted characteristics of StPR1, GUS staining was attenuated in the 35S::StPR1.2 GUS plants as the inoculation day progressed [fig_ref] Figure 4: Cross-kingdom translocation of StPR1 from host to Phytophthora infestans in vivo and... [/fig_ref] , while there was no decrease in the control plant 35S:: GUS [fig_ref] Figure 4: Cross-kingdom translocation of StPR1 from host to Phytophthora infestans in vivo and... [/fig_ref] , suggesting that StPR1 was secreted into the extracellular environment during P. infestans infection. Interestingly, GUS histochemical assays showed that StPR1.2 GUS could be taken up by the hyphae and germ tube in or around the 35S::StPR1.2 GUS plants after 2 days of co-incubation, while the hyphae and sporangia from the control plant 35S::GUS were unstained (Supplementary [fig_ref] Figure 8: Functional characterization of PiSNF1 and the influence of AMPK activation and inhibition... [/fig_ref] StPR1.2 DSPDGH in vitro were stained in blue, while those coincubated only with GUS were unstained [fig_ref] Figure 4: Cross-kingdom translocation of StPR1 from host to Phytophthora infestans in vivo and... [/fig_ref]. Western blot analysis further confirmed the presence of the PR1 protein inside the hyphae [fig_ref] Figure 4: Cross-kingdom translocation of StPR1 from host to Phytophthora infestans in vivo and... [/fig_ref]. StPR1.3/1.8, StPR1.3/1.8 DSP , StPR1.3/1.8 DSPDTM , and StPR1.3/1.8 DSPDGH tagged with GUS were also detected in the P. infestans after 12 h of co-incubation [fig_ref] Figure 4: Cross-kingdom translocation of StPR1 from host to Phytophthora infestans in vivo and... [/fig_ref]. Secretory verification of StPR1 in vivo and in vitro indicated that the host secretory StPR1 could efficiently translocate into P. infestans cells; moreover, the StPR1 DSPDTM could also be taken up by P. infestans, suggesting that once secreted into the external environment, the secretory PR1specific TM motif was not necessary anymore for the crosskingdom translocation. Extracellular vesicles (EVs) represent an important mode of intercellular communication by serving as vehicles for transferring proteins, lipids, and RNA [bib_ref] Extracellular vesicles: Exosomes, microvesicles, and friends, Raposo [/bib_ref]. Tetraspanins, such as AtTET8 and AtTET9, are specific EVs markers. S. tuberosum has 20 TETRASPANIN (TET)-like genes, including the StTET8, which is closely related to AtTET8 [fig_ref] Figure 9: StPR1-PiAMPK signaling pathway is involved in regulating ROS homeostasis and RxLR expression... [/fig_ref]. The transcript levels of StTET8 [fig_ref] Figure 9: StPR1-PiAMPK signaling pathway is involved in regulating ROS homeostasis and RxLR expression... [/fig_ref] and the amount of its corresponding protein increased during P. infestans infection [fig_ref] Figure 4: Cross-kingdom translocation of StPR1 from host to Phytophthora infestans in vivo and... [/fig_ref]. However, it is unclear whether the PR1 proteins are transported through EVs. Hence, the positive correlation between StPR1 and EVs requires further investigation.
## Piampk complex is the target of stpr1 in p. infestans
The above results indicate that the secretory StPR1 proteins have an anti-oomycete activity and can translocate in a crosskingdom manner between S. tuberosum and P. infestans cells. Through multidimensional analysis, target of rapamycin (TOR) signaling pathway was affected mostly by StPR1 proteins, which is a core regulator of cell growth, development, proliferation and death in eukaryotes [bib_ref] TOR complexes and the maintenance of cellular homeostasis, Eltschinger [/bib_ref]. Taking the S. cerevisiae as a reference, the corresponding genes of the TOR signaling pathway in P. infestans were firstly identified (Supplementary . To determine the target proteins of StPR1 in the TOR signaling pathway in P. infestans, StPR1.2 was used as a bait in Y2H assays. Among the examined target proteins, only PiSNF1 could interact with StPR1.2 (Supplementary [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref]. PiSNF1 (PITG_14707) and PiAMPKa (PITG_07910) are two genes encoding the a-catalytic subunit of AMPK in P. infestans.
Phylogenetic analysis showed that PiSNF1 belonged to the same clade of the plant catalytic subunit of AMPK with a kinase and kinase associated domain, while PiAMPKa was independent from the clade of catalytic subunits of AMPK from plants and fungi [fig_ref] Figure 5: PiSNF1 is the target of StPR1 [/fig_ref]. According to Y2H assays, all StPR1 DSP , except for StPR1a DSP , interacted with PiSNF1 [fig_ref] Figure 5: PiSNF1 is the target of StPR1 [/fig_ref] , but not with PiAMPKa [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref] , indicating that StPR1 DSP can bind specif- ically with PiSNF1. The same results were also obtained when using the target proteins as baits [fig_ref] Figure 5: PiSNF1 is the target of StPR1 [/fig_ref]. The interactions of StPR1.2 DSP , StPR1.3 DSP , and StPR1.8 DSP with PiSNF1 were further confirmed by co-immunoprecipitation when co-expressed in S. cerevisiae [fig_ref] Figure 5: PiSNF1 is the target of StPR1 [/fig_ref]. To investigate whether StPR1 proteins and PiSNF1 exhibited similar subcellular distributions, StPR1.2 RFP , StPR1.3 RFP , and StPR1.8 RFP with PiSNF1 GFP were transiently coexpressed in N. benthamiana. Confocal microscopy showed that StPR1.2, StPR1.3, and StPR1.8 displayed the same distribution as PiSNF1 at 3 dpi [fig_ref] Figure 5: PiSNF1 is the target of StPR1 [/fig_ref] , consistent with the observation that they interact with each other.
In AMPK, the catalytic subunit (a) and two regulatory subunits [fig_ref] Figure 6: AMPK kinase complex in Phytophthora infestans is the target of StPR1 [/fig_ref]. Y2H and BiFC assays showed that B-PiSNF1, B-PiAMPKb and PiAMPKc interacted with each other to form a heterotrimer, as in plants and animals [fig_ref] Figure 6: AMPK kinase complex in Phytophthora infestans is the target of StPR1 [/fig_ref] , while the kinase domains in PiSNF1 and glycogen binding domain in PiAMPKb were not involved in the interaction [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref]. Moreover, the secretory StPR1 DSP , except for StPR1a DSP , also interacted with these three subunits through binding domains in Y2H assays [fig_ref] Figure 6: AMPK kinase complex in Phytophthora infestans is the target of StPR1 [/fig_ref] [fig_ref] Figure 6: AMPK kinase complex in Phytophthora infestans is the target of StPR1 [/fig_ref]. It is worth noting that StPR1.5, StPR1.6 and StPR1.7 could also interact with PiSNF1 [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref] , F). Further Y2H assays confirmed that StPR1 DSPDGH could interact with PiAMPKb and PiAMPKc but not PiSNF1, and that StPR1 DSPDTM could not interact with PiAMPKc [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref] , suggesting that GH was the binding motif of StPR1 with PiSNF1, and TM was the binding motif of StPR1 with PiAMPKc.
## Stpr1 affected the piampk phosphorylation of downstream target proteins
To assess the influence of StPR1 on PiAMPK kinase activity, the total protein of P. infestans mutants OE-StPR1.2, OE-StPR1.3 and OE-StPR1.8 were collected for western blot analysis to measure the phosphorylation levels of AMPK catalytic subunit (pAMPKa) in vivo. Surprisingly, overexpression of StPR1 DSP in P. infestans did not change the pAMPKa in vivo [fig_ref] Figure 7: StPR1 did not change the levels of pAMPKa, while affected the AMPK... [/fig_ref] and [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref] , which was further confirmed by ELISA analysis [fig_ref] Figure 7: StPR1 did not change the levels of pAMPKa, while affected the AMPK... [/fig_ref]. To demonstrate whether the TM or GH motif of StPR1 proteins is necessary to affect the pAMPKa, we performed phosphorylation assays in vitro and found that the pAMPKa treated with StPR1.2 DSP and its mutations (StPR1 DSPDTM , StPR1 DSPDGH ) were not significantly different to that of the control [fig_ref] Figure 7: StPR1 did not change the levels of pAMPKa, while affected the AMPK... [/fig_ref] and [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref]. The unchanged levels of pAMPKa were also observed in StPR1.3 DSP , StPR1.8 DSP , and their mutations [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref] , suggesting that the TM and GH motifs of StPR1 proteins were not necessary to alter the AMPKa kinase activity.
Even though the in vitro and in vivo phosphorylation assays demonstrated that StPR1 proteins have no influence on the AMPK kinase activity, it was subsequently found that StPR1 proteins altered the AMPK activation to downstream targets. Acetyl-CoA Carboxylase (ACC) is the conserved target substrate of AMPK kinase and is an appropriate choice for detecting the AMPK kinase activity, which is negatively regulated by AMPK at the site of Ser79 in ACC1 [bib_ref] The AMPK signalling pathway coordinates cell growth, autophagy and metabolism, Mihaylova [/bib_ref]. Compared with the wild-type strain, the pACC:ACC ratios were increased by 66.2%, 40.3%, and 44.5% in OE-StPR1.2, OE-StPR1.3 and OE-StPR1.8 mutants, respectively (P 0.001) [fig_ref] Figure 7: StPR1 did not change the levels of pAMPKa, while affected the AMPK... [/fig_ref]. Further ELISA analysis showed that StPR1.2 DSP , StPR1.2 DSPDTM , and StPR1.2 DSPDGH increased the levels of pACC Ser79 and the pACC:ACC ratio in vitro, while the influence of StPR1.2 DSP on pACC:ACC ratio was stronger than that of StPR1.2 DSPDTM and StPR1.2 DSPDGH [fig_ref] Figure 7: StPR1 did not change the levels of pAMPKa, while affected the AMPK... [/fig_ref]. The same results were obtained in P. infestans treated with StPR1.3 DSP , StPR1.8 DSP , and their mutations [fig_ref] Figure 7: StPR1 did not change the levels of pAMPKa, while affected the AMPK... [/fig_ref] , implying that TM and GH motifs are necessary for StPR1 to change the AMPK phosphorylation to its downstream targets.
## Ampk inhibition in p. infestans restricted vegetative growth and decreased pathogenicity
To gain insights into the function of AMPK activity, overexpression mutants of the catalytic subunit encoding gene (PiSNF1) fused with GFP (OE-PiSNF1) were generated [fig_ref] Figure 8: Functional characterization of PiSNF1 and the influence of AMPK activation and inhibition... [/fig_ref] and Supplementary [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref]. The overexpression of PiSNF1 decreased the colony growth with aerial hyphae growing almost 50% less compared with the wild-type [fig_ref] Figure 8: Functional characterization of PiSNF1 and the influence of AMPK activation and inhibition... [/fig_ref]. Likewise, compared with wild-type P. infestans (WT), the pathogenicity of OE-PiSNF1 was lower on leaves and tubers [fig_ref] Figure 8: Functional characterization of PiSNF1 and the influence of AMPK activation and inhibition... [/fig_ref]. Additionally, the phenotype of OE-PiSNF1 resembled that of the P. infestans treated with the AMPK kinase inhibitor dorsomorphin (800 lM; Selleck, cat no. S7306) [fig_ref] Figure 8: Functional characterization of PiSNF1 and the influence of AMPK activation and inhibition... [/fig_ref]. Western blot analysis revealed that the pAMPKa level of OE-PiSNF1 was lower than that of WT, while the pAMPKa level with dorsomorphin treatment was lower than that with DMSO treatment [fig_ref] Figure 8: Functional characterization of PiSNF1 and the influence of AMPK activation and inhibition... [/fig_ref]. However, the phenotype of P. infestans treated with the AMPK activator A-769662 (800 lM; Selleck, cat no. S2697) did not differ from the WT and the DMSO-treated P. infestans [fig_ref] Figure 8: Functional characterization of PiSNF1 and the influence of AMPK activation and inhibition... [/fig_ref]. Overall, these results indicate that the inhibition of AMPK activity decreased the vegetative growth and pathogenicity of P. infestans. StPR1-PiAMPK disrupted ROS homeostasis and inhibited the expression of RxLR effector-encoding genes in P. Infestans
Oxidative stress is critical for establishing pathogenesis [bib_ref] Host-pathogen interaction between asian citrus psyllid and entomopathogenic fungus (cordyceps fumosorosea) is..., Mqa [/bib_ref] [bib_ref] Temperature-dependent development of asian citrus psyllid on various hosts, and mortality by..., Qasim [/bib_ref]. The H 2 O 2 content was significantly lower in OE-StPR1.2, OE-StPR1.3, OE-StPR1.8, and OE-PiSNF1 than in WT [fig_ref] Figure 9: StPR1-PiAMPK signaling pathway is involved in regulating ROS homeostasis and RxLR expression... [/fig_ref] ; while compared with WT, the enzyme activity of ROS scavenging enzymes SOD, POD and CAT was higher in the mutants [fig_ref] Figure 9: StPR1-PiAMPK signaling pathway is involved in regulating ROS homeostasis and RxLR expression... [/fig_ref]. Apart from the changes in ROS homeostasis, the transcript levels of the RxLR effector-encoding genes were down-regulated in the mutants. Ten effectors have been published in P. infestans, including cystatin-like protease inhibitors (EPIC1, EEY55256; EPIC2, A1L016; EPIC2B, D0NBV3), extracellular serine protease inhibitors (EPI1, G8FQ60; EPI10, Q6PQG3), RxLR cytoplasmic effectors (AVRBLB1, D0P3S7; AVRBLB2, D0P1B2; AVR2, D0NN59; AVR3a, A5YTY8), and CRN cytoplasmic effectors (CRN8, Q2M405). Among them, PiAvr3a, a typical RxLR effector protein, targets and stabilizes the E3 ubiquitin ligase CMPG1 in plant cells to inhibit INF-induced necrosis [bib_ref] Phytophthora infestans effector AVR3a is essential for virulence and manipulates plant immunity..., Bos [/bib_ref] ; PiPexRD2, interacts with MAPKKKe and inhibits its kinase activity to obstruct the transmission of the MAPK immune signaling pathway [bib_ref] Phytophthora infestans RXLR effector PexRD2 interacts with host MAPKKK epsilon to suppress..., King [/bib_ref] ; PiAvr2 could activate the lipid signal transduction pathway and inhibit the PTI defense response induced by INF1 [bib_ref] RXLR effector AVR2 up-regulates a brassinosteroid-responsive bHLH transcription factor to suppress immunity, Turnbull [/bib_ref] ; PiAvrblb2 can block the secretion of C14, a disease resistance protein belonging to cysteine protease [bib_ref] Phytophthora infestans effector AVRblb2 prevents secretion of a plant immune protease at..., Bozkurt [/bib_ref]. Owing to their role in pathogenicity, the RxLR effectorencoding genes Avrblb2 (PITG_04085, PITG_04086), Avr2 (PITG_06077, PITG_08943, PITG_22870), Avr3a (PITG_14371), and PexRD2 (PITG_21422) were selected for quantitative analysis. qRT-PCR analysis showed that compared with WT, all the examined genes were down-regulated by more than 2-fold in the OE-StPR1.2, OE-StPR1.3, OE-StPR1.8, and OE-PiSNF1 [fig_ref] Figure 9: StPR1-PiAMPK signaling pathway is involved in regulating ROS homeostasis and RxLR expression... [/fig_ref] , indicating that inhibiting the expression of effectors was one strategy employed by PR1 proteins to protect plants against pathogen infection. In turn, AVR3a was found to bind with StPR1 through Y2H assays [fig_ref] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum [/fig_ref].
# Discussion
PR1 is an indispensable component of innate immune responses in plants under biotic or abiotic stresses and is activated by salicylic acid (SA) application. Its accumulation has long been used as a marker for SA-mediated disease resistance [bib_ref] Emerging insights into the functions of pathogenesis-related protein 1, Breen [/bib_ref]. Many researchers have reported that the PR1 proteins are involved in plant defenses against Phytophthora sp. [bib_ref] Pathogenesis-Related PR-1 proteins are antifungal (isolation and characterization of three 14-kilodalton proteins..., Niderman [/bib_ref] [bib_ref] The sterol-binding activity of Pathogenesis-Related Protein 1 reveals the mode of action..., Gamir [/bib_ref] , while their function and mode of action are still unclear. During P. infestans infection, the ten PR1 genes in S. tuberosum were up-regulated and their transcript levels were higher in the resistant variety 'E3 0 than those in the susceptible variety 'Desiree' [fig_ref] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo [/fig_ref]. Silencing of PR1 proteins increased the susceptibility to the pathogen, with the disease index caused by secretory PR1 proteins being higher than that caused by non-secretory PR1 proteins [fig_ref] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo [/fig_ref]. More-over, the overexpression of secretory PR1 proteins in potato enhanced its resistance to late blight, suggesting the critical role of secretory PR1 proteins in plant defense. In vitro analysis indicated that the secretory PR1 proteins inhibited the sporangiospore germination significantly [fig_ref] Figure 6: AMPK kinase complex in Phytophthora infestans is the target of StPR1 [/fig_ref] , and the heterologous expression of secretory PR1 proteins in P. infestans inhibited its vegetative growth and pathogenicity [fig_ref] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo [/fig_ref] -H and Supplementary [fig_ref] Figure 5: PiSNF1 is the target of StPR1 [/fig_ref]. These results agree with the findings of Niderman et al. [bib_ref] Pathogenesis-Related PR-1 proteins are antifungal (isolation and characterization of three 14-kilodalton proteins..., Niderman [/bib_ref] and Gamir et al. [bib_ref] The sterol-binding activity of Pathogenesis-Related Protein 1 reveals the mode of action..., Gamir [/bib_ref]. Even though there has been considerable interest in PR1 proteins for several decades, its necessary and cross-kingdom translocations in plant defenses remain poorly understood. Pecenková et al. [bib_ref] Subcellular localization of Arabidopsis pathogenesis-related 1 (PR1) protein, Pečenková [/bib_ref] demonstrated that in A. thaliana, the PR1 (AtPR1) has an unconventional secretion pathway, being transported via phosphatidyl-inositol-3-phosphate (PI(3)P)positive LE/MVB-like vesicles. Besides antimicrobial proteins, EVs also send host sRNAs into pathogens to silence virulent genes and inhibit their pathogenicity [bib_ref] Plants send small RNAs in extracellular vesicles to fungal pathogen to silence..., Cai [/bib_ref]. The EVs marker proteins TET8 and TET9 were up-regulated during P. infestans infection; however, further research is necessary to demonstrate the relationship between PR1 proteins and EVs, and to establish whether PR1 proteins are indeed transported via EVs
Since the identification of CAPE peptides and the establishment of their role in biotic and abiotic stress responses, there have been significant advances in our understanding in relation to the function of PR1 proteins [bib_ref] Emerging insights into the functions of pathogenesis-related protein 1, Breen [/bib_ref]. demonstrated that PR1 proteins possess a sterol-binding activity, and their inhibitory effects on pathogen growth are attributed to the deprivation of sterols from pathogens, including the sterol-auxotroph oomycete analysis. The analysis of pACC:ACC ratio of P. infestans without PR1 treatment was used as a control. Results that were significantly different from control (CK) were indicated by asterisks. One-tailed t-tests were used to assess significances. Error bars indicate standard deviation (SD). In E-H: *, P 0.05; **, P 0.01; ***, P 0.001. pathogens. The CAP domain is responsible for the sterol-binding function. The conserved consensus motif of CAP-derived peptide 1 (CAPE1) is PxGNxxxxxPY, which is a 11 aa peptide cleaved from the C terminus of cysteine-rich secretory proteins PR1 and PR5 in various monocot and dicot plants [bib_ref] Quantitative peptidomics study reveals that a wound-induced peptide from PR-1 regulates immune..., Chen [/bib_ref] [bib_ref] A salt-regulated peptide derived from the CAP superfamily protein negatively regulates salt-stress..., Chien [/bib_ref]. The bioactive CAPE1 can induce defense genes to produce immune responses against herbivores, pathogens, and abiotic stresses [bib_ref] Quantitative peptidomics study reveals that a wound-induced peptide from PR-1 regulates immune..., Chen [/bib_ref] [bib_ref] A salt-regulated peptide derived from the CAP superfamily protein negatively regulates salt-stress..., Chien [/bib_ref]. It is essential for host defense responses by acting as a defense-signaling molecule rather than as a protein with a direct antimicrobe/herbivore function [bib_ref] Emerging insights into the functions of pathogenesis-related protein 1, Breen [/bib_ref]. In our research, the secretory PR1 proteins were found to have a different function by acting as kinase regulators. In particular, they translocated from the host to pathogen cells to target the Ser/Thr protein kinase AMPK in P. infestans, leading to changes in AMPK phosphorylation of downstream target proteins.
The TM and GH motifs are key domains for interacting with AMPK, while the CAP domain is not necessary for the interaction, which probably explains why secretory PR1 proteins have a direct antimicrobial function.
AMPK is a Ser/Thr protein kinase that is widely distributed in eukaryotic cells. It is a highly conserved sensor of energy and metabolism and is one of the key regulators of growth and development. In most species, AMPK exists as an obligate heterotrimer, containing a catalytic subunit (a) and two regulatory subunits (b and c) [bib_ref] The AMPK signalling pathway coordinates cell growth, autophagy and metabolism, Mihaylova [/bib_ref]. Phosphorylation of Thr172 in the activation loop of AMPK is required for AMPK activation, which is directly mediated by the Ser/Thr kinase LKB1 [bib_ref] The tumor suppressor LKB1 kinase directly activates AMP-activated kinase and regulates apoptosis..., Shaw [/bib_ref]. Additionally, AMPK can also be phosphorylated on Thr172 in response to a calcium flux, independent of LKB1, by CAMKK2 (also known as CAMKKb) kinase [bib_ref] The Ca 2+ /Calmodulin-dependent orotein kinase kinases are AMPactivated protein kinase kinases, Hurley [/bib_ref]. In pathogenicity tests, the leaves or tubers infected by wild-type P. infestans were used as positive control (WT). CK, leaves or tubers inoculated with sterile water. Each pathogenicity test contained at least 10 leaves or tubers, respectively. Each experiment was repeated three times. Asterisks indicate significant differences (*P 0.05; **P 0.01; ***P 0.001; t-test).
According to the alignment sequences of the established physiological substrates, the conserved phosphorylation motif øbxxx(S/ T) is the most well-defined substrate recognition motif by AMPK [bib_ref] Protein kinase substrate recognition studied using the recombinant catalytic domain of AMPactivated..., Scott [/bib_ref]. In S. cerevisiae, the a and c subunits are annotated as SNF1 and SNF4, respectively [bib_ref] Crystal structure of the heterotrimer core of Saccharomyces cerevisiae AMPK homologue SNF1, Amodeo [/bib_ref] ; and in fungi, SNF1 is closely related to vegetative growth and pathogenicity [bib_ref] VdSNF1, the sucrose nonfermenting protein kinase gene of Verticillium dahliae, is required..., Tzima [/bib_ref]. However, the function of SNF1 in oomycetes has not been well established. By constructing SNF1 mutants and treating P. infestans with AMPK activators and inhibitors, AMPK was demonstrated to be involved in vegetative growth and pathogenicity, accompanying with regulating the expression of effector-encoding genes and ROS homeostasis. However, instead of affecting the AMPK activity through T172 phosphorylation, PR1 proteins bound to the three subunits of the AMPK complex, thus changed the AMPK-driven phosphorylation to downstream targets. For example, PR1 proteins strengthen the ACC phosphorylation level, which is negatively regulated by AMPK. It is a discovery with regards to the antimicrobial function of the secretory PR1 proteins.
To date, PR1 proteins are regarded as pathogen effector hubs. Three effectors from different fungal plant pathogens have been found to interact with host PR1 proteins, including ToxA, Tox3, and CSEP0055 [bib_ref] Wheat PR-1 proteins are targeted by necrotrophic pathogen effector proteins, Breen [/bib_ref] [bib_ref] A dimeric PR-1-type pathogenesis-related protein interacts with ToxA and potentially mediates ToxA-induced..., Lu [/bib_ref] [bib_ref] Interaction of barley powdery mildew effector candidate CSEP0055 with the defence protein..., Zhang [/bib_ref]. ToxA is a universal effector in Parastagonospora nodorum, Pyrenophora tritici-repentis, and Bipolaris sorokiniana, while Tox3 is unique to P. nodorum [bib_ref] The discovery of the virulence gene ToxA in the wheat and barley..., Mcdonald [/bib_ref]. They all interact with amino acids toward the C terminus of PR1, and the ToxA/ Tox3-PR1 interaction facilitates pathogen infection of the host [bib_ref] Wheat PR-1 proteins are targeted by necrotrophic pathogen effector proteins, Breen [/bib_ref] [bib_ref] Interaction of barley powdery mildew effector candidate CSEP0055 with the defence protein..., Zhang [/bib_ref]. The third effector CSEP0055 that interacted with PR1, was identified in powdery mildew fungus Blumeria graminis [bib_ref] Interaction of barley powdery mildew effector candidate CSEP0055 with the defence protein..., Zhang [/bib_ref] , but no further information has been reported to describe the biological effects of their interaction. In our work, independent Y2H assays showed that the RxLR effector Avr3a from P. infestans could interact with potato PR1 proteins without specificity, and this interaction in turn regulates the resistance or susceptibility of the host. As PR1 has a critical role in plant immunity, the identification of more effectors from various plant pathogens that target these proteins and associated pathways will be of great interest [bib_ref] Emerging insights into the functions of pathogenesis-related protein 1, Breen [/bib_ref]. Additionally, PR1 reversely inhibited the expression of effec-tors. More detailed insights into PR-mediated defense responses will contribute to our understanding of pathogen-plant communication, and shed light on the putative roles of StPR1-PiAMPK for disease control.
# Conclusion
In conclusion, we found that plant extracellular PR1 play a crucial role in the cross-kingdom trafficking between S. tuberosum and the oomycete pathogen P. infestans. Cross-kingdom trafficking events have been reported in Arabidopsis, which result in the secretion of exosomes to deliver host sRNAs into fungal cells and silence pathogenicity-related genes in the fungal pathogen B. cinerea [bib_ref] Plants send small RNAs in extracellular vesicles to fungal pathogen to silence..., Cai [/bib_ref]. However, the same phenomenon was rarely reported in antifungal proteins, such as PR1. Functional analysis showed that secretory PR1 proteins have anti-oomycete activity and can translocate from the host to the pathogen. The translocated PR1 proteins target Ser/ Thr protein kinase AMPK in P. infestans, thereby reducing AMPKdriven phosphorylation of downstream proteins, and inhibiting the vegetative growth and pathogenicity. The binding mechanisms of PR1 with AMPK could be a potential target for developing new biofungicide, while the identification of the molecular function of PR1 could provide effective strategies to precisely control plant diseases in various crops, both at pre-and post-harvest stages.
## Compliance with ethics requirements
This article does not contain any studies with human or animal subjects. (E) qRT-PCR analysis showed that compared with WT, the transcript of RxLR effectors encoding genes were down-regulated in OE-StPR1.2, OE-StPR1.3, OE-StPR1.8, and OE-PiSNF1 mutants. Y-axis indicated the relative normalized expression. Error bars represent standard error from three independent replicates. Asterisks indicate significant differences (*P 0.05; ***P 0.001; t-test).
[fig] Figure 1: Bioinformatic analysis of StPR1 in Solanum tuberosum. (A) Ten genes encode pathogenesis-related protein 1 (PR1) in the genome of S. tuberosum. StPR1.1-1.9 are clustered on chromosome 1 whereas StPR1a is located at chromosome 9. (B) Phylogenetic analysis of PR1 proteins from S. tuberosum was performed with MEGA6.0 using the neighbor-joining method. One thousand bootstrap replicates were used. The PR1 proteins in red represent secretory proteins in S. tuberosum, while those in black indicate non-secretory proteins. The identity between StPR1.8 and other PR1 proteins was analyzed using Clustal X. The amino acid sequences were collected from the following organisms: St, S. tuberosum; At, Arabidopsis thaliana; Tom, Solanum lycopersicum; Tob, Nicotiana benthamiana. (C) The analysis of conserved domains of StPR1.1-1.9 in S. [/fig]
[fig] Figure 2: StPR1 involved in plant defense against Phytophthora infestans in vivo. (A, B) Compared with the wild-type 'Desiree' (WT), the disease resistance to P. infestans in transgenic lines overexpressing StPR1.2, StPR1.3 and StPR1.8 was enhanced. 35S::GUS represent GUS in S. tuberosum under 35S promoter. (C, D) Compared with the wild-type resistant variety 'E3 0 (WT), the RNAi lines RNAi-PR1 and RNAi-SPR1 displayed enhanced susceptibility to P. infestans with higher DI, and the disease symptoms on RNAi-PR1 lines were more serious. RNAi-PR1 indicates that all the PR1 genes were knocked down, and the RNAi-SPR1 indicates that only the secretory PR1 genes were knocked down. The DIs of transgenic lines were measured after inoculating 1 Â 10 7 conidia of wild-type P. infestans for 5 days. (E, F) Heterologous overexpression of StPR1.2, StPR1.3 and StPR1.8 in P. infestans inhibited the colony growth and decreased the growth rate of P. infestans. WT, wild-type P. infestans; OE-StPR1, P. infestans mutants overexpressing StPR1. (G, H) Reduced pathogenicity of OE-StPR1.2, OE-StPR1.3 and OE-StPR1.8 on 'Desiree' leaves with lower DIs after 5 days of inoculation. The leaves infected by WT were used as a control. (I, J) Heterologous expression of StPR1 in P. infestans decreased the infection on tubers after 10 days of inoculation. The tubers infected by WT were used as control. CK in G-J, leaves and tubers inoculated with sterile water. Each experiment has three biological replicates. Each replication of pathogenicity tests contained at least 10 leaves and tubers, respectively. Error bars indicate standard deviation (SD). Asterisks indicate significant differences (*P 0.05; **P 0.01; ***P 0.001; t-test). [/fig]
[fig] Figure 3: The GH and TM motifs are important for the anti-oomycete activity of StPR1 against Phytophthora infestans. (A) Overview of five types of StPR1 mutations (StPR1 DSP , StPR1 DTM , StPR1 DSPDTM , StPR1 DGH and StPR1 DSPDGH ) that were changed at the SP, GH and TM motifs. (B) The significant inhibition of prokaryotic expressed StPR1, StPR1 DSP , StPR1 DSPDGH and StPR1 DSPDTM on sporangia germination of P. infestans. The sporangia germination rates treated with StPR1 DSPDGH and StPR1 DSPDTM were higher than those treated with StPR1 and StPR1 DSP . The sporangia germination treated with GUS was used as a control (CK). (C) Three necrosis grades are defined according to the percentage of the necrosis area developed in the pathogen inoculation area. (D) Phenotypes of plant necrosis triggered by P. infestans in Nicotiana benthamiana leaves transiently expressing PR1, PR1 DSP , PR1 DTM , PR1 DSPDTM , and PR1 DGH indicate that the deletion of SP, GH and TM motifs decreased the anti-oomycete activity of StPR1. The necrosis caused by P. infestans in N. benthamiana expressing GFP was used as a control (CK). (E-G) The disease grades of plant necrosis triggered by P. infestans in N. benthamiana expressing StPR1and mutations. One-tailed t-tests were used to assess statistical significance between means. Error bars indicate standard deviation (SD). In (B, E, F, and G): *, P 0.05; **, P 0.01; ***, P 0.001. [/fig]
[fig] Figure 4: Cross-kingdom translocation of StPR1 from host to Phytophthora infestans in vivo and in vitro. (A, B) The GUS histochemical assays of 35S::StPR1.2 GUS transgenic plants during P. infestans infection (B). The assays in 35S::GUS were used as control (A). The GUS staining was attenuated in 35S::StPR1.2 GUS as the inoculation day progressed, while that was not in the 35S::GUS. (C) GUS histochemical assays showed that the prokaryotic expressed StPR1.2 and its mutations tagged with GUS could translocate into P. infestans cells after 12 h of co-incubation in vitro. The hyphae co-incubated with GUS were used as control. The scale bars represent 25 lm. (D, F) Western blot analysis further confirmed the presence of StPR1.2 (D), StPR1.3 (E), StPR1.8 (F) proteins and their mutations in the hyphae of P. infestans cells after 12 h of co-incubation. Purified prokaryotic expressed StPR1 and mutations are marked with red asterisks. Total proteins of P. infestans used for western blot analysis were extracted after 12 h of co-incubation with indicated proteins. b-actin was used as a loading control. (G) The specific marker protein of extracellular vesicles (EVs), TET8, was detected to accumulate significantly during P. infestans infection by western blot analysis. b-actin was used as a loading control. [/fig]
[fig] Figure 5: PiSNF1 is the target of StPR1. (A) Phylogenetic relationship among catalytic subunits of AMPK from different species. The phylogenetic tree was constructed using MEGA6.0 based on an alignment generated with ClustalX. Sl: Solanum lycopersicum, St: Solanum tuberosum, At: Arabidopsis thaliana, Pi: Phytophthora infestans, Sc: Saccharomyces cerevisiae, Vd: Verticillium dahliae, Fo: Fusarium oxysporum, GRCH: human. (B) StPR1 proteins interacts with PiSNF1 as determined by Y2H assays. PiSNF1 on pGBKT7 (BD) vector was used to confirm the interactions with secretory StPR1 DSP cloned into pGADT7 (AD). Yeast transformants were grown on SD/-Leu/-Trp/-His/-Ade medium with 40 g/mL X-a-gal, and the blue colony indicate interactions. Pictures were taken after 3 d of culture. (C) Y2H confirmed the interaction between PiSNF1 and StPR1 DSP using the target proteins as baits. (D) Co-IP assays demonstrating the interactions of StPR1.2, StPR1.3, and StPR1.8 with PiSNF1. The proteins used for Co-IP were derived from S. cerevisiae co-expressing PiSNF1 Myc with StPR1.2 HA , StPR1.3 HA , and StPR1.8 HA . (E) Co-localization of RFP-labeled StPR1.2, StPR1.3, StPR1.8 and GFP-labeled PiSNF1. The fusion constructs were transiently co-expressed in N. benthamiana and the expressions were analyzed by confocal microscopy at 3 dpi. Bar = 50 lm. The N.benthamiana infiltrated with MMA buffer was used as control. [/fig]
[fig] Figure 6: AMPK kinase complex in Phytophthora infestans is the target of StPR1. (A) Structures of the AMPK complex. AMPK exists as a heterotrimer consisting of a catalytic subunit (a) and two regulatory subunits (b and c). (B, C) Y2H (B) and bimolecular fluorescence complementation (BiFC) (C) assays showed that B-PiSNF1, B-PiAMPKb and PiAMPKc interact with each other to form a heterotrimer as those in plant and animal. (D) B-PiSNF1, B-PiAMPKb and PiAMPKc interact with secretory StPR1 DSP in Y2H assays. (E) The AMPK subunits interact with StPR1.2 DSP , StPR1.3 DSP , and StPR1.8 DSP in BiFC assays. The StPR1.2 DSP , StPR1.3 DSP , and StPR1.8 DSP fused with the N terminus of YFP, and AMPK subunits fused with the C terminus of YFP. The constructs were transiently co-expressed in N. benthamiana and examined by confocal microscopy at 3 d postinfiltration (dpi). Only the nYFP halves and single proteins fused to cYFP alone were used as control. The complementation of fluorescence indicates interaction between assayed proteins. [/fig]
[fig] Figure 7: StPR1 did not change the levels of pAMPKa, while affected the AMPK phosphorylation to downstream target proteins. (A) Western blot of AMPK and phosphorylation of AMPKa (pAMPKa) in wild-type (WT) and PR1 overexpression mutants of P. infestans (OE-StPR1.2, OE-StPR1.3 and OE-StPR1.8). b-actin was used as a loading control. The AMPK expression and pAMPKa level were not different between WT and mutants. (B) Quantification of pAMPKa from WT, OE-StPR1.2, OE-StPR1.3 and OE-StPR1.8 (n = 3) through ELISA analysis. The pAMPKa levels were not significantly different among them. (C, D) Phosphorylation of AMPKa under the involvement of StPR1.2 DSP , StPR1.3 DSP , StPR1.8 DSP , and their mutations through phosphorylation assay in vitro. StPR1.2 DSP , StPR1.3 DSP , StPR1.8 DSP , and their mutations had no influence on the phosphorylation levels of AMPKa. StPR1 DSP : deleting the signal peptide; StPR1 DSPDTM : deleting the signal peptide and TM motif; StPR1 DSPDGH : deleting the signal peptide and GH motif. (E) Quantification of pACC:ACC ratio from WT, OE-StPR1.2, OE-StPR1.3 and OE-StPR1.8 (n = 3) through ELISA analysis. The pACC:ACC ratio was increased significantly in OE-StPR1.2, OE-StPR1.3 and OE-StPR1.8. (F, H) Quantification of pACC:ACC ratio after 48 h of treatment with StPR1 DSP , StPR1 DSPDTM , and StPR1 DSPDGH (n = 3) through ELISA [/fig]
[fig] Figure 8: Functional characterization of PiSNF1 and the influence of AMPK activation and inhibition on the colony growth and pathogenicity of Phytophthora infestans. (A) The colony phenotype of PiSNF1 overexpression mutant (OE-PiSNF1). (B) GFP fluorescence observation of OE-PiSNF1 under a fluorescence microscope. (C) Overexpression of PiSNF1 decreased the colony growth rate of P. infestans. (D, E) Reduced pathogenicity of OE-PiSNF1 on 'Desiree' leaves with lower DI after inoculating for 5 days. (F, G) The pathogenicity of OE-PiSNF1 on tubers was decreased. (H) Western blot analysis of pAMPKa in wild-type P. infestans (WT), wild-type P. infestans treated with DMSO (DMSO), wild-type P. infestans treated with AMPK activator A-769662 (A-769662), wild-type P. infestans treated with AMPK kinase inhibitor dorsomorphin (Dorsomorphin), and OE-PiSNF1. (I) Quantification of pAMPKa in OE-PiSNF1 mutant and under AMPK activator/inhibitor treatment. (J, K) Reduced colony growth of P. infestans following dorsomorphin treatment. The colony growth of P. infestans under the treatment of A-769662 and DMSO was not different from that of wild-type. (L, M) Reduced pathogenicity of Dorsomorphin on 'Desiree' leaves with lower DI after 5 days of inoculation. (N, O) Dorsomorphin decreased the infection of P. infestans on tubers. [/fig]
[fig] Figure 9: StPR1-PiAMPK signaling pathway is involved in regulating ROS homeostasis and RxLR expression in Phytophthora infestans. (A) The H 2 O 2 content was lower in OE-StPR1.2, OE-StPR1.3, OE-StPR1.8, and OE-PiSNF1 than in wild-type P. infestans (WT). (B-D) Compared with WT, the enzyme activities of SOD, POD, and CAT were significantly increased in OE-StPR1.2, OE-StPR1.3, OE-StPR1.8, and OE-PiSNF1 mutants. [/fig]
|
Autocatalytic Time-Dependent Evolution of Metastable Two-Component Supramolecular Assemblies to Self-Sorted or Coassembled State
Materials and MethodsUnless otherwise noted, reagents and solvents were purchased from commercial suppliers without further purification. Compound 2 1 , 5 2 , and 6 2 were prepared according to reported procedures. Ultraviolet-visible absorption spectra were recorded using a quartz cuvette of 1.0 or 0.1 cm path length on a Jasco V-630 spectrophotometer equipped with a Jasco ETCS-761 cell holder for temperature control. Atomic force microscopy (AFM) was performed on a Bruker model MultiMode 8 atomic force microscope under ambient conditions in the scan assist analysis. AFM images were analyzed with Bruker Nanoanalysis.3Supplementary Figures and TablesSupplementaryFigure S1. Absorption spectra of J aggregate (pink), H aggregate (green), and short-slipping J aggregate (orange) of 6 in methylcyclohexane 2 .SupplementaryFigure S2. Time profile of changes in the degree of aggregation of 1 H aggregate 1 . Supplementary Figure S3. Time profile of changes in the degree of aggregation of 6 short-slipping J aggregate 2 . 6 H agg. 6 J agg. ε (10 5 l mol -1 cm -1 ) 0 1.0 350 550 Wavelength (nm) 450 650 3.0 2.0 6 short-slipping J agg. 0 8 2 4 Time (h) H aggregate J aggregate 6 Degree of aggregation 0.2 1.0 0 0.4 0.8 0.6 t 50 0 8 2 4 Time (h) short-slipping J aggregate J aggregate 6 Degree of aggregation 0.2 1.0 0 0.4 0.8 0.6 t 50 4 Thermodynamic analysis of 2/6 J aggregate Supplementary Figure S4. (A) Temperature-dependent absorption spectra of 2/6 mixture (Q band) observed during the cooling process: [2]+[6] = 50 µM in MCH; [6]/{[2]+[6]} = 90%. (B) Change in the degree of aggregation of 2/6 J aggregate as a function of temperature obtained by fitting the apparent absorption coefficients at λ = 561 nm to the isodesmic model (Supplementary Equation 1) 3 in which α Mono is the degree of 2/6 Mono (monomeric 2/6 mixture), ε Mono and ε J are the molar absorption coefficients of 2/6 Mono and 2/6 J aggregate, respectively, and ε (T) is the apparent absorption coefficient at the given temperature. The fitting to the isodesmic model yielded ΔH° and T m as summarized in Supplementary Table 1 (correlation coefficient of 0.999). (C) Degree of polymerization number of 2/6 mixture as a function of the temperature obtained by the isodesmic model (Supplementary Equation 2) 3 . (D) Plot of the natural logarithm of K I as a function of the reciprocal temperature (T −1 ): van't Hoff plot showing a linear relationship (correlation coefficient of 0.997). The standard enthalpy change (ΔH°) and standard entropy change (ΔS°) for the formation of 2/6 J aggregate were determined using Supplementary Equation 3 in which R is the ideal gas constant. Absorbance 0 0.4 0.8 1.2 0.2 0.6
# Materials and methods
Unless otherwise noted, reagents and solvents were purchased from commercial suppliers without further purification. Compound 2 1 , 5 2 , and 6 2 were prepared according to reported procedures. Ultraviolet-visible absorption spectra were recorded using a quartz cuvette of 1.0 or 0.1 cm path length on a Jasco V-630 spectrophotometer equipped with a Jasco ETCS-761 cell holder for temperature control. Atomic force microscopy (AFM) was performed on a Bruker model MultiMode 8 atomic force microscope under ambient conditions in the scan assist analysis. AFM images were analyzed with Bruker Nanoanalysis.
## Supplementary figures and tables
Supplementary . Absorption spectra of J aggregate (pink), H aggregate (green), and short-slipping J aggregate (orange) of 6 in methylcyclohexane 2 .
## Supplementary
[formula] ! = 1 − !"#" = 1 − − ! !"#" − ! = 1 1 + −0.908°− ! ! ! (1) ! = 1 1 − ( ) = 1 2 + 1 2 4 ! ! + 1 (2) ln ! = −°+° (3) [/formula]
Summary of thermodynamic analyses of 2/6 J aggregate . Thermodynamic parameters.
isodesmic model 3 van't Hoff plot x 100 (%) 8
[formula] [6]/{[2]+[6]} (%) ΔH° (kJ mol −1 ) T m (K) ΔH° (kJ mol −1 ) ΔS° (J mol −1 K −1 )ΔH° (kJ mol −1 ) T m (K) ΔH° (kJ mol −1 ) ΔS° (J mol −1 K −1 ) ΔG° [/formula]
[fig] Figure S2, Figure S4: Time profile of changes in the degree of aggregation of 1 H aggregate 1 . Supplementary Figure S3. Time profile of changes in the degree of aggregation of 6 short-slipping J aggregate 2 . (A) Temperature-dependent absorption spectra of 2/6 mixture (Q band) observed during the cooling process: [2]+[6] = 50 µM in MCH; [6]/{[2]+[6]} = 90%. (B) Change in the degree of aggregation of 2/6 J aggregate as a function of temperature obtained by fitting the apparent absorption coefficients at λ = 561 nm to the isodesmic model (Supplementary Equation 1) 3 in which α Mono is the degree of 2/6 Mono (monomeric 2/6 mixture), ε Mono and ε J are the molar absorption coefficients of 2/6 Mono and 2/6 J aggregate, respectively, and ε (T) is the apparent absorption coefficient at the given temperature. The fitting to the isodesmic model yielded ΔH° and T m as summarized in Supplementary Table 1 (correlation coefficient of 0.999). (C) Degree of polymerization number of 2/6 mixture as a function of the temperature obtained by the isodesmic model (Supplementary Equation 2) 3 . (D) Plot of the natural logarithm of K I as a function of the reciprocal temperature (T −1 ): van't Hoff plot showing a linear relationship (correlation coefficient of 0.997). The standard enthalpy change (ΔH°) and standard entropy change (ΔS°) for the formation of 2/6 J aggregate were determined using Supplementary Equation 3 in which R is the ideal gas constant. [/fig]
[fig] SupplementaryFigure S7: (A-D) Time-dependent absorption spectra changes of two-component 5/6 J aggregate : [5]+[6] = 50 µM; [6]/{[5]+[6]} = (A) 0%, (B) 10%, (C) 50%, and (S8. (A) Time-dependent absorption spectra changes of 5/6 H aggregate: [5]+[6] = 50 µM; [6]/{[5]+[6]} = 50%. (B) Time profile of changes in the absorbance at 400 nm and 444 nm. Absorbance at 400 nm and 444 nm are characteristic of the H aggregate and the short-slipping J aggregate, respectively. The pseudo-self-sorted system of H aggregate and short-slipping J aggregate Supplementary Figure S9. (A) Time-dependent absorption spectra changes of mixture of 5 H aggregate and 6 short-slipping J aggregate: [5]+[6] = 50 µM; [6]/{[5]+[6]} = 50%. (B) Time profile of changes in the absorbance at 400 nm and 444 nm. Absorbance at 400 nm and 444 nm are characteristic of the H aggregate and the short-slipping J aggregate, respectively. [/fig]
[table] Table S2: Thermodynamic parameters. [/table]
|
Molecular Neuroimaging of the Dopamine Transporter as a Patient Enrichment Biomarker for Clinical Trials for Early Parkinson's Disease
[bib_ref] Precompetitive data sharing as a catalyst to address unmet needs in Parkinson's..., Stephenson [/bib_ref] [bib_ref] The Parkinson Progression Marker Initiative (PPMI), Marek [/bib_ref] [bib_ref] Precompetitive data sharing as a catalyst to address unmet needs in Parkinson's..., Stephenson [/bib_ref]
# Methods
Regulatory pathway with the EMA The regulatory strategy was designed around the pathway for the qualification of novel methodologies in drug development.This process started with the submission of a letter of intent and briefing package that included: (i) the proposed context of use (COU [fig_ref] Table 1: Description of context of use statement COU component Description General area Enrichment... [/fig_ref] , (ii) a comprehensive analysis plan (e.g., statistical model development and evaluation, assessment of magnitude of motor scores worsening, and enrichment utility), and (iii) target data sets to support the proposed analysis plan. This, in turn, triggered a formal review by the EMA's SAWP, followed by a face-to-face meeting in which the SAWP issued formal scientific advice for optimization and finalization of the COU statement and analysis plan. Subsequently, the modeling analyses were executed, and, upon their completion, a final qualification package was submitted to the EMA for a final review and SAWP meeting. This final meeting was aimed at reaching a final determination if the presented results constituted supporting evidence for the proposed COU. Afterward, the SAWP met with the CHMP, who made the final decision to issue a qualification opinion. 8
## Data sources
Integrated patient-level data from the Parkinson's Progression Markers Initiative (PPMI) and the Parkinson Research Examination of CEP-1347 Trial (PRECEPT) were analyzed. [bib_ref] The Parkinson Progression Marker Initiative (PPMI), Marek [/bib_ref] These data correspond to a total of 672 subjects diagnosed with early-stage PD and a total of 4,521 observations in the baseline to 25-month interval 7 [fig_ref] Figure 1: Integration of data sources and harmonization of motor scores [/fig_ref].
## Definition of biomarker status
Dopamine transporter neuroimaging status was treated as a binary covariate based on visual reads, with individuals classified as either having scans without evidence of dopaminergic deficit (SWEDD; or biomarker-negative) or having scans with evidence of dopaminergic deficit (biomarker-positive).
## Harmonization of motor scores
The transformation of the individual Unified Parkinson's Disease Rating Scale (UPDRS) part III subtotal score in PRECEPT to the respective Movement Disorder Society (MDS)-UPDRS part III subtotal score in PPMI relied on a previously derived method. [bib_ref] Calibration of unified Parkinson's disease rating scale scores to Movement Disorder Society-unified..., Goetz [/bib_ref] This allowed motor score integration across both studies.
Statistical modeling and clinical trial simulation As described by Conrado et al., [bib_ref] Dopamine transporter neuroimaging as an enrichment biomarker in early Parkinson's disease clinical..., Conrado [/bib_ref] an early motor PD progression model was developed using the harmonized MDS-UPDRS part III as the end point. Through this model, the difference in progression rate between SWEDDs and DAT deficit subjects and utility of DAT-based enrichment was determined.
In the PD progression model, the time course of the harmonized motor scores was described using a linear mixedeffects model. Prespecified covariates were the effect of biomarker status in baseline motor scores and progression rate and the effect of study in baseline motor scores to account for potential score differences between the PPMI and PRECEPT populations. Additional exploratory covariates included the effect of age in baseline motor scores and in progression rate given the neurodegenerative nature of PD and the effect of study in progression rate to investigate potential rate differences between the PPMI and PRECEPT. The final model included all the prespecified covariates (Eq. S1, Supplementary Material), and all statistically significant exploratory covariates.
Monte Carlo-based clinical trial simulations were performed to compare the statistical power vs. sample size in trials with and without DAT imaging enrichment. Enriched trials had only subjects with DAT deficit, whereas nonenriched trials included 15% of SWEDD subjects. [bib_ref] Dopamine transporter neuroimaging as an enrichment biomarker in early Parkinson's disease clinical..., Conrado [/bib_ref] The statistical power, defined herein as the probability of detecting a drug effect of 50% reduction in progression rate, was calculated as the proportion of trials for which the drug effect on progression rate was beneficial with a two-tailed P value lower than 0.05.
Additional detailed information on the methods has been published in ref. [bib_ref] Dopamine transporter neuroimaging as an enrichment biomarker in early Parkinson's disease clinical..., Conrado [/bib_ref].
# Results
## Disease progression model
The final linear mixed-effects model included: (i) effect of biomarker status on baseline, (ii) effect of biomarker status on progression rate, (iii) effect of study on baseline, and (iv) effect of age on baseline 7 [fig_ref] Figure 2: Predicted harmonized motor scores [/fig_ref]. Model diagnostics suggested an adequate fit of the longitudinal changes in the harmonized score. [bib_ref] Dopamine transporter neuroimaging as an enrichment biomarker in early Parkinson's disease clinical..., Conrado [/bib_ref] The main findings were:
- The estimated effect of SWEDD on progression rate was −0.13 points/month (90% confidence interval (CI): −0.23 to −0.04; one-tailed P value = 0.01). This means that SWEDDs have an average monthly progression in the harmonized motor scores that is 0.05 (90% CI: −0.04 to 0.13) points/month or 0.13 point/month lower than those with DAT deficit (0.18 points/month; 90% CI: 0.14−0.21). - The estimated effect of SWEDD on baseline was −7.69 (90% CI: −9.4 to −6.04) points; hence, SWEDDs have an average baseline harmonized motor score that is 7.69 points lower than those with DAT deficit. - The estimated effect of year of age on baseline was 0.19 (90% CI: 0.14−0.24) points, which means that on average, the baseline harmonized motor score increases by 0.19 points for each year of age. Thus, the baseline score for a typical 60 year old subject with DAT deficit is expected to be 21.54 points. 7
## Magnitude of motor scores worsening between biomarker statuses
The magnitude of motor scores worsening (defined as change from baseline at 24 months) in DAT deficit and SWEDD subjects was 4.28 (90% CI: 3.45−5.08) and 1.12 Clinical trial simulations and statistical power Simulated trial designs were placebo-controlled, parallel, and crossover with total duration of 12 and 24 months. For each design, 2,000 enriched and nonenriched clinical trials were simulated, yielding a total of 8 scenarios. For such scenarios, DAT imaging-based enrichment strategy was estimated to allow a 20-30% reduction of trial size.
# Discussion
The following were considered key issues to support the regulatory discussion:
1. Biomarker deficit status (SWEDD vs. DAT deficit) as a predictor of disease progression, even when the difference in baseline severity has been accounted for. o Although postural instability is a common feature in PD, based on the inclusion criterion of Hoehn and Yahr Stage I or II, postural instability would not be expected in the target population.
Stage of drug development for use
All clinical stages of early PD drug development, including proof-of-concept, dose-ranging through to confirmatory clinical trials. This is not intended for candidate therapies for more advanced stages of PD, such as drugs to treat L-Dopa-induced dyskinesia.
Intended application Purpose: The objective of this project is to apply DAT imaging as a biomarker tool to enrich subjects for clinical trials in early symptomatic PD by identifying subjects with a DAT deficit for possible inclusion into the study and excluding subjects who are unlikely to progress due to the lack of dopamine deficiency in the brain. The DAT imaging is intended to be used after the clinical criteria for early PD have been satisfied. 1. Potential candidates for PD clinical trials will be evaluated for the presence of at least two motor signs of PD, as described in the target population description of this section (according to the PPMI and PRECEPT criteria). 2. Those individuals are then evaluated according to the UK Brain Bank step 1 Criteria for PD. 3. If the two conditions above are met, subjects will undergo the trial-specific inclusion/exclusion criteria and further clinical assessment for atypical Parkinsonian syndromes. 4. As a final step in the subject-selection process, molecular imaging of DAT will be performed to detect the presence or absence of DAT-deficiency and identify and exclude subjects defined as SWEDDs. 5. Such baseline categorization of DAT-deficiency can be applied as an enrichment biomarker that, in combination with specific clinical signs, can more accurately predict disease progression of motor disability in early PD patients. Such progression will be expressed by the motor scores of the UPDRS or MDS-UPDRS scales, which constitute reliable outcomes of disease progression in PD. 6. Baseline categorization of DAT-deficiency can be applied as a subject selection biomarker to enrich trial populations with patients more likely to progress in the motor scores of UPDRS or MDS-UPDRS scale (parts II and III) over the course of clinical trials, which may be up to 2 years in duration. The purpose is to exclude patients who are unlikely to show disease progression (SWEDD), and consequently to increase the probability of the trial conclusively demonstrating the effect of an effective drug in clinical trials for therapeutic interventions for early PD. Those individuals who are not SWEDDs and who meet all the other selection criteria will be enrolled into the trial and randomized as per the specified study design. 7. The use of DAT imaging would allow the exclusion of subjects unlikely to have the diagnosis of PD and, therefore, prevent them from unjustified exposure to experimental PD-specific therapies with inherent safety and tolerability risks without anticipated benefit. 8. The application is relevant to both symptomatic and disease-modifying candidate therapies for early PD and is independent of the mechanism of action of the new drug. 9. The use of DAT imaging for diagnostic applications are out-of-scope for this proposed COU.
## Critical parameters for the context-of-use
The context-of-use specifies that reductions of DAT, as assessed by SPECT neuroimaging, will be utilized as an adjunct to clinical assessments for the purposes of enriching the patient population with subjects who have increased likelihood of having idiopathic PD. The subjects will have an objectively confirmed motor impairment with alternative identifiable causes of motor impairment appropriately excluded through clinical means prior to the use of DAT neuroimaging. SPECT neuroimaging procedures and methodologic aspects of imaging will be performed qualitatively in accord with the tracer manufacturer's specifications and consistent with the methods currently used in the multisite PPMI study. The proposed analysis of DAT SPECT images is by visual assessment by trained blinded readers and analysis is to be carried out by a single site. Such processes are expected to generate sufficiently accurate, reproducible, and robust assessment of DAT neuroimaging to facilitate clinical trial enrichment. Because the distribution of observed baseline motor scores shows some degree of overlap in the baseline scores between SWEDD and DAT deficit subjects [fig_ref] 3: Representativeness of PPMI and PRECEPT of the external SWEDD and DAT deficit... [/fig_ref] , a baseline-matched subset of the data was created. In this baseline-matched subset, DAT deficit subjects were included only if there was more than one SWEDD subject with the same observed baseline score (rounded to zero decimal places); likewise, SWEDD subjects were included only if there was more than one DAT deficit subject with the same observed baseline score (rounded to zero decimal places). A supplementary statistical analysis was then performed using this baseline-matched subset. Given the association between biomarker status and baseline motor scores, a baselinematched data set decreases the likelihood of confounding effects and helps investigate the separate contribution of baseline and biomarker status on the rate of progression. Results showed a significant difference in progression rate between the SWEDD and DAT deficit groups of −0.19 points/ month (two-tailed P value < 0.05) even after accounting for disease severity at baseline [fig_ref] Table 2: Parameter estimates from the supplementary analysis using the baseline matched subset [/fig_ref].
A second supplementary analysis was performed on totality of data (N = 672), including effect of baseline disease severity on progression rate. Such analysis yielded a difference in progression rate between SWEDD and DAT deficit subjects of −0.24 points/month (P value < 0.05). This provides further evidence for DAT imaging as an actual predictor of disease progression [fig_ref] Table 3: Parameter estimates from the supplementary analysis using the entire data set [/fig_ref].
## 2.
Similarities between SWEDD and DAT deficit subject entry criteria in PPMI and PRECEPT.
The clinical enrollment criteria for SWEDD and DAT deficit subjects were equivalent in PPMI and PRECEPT (i.e., SWEDD subjects were not recruited as a separate cohort and were not identified until after recruitment). In PPMI, DAT imaging was performed after subjects met the clinical criteria (as described in [fig_ref] Table 1: Description of context of use statement COU component Description General area Enrichment... [/fig_ref]. SWEDD subjects were asked to remain enrolled and then longitudinally followed for 2 years after consenting to remain in the study. The baseline characteristics of PPMI and PRECEPT are similar and representative of future clinical trial populations with early motor PD. Subjects enrolled in PPMI and PRECEPT are representative of the population likely to participate in PD clinical trials and the target population to be selected by DAT neuroimaging.
Clinical studies will continue to evaluate treatment response in earlier stages of PD, where it is known that there is greater uncertainty in selecting participants based on clinical criteria alone. Enrolling more homogeneous populations in these studies can help optimize clinical trial design and avoid exposing subjects who are less likely to progress to unknown test drugs.
## Similarity between swedd and dat deficit subject
imaging acquisition in PPMI and PRECEPT.
The technical aspects of the data acquisition of DAT single photon emission computed tomography (SPECT) images were identical between SWEDD and DAT deficit subjects and not a reason for dopaminergic differences between SWEDD and PD subjects. For PPMI, all subjects (SWEDD and DAT deficit) were aligned in terms of their imaging acquisition protocol, including the time interval between injection and SPECT reading (4 hours in duration). For PRECEPT, all imaging was done on a single research SPECT camera, and the data were managed by the core laboratory research group. CPP concluded that the dopaminergic differences between SWEDD and DAT deficit subjects are not due to variations in image acquisition. The potential impact of medications on DAT imaging was discussed. Symptomatic agents have been shown not to impact ligand binding. Drugs that bind to DAT (cocaine and amphetamines) were not permitted for PPMI enrollment. Antidepressants have been investigated, with no impact on the outcome of the visual assessments for DAT deficiency.
## Swedd subjects who experienced progression.
Possible reasons that could explain why some SWEDD subjects in the data set experienced progression: (i) they did not progress as typical patients with PD and were possibly dystonic tremor subjects, (ii) the test performed in these subjects was affected by external conditions that might have led to less signal in the striatum as compared with the occipital cortex and, hence, a false negative. As per the COU statement, diagnostic applications for DAT imaging are out of the scope of this work. DAT imaging could identify subjects with a homogeneous motor decline, allowing trial enrichment and meaningful reduction of sample size, regardless of the ultimate diagnosis. Moreover, as aforementioned, DAT imaging aspects were identical between SWEDD and DAT deficit subjects.
## 6.
Consideration of sensitivity, specificity, and predictive values for DAT imaging in the proposed COU.
As discussed, DAT imaging status is a statistically and clinically significant predictor of disease progression. Clinical trial simulations based on the underlying model allow the estimation of trial-specific DAT imaging-based enrichment magnitudes regardless of the ultimate diagnosis, thus providing a useful drug development tool to optimize decision making. As such, traditional concepts such as sensitivity, specificity, and predictive values, commonly applied to diagnostic biomarkers, are not of relevance in this enrichment context.
# Conclusions
These findings show that a DAT-SPECT finding of integrity of presynaptic dopaminergic terminals in a case of suspected PD is associated with a good prognosis, whatever the ultimate diagnosis. Exclusion of SWEDD subjects from future clinical trials will improve the chance of determining clinical benefit of new drug candidates to treat PD.
The EMA pathway for the Qualification of Novel Methodologies in drug development provides a valuable mechanism for the review and regulatory endorsement of DDTs, such as biomarkers and quantitative drug development tools. The final qualification opinion for DAT-SPECT imaging is publicly available.Regulatory endorsement provides sponsors with the necessary confidence to apply novel approaches to optimize drug development, which is much needed for neurodegenerative conditions such as PD.
Model-informed biomarker qualification is an efficient method to evaluate the utility of biomarker candidates for specific COU statements. The modeling approach presented herein relied on the time course of motor scores worsening. Traditional concepts applied to diagnostic biomarkers (e.g., sensitivity, specificity, and predictive values) were not needed to demonstrate the utility of DAT imaging as an enrichment biomarker for trials in early motor PD.
[fig] 3: Representativeness of PPMI and PRECEPT of the external SWEDD and DAT deficit population. [/fig]
[fig] Figure 2: Predicted harmonized motor scores. Subjects with and without DAT, dopamine transporter (DAT) deficit have an average monthly progression in scores of 0.18 (90% confidence interval (CI): 0.14, 0.21) and 0.05 (90% CI: −0.04, 0.13) points/month, respectively (Image reproduced from ref. 7 https://doi.org/10.1111/ cts.12492, is licensed under CC BY 4.0. ©2017 The authors.). [/fig]
[fig] Figure 1: Integration of data sources and harmonization of motor scores. DAT, dopamine transporter; MDS-UPDRS, Movement Disorder Society-Sponsored Revision of the Unified Parkinson Disease Rating Scale; PRECEPT, Parkinson Research Examination of CEP-1347 Trial.Dopamine Imaging to Enrich Parkinson TrialsRomero et al. [/fig]
[table] Table 1: Description of context of use statement COU component Description General area Enrichment biomarker for clinical trials in early motor PD Target population for usePatients with early motor PD, defined by the UK Brain Bank Criteria 10 as outlined below:• Having at least two of the following: resting tremor, bradykinesia, rigidity (must have either resting tremor or bradykinesia); OR either asymmetric resting tremor or asymmetric bradykinesia. o Based on above criteria, combinations could include: resting tremor/bradykinesia, bradykinesia/rigidity, and resting tremor/rigidity. o Symptom(s) or signs may include bradykinesia, a 4−6 Hz resting tremor, muscle rigidity, or postural instability not caused by primary visual, vestibular, cerebellar or proprioceptive dysfunction. • Hoehn and Yahr stage I or II at baseline. [/table]
[table] Table 2: Parameter estimates from the supplementary analysis using the baseline matched subset (N = 463) NS, indicates two-tailed P value > 0.05; PRECEPT, Parkinson Research Examination of CEP-1347 Trial; SWEDDs, scans without evidence of dopaminergic deficits.The estimated effect of SWEDD on progression rate of −0.19 points/month remains statistically significant even after accounting for the effect of baseline. SWEDDs have a progression rate that is 0.19 points/month lower than that observed for dopamine transporter deficit subjects. *Indicates two-tailed P value < 0.05. [/table]
|
(PS)2: protein structure prediction server
Protein structure prediction provides valuable insights into function, and comparative modeling is one of the most reliable methods to predict 3D structures directly from amino acid sequences. However, critical problems arise during the selection of the correct templates and the alignment of query sequences therewith. We have developed an automatic protein structure prediction server, (PS) 2 , which uses an effective consensus strategy both in template selection, which combines PSI-BLAST and IMPALA, and target-template alignment integrating PSI-BLAST, IMPALA and T-Coffee. (PS) 2 was evaluated for 47 comparative modeling targets in CASP6 (Critical Assessment of Techniques for Protein Structure Prediction). For the benchmark dataset, the predictive performance of (PS) 2 , based on the mean GTD_TS score, was superior to 10 other automatic servers. Our method is based solely on the consensus sequence and thus is considerably faster than other methods that rely on the additional structural consensus of templates. Our results show that (PS) 2 , coupled with suitable consensus strategies and a new similarity score, can significantly improve structure prediction. Our approach should be useful in structure prediction and modeling. The (PS) 2 is available through the website at
# Introduction
In the post-genomics era, one of the major challenges facing the structural biology research community is to determine the biological functions of genes identified through large-scale sequencing efforts. Knowledge of the 3D structure of a protein is crucial for understanding the molecular basis of its function. Unfortunately, the gap between the number of solved protein structures and the number of protein sequences continues to widen rapidly due to the long and expensive processes required for solving structures experimentally. Computational prediction of structures from amino acid sequence is an emerging and promising method that may help to narrow this gap. These methods have great potential to approximate the structure of newly acquired sequences based on known structures of similar sequence available from the rapidly growing number of protein crystal structures.
Comparative modeling generally comprises four main steps: (i) searching and selecting at least one known protein structure (the template) that is similar to the query (target sequence); (ii) alignment of the target sequence and the template(s); (iii) building models based on the chosen template(s); and (iv) evaluating the models. These steps can be reiterated until a satisfactory model structure is achieved. Currently, the first two steps are considered most critical because the accuracy of comparative models often tends to increase with the targettemplate sequence identity and the correctness of the alignment. A number of servers have been developed for automated comparative modeling [bib_ref] Automated prediction of CASP-5 structures using the robetta server, Chivian [/bib_ref] [bib_ref] EsyPred3D: prediction of proteins 3D structures, Lambert [/bib_ref] [bib_ref] An automatic homology modeling method consisting of database searches and simulated annealing, Ogata [/bib_ref] [bib_ref] Automatic consensus-based fold recognition using Pcons, ProQ, and Pmodeller, Wallner [/bib_ref] [bib_ref] PROTINFO: secondary and tertiary protein structure prediction, Hung [/bib_ref] [bib_ref] GenTHREADER: an efficient and reliable protein fold recognition method for genomic sequences, Jones [/bib_ref] [bib_ref] Protein threading using PROSPECT: design and evaluation, Xu [/bib_ref]. Several servers that yield predictions based on a set of different methods have demonstrated that consensus methods are significantly better than individual methods with regard to comparative modeling (3) and fold recognition [bib_ref] Automatic consensus-based fold recognition using Pcons, ProQ, and Pmodeller, Wallner [/bib_ref]. However, these methods have focused on target-template alignments and final model selections.
Here, we report the development of an automatic protein structure prediction server, (PS) 2 , using a consensus strategy applied both in the template search/selection and targettemplate alignment phase. (PS) 2 was tested for all comparative modeling targets (47 targets) in CASP6 (Critical Assessment of Techniques for Protein Structure Prediction) (9). Our consensus procedure is computationally efficient and scalable to a greater number of combinations. Our experimental results demonstrate improved prediction accuracy relative to other automatic servers based on GTD_TS score [bib_ref] LGA: a method for finding 3D similarities in protein structures, Zemla [/bib_ref].
# Methods and implementation
The efficiency of (PS) 2 derives from the ability to use an effective consensus strategy both in template selection [PSI-BLAST (11) and IMPALA [bib_ref] IMPALA: matching a protein sequence against a collection of PSI-BLAST-constructed position-specific score..., Schaffer [/bib_ref] The online version of this article has been published under an open access model. Users are entitled to use, reproduce, disseminate, or display the open access version of this article for non-commercial purposes provided that: the original authorship is properly and fully attributed; the Journal and Oxford University Press are attributed as the original place of publication with the correct citation details given; if an article is subsequently reproduced or disseminated not in its entirety but only in part or as a derivative work this must be clearly indicated. For commercial re-use, please contact [email protected] [fig_ref] Figure 1: Overview of the [/fig_ref]. (PS) 2 comprises the following four steps: template selection, target-template alignment, model building, and model evaluation and visualization. These steps are repeated until a satisfactory model structure is achieved. The (PS) 2 consensus approach uses a set of publicly available tools for template search/selection and target-template alignment to produce the inputs for MODELLER [bib_ref] Comparative protein modelling by satisfaction of spatial restraints, Sali [/bib_ref] , a comparative modeling tool based on the condition that spatial restraints must be satisfied.
For the easy use of the (PS) 2 server, it was designed to function with a minimum of user input, i.e., only the target sequence in FASTA format is needed [fig_ref] Figure 1: Overview of the [/fig_ref] , and to provide 3D structure visualization directly through the web browser [fig_ref] Figure 1: Overview of the [/fig_ref]. The server will automatically select suitable templates based on the consensus outputs of two profile search tools (e.g. PSI-BLAST and IMPALA). Alternatively, the user may specify a template structure. The automated modeling procedure begins when at least one modeling template is available. On the other hand, since comparative modeling procedures can have differential complexity, the (PS) 2 server provides for both interactive and batch modes [fig_ref] Figure 1: Overview of the [/fig_ref]. In the interative mode, users can select different templates [fig_ref] Figure 1: Overview of the [/fig_ref] and visualize the modeled results [fig_ref] Figure 1: Overview of the [/fig_ref] on the (PS) 2 website until a satisfactory model structure is obtained. In the batch mode, (PS) 2 will automatically send the modeled results to users by Email when the automated modeling procedure is complete. The modeling procedure is briefly described in the following subsections.
## Template search/selection
(PS) 2 uses the consensus of PSI-BLAST and IMPALA for the template search. PSI-BLAST and IMPALA are widely used for local sequence alignments with different profile search strategies. PSI-BLAST scans the profile of the query sequence against each of the template sequences in a database. In contrast, IMPALA searches the query sequence against each of the template profiles, which constitute a database of PSI-BLAST-generated position-specific score matrices (PSSMs). The template sequence library of (PS) 2 is extracted from the Protein Data Bank (PDB) [bib_ref] The RCSB Protein Data Bank: a redesigned query system and relational database..., Deshpande [/bib_ref]. Any given pair of sequences in the library has <95% sequence identity. Currently, each template profile in the IMPALA profile library, which included 12 011 sequences, was constructed using PSI-BLAST by searching against the nrdb90 database.
For each protein sequence, (PS) 2 collected 20 templates from both the top 10 templates of PSI-BLAST and IMPALA by searching the template sequence and template profile library, respectively. (PS) 2 utilized a sequence similarity score (S IR ) which is a good template classifier if the optimal sequence alignment could be found. The sequence similarity score is given as S IR ¼ (SI + AP)/2 where SI is the sequence identity and AP is the alignment percentage between the query protein and the template sequence. As both PSI-BLAST and IMPALA are local alignment tools, the AP is important for selecting a right template. The AP is defined as the number of aligned residues divided by the total number of residues of a query protein sequence. Hence, from among the 20 templates (PS) 2 automatically selects the one with the highest S IR , which is aligned by our consensus algorithm [fig_ref] Figure 2: Template selection and target-template alignment in [/fig_ref] using the resulting alignments of PSI-BLAST, IMPALA and T-Coffee (a multiple global sequence alignment tool).
## Target-template alignment
As previous studies [bib_ref] Automated prediction of CASP-5 structures using the robetta server, Chivian [/bib_ref] [bib_ref] EsyPred3D: prediction of proteins 3D structures, Lambert [/bib_ref] [bib_ref] An automatic homology modeling method consisting of database searches and simulated annealing, Ogata [/bib_ref] [bib_ref] Automatic consensus-based fold recognition using Pcons, ProQ, and Pmodeller, Wallner [/bib_ref] [bib_ref] PROTINFO: secondary and tertiary protein structure prediction, Hung [/bib_ref] [bib_ref] GenTHREADER: an efficient and reliable protein fold recognition method for genomic sequences, Jones [/bib_ref] [bib_ref] Protein threading using PROSPECT: design and evaluation, Xu [/bib_ref] indicate, the most persistent problem facing comparative modeling is probably the alignment of the query sequence with the template(s). With (PS) 2 , we attempted to improve comparative modeling by considering alternative and consensus alignments based on the alignments of PSI-BLAST, IMPALA and T-Coffee. To efficiently combine the results of these alignment methods, we designed a consensus sequence algorithm [fig_ref] Figure 2: Template selection and target-template alignment in [/fig_ref] by considering the collective alignments from these tools and then given the target-template aligned-result. We briefly describe these steps as follows: (i) initialize all entries of the consensus matrix to 0; (ii) sum up aligned scores of these three alignments for each position with different scoring weights (IMPALA is 2, PSI-BLAST is 4 and T-Coffee is 3); (iii) take the positions with the highest score as the aligned points to build the final target-template alignment; (iv) identify the unfeasible positions (e.g. an amino acid in the target sequence is aligned with two amino acids in the template); (v) reset the scores of unfeasible positions and the aligned points in the consensus matrix to 0; (vi) repeat steps (iii-v) until all entries are 0; and (vii) output the path with the aligned points as the target-template alignment.
## Model building and model evaluation
The final 3D protein structures were built from the consensus alignment using the homology-modeling package, MODELLER, which automatically calculated a model containing all non-hydrogen atoms using geometric restraints and molecular dynamic annealing. After MODELLER generated a predicted model with no other refinements, the program PROCHECK (16) was used to evaluate the quality of this model based on the G-factor. Finally, the predicted model was displayed by Raster3D [bib_ref] Raster3D: photorealistic molecular graphics, Merritt [/bib_ref] and automatically sent to users. The components of the (PS) 2 server shown in [fig_ref] Figure 1: Overview of the [/fig_ref] were built using PHP and Perl.
## Input format
(PS) 2 is an easy-to-use web server [fig_ref] Figure 1: Overview of the [/fig_ref]. Uses input the query protein sequence in FASTA format and choose a template selection method from 'Both', PSI-BLAST or IMPALA. The default option in template(s) selection is 'Both' which uses a consensus method combining PSI-BLAST and IMPALA. Users are also able to assign a specific PDB code as the template for the query sequence. Moreover, (PS) 2 provides both batch and interactive mode. In the batch mode, (PS) 2 automatically selects the template(s), while in the interative mode (PS) 2 allows the user to assign specific template(s) from a list of candidates [fig_ref] Figure 1: Overview of the [/fig_ref] Finally, the server sends the predicted results to the user's Email address.
## Output format
Typically, the (PS) 2 server yielded a predicted structure within 5 min if the sequence length is $200. The predicted results of the (PS) 2 server consists of the selected template(s), target-template alignment, predicted structure and structure evaluations [fig_ref] Figure 1: Overview of the [/fig_ref]. The server provides the selected template and a list candidates yielded by PSI-BLAST and IMPALA [fig_ref] Figure 1: Overview of the [/fig_ref]. The predicted structure is visualized in PNG format generated by MolScript [bib_ref] MOLSCRIPT:a program to produce both detailed and schematic plots of protein structures, Kraulis [/bib_ref] and Raster3D packages [bib_ref] Raster3D: photorealistic molecular graphics, Merritt [/bib_ref]. If the Chime is installed in a browser, the output will display the predicted structure in the browser [fig_ref] Figure 1: Overview of the [/fig_ref]. The server allows a user to download the predicted structure coordinates in the PDB format; furthermore, the target-template alignment in PIR format and the structure quality factors are also provided.
# Results
The global distance test_total score (GTD_TS) of Ca atoms was used to assess the correctness of the predicted model [bib_ref] LGA: a method for finding 3D similarities in protein structures, Zemla [/bib_ref]. GTD_TS has been commonly used in modeling studies and in the CASP community. GTD_TS is defined as
[formula] GDT_TS ¼ 100 P d GDT d =N 4 ð%Þ d 2 f1‚ 2‚ 4‚ 8g‚ [/formula]
where N is the total number residues of a target, GDT d is the number of aligned residues whose Ca-atom distance between the target and predicted model is less than d Å after superposition of the two structures; and d is 1, 2, 4 or 8 s.
(PS) 2 was evaluated for 47 comparative modeling targets in CASP6 (9) [fig_ref] Figure 3: Comparison [/fig_ref] and the mean GDT-TS score was 66.69. In order to test (PS) 2 on these targets, each profile in the IMPALA profile library, which comprises 9775 sequences obtained from PDB on June 30, 2004, was constructed using PSI-BLAST by searching against the nrpb90 database. This server utilized the similarity score (S IR ) for template selection to improve prediction accuracy. When (PS) 2 used the template with the lowest expected value (E-value) in the hit structures similar to ESyPred3D (3), the mean GDT_TS scores for PSI-BLAST and IMPALA are 57.99 and 62.29, respectively. These scores were improved to 62.19 and 62.72, respectively, when the S IR was applied in template selection. These experimental results show that S IR , combining both the sequence identity and the alignment percentage, is a useful strategy for template selection since a low E-value does always imply a high sequent identity for cases such as T0229, T0231 and T0264. For target T0264, the sequence identity is 11.83% and GDT_TS score is 37.53 when the protein with the lowest PSI-BLAST E-value ($10 À55 ) was used as the template (PDB code 1pjqA). In contrast, the sequence identity is 31.32%, the expected value is $10 À39 , and GDT_TS score is 64.97 when the protein with the highest S IR was selected as the template (PDB code 1vhvA).
(PS) 2 outperformed PSI-BLAST and IMPALA alignments based on mean GDT_TS scores. PSI-BLAST and IMPALA selected the same templates for 32 targets among 47 targets and PSI-BLAST identified 10 better templates than IMPALA for 10 targets. Conversely, IMPALA identified five better templates than PSI-BLAST for five other targets. The experimental results show that the consensus sequence algorithm [fig_ref] Figure 2: Template selection and target-template alignment in [/fig_ref] , combining both local and global multiple sequence alignment mechanisms, could indeed improve the performance. PSI-BLAST and IMPALA help to yield homologous protein sequences and local alignments by utilizing profile alignments, whereas T-Coffee expands local alignments to global alignments. For example, for T0205, the aligned percentages are 77.78% (PSI-BLAST), 79.80% (IMPALA), 100% (T-Coffee) and 100% (consensus method); moreover, the GDT_TS score are 66.94% (PSI-BLAST), 69.09% (IMPALA), 73.93% (T-Coffee) and 75.27% (consensus method).
Using these 47 targets, we compared the prediction accuracy of (PS) [bib_ref] Automated prediction of CASP-5 structures using the robetta server, Chivian [/bib_ref] [Pcons5 [bib_ref] Automatic consensus-based fold recognition using Pcons, ProQ, and Pmodeller, Wallner [/bib_ref] ] and 35.57 [Pcomb2 [bib_ref] Automatic consensus-based fold recognition using Pcons, ProQ, and Pmodeller, Wallner [/bib_ref] ]. For these targets, the mean GDT_TS score using (PS) 2 was superior to these of the 10 automatic servers; moreover, the individual GDT_TS scores from (PS) 2 were comparable. Using mean GDT_TS scores, (PS) 2 obtained 13 predicted structures in the first rank and 8 structures in the second place. These analysis results suggest that the accuracy of (PS) 2 is comparable with those of previous prediction servers.
# Example analysis
(PS) 2 predicted the structure of the F2365 glyoxalase protein (AAT03210) sequence in Listeria monocytogenes [bib_ref] Whole genome comparisons of serotype 4b and 1/2a strains of the food-borne..., Nelson [/bib_ref]. It selected the native structure of glyoxalase I (GlxI) (PDB code 1f9zA) from Escherichia coli (20) as the template. GlxI is the first of two enzymes in the pathway to converts cytotoxic a-keto aldehydes into nontoxic a-hydroxycarboxylic acids. This pathway is important in that an increase in methylglyoxal can produce toxic effects by reacting with DNA, RNA, and proteins. Therefore, GlxI has been utilized in the design of anticancer and antimalarial agents [bib_ref] Mechanism-based competitive inhibitors of glyoxalase i: intracellular delivery, in vitro antitumor activities,..., Kavarana [/bib_ref].
The template shares 23.7% sequence identity with the query sequence and the target-template alignment is shown in . (PS) 2 automatically aligned four important residues together (His5, Glu53, His73 and Glu123 in the query sequence; His5, Glu56, His74 and Glu122 in the template sequence), which are responsible for the binding metal activity of the GlxI family (red blocks in . The superimposing result of the predicted structure (green) and the template structure (blue) also shows that the coordinates of side chains and backbones of these four residues are similar.
This protein sequence (AAT03210) was also submitted to SWISS-MODEL [bib_ref] SWISS-MODEL: an automated protein homology-modeling server, Schwede [/bib_ref] , which is a widely used homologymodeling server and ESyPred3D. SWISS-MODEL is unable to find a suitable template since no sequences above 25% sequence identity are found. On the other hand, ESyPred3D selected a native structure of GlxI [PDB code 1qipD (23)] from Homo sapiens as template. The template shares 17.4% identity with this query sequence. Human GlxI is active in the presence of Zn 2+ ; but E.coli GlxI is inactive in the presence of Zn 2+ and is maximally active with Ni 2+ , as L.monocytogenes GlxI (AAT03210) does. [bib_ref] Whole genome comparisons of serotype 4b and 1/2a strains of the food-borne..., Nelson [/bib_ref]. These analysis results show that the query sequence is more correlated to E.coli GlxI than Human GlxI.
# Conclusion
The key novelty of (PS) 2 is the seamless ability of blending local and global multiple sequence alignment mechanisms to allow them to work cooperatively by a new similarity score (S IR ). The analysis using (PS) 2 was significantly faster because (PS) 2 uses an effective consensus strategy that combines three publicly available tools installed on the same machine; moreover, (PS) 2 is based solely on the consensus sequence and thus is considerably faster than other methods that rely on the additional structural consensus of templates. We believe that (PS) 2 is a fast homology-modeling server and should be useful in structure prediction and modeling.
[fig] Figure 1: Overview of the (PS) 2 using the protein sequence of F2365 glyoxalase protein (AAT03210) in L.monocytogenes as query. (a) Main procedure; (b) The assignments of template selection method, target sequence, and interaction/batch; (c) Template selection (single/multiple templates) using the interaction module based on 20 candidates provided by PSI-BLAST and IMPALA; (d) Evaluation and visualization of the predicted structure. [/fig]
[fig] Figure 2: Template selection and target-template alignment in (PS) 2 . (a) The consensus algorithm with T-Coffee, PSI-BLAST and IMPALA. (b) Example and unfeasible solutions. [/fig]
[fig] Figure 3: Comparison (PS) 2 (black) with 10 automatic servers of the prediction accuracies (GDT_TS scores) on 47 targets in CASP6. The results of these 10 automatic servers are summarized from http://predictioncenter.genomecenter.ucdavis.edu/casp6/Casp6.html. [/fig]
|
Intermittent Hypoxia in Childhood: The Harmful Consequences Versus Potential Benefits of Therapeutic Uses
Intermittent hypoxia (IH) often occurs in early infancy in both preterm and term infants and especially at 36-44 weeks postmenstrual age. These episodes of IH could result from sleep-disordered breathing or may be temporally unrelated to apnea or bradycardia events. There are numerous reports indicating adverse effects of IH on development, behavior, academic achievement, and cognition in children with sleep apnea syndrome. It remains uncertain about the exact causative relationship between the neurocognitive and behavioral morbidities and IH and/or its associated sleep fragmentation. On the other hand, well-controlled and moderate IH conditioning/training has been used in sick children for treating their various forms of bronchial asthma, allergic dermatoses, autoimmune thyroiditis, cerebral palsy, and obesity. This review article provides an updated and impartial analysis on the currently available evidence in supporting either side of the seemingly contradictory scenarios. We wish to stimulate a comprehensive understanding of such a complex physiological phenomenon as intermittent hypoxia, which may be accompanied by other confounding factors (e.g., hypercapnia, polycythemia), in order to prevent or reduce its harmful consequences, while maximizing its potential utility as an effective therapeutic tool in pediatric patients.
# Introduction
Episodes of intermittent hypoxia (IH) are among typical consequences of immature respiratory control. Particularly, the incidence of IH in the infants with low birth weight would increase progressively over the first 4 weeks of postnatal life and reach a subsequent plateau followed by a slow decline beginning at sixth to eighth weeks. Repetitive cycles of hypoxia/reoxygenation often lead to a pro-inflammatory cascade with resultant multisystem morbidity, including retinopathy of prematurity and impaired growth, as well as cardio-respiratory instability and neurodevelopmental defects [bib_ref] Physiologic basis for intermittent hypoxic episodes in preterm infants, Martin [/bib_ref]. Similarly, 15 years ago, Gozal and colleagues had described that IH is the most frequent form of hypoxia occurring in the developing mammal [bib_ref] Respiratory plasticity following intermittent hypoxia: developmental interactions, Gozal [/bib_ref] , because the maturational process of neural, mechanical, pulmonary, and sleep state-dependent factors would all favor the occurrence of IH during early postnatal life. It has been increasingly recognized that hypoxia, even when short lasting, can modify subsequent respiratory responses to hypoxia and induce a variety of genes whose consequences will persist for much longer periods than the duration of the hypoxic stimulus itself, i.e., functional and adaptive plasticities. The dynamic interactions between the severity, overall duration, and repetitive frequency of IH and the level of maturity of the organs and systems at the time of IH will modify the ventilatory, metabolic, and cardiovascular responses to hypoxia and the adaptive (beneficial) or maladaptive (detrimental) consequences after exposure to IH.
In this review article, we primarily focus on providing an impartial overview on the currently available evidence regarding the functional and structural impact of IH in healthy children or pediatric patients. In addition, the interested readers may find many other important aspects of IH research in adult individuals in two monographs that we recently edited.
## Intermittent hypoxia and cardiovascular and metabolic functions obstructive sleep apnea: pathological consequences in pediatric patients
One of the most common examples of the negative health impact of IH is obstructive sleep apnea (OSA), which is characterized by brief, recurrent cycles of hypoxia-reoxygenation, typically <60 s in duration. Such repetitive cycles activate sympathetic nervous system and systemic inflammation. When OSA becomes a chronic condition, it often results in adverse physiological effects such as abnormal gas exchange and/or alteration of sleep patterns that impact on health and development [bib_ref] Neurotransmitters and neuromodulators during early human development, Herlenius [/bib_ref]. In fact, pediatric OSA is not only a very frequent condition affecting 2-4% of all children but also is associated with an increased risk for a variety of endorgan injury and dysfunction (such as accelerated atherosclerosis and endothelial dysfunction) that impose both immediate and potentially long-term morbidities and resultant high healthcare costs [bib_ref] Physiological effects of obstructive sleep apnea syndrome in childhood, Muzumdar [/bib_ref].
There are a number of reports indicating that alterations in autonomic nervous system function occur in children with OSA, including either an increase in sympathetic nervous system tone and/or responsiveness, or the emergence of sympathetic-parasympathetic imbalance [bib_ref] Cardiovascular complications of obstructive sleep apnea syndrome: evidence from children, Bhattacharjee [/bib_ref] [bib_ref] Obesity and obstructive sleep apnea syndrome in children: a tale of inflammatory..., Bhattacharjee [/bib_ref]. A recent prospective cross-sectional study of 26 children with polysomnographyconfirmed OSA (versus 30 children in control group) demonstrated that OSA in children was associated with increased lipid peroxidation in an OSA severity-dependent manner, indicated by a positive correlation between plasma oxidized low-density lipoproteins and the apnea/hypopnea index [bib_ref] Oxidative stress in children with obstructive sleep apnea syndrome, Tauman [/bib_ref].
Interestingly, a recent review by Gozal's group suggested that IH may not be the primary cause of pathological disorders in pediatric patients with OSA [bib_ref] Chemoreceptors, baroreceptors, and autonomic deregulation in children with obstructive sleep apnea, Gozal [/bib_ref]. In addition to the component of IH, OSA is also accompanied by hypercapnia, which may be a main contributor to the development of pathological process. The combination of IH, sleep fragmentation, episodic hypercapnia, and increased intrathoracic pressure swings can separately and together activate or amplify the onset and propagation of endothelial dysfunction, atherogenesis, increased systemic inflammation, oxidative stress, and activation of adhesion molecules and coagulation [bib_ref] CrossTalk proposal: the intermittent hypoxia attending severe obstructive sleep apnoea does lead..., Gozal [/bib_ref] [bib_ref] Circulating microparticles in children with sleep disordered breathing, Kim [/bib_ref]. Tam et al. used a piglet model of infant OSA to evaluate circulating IL-6, TNF-alpha, and C-reactive protein following exposure to acute hypercapnic IH [bib_ref] The effect of acute intermittent hypercapnic hypoxia treatment on IL-6, TNF-alpha, and..., Tam [/bib_ref]. The IH protocol consisted of two 90-min sessions of hypercapnic IH with arterial blood sampled before and after each session. The authors concluded that acute hypercapnic IH caused a transient increase in proinflammatory cytokine -IL-6 levels, which may have implications for the pro-inflammatory status in pediatric OSA.
Substantial evidence also suggested a major role for IH in altering autonomic nervous system control [bib_ref] Chemoreceptors, baroreceptors, and autonomic deregulation in children with obstructive sleep apnea, Gozal [/bib_ref]. Both IH and CO 2 retention may augment sympathetic nerve activity (SNA) via stimulation of both central and peripheral chemoreceptors in adults [bib_ref] Contrasting effects of hypoxia and hypercapnia on ventilation and sympathetic activity in..., Somers [/bib_ref]. SNA is higher in infants and young children, and progressively declines through age 5-7 years, followed by a stable period until the beginning of puberty that is associated with increases in SNA [bib_ref] Autonomic nervous system activity in premature and full-term infants from theoretical term..., De [/bib_ref] [bib_ref] Heart rate variability in infants, children and young adults, Finley [/bib_ref] [bib_ref] Pubertal and gender-related changes in the sympathoadrenal system in healthy children, Weise [/bib_ref]. Until today, little information is available concerning the effect of IH on SNA in children. In addition, severe retinopathy of prematurity was also associated with more variable, longer, and less severe IH events in preterm infants [bib_ref] The relationship between patterns of intermittent hypoxia and retinopathy of prematurity in..., Fiore [/bib_ref]. The increased SNA may be driven by two major and potentially interactive pathways as previously proposed, i.e., (1) peripheral chemoreceptor-and baroreceptor-dependent pathway [bib_ref] Cardiovascular and ventilatory acclimatization induced by chronic intermittent hypoxia: a role for..., Iturriaga [/bib_ref] [bib_ref] Cardioventilatory acclimatization induced by chronic intermittent hypoxia, Iturriaga [/bib_ref] [bib_ref] Mechanisms of sympathetic activation and blood pressure elevation by intermittent hypoxia, Prabhakar [/bib_ref] ; and (2) interactions between peripheral and central nervous system (CNS)-located pathways [bib_ref] Invited review: physiological consequences of intermittent hypoxia: systemic blood pressure, Fletcher [/bib_ref] [bib_ref] Chronic-intermittent hypoxia: a model of sympathetic activation in the rat, Sica [/bib_ref]. In addition, Zhao et al. suggested that IH differentially regulates plasma membrane Na + channels in the developing brain, depending on duration of IH [bib_ref] Intermittent hypoxia modulates Na + channel expression in developing mouse brain, Zhao [/bib_ref].
## Intermittent hypoxia and cardiovascular and metabolic responses in children
As IH episodes are common among preterm infants, the early postnatal chronic IH exposure may lead to long-term alterations in cardio-respiratory control, such as reduction in baroreflex sensitivity. shown that the exposure of rats to chronic IH for the first 30 days of life leads to the substantial structural changes within both nucleus tractus solitarii and ventrolateral medulla [bib_ref] Anatomical changes in selected cardio-respiratory brainstem nuclei following early post-natal chronic intermittent..., Reeves [/bib_ref]. A more recent study by tested the hypothesis that a clustered versus dispersed pattern of repetitive IH during early postnatal life would induce differential long-term alteration in growth and cardiovascular regulation in rat pups from 1 to 7 days of life [bib_ref] Effect of postnatal intermittent hypoxia on growth and cardiovascular regulation of rat..., Pozo [/bib_ref]. They found that exposure to both patterns of repetitive IH-induced early growth restriction and exhibited a sustained decrease in heart rate. By contrast, only the clustered paradigm resulted in a significantly lower BP versus controls, while dispersed IH protocol had no effect on blood pressure. Apparently, the repetitive IH during a critical developmental window with either clustered or dispersed IH exposure paradigm contributed to prolonged changes in sympathovagal balance of cardiovascular regulation [bib_ref] Effect of postnatal intermittent hypoxia on growth and cardiovascular regulation of rat..., Pozo [/bib_ref]. Pediatric OSA is associated with cardiovascular consequences, including accelerated atherosclerosis and endothelial dysfunction in blood vessels [bib_ref] Physiological effects of obstructive sleep apnea syndrome in childhood, Muzumdar [/bib_ref].
Severe and chronic IH as occurs during a number of disease states can induce a series of cellular and molecular responses that result in cell injury and death. In response to IH, a number of signaling pathways are involved in oxygen sensing, oxidative stress, metabolism, catecholamine biosynthesis, and immune responsiveness. The cumulative effect of these processes over time can undermine cell integrity and functionality [bib_ref] Can O2 dysregulation induce premature aging?, Douglas [/bib_ref]. In adult individuals, IH-induced oxidative stress may increase predisposition for metabolic dysfunction by impairing insulin sensitivity and glucose tolerance [bib_ref] Effects of acute intermittent hypoxia on glucose metabolism in awake healthy volunteers, Louis [/bib_ref]. IH increases serum and liver fatty acid levels due to an increase in sterol regulatory element binding protein-1 (SREBP-1), a transcription factor of lipid synthesis. Endothelin-1 is also an important factor in insulin resistance during IH.
## Intermittent hypoxia and cognitive functions
Hypoxic brain damage is one of the most common perinatal injuries of the CNS. Different degrees of susceptibility of each child to damaging factors and stressors would lead to quite variable outcomes in terms of response to IH and its impact on CNS. Several recent studies indicated perinatal hypoxia as a risk factor for psychiatric disorders like schizophrenia. It is thought that hypoxia prior to or during birth may contribute to alterations leading to the protracted clinical manifestation during young adulthood. A review on various evidences suggested an adverse effect of chronic IH on cognition in childhood, more specifically on development, behavior, and academic achievement [bib_ref] The effect of chronic or intermittent hypoxia on cognition in childhood: a..., Bass [/bib_ref].
On the other hand, Lima-Ojeda et al. exposed mice during postnatal day 3-7 to two paradigms of chronic intermittent or continuous hypoxia (10% ambient O 2 ) (32). They found that neither intermittent nor continuous perinatal hypoxia-induced longterm behavioral alterations. It seems that more severe hypoxic conditions and/or the presence of additional factors (such as genetic risk factors) are necessary for generating long-term behavioral abnormalities. A more recent study by Guo et al. revealed the impact of chronic asthma-induced hypoxia on cognitive function in children using an ovalbumin-induced chronic asthma model in immature mice [bib_ref] Chronic asthma results in cognitive dysfunction in immature mice, Guo [/bib_ref]. They reported that chronic asthmatic hypoxia impaired learning and memory ability.
On the other hand, Urschitz and co-workers assessed the association of snoring and IH with poor academic performance in 1,144 of third grade school children [bib_ref] Snoring, intermittent hypoxia and academic performance in primary school children, Urschitz [/bib_ref] [bib_ref] Habitual snoring, intermittent hypoxia, and impaired behavior in primary school children, Urschitz [/bib_ref]. A significant relationship between snoring and poor academic performance was found in children without IH, whereas IH did not show an independent association with poor academic performance.
Cai et al. have shown in developing rats that chronic IH exposures across 2 and 4 weeks led to more reference, working, and total memory errors in the 8-Arm radial maze task, and other negative consequences [bib_ref] Endoplasmic reticulum stress plays critical role in brain damage after chronic intermittent..., Cai [/bib_ref]. Endoplasmic reticulum stress-related enhancement of neuronal apoptosis was implicated as one of the underlying mechanisms of cognitive dysfunction induced by chronic IH. In pediatric patients, Kirkham and Datta prospectively recorded overnight oxyhemoglobin saturation in 18 children with intractable epilepsy, 6 of whom were currently or recently in minor status epilepticus (i.e., a seizure goes on for 30 min or more) [bib_ref] Hypoxic adaptation during development: relation to pattern of neurological presentation and cognitive..., Kirkham [/bib_ref]. Children with minor status were more likely to have an abnormal sleep study often with desaturation of blood oxygen.
## Effects of chronic intermittent hypoxia on development in neonatal mammals
Despite the fact that chronic IH seems to exert much less impact on the physical growth of animals and humans than chronic constant hypoxia, some evidence suggested that severe IH can also adversely affect the function and development of the organism. In 2008, Farahani et al. reported differential effects of chronic IH and chronic constant hypoxia on postnatal growth and development in mice [bib_ref] Differential effects of chronic intermittent and chronic constant hypoxia on postnatal growth..., Farahani [/bib_ref]. The postnatal day 2 mice were exposed to constant hypoxia (11% O 2 in isobaric chamber constantly) or IH (cycles of 4 min 11% O 2 with 4 min intervals) for 4 weeks. They found that the most severe developmental delay observed under constant hypoxia. Slower weight gain resulted in a 12 and 23% lower body weight in the mice exposed to chronic IH and constant hypoxia, respectively, by postnatal day 30. The decrease in liver, kidney, and brain weight were greater in the constant hypoxia group than IH group. By contrast, the heart weight from chronic constant hypoxia and IH groups was 13 and 33% greater than control (P < 0.05), respectively, which was associated with increased size of cardiomyocytes by 12 and 14% (P < 0.001) for constant hypoxia and IH mice [bib_ref] Differential effects of chronic intermittent and chronic constant hypoxia on postnatal growth..., Farahani [/bib_ref]. Similarly, Row et al. (39) exposed rat pups to either room air or IH beginning at postnatal day 10 until day 30 [bib_ref] Impaired spatial learning and hyperactivity in developing rats exposed to intermittent hypoxia, Row [/bib_ref]. The pups exposed to IH displayed significant spatial learning impairments and increased locomotor activity, which indicated that exposure to IH at postnatal age could induce substantial learning impairment and gender-dependent behavioral hyperactivity in the juvenile rats [bib_ref] Impaired spatial learning and hyperactivity in developing rats exposed to intermittent hypoxia, Row [/bib_ref].
Human fetus develops in a profoundly hypoxic environment, which exerts a distant effect on human tolerance to hypoxia that promotes survival advantage under severe hypoxemic stress. The phenomenon and potential mechanisms of neonatal hypoxia tolerance are the subjects for extensive investigations [bib_ref] Concepts in hypoxia reborn, Martin [/bib_ref] [bib_ref] Neonatal tolerance to hypoxia: a comparative-physiological approach, Singer [/bib_ref]. Among the main theories, using a new form of body plethysmograph Cross et al. measured respiratory volume and rate of the newborn infants and first described a positive relationship between body weight and respiratory volume [bib_ref] The respiratory rate and ventilation in the newborn baby, Cross [/bib_ref]. The immaturity of respiratory control in early ontogenesis leads to disruption of metabolic processes in children under hypoxia, which triggers a more ancient way of energy production -anaerobic glycolysis as an adaptive response for surviving in hypoxic conditions [bib_ref] Evolution of human hypoxia tolerance physiology, Hochachka [/bib_ref]. In addition, as early as 1921, Benedict and Talbot reported that the daily metabolism adjusted for body surface increases from birth to 1 year of age by 1.6 times. Other researchers have subsequently found that the coincidence in time between the attainment of maximum intensity of respiration and the formation of thermoregulatory mechanism [bib_ref] Ontogeny of physiological regulations in the rat, Adolph [/bib_ref]. During hypoxia, many newborn mammals, including the human infant, decrease metabolic rate, therefore adopting a strategy common to many living creatures, but usually not adopted by adult humans [bib_ref] Implications of hypoxic hypometabolism during mammalian ontogenesis, Mortola [/bib_ref]. In 1996, Hochachka and colleagues proposed a biochemical basis for the response of hypoxia-tolerant systems to hypoxia that includes defense and rescue phases [bib_ref] Unifying theory of hypoxia tolerance: molecular/metabolic defense and rescue mechanisms for surviving..., Hochachka [/bib_ref]. The first lines of defense against hypoxia include a balanced suppression of both ATP-demand and ATP-supply pathways, leading to a new steady state even though ATP turnover rates are greatly declined. The ATP demands of ion pumping and protein synthesis are downregulated by channel and translational arrests. In hypoxia-tolerant systems, these arrests activate the gene-based metabolic reprograming "rescue" mechanisms under hypoxia [bib_ref] Unifying theory of hypoxia tolerance: molecular/metabolic defense and rescue mechanisms for surviving..., Hochachka [/bib_ref].
The development of respiratory lung function in ontogeny is uneven and heterochronic (i.e., a developmental change in the timing of events, leading to changes in size and shape, depending on morphological rearrangement of the lungs and chest and improvement of regulatory mechanisms. The most important stages in the development of respiratory lung function are: neonatal period, up to 1 year, 2-4, 6-7, and 10-11 years. Differentiation of lung tissue is completed by 8-12 years and the growth of tracheobronchial tree ends with the termination of body growth [bib_ref] The mechanisms of onto-and gerontogenesis, Arshavskii [/bib_ref]. Because of continuous decreasing of specific entropy production, the relative values of lung ventilation and gas exchange (adjusted with body weight or surface) decrease with age [bib_ref] Phenomenological theory of ontogenesis, Zotin [/bib_ref]. One time period of exception is from birth to 1 year, when gradual decrease is not observed, on the contrary, the specific rate of gas exchange and lung ventilation increases [bib_ref] Age and human respiratory responses to hyperoxic and hypoxic gas mixtures, Serebrovskaia [/bib_ref].
Changes in respiratory reactivity concur with cardiovascular reactivity in young children. In both longitudinal and cross-sectional studies, most investigators found developmental changes with increasing age in heart rate and respiratory sinus arrhythmia -a cardiac index of activation in the parasympathetic branch of the autonomic nervous system [bib_ref] Developmental and contextual influences on autonomic reactivity in young children, Alkon [/bib_ref] [bib_ref] Developmental changes in heart period and high-frequency heart period variability from 4..., Bar-Haim [/bib_ref]. Overall, resting HR decreases from infancy to young adulthood and respiratory sinus arrhythmia increases. Autonomic measures at 4 and 8 months of age showed that older infants had more sympathetic activation and less parasympathetic withdrawal in response to stressors than younger infants. On the other hand, a cross-sectional study of older children ages 8-10 and 15-17 years showed no age differences in sympathetic and parasympathetic reactivity [bib_ref] Patterns of sympathetic and parasympathetic reactivity in a sample of children and..., Salomon [/bib_ref]. There is apparently a slowing of age-related changes in autonomic reactivity as children move developmentally closer to adolescence and adulthood.
Measurement of the ventilatory responses to inhaled hypoxia in children is usually quite difficult due to its methodological and ethical problems. Very few reports are available. It has been known that the carotid chemoreceptors of sino-and cardio-aortic areas are formed in humans at sixth week of fetal life and begin to function before birth. During the first hours of neonatal life, infants are able to increase their ventilation when blood oxygen tension drops. At different stages of ontogenesis, the increase of ventilation in response to hypoxia is provided by unequal changes in breathing modes [bib_ref] Respiratory control in children and adolescents, Kolchinskaia [/bib_ref]. In contrast to the adult, hypoxic ventilatory response (HVR) in newborns is short and unstable [bib_ref] Chemical control of respiration in newborn infants, Graham [/bib_ref]. An early investigation from our group in 1977 revealed that inhalation of gas mixture with 14.5% O 2 for 12 min resulted in an increase in lung ventilation (V E ) during the first minute by 18% in adults and 24% in children of 4-5 years, indicating a stronger initial response in children than those in adults [bib_ref] Age and human respiratory responses to hyperoxic and hypoxic gas mixtures, Serebrovskaia [/bib_ref]. Whereas adults demonstrated a classical biphasic HVR pattern with a subsequent maintain of ventilation at elevated level, the primary HVR in children was short-term and followed by a subsequent reduction of ventilation below baseline and a concomitant decrease in gas exchange, indicating a less pronounced ability to maintain homeostasis of ventilation in young children.
Effects of IH on HVR in developing mammals were well described by Gozal's group [bib_ref] Respiratory plasticity following intermittent hypoxia: developmental interactions, Gozal [/bib_ref]. They suggested that despite substantial differences in the acute HVR between adult and immature mammals, IH-induced modifications of HVR are qualitatively similar between immature and adult. Short durations of IH exposures elicit increases in the magnitude of HVR and an attenuated HVR occurs over time if IH exposures are prolonged. However, the neural structures, neurotransmitters, and downstream signaling pathways underlying such biphasic changes in HVR induced by IH are unknown. Using an immature rat model simulating OSA, Moss et al. investigated effects of chronic IH (12% O 2 , 7 h daily) on rats from postnatal day 17 (representing early childhood) through day 33 (representing adolescence) and day 47 (adult) [bib_ref] Long-term recurrent hypoxia in developing rat attenuates respiratory responses to subsequent acute..., Moss [/bib_ref]. They reported that chronic IH produced long-lasting attenuation in respiratory responsiveness to subsequent acute hypoxia.
It is noteworthy that when assigning hypoxic stress on child and adolescent, it is necessary to first determine their individual HVR for better safety. Such rare cases in HVR were seen in our previous studies. For example, among 20 pairs of monozygotic twins of 10-15 years old whom we tested, there was one male twin who demonstrated complete absence of HVR [bib_ref] Evaluation of the degree of genetically determined reactions of the human cardiorespiratory..., Serebrovskaia [/bib_ref] [bib_ref] Hereditary defect of sensitivity to hypoxia in normal sensitivity to hypercapnia, Serebrovskaia [/bib_ref] and the father of the twins had similar defect of sensitivity to hypoxia.
Another early study of our group was conducted on respiration and gas exchange in pediatric patients with chronic pneumonia [bib_ref] Effect of an increased and decreased oxygen content in the air on..., Serebrovskaia [/bib_ref]. A significant increase in physiological dead space and reduction of ratio between alveolar and lung ventilation (V A /V E ) were observed. In the older children (10-14 years), these pathological changes were well compensated by significant hyperventilation that resulted in maintenance of PaO 2 and gas exchange at the levels of healthy children. However, in the younger children (6-8 years), the compensatory increase in ventilation was not observed, whereas in 4-5 years old patients with protracted pneumonia of stage II, hypoventilation was noted and PaO 2 was reduced with significantly lower gas exchange rate than age norms. The development of these sick children was also retarded [bib_ref] Effect of an increased and decreased oxygen content in the air on..., Serebrovskaia [/bib_ref]. A more recent work showed that HVR during exercise differs between children and adults [bib_ref] Oxygen uptake and heart rate responses during hypoxic exercise in children and..., Springer [/bib_ref]. However, when corrected for body weight, children and adults have similar values for lactic acidosis threshold and maximal oxygen consumption (VO 2 max) during normoxia. Hypoxia significantly lowered lactic acidosis threshold and VO 2 max in both children and adults. Metabolic efficiency was similar among the two age groups and unaffected by hypoxia.
Intermittent hypoxia during development has also been implicated as a potent inducer of respiratory plasticity [bib_ref] Developmental plasticity of respiratory control following intermittent hypoxia, Reeves [/bib_ref]. The altered ventilatory pattern induced by IH is distinct from other stimuli and elicits markedly different responses in the developing mammal as compared to the adult. Exposures to either hypoxia or hyperoxia during early postnatal life may lead to significant modifications of neural function during adulthood. For example, suppression of peripheral arterial chemoreceptor activity via exposure to hyperoxia (60% inspired oxygen) during the first month of rat life led to significant reduction in the number of unmyelinated axons in the carotid sinus nerve and petrosal ganglion in the adult rat that are accompanied by substantial attenuation of the HVR at 3-5 months of age but not at 15 months. By contrast, when adult rats are exposed to hyperoxia, no changes in HVR characteristics occur, indicating that the persistent plasticity changes in the pathways underlying HVR are unique to the interaction between environmental stimulus and a critical developmental window [bib_ref] Developmental plasticity of the hypoxic ventilatory response, Ling [/bib_ref].
Furthermore, Paton et al. reported in 1989 that both hypercapnic and hypoxic responses during wakefulness were defective in pediatric patients, 6-11 years of age, with congenital central hypoventilation syndrome (CCHS). These sick children had no subjective sensation of dyspnea or discomfort, but no significant change was found from the baseline levels of ventilation in response to either stimulus. The researchers speculated a defect caused by CCHS in central integration of the central and peripheral chemoreceptor signals [bib_ref] Hypoxic and hypercapnic ventilatory responses in awake children with congenital central hypoventilation..., Paton [/bib_ref]. Thus, long-lasting plasticity of neural networks underlying respiratory control is more likely to occur during early and more plastic stages of development. Application of IH during these critical stages of development likely regulates adaptive processes that could be used to our advantage for preventive or therapeutic purposes as we further elaborate in the following sections.
## Therapeutic uses of intermittent hypoxia in pediatric practice
Hypoxic training or conditioning can be traced back from the traditional medical remedy used in ancient time. For example, in the Carpathian mountain region, children who suffered from asthmatic bronchitis were ranged on foot, during 7 days successively, on a high mountain with ingestion of high-altitude herbal tea. The children had recovered. Similarly, a common yogic treatment of various diseases in India, so called "nisshesha rechaka pranayama, " involves breath holding at residual volume, which produces brief IH that triggers the adaptive mechanisms [bib_ref] Nisshesha rechaka pranayama offers benefits through brief intermittent hypoxia, Malshe [/bib_ref]. Nevertheless, the scientific basis for the observed beneficial effects of the so-called intermittent hypoxic training/therapy (IHT) on human organism remains elusive despite the extensive investigations over the past five decades. In the following sections, our discussions focus mainly on the abundant experience and evidence of IHT implementation in pediatric practice, which have been almost exclusively reported by Ukrainian and Russian physicians and researchers since 1970s [bib_ref] Effect of an increased and decreased oxygen content in the air on..., Serebrovskaia [/bib_ref] [bib_ref] Effects of hypoxic stimulation in experimental animals and in children with bronchial..., Anokhin [/bib_ref] [bib_ref] The use of hypoxic stimulation in chronic and relapsing bronchitis in young..., Fesenko [/bib_ref] [bib_ref] The effect of adaptation to the periodic action of hypoxia on the..., Meerson [/bib_ref] [bib_ref] A method for intermittent hypoxic exposures in the combined treatment of bronchial..., Serebrovskaia [/bib_ref]. Based on these studies, various highly specialized IHT equipment and portable devices such as "Hypoxicator" [bib_ref] Hypoxicators: review of the operating principles and constructions, Lopata [/bib_ref] have been invented and developed, including those for pediatric patients, although there are still a number of remaining issues related to the complexity and nonstandardization of various IHT protocols reported in different studies. In addition, further thorough risk-versus-benefit evaluations of IHT will address some fundamental ethical concerns on the IHT practice in pediatric patients.
## Bronchial asthma
Most abundant information for the therapeutic use of IHT has been found in the treatment of bronchial asthma (BA). For example, a study by Anokhin et al. [bib_ref] Effects of hypoxic stimulation in experimental animals and in children with bronchial..., Anokhin [/bib_ref] applied IHT with a normobaric hypoxic stimulation with four sessions of 5 min 12-15% O 2 , followed by 5 min normoxic interval, for 10 days in 200 children aged 4-14 years who suffered from asthma [bib_ref] Effects of hypoxic stimulation in experimental animals and in children with bronchial..., Anokhin [/bib_ref]. Positive effects were seen in 85% of subjects in the IHT group and only in 25% of the sham control group. In children with mild BA, a complete discontinuance of asthma attacks was observed and a significant improvement was also observed in patients with moderate to severe forms of BA without medication [bib_ref] Effects of hypoxic stimulation in experimental animals and in children with bronchial..., Anokhin [/bib_ref]. The beneficial effects lasted for an average of 4 months after IHT. By contrast, in patients with the severe form of BA, only small or no improvement was found. In the hormone-dependent form of BA, efficacy of IHT was also unsatisfactory [bib_ref] Use of therapeutic normobaric hypoxia for preventing exacerbation of chronic nonspecific salpingooophoritis, Karash [/bib_ref].
Such therapeutic effects of IHT were subsequently confirmed by several other research groups in Ukraine and Russia [bib_ref] Effects of hypoxic stimulation in experimental animals and in children with bronchial..., Anokhin [/bib_ref] [bib_ref] A method for intermittent hypoxic exposures in the combined treatment of bronchial..., Serebrovskaia [/bib_ref] [bib_ref] Efficiency of intermittent and resonance intermittent normobaric hypoxia therapy in patients with..., Chizhov [/bib_ref] [bib_ref] Effectiveness of intermittent normobaric hypoxia in patients with bronchial asthma in various..., Ragozin [/bib_ref]. Among these works, in 1990, Meerson and co-workers studied the effects of adaptation to IH on immune status and neurohumoral regulation in pediatric and adult patients with BA, allergic dermatoses, and autoimmune thyroiditis [bib_ref] The effect of adaptation to the periodic action of hypoxia on the..., Meerson [/bib_ref]. They reported that the IH adaptation facilitated normalization of humoral values of immunity in allergic and autoimmune disorders and led to increased serum levels of immunoglobulins, while the level of circulating immune complexes reduced. These beneficial changes of the immune system were associated with an increase of the reserve potency of the hypothalamo-hypophyseal-adrenal and sympathoadrenal systems as well as a reduction in blood histamine levels.
In recent years, there was a revitalized interest in the use of IHT in children with BA. For example, researchers from Brazil studied 48 adolescents (12-14 years of age) under three conditions: mild intermittent asthma; mild persistent asthma; and control [bib_ref] Analysis of physiological variables during acute hypoxia and maximal stress test in..., Maldonado [/bib_ref]. They concluded that adolescents with mild persistent asthma have a greater capacity to adapt to hypoxia than do those with other types of asthma. In addition, Serebrovskaya et al. used IHT for treatment of children (aged 9-13 years) with persistent atopic BA in moderate form without the signs of respiratory insufficiency. The subjects in experimental group underwent IHT alone with regular drug treatment and the control group received the same medical treatment, but not IHT [bib_ref] Intermittent hypoxia in treatment of bronchial asthma in childhood, Serebrovskaya [/bib_ref]. Before and next day after the 2-week session of IHT, individual cardio-respiratory reactions to hypoxia were investigated where normobaric hypoxia was administered with a portable device "Hypoxotron-Complex, " a modified closed spirometer with CO 2 absorption (68). The initial inspired gas was atmospheric O 2 (20.9%) and inspired O 2 fell to 12% after 60-90 s of rebreathing, and then O 2 was added gradually to the device to maintain inspired O 2 at 12% for the remaining 3.5-4 min with a final arterial O 2 saturation (SaO 2 ) typically 89-92%. All children easily tolerated the hypoxia periods without any untoward effects. Each IHT session consisted of four 5-7 min hypoxic periods, followed by 5 min interval with room air inspiration. A significant decline in breath shortness and feelings of chest congestion were noted in the patients of IHT group, with other symptoms such as cough and attacks of asphyxia diminished or disappeared. No significant changes in airway conductance before and after IHT. By contrast, significant differences in hypoxic ventilatory sensitivity were found as a result of IHT, suggesting that adaptation to IH caused considerable augmentation in ventilatory response to hypoxia, likely due to both central and peripheral mechanisms as previously proposed [bib_ref] Enhanced chemosensitivity after intermittent hypoxic exposure does not affect exercise ventilation at..., Katayama [/bib_ref] [bib_ref] Intermittent hypoxia research in the former soviet union and the commonwealth of..., Serebrovskaya [/bib_ref]. Meanwhile, the heart rate response to hypoxia became less pronounced and SaO 2 fell less at 12% O 2 , indicating IHT improved efficiency of cardiovascular system in supporting oxygen supply during hypoxia [bib_ref] Intermittent hypoxia in treatment of bronchial asthma in childhood, Serebrovskaya [/bib_ref].
Since the development of inflammatory process in lungs is usually associated with activation of free radical oxidation and decline of antioxidant enzymes activity [bib_ref] Oxidative stress and antioxidant defenses in asthmatic murine model exposed to printer..., Konga [/bib_ref] , the role of antioxidant enzymes in adaptation to IHT has been a key subject for investigation in pediatric BA patients. Nesvitalova et al. examined the effect of a 10-day IHT on the mRNA expression and protein content of antioxidant enzymes such as Cu,Zn-superoxide dismutase (Cu,Zn-SOD), catalase (CAT), and glutathione-Stransferase (GST) in blood leukocytes of asthmatic children [bib_ref] Changes in mRNA and protein expression of antioxidant enzymes in children with..., Nesvitailova [/bib_ref]. They reported that following IHT, Cu,Zn-SOD protein content in leukocytes did not change significantly, but Cu,Zn-SOD mRNA expression increased by 33%. Conversely, GST protein synthesis increased by 90%, but its mRNA expression was invariable. Both protein content and mRNA of CAT increased by 37 and 13%, respectively. Furthermore, IHT altered mitochondrial enzymes such as succinate dehydrogenase (SDG) and alphaglycerophosphate dehydrogenase (GPDG) in asthmatic children. In 2002, Kurhaliuk et al. demonstrated a strong correlation between the individual hypoxic sensitivity and mitochondrial enzymes activities of GPDG in rats [bib_ref] Regulation of oxidative phosphorylation by liver mitochondria receptors after adaptation by rats..., Kurhaliuk [/bib_ref]. In 2012, Serebrovskaya et al. have shown that in children with decreased HVR, elevated basal SDG and GPDG activities were observed and greater increase was found after IHT [bib_ref] Intermittent hypoxia in treatment of bronchial asthma in childhood, Serebrovskaya [/bib_ref]. Similarly, they measured SDG and GPDG activities in peripheral lymphocytes of asthmatic children (9-13 years old). Both SDG and GPDG activities significantly increased under IHT by 78 and 42%, respectively.
## Brain function
There are few clinical observations made by Ukrainian and Russian researchers regarding the effects of IHT on CNS in children. Among the recent studies, Yatsenko et al. investigated the effects of IHT on CNS function and cerebral circulation in pediatric patients with cerebral palsy [bib_ref] Effects of intermittent normobaric hypoxia on the state of the CNS and..., Yatsenko [/bib_ref]. The 87 sick children (ranged from 9 months to 12 years of age) were examined before and immediately after the IHT course via inhalation of normobaric hypoxic gas mixture (12% O 2 ). Each cycle included 15 min of hypoxia alternated by 5 min of normoxia. The number of IH cycles gradually increased from one to three per day and the entire course of IHT lasted 10 days on average. The authors observed stable positive effects of IHT on the motor status in 94% of the patients and positive dynamics of spectral EEG components were seen in 70% of the patients. Doppler-detected brain hemodynamics was also normalized in 85% of the children who underwent IHT [bib_ref] Effects of intermittent normobaric hypoxia on the state of the CNS and..., Yatsenko [/bib_ref].
Borukaeva recently reported a study on brain bioelectrical activity, mental fitness, coordination of movements among 250 healthy young individuals (8-21 years of age) [bib_ref] Intermittent hypoxic training in the sanatorium and spa treatment for patients with..., Borukaeva [/bib_ref]. The subjects' EEG was recorded while breathing air or hypoxic gas mixture via "Bio Nova 204" hypoxicators made in Russia. They found that the changes in bioelectric activity of the brain under short-term hypoxia (manifested with the increased index and amplitude of alpha, theta, and delta waves of EEG) were similar among the groups of children (8-12 years), adolescents (13-16 years), and young adults . In the children group, the changes in alpha and theta waves were more sensitive indicators of hypoxia than other EEG waves. These data suggested an increased cortical impact on the background of increasing limbic system activity. On the other hand, the adolescents showed a greater reduction in mental performance and motor coordination during hypoxic exposure as compared with the other age groups. They showed an increased time for the maze passage, the number of touches and goes beyond the maze, decreased concentration and irradiation of excitatory and inhibitory processes, breach of their forces and mobility [bib_ref] Intermittent hypoxic training in the sanatorium and spa treatment for patients with..., Borukaeva [/bib_ref]. These results suggest that under hypoxia the children or young adults usually have increased cortical activity, whereas the adolescents have enhanced activity of subcortical structures. The age dependent-influence on CNS should be taken into account when designing IHT protocols with the most optimal condition, in order to prevent any undesirable complications of IHT.
In addition, a recent study demonstrated that IHT (5-min episodes of 10.5% O 2 with 5 min normoxic intervals, three exposures per week for 10 weeks) upregulates pro-plasticity molecules without evidence for CNS pathology in adult rats [bib_ref] Repetitive acute intermittent hypoxia increases expression of proteins associated with plasticity in..., Satriotomo [/bib_ref]. The authors suggested that IHT can be a useful therapeutic tool in treating disorders that cause respiratory insufficiency, such as spinal injury or motor neuron disease [bib_ref] Repetitive acute intermittent hypoxia increases expression of proteins associated with plasticity in..., Satriotomo [/bib_ref]. However, it remains to be determined whether such beneficial effects of IHT can be reproduced in immature bodies.
## Metabolic effects
The therapeutic use of IHT in treatment of metabolic disorders such as obesity had not been investigated until recent years. For instance, in 2014, Wang et al. designed and initiated a new randomized controlled trial to assess the effectiveness of a 4-week IH exposure plus conventional exercise training and diet intervention for inducing short-and long-term weight loss in obese adolescents (Clinical trial registration No. ChiCTR-TRC-14004106) [bib_ref] The effect of 'sleep high and train low' on weight loss in..., Wang [/bib_ref]. They planned to allocate 40 obese boys and girls (11-15 years old) into control group (sleep in normal conditions) and hypoxia group (sleep in a normobaric hypoxia chamber, simulating the "sleep high and train low" IHT mode). Results obtained from this study would potentially provide important evidence for the potential use of IHT in a weight loss intervention program among obese children and adolescents. Clarification of the mechanisms leading to weight loss in "sleep high and train low" protocol such as appetite regulatory effects could provide new information for the development of new strategies in combating obesity.
## Potential confounding factors for intermittent hypoxic training in children
## Hypoxia-induced polycythemia
An increased production of red blood cells (i.e., polycythemia) is one of the common adaptive responses of the body to hypoxia for improving oxygen transport capacity from the lungs to the target organs/tissues. Hypoxia stimulates the secretion of erythropoietin in kidneys and this hormone in turn stimulates the production of erythrocytes in bone marrow. It was suggested that erythropoiesis is an essential mechanism for long-term acclimatization to hypoxic condition [bib_ref] Erythropoietin in polycystic kidneys, Eckardt [/bib_ref]. In fact, polycythemia was observed in humans following chronic intermittent exposures to hypoxic environment at high altitude [bib_ref] Chronic intermittent hypoxia at high altitude exposure for over 12 years: assessment..., Brito [/bib_ref] [bib_ref] Chronic mountain sickness and the heart, Leon-Velarde [/bib_ref] [bib_ref] High-altitude disorders: pulmonary hypertension: pulmonary vascular disease: the global perspective, Pasha [/bib_ref]. A recent study reported that intermittent mining activity in the Andes at 4000 m altitude (i.e., from 4th to 20th day of each month worked at highaltitude mines followed by a resting period of 14 days at sea level) stimulates the production of red blood cells in bone marrow [bib_ref] Acclimatization to chronic intermittent hypoxia in mine workers: a challenge to mountain..., Farias [/bib_ref]. These results in adult humans are in contrast to an earlier study of Quechua Indians children, living at 4200 m altitude [bib_ref] Lack of prominent compensatory polycythemia in traditional native Andeans living at 4,200..., Garruto [/bib_ref] , which provided no support for the long-held belief that altitude hypoxia provokes a dramatic compensatory polycythemia in healthy adults.
In addition, although polycythemias or erythrocytoses in childhood and adolescence at sea level are very rare, the neonatal infants who are small for gestational age or affected by maternal gestational diabetes are at high risk for developing polycythemia [bib_ref] Neonatal polycythemia and hyperviscosity, Werner [/bib_ref]. Primary polycythemias are characterized by acquired somatic or inherited germ-line mutations expressed within hematopoietic progenitors that cause increased accumulation of red blood cells. Secondary erythrocytoses are driven by hormonal factors (predominantly by erythropoietin) extrinsic to the erythroid compartment [bib_ref] Childhood polycythemias/erythrocytoses: classification, diagnosis, clinical presentation, and treatment, Cario [/bib_ref]. With placental insufficiency, there may be chronic or acute fetal hypoxia resulted from birth asphyxia and hypothermia, neonatal hypoglycemia, polycythemia, and coagulopathy (91). Lee et al. described a family case of inherited CCHS, who was accompanied by hypoxia and hypercapnia and polycythemia with a hematocrit level of 70% [bib_ref] PHOX2B mutation-confirmed congenital central hypoventilation syndrome in a Chinese family: presentation from..., Lee [/bib_ref]. Furthermore, IH resulting from sleep apnea was suspected to induce polycythemia. However, Solmaz et al. recently suggested that OSA rarely causes secondary polycythemia [bib_ref] Is obstructive sleep apnoea syndrome really one of the causes of secondary..., Solmaz [/bib_ref]. Similarly, King et al. commented that OSA does not lead to clinically significant erythrocytosis [bib_ref] Obstructive sleep apnoea does not lead to clinically significant erythrocytosis, King [/bib_ref]. There is also no published evidence suggesting development of polycythemia in the children undergoing IHT for therapeutic purposes, which involves intermittent brief inhalation of hypoxic gas mixtures. Nevertheless, the issue of polycythemia should be considered as a potential confounding factor to be monitored in future practice of IHT in pediatric patients.
## Hypoxia-triggered disturbance in carbon dioxide homeostasis
Carbon dioxide (CO 2 ) is an important gaseous molecule that maintains whole body homeostasis as well as cellular signaling. CO 2 accumulates in the tissues during each episode of airway obstruction in sleep apnea or breath holding leading to acidemia. To the contrary, during IHT (e.g., high altitudes, hypobaric chamber, or inhalation of hypoxic gas mixtures), systemic hypoxia elevates pulmonary ventilation leading to hypocapnea and alkalemia. The physiological consequences of hypocapnea versus hypercapnia during IHT remain partially understood. At cellular levels, such changes in arterial CO 2 levels may affect vascular dynamics via activation or inactivation of vasoactive factors such as nitric oxide, angiotensin II, endothelin, and bradykinin [bib_ref] Sensing, physiological effects and molecular response to elevated CO2 levels in eukaryotes, Sharabi [/bib_ref].
On the other hand, Zhang et al. recently investigated experimental hypocapnia and hypercapnia following 14-day sessions of IHT with 10% inspired O 2 and they reported that the repetitive normobaric IH exposures significantly diminished variations of cerebral perfusion in response to both hypercapnia and hypocapnia without compromising cerebral tissue oxygenation in adult humans [bib_ref] Two-week normobaric intermittenthypoxia exposures enhance oxyhemoglobin equilibrium and cardiac responses during hypoxemia, Zhang [/bib_ref]. Similar results were also communicated recently by . Another study by Snow et al. in rats indicated that hypocapnic but not eucapnic IH increases hematocrit and causes a more profound increase in right ventricular mass than those who underwent eucapnic IH [bib_ref] Differential effects of chronic hypoxia and intermittent hypocapnic and eucapnic hypoxia on..., Snow [/bib_ref]. Taken together, there are very limited data on the role of hypocapnia or hypercapnia in the adaptive process to IH or IHT in adults and virtually no report in pediatric individuals. Therefore, at the present time, we cannot give concrete advice on how to manipulate the CO 2 levels during IHT for avoiding adverse effects and/or enhancing therapeutic efficacy of IHT. There is little doubt that the concomitant changes in arterial CO 2 levels during IH exposures can have profound impact on the neurologic and other cellular outcomes (detrimental or beneficial), considering the critical regulatory role of CO 2 homeostasis in the body systems. This important question should be carefully addressed in future investigations.
## Concluding remarks
Taken together, the above-discussed studies in healthy children and pediatric patients have indicated that IH can profoundly trigger either adaptive or maladaptive responses in childhood, which has impact on the structural and functional development in multiple body organs and systems. Whereas more severe regimens of IH induce pathological outcomes, the well-controlled and proper regimens of IHT can produce therapeutic effects against various chronic diseases in children, including BA, allergic dermatoses, autoimmune thyroiditis, cerebral palsy, and obesity. Many interesting emerging fields of IH research such as stem cell biology should be further explored. For example, Bhaskara et al. recently reported that IH enhanced stem-like characteristics and suppressed differentiation propensities in neuroblastoma cells [bib_ref] Chronic intermittent hypoxia at high altitude exposure for over 12 years: assessment..., Brito [/bib_ref]. Serebrovskaya et al. have shown that IH mobilizes hematopoietic progenitors and augments cellular and humoral elements of innate immunity in adult men [bib_ref] Chronic mountain sickness and the heart, Leon-Velarde [/bib_ref].
It is noteworthy that the different stages of child development and various degrees of maturity have great impact on the individual response to IH. Such an age-dependence is not fully characterized in this review due to the lack of published evidence. The short-and long-term effects of IH on the modulation of neurotransmitter release, receptor binding and expression, intracellular signaling cascades, transcriptional regulation, and gene expression as a function of animal maturity are almost completely unknown. As Waters and Gozal pointed out that the responses to IH vary according to the point within the sequence of a single response where the stimulus interruption occurs [bib_ref] Responses to hypoxia during early development, Waters [/bib_ref]. An intermittent stimulus may be seen as "continuous" if the recurrence frequency exceeds a certain threshold, whereas application of slower cycles below such threshold may elicit discordant recruitment of the compensatory responses. Further delineation of such complex responses to IH may permit the formulation of interventional strategies aiming at reducing the overall vulnerability of the young infant and child to apnea and sudden death [bib_ref] Respiratory plasticity following intermittent hypoxia: developmental interactions, Gozal [/bib_ref].
Finally, it is also important to emphasize that clinical research involving young children and infants should be considered separately from that of adults and require the highest ethical standard for the researchers, because the pediatric subjects are often lack of complete understanding and cannot give informed consent to any procedure. Such ethical issues should be a primary concern for those who conduct IH research with infants and young children [bib_ref] Ethical issues in lung function testing in children, Beardsmore [/bib_ref]. The pediatric patients' safety and welfare should always be the top priority of our research involving hypoxic stimulus. A proper choice of IHT regimens for children should take into account uneven and heterochronical development of different functions, in order to avoid the previously described negative effects of IH while maximizing its therapeutic potential.
May 2015 | Volume 3 | Article 44
Frontiers in Pediatrics | www.frontiersin.org May 2015 | Volume 3 | Article 44
Conflict of Interest Statement:The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The Review Editor Chin Moi Chow declares that, despite ongoing collaborations with the author Lei Xi, the review process was handled objectively.Copyright © 2015 Serebrovskaya and Xi. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, inaccordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. |
Prevalence of Pulmonary Tuberculosis among Adults in a Rural Sub-District of South India
Background: We conducted a survey to estimate point prevalence of bacteriologically positive pulmonary TB (PTB) in a rural area in South India, implementing TB program DOTS strategy since 2002.Methods: Survey was conducted among persons $15 years of age in fifteen clusters selected by simple random sampling; each consisting of 5-12 villages. Persons having symptoms suggestive of PTB or history of anti-TB treatment (ATT) were eligible for sputum examination by smear microscopy for Acid Fast Bacilli and culture for Mycobacterium tuberculosis; two sputum samples were collected from each eligible person. Persons with one or both sputum specimen positive on microscopy and/or culture were labeled suffering from PTB. Prevalence was estimated after imputing missing values to correct for bias introduced by incompleteness of data. In six clusters, registered persons were also screened by X-ray chest. Persons with any abnormal shadow on X-ray were eligible for sputum examination in addition to those with symptoms and ATT. Multiplication factor calculated as ratio of prevalence while using both screening tools to prevalence using symptoms screening alone was applied to entire study population to estimate prevalence corrected for non-screening by X-ray.Results: Of 71,874 residents $15 years of age, 63,362 (88.2%) were screened for symptoms and ATT. Of them, 5120 (8.1%) -4681 (7.4%) with symptoms and an additional 439 (0.7%) with ATT were eligible for sputum examination. Spot specimen were collected from 4850 (94.7%) and early morning sputum specimens from 4719 (92.2%). Using symptom screening alone, prevalence of smear, culture and bacteriologically positive PTB in persons $15 years of age was 83 (CI: 57-109), 152 (CI: 108-197) and 196 (CI :145-246) per 100,000 population respectively. Prevalence corrected for non-screening by X-ray was 108 (CI: 82-134), 198 (CI: 153-243) and 254 (CI: 204-301) respectively.Conclusion: Observed prevalence suggests further strengthening of TB control program.
# Introduction
National Tuberculosis Programme (NTP) based on a cost effective operational strategy was implemented all over India from 1962; after a nation-wide survey during 1955-1958 revealed that tuberculosis (TB) was highly prevalent throughout the country. Surveys carried out thereafter in geographically defined areas revealed that the prevalence of TB continued to be high, though varied, in different parts of the country. Taking cues from a review of NTP, Revised National tuberculosis Control Programme (RNTCP) adopting DOTS (an internationally recommended strategy for TB control) was launched in 1997 and expanded in phases to cover the entire population by 2006. Implementation of RNTCP lead to improvements in case detection and high treatment success rates in most parts of the country [bib_ref] The status and prospects of tuberculosis control in India, Khatri [/bib_ref]. Indeed, a decline of about 50% in prevalence of pulmonary tuberculosis (PTB) was observed from 1999 to 2006 in a sub-division of Thiruvallur district, Tamil Nadu state [bib_ref] Rapid decline in prevalence of pulmonary tuberculosis after DOTS implementation in a..., Subramani [/bib_ref]. Since, this observation pertained to a single geographical site, Government of India identified six institutions to carry out independent surveys in seven other sites (districts/sub-division of a district) located in different geographical regions using a generic protocol. The objective of the survey at each site was to find out the point prevalence of bacteriologically positive PTB among adults, which would provide a baseline data to measure the future trends in different parts of the country by repeat surveys. The personnel to be involved at these sites were trained in actual field conditions at a common site by a team of vastly experienced trainers. One of these surveyed sites was the Nelamangala sub-division in Bangalore rural district, the results of which are reported hereunder.
# Materials and methods
## Study site and setting
Nelamangala is one of the four sub-divisions in Bangalore rural district. This district is located adjacent to Bangalore city which is the capital of Karnataka state in southern India. DOTS strategy is being implemented in the entire district since 2002. Nelamangala was chosen for this survey so that the trends from the past could also be elicited since a similar survey had been carried out there during the year 1975 [bib_ref] Estimation of prevalence of bacillary tuberculosis on the basis of chest x-ray..., Gothi [/bib_ref].
Majority of adults in Nelamangala are engaged in agricultural occupations and a proportion work in factories located in a bordering industrial area in Bangalore city.
## Study population
Persons $15 years of age residing in Nelamangala sub-division for $6 months comprised the study population.
## Sampling
Simple random sampling (SRS) was adopted for selection of clusters. Each cluster corresponded to a group of villages in the jurisdiction of a 'Gram Panchayat' -a local self government consisting of elected members of the community and having responsibility to implement rural development programs. Office bearers of the Panchayats visited during planning phase of the study insisted for political reasons to cover all the villages in their jurisdiction. There were altogether 25 panchyats in Nelamangala sub-division.
Since the exact count of adult population in individual clusters was not known, fifteen (arbitrarily decided number) clusters were selected and the survey was planned to start in the first selected cluster covering all eligible persons and proceed to subsequent clusters in order of their selection till the required sample size was achieved.
## Sample size
Sample size was originally calculated at 47,828 to estimate the prevalence within 20% of the true value at 5% level of significance considering a design effect of 2 to account for cluster sampling. Expected prevalence of bacteriologically positive PTB (positive for AFB on microscopy and/or culture) while using both screening tools -symptom screening and chest X-ray, was arbitrarily considered at 400 per 100,000 population. During the earlier survey in the area in 1975, prevalence of culture positive PTB in the age group of $15 years while using both screening tools was 480 per 100,000 population [bib_ref] Estimation of prevalence of bacillary tuberculosis on the basis of chest x-ray..., Gothi [/bib_ref] ; prevalence of bacteriologically positive PTB was not reported.
Since the mobile X-ray equipment broke down during the course of the survey, sample size was revised during the course of the survey to 68,400 for an expected prevalence of 280 per 100,000 population assuming that 30% of PTB cases would be missed due to non-screening by X-ray. This assumption was based from the data of the previous survey in the same area [bib_ref] Estimation of prevalence of bacillary tuberculosis on the basis of chest x-ray..., Gothi [/bib_ref].
## Key variables
Key study variables were: age, sex, presence of pulmonary symptoms, history of anti-TB treatment (ATT), presence of abnormal shadow on X-ray, result of sputum smear microscopy -spot & early morning specimen and result of culture -spot & early morning specimen.
## Data collection instruments
The following schedules were useful for data collection:i.
Household form: to enlist all residents by age and sex ii. Individual card, for each eligible person: to record study identification number (ID), age, sex, symptom status, X-ray film number, date of sputum collection, results of sputum smear and culture examination iii. Laboratory register to record sputum specimen numbers and results of smear and culture examination iv. X-ray result form
## Field procedures
Field work was carried out during October 2008-June 2010. A planning visit was made to each Panchayat office and each village to familiarize the officials, village leaders and the community with the purpose and procedures of the survey and to seek their cooperation. In each village, a rough sketch of lanes and hamlets showing approximate number of houses in each was drawn, after going around and in discussions with village leaders. Survey in each village began on a mutually agreed date. Enumerators went to each household and recorded the age, sex and resident status of each individual. Each eligible person (15 years or more in age and residing for $6 months in the household or in any other village in Nelamangala sub-division) was registered into an individual card. Subsequently, a symptom elicitor queried each eligible individual for presence of symptoms suggestive of PTB (persistent cough for $2 weeks, fever or chest pain for $1 month, presence of blood in sputum any time during last 6 months) and history of ATT. Field supervisors re-interviewed 10% of eligible individuals, as a quality control mechanism.
Individuals having pulmonary symptoms suggestive of PTB or a positive history of ATT were eligible for sputum examination. A spot sputum specimen was collected at a temporarily setup sputum collection centre within the village, after briefing by the laboratory technician (LT) on how to bring out good sputum sample and spit into a pre-numbered sterilized screw capped sputum cup. After a satisfactory extraction of spot specimen, a pre-numbered empty bottle was given for collecting another sample next morning. Sputum containing bottles marked with ID of each patient were transported in a sputum box on day of collection to the accredited laboratory of the institute.
In six of the clusters, each registered person was also screened by X-ray. A 70 mm photofluorography film of the chest was taken at the temporarily setup centre using a mobile mass miniature radiography (MMR unit); pregnant women and bed ridden individuals were excluded (screening by X-ray could not be undertaken in other clusters due to breakdown of equipment during the middle period of the survey). After processing of exposed MMR rolls in X-ray laboratory of the Institute, each film was read and classified by two trained independent X-ray readers into one of the categories -N (normal)/TI (technically inadequate)/A (lung pathology other than tuberculosis)/B (tuberculosis inactive)/C (tuberculosis active). Persons with their films labeled as TI/A/B/C by any of the two X-ray readers were eligible for sputum examination, in addition to those having symptoms suggestive of PTB or history of ATT.
## Laboratory procedures
Sputum specimens were subjected to smear microscopy for acid fast bacilli (AFB) and culture for M. tuberculosis following standard laboratory procedures, at a national reference laboratory located in National Tuberculosis Institute, Bangalore (NTI).
Two direct smears were made from each specimen on new labeled slides under aseptic conditions in a bio-safety cabinet. Each smear was stained using 0.1% auramine-O and 0.5% potassium permanganate and examined under a fluorescence microscope at a magnification of 2006.
After taking out the amount required for making smears, each specimen was homogenized and transferred to a McCartney bottle. For decontamination, 4% sodium hydroxide was added in a volume twice that of sputum specimen (Modified Petroff's Method) and incubated in a shaker for 20 minutes. Sterile distilled water was then added up to the neck of the bottle and centrifuged at 3000 rpm (revolutions per minute) for 15 minutes. The supernatant was decanted and the deposit was inoculated onto 2 slopes of Lowenstein-Jensen (LJ) medium. Cultures were incubated at 37uC and examined for the presence of mycobacterial colonies every week for 8 weeks. Culture was discarded in case of contamination or no growth at 8 weeks. The growth if observed was subjected to Niacin test and incubation on LJ medium containing p-nitro benzoic acid (PNB) in a concentration of 500 mg/ml. It was labeled as M. tuberculosis, if Niacin test was positive and no growth was observed on PNB containing medium.
Quality assurance of sputum microscopy and culture was undertaken using the existing mechanism of RNTCP.
## Study investigators
Survey was carried out under the overall supervision of the Senior Epidemiologist in the institute. Enumeration of residents in the households, screening by interview, sputum collection and data entry was undertaken by trained contractual field staff and monitored by experienced institutional field staff. Subjecting the study subjects to MMR, reading of X-ray films, examination of sputum by smear and culture and analysis was undertaken by experienced staff of the institute.
# Ethical considerations
Survey was approved by the Institutional Ethics Committee. Written consent for participation was sought from each individual, after explaining procedures of the survey and its benefits to the individuals and community, through material printed in local language and also personally by field staff. No one was compelled to participate.
TB cases detected during the survey were counseled and facilitated to initiate ATT at the nearest RNTCP facility. For the patients who failed to reach the RNTCP facility, repeated visits (upto a maximum of nine) were made to their households by the field staff to persuade them and also some times to accompany them to RNTCP facility in order to initiate ATT. Follow up during treatment was the responsibility of routine staff of RNTCP.
Individuals with symptoms but not having TB were advised to seek treatment at the local health center; however no active intervention was undertaken by survey team for referral or otherwise.
Village leaders were informed of the findings of the study emphasizing their role in raising awareness regarding symptoms of TB and availability of quality services under RNTCP.
## Definitions
Smear Positive case: An individual with at least one sputum specimen found to be positive for AFB on smear microscopy, irrespective of culture result. Culture Positive case: An individual with at least one sputum specimen found to be positive for M. tuberculosis on culture, irrespective of smear result.
Bacteriologically positive case: An individual with at least one sputum specimen found to be positive for AFB on microscopy and/or M. tuberculosis on culture.
# Statistical methods
Digitized data was verified and analyzed using SPSS version 17.0 and STATA version 12.0.
Point prevalence was estimated by three different approaches:
[formula] i) [/formula]
Crude prevalence (P). It was calculated as the total number of smear, culture or bacteriologically positive cases detected during the survey divided by the total number (n) of individuals who were screened by interview and whose both sputum results were available. Standard error (SE) was estimated as Standard deviation (SD)/!n where SD = !P (12P). Confidence intervals (95%) of the estimates were calculated as the mean of the binomial exact 6 2SE. ii) Cluster level analysis. Average cluster level prevalence was calculated as P cluster level = gp i /c, where p i is the crude prevalence in i th cluster and c is the number of clusters. SE was calculated as SD/!c where SD = ![g(p i 2P cluster level ) 2 / c]. Confidence intervals (95%) were calculated as the mean prevalence across clusters 6 2SE. iii) Individual level analysis. To correct for bias introduced by incompleteness of data, individual level analysis was performed using logistic regression model with robust standard error. To include all eligible individuals in analysis, missing value imputation was undertaken for individuals: (a) without symptom screening (b) symptoms present but the result of one or both sputum specimen not available. The method accounted for between-cluster variability and uncertainty in estimating SE, under the assumption that within groups of individuals belonging to same sex, agegroup and TB symptoms, data are missing at random. For imputation of missing values for each variable, starting values were assigned to missing data, which in turn was obtained from a random sample of values from individuals whose data were available. Model was then fitted with this particular variable as outcome variable and other variables as explanatory variables. This was done sequentially, in order of the proportion of data that were missing starting with variables with least missing data. Finally, a logistic regression model with smear/culture/bacteriologically confirmed TB as the outcome variable and sex, age-group and TB symptoms as explanatory variables were fitted. Newly imputed values were used as starting values for next iteration of the process which was undertaken in ten cycles in order to obtain one imputed data set. Five such data sets were imputed and the average of their prevalence was considered as final prevalence. This model takes account of clustering in the survey design, variation in the number of individuals per cluster and the uncertainty introduced by imputation of missing values when estimating the 95% CI for prevalence of PTB.
In six clusters where both screening tools were used, individual level analysis was also performed separately with MMR reading as an additional explanatory variable after imputing its missing values. A multiplication factor was calculated as the ratio of prevalence while using both screening tools to the prevalence using symptoms screening alone. This factor was applied to the entire study population to estimate prevalence corrected for nonscreening by MMR.
Chi-square test with continuity correction was used to test the significance of differences between proportions and p-values,0.05 were considered significant.
## Study participants
A total of 71,874 individuals across 158 villages in 15 clusters were eligible for inclusion in the survey. Of them, 63,362 (88.2%) were screened by interview; 8512 (12%) could not be screened due to not being available despite repeated visits to their households. Non-participation rates were higher in males compared to females in all age groups, difference being more predominant in 15-45 years [fig_ref] Table 1: Age and sex distribution of participants and non-participants [/fig_ref].
Size of eligible population varied from 3620 to 7473 in different clusters. Male to female ratio was about 1:1 in each cluster.
# Results
## I. prevalence of ptb based on symptom screening followed by sputum examination
Of the persons interviewed, 4681 (7.4%) were found to have symptoms. Of them, about 83% had cough of 2 weeks or moreeither alone or in combination with one or more of the other symptoms-chest pain, fever, haemoptysis. The remaining 17% had one or more of the other symptoms [fig_ref] Table 2: Frequency distribution of persons with different pulmonary symptoms [/fig_ref].
Previous history of ATT was present in 653 (1.0%) persons and 56 (0.09%) were currently on ATT. Of the total of 709 persons with history of ATT, 270 had symptoms at the time point of the survey while 439 (0.7% of the total interviewed) did not have symptoms.
Of 5120 persons (4681 with symptoms and an additional 439 with history of ATT) eligible for sputum examinations, spot specimen were collected from 4850 (94.7%), of which 28 were positive for AFB on smear microscopy and 65 on culture.
Overnight specimens were collected from 4719 (92.2%), of which 42 were smear positive and 61 were culture positive. Overall, 51 persons were smear positive and 86 culture positive, on spot and/ or overnight specimen. About 1% of spot and 1.8% of overnight specimen were contaminated.
There were a total of 110 bacteriologically positive patients. Of them, 101 were picked up through screening for presence of symptoms and an additional nine through screening for history of ATT. Among those with symptoms, 89 (88.1%) had cough with or without other symptoms while 12 (11.9%) had one or more symptoms other than cough.
Estimated point prevalence of smear positive, culture positive and bacteriologically positive PTB estimated by three analytical methods (using screening by interview for presence of symptoms) is presented at
## Ii. prevalence of ptb corrected for non-screening by mmr
In six clusters where both screening tools were used, there were a total of 31, 823 eligible persons of whom 28,398 (89.2%) were screened by interview and 26,429 (83.0) by MMR. Using individual level analysis, prevalence of bacteriologically positive PTB by using screening for symptoms (and ATT) was 298 (CI: 327-451) per 100,000 population; it was 389 (CI: 243-354) per 100,000 population when screening by MMR was also considered. Multiplication factor for prevalence to correct for non-screening by MMR was 1.3. Considering the estimates by individual level analysis as the best estimates, prevalence among all persons $15 years of age residing in 15 clusters, corrected for non-screening by X-ray, was 108 (CI: 82-134), 198 (CI: 153-243) and 254 (CI: 204-301) per 100000 population respectively for smear, culture and bacteriologically positive PTB.
## Iii. relationship of prevalence with sex and age
Estimated crude prevalence by symptom screening alone was used to find out the relationship with sex and age. Prevalence was about six times higher among males compared to females and generally increased with age [fig_ref] Table 4: Prevalence [/fig_ref].
# Discussion
Prevalence of bacteriologically positive PTB in persons $15 years of age corrected for non-screening by MMR in the entire sample of selected clusters was estimated at 254 per 100000 population, after imputing the values for missing data. Prevalence of smear positive and culture positive PTB was estimated at 108 and 198 respectively.
Prevalence was not estimated for abacillary (bacteriologically negative but radiologically active) TB, due to low specificity of diagnosis of PTB when based on single X-ray picture especially in field conditions [bib_ref] Interpretation of photofluorograms of active Pulmonary TB patients found in epidemiological surveys..., Gothi [/bib_ref]. Similarly, tools for diagnosing EPTB are too cumbersome to apply in field conditions. A common generic protocol was used for simultaneously carried out surveys during 2008-10 at eight sites in the country including Nelamangala. However, screening by symptoms only was undertaken at five sites; MMR as additional screening tool was used in the study populations at two sites and a proportion of study population in Nelamangala. Using only symptom screening, prevalence of bacteriologically positive TB at these eight sites varied between 129 (CI: 92-165) to 399 (325-469) per 100,000 population. Prevalence at three sites where both screening tools were used varied between 199 (CI: 147-251) to 391 (CI: 352-430).
A number of surveys had earlier been carried out in different parts of India since mid-twentieth century. In 1956, a nation-wide survey using screening by MMR followed by sputum examination by smear and culture among those with abnormal shadow on Xray film revealed the prevalence of bacteriologically positive PTB at 400 per 100,000 population. During subsequent surveys in different geographical locations at different points of time, prevalence of bacteriologically positive TB varied between 182-1270 per 100,000. These surveys not strictly comparable due to variations in definition of symptoms, screening tools (symptoms and/or MMR), case definition and analytical methods, nonetheless revealed that TB continued to be a high burden disease in India. Male to female ratio in these surveys has been found to vary between 2:1 to 5:1. It was 6:1 during the present survey.
The earlier survey in Nelamangla sub-division during 1975 was carried out in 55 villages selected by SRS method [bib_ref] Estimation of prevalence of bacillary tuberculosis on the basis of chest x-ray..., Gothi [/bib_ref]. A total of 12,105 persons $15years of age were screened for pulmonary symptoms as well as by MMR. As in the present survey, a spot and an early morning sputum sample were collected from those with symptoms and/or any abnormal shadow on MMR. However, there were differences in age structure with higher proportion of persons in elder age groups during the present survey. We compared the prevalence at the two surveys in the age structures as they existed at given times since TB prevalence has been found . Prevalence of PTB per 100,000 population, by method of estimation, all clusters-using screening by symptoms alone. : Reason for difference in numbers analyzed by three methods: for crude and cluster level prevalence, only those individuals from whom all actual data is available were considered for analysis; persons missed from screening for symptoms and persons who were eligible for sputum collection but one or both sputum specimen not collected or found contaminated on culture were excluded. doi:10.1371/journal.pone.0042625.t003 [bib_ref] Significant decline in the tuberculosis burden in the Philippines ten years after..., Tupasi [/bib_ref] [bib_ref] Threefold reduction in the prevalence of tuberculosis over 25 years in Indonesia, Soemantri [/bib_ref].
A major limitation of the present survey was that mobile odelca camera based MMR unit broke down mid way and could be put back to use only after a long gap. This underlies the necessity of a back up machine. We worked out a multiplication factor of 1.3 to correct for non-screening by X-ray from the data in a subset of the eligible population in which both screening tools were used. This factor has been assumed to be homogenous across the study population which could be a limitation of this approach given that the prevalence even by symptom screening alone was higher in the six clusters where both screening tools were used compared to other clusters. Incidentally, these six clusters were located in closer proximity to an urban area and had higher population density than the rest. These six clusters were however not representative of the study area. Same correction factor of 1.3 has been estimated in Wardha district from the simultaneously carried out survey in 2008-09 (unpublished data). During earlier surveys in different parts of the country, this correction factor was found to vary between 1.5 J and 1.7 [bib_ref] Yield of pulmonary tuberculosis cases by employing two screening methods in a..., Gopi [/bib_ref] [bib_ref] Tuberculosis in North Arcot District of Tamilnadu-a sample survey, Datta [/bib_ref] [bib_ref] A tuberculosis prevalence survey based on symptoms questioning and sputum examination, Gopi [/bib_ref] [bib_ref] Quality of symptom elicitation in an epidemiological survey on tuberculosis, Gopi [/bib_ref]. This variation may be attributed to variation in quality of symptom screening by field surveyors.
Another limitation could be that missing value imputation might not fully account for the missing cases in the event of higher prevalence among the absentees compared to those surveyed.
Finally, the estimated prevalence reveals that TB is still a major problem considering that much lower levels of prevalence have been observed in countries with good TB control programs in place. To further reduce the disease prevalence, further strengthening of the program may be undertaken to ensure universal access to diagnosis, early detection of all incident TB patients including multi-drug resistant (MDR) patients and to achieve higher rates of treatment success. Repeat surveys if carried out in all eight recently surveyed sites after a reasonable time interval of approximately 7-8 years would provide information on future trends in different parts of the country.
[table] Table 1: Age and sex distribution of participants and non-participants. [/table]
[table] Table 2: Frequency distribution of persons with different pulmonary symptoms. [/table]
[table] Table 4: Prevalence (crude) of bacterologically positive PTB per 100,000 population, by age-group and sex. 95% confidence intervals. doi:10.1371/journal.pone.0042625.t004to increase with age[2]. Thus increase in life expectancy per se can lead to increase in prevalence and can offset any declining trends. We compared the two surveys for crude prevalence of culture positive PTB using symptom screening only as authors of previous survey have reported only crude prevalence of culture positive PTB and that a proportion of study population during present survey was not screened by MMR. This crude prevalence was 311 per 100,000 population in 1975 and 137 (112-161) in the present survey[6]. This denotes a decline in prevalence by 56% between the two survey periods ¥ Longitudinal studies in other areas during NTP era had revealed no significant change in prevalence[2]. Thus assuming that prevalence in the present study area remained steady upto 1990, there was 56% decline in crude prevalence between base year of Millennium Development Goals (MDGs) of the United Nations and 2009 (mid-point of present survey). This decline in Nelamangla can be attributed to overall strengthening of TB control activities due to political commitment, implementation of DOTS strategy with quality assured sputum microscopy and administration of a rifampicin based treatment regimen under observation with high rates of treatment success[3]. During the year 2010, case notification rate of PTB (new+retreatment) under RNTCP in the Nelemangala Sub-division was 203 per 100,000 population in $15 years age group [unpublished quarterly case finding reports of Nelamangala Tuberculosis Unit, published data available only at district level]. This corresponds to 80% of the estimated point prevalence of bacteriologically positive PTB at 254 per 100,000 population. Prior to DOTS, TB programme performed poorly due to lack of adequate infrastructure and treatment with a self-administered regimen as in the rest of the country[12]. The progress towards MDGs from the base year of 1990 was thus far available only from Tiruvallur where three rounds revealed a decline in prevalence by about 50% between 1999 and 2006[5]. Trends in prevalence of PTB in other Asian countries have revealed declines of 32% in China between 1990-2000, 30% in Philippines between 1997-2007 and 67% inIndonesia between 1980Indonesia between -2004 [/table]
|
Prevalence of viral hepatitis B and C in Sierra Leone—current knowledge and knowledge gaps: a narrative review
There are no comprehensive data on viral hepatitis from Sierra Leone; however, a huge disease burden has been observed in different subpopulations. This review summarizes available data on hepatitis B and C virus (HBV and HCV) prevalence in Sierra Leone and identifies knowledge gaps. Despite the non-uniformity of the studies and the lack of systematic case recording, different reports published in recent decades yielded a hepatitis B prevalence of 8.7% among healthcare workers, 11.3% among pregnant women, 15.2% among blood donors and 16.7% in school-age children. The actual HBV prevalence in the general population was reported as 21.7%; similar to what was reported for people living with human immunodeficiency virus (PLHIV). HCV prevalence is 8% and 7% in male and female blood donors, respectively, 4.1% in PLHIV and 2.0% in school children. There are significant knowledge gaps regarding the prevalence of viral hepatitis B and C in Sierra Leone, despite the high burden reported in a few studies. There are limited programmatic interventions on the control and prevention of viral hepatitis in the country. Therefore, well-structured representative studies should provide a solid understanding of the true prevalence of hepatitis B and C to inform best possible public health measures in Sierra Leone.
# Introduction
Globally, in 2015, an estimated 328 million people were living with chronic hepatitis B virus (HBV) and hepatitis C virus (HCV) infections, with >1 million deaths attributable to the sequelae of these infections. Africa and the Western Pacific are the main regions affected by the epidemic of HBV, with an estimated prevalence of 6.1% and 6.2%, respectively.Five main types of viral hepatitis (A-E) have been described.Hepatitis A virus (HAV) and hepatitis E virus (HEV), transmitted by the faecal-oral route, usually cause acute self-limiting inflammation of the liver.Transmitted through exposure to blood, motherto-child transmission (MTCT) or sexual intercourse, HBV and HCV can cause both acute and chronic liver diseases.In contrast to HBV, sexual transmission of HCV is rare and is seen mostly in human immunodeficiency virus (HIV)-positive men who have sex with other men.Like HBV and HCV, hepatitis D virus (HDV) is transmitted through blood or blood products or sexual transmission and causes infection only in individuals co-infected with HBV.HBV and HCV, which account for 96% of all hepatitis-related mortality,are a global public health concern. While the epidemic caused by HCV affects all regions, the highest endemicity for HBV occurs in the African and Western Pacific regions, with recorded prevalences of 6.1% and 6.2%, respectively.According to a modelling study on the prevalence and genotypes of HCV, the prevalence in the West African subregion was estimated at 1.3% in 2015.Despite being a global health threat, <10% of patients with chronic HBV and 20% of patients with chronic HCV were diagnosed in 2015.With an estimated 1.34 million deaths attributable globally to viral hepatitis in 2015, comparable Transactions of the Royal Society of Tropical Medicine and Hygiene to deaths caused by tuberculosis,and the availability of effective treatment for both HBV and HCV in high-income countries, an urgent programmatic response is needed to contain the epidemic in low-and middle-income countries (LMICs).
Sierra Leone is a low-income country in West Africa that has experienced a decade-long civil warand a devastating Ebola epidemic,with a huge impact on its health service delivery. Currently the country has no nationally representative data to guide the diagnosis, management, prevention and control of chronic viral hepatitis. However, data from small studies have demonstrated a substantial burden of viral hepatitis in various subpopulations of Sierra Leone. In this article we review the available literature on HBV and HCV epidemiology in Sierra Leone, identify knowledge gaps and suggest solutions to the viral hepatitis epidemic in Sierra Leone and similar countries.
# Methods
A review of published articles on the prevalence of HBV and HCV was conducted through PubMed and Google Scholar using the search criteria 'Sierra Leone' AND 'hepatitis B' OR 'hepatitis C'. The search was conducted with the publication language restricted to English. Titles of articles were all reviewed, followed by the abstracts and the full articles. All articles with the prevalence of HBV, HCV or both in the relevant subpopulations were considered for inclusion.
# Literature search results
Fifteen articles were published, 12 of which were retrieved through PubMed and 3 through Google Scholar. The literature search was initially conducted between September 2019 and November 2019. An updated search was performed between February and May 2020.
All articles primarily identified, although methodologically limited, were considered relevant. All articles were published between 1997 and 2019.provides study characteristics and main prevalence findings for all studies included.
While nine articles focused on HBV, five articles assessed the burden of both HBV and HCV. One article assessed HCV only. All the articles were on cross-sectional studies, of which five were retrospective. The populations studied were blood donors (four articles), routine hospital patient population (four articles), healthcare workers (two articles), pregnant women (two articles), suspected Ebola cases (one article), people living with HIV (PLHIV; one article) and school children (one article). A total of 67 189 patients were enrolled in these studies. All but one study in school children was carried out in health facilities. While the location for eight of the research articles was Freetown, the capital of Sierra Leone, the research work for two articles was conducted in two of the provinces and the remaining five were conducted in both Freetown and selected provincial areas of Sierra Leone.
## Findings
## Viral hepatitis among healthcare workers in sierra leone
The prevalence of HBV, defined as a positive hepatitis B surface antigen (HBsAg) test conducted using a rapid lateral flow assay and enzyme-linked immmunosorbent assay (ELISA) was 8.7% (39/477) among 447 healthcare workers in two hospitals in Freetown. The prevalence was 9.7% (30/320) among participants who were ≥30 y of age compared with those <30 y of age (7.1% [9/127]).Alanine transaminase (ALT) was twice the upper limit of normal in 34.6% (9/27) of these HBV-positive healthcare workers. A total of 14.3% of healthcare workers received at least one dose of HBV vaccine, but serologic evidence of past immunization was present in only 4.5% of this population.Adequate knowledge on clinical outcomes of HBV, its route of transmission and correct preventive measures was reported among 38.9%, 34.6% and 6.6% of the healthcare workers, respectively. 9
## Status of viral hepatitis among blood donors and non-pregnant patients attending health facilities in sierra leone
A retrospective review of 16 807 blood donors nationwide reported an HBV prevalence, defined as HBsAg positivity determined by a rapid diagnostic test, of 9.7% (1633/16 803) and an HCV seroprevalence, defined as anti-HCV positivity determined by a rapid diagnostic test, of 1.0% (159/16 802).The prevalence was higher among rural blood donors (11% vs 7% for HBV and 1% vs 0.3% for HCV). 10 Similar findings were observed for blood donors in a rural hospital in northern Sierra Leone, where an HBV prevalence of 13% in women and 15% in men and an HCV prevalence of 7% in women and 8% in men were reported.In eastern Sierra Leone, the overall HBV prevalence was 10.9% among blood donors and the general patient population that was tested using a rapid diagnostic test.In this study, the prevalence was higher (20.5%) in male patients than male blood donors (7.4%). A similar trend of HBV prevalence was observed in female patients (20.6%) and female blood donors (9.1%).A study of 43 163 blood donors (utilizing diagnostic tests whose specifics were not described in the publication) in Freetown conducted over 5 y showed an HBV prevalence of 15.2% (6554/43 163). In this study, HBV prevalence was higher in males (5735/37 060 [15.5%]) than females (829/6103 [13.6%]).Compared with a military hospital (4.7%), the HBV prevalence was higher among blood donors in an emergency surgical centre (33.2%).In the general population, an HBV prevalence of 21.7% (43/198) was recorded in patients attending a private laboratory in Freetown.The overall prevalence of HBV was 13.7% (118/860) among febrile patients in a secondary hospital in southern Sierra Leone, with a prevalence among males and females of 15.5% (51/330) and 12.6% (67/530), respectively.In the southern regional hospital, the prevalence of HBV was 21.4% (66/308).The HCV and HBV prevalences in samples of febrile jaundice patients using Ion Torrent semiconductor sequencing with quantitative polymerase chain reaction were 34.38% and 1.04%, respectively.However, using ELISA, HCV antibodies were detected in 0.22% of samples of suspected Ebola cases during the 2014-2016 Ebola epidemic in Sierra Leone. 18
## Viral hepatitis among pregnant women
Among pregnant women in Sierra Leone, 6.2% (28/302) to 11.3% (20/179) had a positive hepatitis B serology conducted using immunochromatographic hepatitis B surface antigen test card,
## Hbv and hcv prevalence in hiv co-infection
A high prevalence of HBV (21.7%) and HCV (4.3%) was demonstrated in a cohort of 211 PLHIV.The prevalence of HIV-HBV co-infection was 6.6% (13/138) among patients attending a private laboratory in Sierra Leone. 14 Among 26 healthcare workers with HBV infection, 1 (3.6%) was co-infected with HIV. 8
# Discussion
This review sought to highlight the knowledge and knowledge gaps on the viral hepatitis epidemic and recommend strategies for its prevention and control in Sierra Leone and other LMICs. While, at the time of writing, there were no reported studies on the prevalence of HCV among healthcare workers in Sierra Leone, the prevalence of HBV is comparatively higher (8.7%) than reported in Tanzania (5.7%)and Brazil (0.8%).Due to their occupational engagement, screening for HBV as well as routine vaccination of uninfected persons is recommended for healthcare workers.However, HBV vaccination coverage among healthcare workers in Sierra Leone is low, with only 14.3% having received at least one dose of HBV vaccine, and serologic evidence of past immunization is present in only 4.5% of this population.Unlike the very low immunization coverage observed among healthcare workers in Sierra Leone, 56.9% of healthcare workers in Tanzania had at least one dose of hepatitis B vaccine and 33.3% were fully vaccinated against hepatitis B. 7 The low immunization coverage could be due to a lack of government policy on the provision of routine immunization services for healthcare workers in Sierra Leone, although it could also result from the limited knowledge related to ineffective behavioural change campaigns targeting chronic viral hepatitis. Evidently, a recent knowledge assessment of healthcare workers regarding viral hepatitis in Sierra Leone showed a significant knowledge gap, with only 38.9%, 34.6% and 6.6% having adequate knowledge on clinical outcomes of HBV, its routes of transmission and correct preventive measures, respectively.In the event of a healthcare worker's exposure to bodily fluids from an HBV-infected patient, booster vaccination and HBV immunoglobulins are recommended to prevent transmission.Even though Sierra Leone developed a guideline on post-exposure prophylaxis for HIV and blood-transmitted viral hepatitis among healthcare workers in 2018, routine post-exposure prophylaxis against HBV is currently not implemented. Furthermore, local evidence on HBV post-exposure prophylaxis among healthcare workers is limited, as there are no published data on the status of viral hepatitis post-exposure prophylaxis among this population in Sierra Leone. This situation calls for more attention to the prevention and control of viral hepatitis among healthcare workers.
Because blood donation is a common route for their transmission, the Ministry of Health and Sanitation of Sierra Leone, through its National Blood Service Program, has been screening all blood donors for HIV, HBV, HCV and syphilis since 2005. This subregional variation in the prevalence of HBV could be attributed to the varying interventions in the prevention and control of the viral hepatitis epidemic or perhaps due to the behaviour of the studied population and methods applied for the detection of viral hepatitis. That notwithstanding, the significant burden of viral hepatitis demonstrated among blood donors in Sierra Leone has serious public health implications for the health and well-being of the population.
A disparity in the prevalence of HBV and HCV was observed between rural and urban residents and for gender. For example, the prevalence in one of the studies was high among rural blood donors (11% vs 7% for HBV and 1% vs 0.3% for HCV).This finding is congruent with the higher burden of HBV and HCV reported in a study on blood donors in a rural hospital in northern Sierra Leone, where an HBV prevalence of 13% in women and 15% in men and an HCV prevalence of 7% in women and 8% in men were reported.However, the prevalence of HCV RNA-positive donors has not been established in any of these studies.
In all the studies, male blood donors (20.5%, 20.6% and 15.5%) were disproportionately more infected than female blood donors (7.4%, 9.1% and 13.6%).Similar disparities were observed between hospitals, with an emergency surgical centre reporting 33.2% and a military hospital reporting 4.7%.The lower prevalence of HBV in the military hospital could be due to routine testing of recruits for HBV infection and provision of treatment to HBV-positive persons entering the military. The lower prevalence could also be explained by preventive interventions for sexually transmitted infections in the military.
Despite the 100% coverage of screening services for transfusion-transmissible infections in Sierra Leone, the limited treatment services in the country pose a serious challenge to the prevention and control of viral hepatitis. More efforts are needed to establish links between infected patients and treatment and support services for blood donors with positive HBV and HCV results. Furthermore, concern about the increasing burden of occult HBV in donated blood 29 that is not usually detected by the currently used rapid screening assays warrants research on the prevalence of occult HBV among blood donors in Sierra Leone. Evidence from such research may inform the use of more sensitive nucleic acid assays for screening blood donors in the country.
Like the burden of HBV among blood donors, the prevalence of HBV in patients attending health facilities was high: 21.7% in a private laboratory in Freetown,.5% among febrile patients in a secondary hospital in southern Sierra Leone 15 and 21.4% in patients attending the southern regional tertiary hospital in Bo.As a regional referral hospital, the southern regional hospital is expected to serve a population with higher infection rates, and this may explain the relatively higher prevalence of HBV in this facility. The HCV and HBV prevalences in samples of febrile jaundice were 34.38% and 1.04%, respectively.However, HCV antibodies were detected in 0.22% of samples of suspected Ebola cases during the 2014-2016 Ebola epidemic in Sierra Leone.This disparity in the prevalence of HBV and HCV among the various population groups indicates the need for a more structured study that will provide a solid definition of the prevalence of HBV and HCV among the general population of Sierra Leone.
## Transactions of the royal society of tropical medicine and hygiene
Two published articles reported HBV prevalence rates of 6.2% and 11.3% among pregnant women in Sierra Leone, 19,20 similar to rates in pregnant women in South Sudan (31/280 [11%]), Gambia (39/424 [9.2%]) and Uganda (47/397 [11.8%]).No data are available in Sierra Leone on HCV prevalence among pregnant women, indicating a significant knowledge gap.
To prevent HBV MTCT, routine screening and appropriate immunisation and treatment are recommended for all pregnant women.Prevention of MTCT of HBV requires the use of immunoglobulins, the birth dose of HBV vaccine and infant HBV vaccination.Some professional guidelines recommend the use of nucleoside/nucleotide analogues as additional tools in the prevention of HBV MTCT, particularly if the HBV viral load is high.However, the World Health Organization (WHO) guidelines do not establish any formal recommendation on the use of antiviral medication for the prevention of MTCT, as there was no consensus on widespread antiviral use in the pregnant population and the lack of conclusive evidence at the time of guideline development.Nonetheless, interventions tackling prevention of MTCT of HBV in sub-Saharan Africa are crucial to achieve elimination targets. Low-resource countries should consider implementation of the birth dose of the HBV vaccine with or without immunoglobulins as recommended in the WHO and other guidelines for HBV-exposed infants within 24 h of birth.Although Sierra Leone is among the countries that provides HBV vaccine to infants as part of its Expanded Program on Immunization (EPI), there are still gaps in the prevention of MTCT of HBV. Most pregnant women in Sierra Leone are not screened for viral hepatitis and infants are not provided with an HBV birth dose, which would be an important cornerstone to reduce future transmission. This situation may largely be due to the lack of policies on the prevention of MTCT of viral hepatitis in the country.
There are evident knowledge gaps regarding viral hepatitis in the community. Most published studies on HBV and HCV prevalence were carried out in health facilities, i.e. either in risk groups or inpatient populations, leading to a patchy and epidemiologically non-representative and distorted picture.
Hodges et al.demonstrated an overall prevalence of 16.7% for HBV and 2% for HCV among school children in Freetown between the ages of 6 and 12 y in 1998, long before the introduction of HBV vaccine in the EPI program in 2007. The prevalence of HBV in this population was 13.7% in males and 4.5% in females. Compared with this high prevalence of hepatitis among children in Sierra Leone, a lower prevalence was observed among children in Nigeria (9/749 [1.2%]), Senegal (3/295 [1.1%]) and South Africa (2/450 [0.45%]).In contrast to the study on HBV prevalence among children in Sierra Leone, studies in other countries were carried out long after incorporation of the EPI services in those countries. This could explain the relatively higher burden of viral hepatitis demonstrated among children in Sierra Leone. However, prospectively collected up-to-date data are needed to understand the current state of viral hepatitis among this population in Sierra Leone.
Although a high prevalence of chronic HBV and HCV infections among sex workers and other high-risk groups has been demonstrated in other LMICs,data are currently lacking for these subpopulations in Sierra Leone.
In contrast, a high prevalence of HBV (21.7%) and HCV (4.3%) was found in a cohort of 211 PLHIV.The prevalence of HIV-HBV co-infection was 6.6% (13/138) among patients attending a private laboratory in Sierra Leone. 14 Among 26 healthcare workers with HBV infection in Sierra Leone, 1 (3.6%) was co-infected with HIV.A systematic review of HCV prevalence among HIV co-infected individuals in four countries in sub-Saharan Africa reported an HCV prevalence of 0.4% a systematic review reported that among 8162 participants in Ghana, there was a pooled prevalence of 13.6% for HBV-HIV co-infection.As HBV co-infection in PLHIV is associated with decreased survival,screening and appropriate vaccination of HBV-negative PLHIV is required for improving survival in HIV-positive populations.
In the 2016-2021 global health sector strategy on viral hepatitis, the WHO highlighted strategies for the detection, management and prevention of viral hepatitis to reduce chronic viral hepatitis by 90% from the 2015 baseline and to reduce viral hepatitisrelated mortality from 1.34 million annual deaths in 2015 to fewer than 500 000 deaths by 2030.The achievement of this elimination target requires a mix of preventive and treatment strategies, including prevention of MTCT, implementing universal HBV vaccination, early detection of viral hepatitis and linkage to care and support services.Catch-up HBV vaccination for the general adult population is not routinely recommended since <5% of infections acquired in adulthood progress to chronicity. However, detection and vaccination of high-risk groups like healthcare workers, people who inject drugs, men who have sex with men, commercial sex workers, seronegative partners of persons with HBV, immunosuppressed persons and people with chronic liver disease are recommended as part of the HBV elimination strategy.Detection of HBV and appropriate vaccination is also recommended for patients on haemodialysis, persons in correctional centres, patients requiring immunosuppressives or chemotherapy and persons with multiple sexual partners or sexually transmitted infections.It is therefore prudent for hepatitis prevention programs to focus on the detection and vaccination of these risk groups. Strategies that enhance community-based screening and treatment of chronic HBV infection may provide a costeffective solution.
HBV and HCV programs should also explore the contribution of traditional practices such as female genital mutilation to the growing burden of chronic hepatitis in LMICs, including Sierra Leone, where 90.9% of women 15-49 y of age have been genitally mutilated. 49
# Conclusions
Despite the lack of nationally representative data on HBV and HCV prevalence, a substantial burden of has been demonstrated in some population groups in Sierra Leone, although most of the studies are limited in extent and confined to healthcare facilities.
There is still a huge knowledge gap and high disparity in the burden of viral hepatitis among the various subpopulations in Sierra Leone. Prospective, nationwide, representative surveys are needed to understand the true burden of viral hepatitis in Sierra Leone and other LMICs. Efforts should also be directed towards understanding the prevalence of HBV and HCV among sex workers, men who have sex with men, people who inject drugs, genitally mutilated women, inmates of correctional services and patients on haemodialysis.
To achieve the 2030 Sustainable Development Goal elimination target for viral hepatitis, there is a need to establish sound elimination strategies. Pregnant women should be screened for HBV in their first trimester or the earliest antenatal contact and, where necessary, provided appropriate immunization services. PLHIV should be routinely screened for HBV and HCV at the time of HIV diagnosis and patients with a negative HBV result should be offered HBV vaccination. As with HIV, national programs on viral hepatitis should consider the implementation of a test-andtreat approach for all patients with HCV and implement universal screening and treatment programs for HBV infection.
Overall, Sierra Leone and other LMICs should develop and implement a policy framework that will aid mobilization of resources for universal access to detection using point-of-care tests and appropriate treatment of patients infected with viral hepatitis. Efforts to improve the traditional preventive measures, such as the use of barrier methods, water-based lubricants and safe handling of sharps, especially among high-risk groups, should be strengthened across LMICs, including Sierra Leone. |
Nanographenes as electron-deficient cores of donor-acceptor systems
In the manuscript the authors describe regioselective amination of perchlorinated large polyaromatic hydrocarbons (nanographenes, NGs) by palladium-catalyzed Buchwald-Hartwig coupling, which selectively occurred at the vertexes of NGs. The authors well chemically and structurally characterized the obtained derivatives and probed their electronic properties. They present a suitable protocol for modulating electron-donating properties of the functionalized NGs. The manuscript is clearly and well written. However, the modulation of electronic properties of NGs by functionalization was already published and therefore the reviewer feels the manuscript is more suitable for specialized chemical journal.Specific points: Figure 1c -x-axis should be labeled, solvent should be specified.It is hard to agree with the claims from DFT calculations aimed at providing insight into intramolecular CT (pg. 6). Analysis based on calculations of ground/excited states and electron density differences maps would provide more insightful information. In addition, an order of orbitals strongly depends on the choice of DFT functional and various functionals may provide different frontier orbitals. How was this issue handled by the authors? Figure 2 -How was the "Absorption" defined? Details about acquiring solid state UV/Vis should be provided at least in ESI. Reviewer #2 (Remarks to the Author):This manuscript has presented synthesis of a series of Donor-Acceptor molecular architectures by amination of perchlorinated nanographenes with anilines, and structures and properties of these molecules, as well as the complexes of these molecules with TTF exhibiting intermolecular charge transfer. These D-A molecular architectures are interesting functional π-molecules with potential applications as organic electronic materials and functional dyes. However, two key findings as claimed in this manuscript have not been supported with solid experimental evidences.1) The reactions of perchlorinated nanographenes 1, 3 and 5 with anilines by palladium-catalyzed Buchwald-Hartwig are claimed as regioselective, and only the products that have the vertex chlorine atoms substituted by nitrogen were reported. However, the yields of compounds 2a-e are 26-32%, the yields of 4a-c are 13-20%, and the yields of 6a-b are as low as 8-19%. What are the other products (68% to 92%) of these reactions? Did amination always occur on the vertex chlorines?2) The structures of compounds 2a, 2c, 2d, 2e, 4a and 4c were unambiguously characterized by X-ray crystallography. It is reasonable to establish the structures of 2b and 4b on the basis of these crystal structures and comparison of 1H NMR spectra of 2b and 4b with those of 2a, 2c, 2d, 2e, 4a and 4c.
reactivity difference of Cl at the vertex and bay sites, they were able to replace all vertex Cl in reasonable yield. They have also shown that the perchlorinated HBCs have an electron-deficient concave core, which could form charge-transfer supramolecular structures with electron-rich donor like TTF. This review will primarily focus on the chemistry part of the work. The manuscript can be accepted for publication, provided that the evaluation on the charge transfer part of the work is deemed significant.
The regioselectivity demonstrated in this work is quite interesting. It takes advantage of the difference in the steric environment to modulate the reactivity of Cl at the vertex and bay sites in the C-N cross-coupling reaction. The identified regents and reaction conditions allowed them to completely replaced 6 vertex chlorine with aniline. This work will be of interest to synthetic organic chemists and materials scientists working on carbon materials.
How general is this method? The 3 examples have the same number of "layers", which gave 6 vertex sites for substitution. What are the challenges when the structures are larger?
In the rest of the ~70% product, what is the distribution of the various substituted products?
The charge transfer complex was done on the H or alkyl substituted aniline. How would other substituents on aniline (eg. OCH3) impact the complex formation and their properties?
Several different terminologies were used to describe their materials, for example, HBC, nanographene, larger graphene, graphene nanoribbons. One should be very careful choosing the correct description of these materials, as the term "graphene" refers to a very special type of material having unique physical properties.
The reaction conditions in [fig_ref] Figure 1: Regioselective thiolation at the vertexes [/fig_ref] are missing the critical aniline compound.
The quality of the spectra could be improved. Several spectra have solvent peaks, silicone grease in addition to water peaks.
Compound labels are missing in several figures. For example, compound 1 in [fig_ref] Figure 1: Regioselective thiolation at the vertexes [/fig_ref] , compound 3 in , compound 5 in .
In the manuscript the authors describe regioselective amination of perchlorinated large polyaromatic hydrocarbons (nanographenes, NGs) by palladium-catalyzed Buchwald-Hartwig coupling, which selectively occurred at the vertexes of NGs. The authors well chemically and structurally characterized the obtained derivatives and probed their electronic properties. They present a suitable protocol for modulating electron-donating properties of the functionalized NGs. The manuscript is clearly and well written. However, the modulation of electronic properties of NGs by functionalization was already published and therefore the reviewer feels the manuscript is more suitable for specialized chemical journal.
## Response:
We appreciate the reviewer for the valuable comments and suggestions. Our work presented the first case of donor-acceptor (D-A) system built by nanographene (NG) acceptor, which offered an alternative and new way to construct D-A conjugates using NG moiety. Moreover, beyond C-N coupling, the concept of the regioselective modification at the vertex of NGs can be extended, which furnishes a series of functional NGs with electron-accepting cores.
[redacted]
Detailed studies on the selective C-S coupling at the vertexes and concave-convex supramolecular assembly are under way and we would like to report these in due course.
These points and the wider impact of this chemical modification render our work significantly different from the reported studies on "the modulation of electronic properties of NGs by functionalization". Even more so since we describe not only the molecular, but also the supramolecular consequences.
Specific points: [fig_ref] Figure 1: Regioselective thiolation at the vertexes [/fig_ref] -x-axis should be labeled, solvent should be specified.
## Response:
Thanks for the comments of reviewer. The x-axis was labeled in [fig_ref] Figure 1: Regioselective thiolation at the vertexes [/fig_ref] and the solvent was specified.
It is hard to agree with the claims from DFT calculations aimed at providing insight into intramolecular CT (pg. 6). Analysis based on calculations of ground/excited states and electron density differences maps would provide more insightful information. In addition, an order of orbitals strongly depends on the choice of DFT functional and various functionals may provide different frontier orbitals. How was this issue handled by the authors?
Response: The excitation states for compounds 2, 4 and 6 were calculated by timedependent density-functional theory (TD-DFT). First, in order to ensure the suitable hybrid functional, we compared the theoretical HOMO-LUMO gap (G) of 2c, 4b, and 6a (all of them were decorated with 4-n-butyl-anilino groups at the vertexes) calculated by B3LYP, CAM-B3LYP, M062X and HSEH1PBE with the experimental optical HOMO-LUMO gap (G opt, exp. ) [fig_ref] Table 1: Comparison of theoretically calculated G and G opt, exp [/fig_ref]. The optimized structure by different functionals was also compared with the experimental crystallographic data, taking 2a as an example .
Accordingly, the HSEH1PBE functional is more appropriate for the theoretical calculations. . Numbering of the carbon atoms of 2a Comparison of the bond lengths (BL) between experimental data (crystal structure) and theoretically optimized structure of 2a and relative deviations (RD).
[formula] C 1 -C 1' (Å) C 1 -C 2 (Å) C 2 -C 3 (Å) C 3 -C 3 ' (Å) C 3 -C 4 (Å) C 4 -C 5 (Å) RD (%) [ [/formula]
## Response:
The "absorption" in the figure 2 was measured in a diffuse-reflectance mode on a UV-Vis-NIR spectrometer (Cary 5000 spectrometer).
The samples for solid state absorption were prepared by mixing and grinding the crystals of NGs and BaSO 4 powder. Using BaSO 4 powder as the blank, the absorption spectra of the sample were obtained automatically by the spectrometer.
The absorption was defined in the caption of [fig_ref] Figure 2 -: How was the "Absorption" defined? Details about acquiring solid state UV/Vis should... [/fig_ref] in the text and the details on acquiring solid state UV-Vis spectra were provided in Supplementary Information.
## Reviewer #2 (remarks to the author):
This manuscript has presented synthesis of a series of Donor-Acceptor molecular architectures by amination of perchlorinated nanographenes with anilines, and structures and properties of these molecules, as well as the complexes of these molecules with TTF exhibiting intermolecular charge transfer. These D-A molecular architectures are interesting functional π-molecules with potential applications as organic electronic materials and functional dyes. However, two key findings as claimed in this manuscript have not been supported with solid experimental evidences.
1) The reactions of perchlorinated nanographenes 1, 3 and 5 with anilines by palladiumcatalyzed Buchwald-Hartwig are claimed as regioselective, and only the products that have the vertex chlorine atoms substituted by nitrogen were reported. However, the yields of compounds 2a-e are 26-32%, the yields of 4a-c are 13-20%, and the yields of 6a-b are as low as 8-19%. What are the other products (68% to 92%) of these reactions? Did amination always occur on the vertex chlorines? Response: We appreciate the reviewer for the valuable comments. We carefully investigated the yield, byproducts and selectivity of the C-N coupling reaction between perchlorinated NGs and anilines, taking 2a as a typical case. Therefore, we tried to measure the yield of 2a by NMR. The detailed procedures were listed below.
100 mg compound 1 reacted with aniline under the reported conditions. After the reaction, the mixture was extracted by dichloromethane (DCM). Then the organic phase was washed with water, dried by MgSO 4 and filtered over silica (100-200 mesh). After evaporation of the solvent, the products were dissolved by 4 ml DCM. Then 20ml methanol was added to precipitate the products. The orange precipitate was collected as the crude products (94 mg) by filtration.
Note that the yield based on the amount of orange crude products was about 70% (the theoretical yield is 130 mg), which meant 30% mass was lost after pretreatments.
Because palladium can catalyze the C-C homocoupling of aryl halides under similar conditions as well , we speculated that C-C homocoupling between 1 acted as a competitive reaction for the C-N coupling between 1 and aniline. The C-C homocoupling between 1 would lead to the oligomerization and produce less soluble or insoluble oligomers, which were easy to be removed by pretreatments, such as extraction and filtration. This speculation was supported by the decreased yield of 2a (15% and 18%) as well as the amount of orange crude products after pretreatments (65 and 80 mg) when the amount of aniline (1 and 1.5 fold of aniline according to the molar of chlorines at the vertexes of 1) was reduced, in comparison with our reported conditions (4 fold of aniline) (See [fig_ref] Table 1: Comparison of theoretically calculated G and G opt, exp [/fig_ref] in Supplementary Information).
TLC analysis of orange crude products (spot 1 in TLC, [fig_ref] Figure 8: TLC analysis of crude products [/fig_ref] showed one dominant spot, indicating 2a as the major product. We used NMR spectroscopy to determine the content of 2a in the orange crude products directly, instead of chromatographic separation.
The mass content of 2a in the crude products by NMR is ~70%, which corresponded to an overall yield of ~50%. Then the crude products were subjected to chromatographic separation as reported, yielding 39 mg of pure 2a, which shows the similar yield (30%).
The less pure components were collected and purified by chromatography again, which afforded additional 10 mg 2a and 15 mg byproduct (spot 4 shown in TLC, [fig_ref] Figure 8: TLC analysis of crude products [/fig_ref].
Therefore, we can obtain 49 mg 2a in total and the isolated yield can be improved to 38%.
The lower isolatable yield of 2a than that determined by NMR was attributed to the strong absorption on silica.
However, considering that six C-N bonds are formed in our reactions, the reported yields (32%-8%) required 83% to 65% yields for each C-N bond formation, that is acceptable for the C-N coupling for aryl chloride. If we considered the yield of 2a measured by NMR (~50%), the yield for each C-N bond formation is about 89%.
According to the comment "Did amination always occur on the vertex chlorines?", the byproduct was analyzed by mass spectroscopy [fig_ref] Figure 9: Mass spectrum of byproduct of 2a [/fig_ref] and assigned as pentakisanilino products [C 42 Cl 13 (NHC 6 H 5 ) 5 and C 42 Cl 12 (NHC 6 H 5 ) 5 O]. Products with more than 6 anilino groups were not observed. and 30 fold excess of aniline did not affect the yield (30%-31%) [fig_ref] Table 1: Comparison of theoretically calculated G and G opt, exp [/fig_ref]. The extreme excess of aniline did not affect the yield of the reaction, which validated the inactivity of chlorines in the bay.
In short, we can identify ~70% of the products as the remaining absorbed 2a (~10-20%), byproducts with less than 6 anilino groups (~10-15%), and some oligomers from the C-C homocoupling (~30%). Taking the isolatable byproduct and yields of control experiments together, the regioselective reactivity of chlorines at the vertexes was confirmed.
The related discussions on the yield, byproducts and control experiments were added in the revised Supplementary Information.
2) The structures of compounds 2a, 2c, 2d, 2e, 4a and 4c were unambiguously characterized by X-ray crystallography. It is reasonable to establish the structures of 2b and 4b on the basis of these crystal structures and comparison of 1H NMR spectra of 2b and 4b with those of 2a, 2c, 2d, 2e, 4a and 4c. However, the claimed structures of 6a and 6b lack solid support from experimental results. On Page 5 of the Supplementary Information, the authors noted the difficulties in conducting elemental analysis and mass spectra. However, without molecular formula as determined by elemental analysis or mass spectra, one cannot even know how many chlorine atoms in compound 5 were substituted with anilines. Without appropriate experimental evidences, the structures of 6a and 6b should not be claimed. Response: Thanks for the valuable comments. The complementary experiments on the structure of compound 6 were carried out. 13 C NMR spectrum of 6a was obtained on a 600M Bruker NMR spectrometer equipped with a ultra-sensitive cryo-probe. The 13 C NMR spectrum of 6a was acquired and accumulated for 12 hours. As shown in [fig_ref] Figure 1: Regioselective thiolation at the vertexes [/fig_ref] , there are 30 and 8 13 C peaks in the arene and alkane region, respectively, which confirmed the expected structure of 6a. However, the insufficient solubility of 6b did not allow the 13 C NMR characterization. [fig_ref] Figure 1: Regioselective thiolation at the vertexes [/fig_ref]. 13 C NMR spectrum of 6a in C 2 D 2 Cl 4. 13 C NMR (151 MHz, C 2 D 2 Cl 4 ) δ 139. .20 ppm. The peaks at 28.87 ppm was assigned to the signal of hexane in the solvent, marked by asterisk.
We did great effort on the mass spectroscopic characterizations of these D-A NGs and found that their mass spectra highly depended on the experimental conditions of MALDI-TOF mass spectrometers, even using the same matrix.
The high resolution mass spectra of 2, 4 and 6 acquired by AB SCIEX 5800 MALDI-TOF mass spectrometer (AB Sciex Pte. Ltd. USA) normally showed a lot of unassignable fragments, although the molecular ion peak can be found [fig_ref] Figure 13: High resolution mass spectra of 6 acquired on a AB SCIEX 5800... [/fig_ref]. However, 2e, 4b, 4c and 6b could show clearer high resolution mass spectra [fig_ref] Figure 13: High resolution mass spectra of 6 acquired on a AB SCIEX 5800... [/fig_ref]. [fig_ref] Figure 1: Regioselective thiolation at the vertexes [/fig_ref]. High resolution mass spectra of 2 acquired on a AB SCIEX 5800 MALDI-TOF mass spectrometer using TCNQ as the matrix in cation mode. The experimental and calculated spectra were represented in red and blue, respectively. This time, we tried a Bruker microflex LRF MALDI-TOF mass spectrometer (Bruker Corporation), which had lower resolution, to measure the mass spectra of these compounds. We can obtain clear mass spectra for these compounds [fig_ref] Figure 14: Mass spectra of 2 acquired on a Bruker microflex LRF MALDI-TOF mass... [/fig_ref]. The byproduct of 2a was also analyzed by a Bruker microflex LRF MALDI-TOF mass spectrometer [fig_ref] Figure 9: Mass spectrum of byproduct of 2a [/fig_ref]. Though the resolution of the spectra acquired by Bruker microflex LRF MALDI-TOF is lower, the structural formulae of 2, 4 and 6 can be assigned. As a supplement, we also tried other softer ionization sources for mass spectroscopy to avoid fragmentation and found that electrospray ionization (ESI) worked for compounds 2a-2d and 4a-4c [fig_ref] Figure 8: TLC analysis of crude products [/fig_ref] In the revised manuscript, we added the mass spectra of 2a-2d and 4a-4c measured by ESI and the mass spectra of 2, 4 and 6 measured by the Bruker microflex LRF MALDI-TOF mass spectrometer in the Supplementary Information.
Since the mass spectra of 6a and 6b, 1 H NMR spectra of 6a and 6b and 13 C NMR spectrum of 6a matched the expected structures, we can validate the structure of 6 as was shown in the manuscript.
In order to improve our manuscript, all the 13 C NMR spectra of D-A NGs (except 6b)
were provided in the Supplementary Information.
Moreover, some typos and grammar mistakes are found in the manuscript, and the writing should be improved. For example, in Line 34 (Page 2), "a few of" should be "a few", and in Line 78 (Page 5), "substituted" should be "substituting". And the exact meaning of the sentence "The unobstructed chlorines in the bays are staggered" in Line 79 (Page 5) is not clear.
## Response:
According to the reviewer's suggestion, we have corrected the wording. For clarity, the sentence "The unobstructed chlorines in the bays are staggered" in Line 79 (Page 5) is corrected into "The remaining chlorines in the bays adopt an up and down conformation".
Therefore, this referee does not recommend this manuscript in the current form to be published in Nature Communications. But a revised manuscript with the above issues addressed appropriately can be reconsidered.
Reviewer #3 (Remarks to the Author):
The authors had previously developed a method to perchlorinate HBCs. In this work, they showed regioselective amination at the vertexes of the perchlorinated HBCs. Taking advantage of the reactivity difference of Cl at the vertex and bay sites, they were able to replace all vertex Cl in reasonable yield. They have also shown that the perchlorinated HBCs have an electron-deficient concave core, which could form charge-transfer supramolecular structures with electron-rich donor like TTF. This review will primarily focus on the chemistry part of the work. The manuscript can be accepted for publication, provided that the evaluation on the charge transfer part of the work is deemed significant.
The regioselectivity demonstrated in this work is quite interesting. It takes advantage of the difference in the steric environment to modulate the reactivity of Cl at the vertex and bay sites in the C-N cross-coupling reaction. The identified regents and reaction conditions allowed them to completely replaced 6 vertex chlorine with aniline. This work will be of interest to synthetic organic chemists and materials scientists working on carbon materials. We tried the selective C-N coupling with aniline even for the larger NGs, such as C 96 Cl 27 H 3 , which contained 9 chlorine atoms at the vertexes. We found that it could react with aniline but did not afford a well-defined product at present. The largest challenge when the structures become larger is that both high reactivity and selectivity are required by the multiple functionalization whereby the larger NGs are normally less reactive and less soluble. Taking our reactions as an example, considering that six C-N bonds are formed in our reactions, the reported overall yields here (32%-8%) required 83% to 65%
yields for each C-N bond formation. Another technical challenge for the larger NGs is the difficulty of purification, originating from their decreased solubility and increased irreversible absorption.
In the rest of the ~70% product, what is the distribution of the various substituted products? Response: We appreciate the reviewer for the valuable comments. We carefully investigated the yield, byproducts and selectivity of the C-N coupling reaction between perchlorinated NGs and anilines, taking 2a as a typical case.
These D-A NGs had a strong absorption on silica, which resulted in the tailing and absorption during the chromatographic separation by silica column. As shown in thin layer chromatography (TLC) analysis, besides the major component (2 spot 2 shown in TLC) which we collected as 2a, the subsequent eluent (spot 3 shown in TLC) contained 2a as well [fig_ref] Figure 8: TLC analysis of crude products [/fig_ref]. The silica column also absorbed a lot of products after separation [fig_ref] Figure 8: TLC analysis of crude products [/fig_ref]. The yield of 2a (31%) was calculated using the mass of pure component and the less pure components were neglected [fig_ref] Figure 8: TLC analysis of crude products [/fig_ref]. methanol was added to precipitate the products. The orange precipitate was collected as the crude products (94 mg) by filtration.
Note that the yield based on the amount of orange crude products was about 70% (the theoretical yield is 130 mg), which meant 30% mass was lost after pretreatments.
Because palladium can catalyze the C-C homocoupling of aryl halides under similar conditions as well , we speculated that C-C homocoupling between 1 acted as a competitive reaction for the C-N coupling between 1 and aniline. The C-C homocoupling between 1 would lead to the oligomerization and produce less soluble or insoluble oligomers, which were easy to be removed by pretreatments, such as extraction and filtration. This speculation was supported by the decreased yield of 2a (15% and 18%) as well as the amount of orange crude products after pretreatments (65 and 80 mg) when the amount of aniline (1 and 1.5 fold of aniline according to the molar of chlorines at the vertexes of 1) was reduced, in comparison with our reported conditions (4 fold of aniline) (See [fig_ref] Table 1: Comparison of theoretically calculated G and G opt, exp [/fig_ref] in Supplementary Information).
TLC analysis of orange crude products (spot 1 in TLC, [fig_ref] Figure 8: TLC analysis of crude products [/fig_ref] showed one dominant spot, indicating 2a as the major product. We used NMR spectroscopy to determine the content of 2a in the orange crude products directly, instead of chromatographic separation.
The mass content of 2a in the crude products by NMR is ~70%, which corresponded to an overall yield of ~50%. Then the crude products were subjected to chromatographic separation as reported, yielding 39 mg of pure 2a, which shows the similar yield (30%).
The less pure components were collected and purified by chromatography again, which afforded additional 10 mg 2a and 15 mg byproduct (spot 4 shown in TLC, [fig_ref] Figure 8: TLC analysis of crude products [/fig_ref].
Therefore, we can obtain 49 mg 2a in total and the isolated yield can be improved to 38%.
The lower isolatable yield of 2a than that determined by NMR was attributed to the strong absorption on silica.
According to the comment "what is the distribution of the various substituted products?", the byproduct was analyzed by mass spectroscopy [fig_ref] Figure 9: Mass spectrum of byproduct of 2a [/fig_ref] and assigned as pentakisanilino products Cl 13 (NHC 6 H 5 ) 5 and C 42 Cl 12 (NHC 6 H 5 ) 5 O]. Products with more than 6 anilino groups were not observed. and 30 fold excess of aniline did not affect the yield (30%-31%) [fig_ref] Table 1: Comparison of theoretically calculated G and G opt, exp [/fig_ref]. The extreme excess of aniline did not affect the yield of the reaction, which validated the inactivity of chlorines in the bay.
In short, we can identify ~70% of the products as the remaining absorbed 2a (~10-20%), byproducts with less than 6 anilino groups (~10-15%), and some oligomers from the C-C homocoupling (~30%). Taking the isolatable byproduct and yields of control experiments together, the regioselective reactivity of chlorines at the vertexes was confirmed.
The related discussions on the yield, byproducts and control experiments were added in the revised Supplementary Information.
The charge transfer complex was done on the H or alkyl substituted aniline. How would other substituents on aniline (eg. OCH3) impact the complex formation and their properties?
## Response:
The charge transfer complex of hexakis-4-methoxyl-anilino chlorinated C 60 (4c) and tetrathiafulvalene (TTF) was obtained and characterized by X-ray diffraction [fig_ref] Figure 9: Mass spectrum of byproduct of 2a [/fig_ref]. The supramolecular complex exhibited a charge transfer peak at 770 nm [fig_ref] Figure 9: Mass spectrum of byproduct of 2a [/fig_ref] , which is similar to that of supramolecular complex between 4a and TTF. [fig_ref] Figure 9: Mass spectrum of byproduct of 2a [/fig_ref]. Structure of 2·4c⊃TTF and UV-Vis-NIR spectra of 4c and 2·4c⊃TTF in the solid state.
The hexakis-4-methoxyl-anilino chlorinated hexa-peri-hexabenzocoronene (2d) cannot form the supramolecular complex with TTF by slow solvent evaporation. The synthesis for the supramolecular complex of 2d and TTF remains to be explored. The hexakis-4methoxyl-anilino chlorinated C 78 (6b) can form the complex with TTF, but the structure must be confirmed due to the small size of the crystals (~10 µm) of the complex.
The solid state UV-Vis-NIR spectrum of the complex 2·4c⊃TTF was added in the Supplementary Information.
Several different terminologies were used to describe their materials, for example, HBC, nanographene, larger graphene, graphene nanoribbons. One should be very careful choosing the correct description of these materials, as the term "graphene" refers to a very special type of material having unique physical properties. Response: We agree with the comments of reviewer. These terminologies were carefully used in the revision, especially for the term "graphene".
The reaction conditions in [fig_ref] Figure 1: Regioselective thiolation at the vertexes [/fig_ref] are missing the critical aniline compound.
1. The authors did try to isolate and analyze the by-products. But they failed to demonstrate that the byproducts with less than 6 anilino groups have only chlorine atoms at the vertex positions substituted by nitrogen atoms. Without evidences to support substitution of vertex chlorine atoms, the palladium-catalyzed Buchwald-Hartwig reaction cannot be claimed as regioselective. Response: We appreciate the reviewer for the valuable comments. We tried to get unambiguous structure of byproducts by single-crystal X-ray diffraction, however, a series of trials to grow the single crystals of byproducts, such as vapor evaporation, vapor diffusion and solvent diffusion, were failed. The difficulty of crystal growth for the byproduct we got was ascribed to its low purity. As shown in its mass spectrum, there are two different molecules found. MALDI-TOF mass spectrometer using tetracyanoquinodimethane as the matrix in cation mode. The experimental and calculated spectra are represented in red, blue and purple, respectively. Obviously, two molecules can be assigned (M 1 and M 2 ) in the spectrum.
Therefore, we can not unambiguously demonstrate the substitution pattern of the byproducts and withdraw the claim on the regioselectivity of palladium-catalyzed Buchwald-Hartwig reaction as the reviewer suggested. In the revised manuscript, we deleted all the claims on regioselectivity of the reaction and only described the isolated and characterized compounds 2 and 4 showing a selective amination at the vertexes.
2. The extra experiments did not provide sufficient evidences to support the structures of structures of 6a and 6b. They authors may report the palladium-catalyzed reaction of 5, but shall not claim the structures of 6a and. Response: Thanks for the reviewer's the valuable comments. The single crystals of 6b
can be obtained but it preferred to form fiber-shaped crystals with a diameter around 10 μm, which is too small for the X-ray diffraction even using the synchrotron X-ray beam.
The present experimental evidences of 6a and 6b based on mass spectra and NMR spectra can conclude the chemical formula as C 78 Cl 20 (NHC 6 H 4 R) 6 (R = C 4 H 9 for 6a and OCH 3 for 6b). Therefore, we follow the suggestion of reviewer to remove detailed substitution pattern of 6 and report that the palladium-catalyzed reaction of 5 afforded the hexakis-aminated chlorinated products 6a-6b C 78 Cl 20 (NHC 6 H 4 R) 6 (R = C 4 H 9 for 6a and OCH 3 for 6b) in the revision.
Therefore, this referee does not recommend this manuscript in the current form to be published in Nature Communications. But a revised manuscript with the above issues addressed appropriately can be reconsidered.
[fig] Figure 4, Figure 5, Figure 6, Figure 7, Figure 7: The electron density differences between the first excitation state and the ground state of 2 (a, 2a, b, 2b, c, 2c, d, 2d and e, 2e). (Blue and red refer to a decrease and an increase in electron density, respectively) The electron density differences between the first excitation state and the ground state of 4 (a, 4a, b, 4b and c, 4c). (Blue and red refer to a decrease and an increase in electron density, respectively) The electron density differences between the first excitation state and the ground state of 6 (a, 6a and b, 6b). (Blue and red refer to a decrease and an increase in electron density, respectively) According to the reviewer's comment on "the choice of DFT functional and various functionals may provide different frontier orbitals.", we compared the electronic structure of 2a calculated by B3LYP, CAM-B3LYP, M062X and HSEH1PBE. Nearly the same frontier molecular orbitals (HOMO-1, HOMO, LUMO and LUMO+1) were revealed by these four hybrid functionals (Selected key frontier molecular orbitals of 2a calculated by four different [/fig]
[fig] Figure 2 -: How was the "Absorption" defined? Details about acquiring solid state UV/Vis should be provided at least in ESI. [/fig]
[fig] Figure 8: TLC analysis of crude products (spot 1), major component collected as 2a (spot 2), subsequent eluent (spot 3), isolatable byproduct (spot 4) and the photo of silica column after separation. The byproduct (spot 4) was obtained by additional chromatographic separation. [/fig]
[fig] Figure 9: Mass spectrum of byproduct of 2a (4 shown in TLC analysis) acquired on a Bruker microflex LRF MALDI-TOF mass spectrometer using tetracyanoquinodimethane (TCNQ) as the matrix in cation mode. The experimental and calculated spectra are represented in red, blue and purple, respectively. Moreover, the reported experimental conditions required a large excess of aniline (4 fold of aniline) to favor the multiple C-N coupling of 1 with aniline and suppress the homocoupling between 1. We performed control experiments with different amounts of aniline (1, 1.5, 4, 15 and 30 fold). We found the decreased amount of aniline resulted in a lower yield of 2a (15% for 1 fold and 18% for 1.5 fold), whereas the experiments with 15 [/fig]
[fig] Figure 12: High resolution mass spectra of 4 acquired on a AB SCIEX 5800 MALDI-TOF mass spectrometer using TCNQ as the matrix in cation mode. The experimental and calculated spectra were represented in red and blue, respectively. [/fig]
[fig] Figure 13: High resolution mass spectra of 6 acquired on a AB SCIEX 5800 MALDI-TOF mass spectrometer using TCNQ as the matrix in cation mode. The experimental and calculated spectra were represented in red and blue, respectively. [/fig]
[fig] Figure 14: Mass spectra of 2 acquired on a Bruker microflex LRF MALDI-TOF mass spectrometer using TCNQ as the matrix in cation mode. The experimental and calculated spectra were represented in red and blue, respectively. [/fig]
[fig] Figure 15: Mass spectra of 4 acquired on a Bruker microflex LRF MALDI-TOF mass spectrometer using TCNQ as matrix in cation mode. The experimental and calculated spectra were represented in red and blue, respectively. [/fig]
[fig] Figure 16: Mass spectra of 6 acquired on a Bruker microflex LRF MALDI-TOF mass spectrometer using TCNQ as matrix in cation mode. The experimental and calculated spectra were represented in red and blue, respectively. [/fig]
[fig] Figure 17: Mass spectra of 2a-2d acquired on a Bruker Esquire HCT mass spectrometer using ESI in anion mode. The experimental and calculated spectra were represented in red and blue, respectively. [/fig]
[fig] Figure 18: Mass spectra of 4a-4c acquired on a Bruker Esquire HCT mass spectrometer using ESI in anion mode. The experimental and calculated spectra were represented in red and blue, respectively. [/fig]
[fig] Figure 1: Regioselective thiolation at the vertexes (a) and crystal structure of hexakisthiophenethiolated chlorinated hexa-peri-hexabenzocoronene (b). DMI = 1,3-dimethyl-2imidazolidinone. [/fig]
[table] Table 1: Comparison of theoretically calculated G and G opt, exp . . [/table]
|
Effects of individualized positive end-expiratory pressure combined with recruitment maneuver on intraoperative ventilation during abdominal surgery: a systematic review and network meta-analysis of randomized controlled trials
Low tidal volume ventilation strategy may lead to atelectasis without proper positive end-expiratory pressure (PEEP) and recruitment maneuver (RM) settings. RM followed by individualized PEEP was a new method to optimize the intraoperative pulmonary function. We conducted a systematic review and network meta-analysis of randomized clinical trials to compare the effects of individualized PEEP + RM on intraoperative pulmonary function and hemodynamic with other PEEP and RM settings. The primary outcomes were intraoperative oxygenation index and dynamic compliance, while the secondary outcomes were intraoperative heart rate and mean arterial pressure. In total, we identified 15 clinical trials containing 36 randomized groups with 3634 participants. Ventilation strategies were divided into eight groups by four PEEP (L: low, M: moderate, H: high, and I: individualized) and two RM (yes or no) settings. The main results showed that IPEEP + RM group was superior to all other groups regarding to both oxygenation index and dynamic compliance. LPEEP group was inferior to LPEEP + RM, MPEEP, MPEEP + RM, and IPEEP + RM in terms of oxygenation index and LPEEP + RM, MPEEP, MPEEP + RM, HPEEP + RM, IPEEP, and IPEEP + RM in terms of dynamic compliance. All comparisons were similar for secondary outcomes. Our analysis suggested that individualized PEEP and RM may be the optimal low tidal volume ventilation strategy at present, while low PEEP without RM is not suggested.Keywords Individualized · Positive end-expiratory pressure · Alveolar recruitment maneuver · Protective lung strategy · Abdominal surgery · Network meta-analysis Abbreviations ARDS Acute respiratory distress syndrome MAP Mean arterial pressure NMA Network meta-analysis PEEP Positive end-expiratory pressure PPC Postoperative pulmonary complication Pr Best Probability of being best RCT Randomized controlled trial RM Recruitment maneuver SUCRA Surface under cumulative ranking curve V T
# Introduction
Pulmonary gas exchange and respiratory mechanics are impaired during general anesthesia, due to the collapse of alveolar [bib_ref] Prevention of atelectasis in morbidly obese patients during general anesthesia and paralysis:..., Reinius [/bib_ref]. Conventional ventilation strategy with high 1 3 tidal volume (V T ) > 10 ml/kg predicted body weight, no positive end-expiratory pressure (PEEP), or routine recruitment maneuver (RM) was once recommended to maintain the alveolar open, but was recently proved ineffective and even related to severe volutrauma and barotrauma [bib_ref] A trial of intraoperative low-tidal-volume ventilation in abdominal surgery, Futier [/bib_ref] [bib_ref] Low-tidal-volume ventilation in the acute respiratory distress syndrome, Malhotra [/bib_ref]. The current view suggests that sufficient PEEP and routine RM based on low V T (6-8 ml/kg predicted body weight) are the key points to maintain the alveolar open [bib_ref] A trial of intraoperative low-tidal-volume ventilation in abdominal surgery, Futier [/bib_ref] [bib_ref] Intraoperative protective mechanical ventilation and risk of postoperative respiratory complications: hospital based..., Ladha [/bib_ref] [bib_ref] Lung stress and strain during mechanical ventilation: any difference between statics and..., Protti [/bib_ref].
However, an excessive high PEEP may lead to barotrauma and hemodynamic instability; thus, the lowest PEEP that keeps the alveoli open is defined as "optimal PEEP". The conclusions of previous studies showed a consistent controversy on the value of optimal PEEP [bib_ref] Effect of intraoperative high positive end-expiratory pressure (PEEP) with recruitment maneuvers vs..., Bluth [/bib_ref] [bib_ref] Intraoperative compliance profiles and regional lung ventilation improve with increasing positive end-expiratory..., Wirth [/bib_ref] [bib_ref] High versus low positive end-expiratory pressure during general anaesthesia for open abdominal..., Hemmes [/bib_ref]. At this background, a new concept named "individualized PEEP" was proposed, aiming at finding out the optimal PEEP according to patients' individual characteristics such as lung dynamic compliance and driving pressure [bib_ref] Individualised perioperative open-lung approach versus standard protective ventilation in abdominal surgery (iPROVE):..., Ferrando [/bib_ref] [bib_ref] Driving pressure during thoracic surgery a randomized clinical trial, Park [/bib_ref].
The superiority of individualized PEEP over fixed PEEP was proved by recent studies in terms of intraoperative oxygenation index and respiratory mechanisms such as driving pressure and dynamic compliance [bib_ref] Individualised perioperative open-lung approach versus standard protective ventilation in abdominal surgery (iPROVE):..., Ferrando [/bib_ref] [bib_ref] Driving pressure during thoracic surgery a randomized clinical trial, Park [/bib_ref] [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref] [bib_ref] Individualized positive end-expiratory pressure in obese patients during general anaesthesia: a randomized..., Nestler [/bib_ref] , indicating less atelectrauma and barotrauma separately. However, the superiority was mostly concluded from the comparisons with moderate fixed PEEP (5-8 cm H 2 O) [bib_ref] Van der Linden P. Does lung compliance optimization through PEEP manipulations reduce..., Van Hecke [/bib_ref] , while meta-analysis involving individualized PEEP is still absent. In addition, hemodynamic instability was another concern in the previous studies as individualized PEEP was usually higher than 10 cm H 2 O [bib_ref] Individualised perioperative open-lung approach versus standard protective ventilation in abdominal surgery (iPROVE):..., Ferrando [/bib_ref] [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref].
We aimed at comprehensively evaluate the effects of individualized PEEP and RM based on low V T ventilation strategy on intraoperative pulmonary function and hemodynamics during abdominal surgery, in comparisons of other PEEP and RM settings. Since PEEP (low, moderate, high, and individualized) and RM (yes or no) has multiple levels, the conventional pairwise comparison meta-analysis is difficult to achieve our purpose. We finally performed this network meta-analysis (NMA) and systematic review of randomized controlled trials (RCTs) to provide a strong evidence for the benefits of individualized PEEP.
# Materials and methods
The protocol for this NMA was registered with PROSPERO prospectively (identifier: CRD42020170614). The findings of this NMA was reported in accordance with the preferred reporting items for systematic reviews and meta-analyses-network meta-analyses (PRISMA-NMA) guidelines [bib_ref] Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, Moher [/bib_ref]. The PRISMA-NMA checklist can be found in Supplemental Digital Content 1.
## Selection criteria
Two authors (ZLN and HLG) independently assessed the eligibility of studies by reading the titles, abstracts, and full texts. The chief investigator (JLC) arbitrated the disagreements and made final decisions. Studies were selected according to the following criteria:
(1) The participants were adult surgical patients undergoing supine position abdominal surgery requiring general anesthesia and low V T ventilation on volume-control mode.
(2) The ventilation strategies were classified by levels of PEEP and RM.
(3) Included studies should report comparisons among two or more different low V T ventilation strategies.
(4) We excluded studies that were not randomized controlled or written in English. Studies containing surgery in the lateral and prone position, and mechanical ventilation conducted by laryngeal mask also were excluded. However, we kept studies related to urology and gynecological surgery in the supine position.
## Data extraction and quality assessment
Two authors (XCL and JW) extracted the following data from the original full texts: first author, publication year, study design, procedure and type of surgery, patients' characteristics [age, gender, body mass index (BMI), ASA class, and sample size], ventilation settings (V T , PEEP, and RMs), and intraoperative pulmonary function and hemodynamic indicators [oxygenation index, dynamic compliance, driving pressure, mean arterial pressure (MAP), and heart rate]. Oxygenation index is calculated by arterial partial pressure of oxygen/inspiratory oxygen fraction, and is determined by the ventilation and gas exchange function of patient's respiratory system. For patients with healthy lungs, the oxygenation index mainly depends on the degree of alveolar opening. Dynamic compliance is calculated by V T /(airway peak pressure -PEEP). Compared with the static compliance [V T /(airway plateau pressure-PEEP)] measured at the end of inspiration period, dynamic compliance represents not only elasticity of lung tissue, but also airway resistance which derives from periodic decruitment/recruitment alveoli and small airways [bib_ref] Effects of different positive end-expiratory pressure titrating strategies on oxygenation and respiratory..., Xu [/bib_ref] [bib_ref] Intraoperative positive end-expiratory pressure evaluation using the intratidal compliance-volume profile, Wirth [/bib_ref]. Thus dynamic compliance is associated with the degree of end-expiration alveolar opening. Driving pressure is calculated by (airway plateau pressure -PEEP), and reflects the degree of ventilator-induced barotrauma [bib_ref] Driving pressure and survival in the acute respiratory distress syndrome, Amato [/bib_ref].
Continuous data were extracted as mean and standard deviation. The continuous data presented as median, interquartile range, and range was transferred to mean and standard deviation according to recommendations from the Cochrane Collaboration: assuming that the median was equivalent to the mean, the interquartile range and range was, respectively, divided by 1.35 and 4 to evaluate the standard deviation. If data were merely available in graphical format, then GetData Graph Digitizer 2.25 (http:// getda ta-graph-digit izer. com/) was used to quantify it.
The Cochrane Collaboration tool containing randomization bias, allocation bias, subjects blinding bias, outcome blinding bias, incomplete data bias, and selective reporting bias was applied to assess the methodological quality independently by two authors.
## Outcomes
The primary outcomes were the intraoperative oxygenation index, dynamic compliance, and driving pressure. Secondary outcomes were the intraoperative heart rate and MAP.
# Statistical analysis
We performed this NMA by STATA13.1 (Stata Corporation, College Station, TX) to compare the effects of different PEEP and RM settings based on low V T ventilation strategies on intraoperative oxygenation, lung dynamic compliance, heart rate, and MAP. The difference of mean, corresponding 95% CI, treatment rankings, probability of being best (Pr Best ), and surface under the cumulative ranking curve (SUCRA) values were estimated using the random-effects model. For indirect comparisons, a node-splitting model was conducted to estimate the degree of inconsistency. Z test was performed to assess the significance of the overall effect size. P < 0.05 was considered statistically significant. To ensure the reliability of the NMA, the number of included RCTs was required to be at least equal to the number of ventilation strategies.
The frequency method was applied to the fitted metaregression model after constructing a heterogeneity matrix. The model treats covariates as the basic parameters and presumes that heterogeneity is independent of the comparison between effect sizes from multi-arm studies. Inconsistency reflects the differences between direct and indirect effects for the same comparison. We estimated the probability of a treatment being ranked at a specific place using "network rank". "Comparison-adjusted" funnel plot was used to evaluate the publication bias. The funnel plot should be symmetrical near the zero line if there is no publication bias. The certainty and hence quality of included studies in terms of within-study bias, reporting bias, indirectness, imprecision, heterogeneity, incoherence, and confidence rating was assessed according to the grading of recommendations assessment, development and evaluation (GRADE) system [bib_ref] Evaluating the quality of evidence from a network meta-analysis, Salanti [/bib_ref] using the CINeMA web app [bib_ref] CINeMA: an approach for assessing confidence in the results of a network..., Nikolakopoulou [/bib_ref].
# Results
## Baseline characteristics of included studies
The initial databases search identified 810 studies. Duplicates (n = 320), unrelated to mechanical ventilation (n = 296), protocol without data (n = 11), and intervention beyond our classification criteria (n = 109) were excluded after reviewing the title and abstract. Forty-six of the remaining 74 studies were available for full text. After fulltext review, non-randomized design (n = 4), non-English full text (n = 1), without clinical outcomes (n = 5), and unrelated ventilation strategies (n = 21) were further eliminated. Fifteen RCTs (n = 3634) met our selection criteria [bib_ref] Effect of intraoperative high positive end-expiratory pressure (PEEP) with recruitment maneuvers vs..., Bluth [/bib_ref] [bib_ref] High versus low positive end-expiratory pressure during general anaesthesia for open abdominal..., Hemmes [/bib_ref] [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref] [bib_ref] Individualized positive end-expiratory pressure in obese patients during general anaesthesia: a randomized..., Nestler [/bib_ref] [bib_ref] Van der Linden P. Does lung compliance optimization through PEEP manipulations reduce..., Van Hecke [/bib_ref] [bib_ref] Effects of positive end-expiratory pressure on intraocular pressure and optic nerve sheath..., You [/bib_ref] [bib_ref] Repeated alveolar recruitment maneuvers with and without positive end-expiratory pressure during bariatric..., Wei [/bib_ref] [bib_ref] The effects of positive end-expiratory pressure at different levels on postoperative respiration..., Ciftci [/bib_ref] [bib_ref] Comparison of the effects of different positive end-expiratory pressure levels on cerebral..., Sargin [/bib_ref] [bib_ref] Open lung approach versus standard protective strategies: effects on driving pressure and..., Ferrando [/bib_ref] [bib_ref] Effect of positive end-expiratory pressure on the sonographic optic nerve sheath diameter..., Chin [/bib_ref] [bib_ref] Effect of recruitment maneuver on arterial oxygenation in patients undergoing robot-assisted laparoscopic..., Ahn [/bib_ref] [bib_ref] Low tidal volume with PEEP and recruitment expedite the recovery of pulmonary..., Pi [/bib_ref] [bib_ref] Positive end-expiratory pressure during laparoscopy: cardiac and respiratory effects, Russo [/bib_ref] [bib_ref] High positive endexpiratory pressure preserves cerebral oxygen saturation during laparoscopic cholecystectomy under..., Kwak [/bib_ref]. The screening and inclusion process is presented in . In particular: one study reported in data for open and laparoscopic surgery separately, so we divided into 2 separate studies in the NMA [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref]. .
In this NMA, three studies were included in the IPEEP + RM group. All of them titrated the individualized PEEP through the RM-decremental titration trial-RM process, which consisted of one RM before and after the decremental PEEP titration trial separately [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref] [bib_ref] Individualized positive end-expiratory pressure in obese patients during general anaesthesia: a randomized..., Nestler [/bib_ref] [bib_ref] Open lung approach versus standard protective strategies: effects on driving pressure and..., Ferrando [/bib_ref]. The detailed steps are as follows: first, patients received an RM, usually with an open-lung pressure of 40 cm H 2 O and a higher one in obese patients, to open the alveoli. Then, a decremental PEEP titration trial was performed by decreasing PEEP step by step from a high level (20 or 25 cm H 2 O) until 5 cm H 2 O. The individualized PEEP was defined as the PEEP with optimal value of the titration parameter (electrical impedance tomography-related parameters in two studies [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref] [bib_ref] Individualized positive end-expiratory pressure in obese patients during general anaesthesia: a randomized..., Nestler [/bib_ref] and dynamic compliance in one study [bib_ref] Open lung approach versus standard protective strategies: effects on driving pressure and..., Ferrando [/bib_ref]. For prevention of the alveoli re-decruitment at the end of titration, another RM was performed before the application of individualized PEEP.
One study was included in the IPEEP group [bib_ref] Van der Linden P. Does lung compliance optimization through PEEP manipulations reduce..., Van Hecke [/bib_ref]. Its titration process was an incremental PEEP titration trial without RM, which had a same definition of individualized PEEP as the decremental trial.
## Quality assessment
The details of the risk-of-bias assessment are summarized in [fig_ref] Figure 2: Risk-of-bias assessment of included studies [/fig_ref]. Fourteen studies reported clear randomization. Eleven studies reported allocation measures. Blinding methods for participants were absent in 7 studies and 6 studies lacked blinding methods for outcome assessors. Incomplete data were identified in 3 researches and selective reporting bias existed in 1 study.
The quality of evidence assessed by the GRADE system was moderate in all primary and secondary outcomes.
## Evidence network
All network plots are presented in . Connecting lines indicated direct comparison, and indirect comparison among interventions can be performed by NMA. The size of nodes represents the overall sample size of each ventilation strategy and the width of connecting lines reflects the number of trials.
The network plots consisted of 5 triangular loops for oxygenation index, 2 triangular loops for dynamic compliance, 3 triangular loops for MAP, and 2 triangular loops for heart rate. For indirect comparisons, a node-splitting model was performed to estimate the degree of inconsistency. Inconsistency was found statistically significant for oxygenation index (P < 0.001 for loop moderate PEEP-moderate PEEP + RM-individualized PEEP + RM) and MAP (P = 0.016 for loop low PEEP-moderate PEEP-high PEEP, P = 0.035 for loop low PEEP-high PEEP + RM-high PEEP).
## Primary outcomes
## Oxygenation index
12 RCTs of 569 patients were included and all 8 groups were available for the intraoperative oxygenation index. The IPEEP + RM group was found out superior to all other groups. The LPEEP group was proved to be inferior to the LPEEP + RM, MPEEP, and MPEEP + RM group. Furthermore, the IPEEP group was shown to be inferior to the MPEEP + RM group. The SUCRA, Pr Best , and mean rank were: 100, 99.9%, and 1.0 for IPEEP + RM; 14.5, 0.0%, and 7.0 for LPEEP; 8.8, 0.0%, and 7.4 for IPEEP [fig_ref] Table 3: SUCRA, Pr Best , and mean rank of all ventilation strategies LPEEP... [/fig_ref]. You et al. [bib_ref] Effects of positive end-expiratory pressure on intraocular pressure and optic nerve sheath..., You [/bib_ref] Robot-assisted lapa-. The LPEEP group was proved to be inferior to all other groups except the HPEEP group. Furthermore, the HPEEP group was shown to be inferior to the MPEEP + RM and LPEEP + RM group, while the MPEEP group was inferior to the LPEEP + RM group. The SUCRA, Pr Best , and mean rank were: 100, 100%, and 1.0 for IPEEP + RM; 0.8, 0.0%, and 7.9 for LPEEP [fig_ref] Table 3: SUCRA, Pr Best , and mean rank of all ventilation strategies LPEEP... [/fig_ref].
## Driving pressure
There were only 4 studies that reported driving pressure, so the data were insufficient for NMA. The conclusion of raw studies suggested that driving pressure in IPEEP + RM group was lower than that in LPEEP + RM (P < 0.001), moderate PEEP (P < 0.001), and moderate PEEP + RM (P < 0.001) [bib_ref] Effect of intraoperative high positive end-expiratory pressure (PEEP) with recruitment maneuvers vs..., Bluth [/bib_ref] [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref] [bib_ref] Individualized positive end-expiratory pressure in obese patients during general anaesthesia: a randomized..., Nestler [/bib_ref] [bib_ref] Repeated alveolar recruitment maneuvers with and without positive end-expiratory pressure during bariatric..., Wei [/bib_ref]. The ventilation strategies, driving pressure, and P values are presented in [fig_ref] Table 4: Driving pressure of included studies LPEEP low PEEP, MPEEP moderate PEEP, HPEEP... [/fig_ref].
## Secondary outcomes
## Mean arterial pressure
In regard to the mean arterial pressure, 10 RCTs of 2413 patients were included and all groups except IPEEP were available. The mean differences of all direct and indirect comparisons showed no significant difference. The SUCRA, Pr Best , and mean rank of all comparisons are presented in [fig_ref] Table 3: SUCRA, Pr Best , and mean rank of all ventilation strategies LPEEP... [/fig_ref].
## Heart rate
In regard to the heart rate, 9 RCTs of 3323 patients were included. Only LPEEP, MPEEP, HPEEP, HPEEP + RM, and IPEEP + RM groups were available. The mean differences of all direct and indirect comparisons showed no significant difference. The SUCRA, Pr Best , and mean rank of all comparisons are presented in [fig_ref] Table 3: SUCRA, Pr Best , and mean rank of all ventilation strategies LPEEP... [/fig_ref].
## Effects of recruitment maneuver
The mean differences between groups with same PEEP but different RM settings were merely significant in IPEEP + RM vs. IPEEP and LPEEP + RM vs. LPEEP, regarding to both intraoperative oxygenation index and dynamic compliance. The SUCRA, Pr Best , and mean rank of both two primary outcomes were superior in all groups with RM than that in groups with same PEEP but without RM [fig_ref] Table 3: SUCRA, Pr Best , and mean rank of all ventilation strategies LPEEP... [/fig_ref].
## Publication bias
The funnel plot of both primary outcomes is presented in [fig_ref] Figure 4: Comparison-adjusted funnel plot for intraoperative oxygenation index [/fig_ref] (oxygenation index) and [fig_ref] Figure 5: Comparison-adjusted funnel plot for intraoperative dynamic compliance [/fig_ref]. The included studies were symmetrically distributed on both sides of the vertical line (x = 0), indicating no significant publication bias.
# Discussion
The present NMA demonstrate that in terms of intraoperative pulmonary function and hemodynamics, individualized PEEP combined with RM may be the currently optimal low V T ventilation strategy, while low PEEP without RM was the worst low V T ventilation strategy. Also, evidence suggested that RM was associated with improvements in oxygenation index and dynamic compliance. The lung-protective ventilation strategy (V T of 6 ml/ kg, appropriate PEEP and RM) was proved superior to the conventional high V T ventilation strategy on the prognosis of acute respiratory distress syndrome (ARDS) patients, as a consequence of reduced volutrauma and barotrauma [bib_ref] Low-tidal-volume ventilation in the acute respiratory distress syndrome, Malhotra [/bib_ref]. Inspired by this result, researchers tried to apply lungprotective ventilation strategy in patients undergoing intraoperative mechanical ventilation. However, there are two main problems exposed in clinical anesthesia. First, a low V T of 6 ml/kg was reasonable for ARDS patients to avoid excessive airway pressure but unnecessary for patients with healthy lungs. In addition, a tidal volume of 6 ml/ kg is too low to be accepted by most anesthesiologists, compared with the traditional tidal volume of 10 ml/kg or higher. Therefore, most studies related to intraoperative ventilation strategy selected 8 ml/kg in groups receiving low tidal volume ventilation. Second, the low V T ventilation strategy also increased the risk of atelectasis due to insufficient PEEP. In the previous studies, low V T combined with low PEEP was associated with more serious atelectrauma and higher 30-day mortality in patients undergoing abdominal surgery and indicated the necessity of "optimal PEEP" setting [bib_ref] Low intraoperative tidal volume ventilation with minimal PEEP is associated with increased..., Levin [/bib_ref]. Despite a series of efforts, no agreement has been reached on the optimal level of PEEP [bib_ref] Effect of intraoperative high positive end-expiratory pressure (PEEP) with recruitment maneuvers vs..., Bluth [/bib_ref] [bib_ref] High versus low positive end-expiratory pressure during general anaesthesia for open abdominal..., Hemmes [/bib_ref] [bib_ref] Low intraoperative tidal volume ventilation with minimal PEEP is associated with increased..., Levin [/bib_ref]. This debate may be attributed to applying one fixed PEEP to all patients, as their individual characteristics required different optimal PEEP. The concept of 'individualized PEEP' was then proposed to explore a more reasonable ventilation strategy [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref].
Our NMA indicated a consistent strength of applying RM followed by individualized PEEP over other fixed PEEP strategies. The recent RCTs mainly revealed the advantages of individualized PEEP over low PEEP and moderate PEEP [bib_ref] Individualised perioperative open-lung approach versus standard protective ventilation in abdominal surgery (iPROVE):..., Ferrando [/bib_ref] [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref] [bib_ref] Setting individualized positive endexpiratory pressure level with a positive end-expiratory pressure decrement..., Ferrando [/bib_ref] , while our NMA extended the conclusion to other fixed PEEP values. Higher oxygenation index and dynamic compliance suggested fewer collapsed alveoli. Based on current studies, two crucial points of the individualized PEEP may contribute to the alveoli opening effect. One is the variability of individualized PEEP. Even in non-obese patients undergoing open abdominal surgery, individualized PEEP could reach over 12 cm H 2 O. Such a high PEEP should be able to maintain most alveoli open. However, one included study revealed that compared with fixed PEEP of 2 cm H 2 O, a fixed PEEP of 12 cm H 2 O combined with RM did not lead to better intraoperative pulmonary function [bib_ref] Effect of intraoperative high positive end-expiratory pressure (PEEP) with recruitment maneuvers vs..., Bluth [/bib_ref]. This contradictory conclusion can be explained by the other feature, large range of the individualized PEEP which is mostly over 10 cm H 2 O. The variability of individualized PEEP was probably caused by the individual characteristics such as chest wall compliance, abdominal pressure, pleural pressure, and surgical position. These inevitable high inter-individual heterogeneities strengthen the importance of individualizing the optimal PEEP.
The other crucial point of individualized PEEP is the correct titration process. In this NMA, all three studies in the IPEEP + RM group selected the RM-decremental titration trial-RM process [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref] [bib_ref] Individualized positive end-expiratory pressure in obese patients during general anaesthesia: a randomized..., Nestler [/bib_ref] [bib_ref] Open lung approach versus standard protective strategies: effects on driving pressure and..., Ferrando [/bib_ref]. The individualized PEEP was defined as the PEEP with optimal value of the titration Network plot of enrolled studies in this network metaanalysis. a Oxygenation index; b dynamic compliance; c mean arterial pressure; d heart rate; LPEEP low PEEP, MPEEP moderate PEEP, HPEEP high PEEP, IPEEP individualized PEEP, RM recruitment maneuverNetwork league table for all ventilation strategies in regard to intraoperative oxygenation index, dynamic compliance, mean arterial pressure, and heart rate parameter (electrical impedance tomography-related parameters in two studies [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref] [bib_ref] Individualized positive end-expiratory pressure in obese patients during general anaesthesia: a randomized..., Nestler [/bib_ref] and dynamic compliance in one study [bib_ref] Open lung approach versus standard protective strategies: effects on driving pressure and..., Ferrando [/bib_ref]. The included study in the IPEEP group used the incremental PEEP titration without RM, with the same definition of individualized PEEP as the decremental trial [bib_ref] Van der Linden P. Does lung compliance optimization through PEEP manipulations reduce..., Van Hecke [/bib_ref]. In our current study, the IPEEP + RM group was superior to IPEEP group in terms of both oxygenation index and dynamic compliance. One previous mathematical model revealed the possible reason that there was a consistent relationship between the PEEP level giving maximum lung compliance and the preset "open lung PEEP" (i.e., optimal PEEP) in the decremental trial with RM. This relationship, however, was inconsistent in the incremental trial without RM [bib_ref] Best compliance during a decremental, but not incremental, positive end-expiratory pressure trial..., Hickling [/bib_ref]. This hypothesis was later confirmed by a clinical trial related to obese patients with ARDS [bib_ref] Lung recruitment in obese patients with acute respiratory distress syndrome, Fumagalli [/bib_ref]. Therefore, we concluded that the individualized PEEP should be titrated under the RM-decremental titration-RM trial for consideration of higher intraoperative oxygenation and lung compliance.
Our NMA showed that RM improved the intraoperative oxygenation index and dynamic compliance in groups with RM especially in individualized PEEP and low PEEP groups, which were in line with previous RCTs [bib_ref] Respiratory function during anesthesia: effects on gas exchange, Hedenstierna [/bib_ref] [bib_ref] Perioperative lung protective ventilation in obese patients, Fernandez-Bustamante [/bib_ref]. RM improved pulmonary function by overcoming the opening Estimates are presented as mean differences (95% confidence interval). Mean differences < 0 favor the column ventilation strategy and mean differences > 0 favor the row ventilation strategy LPEEP low PEEP, MPEEP moderate PEEP, HPEEP high PEEP, IPEEP individualized PEEP, RM recruitment maneuver Mean differences in bold are significantly different pressure, reverse atelectasis, and promote the benefits of PEEP [bib_ref] Intraoperative protective mechanical ventilation for prevention of postoperative pulmonary complications: a comprehensive..., Guldner [/bib_ref]. The PEEP value applied during mechanical ventilation was usually insufficient for alveolar recruitment [bib_ref] Respiratory management of perioperative obese patients, Imber [/bib_ref] ; thus, a large amount of dorsal alveoli near the diaphragm may maintain collapsed without RM. In addition, evidence showed that even if RM was absent, the end-expiratory lung volume still increased along with the elevated level of PEEP [bib_ref] Lung recruitment in obese patients with acute respiratory distress syndrome, Fumagalli [/bib_ref]. That mean applying PEEP without RM ventilated air into the already open alveoli, inducing volutrauma and barotraumas and the injury was more severe when using higher PEEP. As the individualized PEEP is usually at a high level, the importance of RM should be considered seriously. Based on the aforementioned effects of PEEP and RM, it is reasonable to believe that the low PEEP group has the worst oxygenation index. In the case of low V T , insufficient PEEP and lack of RM significantly increase the amount of atelectasis, leading to the impairment of gas exchange and respiratory mechanism. Previous studies also revealed that simply lowering V T without supplementing PEEP and RM would significantly increase the incidence of PPCs and even mortality in surgical patients, in comparison of high V T , zero PEEP, and no RM strategy [bib_ref] Low intraoperative tidal volume ventilation with minimal PEEP is associated with increased..., Levin [/bib_ref] [bib_ref] Evaluation of a ventilation strategy to prevent barotrauma in patients at high..., Stewart [/bib_ref].
Our analysis of MAP and heart rate revealed a limited impact of different PEEP and RM settings on intraoperative hemodynamic, which is similar to previous studies. It is worth noting that hypotension is common during the titration trial of individualized PEEP though proved one-past and harmless [bib_ref] Individual positive end-expiratory pressure settings optimize intraoperative mechanical ventilation and reduce postoperative..., Pereira [/bib_ref] [bib_ref] Individualized positive end-expiratory pressure in obese patients during general anaesthesia: a randomized..., Nestler [/bib_ref]. The possible reason for hemodynamic instability when applying individualized PEEP is that the appropriate level of PEEP keeps most alveoli open, improves oxygenation, and thus reduces pulmonary vascular resistance and right ventricular load [bib_ref] Respiratory and haemodynamic changes during decremental open lung positive end-expiratory pressure titration..., Gernoth [/bib_ref].
This study is a relatively small NMA. Its limitations are as follows: first, we only included studies related to low V T ventilation strategies, without studies about conventional high V T strategy. Second, the analysis of the incidence of PPCs was absent, which should be an issue worth more attention than oxygenation and lung compliance. Of note, individualized PEEP was previously proved to increase mortality in patients with moderate-to-severe ARDS [bib_ref] Effect of lung recruitment and titrated positive end-expiratory pressure (PEEP) vs low..., Cavalcanti [/bib_ref]. Furthermore, the maneuver of RM-titration-RM may induce airway peak pressure to reach even 40 cm H 2 O. It would inevitably lead to alveolar barotrauma, which is fatal for patients with severe lung injury. The data we initially extracted included the incidence of PPCs, but only 5 included studies reported the events of PPCs, and the data were insufficient for a proper NMA or even a conventional pairwise comparison meta-analysis. Therefore, further evidence is necessary before applying individualized PEEP as a routine strategy, and future studies should focus on the incidence of PPCs. Thirdly, two studies have much larger sample sizes than other studies [bib_ref] Effect of intraoperative high positive end-expiratory pressure (PEEP) with recruitment maneuvers vs..., Bluth [/bib_ref] [bib_ref] High versus low positive end-expiratory pressure during general anaesthesia for open abdominal..., Hemmes [/bib_ref] , which may increase the risk of type I and II errors. Fortunately, both two studies lack oxygenation index data, thus not included in the analyses of primary outcome.
# Conclusion
In conclusion, our systematic review and NMA suggested that individualized PEEP combined with recruitment maneuver may be the optimal low V T ventilation strategy in abdominal surgery at present, while low PEEP without recruitment maneuver may be the worst one. Recruitment maneuver is able to improve the intraoperative pulmonary function in all PEEP levels. In our opinion, further research should focus on the direct comparison of the individualized PEEP to the high PEEP, in terms of not only functional but also "hard-core" indicators such as the incidence of PPCs and mortality.
[fig] Fig: 1 PRISMA flow diagram. RCT randomized controlled trial, PRISMA Preferred Reporting Items for Systematic Reviews and Meta-Analyses [/fig]
[table] Table 3: SUCRA, Pr Best , and mean rank of all ventilation strategies LPEEP low PEEP, MPEEP moderate PEEP, HPEEP high PEEP, IPEEP individualized PEEP, RM recruitment maneuver, Pr Best probability of being best, SUCRA surface under cumulating ranking curve [/table]
[table] Table 4: Driving pressure of included studies LPEEP low PEEP, MPEEP moderate PEEP, HPEEP high PEEP, IPEEP individualized PEEP, RM recruitment maneuver a This study reported data of open and laparoscopic surgery separately, so we divided this study into 2 separate ones [/table]
|
The Hollow Adrenal Gland Sign: An Ominous Alert
Teaching Point: The hollow adrenal gland sign is common, and may be specific, in patients with septic shock, and is a predictor of poor prognosis.
## Case report
A 58-year-old male, smoker, with a history of alcohol abuse presented to the emergency department due to fever (38.2°C). The patient had bilateral scleral icterus, hepatomegaly, and bilateral lower limb edema. Laboratory tests revealed elevation of inflammatory parameters, thrombocytopenia, acute kidney injury, hyperbilirubinemia, cytocholestasis, and coagulopathy. An acute-on-chronic liver failure syndrome due to an infection of unknown origin was presumed. His condition deteriorated with septic shock and abdominal pain. Nasogastric entubation revealed faecaloid content. He underwent an abdominopelvic computed tomography with intravenous contrast (80ml 300mg iodine/mL at 4mL/s, followed by a 50mL saline flush) on arterial and venous phases. Adrenal glands showed central hypoenhancement relative to the peripheral zone in the arterial phase (arrows in [fig_ref] Figure 1: Figure 1. [/fig_ref] and homogenous enhancement in the venous phase (arrows in [fig_ref] Figure 2: Figure 2. [/fig_ref] -the hollow adrenal gland sign (HAGS). Images also demonstrated ascites (asterisk in , thickened bowel loops due to submucosal edema and mucosal hyperenhancement involving the small bowel (arrow in and the right colon, splenic hypoenhancement (asterisk in [fig_ref] Figure 2: Figure 2. [/fig_ref] , and some bilateral renal infarcts (not shown). These findings were interpreted as a hypoperfusion complex syndrome. There were also signs of segmental acute intestinal ischemia (not shown), so the patient underwent an emergent laparotomy. There was bad perfusion of the small bowel and right colon, with no signs of irreversible ischaemia, so a peritoneal lavage was performed. Subsequent fluorescence angiography confirmed bowel viability. Unfortunately, the patient died.
## Comment
The HAGS is a pattern recently described in the literature on dual phase contrast-enhanced computed tomography (CT) in patients with septic shock [bib_ref] The hollow adrenal gland sign: A newly described enhancing pattern of the..., Peng [/bib_ref]. It is defined as a less intense enhancement of the central zone of the adrenal gland in the arterial phase, followed by a homogeneously enhancing adrenal gland in the venous phase. There are two types: the hypoenhancement of the central zone is homogeneous, forming a typical Y or I-shaped region; or the central zone has a heterogeneous enhancement creating a mosaic appearance. The sign is present when it appears on either adrenal gland. The HAGS is common, and may be specific, in patients with septic shock. It is associated with severe illness more frequently requiring organ-supportive interventions and it has a good specificity to predict poor prognosis; however, it has a low sensitivity. The HAGS showed excellent reproducibility between different observers, suggesting it is easily recognizable and can be used to identify patients with sepsis who have a higher risk of in-hospital mortality and select the appropriate level of care. A similar but rarer enhancement pattern, the "train-track appearance," is defined as preservation of normal peripheral adrenal enhancement with central low attenuation in a thickened adrenal [bib_ref] Train-track appearance" in early non-traumatic adrenal hemorrhage, Coursier [/bib_ref]. In this cases, auxiliary features were described, namely peri-adrenal fat stranding. This has been reported in early non-traumatic adrenal hemorrhage, which is an uncommon but potentially lifethreatening condition, preceding adrenal hematomas.
[fig] Figure 1: Figure 1. [/fig]
[fig] Figure 2: Figure 2. [/fig]
|
Cytokine Receptor Endocytosis: New Kinase Activity-Dependent and -Independent Roles of PI3K
Type I and II cytokine receptors are cell surface sensors that bind cytokines in the extracellular environment and initiate intracellular signaling to control processes such as hematopoiesis, immune function, and cellular growth and development. One key mechanism that regulates signaling from cytokine receptors is through receptor endocytosis. In this mini-review, we describe recent advances in endocytic regulations of cytokine receptors, focusing on new paradigms by which PI3K controls receptor endocytosis through both kinase activity-dependent and -independent mechanisms. These advances underscore the notion that the p85 regulatory subunit of PI3K has functions beyond regulating PI3K kinase activity, and that PI3K plays both positive and negative roles in receptor signaling. On the one hand, the PI3K/Akt pathway controls various aspects downstream of cytokine receptors. On the other hand, it stimulates receptor endocytosis and downregulation, thus contributing to signaling attenuation.
## Multiple pathways for endocytosis of cytokine receptors
Receptor endocytosis is initiated at the plasma membrane and can be generally divided into clathrin-mediated endocytosis (CME) or clathrin-independent endocytosis (CIE) based on the involvement of the endocytic coat protein clathrin [bib_ref] Forty years of clathrin-coated vesicles, Robinson [/bib_ref] [bib_ref] Molecular structure, function, and dynamics of clathrin-mediated membrane traffic, Kirchhausen [/bib_ref]. In CME, activated receptors recruit clathrin adaptors such as the AP2 complex, inducing the formation of a clathrin coat that stabilizes membrane curvature and drives invagination. Subsequently, vesicles are pinched off from the plasma membrane by the dynamin GTPase [bib_ref] Molecular mechanism and physiological functions of clathrin-mediated endocytosis, Mcmahon [/bib_ref] [bib_ref] Dynamin: functional design of a membrane fission catalyst, Schmid [/bib_ref]. CIE is a composite of several distinct pathways, the best studied being the caveolin-mediated endocytosis [bib_ref] Molecular mechanisms of clathrin-independent endocytosis, Hansen [/bib_ref] [bib_ref] Pathways of clathrin-independent endocytosis, Mayor [/bib_ref]. These pathways, which can be either dynamin dependent or independent [bib_ref] Clathrin-independent endocytosis: mechanisms and function, Sandvig [/bib_ref] , require actin polymerization and either Src-family kinases in the case of caveolin-mediated endocytosis [bib_ref] Tyrosine phosphorylationdependence of caveolae-mediated endocytosis, Sverdlov [/bib_ref] or small GTPases such as RhoA and Rac1 for other CIE pathways [bib_ref] Clathrin-independent pathways of endocytosis, Mayor [/bib_ref].
Both CME and CIE are involved in endocytosis of cytokine receptors [bib_ref] Endocytic regulation of cytokine receptor signaling, Cendrowski [/bib_ref] [bib_ref] Essential role of endocytosis for interleukin-4-receptor-mediated JAK/STAT signalling, Kurgonaite [/bib_ref] [bib_ref] Interleukin 2 receptors and detergent-resistant membrane domains define a clathrin-independent endocytic pathway, Lamaze [/bib_ref] [bib_ref] IL-7 induces rapid clathrin-mediated internalization and JAK3-dependent degradation of IL-7Ralpha in T..., Henriques [/bib_ref] [bib_ref] Janus kinases promote cell-surface expression and provoke autonomous signalling from routing-defective G-CSF..., Meenhuis [/bib_ref]. CME mediates endocytosis of gp130, the shared receptor for IL6 family cytokines, and receptors for prolactin, thrombopoietin, erythropoietin, interferon, IL5 (IL5Rα), IL7 (IL7Rα), and IL36 [bib_ref] Identification of a novel function of the clathrin-coated structure at the plasma..., Chen [/bib_ref] [bib_ref] IL-7 induces rapid clathrin-mediated internalization and JAK3-dependent degradation of IL-7Ralpha in T..., Henriques [/bib_ref] [bib_ref] Signal transduction and intracellular trafficking by the interleukin 36 receptor, Saha [/bib_ref] [bib_ref] Three lysine residues in the common beta chain of the interleukin-5 receptor..., Lei [/bib_ref] [bib_ref] Internalization of the interleukin 6 signal transducer gp130 does not require activation..., Thiel [/bib_ref] [bib_ref] Dimerization, ubiquitylation and endocytosis go together in growth hormone receptor function, Strous [/bib_ref] [bib_ref] Stat-mediated signaling induced by type I and type II interferons (IFNs) is..., Marchetti [/bib_ref] [bib_ref] Polyubiquitination of prolactin receptor stimulates its internalization, postinternalization sorting, and degradation via..., Varghese [/bib_ref] [bib_ref] YRRL motifs in the cytoplasmic domain of the thrombopoietin receptor regulate receptor..., Hitchcock [/bib_ref] [bib_ref] Ligand-induced EpoR internalization is mediated by JAK2 and p85 and is impaired..., Sulahian [/bib_ref]. CIE mediates endocytosis of the common γ chain receptor, and IL2Rβ, IL4Rα, and IL15Rα receptors [bib_ref] Essential role of endocytosis for interleukin-4-receptor-mediated JAK/STAT signalling, Kurgonaite [/bib_ref] [bib_ref] Cortactin and dynamin are required for the clathrin-independent endocytosis of gammac cytokine..., Sauvonnet [/bib_ref] [bib_ref] Caveolar internalization of growth hormone, Lobie [/bib_ref] [bib_ref] Separate endocytic pathways regulate IL-5 receptor internalization and signaling, Lei [/bib_ref] [bib_ref] Clathrin-independent endocytosis and signalling of interleukin 2 receptors IL-2R endocytosis and signalling, Gesbert [/bib_ref] [bib_ref] Rapid endocytosis of interleukin-15 by cerebral endothelia, Stone [/bib_ref]. The same receptor can sometimes utilize both CME and CIE pathways. One example is endocytosis of the common β chain receptor (βc), which is shared by IL3, IL5, and GM-CSF receptors. βc co-localizes with both transferrin receptor (a CME marker) and cholera toxin-B (a CIE marker), but interestingly, signaling complexes mainly partition to the transferrin-containing fraction [bib_ref] Separate endocytic pathways regulate IL-5 receptor internalization and signaling, Lei [/bib_ref]. The signaling dichotomy may involve intersectin 2, which is specifically involved in CME to regulate JAK2 and Akt activation downstream of βc [bib_ref] Identification of a novel function of the clathrin-coated structure at the plasma..., Chen [/bib_ref]. Growth hormone receptor also uses both CME and CIE for its internalization [bib_ref] Dimerization, ubiquitylation and endocytosis go together in growth hormone receptor function, Strous [/bib_ref] [bib_ref] Caveolar internalization of growth hormone, Lobie [/bib_ref] , and perturbation of CIE specifically affects ERK activation downstream of the receptor but not STAT5 (54). Thus, differential use of CME and CIE may allow cells to regulate downstream signaling of cytokine receptors.
Ubiquitination plays an important role in receptor endocytosis through both CME and CIE [bib_ref] The role of ubiquitylation in receptor endocytosis and endosomal sorting, Haglund [/bib_ref]. Through sequential actions of ubiquitin-activating (E1), ubiquitin-conjugating (E2), and ubiquitin-ligating (E3) enzymes, a small protein ubiquitin is covalently attached to lysine residues on target receptors. Because ubiquitin itself contains lysines that can serve as acceptor sites, target proteins can be subjected to mono-ubiquitination, multiubiquitination (mono-ubiquitination on multiple lysines), or poly-ubiquitination. Mono-ubiquitination has been shown to mediate protein trafficking and signaling [bib_ref] Signaling through monoubiquitination, Sigismund [/bib_ref] , whereas polyubiquitination can promote protein degradation [bib_ref] The role of ubiquitylation in receptor endocytosis and endosomal sorting, Haglund [/bib_ref]. Endocytic adaptor proteins and the endosomal sorting complex required for transport (ESCRT) contain ubiquitin-binding domain or ubiquitin-interacting motif (UIM), thereby facilitating their interaction with ubiquitinated receptors. This allows endocytic adaptors to target ubiquitinated receptors to the endocytic machinery and allows the ESCRT complexes to direct budding of ubiquitinated receptors into intraluminal vesicles within endosomes, thereby halting receptor signaling [bib_ref] Ubiquitin-dependent sorting in endocytosis, Piper [/bib_ref].
Endocytosis of cytokine receptors is regulated by ubiquitination. For example, ubiquitination by the E3 ubiquitin ligase SCF (βTrCP) drives endocytosis of growth hormone receptor, prolactin receptor, and the Type I interferon receptor (IFNAR1) [bib_ref] Ubc13 and COOH terminus of Hsp70-interacting protein (CHIP) are required for growth..., Slotman [/bib_ref] [bib_ref] The ubiquitin conjugation system is required for ligand-induced endocytosis and degradation of..., Strous [/bib_ref] [bib_ref] JAK2 is a negative regulator of ubiquitin-dependent endocytosis of the growth hormone..., Putters [/bib_ref] [bib_ref] Negative regulation of prolactin receptor stability and signaling mediated by SCF(beta-TrCP) E3..., Li [/bib_ref] [bib_ref] SCF(HOS) ubiquitin ligase mediates the ligand-induced down-regulation of the interferon-alpha receptor, Kumar [/bib_ref]. Another E3 ligase, c-Cbl, has been implicated in the internalization and/or degradation of the βc, thrombopoietin receptor, and the erythropoietin receptor (EpoR) [bib_ref] Three lysine residues in the common beta chain of the interleukin-5 receptor..., Lei [/bib_ref] [bib_ref] Ubiquitination and degradation of the thrombopoietin receptor c-Mpl, Saur [/bib_ref] [bib_ref] Ubiquitination regulates the internalization, endolysosomal sorting, and signaling of the erythropoietin receptor, Bulut [/bib_ref]. Interestingly, different ubiquitination sites on the EpoR are able to regulate distinct steps in the endocytic process [bib_ref] Ubiquitination regulates the internalization, endolysosomal sorting, and signaling of the erythropoietin receptor, Bulut [/bib_ref].
## Pi3k pathway in cytokine signaling
Class IA PI3K is commonly activated by cytokine receptors [bib_ref] The evolution of phosphatidylinositol 3-kinases as regulators of growth and metabolism, Engelman [/bib_ref]. PI3Ks are lipid kinases that phosphorylate the 3′-hydroxyl group of phosphatidylinositol and its phosphorylated derivatives. At the plasma membrane, class IA PI3Ks phosphorylate phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2] to generate phosphatidylinositol 3,4,5-triphosphate [PI [bib_ref] Phosphatidylinositol 3-phosphates -at the interface between cell signalling and membrane traffic, Marat [/bib_ref] [bib_ref] The phosphoinositide 3-kinase pathway, Cantley [/bib_ref] P3], which recruits PI(3,4,5)P3-binding proteins to activate downstream signaling. One of these downstream proteins is the serine/threonine kinase Akt, and together, the PI3K/AKT pathway regulates a plethora of cellular processes [bib_ref] Ten years of protein kinase B signalling: a hard Akt to follow, Brazil [/bib_ref] [bib_ref] Phosphoinositide 3-kinase in immunological systems, Fruman [/bib_ref]. The other is Rac1, which plays a major role in remodeling the actin cytoskeleton [bib_ref] Phosphoinositide 3-kinase in immunological systems, Fruman [/bib_ref].
Class IA PI3Ks function as heterodimers with a p110 catalytic subunit (p110α, β, or δ) and a p85-like regulatory subunit (p85α, β or their splice variants p55α, p50α, or p55γ) (4). p85 stabilizes and maintains p110 in an inhibited state and directly interacts with phosphorylated cytoplasmic tyrosines in cytokine receptors upon ligand binding. Conformational changes in p85 induced by receptor binding relieve its inhibition of p110 [bib_ref] Structural basis for activation and inhibition of class I phosphoinositide 3-kinases, Vadas [/bib_ref]. Recent evidence suggests that the association and activation of PI3K by cytokine receptors promotes receptor endocytosis in addition to the activation of downstream Akt signaling [bib_ref] Cbl ubiquitination of p85 is essential for Epo-induced EpoR endocytosis, Bulut [/bib_ref] [bib_ref] The signalling factor PI3K is a specific regulator of the clathrin-independent dynamin-dependent..., Basquin [/bib_ref]. Moreover, the contribution of the p85 regulatory subunit in these mechanisms can be PI3K kinase activity independent [bib_ref] Ligand-induced EpoR internalization is mediated by JAK2 and p85 and is impaired..., Sulahian [/bib_ref]. Below, we discuss two new paradigms by which class IA PI3Ks regulate cytokine receptor endocytosis.
## Pi3k and actin-mediated endocytosis of il2 receptor (il2r)
IL2 receptor belongs to the Type I cytokine receptors and is important for T cell immune function [bib_ref] Interleukin-2 receptor signaling: at the interface between tolerance and immunity, Malek [/bib_ref] [bib_ref] The role of interleukin-2 during homeostasis and activation of the immune system, Boyman [/bib_ref]. IL2R is composed of IL2Rα, IL2Rβ, and the common γ chain. Internalized IL2Rα recycles back to the plasma membrane, whereas IL2Rβ and the common γ chain are sorted to the lysosome and degraded [bib_ref] Trafficking of interleukin 2 and transferrin in endosomal fractions of T lymphocytes, Duprez [/bib_ref] [bib_ref] Endocytosis of interleukin 2 receptors in human T lymphocytes: distinct intracellular localization..., Hemar [/bib_ref]. IL2Rβ was among the first cytokine receptors shown to be internalized via CIE [bib_ref] Rapid endocytosis of interleukin 2 receptors when clathrin-coated pit endocytosis is inhibited, Subtil [/bib_ref]. It is constitutively internalized, but internalization is augmented by IL2 binding [bib_ref] The signalling factor PI3K is a specific regulator of the clathrin-independent dynamin-dependent..., Basquin [/bib_ref] [bib_ref] Endocytosis of the beta chain of interleukin-2 receptor requires neither interleukin-2 nor..., Hemar [/bib_ref].
Endocytosis of IL2Rβ is clathrin-and caveolin independent and relies on RhoA, dynamin, Rac1, PAK kinases (p21-activated kinases), and actin polymerization [bib_ref] Interleukin 2 receptors and detergent-resistant membrane domains define a clathrin-independent endocytic pathway, Lamaze [/bib_ref] [bib_ref] Cortactin and dynamin are required for the clathrin-independent endocytosis of gammac cytokine..., Sauvonnet [/bib_ref] [bib_ref] Clathrin-independent endocytosis used by the IL-2 receptor is regulated by Rac1, Pak1..., Grassart [/bib_ref] [bib_ref] Pak1 phosphorylation enhances cortactin-N-WASP interaction in clathrin-caveolin-independent endocytosis, Grassart [/bib_ref]. New studies showed that two rounds of actin polymerization are enlisted for IL2Rβ internalization. The first round relies on WAVE (WASP-family verprolin homologous protein), through a WAVEinteracting sequence in the cytoplasmic tail of IL2Rβ [bib_ref] Membrane protrusion powers clathrin-independent endocytosis of interleukin-2 receptor, Basquin [/bib_ref]. This round of actin polymerization occurs before receptor clustering and is thought to be responsible for receptor recruitment near the base of membrane protrusion to initiate pit formation. The second round occurs just before receptor internalization and involves Pak1 phosphorylation of cortactin, another activator of actin polymerization [bib_ref] Cortactin promotes and stabilizes Arp2/3-induced actin filament network formation, Weaver [/bib_ref] [bib_ref] Activation of Arp2/3 complex-mediated actin polymerization by cortactin, Uruno [/bib_ref] , thereby increasing its association with N-WASP (neuronal Wiskott-Aldrich syndrome protein) [bib_ref] Pak1 phosphorylation enhances cortactin-N-WASP interaction in clathrin-caveolin-independent endocytosis, Grassart [/bib_ref]. Interestingly, dynamin, which mediates vesicle scission in the later stage of IL2Rβ internalization, also controls the transition of WAVE complex and N-WASP recruitments (78).
Sauvonnet's group showed that PI3K plays multiple roles in regulating IL2R CIE [bib_ref] The signalling factor PI3K is a specific regulator of the clathrin-independent dynamin-dependent..., Basquin [/bib_ref]. First, IL2 stimulation activates PI3K, leading to the production of PI(3,4,5)P3 and the recruitment of Vav2, the guanine nucleotide exchange factor that activates Rac1 [bib_ref] Role of substrates and products of PI 3-kinase in regulating activation of..., Han [/bib_ref]. Inhibitors of PI3K kinase activity, knockdown of p85 and Vav2, or overexpression of a mutant p85 devoid of p110-binding domain all inhibit IL2R endocytosis. Second, p85 binds directly to Rac1, with higher affinity for the GTP-bound active form. A model is thus proposed that IL2R activation of PI3K leads to the recruitment of both Vav2 and its substrate Rac1, which can stimulate the Rac1-Pak1-cortactin-N-WASP cascade to promote actin polymerization, driving IL2R internalization [bib_ref] The signalling factor PI3K is a specific regulator of the clathrin-independent dynamin-dependent..., Basquin [/bib_ref]. Because the WAVE complex is a known downstream effector of Rac GTPases [bib_ref] WAVE, a novel WASP-family protein involved in actin reorganization induced by Rac, Miki [/bib_ref] [bib_ref] Sra-1 and Nap1 link Rac to actin assembly driving lamellipodia formation, Steffen [/bib_ref] and PIP3 [bib_ref] PtdIns(3,4,5)P3 binding is necessary for WAVE2-induced formation of lamellipodia, Oikawa [/bib_ref] [bib_ref] WAVE3-mediated cell migration and lamellipodia formation are regulated downstream of phosphatidylinositol 3-kinase, Sossey-Alaoui [/bib_ref] , PI3K may also regulate IL2Rβ CIE through WAVE.
Recently, endophilin and its interacting protein Alix (ALG-2interacting protein X) have also been implicated in CIE of IL2Rβ [bib_ref] ALG-2 interacting protein-X (Alix) is essential for clathrin-independent endocytosis and signaling, Mercier [/bib_ref] [bib_ref] Endophilin marks and controls a clathrin-independent endocytic pathway, Boucrot [/bib_ref]. Endophilin is a Bin/Amphiphysin/Rvs domain protein that is involved in vesicle endocytosis and membrane curvature generation [bib_ref] Mechanism of endophilin N-BAR domain-mediated membrane curvature, Gallop [/bib_ref] [bib_ref] Endophilin BAR domain drives membrane curvature by two newly identified structure-based mechanisms, Masuda [/bib_ref]. This pathway, termed fast endophilinmediated endocytosis (FEME) by the McMahon group, is utilized by IL2R as well as several G-protein-coupled receptors and bacterial Shiga and cholera toxins [bib_ref] Endophilin marks and controls a clathrin-independent endocytic pathway, Boucrot [/bib_ref] [bib_ref] Endophilin-A2 functions in membrane scission in clathrin-independent endocytosis, Renard [/bib_ref]. It is characterized by endophilin-positive uptake structures after ligand-induced receptor activation. Endophilin also works together with dynamin and actin in membrane scission [bib_ref] Endophilin-A2 functions in membrane scission in clathrin-independent endocytosis, Renard [/bib_ref] [bib_ref] Membrane fission is promoted by insertion of amphipathic helices and is restricted..., Boucrot [/bib_ref]. As with the PI3K/Vav2 pathway described above, the FEME pathway depends on dynamin, Rac, Pak1, and actin polymerization [bib_ref] Endophilin marks and controls a clathrin-independent endocytic pathway, Boucrot [/bib_ref] , suggesting that FEME and PI3K/Vav2 mechanisms may be part of the same pathway. Importantly, PI3K kinase activity is required for FEME, because PI(3,4)P2, converted from PI(3,4,5)P3 by SHIP1/2-dependent dephosphorylation, is necessary for lamellipodin-dependent recruitment of endophilin in FEME [bib_ref] Endophilin marks and controls a clathrin-independent endocytic pathway, Boucrot [/bib_ref]. The exact molecular details of this pathway, the degree to which the PI3K/ Vav2 and FEME pathways are distinct or can be employed under different context, and whether PI3K regulates other aspects await future interrogations. In addition, whether other cytokine receptors can also utilize similar endocytic pathways is currently unclear.
## Cbl-dependent ubiquitination of p85 mediates epor endocytosis
The EpoR is another member of the Type I cytokine receptors and is essential to drive red blood cell production [bib_ref] The erythropoietin receptor: structure, activation and intracellular signal transduction, Constantinescu [/bib_ref] [bib_ref] The erythropoietin receptor: molecular structure and hematopoietic signaling pathways, Watowich [/bib_ref]. In contrast to the IL2R, which forms heteromeric receptor complexes and associates with both JAK1 and JAK3 for signaling, EpoR forms homodimers and couples to only JAK2 for signaling. Epo-induced endocytosis is a key element in negative regulation FiGURe 2 | Clathrin-dependent endocytosis of erythropoietin receptor (epoR). Upon Epo stimulation, activated JAK2 phosphorylates EpoR cytoplasmic tyrosines to recruit p85 (step 1). Subsequently, ubiquitinated p85, mediated by c-Cbl (step 2), recruits Epsin-1 (step 3), linking EpoR to the endocytic machinery for downregulation [bib_ref] Ligand-induced EpoR internalization is mediated by JAK2 and p85 and is impaired..., Sulahian [/bib_ref] [bib_ref] Cbl ubiquitination of p85 is essential for Epo-induced EpoR endocytosis, Bulut [/bib_ref].
of Epo signaling [bib_ref] Ligand-induced EpoR internalization is mediated by JAK2 and p85 and is impaired..., Sulahian [/bib_ref] [bib_ref] Both proteasomes and lysosomes degrade the activated erythropoietin receptor, Walrafen [/bib_ref] and controls cellular Epo sensitivity and the level of Epo in the circulation [bib_ref] Erythroid-specific expression of the erythropoietin receptor rescued its null mutant mice from..., Suzuki [/bib_ref] [bib_ref] Cellular trafficking and degradation of erythropoietin and novel erythropoiesis stimulating protein (NESP), Gross [/bib_ref]. Studies in our laboratory have shown that Epo induces internalization of EpoR via CME, and we identified a novel function of p85 in EpoR endocytosis and downregulation [bib_ref] Ligand-induced EpoR internalization is mediated by JAK2 and p85 and is impaired..., Sulahian [/bib_ref] [bib_ref] Cbl ubiquitination of p85 is essential for Epo-induced EpoR endocytosis, Bulut [/bib_ref]. Epo stimulation activates JAK2, resulting in the phosphorylation of multiple EpoR cytoplasmic tyrosine residues, including Y 429 , Y 431 , and Y 479 . These phosphotyrosines serve as mutually redundantly docking sites for binding of the p85 subunit of PI3K to EpoR [bib_ref] Ligand-induced EpoR internalization is mediated by JAK2 and p85 and is impaired..., Sulahian [/bib_ref]. p85 binding activates the catalytic p110 subunit, resulting in PI [bib_ref] Phosphatidylinositol 3-phosphates -at the interface between cell signalling and membrane traffic, Marat [/bib_ref] [bib_ref] The phosphoinositide 3-kinase pathway, Cantley [/bib_ref] P3 production and Akt signaling, which is required for erythroid differentiation. Unexpectedly, Epo-induced EpoR internalization does not require PI3K kinase activity [bib_ref] Ligand-induced EpoR internalization is mediated by JAK2 and p85 and is impaired..., Sulahian [/bib_ref]. Instead, Epodependent ubiquitination of p85 by the E3 ligase c-Cbl recruits the endocytic adaptor protein, Epsin-1, through its UIM. Epsin-1 then connects the EpoR/p85 complex to the clathrin-mediated endocytic machinery for internalization [bib_ref] Cbl ubiquitination of p85 is essential for Epo-induced EpoR endocytosis, Bulut [/bib_ref].
The physiological relevance of this pathway is highlighted by mutated EpoRs found in patients with primary familial and congenital polycythemia (PFCP), a proliferative disorder of the red cell lineage characterized by increased red blood cell mass [bib_ref] Genetically heterogeneous origins of idiopathic erythrocytosis, Percy [/bib_ref] [bib_ref] Advances in understanding the pathogenesis of primary familial and congenital polycythaemia, Huang [/bib_ref]. PFCP patients harbor mutations that delete the C-terminal cyto solic domain of the EpoR, resulting in EpoR truncations lacking all three tyrosines responsible for p85 binding. Mutated EpoRs mimicking those found in PFCP patients cannot bind p85 and are unable to recruit Epsin-1 to engage the endocytic machinery. As a result, these receptor variants do not internalize upon Epo stimulation and exhibit Epo hypersensitivity. Similarly, knockdown of Cbl also causes Epo hypersensitivity in primary erythroid progenitors. Restoring p85 binding to PFCP receptors rescues Epo-induced Epsin-1 co-localization and normalizes Epo hypersensitivity [bib_ref] Ligand-induced EpoR internalization is mediated by JAK2 and p85 and is impaired..., Sulahian [/bib_ref] [bib_ref] Cbl ubiquitination of p85 is essential for Epo-induced EpoR endocytosis, Bulut [/bib_ref]. These results elucidate the molecular mechanism underlying Epo-induced p85-mediated EpoR internalization and demonstrate that defect in this pathway may contribute to the etiology of PFCP. Although still controversial, non-canonical heterodimeric complexes consisting of EpoR and the βc receptor have been implicated in non-hematopoietic tissues [bib_ref] Erythropoietin mediates tissue protection through an erythropoietin and common beta-subunit heteroreceptor, Brines [/bib_ref]. Whether the p85-Cbl pathway plays a role in endocytosis of these complexes is unclear.
PI3K is activated by most cytokine receptors, whereas Cbl also functions downstream of many signaling receptors. Therefore, the p85-Cbl pathway might be utilized more broadly to contribute to endocytosis of other cytokine receptors. In addition, the same molecules may be employed in different ways for receptor endocytosis and downregulation. For example, the thrombopoietin receptor activates PI3K for signaling, and utilizes Cbl for downregulation. However, instead of ubiquitinating p85 as in the case of the EpoR, the thrombopoietin receptor itself is poly-ubiquitinated by Cbl upon stimulation, leading to its degradation [bib_ref] Ubiquitination and degradation of the thrombopoietin receptor c-Mpl, Saur [/bib_ref].
# Conclusion and perspectives
The two new paradigms reviewed here underscore the contribution of PI3K in CME (e.g., EpoR) as well as CIE (e.g., IL2R) of cytokine receptors. Besides class I PI3K discussed here, class II PI3K, which produces PI(3)P and PI(3,4)P2, has also been shown to participate in late stage CME (100). These broaden the roles of PI3K family kinases as fundamental and integral regulators of endocytosis in general.
The mechanisms underlying PI3K's contributions are both kinase activity dependent and -independent. PI3K kinase activity is required to recruit Vav2 and endophilin for IL2R internalization. By contrast, in a PI3K kinase activity-independent manner, p85 recruits activated Rac1 to promote IL2R endocytosis and recruits Cbl/Epsin-1 to promote EpoR internalization. Therefore, PI3K plays both positive and negative roles upon cytokine receptor activation. On the one hand, the PI3K/Akt pathway controls various aspects downstream of cytokine receptors. On the other hand, it stimulates receptor endocytosis and downregulation, thus contributing to signaling attenuation.
These advances also highlight the emerging concept that p85 has functions beyond regulating PI3K kinase activity [bib_ref] Positive and negative roles of p85 alpha and p85 beta regulatory subunits..., Ueki [/bib_ref] [bib_ref] A PI3K activity-independent function of p85 regulatory subunit in control of mammalian..., Garcia [/bib_ref] [bib_ref] Role of the PI3K regulatory subunit in the control of actin organization..., Jimenez [/bib_ref] [bib_ref] Phosphoinositide 3-kinase regulatory subunit p85alpha suppresses insulin action via positive regulation of..., Taniguchi [/bib_ref] [bib_ref] Disrupted RabGAP function of the p85 subunit of phosphatidylinositol 3-kinase results in..., Chamberlain [/bib_ref]. For example, cytokinesis defects observed in p85α-deficient cells are restored by expression of a p85α mutant that does not bind p110 [bib_ref] A PI3K activity-independent function of p85 regulatory subunit in control of mammalian..., Garcia [/bib_ref]. It was also shown that p85 exhibits in vitro GTPaseactivating protein (GAP) activity toward Rab5, which regulates vesicle trafficking and actin remodeling [bib_ref] Rab5 is a signalling GTPase involved in actin remodelling by receptor tyrosine..., Lanzetti [/bib_ref] [bib_ref] Rabs and their effectors: achieving specificity in membrane traffic, Grosshans [/bib_ref]. A p85α mutant with defective GAP activities perturbed PDGF receptor trafficking and caused cellular transformation via a kinase-independent mechanism [bib_ref] Disrupted RabGAP function of the p85 subunit of phosphatidylinositol 3-kinase results in..., Chamberlain [/bib_ref] [bib_ref] Deregulation of Rab5 and Rab4 proteins in p85R274A-expressing cells alters PDGFR trafficking, Chamberlain [/bib_ref]. Whether the GAP activity of p85 or Rab5 contributes to IL2Rβ or EpoR endocytosis is unclear. Moreover, p85 also interacts with dynamin (109), the contribution of this interaction is not known. Other p85-interacting proteins, such as phosphatases (e.g., SHP2) and adaptor proteins (e.g., IRS1), may also contribute to its function [bib_ref] A cross-species study of PI3K protein-protein interactions reveals the direct interaction of..., Breitkopf [/bib_ref] [bib_ref] IRS-1 activates phosphatidylinositol 3'-kinase by associating with src homology 2 domains of..., Myers [/bib_ref].
One last layer of complexity we would like to bring up has risen from recent studies concerning dynamin isoform-specific functions. Normally, vertebrates express three dynamin (Dyn) isoforms: Dyn2 is ubiquitously expressed, whereas Dyn1 and Dyn3 are most highly expressed in specific tissues [bib_ref] Differential distribution of dynamin isoforms in mammalian cells, Cao [/bib_ref] [bib_ref] The dynamins: redundant or distinct functions for an expanding family of related..., Urrutia [/bib_ref]. Under normal conditions, Dyn1 contributes little to CME in non-neuronal cells; however, Reis et al. recently showed that Akt, the canonical kinase downstream of PI3K, activates Dyn1 in epithelial cells to induce accelerated CME with altered dynamics [bib_ref] Crosstalk between Akt/GSK3beta signaling and dynamin-1 regulates clathrin-mediated endocytosis, Reis [/bib_ref]. These results raise the interesting possibility that cytokine receptors may stimulate their endocytosis through Akt-dependent activation of Dyn1, adding to the concept that the endocytic machinery can be specifically adapted by signaling receptors to regulate their own endocytosis. Regulatory controls of endocytic components and mechanisms significantly impact physiology and human diseases. Much of what we know about the cross talk between endocytosis and signaling comes from work done with model receptors such as receptor tyrosine kinases (RTK). Many of these lessons may translate to cytokine receptors, because JAK kinases activate many pathways in common with RTKs. Also, in many cases, JAK kinases are integral partner of cytokine receptors, making receptor/JAK complexes equivalent to RTKs [bib_ref] The N-terminal domain of Janus kinase 2 is required for Golgi processing..., Huang [/bib_ref] [bib_ref] JAKs and cytokine receptors -an intimate relationship, Haan [/bib_ref]. However, signaling is not identical and differences are to be expected. Among the open questions are the following: First, do JAK kinases regulate endocytosis beyond receptor phosphorylation? Can they modulate the endocytic machinery directly? Second, does the PI3K/Akt signaling cascade provide a feedback loop for receptor endocytosis in general? Consistent with this notion, Akt promotes EGF receptor degradation by phosphorylating and activating the PIKfyve kinase (FYVE-containing phosphatidylinositol 3-phosphate 5-kinase), which stimulates vesicle trafficking to lysosomes [bib_ref] AKT facilitates EGFR trafficking and degradation by phosphorylating and activating PIKfyve, Er [/bib_ref]. Third, does the GAP activity of p85 and/or other p85-interacting proteins play a role in cytokine receptor endocytosis?
Fourth, how do cytokine receptors employ the molecular toolbox of signaling and endocytic proteins in different cell types and contexts such as normal vs. disease states? More detailed mechanisms are needed to understand the reciprocal cross talk between endocytosis and signaling, which will help to improve our understanding of the physiological functions of cytokine receptors.
# Author contributions
All the authors contributed to the writing of the review.
# Acknowledgments
The authors are grateful to their colleagues Drs. Sandra Schmid and Peter Michaely, who generously shared their insights. They apologize to the many researchers whose work was not discussed because of space constraints.
# Funding
This study was supported by funding from the National Institutes of Health, USA to LJH (HL089966) and a Taiwan National Science Council Grant (103-2917-I-564-029) to PHC.
## References
|
Adrenocortical Carcinoma Diagnosed by Endoscopic Ultrasound-guided Fine-needle Aspiration
Adrenocortical carcinoma (ACC) is a rare malignancy with a very poor prognosis. A 77-year-old man underwent imaging studies due to poorly controlled hypertension, which revealed a mass measuring 43 mm in diameter near the left adrenal gland. There were no findings indicative of pheochromocytoma. Endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) was performed for the preoperative pathological evaluation, and the findings indicated a possibility of ACC. Based on these results, curative surgery was performed. If the diagnosis of pheochromocytoma is excluded, then EUS-FNA for adrenal lesions is relatively safe. It can also be used for the preoperative diagnosis of ACC.
# Introduction
Adrenocortical carcinoma (ACC) is a rare malignant tumor with a reported frequency of 0.5-2 cases per 1 million individuals annually [bib_ref] Adrenocortical carcinoma, Rodgers [/bib_ref]. It usually occurs in adults in the fourth to fifth decade of life [bib_ref] Long-term survival after complete resection and repeat resection in patients with adrenocortical..., Schulick [/bib_ref]. Additionally, a smaller peak is observed in children aged <5 years [bib_ref] Adrenocortical carcinoma: clinical and laboratory observations, Wajchenberg [/bib_ref]. It occurs more frequently in women than in men; the female to male ratio is between 2.5 and 3 to 1 [bib_ref] Adrenal tumours are more predominant in females regardless of their histological subtype:..., Audenet [/bib_ref]. Approximately 55% of the affected patients present with symptoms indicative of hypersecretion of adrenocortical hormones; 30-40% have symptoms related to the tumor volume, and 10% are asymptomatic [bib_ref] Limited prognostic value of the 2004 International Union Against Cancer staging classification..., Fassnacht [/bib_ref]. The mortality rate associated with ACC is >70% [bib_ref] Adrenal cortical carcinoma, Lack [/bib_ref]. Stages I, II, III and IV accounted for 3%, 29%, 19%, and 50% of the cases, respectively [bib_ref] Adrenal cortical carcinoma. Epidemiology and treatment with mitotane and a review of..., Wooten [/bib_ref]. Approximately 40% of the cases have distant metastases [bib_ref] Extent of disease at presentation and outcome for adrenocortical carcinoma: have we..., Kebebew [/bib_ref]. The median duration of survival is 101 months for stages I and II, whereas it is 15 months for stages III and IV [bib_ref] Long-term survival after complete resection and repeat resection in patients with adrenocortical..., Schulick [/bib_ref]. Advanced cases have a poor prognosis.
ACC is a rare tumor that is sometimes difficult to differentiate when only imaging findings are used. Moreover, owing to its rapid progression, few cases are diagnosed at an operable stage. In this report, we present a case of ACC diagnosed by endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA), which was surgically resectable.
## Case report
The patient was a 77-year-old man who had been treated by another doctor for hypertension for the past 3 years. His hypertension was poorly controlled by oral therapy; therefore, he was referred to a hospital to rule out secondary hypertension. Abdominal computed tomography (CT) revealed a heterogeneous mass, measuring 43 mm in diameter, and abutting the left adrenal gland. The margins of the mass were enhanced on contrast imaging; however, the central part was not enhanced, thus indicating either degeneration or necrosis of the cyst [fig_ref] Figure 1: Computed tomography findings [/fig_ref]. T1-weighted magnetic resonance imaging (MRI) showed that the mass had low signal intensity. It also showed a faint high-signal area within the mass. T2-weighted MRI showed that the mass had a relatively high signal intensity. There was a high-signal region with liquid surface formation inside the mass. Fatsuppressed T2-weighted MRI showed no change in the internal properties of the tumor compared with the normal T2weighted MRI. These results indicated that the lesion was an adrenal mass with cystic degeneration containing blood components [fig_ref] Figure 2: Magnetic resonance imaging [/fig_ref]. The results of endocrinological examinations are shown in [fig_ref] Table 1: Endocrinological Examinations [/fig_ref]. Specifically, under the administration of candesartan cilexeil 12 mg/day, the plasma aldosterone level was 126 pg/mL, the serum potassium level was 3.6 mEq/L, and the plasma renin activity was 0.4 ng/ mL/h. The aldosterone-renin ratio (ARR) was relatively high. However, iodine-131 adsterol scintigraphy showed a normal uptake in both adrenal glands (right, 2.1%; left, 1.4%) [fig_ref] Figure 3: Iodine-131 adsterol scintigraphy findings [/fig_ref]. This suggested that it could be an aldosterone-secreting tumor. The level of serum adrenocorti-cotropic hormone (ACTH) was 10.7 pg/mL and that of serum cortisol was 12.0 μg/dL. A low-dose dexamethasone suppression test revealed subclinical Cushing's syndrome based on the serum ACTH level of 9.8 pg/mL and the unsuppressed serum cortisol level of 7.7 μg/dL. By contrast, the serum dehydroepiandrosterone sulfate (DHEA-S) level was 72 μg/dL, which was within the normal range.
He was referred to our hospital for further diagnosis and treatment. The levels of plasma adrenaline, noradrenaline, and dopamine were 136 pg/mL, 1,100 pg/mL, and 27 pg/ mL, respectively. The level of plasma noradrenaline was relatively high. However, the 24-hour urinary metanephrine was 0.15 mg/day, and normetanephrine was 0.23 mg/day, both of which were not elevated. Iodine-123 adrenal metaiodobenzylguanidine (MIBG) scintigraphy showed no increase in the uptake of this tumor [fig_ref] Figure 4: Iodine-123 adrenal MIBG scintigraphy findings [/fig_ref]. The possibility of pheochromocytoma was considered to be low. The lesion was relatively large with hormone secretion, and there was a possibility of malignancy; therefore, surgery was considered, and a biopsy was performed for the preoperative evaluation. EUS (GF-UCT260; Olympus Optical Tokyo, Japan) detected a 38-mm diameter mass adjacent to the left adrenal gland via a gastric approach. It was not continuous with the tail of the pancreas or the left kidney. The surface was smooth, with cystic degeneration inside; however, there was no obvious calcification. EUS-FNA was performed via the gastric approach, and two passes were made with a 22gauge needle (Acquire; Boston Scientific, Marlborough, USA). Solid lesions of the tumor edges were punctured [fig_ref] Figure 5: Endoscopic ultrasound [/fig_ref]. No fluctuations in blood pressure occurred either during or after the procedure. The histopathological findings were as follows: Atypical cells with round nuclei and mostly eosinophilic cytoplasm were seen in alveolar-like patterns. There were some swollen nuclei; however, the overall nuclear atypia was mild to moderate, and the mitotic image was unclear. Immunohistologically, the tumor cells were positive for steroidogenic factor-1 (SF-1) and negative for S 100 protein, cytokeratin AE1/AE3, and chromogranin A, and the Ki-67 index was approximately 30% [fig_ref] Figure 6: Histological findings of EUS-FNA [/fig_ref]. Based on the above, the possibility of a primary adrenal tumor, especially ACC, was suspected.
Left adrenal resection was performed, and the left adrenal gland and adrenal tumor were excised as a single mass. The excised material was a 9.5-6.5 cm mass containing a cyst with internal hemorrhaging and normal adrenal glands. Histopathologically, there was a diffuse proliferation of atypical cells with eosinophilic cytoplasm in most regions of the tumor. Partially, there were myxoid areas of mucus deposition in the stroma. Overall, the tumor cells had nuclear atypia, and some mitotic images were observed [>5/50 high power field (HPF)], but no atypical mitotic images were observed. Coagulation necrosis was observed in the tumor. The tumor invaded the capsule; however, extracapsular invasion was unclear. Elastica van Gieson (EVG) staining showed venous and sinusoidal capillary invasion; however, no lymphatic invasion was observed. Immunohistologically, SF-1 was positive, and the Ki-67 index was approximately 15% [fig_ref] Figure 7: The excised material and histological findings of surgery [/fig_ref]. Seven of the Weiss criteria (9, 10) [fig_ref] Table 1: Endocrinological Examinations [/fig_ref] were met: nuclear atypia, an increased mitotic rate, the percentage of clear cytoplasm, coagulation necrosis, invasion of venous and sinusoidal structures, and capsular invasion. Considering the presence of mucus in the tumor stroma, it was diagnosed as an ACC of myxoid variant type. The final pathological stage was T2N0M0 stage II. After the surgery, his hormone levels, including aldosterone and renin, normalized [fig_ref] Table 1: Endocrinological Examinations [/fig_ref] , and the antihypertensive medication could also be reduced. We recommended adjuvant therapy with mitotane, but he declined, so he has thereafter been carefully monitored on an outpatient basis.
# Discussion
ACC most commonly presents with features of excessive hormonal secretion or symptoms of compression due to the enlarging mass. An increasing percentage of patients with ACC are diagnosed with incidentaloma during abdominal imaging [bib_ref] Clinical review: adrenocortical carcinoma: clinical update, Allolio [/bib_ref]. In the case of adrenal incidentalomas, it is especially important to differentiate primary benign lesions from malignant lesions and to rule out metastasis for the proper management and staging of each case. Most adrenal incidentalomas are benign, and the determination of the malignant potential of an adrenal mass depends on the size of the lesion, imaging features, and hormonal status [bib_ref] The evaluation of incidentally discovered adrenal masses, Vaidya [/bib_ref]. Tumors measuring greater than 40 mm and heterogeneous enhancement on CT are important discriminators of malignancy in adrenal incidentalomas [bib_ref] Differentiation of malignant from benign adrenal masses: predictive indices on computed tomography, Hussain [/bib_ref]. In this case, the tumor diameter was 43 mm, which was large, and the contrast enhancement on CT was heterogeneous, thus leading to a suspected malignancy. ACC often shows a low-attenuation central region on CT that represents tumor necrosis, irregular contrast enhancement, calcification, and a thin capsule-like margin surrounding the tumor [bib_ref] Primary adrenocortical carcinoma: CT evaluation with clinical correlation, Fishman [/bib_ref]. On ultrasound imaging, ACC usually appears as a rounded or oval well-defined hypoechoic mass, with a few displaying a thick partial or com-plete echogenic rim [bib_ref] Primary adrenocortical carcinoma: sonographic evaluation with clinical and pathologic correlation in 26..., Hamper [/bib_ref]. Although conventional imaging modalities, as mentioned above, are used to rule out adrenal involvement in different malignancies, false-positive and negative-results have been observed in approximately 10% [bib_ref] Clinical impact of endoscopic ultrasound-fine needle aspiration of left adrenal masses in..., Bodtger [/bib_ref]. In addition, ACC has imaging findings similar to those of pheochromocytoma, adrenal adenoma, adrenal metastasis, adrenal lymphoma, ganglioneuroma, adrenal hemorrhaging, adrenal pseudocyst, and adrenal hemangioma [fig_ref] Table 2: Adrenal Lesions That Must Be Differentiated from Adrenocortical Carcinoma [/fig_ref] [bib_ref] Adrenal cortical carcinoma: pathology, genomics, prognosis, imaging features, and mimics with impact..., Ahmed [/bib_ref] [bib_ref] Ultrasonographic findings of 1385 adrenal masses: a retrospective study of 1319 benign..., Gong [/bib_ref] [bib_ref] Sonography of the adrenal glands in the adult, Kim [/bib_ref] [bib_ref] Huge cavernous hemangioma of the adrenal gland: sonographic, computed tomographic, and magnetic..., Xu [/bib_ref] [bib_ref] CT and MRI of adrenal gland pathologies, Wang [/bib_ref] [bib_ref] Imaging features of adrenal masses, Albano [/bib_ref]. Although malignancy was suspected in this case, it was difficult to differentiate ACC from other diseases such as pheochromocytoma, adrenal lymphoma, and adrenal metastasis based on the imaging findings alone. Therefore, endocrinological examinations were necessary. When evaluating some primary adrenal neoplasms, the differential diagnoses can be refined by labora-tory biochemical testing, including serum cortisol, aldosterone, and metanephrine measurements [bib_ref] American Association of Clinical Endocrinologists and American Association of Endocrine Surgeons medical..., Zeiger [/bib_ref]. ACC secretes various hormones, including androgens, cortisol, estrogen, and aldosterone [bib_ref] Primary adrenal malignancy, Reznek [/bib_ref]. Among adult patients with ACC, 30% present with Cushing's syndrome and 20% with virilization. Feminization and hyperaldosteronism are much rarer, accounting for approximately 2% of the cases [bib_ref] Adrenocortical carcinomas: surgical trends and results of a 253-patient series from the..., Icard [/bib_ref]. In the present case, subclinical Cushing's syndrome was revealed by the low-dose dexamethasone suppression test. Although the accumulation of lesions was not clear on iodine-131 adsterol scintigraphy, the plasma aldosterone level and plasma renin activity revealed the aldosterone production capacity of the tumor. However, the DHEA-S level was within the normal limits. Endocrinological examinations were somewhat atypical, but ACC was suspected first, considering the imaging findings. Resection was considered, and a pathological examination was performed to confirm the diagnosis before surgery.
Adrenal FNA for the diagnosis of predominantly solid masses gained attention in the 1970s and has been increasingly used in clinical practice following improvements in imaging technologies and deep-seated FNA techniques [bib_ref] Fine needle aspiration cytology in the diagnosis of solid renal and adrenal..., Orell [/bib_ref]. For decades, adrenal FNA has been primarily performed by interventional radiologists using percutaneous approaches guided by CT. In recent years, EUS-FNA advancements have provided an alternative approach for the biopsy of adrenal mass lesions. EUS-FNA sampling of the adrenal gland started in the 1990s and has become more common with improvements in technology and equipment [bib_ref] Endoscopic ultrasound (EUS) and EUS-guided fine needle aspiration of the left adrenal..., Chang [/bib_ref]. EUS is useful as a diagnostic imaging modality because of its superior resolution, and it also provides direct access to the left and sometimes to the right adrenal glands. Complications of FNA in the adrenal glands, such as adrenal hematoma, abdominal pain, adrenal abscess formation, and tumor recurrence along the needle track, have been reported but are infrequent [bib_ref] Image-guided percutaneous biopsy of the adrenal gland: review of indications, technique, and..., Arellano [/bib_ref]. In a review of 416 patients who underwent EUS-FNA of the adrenal glands, no major complications were reported, except for one case of adrenal hemorrhaging [bib_ref] Endoscopic ultrasound-guided fine-needle aspiration in the diagnosis of adrenal lesions, Patil [/bib_ref]. EUS-FNA of the adrenal gland can be performed safely. However, if the targets of the puncture are pheochromocytomas, extreme caution is required because it might cause hemorrhaging and hypertensive crisis [bib_ref] Unsuspected pheochromocytoma: risk of blood-pressure alterations during percutaneous adrenal biopsy, Casola [/bib_ref]. When performing EUS-FNA for adrenal tumors, it is important to rule out pheochromocytoma. In addition, we always have phentolamine, an alpha-blocker, on standby, to deal with a hypertensive crisis if it occurs. In this case, the iodine-123 adrenal MIBG scintigraphy showed no accumulation in the lesion, and 24-hour urine metanephrine and normetanephrine levels were within the normal limits, suggesting a low probability of pheochromocytoma. However, we took all possible preparations for the examination and were able to perform the procedure safely.
Pathologically, ACC is one of the most difficult tumors to differentiate benign disease from malignancy, and it is often impossible to distinguish a benign tumor from amalignant tumor using only commonly used indices such as nuclear atypia and vascular invasion. A scoring system that combines multiple indices is recommended for the diagnosis of ACC, and the Weiss criteria are the most commonly used (9, 10). It is useful in that it is simple and can diagnose ACC based on the pathological findings alone. ACC can be diagnosed using EUS-FNA; however, it is important to understand that there are certain limitations to this diagnosis. To diagnose ACC accurately, an evaluation of the entire lesion after surgery is necessary. However, EUS is a useful tool to confirm the location of the tumor and adjacent organs, and EUS-FNA allows for pathological evaluations including immunostaining. ACC is often difficult to differentiate from other diseases based on imaging, as it sometimes shows endocrinologically atypical characteristics. Preoperative pathological evaluations are important because it can identify benign and malignant tumors and exclude other malignant tumors such as pheochromocytoma, adrenal metastasis, and adrenal lymphoma. Therefore, it is worthwhile to consider tissue sampling by EUS-FNA before performing surgical resection. ACC is a rare tumor, and there are few reports of cases diagnosed by EUS-FNA. To the best of our knowledge, only three cases of ACC diagnosed using EUS-FNA have so far been reported [bib_ref] Sampling of the adrenal glands by endoscopic ultrasoundguided fine-needle aspiration, Stelow [/bib_ref] [bib_ref] A rare cause of chronic diarrhea diagnosed by endoscopic ultrasound-guided fine-needle aspiration, Dhaliwal [/bib_ref] [bib_ref] Role of endoscopic ultrasound-guided fine-needle aspiration in adrenal lesions: analysis of 32..., Gupta [/bib_ref]. This case is valuable because the ACC was diagnosed using EUS-FNA and then surgically resected.
We encountered a rare case of ACC diagnosed using EUS-FNA. When performing EUS-FNA for adrenal tumors, it is important to rule out pheochromocytoma in advance. There were no major complications at the time of examination, and complete resection was possible based on the diagnosis. An early diagnosis of ACC and surgery is important for improving the prognosis. In conclusion, EUS-FNA is therefore considered to be useful for the preoperative diagnosis of ACC.
The authors state that they have no Conflict of Interest (COI).
[fig] Figure 1: Computed tomography findings. Abdominal plain computed tomography showing a mass measuring 43 mm in diameter in contact with the left adrenal gland (arrow) (a). The margins of the mass were enhanced on contrast; however, the central part was not enhanced, indicating cystic degeneration or necrosis (arrow) (b). [/fig]
[fig] Figure 2: Magnetic resonance imaging (MRI) findings. T1-weighted MRI showed that the mass had low signal intensity. It also contained a faint high-signal area within the mass (arrow) (a). T2-weighted MRI showed that the mass had a relatively high-signal intensity. Inside the mass, there was a region of high-signal intensity with liquid surface formation (arrow) (b). The fat-suppressed T2-weighted MRI showed no change in the internal properties of the tumor compared to the normal T2-weighted MRI (arrow) (c). [/fig]
[fig] Figure 3: Iodine-131 adsterol scintigraphy findings. Iodine-131 adsterol scintigraphy showed normal uptake in both adrenal glands (right, 2.1%; left, 1.4%). a: Anterior, b: Posterior. [/fig]
[fig] Figure 4: Iodine-123 adrenal MIBG scintigraphy findings. Iodine-123 adrenal MIBG scintigraphy showed no increase in the uptake of this tumor. a: Anterior, b: Posterior. [/fig]
[fig] Figure 5: Endoscopic ultrasound (EUS) findings. EUS via a gastric approach detected a 38 mm diameter mass adjacent to the left adrenal gland. The surface was smooth with cystic degeneration; however, there was no obvious calcification (a). EUS-FNA was performed via the gastric approach, and two passes were made using a 22-gauge needle (b). [/fig]
[fig] Figure 6: Histological findings of EUS-FNA. Atypical cells with round nuclei and mostly eosinophilic cytoplasm were seen in alveolar-like patterns (Hematoxylin and Eosin staining, ×200) (a). Immunohistologically, the tumor cells were positive for steroidogenic factor-1 (SF-1) (b), negative for S100 protein (c), cytokeratin AE1/AE3 (d), and chromogranin A (e), and a Ki-67 index of approximately 30% (×200) (f). [/fig]
[fig] Figure 7: The excised material and histological findings of surgery. The excised material consisted of a 9.5×6.5 cm mass containing a cyst with internal hemorrhage and normal adrenal glands (a). Histopathologically, there was a diffuse proliferation of atypical cells with eosinophilic cytoplasm in most areas of the tumor (Hematoxylin and Eosin staining, ×40) (b). Partially, there were myxoid areas of mucus deposition in the stroma [enlarged image of the yellow rectangle in (b). ×200] (c). Immunohistologically, the Ki-67 index was approximately 15% (×200) (d). [/fig]
[table] Table 1: Endocrinological Examinations [/table]
[table] Table 2: Adrenal Lesions That Must Be Differentiated from Adrenocortical Carcinoma (18-23). [/table]
|
What Controls DNA Looping?
The looping of DNA provides a means of communication between sequentially distant genomic sites that operate in tandem to express, copy, and repair the information encoded in the DNA base sequence. The short loops implicated in the expression of bacterial genes suggest that molecular factors other than the naturally stiff double helix are involved in bringing the interacting sites into close spatial proximity. New computational techniques that take direct account of the three-dimensional structures and fluctuations of protein and DNA allow us to examine the likely means of enhancing such communication.Here, we describe the application of these approaches to the looping of a 92 base-pair DNA segment between the headpieces of the tetrameric Escherichia coli Lac repressor protein.The distortions of the double helix induced by a second protein-the nonspecific nucleoid protein HU-increase the computed likelihood of looping by several orders of magnitude over that of DNA alone. Large-scale deformations of the repressor, sequence-dependent features in the DNA loop, and deformability of the DNA operators also enhance looping, although to lesser degrees. The correspondence between the predicted looping propensities and the ease of looping derived from gene-expression and single-molecule measurements lends credence to the derived structural picture.OPEN ACCESSInt. J. Mol. Sci. 2014, 15 15091
# Introduction
The cellular environment introduces remarkable changes in DNA. The long, stiff, naturally straight, double-helical molecule appears to be much more deformable in vivo than in vitro. For example, the control of lactose metabolism in Escherichia coli entails the formation of DNA loops much shorter in length than those detected in aqueous salt solution. DNA tends to be highly extended in solution, with very low chances of closing into a loop having the 92 base-pair (bp) spacing found between the centers of the so-called O [bib_ref] Flexible DNA bending in HU-DNA cocrystal structures, Swinger [/bib_ref] and O 1 lac operator sequences. Formation of such a loop, against the DNA binding headpieces of the Lac repressor protein, is thought to preclude access to the genetic signals needed to transcribe proteins involved in the transport and chemical breakdown of lactose.
The observed expression of the Escherichia coli lac genes thus reflects more than the mechanical properties of DNA. Key molecular components include the aforementioned repressor protein, an assembly of four identical polypeptide chains that bind specifically to the noted operators, and the naturally abundant, histone-like, architectural protein HU. The latter protein binds DNA non-specifically [bib_ref] Nonspecific DNA binding and bending by HUαβ: Interfaces of the three binding..., Koh [/bib_ref] , introduces some of the largest known protein-induced deformations of the double helix [bib_ref] Flexible DNA bending in HU-DNA cocrystal structures, Swinger [/bib_ref] , and stabilizes the formation of biologically functional loops as small as ~50 bp [bib_ref] Lac repressor forms loops with linear DNA carrying two suitably spaced lac..., Krämer [/bib_ref] [bib_ref] Gene repression by minimal lac loops in vivo, Bond [/bib_ref]. The Lac repressor, by contrast, binds to specific operator sites on DNA, introduces lesser distortions in the double helix [bib_ref] Crystal structure of the lactose operon repressor and its complexes with DNA..., Lewis [/bib_ref] , and interconverts between the tightly closed, V-shaped protein architecture observed in the crystalline state and an extended open form with widely separated DNA-recognition headpieces [bib_ref] Conformation of Lac repressor tetramer in solution, Ruben [/bib_ref] [bib_ref] Ligand-induced conformational changes and conformational dynamics in the solution structure of the..., Taraban [/bib_ref]. Moreover, the protein-bound DNA operators undergo sizable deformations with respect to the protein recognition elements, particularly at the 3'-end of the O 3 operator [bib_ref] Weak operator binding enhances simulated Lac repressor-mediated DNA looping, Colasanti [/bib_ref] , and the headpieces themselves may wobble with respect to the globular protein core [bib_ref] Structural dynamics of the Lac repressor-DNA complex revealed by a multiscale simulation, Villa [/bib_ref]. Signals in the DNA loop may also contribute to loop formation. For example, the two hexameric promoter elements, located in the middle and near the 3'-end of the 92-bp loop, include pyrimidine-purine base-pair steps (TA:TA and CA:TG) known to be highly deformable and subject to large-scale conformational rearrangements [bib_ref] DNA sequence-dependent deformability deduced from protein-DNA crystal complexes, Olson [/bib_ref] [bib_ref] A novel roll-and-slide mechanism of DNA folding in chromatin: Implications for nucleosome..., Tolstorukov [/bib_ref]. The site of catabolic activator protein (CAP) binding, found at the 5'-end of the loop, also includes the same pyrimidine-purine elements. The precise orientation of the DNA operators on the protein assembly remains an open question. The operators at the ends of the loops may point toward either the inside or the outside of the repressor, yielding two classes of loops where the operators run in parallel directions and two where the operators run in antiparallel directions [bib_ref] Gal repressosome contains an antiparallel DNA loop, Geanacopoulos [/bib_ref].
The treatment of protein-bound DNA molecules at the level of successive base-pair steps makes it possible to examine the contributions of HU binding, Lac repressor mobility, base-pair sequence, and operator orientation on DNA loop formation. The present article summarizes the computational approaches that we have developed to capture the structures of short DNA loops bound to the headpieces of the Lac repressor protein and presents new insights into protein-mediated DNA looping gained from these studies. The correspondence between the predicted looping propensities and the ease of looping derived from gene-expression [bib_ref] Gene repression by minimal lac loops in vivo, Bond [/bib_ref] [bib_ref] Bacterial repression loops require enhanced DNA flexibility, Becker [/bib_ref] [bib_ref] Effects of nucleoid proteins on DNA repression loop formation in Escherichia coli, Becker [/bib_ref] and single-molecule [bib_ref] Concentration and length dependence of DNA looping in transcriptional regulation, Han [/bib_ref] [bib_ref] Sequence dependence of transcription factor-mediated DNA looping, Johnson [/bib_ref] measurements lends credence to the derived structural picture. Both protein and DNA play critical roles in the simulated looping. Trace amounts of randomly bound HU, deformability in the Lac repressor, and localized flexibility in the repressor-anchored DNA offer molecular rationales that help to account for the unexpectedly short DNA loops implicated in bacterial gene regulation. . Simulated effects of HU on the ease of looping short pieces of DNA between the headpieces of a V-shaped Lac repressor protein assembly. (a) Schematic representation of the molecular components and resulting HU-decorated DNA loops anchored in antiparallel (A1, A2) and parallel (P1, P2) orientations [bib_ref] Gal repressosome contains an antiparallel DNA loop, Geanacopoulos [/bib_ref] on the repressor; (b) Predicted values of the J factor, as a function of operator spacing, for double-helical molecules free of HU or binding one HU on average every 150 or 1000 bp compared, on the top left, with values deduced from the expression of lac genes in Escherichia coli cells containing the wild-type (WT) protein and a mutated strain (∆HU) that cannot express HU [bib_ref] Gene repression by minimal lac loops in vivo, Bond [/bib_ref] [bib_ref] Bacterial repression loops require enhanced DNA flexibility, Becker [/bib_ref] [bib_ref] Effects of nucleoid proteins on DNA repression loop formation in Escherichia coli, Becker [/bib_ref] and, on the lower right, with values deduced from tethered particle motion studies of DNA constructs (E8-09, E8-12) flanked at the 5'-end by a symmetric operator and the 3'-end by the natural O 1 operator [bib_ref] Concentration and length dependence of DNA looping in transcriptional regulation, Han [/bib_ref] [bib_ref] Sequence dependence of transcription factor-mediated DNA looping, Johnson [/bib_ref]. Computed J factors are connected by thin lines. Values deduced from experiments are shown by symbols. Vertical lines denote integral helical repeats of DNA.
# Results and discussion
## Contributions of hu to dna looping
The random binding of HU has striking effects on the looping propensities of short (70-90 bp) DNA duplexes anchored to the V-shaped Lac repressor assembly. Binding of the architectural protein at the level of only one dimer per 1000 bp of DNA increases the calculated J factor, i.e., the relative ease of looping, by two to five orders of magnitude over that predicted for ideal protein-free chains of the same length . Moreover, in contrast to protein-free DNA, where the chances of looping exhibit a striking dependence on the spacing between operator sites, the J factors of short HU-bound loops show limited variation over the same range of chain lengths. Increasing the concentration of HU to one HU dimer per 150 bp of DNA-corresponding to the level of protein present during the exponential growth phase of Escherichia coli, i.e., ~30,000 HU dimers [bib_ref] Growth phase-dependent variation in protein composition of the Escherichia coli nucleoid, Azam [/bib_ref] in the presence of a 4.6 Mbp genome [bib_ref] The complete genome sequence of Escherichia coli K-12, Blattner [/bib_ref] -dampens the amplitude of oscillations in the J factor with chain length and introduces secondary peaks in the looping profile. The latter peaks stem from the enhanced build-up of HU on loops of certain sizes with increase in binding levels (see below). The simulations performed at the higher concentration of HU capture the peaks and valleys in the looping propensities detected in gene-expression studies but overestimate the magnitudes of the J factors deduced from the experiments (points labeled WT in [bib_ref] Gene repression by minimal lac loops in vivo, Bond [/bib_ref] [bib_ref] Bacterial repression loops require enhanced DNA flexibility, Becker [/bib_ref] [bib_ref] Effects of nucleoid proteins on DNA repression loop formation in Escherichia coli, Becker [/bib_ref]. The loops computed at the lower HU level fall within the range of observation but show greater variation with chain length than the in vivo data. The predicted dampening and phase shift in the chain-length dependent variation of the J factor upon addition of HU match trends detected in tethered-particle motion studies of longer Lac repressor-mediated loops [bib_ref] DNA sequence-dependent mechanics and protein-assisted bending in repressor-mediated loop formation, Boedicker [/bib_ref]. The reported twofold enhancement in looping propensities in the latter work, however, falls short of the simulated increase in the J factor and the values deduced from gene-expression studies.
The computations performed in the absence of HU show both the decrease in the J factors derived from gene-expression studies of mutant cells that do not express HU (data labeled ∆HU in and the observed oscillatory variation in loop formation with chain length [bib_ref] Gene repression by minimal lac loops in vivo, Bond [/bib_ref] [bib_ref] Bacterial repression loops require enhanced DNA flexibility, Becker [/bib_ref] [bib_ref] Effects of nucleoid proteins on DNA repression loop formation in Escherichia coli, Becker [/bib_ref]. The computed looping profiles, however, fall substantially below those found upon deletion of the HU gene, and the amplitude of the oscillations in the computed J factor of HU-free DNA greatly exceeds that deduced from experiment. On the other hand, the predicted difficulty in forming Lac repressor-mediated loops in the absence of HU lies within the range of values deduced from tethered particle motion studies of slightly longer loops (110-130 bp) [bib_ref] Concentration and length dependence of DNA looping in transcriptional regulation, Han [/bib_ref] [bib_ref] Sequence dependence of transcription factor-mediated DNA looping, Johnson [/bib_ref]. The system examined in vitro excludes HU and any uncharacterized cellular factors that may contribute to the higher J factors found in vivo. The order-of-magnitude differences in the reported looping propensities, collected at different times with the same DNA construct (points labeled E8-09 and E8-12 in , are thought to reflect the different source of repressor protein in the two sets of experiments. The earlier data (E8-09) coincide with the predicted curve. The later data (E8-12) show lesser variation with chain length and generally fall below the predicted J factors.
The HU concentration reported for the simulations refers to the probability that one HU dimer is taken up, on average, over a given length of linear unconstrained DNA (150 and 1000 bp in . The number of HU dimers accumulated on the DNA loops differs at low and high concentrations of protein and at different chain lengths. The loops take up more HU than the imposed binding levels at all modeled concentrations, and the uptake is greater at chain lengths where the DNA operators are out of phase with the binding sites on the repressor protein. Whereas most 109-bp loops take up a single HU dimer under binding conditions of one HU per 150 bp, the majority of HU-decorated loops of 115 bp take up two architectural proteins . Moreover, the HU-decorated chains anchor to the repressor in different orientations and the build-up of protein is localized. The more easily closed 109-bp chains attach to the repressor in antiparallel orientations and the less easily formed 115-bp loops in parallel orientations. The second HU, also detected in tethered-particle motion studies of less easily close loops [bib_ref] DNA sequence-dependent mechanics and protein-assisted bending in repressor-mediated loop formation, Boedicker [/bib_ref] , helps to align the ends of the DNA with the recognition headpieces on the repressor. The selective accumulation of protein at different chain lengths underlies the dampening of the oscillations in the computed J factors of HU-bound DNA compared to those of bare DNA .
## Effects of the lac repressor on dna loops
Surprisingly, the J factors found to characterize the looping of short, ideal DNA chain fragments between the headpieces of the Lac repressor exhibit a local minimum at 92 bp, the natural spacing between the O 3 and O 1 lac operators (data labeled-HU in . The low probability of forming such a loop reflects both the cost of bending naturally straight DNA in a closed configuration and the twisting needed to bring the terminal bases in correct register with the headpieces of the protein.
The precise contributions to the energy depend upon how DNA is oriented on the protein [fig_ref] Table 1: Elastic energy contributions and populations of 92-bp Lac repressor-mediated DNA loops [/fig_ref]. For example, the bending penalty (E bend ) is lower if the chain attaches to the V-shaped assembly in an antiparallel as opposed to a parallel orientation. Loops of the former type describe gradual U-turns with less straightening of DNA compared to that found in the presence of HU [fig_ref] Figure 3: Effect of the Lac repressor on the minimum-energy configurations of 92-bp DNA... [/fig_ref]. The ease of twisting the DNA (E twist ), however, offsets the cost of bending in one of the two possible parallel loops, the so-called P1 form, in which the chain adopts a more symmetric configuration with a tight turn near the midpoint of the loop. The similar elastic energies of the three kinds of bare DNA loops underlie the nearly even mix of closed forms captured in both Monte Carlo calculations and energy optimization. Occurrences of "wild-type" loops in the alternate P2 parallel form drop precipitously in the absence of HU and are therefore not discussed. The slight discrepancies among the populations of DNA looped forms identified with the two techniques stem, at least in part, from the difficulty in capturing configurational states that meet specific geometric criteria with Monte Carlo sampling. Examples of how slight perturbations of the imposed positions of DNA operators affect the optimized elastic energies of Lac repressor-mediated loops reveal how the approximations associated with random sampling may influence the predicted configurations of the looped structures. Here, the V-shaped protein opens or closes through small changes Δα in the virtual valence angle between the two arms [fig_ref] Figure 3: Effect of the Lac repressor on the minimum-energy configurations of 92-bp DNA... [/fig_ref]. Increments of ±8° limit the movements of the attached DNA to changes comparable to the range of chain displacements allowed in the Monte Carlo search-namely, deviations no more than 15 Å in the distance between terminal base pairs and fluctuations of 11.5° or less in both the global bend angle and the net twist angle. Indeed, these small perturbations in anchoring conditions have marked effects on the relative energies of loops attached in antiparallel versus parallel orientations on the repressor, with the former loops becoming highly favored upon slight closure of the protein and the latter upon slight opening. The changes in protein structure, however, have no noticeable effect on the overall configuration of the DNA-repressor assembly, including the contact interface thought to stabilize the interaction between the two protein arms [bib_ref] Modeling the Lac repressor-operator assembly: The influence of DNA looping on Lac..., Swigon [/bib_ref].
The general correspondence between the optimized energies, i.e., relative statistical weights, of short DNA loops, and the probabilities of linear chain molecules satisfying the same spatial constraints allows us to consider a variety of molecular questions impractical to address with Monte Carlo methods. For example, it is difficult to relate the likelihood of rare configurational events, such as the formation of 92-bp Lac repressor-mediated loops, to specific perturbations in protein structure. We next illustrate how large changes in the angle of opening Δα between the arms of the repressor affect the minimum-energy configurations of DNA loops and how loop orientation influences the degree of protein deformation [fig_ref] Figure 4: Changes in DNA looping associated with large-scale opening of the Lac repressor [/fig_ref]. If fully opened, the long axes of the protein arms would lie roughly perpendicular to the axis of the connecting four-helix bundle. The constraints of DNA deformation, however, limit the degree of protein opening, with loops attached in parallel orientations more easily accommodated on the opened protein than those bound in antiparallel orientations. The changes in anchoring conditions brought about by the movement of protein lower the cost of DNA deformation, particularly the twisting contribution, in the parallel loops. A reduction in free energy comparable to the 7k B T difference in minimum energy found for loops attached to the opened versus closed protein would increase the ease of looping by roughly three orders of magnitude. By contrast, the opening of the repressor introduces a twisting penalty in loops held in antiparallel orientations and the predicted enhancement of looping is only about threefold. The 92-bp loops captured in Monte Carlo simulations bind on average to even more opened forms of the protein than predicted along the minimized energy pathway and raise the J factor ~7000-fold [bib_ref] Interplay of protein and DNA structure revealed in simulations of the lac..., Czapla [/bib_ref]. That is, the arrangements of the repressor in the latter studies include states not considered in the energy optimization.
## Influences of sequence and environment on looped structures
The softening of DNA at the sites of pyrimidine-purine steps along the lac operon, through decrease in selected elastic constants (see Experimental Section), reduces the energy but has little effect on the overall optimized shapes of 92-bp DNA loops constrained to fit against the headpieces of the V-shaped Lac repressor assembly (blue versus gold DNA pathways in [fig_ref] Figure 5: Changes in 92-bp DNA looping brought about by localized changes in the... [/fig_ref]. Although the DNA bends to a greater extent at the soft steps (Δγ values in [fig_ref] Figure 5: Changes in 92-bp DNA looping brought about by localized changes in the... [/fig_ref] , the remaining residues straighten relative to those found at corresponding sites along the ideal DNA loop. The total twist of the modified chain, i.e., the sum of the plotted Δθ values, also drops slightly, with the soft steps uniformly undertwisted and the remaining steps uniformly over-twisted compared to ideal DNA. As expected for a spatially constrained DNA [bib_ref] Calculation of the twist and the writhe for representative models of DNA, White [/bib_ref] , the changes in twist along the locally softened loops are coupled to changes in bending and to the overall configurations of the loops. Individual base pairs move on average by 0.6-0.9 Å and up to 2 Å, depending on loop type, and the writhing numbers, determined along the closed pathways formed by connecting the terminal base pairs with a straight line, differ only slightly (~0.001). The computed decrease in total energy (~6k B T for each of the three looped states) points to an approximately 100-fold increase in the ease of looping via the assumed softening mechanism. By contrast, the overtwisting of DNA, such as found upon the reduction of temperature [bib_ref] Conformational fluctuations of DNA helix, Depew [/bib_ref] or the uptake of salt [bib_ref] Supercoiling in closed circular DNA: Dependence upon ion type and concentration, Anderson [/bib_ref] , has a dramatic effect on the configurations of certain loops (green images in [fig_ref] Figure 5: Changes in 92-bp DNA looping brought about by localized changes in the... [/fig_ref]. The increase of intrinsic twist to 36°, corresponding to a 10-bp helical repeat, displaces points along the optimized antiparallel loops by 27 Å on average and as much as 62 Å from those on the pathways of the DNA loops with an intrinsic 10.5-bp repeat. The relative displacement of base pairs is an order-of-magnitude smaller along the parallel loop, where the lesser change in shape lowers the writhing number by 0.025 (compared to reductions of ~0.25 for the A1 and A2 loops). The uniform uptake of twist along the loops (∆θ values in [fig_ref] Figure 5: Changes in 92-bp DNA looping brought about by localized changes in the... [/fig_ref] follows the expected behavior of an ideal elastic rod. An increase in energy comparable to the difference in computed minimum-energy values (~18k B T) would decrease the likelihood of loop formation by about eight orders of magnitude. A decrease in intrinsic twist should facilitate loop formation.
## Effects of operator structure on looping propensities
The introduction on the V-shaped Lac repressor of various operator structures, representative of the spatial pathways found in NMR studies of different DNA sequences in the presence of the protein headpieces [bib_ref] The solution structure of Lac repressor headpiece 62 complexed to a symmetrical..., Spronk [/bib_ref] [bib_ref] Specificity and affinity of Lac repressor for the auxiliary operators O2 and..., Romanuka [/bib_ref] , also enhances the formation of 92-bp loops. Superposition of the NMR models in a common coordinate frame reveals distortions in DNA, with the double helix flexing to different degrees and in different directions [fig_ref] Figure 6: Relative ease of forming 92-bp DNA loops anchored by various Lac repressor-bound... [/fig_ref]. The changes in overall bending with changes in operator sequence and the variation in apparent DNA deformability stand out in the similarly oriented molecular images. The ends of the DNA structures deduced from NMR measurements of the weak and much less curved O 3 operator exhibit wide variations. The models of the primary O 1 operator and the synthetic symmetric O sym operator, by contrast, show limited deformations in overall structure. The differences among the pathways of the lac operators have striking effects on the predicted ease of DNA loop formation [fig_ref] Figure 6: Relative ease of forming 92-bp DNA loops anchored by various Lac repressor-bound... [/fig_ref]. The computed J factors of the DNA loops anchored to all 121 possible combinations of O sym operator models roughly match the J factor, 3.1 × 10 −11 , found for the 92-bp loops formed between rigid O sym operators on the same Lac repressor assembly. About half of the operator pathways enhance loop closure compared to that found with the rigid operators and about half suppress looping. The J factor increases more than fivefold for a few combinations of operator models and drops, in other cases, to roughly half the value found when the O sym operators are rigid.
The effect of end conditions on loop formation is even more striking for 92-bp loops closed between O 3 and O 1 operators. The J factors of loops anchored to any of the 200 O 3 ···O 1 model combinations exceed the value found with rigid O sym operators [fig_ref] Figure 6: Relative ease of forming 92-bp DNA loops anchored by various Lac repressor-bound... [/fig_ref]. Moreover, the looping enhancement is 20-30 times larger than that from the latter simulations for DNA fragments attached to about two thirds of the operator models and nearly two orders of magnitude larger for loops formed with a quarter of the models. The predicted enhancement of looping is tied to specific operator pathways, notably the straightened O 3 pathways where DNA peels away from the repressor [bib_ref] Weak operator binding enhances simulated Lac repressor-mediated DNA looping, Colasanti [/bib_ref].
## Experimental section
## Protein-decorated dna model
This work takes advantage of methods that we have developed to study the properties of spatially constrained DNA molecules, including Lac repressor-mediated DNA loops. Models of DNA are constructed, one base-pair step at a time, from sets of rigid-body parameters that specify the three-dimensional arrangements of successive base pairs [bib_ref] 3DNA: A software package for the analysis, rebuilding and visualization of three-dimensional..., Lu [/bib_ref] [bib_ref] 3DNA: A versatile, integrated software system for the analysis, rebuilding, and visualization..., Lu [/bib_ref]. Protein-free segments are subject to a potential that allows for elastic deformations of the long, thin molecule from the canonical B-DNA structure [bib_ref] Sequence-dependent effects in the cyclization of short DNA, Czapla [/bib_ref]. The base-pair steps are assigned the elastic properties of an ideal, inextensible, naturally straight DNA helix, with bending deformations (±4.84°) corresponding to a persistence length of nearly 500 Å, fluctuations in twist (±4.09°) compatible with the topological properties of DNA minicircles (that is, twisting 1.4 times more restricted than bending) [bib_ref] Torsional rigidity of DNA and length dependence of the free energy of..., Horowitz [/bib_ref] [bib_ref] Effect of bending strain on the torsion elastic constant of DNA, Heath [/bib_ref] , and a helical rest state with 10.5 bp per turn. The enhanced deformability of the pyrimidine-purine steps introduced in selected loops is described by a model that allows for angular fluctuations ~1.5 times those available to the other base-pair steps, i.e., root-mean-square deviations in the bending and twisting components ~1.5 times those assigned to the canonical helix.
The DNA attached to the Lac repressor is defined in terms of the rigid-body parameters of the base pairs found in known operator-bound structures (described below). The HU-bound fragments are assigned the sets of parameters that describe the pathways of DNA adopted in the four crystal complexes with the Anabaena protein [bib_ref] Flexible DNA bending in HU-DNA cocrystal structures, Swinger [/bib_ref] and are introduced at random along the looped DNA [bib_ref] Interplay of protein and DNA structure revealed in simulations of the lac..., Czapla [/bib_ref]. The probability that a site is occupied depends upon the concentration of free HU, its DNA association constant, and the concentration of DNA base pairs. The DNA is examined, one base pair at a time, in random order, and a decision is made whether or not to place the protein. If the HU is placed, the 14 base-pair steps that make up the binding site are assigned the step parameters in one of the reported crystal structures and the protein is incorporated as a rigid "side group" of the DNA [bib_ref] Effects of the nucleoid protein HU on the structure, flexibility, and ring-closure..., Czapla [/bib_ref].
The Lac repressor is subjected to large-scale motions that preserve its globular elements and follow directions consistent with experimental observations [bib_ref] Conformation of Lac repressor tetramer in solution, Ruben [/bib_ref] [bib_ref] Ligand-induced conformational changes and conformational dynamics in the solution structure of the..., Taraban [/bib_ref]. The opening of the repressor to an extended state is affected via a rotation of one of the two large globular arms of the tetrameric assembly about a "bending" axis perpendicular to the long axes of the two arms and passing through the point of closest approach between those axes [fig_ref] Figure 3: Effect of the Lac repressor on the minimum-energy configurations of 92-bp DNA... [/fig_ref]. The disordered state of the polypeptide backbone in this region of the structure lends support to the idea that the protein may deform by such a mechanism. The reference state is a V-shaped model of the full protein-DNA assembly constructed from known crystallographic components [bib_ref] Weak operator binding enhances simulated Lac repressor-mediated DNA looping, Colasanti [/bib_ref] [bib_ref] Modeling the Lac repressor-operator assembly: The influence of DNA looping on Lac..., Swigon [/bib_ref] , e.g., superposition of the protein atoms in one of the dimeric DNA-bound repressor structures [bib_ref] A closer view of the conformation of the Lac repressor bound to..., Bell [/bib_ref] [bib_ref] Crystallographic analysis of Lac repressor bound to natural operator O1, Bell [/bib_ref] on the corresponding atoms in the tetrameric form of the protein without DNA-binding headpieces [bib_ref] Crystal structure of the lactose operon repressor and its complexes with DNA..., Lewis [/bib_ref]. Here, for simplicity, the opening of the repressor is assumed to be "free", with no energy penalty associated with the disruption of the small contact interface believed to stabilize the V-shaped form [bib_ref] Modeling the Lac repressor-operator assembly: The influence of DNA looping on Lac..., Swigon [/bib_ref]. Introduction of a penalty term proportional to the surface area of the contact interface in the closed complex does not change the general findings.
The subtleties in structure associated with the binding of different DNA sequences are treated by superimposing the solution structures of the Lac repressor headpiece bound to different operators [bib_ref] The solution structure of Lac repressor headpiece 62 complexed to a symmetrical..., Spronk [/bib_ref] [bib_ref] Specificity and affinity of Lac repressor for the auxiliary operators O2 and..., Romanuka [/bib_ref] on V-shaped models of the full protein-DNA assembly [bib_ref] Weak operator binding enhances simulated Lac repressor-mediated DNA looping, Colasanti [/bib_ref] [bib_ref] Modeling the Lac repressor-operator assembly: The influence of DNA looping on Lac..., Swigon [/bib_ref]. Given the uncertainty in the orientation of the headpieces on the tetramer, the complexes are placed in two directions on each arm of the V-shaped models, with the 5'-3' direction of the operator pointing toward either the inside or the outside of the full structure. The GC base pairs at the centers of the natural (O 1 and O 3 ) operator-headpiece complexes [bib_ref] Specificity and affinity of Lac repressor for the auxiliary operators O2 and..., Romanuka [/bib_ref] overlap the corresponding base pair in the O 1 -containing crystal structure [bib_ref] Crystallographic analysis of Lac repressor bound to natural operator O1, Bell [/bib_ref] , and the highly kinked CG base-pair steps at the centers of the symmetric (O sym ) operator complexes [bib_ref] The solution structure of Lac repressor headpiece 62 complexed to a symmetrical..., Spronk [/bib_ref] coincide with the corresponding step in the O sym crystal complex [bib_ref] A closer view of the conformation of the Lac repressor bound to..., Bell [/bib_ref]. The perturbations in these systems affect the positions and orientations of the DNA at the two ends of the tethered loops and the ease of accommodating an intervening length of DNA. The rigid DNA operators introduced in other computations follow the pathway taken by the symmetric operator in the crystalline state [bib_ref] A closer view of the conformation of the Lac repressor bound to..., Bell [/bib_ref].
## Looping simulations
The ends of the protein-bound operator fragments are used as anchoring points in Monte Carlo simulations of the likelihood of DNA loop formation and in the optimization of energetically preferred spatial pathways. As noted above, in the absence of knowledge of the directions in which the operators attach to the arms of the repressor, each operator is placed in two orientations on the protein-binding headpieces and loops of four different types are generated for a given repressor-operator model.
The ease of DNA loop formation is estimated from the fraction of simulated configurations of a linear molecule with terminal base pairs positioned so as to overlap the base pairs at the ends of the repressor-operator assembly. The formation of a successfully closed loop is detected by adding a phantom base pair to the 3'-end of the DNA and checking its coincidence with the first base pair of the loop [bib_ref] Interplay of protein and DNA structure revealed in simulations of the lac..., Czapla [/bib_ref] [bib_ref] Sequence-dependent effects in the cyclization of short DNA, Czapla [/bib_ref]. The joining step incorporates the rigid-body parameters, found in the complex of repressor with operator DNA, that relate the coordinate frames on the anchored base pairs. The latter parameters depend upon the precise structure of the full repressor-operator assembly and thus vary as the protein opens or the DNA operators fluctuate about their rest states. The probability of DNA looping is reported in terms of the Jacobson-Stockmayer J factor, the well-known ratio of the equilibrium constant for polymer ring closure compared to the bimolecular association of a linear molecule of the same length and composition [bib_ref] Intramolecular reaction in polycondensations. I. The theory of linear systems, Jacobson [/bib_ref]. The greater the value of J is, the lower the free energy of DNA is and the greater the likelihood of looping is.
Representative configurations of DNA chains are obtained, as described previously [bib_ref] Sequence-dependent effects in the cyclization of short DNA, Czapla [/bib_ref] , by direct Monte Carlo enumeration using a standard Gaussian random-number generatorand a modification of the Alexandrowicz half-chain pairwise-combination technique [bib_ref] Monte Carlo of chains with excluded volume: A way to evade sample..., Alexandrowicz [/bib_ref]. The random placement of HU on DNA includes corrections for the potential overlap of bound proteins and the generation of partial binding sites on simulated half chains as detailed in full elsewhere [bib_ref] Effects of the nucleoid protein HU on the structure, flexibility, and ring-closure..., Czapla [/bib_ref]. The configurations of DNA chains capable of looping between the headpieces of the repressor are identified from the spatial disposition of terminal base pairs. The six step parameters that relate the first and the last (phantom) base pairs should be null in a perfectly closed loop, and the end-to-end vector r, the global bend angle Γ, and the net twist angle ω should also be zero. Because the chances are very low that the added base pair will superimpose perfectly on the first base pair in any simulated structure, the end conditions are relaxed and only configurations that fall within the following bounds are classified as looped: The effects of specific changes in the protein-DNA model on relative looping propensities are also estimated from the relative statistical weights of the minimum-energy structure that meets the desired conditions compared to the reference state, i.e., e −∆E , where ∆E = E i − E 0 is the difference in energy between the modified state and the ideal isotropic DNA loop anchored to the V-shaped repressor.
## Loop optimization
Minimum-energy looped structures are obtained with a new procedure that makes it possible to optimize the potential energy of elastic deformation of a collection of base pairs, in which the positions and orientations of the first and last pairs are held fixed. Rather than describe the chain in terms of the six standard rigid-body parameters, we introduce a new set of variables that keep track of the vectorial displacements of successive base pairs in a global reference frame. The choice of variables makes it possible to take the imposed spatial constraints into direct account and to use unconstrained numerical optimization methods. In other words, a constrained optimization problem is transformed into an unconstrained one, and the numerical implementation is simpler and more robust. By contrast, the standard rigid-body parameters associated with a DNA base-pair step are expressed in a dimeric frame midway between the reference frames of adjacent base pairs so that the magnitudes of the local translational components are independent of chain direction [bib_ref] Definitions and nomenclature of nucleic acid structure parameters, Dickerson [/bib_ref] [bib_ref] The assessment of the geometry of dinucleotide steps in double-helical DNA: A..., El Hassan [/bib_ref].
We consider a collection of base pairs with imposed positions on the first and last residues. That is, we specify the three components of the end-to-end vector and the three rotational degrees of freedom that describe the relative displacement and orientation of coordinate frames on the first and last base pairs. We replace the set of base-pair step parameters with the equivalent new set of independent variables, termed the step degrees of freedom. The dependence of these degrees of freedom on the base-pair step parameters can be obtained analytically. In addition, the imposed end-to-end vector and end-to-end rotational constraints can be expressed such that the step degrees of freedom for the last base-pair step are functions of the degrees of freedom of all the other steps. In other words, we reduce the dimensionality of the problem by taking into account the boundary conditions (the dimension is reduced from 6(n − 1) to 6(n − 2), where n is the number of base pairs). The gradient of the potential energy of elastic deformation for the collection of base pairs and the Hessian matrix are also obtained analytically. The minimization procedure is implemented as a gradient-based optimization, e.g., conjugate gradient or BFGS (Broyden-Fletcher-Goldfarb-Shanno) optimization, in the reduced space of the step degrees of freedom. Full details are presented elsewhere.
## Loop characterization and manipulation
We describe the local structures of the DNA loops in terms of the bending and twisting of the constituent base pairs. The degree of bending is taken as the angle γ between the normals of successive base pairs, and the twist angle θ is based on a discrete ribbon constructed from the origins and reference frames of successive base pairs [bib_ref] Characterization of the geometry and topology of DNA pictured as a discrete..., Clauvelin [/bib_ref]. In contrast to the twist angle included in the six rigid-body parameters describing the spatial arrangements of successive base pairs [bib_ref] Definitions and nomenclature of nucleic acid structure parameters, Dickerson [/bib_ref] , the twist reported here, the so-called twist of supercoiling, can be combined with the writhing number of a closed structure to obtain the correct linking number [bib_ref] Characterization of the geometry and topology of DNA pictured as a discrete..., Clauvelin [/bib_ref] [bib_ref] Two perspectives on the twist of DNA, Britton [/bib_ref]. The reported loop lengths N correspond to the number of base-pair steps between the centers of bound DNA operators, i.e., seven of the 14 steps attached to each arm of the modeled repressor-operator assembly plus the DNA steps subjected to configurational changes (78 such steps for the 92-bp lac loop). We also take advantage of a feature in the 3DNA software that allows a user to look at multiple structures from a common perspective [bib_ref] Analyzing and building nucleic acid structures with 3DNA, Colasanti [/bib_ref].
# Conclusions
How the naturally stiff DNA double helix forms the short loops implicated in the regulation of bacterial genes has long been a mystery. New computational approaches that take direct account of the known structures and fluctuations of protein and DNA are making it possible to envision ways in which the loops might assemble and function. For example, trace amounts of the nonspecific architectural protein HU enhance the computed likelihood of loop formation several orders of magnitude beyond that predicted for protein-free DNA, to levels comparable in value to those deduced from the expression of lac genes in Escherichia coli. The computed ease of looping DNA between the headpieces of the Lac repressor protein also increases, albeit to a lesser extent, if the tetrameric protein opens from the V-shaped arrangement found in the crystalline state, if parts of the DNA loop soften or unwind, or if the DNA operator fluctuates among the ensemble of structures determined from NMR measurements.
The nearly one-to-one relationship between the relative statistical weights of 92-bp Lac repressor-mediated DNA loops identified through Monte Carlo sampling and energy minimization now makes it possible to address a broader range of systems and problems. We can thus perform rapid optimizations to probe rare looping events or to examine specific contributions of protein and DNA in certain kinds of loops. Here, we showed how a particular type of repressor opening enhances the likelihood of forming 92-bp Lac repressor-mediated loops and how simple modifications of local chain deformability and intrinsic structure change the structures and relative energies of looped DNA. We are in the process of examining how other types of repressor motions contribute to loop formation and how more realistic treatments of DNA fine structure and local deformability, such as anistropic bending or intrinsic curvature, influence the predicted results. We can also take explicit account of the supercoiling of DNA essential for Lac repressor-mediated DNA looping and can introduce loops within larger molecular contexts, such as topologically isolated domains within a long plasmid.
Earlier treatments of the in vivo looping properties of DNA [bib_ref] Analysis of in vivo LacR-mediated gene repression based on the mechanics of..., Zhang [/bib_ref] [bib_ref] Multilevel deconstruction of the in vivo behavior of looped DNA-protein complexes, Saiz [/bib_ref] were phenomenological in the sense that the effects of the molecular system on DNA chain closure were subsumed in the parameters of simple models fitted to the data. For example, DNA treated purely as an ideal elastic rod appears to soften in the presence of HU as gauged by the force constants needed to account for the observed dependence of the J factor upon chain length. By contrast, the present work takes direct account of the structures and fluctuations of protein and DNA and makes it possible to decipher the precise contribution of each component to loop closure. The incorporation of HU allows DNA to adopt spatial pathways normally inaccessible to the stiff, naturally straight duplex, and the propensity of DNA to remain straight restricts the sites of HU-induced deformations. That is, even though HU binds in a sequence neutral fashion to unconstrained linear DNA, the architectural protein must accumulate at specific sites in order to bring the ends of the looped fragment in register with the repressor headpieces. Thus, the DNA stiffens and the HU binds specifically in the sense that the duplex is extended and the HU is localized at particular sites. The predicted rearrangements of DNA in the presence of HU, including the altered spatial disposition of regulatory elements and the potential effects of topologically localized proteins on gene expression, introduce a new perspective on protein-mediated DNA looping worthy of further investigation. In this regard, we are working on improvements to the potential functions that will allow us to include electrostatic interactions of DNA and protein and estimates of the cost of protein deformation in future studies.
[fig] Figure 2, Figure 2: Effects of DNA chain length on the orientation of DNA and the simulated uptake of HU on loops anchored to the Lac repressor. (a) Molecular images illustrating the predominant configurations of DNA and the positions of HU on 109 bp (left) and 115 bp (right) loops. Images rendered in PyMOL[22] with DNA backbones are shown as gold tubes, DNA bases as gold sticks, HU chains as blue ribbons, and repressor chains as pink and cyan ribbons. Views looking down the axis perpendicular to the long axes of the globular arms of the repressor; (b) Frequencies of occurrence σ loop of the four possible looped forms. Note the marked shift in loop orientation (A1/A2→P1/P2) and HU uptake (1→2) with the increase in operator spacing, the straight segments of DNA in the models, and the potential (highly curved) site for a third HU at the center of the P2 loop. Cont [/fig]
[fig] Figure 3: Effect of the Lac repressor on the minimum-energy configurations of 92-bp DNA loops. (a) Molecular images of repressor-mediated loops anchored in antiparallel and parallel orientations on the rigid V-shaped protein assembly. Note the more gradually curved pathways of the dominant types of bare DNA loops compared to the corresponding forms found in the presence of HU (Figure 2a); (b) Variation in the relative weights σ loop of the three looped forms with small changes Δα in the angle between the protein arms, compared with the relative populations of DNA loops found to be attached to the same Lac repressor model through Monte Carlo (MC) sampling[23] (vertical bar to the right of the data obtained by energy optimization). [/fig]
[fig] Figure 4: Changes in DNA looping associated with large-scale opening of the Lac repressor. (a) Molecular images of 92-bp loops anchored in the dominant antiparallel and parallel orientations with the change in the opening angle Δα set to values that minimize the DNA elastic energy. Note the rearrangement of DNA on opened versus closed repressors (Figure 3a); (b) Distribution of the opening angle w (Δα) of Lac repressor proteins anchoring DNA loops subjected to energy minimization and collected in Monte Carlo (MC) sampling. The differences in repressor opening found with the two approaches reflect the different treatments of protein, i.e., precisely imposed routes of conformational change versus spatial forms indirectly captured in the course of DNA loop closure. [/fig]
[fig] Figure 5: Changes in 92-bp DNA looping brought about by localized changes in the DNA model. (a) Superimposed images of the optimized DNA pathways found when the sites of pyrimidine-purine steps along the lac operon are softened (blue) and when the helix is overtwisted (green) compared to ideal DNA (gold). Terminal base pairs are constrained to the spatial positions imposed by the V-shaped Lac repressor assembly. The antiparallel loops are rotated ~45° and the parallel loop ~90° about the vertical axis relative to the images inFigure 3a;(b) Changes in the bending Δγ and twisting Δθ between successive base pairs along the depicted loops, with the data color-coded to match the molecular images. The yellow lines denote the locations of the pyrimidine-purine steps, where the softened DNA shows bursts of bending and untwisting. Note the relative insensitivity of the parallel loop and the responses in the antiparallel loops to imposed overtwisting. [/fig]
[fig] Figure 6: Relative ease of forming 92-bp DNA loops anchored by various Lac repressor-bound O 3 ···O 1 and O sym ···O sym operator models compared to loops terminated by identical rigid, O sym operators. (a) Cartoon images illustrating the variability in the three operator structures. All views looking down the long axis of the kinked CG·CG step (blue) at the center of the DNA fragments; (b) Distribution of the looping propensities, expressed in terms of the logarithmic values of the J factors, for the two classes of flexible anchoring conditions and for the rigid operator pair ( ··· ). Note the enhancement in looping with the natural operators. [/fig]
[table] Table 1: Elastic energy contributions and populations of 92-bp Lac repressor-mediated DNA loops. [/table]
|
Neural attentional-filter mechanisms of listening success in middle-aged and older individuals
Successful listening crucially depends on intact attentional filters that separate relevant from irrelevant information. Research into their neurobiological implementation has focused on two potential auditory filter strategies: the lateralization of alpha power and selective neural speech tracking. However, the functional interplay of the two neural filter strategies and their potency to index listening success in an ageing population remains unclear. Using electroencephalography and a dual-talker task in a representative sample of listeners (N = 155; age=39-80 years), we here demonstrate an often-missed link from single-trial behavioural outcomes back to trial-by-trial changes in neural attentional filtering. First, we observe preserved attentional-cue-driven modulation of both neural filters across chronological age and hearing levels. Second, neural filter states vary independently of one another, demonstrating complementary neurobiological solutions of spatial selective attention. Stronger neural speech tracking but not alpha lateralization boosts trial-to-trial behavioural performance. Our results highlight the translational potential of neural speech tracking as an individualized neural marker of adaptive listening behaviour.
R eal-life listening is characterized by the concurrence of sound sources that compete for our attention [bib_ref] Some experiments on the recognition of speech, with one and with two..., Cherry [/bib_ref]. Successful speech comprehension thus relies on the differentiation of relevant and irrelevant inputs. Here, the concept of neural attentional 'filters' serves as an important and pervasive algorithmic metaphor of how auditory attention is implemented at the neural level [bib_ref] Attention metaphors: how metaphors guide the cognitive psychology of attention, Fernandez-Duque [/bib_ref]. Neural attentional filters can be instantiated by different mechanistic principles and recent studies have predominantly focused on two potential but nonexclusive neural filter strategies originating from distinct research traditions:
From the visual domain stems an influential line of research that supports the role of alpha-band (~8-12 Hz) oscillatory activity in the implementation of controlled, top-down suppression of behaviourally irrelevant information [bib_ref] Shaping functional architecture by oscillatory alpha activity: gating by inhibition, Jensen [/bib_ref] [bib_ref] The role of alpha-band brain oscillations as a sensory suppression mechanism during..., Foxe [/bib_ref] [bib_ref] Alpha oscillations correlate with the successful inhibition of unattended stimuli, Händel [/bib_ref] [bib_ref] Mechanisms of selective inhibition in visual spatial attention are indexed by α-band..., Rihs [/bib_ref]. Importantly, across modalities, it was shown that spatial-attention tasks are neurally supported by hemispheric lateralization of alpha power over occipital, parietal but also the respective sensory cortices [bib_ref] Attentional gain control of ongoing cortical speech representations in a 'cocktail party, Kerlin [/bib_ref] [bib_ref] Dynamic oscillatory processes governing cued orienting and allocation of auditory attention, Ahveninen [/bib_ref] [bib_ref] Lateralized auditory cortical alpha band activity and interregional connectivity pattern reflect anticipation..., Müller [/bib_ref] [bib_ref] Spatiotemporal dynamics of auditory attention synchronize with speech, Wöstmann [/bib_ref] [bib_ref] Opposite effects of lateralised transcranial alpha versus gamma stimulation on auditory spatial..., Wöstmann [/bib_ref] [bib_ref] Probing the limits of alpha power lateralisation as a neural marker of..., Tune [/bib_ref] [bib_ref] Top-down controlled alpha band activity in somatosensory areas determines behavioral performance in..., Haegens [/bib_ref] [bib_ref] Attentional selection of location and modality in vision and touch modulates low-frequency..., Bauer [/bib_ref] [bib_ref] Anticipatory biasing of visuospatial attention indexed by retinotopically specific alpha-band electroencephalography increases..., Worden [/bib_ref] [bib_ref] Increases in alpha oscillatory power reflect an active retinotopic mechanism for distracter..., Kelly [/bib_ref]. This suggests that asymmetric alpha modulation could act as a filter mechanism by modulating sensory gain already in the early processing stages.
In addition, a prominent line of research focuses on the role of low-frequency [bib_ref] Attention metaphors: how metaphors guide the cognitive psychology of attention, Fernandez-Duque [/bib_ref] [bib_ref] Shaping functional architecture by oscillatory alpha activity: gating by inhibition, Jensen [/bib_ref] [bib_ref] The role of alpha-band brain oscillations as a sensory suppression mechanism during..., Foxe [/bib_ref] [bib_ref] Alpha oscillations correlate with the successful inhibition of unattended stimuli, Händel [/bib_ref] [bib_ref] Mechanisms of selective inhibition in visual spatial attention are indexed by α-band..., Rihs [/bib_ref] neural activity in auditory and, broadly speaking, perisylvian cortex in the selective representation of speech input ('neural speech tracking'). Slow cortical dynamics temporally align with (or 'track') auditory input signals to prioritize the neural representation of behaviourally relevant sensory information [bib_ref] Low-frequency neuronal oscillations as instruments of sensory selection, Schroeder [/bib_ref] [bib_ref] Dynamics of active sensing and perceptual selection, Schroeder [/bib_ref] [bib_ref] Frequency modulation entrains slow neural oscillations and optimizes human listening behavior, Henry [/bib_ref] [bib_ref] Neural entrainment and attentional selection in the listening brain, Obleser [/bib_ref] (see also refs. [bib_ref] Rapid transformation from auditory to linguistic representations of continuous speech, Brodbeck [/bib_ref] [bib_ref] Electrophysiological correlates of semantic dissimilarity reflect the comprehension of natural, narrative speech, Broderick [/bib_ref] for the neural tracking of contextual semantic information). In human speech comprehension, a key finding is the preferential neural tracking of attended compared to ignored speech in superior temporal brain areas close to the auditory cortex [bib_ref] Mechanisms underlying selective neuronal tracking of attended speech at a '"cocktail party, Zion Golumbic [/bib_ref] [bib_ref] Emergence of neural encoding of auditory objects while listening to competing speakers, Ding [/bib_ref] [bib_ref] Selective cortical representation of attended speaker in multi-talker speech perception, Mesgarani [/bib_ref] [bib_ref] Attentional selection in a cocktail party environment can be decoded from single-trial..., O'sullivan [/bib_ref] [bib_ref] Suppression of competing speech through entrainment of cortical oscillations, Horton [/bib_ref].
However, with few exceptions [bib_ref] Attentional gain control of ongoing cortical speech representations in a 'cocktail party, Kerlin [/bib_ref] , these two proposed neural auditory filter strategies have been studied independently of one another (but see refs. [bib_ref] Stimulus-driven brain rhythms within the alpha band: the attentional-modulation conundrum, Keitel [/bib_ref] [bib_ref] Spatial attentional selection modulates early visual stimulus processing independently of visual alpha..., Gundlach [/bib_ref] for recent results on visual attention). Also, they have often been studied using tasks that are difficult to relate to natural, conversation-related listening situations [bib_ref] Aging affects the balance of neural entrainment and top-down neural modulation in..., Henry [/bib_ref] [bib_ref] Global dynamics of selective attention and its lapses in primary auditory cortex, Lakatos [/bib_ref].
We thus lack understanding of whether or how modulations in lateralized alpha power and the neural tracking of attended versus ignored speech in the wider auditory cortex interact in the service of successful listening behaviour. Moreover, few studies using more real-life listening and speech-tracking measures were able to explicitly address the functional relevance of the discussed neural filter strategies, that is, their potency to explain behavioural listening success [bib_ref] Selective cortical representation of attended speaker in multi-talker speech perception, Mesgarani [/bib_ref] [bib_ref] Attentional selection in a cocktail party environment can be decoded from single-trial..., O'sullivan [/bib_ref].
As part of an ongoing large-scale project on the neural and cognitive mechanisms supporting adaptive listening behaviour in healthy ageing, this study aims at closing these gaps by leveraging the statistical power and representativeness of our large, agevarying participant sample. We use a dichotic listening paradigm to enable a synergistic look at concurrent single-trial changes in lateralized alpha power and neural speech tracking.
More specifically, our linguistic variant of a classic Posner paradigm 34 emulates a challenging dual-talker listening situation, in which speech comprehension is supported by two different listening cues [bib_ref] Modular reconfiguration of an auditory control brain network supports adaptive listening behavior, Alavash [/bib_ref]. Of particular interest for the present scientific endeavour is the spatial-attention cue that guides auditory attention in space. We additionally manipulated the semantic predictability of upcoming speech via a semantic category cue. While the effects of the semantic cue are of secondary importance for the present research questions, its manipulation still allows insights into whether semantic predictability modulates the engagement of neural attentional filter mechanisms, and how it affects listening success in a large cohort of middle-aged and older adults. Previous research has shown that the sensory analysis of speech and, to a lesser degree, the modulation of alpha power are influenced by the availability of higher-order linguistic information [bib_ref] Effect of informational content of noise on speech representation in the aging..., Presacco [/bib_ref] [bib_ref] Predictive top-down integration of prior knowledge during speech perception, Sohoglu [/bib_ref] [bib_ref] Phase-locked responses to speech in human auditory cortex are enhanced during comprehension, Peelle [/bib_ref] [bib_ref] Suppressed alpha oscillations predict intelligibility of speech and its acoustic details, Obleser [/bib_ref] [bib_ref] The human neural alpha response to speech is a proxy of attentional..., Wöstmann [/bib_ref] [bib_ref] Semantic context enhances the early auditory encoding of natural speech, Broderick [/bib_ref].
Varying from trial to trial, both cues were presented either in an informative or uninformative version. This manipulation allowed us to understand how concurrent changes in the neural dynamics of selective attention and the resulting listening behaviour are connected.
We focus on four main research questions (see . Note that in addressing these, we model additionally known influences on listening success: age, hearing loss, as well as hemispheric asymmetries in speech processing due to the well-known rightear advantage [bib_ref] Accuracy of recognition for speech presented to the right and left ears, Broadbent [/bib_ref] [bib_ref] Cerebral dominance and the perception of verbal stimuli, Kimura [/bib_ref].
First, informative listening cues should increase listening success: these cues allow the listener to deploy auditory selective attention (compared to divided attention), and to generate more specific (compared to only general) semantic predictions, respectively.
Second, we asked how the different cue-cue combinations would modulate the two key neurobiological measures of selective attention -alpha lateralization and neural speech tracking. We aimed to replicate previous findings of increased alpha lateralization and a preferential tracking of the target compared to the distractor speech signal under selective (compared to divided) spatial attention. At the same time, we capitalized on our age-varying sample to quantify the hitherto contested dependence of these neurobiological filters on participants' chronological age and hearing loss [bib_ref] Probing the limits of alpha power lateralisation as a neural marker of..., Tune [/bib_ref] [bib_ref] Hearing impairment is associated with enhanced neural tracking of the speech envelope, Decruy [/bib_ref] [bib_ref] Speech-in-noise representation in the aging midbrain and cortex: effects of hearing loss, Presacco [/bib_ref] [bib_ref] Neural tracking of attended versus ignored speech is differentially affected by hearing..., Petersen [/bib_ref].
Third, an important and often neglected research question pertains to a direct, trial-by-trial relationship of these two candidate neural measures: Do changes in alpha lateralization impact the degree to which attended and ignored speech signals are neurally tracked by low-frequency cortical responses?
Our final research question is arguably the most relevant one for all translational aspects of auditory attention; it has thus far only been answered indirectly when deeming these neurobiological filter mechanisms ʻattentionalʼ: to what extent do alpha lateralization and neural speech tracking allow us to explain the behavioural outcome-that is, listening success-at the individual level and in a single trial?
Here, we show how an attentional cue modulates neural speech tracking and alpha lateralization independently of age and hearing levels. We demonstrate the co-existence of largely independent neural filters that pose complementary neurobiological implementations of selective attention. Stronger neural speech tracking Schematic illustration of addressed research questions. The dichotic listening task manipulated the attentional focus and semantic predictability of upcoming input using two separate visual cues. We investigated whether informative cues would enhance behavioural performance (Q1). In line with (Q2), we also examined the degree to which a spatial (and semantic) cue modulated the two auditory neural measures of interest: neural speech tracking and lateralization of auditory alpha power. Finally, we assessed (Q3) the co-variation of neural measures, and (Q4) their potency in explaining behavioural performance. Furthermore, we investigated the impact of age, hearing loss, and probed ear on listening success and its underlying neural strategies. ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24771-9
but not alpha lateralization increases trial-to-trial listening performance. This emphasizes the potential of neural speech tracking as a diagnostic neural measure of an individual's listening success.
# Results
We recorded and source-localized electroencephalography (EEG) signals in an age-varying sample of healthy middle-aged and older adults (N = 155; age = 39-80 years, see who performed a challenging dichotic listening task. In this linguistic variant of a classic Posner paradigm [bib_ref] Modular reconfiguration of an auditory control brain network supports adaptive listening behavior, Alavash [/bib_ref] , participants listened to two concurrent five-word sentences spoken by the same female talker and were asked to identify the final word in one of the two sentences. Sentence pairs were temporally aligned to the onset of these task-relevant final words which led to slightly asynchronous sentence onsets.
Importantly, sentence presentation was preceded by two visual cues. First, a spatial-attention cue encouraged the use of either selective or divided attention by providing informative or uninformative instructions about the to-be-attended, and thus laterprobed, ear. The second cue indicated the semantic category that applied to both final target words. The provided category could represent a general or specific level, thus allowing for more or less precise prediction of the upcoming speech signal [fig_ref] Figure 2: Experimental design and behavioural benefit from informative cues [/fig_ref]. While this listening task does not tap into the most naturalistic forms of speech comprehension, it still approximates a dualtalker listening situation to probe the neural underpinnings of successful selective listening [bib_ref] Modular reconfiguration of an auditory control brain network supports adaptive listening behavior, Alavash [/bib_ref].
Using generalized linear mixed-effects models on single-trial data, we focus on two key neurobiological instantiations of auditory attention: the lateralization of 8-12 Hz alpha power, emerging from auditory as well as parietal cortex, and the differential neural tracking of attended versus ignored speech by slow (1-8 Hz) auditory cortical responses. We investigate how spatial cues, age and hearing status modulate behaviour and neural filters, whether neural filters operate independently, and to which extent they influence selective listening success.
Informative spatial cues improve listening success. For behavioural performance, we tested the impact of informative versus uninformative cues on listening success. Overall, participants achieved a mean accuracy of 87.8% ± SD 9.1% with a mean reaction time of 1742 ms ± SD 525 ms; as response speed: 0.62 s −1 ± SD 0.17 s -1 .
As expected, behaviour depended on the different combinations of listening cues [fig_ref] Figure 2: Experimental design and behavioural benefit from informative cues [/fig_ref]. Informative compared to uninformative spatial-attention cues yielded a strong behavioural benefit. In selective-attention trials, participants responded more accurately and faster (accuracy: generalized linear mixed-effects model (GLMM); odds ratio (OR) = 3.5, std. error (SE) = 0.12, P < 0.001; response speed: linear mixed-effects model (LMM); β = 0.57, SE = 0.04, P < 0.001; see . That is, when cued to one of the two sides, participants responded on average 261 ms faster and their probability of giving a correct answer increased by 6%.
Also, participants responded generally faster in trials in which they were given a specific, more informative semantic cue (LMM; β = 0.2, SE = 0.03, P < 0.001), most likely reflecting a semantic priming effect that led to faster word recognition. Contrary to our expectations, a more informative semantic cue did not lead to more accurate responses (GLMM; OR = 1.1, SE = 0.11, P = 0.69).
As in a previous fMRI implementation of this task [bib_ref] Modular reconfiguration of an auditory control brain network supports adaptive listening behavior, Alavash [/bib_ref] , we did not find evidence for any interactive effects of the two listening cues on either accuracy (GLMM; OR = 1.3, SE = 0.21, P = 0.36) or response speed (LMM; β = 0.09, SE = 0.06, P = 0.31). Moreover, the breakdown of error trials revealed a significantly higher proportion of spatial stream confusions (6% ± SD 8.3%) compared to random errors (3% ± SD 3.4%; paired t test on logittransformed proportions: t 155 = 6.53, P < 0.001; see Supplementary [fig_ref] Figure 2: Experimental design and behavioural benefit from informative cues [/fig_ref]. The increased rate of spatial stream confusions (i.e., responses in which the last word of the to-be-ignored sentence was chosen) attests to the distracting nature of dichotic sentence presentation and thus heightened task difficulty.
Spatial attention modulates both alpha lateralization and neural speech tracking in the auditory cortex. In line with our second research question, following source projection of EEG data, we probed whether the presence of an informative spatialattention cue would lead to reliable modulation of both alpha power and neural speech tracking within an a priori defined auditory region of interest (ROI; see [fig_ref] Figure 3: Informative spatial cue elicits increased alpha-power lateralization before and during speech presentation [/fig_ref] and Supplementary Methods for details).
For alpha power, we expected attention-induced lateralization due to a decrease in power contralateral and a concomitant increase in power ipsilateral to the focus of attention. For neural speech tracking, we expected stronger neural tracking of attended compared to ignored speech under selective attention but no such systematic difference in the neural tracking of probed and unprobed sentences in divided-attention trials. Accordingly, our analyses of alpha power and neural speech tracking focused on attentional modulation index measures that contrast the relative strength of neural responses to target versus distractor stimuli. In line with previous results, we expected alpha lateralization to be present throughout the auditory sentence presentation but to potentially increase around the task-relevant final word [bib_ref] Spatiotemporal dynamics of auditory attention synchronize with speech, Wöstmann [/bib_ref] [bib_ref] Probing the limits of alpha power lateralisation as a neural marker of..., Tune [/bib_ref] [bib_ref] Nonspatial features reduce the reliance on sustained spatial auditory attention, Bonacci [/bib_ref].
We compared alpha-power changes ipsi-and contralateral to the probed ear to derive a temporally resolved single-trial measure of alpha-power lateralization [alpha lateralization index (ALI) = (α-power ipsi − α-power contra )/(α-power ipsi + α-power contra )] [bib_ref] Top-down controlled alpha band activity in somatosensory areas determines behavioral performance in..., Haegens [/bib_ref].
As shown in [fig_ref] Figure 3: Informative spatial cue elicits increased alpha-power lateralization before and during speech presentation [/fig_ref] , an informative spatial cue-that is, the instruction to pay attention to a given side-elicited pronounced lateralization of 8-12 Hz alpha power within the auditory ROI. Lateralization of alpha power was evident following the spatial cue itself and during dichotic sentence presentation with its strongest peak around final word presentation.
As expected, the statistical analysis of alpha lateralization during sentence presentation (time window: 3.5-6.5 s; see control analysis section below for results on the final-word period) revealed a significant modulation by the attention that was additionally influenced by the probed ear (LMM; spatial cue x probed ear: β = 0.13, SE = 0.02, P < 0.001; [fig_ref] Figure 3: Informative spatial cue elicits increased alpha-power lateralization before and during speech presentation [/fig_ref]. Follow-up analysis showed a significant difference in alpha lateralization between selective-and divided-attention trials when the right ear but not when the left was probed (LMM, right-ear probed: β = 0.12, SE = 0.01, P < 0.001; LMM, left-ear probed: β = 0.016, SE = 0.013, P = 0.55; see . This pattern suggests that when given an uninformative spatial cue, participants presumably payed overall more attention to the left-ear stimulus leading to an increase in alpha lateralization for probed-left compared to probed-right trials.
Notably, we did not find any evidence for modulation by the semantic cue nor any joint influence of the spatial and semantic cue on alpha lateralization during sentence presentation (LMM; semantic cue main effect: β = −0.01, SE = 0.01, P = 0.53, spatial × semantic cue: β = −0.02, SE = 0.02, P = 0.53).
Not least, the extent of overall as well as attention-specific alpha lateralization was unaffected by participants' chronological age and hearing loss (P values > 0.27 for main effects of age, PTA, and their respective interactions with the spatial-attention cue; see also Supplementary for a corresponding analysis of alpha power during the interval of the final word).
In close correspondence to the alpha-power analysis, we investigated whether changes in attention or semantic predictability would modulate the neural tracking of attended versus ignored speech. We used linear backward ('decoding') models to reconstruct the onset envelopes of the to-be-attended and ignored sentences (for simplicity hereafter referred to as attended and ignored) from neural activity in the auditory ROI. Reconstruction models were trained on selective-attention trials only but then utilized to reconstruct attended (probed) and ignored (unprobed) envelopes for both attention conditions (see 'Methodsʼ, [fig_ref] Figure 4: Neural speech tracking of attended and ignored sentences [/fig_ref] and [fig_ref] Figure 4: Neural speech tracking of attended and ignored sentences [/fig_ref] for details).
In line with previous studies [bib_ref] Emergence of neural encoding of auditory objects while listening to competing speakers, Ding [/bib_ref] [bib_ref] Attentional selection in a cocktail party environment can be decoded from single-trial..., O'sullivan [/bib_ref] [bib_ref] Late cortical tracking of ignored speech facilitates neural selectivity in acoustically challenging..., Fiedler [/bib_ref] , the forward-transformed temporal response functions (TRFs) show increased encoding of attended compared to ignored speech in the time window covering the N1 TRF and P2 TRF component (see [fig_ref] Figure 4: Neural speech tracking of attended and ignored sentences [/fig_ref] , left panel). Here, however, this was observed particularly for right-ear inputs processed in the left auditory ROI.
Further attesting to the validity of our reconstruction models, reconstructed attended envelopes were overall more similar to the envelope of the to-be-attended sentence than to that of the to-be-ignored sentence, and vice versa for the reconstructed ignored envelopes (see [fig_ref] Figure 4: Neural speech tracking of attended and ignored sentences [/fig_ref] , right panel).
As shown in [fig_ref] Figure 4: Neural speech tracking of attended and ignored sentences [/fig_ref] , the differential neural tracking of attended and ignored envelopes (probed and unprobed envelopes under divided attention) was modulated by attention. Following an informative spatial cue, the neural tracking index becomes increasingly positive during the second half of sentence presentation with its highest peaks around final-word onset.
The statistical analysis of single-trial index values averaged for the time interval of final-word presentation confirmed this pattern: the difference in the neural tracking of the attended and ignored sentence was generally more pronounced under selective compared to divided attention (see control analysis section below for results on the entire sentence presentation). However, this effect was also modulated by differences in sentence onset: the difference in neural speech tracking between the two attention conditions was reduced when the attended/probed sentence started ahead of the distractor sentence. This effect was driven by an increase in differential neural speech tracking for divided attention in such trials: in absence of an informative spatial cue, participants' attention was captured by the sentence with the earlier onset. Consequently, we observed overall more positive index values when the earlier sentence was probed compared to when it was not probed (LMM, earlier onset x spatial cue: β = −0.05, SE = 0.02, P = 0.049, see [fig_ref] Figure 4: Neural speech tracking of attended and ignored sentences [/fig_ref] and Supplementary for full model details).
We also found a neural correlate of the known right-ear advantage for verbal materials, that is, an overall stronger tracking of left-ear inputs. This effect was independent of spatial-attention cueing (LMM; probed ear main effect: β = −0.03, SE = 0.01, P = 0.023; spatial cue × probed ear: β = 0.02, SE = 0.02, P = 0.54). As for alpha power, we did not observe any modulation of neural tracking by the semantic cue, nor any joint influence of the spatial and semantic cue (LMM; semantic cue main effect: β = 0.01, SE = 0.01, P = 0.53, interaction spatial x semantic cue: β = −0.02, SE = 0.02, P = 0.53).
Again, participants' age and hearing status did not prove significant predictors of neural speech tracking (P values > 0.54 for main effects of age, PTA, and their respective interactions with the spatial-attention cue, see also Supplementary for a corresponding analysis of neural tracking during the entire sentence presentation).
Trial-to-trial neural speech tracking is independent of synchronous alpha lateralization. Our third major goal was to investigate whether neural speech tracking and the modulation of alpha power reflects two dependent neural mechanisms of auditory attention at all. We asked whether neural speech tracking could be explained by auditory alpha lateralization either at the state level (i.e., within an individual from trial to trial) or at the trait level (i.e., between individual mean levels; see 'Statistical analysisʼ for details). If modulations of alpha power over auditory cortices indeed act as a neural filter mechanism to selectively gate processing during early stages of sensory analysis, then heightened levels of alpha lateralization should lead to a more differentiated neural tracking of the attended vs. ignored speech and thus a more positive neural tracking index (cf. [fig_ref] Figure 5: Relationship of alpha lateralization and neural speech tracking [/fig_ref].
However, in the analysis of the task-relevant final-word period, we did not find evidence for an effect of alpha lateralization on neural speech tracking at neither the state nor the trait level . This notable absence of an alpha lateralization-neural speech tracking relationship held irrespective of spatial-attention condition or probed ear (all P values > 0. [bib_ref] Modular reconfiguration of an auditory control brain network supports adaptive listening behavior, Alavash [/bib_ref].
To complement our fine-grained single-trial level investigation into the brain-brain relationship with a coarser, yet time-resolved analysis, we related the temporal dynamics of both neural measures in an exploratory between-subjects cross-correlation Positive values indicate relatively higher alpha power in the hemisphere ipsilateral to the attended/probed sentence compared to the contralateral hemisphere. The shaded grey area shows the time window of sentence presentation. Brain models show the auditory region of interest (red). b ALI during sentence presentation (3.5-6.5 s) shown separately per spatial-cue condition and probed ear (left plot) for N = 155 participants. Coloured dots show trial-averaged individual results, black dots and error bars indicate the grand-average and bootstrapped 95% confidence intervals. Box plots show median centre line, 25th to 75th percentile hinges; whiskers show minimum and maximum within 1.5 × interquartile range. For probed-right trials, there was a significant difference in ALI between selective-and dividedattention trials (right plot). Black dots represent individual mean ALI values with bootstrapped 95% CI error bars. Histogram shows the distribution of differences in ALI for informative vs. uninformative spatial-cue levels. β: slope parameter estimate from the corresponding general linear mixed-effects model; two-sided Wald test (FDR-corrected, ***P = 2.65 × 10 −17 ). Source data are provided as a Source Data file. NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24771-9 ARTICLE NATURE COMMUNICATIONS | (2021) 12:4533 | https://doi.org/10.1038/s41467-021-24771-9 | www.nature.com/naturecommunications analysis. As shown in [fig_ref] Figure 5: Relationship of alpha lateralization and neural speech tracking [/fig_ref] , under selective attention, neural speech tracking and alpha lateralization follow different temporal trajectories with neural speech tracking peaking earlier than alpha lateralization around final-word presentation. The average crosscorrelation of the two neural time courses during sentence presentation confirms a systematic temporal delay with fluctuations in neural speech tracking leading those in alpha power by about 520 ms (see [fig_ref] Figure 5: Relationship of alpha lateralization and neural speech tracking [/fig_ref].
Neural speech tracking but not alpha lateralization explains listening behaviour. Having established the functional independence of alpha lateralization and neural speech tracking at the single-trial level, the final piece of our investigation was to probe their relative functional importance for behavioural outcomes.
Using the same (generalized) linear mixed-effects models as in testing our first research question (Q1 in , we investigated whether changes in task performance could be explained by the independent (i.e., as main effects) or joint influence (i.e., as an interaction) of neural measures. Again, we modelled the influence of the two neural filter strategies on behaviour at the state and trait level [bib_ref] Fixed and random effects models: making an informed choice, Bell [/bib_ref].
For response accuracy, our most important indicator of listening success, we observed an effect of trial-by-trial variation in neural speech tracking both during the presentation of the final word and across the entire sentence: participants had a higher b Left: grand-average (N = 155 participants, 95% confidence interval (CI) error bands) forward-transformed temporal response functions (TRFs) for attended (green) and ignored (yellow) speech in the left and right auditory ROI. Right: single-subject (N = 155 participants; 95% CI error bars) mean Pearson correlation of reconstructed and presented envelopes shown separately for attended (top, green) and ignored speech (bottom, yellow). c Top: grand-average (N = 155 participants) peri-stimulus time course of neural tracking index shown separately for selective (solid dark green curve) and divided attention (dashed lightgreen curve) ±1 SEM error band. Histograms show sentence and final-word onsets. The shaded area indicates the final-word presentation interval used for statistical analysis. Bottom: Single-subject (N = 155 participants) mean attended and ignored neural speech tracking during final-word presentation for selective and divided attention, respectively. d Left: neural tracking index shows per spatial-attention condition and for trials in which cued/probed sentences started ahead of ('probed first') or after the distractor ('probed second'). Coloured dots represent the single-subject average (N = 155 participants), black dots and error bars indicate grand-average and bootstrapped 95% CI. Box plots show median centre line, 25th to 75th percentile hinges, whiskers indicate minimum and maximum within 1.5 × interquartile range. Right: significant difference in neural tracking between selective-and divided-attention trials in probed second trials (top plot), and stronger neural tracking in probed-left trials. Black dots represent the individual mean neural tracking index with bootstrapped 95% CI error bars for N = 155 participants. Histogram shows the distribution of differences in neural tracking in informative vs. uninformative spatial-cue trials, and probed-left vs. probed-right trials, respectively. β: slope parameter estimate from the corresponding general linear mixed-effects model; ***P = 1.22 × 10 −9 , *P = 0.0233 (two-sided Wald test, FDR-corrected). Source data are provided as a Source Data file. chance of responding correctly in trials in which they neurally tracked the cued/probed sentence more strongly than the distractor sentence (see [fig_ref] Figure 6: Neural speech tracking predicts listening behaviour [/fig_ref].
The data held no evidence for any direct effects of trial-to-trial or participant-to-participant variation in alpha lateralization during a sentence or final-word presentation on accuracy (all P values > 0.18; see . We also did not find evidence for any joint effects of alpha power and neural speech tracking extracted from either of the two time windows (all P values > 0.33). Importantly, the absence of an effect did not hinge on differences in neural measures across spatial-cue, or probed-ear levels as relevant interactions of neural measures with these predictors were included in the model (all P values > 0.55).
The observed effects of neural filters on response speed depended on the analysed time window: while participants with relatively higher average levels of neural speech tracking during sentence presentation responded overall faster (LMM, neural tracking (between-subject effect): β = 0.08, SE = 0.03, P = 0.01; see [fig_ref] Figure 6: Neural speech tracking predicts listening behaviour [/fig_ref] , right panel and , we found a combined effect of neural dynamics during final-word presentation. Under selective but not divided attention, response speed depended on a combination of trial-to-trial variation in both alpha lateralization and neural speech tracking (LMM; spatial cue × ALI (within-subject effect) × neural tracking index (withinsubject effect): β = 0.08, SE = 0.03, P = 0.01; see . In short, responses were fastest in trials where relatively elevated levels in either neural speech tracking or alpha lateralization were paired with relatively reduced levels in the respective other neural measures thus highlighting the influence of two independent complementary filter solutions (see also [fig_ref] Figure 5: Relationship of alpha lateralization and neural speech tracking [/fig_ref].
In line with the literature on listening behaviour in ageing adults [bib_ref] Human aging compromises attentional control of auditory perception, Passow [/bib_ref] [bib_ref] A dynamic auditory-cognitive system supports speech-in-noise perception in older adults, Anderson [/bib_ref] , the behavioural outcome was further reliably predicted by age, hearing loss, and probed ear. We observed that participants' performance varied in line with the well-attested right-ear advantage (REA, also referred to as left-ear disadvantage) in the processing of linguistic materials [bib_ref] Cerebral dominance and the perception of verbal stimuli, Kimura [/bib_ref]. More specifically, participants responded both faster and more accurately when they were probed on the last word presented to the right compared to the left ear (response speed: LMM; β = 0.08, SE = Control analyses. We ran additional control analyses to validate our main set of results. First, we asked whether the observed independence of alpha lateralization and neural speech tracking hinged on the precise time window and cortical site from which neural measures were extracted. One set of control models thus included alpha lateralization during spatial cue rather than sentence presentation as a predictor of neural speech tracking during sentence and final-word presentation. Next, we related the two neural filters during the entire sentence presentation rather than only during the final word.
Second, we tested the hypothesis that neural speech tracking might be driven not primarily by alpha-power modulations emerging in auditory cortices, but rather by those generated in domain-general attention networks in the parietal cortex [bib_ref] Oscillatory alpha-band mechanisms and the deployment of spatial attention to anticipated auditory..., Banerjee [/bib_ref]. We, therefore, ran control models including alpha lateralization within the inferior parietal lobules. However, none of these additional analyses found evidence for an effect of alpha lateralization on neural speech tracking (see and .
Third, we asked whether our neural speech tracking results were impacted by the range of time lags used for reconstruction, or by the specific decoder model underlying the neural tracking index. Reconstructing envelopes using a shorter time window (50-250 ms) did not significantly change the resulting neural tracking index values (LMM, β = 0.002, SE = 0.007, P = 0.84; see also . In a separate analysis, we calculated the neural tracking index using only the attended decoder model and probed its influence on behaviour. The results are overall in line with our main conclusions and particularly underscore the impact of neural speech tracking on response accuracy (see .
Finally, we tested whether changes in age or hearing loss would modulate the relationship of neural tracking and alpha lateralization with listening behaviour. However, the inclusion of the respective interaction terms did not further improve the statistical models of accuracy and response speed (all P values > 0.44).
# Discussion
We have utilized the power of a representative sample of middle-aged and older listeners to explicitly address the question of how two eminent neurobiological implementations of attentional filtering, typically studied in isolation, relate to one another, and how they jointly shape listening success. In addition, we leveraged our agevarying sample to ask how chronological age and hearing loss affect the fidelity of neural filter strategies and their influence on behaviour.
In our dichotic listening task, we source-localized the electroencephalogram, and primarily focused on systematic spatial-cuedriven changes within the auditory cortex in alpha lateralization and the neural tracking of attended versus ignored speech. These results provide large-sample support for their suggested roles as neural instantiations of selective attention.
First, an informative spatial-attention cue not only boosted both neural measures but also consistently boosted behavioural performance. Listening behaviour was additionally influenced by both trial-to-trial and individual-to-individual variation in neural speech tracking, with relatively stronger tracking of the target sentence leading to better performance. An informative semantic cue led to faster responses but did not affect the two neural measures, thus most likely reflecting a priming effect speeding up the analysis of response alternatives rather than the differential processing of the sentences themselves.
Second, when related at the single-trial, single-subject level, the two neural attentional filter mechanisms were found to operate statistically independent of each other. This underlines their functional segregation and speaking to two distinct neurobiological implementations. Yet, when related in a coarser, betweensubjects analysis across time, peaks in selective neural speech tracking systematically preceded those in alpha lateralization.
Importantly, while chronological age and hearing loss reliably decreased behavioural performance they did not systematically participants. Left panel shows the predicted group-level fixed effect (green line ± 95% CI) of trial-to-trial variation in neural tracking on accuracy. Grey thin lines indicate estimated subject-specific slopes. Right panel shows the predicted group-level fixed effect of neural tracking at the between-subject level on response speed (green line ± 95% CI). Grey dots indicate subjectlevel model predictions. OR: odds ratio, β: slope parameter estimate from the corresponding general linear mixed-effects model, two-sided Wald test (FDR-corrected). b Summary of results. Black arrows highlight statistically significant effects from (generalized) single-trial linear mixed-effects modelling. Grey arrow shows the effect of additionally modelled influences. Notably, changes in age and hearing loss did not modulate the fidelity of the two key neural measures. Source data are provided as a Source Data file. affect the fidelity of neural filter strategies nor their influence on behaviour.
Neural speech tracking but not alpha lateralization predicts listening success. This study explicitly addressed the often overlooked question of how neural filter states (i.e., fluctuations from trial-to-trial) impact behaviour, here single-trial listening success [bib_ref] Local cortical desynchronization and pupil-linked arousal differentially shape brain states for optimal..., Waschke [/bib_ref] [bib_ref] Beyond establishing involvement: quantifying the contribution of anticipatory α-and β-band suppression to..., Van Ede [/bib_ref] [bib_ref] Cortical entrainment to continuous speech: functional roles and interpretations, Ding [/bib_ref]. Using a sophisticated linear-model approach that probed the impact of both state-and trait-level modulation of neural filters on behaviour, we only found evidence for a direct influence of neural speech tracking but not alpha lateralization on behavioural performance even though all three measures were robustly modulated by the presence of a spatial cue (see [fig_ref] Figure 6: Neural speech tracking predicts listening behaviour [/fig_ref]. What could be the reason for this differential impact of neural measures on behaviour?
To date, the behavioural relevance of selective neural speech tracking is still poorly supported given the emphasis on more naturalistic, yet more complex language stimuli [bib_ref] How to analyse electrophysiological responses to naturalistic language with time-resolved multiple regression, Sassenhagen [/bib_ref]. While these stimuli provide a window onto the most natural forms of speech comprehension, they are not easily paired with fine-grained measures of listening behaviour. This makes it particularly challenging to establish a direct link between differential neural speech tracking and listening success [bib_ref] Rapid transformation from auditory to linguistic representations of continuous speech, Brodbeck [/bib_ref] [bib_ref] Mechanisms underlying selective neuronal tracking of attended speech at a '"cocktail party, Zion Golumbic [/bib_ref] [bib_ref] Emergence of neural encoding of auditory objects while listening to competing speakers, Ding [/bib_ref] [bib_ref] Late cortical tracking of ignored speech facilitates neural selectivity in acoustically challenging..., Fiedler [/bib_ref] [bib_ref] Neural coding of continuous speech in auditory cortex during monaural and dichotic..., Ding [/bib_ref]. Nevertheless, there is preliminary evidence linking stronger neural tracking to improved comprehension when it is tested at a comparably high level 28 (i.e., content questions on longer speech segments). Our current results thus provide important additional fine-grained and temporally resolved support to the functional relevance of selective neural speech tracking for moment-to-moment listening behaviour [bib_ref] Single-trial analyses: why bother?, Pernet [/bib_ref] [bib_ref] Neural entrainment to speech modulates speech intelligibility, Riecke [/bib_ref] [bib_ref] Phase entrainment of brain oscillations causally modulates neural responses to intelligible speech, Zoefel [/bib_ref].
Despite a vast number of studies investigating the role of (lateralized) alpha oscillations in attentional tasks, the circumstances under which their top-down modulation may affect the behavioural outcome are still insufficiently understood [bib_ref] Spatial attentional selection modulates early visual stimulus processing independently of visual alpha..., Gundlach [/bib_ref]. Rather, the presence of a stable brain-behaviour relationship hinges on several factors.
First, the link of neural filter state to behaviour seems to be impacted by age: most evidence linking increased alpha lateralization to better task performance in spatial-attention tasks stems from smaller samples of young adults [bib_ref] Spatiotemporal dynamics of auditory attention synchronize with speech, Wöstmann [/bib_ref] [bib_ref] Top-down controlled alpha band activity in somatosensory areas determines behavioral performance in..., Haegens [/bib_ref] [bib_ref] α-band electroencephalographic activity over occipital cortex indexes visuospatial attention bias and predicts..., Thut [/bib_ref] [bib_ref] The neural markers of an imminent failure of response inhibition, Bengson [/bib_ref]. By contrast, the presence of such a systematic relationship in middle-aged and older adults is obscured by considerable variability in age-related changes at the neural (and to some extent also behavioural) level 14,66-69 (see discussion below).
Second, previous findings differ along (at least) two dimensions: (i) whether the functional role of alpha lateralization is studied during attention cueing, stimulus anticipation, or stimulus presentation 9,66,70 , and (ii) whether the behaviour is related to the overall strength of alpha lateralization or its stimulus-driven rhythmic modulation [bib_ref] Spatiotemporal dynamics of auditory attention synchronize with speech, Wöstmann [/bib_ref] [bib_ref] Probing the limits of alpha power lateralisation as a neural marker of..., Tune [/bib_ref]. Depending on these factors, the observed brain-behaviour relations may relate to different top-down and bottom-up processes of selective auditory attention.
Third, as shown in a recent study by Wöstmann et al. [bib_ref] Alpha oscillations in the human brain implement distractor suppression independent of target..., Wöstmann [/bib_ref] , the neural processing of target and distractor is supported by two uncorrelated lateralized alpha responses emerging from different neural networks. Notably, their results provide initial evidence for the differential behavioural relevance of neural responses related to target selection and distractor suppression, respectively.
In summary, it is still a matter of debate by which mechanistic pathway, and at which processing stage the modulation of alpha power will impact behaviour. While it is (often implicitly) assumed that alpha oscillations impact behaviour via modulation of neural excitability and thus early sensory processing, there is little evidence that shows a direct influence of alpha oscillation on changes in neural excitability and on subsequent behaviour [bib_ref] Spatial attentional selection modulates early visual stimulus processing independently of visual alpha..., Gundlach [/bib_ref].
Lastly, the increase in alpha lateralization around final-word presentation could at least partially reflect post-perceptual processes associated with response selection rather than the perceptual analysis itself [bib_ref] Humans strategically shift decision bias by flexibly adjusting sensory evidence accumulation, Kloosterman [/bib_ref]. The observed combined influence of neural tracking and alpha lateralization on response speed but not accuracy would seem compatible with such an interpretation (but see also ref. [bib_ref] Degradation levels of continuous speech affect neural speech tracking and alpha power..., Hauswald [/bib_ref] for the combined influence of non-lateralized alpha power and neural speech tracking on intelligibility in a nonspatial listening task).
Taken together, our results underscore the impact of prioritized sensory encoding of relevant sounds via selective neural speech tracking on listening performance and highlight the difficulty in establishing a comparable link for a neural signature as multifaceted as alpha oscillations [bib_ref] The many characters of visual alpha oscillations, Clayton [/bib_ref] [bib_ref] Dynamic circuit motifs underlying rhythmic gain control, gating and integration, Womelsdorf [/bib_ref] [bib_ref] Brain networks and α-oscillations: structural and functional foundations of cognitive control, Sadaghiani [/bib_ref].
Are fluctuations in lateralized alpha power and neural speech tracking functionally connected? We investigated attentionrelated changes in two neural filter strategies that (i) involve neurophysiological signals operating at different frequency regimes, (ii) are assumed to support auditory attention by different neural mechanisms, and (iii) are typically studied in isolation [bib_ref] The role of alpha-band brain oscillations as a sensory suppression mechanism during..., Foxe [/bib_ref] [bib_ref] Neural entrainment and attentional selection in the listening brain, Obleser [/bib_ref]. Here, we found both neural filter strategies to be impacted by the same spatial-attention cue which afforded insights into their neurobiological dependence.
There is preliminary evidence, mostly from between-subjects analyses, suggesting that the two neural filter strategies may exhibit a systematic relationship [bib_ref] Attentional gain control of ongoing cortical speech representations in a 'cocktail party, Kerlin [/bib_ref] [bib_ref] Probing the limits of alpha power lateralisation as a neural marker of..., Tune [/bib_ref] [bib_ref] Aging affects the balance of neural entrainment and top-down neural modulation in..., Henry [/bib_ref] [bib_ref] Global dynamics of selective attention and its lapses in primary auditory cortex, Lakatos [/bib_ref] [bib_ref] Acoustic detail but not predictability of taskirrelevant speech disrupts working memory, Wöstmann [/bib_ref]. How the two neural filter strategies may be connected mechanistically is thus still an open question. We here asked whether concurrent changes in neural filter states would imply a neural hierarchy in which alphadriven controlled inhibition modulates the amplification of behaviourally relevant sensory information via selective neural speech tracking [bib_ref] Alpha and gamma oscillations characterize feedback and feedforward processing in monkey visual..., Van Kerkoerle [/bib_ref] [bib_ref] Layerspecific entrainment of gamma-band neural activity by the alpha rhythm in monkey..., Spaak [/bib_ref] [bib_ref] Largescale communication in the human brain is rhythmically modulated through alpha coherence, Chapeton [/bib_ref].
Our in-depth trial-by-trial analysis revealed independent modulation of alpha power and neural speech tracking. At the same time, in our exploratory between-subjects cross-correlation analysis we observed a systematic temporal delay with peaks in neural speech tracking leading those in alpha lateralization. While the direction and duration of this delay were closely in line with previous findings [bib_ref] Spatiotemporal dynamics of auditory attention synchronize with speech, Wöstmann [/bib_ref] [bib_ref] Probing the limits of alpha power lateralisation as a neural marker of..., Tune [/bib_ref] , at this coarser level of analysis, they speak against a hierarchy of neural processing in which lateralized alpha responses govern the differential neural tracking of attended versus ignored speech [bib_ref] Too little, too late, and in the wrong place: alpha band activity..., Antonov [/bib_ref].
Our single-trial results are well in line with recent reports of independent variation in alpha-band activity and steady-state (SSR) or frequency-following responses (FFR) in studies of visualspatial attention [bib_ref] Stimulus-driven brain rhythms within the alpha band: the attentional-modulation conundrum, Keitel [/bib_ref] [bib_ref] Spatial attentional selection modulates early visual stimulus processing independently of visual alpha..., Gundlach [/bib_ref] [bib_ref] Probing cortical excitability using rapid frequency tagging, Zhigalov [/bib_ref]. In addition, the inclusion of single-trial alpha lateralization as an additional training feature in a recent speech-tracking study failed to improve the decoding of attention [bib_ref] EEG decoding of the target speaker in a cocktail party scenario: considerations..., Teoh [/bib_ref]. The results from our most fine-grained single-trial level of analysis thus speak against a consistent, linear relationship of momentary neural filter states. Instead, we observed the co-existence of two complementary but seemingly independent neurobiological solutions to the implementation of auditory selective attention.
How can this finding be reconciled with findings from previous electrophysiological studies 9,32,33 pointing towards a functional trade-off between neurobiological attentional-filter mechanisms? And what could be an advantage to independent neural solutions for selective auditory attention?
Our between-subjects cross-correlation analysis appears to provide at least tentative support for a systematic relationship in which peaks in neural speech tracking precede those in alpha lateralization. A closer inspection of the group-level temporal modulation of neural measures throughout sentence presentation, however, reveals some important differences to previous results.
Whereas earlier studies reported an acyclic waxing and waning of neural entrainment and alpha power in response to rhythmic auditory stimulation [bib_ref] Spatiotemporal dynamics of auditory attention synchronize with speech, Wöstmann [/bib_ref] [bib_ref] Probing the limits of alpha power lateralisation as a neural marker of..., Tune [/bib_ref] [bib_ref] Global dynamics of selective attention and its lapses in primary auditory cortex, Lakatos [/bib_ref] , in this study, the two neural measures show different temporal dynamics: neural speech tracking gradually increases leading up to the final word, while alpha lateralization peaks at the sentence and final-word onset. The temporal dynamics of alpha lateralization, in particular, may point to the strategic intermittent engagement of spatial attention in line with task demands [bib_ref] Nonspatial features reduce the reliance on sustained spatial auditory attention, Bonacci [/bib_ref].
Do these differences in temporal dynamics of the two neural filters challenge the existence of a systematic single-trial brain-brain relationship? Yes, but they also point to a potential benefit of independent neural filter solutions. If the two neural measures of auditory attention were indeed functionally unconnected as suggested by the current results, they would allow for a wider range of neural filter state configurations to flexibly adapt to the current task demands and behavioural goals. The coexistence of two independent but complementary filter mechanisms operating either via the selective amplification of relevant or via the controlled inhibition of irrelevant sounds, enables different modes of auditory attention to serve a listener's goal in the face of complex real-life listening situations [bib_ref] Low-frequency neuronal oscillations as instruments of sensory selection, Schroeder [/bib_ref] [bib_ref] Dynamics of active sensing and perceptual selection, Schroeder [/bib_ref] [bib_ref] Temporal expectations and neural amplitude fluctuations in auditory cortex interactively influence perception, Herrmann [/bib_ref].
Do age and hearing loss affect neural filter strategies? The detrimental effects of increasing age and associated hearing loss on speech comprehension in noisy listening situations are well attestedand borne out by the current results. However, the extent to which the neural implementations of attentional filtering are affected by age and hearing loss, and in how far they may constitute neural markers of age-related speech comprehension problems, remains poorly understood [bib_ref] Human aging compromises attentional control of auditory perception, Passow [/bib_ref].
As in a previous study on a subset of the current sample, we found the fidelity of alpha lateralization unchanged with age [bib_ref] Probing the limits of alpha power lateralisation as a neural marker of..., Tune [/bib_ref].
Other studies on auditory attention, however, have observed diminished and less sustained alpha lateralization for older compared to younger adults that were to some extend predictive of behaviour [bib_ref] Diminished prestimulus alpha-lateralization suggests compromised self-initiated attentional control of auditory processing in..., Dahl [/bib_ref] [bib_ref] Older adults show impaired modulation of attentional alpha oscillations: evidence from dichotic..., Rogers [/bib_ref] [bib_ref] EEG correlates of spatial shifts of attention in a dynamic multi-talker speech..., Getzmann [/bib_ref].
Our observation of preserved neural speech tracking across age and hearing levels only partially agree with earlier findings. They are in line with previous reports of differential neural tracking of attended and ignored speech for hearing-impaired older adults that mirrored the attentional modulation observed for younger or older normal-hearing adults [bib_ref] Hearing impairment is associated with enhanced neural tracking of the speech envelope, Decruy [/bib_ref] [bib_ref] Speech-in-noise representation in the aging midbrain and cortex: effects of hearing loss, Presacco [/bib_ref] [bib_ref] Neural tracking of attended versus ignored speech is differentially affected by hearing..., Petersen [/bib_ref] [bib_ref] Effects of sensorineural hearing loss on cortical synchronization to competing speech during..., Fuglsang [/bib_ref]. As revealed by follow-up analysis (see , however, our data do not provide evidence for a differential impact of hearing loss on the neural tracking of attended or ignored speech as found in some of these studies. We also did not find evidence for overall increased levels of cortical neural tracking with age as observed in earlier studies [bib_ref] Evidence for enhanced neural tracking of the speech envelope underlying age-related speech-in-noise..., Decruy [/bib_ref] [bib_ref] Evidence of degraded representation of speech in noise, in the aging midbrain..., Presacco [/bib_ref].
The discrepancy in results may be explained by differences in (i) the studied populations (i.e., whether groups of younger and older participants were contrasted compared to the modelling of continuous changes in age and hearing loss), (ii) whether natural stories or short matrix sentence speech materials were used 91 , or (iii) by differences in task details. In sum, the results suggest that the commonly observed adverse effects of age and hearing loss on speech-in-noise processing are not readily paired with concomitant changes at the neural level.
In a representative, age-varying sample of listeners, we underscore the functional significance of lateralized alpha power and neural speech tracking to spatial attention. Our results point to the co-existence of two independent yet complementary neural filter mechanisms to be flexibly engaged depending on a listener's attentional goals. However, we see no direct, behaviourally relevant impact of alpha-power modulation on early sensory gain processes.
Only for neural speech tracking, we established a mechanistic link from trial-to-trial neural filtering during the concurrent sound input to the ensuing behavioural outcome. This link exists irrespective of age and hearing status, which points to the potency of neural speech tracking to serve as an individualized marker of comprehension problems in clinical settings and as a basis for translational neurotechnological advances.
This key advance notwithstanding, the notable absence of an association between alpha lateralization and listening behaviour also highlights the level of complexity associated with establishing statistically robust relationships of complex neural signatures and behaviour in the deployment of auditory attention. To understand how the brain enables successful selective listening it is necessary that studies go beyond the characterization of neurobiological filter mechanisms alone, and further jointly account for the variability in both neural states and behavioural outcomes [bib_ref] Behavior needs neural variability, Waschke [/bib_ref].
# Methods
Data collection. The analysed data are part of an ongoing large-scale study on the neural and cognitive mechanisms supporting adaptive listening behaviour in healthy middle-aged and older adults (ʻThe listening challenge: How ageing brains adapt (AUDADAPT)ʼ; https://cordis.europa.eu/project/rcn/197855_en.html). This project encompasses the collection of different demographic, behavioural, and neurophysiological measures across two time points. The analyses carried out on the data aim at relating adaptive listening behaviour to changes in different neural dynamics [bib_ref] Modular reconfiguration of an auditory control brain network supports adaptive listening behavior, Alavash [/bib_ref].
Participants and procedure. A total of N = 155 right-handed German native speakers (median age 61 years; age range 39-80 years; 62 males; see for age distribution) were included in the analysis. Handedness was assessed using a translated version of the Edinburgh Handedness Inventory 93 . All participants had normal or corrected-to-normal vision, did not report any neurological, psychiatric, or other disorders and were screened for mild cognitive impairment using the German version of the 6-Item Cognitive Impairment Test (6CIT 94 ).
During the EEG measurement, participants performed six blocks of a demanding dichotic listening task (see [fig_ref] Figure 2: Experimental design and behavioural benefit from informative cues [/fig_ref] and Supplementary Methods for details on sentence materials).
As part of our the overarching longitudinal study on adaptive listening behaviour in healthy ageing adults, prior to the EEG session, participants also underwent a session consisting of a general screening procedure, detailed audiometric measurements, and a battery of cognitive tests and personality profiling (see ref. [bib_ref] Probing the limits of alpha power lateralisation as a neural marker of..., Tune [/bib_ref] for details). Only participants with normal hearing or ageadequate mild-to-moderate hearing loss were included (see for individual audiograms). As part of this screening procedure, an additional 17 participants were excluded prior to EEG recording due to non-age-related hearing loss or medical history. Three participants dropped out of the study prior to EEG recording and an additional nine participants were excluded from analyses after EEG recording: three due to incidental findings after structural MR acquisition, and six due to technical problems during EEG recording or overall poor EEG data quality. Participants gave written informed consent and received financial compensation (8€ per hour). Procedures were approved by the ethics committee of the University of Lübeck and were in accordance with the Declaration of Helsinki.
Dichotic listening task. In a recently established 35 linguistic variant of a classic Posner paradigm 34 , participants listened to two competing, dichotically presented five-word sentences. They were probed on the sentence-final noun in one of the two sentences. All sentences followed the same sentence structure and had an average length of 2512 ms (range: 2183-2963 ms).
Sentences were spoken by the same female talker. Root mean (mean) square intensity (-26 dB Full Scale, FS) was equalized across all individual sentences and they were masked by continuous speech-shaped noise at a signal-to-noise ratio of 0 dB. Noise onset was presented with a 50 ms linear onset ramp and preceded sentence onset by 200 ms. Each sentence pair was temporally aligned by the onset of the two task-related sentence-final nouns. This, however, led to slight differences in the onset of the individual sentences. Crucially, the range and average sentence onset difference were similar for trials in which the probed (to-be-attended) sentence began earlier and those in which the unprobed (to-be-ignored) sentence began earlier (probed first: range: 0-580 ms, 162.1 ms ± 124.6; unprobed first: 0-560 ms, 180.6 ms ± 127.2). All participants listened to the same 240 sentence pairs but in subject-specific randomized order. In addition, across participants, we balanced the assignment of sentences to the right and left ear, respectively. Details on stimulus construction and recording can be found in the Supplementary Methods.
Critically, two visual cues preceded auditory presentation. First, a spatialattention cue either indicated the to-be-probed ear, thus invoking selective attention, or did not provide any information about the to-be-probed ear, thus invoking divided attention. Second, a semantic cue specified a general or a specific semantic category for the final word of both sentences, thus allowing to utilize a semantic prediction. Cue levels were fully crossed in a 2 × 2 design and the presentation of cue combinations varied on a trial-by-trial level [fig_ref] Figure 2: Experimental design and behavioural benefit from informative cues [/fig_ref]. The trial structure is exemplified in [fig_ref] Figure 2: Experimental design and behavioural benefit from informative cues [/fig_ref].
Each trial started with the presentation of a fixation cross in the middle of the screen (jittered duration: mean 1.5 s, range 0.5-3.5 s). Next, a blank screen was shown for 500 ms followed by the presentation of the spatial cue in the form of a circle segmented equally into two lateral halves. In selective-attention trials, one half was black, indicating the to-be-attended side, while the other half was white, indicating the to-be-ignored side. In divided-attention trials, both halves appeared in grey. After a blank screen of 500 ms duration, the semantic cue was presented in the form of a single word that specified the semantic category of both sentence-final words. The semantic category could either be given at a general (natural vs. manmade) or specific level (e.g. instruments, fruits, furniture) and thus provided different degrees of semantic predictability. Each cue was presented for 1000 ms.
After a 500 ms blank-screen period, the two sentences were presented dichotically along with a fixation cross displayed in the middle of the screen. Finally, after a jittered retention period, a visual response array appeared on the left or right side of the screen, presenting four-word choices. The location of the response array indicated which ear (left or right) was probed. Participants were instructed to select the final word presented on the to-be-attended side using the touch screen. Among the four alternatives were the two actually presented nouns as well as two distractor nouns from the same cued semantic category. Note that because the semantic cue applied to all four alternative verbs, it could not be used to post hoc infer the to-be-attended sentence-final word.
Stimulus presentation was controlled by PsychoPy Standalone v2.0 95 . The visual scene was displayed using a 24" touch screen (ViewSonic TD2420) positioned within an arm's length. Auditory stimulation was delivered using in-ear headphones (EARTONE 3 A) at a sampling rate of 44.1 kHz. Following instructions, participants performed a few practice trials to familiarize themselves with the listening task. To account for differences in hearing acuity within our group of participants, individual hearing thresholds for a 500-ms fragment of the dichotic stimuli were measured using the method of limits. All stimuli were presented 50 dB above the individual sensation level. During the experiment, each participant completed 60 trials per cue-cue condition, resulting in 240 trials in total. The cue conditions were equally distributed across six blocks of 40 trials each (~10 min) and were presented in random order. Participants took short breaks between blocks.
Behavioural data analysis. We evaluated participants' behavioural performance in the listening task with respect to accuracy and response speed. For the binary measure of accuracy, we excluded trials in which participants failed to answer within the given 4-s response window ('timeouts'). Spatial stream confusions, that is trials in which the sentence-final word of the to-be-ignored speech stream were selected, and random errors were jointly classified as incorrect answers. The analysis of response speed, defined as the inverse of reaction time, was based on correct trials only. Single-trial behavioural measures were subjected to (generalized) linear mixed-effects analysis and regularized regression (see 'Statistical analysis').
EEG data analysis. The preprocessed continuous EEG data (see Supplementary Methods for details on data collection and preprocessing) were high-pass-filtered at 0.3 Hz (finite impulse response (FIR) filter, zero-phase lag, order 5574, Hann window) and low-pass filtered at 180 Hz (FIR filter, zero-phase lag, order 100, Hamming window). The EEG was cut into epochs of -2 to 8 s relative to the onset of the spatial-attention cue to capture cue presentation as well as the entire auditory stimulation interval.
For the analysis of changes in alpha power, EEG data were downsampled to f s = 250 Hz. Spectro-temporal estimates of single-trial data were then obtained for a time window of −0.5 to 6.5 s (relative to the onset of the spatial-attention cue) at frequencies ranging from 8 to 12 Hz (Morlet's wavelets; the number of cycles = 6).
For the analysis of the neural encoding of speech by low-frequency activity, the continuous preprocessed EEG were downsampled to f s = 125 Hz and filtered between f c = 1 and 8 Hz (FIR filters, zero-phase lag, order: 8 f s /f c and 2 f s /f c , Hamming window). The EEG was cut to yield individual epochs covering the presentation of auditory stimuli, beginning at noise onset until the end of the auditory presentation.
Following EEG source and forward model construction (see Supplementary Methods for details), sensor-level single-trial data in each of our two analysis routines were projected to source space by matrix multiplication of the spatial filter weights. To increase the signal-to-noise ratio in source estimates and computationally facilitate source-level analyses, source-projected data were averaged across grid points per cortical area defined according to the HCP functional parcellation template [bib_ref] A multi-modal parcellation of human cerebral cortex, Glasser [/bib_ref] [bib_ref] Individual human brain areas can be identified from their characteristic spectral activation..., Keitel [/bib_ref]. This parcellation provides a symmetrical delineation of each hemisphere into 180 parcels for a total of 360 parcels. We constrained the analysis of neural measures to an a priori defined, source-localized auditory region of interest (ROI) as well as one control ROI in the inferior parietal lobule (see Supplementary Methods for details). The described analyses were carried out using the Fieldtrip toolbox (v. 2017-04-28) in Matlab 2016b, and the Human Connectome Project Workbench software (v1.5) as well as FreeSurfer (v.6.0).
Attentional modulation of alpha power. Absolute source power was calculated as the square amplitude of the spectro-temporal estimates. Since oscillatory power values typically follow a highly skewed, non-normal distribution, we applied a nonlinear transformation of the Box-Cox family (power trans = (power p − 1)/P with P = 0.5) to minimize skewness and to satisfy the assumption of normality for parametric statistical tests involving oscillatory power values [bib_ref] Single-trial log transformation is optimal in frequency analysis of resting EEG alpha, Smulders [/bib_ref]. To quantify attention-related changes in 8-12 Hz alpha power, per ROI, we calculated the single-trial, temporally resolved alpha lateralization index as follows [bib_ref] Spatiotemporal dynamics of auditory attention synchronize with speech, Wöstmann [/bib_ref] [bib_ref] Probing the limits of alpha power lateralisation as a neural marker of..., Tune [/bib_ref] [bib_ref] Top-down controlled alpha band activity in somatosensory areas determines behavioral performance in..., Haegens [/bib_ref].
[formula] ALI ¼ ðα-power ipsi À α-power contra Þ=ðα-power ipsi þ α-power contra Þð1Þ [/formula]
To account for overall hemispheric power differences that were independent of attention modulation, we first normalized single-trial power by calculating per parcel and frequency the whole-trial (-0.5-6.5 s) power averaged across all trials and subtracted it from single trials. We then used a robust variant of the index that applies the inverse logit transform [(1/(1 + exp(−x))] to both inputs to scale them into a common, positive-only [0;1]-bound space prior to index calculation. For visualization and statistical analysis of cue-driven neural modulation, we then averaged the ALI across all parcels within the auditory ROI and extracted single-trial mean values for the time window of sentence presentation (3.5-6.5 s), and treated them as the dependent measure in linear mixed-effects analysis (see 'Statistical analysis' below). They also served as continuous predictors in the statistical analysis of brain-behaviour and brain-brain relationships. We performed additional analyses that focused on the ALI in the auditory cortex during presentation of the sentence-final word and spatial-attention cue, respectively. Further control analyses included single-trial ALI during sentence and final-word presentation that were extracted from the inferior parietal ROI.
Estimation of envelope reconstruction models. To investigate how lowfrequency (i.e., <8 Hz) fluctuations in EEG activity related to the encoding of attended and ignored speech, we trained stimulus reconstruction models (also termed decoding or backward models) to predict the onset envelope (see Supplementary Methods for details) of the attended and ignored speech stream from EEG [bib_ref] The multivariate temporal response function (mTRF) toolbox: a MATLAB toolbox for relating..., Crosse [/bib_ref] [bib_ref] Neural responses to uninterrupted natural speech can be extracted with precise temporal..., Lalor [/bib_ref]. In this analysis framework, a linear reconstruction model g is assumed to represent the linear mapping from the recorded EEG signal, r(t,n), to the stimulus features, s(t):ŝ
[formula] t ð Þ ¼ ∑ n ∑ τ g τ; n ð Þrðt þ τ; nÞð2Þ [/formula]
whereŝðtÞ is the reconstructed onset envelope at time point t. We used all parcels within the bilateral auditory ROI and time lags τ in the range of -100 ms to 500 ms to compute envelope reconstruction models using ridge regression [bib_ref] Ridge regression: biased estimation for nonorthogonal problems, Hoerl [/bib_ref] :
[formula] g ¼ R T R þ λmI À Á À1 R T sð3Þ [/formula]
where R is a matrix containing the sample-wise time-lagged replication of the neural response matrix r, λ is the ridge parameter for regularization, I is the identity matrix, and m is a subject-specific scalar representing the mean of the trace of R T R 102,103 . The same grid of ridge parameters (λ = 10 −5 , 10 −4 , …10 10 ) was used across subjects, and m proved to be relatively stable across subjects (387.2 ± 0.18, mean ± SD). The optimal ridge value of λ = 1 was determined based on the average Pearson's correlation coefficient and mean squared error of the reconstructed and actually presented envelope across all trials and subjects. Compared to linear forward ('encoding') models that derive temporal response functions (TRFs) independently for each EEG channel or source, stimulus reconstruction models represent multivariate impulse response functions that exploit information from all time lags and EEG channels/sources simultaneously. To allow for a neurophysiological interpretation of backward model coefficients, we additionally transformed them into linear forward model coefficients [bib_ref] On the interpretation of weight vectors of linear models in multivariate neuroimaging, Haufe [/bib_ref]. All analyses were performed using the multivariate temporal response function (mTRF) toolbox 99 (v1.5) for Matlab (v2016b).
Prior to model estimation, we split the data based on the two spatial-attention conditions (selective vs. divided), resulting in 120 trials per condition. Envelope reconstruction models were trained on concatenated data from selective-attention trials, only. Prior to concatenation, single trials were zero-padded for 600 ms to reduce discontinuity artefacts, and one trial was left out for subsequent testing. On each iteration, two different backward models were estimated, an envelope reconstruction model for the-be-attended speech stream (short: attended reconstruction model), and one for the to-be-ignored speech stream (short: ignored reconstruction model. Reconstruction models for attended and ignored speech signals were trained separately for attend-left and attend-right trials which yielded 120 decoders (60 attended, 60 ignored) per attentional setting. For illustrative purposes, we averaged the forward-transformed models of attended and ignored speech per hemisphere across all participants [fig_ref] Figure 4: Neural speech tracking of attended and ignored sentences [/fig_ref].
Evaluation of neural tracking strength. We analysed how strongly the attended compared to ignored sentences were tracked by slow cortical dynamics by quantifying the envelope reconstruction accuracy for individual trials. To this end, we reconstructed the attended and ignored envelope of a given trial using a leave-oneout cross-validation procedure. The two envelopes of a given trial were reconstructed using the models trained on all but the current trial from the same attention condition. The reconstructed onset envelope obtained from each model was then compared to onset envelopes of the actually presented speech signals using a 248-ms sliding window (rectangular window, step size of 1 (8 ms) sample). The resulting time courses of Pearson correlation coefficients, r attended and r ignored , reflect a temporally resolved measure of single-trial neural tracking strength or reconstruction accuracy 28 (see [fig_ref] Figure 4: Neural speech tracking of attended and ignored sentences [/fig_ref] and [fig_ref] Figure 4: Neural speech tracking of attended and ignored sentences [/fig_ref].
We proceeded in a similar fashion for divided-attention trials. Since these trials could not be categorized based on the to-be-attended and -ignored sides, we split them based on the ear that was probed at the end of the trial. Given that even in the absence of a valid attention cue, participants might still (randomly) focus their attention on one of the two streams, we wanted to quantify how strongly the probed and unprobed envelopes were tracked neurally. We used the reconstruction models trained on selective-attention trials to reconstruct the onset envelopes of divided-attention trials. Sentences presented in probed-left/unprobed-right trials were reconstructed using the attend-left/ignore-right reconstruction models while probed-right/unprobed-left trials used the reconstruction models trained on attend-right/ignore-left trials.
Attentional modulation of neural tracking. In close correspondence to the alpha lateralization index, we calculated a neural tracking index throughout sentence presentation. The index expresses the difference in neural tracking of the to-beattended and ignored sentence (in divided attention: probed and unprobed, respectively) 27 :
Neural tracking index ¼ ðr attended À r ignored Þ=ðr attended þ r ignored Þ ð4Þ
Positive values of the resulting index indicate that the attended envelope was tracked more strongly than the ignored envelope, and vice versa for negative values. Since individual sentences differed in length, for visualization and statistical analysis, we mapped their resulting neural tracking time courses onto a common time axis expressed in relative (percent) increments between the start and end of a given stimulus. We first assigned each sample to one of 100 bins covering the length of the original sentence in 1% increments. We then averaged across neighbouring bins using a centred rectangular 3% sliding window (1% overlap). The same procedure was applied to the time course of alpha-power lateralization following up-sampling to 125 Hz. Single-trial measures for the interval of final-word presentation were averaged across the final 35% of sentence presentation as this interval covered final-word onset across all 240 sentence pairs. We used the single-trial neural tracking index as (in-)dependent variables in our linear mixed-effects analyses (see below).
Statistical analysis. We used (generalized) linear mixed-effect models to answer the research questions outlined in . This approach allowed us to jointly model the impact of listening cues, neural filter strategies and various additional covariates known to influence behaviour. These included the probed ear (left/right), whether the later-probed sentence had the earlier onset (yes/no), as well as participants' age and hearing acuity (pure-tone average across both ears). To arbitrate between state-level (i.e., within-subject) and trait-level (i.e., between-subject) effects, our models included two separate regressors for each of the key neural measures. Between-subject effect regressors consisted of neural measures that were averaged across all trials at the single-subject level, whereas the within-subject effect was modelled by the trial-by-trial deviation from the subjectlevel mean [bib_ref] Fixed and random effects models: making an informed choice, Bell [/bib_ref].
Deviation coding was used for categorical predictors. All continuous variables were z-scored. For the dependent measure of accuracy, we used a generalized linear mixed-effects model (binomial distribution, logit link function). For response speed, we used a general linear mixed-effects model (Gaussian distribution, identity link function). Given the sample size of N = 155 participants, P values for individual model terms are based on Wald t-as-z-values for linear models [bib_ref] Evaluating significance in linear mixed-effects models in R, Luke [/bib_ref] and on z-values and asymptotic Wald tests in generalized linear models. All reported P values are corrected to control for the false discovery rate at q = 5% [bib_ref] Controlling the false discovery rate: a practical and powerful approach to multiple..., Benjamini [/bib_ref].
In lieu of a standardized measure of effect size for mixed-effects models, we report odds ratios (OR) for generalized linear models and standardized regression coefficients (β) for linear models along with their respective standard errors (SE).
All analyses were performed in R (v3.6.1) 107 using the packages lme4 (v1.1-23) [bib_ref] Fitting linear mixed-effects models using lme4, Bates [/bib_ref] , and sjPlot (v2.8.5).
## Model selection.
To avoid known problems associated with a largely data-driven stepwise model selection that includes the overestimation of coefficients 110 or the selection of irrelevant predictors 111 , the inclusion of fixed effects was largely constrained by our a priori defined hypotheses. The influence of visual cues and of neural measures was tested in the same brain-behaviour model. The brain-behaviour model of accuracy and response speed included random intercepts by subject and item. In a data-driven manner, we then tested whether model fit could be further improved by the inclusion of subject-specific random slopes for the effects of the spatial-attention cue, semantic cue, or probed ear. The change in model fit was assessed using likelihood ratio tests on nested models.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.
## Data availability
The complete neural and behavioural data required to reproduce the analyses supporting this work, as well as the auditory stimuli used in this study are publicly available in the study's Open Science Framework repository (https://osf.io/nfv9e/). Source data are provided with this paper.
[fig] Figure 2: Experimental design and behavioural benefit from informative cues. a Visualization of used 2 × 2 design35 . Levels of spatial and semantic cues differed on a trial-by-trial basis. Note that the effects of the semantic cue were of secondary importance to the current analyses. Top row shows the informative [+] cue levels, bottom row the uninformative [-] cue levels. b Schematic representation of the trial structure. Successive display of the two visual cues precedes the dichotic presentation of two sentences spoken by the same female talker. After sentence presentation, participants had to select the final word from four alternative words. c Left: accuracy per cue-cue combination. Coloured dots are individual (N = 155 participants) trial averages, black dots and vertical lines show group means with bootstrapped 95% confidence intervals (CI). Right: Individual cue benefits displayed separately for the two cues (top: spatial cue, bottom: semantic cue). Black dots indicate individual (N = 155) mean accuracy with bootstrapped 95 % CI. Histograms show the distribution of the difference in accuracy for informative vs. uninformative levels. OR: odds ratio parameter estimate from generalized linear mixedeffects models; two-sided Wald test (FDR-corrected); spatial cue: P = 1.36 × 10 -24 ; semantic cue: P = 0.68. d Left: Response speed per cue-cue combination. Coloured dots show individual (N = 155 participants) mean speed, black dots and vertical lines show group means with bootstrapped 95% CI. Right: Individual cue benefits displayed separately for the two cues (top: spatial cue, bottom: semantic cue). Black dots indicate individual (N = 155) mean speed with bootstrapped 95% CI. Box plots in (c) and (d) show median centre line, 25th to 75th percentile hinges, whiskers indicate minimum and maximum within 1.5 × interquartile range. β: slope parameter estimate from general linear mixed-effects models; two-sided Wald test (FDR-corrected); spatial cue: P = 4.49 × 10 -48 ; semantic cue: P = 2.49 × 10 -9 . Source data are provided as a Source Data file. [/fig]
[fig] Figure 3: Informative spatial cue elicits increased alpha-power lateralization before and during speech presentation. a Grand-average (N = 155 participants) whole-trial attentional modulation of 8-12 Hz auditory alpha power. Purple traces show the grand-average alpha lateralization index (ALI) for the informative (solid dark purple line) and uninformative spatial cue (dashed light purple line), each collapsed across semantic cue levels. Error bands indicate ±1 SEM. [/fig]
[fig] Figure 4: Neural speech tracking of attended and ignored sentences. a Schematic representation of linear backward model approach. Linear backward models estimated on selective-attention trials. Onset envelopes are reconstructed via convolution of auditory EEG responses with estimated backward models and compared actual envelopes to assess neural tracking strength and decoding accuracy (see Supplementary Methods). [/fig]
[fig] Figure 5: Relationship of alpha lateralization and neural speech tracking. a Hypothesized relationship of alpha power and neural speech tracking within the auditory region of interest. Changes in alpha lateralization are assumed to drive changes in neural tracking. Schematic representation for an attend-left trial. b Independence of neural speech tracking and alpha lateralization during final-word presentation as shown by the predictions from the same linear mixedeffect model. Plots show the predicted, non-significant effect of within-and between-subject variations in alpha lateralization on selective neural tracking, respectively. Blue lines indicate the respective predicted fixed effects with 95% confidence interval, grey thin lines in the left plot show N = 155 subjectspecific random slopes (included for illustrative purposes only), and grey dots show average predictions per subject. β: slope parameter estimate from the corresponding general linear mixed-effects model, two-sided Wald test (FDR-corrected). c Grand-average time courses of alpha lateralization and neural speech tracking during sentence presentation mapped to the same peri-stimulus time axis. Shown separately for selective attention (darker, solid curves) and divided-attention trials (lighter, dashed curves). Error bands reflect ±1 SEM. Note how the peak in neural speech tracking under selective attention precedes the peak in alpha lateralization. d Mean normalized cross-correlation of trial-averaged neural speech tracking and alpha lateralization time courses. The upper and lower bound of the shaded areas reflect the 97.5th and 2.5th percentile of surrogate data derived from 5000 independently permuted time courses of alpha power and neural speech tracking. Source data are provided as a Source Data file. | (2021) 12:4533 | https://doi.org/10.1038/s41467-021-24771-9 | www.nature.com/naturecommunications 0.013, P < 0.001; accuracy: GLMM; OR = 1.25, SE = 0.07, P = 0.006; see also Supplementary Fig. 6). Increased age led to less accurate and slower responses (accuracy: GLMM; OR = 0.80, SE = 0.08, P = 0.025; response speed: LMM; β =−0.15, SE = 0.03, P < 0.001). In contrast, increased hearing loss led to less accurate (GMM; OR = 0.75, SE = 0.08, P = 0.002) but not slower responses (LMM; β = −0.05, SE = 0.03, P = 0.21, see Supplementary Tables 1-2, and Supplementary Fig. 7). [/fig]
[fig] Figure 6: Neural speech tracking predicts listening behaviour. a Model predictions for the effect of neural tracking on behaviour for N = 155 [/fig]
|
Shotgun drug repurposing biotechnology to tackle epidemics and pandemics
[bib_ref] CANDO and the infinite drug discovery frontier, Minie [/bib_ref] [bib_ref] Combating ebola with repurposed therapeutics using the CANDO platform, Chopra [/bib_ref] [bib_ref] Systems biology-based investigation of cellular antiviral drug targets identified by gene-trap insertional..., Cheng [/bib_ref] [bib_ref] A new computational drug repurposing method using established disease-drug pair knowledge, Saberian [/bib_ref] [bib_ref] CANDO and the infinite drug discovery frontier, Minie [/bib_ref] [bib_ref] Combating ebola with repurposed therapeutics using the CANDO platform, Chopra [/bib_ref] [bib_ref] Network-based drug repurposing for novel coronavirus 2019-nCoV/SARS-CoV-2, Zhou [/bib_ref] [bib_ref] A SARS-CoV-2 protein interaction map reveals targets for drug repurposing, Gordon [/bib_ref] [bib_ref] CANDO and the infinite drug discovery frontier, Minie [/bib_ref] [bib_ref] Combating ebola with repurposed therapeutics using the CANDO platform, Chopra [/bib_ref] [bib_ref] CANDO and the infinite drug discovery frontier, Minie [/bib_ref] [bib_ref] CANDOCK: Chemical atomic network based hierarchical flexible docking algorithm using generalized statistical..., Fine [/bib_ref] [bib_ref] CANDO and the infinite drug discovery frontier, Minie [/bib_ref] [bib_ref] Combating ebola with repurposed therapeutics using the CANDO platform, Chopra [/bib_ref] [bib_ref] CANDOCK: Chemical atomic network based hierarchical flexible docking algorithm using generalized statistical..., Fine [/bib_ref] [bib_ref] Coronavirus puts drug repurposing on the fast track, Harrison [/bib_ref] [bib_ref] High-throughput screening and identification of potent broadspectrum inhibitors of coronaviruses, Shen [/bib_ref] [bib_ref] Repurposing of clinically developed drugs for treatment of Middle East respiratory syndrome..., Dyall [/bib_ref] [bib_ref] Coronavirus puts drug repurposing on the fast track, Harrison [/bib_ref]
# Author contributions
WM, ZF, and RS conceived and implemented all data analysis. WM, ZF, and RS wrote the manuscript. ZF generated the docking images. TM directed the prodrug analysis and provided expert opinion on antiviral drugs. GC substantially edited the manuscript and conceived ideas for data analysis.
## Conflicts of interest
The authors declare no conflicts of interest. |
Multiple Stones in a Single-System Ureterocele in a Child
Presence of multiple calculi in ureterocele is rare in children. A 6-year-old boy presented with hematuria in whom on x-ray and ultrasound multiple calculi were noted in the urinary bladder. At surgery a ureterocele containing multiple calculi was found. The postoperative (99m) Tc-Dimercaptosuccinic acid scan (DMSA) reported normal renal function.
# Introduction
Ureterocele is an abnormal congenital dilatation of the submucosal part of ureter in the bladder. It may present in a single though more commonly observed with the duplex urinary system, in which case it is usually associated with upper moiety. Ureteroceles may cause complications ranging from urinary tract infections to renal scarring and upper pole destruction in the duplex systems. Stones in ureteroceles are less common in the pediatric population. [bib_ref] Bilateral simple orthotopic ureteroceles with bilateral stones in an adult: a case..., Shamsa [/bib_ref] We present a 6-year-old boy with multiple stones in the right ureterocele diagnosed intra-operatively.
## Case report
A 6-years-old boy was admitted with hematuria. There was no previous history of urinary tract infection, hematuria, or abdominal pain. The physical examination was unremarkable and the laboratory tests were reported as normal except for hematuria. A plain abdominal radiograph demonstrated irregular radio opaque shadows in the pelvic area [fig_ref] Figure 1: Plain abdominal radiograph showing radiopaque stones in the pelvis [/fig_ref]. Abdominal ultrasonography showed a 3 cm × 2 cm, mobile, echoic area with an acoustic shadow in the bladder. Bladder wall thickness was 5 mm. No pathology was observed in the kidneys. Open surgery was performed based on the provisional diagnosis of urinary bladder stones; however, no stones were detected in it. On the right side, a ureterocele was found. The edematous and hyperemic right ureterocele had an extremely irregular mucosal surface. A ureteral catheter was placed inside the left ureter, and then the roof of the catheterized right ureterocele was opened with a small incision in order to remove seven stones from the ureterocele [fig_ref] Figure 2: Removed seven stones from the ureterocele [/fig_ref]. Upon analysis of the stones, the composition was calcium oxalate (100%). The postoperative DMSA renal scan was normal. At 2-year follow-up, the patient is asymptomatic with no vesicoureteral reflux or stone recurrence.
# Discussion
The presence of a single stone in a single ureterocele is often reported in adults in 4-39% cases though it is rarely reported in children. [bib_ref] Ureterocele containing a stone in a duplex system, Gharbi [/bib_ref] [bib_ref] Development of small calculi in an infant with bilateral single system ureteroceles, Gilbert [/bib_ref] [bib_ref] Voluminous "cobra-head" stone in a duplex system ureterocele: combined cysto-ureteroscopic and percutaneous..., Scuderi [/bib_ref]. The important factors that predispose to stone formation in the ureterocele are ureteral atony, occlusion, urine stasis in the ureterocele, family history and a urinary tract infection. [bib_ref] Adult bilateral non-obstructing orthotopic ureteroceles with multiple calculi: endoscopic management with review..., Singh [/bib_ref] The stones in ureterocele remain asymptomatic and progressively grow, unless they lead to hematuria or an obstruction. Gilbert et al reported three stones in the left ureterocele in a child; a single stone in the ureterocele was reported in other three pediatric cases reported in literature. [bib_ref] Ureterocele containing a stone in a duplex system, Gharbi [/bib_ref] [bib_ref] Development of small calculi in an infant with bilateral single system ureteroceles, Gilbert [/bib_ref] [bib_ref] Voluminous "cobra-head" stone in a duplex system ureterocele: combined cysto-ureteroscopic and percutaneous..., Scuderi [/bib_ref] In the index case, seven stones were removed from the right ureterocele.
Ultrasound and intravenous pyelogram are diagnostic in 50-70% cases.Extremely irregular and edematous wall and the adequate opening of the ureterocele precluded our radiologist to diagnose it on ultrasound. However, the inability to detect the ureterocele with multiple stones may also be attributed to the radiologist's inexperience. The aim of ureterocele treatment is to prevent urinary tract infections and vesicoureteral reflux while preserving renal function. Transurethral incision of the ureterocele and subsequent removal of the stones are the most commonly suggested treatment options for adult patients; excision of the ureterocele with either open or endoscopic surgery may also be preferred. [bib_ref] Bilateral single-system ureteroceles with multiple calculi in an adult woman, Mizuno [/bib_ref] In conclusion, hematuria may be the initial presentation of a ureterocele with urolithiasis in children. Moreover, in cases with a diagnosis of bladder stones, the possibility of stones inside the ureterocele should be considered.
[fig] Figure 1: Plain abdominal radiograph showing radiopaque stones in the pelvis. [/fig]
[fig] Figure 2: Removed seven stones from the ureterocele. [/fig]
|
Information sharing practices during the COVID-19 pandemic: A case study about face masks
We thank the reviewers for the time spent reviewing our manuscript, and we have edited the manuscript to address the points raised. This includes (but is not limited to): ensuring that we set a theoretical context, 2) making our theoretical and practical contributions clearer, 3) shortening the paper to focus on our key findings.In the following pages we respond to each point raised. Our responses are shown in blue text. We have also provided a copy of the revised manuscript with tracked changes on the submission system.We feel that these changes have significantly improved the paper, so once again thank you for the time spent to date in the review process. We hope that the manuscript is now ready for publication in PLOS ONE.
clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: Yes Reviewer #2: Yes Reviewer #3: Yes 5. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: Dear Author(s) Thank you for the opportunity to read this interesting piece of work. The manuscript is well-written. However, I feel the manuscript is rather long and descriptive falling short of critically engaging with previous research. three major issues need to be addressed. First, the research gap your study attempts to address needs to be expressed more explicitly. Second, the conclusions section failed to engage with previous research. Do your study findings confirm/refute previous research? Finally, the manuscript falls short of discussing the theoretical and practical contributions/implications of study. Good luck with revising your manuscript.
Thank you for these comments, which by addressing have helped improve the quality of our paper. Our contribution and key findings are now much clearer in the abstract of the paper, as well as the introduction, discussion and conclusions. To highlight the research gap, in the abstract, we now say "This research examines tweets in this longitudinal context and therefore provides novel insights into the dynamics of crisis communication in an ongoing crisis event with emerging scientific evidence.". This is re-emphasised in the manuscript.
To focus on the key findings, and therefore the theoretical and practical contributions, we have significantly shortened the paper. Concurrently, the paper has been slightly restructured. We have switched the first and second research questions around, to allow us to discuss the topic modelling and subject categories first and set the scene (i.e. circulation of information includes circulation of scientific evidence and expertise). This also involves restructuring the methodology, results and discussion to keep the correct order. In the discussion section, we have removed text, figures and tables, which we considered to be just descriptive with limited value to our overall arguments (for example, the information on subject sub-categories), as we agree these distracted from the narrative.
To further reduce the length of the paper, we have also merged what were the sections 'changes in the guidance for face masks' and 'knowledge, attitudes, and practices (KAP) towards face masks' into one section: 'Changes in the guidance and use of face masks'. This is now followed by a section entitled 'digital information ecology and crisis communication', which better sets a theoretical context for the work, which we draw upon in the discussion and conclusion sections.
What was the 'Limitations and further work' section has been separated and reduced in length, with the limitations now at the end of the methodology section and further work after the conclusions. This is to further help focus on the key findings and contributions at the end of the paper.
On reflection, we agree that the conclusions did not sufficiently engage with previous work. We have now added references to previous studies and theory in the 'Conclusions and further work' section, in addition to making these clearer in the discussions for each research question. One example of this engagement with previous theory now in our conclusion section is: "Although existing theoretical frameworks have provided useful approaches for conceptualising user's motivations for accessing information , their assessments of source credibility and organisation risk communication strategies , our work highlights that there are specific considerations to be factored in when crisis communication is dependent on emergent scientific evidence. We identified that challenges can arise due to the prolonged circulation out-of-date information, i.e. not strategic misinformation, nor "mis"-information at all, which can have serious ramifications for crisis communication practitioners working to deliver good quality information" Thank you again for your review.
Reviewer #2: The manuscript presents a very interesting topic, of global scope and likely to interest the world of academia, including sociologists, medical staff, marketers and social media experts, but also people not associated with the scientific community (eg., influencers). It has a universal character, which, however, is not a disadvantage, but deserves recognition. It is not written at a high level of generality, on the contrary, the article is very carefully prepared. The manuscript is presented in an intelligible fashion: the language is clear, correct, and unambiguous. No typographical or grammatical errors were found. The research was well thought out and well designed. The manuscript describes a technically sound piece of scientific research with data that supports the conclusions. The authors have fully provided all data on which the conclusions in the manuscript are based. These data have been provided both within the manuscript and its supplementary information.
We thank you for this really positive feedback. We have acknowledged this in the paper: "the universal character of our study contributes to developing understanding of the digital information ecology and provides practical insights for crisis communicators."
Reviewer #3: Review of Information sharing practices during the COVID-19 pandemic: a case study about face masks Submitted to PLOS ONE PONE-D-21-31236
The paper investigates the sharing of information about face masks during the COVID-19 pandamic in the UK. In a general sense, there is much to like about the paper. The paper is well written, and the overall argumentation is well crafted. Nevertheless, I have three substantial issues with respect to the paper that limit the scientific contribution substantially.
Thank you for your detailed comments and the time spend reviewing the manuscript. Please find our responses to each point raised below.
## Major issues
(1) CONTRIBUTION:
a. I think that you have an interesting story. However, I was not really convinced about the contribution of the paper. I believe your findings are important, but I really suggest coming up already in the introduction with the main findings. It takes much too long to see the results.
We have added two new paragraphs to the introduction. The first follows the research questions to outline their purpose. The last paragraph briefly describes the main findings:
"Analysing tweets in the longitudinal context of an ongoing crisis such as COVID-19 provides novel insight and understanding into the dynamics and influences of risk crisis communication . The
## Research, therefore, contributes to developing understanding of the digital information ecology, in particular the circulation of scientific information and expertise in an ongoing crisis context which is subject to uncertainty. this includes insight into when information is shared, including concerns about the recirculation of out-of-date articles, the plurality of information resources and communicators online, including the dominant role of traditional media but also some individuals in information circulation"
We have also made our contribution clearer in the abstract and conclusions section.
b. When reading the abstract, I was wondering about findings and what I can learn from this. You need to be much more specific to "sell" your results.
On reflection, we agree that we could better 'sell' our results. We have therefore re-written the abstract. This includes making our contribution clear in the first sentence and then clearly outlining the key findings in the second paragraph.
c. I also suggest focusing on findings that are (a) new and (b) impactful. For example., I think that it was interesting to discuss the relevance of outdated scientific results (page 32 in the manuscript). This is a key finding! I have no problem if you decide shortening the paper and focus on less but interesting findings.
Thank you for this observation. We have both restructured and shortened the paper to help focus on the key findings i.e. those we consider to be new and impactful. We have switched the first and second research questions around, to allow us to discuss the topic modelling and subject categories first (but in less detail) and set the scene (i.e. circulation of information includes circulation of scientific evidence and expertise). This also involves restructuring the methodology, results and discussion to keep the correct order. In the discussion section, we have removed text which we considered to be just descriptive with limited value to our overall arguments (for example. the information on subject sub-categories), as we agree these distracted from the narrative.
In the discussion section we are clearer about what our new findings are, and the conclusion now highlights three of these: circulation of out-of-date information, points in time that people are seeking information correspond with guidance/policy changes, and the continued dominance of the traditional media despite plurality of voices, although there are examples of high profile scientific organisations and individuals. d. Many findings that you report are not so surprising and lack (causal) support. You report sharing behavior, but you do not control for many drivers that might lead to specific sharing behavior (e.g., discussion of a news issue on TV, sharing of content by a celebrity etc.). Thus, I wonder if you should reduce the storyline and focus on the really new insights.
In addition to our response to point c, we have also shortened the paper to focus on the key findings by 1) merging what were the sections 'changes in the guidance for face masks' and 'knowledge, attitudes, and practices (KAP) towards face masks' into one section: 'Changes in the guidance and use of face masks'. 2) What was the 'Limitations and further work' section has been separated and reduced in length, with the limitations now at the end of the methodology section and further work after the conclusions. This is to further help focus on the key findings and contributions at the end of the paper. 3) Removed some of the figures and tables which were not necessary to support the key/new insights. 4) Deleted several paragraphs of text which were too descriptive and distracted from the overall narrative.
In our 'further work section' we have acknowledged that further work is required to understand the drivers and motivations behind specific sharing behaviour. e. Please add a new table that shows in a top-down fashion an overview of all your studies (data, method, and the subsequent findings).
We have added this suggested table to the beginning of the discussion section. 'Overview of research questions, applicable datasets, content analysed & methods and key findings'. This helps to highlight the key findings and provide a good overview of the studyso thank you for the suggestion.
(2) THEORY: The paper is basically free of theory. While I do not believe that you must add theory to the paper, I suggest presenting insights that are either supported by theory or that substantially challenge theories. Currently, I think your findings are in line with agenda setting theories. Thus, what can we learn? If we cannot learn much for theory testing and building, then I suggest focusing on the managerial impact of your results.
We have restructured the literature review section and now have a section called 'digital information ecology and crisis communication'. We use this section to better provide the theoretical context (with new references), and acknowledge that here are several microenvironments and relevant theories within this digital information landscape:
## "following perreault & perreault's work on the communication ecology of pandemic reporting [38] we acknowledge that there are several microenvironments within the crisis communication ecology, and this study examines how journalistic outputs, citizen commentary, governmental and public health messaging, as the associated communication practices coincide. each warrants consideration, as the motivations and agendas of different actors in the network contribute to the wider communication ecology. despite increased efforts to understand the digitally mediated communication practices in both scientific and crisis communication contexts, there is no systematised theorisation that attends to these overlapping and related concerns. developing this understanding is therefore key for effective crisis communication [39]."
Within this framing we draw upon relevant theoretical perspectives including Agenda setting, as well as Media Dependence studies and the Social-Mediated Crisis Communication model. We now present insights that either support or challenge some of these theories in the discussion and conclusion. Examples include: 1) "While studies have developed theoretical standpoints on how to mitigate the effects of misinformation in public health crises , what our findings show is that specific challenges can arise due to the prolonged circulation out-of-date information, i.e. not strategic misinformation, nor "mis"-information at all."
2) "As our study found that the distribution of retweets attached to organisations compared to individuals was 58.16% and 41.84% respectively, we agree with previous critiques that is it vital for Social-Mediated Crisis Communication theory to focus on how multiple actors engage with crisis response [60,62]"
3) "The shift and increasing focus on guidelines/policy in the 'most retweeted tweets' dataset, particularly after they became mandatory in England's and Scotland's shops, suggests that interest in a topic is linked to the latest news and major eventsas covered in Agenda Setting theory, a policy agenda/change has therefore directed the public's attention ."
(3) ROBUSTNESS: a. I would like to see your findings contrasted to google trends.
We have added google trends data to . Alongside the tweet frequency data, we discuss these trends that show people searching for information at particular points in time, in both the results and discussion.
b. Please also show your trend over time and clearly discuss the events that occurred during that time. This allows the reader to better understand confounding events that might drive search (via Google) or tweeting.
We have combined the Google Trends data with the previous diagram we had showing Tweet frequency which includes annotations of key policy changes . We have added arrows to these figures to show the timeframes used in our analysis i.e. TF1. TF2 and TF3. We have also added these arrows, and additional annotations, to the diagram of re-circulated articles to help the reader better understanding the timeline of events.
c. Please report the reach of the newspapers when discussing the popular domains.
In the section about 'Popular domains' we have added information (and reference to a survey) about the reach of the newspapers. Paragraph beginning 'Within the highest-ranking domains…." After . However, the two highest ranking newspapers in the survey, The Daily Mail and The Sun, only account for 1.17% and 0.96% of total retweets in our 'most retweeted tweets dataset' (ranks 20 th and 24 th ). Therefore, these results suggest that the reach of these newspapers may vary depending on the type of platform being used or topics being discussed" d. I think it would also be interesting to discuss why some other relevant posts have not been shared as much (discussion of non-findings).
We have pin-pointed some examples of some information not being shared as much as other posts. For example in the results we state:
"One external official resource that is noticeable missing from our sample and these conversations about scientific evidence and expertise is SAGE's official website which contains (since 29 May 2020 and with a delay summaries of their meetings and studies feeding into their discussions . Although in some cases, scientific resources such as The Lancet were drawn upon, official government channels of information about the scientific advice was not to the same extent -it is highly likely this is due to accessibility (i.e. readability) of those documents compared to the overviews/interpretations provided by the media.".
When talking about the reach of newspapers, we also acknowledge that "the reach of these newspapers may vary depending on the type of platform being used", as the Daily Mail and the Sun have the highest reach but do not rank as highly in our analysis.
In our further work section, we have recognised that understanding why posts are shared (or not) is a limitation of our work and state:
"Lastly, we (as have previous studies discussed that Twitter users are likely to have different motivations for posting and resharing informationhowever we have not explored these motivations.
To truly understand the digital information ecology, including why some posts are shared more than others, and there are differences between people's original tweets and retweeting behaviours, further work on these agendas is welcomed"
Overall, I am supportive towards this paperbut the authors need to streamline the paper to better show their contribution. I thank you for the opportunity to read this manuscript and I hope that the authors will find the comments helpful and that they will continue to follow this research stream.
6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.
If you choose "no", your identity will remain anonymous but your review may still be made public.
Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: No Reviewer #2: No Reviewer #3: No [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at [email protected]. Please note that Supporting Information files do not need this step.
All figures have been uploaded to PACE and PACE generated files downloaded and used for submission.
[table] Table 8: "Within the highest ranked domains, those categorised as journalism or broadcasting include: The Independent, The Guardian, The Daily Mirror and the BBC. In a survey (conducted April 2019 -March 2020) assessing the monthly reach of leading national newspapers in the United Kingdom [87], The Independent, Guardian and The Daily Mirror are ranked within the top 5 (taking into account phones, [/table]
|
Antibiotic tolerance is associated with a broad and complex transcriptional response in E. coli
Antibiotic treatment kills a large portion of a population, while a small, tolerant subpopulation survives. Tolerant bacteria disrupt antibiotic efficacy and increase the likelihood that a population gains antibiotic resistance, a growing health concern. We examined how E. coli transcriptional networks changed in response to lethal ampicillin concentrations. We are the first to apply transcriptional regulatory network (TRN) analysis to antibiotic tolerance by leveraging existing knowledge and our transcriptional data. TRN analysis shows that gene expression changes specific to ampicillin treatment are likely caused by specific sigma and transcription factors typically regulated by proteolysis. These results demonstrate that to survive lethal concentration of ampicillin specific regulatory proteins change activity and cause a coordinated transcriptional response that leverages multiple gene systems.OPENwith known functions; however, there are seven genes categorized as "function unknown" or "general function prediction only" upregulated during ampicillin treatment, which suggests their function is related to survival or recovery(Fig. 1).The most upregulated category is translation, ribosomal structure and biogenesis (68 genes). While the accumulation of proteins involved in translational machinery has previously been shown in ampicillin treated E. coli 25 , it is important to note that we removed rRNA during the RNA-seq library preparation process and most of the genes in this category encode ribosomal proteins. Thus, we will not go into depth for this particular group to avoid potential biases in our datasets. The remaining groups with more than 20 upregulated genes are amino acid transport and metabolism (28 genes); energy production and conversion (26 genes); and posttranslational modification, protein turnover, chaperones (23 genes). We explore these groups further in the context of the TRN.Analysis of the sigma factor (σ) regulatory network shows that the activity of σ 70 , σ 38 , and σ 32 increase during ampicillin treatment. Sigma factors are the master regulators of gene expression and
When disaster strikes, a coordinated response is required to mitigate the effects and survive in the face of adversity. Cells have developed an array of regulatory networks to coordinate cellular activity and respond to their environment. These regulatory events occur simultaneously and at multiple levels, including transcriptional, translational, and post-translational regulation. The identification of individual regulatory interactions in combination with bioinformatics approaches has resulted in multiple databases detailing regulation throughout the cell, the most detailed databases are currently for E. coli (e.g. RegulonDB 1 , EcoCyc 2 ). The depth and availability of information regarding regulatory dynamics provide a new framework for analyzing sequencing data in the context of the whole cell. As such, these types of databases have been applied to analyze network topology [bib_ref] Functional organisation of Escherichia coli transcriptional regulatory network, Martinez-Antonio [/bib_ref] as well as experimental datasets [bib_ref] Antibiotics disrupt coordination between transcriptional and phenotypic stress responses in pathogenic bacteria, Jensen [/bib_ref] [bib_ref] NetProphet 2.0: mapping transcription factor networks by exploiting scalable data resources, Kang [/bib_ref] [bib_ref] The Escherichia coli transcriptome mostly consists of independently regulated modules, Sastry [/bib_ref] [bib_ref] Independent component analysis of E. coli's transcriptome reveals the cellular processes that..., Tan [/bib_ref] [bib_ref] Engineering regulatory networks for complex phenotypes in E. coli, Liu [/bib_ref]. This data has been used to examine transcriptional regulatory networks (TRNs) at the transcriptomic level, which consider gene expression as a function of regulatory interactions by DNA-binding proteins (transcription factors and sigma factors) [bib_ref] Constructing transcriptional regulatory networks, Blais [/bib_ref]. We focused on RNA-sequencing data in the context of cell growth and antibiotic stress. Antibiotics disrupt microbial systems and cause targeted cell death; bacterial survival of antibiotics is therefore dependent on cellular networks adjusting to circumvent the physiological impact of the antibiotic.
Antibiotic tolerance, and more specifically antibiotic persisters, are a perfect example of how alterations in metabolic activity (i.e. gene expression and protein activity) lead to changes in cell phenotypes, increasing the chance of survival. Persisters are a subpopulation of cells that tolerate antibiotics for long periods of time despite being genetically identical to susceptible cells. After antibiotic treatment, the surviving population may mutate or acquire resistance genes from their environment [bib_ref] Antibiotic tolerance facilitates the evolution of resistance, Levin-Reisman [/bib_ref] [bib_ref] Frequency of antibiotic application drives rapid evolutionary adaptation of Escherichia coli persistence, Van Den Bergh [/bib_ref] [bib_ref] Bacterial persistence promotes the evolution of antibiotic resistance by increasing survival and..., Windels [/bib_ref]. Persistence and antibiotic tolerance have been studied extensively, but the heterogeneity of the tolerant population combined with the variance between antibiotic survival of diverse species and growth conditions has led to a number of different, often conflicting, models regarding the causative factors underlying antibiotic tolerance. Some conflicting reports suggest the presence of subpopulations that are metabolically repressed, particularly with respect to translation, while others have found that active metabolism plays a vital role in antibiotic tolerance [bib_ref] Bacterial persister cell formation and dormancy, Wood [/bib_ref] [bib_ref] Inhibition of stationary phase respiration impairs persister formation in E. coli, Orman [/bib_ref] [bib_ref] Persistent persister misperceptions, Kim [/bib_ref] [bib_ref] Metabolic aspects of bacterial persisters, Prax [/bib_ref] [bib_ref] Enhanced efflux activity facilitates drug tolerance in dormant bacterial cells, Pu [/bib_ref]. A few studies have leveraged [bib_ref] Specialized persister cells and the mechanism of multidrug tolerance in Escherichia coli, Keren [/bib_ref] [bib_ref] Unique transcriptional profile of native persisters in Escherichia coli, Matsumoto [/bib_ref] [bib_ref] Characterization and transcriptome analysis of Mycobacterium tuberculosis persisters, Keren [/bib_ref] RNA-sequencing (RNA-seq) [bib_ref] Enhanced efflux activity facilitates drug tolerance in dormant bacterial cells, Pu [/bib_ref] [bib_ref] Transcriptomic study on persistence and survival of Listeria monocytogenes following lethal treatment..., Wu [/bib_ref] [bib_ref] Identification of novel genes that promote persister formation by repressing transcription and..., Long [/bib_ref] and proteomic data [bib_ref] Selective proteomic analysis of antibiotic-tolerant cellular subpopulations in Pseudomonas aeruginosa biofilms, Babin [/bib_ref] [bib_ref] A proteomic signature of dormancy in the actinobacterium Micrococcus luteus, Mali [/bib_ref] [bib_ref] ATP-dependent dynamic protein aggregation regulates bacterial dormancy depth critical for antibiotic tolerance, Pu [/bib_ref] , and they have identified several individual genes and systems related to antibiotic tolerance.
Other work has used gene knockouts and/or overexpression to study specific elements of antibiotic survival [bib_ref] Ranking of persister genes in the same Escherichia coli genetic background demonstrates..., Wu [/bib_ref] [bib_ref] Transposon mutagenesis identifies novel genes associated with Staphylococcus aureus persister formation, Wang [/bib_ref] [bib_ref] Role of global regulators and nucleotide metabolism in antibiotic tolerance in Escherichia..., Hansen [/bib_ref]. However, knockouts and gene overexpression can lead to disrupted cellular networks, alter growth rates, and impact mechanisms that alter awakening from dormancy (growth rates and dormancy have been well-studied [bib_ref] The frequency of persisters in Escherichia coli reflects the kinetics of awakening..., Joers [/bib_ref]. Nonetheless, these works and others have attributed several different mechanisms to antibiotic tolerance. Many genes identified by these studies are sigma factors or transcription factors, regulatory proteins that affect gene expression [bib_ref] ATP-dependent dynamic protein aggregation regulates bacterial dormancy depth critical for antibiotic tolerance, Pu [/bib_ref] [bib_ref] Role of global regulators and nucleotide metabolism in antibiotic tolerance in Escherichia..., Hansen [/bib_ref] [bib_ref] Bacterial persistence is an active sigmaS stress response to metabolic flux limitation, Radzikowski [/bib_ref] [bib_ref] SoxS and Rob of Escherichia coli: Global regulators of multidrug resistance, virulence..., Duval [/bib_ref] [bib_ref] The multiple antibiotic resistance operon of enteric bacteria controls DNA repair and..., Sharma [/bib_ref]. The significance of these genes to antibiotic survival suggests that a broader regulatory network underlies the antibiotic stress response. Although understanding the role of individual genes in antibiotic survival is important, what is currently lacking is a more holistic understanding at the network level. While it has been a long-standing goal to identify one or more genes unique to antibiotic tolerance, our TRN analysis shows that the transcriptional response to antibiotic treatment is controlled via a whole-cell, network response consisting of smaller, overlapping networks, some of which are redundant.
We have taken a systems biology, top-down approach to explore antibiotic survival of tolerant populations, and obtained deep, RNA-seq data on both antibiotic-treated and -untreated populations emerging from stationary phase. Our results show that cells have an active and coordinated response to lethal antibiotic stress that alters the activity of specific sigma factors and transcription factors (TFs). Existing literature allows us to connect the activity of these regulatory proteins and changes in gene expression to the pathways involved in antibiotic tolerance. Our findings suggest that many of the regulatory proteins activated during antibiotic stress are regulated by protein degradation. Protein aggregation was recently demonstrated in persister cells [bib_ref] ATP-dependent dynamic protein aggregation regulates bacterial dormancy depth critical for antibiotic tolerance, Pu [/bib_ref] , and our previous work shows that interfering with degradation at the protease ClpXP increases antibiotic tolerance in E. coli [bib_ref] Proteolytic queues at ClpXP increase antibiotic tolerance, Deter [/bib_ref]. To test if the transcriptional changes are due to altered proteolysis, we then compared the transcriptome of populations with and without slowed protein degradation using our previously established system [bib_ref] Proteolytic queues at ClpXP increase antibiotic tolerance, Deter [/bib_ref]. The analysis of the TRNs in these populations provides evidence that slowed proteolytic processing of native proteins activates a stress response that promotes recovery once the stress is removed.
# Results and discussion
We hypothesize that bacteria survive antibiotics by a network response where intracellular networks communicate to adjust cellular responses, rather than a single gene or system as the sole agent of antibiotic tolerance. While individual genes or systems can alter a population's ability to survive antibiotics, we assert that the individual genes/systems regulate the response but that multiple pathways play a role in antibiotic survival. Here, we support this hypothesis using transcriptional data from antibiotic-treated and untreated populations, and then compare our results to the current knowledge in the field. We attest that the conflicting data surrounding genes related to antibiotic survival (persistence and tolerance) in the literature is due, in part, to researchers slightly or significantly altering network dynamics through gene knockouts and overexpression, and that a more holistic system biology approach could provide new insights.
RNA-sequencing of ampicillin treated and untreated cultures. We will initially discuss gene expression in three different populations: a stationary phase population, which was then diluted into fresh media and incubated for three hours without ampicillin (untreated, Amp-) or with lethal concentrations of ampicillin (amptreated, Amp+). We used rapid filtration to harvest cell samples followed by flash freezing in liquid nitrogen. This allows us to recover any whole cells present after antibiotic treatment regardless of viability (this includes persisters and cells that are considered "viable but not culturable" (VBNC)). An advantage to this method is that it enables the concentration of cells while minimizing the possible time for RNA degradation. We found that rapid filtration has greater than 2.5 times fewer dead cells during harvesting than pelleting through centrifugation (a commonly used alternative to filtration). The VBNC levels were similar in the untreated and treated populations; we observed that VBNCs levels were 2.7-fold and 2.2-fold higher than culturable cells (counted on agar plates) in stationary phase and ampicillin treated cultures respectively. Recent research has suggested that VBNC cells are physiologically related to persisters; both populations are often found together 34 , increase due to stress [bib_ref] Recent findings on the viable but nonculturable state in pathogenic bacteria, Oliver [/bib_ref] [bib_ref] Viable but nonculturable bacteria: a survival strategy, Colwell [/bib_ref] , appear stochastically in non-stressed environments [bib_ref] Viable but nonculturable and persister cells coexist stochastically and are induced by..., Ayrapetyan [/bib_ref] , and can withstand greater stress than the susceptible population [bib_ref] Relationship between the viable but nonculturable state and antibiotic persister cells, Ayrapetyan [/bib_ref] [bib_ref] Development of a direct viable count procedure for the investigation of VBNC..., Besnard [/bib_ref] [bib_ref] Antibiotic pressure can induce the viable but non-culturable state in Staphylococcus aureus..., Pasquaroli [/bib_ref] [bib_ref] Detection of viable but nonculturable Escherichia coli O157:H7 using propidium monoazide treatments..., Xiao [/bib_ref] [bib_ref] Resuscitation of Escherichia coli VBNC cells depends on a variety of environmental..., Pinto [/bib_ref]. However, the presence of VBNCs demonstrates the heterogeneity of the population that survive antibiotic treatment, but VBNCs are difficult to separate from persister cells. Regardless, we are interested in the tolerant population as a whole and the systems we have identified relating to antibiotic tolerance may be relevant to the whole surviving population or specific subpopulations.
This work builds off our previously reported results with antibiotic tolerance, and uses a similar setup with the same strains of E. coli [bib_ref] Proteolytic queues at ClpXP increase antibiotic tolerance, Deter [/bib_ref]. The sequencing data was analyzed using DESeq. Here we consider significant changes in gene expression to be greater than two-fold with an adjusted p-value < 0.1. We began by examining changes in gene expression between the stationary phase and the untreated populations. Comparing these two populations allows us to identify how the cells adjust to their new environment, as diluting the cells into fresh media is a critical aspect of selection for ampicillin tolerance from stationary phase (ampicillin is most effective on growing cells) [bib_ref] The frequency of persisters in Escherichia coli reflects the kinetics of awakening..., Joers [/bib_ref] [bib_ref] Single-cell protein induction dynamics reveals a period of vulnerability to antibiotics in..., Gefen [/bib_ref] [bib_ref] Persisters: Methods for isolation and identifying contributing factors-a review, Rowe [/bib_ref]. The majority of the genes with significantly different expression in the untreated population decreased from stationary phase (rather than increased), suggesting that cells are expending fewer resources in the new media. This phenomenon is likely due to the introduction of fresh nutrients and dilution of spent media resulting in a less stressful environment. Data for the untreated population provides information on the changes specific to the new environment; thus, we can focus on changes specific to antibiotic stress by identifying differences in gene expression unique to the treated population. These differences provide key insights into the molecular mechanisms of antibiotic tolerance and the role of transcriptional regulatory networks (TRNs) in antibiotic survival and recovery.
Clusters of Orthologous Groups (COGs) during ampicillin treatment. Using a top-down approach, we analyzed changes in gene expression based on the general functional classification, as defined by Clusters of Orthologous Groups (COGs) [bib_ref] The COG database: A tool for genome-scale analysis of protein functions and..., Tatusov [/bib_ref] [bib_ref] Expanded microbial genome coverage and improved protein family annotation in the COG..., Galperin [/bib_ref]. Although the COG classification system is limited to known functions, more qualitative than quantitative, and often used to study evolutionary relationships; it provides a practical method of grouping genes to identify notable trends in the data. When we analyze the COGs for gene expression unique to the amp-treated population, we find that a few categories stand out [fig_ref] Figure 1: Cluster of Orthologous Genes [/fig_ref]. We will focus on groups essential components of transcriptional regulation. These proteins interact with RNA polymerase to initiate transcription, and their activity changes depending on cellular conditions [bib_ref] The essential activities of the bacterial sigma factor, Davis [/bib_ref]. Sigma factors are often regulated at the post-translational level by anti-sigma factors or alterations in protein degradation, and different sigma factors are active under specific circumstances. The activity of some sigma factors can be altered by environmental conditions with little change in their mRNA level [bib_ref] Bacterial sigma factors and anti-sigma factors: Structure, function and distribution, Paget [/bib_ref]. However, we can contextualize changes in gene expression controlled by sigma factors by using TRN analysis. To do so, we used all the available information on the direct regulation of all annotated genes from EcoCyc 2 as defined at the time of analysis (File S1). We compared overall transcriptional activity by mapping changes in expression to specific sigma factors in the context of COGs to get a global picture of gene expression [fig_ref] Figure 2: Stress related sigma factors [/fig_ref]. We are the first (to our knowledge) to use COGs to support a TRN analysis based on experimental data.
Extensive studies on sigma factor activity, both in vivo and in silico, have identified or predicted the sigma factor(s) responsible for transcription at specific promoters throughout the genome [bib_ref] The global transcriptional response of Escherichia coli to induced sigma 32 protein..., Zhao [/bib_ref] [bib_ref] Genome-Wide Transcriptional Response to varying RpoS levels in Escherichia coli K-12, Wong [/bib_ref] [bib_ref] In vitro transcription profiling of the sigmaS subunit of bacterial RNA polymerase:..., Maciag [/bib_ref] [bib_ref] Mutational analysis of Escherichia coli sigma28 and its target promoters reveals recognition..., Koo [/bib_ref] [bib_ref] Characterization of the Escherichia coli sigma E regulon, Dartigalongue [/bib_ref] [bib_ref] A novel method for prokaryotic promoter prediction based on DNA stability, Kanhere [/bib_ref]. We compared both the untreated and amp-treated populations to stationary phase. The total number (count) of genes that changed expression is higher in the treated population than in the untreated population as compared to the stationary phase population [fig_ref] Figure 2: Stress related sigma factors [/fig_ref]. As expected, the amp-treated population had more expression associated with stress related sigma factors, RpoH (σ 32 ), RpoS (σ 38 ), and RpoE (σ 24 ), than the untreated population (p < 0.001) [fig_ref] Figure 2: Stress related sigma factors [/fig_ref]. Many of the genes that increase in expression in response to ampicillin are regulated by both RpoD (σ 70 ) as Only genes that were unique to the treated population (compared to stationary phase) were included; genes that show differential regulation in both the treated and untreated population (compared to stationary phase) were removed. Genes in multiple functional categories were accounted for more than once. Right: Functional categories and codes from REF [bib_ref] The COG database: A tool for genome-scale analysis of protein functions and..., Tatusov [/bib_ref]. Up: increased > twofold gene expression. Down: decreased < twofold gene expression. Genome: number of genes with that functional category based on the annotated genome. www.nature.com/scientificreports/ www.nature.com/scientificreports/ well as another sigma factor. The overlapping regulons means that some genes are accounted for twice, but the increase in expression for many of these genes is probably due to a stress-related sigma factor rather than RpoD. As expected, there was an increase in the expression of many genes regulated by sigma factors related to stress, which have previously been associated with antibiotic tolerance [bib_ref] ATP-dependent dynamic protein aggregation regulates bacterial dormancy depth critical for antibiotic tolerance, Pu [/bib_ref] [bib_ref] Bacterial persistence is an active sigmaS stress response to metabolic flux limitation, Radzikowski [/bib_ref]. Namely, RpoH (σ 32 ; associated with heat shock 55 ), RpoS (σ 38 ; associated with stationary phase and stress [bib_ref] The RpoS-mediated general stress response in Escherichia coli, Battesti [/bib_ref] , and RpoE (σ 24 ; also associated with heat shock 57 ) had more active regulons in the treated compared to the untreated population. Our analysis shows that RpoH transcribes several genes related to one of the COG groups noted earlier, 'posttranslational modification, protein turnover and chaperones' . Genes specifically regulated by RpoH encode the proteins ClpB, GroL, Lon, DnaK, DnaJ, and FtsH, many of which are related to protein degradation, which we will discuss. Meanwhile, the sigma factor RpoS regulates several genes related to energy production and conversion, and RpoE regulates a few genes related to cell wall biogenesis [fig_ref] Figure 2: Stress related sigma factors [/fig_ref]. While these trends are exciting to find in our data, it is clear that other components of transcriptional regulation, namely TFs, are playing a critical role in controlling gene expression and the antibiotic stress response.
Analysis of the transcription factor (TF) regulatory network shows that the activity of global regulators increase during ampicillin treatment. TFs act on a variety of scales, from simultaneously regulating hundreds of genes (e.g. CRP 58 ) to simply a few genes [bib_ref] Local and global regulation of transcription initiation in bacteria, Browning [/bib_ref]. As with sigma factors, the activity of many TFs, especially global TFs, is regulated post-translationally (i.e. at the protein level) 58,60-62 , and therefore directly measuring expression of TFs is not typically informative in regards to their activity. We used TRN analysis to identify the role of TFs that alter gene expression in response to ampicillin.We mapped changes in expression to all possible direct regulatory interactions that could account for that change in RNA levels during ampicillin treatment. Analyzing this network makes it very clear that the regulatory pathways at play in the population's response to ampicillin are highly interconnected and likely redundant [fig_ref] Figure 3: TFs associated with antibiotic tolerance are more active in the ampicillin treated... [/fig_ref].
Several TFs are more active during ampicillin treatment, including all seven global regulators CRP, FNR, IHF, FIS, ArcA, NarL and Lrp. The impact of global TFs on the system is not entirely surprising as they are critical factors for a healthy cell physiology. Global regulators are responsible for the up and down regulation of genes across a heterogenous array of functional groups, and they are an integral facet of the hierarchical structure of the TRN in E. coli [bib_ref] Functional organisation of Escherichia coli transcriptional regulatory network, Martinez-Antonio [/bib_ref]. The altered activity of these TFs in the treated population suggests that the cells are actively regulating their metabolism and responding at the whole-cell level. REF [bib_ref] Role of global regulators and nucleotide metabolism in antibiotic tolerance in Escherichia..., Hansen [/bib_ref] found that global regulation is critical to antibiotic survival and the authors suggest that global regulators control several redundant pathways related to antibiotic stress [bib_ref] Role of global regulators and nucleotide metabolism in antibiotic tolerance in Escherichia..., Hansen [/bib_ref] , which is also supported by our results. Furthermore, both fnr and crp have previously been identified in gene knockout studies as being related to antibiotic tolerance [bib_ref] Genetic basis of persister tolerance to aminoglycosides in Escherichia coli, Shan [/bib_ref] [bib_ref] Impacts of global transcriptional regulators on persister metabolism, Mok [/bib_ref].
Another global TF identified in our analysis and previously related to antibiotic tolerance is Fur. A Δfur strain has been shown to have decreased levels of tolerance to ciprofloxacin but also have an increased chance of acquiring resistance [bib_ref] Perturbation of iron homeostasis promotes the evolution of antibiotic resistance, Mehi [/bib_ref]. Previous work examining expression data of the Δfur strain with and without ciprofloxacin identified entA, entC and sufA expression as being affected by ciprofloxacin and Fur. As in, Fur regulation was critical for expression of these genes and tolerance to ciprofloxacin [bib_ref] Perturbation of iron homeostasis promotes the evolution of antibiotic resistance, Mehi [/bib_ref]. These three genes also show increased expression in our ampicillin treated population and have been identified in other types of tolerance as well, such as n-butanol tolerance [bib_ref] Visualizing evolution in real time to determine the molecular mechanisms of n-butanol..., Reyes [/bib_ref] , which suggests that they are involved as a general stress response mechanism. The operon entCDEBAH is regulated by both CRP and Fur, encodes enterobactins, and is associated with iron homeostasis and the oxidative stress response. Moreover, enterobactins protect intracellularly against oxidative stress independently of iron availability 67 , supporting that the oxidative and antibiotic stress responses are integrated.
Transcription factor (TF) regulatory network shows increased activity of specific, local regulators during ampicillin treatment. When we further examined pathways related to iron-sulfur (Fe-S) clusters and oxidative stress, we identified a group of genes with changes in expression regulated by local regulators OxyR and IscR. Notably, the ampicillin treated population had increased expression of genes required for Fe-S cluster formation: sufS, sufE, sufC, iscA, iscR, iscS, iscU and iscX. Fe-S clusters act as protein co-factors and serve a wide variety of functions in the cell; they are suited to a variety of processes including redox sensing, catalysis, or electron transfer. Notably, the clusters are also part of the oxidative stress response [bib_ref] How is Fe-S cluster formation regulated?, Mettert [/bib_ref]. Fe-S clusters have been associated with increased development of antibiotic resistance to ciprofloxacin [bib_ref] Perturbation of iron homeostasis promotes the evolution of antibiotic resistance, Mehi [/bib_ref] and uptake of aminoglycosides [bib_ref] Fe-S cluster biosynthesis controls uptake of aminoglycosides in a ROS-less death pathway, Ezraty [/bib_ref].
Our findings further support that Fe-S clusters are involved in antibiotic tolerance. This system also highlights a restraint of our study, namely that we are only mapping regulatory events as opposed to deregulatory events. Derepression (removal of repression) is another type of transcriptional regulation that affects antibiotic survival. In this instance, when accounting for genes regulated by IscR, the isc genes were not included because IscR represses these genes and the genes increased in expression in our network. However, increased expression of isc genes is likely due to derepression of IscR, which is supported by the simultaneous increase of the suf genes, which IscR also regulates [bib_ref] How is Fe-S cluster formation regulated?, Mettert [/bib_ref]. The same can be said for the regulation by Fur, which is also associated with iron homeostasis. While Fur was identified as significantly more active in the treated population and primarily acts as a repressor, increased expression resulting from Fur's derepression was not accounted for in the TRN analysis. This particular group of TFs demonstrates how changes in regulatory interactions must be identified contextually with respect to when these changes might occur. The context of regulation is particularly significant for dual-regulators TFs that can activate and repress different genes depending on the state of the cell (e.g. global regulators).
Another group of local TFs that play a role in the antibiotic-treated population is the Mar/Rob/SoxS regulon. These three TFs are closely related, with degenerative consensus sequences that result in considerable overlap of regulated genes, and the regulon is positively associated with antibiotic tolerance and multidrug resistance [bib_ref] SoxS and Rob of Escherichia coli: Global regulators of multidrug resistance, virulence..., Duval [/bib_ref] www.nature.com/scientificreports/ www.nature.com/scientificreports/ Constitutive expression of SoxS has been shown to cause multiple antibiotic resistance in clinical isolates of E. coli [bib_ref] Constitutive soxR mutations contribute to multiple-antibiotic resistance in clinical Escherichia coli isolates, Koutsolioutsou [/bib_ref]. Extensive work has shown that the TF MarA plays a role in antibiotic tolerance, as the operon was named for its role in multiple antibiotic resistance [bib_ref] Genetic and functional analysis of the multiple antibiotic resistance (mar) locus in..., Cohen [/bib_ref]. MarA is expressed stochastically and is primarily regulated by degradation at the Lon protease [bib_ref] Active degradation of MarA controls coordination of its downstream targets, Rossi [/bib_ref] [bib_ref] Stochastic expression of a multiple antibiotic resistance activator confers transient resistance in..., El Meouche [/bib_ref]. SoxS is also degraded by Lon as well as the protease FtsH. The rapid degradation of MarA and SoxS enables the cell to rapidly recover once a stress is removed [bib_ref] Proteolytic degradation of Escherichia coli transcription activators SoxS and MarA as the..., Griffith [/bib_ref]. We will discuss more about the importance of protein degradation to antibiotic survival later. Meanwhile, the TF Rob is in an inactive state until it is induced by binding to specific substrates 74 , a perfect example of how analyzing the mRNA level of a TF is not sufficient to understand protein activity. It is important to note that the level of rob, marA or soxS mRNA may not significantly change, while the proteins' concentration and/or regulatory action do change. This fact underscores the importance of analyzing transcriptional networks rather than focusing on TF expression levels.
Altered proteolytic activity affects transcriptional regulation similarly to ampicillin treatment. In addition to the regulation of MarA and SoxS through degradation, proteolytic processes play a vital role in antibiotic tolerance, as many previous works have suggested [bib_ref] Proteolytic queues at ClpXP increase antibiotic tolerance, Deter [/bib_ref] [bib_ref] Clp chaperones and proteases are central in stress survival, virulence and antibiotic..., Frees [/bib_ref] [bib_ref] Activated ClpP kills persisters and eradicates a chronic biofilm infection, Conlon [/bib_ref]. These effects on the TRN during ampicillin treatment suggest that the activity of many transcription factors is controlled at the proteolytic level and that disruption of degradation during environmental stress leads to an increase in the activity of these TFs.
We have previously demonstrated that interfering with proteolytic activity causes an increase in antibiotic tolerance. To do so, we overexpressed a fluorescent protein tagged with a short amino acid degradation tag that causes it to compete for degradation at the protease complex ClpXP without affecting growth rate [bib_ref] Proteolytic queues at ClpXP increase antibiotic tolerance, Deter [/bib_ref]. This competition results in a proteolytic queue, wherein proteins compete to be processed (i.e. degraded); thus the degradation of native proteins effectively slows due to an influx of new proteins with a high affinity for degradation (i.e. CFP-LAA; [fig_ref] Figure 4: Proteolytic queueing increases survival of ampicillin [/fig_ref] 77-81 . The advantage of using this artificial mechanisms to slow down protease activity is it results in alteration of network dynamics without noticeable effects on growth, unlike tradition protease knockouts [bib_ref] Proteolytic queues at ClpXP increase antibiotic tolerance, Deter [/bib_ref]. We hypothesized that the increase in antibiotic tolerance via proteolytic queueing is due to an amplification of the transcriptional response. We compared the transcriptome of populations diluted from stationary phase into fresh media containing IPTG to induce expression of CFP-LAA. Then we incubated the cells for three hours without ampicillin (queueing population) or with lethal concentrations of ampicillin (queueing-amp population). We will discuss changes in gene expression as compared to stationary phase.
If slowed proteolytic activity leads to an increase in antibiotic tolerance, TRN behavior should be similar in the amp-treated and queueing populations. Indeed, that is the case. The queueing-amp population, which simultaneously responds to the induction of a proteolytic queue and ampicillin treatment, shares most of these similarities as well. The transcriptional response seen in the queuing-amp population seems to be a magnification of the response in either the queueing or amp-treated population alone, with many of the same trends [fig_ref] Figure 5: Proteolytic queueing and ampicillin treatment similarly affect gene expression in E [/fig_ref].
This led us to hypothesize that the queueing is effectively 'quickening' (i.e. causing certain events to happen at a faster timescale) the response to the antibiotic. We then compared the ampicillin treated culture's RNA profile after six hours to the untreated and treated populations. The gene expression profile is similar at three and six hours; however, hundreds of additional genes changed their expression by six hours [fig_ref] Figure 5: Proteolytic queueing and ampicillin treatment similarly affect gene expression in E [/fig_ref]. In general, the trends between gene expression in the amp-treated, queueing, and queueing-amp populations are quite similar. Upregulated genes common to the three populations include genes and systems related to tolerance discussed earlier, including those related to Fe-S clusters, oxidative stress and heat shock. When analyzing the COGs most highly regulated in queueing-tolerance, it appears that the population is heavily regulating processes related to metabolism and energy [fig_ref] Figure 5: Proteolytic queueing and ampicillin treatment similarly affect gene expression in E [/fig_ref]. It has previously been reported that energy levels largely affect antibiotic tolerance and persistence [bib_ref] ATP-dependent dynamic protein aggregation regulates bacterial dormancy depth critical for antibiotic tolerance, Pu [/bib_ref] [bib_ref] ATP-dependent persister formation in Escherichia coli, Shan [/bib_ref] [bib_ref] Persister formation in Staphylococcus aureus is associated with ATP depletion, Conlon [/bib_ref]. Particularly relevant to our work, are the findings of REF [bib_ref] ATP-dependent dynamic protein aggregation regulates bacterial dormancy depth critical for antibiotic tolerance, Pu [/bib_ref] , which demonstrates that lower ATP levels led to protein aggregation, which also occurs in ampicillin treated E. coli. www.nature.com/scientificreports/ Mechanistically, protein aggregation likely results from slow protease activity, since many proteases require ATP to function [bib_ref] Proteases and their targets in Escherichia coli, Gottesman [/bib_ref]. In comparison, proteolytic queueing via overexpression of CFP-LAA interferes with degradation at ClpXP and causes a build-up of proteins normally degraded by the complex. The queue also affects degradation rates by other proteases 81 and leads to protein aggregation. When looking at sigma factors and TFs that affect the transcriptional response, in this case, the alternative sigma factors are more active in the queueingamp population compared to queueing or amp-treated cells [fig_ref] Figure 6: Proteolytic queueing and ampicillin treatment affect specific regulatory proteins in E [/fig_ref].
We propose that the magnified response in the queueing-amp population expediates the transcriptional response to the antibiotic; thus, some trends present in the amp-treated population at three hours become more apparent after six hours and/or in the queueing-amp population, which was only treated for three hours. In the case of sigma factors, we find that RpoS, RpoH and RpoE are more active in every condition compared to the untreated population [fig_ref] Figure 6: Proteolytic queueing and ampicillin treatment affect specific regulatory proteins in E [/fig_ref].
It is important to note that some TFs that were somewhat active in the amp-treated population at three hours are now significantly engaged in the queueing-amp population and after six hours of ampicillin treatment [fig_ref] Figure 6: Proteolytic queueing and ampicillin treatment affect specific regulatory proteins in E [/fig_ref]. For example, DksA was slightly more active in the amp-treated population at three hours than in the www.nature.com/scientificreports/ untreated population, but it is even more active at six hours of ampicillin treatment and in the queueing-amp population. DksA is particularly interesting because it has previously been shown to play a role in antibiotic tolerance and the stringent stress response [bib_ref] Interactions between DksA and stress-responsive alternative sigma factors control inorganic polyphosphate accumulation..., Gray [/bib_ref] , and is typically degraded by ClpXP [bib_ref] Persister heterogeneity arising from a single metabolic stress, Amato [/bib_ref] [bib_ref] Proteomic discovery of cellular substrates of the ClpXP protease reveals five classes..., Flynn [/bib_ref]. The change in gene expression as the duration of antibiotic increases suggests that the cell population is altering its expression of gene networks in response to prolonged stress. It would be interesting in the future to see the expression profile with even longer exposure to antibiotics or a different class of antibiotics.
Proteolytic degradation commonly regulates the sigma and transcription factors involved in the antibiotic stress response. As we previously touched on, many of the sigma factors and TFs involved in the antibiotic stress response are regulated by proteolysis. In the tested stressed populations (queueing and/or ampicillin), the sigma factors RpoH and RpoS are significantly more active than the untreated population based on our TRN analysis [fig_ref] Figure 6: Proteolytic queueing and ampicillin treatment affect specific regulatory proteins in E [/fig_ref]. Increased activity of these proteins is likely due to slowed proteolysis; ClpXP and FtsH rapidly degrade RpoS and RpoH , respectively [bib_ref] Structure-function studies of Escherichia coli RpoH (sigma32) by in vitro linker insertion..., Narberhaus [/bib_ref] [bib_ref] Translational regulation of the Escherichia coli stress factor RpoS: A role for..., Ranquet [/bib_ref]. In fact, there is direct evidence showing that ATP levels regulate RpoS degradation at the protease complex ClpXP, which slows when ATP levels are low [bib_ref] RpoS proteolysis is controlled directly by ATP levels in Escherichia coli, Peterson [/bib_ref]. The induction of CFP-LAA targets queue formation specifically at ClpXP, and an ΔrpoS strain of E. coli is more susceptible to antibiotics [bib_ref] Ranking of persister genes in the same Escherichia coli genetic background demonstrates..., Wu [/bib_ref]. We previously attempted to replicate the queueing-tolerance phenomenon by overexpressing RpoS, www.nature.com/scientificreports/ but high expression levels of RpoS alone did not measurably affect antibiotic tolerance [bib_ref] Proteolytic queues at ClpXP increase antibiotic tolerance, Deter [/bib_ref]. Yet, the TRN analysis of the queueing-amp population suggests that RpoS is playing a role in cell survival. This highlights the importance of looking at networks over individual genes when studying survival mechanisms to antibiotic stress. RpoS likely works in concert with RpoH and transcription factors to mediate a response, and altered proteolytic activity could act as the coordinating factor underlying this stress response. As mentioned earlier, several genes coordinated under RpoH are related to proteolysis. Particularly noteworthy are clpB and dnaK which both have increased expression during ampicillin treatment, and have previously been demonstrated using gene knockouts in E. coli as critical proteins for resuscitation after removal of the antibiotic [bib_ref] ATP-dependent dynamic protein aggregation regulates bacterial dormancy depth critical for antibiotic tolerance, Pu [/bib_ref]. Other genes that would be key to recovering from antibiotic stress are also upregulated in the amp-treated, queueing and queueingamp populations, such as genes for proteases lon and ftsH (both upregulated by RpoH). Increased expression of proteases likely prepares the cell to remove aggregated proteins and recover from the stress. However, both proteases are ATP-dependent; thus, their activity is regulated by the availability of ATP (especially important during recovery [bib_ref] ATP-dependent dynamic protein aggregation regulates bacterial dormancy depth critical for antibiotic tolerance, Pu [/bib_ref] [bib_ref] ATP-dependent persister formation in Escherichia coli, Shan [/bib_ref] [bib_ref] Persister formation in Staphylococcus aureus is associated with ATP depletion, Conlon [/bib_ref]. TRN analysis suggests a multifaceted response to ampicillin treatment. These results show that a coordinated transcriptional response to ampicillin results from the expression of many systems with redundant functions, as exemplified by the overlap between the Rob, MarA, and SoxS regulons. The changes we observed in the TRN under proteolytic and/or antibiotic stress indicate that cells are actively preparing for survival and recovery. The cells are likely transcribing mRNA during antibiotic treatment, especially when we consider that queueing only affects tolerance if the system is actively transcribed [bib_ref] Proteolytic queues at ClpXP increase antibiotic tolerance, Deter [/bib_ref]. Furthermore, RNA levels of specific transcripts increased expression between three hours to six hours of ampicillin treatment, which supports active transcription.
We hypothesize that this transcriptional response to proteolytic queueing and antibiotic stress improve shortterm survival and recovery. Especially when we consider that our previous results demonstrate that proteolytic queueing does not improve long-term survival [bib_ref] Proteolytic queues at ClpXP increase antibiotic tolerance, Deter [/bib_ref]. Other factors are therefore required to survive for extended periods of time. One such possibility is translational regulation. Many previous works have demonstrated that ribosomes are largely inactive in persister cells [bib_ref] Forming and waking dormant cells: The ppgpp ribosome dimerization persister model, Wood [/bib_ref] [bib_ref] Ribosome dependence of persister cell formation and resuscitation, Wood [/bib_ref] [bib_ref] Persister cells resuscitate using membrane sensors that activate chemotaxis, lower cAMP Levels,..., Yamasaki [/bib_ref] [bib_ref] Escherichia coli persister cells suppress translation by selectively disassembling and degrading their..., Cho [/bib_ref] resulting in slowed translation. Therefore, the transcripts we have measured may be slowly building up in the cells and are translated at low levels, if at all. The transcripts would then be readily available to the cell, and once the antibiotic is removed, ribosomes could resume protein biosynthesis. It is also possible that cells are actively translating select proteins, which seems reasonable with the number of genes upregulated during antibiotic treatment.
We also must consider that our samples are heterogeneous populations, and the data is an average of that population, which is informative but not comprehensive. Our results show that the transcriptional profile changes the more prolonged the population is exposed to antibiotics (six compared to three hours); however, it may be more complicated than that. We detected high levels of VBNCs in our populations with and without antibiotics, and our transcriptional profiles consist of a mixture of both persisters and VBNCs. Individual cells may have transcriptional profiles different from the average, and as antibiotic treatment continues, different subgroups are killed. Perhaps the surviving population after six hours of ampicillin treatment are a subpopulation present after three hours, but the expression profile is overshadowed by other subpopulations that are killed later. Understanding transcriptomic heterogeneity in the tolerant population would necessitate further experiments, such as single-cell transcriptomics, which has only recently been demonstrated in bacteria [bib_ref] Single-cell RNA-sequencing reports growth-conditionspecific global transcriptomes of individual bacteria, Imdahl [/bib_ref]. Regardless, future experiments (ribosome profiling, mass spectrometry, single-cell transcriptomics etc.) are required to understand the cellular response. Future experiments must use lethal (bactericidal) concentrations of antibiotics to draw conclusions directly related to antibiotic survival. Future directions to expand our understanding of antibiotic tolerance. We have demonstrated the feasibility of doing transcriptome analysis on cultures treated with lethal (bactericidal) concentrations of antibiotics and using TRN analysis to probe the network response. We used E. coli as our model organism because its regulatory networks are currently the most characterized of known bacteria; however, our analysis is limited to the current state of knowledge. We tested one antibiotic (ampicillin) under a single growth condition (emergence from stationary phase), and some responses we identified could be specific to these experimental conditions. Future work may be able to incorporate new information regarding the regulatory networks and contexts in which TFs and sigma factors regulate gene expression and may allow this type of analysis to be expanded to other strains of E. coli or different microorganisms. Exploring tolerance with a more minimal bacterial genome with less redundancy between networks (e.g. Rob and SoxS, etc.) may clarify how the TRN affects survival and recovery.
Our knowledge of the network response to antibiotics could also be expanded by using different antibiotics, with varying treatment times and growth conditions. Perhaps, the use of multiple antibiotics will lead to a greater understanding of the networks (or TFs) that are most important for survival and recovery. The recovery mechanisms after antibiotic treatment are of therapeutic interest because new drugs may be developed in combination with current antibiotics to prevent recovery, improving the effectiveness of antibiotic treatments. A recent paper highlights the potential to use combinations of antibiotics and staggered treatment of antibiotics where pairing strongly and weakly dependent metabolism antibiotics together effectively kill persisters [bib_ref] Eradicating bacterial persisters with combinations of strongly and weakly metabolismdependent antibiotics, Zheng [/bib_ref].
We have two competing hypotheses: (1) The longer the cells are exposed to lethal levels of antibiotics, the more they alter their transcriptome.The longer the cells are exposed to lethal levels of antibiotics, the more they change cell types (e.g. switched from persisters to VBNCs). The ratio of VBNC to persisters may be responsible for the change in gene expression we observed. This hypothesis could be tested in the future by modifying our procedure for isolating RNA from cells treated with lethal concentrations of antibiotics, with TRN analysis of the www.nature.com/scientificreports/ data, and quantification of VBNC levels over time. Expanding our knowledge of antibiotic tolerance beyond E. coli and ampicillin treatment would significantly improve our understanding of the regulatory networks across strains and species. However, as proteolysis is a highly conserved process, it is likely that the coordination of transcriptional regulation is highly conserved during stress.
# Conclusions
Our TRN analysis and top-down approach show that studying regulatory networks in antibiotic surviving subpopulations is a powerful method to understand the molecular mechanisms that enable cells to survive lethal levels of antibiotics. Gene knockouts have been useful for identifying aspects of these networks. However, our results support that antibiotic tolerance is a whole-cell response that can occur through multiple pathways. This type of systems biology approach is advantageous compared to methods focusing on individual genes and specific mechanisms, which are susceptible to redundant systems and can often lead to contrary results. For example, toxin-antitoxins systems were once thought by many to be the primary effectors of persistence; however, several studies have suggested that they are not directly related to persistence [bib_ref] Reassessing the role of Type II toxin-antitoxin systems in formation of Escherichia..., Goormaghtigh [/bib_ref] [bib_ref] Slow growth determines nonheritable antibiotic resistance in Salmonella enterica, Pontes [/bib_ref]. We instead propose that antibiotic tolerance is a multifaceted response rooted in the general stress response systems. Indeed, in our very recent work, we demonstrated that a bacteria population containing primarily essential genes have both antibiotic tolerant and persister subpopulations to different types of antibiotics. Based on ours and others' work, we propose that a systems biology approach would be fruitful for further study of antibiotic tolerance and persistence. Many of the networks we identify and discuss in this work are universal (e.g. proteolysis, oxidative stress response), which indicates that there are likely systems common to antibiotic tolerance across several species. As such, TRN analysis of transcriptomic data is a good step towards developing a comprehensive, systems-level perspective of antibiotic tolerance and recovery from antibiotic stress.
# Materials and methods
Strains and plasmids. www.nature.com/scientificreports/ from what are considered "viable but not culturable" (VBNC) cells and dying cells (note that all cells during antibiotic treatment are indeed dying cells, and cells only survive because we remove the antibiotic).
DNA-sequencing. DH5αZ1 genomic DNA was harvested and purified using Genomic DNA Purification Kit (ThermoFisher) in accordance with the manufacturer's instructions. The genome was then sequenced using paired-end Illumina sequencing Novogene Ltd. Sequencing libraries were generated from the qualified DNA samples(s) using the NEB Next Ultra DNA Library Prep Kit for Illumina (NEB, USA) following the manufacturer's protocol. Reads were then assembled using Geneious v. 11.0.3 [bib_ref] Geneious Basic: An integrated and extendable desktop software platform for the organization..., Kearse [/bib_ref] RNA-sequencing data analysis. Fastq files were aligned to NC_000913.3 and the plasmid p24KmNB82 using Geneious v. 11.0.3 [bib_ref] Geneious Basic: An integrated and extendable desktop software platform for the organization..., Kearse [/bib_ref]. Raw read counts were then analyzed using DEseq to compare expression (counts were normalized using the default median of ratios method); we used a twofold cutoff and adjusted p-values < 0.1 to identify significant changes in expression [bib_ref] Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2, Love [/bib_ref]. All conditions tested had three biological replicates. Data has been deposited to NCBI GEO at GSE156896.
Network analysis. Information on transcriptional regulation was downloaded using the EcoCyc SmartTables feature 2 . We then used custom Python scripts to generate a network graph made up of directed edges toward any gene with a significant change in expression (> twofold change in expression, adjusted p < 0.01) originating from each sigma factor/TF that may be an effector of the observed change. Visualization of this graph was done using Cytoscape 105 [fig_ref] Figure 2: Stress related sigma factors [/fig_ref]. The total number of genes mapped to a particular sigma factor/TF was compared between populations (i.e. untreated and treated each compared to stationary phase) using a two-sided Fisher's exact test COGs. Data was downloaded from http://www.ncbi.nlm.nih.gov/COG/ Graph, figures, and chart production. All images were produced in-house using Python scripts and open-source libraries, Cytoscape 105 , and/or Adobe Illustrator.
## Data availability
All data that supports the findings of this study are available from the corresponding author upon request (email: [email protected]). www.nature.com/scientificreports/
[fig] Figure 1: Cluster of Orthologous Genes (COGs). Changes in gene expression specific to ampicillin treatment are enriched in a few COG categories but are rare or nonexistent in others. Left: The number of genes (counts) that are differently expressed (adjusted p < 0.1) in the ampicillin treated population organized according to COGs (Top left: upregulated; Bottom left: downregulated). [/fig]
[fig] Figure 2: Stress related sigma factors (σ) are more active in the ampicillin treated population. Sigma factors act as global regulators that control several redundant and overlapping pathways related to antibiotic stress. Genes with increased expression over two-fold (adjusted p < 0.1) were mapped to sigma factors responsible for their transcription. (A) A graphical representation of the sigma regulatory network during antibiotic tolerance inferred from changes in gene expression unique to the ampicillin treated population compared to stationary phase (File S2). Blue lines: increased expression. Large yellow circles: sigma factors. Small grey circles: genes regulated by the sigma factors. (B) Counts of upregulated genes for the untreated and treated populations compared to stationary phase. Untreated (Amp−): cells incubated for 3 h in fresh media without ampicillin. Treated (Amp+): cells incubated for 3 h in fresh media with ampicillin. Genes with multiple possible sigma factors are accounted for each possible sigma factor. (C) Changes in gene expression for each sigma factor based on COG functional group; genes uncharacterized by REF 44 were not included. Scientific Reports | (2021) 11:6112 | https://doi.org/10.1038/s41598-021-85509-7 [/fig]
[fig] Figure 3: TFs associated with antibiotic tolerance are more active in the ampicillin treated population. TFs act as transcriptional regulators that control a several redundant and overlapping pathways related to antibiotic stress. Genes with changes in expression over two-fold (adjusted p < 0.1) were mapped to TFs responsible for their transcription.(A) A graphical representation of the selected TF regulatory network during antibiotic tolerance inferred from changes in gene expression unique to the ampicillin treated population compared to stationary phase (File S3). Red lines: decreased expression. Blue lines: increased expression. Large yellow circles: select TFs. Small grey circles: genes regulated by select TFs. (B) Counts of differentially regulated genes for the untreated and treated populations compared to stationary phase. Selected TFs displayed a significant difference (p < 0.05) in the number of genes that changed expression when comparing untreated (Amp-) and treated (Amp+) groups based on the Fisher's exact test. Amp-(untreated): cells incubated for 3 h in fresh media without ampicillin. Amp+ (treated): cells incubated for 3 h in fresh media with ampicillin. Genes with multiple possible TFs are accounted for each possible TF. (C) Changes in gene expression for the select TFs based on COG functional group; genes uncharacterized by REF 44 were not included. A high number of genes differently expressed are specific to a few COGs. Scientific Reports | (2021) 11:6112 | https://doi.org/10.1038/s41598-021-85509-7 [/fig]
[fig] Figure 4: Proteolytic queueing increases survival of ampicillin. (A) We obtained RNA-sequencing (RNA-seq) data on populations with (Q+) or without (Q-) induction of a queue at the protease ClpXP via overexpression of CFP-LAA, and populations with (Amp+) or without (Amp-) ampicillin treatment (100 μg/ml). (B) Percent survival of cells after ampicillin treatment for three hours without a queue (Amp-treated) or with a queue (Queueing-amp). Scientific Reports | (2021) 11:6112 | https://doi.org/10.1038/s41598-021-85509-7 [/fig]
[fig] Figure 5: Proteolytic queueing and ampicillin treatment similarly affect gene expression in E. coli. Changes in gene expression compared to stationary phase (adjusted p < 0.1; with > twofold change) for each population with/without queueing (Q) and/or ampicillin (Amp). Not shown are changes in gene expression present in the untreated population. Cultures treated with ampicillin were treated for three hours (3 h) or six hours (6 h). (A) Venn diagram comparing each population. Left: Upregulated. Right: Downregulated (B) Changes in gene expression sorted in Clusters of Orthologous (COG) groups for each population. Genes in multiple categories were counted for each category. Scientific Reports | (2021) 11:6112 | https://doi.org/10.1038/s41598-021-85509-7 [/fig]
[fig] Figure 6: Proteolytic queueing and ampicillin treatment affect specific regulatory proteins in E. coli. Changes in gene expression compared to stationary phase (adjusted p < 0.1; with > twofold change) for each population with/without queueing and/or ampicillin. Not shown are changes in gene expression present in the untreated population. The amp treated population was sequenced after three (3 h) and six (6 h) hours of ampicillin treatment. (A) Changes in gene expression mapped to sigma factors (σ) for each population. Genes transcribed by multiple sigma factors were counted for each sigma factor. (B) Changes in gene expression mapped to transcription factors (TFs) for each population. Genes transcribed by multiple TFs were counted for each TF. [/fig]
|
Chirped Laser Dispersion Spectroscopy for Remote Open-Path Trace-Gas Sensing
In this paper we present a prototype instrument for remote open-path detection of nitrous oxide. The sensor is based on a 4.53 μm quantum cascade laser and uses the chirped laser dispersion spectroscopy (CLaDS) technique for molecular concentration measurements. To the best of our knowledge this is the first demonstration of open-path laser-based trace-gas detection using a molecular dispersion measurement. The prototype sensor achieves a detection limit down to the single-ppbv level and exhibits excellent stability and robustness. The instrument characterization, field deployment performance, and the advantages of applying dispersion sensing to sensitive trace-gas detection in a remote open-path configuration are presented.
# Introduction
Laser technology offers unique remote non-invasive chemical sensing capabilities, and plays an important role in a broad range of applications such as industrial monitoring and emission control [bib_ref] Open-path ozone detection by quantum-cascade laser, Taslakov [/bib_ref] [bib_ref] Ultrasensitive detection of nitric oxide at 5.33 μm by using external cavity..., Lewicki [/bib_ref] exhaled breath analysis for medical diagnostics [bib_ref] Recent advances of laser-spectroscopy-based techniques for applications in breath analysis, Mccurdy [/bib_ref] [bib_ref] Real time ammonia detection in exhaled human breath using a distributed feedback..., Lewicki [/bib_ref] , or security and safety applications [bib_ref] Remote explosive and chemical agent detection using broadly tunable mid-infrared external cavity..., Rayner [/bib_ref]. One of the most dynamically developing areas of laser-based spectroscopic sensing is focused on environmental monitoring. Despite the environmental importance of various atmospheric trace-gases (CO 2 , N 2 O, NO, CH [bib_ref] Real time ammonia detection in exhaled human breath using a distributed feedback..., Lewicki [/bib_ref] , O 3 , etc.), exact knowledge about the location and magnitude of their emission OPEN ACCESS sources and sinks is still very limited. In this respect, remote sensing technologies can offer a convenient and cost-effective solution for field deployed sensors that can monitor large areas while providing good temporal and spatial resolution of precise trace-gas concentration data.
Most of the currently available remote sensors are based on optical detection that utilizes measurements of light absorption to determine chemical composition of the sample. At shorter wavelengths in the ultra-violet and visible spectral range, laser detection and ranging (LIDAR) technologies that rely on backscattering of light from particles in the atmosphere can also be used for chemical detection if spectroscopic information is being retrieved. Differential techniques such as differential absorption LIDAR (DIAL) utilize measurements at different optical frequencies for spectroscopic information retrieval, but the sensitivity, and most importantly, the specificity of those systems is moderate. Particularly important for chemical sensing is the mid-infrared (mid-IR) spectral region, where most chemicals in the gas phase exhibit their strongest ro-vibrational absorption bands. However, scattering in the atmosphere is less effective at longer wavelengths and remote sensing is usually implemented in a hard-target LIDAR configuration, often with retroreflectors used as hard targets to enhance return optical signals. A majority of remote sensing spectroscopic LIDAR/DIAL systems use a direct laser absorption spectroscopy (DLAS) technique or its derivatives based on wavelength modulation spectroscopy (WMS). In DLAS the measured signal is typically detected as a small intensity change over a large background of the total transmitted laser intensity. Due to transmission variations, which are likely to occur in remote sensing configurations, the DLAS baseline is prone to significant fluctuations that might be misinterpreted as a molecular absorption. WMS is known to suppress baseline and 1/f noise and thus can improve detection sensitivity. However, the WMS signal is still intensity-dependent and requires normalization by the received power, which especially in low-light conditions might be a source of a significant measurement error. Such a power normalization using the first harmonic of the WMS signal, or through direct power level monitoring, was successfully demonstrated in the near-infrared spectral region (near-IR) where excellent photodetectors with negligible nonlinearities are available [bib_ref] Precise isotope abundance ratio measurement of nitrous oxide using diode lasers, Uehara [/bib_ref] [bib_ref] Calibration-free wavelength-modulation spectroscopy for measurements of gas temperature and concentration in harsh..., Rieker [/bib_ref]. In the mid-infrared (mid-IR) photodetectors are usually strongly nonlinear, which makes such normalization far more challenging or impossible. Therefore there is a clear need for remote sensing technologies that could mitigate those limitations of DLAS/WMS while offering cost effective solutions for remote chemical sensing over large areas with high sensitivity and chemical selectivity.
Recently we have introduced the concept of chirped laser dispersion spectroscopy (CLaDS), a novel technique for quantitative trace gas detection based on molecular dispersion measurements. In contrast to most gas sensing methods, CLaDS does not measure absorption but instead it detects refractive index changes that occur in the vicinity of a molecular transition. Because CLaDS is essentially a phase detection technique, the CLaDS measurement is highly immune to variations of optical power at the photodetector. This makes it particularly suitable for remote open-path sensing, where the received power level may strongly fluctuate. In this paper a CLaDS-based system developed for open-path remote sensing of nitrous oxide (N 2 O) is discussed. A complete system characterization and demonstration of open-path N 2 O sensing, both in the laboratory and in the field, have been performed.
## Clads basic operation principles
When the frequency of an electromagnetic wave coincides with a resonance of the irradiated medium (e.g., rotational, vibrational or electronic) absorption and dispersion of the transmitted light wave occurs. The amount of dispersion (measured as a change in refractive index per unit frequency) is proportional to the concentration of the absorbing species, thus it can be used for quantitative concentration measurements. Attempts to use refractive index measurements for molecular detection had already been made in 1902 [bib_ref] The anomalous dispersion of sodium vapour, Wood [/bib_ref] , but only a few techniques have been implemented to-date for trace gas detection [bib_ref] Measurements of the anomalous dispersion of HF in absorption, Gross [/bib_ref] [bib_ref] Measurement of the refractive index dispersion around an absorbing line, Marchetti [/bib_ref] [bib_ref] Highly sensitive dispersion spectroscopy by probing the free spectral range of an..., Schmidt [/bib_ref] , and none of them is suitable for remote sensing. In contrast, the ClaDS technique is a promising dispersion-based method that can be used in a remote-sensing configuration and may become an attractive alternative to absorption-based techniques that rely on the Beer-Lambert law.
The theory of CLaDS has been discussed in detail in Refs. [bib_ref] Molecular dispersion spectroscopy-New capabilities in laser chemical sensing, Nikodem [/bib_ref]. Here we provide only the necessary details needed for a self-contained description. A schematic diagram of the mid-IR CLaDS setup based on a distributed feedback (DFB) quantum cascade laser (QCL) is shown in [fig_ref] Figure 1: Schematic diagram of mid-IR CLaDS setup [/fig_ref]. In general, the output radiation from a frequency-chirped single-mode laser is directed through an optical frequency shifter (e.g., acousto-optical modulator-AOM). Electrically pumped semiconductor laser sources, especially QCLs operating in the mid-IR, are particularly suitable for this application due to their flexibility and convenience in terms of laser frequency modulation and control. However, any other single-mode laser source that allows for controlled frequency chirping can be used in CLaDS. Both the fundamental and the frequency shifted wave are combined into a single dual-frequency laser beam (in case of the AOM shifter the Mach-Zehnder interferometer arrangement is used). The dual-frequency beam is sent through a gas sample and is focused onto a fast photodetector that extracts the heterodyne beat-note between both frequency components. When the dual-frequency beam interacts with a molecular transition both wavelengths experience slightly different refractive indices. As the laser is frequency-chirped, the difference in propagation velocities impacts the instantaneous frequency difference between the two optical waves, and this effect (that is proportional to the chirp rate) can be directly measured as an instantaneous frequency of the heterodyne beat-note signal. This signal enhancement due to the frequency chirp rate is central to the concept of CLaDS signal generation. The dispersion spectrum is obtained through frequency demodulation of the heterodyne beat-note and the molecular concentration can be retrieved through spectroscopic modeling based on parameters available from spectral databases or by using prior calibration of the system with reference gases.
In this basic arrangement (as shown in [fig_ref] Figure 1: Schematic diagram of mid-IR CLaDS setup [/fig_ref] by applying appropriate modulation to the laser injection current a linear chirp across the target molecular transition can be obtained and the dispersion spectrum is directly recorded using FM-demodulation of the heterodyne signal. Such a configuration and data processing is defined as direct CLaDS. The most significant advantage of the CLaDS technique over a conventional absorption based measurement is the immunity of the molecular dispersion measurement (performed in the frequency domain) to fluctuations of the received optical power. As a consequence, the large baseline that is typically observed in DLAS is not present in CLaDS. Due to exponential character of the Beer-Lambert law, the non-linear dependence of the absorption on sample concentration represents another important drawback of DLAS, especially for samples with absorption larger than 10%. In contrast, the dispersion signal increases linearly with concentration, resulting in an extended dynamic range for concentration measurements [bib_ref] High Dynamic Range Laser Dispersion Spectroscopy of Saturated Absorption Lines, Franz [/bib_ref] that also allows for longer optical distances to be used in path-integrated molecular studies. The latter can significantly benefit remote sensing applications that rely on multiple optical paths of different lengths to perform tomographic measurements of atmospheric constituents [bib_ref] Model studies of laser absorption computed tomography for remote air pollution measurement, Wolfe [/bib_ref]. Despite the fundamental advantages of direct CLaDS, previous studies have revealed some limitations in this technique. Although theory suggests that the CLaDS signal is baseline-free, in practice a weak baseline can be observed when the interferometer in the direct CLaDS setup is not perfectly balanced and the chirp rate is not truly linear (both can occur in practical implementations). Moreover, with chirp rates typically between 10 12 and 10 15 Hz/s, the acquisition bandwidth (BW) required to capture the dispersion spectrum must be very broadband, which leads to an increase in the FM-demodulation noise (proportional to BW 2 ). On the other hand, with lower chirp rates the CLaDS signal amplitude is proportionally reduced and the signal-to-noise ratio (SNR) decreases [bib_ref] Signal-to-noise ratio in chirped laser dispersion spectroscopy, Nikodem [/bib_ref]. Hence, direct CLaDS requires the chirp rate and acquisition bandwidth to be carefully optimized for the best performance. In order to overcome the identified limitations in direct CLaDS, a CM-CLaDS technique that is based on a harmonic detection scheme has been developed [bib_ref] Chirped laser dispersion spectroscopy with harmonic detection of molecular spectra, Nikodem [/bib_ref]. In the QCL-based CM-CLaDS the laser frequency is chirped with a fast sinusoidal waveform added to the injection current. The heterodyne detection followed by the FM-demodulation is similar as in direct CLaDS. However, the CM-CLaDS signal is recovered in the post-processing of the time series of the FM-demodulated beat-note frequency. The beat-note frequency time-series undergoes Fourier analysis and the amplitude of the 2nd harmonic component of the chirp modulation frequency is recorded as the CM-CLaDS signal. Narrowband analysis of the 2f signal allows reduction of the acquisition bandwidth, which leads to reduction of the FM-noise and increases the SNR. Moreover, it has been shown in [bib_ref] Chirped laser dispersion spectroscopy with harmonic detection of molecular spectra, Nikodem [/bib_ref] that in CM-CLaDS only the first harmonic is affected by the imbalance of the interferometer arms. Therefore, the 2nd harmonic signal is baseline-free which allows for convenient operation of CM-CLaDS system in a line-locked mode (laser frequency is locked to the peak of the transition) without the need for laser frequency scanning and spectroscopic data fitting for concentration retrieval.
A transition from direct CLaDS to CM-CLaDS is analogous to the transition from DLAS to WMS that also allows for continuous monitoring of the sample concentration in a line-locked mode. However, in WMS the 2f signal must be corrected for the laser power/sample transmission fluctuations [bib_ref] Digital, phasesensitive detection for in situ diode-laser spectroscopy under rapidly changing transmission..., Fernholz [/bib_ref] [bib_ref] Theoretical description based on Fourier analysis of wavelength-modulation spectrometry in terms of..., Kluczynski [/bib_ref] [bib_ref] Measurements of CO 2 concentration and temperature at high pressures using 1f-normalized..., Farooq [/bib_ref]. Such corrections might be challenging or even not possible in the presence of large changes in the received optical power levels, especially when using mid-infrared photodetectors that usually have nonlinear response. Moreover, such nonlinearities can create a baseline in the WMS spectrum that generates additional errors in 2f/1f normalization. These issues are not affecting CM-CLaDS signal due to its immunity to optical power variations [bib_ref] Chirped laser dispersion spectroscopy with harmonic detection of molecular spectra, Nikodem [/bib_ref].
## Remote sensing with cm-clads
A preliminary test of remote open-path sensing with CM-CLaDS was performed using nitrous oxide as the target molecule. N 2 O is an important greenhouse gas and atmospheric pollutant. Agricultural emissions are the primary source of a gradual increase of N 2 O concentration in the atmosphere at almost 1 ppbv/year rate [bib_ref] Increase in the atmospheric nitrous oxide concentration during last 250 years, Machida [/bib_ref] [bib_ref] Variations of atmospheric nitrous oxide concentration in the northern and western Pacific, Ishijima [/bib_ref]. Unfortunately, the global budget of N 2 O is still poorly characterized. In order to analyze changes in atmospheric N 2 O concentration, an instrument that provides precision and accuracy at a single ppbv level is required. Such performance can be achieved with mid-IR QCLs that give access to the strongest N 2 O transitions around 2,200 cm −1) [bib_ref] High precision measurements of atmospheric nitrous oxide and methane using thermoelectrically cooled..., Nelson [/bib_ref] [bib_ref] Simultaneous Open-Path Detection of Atmospheric N 2 O and CO, Tao [/bib_ref]. The N 2 O P18e transition located at 2,207.62 cm −1 exhibits very small interference from other atmospheric constituents (such as H 2 O or CO) which makes it ideal for sensing in ambient conditions. The experimental arrangement of the prototype CM-CLaDS sensor for remote N 2 O detection is shown schematically in [fig_ref] Figure 3: Optical setup of CM-CLaDS sensor for remote sensing [/fig_ref]. It was configured as a transportable system with optics placed on a 24″ × 36″ breadboard and installed with all the necessary electronics on a cart. The total power consumption of this transportable platform is ~500 W (primarily limited by an off-the-shelf RF spectrum analyzer consuming ~300 W). A thermoelectrically cooled (TEC) DFB QCL (provided by Corning Inc., Corning, NY, USA) generated up to 5 mW of optical power at 4.53 µm. By changing the laser temperature and bias current multiple N 2 O transitions between 2,207.6 and 2,211.4 cm −1 could be accessed. A CaF 2 wedge plate was used to out-couple a small fraction of the laser power for active laser frequency locking. The main laser beam was divided with 50/50 beam splitter and the two beams were directed into the AOM. Both incident angles on the AOM were set to provide two Bragg reflected beams with one of them being up-shifted and the second down-shifted in frequency by Ω = 50 MHz. Both beams were re-combined into one beam using the Mach-Zehnder arrangement. The laser power available for measurement of N 2 O at 2,207.62 cm −1 was only 250 μW due to losses on multiple optical components placed after the laser (AOM, beam splitters, several mirrors), but this was still sufficient to achieve a satisfactory SNR. This power was directed towards a distant retroreflector (diameter of 65 mm). Due to the divergence of the laser beam, a 6 inch concave mirror was used to collect the returning radiation and to focus it onto high-speed TEC MCT photodetector (PVI-3TE-10.6 from Vigo Systems S.A., Warsaw, Poland). The optical configuration of the collection optics can also be used as a standard Newtonian telescope to collect the scattered optical radiation from objects/hard-targets rather than a specular reflection from a retroreflector. The laser was frequency-chirped using a sinusoidal waveform at a modulation frequency of f 1 = 101 kHz and with an optimized optical frequency modulation depth [bib_ref] Chirped laser dispersion spectroscopy with harmonic detection of molecular spectra, Nikodem [/bib_ref]. Due to imperfections of the antireflection coatings on the AOM facets the signal contained narrowband optical fringes created by an etalon formed between the AOM facets. In order to suppress these etalon fringes an additional small-amplitude modulation (at frequency f 2 = 1 kHz) was applied to the laser current, which allowed suppression of the fringe pattern through averaging. The system could be operated in two modes: (1) a spectral measurement mode with a slow scan of the laser bias current to acquire a full 2f CM-CLaDS spectrum, or (2) a line-locked mode which enables continuous monitoring of N 2 O by CM-CLaDS. The latter was implemented using a proportional-integral (PI) controller to actively adjust the laser bias current using the 3f WMS as an error signal to keep the laser frequency at the center of the target transition. The heterodyne signal recorded by the photodetector was analyzed using an RF spectrum analyzer (RSA6106A from Tektronix, Beaverton, OR, USA). The time series of the beat-note frequency was digitally filtered in order to retrieve a 2f 1 component that is proportional to the N 2 O concentration.
## System characterization
Characterization of the CM-CLaDS instrument was conducted in the laboratory using two setups: a remote sensor arrangement with a retroreflector positioned ~17 m away, or a setup with a 10 cm long gas cell filled with 151 ppmv N 2 O in N 2 at 300 Torr pressure for calibration purposes. A significant improvement in CLaDS system performance was achieved through suppression of the parasitic optical fringes originating from the AOM. Because the length of the AOM frequency shifter is >5 cm (which for germanium corresponds to an optical length of >20 cm), the free spectral range of the observed parasitic fringes is much smaller than the line width of the target spectral feature (>0.1 cm −1 ). Hence, additional non-synchronized sinusoidal modulation was very effective at fringe suppression. The effect of the additional modulation of the laser current at frequency f 2 is clearly visible in [fig_ref] Figure 4: An example of the 2f CM-CLaDS spectrum recorded in the remote configuration... [/fig_ref] , which compares N 2 O spectra acquired with and without the fringe suppression. The CLaDS signal is strongly immune to changes in the received optical power [bib_ref] Chirped laser dispersion spectroscopy with harmonic detection of molecular spectra, Nikodem [/bib_ref]. This is especially important in open-path sensing over long distances when mechanical drifts in the system, turbulence in the atmosphere, or environmental conditions (dust, fog) can significantly affect the amount of power that is received by the photodetector. This property was discussed in detail for direct CLaDS in [bib_ref] Signal-to-noise ratio in chirped laser dispersion spectroscopy, Nikodem [/bib_ref] and has been experimentally investigated in this work for CM-CLaDS by measuring the amplitude of 2f signal while the received power was varied. The laser wavelength was locked to the peak of the transition for continuous N 2 O detection, and the beat-note (carrier) level was varied through partial obstruction of the primary mirror. The concentration measurements and the variations of the RF signal power are shown in [fig_ref] Figure 5: A continuous measurement of N 2 O concentration [/fig_ref]. In FM-demodulation there is a slight dependence of the output noise on the carrier-to-noise ratio, which affects the SNR of the system. This is also observed in our instrument. Nevertheless, it is clear that despite power changes by more than 20 dB (100 times), the mean value of the CLaDS signal was not affected. To test the stability of the CM-CLaDS system, an arrangement with a gas cell containing a calibrated gas mixture instead of a remote sensing setup was used. A gas mixture of 151 ppmv N 2 O in N 2 was flowed through the cell with a fixed pressure of 300 Torr. Due to the limited data acquisition capabilities of the spectrum analyzer, an acquisition of a continuous data stream could not be performed. Hence, the signal was recorded for only 50 ms with an interval of 1 s between the consecutive acquisitions (5% duty cycle; the remaining 950 ms in the cycle were used for data processing). This has resulted in a 20-fold reduction in the effective integration time, but at the same time it enabled long-term stability analysis with a quasi-continuous signal acquisition. Both the CM-CLaDS signal as well as the RF beat-note power were simultaneously recorded and are shown in [fig_ref] Figure 6: The Allan variance analysis [/fig_ref]. Allan variance plots were calculated for both data series. The received laser power has changed by more than 5% during the measurement (mainly due to line-locking feedback that adjusted the laser current to compensate for laser frequency drift caused by ambient temperature fluctuations). This drift is clearly visible in the Allan Variance calculated for the beat-note power in [fig_ref] Figure 6: The Allan variance analysis [/fig_ref]. However, due to the CLaDS signal immunity to power variations, those drifts did not affect stability of the CM-CLaDS system. Based on the Allan plot the CM-CLaDS signal can be effectively integrated for up to one thousand seconds and it does not show a drift that is clearly observed in the Allan plot for the RF power (power drift starts dominating at integration times longer than 20 s). The ultimate SNR of 1,000 that is obtained for the maximum integration time tested (1,000 s) yields a minimum detection limit of 15 ppbv×m. It should be noted that the system performance in its current configuration is far from optimal because the frequency split achievable with commercially available AOMs (<100 MHz) is not sufficient to match the observed transition line width (~3 GHz at atmospheric pressure). The signal loss due to non-optimal frequency splitting (100 MHz in this work) reduces the signal by ~7 times with respect to an optimized signal at 300 Torr and by ~16 times with respect to an optimized signal at atmospheric pressure. In addition to this signal loss, the 5% data acquisition duty cycle results in ~ times higher noise. Based on this stability test a bandwidth-and optical-path-normalized detection limit to N 2 O interpolated for conditions relevant for remote open-path atmospheric sensing is ~1.2 ppmv×m/Hz 1/2 (current system specifications have been used). This can be improved to below 50 ppbv×m/Hz 1/2 with an optimum frequency shifter and a 100% duty cycle data acquisition. However, despite the performance limitations, the current system can still achieve the single-ppbv N 2 O detection limits required for atmospheric studies when implemented with longer (>50 m) optical paths.
## Ambient n 2 o sensing
To demonstrate capabilities suitable for environmental N 2 O monitoring, the CM-CLaDS prototype setup was deployed in a field environmental measurement site. The system was installed at a measurement site on the University of Maryland Baltimore County (UMBC) campus, about 8 km from Baltimore, MD, USA [fig_ref] Figure 7: Picture of the environmental measurement site on the UMBC campus in Maryland [/fig_ref]. The CM-CLaDS sensor was placed inside a non-air-conditioned building and was operated almost continuously for 5 days in October 2011 with only one significant interruption on 26 October due to a power outage. The laser beam was directed through an open window towards a distant retroreflector placed outside the building, yielding a total optical path of 71 m above a neighboring meadow. The laser temperature and current were set to 25 °C and 150mA, respectively, which allowed targeting the optimal N 2 O transition at 2,207.62 cm −1 . The signal processing was further optimized and enabled acquisition with a higher duty cycle of 12.5% (250 ms of data with 2 s intervals). The CLaDS signal was calibrated using a certified gas mixture composed of 151 ± 1 ppmv of N 2 O in dry N 2 at pressure of 760 Torr, which determines the ultimate relative accuracy of concentration measurements to ±0.66% (approximately ±2.2 ppbv at the expected ambient conditions). The atmospheric N 2 O concentration time series was recorded and is shown in [fig_ref] Figure 8: Top plot [/fig_ref].
Atmospheric temperature and pressure data were measured with a close-by weather station, and both data sets have been used to correct the measured molecular concentration based on the ideal gas law and spectroscopic parameters of the transition. A correction for water vapor in the atmosphere was also applied using the data from a local hygrometer. Each point (black dot) in [fig_ref] Figure 8: Top plot [/fig_ref] is an average of data recoded within 5 min (150 points with 2 s interval and effective duty cycle of 12.5%).
The RF beat-note power was recorded simultaneously with the CM-CLaDS signal. Due to mechanical drift in the optical arrangement (temperature fluctuations etc.) during deployment, the beat-note power was slowly decreasing from −16 dBm on the first day down to slightly below −20 dBm by 30 October. There were also additional changes in the amount of received optical power caused by changing environmental conditions. The most significant occurred on 10/27 when the beat-note power dropped by more than 20 dB due to atmospheric water condensation on the retroreflector. The signal power was recovered after cleaning the retroreflector. Some other smaller power variations visible in the data were caused primarily by precipitation events.
Similar to the laboratory tests shown in [fig_ref] Figure 5: A continuous measurement of N 2 O concentration [/fig_ref] , the CM-CLaDS signal measured in the field also shows high immunity to changes in the received optical power. Continuous monitoring of As shown in , the noise dependence on the received power level observed in the field [fig_ref] Figure 8: Top plot [/fig_ref] is in good agreement with the performance tested in the laboratory (based on data from [fig_ref] Figure 5: A continuous measurement of N 2 O concentration [/fig_ref]. In both cases the noise increases by less than 4 dB per every 10 dB drop in the beat-note power, which is in good agreement with our previous studies on direct CLaDS [bib_ref] Signal-to-noise ratio in chirped laser dispersion spectroscopy, Nikodem [/bib_ref]. The data acquisition in the field was performed by averaging 150 consecutive points (instead of 50 acquired in the laboratory test) and with a higher duty cycle (12.5% vs. 5% in the laboratory), which results in an effective acquisition time of 37.5 s (vs. 10 s in the laboratory). Therefore, to directly compare both results the field characteristics should be multiplied by a factor of (37.5/10) 1/2 1.936. Indeed the experimental data presented in show the same slope for both data sets, except the field results are shifted down by a constant value that corresponds to 1.911 times reduction in noise. This is in excellent agreement with the predicted factor of 1.936 and confirms that in both tests the system was operating within the same noise limit, indicating that no other unpredictable noise sources have been introduced in the less-controlled environment in the field. . The CM-CLaDS N 2 O concentration data noise vs. carrier power level based on laboratory (red) and field (grey) measurements in [fig_ref] Figure 5: A continuous measurement of N 2 O concentration [/fig_ref] , respectively. For field data the noise is calculated as the standard deviation of 150 consecutive data points and each point corresponds to a 250 ms long data acquisition of the optical signal. This yields an effective averaging time of 37.5 s. For laboratory data from [fig_ref] Figure 5: A continuous measurement of N 2 O concentration [/fig_ref] (measured with an effective averaging time of 10 s) the noise was calculated as standard deviation of 200 consecutive data points and each point corresponds to a 50 ms long data acquisition of optical signal. Both data sets were fitted by linear functions shown in black (for field data) and in magenta (for laboratory data).
During the field deployment several significant precipitation events occurred [fig_ref] Figure 8: Top plot [/fig_ref] , and each time a subsequent increase of N 2 O concentration was measured. For the three major rainfalls an increase in concentration is indicated by blue arrows. Although careful analysis of these events is beyond the scope of this work, such correlation has already been observed by other groups [bib_ref] N 2 O exchange over managed grassland: Application of a quantum cascade..., Neftel [/bib_ref] [bib_ref] Impact of extreme precipitation and water table change on N 2 O..., Zona [/bib_ref]. N 2 O is naturally produced in soil by the processes of nitrification and denitrification, which can be triggered by rain, causing an increase of the atmospheric N 2 O concentration after precipitation. As a validation of the instrument performance on 29 October (when one of these events occurred) we have performed a cross-comparison of the continuous N 2 O concentration measurement performed by the remote CM-CLaDS sensor with another experimental N 2 O sensing system co-located at the same measurement site. The second system was an open-path point sensor based on the WMS technique developed by Zondlo et al. [bib_ref] Simultaneous Open-Path Detection of Atmospheric N 2 O and CO, Tao [/bib_ref] [bib_ref] Simultaneous detection of atmospheric nitrous oxide and carbon monoxide using a quantum..., Khan [/bib_ref]. It employed a DFB QCL operating around 2,203 cm −1 as a spectroscopic source and an open path multi-pass cell has been set-up to probe the N 2 O directly at atmospheric conditions. The N 2 O concentration data sets acquired by both instruments are shown in [fig_ref] Figure 1: Schematic diagram of mid-IR CLaDS setup [/fig_ref].
Both sensors recorded a similar average trend in N 2 O concentration, including an increase after the precipitation event. The small short-term discrepancies shown in [fig_ref] Figure 1: Schematic diagram of mid-IR CLaDS setup [/fig_ref] are primarily attributed to the different active optical paths used in both cases to interact with the atmosphere: long path-integrated CLaDS measurement vs. point sensing with WMS system. This cross-comparison result is very encouraging because the CM-CLaDS system was calibrated only once in the laboratory (one week before this test) and after four days of almost undisturbed operation an excellent agreement between the two different sensor systems has been found. This confirms the excellent long term stability and shows a promising robustness of this technology as well as its capability of unattended operation in field settings. [fig_ref] Figure 1: Schematic diagram of mid-IR CLaDS setup [/fig_ref]. N 2 O concentration data measured using the CM-CLaDS instrument (concentration data in gray, 30 s average of data points recorded in 2 s intervals) and the WMS sensor (red, 30 s average of data points recorded in 0.1 s intervals).
## Conclusions and future outlook
In this work a transportable prototype QCL-based CM-CLaDS instrument designed for open-path sensing of atmospheric N 2 O concentration was developed. The CM-CLaDS system is based on molecular dispersion spectroscopy and provides sensitive and selective remote detection of N 2 O (1.2 ppmv×m/Hz 1/2 at ambient conditions). The laboratory and field tests performed with this instrument confirmed unique properties of CLaDS technique i.e., baseline-free nature and no dependence of CLaDS signal on received optical power with only minor effect on SNR. The prototype sensor was deployed in the field setting to perform remote monitoring of atmospheric N 2 O. With only 250 μW of optical power being emitted towards retroreflector, the CM-CLaDS system enabled detection of N 2 O concentration changes at single ppbv levels. The operation of the sensor was not interrupted despite large (>20 dB) transmission changes caused by the atmospheric water condensation and several precipitation events. A cross-comparison measurement with another N 2 O sensing instrument confirmed the consistent performance of CM-CLaDS and its capability of reliable long-term open-path trace-gas detection. It is worth noting that the CM-CLaDS 2f signal amplitude is directly proportional to molecular number density and does not require any power normalization or baseline correction. This is a significant advantage over other optical techniques such as LAS or WMS, which rely on accurate normalization with respect to the received optical power and can suffer from baseline and/or signal drifts due to laser intensity variations or photodetector nonlinearities. This particular property of CM-CLaDS assures strong immunity to changes in environmental conditions and provides better long-term stability of the system. The performance of the present system can be improved by using optical frequency shifters with larger (~1 GHz) frequency shift for optimum signal enhancement and better AR coatings for fringe noise suppression. Application of a more powerful QCL laser and a more sensitive photodetector can additionally improve the CLaDS SNR by increasing the carrier-to-noise ratio at the input of the FM demodulator, thus reducing demodulation noise. A long-term deployment of the CM-CLaDS system that will focus on environmental studies of N 2 O concentration, effects of precipitation on N 2 O emission, and on seasonal trends is currently being planned.
[fig] Figure 1: Schematic diagram of mid-IR CLaDS setup (M-mirror, BS-beam splitter). [/fig]
[fig] Figure 2: (A) A HITRAN simulation showing that the N 2 O absorption band near 2,200 cm −1 is significantly stronger than absorption bands at other wavelengths (inset shows magnification of the near-IR transitions); (B) A simulated absorption spectra for four atmospheric molecules shows that the transition at 2,207.6 cm −1 is optimal for N 2 O detection with minimum interference from water vapor, carbon monoxide, and carbon dioxide. [/fig]
[fig] Figure 3: Optical setup of CM-CLaDS sensor for remote sensing (PD-photodetector). [/fig]
[fig] Figure 4: An example of the 2f CM-CLaDS spectrum recorded in the remote configuration with and without additional small-amplitude wavelength modulation at f 2 (the sample is approximately 35 m of lab air at atmospheric conditions). The narrow fringes (red) are effectively suppressed with the modulation turned on (black). [/fig]
[fig] Figure 5: A continuous measurement of N 2 O concentration (upper plot) performed in the laboratory using an open-path configuration with a retroreflector placed ~17 m away from the instrument. The carrier (beat-note) power (lower plot) is varied through changes of received optical power through partial obstruction of the telescope. [/fig]
[fig] Figure 6: The Allan variance analysis (A) was performed for: (B) a time series of CM-CLaDS signal (normalized to the first data point) acquired for 151 ppmv N 2 O in N 2 mixture in a 10 cm gas cell at 300 Torr, and (C) a received RF power series that was acquired simultaneously with data in (B). Active locking of the laser frequency to the peak of the transition at 2,207.6 cm −1 was enabled during this measurement. A dashed magenta line in (A) shows a theoretical performance of a random noise-limited system. [/fig]
[fig] Figure 7: Picture of the environmental measurement site on the UMBC campus in Maryland. Locations of the sensor and the retroreflector are labeled. [/fig]
[fig] Figure 8: Top plot: A long-term atmospheric N 2 O concentration measurement during the field deployment of the CM-CLaDS sensor (5 min. average of data points recorded in 2 s intervals). Bottom plot: The corresponding heterodyne beat-note power indicates optomechanical drift of the setup and fluctuations in atmospheric transmission due to varying environmental conditions. Center plot: A precipitation rate recorded using an on-site weather station. Each rainfall event is followed by an increase in N 2 O concentration (indicated with arrows) which most likely is a result of nitrification and denitrification processes in soil. [/fig]
|
Cancer in children born after frozen-thawed embryo transfer: A cohort study
AbstractAU : PleaseconfirmthatallheadinglevelsarerepresentedcorrectlyBackgroundThe aim was to investigate whether children born after assisted reproduction technology (ART), particularly after frozen-thawed embryo transfer (FET), are at higher risk of childhood cancer than children born after fresh embryo transfer and spontaneous conception.Methods and findingsWe performed a registry-based cohort study using data from the 4 Nordic countries: Denmark, Finland, Norway, and Sweden. The study included 7,944,248 children, out of whom 171,774 children were born after use of ART (2.2%) and 7,772,474 children were born after spontaneous conception, representing all children born between the years
[formula] a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 [/formula]
# Introduction
Recently, a substantial increase in use of frozen-thawed embryo transfers (FETs) in in vitro fertilization has occurred worldwide. In the United States of America, the FET rate has doubled since 2015 and comprised 78.8% of all embryo transfers using non-donor assisted reproductive technology (ART) in 2019. A similar pattern is observed in Australia, New Zealand, and Europe [bib_ref] 20 years of the European IVF-monitoring Consortium registry: what have we learned?..., Geyter [/bib_ref]. The main reason for the increase in FET is improved embryo survival and the high pregnancy/live birth rates after transfer of vitrified/thawed blastocysts compared to the previously used technique with transfer of slow frozen-thawed cleavage stage embryos [bib_ref] A randomized controlled study of human Day 3 embryo cryopreservation by slow..., Balaban [/bib_ref] [bib_ref] Oocyte, embryo and blastocyst cryopreservation in ART: systematic review and meta-analysis comparing..., Rienzi [/bib_ref]. A freeze-all policy (freezing of all embryos from a treatment cycle and no fresh embryo transfer) is currently being implemented in many parts of the world [bib_ref] 20 years of the European IVF-monitoring Consortium registry: what have we learned?..., Geyter [/bib_ref] , despite indications of increased birth weight and risk of hypertensive disorders in pregnancy [bib_ref] Fresh versus frozen embryo transfers in assisted reproduction, Zaat [/bib_ref] and without careful consideration of benefits and harms. Six large randomized controlled trials have investigated the differences in live birth rate following fresh embryo transfer and FET in freeze-all cycles [bib_ref] Fresh versus Frozen Embryos for Infertility in the Polycystic Ovary Syndrome, Chen [/bib_ref] [bib_ref] Transfer of Fresh versus Frozen Embryos in Ovulatory Women, Shi [/bib_ref] [bib_ref] Freeze-all versus fresh blastocyst transfer strategy during in vitro fertilisation in women..., Stormlund [/bib_ref] [bib_ref] Frozen versus fresh single blastocyst transfer in ovulatory women: a multicentre, randomised..., Wei [/bib_ref] [bib_ref] IVF Transfer of Fresh or Frozen Embryos in Women without Polycystic Ovaries, Vuong [/bib_ref] [bib_ref] Elective freezing of embryos versus fresh embryo transfer in IVF: a multicentre..., Maheshwari [/bib_ref]. The first trial, published in 2016 [bib_ref] Fresh versus Frozen Embryos for Infertility in the Polycystic Ovary Syndrome, Chen [/bib_ref] , showed a significantly higher live birth rate in freeze-all groups than fresh embryo transfer groups in anovulatory women. In ovulatory women, most trials show similar ongoing pregnancy and live birth rates in a freeze-all group (either cleavage stage embryos or blastocysts) compared with a fresh embryo transfer group [bib_ref] Transfer of Fresh versus Frozen Embryos in Ovulatory Women, Shi [/bib_ref] [bib_ref] Freeze-all versus fresh blastocyst transfer strategy during in vitro fertilisation in women..., Stormlund [/bib_ref] [bib_ref] IVF Transfer of Fresh or Frozen Embryos in Women without Polycystic Ovaries, Vuong [/bib_ref] [bib_ref] Elective freezing of embryos versus fresh embryo transfer in IVF: a multicentre..., Maheshwari [/bib_ref]. Importantly, freezing has reduced multiple pregnancies by facilitating single embryo transfer [bib_ref] Elective single-embryo transfer versus double-embryo transfer in in vitro fertilization, Thurin [/bib_ref] , and the freeze-all strategy has almost eliminated ovarian hyperstimulation syndrome [bib_ref] Fresh versus frozen embryo transfers in assisted reproduction, Zaat [/bib_ref] [bib_ref] Fresh versus elective frozen embryo transfer in IVF/ICSI cycles: a systematic review..., Roque [/bib_ref] , a potentially life-threatening complication in ART [bib_ref] Maternal death related to IVF in the Netherlands, Braat [/bib_ref]. Currently, up to 7.9% of children in Europe and 5.1% in the United States are born after ART, making health of children born after ART a topic of public health importance [bib_ref] ART in Europe, 2016: results generated from European registries by ESHRE, Wyns [/bib_ref] [bib_ref] Assisted Reproductive Technology Surveillance-United States, Sunderam [/bib_ref].
In many countries, the number of FET-conceived children has now exceeded the number born after fresh embryo transfer .
Childhood cancer includes a wide array of diagnoses, some of them very rare. Often the diagnoses are seen only in children but also cancer diseases common in adults occur. Leukemia is the most common neoplasm followed by various forms of tumors in the central nervous system (CNS). The incidence peaks during the first years of life [bib_ref] Cancer Progress and Priorities: Childhood Cancer, Lupo [/bib_ref]. The overall incidence in Northern Europe increased slightly up to the turning of the century, but later on, a stabilization has followed [bib_ref] Changing geographical patterns and trends in cancer incidence in children and adolescents..., Steliarova-Foucher [/bib_ref]. Studies on risk of childhood cancer after ART show conflicting results. Most large observational studies indicate similar overall cancer risk in children born after ART and in children in the general population [bib_ref] Cancer risk among children born after assisted conception, Williams [/bib_ref] [bib_ref] Cancer in children and young adults born after assisted reproductive technology: a..., Sundh [/bib_ref] [bib_ref] Association Between Fertility Treatment and Cancer Risk in Children, Hargreave [/bib_ref] [bib_ref] Assessment of Birth Defects and Cancer Risk in Children Conceived via In..., Luke [/bib_ref] , but a higher risk for both any cancer [bib_ref] Association of In Vitro Fertilization With Childhood Cancer in the United States, Spector [/bib_ref] [bib_ref] Cancer risk in children and young adults (offspring) born after medically assisted..., Chiavarini [/bib_ref] [bib_ref] Cancer risk among children conceived by fertility treatment, Wang [/bib_ref] [bib_ref] Possible association between in vitro fertilization technologies and offspring neoplasm, Bal [/bib_ref] and specific malignancies [bib_ref] Cancer risk among children born after assisted conception, Williams [/bib_ref] [bib_ref] Cancer in children and young adults born after assisted reproductive technology: a..., Sundh [/bib_ref] [bib_ref] Association of In Vitro Fertilization With Childhood Cancer in the United States, Spector [/bib_ref] [bib_ref] Cancer risk in children and young adults (offspring) born after medically assisted..., Chiavarini [/bib_ref] [bib_ref] Cancer risk among children conceived by fertility treatment, Wang [/bib_ref] has also been reported. In a Danish populationbased registry study [bib_ref] Association Between Fertility Treatment and Cancer Risk in Children, Hargreave [/bib_ref] , a higher risk of any childhood cancer was found after FET compared to spontaneous conception, but the finding was based on a limited number of cases.
In this large population-based registry study from 4 Nordic countries, we estimated the risk of childhood cancer in an unselected ART-conceived population, with special focus on children born after FET, and compared it to the risk in children born after fresh embryo transfer and spontaneous conception during the same period.
# Methods
## Study population and data collection
Data were obtained for Denmark, Finland, Norway, and Sweden from the CoNARTaS (Committee of Nordic ART and Safety) cohort [bib_ref] Data Resource Profile: Committee of Nordic Assisted Reproductive Technology and Safety (CoNARTaS)..., Opdahl [/bib_ref] , established to study short-and long-term health consequences of ART treatment in children and their mothers. Data on maternal and perinatal health in all deliveries were obtained from nationwide Medical Birth Registries in each country [bib_ref] The Nordic medical birth registers-a potential goldmine for clinical research, Langhoff-Roos [/bib_ref] and cross-linked with data from the national cancer registries, national patient registries, the national cause of death registries, and socioeconomic data retrieved from the population registries in each country. The unique personal identity number assigned to each resident in the Nordic countries enabled individual-level data linkage between registries and between children and their mothers.
All Nordic cancer registries are population based and nationwide. The respective cancer registries were founded in 1942 in Denmark, 1952 in Finland, 1951 in Norway, and 1958 in Sweden. Notification of cancer is mandatory in all Nordic countries. A high degree of completeness and accuracy of the registered data and comparability between countries has been documented [bib_ref] Nordic Cancer Registries-an overview of their procedures and data comparability, Pukkala [/bib_ref].
ART conception was determined from reports to the Medical Birth Registry (Finland), notifications from fertility clinics regarding all ongoing ART-conceived pregnancies in gestational weeks 6 to 7 (Norway) or the National Board of Health and Welfare (Sweden until 2007), or through linkage with cycle-based ART registries (Denmark, Sweden from 2007) (S1 . Details on the cohort and registries are given elsewhere [bib_ref] Data Resource Profile: Committee of Nordic Assisted Reproductive Technology and Safety (CoNARTaS)..., Opdahl [/bib_ref].
Inclusion criteria were all live-born singletons, twins, and higher order multiples born after ART and spontaneous conception (here defined as any conception without ART) during the study period.
## Outcome variables and follow-up
The primary outcomes were any cancer diagnosed before age 18 years after use of any ART and specifically after FET. Secondary outcomes were cancer diagnosis groups according to the International Classification of Childhood Cancer (ICCC-3). The ICCC-3 is based on the World Health Organization (WHO) classification system for cancer morphology and allows comparison of broad categories of neoplasms in continuity with previous classifications. In ICCC-3, the diagnoses are grouped into 12 main categories according to the morphology code, the topographic code, and the behavior of the tumor, i.e., benign or malignant (S2 . We grouped all patients into ICCC-3 categories. The 12 groups each include a defined set of morphology codes, and occasionally, the additional use of topography codes was used. In older patients topography codes according to the International Statistical Classification of Diseases (ICD) and Related Health problems were transferred to the latest version, ICD-10, by an algorithm used by the cancer registries. ICCC-3 only groups tumors with a malignant diagnosis except for tumors located in the CNS. Consequently, other benign or borderline tumors were not included in this report. Although there are discrepancies, due mainly to different traditions in cancer registration between the countries [bib_ref] Data Resource Profile: Committee of Nordic Assisted Reproductive Technology and Safety (CoNARTaS)..., Opdahl [/bib_ref] [bib_ref] Nordic Cancer Registries-an overview of their procedures and data comparability, Pukkala [/bib_ref] , pooling of data was possible because all use the WHO classification system.
Macrosomia was defined as birth weight �4,000 g. Birth defects and chromosomal aberrations were defined according to ICD-9 (740-759) or ICD-10 (Q00-99) code. Major birth defects were defined according to the EUROCAT classification system (S1 .
This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 STROBE Checklist). Our analyses were planned in advance of the research team accessing any data, and our study protocol is provided (S1 Text). The CoNARTaS project is also registered in the ISRCTN registry (ISRCTN11780826).
# Ethical approval
Ethical approval was obtained from Ethical Committee in Gothenburg, Sweden. In Norway, approval was given by the Regional Committee for Medical and Health (REK-Nord, 2010/ 1909). There are no requirements for ethical approval for registry-based studies in Denmark and Finland. All registry-keeping organizations gave permission to use their data in this study.
# Statistical analysis
We used Cox proportional hazards models to estimate the risk of any cancer, with age as the time scale. We computed each child's time at risk from date of birth until whichever event occurred first: diagnosis of any cancer, emigration (available through 2014 for Denmark, through 2015 for Sweden and Norway, and not available for We compared risk of cancer between children born after ART and spontaneous conception, between children born after FET and fresh embryo transfer, and between children born after FET and spontaneous conception, for any cancer and the 12 different cancer groups. In all analyses, only the first diagnosed cancer type was considered. Finland was not included in the analysis of FET since the Finnish registration does not differentiate between different assisted reproduction methods. We further analyzed risk of any cancer for singletons and multiples separately.
We estimated crude and adjusted hazard ratios (HRs) with 95% confidence intervals (CIs). The significance level was set to 5%. A number of <10 events in any group was considered too small to calculate a stable estimate. Adjustments were made for selected covariates. Selection of covariates was primarily based on medical knowledge and previous studies. We searched literature for identification of covariates [bib_ref] Risk of Cancer in Children Conceived by Assisted Reproductive Technology, Reigstad [/bib_ref]. The variables included the following child and maternal characteristics: calendar year of birth (continuous variable), country of birth (Denmark, Finland, Norway, Sweden), maternal age at delivery (continuous variable), parity (nulliparous/parous), sex, and plurality (singletons/multiples). Calendar year at birth and country of birth both influence cancer incidence as well as likeliness of having been conceived by ART. Both maternal age and birth order have been shown to be associated with cancer in offspring [bib_ref] Parental age and risk of childhood cancer: a pooled analysis, Johnson [/bib_ref] [bib_ref] Birth order and risk of childhood cancer: a pooled analysis from five..., Behren [/bib_ref] and are also associated with ART (ART mothers are older and of lower parity than spontaneous conception mothers). Risk of certain cancers is different among males and females [bib_ref] Maternal and birth anthropometric characteristics in relation to the risk of childhood..., Petridou [/bib_ref] , and some ART methods (transfer of blastocysts) may alter the sex ratio [bib_ref] Obstetric and perinatal risks in 4601 singletons and 884 twins conceived after..., Spangmose [/bib_ref] , and therefore, sex was included as a covariate. Furthermore, an association with multiple birth and cancer (leukemia) has been found [bib_ref] Birth characteristics and childhood leukemia in Switzerland: a register-based case-control study, Lupatsch [/bib_ref] and multiple birth is more common after ART, and plurality was therefore also included as a covariate.
In a sensitivity analysis, we also included maternal smoking during pregnancy (yes/no) as a covariate. In an additional sensitivity analysis, maternal highest educational level achieved during the study period (low, medium, high) was included as a covariate. This analysis included data from Denmark, Finland, and Sweden because data on education were not available from Norway.
In the main regression analysis where adjustment was performed for year of birth, country of birth, maternal age at birth, parity, sex, and plurality, the percentage of missing data was small. In the sensitivity analyses where adjustment was performed for maternal smoking or educational level, missing data for these variables were substantial. Participants with missing data were excluded from these models. No imputations were made.
Macrosomia and major birth defects have been associated with childhood cancer [bib_ref] Birth weight as a risk factor for childhood leukemia: a meta-analysis of..., Hjalgrim [/bib_ref] [bib_ref] Anthropometrics at birth and risk of a primary central nervous system tumour:..., Georgakis [/bib_ref] and are also associated with ART [bib_ref] Assessment of Birth Defects and Cancer Risk in Children Conceived via In..., Luke [/bib_ref] [bib_ref] Is frozen embryo transfer better for mothers and babies? Can cumulative meta-analysis..., Maheshwari [/bib_ref]. To investigate macrosomia and major birth defects, as possible mechanisms of an increased risk of cancer in children born after FET, separate exploratory analyses were performed with additional adjustment for macrosomia (yes/no) and major birth defects (yes/no). A similar analysis was also performed with birth weight as a continuous variable. Finally, as an indicator of embryo quality, we additionally adjusted for embryo stage, i.e., cleavage stage or blastocyst in a separate exploratory analysis comparing conception after FET and fresh embryo transfer.
Collinearity was assessed via the post-estimation command estat variance-covariance matrix of the estimators (VCE) in Stata, giving the covariances/correlations between the different covariates in the Cox proportional hazards model. No major issues with multicollinearity were identified in our analyses.
The proportional hazards assumption was tested with Schoenfeld residuals, and there were no clear violations. All analyses were performed in Stata, version 16.
## Patient and public involvement
Children or parents were not involved in the design, outcome measures, or planning of the study, and they were not asked to give advice on interpretation of results. The results of the research will be disseminated to the public through broadcasts, popular science articles, and newspapers.
# Results
## Child and maternal characteristics
The study population included 171,774 children born after use of ART and 7,772,474 children born after spontaneous conception (S1 Fig). Child and maternal characteristics are presented infor any ART method and spontaneous conception and in S3 Table for FET, fresh embryo transfer, and spontaneous conception. Overall, 25.9% and 2.6% of the children born after ART and spontaneous conception were part of a multifetal pregnancy. Preterm birth (<37 weeks) and low birth weight (<2,500 g) occurred among 16.1% and 13.0% of the children born after ART and among 5.6% and 3.5% of children born after spontaneous conception. Mean maternal age at delivery was 33.9 and 29.7 years in the ART and spontaneously conceived population, and 68.1% and 41.8% of the mothers were primiparous.
## Risk of cancer after art-conception
The total follow-up time was 1,705,772 person-years for the ART group (mean [SD] 9.9 years) and 97,027,051 person-years for the spontaneously conceived group (mean [SD] 12.5 [5.9] years). Cancer was diagnosed before age 18 years in 329 children in the ART group (incidence rate (IR) 19.3 per 100,000 person-years, and in 16,183 spontaneously conceived children (IR 16.7/100,000 person-years). The mean age at first cancer diagnosis was 6.0 years after ART and 6.8 years after spontaneous conception, and the distribution of age at first cancer diagnosis [fig_ref] Fig 1: ProportionalAU [/fig_ref] reflected the longer mean follow-up after spontaneous conception. Agespecific hazard rates were slightly higher among ART-conceived compared to spontaneously conceived children from approximately 5 to 12 years of age [fig_ref] Fig 2: Age-specific hazard rates of first cancer [/fig_ref] , corresponding to unadjusted cumulative hazards that were similar up to about 6 years of age and diverged slightly [fig_ref] Fig 3: Cumulative hazard of first cancer [/fig_ref]. After adjustment, no statistically significant difference in any cancer risk was found for children born after ART versus spontaneous conception (aHR 1.08, 95% CI 0.96 to 1.21, p = 0.18) . The 2 most common cancer types were leukemia and CNS tumors . There were 111 cases of leukemia among children born after ART (IR 6.5/100,000 person-years) and 4,921 cases after spontaneous conception (IR 5.1/100,000 person-years) (aHR 1.09, 95% CI 0.89 to 1.33, p = 0.40). The rates of any chromosomal aberration among children with leukemia were 4.5% in ART and 2.2% in the spontaneous conception group.
## . characteristics of study population by mode of conception defined as art or sc and by country of birth in children born in
CNS tumors occurred in 87 children born after ART and 4,080 after spontaneous conception (IR 5.1 and 4.2/100,000 person-years, respectively) (aHR 1.22, 95% CI 0.97 to 1.52, p = 0.09). A higher risk of epithelial tumors and melanoma was found in children born after ART (22 cases, IR 1.3/100,000 person-years) compared with children born after spontaneous conception (812 cases, IR 0.8/100,000 person-years) (aHR 1.89, 95% CI 1.20 to 2.97, p = 0.01).
## Plos medicine
Cancer in children born after frozen-thawed embryo transfer: A cohort study No significant differences were observed for other types of cancer where statistical comparisons were performed.
The IRs for any cancer and different cancer types by country of birth are presented in S4 Sensitivity analyses, including adjustment for any smoking during pregnancy or highest maternal educational level, only marginally changed the association (aHR 1.02, 95% CI 0.90 to 1.15, p = 0.75 and aHR 1.08, 95% CI 0.95 to 1.22, p = 0.27, respectively). [fig_ref] Fig 1: ProportionalAU [/fig_ref].
## Table 2. ir of any cancer before 18 years of age by mode of conception and country of birth in children born in denmark, finland, norway, or sweden (denmark
## Art spontaneous conception all
## Plos medicine
Cancer in children born after frozen-thawed embryo transfer: A cohort study No statistically significant differences for any cancer between ART-conceived and spontaneously conceived singletons (aHR 1.05, 95% CI 0.92 to 1.20, p = 0.48) or multiples (aHR 1.16, 95% CI 0.92 to 1.47, p = 0.22) were found.
## Risk of cancer after frozen-thawed embryo transfer
There were 48 cases of cancer in children born after FET (IR 30.1/100,000 person-years . Children born after FET had a higher risk of any cancer compared both to children born after fresh embryo transfer (227 cases, IR 18.8/100,000 person-years, aHR 1.59, 95% CI 1.15 to 2.20, p = 0.005) and children born after spontaneous conception (aHR 1.65, 95% CI as did adjusting for birth weight as a continuous [fig_ref] Fig 1: ProportionalAU [/fig_ref]. [fig_ref] Fig 1: ProportionalAU [/fig_ref] Adjusted for sex, plurality, year of birth, country of birth, maternal age at birth, parity, and birth weight (continuous variable).
## Plos medicine
## . ir and risk of any cancer and type of cancer according to iccc-3 categories before 18 years of age by first diagnosis and mode of conception in children born in denmark, finland, norway, or sweden
## Cancer type (iccc-3 category) a art
f Adjusted for sex, plurality, year of birth, country of birth, maternal age at birth, parity, and major birth defects.
g Adjusted for sex, plurality, year of birth, country of birth, maternal age at birth, parity, and embryo stage.
h Adjusted for sex, year of birth, country of birth, maternal age at birth, and parity.
i Adjusted for sex, year of birth, country of birth, maternal age at birth, parity, and macrosomia (�4,000 g).
j Adjusted for sex, year of birth, country of birth, maternal age at birth, and birth weight (continuous variable).
k Adjusted for sex, year of birth, country of birth, maternal age at birth, parity, and major birth defects.
l Adjusted for sex, year of birth, country of birth, maternal age at birth, parity, and embryo stage.
aHR, adjusted hazard ratio; ART, assisted reproduction technology; CI, confidence interval; FET, frozen-thawed embryo transfer; HR, hazard ratio; IR, incidence rate. https://doi.org/10.1371/journal.pmed.1004078.t004
variable instead of macrosomia. In the FET versus fresh embryo group, adjustment for embryo stage slightly strengthened the association . Risks of specific types of cancer in children born after FET versus fresh embryo transfer and versus spontaneous conception are presented in S5 A higher risk was observed for leukemia in children born after FET (23 cases, IR 14.4/100,000 person-years) versus fresh embryo transfer (75 cases, IR 6.2/100,000 person-years) (aHR 2.25, 95% CI 1.38 to 3.68, p = 0.001) and in children born after FET versus spontaneous conception (aHR 2.22, 95% CI 1.47 to 3.35, p < 0.001). Further adjustment for macrosomia or major birth defects only attenuated the association slightly (S5 . The rates of any chromosomal aberration among children with leukemia were 0% in the FET, 4.0% in the fresh, and 2.2% in the spontaneous conception group. In the FET versus fresh embryo group adjustment for embryo stage slightly strengthened the association (S5 .
The HRs for covariates included in the regression analyses are illustrated in Figs 4 and 5 and S2.
# Discussion
The main finding in this large cohort study, based on nationwide registries and including 171,774 children born after use of any ART, was that while no increase in any childhood cancer was found after any ART, a higher risk was observed in children born after FET. The estimates were robust and changed only marginally after adjustment for relevant confounders. For specific cancer types, a significantly higher risk was found for epithelial tumors and melanomas in children born after ART versus spontaneous conception and for leukemia in children born after FET versus fresh embryo transfer and spontaneous conception. Further adjustments for either macrosomia, continuous birth weight, or birth defects only marginally attenuated these associations while adjusting for embryo morphology slightly strengthened the association. Associations for FET were weaker for singletons than for multiples.
The reason for a possible higher risk of cancer in children born after FET is not known. Each childhood cancer type has its own risk factor profile, but many childhood cancers are thought to derive from embryonic accidents and originate in utero [bib_ref] Cancer Progress and Priorities: Childhood Cancer, Lupo [/bib_ref]. High birth weight has been associated with higher childhood cancer risk, and epigenetic alterations have been proposed as a possible explanation [bib_ref] Birth weight as a risk factor for childhood leukemia: a meta-analysis of..., Hjalgrim [/bib_ref] [bib_ref] Anthropometrics at birth and risk of a primary central nervous system tumour:..., Georgakis [/bib_ref]. Recent studies suggest changes in the epigenetic control in newborns after use of different ARTs [bib_ref] Integrated multi-omics reveal epigenomic disturbance of assisted reproductive technologies in human offspring, Chen [/bib_ref] [bib_ref] Do frozen embryo transfers modify the epigenetic control of imprinted genes and..., Barberet [/bib_ref]. A population-based US study found that among children with birth defects, particularly birth defects of chromosomal origin, those conceived via ART were at greater risk of developing cancer compared with spontaneously conceived children [bib_ref] Assessment of Birth Defects and Cancer Risk in Children Conceived via In..., Luke [/bib_ref]. Although in our study, a major birth defect was an independent predictor of cancer in the analysis of children born after FET versus spontaneous conception, the association changed only marginally after adjustment for major birth defects, as did analyses with adjustment for macrosomia or birth weight. However, these analyses should be interpreted with caution due to the possibility of confounding from factors influencing both birth weight, major birth defects, and cancer risk [bib_ref] Conditioning on intermediates in perinatal epidemiology, Vanderweele [/bib_ref] [bib_ref] Confounding, causality, and confusion: the role of intermediate variables in interpreting observational..., Ananth [/bib_ref]. Thus, further adjustment, separating chromosomal and non-chromosomal aberrations was not performed.
Higher risks of preterm birth, low birth weight, and birth defects in singletons after ART have been repeatedly found both in large cohort and registry-based studies and in systematic reviews and meta-analyses [bib_ref] Do the children born after assisted reproductive technology have an increased risk..., Zhao [/bib_ref] [bib_ref] Worldwide prevalence of adverse pregnancy outcomes among singleton pregnancies after in vitro..., Qin [/bib_ref]. For children born after FET compared to children born after fresh embryo transfer, a lower risk of preterm birth and low birth weight, but a higher risk of macrosomia is apparent [bib_ref] Is frozen embryo transfer better for mothers and babies? Can cumulative meta-analysis..., Maheshwari [/bib_ref]. Studies on long-term outcomes in ART-conceived children are more limited. Divergent results have been published concerning childhood cancer after ART. Most large observational studies show similar cancer risk for children born after ART compared to the general population [bib_ref] Cancer risk among children born after assisted conception, Williams [/bib_ref] [bib_ref] Cancer in children and young adults born after assisted reproductive technology: a..., Sundh [/bib_ref] [bib_ref] Association Between Fertility Treatment and Cancer Risk in Children, Hargreave [/bib_ref] [bib_ref] Assessment of Birth Defects and Cancer Risk in Children Conceived via In..., Luke [/bib_ref]. In a large cohort study in the United Kingdom, including 106,013 ART children [bib_ref] Cancer risk among children born after assisted conception, Williams [/bib_ref] , 108 children with cancer were identified, compared to 109.7 expected cancers (standardized IR 0.98, 95% CI 0.81 to . Higher risks were detected for certain malignancies such as hepatoblastoma and rhabdomyosarcoma. For children born after FET, the risk was similar to that in children born after fresh embryo transfer. Also, 2 earlier Nordic studies including 91,796 and 25,782 ART children, respectively, partly overlapping the present study, did not indicate any higher cancer risk in ART children (aHR 1.08, 95% CI 0.91 to 1.27 and aHR 1.21, 95% CI 0.90 to 1.63, respectively) [bib_ref] Cancer in children and young adults born after assisted reproductive technology: a..., Sundh [/bib_ref] [bib_ref] Risk of Cancer in Children Conceived by Assisted Reproductive Technology, Reigstad [/bib_ref]. In a large observational study in the USA, a slightly higher risk of cancer among children born after ART was observed (HR 1.17, 95% CI 1.00 to 1.36) [bib_ref] Association of In Vitro Fertilization With Childhood Cancer in the United States, Spector [/bib_ref]. The study identified 321 children with cancer in an ART population of 275,686 children, but no difference in risk was found for children born after FET. In contrast, a Danish population-based registry study found higher risk of any childhood cancer after FET than after spontaneous conception, but the result was based on only 14 cases (HR 2.43, 95% CI 1.44 to 4.11) [bib_ref] Association Between Fertility Treatment and Cancer Risk in Children, Hargreave [/bib_ref]. Studies on childhood cancer after ART were recently summarized in a systematic review [bib_ref] The association between fertility treatments and the incidence of paediatric cancer: A..., Zhang [/bib_ref] , concluding that FET may be related to a higher risk of pediatric cancer. In an even more recent study from Israel, with a limited number of children, a higher risk of cancer was found in children born after fresh transfer [bib_ref] Possible association between in vitro fertilization technologies and offspring neoplasm, Bal [/bib_ref]. The conflicting results may partly be due to limited study sizes with few events, differences in cancer registration, and various completeness of registries.
The main strengths of the present study are the large sample size, including unselected ART and spontaneously conceived populations born during a period of up to 3 decades in 4 Nordic countries and the use of high-quality validated population-based registries [bib_ref] Nordic Health Registry-Based Research: A Review of Health Care Systems and Key..., Laugesen [/bib_ref]. Individual data linkage between population-based registries made adjustments for potential confounders possible.
The main limitation is the number of children with cancer in the FET group. Although including a large cohort, this study cannot give a definite answer if FET is associated with an increased risk of cancer in childhood. It was not possible to include Finland in the FET analysis due to missing information on ART method. Further, there was also lack on information on emigration from Finland. Adjustment for race/ethnicity was not possible since registration on race/ethnicity is not allowed in the Nordic countries. It has been reported that non-white children and young adults might have lower rates of some childhood cancers [bib_ref] Racial and ethnic disparities in pediatric cancer incidence among children and young..., Marcotte [/bib_ref]. The percentage of mother's country of birth being outside the Nordic countries was however low and similar in ART and spontaneous conception in an earlier publication from CoNARTaS [bib_ref] School performance in singletons born after assisted reproductive technology, Norrman [/bib_ref].
Furthermore, all data are observational, and residual confounding by factors such as genetics, parental preconception health, and lifestyle cannot be excluded.
We were not able to exclude other medically assisted reproduction methods such as intrauterine insemination or ovulation induction from spontaneously conceived children. Although today such cycles, at least in Denmark are substantial, they only accounted for a small proportion of the spontaneous conception cohort. This misclassification might have attenuated the associations.
In the present study, only patients performing ART and delivering in their home countries are included. Although fertility tourism, meaning that patients go abroad or coming from abroad for fertility treatment today is rather common in some Nordic countries, this was uncommon during the study period. Such cycles are further impossible to correctly identify.
It might be argued that selecting the best quality embryos for fresh embryo transfer while cryopreserving less good quality could represent 2 morphologically different populations of embryos with different risks of any adverse outcome. Although numerous studies have found an association between embryo quality and pregnancy and live birth rates, there are at present no indication of more adverse outcome in children born from poor quality embryos [bib_ref] Congenital malformations, chromosomal abnormalities and perinatal results in IVF/ICSI newborns resulting from..., Mendoza [/bib_ref]. In addition, more FET pregnancies were conceived after blastocyst transfers which were considered having higher quality than cleavage stage embryos and when adjusting for embryo stage as an indicator of embryo quality, the association between FET and cancer slightly strengthened. Furthermore, a vast majority of FET cycles were performed after a failed fresh cycle from the same oocyte retrieval and cryopreservation of surplus embryos while the freeze-all concept was hardly used. In fact, the FET population probably represents more good prognosis patients since it was possible to freeze embryos in the same cycle in addition to the fresh embryo transfer. Even though the present study included both children born after slow freezing of cleavage stage embryos and the more recent technique with vitrification of blastocysts, which could be associated with different risks, a recent large study comparing these techniques has not shown any major differences in perinatal outcome between these groups [bib_ref] Perinatal and maternal outcome after vitrification of blastocysts: a Nordic study in..., Ginströ M Ernstad [/bib_ref].
It is not clear if the results of this study can be broadly generalizable; however, the study population represents an unselected ART as well as spontaneously conceived cohort from 4 Nordic countries covering a long time period.
## Conclusions and further implications
In conclusion, while risk of any cancer was not higher in children born after use of ART, we found that children born after FET had a higher risk of childhood cancer than children born after fresh embryo transfer and spontaneous conception. The results should be interpreted cautiously based on the limited number of children with cancer. Although the absolute risk is low, these findings are important considering the increasing use of the freeze-all strategy. Future research should elucidate these results and the mechanisms behind.
## Supporting information
[fig] Fig 1: ProportionalAU : AbbreviationlistshavebeencompiledforthoseusedinFigs1 À 5andTables2 À 4:Pleaseverifythatallentriesar distribution of age at first cancer (any type) among spontaneously and ART-conceived children born in Denmark (1994-2014), Finland (1990-2014), Norway (1984-2015), and Sweden (1985-2015) and diagnosed with cancer before age 18 years. ART, assisted reproduction technology. https://doi.org/10.1371/journal.pmed.1004078.g001 [/fig]
[fig] Fig 2: Age-specific hazard rates of first cancer (any type) among spontaneously and ART-conceived children born in Denmark (1994-2014), Finland (1990-2014), Norway (1984-2015), and Sweden (1985-2015) and diagnosed with any cancer before age 18 years. ART, assisted reproduction technology; CI, confidence interval. https://doi.org/10.1371/journal.pmed.1004078.g002 [/fig]
[fig] Fig 3: Cumulative hazard of first cancer (any type) up to 18 years for spontaneously and ART-conceived children born in Denmark (1994-2014), Finland (1990-2014), Norway (1984-2015), and Sweden (1985-2015). Crude hazard ratio 1.13; 95% CI 1.01 to 1.26, p = 0.03. ART, assisted reproduction technology; CI, confidence interval. https://doi.org/10.1371/journal.pmed.1004078.g003 [/fig]
[fig] Fig 4: HRs with 95% CI for independent covariates including macrosomia for risk of cancer in children born after FET versus fresh embryo transfer. CI, confidence interval; FET, frozen-thawed embryo transfer; HR, hazard ratio. https://doi.org/10.1371/journal.pmed.1004078.g004 [/fig]
[fig] Fig 5: HRs with 95% CI for independent covariates including macrosomia for risk of cancer in children born after FET versus spontaneous conception. CI, confidence interval; FET, frozen-thawed embryo transfer; HR, hazard ratio. https://doi.org/10.1371/journal.pmed.1004078.g005 [/fig]
|
Intermittent subacute obstruction of small bowel by Giant Meckel's Diverticulum- a case report
a b s t r a c t Meckel's diverticulum is the most common congenital anomaly of gastrointestinal tract, which results from incomplete involution of the proximal part of the vitelline duct during weeks 5-7 of fetal development. Giant Meckel's diverticulum more than 5 cm in length is relatively rare and may be associated with small bowel obstruction. Here we report Giant Meckel's diverticulum in 11 year old male who presented with the symptoms of acute abdomen with intermittent subacute small bowel obstruction. Patient presented with muscle guarding and tenderness. CT abdomen revealed a large outpouching round lesion in distal ileum intermittently compressing the small bowel that was suggestive of Giant Meckel's diverticulum. The morphology, pathology and complications of Giant Meckel's diverticulum is important for radiologist and surgeons for successful management of patient.
# Introduction
Meckel's diverticulum is the most common congenital anomaly of gastrointestinal tract with an average length of 3 cm . This prevailing abnormality usually varies in size from 1 to 10 cm [bib_ref] T-shaped Meckel's diverticulum: a rare anatomical variant complicating small bowel volvulus, Paul [/bib_ref]. Cases of Giant Meckel's diverticulum more than 5 cm in length are rare and linked with serious complications, mostly obstruction . Meckel's diverticulum com-✩ Postal Address-A-1, Nirmal Prabha Residency, Vishalgadkar House Compound, Nagala Park, Kolhapur, 416003 ✩✩ Consent for Publication-Written informed consent was obtained from the patient for publication of this case report and any accompanying images. ★ Conflict of Interest-The authors have disclosed no potential conflicts of interest, financial or otherwise. monly exhibits as gastrointestinal bleeding in pediatric population and as intestinal obstruction in adults [bib_ref] Giant Meckel's diverticulum compressing root of mesentry-A rare cause of ileal gangrene-Case..., Mohammed [/bib_ref]. Although Meckel's diverticulum is most commonly diagnosed congenital gastrointestinal tract anomaly, it affects only 2% of population worldwide [bib_ref] Giant Meckel's diverticulum: An exceptional cause of intestinal obstruction, Sami [/bib_ref]. Surgical treatment of Meckel's diverticulum by both open and laparoscopic procedures has led to the clinical "rule of 2" for symptomatic cases, by which the anatomical deformity is most often located 2 feet from the ileocecal junction and is 2 inches long [bib_ref] Meckel's diverticulum: a systematic review, Sagar [/bib_ref]. Distal small bowel diverticulosis is an atypical finding, and most are in the duodenum and jejunum. Large ileal diverticula are intensely rare [bib_ref] Khor Jen lock , Chew Min Hoe . An unusual case of..., Koh Fang [/bib_ref]. Both the pediatric and adult patients with complex Meckel's diverticulum may develop small bowel obstruction and present with colic abdominal pain, vomiting and distension. The process of obstruction can be intussusception of an inverted Meckel's diverticulum, volvulus or strangulation of distal ileum by the fibrous band connecting the diverticulum keeping retained foreign objects, enteroliths or a tumor [bib_ref] Small bowel obstruction caused by Giant Meckel's Diverticulum, Osman [/bib_ref].
Here we report an unusual case of Giant Meckel's diverticulum presenting with intermittent subacute obstruction in an 11 year old male.
## Case report
An 11 year old male patient was referred for CT imaging for an acute abdomen. He complained of recurrent episode of pain in abdomen, 8-10 episodes of vomiting and mild distension of the abdomen. Most of the episodes responded to conservative management. However, this time the symptoms were more acute and associated with fever.
On examination, there was mild distension of abdomen and tenderness on palpation with muscular guarding. Bowel sounds were normal, and no signs of peritonitis were present. The patient was febrile, laboratory testing showed mild elevated count.
CT scan of abdomen was performed on 16-slice MDCT Siemens Scanner. 100 mL nonionic iodine contrast was given orally and then non-contrast and contrast enhanced images were obtained. Additional images were obtained after positive oral contrast administration and delayed images were obtained after 90 minutes.
A CT scan of abdomen and pelvis revealed a round lesion that was outpouching through small intestinal wall in distal ileum, which was suggestive of diverticulum. The lesion was placed in pelvis with well-defined enhancing thin wall and air fluid levels within. Mild ascites was noted [fig_ref] Figure 1 -: Post contrast axial image showing oval lesion in pelvis denoted by red... [/fig_ref].
Post-oral contrast image showed opacification of the lesion with positive contrast confirming connectivity with the bowel loops with a wide neck of communication [fig_ref] Figure 3 -: Sagittal reconstructed image after oral positive contrast administration showing large diverticulum with... [/fig_ref]. These findings were suggestive of Giant Meckel's diverticulum A diagnosis of Giant Meckel's diverticulum with intermittent subacute small bowel obstruction was confirmed.
Emergency laparotomy was performed under general anesthesia, which revealed oedema throughout the entire small bowel. There were few dilated bowel segments present.
A Giant ileal diverticulum about 7.2 cm in length and 5.2 cm of circumference was present on the antimesenteric border of small bowel at 70 cm proximal to the ileocecal valve. The diverticulum was not adherent to the parietal peritoneum of the abdominal wall at the pelvis but was adherent to few segments of small bowel. The diverticulum was compressing on few loops of small bowel due to its giant size. The compression was released, the diverticulum was freed from the adhe-
## Fig. 4 -coronal reformat image of oral contrast phase showing outline of lesion (denoted by red arrows) and its connection with ileal loops in pelvis (green arrow demonstrates ileal loop). also, note the wide neck of communication (denoted by black arrows).
## Fig. 5 -showing the post-operative image of giant meckel's diverticulum with a small segment of ileum.
sions of small bowel and the Giant Meckel's diverticulum was resected along with a small segment of ileum .
The remaining intestinal segment was restored. The patient was discharged on postoperative day 8 without any complications. Histopathological examination of the specimen was reported as Giant Meckel's diverticulum with hyperplasia of lymphoid tissue.
The patient was followed up one week after discharge and had no complaints.
## Embryological aspect
After 7th week of intrauterine life, the vitello-intestinal duct and the secondary yolk sac (omphalomesenteric duct), undergo regression and complete reabsorption. Both these structures lie in the extraembryonic coelom, surrounded by embryonic mesenchyme.
Ultimately, the secondary yolk sac is compressed onto the chorionic surface of the placental plate by expanding the amnion and its fusion with the chorion. At the distal end of the vitelline duct, the secondary yolk sac progressively shrinks in size and is seen as a discrete pale yellow discoid tissue mass on the amniotic surface of the placental plate, near the insertion site of the umbilical cord [bib_ref] T-shaped Meckel's diverticulum: a rare anatomical variant complicating small bowel volvulus, Paul [/bib_ref].
# Discussion
Meckel's diverticulum is a true diverticulum, which was first described by Johann Friedrich Meckel in 1809. It is typically situated up to 100 cm away from the ileocecal valve and occurs in 2% of general population with an average size of 3 cm. Giant Meckel's diverticula are defined as diverticula measuring more than 5 cm [bib_ref] Chris Braumann; small bowel obstruction due to a Giant Meckel's Diverticulum, Luu [/bib_ref]. In our case, there was Giant Meckel's diverticulum as it was measuring about 7.2 cm in length and 5.2 cm in circumference.
Males are prone to suffer from complications, which occur twice as common when compared with females. 50% of these complications are observed in people under the age of ten years and it seems as the frequency declines with age [bib_ref] T-shaped Meckel's diverticulum: a rare anatomical variant complicating small bowel volvulus, Paul [/bib_ref]. In our case patient was male of 11 years.
Long Meckel's diverticulum has been affiliated with enteroliths and are susceptible for obstructions [bib_ref] Giant Meckel's diverticulum compressing root of mesentry-A rare cause of ileal gangrene-Case..., Mohammed [/bib_ref]. In our case patient had intermittent subacute obstruction of small bowel.
In adult case of Meckel's diverticulum obstruction is the most often reporting presenting symptom. Though first hypothesized in 1902 these possible reasons for Meckel's diverticulum caused intestinal obstruction to remain the hallmark by which Meckel's diverticulum cases are classified. The obstruction allied with a free or unattached diverticulum or having only one attachment to the intestine represent first Meckel's diverticulum type and obstructions allied with an attached diverticulum, including through its terminal ligament, to the abdominal wall or intestinal viscus represents the second type. Between these two types, former is much rarer [bib_ref] Giant Meckel's diverticulum: An exceptional cause of intestinal obstruction, Sami [/bib_ref]. In our case, it was Giant Meckel's diverticulum and can be categorised into first type, as it was unattached diverticulum to the abdominal wall.
Congenital diverticula are true diverticulum that consists of all layers of the intestinal wall, including the muscular layer. This contrasts with false diverticula, which occur when only the mucous membrane and submucosa protrude through a weak point in the intestinal wall .
Complicated Meckel's diverticulum may be clinically inappreciable from variety of other intraabdominal diseases such as acute appendicitis, inflammatory bowel disease or other causes of small bowel obstruction. Process of the obstruction comprises, enlargement of the small bowel around a fibrous band attached to the umbilicus, entrapment of an intestinal loop, intussusception, volvulus or stenosis [bib_ref] Small bowel obstruction caused by Giant Meckel's Diverticulum, Osman [/bib_ref].
In our case, the Giant Meckel's diverticulum presented with wide neck and was causing intermittent subacute small bowel obstruction. Other complications such as ischaemia of intestinal wall, volvulus, stenosis or intussception was not present. Hence, the patient underwent laparotomy and diverticulectomy was done.
# Conclusion
Our case exemplifies a rare case of Giant Meckel's diverticulum leading to intermittent subacute small bowel obstruction. Comprehension of morphology, pathology and symptoms caused by Giant Meckel's diverticulum is of importance to surgeons, gastroenterologist, pathologist and radiologist. Our case describes novel insights of Giant Meckel's diverticulum and its successful management by surgical resection.
## Patient consent
Patient have no objection to any of the above and give permission for the same.
[fig] *: Corresponding author: (N. Vasudha) E-mail address: [email protected] (N. Vasudha). [/fig]
[fig] Figure 1 -: Post contrast axial image showing oval lesion in pelvis denoted by red arrows. [/fig]
[fig] Figure 2 -: Post contrast coronal reformat showing the oval lesion in pelvis with well-defined enhancing thin wall (denoted by red arrows). The lesion shows air fluid levels within. Mild ascites is denoted by green arrow. [/fig]
[fig] Figure 3 -: Sagittal reconstructed image after oral positive contrast administration showing large diverticulum with communication with bowel loop. Note air fluid level in anterior portion of diverticulum. measuring about 7.2 cm in length with a circumference of 5.3 cm. Due to its giant size and mass effect; it was probably producing intermittent obstruction of small bowel as well as a source of infection. [/fig]
|
A successful case of anti-NMDAR encephalitis without tumor treated with a prolonged regimen of plasmapheresis
Anti-NMDA receptor encephalitis is a severe but treatable autoimmune disease of the CNS. However, the use of immunotherapy and long-term outcomes have yet to be defined for this disease. We describe a case of an 18-year-old male diagnosed with anti-NMDAR encephalitis not associated with tumor, which did not respond to initial treatment with immunoglobulin, followed by corticosteroids, cyclophosphamide and evolved with significant clinical improvement after a prolonged course of plasmapheresis. Although it is not possible to affirm the good outcome was due solely to the prolonged plasmapheresis regimen, recently published data shows that improvement may take weeks or months to occur. This case discloses another therapeutic possibility for patients with refractory disease who fail to respond to recommended first-line and second-line therapy.
## Resumo.
A encefalite antirreceptor de NMDA é uma grave doença autoimune do sistema nervoso central, mas responsiva a tratamento. No entanto o uso da imunoterapia e o prognóstico a longo prazo ainda não foram estabelecidos para esta doença. Aqui nós descrevemos o caso de um homem de 18 anos de idade diagnosticado com encefalite antirreceptor de NMDA, não associada a tumor, o qual provavelmente não respondeu à terapia inicial com corticosteroides, ciclofosfamida e, o qual evoluiu com melhora clínica importante após um curso prolongado de plasmaférese. Embora não seja possível atribuir o desfecho favorável apenas ao regime de plasmaférese prolongado, uma vez que publicações recentes demonstram que a resposta à imunoterapia pode levar semanas à meses, nosso caso abre uma possibilidade terapêutica nova para casos com doença refratária, não responsivos ao tratamento de primeira e segunda-linha preconizados Palavras-chave: doença autoimune, encefalite, receptor anti-NMDA, plasmaferese, sintomas psiquiátricos.
## Case presentation
A n 18-year-old male presented with numbness to the left side of his face and insomnia. In the days that followed, behavioral and psychiatric symptoms consisting of agitation, irritability and delusional thoughts manifested. He was initially assessed by a neurologist, who prescribed clonazepam and sertraline. Because psychomotor agitation worsened, treatment was changed to valproate, risperidone and promethazine. After eleven days, focal motor seizures developed. On day 15, he presented with a decreased level of consciousness, evolved with urinary incontinence and hyperthermia. He was admitted to another hospital on the 17 th day. In the initial investigation workup, cerebrospinal fluid (CSF) revealed mild lymphocytic pleocytosis (25 mm 3 ), normal protein concentration (39 mg per dL), normal glucose concentration (59 mg per dL), immunologic reactions for syphilis and toxoplasmosis were negative, polymerase chain reaction for Varicella Zoster, Herpes simplex virus 1 and 2 tested negative, and IgG was slightly elevated (12.8%). Brain magnetic resonance imaging (MRI) showed unremarkable findings. Electroencephalogram (EEG) showed disorganization of background activity and slow wave paroxysms. Acyclovir was then initiated. Hyperthermia persisted, so malignant neuroleptic syndrome was then suspected and he was given dantrolene. On day 19 he was submitted to endotracheal intubation for barbiturate-induced coma to treat refractory epileptic status. On day 25, he was transferred to our hospital. The patient was empirically started on doxycycline, sulphamethoxazoletrimethoprim, ampicillin and dexamethasone. Anticonvulsants were adjusted. Investigation workup was broadened to include autoimmune and paraneoplastic causes of encephalitis. Blood samples for anti-NMDA receptor were collected and sent for analysis (CSF samples were not sent). Chest, abdominal and pelvic computed tomography (CT) scans showed an enlarged axillary lymph node. Biopsy was performed but pathologic findings suggested reactive hyperplasia. The paraneoplastic panel was negative (only serum sample was analysed) for anti-NMDA; Anti-Neuronal Nuclear Antibody 1 (ANNA1), ANNA-2, ANNA-3; Anti-Glial/Neuronal Nuclear Ab Type 1 (AGNA-1); Purkinje cell cytoplasmic antibody type 1 (PCA-1), PCA-2, PCA-Tr; Amphysine antibodies; Collapsin response mediator protein 5 (CRMP-5); anti-voltage-gated calcium channel (VGCC); antivoltage-gated potassium channel (anti-VGKC); antimyelin-associated glycoprotein (anti-MAG); Ganglionic acetylcholine receptor autoantibody and acetylcholine autoantibodies. Borrelia Burgdorferi enzyme-linked immunosorbent assay (ELISA) tested positive, but Western blot test was negative.
Because clinical features were highly suggestive of anti-NMDA receptor encephalitis, a diagnostic test was repeated, this time including CSF samples. Results disclosed antibodies against anti-NMDA receptor detected only in the CSF.
Intravenous and continuous midazolam, propofol, and ketamine were started to wean patient from barbituric burst-suppression pattern on EEG. On day 29, intravenous immunoglobulin (0.4 mg/Kg/day for five days) was started empirically. In the ensuing days, there was a slight improvement in the neurological conditions; the patient was submitted to tracheostomy. Antiepileptic drugs were changed, midazolam, propofol and ketamine were weaned, but the patient was still comatose and although non-convulsive status epilepticus was resolved, he still had sporadic seizures. On day 36, the patient was still comatose and presenting autonomic instability, when methylprednisolone was initiated. On day 39, he received cyclophosphamide. The patient began to present spontaneous eye opening and followed the examiner, although he did not obey verbal commands and showed no response to painful stimuli. On day 42, he began obeying verbal commands, showed a slight improvement in neurological status and no seizures, but oro-lingual-facial dyskinesias were still present. On day 45, his level of consciousness deteriorated again. Hence, the patient was submitted to the first plasmapheresis session. After the third plasmapheresis session, on day 48, the patient was able to talk (assisted by a phonation valve) and exhibited a significant improvement in level of consciousness. A regular plasmapheresis regimen (at least two times a week) was maintained and the patient gradually recovered: antiepileptic drugs doses were reduced, hyperthermia and tachycardia resolved and decannulation was achieved.
After marked recovery, he was discharged, twelve days after the 19 th plasmapheresis session. At discharge, his mental examination was normal and neurologic examination showed bilateral foot drop and absent ankle reflex, where these findings were considered to be due to critical care neuropathy.
# Discussion
In 2005, a syndrome affecting young women with ovarian teratoma characterized by prominent psychiatric symptoms, memory dysfunction, decreased level of consciousness and central hypoventilation was described. [bib_ref] Paraneoplastic encephalitis, psychiatric symptoms, andhypoventilation in ovarian teratoma, Vitaliani [/bib_ref] Shortly after, the discovery of a specific autoantibody to N-methyl D-aspartate receptor (NMDAR) in serum and CSF of these 4 patients, and of another 8 patients with similar neurologic symptoms (7 with associated ovarian teratoma), changed the paradigm in CNS autoimmunity 2 neuropathological findings, tumors, and serum/ cerebrospinal fluid antibodies using rat tissue, neuronal cultures, and HEK293 cells expressing subunits of the N-methyl-D-aspartate receptor (NMDAR).
It is believed to be the second-most-common cause of autoimmune encephalitis after acute demyelinating encephalomyelitis (ADEM). [bib_ref] Causes of encephalitis and differences in their clinical presentations in England: a..., Granerod [/bib_ref] Care must be taken in performing diagnostic tests and antibody studies should be done in both serum and CSF because NMDAR antibodies might go undetected in serum after immunotherapy initiation or delayed diagnosis 4 the encephalitis associated with antibodies against the N-methyl-D-aspartate receptor (NMDAR). In our case antibodies were detected only in CSF.
Although severe, this form of autoimmune encephalitis is potentially responsive to treatment, which centers on two principles: immunotherapy (first and second line) and tumor removal, if applicable. A recent observational worldwide cohort study, in which 577 patients were enrolled, demonstrated that the occurrence of an underlying tumor was greater in female patients aged 12 years or older than in young children and male patients, and if present is almost always an ovarian teratoma containing nervous tissue and expressing NM-DAR (94% of all tumors were ovarian teratomas). [bib_ref] Clinical experience and laboratory investigations in patients with anti-NMDAR encephalitis, Dalmau [/bib_ref] Firstline therapy consists of methylprednisolone plus IVIg or plasmapheresis while second-line therapy is rituximab, cyclophosphamide or both.
In the study, 472 patients were treated with first-line immunotherapy (steroids, intravenous immunoglobulin, plasmapheresis) or tumor removal, 253 patients (53%) showed symptom improvement within 4 weeks of treatment; 125 patients (57%) received second-line immunotherapy (rituximab or cyclophosphamide), and of these 84 had an mRS score of 0-2 during the first 24 months.
Plasmapheresis is a first-line therapy option 4,5 the encephalitis associated with antibodies against the N-methyl-D-aspartate receptor (NMDAR, however evidence for its efficacy is very limited, mostly derived from case reports and expert opinion. [bib_ref] Therapeutic plasma exchange for the treatment of anti-NMDA receptor encephalitis, Pham [/bib_ref] [bib_ref] Response of anti-NMDA receptor encephalitis without tumor to immunotherapy including rituximab, Ishiura [/bib_ref] [bib_ref] Successful treatment of anti-N-methyl-D-aspartate receptor encephalitis presenting with catatonia, Schimmel [/bib_ref] [bib_ref] Successful treatment of antiN-methyl-d-aspartate receptor limbic encephalitis in a 22-monthold child with..., Agrawal [/bib_ref] It has been used for other paraneoplastic neurological syndromes (subacute cerebellar degeneration, paraneoplastic encephalomyelitis, paraneoplastic opsoclonus/myoclonus, and cancer-associated retinopathy). The findings of specific autoantibodies in these syndromes led to the use of plasmapheresis in their management. [bib_ref] Adjunct therapeutic plasma exchange for anti-N-methyl-D-aspartate receptor antibody encephalitis:a case report and..., Mirza [/bib_ref] For these syndromes, the American Society for Apheresis assigns a category III (grade 2 C) recommendation. [bib_ref] Guidelines on the use of therapeutic apheresis in clinical practice:evidence-based approach from..., Szczepiorkowski [/bib_ref] However, effects of individual treatments (e.g. Plasma exchange vs. IVIg) cannot be compared side-by-side statically. [bib_ref] Clinical experience and laboratory investigations in patients with anti-NMDAR encephalitis, Dalmau [/bib_ref] Although it is difficult to establish an unequivocal causal relationship between plasmapheresis and improvement in this case because other modalities of immunotherapy were used beforehand, our patient showed a significant clinical improvement concomitant with plasmapheresis initiation. |
Mediterranean Quarantine Regulations
## 207
The reporter reminds the Academy, that when Dupuytren demanded of the young Egyptians who were brought to that assembly by Clot-Bey, after they had obtained the title of doctors in medicine from the Faculty of Paris, what was the opinion of the most enlightened men in Egypt on the origin of the plague, they answered, " the plague proceeds from the ground meaning thereby, that a moist marshy soil more or less covered with animal and vegetable matters in putrefaction, is a powerful cause of changing the air, and thereby of inducing plague.
To show the influence of terrestrial conditions in the production of the disease, the topography of Damietta is contrasted with that of Fayoum, from the account of Dr Rossi. Fayoum is raised above the level of the sea; Damietta borders the sea-coast. At Damietta the air is hot and moist; at Fayoum it is hot but dry. Fayoum is exempt from marshes ; Damietta is surrounded by marshy pools formed of fresh and salt water. At Damietta the cemeteries are within the town ; at Fayoum they are remote from the dwellings of the inhabitants. At Fayoum, the water, though not exquisite, may be drunk without inconvenience, by reason of the nitre which it contains in abundance; at Damietta, the fresh water is mixed with sea-water, or it is defiled by excrementitial products, and by putrefying animal and vegetable matters. Fayoum is surrounded by the desert of Lybia ; Damietta is surrounded by rice-fields, and situated in front of the formidable Delta. Now, though these two towns belong to the same country, are under the same government and laws, and are inhabited by the same people; yet in one, viz. Damietta, the plague is endemial, while the other is exempt from the disease, though often persons affected with plague come thither to die.
It is scarcely possible to imagine a stronger proof of the fact, that the formation of plague depends on local causes, much more than on contagion or importation. If, indeed, to this be added the analogous fact, that plague is never observed to be developed in Upper Egypt, it must be admitted that strong reasons are afforded for calling in question the influence of the latter cause.
Atmospherical variations exercise great influence on the development and progress of pestilential epidemics. Larrey, Pugnet, and all the physicians who succeeded them in Egypt, observed that the attacks are more numerous, and the mortality greater, when the air is hot and moist, than when the weather is tempestuous.
At Constantinople, the same causes produce the same effects.
The reporter also adverts to the uniform and long continued condition of the meteorological constitution, noticed by Humboldt and many other observers. Thus, in speaking only of plague, it is after long continuance of the same winds, the same tempera-2()8 ture, that observers have most commonly observed epidemic plague spring up in Egypt, in Syria, and at Constantinople. It may be conceived, he argues, that when a population have lived for a long period in the same conditions of climate, atmosphere, alimentation, and other physical conditions, the organisms become deeply modified in the same manner, and are disposed to receive or even to develope the disease.
It may be further observed, that both with regard to epidemic visitations of plague and yellow fever, the atmosphere is unusually still, calm, and little disturbed by winds; and as to plague, at least in general, when it is most violent and general in its attacks, the air is often little disturbed by currents for weeks. When at length the Etesian, or north winds, have begun to blow at the _ stated 24th of June, the disease suddenly abates greatly; new cases are no longer heard of, and in a few days it is extinct.
Though the same accuracy is not observed as to yellow fever, yet it has been remarked, that so long as the atmosphere remains calm, still, and little disturbed, cases continue to take place, and mortality proceeds. If, however, a hurricane or tornado arise, a sudden stop is put to the disease ; some of the sick speedily become well; and, for a time at least, the progress and violence of the epidemic are temporarily suspended. It is only, however, when the temperature falls considerably, a change of wind sets in, or rain descends, that the disease begins to subside, and disappears for that season.
To this influence and modification thereby of the organisms of the population, the reporter attaches very great importance; in short, a degree of importance in which, we think, he will get few dispassionate observers to follow him. He seems to think, for instance, that the organisms of the natives of different countries are so much and so deeply imbued with this influence, that when a pestilential epidemic is prevailing in any given country, it may affect the systems of the natives of that country, though at a distance and in different conditions. It seems as if he believed that there was, between the epidemic constitution and the temperaments of the inhabitants of a given country, some secret chain or link of communication, by which the disease was made to affect these persons, and these only, amidst many other masses of human beings. " One remarkable fact," he says, " observed in several epidemics, and which has been denied because it was impossible to explain it, might thus be explained, and might perfectly confirm this truth, namely, that individuals who have long been subjected to the same influences may be attacked by the same disease at a given period, though at this time they might be stationed in different countries and placed in different conditions.'" Diemerbroeck knew several families whose members, though remote from each other, had been attacked by plague at the time Plague, its Origin and Propagation. 209 at which this disease was prevailing, in places where these families were dwelling. He specifies this coincidence with regard to the family Von Dams of Nimeguen. ''The father, dreading the plague as to two of his children, sent them to Gorcum in Holland.
The third remained with him at Nimeguen. The two children who were at Gorcum, where there was no plague, remained for three months in perfect health. But all at once they were attacked with plague, and died at a period not distant from that at which the father and the third child fell under the disease at Nimeguen.1, In this fact, if it happened as related, we see nothing very remarkable. The latent period, or what may be called the period of incubation in plague, is not well known. But there is reason to believe that it varies in different individuals ; and that, supposing the disease to depend on the operation of a poison developed at, or evolved from a particular locality, it may take in one person one month, in another six weeks, and in a third two months, to mature itself so as to form the complete disease. There is surely nothing very extraordinary in observing, that two children, who had been exposed to the air of a locality where plague was prevailing, should, when removed from that locality, have been, after the lapse of a certain time, attacked by the disease. It is quite possible that, before quitting Gorcum, these children had imbibed a sufficient dose of the poison to have their systems affected, whether they had been removed from, or had remained in the locality. The susceptibility of a certain family or temperament may be admitted, without in any degree supposing that there is in the organisms of the natives such a condition, as renders plague more likely to attack them, even in a distant locality, than the persons among whom they are residing.
The reporter, however, has other anecdotes which he adduces as facts confirming the same inference. Evagrius, he tells us, speaking of an epidemic of plague which reigned at Antioch, thus speaks : " A circumstance truly astonishing is, that when the inhabitants of a city desolated by the epidemic were absent, and in places where the disease was not prevailing, they alone were attacked by it." Procopius also, treating of the plague of 542, thus expresses himself: It infected with its poison, in a healthy city, persons who were born in that in which it was prevailing. Senac also states that the English, at a season when plague was prevailing in England, were attacked by it even in foreign countries. This fact he tries to -explain by stating, that the relation of parentism is a sort of contagion. These facts, if such they are to be viewed, it would be idle to explain, until we know that they are facts. Their accuracy we altogether doubt; and it appears to us no improper stretch of incredulousness to do so, unless better evidence of their truth is afforded. We decline, indeed, attaching importance to any state-Mediterranean Quarantine Regulations. ment made by Evagrius and Procopius, unless supported by collateral authority.
Of a similar tendency is another statement given by the reporter. The epidemic action varies in certain circumstances according to the race. Thus he tells us that negroes have a peculiar predisposition to be attacked by plague. The following statement is quoted from a report by M. Penay to the Council of Health at Cairo on the plague of 1841. " Towards the close of the month of May 1841, a passage-boat descending the Nile had ten black slaves at Nadder, while plague was raging there. The day after their arrival most of the slaves were ill; some died the same evening; others during the night. The master quitted the village of Nadder, carrying with him the rest of the slaves, who died before debarkation at Kafr-Regard. It deserves attention, that neither the crew nor the other passengers who loaded this vessel, and were mixed indiscriminately with the pestilential persons, were attacked by the disease."
Not one syllable of information is here given as to the state of the slaves ; the manner in which they were fed or sheltered; their clothing ; their general condition ; or the moral state of persons forcibly and violently dragged from their homes and relatives. On all these points a decent and profound silence is observed. Of one thing only we may be certain. The negroes, Nubians, and Arabs, it is added, were living nearly in the same hygienic conditions. All were in free pratique.
As to the rest of the population of Alexandria, the conclusion to be drawn from the numbers is less certain, he allows, considering that the conditions of isolation and hygienic position were not the same for all. Here, however, are the numbers :
The Greeks lost Jt is quite idle to reason on these numerical proportions and statements, without knowing much more of the relative position, social and physical, of the members of these different nations. So much depends on modes of living, means of accommodation, the district in which the persons reside and sleep, and other circum-stances, that, without very ample information on all these points, it is evidently absurd to speak of the influence of the pestilential poison in different nations and tribes of men. Only one remark, we add, on the alleged hygienic conditions of the Negroes, Nubians, Maltese, and Arabs. They are all almost, with little exception, labourers and porters, working in the sun, at times very hard and laboriously ; getting little for their labour, and that not steadily ; living badly; sleeping in wretched situations, often on the ground and near the quays; and drinking strong liquors when they can procure them.
The converse of the proposition already stated is however also observed. Thus strangers passing through or residing in a city visited by plague, remain free from its attacks. Thus, in a plague which prevailed at Copenhagen, the English, Dutch, and Germans escaped the disease. In other circumstances, again, strangers are more affected than natives. In the year VII. (1798,) when plague reigned at Damietta, the Turks, who then constituted two-thirds of the population of the town, underwent only 8 attacks among 100, which affected the French and the Greeks.
The resisting force to the action of epidemic causes of plague varies according to family, sex, age, professions, mode of life.
The general conclusions are the following.? 1. Epidemic plague attacks many individuals at once. 2. In the different latitudes in which the disease has appeared, it has presented periods of commencement, establishment or height, and of decline.
3. During its prevalence, other diseases become greatly more rare than usual, and are almost always stamped by its characters.
4. In periods of plague, the epidemic influence is felt by persons who have formerly had the disease, and even by those who have not had it, and who retain health. 5. In Egypt, Syria, and at Constantinople, it has been possible to determine the periods of invasion of pestilential epidemics, that of their increase, and that of their termination.
6. Epidemic plague has been often preceded, and its approach, as it were, announced by other epidemic disorders.
Whatever may be the fact as to the existence and operation of contagion, it is certain that the disease, in epidemic constitutions, is often propagated without any evidence of that agent. It must be admitted that the atmosphere of certain localities becomes, as it were, impregnated with the active principles of the distemper; and that, for some time at least, this atmosphere alone is capable of communicating the disease. This it does, not by causing the disease to pass from one person to another, but by affecting several persons, not in communication, simultaneously or successively, according to their susceptibility, and according to the doses of the poison which they have respectively imbibed.
vol. lxviii. no. 172. p The local atmosphere, in such circumstances, is an epidemic focus of infection ; and every individual remaining in it is liable to contract the disease. Numerous and authentic facts observed in Egypt in 1835 and 1841, proved that the most complete isolation, the most rigorous quarantine, do not always preserve those who are subjected to it. The same remark was made as certainly at Marseilles and Toulon, during the great plague of 1721.
If this be the fact, it follows that removal from the impregnated atmosphere, if early practised, must save those who do remove from it. Numerous facts accordingly are recorded by writers on plague epidemics, of persons who, having quitted plague-stricken cities, have escaped and not been attacked by the disease. It is almost idle to refer to the instances of persons who, in the various plague epidemics which affected London, withdrew from that city to the country, and remained free from the distemper; because it may be said that they were not susceptible, or had not been previously in communication with persons labouring under the disease. Yet it is a well established fact, that, though free communication was maintained at these periods between London and the neighbouring towns ; and, indeed, it was impracticable to isolate London, or put its inhabitants under quarantine segregation ; yet the disease was never imported into these towns, and if carried, it did not spread. 213 tained an ordinance that this regiment should be sent to a healthy place where plague was not prevailing. From this time the regiment presented no more cases of plague.
In April 1841, when the 3d regiment of cavalry was in garrison at Neguille, in Lower Egypt, they were suffering from plague as other inhabitants of the locality. At the commencement of May they received orders to proceed to Zagazig, a village situate in the Delta, near the right mouth of the Nile. The plague patients, amounting to twelve, of whom several were dangerously ill, were placed in boats ; but, in consequence of the penury of the means of transport, it was impossible to isolate them from the other soldiers ; and during the whole voyage they were in contact with the workmen of the regiment and the soldiers1 wives, who encumbered the vessels. At this period, though the intensity of the disease was abating, several plague patients were severely ill, and the condition of four of these made M. Penay, the physician to the regiment, afraid, lest they should die during the passage. Fortunately this apprehension was not realized. As soon as the plague patients began to remove from the sickly locality, they underwent a marked improvement; and by the time they arrived at Zagazig, after a voyage of ten days, almost all were convalescent.
Only one died during the voyage. Those who reached Zagazig completely recovered. They met there the favourable circumstances in which they were placed after their departure from Neguille, that is, they found a region in which the epidemic constitution had not been developed. He concludes with the following aphorism. When plague ravages violently in Africa, Asia, or Europe, it always shows itself with the principal characters of epidemic diseases.
The reporter next shows, from a comparative view of the characters of sporadic and epidemic plague, that the former differs from the latter not only by the small number of persons attacked, but also and especially by its not presenting the characters peculiar to epidemic diseases.
A question of considerable difficulty next occupies attention.
Is plague propagated after the manner of epidemic diseases, that is, by the migration of certain atmospheric influences, and independently of the action which the plague patients may exercise ? This question is not only not free from difficulty, but it is stated in such terms, that many persons will be disposed to object to it at the very outset. What is the mode, it will be asked, in which epidemic diseases are propagated P Is this positively ascertained ? Is it understood that the persons of those affected have no influence whatever in propagating the disease ? Can it be imagined that when ten, fifteen, or twenty thousand persons are labouring under any acute disease of the nature of fever or plague, the exhalations from the persons of these patients are to have no influence in diffusing and extending the disorder ? Such will be the language held by many, we do not pretend to say, with how much or how little reason. Many who allow that yellow fever is not originally a contagious disease, and arises mostly from terrestrial emanations, acted on by particular atmospheric conditions, contend, nevertheless, that, as the disease advances and affects many persons congregated in one locality, and accumulated towards one or more points, it acquires contagious properties, and thus is extended and diffused in the ulterior course of its visitation.
The correctness of this view we have always held doubtful; yet we believe that many respectable members of the profession conscientiously believe it to be well founded. As to plague, we are much at a loss what opinion to entertain. We have always thought that more was made of the supposed power of plague patients in diffusing the disease than the facts of the case could warrant.
At present, the question must be regarded as open to inquiry, and for a considerable time to continue so. We have merely to observe that the reporter, evidently in his mode of stating the question above given, presumes that the persons of the sick, and the alleged emanations from them, have no concern in the propagation and diffusion of the disease. This presumption, however, is at variance, as we shall see, with the views of some of those observers even, who deny the existence and influence of contagion.
The reporter, in order to illustrate the question now stated, adduces the following facts and considerations.
Localities adjoining to that in which epidemic plague has sprung up, often remain free from any attack. Plague has often, for example, prevailed in Trebizond without appearing at Platana, a port with which the inhabitants of Trebizond have constant communication.
In 1834, plague raged violently at Constantinople, while the Isle of Princes and Buyukdere had not one single case of the disease, notwithstanding communication at all times. In 1835, when Cairo was a prey to a pestilential epidemic which carried off one-third of the inhabitants, the town of Abouzabel, which is only four leagues from the capital, and communicated freely with the latter, was for near two months exempt from the disease.
Sometimes, on the contrary, epidemic plague attacks successively and gradually the city and the neighbouring villages. Frequently it strikes towns remote from each other, avoiding Intermediate points. Does epidemic plague, by aid of the atmosphere alone, cross seas and pass from one continent to another; for instance, passing the Mediterranean to go from Alexandria to Marseilles ? Such is the opinion of Clot-Bey and M. Aubert-Roche.
Epidemic plague, however, may meet, not far from the place where it arises, insurmountable barriers; at least limits beyond which it does not extend. Thus plague rising in Lower Egypt Plague, its Origin and Propagation. 215 never passes the first cataract. Some places raised at a considerable height above the level of the sea, appear to enjoy complete immunity. At five leagues1 distance from Constantinople is a village situate on the mountain of Alem-Daghe, at the elevation of about 500 metres = 1640 feet above the level of the sea.
A ccording to Dr Brayer, plague never appears there; and during the period of epidemics, this village is a place of refuge to the inhabitants of the capital. On the same mountain, but lower down, is another village, which is by no means exempt from visitations. In Malta, also, is one spot which, from this circumstance, has hitherto been inaccessible to plague. Lastly, Desgenettes and Clot-Bey ascertained that in plague epidemics which affected the city of Cairo, the citadel had always escaped any attack.
An important inquiry would be, to determine when a pestilential epidemic prevails in a town, how many patients owe their disease to the epidemic constitution, and how many are affected in consequence of absorption of miasms, or exhalations which have escaped from the persons of the sick, either by their direct or indirect contact.
On this point M. Lacheze ascertained that at Alexandria and at Cairo, epidemic influence had struck persons completely segregated, so as to cause the death of one person among 400 ; while plague had destroyed one person among three among a population left in free pratique, that is, exposed at once to the influence of the epidemic constitution, and to that of miasms which might be extricated from the plague patients ; in short, to contact direct or indirect with persons infected.
The reporter justly observes, that it is easy to see that the data here given are not much to be trusted to. Without calling in question the numerical statements of M. Lacheze, it is possible easily to give them a different explanation.
It is said that it is sufficient that persons who have performed quarantine, and those who are in free pratique, should be in opposite hygienic conditions, in order that the epidemic should attack them in very different proportions, and consequently, in order that the pestilential miasma, or the contact of plague patients, should not have performed the part ascribed to them. The reporter thinks it certain that persons who are placed in quarantine enjoy greatest ease, and take most care of their health. This point appears to us very doubtful, and can only be understood of those who withdraw altogether from the scene and locality of plague. Those who remain, and who are either placed or place themselves in quarantine, are subjected to so much restraint and confinement in most situations, that the remedy appears often worse than the disease. They may take as much care of their health as such a position allows; but it is certain that confinement, detention, and restrictions on personal liberty, are by no means favourable to the maintenance of health.
216 Mediterranean Quarantine Regulations.
The reporter, however, to place the terms of comparison more on a footing of equality, tells us that they endeavoured to ascertain what was, whether at Alexandria or Cairo, the great establishment placed in quarantine, and containing a population in conditions as similar as possible to those in which the population left in free pratique were living. The Arsenal of Alexandria, which, during the epidemic of 1835, contained 6000 workmen at least, appeared in this respect to deserve consideration.
In that establishment, no attack can be ascribed to an accumulation of pestilential miasms, because such accumulation did not exist, as, whenever a patient was recognized to be affected by plague, he was immediately conveyed to an hospital situate without the arsenal. It is equally impossible to allege as a cause the contact of pestiferous patients, considering that, either in consequence of the patient's being removed at the commencement of the attack, or from some other cause, those next to the individuals attacked by plague, and those who had touched the latter, were never attacked by the distemper. The number of the workmen of the arsenal conveyed to hospital on account of plague furnishes, therefore, the number of cases owing to epidemical influence alone, in the class least comfortable. 300 workmen having been attacked by plague among a 217
In conclusion, the reporters think the following proposition . How long may the infection of plague remain dormant in an infected person before it declares itself by evident symptoms ? 6. How long will the contagious matter of plague, when lodged in inanimate substances, retain its infecting power ? 7. What are the means by which substances containing the contagion of plague may be purified ? Answers of different kinds were received from fifteen physicians. Seven gave answers decidedly in favour of plague being a contagious disease;?Dr Floquin of Smyrna, Dr M'Carthy of Constantinople, Dr Raffinesque, Dr lcare, Dr Gaetani Bey, Dr Del Signore, and Dr Grassi, Proto-Medico of Health in Egypt, at Alexandria. One, Dr Pruner of Cairo, allowed that plague arises spontaneously in Egypt; that it may be either sporadic or epidemic ; and that in the first case it is scarcely contagious, but that, when epidemic, it may probably become contagious nearly in the same degree, under the same conditions, and in the same manner as typhus, malignant dysentery, hospital gangrene, and similar diseases, which may be developed by miasms, and may be propagated by infection in confined situations, as in prisons, barracks, and hospitals.
Five physicians maintained that the disease was not propagated by contagion, but by endemic or epidemic influence ; Clot-Bey, Dr Abbot, M. Seisson, Dr Fischer, and Dr Laidlaw; and other three, M. Perron, Dr Duvigneau, and Dr Robertson, speak rather doubtfully, and, excepting the last, with some confusion, yet more against than in favour of contagion ; allowing that its existence is either not proved, or that the phenomena cannot be explained by its existence and influence.
Of these, Clot-Bey denies any thing like contagion ; contends that the disease is never communicated by this principle ; maintains that the disease is the offspring of the soil and atmosphere, and local causes generally; that the atmosphere, however, is the principal, if not the sole, agent by which the disease is formed and developed, and admits that it is propagated only by epidemic influence. Dr Abbot thinks that actual contact is not necessary, though all close approaching is dangerous ; but he maintains the omnipotence of epidemic influence, and the inutility of quarantine restrictions. He thinks the disease is not communicated by goods or clothes.
M. Seisson speaks doubtfully of contagion. If the disease be contagious, he argues, it is not so in the manner of diseases reputed and generally known to be contagious. The influence of contagion is in short not proved. On the contrary, the disease is certainly communicated otherwise than by contagion. It springs up spontaneously in countries where it is endemic, as in Egypt and Turkey; and when epidemic influence is once manifested, segregation and quarantine restrictions are utterly useless. He thinks it not impossible that in extraordinary circumstances, and independent of importation, epidemic influence may be manifested spontaneously in countries where the disease is not endemic, yet which have some analogy in climate and geological characters with those regions in which it is. The contact of an infected person is not indispensable to the communication of the disease, but it adds to the danger of epidemic influence that of the infecting influence of patients under the disorder.
M. Perron speaks much in the same manner. After perplexing himself about contagion, he at once maintains that the existence of contagion, as usually understood, that is a virus transmissible to distances more or less remote, either by individuals or by articles, is to him any thing but proved ; in short it is not proved.
He admits that persons remaining in a place where plague is prevailing may contract and do thus contract the disease ;?in short he allows the existence and operation of a locally tainted atmosphere. He allows also that if patients are too much approximated to each other, their persons may infect the atmosphere. He recognizes the existence of epidemic influence. Plague is an endemic disease, and requires for its production the climate and soil of Egypt, and the accessary habits of its inhabitants; and it is quite impossible, physically impossible, he thinks, to export plague to countries in which neither the climate nor the soil of Egypt are found.
The epidemic constitution of the atmosphere he also admits ; and as it is impossible to say that this may not extend over different parts of the globe, or may be developed in some parts or other, he does not deny altogether the utility of quarantine restrictions, but he maintains that they might be and must 219 be, in order to be useful and not hurtful, very different from those at present in use. He thinks, however, that, in spite of every precaution, there may be established in Europe, and over the whole extent of the globe, general circumstances in which plague may find what is favourable to its formation and diffusion, and, in that case, no means in the world will prevent this distemper from travelling in various directions, and committing the usual ravages in different countries.
Dr Duvigneau gives a similar opinion, though encumbered with various contradictions. He thinks that to contagion too much influence is ascribed. In certain epidemics, as in the plague of 1839 at Cairo, the first cases looked like the effect of contagion ; but such coincident accidents are observed in all epidemics. Sometimes one side of a street, one district alone is attacked. Plague takes place at Cairo and none at Boulacy ; villages continue in a state of perfect health amidst villages suffering under the distemper. Plague is not always contagious. Sporadic plague is certainly not. Of this fact Lower Egypt presents annual examples. Some persons only are attacked ; and the disease arises spontaneously. When the disease is contagious, so far as it may be so, the co-existence of certain peculiar circumstances, a certain disposition of the individual, perhaps a change in the magnetic state of the ground are requisite. When this state ceases to exist the contagious power is extinct. During, and at the close of the plague, the effects of the dead, their mattresses, coverings, linen, and all are put in use, often without the previous precaution of any measure of salubrity, or even of cleanliness. Meanwhile the ravages of the distemper cease; the distemper itself completely disappears ; because the general influence having ceased, the contagious matter has lost its action.
The following anecdote is given by Perron and Duvigneau, and is repeatedly quoted by others, in order to show the small power of contagion. We know not whether its indications be so strong as the authors imagine. But it may be received for what it is worth.
There was established at Abouzabel, four leagues from Cairo, a sort of quarantine for the hospital. Tents were placed in the desert at a short distance, for keeping the patients under observation for several days, before they were made to enter the hospital. The majority of patients were furnished by the hospital at Cairo, which was crowded with sick and plague patients. Their number was very soon too great for the space assigned to receive them ; and they could often remain only three days, two days, and even twenty-four hours under observation. Five physicians, of which Dr Duvigneau was one, were attached to the service of the hospital. They performed daily two or three visits to the plague patients, whose tents were removed only two hundred paces from the first. Forty inspections were made at periods shortly after death.
Three bodies were inspected 15, 20, and 25 minutes after death. The physicians never adopted any other precautions except that of washing the hands with water and vinegar, then with water and soap, when the former was wanting. They never changed their dress before entering the hospital, where they made daily two visits, approaching and touching the patients as usual. Notwithstanding all these means for facilitating propagation, not one of the physicians experienced the slightest accident; not a single patient of the hospital was attacked with plague; and yet the number of those treated during the months of April, May, and June was considerable, as it amounted to one thousand persons.
By the advocates for contagious influence, it would be said, first, that one negative fact proves nothing amidst several or many positive'facts; and secondly, that the persons so exposed might not be susceptible. It cannot be denied, nevertheless, that where none were attacked under circumstances apparently of considerable exposure, the disease cannot be very contagious.
Dr Duvigneau allows, however, that the persons of plague patients may furnish exhalations, which, under the action of a peculiar inexplicable influence, may communicate plague, directly by contact with the patient, indirectly by contact with articles which have been used by them, and impregnated with their exhalations.
The most extraordinary testimony perhaps is that given by Dr Fischer, Professor of Anatomy at the School of Medicine at Abouzabel.
He first perplexes himself a little about the nature of contagion, which he states to be of three sorts. 1st, Contagion properly so called, which produces itself by immediate contact either of person or of articles impregnated by contagious matter, as in itch. 2d, Contagion by inoculation, in which a poisonous material is introduced into the human body, and taken into the blood-vessels, as in rabies. And 3d, Cantagion by infection, when the noxious elements are received by absorption in the mucous membrane of the respiratory organs, or of the intestinal tube.
He is not in possession of facts to prove that glandular plague is propagated by contagion properly so called ; and, notwithstanding various facts which seem to prove this mode of propagation, he is of opinion that glandular plague is not contagious, but propagates itself only by infection under the influence of particular causes. This opinion he rests on the facts, that cases of glandular plague are observed every year in Egypt; that those affected sporadically present all the usual symptoms of the disease, as during the prevalence of an epidemic ; that these persons die speedily ; that they are seen and touched by many persons without the disease Plague, its Origin and Propagation. 12*21 being diffused : and their beds and other articles are used by other persons without thus extending the disease; that various villages remain exempt while others are attacked, for instance the exemption of Abouzabel already mentioned ; while, on the other hand, during epidemic visitations, persons are attacked and cut off, notwithstanding the strictest isolation and the most rigorous quarantine.
Dr Laidlaw of Alexandria is of opinion that plague is endemic in Egypt, and in the Turkish dominions. He has seen sporadic cases repeatedly arise, and destroy the individuals without any other person being attacked in consequence; and he is sure, from multiplied instances, that the disease is never so communicated.
As to the infectious nature of the disease, by which he understands the property of spreading itself from man to man, though he feels that he is not warranted in saying, that the disease is in no circumstances communicated from individuals; yet he sees in this no ground for the inference, that it is the cause of the disease spreading epidemically. He observes, however, that whenever any number of plague patients are assembled together, they seem to create an infectious atmosphere, if ventilation be not particularly attended to; for in general, when many plague patients are collected in one point, all the attendants are attacked. Plague depends in a great degree, though not wholly, on the season of the year. It is not epidemic annually. Its epidemic visitations evidently depend by plague are crowded together, their presence may contaminate the atmosphere, and create an epidemic constitution or influence, which shall affect all that enter the limits of that atmosphere. This Dr Gaetani Bey calls mediate contact. Persons affected with plague in various epidemics have fled from Lower Egypt to Upper Egypt, and there have died. Yet in no instance has it been found that the disease appeared in Upper Egypt. The soil and climate are unsuitable.-In Lower Egypt also, many populous villages not subjected to any quarantine restrictions, even of the slightest nature, though surrounded by infected places laid waste by the epidemic, were never attacked by the distemper. If it be asked, what conclusion is to be drawn from these well ascertained and remarkable facts, Dr Gaetani Bey thinks that it must be, that plague did not penetrate to Upper Egypt, and enter these villages, solely because epidemic causes were wanting and did not operate.
He finally concludes ; 1st, that epidemic causes, though, under some circumstances, they may of themselves produce the disease, yet generally do no more than predispose the human frame to its attacks ; and, indeed, plague, lie thinks, has rarely been known to arise from these causes acting alone, while it has very frequently been observed where they have acted conjointly with the latter. 2d, That the two kinds of contact, though insufficient to produce the disease, where the general pathogenic causes do not exist, are yet capable of determining its developement, wherever the presence of these causes has predisposed the individual to the attack.
Dr Del Signore has no doubt, that plague is communicated by contagion, and he knows no other mode of communication. Contact with the infected person is necessary for the communication of plague, and simple proximity is not sufficient. Consequently, a medical attendant runs no risk of contracting the disease, if he visits a body of infected individuals, investigates their pathological condition, and examines the condition of every symptom, so long as he avoids only touch.
Dr J. Robertson, Deputy-Inspector-General of Hospitals, had occasion to observe plague at Jaffa in February 1841, shortly after at Beyrout, where it prevailed from the 27th February to the 23d June 1841, and, as we stated in last article, it was prevailing at the same time at Acre and Sidon, in short, over much of the coast of Syria. Dr Robertson paid much attention to the mode in which the disease appeared and arose in the former places; and as to the two latter he had accurate information. On the 4th of May 1841, he inspected all the houses at Beyrout known to be infected ; and as these towns are not large, the means of information may be allowed to be more applicable to the purposes of the etiological inquirer. The following are the conclusions at which he arrives. Two of his facts, 10 and 11, are premised. tion ; in that case, a stray dog or cat, a bird or insect, is accused as the infecting medium, or some contaminated piece of cotton or paper has been blown in at some unprotected aperture. It would appear, therefore, from what I have stated, that numerous persons who have been in intimate communication with the diseased, escape ; and that frequently those who most carefully seclude themselves fall victims, whose attack can only be accounted for by the contagionists on the most improbable grounds.
" Upon a review of all the circumstances, the only conclusion I can arrive at is, that plague arises from local causes, although the nature of these, and the peculiar circumstances under which they are generated, are, and probably will for ever remain unknown. The endemic nature of the disease in Syria is proved by the fact, that it is often developed in the interior of the country, in remote mountain villages, distant from the great route of communication with those countries where plague is known to exist. It existed at Zeb-d-Eni, in Anti-Lebanon, during the last summer,?a season when it is never known on the sea coast. " An intelligent French surgeon, who was in the Egyptian service, stated that for two years preceding 1841, plague had existed in some of the mountain villages of Lebanon. The disease produced great mortality at Acre, Caiffa, Tyre, and Sidon ; at the latter place 1800 are said to have perished out of a population of 9000 ; 400 cases occurred among the troops, and 190 proved fatal. Contrast this result with what occurred at Beyrout, when the total number of cases was 121, notwithstanding there were abundant opportunities for the disease to propagate itself by contagion. Further north than Beyrout, the disease was even more rare ; at Tripoli it did not appear at all; and Aleppo, according to information 1 received in my inquiries there, had not been visited by plague for thirteen years; yet all these places were not protected by any sanitary regulations whatever, nor indeed could any ever be established that would be efficient in a country like Syria, to cut off communication between diseased and healthy places. The fact seems to be, that the causes of plague, whatever may be their nature, are sometimes confined to certain localities of a country, sparingly generated and acting with little intensity, only affecting a few and probably highly predisposed persons ; at other times the deleterious influence is widely developed, and acts with extreme energy, producing an epidemic, with frightful mortality. Want of cleanliness, heat and moisture, the decomposition of animal and vegetable matter, have been assigned as causes of plague ; but all these, if they have any influence, it is only of a secondary nature ; and some peculiar circumstances, as yet beyond the reach of science, are necessary to its development. Some of the above causes do not seem to have much influence even on the progress of the disease. which it will be seen that at no great distance there is a low irrigated ground, highly productive of miasma. In the neighbourhood of this ground 1 admit the attacks were most numerous, the scanty population considered.
" Temperature certainly has great influence on the disease, which never, or scarcely ever, appears until after the great heat of summer is over, and disappears in June when the hot weather returns. Hot days and cold nights seem peculiarly favourable to its production.
face of the ground and the atmosphere together, have arrived at that state, whatever it may be, which may be called plague-point.
When this is the case, no measure but the removal of materials, the ejection and expulsion of the human beings from the region, can be of the smallest benefit.
In the third part of their Report, the reporters enter on the question of the communicability of plague either by means of epidemic foci, or without these adjuncts.
And they first consider the question of the communicability of plague by inoculation by blood drawn from the vein of a plaguepatient, purulent matter taken from a bubo, or serum extracted from the phlyctaena of a pestilential carbuncle.
Of all such experiments, which appear to us both disgusting and not conclusive, it is sufficient to say that not one has furnished any useful information whatever. It is to be observed that these experiments are inconclusive, because, being performed in situations and at seasons when plague is or was prevailing, the disease, when it does take place, might have been owing, not to the inoculation, but general causes, as infection or epidemic influence.
At the very first we meet with a story which shows the lameness of the cause which is to be supported by such means, and the gross credulity or blind folly of those who believe such stories. It is pretended that Willis inoculated himself with plague in 1665; and died of the disease. This story the reporters call in question. They mayjustly do so. Willis died at the age of 55, not in 1665, as this story implies, but on the 11th day of November 1675, not of plague, but of bronchitis, terminating in pneumonia and pleurisy. That was further not a plague year.* But had this lancet under the epidermis at two points of the inner surface of the arm of one of the condemned persons.
On the 21st April, this convict had plague, which M. Lacheze calls confirmed, and which M. Clot-Bey regarded as doubtful. On the 26th April, the patient was convalescent.
On the 22d April, M. Lacheze performed another inoculation by means of pestiferous blood. It produced no result.
On the 30th of April, the same attempt was made on another subject, and was equally ineffectual.
Lastly, a young man, aged 18, who had mild plague, after having slept on the 15th April previously on the clothes of a bed recently left by a plague patient severely attacked, was inoculated with the blood of a plague patient on the 13t.h of May. This inoculation was equally without result.
The same person had undergone, eight days previously, and with the same immunity, an inoculation in the groin and in the arm-pit, with the serum taken from the vesicle of a carbuncle.
Another condemned person, aged 16 years, was inoculated on the 20th April with serum taken from a carbuncle. The result was negative. On the 30th of April, the same person was inoculated in the arm-pit and in the groin of the right side with the matter of a bubo, which had just been opened. This inoculation was without effect.
From these experiments it follows, that among four persons inoculated with the blood of a plague patient, one only had an attack of mild plague; while two persons inoculated with serum taken from a pestilential carbuncle, and a third inoculated with the matter of a bubo recently opened, experienced no symptom of disease. The reporters add, that if it be remarked, that the individual who had mild plague in consequence of inoculation by blood, was then, as all the inhabitants of Cairo, under the epidemic influence ; that, further, he was for three days in an hospital which contained many plague patients under treatment, and which hospital had become the source of pestilential infection, it will easily be allowed, that doubts may be entertained on the real cause of the disease.
At a later period, Clot-Bey inoculated himself with the blood of a plague patient. By a lancet dipped in this blood, he made six punctures sufficiently deep, three of which were at the anterior part of the left fore-arm, and three at the fold of the right groin. No symptom of plague followed. Some days after this ineffectual inoculation, Clot-Bey inoculated himself with matter from a pestilential bubo, by three punctures made on the internal surface of the left arm. This trial was followed by slight feelings of uneasiness, which the experimentalist ascribed to absorption of purulent matter, and which presented no analogy with the symptoms of plague.
vol. lxviii. no. 172. Q Lastly, M. Pruner, in a letter addressed in 1829 to the consulgeneral of England, residing at Alexandria, asserts that the blood of plague patients and the matter of buboes have produced, by their inoculation, nothing like plague.
The only inferences which can be justly drawn from experiments of the kind now mentioned are, that they furnish at best doubtful results ; and if we adhere strictly to their interpretation, they would show that plague is a disease not communicable by inoculation, either by pestilential blood or the matter of buboes.
The next question which demands attention is, whether plague is communicable by contact of patients, that is, by immediate or direct contact. This is the mode which has been believed by the whole body of contagionists, wherever the doctrine of contagion has been taught, to be least objectionable, most perfectly established, and that by which plague was most usually and frequently propagated. It is needless to trace the history of the origin of this doctrine from Fracastorius downwards, which has been already often done in this .Journal. It is enough to say, that, though from time to time, in the course of nearly three centuries, individuals appeared, who thought for themselves, yet, on the whole, it was the accredited doctrine in all the medical schools of Europe ; and any one who ventured to express a doubt of its truth, or to differ in opinion, was regarded as an eccentric unreasonable person, with whom it was idle and dangerous to hold any communication.
Matters were in this state when, in 1835, a considerable number of European physicians had occasion to observe the extensive epidemic which prevailed over Egypt. The result of this was, that their previous opinions were very much shaken; and while M. Brayer and Cholet, who had observed plague at Constantinople in 1819, 1826, and 1834, changed their opinions as to the contagious character of the distemper ;?MM. Aubert-Roche, Clot-Bey, and others, either took the view that the distemper was not contagious at all, or very much modified their opinions as to this property. The facts which have contributed to effect this revolution are various and multiplied. A few only can be noticed.
Several have been already stated.
In the plague of 1824, at Cairo, 30,000 individuals were destroyed. In Alexandria, only two or three cases took place, and the disease never spread, although the communications were positively unrestricted. Sometimes plague has raged in the suburb of Boulac, and yet has not entered Cairo, though no precaution has been adopted.
Two harbours situate on the Red Sea, Suez and Cosseir, derive all their supplies of food from two towns in the interior. Suez receives its subsistence from Cairo, and Cosseir from Keneh, a town of Upper Egypt. In 1835, plague broke out at Keneh almost at the same time as at Cairo. Suez was attacked by the disease; while Cosseir, left in free communication with Keneh, had not one single attack. The temperature of Suez is lower than that of Cosseir, which, surrounded with dry naked mountains, is founded on rocks ; while, on the contrary, Suez is surrounded by stagnant waters and marshes of salt water, circumstances which give it the atmospheric and physical condition of the towns of the Delta.
During the epidemic of 1835, Djedda, Yambo, and Moka enjoyed the same immunity as Cosseir, though plague patients, proceeding from Suez or other infected places, died in the middle of the inhabitants of these towns.
It has been observed that while Syria and its towns are ravaged from time to time by plague, yet Arabia, with which there is the freest communication, has never been affected. Nubia, Sennaar, and Abyssinia have also remained exempt, notwithstanding their free and uninterrupted relations with Egypt. Now, if in regard to Arabia, Sennaar, and Nubia, it may be said that the high temperature prevents the condensation of pestilential miasmata, the same reason cannot be applied to Abyssinia, a temperate country, and where the thermometer varies from 15? to 25 centigrade above zero. Here the salubrity of the country can alone repel the disease. Abyssinia is a country of mountains and elevated table-lands arranged like a series of inclined planes, where are neither marshes nor stagnant waters.
One of the least equivocal characters of the non-contagion of a disease, however, is the fact that either attacks take place independently of each other; that is, when two or more persons are attacked, there is no evident communication between them ; or if this take place about the same time, all are attacked within so short a time of each other, that it is impossible to believe that the distemper could be transmitted from one to the other, and that it is not less impossible to doubt that all have become ill from one general and common cause. Of these characters the following are examples.
Three soldiers of the 5th regiment were attacked by confirmed plague. They belonged to three different companies. As they were left in free communication with their comrades till the time when the nature of the disease was established, the whole regiment nearly was found in the state called compromised, that is to say, suspected, yet no other case took place among the corps.
At the same time, M. Penay, physician to the 5th regiment of cavalry in garrison at Neguille, in Lower Egypt, ascertained that among 40 plague patients only 2 came from the same tent; all the others had been attacked in different tents.
Though the hospital at Neguille received many patients, yet none of the persons who were in frequent contact with the patients, such as the nurses or servants of the hospital, the apothecary, the physician, experienced the smallest symptom of plague.
At the commencement of the epidemic of 1841, the Plague Hospital at Cairo was not yet established, and plague patients were received into the Military Hospital and treated along with those attacked by other disorders. The latter class of patients did not contract plague. One hundred and eighty-two plague patients were treated in the hospital. The numerous staff of attendants of all ranks attached to the establishment presented no attack. The ordinary patients and the attendants in this hospital, nevertheless, amounted to 1200 persons.
Among the body of officers of health, consisting of 92 Europeans and 300 Arabs, no one performed quarantine. All zealously attended plague patients without observing any precaution. Among the European officers of health, three cases of plague took place, of which two recovered. Among the Arabs, three cases also took place, but all died.
M. Chedufau, the chief physician to the Military Hospital at Cairo, attended in the city 61 persons ill of plague, in whom he opened buboes, dressed carbuncles, and to whom he devoted every sort of attention.
He performed 17 morbid inspections.
During the whole epidemic he never ceased to be in communication with his family, and without employing any preservative ; yet neither he nor any of his family presented any symptom of the disease.
The general conclusion at which the reporters arrive from these and similar facts is, that attentive and rigorous examination of facts shows, on the one hand, that the immediate contact of thousands of plague patients is free from danger to those who practise this contact in open air, or in well ventilated situations ; and, on the other hand, that no correct observation proves the communicability of plague by the mere contact of patients. It has been very uniformly and confidently maintained by European physicians of all countries, that plague is transmissible in articles of clothing and such materials as have been near or in contact with the persons of plague patients; that articles of this kind might imbibe or be impregnated with the peculiar pestilential virus; that this they might retain for years, and then, by giving it out when opened or used, engender plague, and propagate the disease in situations where it was either unknown or had not for a long time appeared. Such articles have been denominated fomites of contagion by all the teachers in the European schools ; and against the introduction of plague, by means of these fomites, many, if not most of the restrictions, rules, observances, and expensive and sometimes destructive processes of the quarantine laws were directed. As some articles were conceived highly capable of being impregnated with this material cause of plague, and others less so or not at all, this led to the distinction of articles into susceptible, highly susceptible, little susceptible, and insusceptible ; and it is utterly impossible for the most fertile imagination to conceive any thing more wild, more irregular, or more unreasonable than the division of articles into these different classes and orders, and the variety of rules and modes of disinfection thereby suggested. To speak of principle, or seek for some system, is quite idle and foolish ; for no principle can be found. All is arbitrary confusion, unreasonable and futile distinction, and only shows what extravagances human reason may perpetrate when under no guidance but that of fear, ignorance, and the most gross and confident presumption.
Of all this doctrine of transmissibility by articles of clothing and similar substances, no trace existed in the Levant till within the last twenty-five or thirty years, when learned Europeans began strict ions, have convinced me not only of the barbarity, but of the absolute folly of such measures. I was myself a Member of the Board of Health at that time, and it has been a consolation to me to reflect that I had contributed to induce the government to discontinue that practice for the time.
" I am aware that the cessation of the plague at Malta, in 1813, and latterly at Odessa, is ascribed by some individuals to the employment of similar means. In both cases, however, the disease ran its course for several months, and when it finally ceased, there is good ground for believing that its suppression was effected, not so much by isolation, as by some change in the atmosphere, proceeding, as in Egypt, from natural causes, or by the enforcement of greater attention to the means of diminishing its pestilential condition, and by supplying the poorer classes with nourishing food."?P.
## 510, 511.
Observe next what is stated on the same subject by the Cairo Board of Health, including several physicians expressly appointed to give a judgment on the merits of quarantine restrictions. " Reforms of which the Quaratitine Laws are susceptible.?Quarantines being established for the sole object of preventing the plague from the penetrating into Egypt, the Commission is of opinion, that if the plague is endemic in this country, they are of no use, since, even admitting the supposed importation of the disease, there would be required for its development the concurrence of the general influences of which we have before spoken, and which of themselves would be sufficient for its development without importation, for there are always in the country more elements of production that could be brought from abroad, viz. contaminated effects, and cases of plague. The inutility is still farther confirmed with respect to the importation from Syria, since it appears to be proved by the experience of every age, that the plague has never been introduced into Egypt from Syria. " Seclusion or shutting up of the Houses.?The Commission declares that the measures of seclusion are not only useless, but pernicious and inhuman, inasmuch as they spread terror among those who are exposed to them, and cause the people to evade them by violating all the laws of medical police. It is, moreover, sufficiently proved, that seclusion has never stopped the progress of an epidemic.
Seclusion is one of the causes that promote the development of typhus, as it has the effect of keeping the individuals thus shut up exposed to the effects of a vitiated atmosphere. The case of Giglio, so often brought forward in support of contagion, has clearly manifested the fatal consequence of seclusion, since, among the numerous individuals who were in communication with the first one attacked, only those who were shut in the house fell victims to the disease, whilst two of his brothers, who had made their escape in order to evade seclusion, did not experience the least sickness."?P.
## 512.
Then follows the recommendation. The Commission being persuaded of the good intentions of the Plague, its Origin n/ul Propagation. 233 consuls, is induced to hope for, if not the total abolition of quarantine in Egypt, at least the suppression of the more rigorous sanitary restrictions, and that we shall no longer see vessels, coming from ports where no plague exists, subjected to quarantine at Alexandria during the prevalence of an epidemic there."?P. 512.
If these statements do not demonstrate most unequivocally, not only the inutility, but the pernicious effects of quarantine restrictions, we despair of the intellectual faculties of those who have the authority, ever becoming enlightened or amenable to reason. This doctrine of contagion by fomites is giving way with that of personal contagion, before observation and actual experience.
Among the physicians to whom the interrogations of the English government were circulated, the greater number doubt entirely, or call in question, this mode of communication ; and among those who admit it in certain articles, more than one allows that the distinctions of susceptible and unsusceptible articles are founded neither on reason nor on experience.
Dr Floquin tells us that articles of clothing?woollen, cotton, silk, fur, and paper and leather, are all conductors of plague contagion ; and that hair, wood, metal, glass, are non-conductors.
How can fur be a conductor, it may be asked, and hair a nonconductor ? Dr M'Carthy agrees with the quarantine authorities in considering wool, cotton, fur, silk, and all articles manufactured from them ; also, leather, paper, and metals with a rough surface, and money or currency, as conductors and susceptible.
Dr Grassi regards all substances capable of receiving infection, but not of retaining it. Metals he allows capable of receiving, but not in all cases of retaining it. He admits that it is not possible to assign any reason, why wool, cotton, and skins are more fitted to receive and transmit contagion than wood and metals. He finds horse-hair not on the list of conductors or susceptible; and also amadou, which last is highly susceptible. Dr Clot-Bey allows, that the opinion most generally received affirms, that plague may be communicated by articles which have been in contact with the persons of plague patients. He sees, however, no reason to believe that this opinion rests on well established facts. He does not absolutely deny that articles which have been long in contact with the persons of plague patients may be impregnated by miasms or some similar subtile substances issuing from these persons ; but he repeats, that this does not rest on anything like demonstration by incontestable facts. The fashion of ascribing the communication of the disease to these and similar articles, he thinks, has arisen from indifference to inquire after the truth.
It is easy and simple to ascribe any given attack of plague to contact real or supposed of articles suspectcd to be iro-bued with pestilential matter, with those who think it difficult to comprehend how plague can arise from atmospherical causes.
Numerous facts prove, on the other hand, that the closest contact with suspected or used articles is not followed by plague in those who afterwards use them. The director of the hospital at Cairo neglected to cause to be washed the bedding articles, coverings, mattresses, and pillows, which had been used by hundreds of plague patients; and gave them these without previous purification, to be used by ordinary patients. Not one contracted plague. During the epidemic of 1835, more than 500 houses whose inhabitants had died were shut up. The furniture, beds, sofas, were left in these abodes, exactly as they had been when the dead bodies were removed. Towards the'end of June and beginning of July, all these houses were opened ; the effects were sold, and dispersed through the city and beyond it, without having undergone any sort of disinfection. Not one of the persons who first entered these houses, or who used these articles, were attacked by plague.
Dr Abbot has never known the malady to be thus communicated.
" In the Egyptian navy in 1835, quantities of effects of every kind, such as sheets, blankets, red night-caps, shoes, &c., were received on board the vessels of war, two or three days before their departure for Candia, from the stores of the arsenal, where the plague was raging. These articles certainly could not have been transported from the magazines without being touched by some infected person. In fact, I know positively that many persons ill with the plague were concealed in those same store rooms, that they might not be seen by the quarantine medical officers. The day before our departure, the wives and families of the sailors were permitted to come on board the ships, to take leave; and many of the sailors of the ship I was in were then on rafts, busy in caulking the sides, and in other work. They not only received several articles from the hands of their relatives, but I myself, and several other officers, saw them embrace their wives and children. Many of the most susceptible objects were taken on board, and yet I am not aware that the least case of plague broke out in any ship after we sailed from Alexandria. Probably if a dress worn by a sick person should be immediately put on by one in health, while saturated with the perspiration and exhalations of the infected person, the disease might be communicated ; but if the article were exposed for an hour to the sun and wind, I do not think it would communicate the disease to the wearer."?P. 471.
Dr Seisson thinks positively that inert bodies do not communicate plague. Egypt furnishes a remarkable fact which proves the negative, whatever may be stated on the affirmative side by many authors. The fact is this. Plague ceases at the end of June, at which time Cairo contains a multitude of articles of all sorts, which have belonged to the victims of the epidemic; and a great part of the population are in fact clothed in the spoils of the dead; for as the Arab does not undress to go to bed, most of the patients have passed their whole disease and died in the same garments. In the hospitals, even the articles of plague patients are preserved without distinction between them and those of other patients. No one suffers from this indiscriminate and unpurified use of articles reputed suspected.
The classification of articles which ought to be founded on physical properties and chemical affinities seems unreasonable and frivolous, arbitrary and useless.
Dr Perron admits not one of the marvellous stories of alleged conveyance of contagion and disease by boxes, threads, bales, handkerchiefs, and other modes. All are equally incredible, and all equally silly. Dr Duvigneau allows that articles that have been in contact with infected persons may communicate the disease, more especially articles of cotton or woollen, by imbibing and transporting miasmata.
Dr Fischer has not had experience to enable him to speak positively. But he is rather opposed to the belief of contagion by fomites on the usual grounds.
Dr Raffinesque believes that the disease may be communicated by means of the articles of clothing of plague patients. The substances most susceptible are linen, woollen, cotton, silk, and various tissues ; woolly hair of the hare, deer, camel; and paper. Wood, metal, horse-hair, vegetables, fruits in general, and all substances smooth and polished are not susceptible.
The answer of Dr Icare of Smyrna is quite similar to that of Dr Raffinesque. Dr Laidlaw is not prepared to deny that, under the circumstances influencing the communication of the disease by infection, communication by clothing of the sick may be possible ; but, so far as his own experience goes, he has no evidence to offer that it actually takes place. He is in all other respects inclined rather to doubt and to deny all such communication. His observations and facts are extremely interesting, and we may take a subsequent opportunity of placing them at length before our readers.
Dr Gaetani Bey allows that certain objects which have been long in contact with the sick may communicate the disease, because, being impregnated with morbid exhalations, and possibly morbid products of every kind, both being absorbed, may, by the concurrence of epidemic causes, give rise to the distemper. The articles most likely to be impregnated and to communicate the disease are hair, feathers, hemp, cotton, silk, furs; articles unlikely are compact smooth bodies, as metals and inorganic substances ; and some vegetable products, as wood, leaves, fruits, &c.
## Mr ormerocts clinical collections
A number of facts and arguments very similar are given by the reporters. Those that we have now adduced, however, are sufficient to show the reasoning on both sides of the question. We have only to close this part of the statement with the conclusion given by the reporters, which is;?That numerous facts prove that goods and articles of clothing which have been used by plague patients have not communicated the disease to those persons who have used such articles without purification.
We should now proceed to consider another division of this part of the subject, namely, the question as to plague being transmitted by articles of merchandize supposed to contain pestilential materials. This question is closely connected with the previous one.
The examination in the report is short though decided, and, on the whole, not unsatisfactory. The fullest statement of the question is given by Dr Laidlaw in his answers, and this we should subjoin, but for want of space at present. We hope on another occasion to be able to give his views at greater length.
Meanwhile, we subjoin merely the general conclusion contained in the report. This is, that the communicability of plague by merchandize in countries in which the disease is endemic or epidemic, is by no means a demonstrated fact.
This conclusion is, it must be confessed, merely negative. But it must be remembered that the affirmative appears at present altogether in the form of an assumption. In short, when we consider the facts already mentioned of the cessation of the plague in Egypt at a definite time ; the perfect immunity which the inhabitants enjoy between June and the beginning or end of December at all times; and the immunity which they occasionally enjoy for years; yet, with all the articles used by plague patients in full employment; it seems unreasonable to imagine that the disease is ever communicated in this way ; and it seems not unreasonable to think that all the instances of alleged communication by suspected articles must have originated from some other causes. As to the alleged distinctions of susceptible and unsusceptible articles, it appears as if the persons who made them had never used their observation or intellectual faculties at all.
[table] established ?: Plague, abstracting from the influence which may be exerted by plague patients, is propagated after the manner of epidemic diseases ; that is, by the action of general causes.Let us now see, before finally quitting this part of the subject, what information the evidence given by the physicians interrogated by the English government furnishes on this question. [/table]
|
Lingual leiomyomatous hamartoma with bifid tip and ankyloglossia in a patient without oral-facial-digital syndrome: a case report and literature review
Here is a rare case of lingual leiomyomatous hamartoma (LLH) with bifid tongue tip and tongue-tie in a patient with non-oral-facial-digital syndrome (OFDS). A 29-year-old male consulted for a painless tumor over the midline of the tongue dorsum measuring 2 × 1.5 cm. The tumor was excised and the tongue-tie was corrected. Diagnosis of LLH was based on histo-pathologic and immuno-histochemical studies. The epidemiologic data and differential diagnosis of LLH, as well as related literature, are discussed. To date, only 14 cases of LLH have been reported in English literature. This may be the first reported case of LLH with bifid tip and ankyloglossia in a non-OFDS patient.
# Background
Hamartoma was first defined by Alberechtin 1904 as a benign, tumor-like malformation composed of a disordered mixture of mature tissues that normally occur in the affected part, but with a predominance of one particular tissue. Hamartomas often occur in the liver, kidneys, lungs, spleen, pancreas, and in the oral region and are classified into many subtypes according to their main component.
Leiomyomatous hamartoma (LH) is composed mainly of smooth muscle tissue intermixed with other adjacent tissue and often occurs in the lungs and kidneys. Cases involving the oral cavity are extremely rare and most have occurred in the Japanese, with few cases among Caucasian and Latin Americans of young age [bib_ref] Leiomyomatous hamartoma of the tongue: a case report, Kobayashi [/bib_ref]. Only 14 cases of LH in the tongue have been reported, most of which are in young children; these have been reported individually or as part of a genetic syndrome, especially oral-facial-digital syndrome (OFDS).
The OFDS is a complex syndrome with 10 different subtypes that include oral, facial, and digital abnormalities. LH with a lobulated tongue tip or short frenulum has been reported in some types of OFDS [bib_ref] Leiomyomatous hamartoma of the midline maxillary gingival presenting as a congenital epulis:..., Zhang [/bib_ref]. This report is of a case of lingual leiomyomatous hamartoma (LLH) with a bifid tip and ankyloglossia in a 29-year-old man without OFDS. The current English literature on LLH is also reviewed.
## Case presentation
A 29 year-old man was referred in June 2008 for a painless, slow-growing mass over the tongue dorsum that had been noted since childhood. Upon examination, there was evidence of an exophytic, reddish, non-tender mass, measuring 2 × 1.5 cm over the midline of the tongue dorsum. There was also a bifid tongue tip and tongue-tie [fig_ref] Figure 1: The preoperative and postoperative appearance of tongue tumor [/fig_ref]. He denied any discomfort associated with this mass. The mass was excised and the tongue-tie was corrected smoothly under general anesthesia.
Microscopically, the polypoid tumor was nonencapsulated. It was composed of smooth muscle cells with cigar-shaped nuclei mainly arranged in irregular thick bundles intermingled with blood vessels, nerves, and adipose tissue within fibrous stroma [fig_ref] Figure 2: The histopathological study of LLH in hematoxylin and eosin [/fig_ref]. Immuno-histochemical study [fig_ref] Figure 3: The histopathological study of LLH in immuno-histochemical stain [/fig_ref] revealed a positive reaction with smooth muscle actin (SMA) and muscle actin (HHF 35) in the smooth muscle bundles and vascular walls. The nerves and adipose tissue were also immunoreactive to the S-100 protein. These features were all characteristic of LH.
The patient was followed up uneventfully for four years after surgical intervention. The surgical wound was wellhealed, with mild scar formation [fig_ref] Figure 1: The preoperative and postoperative appearance of tongue tumor [/fig_ref]. There was no evidence of recurrence.
# Discussion
This report is the fourteenth known case of lingual LLH but is perhaps the first case of a non-OFDS patient presenting with a bifid tip and ankyloglossia. Of the 14 cases with LLH reported to date [fig_ref] Table 1: Summary of cases of lingual leimyomatous hamartomas reported in English literature [/fig_ref] , ten (71.4%) are male and four (28.6%) are female patients. Most cases have been observed at birth, but age at diagnosis has a range of 8 days to 61 years. The LLH is usually a congenital, selflimited, and slow-growing tumor, such that the patient may not have any discomfort unless the mass suddenly becomes enlarged. In the review of cases, most have been diagnosed at an age <6 years and most lesions are <2 cm in size. Most cases involving the oral region have occurred in the Japanese [bib_ref] A case of hamartoma occurring in tongue and upper gingiva, Kanekawa [/bib_ref] [bib_ref] Congenital tongue mass associated with heterotopic smooth muscle: report of a case, Tamaki [/bib_ref] , although according to the report of Nava-Villalba, 50% of patients with LLH are Caucasian and only three patients are Japanese. To date, the present case may be the first reported Chinese incidence in the world [fig_ref] Table 1: Summary of cases of lingual leimyomatous hamartomas reported in English literature [/fig_ref].
Kobayashi et al. [bib_ref] Leiomyomatous hamartoma of the tongue: a case report, Kobayashi [/bib_ref] have mentioned that the tumor might become apparent when it enlarges rapidly after recurrent trauma or during the pubertal growth spurt. This might explain the huge LLH (3.6 × 3.0 × 2.0 cm) in De Faria and Batista' [bib_ref] Giant leiomyomatous hamartoma of the tongue, De Faria [/bib_ref] report. Kobayashi et al. [bib_ref] Leiomyomatous hamartoma of the tongue: a case report, Kobayashi [/bib_ref] also mention that oral hamartoma usually arises from the foramen cecum during embryonic development, which may explain why most LLH are in the midline of the tongue. Furthermore, most LLH are solitary and only two cases have multiple lesions [bib_ref] Giant leiomyomatous hamartoma of the tongue, De Faria [/bib_ref].
The normal tongue is composed predominantly of skeletal muscle with an overlying, keratinizing, stratified, squamous mucosa. Smooth muscle is part of the normal lingual vasculature but does not occur independently. A hamartoma is a proliferation of normal tissues that are considered endogenous to the site of occurrence. Typically the tissues in hamartomas appear disorganized and ill-defined, merging with the normal surrounding tissues. Within the tongue, endogenous elements that might result in a hamartoma would include vessels, nerves, lymphatics, muscle, fat, and salivary gland tissue. The etiology of the combination of developmental variations: hamartoma + bifid tongue + ankyloglossia (tongue-tie) in our report remains obscure. The sporadic mutation in the transcription factor may occur during the tongue development at the fourth embryonic week, causing the abnormality of morphogenesis at the end stage of the lingual swelling mergence. As poor fusion between the lateral and the medial lingual swelling occurs, the tongue tumor, combined with the bifid tongue and ankyloglossia may occur together on the tip of the tongue midline. Environmental factors, for example, poor nutrition; overuse of vitamins; use of drugs, alcohol or heroin; smoking; and exposure to thalidomide or pathogens may cause sporadic mutation with abnormality of morphogenesis during embryologic development.
In patients with a congenital midline tongue mass, the differential diagnosis includes choristoma, leiomyoma, and benign mesenchymoma. Choristoma [bib_ref] Multiple leiomyomatous hamartoma in the oral cavity: case report, Iida [/bib_ref] [bib_ref] Hamartomatous tongue lesions in children, Kreiger [/bib_ref] is a well-organized normal tissue proliferation in abnormal locations, usually occurring in elderly patients. Leiomyomamost commonly arises from the smooth muscle of the female genital tract, uterus, or the esophagus, and rarely occurs in the tongue. Moreover, it is often found in the fourth and fifth decades of life. Benign mesenchymoma [3,7-9] is a circumscribed, unencapsulated benign tumor composed of two or more mesenchymal lineage tissues without any single predominant tissue. No nervous tissues are noted in mesenchymoma [bib_ref] Hamartomatous tongue lesions in children, Kreiger [/bib_ref] even though some may show adjacent tissue infiltration. In LLH, there is disorganized tissue growth and smooth muscle predominance but without local tumor infiltration. These often tend to occur in young patients. However, there are other rarer tumors that should be put into the differential diagnosis of LLH. First, congenital epulis [bib_ref] Leiomyomatous hamartoma of the midline maxillary gingival presenting as a congenital epulis:..., Zhang [/bib_ref] which usually presents on alveolar gingival mucosa at birth is the most common congenital intra-oral, benign, soft-tissue tumor, and rarely occurs in the dorsal tongue. Second, the congenital genetic syndrome, OFDS, should also be among the differential diagnoses of LLH. The OFDS usually includes malformations of the mouth, teeth, jaw, facial bones, and limbs, together with varying degrees of mental retardation. An LH combined with lobed tongue and mouth frenulum may occur in some types of OFDS. The unusual combination of LLH with bifid tongue tip and ankyloglossia in a non-OFDS patient, as in the present case, has not been previously reported. Third, teratoma is more commonly associated with other dysmorphism than LLH. For example, Andrade and Raikwar [bib_ref] Congenital benign teratoma of tongue with bifid tip, ankyloglossia and polydactyly: report..., Andrade [/bib_ref] mentioned a case of congenital benign teratoma of the tongue with bifid tip, ankyloglossia and polydactyly. The etiology of both hamartoma and teratoma combined with bifid tongue tip and ankyloglossia are still unknown; however, genetic or embryonic events might explain the coincidental finding.
Immunohistochemical characteristics are essential in differentiating soft-tissue neoplasms. The SMA and S-100 protein are the most commonly used markers in confirming the histomorphologic findings in LLH [bib_ref] Leiomyomatous hamartoma of the tongue: a case report, Kobayashi [/bib_ref] [bib_ref] Leiomyomatous hamartoma of the midline maxillary gingival presenting as a congenital epulis:..., Zhang [/bib_ref] [bib_ref] Giant leiomyomatous hamartoma of the tongue, De Faria [/bib_ref] [bib_ref] Multiple leiomyomatous hamartoma in the oral cavity: case report, Iida [/bib_ref]. However, use of vimentin, desmin, and HHF35 has been reported in other articles [bib_ref] Leiomyomatous hamartoma of the tongue: a case report, Kobayashi [/bib_ref] [bib_ref] Leiomyomatous hamartoma of the midline maxillary gingival presenting as a congenital epulis:..., Zhang [/bib_ref] [bib_ref] Giant leiomyomatous hamartoma of the tongue, De Faria [/bib_ref] [bib_ref] Multiple leiomyomatous hamartoma in the oral cavity: case report, Iida [/bib_ref]. SMA is a specific smooth-muscle immunomarker that can be seen in smooth-muscle bundles and vessel walls, in the present case [fig_ref] Figure 3: The histopathological study of LLH in immuno-histochemical stain [/fig_ref]. The S-100 protein is normally present in cells derived from the neural crest, chondrocytes, adipocytes, myoepithelial cells, macrophages, and keratinocytes. Adipose tissue and nerve fibers may also have positive staining but smooth muscles have negative staining [fig_ref] Figure 3: The histopathological study of LLH in immuno-histochemical stain [/fig_ref]. So, S-100 protein can be a key immunohistochemical marker to differentiate LLH from vascular or solid leiomyoma. HHF35 reacts with all kind of muscle cells, pericytes, and myoepithelial cells, but is nonreactive with endothelial-, epithelial-, neural-, or connective-tissue cells; this can also confirm the smooth muscle predominant diagnosis in our article [fig_ref] Figure 3: The histopathological study of LLH in immuno-histochemical stain [/fig_ref].
Most case reports advocate surgical excision with a safe margin as the treatment of choice for LLH [bib_ref] A case of hamartoma occurring in tongue and upper gingiva, Kanekawa [/bib_ref] [bib_ref] Hamartomatous tongue lesions in children, Kreiger [/bib_ref] [bib_ref] Congenital benign teratoma of tongue with bifid tip, ankyloglossia and polydactyly: report..., Andrade [/bib_ref] [bib_ref] Leiomyomatous hamartoma of the posterior tongue: a case report, Goldsmith [/bib_ref] [bib_ref] Leiomyomatous hamartoma of the anterior tongue: report of a case and review..., Rosa-García [/bib_ref]. Prognosis is generally quite good and no post surgical recurrence has been reported in the current literature.
# Conclusions
Here is a special case of LLH combined with bifid tongue tip and tongue-tie in a patient without OFDS. To date, this may be the first such case reported in English literature. It is a reminder that when confronted with a smooth-mass lesion over the midline of the tongue dorsum, leiomyomatous hamartoma should be considered in the differential diagnosis, especially when the patient is a young adult and without OFDS.
## Consent
Written informed consent was obtained from the patient for publication of this case report and any accompanying images (IRB number: KMUH-IRB-20120262). A copy of the written consent is available for review by the Editorin-Chief of this journal.
[fig] Figure 1: The preoperative and postoperative appearance of tongue tumor. (A) Preoperative view of an exophytic, well-defined, reddish, smooth, non-tender mass midline over the tongue dorsum measuring 2 × 1.5 cm in diameter. An M-shaped (bifid) tongue tip and tongue-tie were also noted. (B) Four years postoperative the surgical wound was well-healed, with only mild scar formation. [/fig]
[fig] Figure 2: The histopathological study of LLH in hematoxylin and eosin (H&E) stain. (A) Muscle bundles, blood vessels, and adipose tissue in the polypoid tumor covered by stratified squamous epithelium (H&E stain; original magnification × 2). (B) The non-encapsulated tumor was composed of smooth muscle bundles, blood vessels, and adipose tissue within fibrous stroma (H&E stain; original magnification × 4). (C) Smooth muscle bundles admixed with a nerve (arrow) and blood vessels (H&E stain; original magnification × 20). [/fig]
[fig] Figure 3: The histopathological study of LLH in immuno-histochemical stain. (A) Positive reaction with smooth muscle actin in the muscle bundles, but negative in the nerve (arrow). (Immuno-histochemical stain; original magnification × 20). (B) Positive reaction with muscle actin (HHF 35) in the muscle bundles, but negative in the nerve (arrow). (Immuno-histochemical stain; original magnification × 20). (C) Positive reaction with S-100 protein in the nerve (arrow), but negative in smooth muscle bundles. (Immuno-histochemical stain; original magnification × 20). [/fig]
[table] Table 1: Summary of cases of lingual leimyomatous hamartomas reported in English literature [/table]
|
Host Factor SAMHD1 Restricts DNA Viruses in Non-Dividing Myeloid Cells
SAMHD1 is a newly identified anti-HIV host factor that has a dNTP triphosphohydrolase activity and depletes intracellular dNTP pools in non-dividing myeloid cells. Since DNA viruses utilize cellular dNTPs, we investigated whether SAMHD1 limits the replication of DNA viruses in non-dividing myeloid target cells. Indeed, two double stranded DNA viruses, vaccinia and herpes simplex virus type 1, are subject to SAMHD1 restriction in non-dividing target cells in a dNTP dependent manner. Using a thymidine kinase deficient strain of vaccinia virus, we demonstrate a greater restriction of viral replication in nondividing cells expressing SAMHD1. Therefore, this study suggests that SAMHD1 is a potential innate anti-viral player that suppresses the replication of a wide range of DNA viruses, as well as retroviruses, which infect non-dividing myeloid cells. (BK) . These authors contributed equally to this work.
# Introduction
It is becoming increasingly evident that host cells employ metabolic regulatory mechanisms in order to restrict the life cycle of pathogens [bib_ref] Metabolic pathways as regulators of HIV infection, Craveiro [/bib_ref] [bib_ref] The interaction of metabolic factors with HCV infection: does it matter, Bugianesi [/bib_ref] [bib_ref] Dynamics of the cellular metabolome during human cytomegalovirus infection, Munger [/bib_ref] [bib_ref] Systems-level metabolic flux profiling identifies fatty acid synthesis as a target for..., Munger [/bib_ref]. The recent discovery of sterile alpha motif (SAM) domain and histidine-aspartic (HD) domain-containing protein 1 (SAMHD1) has contributed to our understanding of the metabolic regulation of deoxynucleoside triphosphates (dNTPs), the substrate for cellular DNA polymerases to synthesize and repair host DNA. SAMHD1 expression limits proviral DNA synthesis in lentiviruses particularly in non-dividing myeloid cells such as macrophages and dendritic cells (DCs) [bib_ref] Vpx relieves inhibition of HIV-1 infection of macrophages mediated by the SAMHD1..., Hrecka [/bib_ref] [bib_ref] Tight interplay among SAMHD1 protein level, cellular dNTP levels, and HIV-1 proviral..., Kim [/bib_ref] [bib_ref] SAMHD1 is the dendritic-and myeloid-cell-specific HIV-1 restriction factor counteracted by Vpx, Laguette [/bib_ref] [bib_ref] SAMHD1 restricts the replication of human immunodeficiency virus type 1 by depleting..., Lahouassa [/bib_ref]. SAMHD1 is a dNTPs triphosphohydrolase, and functions by hydrolyzing dNTPs into dNs and triphosphates [bib_ref] HIV-1 restriction factor SAMHD1 is a deoxynucleoside triphosphate triphosphohydrolase, Goldstone [/bib_ref] [bib_ref] The Aicardi-Goutieres syndrome gene and HIV-1 restriction factor SAMHD1 is a dGTP-regulated..., Powell [/bib_ref] , thus leading to the reduction of cellular dNTP concentrations [bib_ref] Vpx relieves inhibition of HIV-1 infection of macrophages mediated by the SAMHD1..., Hrecka [/bib_ref] [bib_ref] Tight interplay among SAMHD1 protein level, cellular dNTP levels, and HIV-1 proviral..., Kim [/bib_ref]. This in turn can impact the kinetics of cellular, viral, and parasitic DNA polymerization by reducing the availability of dNTP substrate for the enzyme.
Cellular dNTP concentrations are significantly varied among cell types [bib_ref] Modification of human immunodeficiency virus type 1 reverse transcriptase to target cells..., Jamburuthugoda [/bib_ref]. Due to the close link between S phase-dependent dNTP biosynthesis and cellular DNA replication, dividing cells harbor an abundant amount of dNTPs compared to non-dividing cells [bib_ref] Physiological concentrations of purines and pyrimidines, Traut [/bib_ref]. Indeed, we previously reported that terminally differentiated/non-dividing monocyte-derived macrophages (MDMs), which are a HIV target cell type [bib_ref] Macrophage tropism of HIV-1 depends on efficient cellular dNTP utilization by reverse..., Diamond [/bib_ref] , have 22-320 times lower dNTP concentrations compared to actively dividing CD4 + T cells [bib_ref] Macrophage tropism of HIV-1 depends on efficient cellular dNTP utilization by reverse..., Diamond [/bib_ref] [bib_ref] rNTPs as substrate of human immunodeficiency virus type 1 reverse transcriptase in..., Kennedy [/bib_ref]. Even though lentiviral reverse transcriptases (RT) have evolved to function at low dNTP concentrations, the limited dNTP availability contributes to a significant delay in proviral DNA synthesis in macrophages as compared to activated CD4 + T cells [bib_ref] Macrophage tropism of HIV-1 depends on efficient cellular dNTP utilization by reverse..., Diamond [/bib_ref] [bib_ref] Intracellular nucleotide levels and the control of retroviral infections, Amie [/bib_ref]. However, some lentiviruses, such as HIV-2 and SIVsm, encode an accessory protein called viral protein X (Vpx) that overcomes the SAMHD1-induced dNTP depletion in non-dividing target cells [bib_ref] Vpx relieves inhibition of HIV-1 infection of macrophages mediated by the SAMHD1..., Hrecka [/bib_ref] [bib_ref] SAMHD1 is the dendritic-and myeloid-cell-specific HIV-1 restriction factor counteracted by Vpx, Laguette [/bib_ref]. Upon infection, virally co-packaged Vpx promotes proteasomal degradation of SAMHD1 [bib_ref] The human immunodeficiency virus type 2 Vpx protein usurps the CUL4A-DDB1 DCAF1..., Bergamaschi [/bib_ref] [bib_ref] Characterization of simian immunodeficiency virus SIVSM/human immunodeficiency virus type 2 Vpx function..., Goujon [/bib_ref] , leading to the rapid elevation of cellular dNTP concentrations and ultimately the acceleration of proviral DNA synthesis [bib_ref] Tight interplay among SAMHD1 protein level, cellular dNTP levels, and HIV-1 proviral..., Kim [/bib_ref] [bib_ref] SAMHD1 restricts the replication of human immunodeficiency virus type 1 by depleting..., Lahouassa [/bib_ref]. Both the Vpx-induced dNTP pool elevation and the promotion of viral reverse transcription were observed in several non-dividing viral target cell types which include macrophages [bib_ref] Vpx relieves inhibition of HIV-1 infection of macrophages mediated by the SAMHD1..., Hrecka [/bib_ref] [bib_ref] Tight interplay among SAMHD1 protein level, cellular dNTP levels, and HIV-1 proviral..., Kim [/bib_ref] [bib_ref] SAMHD1 is the dendritic-and myeloid-cell-specific HIV-1 restriction factor counteracted by Vpx, Laguette [/bib_ref] [bib_ref] SAMHD1 restricts the replication of human immunodeficiency virus type 1 by depleting..., Lahouassa [/bib_ref] , DCs [bib_ref] SAMHD1 restricts HIV-1 infection in dendritic cells (DCs) by dNTP depletion, but..., St Gelais [/bib_ref] [bib_ref] Efficient transduction of myeloid cells by an HIV-1-derived lentiviral vector that packages..., Bobadilla [/bib_ref] and resting CD4 + T cells. Moreover, all these cell types play a significant role in lentiviral pathogenesis. In addition, HIV-1 replicated more efficiently in monocytes isolated from Aicardi-Goutières Syndrome patients, who have mutations in SAMHD1 [bib_ref] SAMHD1-deficient CD14+ cells from individuals with Aicardi-Goutieres syndrome are highly susceptible to..., Berger [/bib_ref]. The enhanced HIV-1 replication likely resulted from the elevated cellular dNTP pools due to loss of phosphohydrolase activity of mutated SAMHD1 [bib_ref] Mutations involved in Aicardi-Goutieres syndrome implicate SAMHD1 as regulator of the innate..., Rice [/bib_ref]. A recent study also reported that other retroviruses such as feline immunodeficiency virus, bovine immunodeficiency virus, N-tropic and B-tropic murine leukemia viruses and equine infectious anemia virus were subject to restriction by SAMHD1 in macrophages, and this restriction was counteracted by the expression of Vpx [bib_ref] Contribution of SAM and HD domains to retroviral restriction mediated by human..., White [/bib_ref]. These recent SAMHD1/Vpx studies support the hypothesis that SAMHD1 imposes a strong evolutionary selective pressure against lentiviral proviral DNA synthesis in non-dividing target cells by limiting dNTPs, the essential metabolic building blocks for DNA.
Interestingly, large double stranded DNA (dsDNA) viruses such as vaccinia virus, herpes simplex virus (HSV) and cytomegalovirus also infect non-dividing cells such as macrophages during the course of infection [bib_ref] Stunned silence: gene expression programs in human cells infected with monkeypox or..., Rubins [/bib_ref] [bib_ref] Induction of Noxa-mediated apoptosis by modified vaccinia virus Ankara depends on viral..., Ferrer [/bib_ref] [bib_ref] Immediate-early viral gene regulation and function, Stinski [/bib_ref] [bib_ref] Characteristics of a macrophage culture persistently infected with herpes simplex virus type..., Bustos [/bib_ref] [bib_ref] Role of interferon in persistent infection of macrophages with herpes simplex virus, Domke-Opitz [/bib_ref] [bib_ref] Nitric oxide production induced by herpes simplex virus type 1 does not..., Lopez-Guerrero [/bib_ref]. However, unlike lentiviruses, these large dsDNA viruses encode dNTP biosynthesis proteins such as ribonucleotide reductase (RNR) and thymidine kinase (TK) that supply essential dNTP substrates for the viral DNA polymerase. Both of these genes are dispensable for HSV-1 viral replication in dividing cells, but are essential for replication under serum-starvation/non-dividing conditions where dNTP pools are limited [bib_ref] Replication of thymidine kinase deficient herpes simplex virus type 1 in neuronal..., Rubenstein [/bib_ref] [bib_ref] A herpes simplex virus ribonucleotide reductase deletion mutant is defective for productive..., Jacobson [/bib_ref]. Thus, it is plausible that the dNTP biosynthesis machinery of dsDNA viruses promotes efficient replication in both dividing and non-dividing target cell types.
In this study we examined whether SAMHD1 affects the ability of vaccinia virus and HSV-1 to replicate in non-dividing cells. We observed that SAMHD1 controls the replication capacity of these dsDNA viruses by limiting the dNTP concentration.
# Results
Vaccinia virus replication in primary human monocytederived macrophages was enhanced by Vpx-mediated degradation of SAMHD1
We investigated whether SAMHD1 affects the replication of vaccinia virus, the prototypical poxvirus, in primary human monocyte-derived macrophages (MDMs). Vaccinia is a large dsDNA virus that replicates entirely in the cytoplasm of the cell and has staged expression with early, intermediate, and late gene expression. Early genes are transcribed in the core of the virion upon entering the cytoplasm, whereas intermediate and late gene expression requires uncoating and replication of the dsDNA genome. For these studies, we utilized the Western Reserve (WR) strain of vaccinia virus that has the viral core protein (A4) fused to YFP (vA4-YFP), [bib_ref] Visualization and characterization of the intracellular movement of vaccinia virus intracellular mature..., Ward [/bib_ref]. A4 is a late gene and is expressed after the viral genome has been replicated. It has been shown that the delivery of Vpx via lentiviral generated virus-like particles (VLPs) reduced the levels of SAMHD1, and increased the concentrations of dNTPs in MDMs [bib_ref] Vpx relieves inhibition of HIV-1 infection of macrophages mediated by the SAMHD1..., Hrecka [/bib_ref] [bib_ref] Tight interplay among SAMHD1 protein level, cellular dNTP levels, and HIV-1 proviral..., Kim [/bib_ref] [bib_ref] SAMHD1 is the dendritic-and myeloid-cell-specific HIV-1 restriction factor counteracted by Vpx, Laguette [/bib_ref] [bib_ref] SAMHD1 restricts the replication of human immunodeficiency virus type 1 by depleting..., Lahouassa [/bib_ref] [bib_ref] Efficient transduction of myeloid cells by an HIV-1-derived lentiviral vector that packages..., Bobadilla [/bib_ref] [bib_ref] The Vpx lentiviral accessory protein targets SAMHD1 for degradation in the nucleus, Hofmann [/bib_ref] [bib_ref] A simple, versatile and efficient method to genetically modify human monocyte-derived dendritic..., Berger [/bib_ref].
First, MDMs were pretreated with VLPs (containing or not containing Vpx) for 24 h, and then infected with virus. At 24 hpi, samples were collected and analyzed by various assays. As shown in [fig_ref] Figure 1: Vaccinia replication in MDMs [/fig_ref] for a representative donor of MDMs, the productive infection frequency was monitored by examining the level of A4-YFP expression by flow cytometry. We then plotted the data for three independent MDM donors [fig_ref] Figure 1: Vaccinia replication in MDMs [/fig_ref] and observed a measurable 1.8-to 2.3-fold increase in the frequency of YFP+ cells treated with Vpx+ VLPs as compared to the Vpx2 VLP treated group [fig_ref] Figure 1: Vaccinia replication in MDMs [/fig_ref]. Six MDM donors were infected with A4-YFP WR virus, analyzed by flow cytometry for infection frequency and plotted in [fig_ref] Figure 1: Vaccinia replication in MDMs [/fig_ref]. The means (p,0.05; Mann-Whitney test) for the Vpx2 versus Vpx+ VLP groups was 59% and 76%, respectively. Finally, we measured plaque-forming units (PFUs) for the different treatment groups, and found that Vpx+ VLP treatment led to a 1.8-to 4-6-fold increase in the amount of infectious viral particles produced as compared to Vpx2 VLP treatment [fig_ref] Figure 1: Vaccinia replication in MDMs [/fig_ref] , To further confirm that the increases observed in infection frequency and PFUs were solely due to Vpx-mediated SAMHD1 degradation, a composite of all the experiments is presented in [fig_ref] Figure 2: Analysis of WRDTK replication in MDMs [/fig_ref] , showing that the no VLP treatment group was comparable to Vpx2 VLP treatment group in all parameters.
Vaccinia thymidine kinase (TK; J2R gene) is an early viral gene and is dispensable for viral replication in tissue culture; but it is important for virulence in animal models [bib_ref] Decreased virulence of recombinant vaccinia virus expression vectors is associated with a..., Buller [/bib_ref] [bib_ref] Development of a replication-selective, oncolytic poxvirus for the treatment of human cancers, Zeh [/bib_ref]. TK is a central enzyme in the nucleotide salvage pathway and catalyzes the addition of a monophosphate to deoxythymidine (dT) [bib_ref] Enzymatic regulation of cytosolic thymidine kinase 1 and mitochondrial thymidine kinase 2:..., Munch-Petersen [/bib_ref]. To look at the involvement of viral TK in the infection of MDMs, the TK gene was removed and replaced with F13L-GFP fusion protein to generate, the WRDTK (DJ2R) strain. The F13 gene is expressed late during infection and MDMs only become GFP+ when productively infected. To validate that the TK deletion, DJ2R, did not compromise viral growth, HeLa cells were pretreated with VLPs and then infected with WRDTK (vDJ2R/ F13L-GFP) or parental WR (vA4-YFP) strain. We found that both strains infected HeLa cells comparably using an MOI of 0.5 PFU/ cell, and did not require Vpx+ VLPs for enhancing infectivity [fig_ref] Figure 3: Expression of SAMHD1 in vaccinia-infected MDMs [/fig_ref] , which is consistent with published data showing that TK is not essential for infection in tissue culture [bib_ref] Decreased virulence of recombinant vaccinia virus expression vectors is associated with a..., Buller [/bib_ref]. To test the importance of viral TK, MDMs were pretreated with VLPs for 24 h and subsequently infected with WRDTK virus. At 24 hpi, we quantified the frequency of GFP+ MDMs by flow cytometry [fig_ref] Figure 2: Analysis of WRDTK replication in MDMs [/fig_ref] , representative donor shown) and plotted the percentages of GFP+ MDMs for three independent donors [fig_ref] Figure 2: Analysis of WRDTK replication in MDMs [/fig_ref]. Similar to the parental WR strain, MDMs pretreated with Vpx+ VLP showed a 2.8-to 6.1-fold increase in the percentage of MDMs expressing the late protein F13L-GFP in the absence of the viral TK gene, indicating an enhancement of viral replication. Six MDM donors were infected with WRDTK virus, analyzed by flow cytometry and plotted in [fig_ref] Figure 1: Vaccinia replication in MDMs [/fig_ref]. The mean percentages for viral infection were 9% and 24% (p,0.05) for the Vpx2 VLP versus Vpx+ VLP groups, respectively. Finally, we observed 3.5-to 27.1fold increase in PFUs between Vpx+ VLP versus Vpx2 VLP treatments [fig_ref] Figure 2: Analysis of WRDTK replication in MDMs [/fig_ref]. PFUs were determined for differentiated THP-1 cells, which were treated with +/2Vpx VLPs and then infected with either WR or WRDTK virus. THP1 cells showed similar trends to treatment as the primary human MDMs [fig_ref] Figure 4: HIV-1 RT primer extension assay monitors dNTP levels [/fig_ref]. Collectively, data from experiments using WR and WRDTK viruses suggest that Vpx-mediated SAMHD1 degradation enhances vaccinia virus infection in MDMs, with an even greater
## Author summary
Various viral pathogens such as HIV-1, herpes simplex virus (HSV) and vaccinia virus infect terminally-differentiated/ non-dividing macrophages during the course of viral pathogenesis. Unlike dividing cells, non-dividing cells lack chromosomal DNA replication, do not enter the cell cycle, and harbor very low levels of cellular dNTPs, which are substrates of viral DNA polymerases. A series of recent studies revealed that the host protein SAMHD1 is dNTP triphosphohydrolase, which contributes to the poor dNTP abundance in non-dividing myeloid cells, and restricts proviral DNA synthesis of HIV-1 and other lentiviruses in macrophages, dendritic cells, and resting T cells. In this report, we demonstrate that SAMHD1 also controls the replication of large dsDNA viruses: vaccinia virus and HSV-1, in primary human monocyte-derived macrophages. SAMHD1 suppresses the replication of these DNA viruses to an even greater extent in the absence of viral genes that are involved in dNTP metabolism such as thymidine kinase. Therefore, this study supports that dsDNA viruses evolved to express enzymes necessary to increase the levels of dNTPs as a mechanism to overcome the restriction induced by SAMHD1 in myeloid cells.
## Vaccinia virus infection does not promote samhd1 degradation in mdms
Several lentiviruses encode Vpx. It promotes the degradation of SAMHD1, leading to an increase in the dNTP pool and faster proviral DNA replication [bib_ref] The Aicardi-Goutieres syndrome gene and HIV-1 restriction factor SAMHD1 is a dGTP-regulated..., Powell [/bib_ref] [bib_ref] Restricted access to myeloid cells explained, Planelles [/bib_ref] [bib_ref] SAMHD1: a new insight into HIV-1 restriction in myeloid cells, St Gelais [/bib_ref]. We tested whether vaccinia virus infection promoted degradation of SAMHD1 in MDMs. Cells were pretreated with Vpx+ or Vpx2 VLPs for 24 h and were then infected with either WR or WRDTK vaccinia viruses at 1 PFU/cell. The following day, MDMs were harvested and lysed to detect SAMHD1 levels by Western blot analysis. Vpx+ VLP treatment showed a substantial decrease in the amount of SAMHD1 (85-99% reduction in protein level as compared to uninfected MDM or infected Vpx2 VLP treated MDMs [fig_ref] Figure 3: Expression of SAMHD1 in vaccinia-infected MDMs [/fig_ref] ; representative donor shown). SAMHD1 expression level data for the three donors are plotted in [fig_ref] Figure 3: Expression of SAMHD1 in vaccinia-infected MDMs [/fig_ref]. To monitor viral replication, cell lysates were incubated with a GFP antibody, which recognizes both A4-YFP and F13L-GFP (late gene protein products). Immunoblots showed that the Vpx+ VLP pretreated groups for either WR or WRDTK virus had more protein detected than the Vpx2 VLP pretreated groups [fig_ref] Figure 3: Expression of SAMHD1 in vaccinia-infected MDMs [/fig_ref] , indicating more viral genome replication had occurred in the Vpx+ VLP treatment groups. These data show that vaccinia virus does not promote SAMHD1 degradation in MDMs, and that the enhanced viral replication capability in MDMs was mediated by Vpx+ VLP treatment.
## Vaccinia infection increases dntp levels, which is further increased by vpx+ vlp treatment
Like other large dsDNA viruses, vaccinia virus encodes several dNTP biosynthesis proteins such as TK and RNR [bib_ref] Control of expression of the vaccinia virus thymidine kinase gene, Hruby [/bib_ref] [bib_ref] Vaccinia virus-encoded ribonucleotide reductase subunits are differentially required for replication and pathogenesis, Gammon [/bib_ref]. Therefore, we examined cellular dNTP concentrations in MDMs infected with vaccinia virus in the presence and absence of Vpx treatment. To do this, we used a highly sensitive HIV-1 RT-based primer extension assay [bib_ref] Macrophage tropism of HIV-1 depends on efficient cellular dNTP utilization by reverse..., Diamond [/bib_ref] to monitor changes in dNTP levels for each of the four dNTPs. In this assay [fig_ref] Figure 4: HIV-1 RT primer extension assay monitors dNTP levels [/fig_ref] , the level of the extended primer product (Primer +1) is indicative of the dNTP level. Both fold increases and absolute dNTP concentrations for dATP, dGTP, dCTP and dTTP were determined for the various treatment groups [fig_ref] Figure 4: HIV-1 RT primer extension assay monitors dNTP levels [/fig_ref]. As shown in lane 3 of [fig_ref] Figure 4: HIV-1 RT primer extension assay monitors dNTP levels [/fig_ref] , uninfected MDMs displayed very low levels of dNTPs, indicative of the low cellular dNTP level in MDMs. MDMs infected with WR [fig_ref] Figure 4: HIV-1 RT primer extension assay monitors dNTP levels [/fig_ref] showed elevated levels for dATP (8.2-fold), dGTP (6.3-fold) and dTTP (3.7-fold), while dCTP (1.0-fold) remained unchanged. Our results demonstrate that virus encoded dNTP biosynthesis machinery was able to elevate the cellular dNTP levels in MDMs. However, when we analyzed the dNTP samples extracted from MDMs infected with WRDTK virus only dGTP (7.1-fold) and dATP (4.5-fold) were elevated while dCTP (0.8-fold) and dTTP (0.4-fold) were reduced as compared to uninfected MDMs. This demonstrates that WRDTK is deficient in TK activity. These data indicate that the increased replication of WR strain as compared to WRDTK strain in MDMs is the result, in part, of increased dTTP biosynthesis.
When MDMs were pretreated with Vpx+ VLP and were infected with either WR or WRDTK viruses, all four dNTPs increased [fig_ref] Figure 4: HIV-1 RT primer extension assay monitors dNTP levels [/fig_ref] , lane 5 compared to lane 6 for WR, and lane 8 compared to lane 9 for WRDTK). Comparing Vpx+ VLP versus Vpx-VLP treated MDMs, dTTP was increased by 4.5-fold; dATP was increased by 65.8-fold and 46.7-fold; dGTP was increased by 9.0-fold and 8.2-fold; dCTP had increased by 6.9-fold and 2.8-fold for WR and WRDTK viruses, respectively. Collectively, vaccinia infection of MDMs alone only modestly elevated the cellular dNTP pools, which was further enhanced by Vpx+ VLP treatment. However, the measured dNTP pool levels for vaccinia-infected MDMs may be underestimated, since the viral DNA polymerase utilizes dNTPs in order to replicate its genome.
SAMHD1 also controls infection of HSV-1 viral replication in THP-1 cells Next we validated that SAMHD1 controls replication of another large dsDNA virus, HSV-1. For these studies, we infected a human monocytic cell line, THP-1 cells, with an HSV-1 strain (HSV-1 KOS). THP1 cells undergo differentiation to a macrophage-like cell by Phorbol 12-myristate 13-acetate (PMA) treatment. In addition, we also used PMA-treated THP-1 cell lines that stably express SAMHD1-specific shRNA (shSAMHD1) or control shRNA (shControl). As previously reported [bib_ref] The Vpx lentiviral accessory protein targets SAMHD1 for degradation in the nucleus, Hofmann [/bib_ref] , we confirmed knockdown of SAMHD1 in the differentiated THP-1 cells expressing SAMHD1 specific shRNA with HSV-1 infections at various MOIs, but not in cells expressing the control shRNA (MOI ''0'' in [fig_ref] Figure 5: SAMHD1 inhibits HSV replication [/fig_ref]. In addition, the SAMHD1 level remained unchanged even with a MOI of 1 for HSV-1 KOS in the shControl THP-1 cells, indicating that like vaccinia virus, HSV-1 does not down-regulate SAMHD1 expression. Next, we moni- To further validate that an increase in viral production occurred, differentiated THP-1 cells were pretreated with ganciclovir, a nucleoside analogue that blocks HSV-1 DNA replication. As shown in [fig_ref] Figure 5: SAMHD1 inhibits HSV replication [/fig_ref] , we detected inhibition of HSV-1 DNA replication by real-time PCR in both shControl and shSAMHD1 THP-1 cells pretreated with 20 mM ganciclovir, which is sufficient to inhibit viral replication in dividing cells [bib_ref] Anti-herpesvirus activity of the acyclic nucleoside 9-(1,3-dihydroxy-2-propoxymethyl)guanine, Smee [/bib_ref]. However, in the absence of drug, we observed a 39-fold increase in DNA replication in the shControl cells and a 737-fold increase in the shSAMHD1 THP-1 cells [fig_ref] Figure 5: SAMHD1 inhibits HSV replication [/fig_ref]. Without ganciclovir treatment, a 19-fold difference was observed between shControl cells versus shSAMHD1 THP-1 cells. These data support the notion that SAMHD1 controls the DNA replication of HSV-1 KOS in differentiated THP-1 macrophages.
## Samhd1 inhibition of hsv-1 infection requires differentiation of thp-1 cells
Next, we tested whether the SAMHD1-mediated suppression of HSV-1 KOS requires the differentiation of the THP-1 cells. For this test, THP-1 cells were infected with HSV-1 KOS at a MOI of 0.1 for 48 h in the presence of PMA (maturated, non-dividing cells) or in the absence of PMA (dividing cells). We found that the copy number of HSV-1 DNA was higher in PMA differentiated shSAMHD1 cells than in shControl cells by 488-fold [fig_ref] Figure 6: Vpx-mediated down regulation of SAMHD1 augments HSV-1 KOS infection of THP-1 cells [/fig_ref]. In contrast, without differentiation, only a 3.5-fold increase in HSV-1 KOS DNA replication was observed in shSAMHD1 cells as compared to shControl cells [fig_ref] Figure 6: Vpx-mediated down regulation of SAMHD1 augments HSV-1 KOS infection of THP-1 cells [/fig_ref]. These data suggest that SAMHD1-mediated inhibition of HSV-1 infection is dependent on the differentiation stage of the cells.
Next, we examined dATP concentration in differentiated and undifferentiated THP-1 cells that were either infected or uninfected with HSV-1 KOS virus. As shown in [fig_ref] Figure 6: Vpx-mediated down regulation of SAMHD1 augments HSV-1 KOS infection of THP-1 cells [/fig_ref] , both the uninfected and infected differentiated shSAMHD1 cells showed elevated dATP pools (3.6-fold and 4.5-fold, respectively) as compared to the shControl THP-1 cells, suggesting that HSV-1 infection alone did not promote a significant change in cellular dATP levels. Between undifferentiated shControl and shSAMHD1 THP-1 cells, no difference in dATP concentration was observed for either NI (0.9-fold) or infected (1.1-fold) groups. These data suggest that the dNTP depletion by SAMHD1 can negatively regulate the dNTP pool size even in the presence of HSV-1 dNTP biosynthesis in differentiated THP-1 cells.
## Vpx-dependent degradation of samhd1 augments hsv-1 infection in pma-differentiated thp-1 cells
Next we tested whether Vpx treatment could enhance HSV-1 replication in differentiated THP-1 cells. First, the shControl and shSAMHD1 THP-1 cells were treated with Vpx2 or Vpx+ VLPs, and the SAMHD1 level was monitored by Western blot analysis. As shown in [fig_ref] Figure 6: Vpx-mediated down regulation of SAMHD1 augments HSV-1 KOS infection of THP-1 cells [/fig_ref] , Vpx+ VLP treatment effectively degraded SAMHD1 in the shControl THP-1 cells, while shSAMHD1 THP-1 cells showed no detectable SAMHD1 expression regardless of Vpx+ VLP treatment. Using real-time PCR, we found that Vpx+ VLP treatment elevated HSV-1 replication by 6.5-fold for differentiated shControl THP-1 cells [fig_ref] Figure 6: Vpx-mediated down regulation of SAMHD1 augments HSV-1 KOS infection of THP-1 cells [/fig_ref] , while Vpx-VLP treatment was comparable to no VLP treatment (1-fold). No change in viral DNA copy number was detected after Vpx+ VLP treatment in the differentiated shSAMHD1 THP-1 cells [fig_ref] Figure 6: Vpx-mediated down regulation of SAMHD1 augments HSV-1 KOS infection of THP-1 cells [/fig_ref] , which is likely due to nearly complete inhibition of SAMHD1 by the shRNA. We observed a 1.7-fold increase in PFUs for the Vpx+ VLP treated shControl THP-1 cells as compared to NI control cells [fig_ref] Figure 6: Vpx-mediated down regulation of SAMHD1 augments HSV-1 KOS infection of THP-1 cells [/fig_ref] and a slight decrease in PFUs for VLP treatment shSAMHD1 groups. Collectively, these data show that Vpx+ VLP treatment can enhance both the viral genome copy number and PFUs in differentiated shControl THP-1 cells, while have no significant increase in these two parameters in the differentiated shSAMHD1 THP-1 cells, which already have greatly diminished SAMHD1 protein level.
## Vpx promotes hsv-1 replication in primary human dcs and mdms
HSV-1 also productively replicates in primary DCs, which are non-dividing cells [bib_ref] The virion host shutoff protein of herpes simplex virus 1 blocks the..., Cotter [/bib_ref] [bib_ref] Herpes simplex virus type 1 induces CD83 degradation in mature dendritic cells..., Kummer [/bib_ref]. We recently reported that Vpx enhances HIV-1 replication in DCs by targeting SAMHD1 for degradation [bib_ref] SAMHD1 restricts HIV-1 infection in dendritic cells (DCs) by dNTP depletion, but..., St Gelais [/bib_ref]. Thus, we investigated whether the degradation of SAMHD1 also augments HSV-1 infection in human primary DCs. Cells were infected with HSV-1 KOS after the addition of Vpx2 or Vpx+ VLPs. Similarly to PMA-differentiated THP-1 cells, the addition of Vpx+ VLPs caused a decrease in SAMHD1 protein levels [fig_ref] Figure 7: SAMHD1 augments HSV-1 infection in human primary DCs and MDMs [/fig_ref] , and enhanced HSV-1 KOS infection by 10.4-fold as compared to primary DCs treated with Vpx2 VLPs [fig_ref] Figure 7: SAMHD1 augments HSV-1 infection in human primary DCs and MDMs [/fig_ref].
Finally, in order to generalize the interplay between SAMHD1 and HSV-1 in macrophages, we employed another HSV-1 strain, HSV-1 F strain [bib_ref] Regulation of alpha genes of herpes simplex virus: expression of chimeric genes..., Post [/bib_ref] and conducted studies using MDMs infected at 0.5 PFU/cell. Vpx+ VLP treatment of MDMs demonstrated 1.7-fold and 5.2-fold enhancement at 48 and 72 hpi, respectively, for HSV-1 (F) infection by intracellular straining of the viral gB protein [fig_ref] Figure 7: SAMHD1 augments HSV-1 infection in human primary DCs and MDMs [/fig_ref]. Collectively, these data show that SAMHD1 suppresses HSV-1 replication in primary human DCs and MDMs, and that SAMHD1 degradation promotes HSV-1 replication in these two mature myeloid cell types.
# Discussion
Several studies have shown the necessity of both innate and adaptive immune responses in order to control both herpes and poxvirus infections [bib_ref] Innate immune recognition of poxviral vaccine vectors, Lousberg [/bib_ref] [bib_ref] Vaccinia viruses: vaccines against smallpox and vectors against infectious diseases and tumors, Walsh [/bib_ref] [bib_ref] HSV-1 infection through inhibitory receptor, Satoh [/bib_ref] [bib_ref] How vaccinia virus has evolved to subvert the host immune response, Bahar [/bib_ref]. Macrophages are part of the innate immune response and are rapidly recruited to sites of infection. It is therefore not surprising that these cells have developed strategies to limit pathogen replication within them. A series of recent studies highlighted SAMHD1, a cellular dNTP triphosphohydrolase, as an anti-lentiviral restriction factor predominantly found in myeloid cells [bib_ref] Vpx relieves inhibition of HIV-1 infection of macrophages mediated by the SAMHD1..., Hrecka [/bib_ref] [bib_ref] Tight interplay among SAMHD1 protein level, cellular dNTP levels, and HIV-1 proviral..., Kim [/bib_ref] [bib_ref] SAMHD1 is the dendritic-and myeloid-cell-specific HIV-1 restriction factor counteracted by Vpx, Laguette [/bib_ref] [bib_ref] SAMHD1 restricts the replication of human immunodeficiency virus type 1 by depleting..., Lahouassa [/bib_ref] [bib_ref] Contribution of SAM and HD domains to retroviral restriction mediated by human..., White [/bib_ref] [bib_ref] SAMHD1: a new insight into HIV-1 restriction in myeloid cells, St Gelais [/bib_ref] [bib_ref] SAMHD1 Restricts HIV-1 Cell-to-Cell Transmission And Limits Immune Detection In Monocyte-Derived Dendritic..., Puigdomenech [/bib_ref]. Since dNTPs are universal substrates for cellular, parasitic, and viral DNA polymerases, it seems likely that cellular dNTP concentrations could influence the genome replication of all dsDNA viruses. We attempted to expand the role of SAMHD1 as a restriction factor against large dsDNA viruses such as vaccinia virus and HSV-1. Indeed, the results presented in this study support the theory that SAMHD1 suppresses the replication capacity of these dsDNA viruses in a similar manner observed with HIV-1 and HIV-2 lentiviruses, by depleting the cellular dNTP pools and thus limiting the substrate availability for viral DNA polymerases. Viral replication occurs in different compartments for vaccinia virus and HSV-1. Like retroviruses, vaccinia virus undergoes DNA synthesis outside the nucleus and sets up a cytoplasmic viral factory for replication of its genomic DNA [bib_ref] Nuclear localization of a double-stranded RNA-binding protein encoded by the vaccinia virus..., Yuwen [/bib_ref]. In contrast, the HSV-1 genome is replicated in the host cell nucleus [bib_ref] Chromatin dynamics during herpes simplex virus-1 lytic infection, Placek [/bib_ref] and is dependent on nuclear import [bib_ref] Nuclear delivery mechanism of herpes simplex virus type 1 genome, Liashkovich [/bib_ref]. This difference in cellular localization of viral genome replication may contribute to the complexity in understanding each virus; but clearly from our data, high cellular dNTP pools appears to be critical for efficient viral genome replication regardless of the cellular compartment in which it takes place in. It is surprising that neither vaccinia virus nor HSV has evolved to target SAMHD1 for degradation. Since cellular nucleotides are dispersed throughout the cell [bib_ref] Physiological concentrations of purines and pyrimidines, Traut [/bib_ref] , SAMHD1 can deplete dNTPs and restrict viral replication of DNA viruses regardless of their replication site within a cell [bib_ref] Tetramerization of SAMHD1 is required for biological activity and inhibition of HIV..., Yan [/bib_ref].
HSV-1 and vaccinia virus have been reported to hamper adaptive immunity upon infection of myeloid cells. HSV-1 infection of DCs promotes the down regulation of CD83 [bib_ref] Herpes simplex virus type 1 induces CD83 degradation in mature dendritic cells..., Kummer [/bib_ref] , a co-stimulatory molecule. This in turn may limit T cell activation against this pathogen. Vaccinia virus infection inhibits MHC class II presentation on the surface of macrophages [bib_ref] Vaccinia virus infection induces dendritic cell maturation but inhibits antigen presentation by..., Yao [/bib_ref] and DCs [bib_ref] Protein kinase C overexpression does not enhance immune-stimulatory surface markers of vaccinia-infected..., Huemer [/bib_ref]. Therefore, the antiviral role of SAMHD1 in myeloid cells may be to limit viral replication, in an environment of decreased antigenpresentation by myeloid cells.
Interestingly, unlike retroviruses, these large dsDNA viruses are equipped with their own dNTP biosynthesis machinery such as TK and RNR, to aid in genome replication when dNTP concentrations are suboptimal. However, TKs from vaccinia and HSV-1 are most similar in sequence and structure to human TK1 (hTK1) [bib_ref] Biophysical characterization of vaccinia virus thymidine kinase substrate utilization, Smith [/bib_ref]. hTK1 is also cell cycle regulated [bib_ref] Hiding human thymidine kinase 1 from APC/C-mediated destruction by thymidine binding, Ke [/bib_ref] , being degraded in G0 cells, such as MDMs. Moreover, the activity of human RNR is coupled to DNA replication [bib_ref] Ribonucleotide reductase activity is coupled to DNA synthesis via proliferating cell nuclear..., Salguero [/bib_ref]. Collectively, MDMs have very low dNTPs, making them very restrictive to infection. Indeed, infection with vaccinia virus alone promoted a measurable increase in dTTP, dATP, and dGTP concentrations in MDMs [fig_ref] Figure 4: HIV-1 RT primer extension assay monitors dNTP levels [/fig_ref]. With the exception of dTTP, WRDTK virus also induced a similar increase, albeit not to the same levels. The increase in dNTPs was not enough to promote a robust infection of MDMs by vaccinia virus. Indeed, the antiviral restriction potential of SAMHD1 became greater when using WRDTK (vDJ2R/F13L-GFP) virus to infect MDMs. This observation proposes that dsDNA viruses may have evolved to encode their own viral dNTP biosynthesis enzymes to change the dNTP pool ratios in their favor, since vaccinia contains a high 66.6% AT rich genome [bib_ref] The complete DNA sequence of vaccinia virus, Goebel [/bib_ref] , whereas HSV-1 contains a 68% GC rich genome [bib_ref] High G+C Content of Herpes Simplex Virus DNA: Proposed Role in Protection..., Brown [/bib_ref]. However, both viruses have not adapted to escape from the replicative restriction pressure induced by SAMHD1 in macrophages and DCs, since they do not degrade SAMHD1 [fig_ref] Figure 3: Expression of SAMHD1 in vaccinia-infected MDMs [/fig_ref] and 5A). Collectively, our findings validate that SAMHD1 acts as a host restriction factor against large dsDNA viruses in nondividing target cells.
# Materials and methods
# Ethics statement
For vaccinia virus and HSV-1 F strain experiments, primary human primary monocytes were obtained from human buffy coats (New York Blood Services, Long Island, NY). These are preexisting materials that are publicly available, and there is no subject-identifying information associated with the cells. As such, the use of these samples does not represent human subjects research because: 1) materials were not collected specifically for this study, and 2) we are not able to identify the subjects. For HSV-1 KOS experiments, monocytes were obtained from the buffy coats of healthy volunteers with the approval of the Kyoto University ethics committee. Written informed consent forms were obtained from all participants.
## Cells
Primary human monocytes were isolated from the peripheral blood buffy coats by positive selection using MACS CD14+ beads as previously described [bib_ref] Leishmania induces survival, proliferation and elevated cellular dNTP levels in human monocytes..., Mock [/bib_ref] or by a Dynabeads Untouched Human Monocytes kit according to the manufacturer's protocol. For macrophage differentiation, monocytes were matured for seven days in RPMI medium containing 10% FCS, Pen/Strep antibiotics and 5 ng/ml human recombinant GM-CSF (R&D Systems) before use in experiments. For dendritic cell differentiation, monocytes were matured for 5 days in RPMI medium containing 10% FCS, Pen/Strep antibiotics, 100 ng/mL human recombinant GM-CSF, and 100 ng/mL human recombinant IL-4 (R&D Systems). THP-1 cells containing shControl and shSAMHD1 were kindly provided by Dr. Nathaniel R. Landau (New York University School of Medicine). THP-1 cell lines were maintained in RPMI medium with 10% FCS under 0.5 mg/mL puromycin selection before maturation with 50 nM PMA. HeLa cells were maintained with RPMI medium containing 10% FCS and Pen/Strep antibiotics. Vero cells were maintained with DMEM medium containing 10% FCS and Pen/Strep antibiotics and used to determine the PFU for the HSV-1 KOS virus. BSC-40 cells were maintained as previously described and used to determine PFU for WR (vYFP-A4) and WRDTK (vDJ2R/F13L-GFP) vaccinia viruses [bib_ref] Visualization and characterization of the intracellular movement of vaccinia virus intracellular mature..., Ward [/bib_ref].
## Viruses and infection
Vaccinia virus (VV) vYFP-A4 and vTF7.3 were kind gifts of Bernard Moss. To construct the VV WRTK-reporter virus, the F13L gene was replaced with the coding sequence of F13L-GFP in the recombinant VV vTF7.3 using a strategy that has already been described [bib_ref] Vaccinia virus F13L protein with a conserved phospholipase catalytic motif induces colocalization..., Husain [/bib_ref]. The resulting recombinant vaccinia virus, vDJ2R/F13L-GFP, has the TK gene (J2R) removed and expresses the late protein F13L fused to GFP. HSV-1 KOS was kindly provided by Dr. Sandra Weller (University of Connecticut Health Center). Wild type HSV-1 F strain was kindly provided by Dr. Bernard Roizman (University of Chicago; [bib_ref] Regulation of alpha genes of herpes simplex virus: expression of chimeric genes..., Post [/bib_ref]. For HSV-1 KOS infection experiments, THP-1 cells were stimulated with 50 ng/mL of PMA overnight, washed on day two with PBS, and replaced with fresh medium. On the third day, HSV-1 KOS adsorption was carried out for one hour at 37uC at the indicated MOI, cells were washed twice with PBS, replaced with fresh media and infection was allowed to continue for the indicated time points. Unless indicated otherwise, a MOI 0.1 was used for all HSV-1 experiments.
## Vlp generation
Six T225 flasks containing 293T cells were transfected with 40 mg of pVpx2 VLP or pVpx+ VLP (kindly provided by Drs. Florence Margottin-Goguet and Nathaniel Landau) and 20 mg of pVSVg at a ratio of 1 mg of DNA to 3 ml of polyethylenimine (1 mg/ml). The following day, medium were discarded and replaced with fresh DMEM medium (5% FBS and antibiotics). On days 2-3 after transfection, the medium was collected and replaced with fresh medium. On the day of collection, medium was centrifuged at 1200 rpm for 5 min to remove cells. Supernatant was subsequently filtered through a 0.45-mm membrane (Corning Inc.). Supernatant was overlaid on top of 5 ml of a 25% sucrose cushion (25% (w/v) sucrose, 10 mM Tris-HCl [pH 7.5], 0.1 M NaCl and 1 mM EDTA). VLPs were concentrated at 28,000 rpm for 90 min by ultracentrifugation. Supernatant was aspirated, and pellets were suspended in 600 ml of serum-free DMEM. Supernatant was centrifuged for 1 min at 14K rpm to remove debris. Aliquots (50 ml) were stored at 280uC. The p27 antigen level was determined using an ELISA kit (Advanced BioScience Laboratories, Inc., Rockville MD). A minimum of 145 ng of p27/million cells was used.
## Viral plaque assay
The procedure for the vaccinia plaque assay used has been described previously [bib_ref] Visualization and characterization of the intracellular movement of vaccinia virus intracellular mature..., Ward [/bib_ref]. Briefly, confluent monolayers of BSC-40 cells were infected with the indicated viruses. At 2 hpi, the inoculum was removed and cell monolayers were overlaid with semisolid medium. Three days after infection, cell monolayers were stained with crystal violet and imaged. For the HSV-1 plaque assay, supernatants obtained from HSV-1-infected THP-1 cells were serially diluted and then inoculated onto Vero cells seeded on a 6-well polystyrene plate. After a one hour adsorption at 37uC, cells were replaced with fresh medium containing the heavy chain of immunoglobulin G. After 3 dpi, the cells were fixed with 100% methanol for 5 min at room temperature and then stained with crystal violet solution (20% ethanol and 1 mg/ml of crystal violet) for 20 min at room temperature followed by washing with distilled water to remove the staining solution. Visible plaques were quantified.
## Primer extension assay
MDMs were lysed with 60% cold methanol. Cellular debris was cleared by 14K rpm centrifugation. Supernatant was dried using a speedvac. Pellets were resuspended in 20 ml water. Two microliters of sample were used in the primer extension assay. 59 32 P-endlabeled primer (''P''; 59-GTCCCTCTTCGGGCGCCA-39) was individually annealed to one of four different templates , and this template:primer complex was extended by HIV-1 reverse transcriptase generating one additional nucleotide extension product (''P+1'') for one of four dNTPs contained in the dNTP samples extracted form the cells. In this assay, the molar amount of the P+1 product is equal to that of each dNTP contained in the extracted samples, which allows us to calculate and compare the dNTP concentrations for the different treatments [bib_ref] Macrophage tropism of HIV-1 depends on efficient cellular dNTP utilization by reverse..., Diamond [/bib_ref].
# Western blot analysis
Samples were processed in RIPA buffer containing 1 mM DTT, 10 mM PMSF, 10 ml/ml phosphatase inhibitor (Sigma) and 10 ml/ ml protease inhibitor (Sigma). The cells were sonicated with 36, 5 second pulses, to ensure compete lysis. Cellular debris was removed by 15K rpm centrifugation for 10 min. Supernatants were stored at 280uC before use. Cell lysates were resolved on 4-12% Bis-Tris NuPAGE gels (Invitrogen) and were transferred to nitrocellulose membranes and detected as described in the Figure legends using the following antibodies: mouse anti-GFP mAb (Roche), mouse anti-SAMHD1 mAb (Abcam), mouse anti-SAMHD1 mAb (kindly provided by Dr. Oliver Schwartz, [bib_ref] SAMHD1 Restricts HIV-1 Cell-to-Cell Transmission And Limits Immune Detection In Monocyte-Derived Dendritic..., Puigdomenech [/bib_ref] , mouse anti-tubulin mAb (Sigma), mouse anti-HSV-1 ICP4 mAb (Virusys), anti-UL27 and anti-ß-actin mouse mAb. HRP and Cy-5 conjugated, anti-mouse and anti-rabbit secondary antibodies were purchased from Jackson ImmunoResearch Laboratories and Cell Signaling. HRP was detected using chemiluminescent reagents (Pierce) following the manufacturers instructions. The fluorescent and chemiluminescent signals were captured using a Kodak Image Station 4000 mmPro (Carestream Health), outfitted with appropriate filters, and a Fujifilm LAS4000. Blot was striped and reprobed for actin. Images were captured using BioRad ChemiDoc Imager.
# Facs analysis
Samples were fixed with 4% formaldehyde and data collected using an Accuri C6 flow cytometer to monitor GFP/YFP expression at 24 h post infection. MDMs infected with HSV-1 (F) were fixed for 20 min at RT and then stained with primary anti-gB antibody (ab6506). Cells were washed and stained with secondary goat anti-mouse-PE (Southern BioTech). Samples were collected using Accuri C6 flow cytometer. FCS data files were analyzed using FlowJo software (TreeStar).
## Real-time pcr
Nucleic acids were extracted by the urea lysis method as previously described [bib_ref] HIV-1 entry into quiescent primary lymphocytes: molecular analysis reveals a labile, latent..., Zack [/bib_ref]. For HSV-1 DNA quantitation, 100 ng of nucleic acids was used for real-time PCR and the HSV-1 genome copy number was calculated based on a standard curve generated using a plasmid containing the UL27 gene. Thermocycler conditions for the real-time PCR were as follows: 95uC for 10 min and 95uC for 15 s plus 60uC for 1 min for 40 cycles. The primers used for real-time PCR are as follows: HSV-1 UL27 Forward (59-TCGCCTTTCGCTACGTCAT-39), HSV-1 UL27 Reverse (59-GGTTCTTGAGCTCCTTGGTGG-39), GAPDH Forward (59-GCAAATTCCATGGCACCGT-39), and GAPDH Reverse (59-TCGCCCCACTTGATTTTGG-39).
## Graphing and statistical analysis
Prism software was used for plotting the data. All the data sets greater than three samples were first compared using ANOVA for significant differences with the means. Post-analysis was the done using Mann-Whitney test to determine significant difference between two groups. [fig_ref] Figure 4: HIV-1 RT primer extension assay monitors dNTP levels [/fig_ref] Analysis of PFUs generated by differentiated THP-1 cells. One million THP-1 cells were differentiated with 50 nM PMA overnight in 6-well dishes followed by VLP treatments. Twenty-four hours later, cells were infected with either WR or WRDTK virus. Cells and supernatant were collected 24 hpi and then analyzed for PFUs. Data was performed in replicates and plotted. (TIFF)
## Supporting information
[fig] Figure 1: Vaccinia replication in MDMs. (A) Human primary monocyte-derived macrophages (MDMs) were pretreated with VLPs: Vpx+ and Vpx2 for 24 h prior to infection with Western Reserve (WR) stain of vaccinia virus expressing the viral core protein A4 fused to YFP. A representative FACS dot blot shows the percent of infected MDMs at 24 hpi. (B) The means of three independent donors were graphed and are displayed as the percentages of YFP+ MDMs for the different VLPs treatment groups. (C) Plaque forming units (PFUs) were determined for three donors as described in Experimental Procedures. The fold changes between Vpx+ VLP and Vpx2 VLP groups are displayed. doi:10.1371/journal.ppat.1003481.g001 [/fig]
[fig] Figure 2: Analysis of WRDTK replication in MDMs. (A) MDMs pretreated with Vpx+ or Vpx2 VLPs for 24 h and then infected with a thymidine kinase deficient vaccinia virus expressing the late viral F13 protein fused to GFP (WRDTK). (A) The percentage of GFP+ MDMs of three independent donors was determined using FACS, and the fold change between Vpx+ VLP and Vpx2 VLP treatment was plotted. (B) The means of three independent donors were graphed showing the percentage of GFP+ MDMs for the different VLPs treatment groups. (C) PFUs were determined and plotted for the three donors infected with the WRDTK virus. Fold changes were plotted comparing the Vpx+ VLP versus Vpx2 VLP groups. doi:10.1371/journal.ppat.1003481.g002 enhancement of infection in the absence of the viral encoded TK gene. [/fig]
[fig] Figure 3: Expression of SAMHD1 in vaccinia-infected MDMs. (A) MDMs treated with Vpx+ or Vpx2 VLP were infected (1 PFU/cell) with either WR or WRDTK virus (NI: no infection). Cell lysates were generated 24 hpi and analyzed by Western blot using a-SAMHD1, a-actin, or a-GFP antibodies. (A) Representative immunoblots from one donor are shown. Immunoblots for SAMHD1 (B) and GFP (C) were quantified for three independent donors. Data were normalized to the amount of actin signal and were plotted for the amount of signal in the NI lane (set to 1.0) for SAMHD1 and the Vpx2 VLP lane (set to 1.0) for GFP. doi:10.1371/journal.ppat.1003481.g003 [/fig]
[fig] Figure 4: HIV-1 RT primer extension assay monitors dNTP levels. (A) MDM sample extracts were analyzed using a dNTP assay. 59 32 P-endlabeled primer (''P'') was individually annealed to one of four different templates. The molar amount of the nucleotide extension product (''P+1'') is equal to that of each dNTP contained in the extracted samples, which allows us to calculate and compare the dNTP concentrations for the different treatments[13]. Representative gels are shown for one of three MDM donors monitoring the levels of dNTPs under various conditions. The lane key indicates the different treatment conditions used for each lane. (B) Quantitation of the raw data was plotted as bar graphs for each of the dNTP concentrations. Fold changes in dNTPs between different MDMs treatment groups were plotted. doi:10.1371/journal.ppat.1003481.g004 [/fig]
[fig] Figure 5: SAMHD1 inhibits HSV replication. (A) Western blot analysis of lysates obtained from PMA-stimulated THP-1 shControl and shSAMHD1 cells, infected with varying MOIs of HSV-1 KOS, was performed to determine the levels of SAMHD1, the HSV-1 viral proteins UL-27 and ICP4 (denoted by asterisks), and a tubulin loading control. (B) Real-time PCR was carried out to determine HSV-1 DNA levels after HSV-1 KOS infection in PMAstimulated shControl (solid line) and shSAMHD1 (dashed line) THP-1 cells at an MOIs of 0.01 over 72 h. (C) PFUs were measured for THP-1 cells infected at an MOI of 0.01. (D) Real-time PCR measured HSV-1 DNA levels for THP-1 cells infected at an MOI of 0.1. (E) PFUs were measured for THP-1 cells infected at an MOI of 0.1. (F) HSV-1 DNA levels were determined after HSV-1 infection of PMA-stimulated shControl and shSAMHD1 THP-1 cells in the presence or absence of 20 mM ganciclovir. The results are representative of two independent experiments performed in duplicate. doi:10.1371/journal.ppat.1003481.g005 [/fig]
[fig] Figure 6: Vpx-mediated down regulation of SAMHD1 augments HSV-1 KOS infection of THP-1 cells. (A and B) shControl and shSAMHD1 THP-1 cells were infected by HSV-1 (KOS) at a MOI of 0.1 under non-dividing or dividing conditions for 48 h. These results are representative data of three independent experiments performed in duplicate. (A) HSV-1 DNA copy numbers were analyzed by real-time PCR and (B) the level of dATP in the cells was determined by the HIV-1 RT primer extension assay as shown in Fig. 4. (C) PMA-stimulated shControl and shSAMHD1 THP-1 cells were inoculated with medium containing vehicle, Vpx-VLP, or Vpx+ VLP for 24 h and infected with HSV-1 WT virus at a MOI of 0.1. Cell lysates were generated 48 hpi and protein levels of SAMHD1 and tubulin were determined by western blot analysis. The results are representative of three independent experiments performed in duplicate. (D) Real-time PCR was used to monitor HSV-1 DNA levels. (E) PFUs for shControl and shSAMHD1 THP-1 cells were measured for the different VLP treatment groups. Fold changes between the different groups were determined and displayed. doi:10.1371/journal.ppat.1003481.g006 [/fig]
[fig] Figure 7: SAMHD1 augments HSV-1 infection in human primary DCs and MDMs. Primary mature DCs were inoculated with medium containing vehicle, Vpx2 VLP, or Vpx+ VLP for 48 h and then infected with HSV-1 KOS at a MOI of 0.1. Cell lysates were generated at 48 hpi and (A) western blot analysis for SAMHD1 and tubulin were performed. (B) Real-time PCR for HSV-1 DNA levels were determined 48 hpi. (C) Primary human MDMs were pretreated with VLPs before infecting with HSV-1 F strain (0.5 PFU/cell). Infection was monitored for intracellular staining of viral gB protein by FACS analysis and the data graphed as a percentage of gB positive MDMs. doi:10.1371/journal.ppat.1003481.g007 [/fig]
[fig] Figure S1: Infection of MDMs donors with different viruses. A total of six MDMs donors were analyzed for the ability of Vpx VLP to enhance the infection of WR and WRDTK strains. Mann-Whitney test was performed and significant differences indicated with * and p,0.05. (TIFF) Figure S2 Direct comparison of WR and WRDTK viruses for one donor. FACS data were plotted for the percentage of (A) WR (YFP+) and (B) WRDTK (GFP+) cells. PFU for (C) WR and (D) WRDTK infected MDMs. The non-infected (NI) MDMs had zero PFUs detected. (E) Immunoblot analysis for SAMHD1 expression at 48 h after VLP treatment and 24 hpi. The NI MDMs had comparable SAMHD1 expression as compared to the Vpx2 VLP treated MDMs. (TIFF) Figure S3 HeLa cells were pretreated for 24 h with VLPs prior to infection with either WR or WRDTK virus (MOI of 0.5 PFU/cell). At 24 hpi, cells were analyzed for YFP (WR virus) or GFP (WRDTK virus) expression by flow cytometry. Data were normalized to 1.0 for the Vpx2 VLP treatment groups and were graphed. Data shows that HeLa cells are very permissive to vaccinia infection by both WR and WRDTK viruses. (TIFF) [/fig]
|
Cutaneous mastocytosis with a mutation in the juxtamembrane domain of c-kit in a young laboratory beagle dog
Cutaneous mastocytosis, which resembles a subset of urticaria pigmentosa in humans, is rare in dogs. We herein report unrepresentative neoplastic proliferation of mast cells in ventral skin removed routinely from a nine-month-old female laboratory beagle dog at necropsy. A histological examination revealed diffuse extensive cellular infiltration from the superficial to deep dermis in most parts of the skin around the fourth and fifth mammary papilla without nodule formation. Tumor cells were fairly monomorphic, well-differentiated mast cells with round nuclei of small distinct nucleoli and moderate to abundant, slightly eosinophilic and granular cytoplasm. A perivascular arrangement of mast cells was noted at the margin of the lesions. Infiltration of eosinophils and degeneration of collagen were not observed in the dermis. Cutaneous mastocytosis was diagnosed based on these features. A sequence analysis of lesions revealed the deletion of Gln 555 to Ile 570 within the juxtamembrane domain of c-kit (exon 11).
Cutaneous mastocytosis (CM) is a rare disease in dogs. Some dogs have been diagnosed with primary CM resembling urticarial pigmentosa (UP) in humans [bib_ref] Cutaneous mastocytosis in a dog, Davis [/bib_ref] [bib_ref] Urticaria pigmentosa-like disease in the dog, Jeromin [/bib_ref]. We herein report unrepresentative neoplastic proliferation of mast cells in the skin of a young laboratory beagle dog. To the best of our knowledge, this is the first report of a sequence analysis of c-kit from canine CM and we also revealed a ckit mutation in the juxtamembrane domain (exon 11).
A nine-month-old female Marshall beagle dog (Marshall BioResources, North Rose, NY, USA), which was assigned to a test-article treatment group, was sacrificed at the end of a three-month repeated-dose oral toxicity study. Cutaneous lesions were considered to have developed spontaneously because no similar lesions were found in the other dogs given the compound and no pharmacological or biological activity was found for mast cells by a panel assay. The dog had been housed in a commodious cage in an environmentally controlled room (room temperature, 23 ± 3°C; relative humidity, 30-70%; lighting cycle, 12-h light/dark) and supplied expanded food once daily and tap water ad libitum. All experimental procedures were conducted after approval for the study was obtained from the Animal Care and Use Committee of Shionogi Research Laboratories.
Macroscopic changes were not detected in any organs including the skin. The skin on the right side around the fourth to fifth mammary papilla of the dog was routinely fixed with 10% neutral buffered formalin, processed, and embedded in paraffin. Paraffin-embedded sections were then stained with hematoxylin and eosin (HE) or toluidine blue stain. A CD117 antibody (1:50; Dako, Glostrup, Denmark) and antibodies against histamine (1:400; Merck Millipore, Billerica, MA, US), mast cell tryptase (1:100; AbD Serotec, Oxford, UK), and Ki-67 (clone MIB-1; 1:100; Dako, Glostrup, Denmark) were selected for an immunohistochemical study. Heat-induced antigen retrieval was performed with citrate buffer for histamine, mast cell tryptase, and MIB-1 antibodies or with ethylenediaminetetraacetate (pH=8) for the CD117 antibody.
DNA was isolated from a formalin-fixed, paraffin-embedded skin sample. Polymerase chain reaction (PCR) amplification and a sequence analysis of c-kit was performed for the entire coding regions of c-kit exons 8, 11, 17, and 18 and for some of the region (1445-1537) of exon 9. As a sequence analysis of c-kit from canine CM has never been reported, we used the findings from canine mast cell tumor and human CM as a reference for selection of the analysis region for c-kit.
Microscopically, diffuse extensive cellular infiltration was observed from the superficial (nearly the epidermal-dermal junction) to deep dermis in most parts of the skin around the fourth and fifth mammary papilla without nodule formation [fig_ref] Figure 1: Histological appearance of cutaneous mastocytosis [/fig_ref]. Tumor cells were fairly monomorphic, well-differentiated mast cells with round nuclei of small distinct nucleoli and moderate to abundant, slightly eosinophilic and granular cytoplasm [fig_ref] Figure 1: Histological appearance of cutaneous mastocytosis [/fig_ref]. A perivascular arrangement of mast cells was noted at the margin of the lesions, and these neoplastic cells were slightly small, like normal mast cells [fig_ref] Figure 1: Histological appearance of cutaneous mastocytosis [/fig_ref]. Mitotic figures were rare, and the number of MIB-1-positive mast cells observed was smaller than that of cutaneous basal cells. Infiltration of eosinophils and collagen degeneration was not observed in the dermis. Furthermore, infiltration of mast cells was not present in any organs or tissues including bone marrow and peripheral blood, except for the affected skin. Toluidine blue staining revealed the typical metachromatic characteristics of cytoplasmic granules [fig_ref] Figure 1: Histological appearance of cutaneous mastocytosis [/fig_ref]. Infiltrated mast cells were positive for CD117, histamine, and mast cell tryptase [fig_ref] Figure 1: Histological appearance of cutaneous mastocytosis [/fig_ref].
A sequence analysis of c-kit revealed the deletion of Gln 555 to Ile 570 within the protein sequence of the juxtamembrane domain in exon 11 (c.1663_1710del) [fig_ref] Figure 2: Sequence of the canine c-kit gene in exon 11 [/fig_ref]. No mutation was found in exon 8, 9, 17, or 18 in this case.
In children during infancy or early childhood, the skin lesions of UP, the most common form of CM, exhibit typical clinical findings including lesions that urticate when stroked (Darier's sign) and also show intraepidermal accumulation of melanin pigment, which is histopathological proof of abnormal mast cell infiltration of the dermis 3 . Canine CM was previously reported to be similar to UP in children 1, 2 . Dogs present with partially erythematous small papules to plaques on the head, neck, trunk, perineum, and legs 4 . Affected dogs are mostly under 1 year of age 1, 2, 4 . Histopathologically, the lesions are characterized by moderate to severe perivascular to diffuse infiltration of well-differentiated mast cells in the dermis 1, 2, 4 . Similar to UP in children, early onset CM in dogs may regress spontaneously; however, progression from CM to systemic mastocytosis has also been reported [bib_ref] Cutaneous mastocytosis in a dog, Davis [/bib_ref] [bib_ref] Urticaria pigmentosa-like disease in the dog, Jeromin [/bib_ref] [bib_ref] Mast cell tumors, Thelma [/bib_ref].
The diagnosis of CM in this case was based on the following characteristics: lack of typical clinical findings, absence of cutaneous nodules, the afflicted dog's young age, the apparent restriction of mast cell infiltration to the skin, diffuse extensive aggregation of well-differentiated mast cells from the superficial (nearly the epidermal-dermal junction) to deep dermis, slightly small mast cells infiltrated into the marginal zone of the lesions (perivascular arrangement), and no infiltration of eosinophils or collagen degeneration. We considered this clinical information and the histopathological features to be consistent with previously reported case of human and canine CM rather than mast cell tumors; however, it was difficult to differentiate its features from those of early stage of mast cell tumors because the lesions lacked the typical clinical findings of CM.
KIT is a type III receptor tyrosine kinase encoded by the c-kit gene, which includes an extracellular domain (encoded by exons 1-9), transmembrane domain (exon 10), and intracellular domain (exons 11-21) 5, 6 . The intracellular domain is further divided into a negative regulatory juxtamembrane domain (exons 11 and 12) and cytoplasmic tyrosine kinase domain that is spilt by an insert into ATP-binding (exon 13) and phosphotransferase lobes (exon 17) [bib_ref] Gain-of-function mutations in the extracellular domain of KIT are common in canine..., Letard [/bib_ref]. Mutations in the kinase domain or neighboring juxtamembrane domain have been shown to cause constitutive activation of c-kit in the absence of ligand binding, and they may result in the development and/or proliferation of certain canine mast cell tumors [bib_ref] Imatinib responsiveness in canine mast cell tumors carrying novel mutations of c-KIT..., Nakano [/bib_ref] [bib_ref] Constitutively activating mutations of c-kit receptor tyrosine kinase confer factor-independent growth and..., Kitayama [/bib_ref] [bib_ref] Mutations in the juxtamembrane domain of c-KIT are associated with higher grade..., Zemke [/bib_ref] [bib_ref] Clustering of activating mutations in c-KIT's juxtamembrane coding region in canine mast..., Ma [/bib_ref]. For example, the most common somatic mutation in human (especially adult) CM, Asp816Val, is located in the kinase domain of KIT in exon 17 and results in augmented mast cell proliferation [bib_ref] Mastocytosis. In: WHO classification of tumours of haematopoietic and lymphoid tissues, Horny [/bib_ref]. Previous studies on canine mast cell tumors identified mutations in exons [bib_ref] Mutations in the juxtamembrane domain of c-KIT are associated with higher grade..., Zemke [/bib_ref] [bib_ref] Clustering of activating mutations in c-KIT's juxtamembrane coding region in canine mast..., Ma [/bib_ref] , and 17 of KIT [bib_ref] Gain-of-function mutations in the extracellular domain of KIT are common in canine..., Letard [/bib_ref]. Most alterations within exon 11 were internal tandem duplications (ITDs) located in the distal part between residues 571 and 590, while other types, such as point mutations, deletions, and insertions in the proximal part between residues 553-562, have also been reported [bib_ref] Imatinib responsiveness in canine mast cell tumors carrying novel mutations of c-KIT..., Nakano [/bib_ref] [bib_ref] Gain-of-function mutations in the extracellular domain of KIT are common in canine..., Letard [/bib_ref] [bib_ref] Mutations in the juxtamembrane domain of c-KIT are associated with higher grade..., Zemke [/bib_ref]. In the present study, deletion of Gln 555 to Ile 570 was observed in the proximal part of exon 11, and this may have resulted in augmented mast cell proliferation. Imatinib, a small molecule tyrosine kinase inhibitor, has been reported to be effective for treatment of canine mast cell tumors with c-kit mutations in exon 11; therefore, our case report indicates that imatinib is possibly effective for canine CM (in small animal practice) 5, 10 .
[fig] Figure 1: Histological appearance of cutaneous mastocytosis. Hematoxylin and eosin staining (A-D). Toluidine blue staining (E). Immunostaining of mast cell tryptase (F). Diffuse extensive cellular infiltration of well-differentiated mast cells from the superficial to deep dermis in most parts of the skin around the fourth and fifth mammary papilla without nodule formation (A-B). Tumor cells were fairly monomorphic, well-differentiated mast cells with round nuclei of small distinct nucleoli and moderate to abundant, slightly eosinophilic and granular cytoplasm (C). A perivascular arrangement of mast cells was noted at the margin of the lesions, and these neoplastic cells were slightly small, like normal mast cells (D). Mast cells that infiltrated the dermis had cytoplasmic granules with metachromatic staining (E) and were immunopositive for mast cell tryptase (F). Bars: 1 mm (A), 500 μm (B), 100 μm (D-F), 50 μm (C). [/fig]
[fig] Figure 2: Sequence of the canine c-kit gene in exon 11. The deletion of Gln 555 to Ile 570 (c.1663_1710 del) in this case is denoted by a dotted dashed line. [/fig]
|
Autophagy-independent enhancing effects of Beclin 1 on cytotoxicity of ovarian cancer cells mediated by proteasome inhibitors
Background: The ubiquitin-proteasome system and macroautophagy (hereafter referred to autophagy) are two complementary pathways for protein degradation. Emerging evidence suggests that proteasome inhibition might be a promising approach for tumor therapy. Accumulating data suggest that autophagy is activated as a compensatory mechanism upon proteasome activity is impaired.Method: Autophagy activation was measured using acridine orange staining and LC3 transition. Cell viability and apoptosis were measured using MTT assay and flow cytometry, respectively. Beclin 1 expression vectors or shRNA against Beclin 1 (shBeclin 1) were transfected to investigate the role of Beclin 1 in autophagy activation and cytotoxicity of ovarian cancer cells induced by proteasome inhibitors. Results: Proteasome inhibitors suppressed proliferation and induced autophagy in ovarian cancer cells. Neither phosphoinositide 3-kinase (PI3K) inhibitors nor shRNA against Beclin 1 could abolish the formation of acidic vacuoles and the processing of LC3 induced by proteasome inhibitors. Moreover, Beclin 1 overexpression enhanced anti-proliferative effects of proteasome inhibitors in ovarian cancer cells.Conclusions: For the first time, the current study demonstrated that proteasome inhibitors induced PI3K and Beclin 1-independent autophagy in ovarian cancer cells. In addition, this study revealed autophagy-independent tumor suppressive effects of Beclin 1 in ovarian cancer cells.
# Background
The ubiquitin-proteasome system serves as a major intracellular pathway for protein degradation in mammalian cells [bib_ref] Protein degradation and protection against misfolded or damaged proteins, Goldberg [/bib_ref]. Many proteins involved in cancer cell growth and survival are regulated by proteasomal degradation [bib_ref] The proteasome: a suitable antineoplastic target, Adams [/bib_ref]. In this connection, proteasome inhibitors constitute a novel class of anti-tumor agents with pre-clinical and clinical evidence of activity against hematologic malignancies and solid tumors [bib_ref] Bortezomib: proteasome inhibition as an effective anticancer therapy, Richardson [/bib_ref]. Macroautophagy (hereafter is referred as autophagy) is an evolutionarily conserved catabolic process by which cell destructs its cytoplasmic content and organelles through the lysosomal machinery [bib_ref] Autophagy fights disease through cellular self-digestion, Mizushima [/bib_ref]. Autophagy is initiated by the formation of a double-membrane bound vacuole (autophagosome), which sequesters cytosolic proteins and organelles such as mitochondria, endoplasmic reticulum. Autophagosomes are short-lived organelles that fuse with acidic lysosomes to produce autolysosomes where the sequestered content is degraded by lysosomal enzymes, and amino acids and sugars are recycled into the cytosol for reuse. Morphologically, autophagy is characterized by the formation of LC3+ double-membrane bound autophagosomes, the accumulation of acidic vesicular organelles and autolysosomes in the cytoplasm [bib_ref] LC3, a mammalian homologue of yeast Apg8p, is localized in autophagosome membranes..., Kabeya [/bib_ref] [bib_ref] A novel response of cancer cells to radiation involves autophagy and formation..., Paglin [/bib_ref] [bib_ref] A novel assay to study autophagy: regulation of autophagosome vacuole size by..., Munafo [/bib_ref]. Autophagy was originally recognized as a crucial prosurvival mechanism to supply the cell with nutrients under unfavorable grown conditions [bib_ref] Autophagy fights disease through cellular self-digestion, Mizushima [/bib_ref]. It is now clear that autophagy plays a crucial role in development, programmed cell death and aging [bib_ref] Autophagy fights disease through cellular self-digestion, Mizushima [/bib_ref] [bib_ref] Autophagy in the pathogenesis of disease, Levine [/bib_ref] [bib_ref] Regulation of the aging process by autophagy, Salminen [/bib_ref] [bib_ref] Role and regulation of autophagy in cancer, Chen [/bib_ref]. Dysregulation of autophagy has been involved in many human diseases including cancers. The fact that autophagy can have both suppressive and promoting roles in carcinogenesis makes it an attractive target in cancer research [bib_ref] Role and regulation of autophagy in cancer, Chen [/bib_ref]. As a tumor suppressing mechanism, autophagy serves as an alternative to apoptosis to eliminate transformed cells [bib_ref] Autophagy fights disease through cellular self-digestion, Mizushima [/bib_ref]. Moreover, tumorigenesis is often associated with a reduced autophagy while genes that are involved in the execution of autophagy are found to be tumor suppressors [bib_ref] Autophagy fights disease through cellular self-digestion, Mizushima [/bib_ref]. On the other hand, autophagy may facilitate tumor growth and survival by providing tumor cells a selective advantage to therapy resistance and aggressiveness [bib_ref] Autophagy fights disease through cellular self-digestion, Mizushima [/bib_ref] [bib_ref] Role and regulation of autophagy in cancer, Chen [/bib_ref]. As two important intracellular pathways for protein degradation in mammalian cells, autophagy functions complementarily with the ubiquitin-proteasome system [bib_ref] Protein degradation and protection against misfolded or damaged proteins, Goldberg [/bib_ref] [bib_ref] The roles of intracellular protein-degradation pathways in neurodegeneration, Rubinsztein [/bib_ref] , and suppression of UPS can activate autophagy [bib_ref] The role of autophagy-lysosome pathway in neurodegeneration associated with Parkinson's disease, Pan [/bib_ref] [bib_ref] Autophagy induction rescues toxicity mediated by proteasome inhibition, Rubinsztein [/bib_ref] [bib_ref] Linking of autophagy to ubiquitin-proteasome system is important for the regulation of..., Ding [/bib_ref] [bib_ref] The formation of peripheral myelin protein 22 aggregates is hindered by the..., Fortun [/bib_ref] [bib_ref] The role of ATF4 stabilization and autophagy in resistance of breast cancer..., Milani [/bib_ref] [bib_ref] Induction of autophagy by proteasome inhibitor is associated with proliferative arrest in..., Wu [/bib_ref] [bib_ref] Role of the aggresome pathway in cancer: targeting histone deacetylase 6-dependent protein..., Rodriguez-Gonzalez [/bib_ref] [bib_ref] Apoptotic and autophagic cell death induced by histone deacetylase inhibitors, Shao [/bib_ref] [bib_ref] Domainselective small-molecule inhibitor of histone deacetylase 6 (HDAC6)-mediated tubulin deacetylation, Haggarty [/bib_ref]. Emerging evidence shows that autophagy is important in the regulation of cancer development and progression [bib_ref] Role and regulation of autophagy in cancer, Chen [/bib_ref]. However, the role of autophagy is complicated and autophagy may have opposing consequences in cells. On one hand, autophagy may protect tumor cells from nutrient deprivation and hypoxia; on the other hand, autophagy defect is associated with the development of cancer [bib_ref] Autophagy in the pathogenesis of disease, Levine [/bib_ref] [bib_ref] Defective autophagy leads to cancer, Edinger [/bib_ref].
Beclin 1 is a tumor suppressor gene product that allosterically activates the class III phosphatidylinositol 3kinase (PI3KC3), which is essential for the recruitment of other autophagy-related gene (Atg) proteins to the phagophore assembly site (PAS) to initiate autophagosome formation [bib_ref] Induction of autophagy and inhibition of tumorigenesis by beclin 1, Liang [/bib_ref] [bib_ref] Beclin-phosphatidylinositol 3-kinase complex functions at the trans-Golgi network, Kihara [/bib_ref]. The BH3 binding groove of Bcl-XL/Bcl-2 binds the BH3 helix of Beclin1, preventing Beclin1 from recruitment to the PI3KC3 complex [bib_ref] Crystal structure of the Bcl-XL-Beclin 1 peptide complex: Beclin 1 is a..., Oberstein [/bib_ref] [bib_ref] Molecular basis of Bcl-xL's target recognition versatility revealed by the structure of..., Feng [/bib_ref]. Recently, accumulating studies suggest that autophagy can also occur in a Beclin1-independent manner and in this case PI3K inhibitors fails to suppress it [bib_ref] Role of non-canonical Beclin 1-independent autophagy in cell death induced by resveratrol..., Scarlatti [/bib_ref] [bib_ref] Beclin 1-independent autophagy induced by a Bcl-XL/Bcl-2 targeting compound, Z18, Tian [/bib_ref] [bib_ref] Regulation of autophagy by extracellular signal-regulated protein kinases during 1-methyl-4-phenylpyridinium-induced cell death, Zhu [/bib_ref] [bib_ref] Beclin 1-independent pathway of damageinduced mitophagy and autophagic stress: implications for neurodegeneration..., Chu [/bib_ref] [bib_ref] Non-canonical autophagy: an exception or an underestimated form of autophagy, Scarlatti [/bib_ref].
Here we reported that proteasome inhibitors induced cell death and autophagy in ovarian cancer cells. It was of note that MG132-induced autophagy was accompanied by a reduction of Beclin 1. In addition, we reported that proteasome inhibitors elicited autophagy even in shRNA against Beclin 1 (shBeclin 1) transfected cells, or in the presence of PI3Ks inhibitors, indicating that proteasome inhibitors caused Beclin 1/PI3Ks-independent autophagy. Furthermore, we demonstrated that Beclin 1 overexpression enhanced proteasome inhibitors-mediated cell death of ovarian cancer cells. Collectively, these data suggested that Beclin 1 sensitized ovarian cancer cells to proteasome inhibitors in an autophagy-independent manner.
# Methods
Culture of multiple cancer cell lines SKOV3, OVCAR3 and A2870 ovarian cancer cell lines were maintained in DMEM (Sigma-Aldrich, Saint Louis, MO) supplemented with 10% fetal bovine serum (FBS, Sigma-Aldrich, Saint Louis, MO).
## Chemicals
MG132, epoxomicin, PSI and lactacystin were purchased from Calbiochem (La Jolla, CA). 0.02% DMSO was used as vehicle control.
## Cell viability assays
For cell viability assays, cells were plated in 96-well dishes (1 × 10 4 cells per well) and the next day were treated with or without apoptosis inducing agents in 10% FBS-containing media and grown over a 24-h period. Cell viability was assessed using the 3-(4,5-dimethylthiazol-2thiazolyl)-2,5-diphenyl tetrazolium bromide (MTT) assay (Chemicon, Bedford, MA) according to the manufacturer's instruction.
## Detection of apoptotic cell death
For cell death assays, cells were washed twice in phosphate-buffered saline and then stained with Annexin V-FITC (Biovision, Mountainview, CA) and propidium iodide (PI, Sigma-Aldrich) according to the manufacturer's instructions. After staining with Annexin V-FITC and PI, samples were analyzed by fluorescence-activated cell scanner (FACScan) flow cytometer (Becton Dickinson, Franklin Lakes, NJ).
## Acridine orange staining for acidic vesicular organelles
Acridine orange was added at a final concentration of 1μg/ml for a period of 15 min. Pictures were obtained with a fluorescence microscope (Olympus) equipped with a digital camera (Olympus).
# Western blot analysis
Cells were lysed in lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 2 mM EDTA, 1% Triton-X100 and protease inhibitor cocktail (Sigma-Aldrich, Saint Louis, MO). Cell extract protein amounts were quantified using the BSA (See figure on previous page.) [fig_ref] Figure 1: See legend on next page [/fig_ref] Growth inhibition and apoptosis of ovarian cancer cells induced by proteasome inhibitors. A, SKOV3, OVCAR3 or A2870 cells were treated with the indicated concentrations of MG132 for 24 h, and cell viability was measured using MTT assay. B, SKOV3, OVCAR3 or A2870 cells were treated with 5 μM of MG132 for 24 h, and nuclear morphology was analyzed using Hoechst 33258 staining. C, SKOV3, OVCAR3 or A2870 cells were treated with vehicle, bortezomib (BZ), epoxomicin (Epox), lactacystin (Lacta), or MG132 for 24 h, and Western blot analysis was performed using the indicated antibodies. D, SKOV3, OVCAR3 or A2870 cells were treated with the indicated concentrations of MG132 for 24 h, and apoptotic cells were measured using Annexin V and PI double staining followed by flow cytometry. *, P<0.01.
protein assay kit. Equivalent amounts of protein (25 μg) were separated using 12% SDS-PAGE and transferred to PVDF membrane (Millipore Corporation, Billerica, MA).
## Caspase-3 activity assay
For caspases-3 enzymatic assays, 50 μg whole-cell extract was added to reaction buffer containing 25 mm HEPES (pH 7.5), 4 mm CHAPS, 1mm dithiothreitol (DTT), 1 mm phenylmethylsulfonyl fluoride (PMSF), 2 μg/ml aprotinin, 1 μg/ml leupeptin, and 2 μg/ml pepstatin, to achieve a total reaction volume of 500 μl. Ac-DEVD-AMC (Ac-Asp-Glu-Val-Asp-7-amino-4-methylcoumarin;
Alexis Biochemicals, San Diego, CA) was added to the mixture at a concentration of 100 μM and incubated for 1 h at 37°C. Cleavage of the substrate was measured by fluorescence spectrometer (HTS 7000; PerkinElmer, Boston, MA) using an excitation and emission wavelength of 360 and 465 nm, respectively. The activities were expressed as fluorescence increase per microgram of protein.
## Dna construction and transfection
Beclin 1 plasmid was constructed by PCR and cloned into pcDNA3.1 vector. The construct was verified by Induction of autophagy by proteasome inhibitors in ovarian cancer cells. A, SKOV3, OVCAR3 or A2870 cells were treated with vehicle or 5 μM of MG132 for 24 h, and the formation of acidic vacuoles were analyzed using AO staining. B, SKOV3, OVCAR3 or A2870 cells were treated with the indicated concentrations of MG132 for 24 h, Western blot analysis was performed to investigate LC3 transition. C, OVCAR3 were treated with vehicle, BZ, Epox, Lacta or MG132 in the absence or presence of bafiloymycin A1 for 24 h, and LC3 transition was measured using Western blot analysis. D, OVCAR3 were transfected with a scramble shRNA or shRNA specific against Atg7 (shAtg7) for 24 h, then treated with 5 μM of MG132 for additional 24 h, and Western blot was performed.
DNA sequencing. Short hairpin RNA (shRNA) against Beclin 1 (shBeclin 1) or Atg7 (shAtg7) was purchased from Open Biosystems. Cells were transfected with Lipofectamine 2000 reagent (Invitrogene) as instructed by the supplier.
## Statistics
The statistical significance of the difference was analyzed by ANOVA and post hoc Dunnett's test. Statistical significance was defined as p<0.05. All experiments were repeated three times, and data were expressed as the mean±SD (standard deviation) from a representative experiment.
# Results
## Proteasome inhibitors inhibited proliferation and induced apoptosis in ovarian cancer cells
To study the effect of blockade of ubiquitin-proteasome system on proliferation of ovarian cancer cells, SKOV3, OVCAR3 and A2870 cells were treated with proteasome inhibitor MG132 at concentrations ranging from 0 to 10 μM for 24 h, the cell viability was determined using the MTT assay. MG132 significantly reduced cell proliferation in these cell lines in a concentrationdependent manner [fig_ref] Figure 1: See legend on next page [/fig_ref]. To determine the incidence of apoptosis morphologically, we stained the nuclei of 5 μM MG132 treated SKOV3, OVCAR3 and A2870 cells with Hoechst 33258. Apoptotic morphological characteristics such as chromatin condensation and nuclear fragmentation were detected in these ovarian cancer cells treated with 5 μM of MG132 [fig_ref] Figure 1: See legend on next page [/fig_ref]. Western blot confirmed that proteasome inhibitors including MG132, epoxomicin (Epox), Lactacystin (Lacta) and bortezomib (BZ) elicited cleavage of PARP in SKOV3, OVCAR3 and A2870 cells [fig_ref] Figure 1: See legend on next page [/fig_ref]. Annexin V-FITC and PI double staining followed by flowcytometry also confirmed that 5 μM of MG132 caused apoptosis of SKOV3, OVCAR3 and A2870 cells [fig_ref] Figure 1: See legend on next page [/fig_ref].
## Proteasome inhibitors induced autophagy in ovarian cancer cells
Under the light microscope, it was apparent that MG132 induced the formation of large vacuoles in the cytoplasm of ovarian cancer cells (data not shown). To determine the effect of MG132 on autophagy, we analyzed the accumulation of acidic vesicular organelles using the acridine orange (AO) staining. AO emitted bright red fluorescence in acidic vesicles but fluoresced green in cytoplasm and nucleus [bib_ref] A novel response of cancer cells to radiation involves autophagy and formation..., Paglin [/bib_ref]. Vital staining of SKOV3, OVCAR3 or A2870 cells with AO revealed the appearance of acidic vesicular organelles after 5 μM of MG132 treatment . Since the conversion of LC3 protein from LC3-I (the cytosolic form) to LC3-II (the membrane bound form) correlates with the extent of autophagy [bib_ref] LC3, a mammalian homologue of yeast Apg8p, is localized in autophagosome membranes..., Kabeya [/bib_ref] , we also analyzed the conversion of cytosolic LC3-I into LC3-II Western blot analysis. Results showed that MG132 induced LC3 transition in a dose-dependent manner in SKOV3, OVCAR3 or A2870 cells, respectively . Similar like MG132, Western blot analysis also demonstrated that other proteasome inhibitors including BZ, Epox and Lacta also induced LC3 transition in OVCAR3 cells . Since both autophagy induction and impaired autophagic degradation ascribes to accumulation of LC3-II [bib_ref] Guidelines for the use and interpretation of assays for monitoring autophagy in..., Klionsky [/bib_ref] , the effect of inhibiting lysosomal turnover of autophagosome contents by bafilomycin A1 [bib_ref] Guidelines for the use and interpretation of assays for monitoring autophagy in..., Klionsky [/bib_ref] were also examined. Preventing lysosomal degradation by bafilomycin A1 cotreatment significantly increased LC3-II transition elicited by proteasome inhibitors . In addition, knockdown of Atg7, a well-known autophagy essential gene, blocked LC3-II transition elicited by MG132 .
## Autophagy demonstrated little effects on mg132mediated cytotoxicity in ovarian cancer cells
To investigate the potential role of autophagy in cytotoxicity induced by proteasome inhibitors, OVCAR3 cells were transfected with shRNA against Atg7 (shAtg7) to suppress autophagy at the early stage. MTT assay demonstrated that shAtg7 demonstrated little effects on viability of OVCAR3 upon MG132 exposure [fig_ref] Figure 3: Autophagy demonstrated little effects on cytotoxicity induced by MG132 in ovarian cancer... [/fig_ref]. Cotreatment with chloroquine or bafilomycin A1 to suppress autophagy at the late stage also demonstrated little effect on MG132-induced cytotoxicity of ovarian cancer cells [fig_ref] Figure 3: Autophagy demonstrated little effects on cytotoxicity induced by MG132 in ovarian cancer... [/fig_ref].
## Wortmannin (wt) and 3-ma demonstrated no effect on proteasome inhibitors-induced autophagy in ovarian cancer cells
To investigate the role of autophagy in proteasome inhibitors-mediated cytotoxicity of ovarian cancer cells, we managed to suppress autophagy activated by proteasome inhibitors using PI3Ks inhibitors.
Unexpectedly, AO staining demonstrated that neither WT nor 3-MA suppressed MG132-induced accumulation of acid vacuoles in SKOV3, OVCAR3 and A2870 cells . Western blot analysis confirmed that WT or 3-MA could not inhibit the conversion of LC3-I to LC3-II in MG132-treated SKOV3, OVCAR3 or A2870 cells . To investigate whether proteasome inhibitors generally induced PI3Kindependent autophagy, we tested some other proteasome inhibitors including BZ, Epox and Lacta in OVCAR3 cells. AO staining demonstrated that all these proteasome inhibitors induced accumulation of acidic vacuoles, neither WT nor 3-MA could block acidic vacuoles accumulation . Western blot also confirmed that neither WT nor 3-MA could block transition of LC3-I to LC-II elicited by these proteasome inhibitors .
## Proteasome inhibitors elicited beclin 1-independent autophagy in ovarian cancer cells
As Beclin 1 is essential for the PI3K complex [bib_ref] Regulation of macroautophagy by mTOR and Beclin 1 complexes, Pattingre [/bib_ref] , observations that neither WT nor 3-MA was able to inhibit the increase in autophagosomes induced by proteasome inhibitors prompted us to confirm the role of Beclin 1 in proteasome inhibitors-induced autophagy. Western blot analysis demonstrated that MG132 reduced Beclin 1 expression in a dose-dependent manner in SKOV3, OVCAR3 and A2870 cells . Real-time RT-PCR found that MG132 had no obvious effects on Beclin 1 mRNA expression , suggesting that MG132 suppresses Beclin 1 at the translational or posttranslational level. To confirm the involvement of Beclin 1 in autophagy elicited by proteasome inhibition, Beclin 1 expression levels were further reduced by shRNA specific against Beclin 1 (shBeclin 1) in OVCAR3 cells. Western blot analysis confirmed that with some different extents, proteasome inhibitors reproducibly reduced Beclin 1 expression . Specific shRNA against Beclin 1 (shBeclin 1) effectively reduced Beclin 1 levels under basal condition or upon exposure to proteasome inhibitors . Importantly, transition of LC3-I to LC3-II and acidic vacuoles formation elicited by proteasome inhibitors was not affected by shBeclin 1.
## Overexpression of beclin 1 enhanced cytotoxicity of ovarian cancer cells induced by proteasome inhibitors
To determine the influence of Beclin 1 in cytotoxicity of ovarian cancer cells induced by proteasome inhibitors, OVCAR3 cells were transfected with Beclin 1 eukaryotic expression vector. Compared to parental and pcDNA3.1 vector-transfected controls, a higher expression of Beclin 1 protein was detected in the Beclin 1-transfected OVCAR3 cells, and reduction of Beclin1 protein by proteasome inhibitors was suppressed by Beclin 1 transfection [fig_ref] Figure 6: See legend on next page [/fig_ref]. Overexpression of Beclin 1 significantly enhanced proteasome inhibitors-induced cytotoxicity of ovarian cancer cells, as assessed by cleavage of PARP [fig_ref] Figure 6: See legend on next page [/fig_ref] , MTT assay [fig_ref] Figure 6: See legend on next page [/fig_ref] , nuclei staining with Hoechst 33258 [fig_ref] Figure 6: See legend on next page [/fig_ref] , and caspase 3 activity assay [fig_ref] Figure 6: See legend on next page [/fig_ref]. In addition, analysis of PARP cleavage [fig_ref] Figure 6: See legend on next page [/fig_ref] and MTT assay [fig_ref] Figure 6: See legend on next page [/fig_ref] demonstrated that Beclin 1 overexpression also increased MG132induced cytotoxicity of SKOV3 cells and A2870 cells.
# Discussion
Two major proteolytic systems for the clearance of proteins are conserved in eukaryotic cells: the first is the UPS and the second is autophagy. Both UPS and autophagy are involved in most aspects of normal physiology and development. They are also implicated in multiple pathological states, such as cancer, neurodegeneration and aging. Although UPS and autophagy are generally thought to be independent from each other, recent investigations now support that the two proteolytic systems are functionally linked, and autophagy is activated and plays a compensatory role when UPS function is impaired [bib_ref] The role of autophagy-lysosome pathway in neurodegeneration associated with Parkinson's disease, Pan [/bib_ref] [bib_ref] Autophagy induction rescues toxicity mediated by proteasome inhibition, Rubinsztein [/bib_ref] [bib_ref] Linking of autophagy to ubiquitin-proteasome system is important for the regulation of..., Ding [/bib_ref] [bib_ref] The formation of peripheral myelin protein 22 aggregates is hindered by the..., Fortun [/bib_ref] [bib_ref] The role of ATF4 stabilization and autophagy in resistance of breast cancer..., Milani [/bib_ref] [bib_ref] Induction of autophagy by proteasome inhibitor is associated with proliferative arrest in..., Wu [/bib_ref] [bib_ref] Role of the aggresome pathway in cancer: targeting histone deacetylase 6-dependent protein..., Rodriguez-Gonzalez [/bib_ref] [bib_ref] Apoptotic and autophagic cell death induced by histone deacetylase inhibitors, Shao [/bib_ref] [bib_ref] Domainselective small-molecule inhibitor of histone deacetylase 6 (HDAC6)-mediated tubulin deacetylation, Haggarty [/bib_ref]. Autophagy is frequently activated in cancer cells in response to chemo-or radiotherapy [bib_ref] Autophagic cell death: the story of a misnomer, Kroemer [/bib_ref] [bib_ref] Autophagy in tumour suppression and promotion, Brech [/bib_ref]. The contribution of autophagy to cell death induced by therapy generally remains controversial as autophagy protects some cancers against chemotherapy yet sensitizes others to chemotherapy-mediated cytotoxicity [bib_ref] Autophagic cell death: the story of a misnomer, Kroemer [/bib_ref] [bib_ref] Autophagy in tumour suppression and promotion, Brech [/bib_ref].
In the current study, we confirmed that proteasome inhibitors activated autophagy in ovarian cancer cells, as evidenced by accumulation of acidic vacuoles, increase in LC3-II transition. However, suppression of autophagy at the early stage by knockdown of Atg7, as well as at the late stage by cotreatment with chroroquine or bafilomycin A1 demonstrated little effects on cytotoxicity of ovarian cancer cells mediated by proteasome inhibition. The different role of autophagy in chemotherapy-induced cytotoxicity might represent cellspecific and/or stress-specific response. The dual roles of autophagy in survival and cell death require further clarification. Beclin 1, the mammalian homologue of the yeast Atg6 was initially identified as a Bcl2-interacting tumor suppressor [bib_ref] Protection against fatal Sindbis virus encephalitis by beclin, a novel Bcl-2-interacting protein, Liang [/bib_ref]. It is now known that Beclin 1 cooperates with several cofactors to activate lipid kinase PI3KC3, which is essential for recruitment of other Atg proteins to form autophagic vacuoles or autophagosomes [bib_ref] Regulation of macroautophagy by mTOR and Beclin 1 complexes, Pattingre [/bib_ref]. However, several recent studies have demonstrated that some stimuli can also induce (See figure on previous page.) Beclin 1-independent autophagy induced by proteasome inhibitors in ovarian cancer cells. A, Ovarian cancer cells were treated with the indicated concentrations of MG132, and Western blot was performed using the indicated antibodies. B, Ovarian cancer cells were treated with the indicated concentrations of MG132, Beclin 1 mRNA was measured using real-time PCR. C, OVCAR3 cells were transfected with control shRNA or shRNA against Beclin 1 (shBeclin 1), then treated with the indicated proteasome inhibitor, and LC3 transition was analyzed using Western blot analysis. D, OVCAR3 cells were transfected with control shRNA or shRNA against Beclin 1 (shBeclin 1), then treated with the indicated proteasome inhibitor, and AVs formation was measured using AO staining. PI3KC3 and Beclin 1-independent autophagy, so named as non-canonical autophagy [bib_ref] Role of non-canonical Beclin 1-independent autophagy in cell death induced by resveratrol..., Scarlatti [/bib_ref] [bib_ref] Beclin 1-independent autophagy induced by a Bcl-XL/Bcl-2 targeting compound, Z18, Tian [/bib_ref] [bib_ref] Regulation of autophagy by extracellular signal-regulated protein kinases during 1-methyl-4-phenylpyridinium-induced cell death, Zhu [/bib_ref] [bib_ref] Beclin 1-independent pathway of damageinduced mitophagy and autophagic stress: implications for neurodegeneration..., Chu [/bib_ref] [bib_ref] Non-canonical autophagy: an exception or an underestimated form of autophagy, Scarlatti [/bib_ref]. For example, resveratrol, Parkinsonian neurotoxin MPP + and a small compound targeting the BH3 binding groove of Bcl-XL has been shown to activate autophagy in a Beclin 1-independent manner in breast cancer MCF7 cells, neuroblastoma cells and HeLa cells, respectively [bib_ref] Role of non-canonical Beclin 1-independent autophagy in cell death induced by resveratrol..., Scarlatti [/bib_ref] [bib_ref] Beclin 1-independent autophagy induced by a Bcl-XL/Bcl-2 targeting compound, Z18, Tian [/bib_ref] [bib_ref] Regulation of autophagy by extracellular signal-regulated protein kinases during 1-methyl-4-phenylpyridinium-induced cell death, Zhu [/bib_ref] [bib_ref] Beclin 1-independent pathway of damageinduced mitophagy and autophagic stress: implications for neurodegeneration..., Chu [/bib_ref] [bib_ref] Non-canonical autophagy: an exception or an underestimated form of autophagy, Scarlatti [/bib_ref]. In the current study, for the first time, we reported that proteasome inhibitors elicited PI3KC3 and Beclin 1-independent autophagy in ovarian cancer cells, as evidenced by neither PI3Ks inhibitor wortmannin or 3-MA, nor shRNA against Beclin 1 could block accumulation of acidic vacuoles and increase in LC3-II transition induced by proteasome inhibitors. The mechanisms by which autophagosome formation can bypass the Beclin 1-PI3KC3 pathway remain to be clarified in the future. Genetic analysis has revealed that Beclin 1 is implicated in tumorigenesis and plays a role in cellular proliferation [bib_ref] Autophagy in the pathogenesis of disease, Levine [/bib_ref] [bib_ref] Induction of autophagy and inhibition of tumorigenesis by beclin 1, Liang [/bib_ref] [bib_ref] Inhibition of macroautophagy triggers apoptosis, Boya [/bib_ref] [bib_ref] Promotion of tumorigenesis by heterozygous disruption of the beclin 1 autophagy gene, Qu [/bib_ref]. It has been reported that overexpression of Beclin 1 activates autophagy and reduces the tumorigenetic potential of breast cancer cells [bib_ref] Induction of autophagy and inhibition of tumorigenesis by beclin 1, Liang [/bib_ref]. In addition, overexpression of Beclin 1 has been shown to enhance the sensitivity of cervix and gastric cancer cells to chemotherapeutic drugs [bib_ref] Over-expression of the Beclin1 gene upregulates chemosensitivity to anti-cancer drugs by enhancing..., Sun [/bib_ref] [bib_ref] Beclin 1 augmented cisdiamminedichloroplatinum induced apoptosis via enhancing caspase-9 activity, Furuya [/bib_ref]. On the contrary, heterozygous disruption of Beclin 1 in mice increases cellular proliferation and results in spontaneous malignancies [bib_ref] Promotion of tumorigenesis by heterozygous disruption of the beclin 1 autophagy gene, Qu [/bib_ref]. Consistent with previous reports [bib_ref] Over-expression of the Beclin1 gene upregulates chemosensitivity to anti-cancer drugs by enhancing..., Sun [/bib_ref] [bib_ref] Beclin 1 augmented cisdiamminedichloroplatinum induced apoptosis via enhancing caspase-9 activity, Furuya [/bib_ref] , in the current study, flowcytometry analysis and caspase 3 activity assay indicated that a greater in crease in apoptosis was observed in Beclin 1-transfected cells than the mocktransfected cells. Controversially, Beclin 1 knockdown has been shown to promote apoptosis induced by doxorubicin in HepG2 cells [bib_ref] Partial Beclin 1 silencing aggravates doxorubicin-and Fas-induced apoptosis in HepG2 cells, Daniel [/bib_ref]. These reports therefore suggest that Beclin 1 may modulate apoptosis in cell-specific and stimuli-specific patterns. It has been generally believed that Beclin 1 functions as a haploinsufficient tumor suppressor via autophagy activation [bib_ref] Induction of autophagy and inhibition of tumorigenesis by beclin 1, Liang [/bib_ref] [bib_ref] Cloning and genomic organization of beclin 1, a candidate tumor suppressor gene..., Aita [/bib_ref]. However, in the current study, we found that proteasome inhibitors activated autophagy in a Beclin 1-independent manner. In addition, suppression of autophagy both at the early stage and at the late stage had no obvious effects on cytotoxicity mediated by proteasome inhibitors. On the contrary, Beclin 1 overexpression enhanced responsiveness of ovarian cancer cells to proteasome inhibitorsmediated cytotoxicity, indicating that Beclin 1 exerts autophagy-independent tumor suppressive effect in ovarian cancer cells upon exposure to proteasome inhibitors. Therefore, mechanisms underlying enhancing effects of Beclin 1 on chemosensitivity may be multifactorial, and the mechanisms by which Beclin 1 sensitizes ovarian caner cells to proteasome inhibition require further investigation.
(See figure on previous page.) [fig_ref] Figure 6: See legend on next page [/fig_ref] Sensitizing effects of Beclin 1 overexpression on proteasome inhibitors-mediated cytotoxicity of ovarian cancer cells. A, OVCAR3 cells were transfected with mock or Beclin 1 eukaryotic plasmid for 24 h, then treated with the indicated proteasome inhibitors for additional 24 h, and Western blot was performed using the indicated antibodies. A representative image was presented, and the ratios vs that of control (normalized by GAPDH) was noted at the bottom of the Beclin 1 image. B, OVCAR3 cells were transfected with mock or Beclin 1 eukaryotic plasmid for 24 h, then treated with the indicated proteasome inhibitors for additional 24 h, and cell viability was measured using MTT assay. C, OVCAR3 cells were transfected with mock or Beclin 1 eukaryotic plasmid for 24 h, then treated with the indicated proteasome inhibitors for additional 24 h, and nuclei morphology was measured using Hoechst 33258 staining. D, OVCAR3 cells were transfected with mock or Beclin 1 eukaryotic plasmid for 24 h, then treated with the indicated proteasome inhibitors for additional 24 h, and caspase-3 activity was measured. E, SKOV3 and A2870 cells were transfected with mock or Beclin 1 eukaryotic plasmid for 24 h, then treated with 5 μM of MG132 for additional 24 h, and Western blot was performed. F, SKOV3 and A2870 cells were transfected with mock or Beclin 1 eukaryotic plasmid for 24 h, then treated with 5 μM of MG132 for additional 24 h, and cell viability was analyzed using MTT assay. *, P<0.01.
[fig] Figure 1: See legend on next page.) [/fig]
[fig] Figure 3: Autophagy demonstrated little effects on cytotoxicity induced by MG132 in ovarian cancer cells. A, OVCAR3 were transfected with scramble shRNA or shRNA specific against Atg7 (shAtg7) for 24 h, then treated with 5 μM of MG132 for additional 24 h, and cell viability was measured using MTT assay. B, SKOV3, OVCAR3 or A2870 cells were treated with 5 μM of MG132 in the presence of chloroquine or bafilomycin A1 for 24 h, and cell viability was measured using MTT assay. N.S., p>0.05. [/fig]
[fig] Figure 4, Figure 5: PI3KC3-independent autopahgy induced by proteasome inhibitors in ovarian cancer cells. A, Ovarian cancer cells were treated with 5 μM of MG132 in the absence or presence of wortmannin (WT) or 3-MA, and AVs formation was analyzed using AO staining. B, Ovarian cancer cells were treated with 5 μM of MG132 in the absence or presence of WT or 3-MA, and LC3 transition was analyzed using Western blot analysis. C, OVCAR3 cells were treated with the indicated proteasome inhibitor in the absence or presence of WT or 3-MA, and AVs formation was measured using AO staining. D, OVCAR3 cells were treated with the indicated proteasome inhibitor (PI) in the absence or presence of WT or 3-MA, and LC3 transition was analyzed using Western blot analysis. See legend on next page.) [/fig]
[fig] Figure 6: See legend on next page.) [/fig]
|
Prevalence of psychoactive substance use among acutely hospitalised patients in Oslo and Moscow: a cross-sectional, observational study
Hospital. We prepared the samples by 96-well supported liquid extraction, which were then analyzed by liquid chromatography -tandem mass spectrometry (LC-MS/MS), similar to method used by Kristoffersen et al (19). Stable isotope-labeled internal standards were used for all compounds. Two multiple reaction monitor (MRM) transitions were used for each compound and each internal standard. The detection limits chosen to be used (cut-off limit) were based on validation data (precision, RSD ≤ 20 % and accuracy ≤ ±20 %) and signal/noise (S/N) ratio of the quantifier MRM transition (S/N ≥ 10). For THC S/N ≥ 3 was used. The Russian samples were analyzed at the Moscow Research Center for Addiction. Blood samples were prepared by protein precipitation and analyzed by UHPLC-MS/MS. Two MRM transitions were used for each compound. The cut-off level for each analyte was defined as a signal corresponding to S/N = 1000 for the quantifier MRM transition.Samples in which the patient was administered morphine or diazepam prior to blood sampling were omitted from the results. In order to distinguish medicinal morphine use from illicit use, a morphine/codeine-ratio cut-off value was utilized.QuestionnaireUpon enrolment, every patient filled out a questionnaire containing demographic data: age ranges (18-40, 41-60, 61-70, and ≥71 years), gender (male/female), marital status (single, married or widowed), and employment status (employed, retired, unemployed, or student). Psychological distress was measured using the Symptom Checklist 5 (SCL-5), a short-form screening tool which performs almost as well as the full version (20). A value above >1.75 points indicates psychological distress. Self-reported alcohol use was measured through the Alcohol Use Disorder Identification Test -4 (AUDIT-4), which has been validated as a simple, yet effective, screening tool for identifying alcohol use disorders (21). A value of 7 points or more for men and 5 points or more for women indicates at-risk alcohol consumption patterns. Questionnaires were manually punched, controlled, and later transferred to the project database.Psychoactive substancesOur primary outcome was the presence of a psychoactive substance, determined through analytical cut-off values (shown in tables 1 and 2). Psychoactive prescription medication was divided into opioids, benzodiazepines and z-hypnotics. The specific drugs were selected based on prescription practices among primary care and hospital physicians, as well as hospital usage rates. Due to differences in prescription practices, certain substances were omitted at the Moscow site, Page 8 of 28For peer review only -http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Neutel CI, Skurtveit S, Berg C. What is the point of guidelines? Benzodiazepine and zhypnotic use by an elderly population. Sleep medicine. 2012;13(7):893-7. 3.The use of psychoactive prescription medication is increasing. This is a cause for concern, particularly among the elderly, as age-related changes in physiology increase the risk for adverse effects from psychoactive drug use. Previous studies have typically been population-based, and have assessed the frequency of psychoactive drug use through the use of registry based-data. Using blood sample analysis, we sought to accurately assess the prevalence of psychoactive drugs in patients acutely admitted to Departments of Internal Medicine, along with demographic data, concurrent use of alcohol and level of psychological distress.
# Methods
A total of 5839 patients aged 18 years and above participated, and were evenly distributed among two hospital sites, one in Oslo, Norway and one in Moscow, Russia. The presence of psychoactive substances was determined through the use of blood sample analysis utilizing liquid chromatography-mass spectrometry. In addition, questionnaire data comprising demographic data, level of psychological distress, and self-reported illicit drug and alcohol use, was manually punched.
# Results
Of the 2830 patients included in Oslo and the 3009 patients included in Moscow, 32.3 % and 12 % were positive for one or more psychoactive medicinal drugs, respectively. In Oslo, medicinal drug use was associated with increasing age and psychological distress. In Moscow, psychoactive medicinal drug use was more common among patients aged 18 to 40 years compared to those aged 41 years and older.
# Conclusion
This study demonstrates that a significant proportion of admitted patients use one or more psychoactive medicinal drugs, in particular benzodiazepines (Oslo and Moscow) and opiates (Oslo). We suggest formalized screening for inappropriate prescription drug use, prudent appraisals of the risk for adverse events related to such use, and increased adherence to clinical prescription guidelines.
Word count: 3680
# Strengths and limitations
-Blood sample analysis was utilized in addition to self-reporting to accurately assess the prevalence of psychoactive substance use -Large sample size divided across two sites, with an inclusion period lasting one year -Differences in laboratory methods, legislation and prescription practices for psychoactive substances between the selected sites limit comparability
# Introduction
Background While the deleterious effects of harmful alcohol use and tobacco smoking are well known, several studies have also called attention to increasing rates of psychoactive prescription drug use in Western countries, especially among the elderly. These drugs include opioids, benzodiazepines and z-hypnotics. In addition to increased usage rates, there seems to be an apparent increase in the non-medical use and abuse of psychoactive medication (4), as well as inappropriate prescription practices and diversion from clinical guidelines.
Several studies have prospectively assessed the impact and usage rates of alcohol, psychoactive medication and illicit substances in the emergency department . Alcohol-related presentations to the emergency department have been quantified both through the use of questionnaires, and classification through the International Statistical Classification on Diseases and Related Health Problems-system (ICD-10). However, patients may under-report alcohol consumption when presented with a questionnaire, and self-reporting may be limited by recall bias . Potentially harmful, but not clinically apparent alcohol-related presentations may also remain undetected by health professionals, and misclassified. These limitations may attenuate the accuracy of the prevalence data in such studies. A study published in 2011 utilized blood sample analysis using highly selective methods to assess the prevalence of psychoactive substances among patients acutely admitted due to injury -this provided accurate and reliable prevalence data. However, blood sample analysis has not been employed to assess the prevalence of psychoactive substances among patients acutely admitted to Departments of Internal Medicine.
Patients admitted to Departments. of Internal Medicine are likely to be older than the general population(11). Due to age-related changes in the pharmacodynamics and pharmacokinetics properties of various drugs, the elderly are also more vulnerable to any potential adverse effects from psychoactive medication use, as well as alcohol . Changes in physiology and body composition related to aging increases the susceptibility to adverse effects from alcohol consumption , and the concurrent use of psychoactive drugs and alcohol also exacerbates the respective adverse effects of each substance. Drowsiness and reduced postural stability are common effects of psychoactive drugs, and are enhanced by alcohol consumption. Consequently, both drug-drug and drug-alcohol interactions may cause clinically significant adverse reactions in the elderly, such as falls, injury and infections . These events may be precipitated even by low-tomoderate doses of the substances mentioned. Polypharmacy -the concurrent use of multiple medications -increases in frequency with age, with rates ranging from 12 % to 35 % among the elderly in Western countries . This may contribute to the additive risk for adverse drug-drug interactions. Considering the high frequency of use among older adults, knowledge regarding the prevalence of these substances among acutely hospitalized patients may be of clinical significance.
Contrary to alcohol, which is usually independently procured, and where ingestion is userinitiated, the distribution of psychoactive medication is generally mediated by health professionals. In light of increased rates of non-medical use and abuse of these drugs , the presence of reliable and accurate prevalence data may mediate prudence regarding prescription practices, and lead to more critical appraisals of when these drugs are necessary, or should be discontinued.
## Study aims
In this study, we sought to assess the prevalence of psychoactive substances among patients acutely hospitalized to Departments of Internal Medicine.
Specifically, we wished to: 1) Measure the rate of psychoactive substance use through the use of blood sampling and highly selective analytical methods 2) Investigate patient characteristics when a psychoactive substance is detected; namely age, gender, employment status, marital status, degree of psychological distress, and selfreported alcohol consumption. 3) Investigate the concurrent rate of alcohol and illicit drug use when psychoactive substances are detected
This study is part of larger collaborative project between Oslo University Hospital and the Moscow Research Center on Addiction.
## Study design, setting and participants
Design and site selection We have employed an observational, cross-sectional approach. We chose two study sites -Lovisenberg Diaconal Hospital (LDS) in Oslo, Norway (site 1) and Hospital No. 68 in Moscow, Russia (site 2). Both sites are medium-sized urban hospitals. The mean life expectancy in Norway is 84.2 years for women and 80.6 years for men , which is significantly higher than in Russia, where the mean life expectancy is 72.1 years (17). However, due to socio-economic disparities within the various districts in Oslo, the life expectancy in several populations belonging to the catchment area for LDS is similar to that in Russia, ranging from 72 -76 years for men and 79 -81 years for men and women, respectively . Due to differences in laboratory techniques, prescription practices, as well as guidelines and legislation for psychoactive drug use in each nation, results for each site will be presented separately.
## Inclusion and exclusion criteria
Emergency Department(ED)-nurses recruited patients consecutively upon admission in Oslo, while in Moscow patients were recruited by medical doctors upon arrival to the ward. The ED in Oslo and medical wards in Moscow were both located at the Department of Internal Medicine, and therefore did not admit patients with injuries or surgical conditions. Only patients aged 18 years and above and able to give informed consent were included. Patients unable to consent upon admission due reversible or transient causes (such as intoxication) were approached at a later time by dedicated research assistants, and were offered to participate when they were able to consent. The inclusion period lasted from November 2016 to December 2017, and patients were included at all hours of the day, to account for any seasonal or diurnal variations. We employed the following exclusion criteria: 1) Permanently unable to give an informed consent. 2) Elective transfer from other hospitals. 3) Limited or no ability to read or write the national language, e.g. Russian or Norwegian, as translated copies of both the consent form and accompanying questionnaire were unavailable.
## Patient and public involvement
There was no patient or public involvement in the design and implementation of this study. In addition to dissemination results to clinicians and policy makers, we intend to inform and involve patient interest organizations in the field of substance use disorders.and vice versa for the Oslo site. Specifically, the use of z-hypnotics in Russia is uncommon, and phenazepam, a commonly prescribed benzodiazepine in Russia, is not marketed in Norway. The presence of alcohol was determined through direct analysis of blood alcohol content. Based on population studies, the most common illicit drugs and their metabolites were analyzed -these comprise amphetamines, cocaine, MDMA/ecstasy, heroin and tetrahydrocannabinol. So-called "new psychoactive substances (NPS)" were not analyzed -the rapid introduction of new compounds poses significant analytical challenges, and selecting which substance to analyze is complicated by limited knowledge regarding their availability in the general population . Polyuse was defined as the presence of two or more psychoactive substances (prescription medication, positive blood alcohol content, or illicit drugs) in any given blood sample.
# Statistical analysis
We analyzed the data using IBM SPSS 25.0 (Armonk, NY). The prevalence of various psychoactive substances is shown through descriptive tables. We employed bi-variate cross tables to analyze associations between the presence of a psychoactive substance and sample characteristics. Chi-square statistics with odds ratios (ORs) were used to assess statistical significance. Our primary dependent variables were the presence of psychoactive substances, illicit drugs, and polyuse. Our co-variates were age ranges, marital status, employment status, positive blood alcohol content, presence of illicit drugs, polyuse, presence of psychological distress, and selfreported alcohol use. The level of significance was set at p < 0.05. Continuous data are presented as means, with 95 % confidence intervals (CIs).
## Page 9 of 28
For peer review only -http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open
## Participants
A total of 2890 patients in the Oslo site and 3009 patients in the Moscow site were included, with a participation rate of 81 % and 90 % respectively -52.4 % were women in Oslo and 52.9 % were women in Moscow. Distribution across age ranges was similar at both sites (tables 1 and 2).
## Prevalence data and sample characteristics prevalence data
Tables 1 and 2 depict the main findings in our study, showing prevalence data for all psychoactive medicines analyzed and their analytical cut-off values. In Oslo, psychoactive substances were detected in 32.3 % of analyzed samples. The most common drug class detected was opioids (19.5 %), followed by benzodiazepines . Z-hypnotics were found in 6.6 % of samples. In the Moscow site, psychoactive substances were detected in 12 % of samplesbenzodiazepines were the most frequently detected drug class (9.7 %). The rates for other drug classes were low, with opioids found in 0.9 %.
In Oslo, illicit drugs were detected in 6.4 % of all samples, with tetrahydrocannabinol (3.7 %) and amphetamines (2.1 %) being the most common. In the Moscow site, the total rate for illicit drug use was lower compared to the Oslo site, at 1.4 %, with THC being the most common (1.1 %). The rates for the remaining substances in both sites were low.show the un-adjusted distribution for psychoactive substance users and non-users across all co-variates measured in Oslo and Moscow respectively. Psychoactive substance use was evenly distributed among genders at both sites. In Oslo, the highest proportion of psychoactive drug use was found among patients aged older than 40 years, as well as among those being retired, and/or in the lowest AUDIT-4 ranges. In Moscow, medicinal drug use was more evenly distributed across all age groups, with the highest proportion of use among patients that were non-active in the workforce.depicts logistic regression analysis for various co-variates in the Oslo site. Compared to patients aged 18-40, the likelihood for medicinal drug use was higher among patients aged 61-70 and 71 years or older. There was no significant gender difference among medicinal drug users. The likelihood for polyuse was higher among non-employed patients compared to patients being active in the workforce -this was also observed in regards to illicit substances as well as medicinal drug use. In the Oslo site, female patients were also significantly less likely to use illicit drugs than male patients. Patients aged 18-40 were more likely to use illicit drugs than other age groups. The use of psychoactive substances, illicit drugs, and polyuse was significantly more likely in patients scoring above the cut-off value for SCL-5 (>1.75), similar trends for SCL-5 were observed, where patients scoring above the cut-off value were more likely to use illicit drugs and psychoactive substances. However, there was no significant difference in the likelihood for polyuse. Contrary to the Oslo site, the presence of psychoactive substances and polyuse was significantly less likely in patients aged >40 years. Being retired increased the likelihood for both psychoactive substance use and polyuse compared to patients being actively employed -the same association was not observed for patients being nonactive in the workforce.shows the distribution of psychoactive substances within different AUDIT-4 ranges in Oslo. The highest prevalence of benzodiazepine-use is found in the highest AUDIT-4 range (9-16 points), with a rate of 31 %, compared with 18.5 % for the whole study population. The use of opiates is more prevalent at the lowest and highest AUDIT-4 ranges. The use of THC, stimulants and zhypnotics is evenly distributed along all AUDIT-4 ranges. In Moscow, the highest rate of benzodiazepine use is also found among the highest AUDIT-4 range, while the usage rate for other psychoactive substances is uniformly low in all other AUDIT-4 ranges.
## Sample characteristics
# Discussion
Our study demonstrates that the use of psychoactive substances among patients admitted to Departments of Internal Medicine is highly prevalent, while simultaneously highlighting significant differences in sample characteristics between Oslo and Moscow. Based on previous studies, the volume of participants and rate of inclusion provide representative, accurate data. The number of positive sample for psychoactive substances was far higher in Oslo than in Moscow. The use of psychoactive substances in Oslo was significantly associated with older age, psychological distress, and being unemployed. Previous population-based studiesshow an apparent increase in the use of psychoactive prescription medication, particularly among the elderly. However, the drug use statistics were extracted from national prescription databases, and whether or not the patient took the prescribed drug as well as the duration of drug use was not measured. Additionally, substance abuse disorders are more prevalent among non-participants in health surveys. These factors may underestimate the prevalence of psychoactive substance use.
The high rate of psychoactive substance use among the elderly is a cause for concern. The changes in pharmacodynamics and pharmacokinetics associated with senescence increase complexity when prescribing medications to the elderly. Inter-individual variability to drugs increases with age, partly explained by reduced physiological reserves during stressors such as inter-current illness or disease. Half-life of elimination is prolonged by changes in body composition, affecting the volume of distribution of lipid-soluble drugs, and reduced renal and hepatic clearance . These factors alter and increase the sensitivity to psychoactive substances. Elderly patients are particularly sensitive to the effect of benzodiazepines on the central nervous system. Sedation may occur at lower doses than in younger patients. Drowsiness and decreased motor skills, combined with the age-related reduction in functional reserves, increase the risk for immobility and falls (29).Physiological reserves in many elderly people may be adequate at normal homeostasis. However, the introduction of a stressor, such as an illness requiring acute hospitalization, may deplete these reserves, even after the precipitating cause has been corrected. Concurrent use of psychoactive substances and the increased vulnerability to their adverse effects may exacerbate this course. Additionally, conditions which increase in frequency and severity with old age, such as chronic obstructive pulmonary disease, have increased rates of adverse events among users of benzodiazepines and opioids. Increasing age also increases morbidity and multi-morbidity, which in turn is associated with polypharmacy (32), with a concurrent increased risk for drug-drug interactions and other adverse effects.
## Clinical implications
The above considerations, combined with the findings from our study, should prompt an increased awareness regarding the use of psychoactive substances among acutely hospitalized patients. A reduction in inappropriate use or misuse of psychoactive substances may have a positive impact on hospitalization rates, risk of re-admittance, and health care costs. Specifically, strengthening and implementing the following practices may contribute to improved patient outcomes: -Screening for inappropriate or harmful psychoactive substance use -Critically assessing the indications for psychoactive drug use and considering discontinuation or tapering if inappropriate prescribing is present -Assess the degree of increased risk for adverse events in specific conditions when using psychoactive substances, such as chronic obstructive pulmonary disease -Assess concomitant alcohol use if psychoactive substance use is present Hospital admissions may serve as opportunities to screen for unapparent harmful use or misuse of psychoactive substances, but this practice is less likely to be performed among elderly patients than among younger individuals. Health professionals may experience hesitation in screening for unhealthy psychoactive substances use due to the perceived sensitive nature of the subject (34), but guidelines for proper communicative techniques exist, which may ameliorate any associated stigma. Due to the commonality of substance use disorders among acutely admitted patients, clinicians may benefit from assessing psychoactive medication use in addition to alcohol and illicit drugs, particularly among the elderly. Similar to screening for alcohol, validated singlequestion screening tools exist for illicit drug use and non-medical use of psychoactive medication. However, these tools were validated in a primary care setting. Older adults may also perceive symptoms of harmful psychoactive substance use as symptoms of aging, and adverse effects stemming from psychoactive drug use may be similar to manifestations of other conditions. Single-question screening tools may therefore better serve as an adjunct to the general clinical assessment of every admitted patient.
The use of psychoactive medication is usually mediated by health care professionals, which mandates adherence to prescription guidelines. However, based on prescription databases, inappropriate use of benzodiazepines and z-hypnotics is widespread. Hospital clinicians have the opportunity to prudently assess medication use among admitted patients, and are in a position to inform and co-operate with primary care physicians or other health professionals. Any identified instances of harmful or inappropriate use may then be managed by targeted interventions, ordiscontinuation or drug-tapering. In the event of specific illnesses, it is important to consider whether harmful, inappropriate or even therapeutic use may have precipitated an adverse event. As several illnesses that are more commonly occurring in old age progress, physiological reserves may be depleted. For instance, respiratory depression associated with benzodiazepine use (40) may cause adverse respiratory events in patients with chronic pulmonary obstructive disease if pulmonary function continues to decrease.
Brief interventions for substance use disorders have shown varying degrees of efficacy in the reduction of harmful alcohol useand illicit drug use, with some studies showing a reduction in at-risk drinking, while brief interventions have shown no increased efficacy in reducing illicit drug misuse compared to control groups. These studies also did not include elderly patients (above 64 years of age), nor the critically ill. Additionally, study participation and the screening process itself may cause a reduction in harmful substance use in both the control group and the groups receiving brief interventions.
While the above recommendations may be valid for Norway and other similar countries in the West, the number of samples positive for psychoactive substances was markedly lower across all age groups in Moscow, at 12 %. Plausible causes may be differences in prescription guidelines and legislative practices. Additionally, there were significant differences in the cut-off values utilized at both sites -in Moscow, the S/N-ratio cut-off was set at 1000, versus a S/N-ratio at 10 in Oslo -this may have underestimated the number of positive samples. However, the usage frequency of psychoactive medications in a Russian ED-population may still warrant prudence regarding adherence to prescription guidelines and adverse effects. Their use may also reflect a different selection of illnesses in that particular population -further discussion regarding these elements is beyond the scope of this article.
# Limitations
Previous studies have utilized both self-reporting and registry based data to assess substance use among acutely hospitalized patients -a major strength in our study is the addition of blood sample analysis. The analytical methods employed have allowed us to identify individual substances, thus increasing the descriptive accuracy in our data. Our inclusion criteria were broad, increasing the likelihood for a representative sample size -this is also strengthened by a high participation rate. The use of two to different hospital sites in two different nations increased our sample size; however, external generalizability to other ED-populations is limited to due to differences in medical practice and cultural norms. Differences in analytical cut-off values between Oslo and Moscow also limit the comparability between the two data sets. In order to satisfy necessary ethical considerations, patients unable to give informed consent were excluded. Their cognitive limitations may have been caused by or exacerbated by psychoactive substances or alcohol use -consequently, prevalence data for this patient group remains unavailable. The prevalence of substance use among patients who declined to participate is also a potential bias, as they may have been unwilling to participate due to concerns regarding the disclosure of substance or alcohol use, or may have been incapacitated due to illicit drug use.
# Conclusion
The use of psychoactive substances among patients acutely admitted to Departments of Internal Medicine is common, and may cause both readily identifiable as well as oblique adverse effects. Health professionals should be aware that the use of benzodiazepines and opiates is highly prevalent, particularly among elderly patients, and should exact prudence when prescribing these drug classes. Implementing screening practices may improve the identification of inappropriate psychoactive substance use.
# Author disclosures
Contributorship statement DG drafted the manuscript. BJ and SK did the initial data analysis. BJ, SK, VV, DG, AL, GN and STB organized or contributed to the Oslo data collection. EB, EK, AN, ET, AK, AP, SP and ED organized or contributed to the Moscow data collection. All authors were responsible for study design, interpretation of the results, as well as final revision and approval of the manuscript.
## Competing interests
## No conflict of interest declared
# Funding
The study was supported by the Norwegian Ministry of Health's Funding for Norwegian-Russian Collaboration in Health and Social issues grant B-1408. The Norwegian Ministry of Health had no role in the design of the study, in the collection, analysis or interpretation of the data, nor in writing the manuscript or in the decision to submit the article for publication.
## Data sharing
No additional data will be shared.a: M/C = morphine/codeine-ratio: a: In Moscow, several other substances were analyzed in addition to those listed in table 2 -these were omitted due to lack of positive samples. b: In Moscow, the detection limit was based on a signal/noise -ratio above 1000. In Oslo, absolute lower detection limits were utilized. Objectives: The use of psychoactive prescription medication is increasing in the general population. This is a cause for concern, particularly among the elderly, where physiological changes related to senescence increase the risk for adverse effects. While previous studies regarding psychoactive substance use (including illicit drugs and alcohol) have generally been population based, we sought to determine the frequency of such use among acutely hospitalized patients.
## Figure legends
# Strengths and limitations
-Blood sample analysis was utilized in addition to self-reporting to accurately assess the prevalence of psychoactive substance use -Large sample size divided across two sites, with an inclusion period lasting one year -Differences in laboratory methods, legislation and prescription practices for psychoactive substances between the selected sites limit comparability
# Introduction
Background While the deleterious effects of harmful alcohol use and tobacco smoking are well known, several studies have also called attention to increasing rates of psychoactive prescription drug use in Western countries, especially among the elderly (1-3). These drugs include opioids, benzodiazepines and z-hypnotics. In addition to increased usage rates, there seems to be an apparent increase in the non-medical use and abuse of psychoactive medication (4), as well as inappropriate prescription practices and diversion from clinical guidelines.
Several studies have prospectively assessed the impact and usage rates of alcohol, psychoactive medication and illicit substances in the emergency department . Alcohol-related presentations to the emergency department have been quantified both through the use of questionnaires, and classification through the International Statistical Classification on Diseases and Related Health Problems-system (ICD-10). However, patients may under-report alcohol consumption when presented with a questionnaire, and self-reporting may be limited by recall bias . Potentially harmful, but not clinically apparent alcohol-related presentations may also remain undetected by health professionals, and misclassified. These limitations may attenuate the accuracy of the prevalence data in such studies. A study published in 2011 (10) utilized blood sample analysis using highly selective methods to assess the prevalence of psychoactive substances among patients acutely admitted due to injury -this provided accurate and reliable prevalence data. However, blood sample analysis has not been employed to assess the prevalence of psychoactive substances among patients acutely admitted to Departments of Internal Medicine.
Patients admitted to Departments. of Internal Medicine are likely to be older than the general population(11). Due to age-related changes in the pharmacodynamics and pharmacokinetics properties of various drugs, the elderly are also more vulnerable to any potential adverse effects from psychoactive medication use, as well as alcohol . Changes in physiology and body composition related to aging increases the susceptibility to adverse effects from alcohol consumption , and the concurrent use of psychoactive drugs and alcohol also exacerbates the respective adverse effects of each substance. Drowsiness and reduced postural stability are common effects of psychoactive drugs, and are enhanced by alcohol consumption. Consequently, both drug-drug and drug-alcohol interactions may cause clinically significant adverse reactions in the elderly, such as falls, injury and infections (1). These events may be precipitated even by low-tomoderate doses of the substances mentioned. Polypharmacy -the concurrent use of multiple medications -increases in frequency with age, with rates ranging from 12 % to 35 % among the elderly in Western countries . This may contribute to the additive risk for adverse drug-drug interactions. Considering the high frequency of use among older adults, knowledge regarding the prevalence of these substances among acutely hospitalized patients may be of clinical significance. Psychoactive medication, specifically benzodiazepines and z-hypnotics, is also often used to treat psychological distress and sleep disorders, which are common among the elderly . Contrary to alcohol, which is usually independently procured, and where ingestion is user-initiated, the distribution of psychoactive medication is generally mediated by health professionals. In light of increased rates of non-medical use and abuse of these drugs (4), the presence of reliable andaccurate prevalence data may mediate prudence regarding prescription practices, and lead to more critical appraisals of when these drugs are necessary, or should be discontinued.
## Study aims
In this study, we sought to assess the prevalence of psychoactive substances among patients acutely hospitalized to Departments of Internal Medicine.
Specifically, we wished to: 1) Measure the rate of psychoactive substance use through the use of blood sampling and highly selective analytical methods 2) Investigate patient characteristics when a psychoactive substance is detected; namely age, gender, employment status, marital status, degree of psychological distress, and selfreported alcohol consumption. 3) Investigate the concurrent rate of alcohol and illicit drug use when psychoactive substances are detected
This study is part of larger collaborative project between Oslo University Hospital and the Moscow Research Center on Addiction.
# Methods
## Study design, setting and participants
Design and site selection We have employed an observational, cross-sectional approach. We chose two study sites -
. Due to differences in laboratory techniques, prescription practices, as well as guidelines and legislation for psychoactive drug use in each nation, results for each site will be presented separately.
## Ethics
This study was approved by the Regional Ethics Committee for South Eastern Norway (2015/2404), and by the Moscow Local Ethical Committee of MRPCA.
## Inclusion and exclusion criteria
Patients were generally either recruited by emergency department(ED)-nurses while awaiting ward transfer or examinations such as radiographic imaging, or shortly after arrival to their respective ward in Oslo, while in Moscow, patients were recruited upon arrival to the ward by physicians serving as dedicated study and recruitment coordinators. . The EDs in Oslo and Moscow were both located at the Department of Internal Medicine, and therefore did not admit patients with injuries or surgical conditions. Patients were transferred to an appropriate ward after an initial evaluation, which in Oslo was General Internal Medicine, Infectious Diseases, Cardiology, Pulmonary Medicine, Cerebravascular and Geriatric conditions, Hematology, or Gastroenterology, and in Moscow General Internal Medicine, Cardiology, Pulmonary Medicine, or Neurology.Only patients aged 18 years and above and able to give informed consent were included. Patients unable to consent upon admission due reversible or transient causes (such as intoxication or severe illness) were approached at a later time by dedicated research assistants, and were offered to participate when they were able to consent. The inclusion period lasted from November 2016 to December 2017, and patients were included at all hours of the day, to account for any seasonal or diurnal variations. We employed the following exclusion criteria: 1) Permanently unable to give an informed consent. 2) Elective transfer from other hospitals. 3) Limited or no ability to read or write the national language, e.g. Russian or Norwegian, as translated copies of both the consent form and accompanying questionnaire were unavailable.detail the flow of patient recruitment.
## Patient and public involvement
There was no patient or public involvement in the design and implementation of this study. In addition to disseminating results to clinicians and policy makers, we intend to inform and involve patient interest organizations in the field of substance use disorders.
## Variables
Study design, variables and analytical methods were jointly developed by the research teams at both study sites, in order to maintain consistency in data collection and collation. The study sites were also visited by the respective study coordinators after study commencement in order to ensure methodological continuity.
# Blood analysis
Whole blood was drawn from every study participant and secured in 5 ml BD-Vacutainer blood collection glass tubes (BD Vacutainer Systems, Franklin, NJ, USA). The blood samples were stored at 4°C for a maximum of 7 days before analysis. In Oslo, the samples were analyzed at the Division of Laboratory Medicine, Section of Drug Abuse Research, Oslo University Hospital. We prepared the samples by 96-well supported liquid extraction, which were then analyzed by liquid chromatography -tandem mass spectrometry (LC-MS/MS). This is the same sample preparation and LC-MS/MS-method used by Kristoffersen et al . Stable isotope-labeled internal standards were used for all compounds. Two multiple reaction monitor (MRM) transitions were used for each compound and each internal standard. The chosen detection limits (cut-off limit) were based on validation data and signal/noise (S/N) ratio of the quantifier MRM transition (S/N ≥ 10). For THC S/N ≥ 3 was used. The Russian samples were analyzed at the Moscow Research Center for Addiction. Blood samples were prepared by protein precipitation and analyzed by UHPLC-MS/MS. Two MRM transitions were used for each compound. The cut-off level for each analyte was defined as a signal corresponding to S/N = 1000 for the quantifier MRM transition. We have provided additional methodological details and validation data for the Oslo site in.
Samples in which the patient was administered morphine or diazepam prior to blood sampling were omitted from the results. In order to distinguish medicinal morphine use from illicit use, a morphine/codeine-ratio cut-off value was utilized.
## Questionnaire
Upon enrolment, every patient filled out a questionnaire containing demographic data: age ranges (18-40, 41-60, 61-70, and ≥71 years), gender (male/female), marital status (single, married or widowed), and employment status (employed, retired, unemployed, or student). Psychological distress was measured using the Symptom Checklist 5 (SCL-5), a short-form screening tool which performs almost as well as the full version. A value above >2 points indicates psychological distress. Self-reported alcohol use was measured through the Alcohol Use Disorder Identification Test -4 (AUDIT-4), which has been validated as a simple, yet effective, screening tool for identifying alcohol use disorders. A value of 7 points or more for men and 5 points or more for women indicates at-risk alcohol consumption patterns. Both questionnaires have been validated for use in their respective nations . The questionnaires were generally filled out by the patients themselves, with occasional aid from ER-nurses if questions arose. The ER nurses were also instructed beforehand by the study coordinator, and a data collection pilot preceded the inclusion period. We also included self-reported use of benzodiazepines and/or barbiturates (termed sedatives`), z-hypnotics (termed `sleeping medication`), opioids (termed `painkillers`) and illicit drugs (termed `narcotic substances`). All groups had example trade or colloquial names to improve question comprehension. Questionnaires were manually entered, controlled, and later transferred to the electronic project database.
## Psychoactive substances
Our primary outcome was the presence of a psychoactive substance, determined through analytical cut-off values (shown in tables 1 and 2). Psychoactive prescription medication was divided into opioids, benzodiazepines and z-hypnotics. The specific drugs were selected based on prescription practices among primary care and hospital physicians, as well as hospital usage rates. Due to differences in prescription practices, certain substances were omitted at the Moscow site, and vice versa for the Oslo site. Specifically, the use of z-hypnotics in Russia is very uncommon, and phenazepam, a commonly prescribed benzodiazepine in Russia, is not marketed in Norway.
Barbiturates are also commonly used as a sedative in Russia, but this use is uncommon in Norway. The presence of alcohol was determined through direct analysis of blood alcohol content. Based on population studies, the most common illicit drugs and their metabolites were analyzed -these comprise amphetamines, cocaine, MDMA/ecstasy, heroin and tetrahydrocannabinol. So-called "new psychoactive substances (NPS)" were not analyzed -the rapid introduction of new compounds poses significant analytical challenges, and selecting which substance to analyze is complicated by limited knowledge regarding their availability in the general population. Polyuse was defined as the presence of two or more psychoactive substances (prescription medication, positive blood alcohol content, or illicit drugs) in any given blood sample.
# Statistical analysis
We analyzed the data using IBM SPSS 25.0 (Armonk, NY). The prevalence of various psychoactive substances is shown through descriptive tables. We employed bi-variate cross tables to analyze associations between the presence of a psychoactive substance and sample characteristics. Chi-square statistics with odds ratios (ORs) were used to assess statistical significance. Our primary dependent variables were the presence of psychoactive substances, illicit drugs, and polyuse. Our co-variates were age ranges, marital status, employment status, positive blood alcohol content, presence of illicit drugs, polyuse, presence of psychological distress, and selfreported alcohol use. We performed logistic regression analyses to test the association between our dependent values, which were medicinal drug use, illicit drug use and polyuse, and independent covariates, which were age, gender, SCL-5-values, marital status, employment status and AUDIT-4zones. Our reference categories for each variable were age 18-40 years, male gender, SCL-5 score <2, married, active in the workforce and an AUDIT-score of 0-3. The level of significance was set at p < 0.05. Continuous data are presented as means, with 95 % confidence intervals (CIs).
## Participants
A total of 2874 patients in the Oslo site and 3009 patients in the Moscow site were included, with a participation rate of 81 % and 90 % respectively -52.4 % were women in Oslo and 52.9 % were women in Moscow. Distribution across age ranges was similar at both sites (tables 1 and 2).
## Prevalence data and sample characteristics prevalence data
Tables 1 and 2 depict the main findings in our study, showing prevalence data for all psychoactive medicines analyzed and their analytical cut-off values. In Oslo, psychoactive substances were detected in 32.3 % of analyzed samples. The most common drug class detected was opioids (19.5 %), followed by benzodiazepines (18.5 %). Z-hypnotics were found in 6.6 % of samples. In the Moscow site, psychoactive substances were detected in 12 % of samplesbenzodiazepines were the most frequently detected drug class (9.7 %). The rates for other drug classes were low, with opioids found in 0.9 %.
In Oslo, illicit drugs were detected in 6.4 % of all samples, with tetrahydrocannabinol (3.7 %) and amphetamines (2.1 %) being the most common. In the Moscow site, the total rate for illicit drug use was lower compared to the Oslo site, at 1.4 %, with THC being the most common (1.1 %). The rates for the remaining substances in both sites were low.
Cut-off values at each of the two sites were not identical due to differences in analytical methods, with a lower signal/noise-ratio at the Oslo site. We have therefore provided mean concentrations for the substances detected at the Oslo site, which in the majority of substances are well above the analytical cut-off.show the un-adjusted distribution for psychoactive substance users and non-users across all co-variates measured in Oslo and Moscow respectively. Psychoactive substance use was evenly distributed among genders at both sites. In Oslo, the highest proportion of psychoactive drug use was found among patients aged older than 40 years, as well as among those being retired, and/or in the lowest AUDIT-4 ranges. In Moscow, medicinal drug use was more evenly distributed across all age groups, with the highest proportion of use among patients that were non-active in the workforce.depicts logistic regression analysis for various co-variates in the Oslo site. Compared to patients aged 18-40, the likelihood for medicinal drug use was higher among patients aged 61-70 and 71 years or older. There was no significant gender difference among medicinal drug users. The likelihood for polyuse was higher among non-employed patients compared to patients being active in the workforce -this was also observed in regards to illicit substances as well as medicinal drug use. In the Oslo site, female patients were also significantly less likely to use illicit drugs than male patients. Patients aged 18-40 were more likely to use illicit drugs than other age groups. The use of psychoactive substances, illicit drugs, and polyuse was significantly more likely in patients scoring above the cut-off value for SCL-5 (>2).
## Sample characteristics
In the Moscow site, similar trends for SCL-5 were observed, where patients scoring above the cut-off value were more likely to use illicit drugs. However, there was no significant difference in the likelihood for polyuse and medicinal drugs. Contrary to the Oslo site, the presence of psychoactive substances and polyuse was significantly less likely in patients aged >40 years. Being retired increased the likelihood for both psychoactive substance use and polyuse compared to patients being actively employed -the same association was not observed for patients being nonactive in the workforce.show the distribution of self-reported use if a sample is positive for either benzodiazepines, opioids, z-hypnotics and illicit drugs, with both benzodiazepines and barbiturates categorized as `sedatives` in Moscow. If a sample is negative, this is generally mirrored in questionnaires regarding self-reported use. The highest concordance between positive samples and confirmatory self-reported use in Oslo was for z-hypnotics (75.8 %), followed by opioids (62.6 %), illicit drugs (49.0) and benzodiazepines . In Moscow, there was a high concordance between negative samples for opioids and illicit drugs, and self-reported non-use. This percentage was lower for sedatives (83.2 % for barbiturates and 83.8 % for benzodiazepines). For positive samples, the percentage of self-reported use ranged from 11.6 % for illicit drugs, to 55.9 % for barbiturates, 32.0 % for benzodiazepines and 7.1 % for opioids.shows the distribution of psychoactive substances within different AUDIT-4 ranges in Oslo. The highest prevalence of benzodiazepine-use is found in the highest AUDIT-4 range (9-16 points), with a rate of 31 %, compared with 18.5 % for the whole study population. The use of opioids is more prevalent at the lowest and highest AUDIT-4 ranges. The use of THC, stimulants and zhypnotics is evenly distributed along all AUDIT-4 ranges. In Moscow, the highest rate of benzodiazepine use is also found among the highest AUDIT-4 range, while the usage rate for other psychoactive substances is uniformly low in all other AUDIT-4 ranges.
# Discussion
Our study demonstrates that the use of psychoactive substances among patients admitted to Departments of Internal Medicine is highly prevalent, while simultaneously highlighting differences in sample characteristics between Oslo and Moscow. Based on previous studies, the volume of participants and rate of inclusion provide representative, accurate data. The number of positive sample for psychoactive substances was far higher in Oslo than in Moscow. The use of psychoactive substances in Oslo was significantly associated with older age, psychological distress, and being unemployed. The latter two characteristics are known to be associated with substance.There was no significant gender disparity among positive samples -this may be explained by a higher percentage of positive samples among younger males compared to females, and vice versa for older females compared to older males. Previous population-based studies also show an apparent increase in the use of psychoactive prescription medication in the general population, particularly among the elderly. However, the drug use statistics were extracted from national prescription databases, and whether or not the patient took the prescribed drug as well as the duration of drug use was not measured. Additionally, substance abuse disorders are more prevalent among non-participants in health surveys. These factors may underestimate the prevalence of psychoactive substance use -as far as we are aware, this is the first study to demonstrate the frequency of substance use among hospitalized patients.
The high rate of psychoactive substance use among the elderly is a cause for concern. The changes in pharmacodynamics and pharmacokinetics associated with senescence increase complexity when prescribing medications to the elderly. Inter-individual variability to drugs increases with age, partly explained by reduced physiological reserves during stressors such as inter-current illness or disease. Half-life of elimination is prolonged by changes in body composition, affecting the volume of distribution of lipid-soluble drugs, and reduced renal and hepatic clearance . These factors alter and increase the sensitivity to psychoactive substances. Elderly patients are particularly sensitive to the effects of benzodiazepines on the central nervous system. Sedation may occur at lower doses than in younger patients. Drowsiness and decreased motor skills, combined with the age-related reduction in functional reserves, increase the risk for immobility and falls. Physiological reserves in many elderly people may be adequate at normal homeostasis. However, the introduction of a stressor, such as an illness requiring acute hospitalization, may deplete these reserves, even after the precipitating cause has been corrected. Concurrent use of psychoactive substances and the increased vulnerability to their adverse effects may exacerbate this course. Additionally, conditions which increase in frequency and severity with old age, such as chronic obstructive pulmonary disease, have increased rates of adverse events among users of benzodiazepines and opioids. Increasing age also increases morbidity and multi-morbidity, which in turn is associated with polypharmacy (37), with a concurrent increased risk for drug-drug interactions and other adverse effects. While physicians may be aware of these factors, there is evidence that diversion from clinical guidelines for the prescription of psychoactive medication, in particular benzodiazepines, is widespread (2).
Regarding self-reported use of benzodiazepines, opioids, z-hypnotics and illicit drugs, there appears to be a high negative predictive value if a patient reports non-use. However, if a sample is positive, under-reporting seems to be prevalent. This differs based on substance category. Patient uncertainty regarding medication list may contributeand lead to under-reporting, as well as any potential stigma associated with illicit substance or psychoactive medication use.
The co-occurrence of harmful alcohol use and non-medical prescription drug use is also well established.. Our study demonstrates primarily increased rates of benzodiazepine-use in the upper AUDIT-4 ranges. Whether this reflects co-occurring non-medical prescription drug use or as a reflection of treatment of symptoms related to harmful alcohol use is uncertain.
## Clinical implications
The above considerations, combined with the findings from our study, should prompt an increased awareness regarding the use of psychoactive substances among acutely hospitalized patients. A reduction in inappropriate use or misuse of psychoactive substances may have a positive impact on hospitalization rates, risk of re-admittance, and health care costs. Specifically, strengthening and implementing the following practices may contribute to improved patient outcomes: -Screening for inappropriate or harmful psychoactive substance use -Critically assessing the indications for psychoactive drug use and considering discontinuation or tapering if inappropriate prescribing is present -Assess the degree of increased risk for adverse events in specific conditions when using psychoactive substances, such as chronic obstructive pulmonary disease -Assess concomitant alcohol use if psychoactive substance use is present Hospital admissions may serve as opportunities to screen for unapparent harmful use or misuse of psychoactive substances, but this practice is less likely to be performed among elderly patients than among younger individuals (29). Health professionals may experience hesitation in screening for unhealthy psychoactive substances use due to the perceived sensitive nature of the subject (41), but guidelines for proper communicative techniques exist, which may ameliorate any associated stigma. Due to the commonality of substance use disorders among acutely admitted patients, clinicians may benefit from assessing psychoactive medication use in addition to alcohol and illicit drugs, particularly among the elderly. Similar to screening for alcohol, validated singlequestion screening tools exist for illicit drug use and non-medical use of psychoactive medication. However, these tools were validated in a primary care setting. Older adults may also perceive symptoms of harmful psychoactive substance use as symptoms of aging, and adverse effects stemming from psychoactive drug use may be similar to manifestations of other conditions (29). Single-question screening tools may therefore better serve as an adjunct to the general clinical assessment of every admitted patient.
The use of psychoactive medication is usually mediated by health care professionals, which mandates adherence to prescription guidelines. However, based on prescription databases, inappropriate use of benzodiazepines and z-hypnotics is widespread. Hospital clinicians have the opportunity to prudently assess medication use among admitted patients, and are in a position to inform and co-operate with primary care physicians or other health professionals. Any identified instances of harmful or inappropriate use may then be managed by targeted interventions, or discontinuation or drug-tapering. In the event of specific illnesses, it is important to consider whether harmful, inappropriate or even therapeutic use may have precipitated an adverse event. As several illnesses that are more commonly occurring in old age progress, physiological reserves may be depleted. For instance, respiratory depression associated with benzodiazepine use (47) may cause adverse respiratory events in patients with chronic pulmonary obstructive disease if pulmonary function continues to decrease.
Brief interventions for substance use disorders have shown varying degrees of efficacy in the reduction of harmful alcohol use (48-50) and illicit drug use, with some studies showing aF o r p e e r r e v i e w o n l y reduction in at-risk drinking, while brief interventions have shown no increased efficacy in reducing illicit drug misuse compared to control groups. These studies also did not include elderly patients (above 64 years of age), nor the critically ill. Additionally, study participation and the screening process itself may cause a reduction in harmful substance use in both the control group and the groups receiving brief interventions (53).
While the above recommendations may be valid for Norway and other similar countries in the West, the number of samples positive for psychoactive substances was markedly lower across all age groups in Moscow, at 12 %. Plausible causes may be differences in prescription guidelines and legislative practices. Additionally, there were significant differences in the cut-off values utilized at both sites -in Moscow, the S/N-ratio cut-off was set at 1000, versus an S/N-ratio set at 10 in Oslothis may have underestimated the number of positive samples. However, the usage frequency of psychoactive medications in a Russian ED-population may still warrant prudence regarding adherence to prescription guidelines and adverse effects. Their use may also reflect a different selection of illnesses in that particular population -further discussion regarding these elements is beyond the scope of this article.
# Limitations
Previous studies have utilized both self-reporting and registry based data to assess substance use among acutely hospitalized patients -a major strength in our study is the addition of blood sample analysis. The analytical methods employed have allowed us to identify individual substances, thus increasing the descriptive accuracy in our data. Our inclusion criteria were broad, increasing the likelihood for a representative sample size -this is also strengthened by a high participation rate. The use of two to different hospital sites in two different nations increased our sample size; however, external generalizability to other ED-populations is limited to due to differences in medical practice and cultural norms. Differences in analytical cut-off values between Oslo and Moscow also limit the comparability between the two data sets. In order to satisfy necessary ethical considerations, patients unable to give informed consent were excluded. Their cognitive limitations may have been caused by or exacerbated by psychoactive substances or alcohol use -consequently, prevalence data for this patient group remains unavailable. The prevalence of substance use among patients who declined to participate is also a potential bias, as they may have been unwilling to participate due to concerns regarding the disclosure of substance or alcohol use, or may have been incapacitated due to illicit drug use.The use of psychoactive substances among patients acutely admitted to Departments of Internal Medicine is common, and may cause both readily identifiable as well as oblique adverse effects. Health professionals should be aware that the use of benzodiazepines and opiates is highly prevalent, particularly among elderly patients, and should exact prudence when prescribing these drug classes. Implementing screening practices may improve the identification of inappropriate psychoactive substance use.
# Conclusion
# Author disclosures
Contributorship statement DG drafted the manuscript. BJ and SK did the initial data analysis. TB performed supplementary data analysis. BJ, SK, VV, DG, AL, GN and STB organized or contributed to the Oslo data collection. EB, EK, AN, ET, AK, AP, SP and ED organized or contributed to the Moscow data collection. All authors were responsible for study design, interpretation of the results, as well as final revision and approval of the manuscript.
## Competing interests
## No conflict of interest declared
# Funding
The study was supported by the Norwegian Ministry of Health's Funding for Norwegian-Russian Collaboration in Health and Social issues grant B-1408. The Norwegian Ministry of Health had no role in the design of the study, in the collection, analysis or interpretation of the data, nor in writing the manuscript or in the decision to submit the article for publication.
## Data sharing
No additional data will be shared.a: Cocaine-positive is defined as a sample positive for either cocaine or benzoylecgonine. Because the recovery of benzoylecgonine is very low using SLE extraction, a patient may be positive for cocaine, but negative for benzoylecgonine b: M/C = morphine/codeine-ratio c: Alcohol-positive is defined as blood alcohol content >0,2 gram/kg: a: In Moscow, several other substances were analyzed in addition to those listed in table 2 -these were omitted due to lack of positive samples. b: In Moscow, the detection limit was based on a signal/noise -ratio above 1000, which applies for all substances listed. In Oslo, absolute lower detection limits were utilized. c: Alcohol-positive is defined as blood alcohol content >0,2 gram/kg:
Medicinal drugs are defined as the presence of benzodiazepines, z-hypnotics, or opioids: Medicinal drugs are defined as the presence of benzodiazepines, z-hypnotics, or opioidsA: Medicinal drugs defined as benzodiazepines, z-hypnotics or opioidsB: Medicinal drugs defined as benzodiazepines, z-hypnotics, barbiturates or opioidsZ-hypnotics are almost completely unavailable in Russia due to various regulatory practices, and there are no drug classes typically marketed as "sleeping medication". Barbiturates and benzodiazepines are generally used to treat both anxiety as well as sleep disorders.Abstract Objectives: The use of psychoactive prescription medication is increasing in the general population. This is a cause for concern, particularly among the elderly, where physiological changes related to senescence increase the risk for adverse effects. While previous studies regarding psychoactive substance use have generally been population based, we sought to determine the frequency of such use among acutely hospitalized patients. Setting: Two emergency departments admitting patients to Departments of Internal Medicine -one in Oslo, one in Moscow. Participants: 5583 patients aged ≥18 years participated, distributed evenly between genders and study locations. Patients unable to give informed consent were excluded. The study sites did not admit patients with surgical conditions and/or injuries. Primary and secondary outcomes: The presence of psychoactive substances was determined through the use of blood analysis utilizing liquid chromatography-mass spectrometry. Secondary outcomes comprised demographic data (including age, gender, employment and marital status), degree of psychological distress, concurrent alcohol use, and self-reported alcohol, psychoactive drug and illicit substance use.
# Strengths and limitations
-Blood sample analysis was utilized in addition to self-reporting to accurately assess the prevalence of psychoactive substance use -Large sample size divided across two sites, with an inclusion period lasting one year -Differences in laboratory methods, legislation and prescription practices for psychoactive substances between the selected sites limit comparabilityn l y
# Introduction
Background While the deleterious effects of harmful alcohol use and tobacco smoking are well known, several studies have also called attention to increasing rates of psychoactive prescription drug use in Western countries, especially among the elderly. These drugs include opioids, benzodiazepines and z-hypnotics. In addition to increased usage rates, there seems to be an apparent increase in the non-medical use and abuse of psychoactive medication (4), as well as inappropriate prescription practices and diversion from clinical guidelines.
Several studies have prospectively assessed the impact and usage rates of alcohol, psychoactive medication and illicit substances in the emergency department . Alcohol-related presentations to the emergency department have been quantified both through the use of questionnaires, and classification through the International Statistical Classification on Diseases and Related Health Problems-system (ICD-10). However, patients may under-report alcohol consumption when presented with a questionnaire, and self-reporting may be limited by recall bias . Potentially harmful, but not clinically apparent alcohol-related presentations may also remain undetected by health professionals, and misclassified. These limitations may attenuate the accuracy of the prevalence data in such studies. A study published in 2011 (10) utilized blood sample analysis using highly selective methods to assess the prevalence of psychoactive substances among patients acutely admitted due to injury -this provided accurate and reliable prevalence data. However, blood sample analysis has not been employed to assess the prevalence of psychoactive substances among patients acutely admitted to Departments of Internal Medicine.
Patients admitted to Departments. of Internal Medicine are likely to be older than the general population(11). Due to age-related changes in the pharmacodynamics and pharmacokinetics properties of various drugs, the elderly are also more vulnerable to any potential adverse effects from psychoactive medication use, as well as alcohol . Changes in physiology and body composition related to aging increases the susceptibility to adverse effects from alcohol consumption , and the concurrent use of psychoactive drugs and alcohol also exacerbates the respective adverse effects of each substance. Drowsiness and reduced postural stability are common effects of psychoactive drugs, and are enhanced by alcohol consumption. Consequently, both drug-drug and drug-alcohol interactions may cause clinically significant adverse reactions in the elderly, such as falls, injury and infections (1). These events may be precipitated even by low-tomoderate doses of the substances mentioned. Polypharmacy -the concurrent use of multiple medications -increases in frequency with age, with rates ranging from 12 % to 35 % among the elderly in Western countries . This may contribute to the additive risk for adverse drug-drug interactions. Considering the high frequency of use among older adults, knowledge regarding the prevalence of these substances among acutely hospitalized patients may be of clinical significance. Psychoactive medication, specifically benzodiazepines and z-hypnotics, is also often used to treat psychological distress and sleep disorders, which are common among the elderly . Contrary to alcohol, which is usually independently procured, and where ingestion is user-initiated, the distribution of psychoactive medication is generally mediated by health professionals. In light of increased rates of non-medical use and abuse of these drugs (4), the presence of reliable andaccurate prevalence data may mediate prudence regarding prescription practices, and lead to more critical appraisals of when these drugs are necessary, or should be discontinued.
## Study aims
In this study, we sought to assess the prevalence of psychoactive substances among patients acutely hospitalized to Departments of Internal Medicine.
Specifically, we wished to: 1) Measure the rate of psychoactive substance use through the use of blood sampling and highly selective analytical methods 2) Investigate patient characteristics when a psychoactive substance is detected; namely age, gender, employment status, marital status, degree of psychological distress, and selfreported alcohol consumption. 3) Investigate the concurrent rate of alcohol and illicit drug use when psychoactive substances are detected
This study is part of larger collaborative project between Oslo University Hospital and the Moscow Research Center on Addiction.
## Study design, setting and participants
## Design and site selection
We have employed an observational, cross-sectional approach. We chose two study sites -Lovisenberg Diaconal Hospital (LDS) in Oslo, Norway (site 1) and Hospital No. 68 in Moscow, Russia (site 2). Both sites are medium-sized urban hospitals. The mean life expectancy in Norway is 84.2 years for women and 80.6 years for men (17), which is significantly higher than in Russia, where the mean life expectancy is 72.1 years . However, due to socio-economic disparities within the various districts in Oslo, the life expectancy in several populations belonging to the catchment area for LDS approaches that in Russia, ranging from 72 -76 years for men and 79 -81 years for women, (19). Due to differences in laboratory techniques, prescription practices, as well as guidelines and legislation for psychoactive drug use in each nation, results for each site will be presented separately.
## Ethics
This study was approved by the Regional Ethics Committee for South Eastern Norway (2015/2404), and by the Moscow Local Ethical Committee of MRPCA.
## Inclusion and exclusion criteria
Patients were generally either recruited by emergency department(ED)-nurses while awaiting ward transfer or examinations such as radiographic imaging, or shortly after arrival to their respective ward in Oslo, while in Moscow, patients were recruited upon arrival to the ward by physicians serving as dedicated study and recruitment coordinators. The EDs in Oslo and Moscow were both located at the Department of Internal Medicine, and therefore did not admit patients with injuries or surgical conditions. Patients were transferred to an appropriate ward after an initial evaluation, which in Oslo was General Internal Medicine, Infectious Diseases, Cardiology, Pulmonary Medicine, Cerebrovascular and Geriatric conditions, Hematology, or Gastroenterology, and in Moscow General Internal Medicine, Cardiology, Pulmonary Medicine, or Neurology. Only patients aged 18 years and above and able to give informed consent were included. Patients unable to consent upon admission due reversible or transient causes (such as intoxication or severe illness) were approached at a later time by dedicated research assistants, and were offered to participate when they were able to consent. The inclusion period lasted from November 2016 to December 2017, and patients were included at all hours of the day, to account for any seasonal or diurnal variations. We employed the following exclusion criteria: 1) Permanently unable to give an informed consent. 2) Elective transfer from other hospitals. 3) Limited or no ability to read or write the national language, e.g. Russian or Norwegian, as translated copies of both the consent form and accompanying questionnaire were unavailable.detail the flow of patient recruitment.Patient and public involvement There was no patient or public involvement in the design and implementation of this study. In addition to disseminating results to clinicians and policy makers, we intend to inform and involve patient interest organizations in the field of substance use disorders.
## Variables
Study design, variables and analytical methods were jointly developed by the research teams at both study sites, in order to maintain consistency in data collection and collation. The study sites were also visited by the respective study in order to ensure methodological continuity.
# Blood analysis
Whole blood was drawn from every study participant and secured in 5 ml BD-Vacutainer blood collection glass tubes (BD Vacutainer Systems, Franklin, NJ, USA). The blood samples were stored at 4°C for a maximum of 7 days before analysis. In Oslo, the samples were analyzed at the Division of Laboratory Medicine, Section of Drug Abuse Research, Oslo University Hospital. We prepared the samples by 96-well supported liquid extraction, which were then analyzed by liquid chromatography -tandem mass spectrometry (LC-MS/MS). This is the same sample preparation and LC-MS/MS-method used by Kristoffersen et al . Stable isotope-labeled internal standards were used for all compounds. Two multiple reaction monitor (MRM) transitions were used for each compound and each internal standard. The chosen detection limits (cut-off limit) were based on validation data and signal/noise (S/N) ratio of the quantifier MRM transition (S/N ≥ 10). For THC S/N ≥ 3 was used. The Russian samples were analyzed at the Moscow Research Center for Addiction. Blood samples were prepared by protein precipitation and analyzed by UHPLC-MS/MS. Two MRM transitions were used for each compound. The cut-off level for each analyte was defined as a signal corresponding to S/N = 1000 for the quantifier MRM transition. We have provided additional methodological details and validation data for the Oslo site inAny sample in which the patient was administered a psychoactive medicinal drug by health care personnel prior to blood sampling was omitted from the results -specifically, morphine and diazepam at the Oslo site, and diazepam, phenazepam, and tramadol at the Moscow site. In order to distinguish medicinal morphine use from heroin use, a morphine/codeine-ratio cut-off value of >1 was utilized. Similarly, concentration ratios of methamphetamine to amphetamine were used to distinguish methamphetamine use from various mixtures of the two stimulants.
## Questionnaire
Upon enrolment, every patient filled out a questionnaire containing demographic data: age ranges (18-40, 41-60, 61-70, and ≥71 years), gender (male/female), marital status (single, married or widowed), and employment status (employed, retired, unemployed, or student). Psychological distress was measured using the Symptom Checklist 5 (SCL-5), a short-form screening tool which performs almost as well as the full version (23). A value above 2 points indicates psychological distress. Self-reported alcohol use was measured through the Alcohol Use Disorder IdentificationTest -4 (AUDIT-4), which has been validated as a simple, yet effective, screening tool for identifying alcohol use disorders. A value of 7 points or more for men and 5 points or more for women indicates at-risk alcohol consumption patterns. Both questionnaires have been validated for use in their respective nations. The questionnaires were filled out by the patients themselves, with occasional aid from ER-nurses at the Oslo site if questions arose. The ER nurses were also instructed beforehand by the study coordinator, and a data collection pilot preceded the inclusion period. We also included self-reported use of benzodiazepines and/or barbiturates (termed `sedatives`), zhypnotics (termed `sleeping medication`), opioids (termed `painkillers`) and illicit drugs (termed `narcotic substances`). All groups had example trade or colloquial names to improve question comprehension. Questionnaires were manually entered, controlled, and later transferred to the electronic project database.
## Psychoactive substances
Our primary outcome was the presence of a psychoactive substance, determined through analytical cut-off values (shown in tables 1 and 2). Psychoactive medicinal drugs were defined as opioids, benzodiazepines, z-hypnotics, or barbiturates. The specific drugs were selected based on prescription practices among primary care and hospital physicians, as well as hospital usage rates. Due to differences in prescription practices, certain substances were omitted at the Moscow site, and vice versa for the Oslo site. Specifically, the use of z-hypnotics in Russia is very uncommon, and phenazepam, a commonly prescribed benzodiazepine in Russia, is not marketed in Norway (29). Barbiturates are also commonly used as a sedative in Russia, but this use is uncommon in Norway. The presence of alcohol was determined through direct analysis of blood alcohol content. Based on population studies, the most common illicit drugs and their metabolites were analyzed -these comprise amphetamines, methamphetamines, cocaine, MDMA/ecstasy, heroin and tetrahydrocannabinol. So-called "new psychoactive substances (NPS)" were not analyzed -the rapid introduction of new compounds poses significant analytical challenges, and selecting which substance to analyze is complicated by limited knowledge regarding their availability in the general population. Polyuse was defined as the presence of two or more psychoactive substances (prescription medication and/or illicit drugs) in any given blood sample.
# Statistical analysis
We analyzed the data using IBM SPSS 25.0 (Armonk, NY). The prevalence of various psychoactive substances is shown through descriptive tables. We employed bi-variate cross tables to analyze associations between the presence of a psychoactive substance and sample characteristics. Chi-square statistics with odds ratios (ORs) were used to assess statistical significance. Our primary dependent variables were the presence of psychoactive substances, illicit drugs, and polyuse. Our co-variates were age ranges, marital status, employment status, positive blood alcohol content, presence of illicit drugs, polyuse, presence of psychological distress, and selfreported alcohol use. We performed logistic regression analyses to test the association between our dependent values, which were medicinal drug use, illicit drug use and polyuse, and independent covariates, which were age, gender, SCL-5-values, marital status, employment status and AUDIT-4zones. Our reference categories for each variable were age 18-40 years, male gender, SCL-5 score ≤2, married, active in the workforce and an AUDIT-score of 0-3. The level of significance was set at p < 0.05. Continuous data are presented as means, with 95 % confidence intervals (CIs).
## Participants
A total of 2874 patients in the Oslo site and 3009 patients in the Moscow site were included, with a participation rate of 81 % and 91 % respectively -47.6 % were women in Oslo and 52.9 % were women in Moscow. Distribution across age ranges was similar at both sites (tables 3 and 4).
## Prevalence data and sample characteristics prevalence data
Tables 1 and 2 depict the main findings in our study, showing prevalence data for all psychoactive medicines analyzed and their analytical cut-off values. In Oslo, psychoactive substances were detected in 32.3 % of analyzed samples. The most common drug class detected was opioids (19.5 %), followed by benzodiazepines . Z-hypnotics were found in 6.6 % of samples. In the Moscow site, psychoactive substances were detected in 12 % of samplesbenzodiazepines were the most frequently detected drug class (9.7 %). The rates for other drug classes were low, with opioids found in 0.9 %.
In Oslo, illicit drugs were detected in 6.4 % of all samples, with tetrahydrocannabinol (3.7 %) being the most common. In the Moscow site, the total rate for illicit drug use was lower compared to the Oslo site, at 1.4 %, with THC being the most common (1.1 %). The rates for the remaining substances in both sites were low.
Cut-off values at each of the two sites were not identical due to differences in analytical methods, with a lower S/N-ratio at the Oslo site. We have therefore provided mean concentrations for the substances detected at the Oslo site, which in the majority of substances are well above the analytical cut-off.show the un-adjusted distribution for psychoactive medicinal drug users and nonusers across all co-variates measured in Oslo and Moscow respectively. Medicinal drug use was evenly distributed among genders at both sites. In Oslo, the highest proportion of medicinal drug use was found among patients aged older than 40 years, as well as among those being retired, and/or in the lowest AUDIT-4 ranges. In Moscow, medicinal drug use was more evenly distributed across all age groups, with the highest proportion of use among patients that were non-active in the workforce.depicts logistic regression analysis for various co-variates in the Oslo site. Compared to patients aged 18-40, the likelihood for medicinal drug use was higher among patients aged 61-70 and 71 years or older. There was no significant gender difference among medicinal drug users. The likelihood for polyuse was higher among non-employed patients compared to patients being active in the workforce -this was also observed in regards to illicit substances as well as medicinal drug use. In the Oslo site, female patients were also significantly less likely to use illicit drugs than male patients. Patients aged 18-40 were more likely to use illicit drugs than other age groups. The use of psychoactive substances, illicit drugs, and polyuse was significantly more likely in patients scoring above the cut-off value for SCL-5 (>2).
## Sample characteristics
In the Moscow site, similar trends for SCL-5 were observed, where patients scoring above the cut-off value were more likely to use illicit drugs. However, there was no significant difference in the likelihood for polyuse and medicinal drugs. Contrary to the Oslo site, the presence of psychoactive substances and polyuse was significantly less likely in patients aged >40 years. Being retired increased the likelihood for both psychoactive substance use and polyuse compared to patients being actively employed -the same association was not observed for patients being nonactive in the workforce.show the distribution of self-reported use if a sample is positive for either benzodiazepines, opioids, z-hypnotics and illicit drugs, with both benzodiazepines and barbiturates categorized as `sedatives` in Moscow. If a sample is negative, this is generally mirrored in questionnaires regarding self-reported use. The highest concordance between positive samples and confirmatory self-reported use in Oslo was for z-hypnotics (75.8 %), followed by opioids (62.6 %), illicit drugs (49.4 %). and benzodiazepines (48.9 %). In Moscow, there was a high concordance between negative samples for opioids and illicit drugs, and self-reported non-use. This percentage was lower for sedatives (83.2 % for barbiturates and 83.8 % for benzodiazepines). The percentage of sample-positive self-reported use ranged from 7.1 % for opioids to 55.9 % for barbiturates.shows the distribution of psychoactive substances within different AUDIT-4 ranges in Oslo. The highest prevalence of benzodiazepine-use was found in the highest AUDIT-4 range (9-16 points), with a rate of 31 %, compared with 18.5 % for the whole study population. The use of opioids is more prevalent at the lowest and highest AUDIT-4 ranges. The use of THC, stimulants and z-hypnotics is evenly distributed along all AUDIT-4 ranges. In Moscow, the highest rate of benzodiazepine use is also found among the highest AUDIT-4 range, while the usage rate for other psychoactive substances is uniformly low in all other AUDIT-4 ranges.
# Discussion
Our study demonstrates that the use of psychoactive substances among patients admitted to Departments of Internal Medicine is highly prevalent, while simultaneously highlighting differences in sample characteristics between Oslo and Moscow. Based on previous studies, the volume of participants and rate of inclusion provide representative, accurate data. The number of positive sample for psychoactive substances was far higher in Oslo than in Moscow. The use of psychoactive substances in Oslo was significantly associated with older age, psychological distress, and being unemployed. The latter two characteristics are known to be associated with substance use.There was no significant gender disparity among positive samples -this may be explained by a higher percentage of positive samples among younger males compared to females, and vice versa for older females compared to older males. Previous population-based studiesalso show an apparent increase in the use of psychoactive prescription medication in the general population, particularly among the elderly. However, the drug use statistics were extracted from national prescription databases, and whether or not the patient took the prescribed drug as well as the duration of drug use was not measured. Additionally, substance abuse disorders are more prevalent among non-participants in health surveys. These factors may underestimate the prevalence of psychoactive substance use -as far as we are aware, this is the first study to demonstrate the frequency of substance use among internal medicine patients.
The high rate of psychoactive substance use among the elderly is a cause for concern. The changes in pharmacodynamics and pharmacokinetics associated with senescence increase complexity when prescribing medications to the elderly. Inter-individual variability to drugs increases with age, partly explained by reduced physiological reserves during stressors such as inter-current illness or disease. Half-life of elimination is prolonged by changes in body composition, affecting the volume of distribution of lipid-soluble drugs, and reduced renal and hepatic clearance . These factors alter and increase the sensitivity to psychoactive substances. Elderly patients are particularly sensitive to the effects of benzodiazepines on the central nervous system. Sedation may occur at lower doses than in younger patients. Drowsiness and decreased motor skills, combined with the age-related reduction in functional reserves, increase the risk for immobility and falls. Physiological reserves in many elderly people may be adequate at normal homeostasis. However, the introduction of a stressor, such as an illness requiring acute hospitalization, may deplete these reserves, even after the precipitating cause has been corrected. Concurrent use of psychoactive substances and the increased vulnerability to their adverse effects may exacerbate this course. Additionally, conditions which increase in frequency and severity with old age, such as chronic obstructive pulmonary disease, have increased rates of adverse events among users of benzodiazepines and opioids. Increasing age also increases morbidity and multi-morbidity, which in turn is associated with polypharmacy (40), with a concurrent increased risk for drug-drug interactions and other adverse effects. While physicians may be aware of these factors, there is evidence that diversion from clinical guidelines for the prescription of psychoactive medication, in particular benzodiazepines, is widespread (2). Regarding self-reported use of benzodiazepines, opioids, z-hypnotics and illicit drugs, there appears to be a high negative predictive value if a patient reports non-use. However, if a sample is positive, under-reporting seems to be prevalent. This differs based on substance category. Patient uncertainty regarding medication their list may contributeand lead to under-reporting, as well as any potential stigma associated with illicit substance or psychoactive medication use.
The co-occurrence of harmful alcohol use and non-medical prescription drug use is also well established.. Our study demonstrates primarily increased rates of benzodiazepine-use in the upper AUDIT-4 ranges. Whether this reflects co-occurring non-medical prescription drug use or as a reflection of treatment of symptoms related to harmful alcohol use is uncertain.
## Clinical implications
The above considerations, combined with the findings from our study, should prompt an increased awareness regarding the use of psychoactive substances among acutely hospitalized patients. A reduction in inappropriate use or misuse of psychoactive substances may have a positive impact on hospitalization rates, risk of re-admittance, and health care costs. Specifically, strengthening and implementing the following practices may contribute to improved patient outcomes: -Screening for inappropriate or harmful psychoactive substance use -Critically assessing the indications for psychoactive medicinal drug use and considering discontinuation or tapering if inappropriate prescribing is present -Assess the degree of increased risk for adverse events in specific conditions when using psychoactive substances, such as chronic obstructive pulmonary disease -Assess concomitant alcohol use if psychoactive substance use is present Hospital admissions may serve as opportunities to screen for unapparent harmful use or misuse of psychoactive substances, but this practice is less likely to be performed among elderly patients than among younger individuals. Health professionals may experience hesitation in screening for unhealthy psychoactive substances use due to the perceived sensitive nature of the subject (44), but guidelines for proper communicative techniques exist, which may ameliorate any associated stigma. Due to the commonality of substance use disorders among acutely admitted patients, clinicians may benefit from assessing psychoactive medication use in addition to alcohol and illicit drugs, particularly among the elderly. Similar to screening for alcohol, validated singlequestion screening tools exist for illicit drug use and non-medical use of psychoactive medication (49). However, these tools were validated in a primary care setting. Older adults may also perceive symptoms of harmful psychoactive substance use as symptoms of aging, and adverse effects stemming from psychoactive drug use may be similar to manifestations of other conditions. Single-question screening tools may therefore better serve as an adjunct to the general clinical assessment of every admitted patient.
The use of psychoactive medication is usually mediated by health care professionals, which mandates adherence to prescription guidelines. However, based on prescription databases, inappropriate use of benzodiazepines and z-hypnotics is widespread. Hospital clinicians have the opportunity to prudently assess medication use among admitted patients, and are in a position to inform and co-operate with primary care physicians or other health professionals. Any identified instances of harmful or inappropriate use may then be managed by targeted interventions, or discontinuation or drug-tapering. In the event of specific illnesses, it is important to consider whether harmful, inappropriate or even therapeutic use may have precipitated an adverse event. As several illnesses that are more commonly occurring in old age progress, physiological reserves may be depleted. For instance, respiratory depression associated with benzodiazepine use (50) may cause adverse respiratory events in patients with chronic pulmonary obstructive disease if pulmonary function continues to decrease.
Brief interventions for substance use disorders have shown varying degrees of efficacy in the reduction of harmful alcohol use (51-53) and illicit drug use (54, 55), with some studies showing a reduction in at-risk drinking, while brief interventions have shown no increased efficacy in reducing illicit drug misuse compared to control groups. These studies also did not include elderly patients (above 64 years of age), nor the critically ill. Additionally, study participation and the screening process itself may cause a reduction in harmful substance use in both the control group and the groups receiving brief interventions (56).
While the above recommendations may be valid for Norway and other similar countries in the West, the number of samples positive for psychoactive substances was markedly lower across all age groups in Moscow, at 12 %. Plausible causes may be differences in prescription guidelines and legislative practices. Additionally, there were significant differences in the cut-off values utilized at both sites -in Moscow, the S/N-ratio cut-off was set at 1000, versus an S/N-ratio set at 10 in Oslothis may have underestimated the number of positive samples. However, the usage frequency of psychoactive medications in a Russian ED-population may still warrant prudence regarding adherence to prescription guidelines and adverse effects. Their use may also reflect a different selection of illnesses in that particular population -further discussion regarding these elements is beyond the scope of this article.
# Limitations
Previous studies have utilized both self-reporting and registry based data to assess substance use among acutely hospitalized patients -a major strength in our study is the addition of blood sample analysis. The analytical methods employed have allowed us to identify individual substances, thus increasing the descriptive accuracy in our data. Our inclusion criteria were broad, increasing the likelihood for a representative sample size -this is also strengthened by a high participation rate. The use of two to different hospital sites in two different nations increased our sample size; however, external generalizability to other ED-populations is limited to due to differences in medical practice and cultural norms. Differences in analytical cut-off values between Oslo and Moscow also limit the comparability between the two data sets. In order to satisfy necessary ethical considerations, patients unable to give informed consent were excluded. Their cognitive limitations may have been caused by or exacerbated by psychoactive substances or alcohol use -consequently, prevalence data for this patient group remains unavailable. The prevalence of substance use among patients who declined to participate is also a potential bias, as they may have been unwilling to participate due to concerns regarding the disclosure of substance or alcohol use, or may have beenincapacitated due to illicit drug use. Finally, substances with a long half-life (such as certain benzodiazepines and z-hypnotics) may have a detection window longer than 24 hours -this may result in a substance-positive sample, but a negative reply regarding self-reported ingestion the past 24 hours.
# Conclusion
The use of psychoactive substances among patients acutely admitted to Departments of Internal Medicine is common, and may cause both readily identifiable as well as oblique adverse effects. Health professionals should be aware that the use of benzodiazepines and opiates is highly prevalent, particularly among elderly patients, and should exact prudence when prescribing these drug classes. Implementing screening practices may improve the identification of inappropriate psychoactive substance use.
# Author disclosures
# Contributorship statement
DG drafted the manuscript. BJ, AAN and SK did the initial data analysis. TB performed supplementary data analysis. BJ, SK, VV, DG, AL, GN and STB organized or contributed to the Oslo data collection. EB, EK, AN, ET, AK, AP, SP and ED organized or contributed to the Moscow data collection. All authors were responsible for study design, interpretation of the results, as well as final revision and approval of the manuscript.
## Competing interests
## No conflict of interest declared
# Funding
The study was supported by the Norwegian Ministry of Health's Funding for Norwegian-Russian Collaboration in Health and Social issues grant B-1408. The Norwegian Ministry of Health had no role in the design of the study, in the collection, analysis or interpretation of the data, nor in writing the manuscript or in the decision to submit the article for publication.
## Data sharing
No data are available due to institutional policies.a: Cocaine-positive is defined as a sample positive for either cocaine or benzoylecgonine. Because the recovery of benzoylecgonine is very low using SLE extraction, a patient may be positive for cocaine, but negative for benzoylecgonine b: M/C = morphine/codeine-ratio >1 c: Alcohol-positive is defined as blood alcohol content >0,1 gram/kg: a: In Moscow, several other substances were analyzed in addition to those listed in table 2 -these were omitted due to lack of positive samples. b: In Moscow, the detection limit was based on aS/N-ratio above 1000, which applies for all substances listed. In Oslo, absolute lower detection limits were utilized. c: Alcohol-positive is defined as blood alcohol content >0,1 gram/kg d: Cocaine-positive is defined as a sample positive for either cocaine or benzoylecgonine. Because the recovery of benzoylecgonine is very low using SLE extraction, a patient may be positive for cocaine, but negative for benzoylecgonine e: In instances of substances being detected in a very small number of samples, percentages are presented with 2 decimal points.: A: Medicinal drugs are defined as the presence of benzodiazepines, z-hypnotics, or opioids: A: Medicinal drugs are defined as the presence of benzodiazepines, z-hypnotics, barbiturates or opioidsA: Medicinal drugs defined as benzodiazepines, z-hypnotics or opioidsA: Medicinal drugs defined as benzodiazepines, barbiturates or opioidsA: Z-hypnotics are almost completely unavailable in Russia due to various regulatory practices, and there are no drug classes typically marketed as "sleeping medication". Barbiturates and benzodiazepines are generally used to treat both anxiety as well as sleep disorders.F o r p e e r r e v i e w o n l yf For some compounds the calibrator concentration range did not include concentration as low as cut-off level, but QC samples with a concentration at cut-off was included for all compounds.
## Page 37 of 37
For peer review only -http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ OpenQuantitative variables 11 Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and why(p. 6-7)
(a) Data was analysed using SPSS. Bi-variate cross tables were used to analyse associations. Chi-square statistics with odds-rations were used to assess statistical significance. (p. 8)
# Statistical methods 12
Continued on next page Generalisability 21 Detailed in "Discussion"-chapter and "Conclusion (p. 9-12)
Other information/author disclosures The study was supported by the Norwegian Ministry of Health's Funding for Norwegian-Russian Collaboration in Health and Social issues grant B-1408. The Norwegian Ministry of Health had no role in the design, data collection, analysis or interpretation of the data, nor in the writing of the report or decision to submit the article for publication (p.12) *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and unexposed groups in cohort and cross-sectional studies. |
Comparison of dynamic and static spacers for the treatment of infections following total knee replacement: a systematic review and meta-analysis
Background: Revision surgery is the most common treatment for patients who develop infection after total knee arthroplasty (TKA). Two types of spacers are often used in revision surgery: dynamic spacers and static spacers. The comparative efficacy of these two types of spacers on knee prosthesis infections is not well established. Therefore, we carried out a systematic evaluation and meta-analysis with the aim of comparing the difference in efficacy between dynamic and static spacers.Methods:We conducted the literature search in PubMed, Web of Science, Cochrane Library, and Embase databases. The articles searched were clinical study comparing the difference in efficacy between dynamic spacers and static spacers for the treatment of prosthetic infections occurring after total knee arthroplasty.Results:We conducted a literature search and screening based on the principles of PICOS. Ultimately, 14 relevant clinical studies were included in our current study. We use infection control rate as the primary evaluation indicator. The KSS knee scores (KSSs), KSS functional scores, bone loss and range of motion (ROM) are secondary indicators of evaluation. Thirteen of these included studies reported the infection control rates, with no significant difference between dynamic and static shims (RR: 1.03; 95% Cl 0.98, 1.09; P = 0.179 > 0.05). The KSSs were reported in 10 articles (RR: 5.98; 95% CI 0.52, 11.43; P = 0.032 < 0.05). Six articles reported the KSS functional scores (RR: 13.90; 95% CI 4.95, 22.85; P = 0.02 < 0.05). Twelve articles reported the ROM (RR: 17.23. 95% CI 10.18, 24.27; P < 0.0001). Six articles reported the bone loss (RR: 2.04; 95% CI 1.11, 3.77; P = 0.022 < 0.05).Conclusion: Current evidence demonstrates that dynamic spacers are comparable to static spacers in controlling prosthetic joint infection. In terms of improving the functional prognosis of the knee joint, dynamic spacers are more effective than static spacers.
# Background
For patients with severe knee injuries, total knee arthroplasty (TKA) can effectively improve the function of the knee joint, relieve knee pain, and improve the quality of life of patients. Knee prosthesis joint infection (PJI) is one of the most terrible complications after the total knee arthroplasty, and it is often the main reason for the failure of the total knee arthroplasty [bib_ref] Infection and periprosthetic fracture are the leading causes of failure after aseptic..., Meyer [/bib_ref] [bib_ref] Periprosthetic joint infection is the main cause of failure for modern knee..., Koh [/bib_ref]. The probability of prosthesis joint infection after primary knee replacement varies from 0.4 to 3% [bib_ref] Infection risk assessment in patients undergoing primary total knee arthroplasty, Poultsides [/bib_ref] [bib_ref] Functional recovery after twostage short-interval revision of chronic periprosthetic knee joint infection, Madarevic [/bib_ref]. There are many risk factors for PJI in knee prosthesis infection. Age, duration of surgery, diabetes, urinary tract infection and rheumatoid arthritis are all important risk factors for PJI after TKA surgery [bib_ref] Risk factors for periprosthetic joint infection following primary total hip or knee..., Kong [/bib_ref] [bib_ref] The Otto Aufranc Award: modifiable versus nonmodifiable risk factors for infection after..., Maoz [/bib_ref]. Currently, the number of TKA operations worldwide is increasing year by year. However, due to the trend of aging population and rising obesity rate, the incidence of PJI after TKA is also increasing year by year [bib_ref] Measuring surgical site infection from linked administrative data following hip and knee..., Lethbridge [/bib_ref] [bib_ref] Risk factors associated with revision for prosthetic joint infection following knee replacement:..., Lenguerrand [/bib_ref].
The diagnosis and treatment of infections in artificial joint prostheses are still challenging, especially in the early stages. The latest diagnosis of knee prosthesis joint infection is largely based on the diagnostic criteria proposed by the Musculoskeletal Infection Society (MSIS), and PJI is diagnosed when one of the following three conditions is met: (1) the presence of a sinus tract that communicates with the prosthesis; (2) pathogens isolated from two or more separate tissue or fluid samples obtained from the affected prosthetic joint; and (3) four of the six criteria identified are met: 1. the elevation of erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) concentrations; 2. the elevation of white blood cells (WBC); 3. the elevation of the percentage of mesophilic polymorphonuclear leukocytes (PMN); 4. septicemia: septicemia of the affected joint; 5. isolation of microorganisms in cultures in histological analysis at 400 × magnification; 6. more than 5 neutrophils present in each of the 5 high magnification fields [bib_ref] Multidrug-resistant organisms in the setting of periprosthetic joint infection-diagnosis, prevention, and treatment, Siljander [/bib_ref] [bib_ref] Definition of successful infection management and guidelines for reporting of outcomes after..., Fillingham [/bib_ref] [bib_ref] The 2018 international consensus meeting minor criteria for chronic hip and knee..., Abdelaziz [/bib_ref]. Early and definitive diagnosis is the key to treating PJI. Once the diagnosis of PJI is made, the treatment options include conservative treatment with preservation of the joint prosthesis or debridement, and one or two stage prosthetic replacement surgery. The two-stage arthroprosthetic revisions is the gold standard for the treatment of PJI and is the most common treatment in clinical practice. In the tow-stage of prosthetic revision, the infected prosthesis is removed and a joint spacer with antibiotics is placed into the joint to provide continuous anti-infection treatment and then a new prosthesis is reinserted once the infection has been controlled. The two-stage of prosthetic revision surgery has achieved satisfactory results in current clinical practice [bib_ref] The management of an infected total knee arthroplasty, Gehrke [/bib_ref] [bib_ref] Complications in the treatment of prosthetic joint infection, Hartzler [/bib_ref] [bib_ref] Two-stage exchange protocol for periprosthetic joint infection following total knee arthroplasty in..., Petis [/bib_ref].
In the two-stage of the revision surgery, the two types of spacers that are frequently used include dynamic spacers and static spacers [bib_ref] Spacer design options and consideration for periprosthetic joint infection, Sporer [/bib_ref]. For static spacer, it is easier to design and the production costs of it are cheaper. In addition, static spacers are easier to implant in the joint during surgery. However, the mobility of the patient's knee joint will reduce after static spacer implantation. Static spacers are also prone to a number of complications such as causing bone loss and soft tissue damage [bib_ref] Pearls: mind the gap! the art of the static antibiotic-laden cement spacer, Abdeen [/bib_ref] [bib_ref] Point/counterpoint: static vs articulating spacers-static spacers for resection arthroplasty of the knee, Bowman [/bib_ref]. In contrast, dynamic spacers can improve the recovery of patient's knee function and reduce the incidence of associated complications. The use of spacers can provide good flexion and extension after the implantation of a new prosthesis. However, dynamic spacers are more expensive to produce. And they are not as good as static spacers in maintaining joint stability [bib_ref] Loading capacity of dynamic knee spacers: a comparison between hand-moulded and COPAL..., Chong [/bib_ref] [bib_ref] The effect of time spent with a dynamic spacer on clinical and..., Golgelioglu [/bib_ref]. In terms of infection control, some reports suggest that dynamic spacers and static spacers are similarly effective, while others suggest that dynamic spacers are less effective than static spacers in controlling infection [bib_ref] Articular spacers in two-stage revision arthroplasty for prosthetic joint infection of the..., Craig [/bib_ref] [bib_ref] Use of static or articulating spacers for infection following total knee arthroplasty:..., Voleti [/bib_ref] [bib_ref] The fate of spacers in the treatment of periprosthetic joint infection, Gomez [/bib_ref].
In order to further investigate whether there are any significant differences between dynamic and static spacers for the treatment of knee prosthesis infections in terms of efficacy, complications and functional impact, we conducted a comprehensive search of the literature and collated the data of inclusion literatures, then we did a corresponding meta-analysis and obtained the final results.
# Methods
## Study design and search strategy
We followed the principles of The Population-Intervention-Comparators-Outcomes-Study design (PICOS) strictly when conducting our literature search. The databases used for literature searches include PubMed, Embase, Web of Science and Cochrane library databases. The time frame for the search was from database creation to 10 March 2022 and the article type were Clinical randomized controlled studies, retrospective case-control studies or prospective cohort studies of dynamic spacers and static spacers for the treatment of knee prosthesis infections. The following keywords were used in the search process: "static spacer" and "dynamic spacer" and "Knee prosthesis infection", "static spacer" and "Articulating spacer" and "PJI". We had no restrictions on region or population, and all study subjects were human. All patients included in the studies met the diagnostic criteria for joint prosthesis infection as defined in the Infectious Diseases Society of America (IDSA), Musculoskeletal Infection Society (MSIS) [bib_ref] Comparison of diagnostic accuracy of periprosthetic tissue culture in blood culture bottles..., Yan [/bib_ref] [bib_ref] The clinical presentation of prosthetic joint infection, Barrett [/bib_ref]. All studies that did not meet this diagnostic criterion were excluded. The specific search process is shown in [fig_ref] Figure 1: The [/fig_ref].
## Exclusion criteria
In addition to the inclusion criteria described above, we have also developed exclusion criteria. Articles were excluded when one of the following conditions existed: (I) only patients using dynamic spacers or static spacers in the article; (II) there was a lack of follow-up time in the article or the follow-up time was less than 12 months; (III) the outcome indicators and statistical methods associated with the article were not uniform; (IV) literature reviews, case reports or experimental animal studies; (V) the diagnostic criteria for subjects were not referenced to the Infectious Diseases Society
## Data extraction and management
After determining the final inclusion literatures, we had two researchers responsible for extracting and summarizing the data from the articles, and the inclusion and exclusion of literature strictly followed the criteria described above. Articles were included and excluded by reading the title, abstract and content. If there was a disagreement in the screening and data extraction process, a third researcher was responsible for negotiating a resolution. After the inclusion of literatures, data on infection control rates, KSS scores, KSS functional scores, bone loss, ROM and other relevant indicators were extracted from the literature. We extracted and integrated the data and then performed a meta-analysis.
In addition, after identifying the 14 pieces of literature included, we assessed and scored the methodological quality of each study independently on the Newcastle-Ottawa Scale (NOS) [bib_ref] Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality..., Stang [/bib_ref]. We evaluated the quality of articles based on the NOS scale for the following items: (I) selection of study population (group), (II) comparability, and (III) how to determine which spacers to use. With a full score of 9 stars, the study is considered to be of high quality when the following conditions exist: (I) 3 or 4 stars in the selection domain, (II) 1 or 2 stars in the comparability domain, and (II) 2 or 3 stars in the outcome or exposure domain. If the NOS score ≥ 7 stars is considered high quality, ≥ 4 and < 7 stars is considered medium quality, and < 4 stars is considered low quality.
## Outcome measurements
Our primary outcome indicator for this meta-analysis was the rate of infection control in the dynamic spacer and static spacer groups. Secondary outcome indicators included KSSs scores, KSS functional scores, bone loss and ROM after treatment with the respective spacers.
## Statistical analysis and data synthesis
After completing the data extraction and classification summary, we used Stata SE-64 to analyze the data and finally obtained the corresponding forest plots for each group. If P < 0.05 indicates that the results are statistically significant and there is a significant difference between the two groups. If P > 0.05 indicates that the results are not statistically significant and there is no difference between the two groups. The variables obtained included dichotomous and continuous variables, and the variables were assessed so that relative risks (RR) and 95% confidence intervals (CI) were used. In terms of article heterogeneity, we used the I 2 test for analysis. When I 2 < 50% indicates little heterogeneity between the literature, a fixed effects model (FE) was chosen to analyze the data. When I 2 > 50% indicates significant heterogeneity between the literature, a random effects model (RE) was chosen to analyze the data.
# Result
## Identification of included studies
After a literature search in PubMed, Embase, Web of Science and Cochrane library databases, a total of 353 relevant articles were initially retrieved. We first read the titles and abstracts of the articles to eliminate duplicates or articles with inconsistent content, leaving 62 articles. The content of the remaining 62 articles was carefully read by two of our researchers, and after excluding 48 of the non-compliant articles, we were left with 14 articles that met the requirements of our study [bib_ref] Do dynamic cement-on-cement knee spacers provide better function and activity during two-stage..., Jaekel [/bib_ref] [bib_ref] Functional advantage of articulating versus static spacers in 2-stage revision for total..., Freeman [/bib_ref] [bib_ref] Articulating versus static spacers in revision total knee arthroplasty for sepsis. The..., Fehring [/bib_ref] [bib_ref] A randomized trial of static and articulating spacers for the treatment of..., Nahhas [/bib_ref] [bib_ref] Comparison of static and mobile antibiotic-impregnated cement spacers for the treatment of..., Park [/bib_ref] [bib_ref] Antibiotic-loaded cement articulating spacer for 2-stage reimplantation in infected total knee arthroplasty:..., Hsu [/bib_ref] [bib_ref] Static vs articulating spacers for two-stage revision total knee arthroplasty: minimum five-year..., Vasarhelyi [/bib_ref] [bib_ref] Are prosthetic spacers safe to use in 2-stage treatment for infected total..., Choi [/bib_ref] [bib_ref] Initial results of managing severe bone loss in infected total joint arthroplasty..., Incavo [/bib_ref] [bib_ref] Comparison of a static with a mobile spacer in total knee infection, Emerson [/bib_ref] [bib_ref] Minimizing dynamic knee spacer complications in infected revision arthroplasty, Johnson [/bib_ref] [bib_ref] Spacer prostheses in two-stage revision of infected knee arthroplasty, Jämsen [/bib_ref] [bib_ref] Application of 3D printing-assisted articulating spacer in two-stage revision surgery for periprosthetic..., Kong [/bib_ref]. The specific search and selection process for our included articles is referenced in [fig_ref] Figure 1: The [/fig_ref].
## Quality assessment of included studies
We finally included 14 relevant clinical studies. A total of 799 patients were included in all articles, with a total of 407 patients in the treatment group using dynamic spacers and 392 patients in the group using static spacers. The average follow-up time for patients in both treatment groups was more than 12 months. Among the included studies, 9 studies were graded as high quality (≥ 7 stars) and 5 studies were graded as moderate quality (≥ 4 and < 7 stars). The specific characteristics and ratings of each article are shown in .
## Infection control rate
The control rate of knee prosthesis infection was the main evaluation index in this meta-analysis. In total, 13 of the 14 included literatures had reported the specific control rates of knee prosthesis infection [bib_ref] Functional advantage of articulating versus static spacers in 2-stage revision for total..., Freeman [/bib_ref] [bib_ref] Articulating versus static spacers in revision total knee arthroplasty for sepsis. The..., Fehring [/bib_ref] [bib_ref] A randomized trial of static and articulating spacers for the treatment of..., Nahhas [/bib_ref] [bib_ref] Comparison of static and mobile antibiotic-impregnated cement spacers for the treatment of..., Park [/bib_ref] [bib_ref] Antibiotic-loaded cement articulating spacer for 2-stage reimplantation in infected total knee arthroplasty:..., Hsu [/bib_ref] [bib_ref] Static vs articulating spacers for two-stage revision total knee arthroplasty: minimum five-year..., Vasarhelyi [/bib_ref] [bib_ref] Are prosthetic spacers safe to use in 2-stage treatment for infected total..., Choi [/bib_ref] [bib_ref] Initial results of managing severe bone loss in infected total joint arthroplasty..., Incavo [/bib_ref] [bib_ref] Comparison of a static with a mobile spacer in total knee infection, Emerson [/bib_ref] [bib_ref] Minimizing dynamic knee spacer complications in infected revision arthroplasty, Johnson [/bib_ref] [bib_ref] Spacer prostheses in two-stage revision of infected knee arthroplasty, Jämsen [/bib_ref] [bib_ref] Application of 3D printing-assisted articulating spacer in two-stage revision surgery for periprosthetic..., Kong [/bib_ref]. From the forest plots, we can see that there is no statistical difference between dynamic spacers and static spacers in terms of prosthetic infection control rates. Such results suggest that dynamic spacers and static spacers are similar in controlling prosthetic infection (RR: 1.03; 95% Cl 0.98, 1.09; P = 0.179), as shown in [fig_ref] Figure 2: Forest plot of the infection control rate [/fig_ref]. Such results suggest that dynamic spacers are no less effective in controlling prosthetic infection than static spacers. The funnel plot is shown in [fig_ref] Figure 3: Funnel [/fig_ref]. The Egger's test result is P = 0.015, suggesting some publication bias. For this reason, we assessed this further by using the trim and fill method. And the results remained consistent with those obtained earlier after including the data from five dummy studies. This indicating that the result we derived is steady.
## Kss knee score (ksss)
Of the included articles, 10 reported on the KSS Knee Scores (KSSs) of patients after receiving spacer implantation . Based on the results of the forest plots obtained, it is known that the dynamic spacer was more effective than the static spacer in terms of KSSs scores in the knee joint after spacer implantation, with a statistically significant difference in treatment effect between the two groups (RR: 7.34; 95% CI 1.89, 12.79; p = 0.032). A specific forest plot is shown in .
## Kss functional score
Among the secondary outcome indicators in this metaanalysis, the KSS functional score was one of the indicators to judge the effect of the spacer implantation on the patient's later functional recovery of the knee. Of the 14 articles included, a total of 6 articles reported on the KSS functional scores of patients after receiving spacer implantation . According to the forest plot we can know that the dynamic spacer was significantly better than the static spacer in improving the patients' recovery of knee function, with statistically significant results between the two (RR: 11.16; 95% CI 4.18, 18.13; P = 0.02). The forest plot is shown in .
## Bone loos
Bone loss is one of the secondary outcome indicators in this meta-analysis, and it is an indicator of the occurrence of bone loss or deficiency in and around the joint after spacer implantation. A total of six articles reported on bone loss after patients received treatment [bib_ref] Articulating versus static spacers in revision total knee arthroplasty for sepsis. The..., Fehring [/bib_ref] [bib_ref] Comparison of static and mobile antibiotic-impregnated cement spacers for the treatment of..., Park [/bib_ref] [bib_ref] Minimizing dynamic knee spacer complications in infected revision arthroplasty, Johnson [/bib_ref] [bib_ref] Spacer prostheses in two-stage revision of infected knee arthroplasty, Jämsen [/bib_ref] [bib_ref] Application of 3D printing-assisted articulating spacer in two-stage revision surgery for periprosthetic..., Kong [/bib_ref]. The results suggest that after spacer implantation in our patients, significantly fewer patients with dynamic spacers experienced bone loss than those with static spacers. There was a statistically significant difference between the two groups (RR: 2.04; 95% CI 1.11, 3.77; p = 0.018). The specific forest plot is shown in .
## Rom
A total of 12 of these articles reported on the ROM of the knee joint after patients received spacer implantation [bib_ref] Functional advantage of articulating versus static spacers in 2-stage revision for total..., Freeman [/bib_ref] [bib_ref] Articulating versus static spacers in revision total knee arthroplasty for sepsis. The..., Fehring [/bib_ref] [bib_ref] A randomized trial of static and articulating spacers for the treatment of..., Nahhas [/bib_ref] [bib_ref] Comparison of static and mobile antibiotic-impregnated cement spacers for the treatment of..., Park [/bib_ref] [bib_ref] Antibiotic-loaded cement articulating spacer for 2-stage reimplantation in infected total knee arthroplasty:..., Hsu [/bib_ref] [bib_ref] Static vs articulating spacers for two-stage revision total knee arthroplasty: minimum five-year..., Vasarhelyi [/bib_ref] [bib_ref] Are prosthetic spacers safe to use in 2-stage treatment for infected total..., Choi [/bib_ref] [bib_ref] Initial results of managing severe bone loss in infected total joint arthroplasty..., Incavo [/bib_ref] [bib_ref] Minimizing dynamic knee spacer complications in infected revision arthroplasty, Johnson [/bib_ref] [bib_ref] Spacer prostheses in two-stage revision of infected knee arthroplasty, Jämsen [/bib_ref] [bib_ref] Application of 3D printing-assisted articulating spacer in two-stage revision surgery for periprosthetic..., Kong [/bib_ref]. The result of the meta-analysis suggested to us that the patients with dynamic spacers had significantly better ROM than those with static spacers, (RR: 18.73; 95% CI 11.67, 25.79; P < 0.0001). The specific forest plot is shown in [fig_ref] Figure 7: Forest [/fig_ref].
# Discussion
With the increase of total knee arthroplasty surgery year by year, the incidence of prosthetic joint infection is also increasing. When a joint prosthesis becomes infected, a poorly controlled infection can lead not only to failure of the joint arthroplasty, but in severe cases even to amputation [bib_ref] Staphylococci and implant surfaces: a review, Harris [/bib_ref] [bib_ref] What are the frequency, associated factors, and mortality of amputation and arthrodesis..., Son [/bib_ref]. Two stage revision surgery is the most commonly used method in clinical practice and it can offer the best results in terms of treatment. In the two stage revision surgery, the commonly used spacers include dynamic spacers and static spacers and each has its own characteristics. The structure of the dynamic spacer is match to the anatomy of the knee joint, thus it can reduce adhesions and scarring of the soft tissues surrounding the knee joint. The dynamic spacer can also improve the recovery of knee function after revision surgery and reduce the incidence of some complications. However, there are also some reports suggesting that dynamic spacers are less effective in controlling prosthetic joint infection. In contrast, static spacers are less prone to dislocation during fixation and it can also provide good joint stability. In addition, the static spacer can provide a high concentration of antibiotics for better infection control while maintaining limb length. Of course, there is still controversy in clinical and related research regarding the difference between dynamic and static spacers in controlling prosthetic infection and improving the prognosis of revision surgery. In the current meta-analysis, we collected relevant clinically controlled studies and performed a meta-analysis of these. Our aim is to further investigate the differences between dynamic spacers and static spacers in terms of therapeutic effect and impact on knee function. A total of 14 articles were included in this meta-analysis, and we grouped one primary and four secondary outcome indicators according to the indicators in each article. Finally, we did the corresponding meta-analysis for each of the five outcome indicators. The rate of infection control in prosthetic joint infection is the most important indicator of the effectiveness of revision surgery treatment. Based on the final results we can find that the treatment effect between dynamic spacers and static spacers is the same in terms of the main outcome indicator of prosthetic infection control, with no significant difference between the two groups (p = 0.179). This suggests us that dynamic spacers are no less effective than static spacers for prosthetic infection control and it can also achieve the same clinical outcomes. This result is consistent with the results of many current clinical studies and further validates the efficacy of joint spacers [bib_ref] 1 complications and outcomes associated with a novel, prefabricated, articulating spacer for..., Buller [/bib_ref] [bib_ref] Intraoperative molds to create an articulating spacer for the infected knee arthroplasty, Van Thiel [/bib_ref] [bib_ref] Two-stage treatment for total knee arthroplasty infection utilizing an articulating prefabricated antibiotic..., Warth [/bib_ref]. According to relevant literature reports, the concentration of antibiotics and the duration of antiinfection therapy are important factors in the outcome of two-stage revision surgery for prosthetic infections. The use of adequate antibiotic concentrations and duration of anti-infection therapy not only results in more satisfactory infection control, but also better reduces the recurrence of prosthetic infections [bib_ref] Periprosthetic joint infection, Kapadia [/bib_ref] [bib_ref] Antibiotic elution from hip and knee acrylic bone cement spacers: a systematic..., Anagnostakos [/bib_ref]. In addition, revision surgery at an early stage of prosthetic infection can also provide better control of prosthetic infection, and early anti-infection treatment is one of the keys to treatment and reduces the risk of surgery [bib_ref] Factors related to outcome of early and delayed prosthetic joint infections, Ascione [/bib_ref] [bib_ref] The diagnosis and management of prosthetic joint infections, Moran [/bib_ref]. Therefore, not only does the correct use of spacers will influence the outcome of treatment, but timely and adequate anti-infective treatment is also crucial to the success of treatment.
Of the four secondary outcome indicators, KSSs scores, KSS functional scores and ROM were used to evaluate the impact on knee function after patients received spacer implantation, all as an indicator of post-operative recovery and efficacy. In the results of our meta-analysis, the results for all three indicators were statistically significant (p < 0.05). This indicates a significant difference between the results of the dynamic spacer group and the static spacer group in these three indicators. In terms of improved prognosis for knee function, patients with dynamic spacers had significantly better functional scores and knee mobility than those with static spacers. These results once again verified the superiority of the dynamic spacers over the static spacers, in comparison, the dynamic spacer could better improve the patient's motion function and range of motion [bib_ref] A two-stage approach to primary TKA using articulating antibiotic-loaded spacers improve function..., Pietsch [/bib_ref] [bib_ref] Articulated spacer provides long-term knee improvement after two-stage reimplantation, Vasso [/bib_ref] [bib_ref] Systematic review comparing static and articulating spacers used for revision of infected..., Pivec [/bib_ref]. Bone loss is a common complication after spacer implantation and the patient's bone loss is a judgment indicator to evaluate the impact of the spacer on the patient's bone health. Therefore, bone loss was also one of the reference indicators in our current meta-analysis. Based on the results of the meta-analysis, it is known that the patients with dynamic spacers experienced significantly less bone loss than those with static spacers (p < 0.05). There was a significant difference in the results between the two groups. This is also in line with some current findings that patients using static spacers are more likely to experience bone loss [bib_ref] Küntscher nails with static cement spacer: a simple technique in periprosthetic knee..., Presti [/bib_ref].
Combining these results, we can see that the dynamic spacer can achieve similar results to the static spacer in terms of infection control of the prosthesis. The performance of dynamic spacers is better than that of static spacers in improving the prognosis of knee function and preventing bone loss. For patients requiring prosthetic revision surgery, the use of dynamic spacers may provide a better prognosis and recovery of joint function.
Of course, there are some limitations to our current meta-analysis. Firstly, although we included a total of 14 relevant literatures, the total number of patients studied was only 799, which may not be a large enough sample size. Perhaps we need more clinical studies with larger samples to further confirm our results. Secondly, the articles we included were not clinical randomized controlled studies, but rather retrospective clinical studies and prospective trials. The use of blinding was also not reported for the assessment of outcomes and scores related to knee function, which lacks a certain degree of concealment. In addition, there is a lack of uniformity in the assessment and the associated results may be influenced by subjective factors. Thirdly, the possible influence of relevant factors on the treatment outcomes (e.g., gender, age, height, weight, etc.) was not adequately considered in determining patient groupings.
# Conclusion
Based on the results of the meta-analysis, we finally obtained, we can know that there was no significant difference in the rate of infection control of the prosthesis between the dynamic spacer group and the static spacer group, similar results can be obtained with both types of spacers. And there are significant differences between the dynamic spacer group and the static spacer group in terms of knee function related scores and ROM. The patients using dynamic spacers can achieve better joint mobility and function. There is also a significant difference between the two groups in terms of bone loss. Patients with dynamic spacers experienced significantly less bone loss than those with static spacers.
[fig] Figure 1: The [/fig]
[fig] Figure 2: Forest plot of the infection control rate [/fig]
[fig] Figure 3: Funnel [/fig]
[fig] Figure 4, Figure 5: Forest Forest plot of the KSS functional scoreFig. 6 Forest plot of the bone loose [/fig]
[fig] Figure 7: Forest [/fig]
|
Intensive follow-up for women with breast cancer: review of clinical, economic and patient’s preference domains through evidence to decision framework
Background: Women treated for breast cancer are followed-up for monitoring of treatment effectiveness and for detecting recurrences at an early stage. The type of follow-up received may affect women's reassurance and impact on their quality of life. Anxiety and depression among women with breast cancer has been described, but little is known about how the intensity of the follow-up can affect women's psychological status. This study was undertaken to evaluate the effects of intensive vs. less-intensive follow-up on different health outcomes, to determine what are women's preferences and values regarding the follow-up received, and also assess the costs of these different types of follow-up. Methods: A systematic review following standard Cochrane Collaboration methods was carried out to assess the efficacy of intensive follow-up versus non-intensive follow-up in breast cancer patients. Two additional reviews on women's preferences and economic evidence were also carried out. The search was performed up to January 2016 in: MEDLINE, EMBASE, PDQ, McMaster Health Systems Evidence, CENTRAL, and NHS EED (through The Cochrane Library). The quality of evidence was assessed by GRADE (for quantitative studies) and CerQUAL (for qualitative studies). Several outcomes including mortality, breast cancer recurrences, quality of life, and patient satisfaction were evaluated. Results: Six randomised trials (corresponding to 3534 women) were included for the evaluation of health outcomes; three studies were included for women's values and preferences and four for an economic assessment. There is moderate certainty of evidence showing that intensive follow-up, including more frequent diagnostic tests or visits, does not have effects on 5-or 10-year overall mortality and recurrences in women with breast cancer, compared with less intensive follow-up. Regarding women's preferences and values, there was important variability among studies and within studies (low confidence due to risk of bias and inconsistency). Furthermore, intensive follow-up, as opposed to less intensive follow-up, is not likely to be cost-effective. Conclusions: Less intensive follow-up appears to be justified and can be recommended over intensive follow-up. Resources could thus be mobilised to other aspects of breast cancer care, or other areas of healthcare.
# Background
Breast cancer is the most frequently diagnosed cancer and among the leading causes of cancer death among females [bib_ref] Epidemiology of breast cancer, Ban [/bib_ref] [bib_ref] Breast cancer: epidemiology and etiology, Tao [/bib_ref]. Due to significant improvements in screening, early diagnosis, and treatment in the recent decades, breast cancer mortality has decreased worldwide [bib_ref] Cancer survival in five continents: a worldwide population-based study (CONCORD), Coleman [/bib_ref] [bib_ref] Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN..., Ferlay [/bib_ref] [bib_ref] Cancer statistics, Siegel [/bib_ref]. This leads to a situation where the total number of prevalent breast cancer cases is increasing, and therefore a growing number of women needing follow-up care. Worldwide and European estimates of women with a diagnosis of breast cancer occurring in the last 5 years correspond to 6.2 and 1.8 million, respectively.
Women treated for breast cancer are followed-up for monitoring treatment effectiveness and complications, and for detecting recurrences at an early stage or new primary contralateral breast cancer. Follow-up includes clinical and test examinations such as routine haematological and liver function tests, tumour markers, chest X-ray, mammography and bone and liver scans [bib_ref] Breast cancer follow-up and management after primary treatment, Khatcheressian [/bib_ref]. The diversity in frequency and in the type of examination results in many different follow-up practices, the intensity of which can be defined by the frequency of clinical visits and/or physical examinations (e.g. intensive, standard, patient-initiated or low intensity). There is also evidence of variability [bib_ref] Screening for breast cancer, Elmore [/bib_ref] in the way follow-up is implemented in clinical practice.
Despite doubts that intensive follow-up care can improve survival in these patients, intensive follow-up is quite common in clinical practice and represents a significant workload for radiotherapy, surgical and oncologic departments [bib_ref] It is now the age to define the appropriate follow-up of primary..., Loprinzi [/bib_ref].
For a long time, the scientific community has focused on the relationship between the type of follow-up (i.e. intensity) and health outcomes, such as long-term mortality and morbidity, but also quality of life. A recent Cochrane review that analysed randomised control trials with almost 20 years of follow-up gave the following suggestion: "follow-up programs based on regular physical examinations and yearly mammography alone are as effective as more intensive approaches based on regular performance of laboratory and instrumental tests in terms of timeliness of recurrence detection, overall survival and quality of life" [bib_ref] Follow-up strategies for women treated for early breast cancer, Moschetti [/bib_ref].
Despite the importance of health outcomes in terms of mortality and morbidity, it is also important to take into account the women's perspective, including psychosomatic symptoms and diseases, which could be manifested as preference towards one or another type of follow-up scheme [bib_ref] The prevalence of long-term symptoms of depression and anxiety after breast cancer..., Maass [/bib_ref]. Similarly, economic evidence in healthcare is becoming increasingly important, not only in the form of cost-effectiveness or cost-utility analyses, which are the most common mechanism for generating economic evidence in decision making, but also in the form of cost-minimisation, cost-consequences or cost-benefit analyses or total budget impact estimates. This is another key point to be considered in recommending a certain type of follow-up protocol.
Given that all these aspects should be considered together to make decisions in healthcare, there is an urgent need to use up-to-date and user-friendly evidencepresentation formats, in order to improve the communication of evidence-based healthcare recommendations, addressing communication needs of guideline users and decision-makers [bib_ref] Developing and evaluating communication strategies to support informed decisions and practice based..., Treweek [/bib_ref]. The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) methodology [bib_ref] An emerging consensus on rating quality of evidence and strength of recommendations, Guyatt [/bib_ref] [bib_ref] GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables, Guyatt [/bib_ref] [bib_ref] Use of GRADE for assessment of evidence about prognosis: rating confidence in..., Iorio [/bib_ref] in combination with Evidence to Decision (EtD) Frameworks provides an assessment and a summary of alternative strategies on three key elements: patient-important outcomes, patients' values and preferences, and economic evidence. Information regarding acceptability and feasibility of the analysed strategies, and their impact on health equity is also included [bib_ref] An emerging consensus on rating quality of evidence and strength of recommendations, Guyatt [/bib_ref] [bib_ref] Assessing and presenting summaries of evidence in Cochrane reviews, Langendam [/bib_ref].
The aim of the present paper is to evaluate the available research evidence on the clinical question about whether intensive follow-up should be provided for breast cancer patients treated with curative intent. The evidence is assessed and summarised according to GRADE and the EtD framework; the recommendations made in this manuscript are based on the authors judgements and should only be considered as the authors' recommendations and not as recommendations made by a guideline panel. Nevertheless, they are useful to facilitate the further decision-making process carried out by guideline panels in charge of issuing clinical recommendations.
# Methods
Systematic review on the evidence of effects of intensive follow-up on breast cancer outcomes
The research question was addressed by means of a systematic review of the literature on the evidence of health outcomes related to the alternative strategiesintensive and less intensive follow-up. An operational definition was used for intensive follow-up, where intensive was defined in comparison with a less intensive follow-up schedule or a patient-initiated approach. The review protocol is available upon request. Standard Cochrane Collaboration methods were followed. For the evaluation of the importance of the outcomes and for the assessment of the quality of evidence, the GRADE system was used.
Research question: the clinical question was structured following the PICO (Patient, Intervention, Comparison, Outcomes) format:
Population: breast cancer patients, treated with curative intent; Intervention: intensive follow-up schedule; Comparison: non-intensive follow-up; Outcomes: 5-and 10-year mortality due to breast cancer; 5 and 10-year breast cancer recurrences (loco-regional and distant separately); 5-and 10-year breast cancer specific survival; quality of life at 2 and 5 years after diagnosis; women's satisfaction with follow-up (measured by reassurance of women with the intensive follow-up and convenience by the women of intensive follow-up).
Critical outcomes included mortality due to breast cancer, breast cancer recurrences and breast cancer specific survival. Quality of life and satisfaction were considered important outcomes.
## Inclusion and exclusion criteria
Following the WHO Handbook for Guidelines Developmentas guidance, existing relevant systematic reviews of observational and experimental evidence were included as a source of individual studies; additional individual studies were searched, to update the body of evidence. Temporal or language restrictions were not applied. Studies in which the effects of follow-up intensity were not assessed, or when the outcomes were out of the scope of the clinical question, were excluded.
## Search strategy
Systematic reviews were identified by introducing a combination of controlled vocabulary and search terms (e.g., follow-up, breast neoplasms, mortality, recurrences, quality of life, satisfaction, cost, healthcare resources, survival) in The Cochrane Database of Systematic Reviews (2015, issue 11), The Database of Abstracts of Reviews of Effects (DARE), and PubMed limiting the search to the subset "systematic [sb]".
Original studies were searched in MEDLINE (through PubMed; from 1946 to January 2016), EMBASE (through Ovid; from 1980 to November 2015), PDQ, McMaster Health Systems Evidence, CENTRAL, and NHS EED (through The Cochrane Library; January 2016).
The complete search algorithms designed for each database, the hits retrieved, and the reasons for exclusion are presented in Additional file 1 and [fig_ref] Figure 1: PRISMA flowcharts [/fig_ref].
One reviewer screened the search results based on title and abstract. This process was subjected to a quality process, by reviewing 20% of the references by a second reviewer. Two reviewers independently confirmed eligibility, based on the full text of the relevant articles. In case of disagreement between reviewers the inclusion of studies was determined by consensus.
## Data extraction
Data extraction was conducted by one reviewer. As quality control, another reviewer went through 20% of the data for accuracy.
## Risk of bias
The assessment was carried out by one reviewer. As quality control, another reviewer went through 20% of the data for accuracy. For each study, the risk of bias was rated for each domain as low, high or unclear risk of bias.
## Effect measures
Odds ratios (OR), risk ratios (RR) and hazard ratios (HR) were extracted, with their 95% confidence intervals (CI). If available, only adjusted effect measures were collected. Data from any estimation of effect provided (percentages, means, medians) were also collected.
## Quality of the evidence evaluation
The quality of evidence per outcome was rated from high to very low considering the standard GRADE domains: risk of bias, imprecision, inconsistency, indirectness and publication bias [bib_ref] GRADE guidelines: 3. Rating the quality of evidence, Balshem [/bib_ref] [bib_ref] GRADE handbook for grading quality of evidence and strength of recommendations, Schünemann [/bib_ref].
# Data analysis
A pooled analysis was conducted applying the inversevariance method under the random-effects model [bib_ref] Meta-analysis in clinical trials, Dersimonian [/bib_ref] ; the analysis was performed through the Software Review Manager v. 5.3. Heterogeneity was assessed using the I 2 statistic.
The report of results of the meta-analysis adhered to the guidelines articulated in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [bib_ref] Preferred reporting items for systematic review and metaanalysis protocols (PRISMA-P), Moher [/bib_ref].
## Review on women's values and preferences
A review about women's values and preferences for intensive follow-up versus non-intensive follow-up after breast cancer treatment was undertaken.
## Inclusion and exclusion criteria
In a first stage, after conducting the systematic search of the literature, the screening of references was carried out, initially prioritising the identification of systematic reviews. In a second stage, individual studies were retrieved (e.g., qualitative studies, surveys, utility elicitation studies). Studies in English, French, German, Greek, and Spanish, carried out in the Organisation for Economic Co-operation and Development (OECD) Region, were included. Included studies were: examining women's preferences for follow-up strategies after breast cancer treatment; evaluating how women value the main outcomes associated with follow-up strategies after breast cancer treatment; examining the choices women make when informed about the desirable and undesirable effects associated with follow-up strategies after breast cancer treatment.
Studies assessing only women's knowledge, views, perceptions, attitudes and expectations regarding follow-up strategies after breast cancer treatment were excluded; similarly, studies assessing barriers to follow-up strategies after breast cancer treatment were not included.
## Search strategy
A search strategy was designed to identify relevant studies in MEDLINE (accessed through Ovid). For systematic reviews, there were no time restrictions. For primary studies, only studies published after 2006 were included. The complete search strategy can be found in Additional file 2.
One reviewer screened the search results based on the title and abstract. Two reviewers independently confirmed eligibility based on the full text of the relevant articles. In case of disagreement between researchers, the inclusion of studies was decided by consensus [fig_ref] Figure 1: PRISMA flowcharts [/fig_ref].
## Data extraction
One reviewer extracted the main characteristics of the included studies and their findings in a tabulated format. A second reviewer checked the extracted data for accuracy.
## Quality of the evidence evaluation
The quality of evidence was rated with GRADE. In the case of qualitative research, the Confidence in the Evidence from Reviews of Qualitative research (CERQual) approach was used [bib_ref] Using qualitative evidence in decision making for health and social interventions: an..., Lewin [/bib_ref].
## Review on economic evidence
A review about the economic evidence for intensive follow-up versus non-intensive follow-up after breast cancer treatment was carried out.
## Inclusion and exclusion criteria
Screening of literature and study selection was done in a step-by-step approach. Firstly, the search focused on studies that addressed economic aspects directly related to the PICO question. Then, recent European costeffectiveness or cost-utility analyses related to the PICO question were looked for. Only studies in English were included.
## Search strategy
Search strategies were designed to identify relevant studies in MEDLINE (through Ovid, January 2016) and in the NHS Economic Evaluation Database (through The Cochrane Library, January 2016). The complete search strategies are included in Additional file 3. Study design filters were applied to retrieve relevant studies. The selection process is presented in [fig_ref] Figure 1: PRISMA flowcharts [/fig_ref].
## Data extraction
Main characteristics of included studies were described in a tabulated format, including the following data: author and publication year, country, type of economic analysis, perspective of the analysis, time horizon and discounting, relevant outcomes and costs included, sources of information (baseline outcomes, relative intervention effects, resource use and costs), Quality Adjusted Life Years (QALY), Incremental Cost Effectiveness Ratio (ICER), sensitivity analysis and conflict of interest.
## Quality of evidence
The quality of evidence for the resource requirements was rated according to GRADE [bib_ref] GRADE guidelines: 10. Considering resource use and rating the quality of economic..., Brunetti [/bib_ref]. The NICE methodology checklist for economic evaluationswas used to assess the risk of bias and decide whether to include the studies. Included studies were of low risk of bias and were considered applicable to the European context.
## Evidence to decision framework
To summarise the evidence, and in accordance to the GRADE methodology [bib_ref] GRADE handbook for grading quality of evidence and strength of recommendations, Schünemann [/bib_ref] and the interactive Evidence to Decision framework guidance, an EtD Framework was developed. The authors covered the role of the panel with respect to the EtD framework.
# Results
## Evidence of effects of intensive follow-up on breast cancer outcomes
Five systematic reviews were included for the evaluation of health outcomes [bib_ref] Follow-up care of patients treated for breast cancer: a structured review, Collins [/bib_ref] [bib_ref] Follow-up in breast cancer: does routine clinical examination improve outcome? A systematic..., Montgomery [/bib_ref] [bib_ref] Surveillance mammography for detecting ipsilateral breast tumour recurrence and metachronous contralateral breast..., Robertson [/bib_ref] [bib_ref] Follow-up strategies for women treated for early breast cancer, Rojas [/bib_ref]. These systematic reviews were used as a source to identify primary studies. Eight papers, referring to six randomised clinical trials for a total of 3534 randomised women [bib_ref] Intensive diagnostic follow-up after treatment of primary breast cancer. A randomized trial...., Del Turco [/bib_ref] [bib_ref] Popularity of less frequent follow up for breast cancer in randomised study:..., Gulliford [/bib_ref] [bib_ref] Intensive vs clinical follow-up after treatment of primary breast cancer: 10-year update..., Palli [/bib_ref] [bib_ref] Patient initiated follow up of breast cancer, Brown [/bib_ref] [bib_ref] Role of chest X-ray in diagnosis of the first breast cancer relapse:..., Kokko [/bib_ref] [bib_ref] Follow-up cost of breast cancer patients with localized disease after primary treatment:..., Kokko [/bib_ref] [bib_ref] Cost-benefit analysis of a follow-up program in patients with breast cancer: a..., Oltra [/bib_ref] , were retrieved and included. These studies are summarised in [fig_ref] Table 1: Summary and short description of the six included randomised clinical trials phy,... [/fig_ref] The included studies had different definitions for intensive follow-up. In four studies, intensive follow-up referred to a greater number of diagnostic tests compared to regular follow-up [bib_ref] Intensive diagnostic follow-up after treatment of primary breast cancer. A randomized trial...., Del Turco [/bib_ref] [bib_ref] Role of chest X-ray in diagnosis of the first breast cancer relapse:..., Kokko [/bib_ref] [bib_ref] Cost-benefit analysis of a follow-up program in patients with breast cancer: a..., Oltra [/bib_ref] , while in two studies it referred to more frequent visits without modification in the number of diagnostic tests [bib_ref] Popularity of less frequent follow up for breast cancer in randomised study:..., Gulliford [/bib_ref] [bib_ref] Patient initiated follow up of breast cancer, Brown [/bib_ref]. Three studies compared an intensive versus a standard follow-up [bib_ref] Intensive diagnostic follow-up after treatment of primary breast cancer. A randomized trial...., Del Turco [/bib_ref] [bib_ref] Cost-benefit analysis of a follow-up program in patients with breast cancer: a..., Oltra [/bib_ref] , while the other three compared a low-intensity patient-initiated versus a standard follow-up [bib_ref] Popularity of less frequent follow up for breast cancer in randomised study:..., Gulliford [/bib_ref] [bib_ref] Patient initiated follow up of breast cancer, Brown [/bib_ref] [bib_ref] Role of chest X-ray in diagnosis of the first breast cancer relapse:..., Kokko [/bib_ref]. Five studies specified that patients (including the nonintensive follow-up group) underwent an annual mammography [bib_ref] Intensive diagnostic follow-up after treatment of primary breast cancer. A randomized trial...., Del Turco [/bib_ref] [bib_ref] Popularity of less frequent follow up for breast cancer in randomised study:..., Gulliford [/bib_ref] [bib_ref] Intensive vs clinical follow-up after treatment of primary breast cancer: 10-year update..., Palli [/bib_ref] [bib_ref] Cost-benefit analysis of a follow-up program in patients with breast cancer: a..., Oltra [/bib_ref]. No studies provided information about specific breast cancer mortality or survival. Among all studies, only the one carried out by the GIVIO groupreported the expected 5-year relative mortality reduction used for the calculation of sample size, i.e. 20% reduction; this threshold may be considered as the clinically significant mortality reduction expected.
Results and pooled analysis are provided when possible. Quantitative estimates are available only for the following outcomes, presented in [fig_ref] Figure 2: Estimates of effect of intensive vs [/fig_ref] :
10-year overall mortality: one trial [bib_ref] Intensive vs clinical follow-up after treatment of primary breast cancer: 10-year update..., Palli [/bib_ref] , that compared intensive vs. standard follow-up in 1243 women, found a RR = 1.05 (95%CI: 0.90 to 1.22). The quality of evidence was high because non-blinding was not considered a cause of risk of bias for this outcome.
5-year overall mortality: three trials [bib_ref] Intensive diagnostic follow-up after treatment of primary breast cancer. A randomized trial...., Del Turco [/bib_ref] [bib_ref] Role of chest X-ray in diagnosis of the first breast cancer relapse:..., Kokko [/bib_ref] , on 3035 patients, that compared intensive vs. standard follow-up found a RR 1.00 (95%CI: 0.86 to 1.16; I 2 = 0%). The 5-year mortality RR lower 95%CI did not reach the expected value for clinical significance either (vs. an expected 20% mortality reduction, as per GIVIO group outcome). The quality of evidence was high because non-blinding was not considered a cause of risk of bias for this outcome.
5-year breast cancer recurrences (any loco-regional and distant): three trials [bib_ref] Intensive diagnostic follow-up after treatment of primary breast cancer. A randomized trial...., Del Turco [/bib_ref] [bib_ref] Role of chest X-ray in diagnosis of the first breast cancer relapse:..., Kokko [/bib_ref] , on 3035 patients, that compared intensive and standard follow-up, resulted in a RR = 1.08 (95%CI: 0.89 to 1.30; I 2 = 60%). The quality of evidence was moderate because it was downgraded for risk of bias (the outcome assessment was not blinded). From the clinical point of view, when the patient is diagnosed with recurrence, there will be an initiation of new treatments or change in the treatment, so we considered that non-blinding is not an issue in this case.
Breast cancer recurrences at any time: five trials [bib_ref] Intensive diagnostic follow-up after treatment of primary breast cancer. A randomized trial...., Del Turco [/bib_ref] [bib_ref] Patient initiated follow up of breast cancer, Brown [/bib_ref] [bib_ref] Role of chest X-ray in diagnosis of the first breast cancer relapse:..., Kokko [/bib_ref] [bib_ref] Cost-benefit analysis of a follow-up program in patients with breast cancer: a..., Oltra [/bib_ref] , on 3217 women, found a RR = 1.10 (95%CI: 0.95 to 1.27; I 2 = 23%, when comparing intensive vs. standard follow-up. The quality of evidence was moderate because it was downgraded for risk of bias (the outcome assessment was not blinded). As discussed above, non-blinding was not considered as an issue.
Satisfaction of women with the type of follow-up: satisfaction was measured as reassurance (capacity of the type of follow-up to clear patients' doubts or fears) in two studies [bib_ref] Popularity of less frequent follow up for breast cancer in randomised study:..., Gulliford [/bib_ref] [bib_ref] Patient initiated follow up of breast cancer, Brown [/bib_ref] on 245 patients, and as convenience (suitability of the follow-up to the woman's life circumstances) in one study [bib_ref] Patient initiated follow up of breast cancer, Brown [/bib_ref] on 61 women. The results on reassurance were in favour of intensive follow-up (RR 1.28, 95%CI: 1.07 to 1.54; I 2 = 90%). The overall quality of evidence was very low due to risk of bias (the studies were not blinded) and imprecision (the number of events is small). The results on convenience favoured non-intensive follow-up (RR 0.04, 95%CI: 0.01 to 0.31). The overall quality of evidence was low because evidence needed to be downgraded for risk of bias (the outcome assessment was not blinded) and imprecision (the number of events is very small).
For each studied outcome, the evidence profile is reported in [fig_ref] Table 2: Evidence profiles for selected health outcomes related to intensive vs [/fig_ref].
## Women's values and preferences
For women's values and preferences, three European studies were included [bib_ref] Popularity of less frequent follow up for breast cancer in randomised study:..., Gulliford [/bib_ref] [bib_ref] The reality in the surveillance of breast cancer survivors-results of a patient..., Stemmler [/bib_ref] [bib_ref] Follow-up after treatment for breast cancer: one strategy fits all? An investigation..., Kimman [/bib_ref] [fig_ref] Table 3: Summary and short description of the three included studies on women's preferences... [/fig_ref].
Gulliford et al. [bib_ref] Popularity of less frequent follow up for breast cancer in randomised study:..., Gulliford [/bib_ref] compared experiences of 193 patients with breast cancer, randomised into a group with a conventional schedule of clinic visits, and a group of Kimman et al. [bib_ref] Follow-up after treatment for breast cancer: one strategy fits all? An investigation..., Kimman [/bib_ref] conducted a multicentre discretechoice experiment survey to measure the strength of preferences for several characteristics of breast cancer follow-up. The results of these three studies were inconsistent: in the first study, women appeared to prefer non-intensive follow-up schedules, while in the other The quality of evidence was downgraded because studies were not blinded; 2
The quality of evidence was downgraded due to important heterogeneity two the preferences favoured intensive schedules. However, important variability was present among studies and within studies. There was low confidence in the evidence due to risk of bias and inconsistency. The results of the review indicated that most of the regularly scheduled follow-up visits used further extensive laboratory and imaging procedures exceeding the quantity of examinations recommended in most of the current follow-up guidelines.
## Economic evidence
Four studies [bib_ref] Follow-up cost of breast cancer patients with localized disease after primary treatment:..., Kokko [/bib_ref] [bib_ref] Cost-benefit analysis of a follow-up program in patients with breast cancer: a..., Oltra [/bib_ref] [bib_ref] Economic evaluation of four follow-up strategies after curative treatment for breast cancer:..., Kimman [/bib_ref] [bib_ref] The clinical effectiveness and costeffectiveness of different surveillance mammography regimens after the..., Robertson [/bib_ref] assessed resources used, costs and cost-effectiveness of intensive follow-up strategies.
Robertson et al. [bib_ref] The clinical effectiveness and costeffectiveness of different surveillance mammography regimens after the..., Robertson [/bib_ref] conducted a cost-utility analysis in the UK and provided estimated costs (in 2008 value) for different mammographic surveillance regimens in women after breast cancer surgery. By assuming the cost of a mammography and of a clinical follow-up visit to be 71 and 110 €, respectively, in a cohort of 10,000 UK women with a mean age of 57, total costs varied from 3.27 million € (mammographic surveillance every 2 years) to 16.8 million € (yearly mammographic and clinical follow-up) for a 10-year surveillance period. The study used a Markov model and found that the most costeffective strategy was surveillance with mammography alone, provided every 12 months since the incremental cost-effectiveness ratio (ICER) for this strategy compared to no surveillance was € 6051 per QALY gained.
A cost-utility evaluation conducted in The Netherlands [bib_ref] Economic evaluation of four follow-up strategies after curative treatment for breast cancer:..., Kimman [/bib_ref] analysed data (costs in 2008 value) on 299 patients randomised into four groups: (1) hospital follow-up; [bib_ref] Epidemiology of breast cancer, Ban [/bib_ref] nurse-led telephone follow-up; (3) hospital follow-up plus a short educational group programme (EGP); and (4) nurse-led telephone follow-up plus EGP. Hospital followup plus EGP had an ICER of 236 € per QALY compared to the next best alternative nurse-led telephone follow-up plus EGP. The other two strategies were dominated (higher costs and fewer QALYs). The authors concluded that nurse-led telephone follow-up combined with a short EGP could be a cost-effective option. However, they did not estimate the ICER of this strategy compared to standard follow-up. Furthermore, the time horizon of the study (one year) was clearly insufficient to evaluate the costeffectiveness of compared alternatives.
The study of Oltra et al. [bib_ref] Cost-benefit analysis of a follow-up program in patients with breast cancer: a..., Oltra [/bib_ref] found that an intensive follow-up characterised by multiple laboratory and imaging tests triples average costs of the standard clinical follow-up without differences in early detection of relapses during the three years of follow-up. The study of Kokko et al. [bib_ref] Follow-up cost of breast cancer patients with localized disease after primary treatment:..., Kokko [/bib_ref] found that the most expensive strategy doubled the costs of the cheapest one without important differences in breast cancer recurrences among them.
The quality of the evidence on economic evidence was moderate due to indirectness. The study [bib_ref] French women's representations and experiences of the post-treatment management of breast cancer..., Laporte [/bib_ref] considered in the evidence to decision framework was conducted in the UK, and the results may not be applicable to other European countries.
The EtD framework was applied to conclude the assessment. The research question is summarised in [fig_ref] Table 5: Summary of the assessment on the research question [/fig_ref] represents the assessment, carried out in its 12 domains: among others, the certainty of evidence (e.g., no statistically significant differences in The most preferred person to perform follow-up was the medical specialist, but a combination of the medical specialist and breast care nurse was also acceptable to patients. Face-to-face contact was strongly preferred over telephone contact. Follow-up visits every three months were preferred over visits every four, six, or 12 months.
Moderate risk of bias mortality) between different types of follow-up), the important uncertainty and variability in women's values, and the cost-effectiveness of the intervention (which favours non-intensive schedules) are crucial elements in drawing conclusions.
Finally, reports the conclusions summarised by the authors, in the form of a suggestion to perform breast cancer follow-up once a year with a mammography visit, as opposed to other types of regimens.
# Discussion
## Main findings
Our results showed that intensive follow-up, compared with less intensive follow-up including more frequent diagnostic test or visits, does not have beneficial effects on 5-or 10 -year overall mortality or recurrences in women with breast cancer. This finding was consistent between the studies included, and the quality of the evidence was moderate. Among the included studies, two randomised trials showed that intensive follow-up appeared to increase reassurance in patents (data on 250 women; RR 1.28, 95% CI from 1.07 to 1.54) [bib_ref] Popularity of less frequent follow up for breast cancer in randomised study:..., Gulliford [/bib_ref] [bib_ref] Patient initiated follow up of breast cancer, Brown [/bib_ref]. However, the quality of the studies was downgraded due to the inconsistency of studies. The cost of different regimens of follow-up is variable, with more intensive regimes being more expensive but without increases in health benefits; thus less intensive regimes are favoured. From one cost-utility analysis [bib_ref] The clinical effectiveness and costeffectiveness of different surveillance mammography regimens after the..., Robertson [/bib_ref] , an annual visit with mammography results in moderate costs, can be considered cost-effective compared to no surveillance, and is likely to be feasible.
## Our results in the context of previous results
The European Society of Medical Oncology (ESMO) Guidelines on breast cancer recommend regular visits every 3-4 months for the first 2 years after treatment (and gradually decreasing thereafter) in addition to an annual mammography [bib_ref] Primary breast cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up, Senkus [/bib_ref]. American Cancer Society/ American Society of Clinical Oncology Guidelines [bib_ref] Breast cancer follow-up and management after primary treatment, Khatcheressian [/bib_ref] also recommend detailed cancer-related history and physical examination every 3 to 6 months for the first 3 years after primary therapy (and thereafter decreasing) in addition to a yearly mammography. Their recommendations would fall under the definition of a "less intensive follow-up" that, in the majority of the studies included in our review, would include at least a clinical visit and mammography once a year. However, intensive follow-up is still also quite common in clinical practice [bib_ref] indicators based on registers and administrative data for breast cancer: routine evaluation..., Andreano [/bib_ref] [bib_ref] Follow-up care for breast cancer survivors: improving patient outcomes, Chopra [/bib_ref] [bib_ref] Evaluation of the adherence to follow-up care guidelines for women with breast..., Grandjean [/bib_ref] [bib_ref] Heterogeneity of cancer surveillance practices among medical oncologists in Washington and Oregon, Richert-Boe [/bib_ref] and represents a significant workload for radiotherapy, surgery and oncology professionals [bib_ref] It is now the age to define the appropriate follow-up of primary..., Loprinzi [/bib_ref] , in addition to being a costly process.
A recently published systematic review on the effects of breast cancer follow-up showed that standard approaches are as effective as intensive ones; moreover, no differences in quality of life were documented. While considering the health outcomes, including mortality and recurrences, our results confirm the already reported results.
# Limitations and strengths
Many of the studies included in our review were carried out in previous decades, and their results might be slightly outdated, given the recent substantial changes in breast cancer care [bib_ref] Recommendations for breast imaging follow-up of women with a previous history of..., Bucchi [/bib_ref] ; However, our study also took into account further perspectives, by including also women's preferences and values, and economic aspects, as adopted in the GRADE approach. The reviews on women's values and preferences and economic evidence were, however, limited to English and for the last ten years and Medline only, and results would have been Summary of the research question Should women be followed intensively after breast cancer treatment?
## Problem:
Women treated for breast cancer are followed-up for monitoring treatment effectiveness and for detecting recurrences at an early stage, but the frequency of follow-up is under discussion.
Background: Women treated for breast cancer are followed up for monitoring treatment effectiveness and for detecting recurrences at an early stage. Follow-up includes clinical and test examinations like routine haematological and liver function tests, tumour markers, chest X-ray, mammography, bone and liver scans. There is variability in the frequency of medical visits and the tests performed.
Option:
Intensive follow-up.
Comparison: Non-intensive follow-up.
Main outcomes: 1. 10-year mortality due to breast cancer. 2. 5-year mortality due to breast cancer. 3. 5 (or 10)-year breast cancer specific survival. 4. 10-year breast cancer recurrences (loco-regional and distant separately). 5. 5-year breast cancer recurrences (logo-regional and distant separately). 6. Quality of life of breast cancer patients 2 (or 5) years after diagnosis. 7. Patient satisfaction with follow-up.
## Setting:
Breast cancer centres/other healthcare services.
Perspective: Population.
Legend: this table represents the first part of the Evidence to Decision framework Despite the doubts that intensive follow-up care could improve survival in patients after breast cancer, intensive followup is quite common in clinical practice and represents a significant workload for radiotherapy, surgical and oncologic departments [bib_ref] It is now the age to define the appropriate follow-up of primary..., Loprinzi [/bib_ref] , and it is also costly. The evidence showed uncertain differences in overall mortality at 5 and 10-year follow-up (high quality evidence), and uncertain differences in recurrences at 5 years of follow-up (moderate quality evidence).
The evidence showed significant differences in reassurance of women in favour of intensive follow-up (very low quality evidence), and convenience in favour of non-intensive followup (low quality evidence).
There was missing research evidence in respect to the outcomes: 5 and 10-year breast cancer specific survival, 10-year breast cancer recurrences and quality of life of breast cancer patients 2 or 5 years after diagnosis. The evidence on 5-and 10-year overall mortality was of high quality, and did not favour intensive versus standard follow-up. The evidence on 5-year cancer recurrences was of moderate quality, and there were uncertain differences between intensive and standard follo-up; similar conclusions apply to cancer recurrences at any time. The evidence of women satisfaction was of very low quality (reassurance domain) and of moderate quality (convenience domain). The evidence on values for women was of low quality (inconsistency among studies). The evidence on economic evaluations was of high quality, and favoured non-intensive follow-up. In the base-case scenario of a cost-utility analysis of different follow-up strategies carried out in the UK, the strategy with the highest net benefit, and most likely to be considered costeffective, was surveillance mammography alone every 12 months at a societal willingness to pay for a qualityadjusted life year of either £20,000 or £30,000. The incremental cost-effectiveness ratio for surveillance mammography alone every 12 months compared with no surveillance was € 6051 (2008 value) more robust if such reviews were carried out with a broader scope. Moreover, the suggestion for less intensive follow-up was built by using the EtD: this is a new approach in the clinical oncology field, but has been previously used already in breast cancer screening, colon cancer screening [bib_ref] National Guidelines for colorectal cancer screening in Saudi Arabia with strength of..., Alsanea [/bib_ref] , as well as in other contexts [bib_ref] Conceptual frameworks and key dimensions to support coverage decisions for vaccines, Gonzalez-Lorenzo [/bib_ref]. The EtD explicitly takes into account factors related, among others, to the quality of evidence, desirable and undesirable effects, values, resources and feasibility, that altogether constitute a comprehensive approach to a decision-making exercise. The suggestion reported in this paper was made by a multidisciplinary group of authors, but it should not be considered as a recommendation from a guideline panel.
## Implications for practice and research
The main expectation from an intensified follow-up from a women's perspective was reassurance and increased sense of security. This finding raises the need to better inform women on the lack of evidence of effect of intensive follow-up on clinical outcomes of mortality and recurrences. However, it needs to be considered that the follow-up visit may also have additional aims than detection of recurrence, such as motivating women to continue endocrine treatment during the follow-up period, providing information about long-term adverse effects of treatment, and helping in their management, as well as providing psychosocial support [bib_ref] French women's representations and experiences of the post-treatment management of breast cancer..., Laporte [/bib_ref]. These other aims of follow-up are very important in the light of the high prevalence of e.g., depression (varying from 9.4% to 66.1%), and anxiety (varying from 17.9% to 33.3%) among breast cancer survivors [bib_ref] The prevalence of long-term symptoms of depression and anxiety after breast cancer..., Maass [/bib_ref]. These additional aspects should not be neglected and they should be better explored while evaluating the effects of different follow-up strategies. Hence, further well-designed studies should be performed. There is a need to balance and prioritise these different outcomes, including also additional patientcentred endpoints described above, as well as including undesirable effects of more frequent investigations. Moreover, organisational aspects related to the coordination of follow-up activities (i.e. nurse-led and GP-led activities, etc.) are only analysed in few studies [bib_ref] Randomized trial of long-term follow-up for early-stage breast cancer: a comparison of..., Grunfeld [/bib_ref] and should be better explored, as they may impact on the acceptance of the protocol by women, healthcare providers, etc. as well as on costs and feasibility. From the clinical point of view, annual mammography is well justified to detect potential new primary or local recurrences. On the contrary more intensive follow-up schedules including additional diagnostic tests, such as breast MRI, liver ultrasound or bone scans could result in large costs without sufficient evidence regarding their benefits or harms.
In summary, based on these findings, less intensive follow-up could be recommended, although the exact format of the follow-up visit would need to be further clarified, as the studies used quite different follow-up schedules and tests. The treatment of breast cancer has become increasingly individualised [bib_ref] Personalized medicine in breast cancer: a systematic review, Cho [/bib_ref] as the risk of breast cancer recurrences is very variable and is related, among other variables, to genetic predisposition of individual women, breast cancer characteristics and its treatment. Therefore, also the follow-up should be individualised based on the risk estimates, and on women's perceptions and values. A "one size fits all" approach may not be relevant.
# Conclusion
Based on the evaluation of clinical and economical outcomes carried out, a less intensive follow-up could be recommended. Patients should be provided with accurate information on the benefits (or lack of those) and harms of intensive follow-up. Resources could thus be mobilised to other aspects of breast cancer care, or other areas of healthcare, potentially increasing equity in society.
## Additional files
[fig] Figure 1: PRISMA flowcharts. Legend: Flowcharts representing the selection of studies for health outcomes (a), values and preferences (b), and resource utilisation and costs (c) [/fig]
[fig] Figure 2: Estimates of effect of intensive vs. standard follow-up on breast cancer outcomes [/fig]
[fig] Additional file 1: Search strategy for the evidence of effects. (DOCX 100 kb) Additional file 2: Search strategy for women's values and preferences evidence. (DOCX 95 kb) Additional file 3: Search strategy for economic evidence. (DOCX 97 kb) Abbreviations CERQual: Confidence in the Evidence from Reviews of Qualitative research; CI: Confidence intervals; EtD: Evidence to Decision; GRADE: Grading of Recommendations Assessment, Development, and Evaluation; HR: Hazard ratios; ICER: Incremental Cost Effectiveness Ratio; OR: Odds ratios; PICO: Patient, Intervention, Comparison, Outcomes; QALY: Quality Adjusted Life Years; RR: Risk ratios [/fig]
[table] Table 1: Summary and short description of the six included randomised clinical trials phy, chest X-ray, and bone scan. Patient-initiated follow-up: Patients received written information on the signs and symptoms of recurrence, and the invitation [/table]
[table] Table 2: Evidence profiles for selected health outcomes related to intensive vs. standard follow-up in breast cancer patients [/table]
[table] Table 3: Summary and short description of the three included studies on women's preferences and values [/table]
[table] Table 5: Summary of the assessment on the research question [/table]
|
Comparison of the sixth and the seventh editions of the UICC classification for intrahepatic cholangiocarcinoma
Background: The current seventh edition of the TNM classification for intrahepatic cholangiocarcinoma (ICC) includes tumor number, vascular invasion, lymph node involvement but no longer the tumor size as compared to the sixth edition. The impact of the seventh edition on stage-based prognostic prediction for patients with ICC was evaluated.Methods: Between 03/2001 and 02/2013, 98 patients with the diagnosis of an ICC were surgically treated at our center. Median survival times were calculated for these patients after separate classification by both sixth and seventh editions.Results: Median overall survival was increased in patients classified to the lower tumor stages I and II using the seventh as compared to the sixth edition: stage I (54.9 vs. 47.3 months), stage II (19.9 vs. 18.9 months), stage III (17.2 vs. 19.9 months), and stage IV (23.2 vs. 15.3 months), respectively. The seventh edition definition of the T category resulted in an increased median survival regarding the T1 (50.4 vs. 47.3 months) as well as the T2 category (19.9 vs. 15.6 months) and revealed a reduced median survival of patients within the T3 (21.6 vs. 24.8 months) as well as the T4 category (19.9 vs. 27.0 months).Conclusions:The UICC seventh edition TNM classification for ICC improves separation of patients with intermediate stage tumors as compared to the sixth edition. The prognostic value of the UICC staging system has been improved by the seventh edition.Trial registration The data for this study have been retrospectively registered and the study has been approved by the ethic committee of the medical faculty of the University Hospital of Essen, Germany (license number 15-6353-BO).
# Background
Cholangiocarcinoma is a primary cancer of the bile ducts that arises from malignant transformation of bile duct epithelium [bib_ref] Multidisciplinary care of patients with intrahepatic cholangiocarcinoma: updates in management, Lafaro [/bib_ref]. This tumor entity can be classified into intrahepatic cholangiocarcinoma (ICC) and extrahepatic tumors (hilar and distal bile duct) depending on its location within the biliary tree [bib_ref] Guidelines for the diagnosis and management of intrahepatic cholangiocarcinoma, Bridgewater [/bib_ref]. Ten to fifteen percent of all primary liver cancers are ICCs and these neoplasms are the second most common primary liver cancers following hepatocellular carcinomas. Different incidences have been reported ranging from > 80 per 100,000 population in Thailand to lower incidences in the western world, e.g., Canada with 0.3 per 100,000 population [bib_ref] Guidelines for the diagnosis and management of intrahepatic cholangiocarcinoma, Bridgewater [/bib_ref] [bib_ref] Cholangiocarcinoma: epidemiology and risk factors, Bragazzi [/bib_ref] [bib_ref] Prognostic assessment of three single-nucleotide polymorphisms (GNB3 825C>T, BCL2-938C>A, MCL1-386C>G) in extrahepatic..., Fingas [/bib_ref]. Current studies report that the incidence of ICC is increasing [bib_ref] A comparison of trends in the incidence of hepatocellular carcinoma and intrahepatic..., Mcglynn [/bib_ref] [bib_ref] Hedgehog inhibition promotes a switch from type II to type I cell..., Kurita [/bib_ref]. Comparable to hilar cholangiocarcinoma [bib_ref] Comparison of the sixth and the seventh editions of the UICC classification..., Juntermanns [/bib_ref] , surgery is the only effective curable treatment up to now [bib_ref] Multimodality treatment of intrahepatic cholangiocarcinoma: a review, Simo [/bib_ref] [bib_ref] Operations for intrahepatic cholangiocarcinoma: single-institution experience of 158 patients, Lang [/bib_ref]. After surgical therapy, the reported 5-year survival rates range between 30 and 35% [bib_ref] Operations for intrahepatic cholangiocarcinoma: single-institution experience of 158 patients, Lang [/bib_ref] [bib_ref] Extended liver resection for intrahepatic cholangiocarcinoma: a comparison of the prognostic accuracy..., Jonas [/bib_ref] [bib_ref] Indications for surgical treatment of intrahepatic cholangiocarcinoma with lymph node metastases, Uenishi [/bib_ref]. Therefore, a staging system that exactly separates patients suffering from hilar ICC into prognostic groups is desirable to support patient stratification for treatment with regard to future multimodal perioperative therapeutic strategies. The current seventh edition of the American Joint Committee on Cancer (AJCC) staging manual has introduced a new distinct staging system for ICC based on prognostic factors including tumor number, vascular invasion, lymph node involvement but excluded the tumor size in comparison to the sixth edition. Several studies have focused on the comparison of the sixth and seventh AJCC manual for ICC and found the seventh edition quite suitable for estimation on patients' prognosis [bib_ref] Comparison of the prognostic accuracy of the sixth and seventh editions of..., Ribero [/bib_ref] [bib_ref] A proposed staging system for intrahepatic cholangiocarcinoma, Nathan [/bib_ref]. Comparable to our previous hilar cholangiocarcinoma study [bib_ref] Comparison of the sixth and the seventh editions of the UICC classification..., Juntermanns [/bib_ref] , the purpose of the present work was to compare the prognostic accuracy of the sixth and the new seventh edition of the AJCC staging systems to predict survival after liver resection for ICC in patients treated at the West German Cancer Center, as one of the largest hepatobiliary centers in Germany.
# Methods
## Patients and specimens
Between 03/2001 and 02/2013, 98 patients with the diagnosis of an ICC were surgically treated at our center. This study was conducted in accordance with the Helsinki Declaration of 1975 and approved by the ethic committee of the medical faculty of the University Hospital of Essen, Germany (license number 15-6353-BO). Routine histopathological workup was conducted for all resected tumors by the Department of Pathology of the University Hospital Essen confirming the diagnosis of an ICC. Our cohort included 42 male (42.86%) and 56 female (57.14%) patients, with an average age of 62.9 (± 11.5) years. All types of resection margins (R0, R1, and R2) and all cases of irresectable disease were included in the study cohort. Follow-up data were prospectively recorded until February 2014.
## Histopathological processing
Surgical specimens were placed in 4% neutral-buffered formalin prior to histopathological processing, dehydrated, and then cleared using an automated standard procedure (Shandon Pathcentre, Thermo Fisher Scientific Inc., USA) before paraffin embedding in Paraplast (McCormick Scientific, USA). From each paraffin block, 3-to 5-μm sections were cut and Hematoxylin & Eosin slides were prepared adherent to in-house standards. Histopathology reports were available for every case and included macroscopic and microscopic tumor evaluations, in a continuous text summary, and the epicritical report including the TNM (Classification of Malignant Tumors). Data including operative reports and surgical pathology reports of all patients were entered prospectively into a computer database. Cases were stratified according to the sixth and seventh editions of the AJCC/ UICC (International Union Against Cancer) TNM classification algorithm for ICC. The sixth and seventh editions of the AJCC/UICC TNM classification algorithm for ICC are compared in [fig_ref] Table 1: Sixth and seventh editions of the AJCC/UICC TNM classification algorithm for intrahepatic... [/fig_ref].
# Statistical analysis
Changes in the distribution of TNM classifications and UICC stages were compared between the sixth and seventh editions, and median survival and survival ranges were calculated independently for each classification. Additionally, 1-, 3-, and 5-year survival rates were calculated. Overall survival was defined as the time from the date of surgery to the date of ICC-specific death or the date of last follow-up. Kaplan-Meier survival plots were generated and comparisons made using the log-rank statistic. All parameters that revealed significant results in univariate Kaplan-Meier analysis were subjected to multivariate Cox proportional hazards regression survival analysis to evaluate the prognostic value of the TNM categories and UICC stages derived from the sixth and seventh editions. Differences of p < 0.05 were considered to be statistically significant.
# Results
A total of 98 patients suffering from ICC were surgically treated at our center between 03/2001 and 02/2013. In this retrospective study, we compared the impact of applying either the sixth or seventh editions of UICC tumor staging to stratify median patient survival or predict prognosis in this patient cohort. A summary of differences between the sixth and seventh editions of UICC staging of ICC and the respective TNM categories is presented in [fig_ref] Table 1: Sixth and seventh editions of the AJCC/UICC TNM classification algorithm for intrahepatic... [/fig_ref]. We compared the influence of tumor staging using either the sixth or seventh UICC editions on the 1-, 3-, and 5-year survival of patients in this cohort [fig_ref] Table 2: Median survival by UICC stage [/fig_ref].
The median overall survival for patients was staged according to the sixth or seventh editions and broken down by tumor stage: stage I (47.3 or 54.9 months), stage II (18.85 or 19.9 months), stage III (19.9 or 17.2 months), and stage IV (15.3 or 23.15 months), respectively. Staging according to the seventh edition resulted in an increased median overall survival for patients suffering from ICC in the lower UICC tumor stages I and II. A change in tumor staging occurred in 45 out 98 patients (45.9%) by using the current classification. We also compared the ability of the sixth and seventh editions of the UICC classification to accurately predict patient prognosis based on UICC tumor stage. Kaplan-Meier survival analysis for patients with different UICC tumor stages revealed that the seventh edition more accurately separates these groups [fig_ref] Figure 1: Comparison of survival prediction [/fig_ref]. The prognostic values of the sixth and seventh Lymph nodes positive for cancer can either be defined as regional spreading of the tumor or as metastases. The definition of regional lymph nodes has been modified within the N (lymph nodes) and M (metastases) categories of the seventh edition of the UICC classification. The overall impact of a positive lymph node on the extent of the disease is emphasized by the seventh edition, such that involvement of any lymph node results in tumor stage IV which has been a tumor stage III in the sixth edition. Applying the seventh edition resulted in reclassification of five patients as M1 based on histological positive lymph node and, therefore, a reclassification from stage IIIc into stage IVa in the current edition. Median survival for patients classified for the N category using the sixth and seventh editions was also compared.
# Discussion
In this single-institution study, we found that the seventh edition of the UICC TNM classification for ICC slightly more accurately stratifies patients suffering from this neoplasm. Particularly, the guidelines for the current seventh edition AJCC staging manual has introduced a new distinct staging system for ICC based upon prognostic factors including tumor number, vascular invasion, lymph node involvement but no longer the tumor size as opposed to the sixth edition. This redefinition enables a more appropriate patient stratification in our cohort. Another novelty is the new definition of distant lymph node metastases resulting in several tumor stage upgrades in our cohort with a change of median survival regarding UICC tumor stage IV.
Tumor staging according to AJCC/UICC seventh edition was accompanied by an increased survival in tumor stages I and II allowing a more appropriate prognosis stratification. The reason for this was an altered classification of 45 out of 98 patients using the current classification. Thus, a total of 67 (seventh edition) instead of 51 (sixth edition) patients is classified to the lower tumor stages I and II in this study. This finding is consistent with a multicenter study conducted by Ribero et al. [bib_ref] Comparison of the prognostic accuracy of the sixth and seventh editions of..., Ribero [/bib_ref]. This group also reported that the majority of their patients displayed a T1 or a T2 tumor stage and an UICC/AJCC stage I or II. In our opinion, this new distribution allows a better stratification within our cohort and, therefore, a better selection for adjuvant therapy strategies.
We also found several changes regarding the T status applying the seventh edition. In our cohort, the new T categories enabled a more appropriate patient stratification, especially for T1 and T2. For T2, this fact is due to the loss of tumor size evaluation causing a downstaging for multiple tumors from T3 to T2b. Nathan et al. [bib_ref] A proposed staging system for intrahepatic cholangiocarcinoma, Nathan [/bib_ref] could demonstrate that the tumor size does not correlate with any additional prognostic value by analyzing the survival data of 598 patients following resection for ICC. Thus, they concluded that the sixth edition AJCC/ UICC T classification failed to stratify T2 and T3 cohorts into two distinct prognostic groups. Our data support this finding as the prognostic value of the seventh edition T category staging reached a higher level of significance as compared to the sixth edition T category staging (p = 0.004 vs. 0.02). Another novelty in the seventh AJCC/UICC staging system is the introduction of a separate classification of the T4 category "periductal tumor infiltration. " In this study, only two patients were found to show periductal infiltration. Both patients also displayed positive lymph node metastasis and were therefore staged as AJCC/UICC IVa. Thus, the prognostic significance of this new T4 category remains uncertain in our cohort and needs to be investigated more specifically in upcoming studies with larger patient numbers.
The seventh edition also re-evaluates the presence of regional lymph node involvement regarding the N and M classifications emphasizing the importance lymph nodes infiltrated with cancer for UICC tumor staging. This resulted in the reclassification of patients with lymph node metastases into UICC stage IV by the seventh edition (formerly stage III in the sixth edition). Additionally, distal lymph nodes were reclassified in the current classification manual. This novelty affected a total of five patients in our cohort categorized as stage III according to the sixth edition of the TNM classification. Due to a celiac or mesenteric lymph node infiltration (N1), the tumors were upstaged to metastatic (M1, stage IV) disease according to the seventh edition. Similar to our previous study of perihilar cholangiocarcinoma [bib_ref] Comparison of the sixth and the seventh editions of the UICC classification..., Juntermanns [/bib_ref] , this novelty did not significantly affect patient survival comparing both groups. Interestingly, the new M staging in our collective caused a prolonged survival in the M1 group. This finding raises the question, if the upgrade from N1 to M1 really mirrors an altered prognosis as intended by the AJCC/UICC. Nevertheless, Farges et al. [bib_ref] AJCC seventh edition of TNM staging accurately discriminates outcomes of patients with..., Farges [/bib_ref] recommended that a routine lymphadenectomy at the time of surgery for ICC should become the standard of care, which again emphasizes the importance of lymph node involvement.
# Conclusion
In conclusion, based on this study of 98 patients treated for ICC at a single institution, the categorization of UICC tumor stages by the seventh UICC edition enables a better patient stratification than the sixth edition. The current edition emphasizes the importance of lymph node involvement and periductal infiltration instead of tumor size. Thus, the seventh edition more appropriate separates intermediate tumor stages as reflected by the median patient survival for this cohort and confers a higher prognostic value to the tumor stage. This should facilitate the stratification of patients diagnosed with ICC into different risk groups resulting in better customized multimodal perioperative treatment strategies. Our data show that particularly the T categories of the seventh edition enable a more appropriate prediction of patient survival than the corresponding categories from the sixth edition. A considerable number of patients were staged differently by the seventh edition in the present study. This fact should be considered when comparing previous studies employing the sixth edition with new data from ICC patients. Finally, the T4 category (periductal infiltration) and the reclassification of distant lymph nodes need to be re-evaluated in upcoming studies with larger patient numbers.
[fig] Figure 1: Comparison of survival prediction (n = 98) after surgery using the sixth (left) and seventh (right) editions of the UICC tumor classification. Kaplan-Meier analysis was based on tumor stage (a) T category, (b) N category, (c) and M category. (d) Significant survival differences (p values) were assessed using the log-rank test Juntermanns et al. Eur J Med Res (2018) 23:29 [/fig]
[table] Table 1: Sixth and seventh editions of the AJCC/UICC TNM classification algorithm for intrahepatic cholangiocarcinoma (ICC) [/table]
[table] Table 2: Median survival by UICC stage (n = 98) using the sixth and seventh editions of the TNM classificationSurvival was calculated using the Kaplan-Meier method and compared using the log-rank test. Cox proportional hazards regression analysis was performed to evaluate the prognostic value of the UICC stage according to the sixth and seventh editions tumor) category, we compared median survival of patients based on either the sixth or seventh edition T category classification (Table 3). The seventh edition definition of the T category resulted in increased median survival in the T1 (50.4 vs. 47.3 months) as well as the T2 category (19.88 vs. 15.6 months) and revealed a reduced median survival in patients within the T3 category (21.62 vs. 24.8 months) and the T4 category (19.86 vs. 26.95 months). The prognostic value of the seventh edition T category staging reached a higher level of significance as compared to the sixth edition T category staging (p = 0.004 vs. 0.02 by log-rank test; Table 3). Thus, the distribution shift of patients suffering from multiple tumors from T3 category in T2b category in the seventh classification clearly improved patient stratification particularly for this intermediate tumor stage. [/table]
[table] Table 3: Median survival of patients classified N0 did not change (41.26 months) and also remained the same for N1 (22.7 months). Log-rank test confirmed significance for both editions (p = 0.05) regarding the N status. [/table]
|
Minimally Invasive Postmortem Intestinal Tissue Sampling in Malnourished and Acutely Ill Children Is Feasible and Informative
Background.Intestinal disorders such as environmental enteric dysfunction (EED) are prevalent in low-and middle-income countries (LMICs) and important contributors to childhood undernutrition and mortality. Autopsies are rarely performed in LMICs but minimally invasive tissue sampling is increasingly deployed as a more feasible and acceptable procedure, although protocols have been devoid of intestinal sampling to date. We sought to determine (1) the feasibility of postmortem intestinal sampling, (2) whether autolysis precludes enteric biopsies' utility, and (3) histopathologic features among children who died during hospitalization with acute illness or undernutrition.Methods.Transabdominal needle and endoscopic forceps upper and lower intestinal sampling were conducted among children aged 1 week to 59 months who died while hospitalized in Blantyre, Malawi. Autolysis ratings were determined for each hematoxylin and eosin slide, and upper and lower intestinal scoring systems were adapted to assess histopathologic features and their severity.Results. Endoscopic and transabdominal sampling procedures were attempted in 28 and 14 cases, respectively, with >90% success obtaining targeted tissue. Varying degrees of autolysis were present in all samples and precluded histopathologic scoring of 6% of 122 biopsies. Greater autolysis in duodenal samples was seen with longer postmortem interval (Beta = 0.06, 95% confidence interval, 0.02-0.11). Histopathologic features identified included duodenal Paneth and goblet cell depletion. Acute inflammation was absent but chronic inflammation was prevalent in both upper and lower enteric samples. Severe chronic rectal inflammation was identified in children as young as 5.5 weeks.Conclusions. Minimally invasive postmortem intestinal sampling is feasible and identifies histopathology that can inform mortality contributors.
Undernutrition underlies 45% of child deaths and remains prevalent in low-and middle-income countries, including Malawi, where the under-5 year mortality rate is 42 per 1000 live births: 11% above the global average. Causes of undernutrition include food insecurity, recurrent infections, and environmental enteric dysfunction (EED). A largely asymptomatic condition, EED is prevalent in settings with inadequate sanitation and hygiene facilities. It is characterized by intestinal inflammation and villus blunting, malabsorption, permeability facilitating microbe translocation, and systemic inflammation.
Accelerating mortality reductions requires improved cause of death (COD) understanding, particularly within populations in which undernutrition and infections are the leading mortality risk factors. Full autopsy is the most comprehensive and accurate method for COD assignment, but is often infeasible because of resource constraints or societal unacceptance. Minimally invasive tissue sampling (MITS) uses transcutaneous needle organ and fluid sampling and is increasingly used as a validated, nondisfiguring, and more acceptable alternative to full autopsy. However, MITS lacks enteric sampling because intestines, largely untethered within the abdominal cavity, likely enable needles to push away, potentially hindering sampling of this organ. Endoscopy may afford visualization and sampling but has rarely been used, and to our knowledge, never attempted in pediatric postmortem studies. Several potential barriers could preclude this approach, including jaw, pyloric sphincter or anal rigor mortis restricting intubation, and stomach contents or stool interfering with visualization and sampling. Tissue autolysis, driven by microbes and cellular enzymes, may render biopsies uninformative because intestine is the first organ to degrade postmortem.
Lack of gastrointestinal interrogation impedes understanding of intestinal infections, EED, and other enteric contributions to childhood deaths. The MITS in Malawi (MiM) study sought to assess postmortem endoscopic intestinal sampling feasibility, determine whether autolysis precludes enteric biopsy utility, and examine enteric histopathologic features among children who died during hospitalization with acute illness or undernutrition.
# Methods
## Study design, setting, participant selection
MiM was conducted at Queen Elizabeth Central Hospital (QECH) in Blantyre, Malawi, a national referral and teaching hospital. MiM started as a substudy at the QECH site of the Childhood Acute Illness & Nutrition (CHAIN) study, which assessed risk factors for mortality among children hospitalized with acute illness or undernutrition in low-and middleincome countries. Because of low enrollment and case fatality among CHAIN-enrolled patients, MiM recruitment expanded to 2 other QECH-based studies and the general pediatric wards. Exclusion criteria were: known terminal illnesses, congenital syndromes, injuries, surgical conditions, and age <1 week or >59 months. Recruitment was August 20, 2018-April 9, 2020.
Study staff approached parents/guardians of eligible children after a respectful period following death. Written informed consent was obtained from parents/guardians. Assistance with coffin purchase, transportation, and grief support was offered to all approached parents/guardians, regardless of consent to participation.
## Sampling procedures
MITS procedures commenced as soon as feasible after obtaining consent. A study gastroenterologist or endoscopist used Olympus GIF-P140, GIF XP-160, or GIF-160 gastroscopes and Olympus FB-231K.A or FB-230.K.A forceps. Gastric and duodenal fluid were aspirated, either prebiopsy (preferred) or postbiopsy after flushing with buffered saline. Biopsies were obtained from stomach, the first (D1), second (D2), and most distal accessible duodenal segments (D3/D4), rectum; and colon last as proximally as possible. Postendoscopy, standard MITS procedures were performed.
Transabdominal intestinal needle sampling was initially considered impractical, but was adopted midstudy to assess feasibility. Bard Monopty 16G 100-mm needles were introduced just inferior to the umbilicus and right-side posterolateral for small bowel and ascending colon samples, respectively. All tissues were paraffin embedded, sectioned, and stained with hematoxylin and eosin.
## Data collection
Antemortem clinical data were collected during admission for coenrolled patients and otherwise extracted from medical records. Antemortem clinical test results were abstracted from laboratory reports. Because of incomplete antemortem anthropometric data, postmortem anthropometry was used, as is standard in MITS studies. Two study personnel measured mid-upper arm circumference, weight, and length and where discrepant, measured a third time. The average of 2 closest measurements was used. Z scores were calculated using World Health Organization Anthro software. Severe acute malnutrition (SAM) was defined by mid-upper arm circumference <11.5 cm (among ≥6-month-old children), weight-forlength z score <-3, or nutritional edema.
A gastrointestinal pathologist (T-C.L.) rated each intestinal hematoxylin and eosin slide for autolysis, categorized as: (0) no autolysis; (1) <50% of tissue autolyzed; (2) 50%-75%; (3) >75% but <100%; and (4) complete autolysis. All sufficiently intact duodenal slides were scored using the EED Biopsy Consortium histology index, whereby 10 histologic parameters are assigned semiquantitative scores from 0 (normal) to 3 or 4 (severe pathology). Lower intestinal histology was similarly scored based on 4 ulcerative colitis histological parameters. Because slide preparation quality or autolysis may render certain features unscorable, we calculated total score percent (total score divided by maximum possible among scored parameters) to describe overall histopathologic severity.
# Data analysis
Analysis was primarily descriptive because this study was designed to examine the feasibility of adding intestinal sampling to standard MITS procedures and not powered to comprehensively test associations with intestinal autolysis or histopathology. However, once feasibility was determined, we leveraged obtained tissues to explore such associations.
We examined autolysis variability by anatomic location and sampling method (endoscopic vs transabdominal) using paired t tests. Autolysis and histologic score intraindividual variability was determined with intraclass correlation coefficients (ICC) using a 2-way mixed effects model. We used univariate linear regression to explore the effect on autolysis by age, sex, admission diagnosis (sepsis, gastroenteritis), nutritional status (per postmortem anthropometry), human immunodeficiency virus (HIV) status, exposure to antibiotics that are potentially more gut microbiome disruptive (ceftriaxone, ciprofloxacin) versus less so (amikacin, amoxicillin, flucloxacillin, fluconazole, gentamicin, metronidazole, penicillin, trimethoprim-sulfamethoxazole), refrigeration duration, postmortem interval (PMI; time to intestinal sampling), and endoscopically visualized stomach contents (restricted to autolysis of D1 because of proximity). All bodies were refrigerated. Refrigeration time highly correlates with PMI, so refrigeration time was calculated as percentage of PMI. Because the PMI/refrigeration duration and autolysis relationship was of primary a priori interest, they were assessed together in a multivariate model.
We used univariate linear regression to explore influences on histologic severity by age, sex, admission diagnosis (sepsis, gastroenteritis, anemia [can be caused by malabsorption]), HIV status, nutritional status, length of hospital stay, higher dysbiotic potential antibiotics (described previously), and hepatic steatosis (indicator of severe malnutrition).
Data were insufficient to relate blood culture results, blood gas acidosis, antibiotics (any use), and stomach pH to autolysis, and malaria diagnosis and acidosis to histologic severity. Analyses were conducted using Stata/SE 16.1. Significance was defined by 2-tailed alpha = .05. Graphics were generated using RStudio version 1.2.5019.
# Ethical approval
The Malawi National Health Sciences Research Committee and Oxford University Ethics Committee (OxTREC 34-16) provided ethical approval. The University of Washington Institutional Review Board exempted the study from review (STUDY00003689).
# Results
Seventy-five children eligible for this study died on enrollment days. Reasons for not seeking consent included: inability to mobilize a key study member within an appropriate time (n = 8), lack of notification by ward staff (n = 5), or unknown reasons (n = 4). Of 58 caregivers approached, 29 (50%) consented. Reasons for refusal were: lack of perceived benefit (n = 9), preference to immediately take body home (n = 7), cultural/religious concerns (n = 5), poor relationship with healthcare staff (n = 2), consenters unavailable (n = 2), child too young (n = 1), or unknown reasons (n = 3).summarizes participant characteristics.
## Procedure feasibility
Upper and lower endoscopic sampling was attempted for all but the final case because of corona virus disease 2019-related circumstances. Upper and lower endoscopy mean (range) duration were 34.4 (20-57) and 10.5 (6-20) minutes, respectively.
At least 1 intestinal biopsy was endoscopically obtained from each of 28 cases, including: 21 D1 (75%), 20 D2 (71%), 23 D3/ D4 (82%), 5 colon (18%), and 26 rectum (93%).
Presence of stomach contents (eg, milk) and stool often caused insufficient equipment insufflation by procedure end. This explains colon sampling paucity because colon was the final sampling location, and twice interfered with rectal sampling. One case lacked duodenal sampling because of difficulty negotiating the C-loop. Pyloric sphincter constriction precluded duodenal intubation in 3 cases (PMIs between 2 and 14 hours). Transabdominal needle sampling of small and large intestine was successful in 14 (100%) and 13 (93%) of the last 14 cases, respectively. Across all participants, 26 (90%) and 29 (100%) had at least 1 successful small and large bowel biopsy, respectively, regardless of method.
## Tissue autolysis
All intestinal samples displayed some autolysis. Thirty-two biopsies (26%) had <50% autolysis, 59 (49%) had 50%-75%, and 31 (25%) had >75% autolysis. None were completely autolyzed. The mean autolysis rating was 1.9, corresponding to slightly <50%-75% of tissue autolyzed. In contrast, only 1 case demonstrated extraintestinal tissue autolysis, in the brain. Autolysis grading was inconsistent across biopsies within the same individual (ICC = 0.16), but did not differ by gut location. Among cases with both sampling methods, mean autolysis rating was higher in transcutaneous (2.3, standard deviation [SD] = 0.8) than endoscopic biopsies (1.9, SD = 0.7) (P = .040). Upper intestinal autolysis was 0.4 points (out of 4) higher with a 6-hour increased PMI (95% confidence interval [CI]. 0.1-0.7, P = .01), without meaningful change after adjusting for refrigeration (as percent of PMI) , . No clear relationship was appreciated between PMI and lower intestinal autolysis.
Rectal autolysis was 0.5 points higher among children diagnosed with sepsis (95% CI, 0.1-0.9, P = .032). Endoscopicallyobtained small intestine autolysis was 0.6 points lower in those with SAM (95% CI, -1.1 to -0.1, P = .014). Antibiotic administration during hospitalization was universal, but those who Autolysis ratings were coded as following: (0) no autolysis, (1) <50% of tissue autolyzed, (2) 50%-75%, (3) >75% but some intact tissue, and (4) complete autolysis. c N represents the number of biopsies that could be examined for histopathologic features (ie, autolysis did not preclude scoring of at least 1 histologic feature). d Some tissues were unable to be scored because extensive autolysis impeding the identification of histologic features. e Autolysis rendered some histologic features nonscorable; therefore, total score percent (total score divided by maximum possible score of available criteria) is used to summarize histopathologic findings. Only tested for association with D1 because of proximity to gastric contents.
received higher dysbiotic potential antibiotics had 0.5-point lower autolysis than those receiving other antibiotics, although this finding was limited to endoscopically obtained duodenal samples (95% CI, -1.0 to -0.03, P = .037). Those with endoscopically visible gastric contents had 0.5-point lower autolysis (only assessed in D1) (95% CI, -1.0 to -0.1, P = .031).
## Intestinal histopathologic disease severity
Autolysis precluded scoring of 7 (6%) slides. At least 1 slide per case, totaling 115 slides (94%), were scored: 20 D1 (95%), 20 D2 (100%), 22 D3/D4 (96%), 5 colon (100%), and 26 rectal (100%) endoscopic biopsies and 11 transabdominal small (79%) and 11 large intestine samples (85%),. Scores were inconsistent across tissues within individuals (ICC = 0.22 for all biopsies, ICC = 0.39 restricted to upper intestine only). Mean scores did not differ by enteric location, . Transabdominally sampled small bowel had lower mean scores than endoscopically obtained (mean total score percent = 6% and 20%, respectively, P = .006).
Of upper intestinal histology features, Paneth cell depletion was most severely abnormal (mean score = 2 of 3), followed by chronic inflammatory infiltration (mean score = 1.8 of 3) and goblet cell depletion (mean score = 2 of 4) ,. Autolysis frequently precluded assessment of foveolar metaplasia, villus architecture, intraepithelial lymphocytes, and epithelial detachment. When assessable, mean scores were low (0, 0.7, 1, 1.3, and 1.5, respectively). Eosinophilic and neutrophilic infiltration were largely assessable, but uncommon (average scores 0.2 and 0, respectively). Lower intestine lacked evidence of neutrophilic infiltration or ulcerations (although the latter was infrequently assessable). Chronic rectal inflammation was prevalent and often severe (mean score = 2 of 3), including among 29% (n = 4) of infants younger than 6 months . We explored factors that might have influenced this trend, but no relationships were identified. We found no relationships between a priori defined characteristics and histology scores.
# Discussion
To our knowledge, this is the first pediatric use of minimally invasive intestinal sampling. In 22 adults from Hong Kong, Fan et al assessed postmortem laparoscopic and thoracoscopic approaches to various organs and endoscopy of stomach and great vessels. They also visualized (without biopsying) the duodenum endoscopically in 1 patient. Postmortem upper endoscopy was attempted by Denzer et al in 20 German adults. They reached D2 in 17 cases (stomach contents, pyloric atony, and anatomic issues precluded intestinal intubation in the others). However, intestinal sampling was only attempted (successfully) in 1 case. We are unaware of postmortem lower bowel endoscopy reports. The environmental enteric dysfunction (EED) Biopsy Consortium initially published the use of 11 upper intestinal criteria; however, has since identified that the Brunner glands parameter has a negative relationship with EED, hence we excluded this parameter, but retained the other 10 original parameters. In our study, jaw rigor mortis did not inhibit oral intubation, though occasionally instrumented jaw opening was required. Anal rigor mortis, when present, was easily overcome. Pyloric sphincter rigor mortis impeded duodenal intubation in 3 cases. This phenomenon arises 2 hours postmortem and resolves by 12-14 hours in uncooled bodies. It is encouraging that this obstruction was not more prevalent in this study, where most procedures commenced within 14 hours and all bodies underwent refrigeration. Stomach contents did not preclude endoscopic duodenal sampling. However, food/stool contents frequently blocked insufflation of endoscopy equipment by colon sampling at procedure end. We conclude that if a single endoscope is used per both oral and anal routes, routine colonic sampling is not feasible. Duodenal and rectal sampling is feasible, although needed equipment and clinical expertise may limit implementation to centers with requisite capacity.
The standard MITS procedure uses transcutaneous needle biopsies, but to our knowledge, this method had not been applied intestinally. We included transabdominal small and large bowel sampling for the last 14 cases with near universal success. Rapid needle firing likely facilitated intestinal tissue acquisition. Although likely more scalable than endoscopy, location within the upper intestine cannot be determined by blind biopsy nor can mucosa be grossly visualized. Preliminary data from an ongoing study suggest this approach unreliably yields samples in adults (unpublished). Moreover, preservation tended to be better in endoscopically sampled tissues than transabdominally. Therefore, endoscopic approach may be preferred, except in settings where precluded by resource strain.
Sampling utility is an important consideration even if feasible and scalable. Intestine autolyzes early. Microbial colonization is less abundant in the duodenum but likely an important driver of large bowel autolysis. Digestive enzymes may be more involved in the duodenum, particularly D2 where pancreatic secretions drain. Although upper and lower intestine autolysis did not differ in this study, we noted a tendency toward tissue preservation with exposure to higher dysbiotic potential antibiotics (limited to small bowel). Less autolysis in children with SAM might be attributed to microbiome alterations or reduced pancreatic enzymatic activity associated with this condition. Less duodenal autolysis was found in the presence of stomach contents, perhaps because of protection from or competition for enzymatic degradation. Because sepsis was diagnosed by admitting clinician, its relationship with autolysis likely represents a relationship between illness severity at admission and autolysis rather than sepsis specifically.
Greater small bowel autolysis was seen with longer PMI, corroborating prior studies suggesting tissues become histologically uninformative beyond 12-24 hours. Our data suggest that sampling is best performed within 6 hours postmortem as no tissues had ≥50% autolysis before this interval (Supplementary . However, even samples obtained between 6 and 12 hours often yielded informative histology. Post-12 hours, no tissues had <50% autolysis.
Despite autolysis prevalence, many histologic features were discernible. As expected, epithelial structure was generally undiscernible because luminal surfaces degrade first. However, inflammatory response was scorable in 94% of slides. Duodenal chronic inflammation was prevalent. All tissues were devoid of acute inflammation. This is consistent with findings from biopsies of Zambian and Pakistani children with EED, as was our finding of reduced Paneth and goblet cell density which are important in antimicrobial activity and mucin production, respectively. Depletion may be from rapid cell turnover or impaired stem cell differentiation. EED histopathology investigations have focused on small bowel. Lower intestinal scrutiny has been limited. Rectal histology from this study demonstrated an absence of acute inflammation; however, chronic inflammation was identified in 96% of samples and severe infiltration seen in children as young as 5.5 weeks old. Although Chacko et al's seminal autopsy investigation lacked inflammation assessment, they demonstrated generally normal fetal and neonatal villus architecture, whereas blunting was noted as early as age 8 weeks and severity increased with age. Although our data are insufficient to explore further, factors including exposure to breastmilk substitutes or contaminated fluids could explain these findings. We hope to explore potential pathogen explanations in future analyses.
Transabdominally obtained small intestine had fewer histologic abnormalities than endoscopically obtained. Histopathology is a presampling phenomenon, so sampling method should not affect this. EED is thought to be a patchy duodenal disorder, as supported by our findings of intraindividual variation, possibly explained by sampling method difference. Because transabdominal sampling is blind, we cannot determine upper intestinal anatomic location and we may be scoring regions less affected than duodenum.
We hoped to explore relationships with histopathology to further elucidate the role of EED and intestinal pathology in child mortality. Though EED is considered an underlying cause of undernutrition, our study was unable to detect such relationship, likely from lack of statistical power. Further, we did not identify relationships between histopathology and demographics, clinical interventions, or admission diagnoses. Pending final COD determination based on all MITS samples could enhance this exploration.
We acknowledge study limitations. Sample size was constrained because MiM was designed to assess feasibility; therefore, our assessment of relationships with autolysis and intestinal histopathology are exploratory. Additionally, participants died while at a referral hospital, where the majority of patients are from lower socioeconomic backgrounds and reside near Blantyre. Results may not generalize to community-based deaths or other populations.
Despite these limitations, our study demonstrates that minimally invasive postmortem intestinal sampling is feasible and histologically informative. Biopsy ascertainment within 12 hours of death appears to improve tissue preservation. Preservation may also be affected by patient characteristics (eg, nutritional status, clinical interventions). Histologic interrogation successfully identified and quantified features of EED and rectal chronic inflammation. Future studies with larger sample sizes should elucidate factors affecting autolysis and enteric disease and its role in child mortality and undernutrition.
## Supplementary data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
## Notes
|
Blood-Based Gene Expression Profiles Models for Classification of Subsyndromal Symptomatic Depression and Major Depressive Disorder
Subsyndromal symptomatic depression (SSD) is a subtype of subthreshold depressive and also lead to significant psychosocial functional impairment as same as major depressive disorder (MDD). Several studies have suggested that SSD is a transitory phenomena in the depression spectrum and is thus considered a subtype of depression. However, the pathophysioloy of depression remain largely obscure and studies on SSD are limited. The present study compared the expression profile and made the classification with the leukocytes by using whole-genome cRNA microarrays among drugfree first-episode subjects with SSD, MDD, and matched controls (8 subjects in each group). Support vector machines (SVMs) were utilized for training and testing on candidate signature expression profiles from signature selection step. Firstly, we identified 63 differentially expressed SSD signatures in contrast to control (P, = 5.0E-4) and 30 differentially expressed MDD signatures in contrast to control, respectively. Then, 123 gene signatures were identified with significantly differential expression level between SSD and MDD. Secondly, in order to conduct priority selection for biomarkers for SSD and MDD together, we selected top gene signatures from each group of pair-wise comparison results, and merged the signatures together to generate better profiles used for clearly classify SSD and MDD sets in the same time. In details, we tried different combination of signatures from the three pair-wise compartmental results and finally determined 48 gene expression signatures with 100% accuracy. Our finding suggested that SSD and MDD did not exhibit the same expressed genome signature with peripheral blood leukocyte, and blood cell-derived RNA of these 48 gene models may have significant value for performing diagnostic functions and classifying SSD, MDD, and healthy controls.
# Introduction
Depression affects about 10% of the population at some point in their life and is the leading cause of disability across the world [bib_ref] Mood Disorders:Epidemiology, Blazer [/bib_ref]. Lacking specific objective findings, depression is often missed or undiagnosed [bib_ref] The underrecognition and undertreatment of depression: what is the breadth and depth..., Davidson [/bib_ref] and studies have focused on subthreshold depressive [bib_ref] A clinical significance criterion is essential for diagnosing subthreshold depression, Baumeister [/bib_ref] [bib_ref] From depressive symptoms to depressive disorders: the relevance of thresholds, Ayuso-Mateos [/bib_ref] [bib_ref] Subthreshold depression based on functional impairment better defined by symptom severity than..., Karsten [/bib_ref]. At present, some types of subthreshold depressive, including dysthymia, minor depression (MinD) and recurrent brief depression (RBD), are described in the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV). However, approximately two-thirds to three-fourths of all subthreshold depressive patients with psychosocial functional impairment did not meet any criteria of DSM-IV [bib_ref] Subsyndromal symptomatic depression: a new mood disorder?, Judd [/bib_ref]. Consequently, the concept of subsyndromal symptomatic depression (SSD) was introduced by Judd in 1994, which is characterized by two or more depressive symptoms, but without depressed mood or anhedonia, lasting for at least 2 weeks accompanied with social dysfunction, and does not meet the criteria for MDD, dysthymia, MinD or RBD [bib_ref] Subsyndromal symptomatic depression: a new mood disorder?, Judd [/bib_ref] [bib_ref] The role and clinical significance of subsyndromal depressive symptoms (SSD) in unipolar..., Judd [/bib_ref]. Convergent evidence has identified that SSD is a common depressive status that affects different ethnic populations [bib_ref] Subsyndromal symptomatic depression: a new mood disorder?, Judd [/bib_ref] [bib_ref] Subsyndromal depression in old age: clinical significance and impact in a multi-ethnic..., Chuan [/bib_ref] [bib_ref] A three-year follow-up of major depression, dysthymia, minor depression and subsyndromal depression:..., Forsell [/bib_ref] [bib_ref] Subsyndromal depression: prevalence, use of health services and quality of life in..., Goldney [/bib_ref] and to which we must pay more attention. However, litter research has been conducted on the biological basis of SSD.
Although the pathophysioloy of depression spectrum remain largely obscure, it has been reported that patients with SSD and MDD have similar family history, and their first-degree relatives have a high risk of comorbidity of depression and alcohol dependence, which implies that these two disorders could share genetic bases [bib_ref] Subthreshold depression and depressive disorder.clinical characteristics of general medical and mental health..., Sherbourne [/bib_ref]. Furthermore, several follow-up studies have suggested that SSD is a transitory phenomena in the depression spectrum and is thus considered a subtype of depression [bib_ref] A three-year follow-up of major depression, dysthymia, minor depression and subsyndromal depression:..., Forsell [/bib_ref] [bib_ref] A prospective 12-year study of subsyndromal and syndromal depressive symptoms in unipolar..., Judd [/bib_ref] [bib_ref] The relationship between major and subthreshold variants of unipolar depression, Maier [/bib_ref]. In addition, previous twin data supported that unipolar depression had a modest heritability [bib_ref] Genetic epidemiology of major depression: review and meta-analysis, Sullivan [/bib_ref]. SSD and MDD, which have different depressive symptoms, may be different subtypes of depression and have different phenotype at gene expression levels.
With the sequence of the human genome being publicly available since February 2001, an array of novel research tools, such as gene expression microarray, have become available that may yield unbiased, hypothesis-free insight into the pathophysiologic underpinnings of this disorder [bib_ref] Gene expression studies in major depression, Mehta [/bib_ref]. The application of high-throughput gene expression profiling to MDD in humans has mostly been restricted to postmortem brain tissue, typically sampled many decades after the critical time frame during which the initial molecular processes underlying the onset and development of disease have occurred, with methodological challenges including decades of cumulative drug exposure and postmortem artifacts [bib_ref] Methodological considerations for gene expression profiling of human brain, Atz [/bib_ref] [bib_ref] Effect of RNA quality on transcript intensity levels in microarray analysis of..., Popova [/bib_ref] [bib_ref] Altered expression of genes involved in inflammation and apoptosis in frontal cortex..., Shelton [/bib_ref] [bib_ref] Gene expression profiling in postmortem prefrontal cortex of major depressive disorder, Kang [/bib_ref] [bib_ref] Dysregulation of the fibroblast growth factor system in major depression, Evans [/bib_ref]. Convincing evidences indicated that depression affects the entire organ systems, including endocrinological, immunological and autonomic nervous systems, through the interaction between the brain and the body [bib_ref] Mood Disorders:Neurobiology, Thase [/bib_ref]. Circulating blood comprises a highly complex system that communicates with every tissue and organ in the body. Peripheral blood cells share more than 80% of the transcriptome with nine tissues: brain, colon, heart, kidney, liver, lung, prostate, spleen, and stomach, and the expression levels of many classes of biological processes have been shown to be comparable between whole blood and prefrontal cortex [bib_ref] The peripheral blood transcriptome dynamically reflects system wide biology: a potential diagnostic..., Liew [/bib_ref] [bib_ref] Evaluating the comparability of gene expression in blood and brain, Sullivan [/bib_ref]. Indeed there is considerable communication between the immune system and the central nervous system (CNS). Many cytokine receptors have been located within the CNS, and interleukin-2 mRNA and T-cell receptors have been specifically detected in neurons [bib_ref] Interleukin-2 enhances the viability of primary cultured rat neocortical neurons, Shimojo [/bib_ref]. Lymphocytes also express several neurotransmitter and hormone receptors, including dopamine, cholinergic, and serotonergic receptors and glucocorticoid and mineralocorticoid receptors and their chaperones [bib_ref] Lymphocytes as a neural probe: potential for studying psychiatric disorders, Gladkevich [/bib_ref]. Lymphocytes are directly influenced by glucocorticoids and catecholamines, and these two systems are perturbed in MDD [bib_ref] Glucocorticoid receptors in major depression: relevance to pathophysiology and treatment, Pariante [/bib_ref]. The circulating blood may act as a ''sentinel tissue'' that can reflect states of health or disease within the body. Some studies successfully discriminated between control subjects and physical disease patients via detection of the expression of ''tissue-specific'' genes in circulating blood [bib_ref] Gene profiling identifies secreted protein transcripts from peripheral blood cells in coronary..., Ma [/bib_ref] [bib_ref] Gene microarray analysis of peripheral blood cells in pulmonary arterial hypertension, Bull [/bib_ref] [bib_ref] Signatures of environmental exposures using peripheral leukocyte gene expression: Tobacco smoke, Lampe [/bib_ref]. Blood based gene expression diagnostics could be applied to the study of psychiatric disorders for which human brain tissue biopsy samples are unavailable.
Disease development is a systematic and dynamic processes influenced by environment factors and genetic factors, together. Computational and systems biology have greatly facilitate the disease studies from transcriptomes by using microarray technology [bib_ref] DNA chips: An array of possibilities, Marshall [/bib_ref]. Based on gene expression profiles, thousands of genes can be featured simultaneously in different conditions or clinical phenotypes [bib_ref] Multiplexed biochemical assays with biological chips, Fodor [/bib_ref]. Scientists have utilized high-throughput technology and computational approach to built disease models and classify disease state.
With gene features made on microarray accumulated by technology developing, many psychiatric disorder studies are also applied by the high-throughput technology with bioinformatics analysis. Tsuang et.,al have assessed the validity of blood-based gene expression profiles for the classification of schizophrenia and bipolar disorder [bib_ref] Assessing the validity of blood-based gene expression profiles for the classification of..., Tsuang [/bib_ref]. Segman et.,al found gene expression signatures that could differentiate between women prone to postpartum depression [bib_ref] Blood mononuclear cell gene expression signature of postpartum depression, Segman [/bib_ref]. Le-Niculuscu et.,al demonstrated that peripheral blood gene expression profiles could offer an unexpectedly informative insight into brain function and disease state [bib_ref] Identifying blood biomarkers for mood disorders using convergent functional genomics, Le-Niculescu [/bib_ref]. Most recently, Spijker et.,al also found that gene expression profiles could be used as a blood marker of MDD, and careful independent validation has been carried out to prove their results [bib_ref] Stimulated gene expression profiles as a blood marker of major depressive disorder, Spijker [/bib_ref].
Thus, in order to develop the potential peripheral blood lymphocytes gene expression signature models which can classify MDD, SSD, and healthy controls, whole-genome cRNA microarray analysis of lymphocytes were performed in this study.
# Results
## Pathway analysis and go analysis results for ssd gene expression signatures
For SSD gene expression signatures, we detected 1,456 differential expressed genes between SSD and healthy controls, in which 753 genes are up regulated and 703 genes are down regulated (adjusted p,0.01), which enriched in 47 pathways (P,0.01). Most of genes involved in several functional related to signaling pathways, including neuroactive ligand receptor interaction, JAK and STAT signaling pathway, G protein signaling, calcium signaling pathway, insulin signaling pathway, GNRH signaling pathway, Wnt signaling pathway and MAPK signaling pathway etc. Cellular communication and cell structure organization were also important in SSD process, such as apoptosis, cell adhesion molecules, tight junction, focal adhesion. The DEG also act in several biosynthesis and metabolism pathways, like oxidative phsphorylation, metabolism of xenobiotic by cytochrome P450, purine metabolism, glycerlipid metabolism, glycan structures biosynthesis, glycerolipid metabolism, starch and sucrose metabolism. We also found that SSD signatures participate in immunity process, antigen processing, leukocyte transendothelial migration, natural killer cell mediated cytototoxicity and cytokine-cytokine receptor interaction [fig_ref] Figure 1: Functional annotation of the DEGs in SSD and MDD [/fig_ref]. GO analysis indicate that SSD gene signatures correlate with cerebellar cortex morphogenesis, cerebellar granular layer development, hydrolase activity, GTPase and ATPase activity, S phase and M phase of mitotic cell cycle and tissue regeneration, etc. [fig_ref] Figure 1: Functional annotation of the DEGs in SSD and MDD [/fig_ref].
## Pathway analysis and go analysis results for mdd gene expression signatures
Based on pre-processed microarray profile, we identified 149 differential genes between MDD patients and controls with 95 upregulated and 54 down-regulated (adjusted P,0.01), 20 of which were identified between SSD and control. These differential genes enriched in 53 pathways, 2 of which also were identified in SSD. Signaling pathways active in MDD include activation of ATR in response to replication stress, NRIF signals cell death from the nucleus, fas signaling pathway, p53-Independent G1/S DNA damage checkpoint and Nicotinamide salvaging, and EGF signaling pathway etc. We noticed that many MDD signatures involves several immunity process, such as T cell receptor signaling pathway and JNK signaling in the CD4+ TCR pathway, IL2mediated signaling events, IL1 signaling and IL6-mediated signaling events and Calcium signaling in the CD4+ TCR pathway and TCR signaling in CD4+ T cells. In MDD subjects, more biosynthesis and metabolism pathway are identified, involveing Vitamin B5 (pantothenate) metabolism, coenzyme A biosynthesis and metabolism of water-soluble vitamins and cofactors. Comparing with former results of SSD signatures, we found that several pathways were shared in MDD and SSD process, including cell cycle controls and Cell Cycle Checkpoints, like G2/M Checkpoints and Wnt signaling. We noticed that MDD-specific functions or pathways compared with SSD were activation of ATR pathway in response to replication stress, NRIF signals cell death from the nucleus, fas signaling pathway, immunity pathway about IL2 signaling events mediated by PI3K and IL1 singaling events. Meanwhile, SSD-specific pathways contain cytokine-cytokine receptor interaction, GPCRDB class A rhodopsin like, MAPK signaling pathway, neuroactive ligand receptor interaction, calcium signaling pathway, breast cancer estrogen signaling pathway, purine metabolism, insulin signaling pathway, cell adhesion molecules and Toll like receptor signaling pathway [fig_ref] Figure 1: Functional annotation of the DEGs in SSD and MDD [/fig_ref]. Alternatively, GO analysis for MDD signatures, 29 of which also were identified in SSD, convinced us that most significant functions (P,0.01) are active in immunity reactions involving pro-B cell differentiation, negative regulation of antigen processing, positive regulation of leukocyte migration and plasminogen activation. Other functions of these genes involve engulfment of apoptotic cell, fibrinogen binding, lymphoid progenitor cell differentiation and immunoglobulin V(D)J recombination and mitotic cell cycle controls (S phase), DNA ligation involved in DNA repair and somatic cell DNA recombination, etc. al. [fig_ref] Figure 1: Functional annotation of the DEGs in SSD and MDD [/fig_ref].
## Gene expression profiles for classification of subsyndromal symptomatic depression and major depressive disorder
In order to filter out most of false positives and select most potential biomarkers, we applied strict threshold on the same pairwise comparisons among SSD, MDD and controls, and identified 63 differentially expressed SSD signatures in contrast to controls (adjusted P, = 1.0E-4) and 30 differentially expressed MDD signatures in contrast to controls (adjusted P, = 5.0E-4), respectively. Then, 123 gene signatures were identified with significantly differential expression level between SSD and MDD (adjusted P, = 1.0E-4). Unsupervised hierarchal clustering analysis by using Euclidean distance and complete linkage clustering method was conducted on three more potential groups of DEGs. The results showed clearly that genes differentially expressed in the peripheral blood lymphocytes were capable of differentiating MDD group, SSD group and healthy controls, separately [fig_ref] Figure 2: Biomarkers differentiation efficiency among MDD group, SDD group and HC [/fig_ref].
To evaluate the predictive performances of SSD and MDD signatures respectively, we utilized SVM with Linear Kernel to build disorder models. For 63 SSD signatures (adjusted P, = 1.0E-4), total instances (8 SSD and 8 control instances, respectively) were correctly classified. Similarly, for 30 MDD signatures (adjusted P, = 5.0E-4), and 123 DEGs signatures (adjusted P, = 1.0E-4) can also clearly classified 8 MDD and 8 SSD instances.
For SSD gene signatures, we detected 1,456 differential expressed genes between SSD and control; For MDD gene signatures, we identified 149 differential genes between MDD patients and controls d. Among these genes, there are only 20 different genes between SSD and MDD.
Furthermore, in order to conduct priority selection for biomarkers for SSD and MDD together, we selected top gene signatures from each group of pair-wise comparison results, and merged the signatures together to generate better profiles used for clearly classify SSD and MDD states in the same time. In details, we tried different combination of signatures from top ranked signatures in the three pair-wise compartmental results and finally determined 48 gene expression signatures [fig_ref] Table 1: Mostly potential signatures of gene expression profiles models for classification of MDD... [/fig_ref]. To maintain the robustness of SSD-MDD disorder model, the predictive power was evaluated using cross validation, which randomly took 9/10 samples used for training and remaining 1/10 as internal testing validation.
When 54 probesets (48 genes) were chosen as biomarkers, we obtained the best predictive performances with 100% accuracy and 100% TPR (leave-one-off validation). Leave-one-off validation refers to that we used n-1 sample to train model and used another sample to test the model. Total 24 MDD, SSD and control samples were separated into train and test profiles in 24 times and train/test the data using leave-one-off method. Finally, we collected the predictions for each sample, and obtained 100% predictive performance [fig_ref] Table 2: Predictive performances of disorder model [/fig_ref].
Then, we interrogated the pathways and biological functions about these mostly potential biomarkers for MDD and SSD, together. Pathway analysis demonstrates that the potential 48 gene biomarkers involved in insulin signaling pathway, signaling by NGF, ErbB signaling pathway, neurotrophin signaling pathway, cell surface interactions at the vascular wall, NRAGE signals death through JNK, Rho GTPase cycle, and G alpha signaling pathway (P value,0.05) [fig_ref] Table 3: Mostly potential pathways of disorder model biomarker [/fig_ref]. Also, GO analysis shows consistency with pathway analysis results. Besides, PURA and TERF2 both function about telomeric DNA binding and single strand DNA binding, and DNA replication. SLC16A3 and CTNS act in the directed movement of carboxylic acids into, out of, within or between cells. FGD3 and KALRN participate in Stimulates the exchange of guanyl nucleotides by a GTPase. Under normal cellular physiological conditions, the concentration of GTP is higher than that of GDP, favoring the replacement of GDP by GTP in association with the GTPase. Also, FGD3, KALRN and RHOQ involve in Rho GTPase cycle. Other signatures also correlated with Cell death signalling via NRAGE, NRIF and NADE, Jak-STAT signaling pathway, B cell receptor signaling pathway, and p75 NTR receptor-mediated signaling [fig_ref] Table 4: Mostly potential GO functions of disorder model biomarker [/fig_ref].
## De novo cis-regulatory element analysis results for candidate biomarkers of mdd and ssd
In order to investigate how the signatures for classifying three groups are regulated, we analyzed the cis-regulaotry elements co-occurring on the promoters of these genes. In details, STAT1 and STAT2 factor' binding motifs were detected on five MDD signatures' promoters (e.g. BDNF, MYB, THBS1, SORBS1, and SH3BGRL). In addition, we identified SRF binding motif on three gene promoters (e.g. THBS1, EGR1 and PODN). For SSD signatures, we identified transcriptional factor SREBP1 was correlated with eleven SSD signature genes (GNAS, MLL5, TOM1L1, DLGAP4, PTMA, NF1, ATP2B2, UNC13D, PDP2, CORO1A, and INPP4A). Most of these transcriptional factors are related with depression disorders as discussed below.
# Discussion
To our knowledge, this is the first study to compare the expression profile and make the classification with the leukocytes by using whole-genome cRNA microarrays among patients with SSD, major depressive disorder (MDD) and controls. We found that SSD and MDD had different blood-based gene expression signature, and the differential expressed genes of SSD were about 10 times of MDD, but there are only 20 overlapping differential expressed genes between SSD and MDD. Pathway analysis for SSD gene signatures showed that differential expressed genes enriched in 47 pathways, and most pathways were involved in regulation of DNA replication, IL2 signaling events mediated by STAT5, and Wnt signaling pathway, etc. For MDD gene signatures, the results of pathway analysis suggested that differential expressed genes enriched in 53 pathways, 2 of which also were identified in SSD, including MAPK signaling pathway and Wnt signaling pathway. Although the relationship between SSD and MDD is unclear, previous follow-up studies have showed that SSD was a subtype of depression and a transitory phenomenon in depression spectrum with a high likelihood of transition to MDD [bib_ref] A three-year follow-up of major depression, dysthymia, minor depression and subsyndromal depression:..., Forsell [/bib_ref] [bib_ref] Subthreshold depression and depressive disorder.clinical characteristics of general medical and mental health..., Sherbourne [/bib_ref] [bib_ref] A prospective 12-year study of subsyndromal and syndromal depressive symptoms in unipolar..., Judd [/bib_ref]. It indicates that the genes involving in these two pathways maybe point to pathogenetically relevant underlying molecular processes of depression. Genetic manipulation of the MAPK pathway, one of the neurotrophin signaling pathways, has received much attention, which postulated that the dysfunction of this pathway played a key role in the pathophysiology of mood disorders [bib_ref] Postmortem studies in mood disorders indicate altered numbers of neurons and glial..., Rajkowska [/bib_ref] , especially in depression-like behavior [bib_ref] Brainderived neurotrophic factor produces antidepressant effects in behavioral models of depression, Shirayama [/bib_ref]. Previous data also have shown that acute systemic blockade of MAPK signaling contributes to a depressive-like phenotype and blocks actions of antidepressants in animal models of depression [bib_ref] A role for MAP kinase signaling in behavioral models depression and antidepressant..., Duman [/bib_ref].
Wnt signaling pathways have been implicated in various physiological functions, such as cell fate determination, cell and tissue polarity, synaptogenesis, dendritic morphogenesis, and axon remodeling. Moreover, abnormal Wnt signaling has been implicated in mood disorder. The relationship between Wnt signaling pathway genes and mood disorders has been reported in several genetic association studies. A study showed that alteration of hippocampal microRNA levels following chronic treatment with mood stabilizers is caused by effectors in the canonical Wnt signaling pathway. Gene expression-profiling of hippocampal subfields has also revealed altered expression of several genes related to Wnt signaling in bipolar disorder patients. Another study supports that the canonical Wnt signaling pathway and related substrates play a role in MDD. Wnt signaling pathway also has been considered relevant to the antidepressant effects, and Wnt2 expression and signaling is a common target of antidepressants and that increased Wnt2 is sufficient to produce antidepressant effects.
Moreover, patients with MDD have depressed mood or anhedonia but SSD have not, so differential expression of genes involving in other 51 pathways in MDD may be correlate with the underlying pathological mechanism of the symptom of depressed mood or anhedonia. Our unsupervised hierarchal clustering analysis showed obviously that each disease state exhibited a unique expressed genome signature except the genes involving in MAPK and Wnt pathways, which suggesting that these two diseases may be two different phenotypes in depression spectrum by respective gene signatures. Furthermore, genes differential expression among SSD group, MDD group, and healthy controls allowed us to discriminate among these three groups. It suggested that blood-derived RNA may potentially be used as a diagnostic tool for SSD and/or MDD, as long as the correct subsets of genes are employed. Blood profiling may also allow identification of differentially expressed genes that are involved in the pathophysiology of these disorders. To select the most potential biomarkers for differentiating these three groups, we combined top differential expressed genes from each set of gene expression signatures, then trained and tested the multiple combinatorial gene signatures from pair-wise comparison groups by using support vector machine classifier. Finally 48 gene expression signatures were determined. Samples can be grouped together according to the similarity of the expression levels of these 48 genes which suggested that different levels of gene expression may reflect different disease states. Among differential genes, BDNF, COX5B, GRK6 are the most significantly differential genes.
We comprehensively analyzed gene functions and pathway for the candidate biomarkers of SSD and MDD and found that potential biomarkers act in some pathways which have been found associated with function of CNS and implicated in depression, including insulin signaling pathway, signaling by NGF, ErbB signaling pathway and neurotrophin signaling pathway. We also found most of them were not reported the relationship with depression, such as cell surface interactions at the vascular wall, NRAGE signals death through JNK, Rho GTPase cycle, and G alpha signaling pathway, etc. al.
Some studies showed that there was a positive association between depressive disorder and insulin resistance due to dysregulation of insulin secretion or insulin receptor signaling. Otherwise, various functions for insulin receptor signaling in the brain have been suggested in normal neurophysiology, such as insulin receptor signaling maybe play a important role in synaptic plasticity and cognitive function,and several lines of work in both laboratory animals and humans suggest that when neurons in cognitive brain regions such as the hippocampus and cerebral cortex do not make enough insulin or cannot respond to insulin properly, everything from very mild memory loss to severe neorodegenerative diseases can result. Dysregulation of insulin secretion or insulin receptor signaling has also been reported in serious mental illnesses, such as Alzheimer's disease. Patients with depression also have some cognitive function problems and maybe have differential expression of genes involving in insulin signaling pathway.
It has been suggested that neuronal atrophy or destruction in the hippocampus and cortex is involved in the pathogenesis of depression. The neurotrophin systems modulate neuronal plasticity, inhibit cell death cascades and increase cell survival proteins that are responsible for proliferation and maintenance of central nervous system neurons. Thus the dysregulation of the neurotro- phin systems, such as differential expression of genes involving in signaling by NGF (nerve growth factorA) and neurotrophin signaling pathway, may be involved in the pathophysiology of depression.
Transgenic mouse experiments have confirmed that the block of erbB signaling pathway will result in the change of OL number and morphology, reducing the thickness of myelin and the transmission rate of CNS axons [bib_ref] Loss of erbB signaling in oligodendrocytes alters myelin and dopaminergic function, a..., Roy [/bib_ref]. The abnormal expression of ERBB (epidermal growth factor receptor, EGFR, epidermal growth factor receptor) signaling pathway can lead to oligodendrocytes (OL) abnormalities, which results in dopaminergic dysfunction, and it may be associated with depression [bib_ref] Oligodendroglial abnormalities in schizophrenia, mood disorders and substance abuse. Comorbidity, shared traits,..., Sokolov [/bib_ref] [bib_ref] Subcortical oligodendrocyte-and astrocyte-associated gene expression in subjects with schizophrenia, major depression and..., Barley [/bib_ref].
The results of analysis of the cis-regulaotry elements cooccurring on the promoters of these genes showed that STAT1 and STAT2 factors were detected on five MDD signatures' promoters (e.g. BDNF, MYB, THBS1, SORBS1, and SH3BGRL). Especially, STAT1 mediates the autoimmune and inflammatory functions, and STAT2 mediates the virus protection function. From previous investigation about the immune cell specificity of activation programs induced by a major component of cell-mediated immunity, the transcriptional activators STAT1 were significantly induced in CD4+ and CD8+ T cells, B cells and monocytes [bib_ref] ) pSTAT1, pSTAT3, and T-bet as markers of disease activity in chronic..., Madia [/bib_ref]. Depression phenotypes are also correlated with immunity reactions reflected from blood transcriptomes [bib_ref] Stress, depression, the immune system, and cancer, Reiche [/bib_ref]. In addition, we identified SRF binding motif on three gene promoters (e.g. THBS1, EGR1 and PODN). Up to now, there was no study about the relationship between SRF binding motif and depression.
For SSD signatures, we identified transcriptional factor SREBF1 was correlated with eleven SSD signature genes (GNAS, MLL5, TOM1L1, DLGAP4, PTMA, NF1, ATP2B2, UNC13D, PDP2, CORO1A, and INPP4A). Several studies have reported the importance of SREBF1 and SREBF2 factors in the lipid biosynthesis and their possible involvement in antipsychotic drug effects and the genetic variants of SREBF1 and/or SREBF2 could affect schizophrenia susceptibility [bib_ref] Stress, depression, the immune system, and cancer, Reiche [/bib_ref] [bib_ref] Polymorphisms in SREBF1 and SREBF2, two antipsychotic-activated transcription factors controlling cellular lipogenesis,..., Hellard [/bib_ref]. HapMap-based association study in a large German sample identified association between schizophrenia and five markers in SREBF1 and five markers in SREBF266. Additionally, scientists have demonstrated in glial cell lines that antipsychotic drugs induce the expression of genes involved in cholesterol and fatty acids biosynthesis through activation of the sterol regulatory element binding protein (SREBP) transcription factors, encoded by the sterol regulatory element binding transcription factor 1 (SREBF1) and sterol regulatory element binding transcription factor 2 (SREBF2) genes [bib_ref] Polymorphisms in SREBF1 and SREBF2, two antipsychotic-activated transcription factors controlling cellular lipogenesis,..., Hellard [/bib_ref].
The results presented were limited by a modest sample size and required more samples to replicate. Quantitative reverse transcription-polymerase chain reaction were required to exam the expression levels of 48 genes, which were found differentially expressed in our pilot study, in a larger sample of SSD and MDD. Additional studies were required to further explore the roles of these 48 genes in pathophysiology of SSD and MDD.
In conclusion, our study demonstrated that SSD and MDD exhibited a unique expressed genome signature with peripheral blood leukocyte, and blood cell-derived RNA may have significant value for performing diagnostic functions and identifying disease biomarkers in SSD and MDD.
# Materials and methods
The study was conducted at the Division of Mood Disorders, Shanghai Mental Health Center, Shanghai Jiao Tong University School of Medicine between Jan 2007 and Dec 2009. Outpatients were recruited from the clinic and ward of Shanghai Mental Health Center. All procedures were reviewed and approved by Institutional Review Boards of Shanghai Mental Health Center. Written informed consent was obtained from each subject before any study-related procedures were performed.
## Subjects
Inclusion criteria for SSD group were: two or more depressive symptoms for at least 2 weeks with social dysfunction but without depressed mood or anhedonia, and having a total score of 17-item Hamilton Rating Scale for Depression (HRSD-17) from 8 to 16. Patients were included into MDD group who met DSM-IV criteria for MDD and had the total score of HRSD-17 $17. Patients were excluded if they had substance dependence, severe medical illness, organic brain disease, pregnancy. Healthy control subjects have a score 7 or lower on the HRSD-17, and did not have any major Axis I disorders (including substance dependence, psychotic disorders, mood disorders and anxiety disorders), family history of mental disorder or severe physical diseases (hypertension, diabetes, cancer).
For the gene expression microarray analysis, this study enrolled eight drug-free Chinese Han patients with their first episode of subsyndromal symptomatic depression, eight previously untreated patients presenting with their first episode of major depression disorder, and eight healthy controls. All groups were matched with sex and age (shown in [fig_ref] Table 5: Demographic data for patients and healthy controls [/fig_ref].
All subjects were screened by the Structured Clinical Interview for DSM-IV (SCID) and assessed through HRSD-17 score by two experienced psychiatrists (inner coherence, Kappa = 0.87).
## Peripheral blood lymphocytes collection and rna processing
Total 20 ml venous peripheral blood from fasting patients and healthy controls were collected during 7am to 9am. Peripheral blood lymphocytes were separated by Ficoll gradient centrifugation using Ficoll-PlaqueTM Plus (GE, Sweden) [bib_ref] Brain-Derived Neurotrophic Factor Gene Expression in Pediatric Bipolar Disorder: Effects of Treatment..., Pandey [/bib_ref].Total RNA was extracted from lymphocytes using Trizol reagent (Invitrogen) according to the manufacturer's protocol. RNA quality was determined by Nanodrop ND-1000 (Nanodrop Technologies, Wilmington, DE) and degradation of mRNA was assessed by denaturing agarose gel electrophoresis and evaluated the sharpness of 28 S and 18 S rRNA bands.
Microarray data pre-processing 24 samples were profiled on affymetrix U133 Plus2.0 GeneChip oligonucleotide arrays (Affymetrix, Santa Clara, CA), which is comprised of more than 22,000 probe sets and can analyze the expression level of 18,400 transcripts and variants (approximately 11,000 genes). The preparation of cRNA hybridization, signal scanning, data acquisition, and preliminary analysis were performed at the National Engineering Center for Biochip at Shanghai according to the standard protocols recommended by Affymetrix (Affymetrix, Santa Clara, CA, USA). Raw data generated from affymetrix Human U133Plus2.0 were processed and normalized by RMA method with Gene Spring Software 11.0 (Agilent technologies, Santa Clara, CA, US), then the values were log2 transformed. Differential gene analysis was preliminarily performed using Welch t test and then P value adjustment under multiple hypothesis testing was implemented with multtest package in Bioconductor under the adjustment method of Bonferroni. We used Welch t test and boost strap resampling approach (B = 100,000) to compute t statistics and p values. The threshold for differential expressed genes (DEGs) was chosen as 0.01. All data is MIAME compliant and that the raw data has been deposited in a MIAME compliant database (E.g. ArrayExpress, GEO), as detailed on the MGED Society website http://www. mged.org/Workgroups/MIAME/miame.html. The accession numbers is GSE32280.
## Disease model and classification
To select the smallest size of biomarkers with robust predictive power and fewer potential false positives, more stringent thresholds were used to identify genes with even greater reliability. Firstly, the thresholds for differentially expressed SSD and MDD signatures compared to control and differentially expressed signatures between SSD and MDD were set as 1.0E-4, 5.0E-4 and 1.0E-4 respectively. Alternatively, P values in combination with foldchange values were used to identify potential biomarker genes to limit the likelihood of false positive results. Secondly, these signatures from 3 pair-wise comparisons were ranked according to their adjusted P values and the top N signatures were merged directly (to obtain a small size of biomarkers comparatively and a better classification performance, the top 10, 15, 18, 20, 25 and 30 signatures from each group were merged respectively). Then, we applied SVM (Support vector machines) on each of candidate expression profiles to search better combination of biomarkers with robust prediction performances (accuracy, sensitivity or specificity). Finally, leave-one-off method was used to validate the biomarkers. Leave-one-off validation involves using a single observation from the original sample as the validation data, and the remaining observations as the training data. This was repeated such that each observation in the sample was used once as the validation data.
# Gene ontology analysis
Standard methods for testing over-representation of a GO category assume that, under the null hypothesis, each gene has equal probability of being detected as DEG (differential expressed gene) [bib_ref] Variations in myelin and oligodendrocyte-related gene expression across multiple brain regions in..., Katsel [/bib_ref]. Under this assumption, the number of genes associated with a category that overlap with the set of DEG follows a hypergeometric distribution. Hence the GO test can be conducted using Fisher's exact test, which uses the hypergeometric distribution, or Pearson's chi-square test, which is a computationally convenient approximation.
## Network and pathway analysis
Pathway was analysis using human pathways from KEGG, biocarta, and metabolism pathway databases [bib_ref] Schizophrenia gene networks and pathways and their applications for novel candidate gene..., Sun [/bib_ref]. Scoring the prioritation of network/pathways according to the relevance to input data. In cases of SSD and MDD experiments result, we analysis how different pathways and networks modules can be prioritized based on their statistical significance with respect to such experimental datasets. Significance is evaluated based on the size of the intersection between differential expressed gene signatures and set of genes/proteins corresponding to a network module/pathway curated in pathway database. This problem can be cast as selection without replacement and the probability to randomly obtain intersection of certain size between differential expressed gene signatures and a network/pathway follows hypergeometric distribution When considering a set of DEG signatures (I), invariable number r of DE signatures among the N nodes of the pathway/network module. The probability of a subset of size n to include r DE genes provided that n and R are unrelated (null-hypothesis) follows the hypergeometric distribution.
## Multiclass svm implementation
In order to classify SSD and MDD from healthy control simultaneously, support vector machines (SVMs) was utilized for training and testing on candidate signature expression profiles from signature selection step. SVMs which represents an extension to nonlinear models of the generalized portrait algorithm developed by Vladimir Vapnik is a group of supervised learning methods that can be applied to classification or regression [bib_ref] Combining gene expression, demographic and clinical data in modeling disease: a case..., Struyf [/bib_ref]. The SVM takes a set of input data, and predicts, for each given input, which of two possible classes the input is a member of, which makes the SVM a non-probabilistic binary linear classifier.
Since an SVM is a classifier, then given a set of training examples, each marked as belonging to one of two categories, an SVM training algorithm builds a model that predicts whether a new example falls into one category or the other. The original SSD, MDD and control problem may be stated in a finite dimensional space, but it often happens that in that space the sets to be discriminated are not linearly separable. For this reason it was proposed that the original finite dimensional space be mapped into a much higher dimensional space presumably making the separation easier in that space. In order to clearly classify SSD and MDD from controls, multiclass SVM were also used in aims to assign labels to instances by using support vector machines. The multiclass approach for conducting this is to reduce the single multiclass problem into multiple binary classification problems. Each of the problems yields a binary classifier, which is assumed to produce an output function that gives relatively large values. In end, polynomial kernel was applied with the best predictive performances for combinatorial gene signatures from the three groups.
## De novo cis-regulatory element analysis
Cis-regulatory motifs are essential elements for gene transcription [bib_ref] Cloning of human neurotensin/neuromedin N genomic sequences and expression in the ventral..., Bean [/bib_ref]. We also interrogated the over-representative motifs on promoter sequences collected from UCSC (www.genome.ucsc. edu/). Two thousand bps sequences around TSS for SSD and MDD signatures and biomarkers for classifying three groups (MDD, SSD and controls, together) were all considered for in promoter-based de novo motif analysis.
[fig] Figure 1: Functional annotation of the DEGs in SSD and MDD. (A) and (C) Pathway analysis of SSD-associated and MDD-associated genes respectively. The y-axis shows the KEGG Pathway terms, and the x-axis shows the enrichment significance P-values for the top 10 enriched Pathway terms. (B) and (D) GO analysis of SSD-associated and MDD-associated genes respectively. The y-axis shows the GO terms, and the x-axis shows the enrichment significance P-values for the top 10 enriched GO terms. Term GO:0004719 remarks the function of protein-L-isoaspartate (D-aspartate) Omethyltransferase activity. MDD: Major depression disorder; SSD: Subsyndromal symptomatic depression. doi:10.1371/journal.pone.0031283.g001 [/fig]
[fig] Figure 2: Biomarkers differentiation efficiency among MDD group, SDD group and HC. (A) Complete linkage clustering analysis with 16 samples using 30 biomarkers under the criteria of adjusted.P, = 5E-4 between MDD and HC. (B) Complete linkage clustering analysis with 16 samples using 63 biomarkers under the criteria of adjusted.P, = 1E-4 between SSD and HC. (C) Complete linkage clustering analysis with 16 samples using 123 biomarkers under the criteria of adjusted.P, = 1E-4 between SSD and MDD. MDD: Major depression disorder; SSD: Subsyndromal symptomatic depression; HC: Healthy controls. doi:10.1371/journal.pone.0031283.g002 [/fig]
[table] Table 1: Mostly potential signatures of gene expression profiles models for classification of MDD and SSD. [/table]
[table] Table 2: Predictive performances of disorder model. [/table]
[table] Table 3: Mostly potential pathways of disorder model biomarker. [/table]
[table] Table 4: Mostly potential GO functions of disorder model biomarker. GC-rich DNA binding, or with a specific sequence motif or type of DNA e.g. promotor binding or rDNA binding. doi:10.1371/journal.pone.0031283.t004 [/table]
[table] Table 5: Demographic data for patients and healthy controls. MDD: Major depression disorder; SSD: Subsyndromal symptomatic depression; HC: Healthy controls. doi:10.1371/journal.pone.0031283.t005 [/table]
|
Multi-well capture zones in strip-shaped aquifers
Capture zone equations for a multi-well system in strip-shaped confined and unconfined aquifers with and without regional flow are presented. The aquifer is limited by two parallel boundaries that are either no flow (barrier) or inflow (variable head) so that aquifers with four possible boundary configurations are formed. The wellfield includes any number of extraction or injection wells or a combination of both types. The flow field in the strip-shaped aquifer was converted to its equivalent extensive aquifer using conformal mapping and image well methods. To delineate the capture envelope, the potential, streamline and stagnation point equations were derived using velocity potential theory. The solution permits rapid determination of the effect of number, position and extraction/injection rate of wells, boundary type and direction, and rate of regional flow on the size, shape and pattern of well capture zones. The derived equations are readily extended to water quality and quantity management simulations, as shown by embedding the equations within two optimization schemes, viz., Particle Swarm Optimization (PSO) and Genetic Algorithm (GA), to automatically determine the most efficient wellfield designs for pump-and-treat remediation, contaminant plume containment and pumping policy projects.
# Introduction
Demands on water resources are increasing due to climate-induced changes and population growth, emphasizing the important role of managing the quantity and quality of water resources. In pumped groundwater systems, knowledge of capture zones is a fundamental element of groundwater management.
The well capture zone within an aquifer is defined as the region from which water is withdrawn by one or more pumping wells. After pumping is initiated, the capture zone grows with time and reaches its maximum size at steady state, which defines the capture envelope. Capture zones are important for aquifer management, groundwater remediation projects, surface-groundwater interactions, well head protection, water rights, and in delineating impacts on transboundary aquifers. For over-exploited aquifers (where annual extraction exceeds recharge), identification of capture zones underpins optimal pumping plans to recover and sustain the depleted storage. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111
Capture zone determination for confined aquifers is long established in groundwater engineering. For instance, a two dimensional (2-D) pump-and-treat system was presented by Javandel and Tsangincluding type curves to determine the number and pumping rate of wells to contain a contamination plume. Three-dimensional (3-D) capture zones are available for both horizontal drains and vertical wells in homogeneous, anisotropic aquifers. studied the effects of anisotropy on capture zones for partially penetrating wells. Analytical and semi-analytical expressions for multiple wells and extraction rates and locations based on complex potential theory and superposition were presented by Christ and Goltz. Their solution was extended by Fienen et al.to the computation of stagnation point locations in a multi-well system. Lu et al.determined stagnation points in a flow field of a single pumping well located at the center of a recharge area.
The above-mentioned studies focused on simple aquifer geometries and did not account for different boundary configurations or unconfined flow. Efforts to extend analytical results include the case of a single well between two parallel streams, and multi-well systems in wedge-shapedand peninsula-shaped confined and unconfined aquifers. Capture zones for multi-well systems in rectangular bounded aquifers are also available, both for confined and unconfined aquifers. Such analytical solutions are used in optimization schemes. For example, Mantogloudeveloped analytical models of saltwater intrusion in coastal aquifers of finite size pumped by several wells that were embedded within an optimization algorithm to find the optimal pumping strategy.
The approach most commonly used to obtain analytical results is the image well methodin conjunction with complex discharge potential theory. For complex boundary conditions, the number of imaginary wells can be large, and the method can become unwieldy. In this situation, however, conformal mapping can be used, since the number of imaginary wells for individual real single well is then limited to three or less for any aquifer boundary configuration.
Conformal mapping is commonly employed in areas of physics and engineering where Laplace's equation applies. In groundwater flow problems, the modeling of seepage flow was solved by Fukuo and Kaihotsu, followed by other applications, e.g., Anderson. Fittsobtained an exact solution for two-dimensional flow to a well in an anisotropic domain. Chaharused conformal mapping to provide an analytical solution to the problem of seepage from a soil channel with a curvilinear bottom. Klammler et al.used the conformal mapping method to obtain the solution for flow fields near drain-and-gate reactive barriers. Using the same method, Klammler et al.also developed an approximate analytical solution to the permeable reactive barrier (PRB) capture and release behaviour considering doubly-symmetric funnel-and-gate as well as drain-and-gate configurations. Similarly, Lu et al.derived analytical solutions for pumping in a fully bounded rectangular aquifer. This approach was used by Strackto obtain analytical results for flow to a well between two parallel rivers or rivers that are at a fixed angle. These results are generalized here.
Our purpose is to obtain a general solution for the capture zone of a multi-well system in bounded strip-shaped aquifers with or without uniform regional flow. The well system includes any number of arbitrarily located extraction or injection wells. The strip-shaped aquifer is located between parallel inflow (variable head) boundaries (surface water bodies like streams, rivers, canals, lakes, and seas) and no-flow (barrier) boundaries (bedrock, fault blocks, glacial tills or impervious rocks). Conformal mapping is applied to determine the complex discharge potential and streamline equations for determining the capture envelope. The method permits efficient and rapid modeling of capture envelopes, as shown by examples where the effects of number, position and extraction rate of wells and regional flow direction and rate on the capture envelope are investigated. We demonstrate the versatility of the approach by its application to groundwater quantity and quality management problems. Specifically, the capture zone model is embedded within a Particle Swarm Optimization (PSO)and Genetic Algorithm (GA)schemes to optimize design of pump-and-treat remediation, contaminant plume containment and pumping projects.shows a schematic plan view of different aquifer (confined or unconfined) configurations. Each case shows an aquifer with a fully penetrating well bounded by two parallel boundaries of infinite extent forming a strip-shaped aquifer. The aquifer is isotropic and homogeneous with uniform thickness. Steady state, 2-D flow is considered. Inthe boundaries are fully penetrating streams (variable head boundary) having no hydraulic resistance with the aquifer. An extraction or injection well is located at (x w ,y w ). Note that the solution is for a multi-well system with arbitrary extraction/injection rates, and steady regional flow at an arbitrary angle (β) relative to positive x-axis. On physical grounds, the flow direction must be consistent with the hydraulic properties of the boundaries in that if the direction of regional flow is perpendicular to streams (β = π/2 or 3π/2), they are treated as constant head boundaries. In, one boundary is an inflow stream (variable head boundary) and another one is an impervious (no-flow) boundary. The aquifer depicted inis different from that ofto the position of wells relative to boundaries and vice versa. As a result, they should be treated as two separate well-aquifer systems. There are two no-flow boundaries inIn this case, on physical grounds, the regional flow direction must be parallel to the impermeable boundaries.
## Conceptual model
## Mathematical formulation
## Capture zone
Using conformal mapping, the conceptual model ofis first mapped onto the upper half-plane in the z 2 −plane using z 2 = exp(πz 1 /d) (see S1 File for details). When the boundaries are a combination of barrier-inflow and inflow-barrier, the z 1 −plane is transformed into the second auxiliary plane, the z 3 −plane, using the transformation z 3 ¼ z 2 1=2 . Transforming the zones inTransformation from the z 1 −plane to the z 3 −plane is given by:
[formula] z 3 ¼ exp pz 1 2d � �ð1Þ [/formula]
The complex potential for flow towards a single well at position z in the z 2 −plane with the regional uniform flow in an aquifer bounded by two parallel streams is:
[formula] O z 2 ð Þ ¼ � 0 À dq 0 p lnðz 2 ÞexpðÀ ibÞ þ Q w 2p ln z 2 À z z 2 À z !ð2Þ [/formula]
where, Q w is the pumping or injection rate (L 3 T -1 ), q 0 is the uniform flow rate in the direction β (radians) relative to the positive x axis, ϕ 0 is the arbitrary potential in the absence of extraction and regional flow. Writing Eq (2) in the z 3 −plane gives:
[formula] O z 3 ð Þ ¼ � 0 À 2dq 0 p lnðz 3 ÞexpðÀ ibÞ þ c Q w 2p ln ðz 3 À dÞðz 3 þ dÞ ðz 3 À dÞðz 3 þ dÞ [/formula]
The parameter c is +1 for extraction wells and -1 for injection wells, respectively. Generalizing Eq (3) for the four boundary configurations ofin:
[formula] O z 3 ð Þ ¼ � 0 À 2dq 0 p lnðz 3 ÞexpðÀ ibÞ þ c Q w 2p J 1 lnðz 3 À dÞ þ J 2 lnðz 3 þ dÞ þ J 3 lnðz 3 À dÞ þ J 4 lnðz 3 þ dÞ � �ð4Þ [/formula]
where parameters J 1 to J 4 take values as given in. Because of the linearity of Laplace's equation, Eq (3) may be extended for a multi-well system using the principle of superposition:
[formula] O z 3 ð Þ ¼ � 0 À 2dq 0 p lnðz 3 ÞexpðÀ ibÞ þ P N j¼1 c j Q wj 2p ½J 1 lnðz 3 À d j Þ þ J 2 lnðz 3 þ d j Þ þ J 3 lnðz 3 À d j Þ þ J 4 lnðz 3 þ d j Þ�ð5Þ [/formula]
where N is the number of wells. The substitution of z 3 for z 1 gives:
[formula] O z 1 ð Þ¼ � 0 À q 0 z 1 expðÀ ibÞ þ P N j¼1 c j Q wj 2p ( J 1 ln exp pz 1 2d � � À exp pz wj 2d � � h i þJ 2 ln exp pz 1 2d � � þ exp pz wj 2d � � h i þ J 3 ln exp pz 1 2d � � À exp pz wj 2d � � � � þ J 4 ln exp pz 1 2d � � þ exp pz wj 2d � � � � )ð6Þ [/formula]
Using the following dimensionless quantities:
[formula] z 1D ¼ z 1 d z wDj ¼ z wj d Q wDj ¼ Q wj Kbd q 0D ¼ q 0 Kb O D ¼ O Kbd � D ¼ � Kbd c D ¼ c Kbdð7Þ [/formula]
Eq (6) becomes:
[formula] O D ¼ � 0D À q 0D x D þ iy D ð Þexp À ib ð Þ þ P N j¼1 c j Q wDj 2p ½J 1 lnðf 1Dj þ if 2Dj Þ þ J 2 lnðg 1Dj þ ig 2Dj Þ þ J 3 lnðf 1Dj þ ig 2Dj Þ þ J 4 lnðg 1Dj þ if 2Dj Þ�ð8Þ [/formula]
with f 1Dj , f 2Dj , g 1Dj and g 2Dj as given in Appendix A. The real part of Eqgives the dimensionless potential, ϕ D :
[formula] � D ¼ � 0D À q 0D ðx D cosbþy D sinbÞ þ P N j¼1 c j Q wDj 4p ½J 1 lnðf 1Dj 2 þ f 2Dj 2 Þ þ J 2 lnðg 1Dj 2 þ g 2Dj 2 Þ þ J 3 lnðf 1Dj 2 þ g 2Dj 2 Þ þ J 4 lnðg 1Dj 2 þ f 2Dj 2 Þ� ð9Þ [/formula]
## Plos one
while the imaginary part gives the dimensionless stream function, ψ D :
[formula] c D ¼ À q 0D y D cosb À x D sinb ð Þ þ P N j¼1 c j Q wDj 2p J 1 tan À 1 f 2Dj f 1Dj ! þ J 2 tan À 1 g 2Dj g 1Dj ! þ J 3 tan À 1 g 2Dj f 1Dj ! þ J 4 tan À 1 f 2Dj g 1Dj ! " #ð10Þ [/formula]
Plotting Eqs (9) and (10) results in a flow net that illustrates the capture curves of a multiwell system in a strip-shaped aquifer in dimensionless form. The flow net is conveniently plotted using the contour syntax in MATLAB. The stream function has a branch cut for each well that may be removed with the unwrap syntax in MATLAB.
In unconfined aquifers, the saturated thickness varies, so the potential equals to:
[formula] � ¼ 1 2 Kh 2ð11Þ [/formula]
Eqs (8-10) can be rewritten for unconfined aquifer by replacing the product Kb with Kh 0 (discharge per unit width) and using the following dimensionless terms:
[formula] Q wDk ¼ Q wk Kh 0 d q 0D ¼ q 0 Kh 0 O D ¼ O Kh 0 d � D ¼ � Kh 0 d c D ¼ c Kh 0 dð12Þ [/formula]
## Stagnation points
The capture envelopes that pass through the stagnation points (where the flow velocity is zero) in the flow field separate the regions of flow. To find the position of stagnation points, the derivative of the complex potential (Eq 5) is taken with respect to z 3 and the result set equal to zero:
[formula] dOðz 3 Þ dz 3 ¼ À 2dq 0 pz 3 exp À ib ð Þ þ P N j¼1 c j Q wj 2p J 1 z 3 À d j þ J 2 z 3 þ d j þ J 3 z 3 À d j þ J 4 z 3 þ d j " # ¼ 0ð13Þ [/formula]
The roots of Eq (13) are the locations of stagnation points in the z 3 −plane, which are easily transformed to the z 1 −plane. Eq (13) was solved using MATLAB. The values of the stream function at the stagnation points are calculated (Eq 10) and the capture envelopes are drawn by the stream line passing through the stagnation points.
## Drawdown
The steady state drawdown (s) in a confined aquifer from a fully penetrating extraction well is:
[formula] s ¼ h 0 À h ¼ Q w 2pKb ln r 0 r � �ð14Þ [/formula]
where h 0 and h are hydraulic heads at distance r 0 and r from the well, respectively. Since the
## Plos one
velocity potential is ϕ = Kh and the dimensionless head is h D = h/d, the dimensionless drawdown (s D ) can be written in terms of dimensionless potential as:
[formula] s D ¼ � 0D À � Dð15Þ [/formula]
where potentials ϕ 0D and ϕ D are calculated using Eq. The steady state drawdown (s) in an unconfined aquifer at a distance (r) from a fully penetrated extraction well with radius (r w ) is:
[formula] h 2 0 À h 2 ¼ Q w pK ln r 0 r � �ð16Þ [/formula]
Using Eq (11) the dimensionless head in an unconfined is written as:
[formula] h 2 0 ¼ 2� 0D h 0 d ; h 2 ¼ 2� D h 0 dð17Þ [/formula]
From Eqs (16) and (17), the dimensionless drawdown in an unconfined aquifer is:
[formula] s D ¼ 2� 0D h 0 d � � 1 2 À 2� D h 0 d � � 1 2ð18Þ [/formula]
Eqs (15) and (18) can be used to demonstrate the application of proposed model to water quantity management (see Water quantity management projects Section below).
# Results and discussion
In the following sections, Eqs (9), (10) and (13) are solved for strip-shaped aquifers with different boundary configurations and flow conditions, after which we plot the capture envelopes of the well(s). First, the effects of various boundary configurationson the shape and properties of the capture zones are investigated, then the effect of extraction rates and regional flow direction and rate are explored. Due to numerous parameters (number, type, location and extraction/injection rates of wells, and the rate and direction of regional flow) that influence the flow field, many capture envelopes can be simulated. Here, representative examples of capture zones in each parallel boundary configuration are presented, along with the water quantity management and optimized remediation schemes for a pump-and-treat method. The dimensionless values of parameters (i.e., x wD , y wD , Q wD ) can be calculated using Eq.
https://doi.org/10.1371/journal.pone.0229767.t002
## Plos one
## Effects of boundary configuration on the capture zones
Capture zone of well(s) in a strip-shaped aquifer with inflow-inflow boundaries. In this example, we consider a strip-shaped aquifer bounded with inflow-inflow boundariesin which the direction β and the rate q 0D of the uniform regional flow are 0 rad and 0.001, respectively. The parameter J is defined in. Eqs (9), (10), and (13) are solved for 1, 2, 3 and 5 wells (the position and pumping rate of extraction wells are defined in.illustrate, respectively for the four cases, their potential, streamlines, stagnation points and capture envelopes. The black diamond shows the well position and the green diamond illustrates the position of the stagnation point. The dashed lines indicate the velocity potential and the solid lines represent the streamline reaching the location of the extraction well. Inthe well extraction rate is provided from regional flow and two inflow boundaries and the capture envelope extends symmetrically. Due to the direction of regional flow, the capture envelope trends toward the west (left) side. Inwell (2) extracts water from the streams as well as the uniform regional flow. A sharp groundwater divide, effectively a barrier, is established between two wells so that the uniform regional flow provides no water to well (1). Streamlines run parallel along the divide. Inwell (3) has the largest capture envelope because it has the maximum extraction rate. The three wells gain water from both boundaries. Two groundwater divides are also formed, which separates the capture zone of well (2) from those of the other wells. Ineach well has its own capture envelope. Wells (1) and (2) gain water from the north and south side boundaries, respectively, whereas well (3) is recharged by the two boundaries and regional flow. Its capture zone shape shows the influence of the other wells. Well (4) is fully supplied by the northern boundary. A groundwater divide separates the capture zone of well (5) from those of the others while both boundaries supply water to it. In both, a stagnation point exists for wellsand (5) in S1 File.
Capture zone of well(s) in a strip-shaped aquifer with inflow-barrier boundary conditions. For this boundary configuration, the aquifer thickness, hydraulic conductivity and the direction and regional flow rate are the same as the previous examples, as specified inwith J values as given in. Eqs (9), (10) and (13) are solved for 1, 2, 3 and 5 wells and the resulting capture curves are plotted inthe case of 1 well, the extracted water is provided from both the regional flow and the inflow boundary. The streamlines are parallel to the no-flow boundary as expected. Inis mainly supplied by the stream, while well (2) gains water from regional flow and the stream. Inhas the largest extraction rate, with water supplied by the stream and the regional flow. A small portion of regional flow reaches well. The capture zone of well (1) extends towards the stream as well as the west and east sides of the aquifer. The capture envelope of well (2) is smaller, due to its short distance from the north boundary and lower extraction rate.the competition among wells in gaining water from the north boundary based on their pumping rate and distance from the boundaries. Only well (2) benefits from regional flow and its capture zone is limited by the no-flow boundary and manages to gain a small portion of water from the stream along three narrow zones: a) between the capture envelopes of wells (3) and (5), b) the capture envelopes of wells (1) and (3) and c) the capture zone of well (5) and the noflow boundary.
Capture zone of well(s) in a strip-shaped aquifer with barrier-inflow boundaries.results computed using Eqs (9),and (13), solved for the configuration shown inwith parameters as given in. The aquifer thickness, hydraulic conductivity and direction and rate of regional flow are the same as in the previous examples.is supplied by the stream as well as regional flow. Inwells (1) and (2) gain water from the stream and well (2) captures the main portion of regional flow and allows a smaller portion to reach well (1). Inthe stagnation points of wells (2) and (3) are positioned on the barrier boundary and form a small stagnation zone. Well (3) prevents regional flow to reach wells (1) and (2) and encompasses the capture envelope of well (1). As a result, wells (2) and (1) are only fed by the inflow boundary.(1) is recharged by regional flow only and wells (2) and (5) by the stream. Water reaches well (4) via a narrow pathway connecting the well to the upstream regional flow and via two sections of the inflow boundary. Well (3) is fed by the stream and its capture envelope encompasses that of well (2).
## Plos one
## Capture zone of well(s) in a strip-shaped aquifer with barrier-barrier boundaries.
The aquifer in this example corresponds to the case in previous section. The difference is that the in-flow boundary is replaced by barriers. According to, for this boundary configuration the value of parameters J 1 , J 2 , J 3 and J 4 are all equal to +1. As above, the regional flow direction is from west to east (β = 0). Eqs (9), (10) and (13) are solved for 1, 2, 3 and 5 well systems-potential and streamline functions are plotted inAs shown in the figure, the streamlines are parallel to the barrier boundary before merging to the wells and equipotential lines are perpendicular to the barrier boundaries, as expected. Inthe extraction rate is provided by regional flow from the west side of the aquifer and also from the infinite boundary in the east. As a result, the west and east flowing fronts intersect and two stagnation points form on the boundaries. Inregional flow is fully captured by well (2) and well (1) gains water from the east infinite side. Inthe capture envelopes of well (3) and wellextend to the west and east sides, respectively, while that of well (1) extends to both sides along two narrow channels parallel to the barrier boundary. Inand (2) capture water from the west side of the aquifer and wells (4) and (5) from the east side while well (3) captures water from both sides. Well (3) forms two stagnation points each located on one barrier boundary. Comparing this figure with Figs 4-6, one can see how boundary types, pumping rates, and well(s) position control the shape, size and pattern of capture envelopes. Wells interact differently with each other and with the boundaries. In a regional pumping scheme, these differences are important for water quantity management, allocation of water rights, and for pumping rate permits.
## Plos one
## Effects of regional flow direction and rate on the capture zones
In the previous examples, we assumed that the direction of regional flow was from west to east (β = 0). To demonstrate the effect of regional flow direction on the capture envelopes, Figs 4D, 5D, 6D and 7D are replotted for β = π inare replotted for β = π/2 in in S1 File. The difference due to the flow direction is profound, as readily seen in the shape, size and pattern of capture envelopes and the interaction of wells and boundaries. Inin contrast to its correspondingfully captures and prevents regional flow to reach other wells. As the result, it is fed along a shorter segment of the north boundary; the capture zone of well (1) leans somewhat to the west and its stagnation point is displaced further south; well (2) is only fed by the southern stream; well (3) is fed along a longer section of the northern stream and a shorter segment of the southern stream while the capture envelope of well (4) slopes further to the east. Inregional flow is mainly captured by wellalthough and a small portion of it is captured by well (2) through a narrow flow channel along the no-flow boundary. Inthe pattern of capture envelopes changes considerably and regional flow is captured by wells (4) and (3) instead; well (1) reaches the stream and gain water from it; the capture envelope of well (4) encompasses that of well, which moves further to the east. Ingreatly change in size and shape and pattern; the capture envelope of well (1) becomes larger than that of well, which extends to the east by encompassing that of well (5); well (2) is only supplied by the southern boundary. The dimensionless regional flow is increased to 0.02 compared to that of the cases presented in Figs 4-8 (which was 0.001), with the new results given in in S1 File. All the capture envelopes in these figures extend in the opposite sense to the regional flow direction, i.e., the regional flow contributes to the extraction rate of all wells. It is clear that the size, shape, dimensions and pattern of well capture zones and the interaction of boundaries and wells are greatly changed compared to Figs 4-8.
Figs 4-8 and in S1 File demonstrate that the capture zone model developed above can help delineate the interaction between surface-subsurface flows in water resources management projects such as allocating water rights, issuing well drilling permits, pumping rate permits, etc., and for the sustainable development of water resources in a region.
## Validation of the solution
To validate the developed capture zone solutions, the capture zones of 5 wells ofwere generated by MODFLOW 2000and MODPATHand plotted inFor the four plots shown inthe well coordinates as well as their extraction/injection rates were set to those ofThe capture zones and the pattern of streamlines generated by the numerical models are comparable to the analytical model capture curves (i.e.,. The slight difference is mainly due to the approximation inherent in the numerical models. It is worth mentioning that the usage of the capture envelope equations developed in this paper is far easier, less time consuming, and more accurate than using the abovementioned numerical models, although numerical models are more flexible in dealing with complicated boundary conditions. Figs 8 and 9 demonstrate the potential of developed analytical capture zone models to verify the accuracy of numerical models.
## Effects of the well distance from boundaries on the capture zones
The positions of the wells are kept constant relative to the south side stream as inand the northern boundary is moved away so that its distance to the wells increases by a factor of 2.that the capture zones are reshaped such that well (1) no longer is fed by the north side stream, but is instead fully fed by regional flow, while well (3) gains less water from the north side stream. Well (4), which was fully supplied with the north side stream, now gains water from both streams. Well (5), which was fed by both streams, is now recharged by the southern stream only. Inthe distance between wells and the northern stream is increased, so that the capture zone of wells (1), (3) and (5) do not reach it, while well (1) is fed only by regional flow. Wells (3) and (5) are no longer supplied by the northern stream. The contribution of the north side stream to well (4) is also negligible. Practically speaking, for this case, the strip-shaped aquifer acts as an aquifer located within a semi-infinite spatial domain.
## Groundwater remediation scheme design
In this section, use of dimensionless capture type curves in water quality management such as groundwater remediation projects is demonstrated. A variety of remediation technologies can be used to ensure contaminant plumes are either contained, removed and/or treated so they no longer pose a threat to water resources. In a typical in situ remediation project such as bioremediation, pump-and-treat, plume containment, etc., polluted groundwater is extracted using one or more wells, treated and possibly reinjected. In such projects the main
## Plos one
elements of an efficient and cost-effective design are determination of the capture zone, the optimal number of wells, the optimal injection/extraction rates, and the layout of wells with respect to the plume .
We consider a confined aquifer with the boundary configuration of type (a) inThe problem setup includes known dimensions and hydraulic properties (aquifer thickness, length and width are 20 m, 1000 m and 500 m, respectively; K = 5 m/d and q 0 = 0.1 m 2 /d). A contaminant plume is located along the direction of regional groundwater flow (from the west to the east, β = 0). The aim is to contain the plume hydraulically and prevent its extension downgradient using two or three wells. A logical step is to position an injection well (to inject treated water) in the vicinity of the contaminant source and to place one or more extraction wells at the leading edge of the plume. A solution could be determined manually by using Eqs (9) andto generate a set of capture zone curves for various well layouts and extraction/injection rates. This approach, although cumbersome, could achieve a satisfactory plume capture design.
Two automated parameter optimization schemes are used instead, i.e., Eqs (9) and (10) are embedded into the PSO algorithm (Appendix B) and GA (Appendix C).the optimal capture envelopes for two-and three-well remediation schemes, with the corresponding well positions and extraction/injection rates given in . As is evident inboth systems involve a closed region that contains and encompasses the plume. shows that both optimization schemes result in similar optimal solutions. This approach can easily be extended to include financial constraints associated with well installation and operation. Inthe three-well solution has a lower total injection/extraction rate, so it represents the optimal solution if the remediation scheme is dominated by pumping costs.also generated by the abovementioned numerical models and presented in in S1 File. In this simulation, if the well positions and flow rates were unknown, finding the optimum plan for the plume containment would be difficult to determine manually.
## Water quantity management projects
The availability and sustainability of groundwater in many major aquifers in the world are threatened due to excessive depletion. To avoid this, a typical management strategy is to allocate well pumping rates in such a way that the drawdown in the aquifer does not exceed a certain value (permissible drawdown). As an example, we consider a confined aquifer with the boundary configuration of type (a) inThe problem setup includes known dimensions and hydraulic properties (aquifer thickness, length and width are 20 m, 1000 m and 500 m, respectively; K = 5 m/d and q 0 = 0.1 m 2 /d). Further, it is assumed that the aquifer is pumped by five wells with positions given in . The objective of this example is to calculate extraction rates such that the maximum drawdown is less than 4 m. . Extraction/injection rates and well positions in two remediation scenarios.
## Pump-and-treat remediation scenarios
Type of well PSO GA
x wD y wD Q wD x wD y wD Q wD
## Plos one
Based on Eq (7), the wells and the aquifer parameter values are converted to dimensionless values and Eq (15) is solved using the abovementioned optimization algorithms. The calculated optimum pumping rates by both optimization algorithms are presented in with similar results. The capture envelope patterns of the 5 wells for the optimum solution are presented inthe corresponding dimensional drawdown contours calculated by Eqare shown inNote that the permissible 4-m drawdown is not exceeded.presents the definition of variables in the equations.
# Conclusions
Analytical models of the capture zone (stream function, velocity potential and stagnation point) for a multi-well system with any number of extraction/injection wells in confined and unconfined strip-shaped aquifers, with/without uniform regional flow are developed from the theory of complex velocity potentials and the Schwarz-Christoffel conformal mapping. The solution is provided for four strip boundary configurations: barrier-barrier, barrier-inflow, inflow-barrier and inflow-inflow. The solution, which is straightforward to compute, can be used to generate dimensionless capture type curves. The applicability of the developed model to evaluate surface-subsurface water interaction, well-head protection plans and water rights adjudication-which are key factors in the management of water resources-were demonstrated through numerical examples. These examples also displayed that the size and pattern of the capture envelopes are controlled by the extraction/injection rate, distance of wells from the boundaries, distances between wells, type and number of wells, the rate andtwo-well and b) three-well system. "I" shows the injection wells and "E" the extraction wells. The colored area is the original contaminant plume. The thick black line surrounding the plume is the dividing streamline that defines the capture zone that separates the plume from the rest of the aquifer and β = 0.
https://doi.org/10.1371/journal.pone.0229767.g011 . Well positions and the calculated extraction rates of wells in the quantity management scenario.
## Well number
Well coordinates PSO GA PLOS ONE direction of regional flow and the initial head along the boundaries. Further, the derived solution was embedded with two optimization schemes (PSO and GA) for the determination of optimal plume capture schemes for pump-and-treat remediation projects and for determination of the optimal pumping policy for sustainable water extraction schemes, with similar results obtained by both methods. The solution was validated by numerical models, and is thus suitable for quantifying the accuracy of such models for complex water injection/extraction scenarios. In general, the proposed models can be used by hydrogeologists, engineers and planners to design optimal layouts of groundwater remediation projects (e.g., in situ bioremediation or chemical oxidation schemes) that involve pump-and-treat and plume containment operations, to design the most cost-effective pumping schemes for sustainable development of water resources, to study the interaction of surface and groundwater resources and to verify numerical models. (B-6)where c 1 and c 2 are acceleration constants (learning factors), r 1 and r 2 are random numbers uniformly distributed between 0 and 1, subscripts p and g respectively indicate the pbest and gbest values of the particles, ω is the inertia weight used to control the impact of the previous history of velocities on the current value, χ is the constriction coefficient, applied to restrain the velocity (v), number of iterations (t), well rates (Q w ) and the particle position (z w ).
## 5.
Repeat steps 3 and 4 until the fitness criterion or the maximum number of iterations is reached.
The above procedure was performed within MATLAB.
For water quantity management projects, the aim is to optimize extraction rates contingent upon a certain value of drawdown (permissible drawdown, s p ) where the objective function and constraints are:
Maximize Q wjD ðB À 7Þ (B-7)subject to:
Q jmin � Q wjD � Q jmax s max � s p ðB À 8Þ
(B-8)where, Q wjD is the extraction rates based on Eq (8), Q jmin and Q jmax are userdefined minimum and maximum extraction rates for the j th well.
Supporting information S1 File. Mapping the conceptual model from the physical plane z 1 to the imaginary planes z 2 and z 3 , and effect of regional flow rate and direction on the shape and orientation of well capture zones.
(DOCX) |
Quantitative approaches are now widely used to study the genetic architecture of complex traits. However, most studies have been conducted in single mapping populations, which sample only a fraction of the natural allelic variation available within a gene pool and can identify only a subset of the loci controlling the traits. To enable the progress towards an understanding of the global genetic architecture of a broad range of complex traits, we have developed and characterised six new Arabidopsis thaliana recombinant inbred populations. To evaluate the utility of these populations for integrating analyses from multiple populations, we identiWed quantitative trait loci (QTL) controlling Xowering time in vernalized plants growing in 16 h days. We used the physical positions of markers to align the linkage maps of our populations with those of six existing populations. We identiWed seven QTL in genomic locations coinciding with those identiWed in previous studies and in addition a further eight QTL were identiWed. Communicated by A. Charcosset.Electronic supplementary materialThe online version of this article (
# Introduction
Natural genetic variation can be exploited to advance studies in many aspects of plant biology and is central to evolutionary biology and plant breeding [bib_ref] The impact of genomics on the study of natural variation in Arabidopsis, Borevitz [/bib_ref]. Molecular dissection of the natural genetic variation controlling quantitative developmental traits is made diYcult by their complex nature. Genes that contribute to variation in quantitative traits are diYcult to identify as each locus is typically responsible for only a small proportion of the total phenotypic variation [bib_ref] Finding genes that underlie complex traits, Glazier [/bib_ref]. Such loci are termed quantitative trait loci (QTL) and they can be identiWed and mapped by quantitative genetic approaches. These provide estimates of the numbers of loci, the type and magnitude of their eVects, interactions between genes and gene by environment interactions; collectively referred to as the genetic architecture of a trait. However, QTL can only be identiWed where the alleles present at a locus diVer between the parents of the population. Consequently, the sets of QTL identiWed will often be population-speciWc. Therefore, any description of the genetic architecture of a trait made from a single population, either natural or experimental, is likely to describe only a small part of the "global genetic architecture" for a trait within a species [bib_ref] Mapping quantitative trait loci in multiple populations of Arabidopsis thaliana identiWes natural..., Symonds [/bib_ref]. In order to understand the complete range of genetic variation that is responsible for the quantitative control of trait variation, many experimental segregating populations, derived from a diverse array of accessions, are needed [bib_ref] Gene function beyond the single trait: natural variation, gene eVects and evolutionary..., Tonsor [/bib_ref].
In addition to the availability of segregating populations, QTL analyses require the construction of linkage maps for those populations. Sequence-based markers are particularly valuable in species for which complete genome sequences are available, such as A. thaliana . This is because they can be anchored to the genome sequence, allowing precise alignment of linkage maps that have been constructed using diVering sets of markers. Simple sequence repeat (SSR) and insertion/deletion (InDel) markers are two of the most widely used types of sequence-based markers. SSRs are sites where one or a few bases are tandemly repeated for a varying number of times. They are found in both coding and non-coding regions, highly polymorphic, co-dominant and abundant across the A. thaliana genome, and can be easily assayed by PCR [bib_ref] Distribution of microsatellites in relation to coding sequences within the Arabidopsis thaliana..., Casacuberta [/bib_ref]. There are many publicly available SSR markers for use in A. thaliana (http:// www.arbidopsis.org; http://www.arbidopsis.org/Cereon/ index.jsp; http://www.inra.fr/vast/). InDel markers are highly abundant, exist mostly in only two co-dominant variants and they are predominantly phenotypically neutral [bib_ref] Genetic mapping with SNP markers in Drosophila, Berger [/bib_ref]. Both SSR and InDel markers are suited to high throughput genotyping using capillary sequencers. Several genomic screens have been conducted for InDel and other sequence variations between A. thaliana accessions [bib_ref] Arabidopsis map-based cloning in the post-genome era, Jander [/bib_ref] [bib_ref] Large scale identiWcation and analysis of genome-wide single nucleotide polymorphisms for mapping..., Schmid [/bib_ref] [bib_ref] The pattern of polymorphism in Arabidopsis thaliana, Nordborg [/bib_ref] , resulting in many publicly available markers (http://www.arabidopsis.org/Cereon/index.html; http://www.mpiz-koeln.mpg.de/masc/; http://www.walnut. usc.edu/2010/).
Arabidopsis thaliana is a largely autogamous species. Consequently, the most favoured approach to the development of "immortal" mapping populations has been the development of recombinant inbred (RI) populations. These are produced by single seed descent to at least the F8 generation, at which each line is practically homozygous and genetically identical individuals can be propagated indeWnitely. This gives RI populations several important advantages over, for example, F2 populations. The principle advantages are that extensive linkage maps can be constructed and the lines can be analysed, in replicated experiments, for a wide range of phenotypic characteristics. The research potential of RI populations for the study of complex traits in A. thaliana has been demonstrated by the two most frequently used populations Ler £ Col [bib_ref] Recombinant inbred populations for mapping RFLP and phenotypic markers in Arabidopsis thaliana, Lister [/bib_ref] and Ler £ Cvi [bib_ref] Development of an AFLP based linkage map of LER, Col and Cvi..., Alonso-Blanco [/bib_ref]. Both have been publicly available for a number of years and in that time the Ler £ Cvi population has been analysed for in excess of forty traits [bib_ref] New Arabidopsis recombinant inbred line populations genotyped using SNPWave and their use..., El-Lithy [/bib_ref] [bib_ref] Evidence for large-scale population structure of Arabidopsis thaliana from genome-wide single nucleotide..., Schmid [/bib_ref] [bib_ref] Common sequence polymorphisms shaping genetic diversity in Arabidopsis thaliana, Clark [/bib_ref] , which aimed to explore the phylogenetic relationships of a broad range of accessions.
Arabidopsis has a wide distribution throughout the northern hemisphere and grows in a range of environments. It is therefore an excellent system to use in understanding the molecular variation underpinning adaptive variation. The timing of the transition from vegetative to reproductive development has signiWcant consequences for the reproductive success of plants and is the result of the integration of both environmental and endogenous cues. A complex regulatory network has evolved, consisting of multiple pathways that quantitatively regulate a set of genes (the Xoral pathway integrators) whose activity causes the transition of the meristem to reproductive development [bib_ref] Arabidopsis, the Rosetta stone of Xowering time?, Simpson [/bib_ref] [bib_ref] Control of Arabidopsis Xowering: the chill before the bloom, Henderson [/bib_ref] [bib_ref] Remembering winter: toward a molecular understanding of vernalization, Sung [/bib_ref]. Two particularly important components of this network are the genes at the FRI (FRIGIDA) and FLC (FLOWERING LOCUS C) loci. The range of natural variation in these genes was investigated using many diVerent Arabidopsis accessions [bib_ref] Analysis of the molecular basis of Xowering time variation in Arabidopsis accessions, Gazzani [/bib_ref] [bib_ref] Role of FRIGIDA and FLOWERING LOCUS C in determining variation in Xowering..., Shindo [/bib_ref] [bib_ref] Variation in the epigenetic silencing of FLC contributes to natural variation in..., Shindo [/bib_ref]. Another important gene involved in the control of Xowering time MAF1 (MADS AFFECTING FLOW-ERING 1) has also been found to be a target of natural selection . Light has a major role in modulating Xowering time and allelic variation at the locus for the photoreceptor PHYC (PHYTOCROME C) was shown to aVect this trait [bib_ref] The PHYTOCHROME C photoreceptor gene mediates natural variation in Xowering and growth..., Balasubramanian [/bib_ref]. Investigations of the genetic control of Xowering time have been conducted using a range of RI populations [bib_ref] Recombinant inbred populations for mapping RFLP and phenotypic markers in Arabidopsis thaliana, Lister [/bib_ref] [bib_ref] Development of an AFLP based linkage map of LER, Col and Cvi..., Alonso-Blanco [/bib_ref] [bib_ref] Bay-0 £ Shahdara recombinant inbred line population: a powerful tool for the..., Loudet [/bib_ref] [bib_ref] Quantitative trait locus analysis of growth-related traits in a new Arabidopsis recombinant..., El-Lithy [/bib_ref] and diVering numbers of QTL identiWed. Many, but not all, of these colocalise in diVerent populations.
We report the development and characterisation of six new RI populations, each comprising 94 independent lines. We also demonstrate their applicability to QTL mapping, using Xowering time as an example. The populations were developed on the basis of the parent accessions diVering in seed lipid composition and representing a broad range of geographic origins [bib_ref] Natural variation for seed oil composition in Arabidopsis thaliana, O'neill [/bib_ref]. The populations have been deposited with the Arabidopsis Biological Resources Centres (http://www.arabidopsis.info/) for public distribution (catalogue numbers to follow).
# Materials and methods
## Generation of the recombinant inbred populations
Seeds for each of the accessions used for the development of the populations were received from the Nottingham Arabidopsis Stock Centre (NASC; http://www.arabidopsis.info/). Their geographic origins and catalogue numbers, the combinations crossed and population names are listed in. The line JIC240#14 was isolated as an apparent contaminant in the seeds we received of the Mz-0 accession. Fingerprinting all 364 available accessions for a range of markers revealed no matches, therefore the line has been submitted to NASC as a new accession (catalogue number to follow).
Before the crosses used to generate the RI populations were made, a single plant from each accession was grown and selfed. These selfed seed stocks were used for crossing, by standard emasculation procedures, to produce unidirectional F1 plants. The F1 plants were conWrmed to be hybrids by molecular marker analysis and selfed. The RI populations were developed from the F2 lines by single seed descent to F7. The process was accelerated and the likelihood of crossing reduced, by removing side branches. Plants of the F8 generation were allowed to fully develop. Plastic sleeves were placed around the individual plants to prevent cross fertilisation and facilitate seed collection.
## Plant growth conditions for ri population generations
Seeds were sown on to Arabidopsis soil mix: Levington's M2 compost with 4 mm grit (8:1) and intercept (to protect against aphids) in P40 trays [bib_ref] Natural variation for seed oil composition in Arabidopsis thaliana, O'neill [/bib_ref]. Plants were vernalized at the seed stage at every generation. To achieve this, they were placed, directly after sowing, in a 5°C chamber with an 8 h photoperiod for 6 weeks. After this time, the trays were moved to a temperature-controlled glasshouse at 22°C (day) and 18°C (night) temperature, with supplementary sodium lighting to provide a 16 h photoperiod.
Plant growth conditions for Xowering time analysis F8 seeds were sown to Arabidopsis soil mix and vernalized for 6 weeks as described above. The plants were then moved to a controlled environment room (CER) or a glasshouse (GH). Both provided a 22°C day/18°C night, 16 h photoperiod with supplementary light provided by sodium lamps. When the Wrst true leaves were fully developed, seedlings were pricked out to Arabidopsis soil mix in 7 cm pots. Flowering time was recorded as the number of rosette leaves present when the bolt had extended 1.5 cm. All populations were grown in replicated experimental designs. Three of the populations SG, KB and TJ were grown under GH conditions in a complete randomised block design. The plant positions within trays were randomised using GenStat [bib_ref] Mapping of QTL for resistance against the crucifer specialist herbivore Peris brassicae..., Payne [/bib_ref]. Three replicate plants were observed per line. Two of the populations CA and WC were grown in the CER as complete randomised blocks with two replicate plants. The blocks were split into two equal sub-blocks and grown in two separate occasions. The sixth population, NG, was grown under CER conditions as a complete fully randomised block with two replicate plants. To minimise positional eVects, plant trays were rotated 180 o and moved along the bench, one tray position each week.
## Marker development and assay
The markers used in this work were identiWed from four diVerent resources: 48 SSRs were identiWed from the TAIR database (http://www.arabidopsis.org), 28 InDels were selected from the Monsanto database (http://www.arabidopsis.org). Another 37 markers were developed at the John Innes Centre for this work. Twenty-three of these were SSRs identiWed by using repeat sequences in BLASTN (http://www.blast.wustl.edu) alignments against the Arabidopsis genome sequence. The remaining 14 were InDel markers, which were identiWed after re-screening the raw sequence data generated in the Nordborg SNP generation/ sequencing programme (http://www.walnut.usc.edu/2010/) for InDels. The primer combinations used for the 37 new markers developed are shown in [fig_ref] Table 2: The 37 newly developed SSR and InDel markers [/fig_ref]. The complete marker set used for the individual populations is available as supplementary materials online, or from http://www.jic. bbsrc.ac.uk/staff/ian-bancroft/ (Supplementary.
Genomic plant DNA was extracted using the high throughput Qiagen Dneasy 96 preparation method by following the manufacturer's instructions (Qiagen GmbH, Hilden, Germany). The markers were ampliWed using PCR and the products were resolved on an ABI 3700 capillary sequencer in order to determine allele sizes (Genome Labo- ratory Services, John Innes Centre). Markers identiWed as being polymorphic between the parents of a particular population were used to genotype the 94 individuals in each population.
## Genetic mapping and qtl analysis
Genetic linkage maps for each recombinant inbred population were constructed using the Kosambi mapping function in the JoinMap program version 3 (Van Ooijen and Voorrips 2001) based on the polymorphic marker data generated for each population. The segregation pattern of parental alleles was determined for all loci on each linkage group using the JoinMap program. The signiWcance values of the segregation ratios for all marker alleles were tested using a 2 test. Flowering time was analysed separately for each experiment by analysis of variance using GenStat and the mean line values used for QTL mapping of each population independently and to estimate heritabilities (h 2 b ). To identify the regions of the genome containing QTL, we initially used interval mapping as implemented by the software package Map QTL version 4 [bib_ref] Quantitative trait locus mapping and DNA array hybridization identify an FLM deletion..., Van Ooijen [/bib_ref]. These were further analysed using composite interval mapping (MQM) using markers identiWed as cofactors by the automatic cofactor selection function in MapQTL. Genome wide signiWcance thresholds of 5% for the LOD values were determined using the permutation command within MapQTL. The number of permutations was set at 10,000.
# Results
## Generation of the ri populations
Six new RI populations derived from the twelve accessions described inwere developed by single seed descent to the F8 generation. To reduce the eVect of the segregation for late Xowering and seed dormancy, all lines, at each generation, were vernalized at 5°C at the seed stage for 6 weeks. Each population consisted of 94 lines, which are available from the Nottingham Arabidopsis Stock Centre (NASC; http://www.arabidopsis.info/). We refer to the populations by names derived from the Wrst letters of the name of the parental accessions, as listed in.
## Linkage map development
Linkage maps are required for populations if they are to be used for QTL analysis. We developed a panel of 113 polymorphic molecular markers, derived from four diVerent resources, as described in the material and methods section. Details of this marker set, and the linkage maps constructed for the six populations, are available as supplementary materials online and from http://www.jic.bbsrc.ac.uk/staff/ ian-bancroft/. The linkage maps for the individual populations consisted of between 47 and 58 markers. The maps ranged in length from 297 cM for SG to 470 cM for the WC population. These are comparable to the lengths of linkage maps reported for other RI populations [bib_ref] Development of an AFLP based linkage map of LER, Col and Cvi..., Alonso-Blanco [/bib_ref] [bib_ref] Bay-0 £ Shahdara recombinant inbred line population: a powerful tool for the..., Loudet [/bib_ref] [bib_ref] New Arabidopsis recombinant inbred line populations genotyped using SNPWave and their use..., El-Lithy [/bib_ref]. All of the markers used in the genetic mapping were collinear with their expected physical positions, i.e. relative to the genome sequence. The linkage maps developed had average distances of 6.2-9.6 cM between markers. However, all populations contained larger intervals. The maximum interval in CA, 41 cM, was found on chromosome 1 between the markers at 66 and 107 cM. In the TJ population, the largest interval, 38 cM, was found between 19 and 57 cM on chromosome 3. That in the WC population was located on chromosome 1 between 49 and 81 cM and a similar sized gap was also found on chromosome 1 in the KB population between 33 and 61 cM. In NG, the largest gap occurred on chromosome 5, between 33 and 55 cM and in SG the largest gap was at the top of chromosome 5, between 0 and 18 cM.
The ratio of physical distance (kb) to genetic distance (cM) provides a measure of the frequency with which recombination events have occurred during the development of the populations. The data given inrepresents the whole-genome mean ratios for each of the populations. The overall ratios for the populations varied from a value of 326 Kb/cM for SG to a value of 236 Kb/cM for CA. These data could be indicative of variation in recombination rate during the development of the diVerent RI populations. It has been reported that structural chromosome variants between accessions are a common occurrence ) and cytogenetic studies have suggested the existence of diVerences in local recombination rates [bib_ref] Variation in chiasma frequency among eight accessions of Arabidopsis thaliana, Sanchez-Moran [/bib_ref]. However, to investigate this in the populations generated here, a much larger number of markers would be required. It should be noted that extensive suppression of recombination in parts of a genome will aVect the ability to conduct map based cloning in such regions. The proportion of the expected ca. 120 Mb genome represented by the linkage maps was calculated, based upon the positions of the terminal markers on each linkage group. Only the map for SG, which spanned 97 Mb, failed to represent at least 85% of the genome. This shortfall was largely because we were unable to identify polymorphic markers at the distal ends of chromosomes 1 and 5.
The average frequencies of heterozygosity for all alleles in the six populations described were examined and found to be very close to 1.0% or less. It is expected that by the F8 generation RIs will be essentially homozygous as the theoretical chance of any given marker being heterozygous is only 0.78% [bib_ref] Development of an AFLP based linkage map of LER, Col and Cvi..., Alonso-Blanco [/bib_ref].
Where, no bias has occurred in the selection of individual plants during the development of a population, it is expected that the segregation ratio for each marker would be 1:1 for the parental alleles. [fig_ref] Figure 1: Allele segregation ratios for the six populations [/fig_ref] presents the allele segregation ratios, with the female allele expressed as a percentage of the male and female alleles combined, for the six populations we developed. SigniWcant distortion from the expected 1:1 at P < 0.001 was observed in all of the populations, as summarised in Supplementary Table 2. There were regions of very unbalanced allele inheritance in populations CA, WC, KB and NG. In the KB population this was found on the linkage groups 1, 4 and 5, where in all cases the Br-0 alleles were in excess. The most signiWcant deviation was in the interval 5-33 cM on chromosome 1, with the peak at the aths0392 marker, where only 22% of the alleles were Kondara in origin. A similar ratio was found at the marker t6k21ind15-15 on chromosome 4 in a region spanning 25-46 cM. The NG population displayed a very marked distortion on chromosome 1, where only 11% of the alleles for the f9hind26-26 marker were from Nok-3. This was part of the region, 25-79 cM, in which all of the markers displayed distorted inheritance. The Nok-3 alleles were under represented at this location and also on chromosome 2 at pls1, however, the reverse was found on chromosome 4 at t19j18ind30-30 (6 cM), where the Nok-3 was in excess. In CA, unbalanced allele inheritance was found on four of the linkage groups with the most severe being on 1 and 5 at the markers jconnchr1_9.9 and f14f1847646 with only 8 and 13%, respectively, of the alleles coming from Cvi-0. In all cases, the Ag-0 alleles were in excess. The region aVected on chromosome 1 lay between 38 and 66 cM and those on chromosome 5 lay between 0 and 30 cM and 69-76 cM. In WC, unbalanced segregation ratios were observed on three of the chromosomes 1, 3 and 4. The distorted regions on chromosome 1 lay between the markers nga59-f8k728985 (0-29 cM) and aths0392-f1404682 (41-49 cM). On chromosome 3, the region between the markers mzn14ind29-29-t16k521877 (50-70 cM) displayed allele distortion. On both chromosomes 1 and 3, allele inheritance was biased in favour of Wt-5. Two separate regions occurred on chromosome 4, 0-18 and 60-65 cM. In the top region the jv30-31 marker had predominantly Ct-1 alleles (78%) but the reverse was the case at the f27g1959898 marker (68%) at 65 cM. Some of the regions with distorted allele inheritance at least partially overlapped between some of the populations. This was found to be the case for chromosome 1 between CA, WC, KB and NG. Another region on chromosome 4 was common to both CA and WC. A region identiWed as having unbalanced allele composition towards the top of chromosome 5 in CA was also picked up in KB and TJ. Other regions identiWed as having distorted allele inheritance appeared to be population speciWc [fig_ref] Figure 2: QTL for Xowering time for 6 RIL Arabidopsis populations [/fig_ref]. On examination of the total number of alleles mapped in each population it was found that an unbalanced pattern of inheritance, as tested by a 2 test (P < 0.01), occurred in two of the populations, KB and CA. The maternal parents in these crosses were Kondara and Cvi-0, respectively, and accounted for only ca. 41% of the total alleles studied in those populations. The levels of the maternal progenitor genome in the other four crosses were closer to the expected value of 50%.
## Mapping qtl controlling xowering time
Under the conditions used in our experiments, subsets of lines in all the populations Xowered at times beyond the ranges of their parental accessions, a phenomenon termed transgressive segregation. For example, the accessions Kondara and Br-0 bolted after 11 and 10 rosette leaves, respectively, while the individual RI lines in the KB population displayed leaf numbers ranging from 7 to 17. Flowering time is known to have a large genetic component, making it a highly heritable trait. This was found to be the case in Wve of our six populations, where the values for heritability (h 2 b ) of Xowering time varied between 0.6 and 0.8 [fig_ref] Table 4: QTL aVecting Xowering time in the six populations and heritability in the... [/fig_ref]. However, in the case of CA, it was very low at 0.2. This was most likely the result of the variation between replicates being relatively high.
Our analyses revealed between 1 and 5 QTL in each population, as summarised in [fig_ref] Table 4: QTL aVecting Xowering time in the six populations and heritability in the... [/fig_ref]. SigniWcance for the LOD values, estimated through permutation testing, indicated that nine of these were signiWcant at P < 0.001 (LOD values >3.8), four were signiWcant at P < 0.01 (LOD values >3) and two were signiWcant at P < 0.05 (LOD values >2.3). Individually, these 15 QTL accounted for between 9 and 58% of the phenotypic variation for Xowering time found in the populations. Only one QTL, SG.1, (LOD = 3.8) was detected on chromosome 1 this was positioned around the marker t12p18ind8-8 at 74.5 cM and accounted for 14% of the phenotypic variation. The paternal Gy-0 allele was responsible for increasing the Xowering time at this QTL. Similarly, just one QTL was found on chromosome 2. This QTL, KB.1, had a LOD value of 13 (P < 0.001) at 30.7 cM and accounted for 58% of the phenotypic variation in that population. In this case the allele responsible for delaying Xowering time arose from the maternal parent Kondara. No QTL was found on chromosome 3 in any of the populations. In four of the populations NG, TJ, CA and SG, QTL were detected at the top of the upper arm of chromosome 4. Four QTL in the diVerent populations mapped between 0.5 and 14 cM, and each accounted for between 9 and 38% of the phenotypic variation.
Of all of the chromosomes, the largest number of QTL, eight, were found to map to chromosome 5 and were distributed along its length. These QTL varied in size but the three strongest examples were found to map to the lower arm in both the WC (WC.1, WC.2) and TJ (TJ.3) populations. WC.1 and WC.2 had LOD values of 8.9 and 10.2, respectively, and each accounted for 37 and 41% of the variation found. The 2 LOD intervals for WC.1 lies between 93 and 103 cM, whilst WC.2 was lower down at 107 cM through to the end of that chromosome. The increasing alleles at both loci came from the maternal parent Wt-5. The QTL TJ.3 had a LOD score of 7.4 and the 2 LOD intervals for this QTL spanned from 74 to 97 cM. TJ.3 accounted for 37% of the phenotypic variation in this population. The 240 #14 allele was the increasing allele in this cross. A second but weaker QTL also mapped in the TJ population to chromosome 5 but to the upper arm at 13.7-26 cM. This QTL, TJ.2, was a weaker QTL with a LOD score of 3.1
Analysis with CIM showed that when marker nga8 (SG.2) was used as a cofactor, SG.1 increased in signiWcance from LOD 2.4 to 3.8 and when marker t12p18ind8-8 (SG.1) was used as a cofactor, SG.2 increased from LOD 3.4 to 4.5. Similar eVects were observed for QTL NG.1 and NG.3, and TJ.1 and TJ.3. In the case of NG.1, it just reached the 5% signiWcance level, when mrg2138439 (NG.3) was selected as a cofactor, while NG.3 also showed increased signiWcance, increasing from LOD 4.7 to 5.4 with t19j18ind30-30 (NG.1) as a cofactor. Using nga8 (TJ.1) as a cofactor, TJ.3 increased from LOD 7.4 to 8.7, while selecting both jv65-66 and k8a1022396 as cofactors (TJ.3), TJ.1 increased from non-signiWcance to a LOD of 2.9. In SG, when nga151 (SG.4) was used as a cofactor, a further QTL not previously identiWed through interval mapping, and designated SG.3, was found with a LOD score of 3.8 between 16 and 38 cM. This QTL was not visible when any of the other QTL cofactors were examined.
# Discussion
In this paper, we describe six new RI populations, which are communally available through the Arabidopsis Biological Research Centres. These populations will provide an additional resource for the large research community focussed on A. thaliana and support investigations on a broad range of plant biology disciplines. Genetic polymorphisms between the parental genotypes are crucial for linkage mapping while allelic variation is necessary for QTL detection. In an eVort to maximise genetic variation in the populations generated, the accessions combined in each of the six original crosses were from geographically diverse locations. Eleven of the accessions were screened in two large-scale studies [bib_ref] New Arabidopsis recombinant inbred line populations genotyped using SNPWave and their use..., El-Lithy [/bib_ref] , which investigated the levels of polymorphism between a wide range of accessions and subsequently calculated their relatedness. From these data, it is clear that the accessions used in each of the parental pairing combinations, with the exception of TJ, could be discriminated. It therefore appears that the strategy behind the selection of suitable crosses was appropriate; geographical location can be used as a means to select for genetic divergence among accessions.
To determine the complete genetic architecture for any trait, it must be analysed in many diVerent genetic backgrounds over many occasions under diVerent environmental conditions [bib_ref] Mapping quantitative trait loci in multiple populations of Arabidopsis thaliana identiWes natural..., Symonds [/bib_ref]. If this objective is to be met access to a large number of "immortal" mapping populations such as RIs is necessary. The genetic linkage maps constructed for the six populations described here are comparable in length to previous ones described in the literature [bib_ref] Development of an AFLP based linkage map of LER, Col and Cvi..., Alonso-Blanco [/bib_ref] [bib_ref] Bay-0 £ Shahdara recombinant inbred line population: a powerful tool for the..., Loudet [/bib_ref] [bib_ref] New Arabidopsis recombinant inbred line populations genotyped using SNPWave and their use..., El-Lithy [/bib_ref]. Our aim was to develop maps with an optimal distance of ca. 10 cM between the markers. Though the average distance between markers in the populations ranged between 6.2 and 9.6 cM, there were also numerous gaps in each that exceeded this 10 cM value. DiYculties with the identiWcation of polymorphic markers at the time of the screens meant that some large gaps remain in the genetic maps generated. The presence of large intervals will hinder QTL mapping in these regions. However, with the advances in marker development, it should be possible for future users to identify suitable markers to Wll in any gaps in a population map of interest. Despite this limitation, we have developed linkage maps that, although rudimentary, are fully anchored to the A. thaliana genome sequence. By keeping each population size to 94 lines, it enables high throughput molecular screening to be conducted. The markers used in this work routinely gave reproducible results.
In any QTL mapping work, the quality of the populations used is fundamental. The average frequencies of heterozygosity for all alleles in the six populations described here were found to be ca. 1.0% or less. This value is very much in line with the expectation that by the F8 generation RIs are essentially homozygous and there is only a 0.78% theoretical chance of any marker being heterozygous [bib_ref] Development of an AFLP based linkage map of LER, Col and Cvi..., Alonso-Blanco [/bib_ref]. The segregation ratios across the six populations were found to be largely below 1:2, though in each of the populations there were exceptions to this with some regions displaying severely distorted allele inheritance pattern. Such regions are likely to aVect the detection of QTL, which lie within them. In four of the populations, CA, KB, WC and NG regions with biased segregation ratios for marker alleles on chromosome 1, at least partially overlapped between 2.8 and 23.5 Mb. A number of previous studies on diVerent RI populations also found unbalanced inheritance in this region of the genome [bib_ref] New Arabidopsis recombinant inbred line populations genotyped using SNPWave and their use..., El-Lithy [/bib_ref] [bib_ref] Mapping of QTL for resistance against the crucifer specialist herbivore Peris brassicae..., Payne [/bib_ref]. In work conducted by [bib_ref] Analysis of natural allelic variation at seed dormancy loci of Arabidopsis thaliana, Alonso-Blanco [/bib_ref] on the Ler £ Cvi population a delay of germination (DOG) QTL was mapped onto chromosome 1 at 7.3 Mb. The Cvi allele at this locus increased the time of seed storage required to allow seed germination to occur. In the CA population, at marker position 9.9 Mb, we found that the level of Cvi-0 alleles dropped to 7.8% of the total alleles at that locus while Kondara contributed just 20% of the alleles at for the marker positioned at 10.9 Mb in the KB population. It was observed at the F3 generation that these two populations exhibited reduced germination rates compared to the other four populations with only 85 and 79%, respectively, of the seeds germinating. It would be expected that the 6-week vernalisation treatment would overcome any over-ripening requirements. However, these results suggest that this may not be the case.
Flowering time represents a suitable trait with which to test the suitability of mapping populations for QTL analysis, partly because many of the major genes involved in its regulation have already been identiWed. A number of Arabidopsis RI populations have previously been used to study Xowering-time [bib_ref] Development of an AFLP based linkage map of LER, Col and Cvi..., Alonso-Blanco [/bib_ref] [bib_ref] Bay-0 £ Shahdara recombinant inbred line population: a powerful tool for the..., Loudet [/bib_ref] [bib_ref] Quantitative trait locus analysis of growth-related traits in a new Arabidopsis recombinant..., El-Lithy [/bib_ref] [bib_ref] The PHYTOCHROME C photoreceptor gene mediates natural variation in Xowering and growth..., Balasubramanian [/bib_ref]. Understanding the allelic variation at the loci identiWed and the subsequent eVects on the Xowering phenotype is now the major challenge [bib_ref] Analysis of the molecular basis of Xowering time variation in Arabidopsis accessions, Gazzani [/bib_ref] [bib_ref] Variation in the epigenetic silencing of FLC contributes to natural variation in..., Shindo [/bib_ref]. By using markers physically anchored to the Arabidopsis genome we are able to compare the results for Xowering time from our six new populations against the results generated in populations studied previously [bib_ref] Bay-0 £ Shahdara recombinant inbred line population: a powerful tool for the..., Loudet [/bib_ref] [bib_ref] Quantitative trait locus analysis of growth-related traits in a new Arabidopsis recombinant..., El-Lithy [/bib_ref] as displayed in [fig_ref] Figure 3: Physical positions of QTL aVecting Xowering time across 12 populations [/fig_ref]. The positions of seven QTL identiWed in this work coincided with QTL identiWed in earlier studies using diVerent populations: Bay-0 £ Shahdara [bib_ref] Bay-0 £ Shahdara recombinant inbred line population: a powerful tool for the..., Loudet [/bib_ref] , Ler £ Sha ), Ler £ An-1, Ler £ Kas-2, and Ler £ Kondara [bib_ref] New Arabidopsis recombinant inbred line populations genotyped using SNPWave and their use..., El-Lithy [/bib_ref]. These previous studies were also carried out under long day conditions but unlike the work described in this paper the plant material was not exposed to a vernalization treatment. Of the seven QTL found to be in common, the Wrst SG.1 was positioned around 24 Mb on chromosome 1 close to the position of the Xowering locus T-gene at 24.3 Mb and MAF1 at 28.8 Mb, which was detected in the Nd-1 £ Col population . Another QTL found in common between the data sets was identiWed at the top of chromosome 4. This QTL was picked up in Wve of the previous populations and in the CA population in this work. It is possible to speculate that the gene responsible for the phenotype is FRI. [bib_ref] Analysis of the molecular basis of Xowering time variation in Arabidopsis accessions, Gazzani [/bib_ref] showed that Cvi-0 carries an FRI allele with an in-frame translation stop codon within exon 1 and F1 analysis suggested that this allele is non-functional. We found that the allele responsible for increasing Xowering time in the CA population came from Ag-0. On chromosome 5 both NG.2 and SG.4 were positioned around the nga151 marker at 4.7 Mb. This is very close to the Xowering time genes CO at 5.1 Mb, FY at 4.3 Mb and FLC at 3.2 Mb. FLC has been shown to be involved in the vernalisation pathway controlling Xowering time and given that all the material used in this work was vernalised, there is a reduced chance that this is the gene responsible for these QTL. QTL in this region were also found in the Ler £ An-1, Ler £ Kas and Ler £ Kon populations. Large QTL were identiWed in the TJ and WC populations towards the end of chromosome 5 at ca. 20 and 24 Mb. QTL in this region were also detected in the Ler £ Sha ) and the Ler £ Kondara and Ler £ Kas populations [bib_ref] New Arabidopsis recombinant inbred line populations genotyped using SNPWave and their use..., El-Lithy [/bib_ref]. In addition highly signiWcant QTL were also identiWed in the NG and SG populations in this region. Major genes controlling Xowering time in this region of the genome are LFY and the MAF genes 2, 3, 4 and 5.
Of the QTL unique to the new populations, the most sig-niWcant one was picked up on chromosome 2 in the KB population at 11.6 Mb. This QTL accounted for 58% of the phenotypic variation with a LOD value of 13. A possible candidate for this QTL is ELF3, which controls rhythmic circadian outputs under constant light conditions and is found at 11 Mb.
In the SG population, SG.1 (on chromosome 1) spans the region containing FT and SG.3 (on chromosome 4) spans the region containing TSF, which is a homologue of FT. SG.3 was only detected when the marker selected as a cofactor for SG.4 was applied in the MQM mapping. This suggests that SG.4, which maps in the region of CO, was masking the eVect of SG.3. It should be noted, however, that the genetic map for chromosome 1 in SG did not provide complete coverage to the end of the physical chromosome, where another gene involved in the control of Xowering, MAF1, is located.
The experiments described in this work were carried out under a deWned set of growing conditions: long days following 6-weeks vernalisation. It might therefore be expected that allelic diVerences in genes involved in the vernalisation pathway would not be detected. This appears largely to be the case, with one exception, CA.1, for which FRI is a possible candidate. It should be noted that for all of the QTL, any clearer identiWcation of the candidate genes requires further mapping beyond the preliminary results that we have presented. The involvement of loci on chromosome 5 in the control of Xowering seems to be extensive, with 15 of the 41 QTL identiWed in the 12 populations mapping to this chromosome [fig_ref] Figure 3: Physical positions of QTL aVecting Xowering time across 12 populations [/fig_ref]. For four of our six populations, two QTL were detected on chromosome 5. It is impossible to assign candidate genes to these all of these QTL, but it does suggest that this chromosome deserves special attention.
In this work, we have developed and characterised six new Arabidopsis thaliana recombinant inbred populations and shown the utility of these populations for the study of the complex trait Xowering time. By using the physical marker positions we aligned the linkage maps of our populations with those of six existing populations and were therefore able to compare QTL for Xowering time across 12 populations. The ability to conduct such comparative work between immortal populations will provide the research community with a powerful tool to aid in the understanding of the global genetic architecture of a broad range of complex traits. [bib_ref] Bay-0 £ Shahdara recombinant inbred line population: a powerful tool for the..., Loudet [/bib_ref] , [bib_ref] Quantitative trait locus analysis of growth-related traits in a new Arabidopsis recombinant..., El-Lithy [/bib_ref] and
[fig] Figure 1: Allele segregation ratios for the six populations. The female alleles are expressed as a percentage of the male and female alleles combined. The sig-niWcance threshold P < 0.001 (i.e. between 38 and 62%) is indicated by the horizontal dashed lines [/fig]
[fig] Figure 2: QTL for Xowering time for 6 RIL Arabidopsis populations. The position of the QTL are represented as 2 LOD intervals where suYciently large, solid lines, otherwise the 5% signiWcance LOD threshold is used, dashed lines. The position of peak LOD scores are indicated by closed circles. QTL detected by interval mapping are indicated in red, while QTL detected only by CIM mapping are indicated in blue. Upward arrows indicate the maternal parent as the source of the increasing allele, downward arrows indicate the paternal parent as the source of the increasing allele [/fig]
[fig] Figure 3: Physical positions of QTL aVecting Xowering time across 12 populations. QTL identiWed in the populations reported here are represented by Wlled triangle. QTL identiWed in the populations studied by [/fig]
[table] Table 2: The 37 newly developed SSR and InDel markers [/table]
[table] Table 4: QTL aVecting Xowering time in the six populations and heritability in the individual populations [/table]
|
Impact of a theory-informed and user-centered stroke information campaign on the public’s behaviors, attitudes, and knowledge when facing acute stroke: a controlled before-and-after study
Background: Public awareness of stroke symptoms is a key factor to ensure access to reperfusion strategies in due time. We designed and launched a regional theory-informed and user-centered information campaign and assessed its impact on emergency medical services (EMS) calls for stroke suspicion, time-to-call, and public attitudes and awareness concerning stroke.Methods: A controlled before-and-after study was conducted during 3 sequential time-periods in 2 separate counties. Key messages of the campaign were underpinned by stroke representations and the theory of planned behavior, and focused on recognition of stroke warning signs and the need to call EMS urgently. The campaign included posters, leaflets, adverts and films displayed in bus and subway stations, internet, social networks, and local radio. Outcome measures on behavior, attitudes, and knowledge were assessed before the launch of the campaign, at 3 months, and 12 months.Results: The number of EMS calls for stroke suspicion increased by 21% at 12 months in the intervention county and this change was significantly different to that observed in the control county (p = 0.02). No significant changes were observed regarding self-reported attitudes in case of stroke. An 8% significant increase in recognizing at least 2 stroke warning signs was observed in the intervention county (p = 0.04) at 3 months, while it did not change significantly in the control county (p = 0.6). However, there was no significant difference in warning sign recognition between both counties (p = 0.16). Conclusion: The campaign significantly improved public's behavior of calling EMS, although stroke knowledge was not improved as much as expected. Repeating these campaigns over time might further help improve timeliness and access to reperfusion strategies.Trial registration: Clinical Trial Registration-URL: http://www.clinicaltrials.gov. Unique identifier: NCT02846363.
# Background
Knowledge concerning stroke symptoms is poor in the general population [bib_ref] Perception of stroke symptoms and utilization of emergency medical services, Hawkes [/bib_ref] [bib_ref] Stroke knowledge and awareness: an integrative review of the evidence, Jones [/bib_ref] [bib_ref] Temporal trends in public awareness of stroke: warning signs, risk factors, and..., Kleindorfer [/bib_ref]. Patients and relatives thus encounter difficulties in symptom recognition, and response is often inappropriate. While it has been demonstrated that calling emergency medical services (EMS) immediately constitutes the best choice in case of stroke, the majority of patients delay their contact with medical services by going directly to emergency departments (ED) or general practitioners [bib_ref] Translation of the 'time is brain' concept into clinical practice: focus on..., Ragoschke-Schumm [/bib_ref] [bib_ref] Patient behaviour at the time of stroke onset: a cross-sectional survey of..., Mellon [/bib_ref]. These increased delays are incompatible with access to an effective stroke treatment and increase stroke morbidity and mortality [bib_ref] Prehospital stroke care: potential, pitfalls, and future, Mazighi [/bib_ref]. In a previous cohort study in the Rhône county (France), less than 1 patient in 2 called EMS in case of stroke symptoms, 80% were managed in ED with a median time between symptom onset and admission over 2 h, and 8% were thrombolyzed [bib_ref] Effectiveness of thrombolysis in the Rhône region, France: a prospective population-based study, Porthault-Chatard [/bib_ref]. With the aim of increasing access to thrombolysis, a number of actions targeting professionals and care organization have been carried out [bib_ref] Improving access to thrombolysis and Inhospital management times in ischemic stroke: a..., Haesebaert [/bib_ref]. Targeting the general population however, would also help improve stroke management by decreasing time from symptom onset to first medical contact and increase EMS calls.
Information campaigns may positively impact stroke knowledge, but they often lack theoretical framework, and their impact on behaviors remains low [bib_ref] Systematic review of mass media interventions designed to improve public recognition of..., Lecouturier [/bib_ref] [bib_ref] Stroke warning campaigns: delivering better patient outcomes? A systematic review, Mellon [/bib_ref] [bib_ref] Development of complex interventions in stroke care: a systematic review, Redfern [/bib_ref]. To develop a more effective campaign that fits the needs of the targeted population, we first carried out a study to assess the population's representations of stroke [bib_ref] Why patients delay their first contact with health services after stroke? A..., Bonniec [/bib_ref]. Based on the results of this study and informed by the theory of planned behavior (TPB) [bib_ref] The theory of planned behavior, Ajzen [/bib_ref] , we developed the ReACT campaign which aimed to improve the skills of the general population in recognizing the first signs of acute stroke and to understand the need to promptly call EMS. The main objective of the present study was to assess the impact of the campaign on the number of calls to EMS for suspected stroke. The secondary objectives were to assess the effect of the campaign on time from symptom onset to EMS call, and the public's knowledge of stroke in the intervention county compared to a control county.
# Methods
## Study design
We conducted a quasi-experimental controlled beforeand-after study, comparing the Rhône county where the campaign was implemented (intervention county) to the Isère county (control county), a nearby region with no local specific campaign during the study period. The ReACT campaign was launched in the Rhône county from October 29th 2014 (World Stroke Day), for a 2month period. The evaluation process occurred during three 2-month evaluation periods: before the launch of the campaign on World Stroke Day (T0); at 3 months (T1; short term); and at 12 months (T2; long term). Reporting of the results follows the TREND Statement [bib_ref] Improving the reporting quality of nonrandomized evaluations of behavioral and public health..., Jarlais [/bib_ref]. No ethics approval was deemed necessary according to French regulation to conduct the study which was classified as non-interventional (3°of Article L.1121-1 of the French Public Health Code).
## Content of the campaign
The ReACT campaign was an information campaign which aimed to improve stroke recognition in the general population and prompt call to EMS. The content and the form of the messages were designed by a multidisciplinary team involving stroke unit (SU) neurologists, EMS physicians, public health researchers, pharmacists, psychologists, communication professionals, and representatives of patients' associations, using a user-centered approach. The content of the campaign targeted three dimensions: stroke symptoms, urgency of the situation and need to call EMS. Information on stroke symptom recognition targeted the 3 FAST-symptoms (Face dropping, Arm weakness, Speech disturbance). The urgency of the situation and the need to call EMS were indicated in all campaign media. The content of the campaign was informed by the TPB [bib_ref] The theory of planned behavior, Ajzen [/bib_ref] and based on public stroke representations [bib_ref] Why patients delay their first contact with health services after stroke? A..., Bonniec [/bib_ref] highlighted in a previous qualitative study conducted by our team which results showed a strong feeling of fatalism and a low perceived selfefficacy to act appropriately. It also pointed out the importance of lay knowledge and shared experiences to gain understanding concerning stroke. Based on these results, we designed the campaign presenting stroke symptoms in everyday-life using characters and situations with which everyone could identify. In line with TPB, the content of the campaign targeted social norms and perceived behavioral control, using a vicarious model. To target subjective norms, we presented stroke victims as well as non-professional rescuers, as both men and women, with various ages (from grandchildren to grandfather). To improve perceived behavioral control and self-efficacy, our campaign slogan was 'Your call can save life' and messages displayed mentioned that everyone can effectively act by just a phone call. All media were developed using 3 visuals: one young student female presenting sudden arm weakness, one middle aged working male presenting sudden speech disturbance, and one grand-father presenting sudden face weakness. In each, one relative effectively acted by immediately calling EMS which enabled the patient to be saved. The main diffusion media were posters, leaflets, and videos on a dedicated internet website. The leaflet included information on stroke risk factors, symptoms, acute management, and treatments at acute and chronic phase. We assessed its understandability, content and form in a random sample of patients identified through several general practitioners. The messages in the posters focused on self-efficacy reinforcement ("he/she saves his/her life", "he/she just call the 15" -15 being the French equivalent for 911-"Your phone call can save lives"). In the movies, the same situation as presented in the posters were played, and the listeners could choose the end of the film by choosing whether they wanted to wait and see or immediately call EMS, such as they could understand benefit in rapid stroke management at acute phase. All campaign materials are provided in Appendix 1.
## Conduct of the campaign
A communication plan was designed specifically for our stroke campaign by communication professionals from a communication school. The campaign was led by the steering committee in partnership with the communication school. Several service providers were responsible for the distribution of the materials (radio campaign, poster campaign, website design). The diffusion strategy consisted in one public event on World Stroke Day in the city-center of Lyon, the biggest city of the county, followed by a 2month multi-media campaign. The event proposed a press conference and an information booth animated by representatives of different health professionals involved in stroke management as well as stroke patients. A free riskfactor screening by a prevention nurse was also proposed in a bus, which included a brief individual interview, information, measure of arterial blood pressure, and capillary blood testing for glucose and cholesterol. On the street, actors performed short interactive theatrical scenes to engage people in discussing the response to adopt in case of stroke. The multimedia campaign following the event was composed of a poster campaign on 233 subway and bus stations of the Rhône county, a one-week advert broadcasting on the two main local radios, an information website, and a social network group on Facebook. The posters and leaflets of the campaign were sent to all the 3600 ambulatory physicians (general practitioners and all medical specialties) of the county and to 60 pharmacists. The posters were also displayed in the four university hospitals of the Rhône county.
## Patient involvement
Two patients' representatives, from the local stroke patient association, were involved in the research, since the first step of the study, just after grant approval. They were involved in the design of the intervention, in the choice of the content of the information leaflet and campaign, to include their experience and preferences; they also participated in organizing the event on world stroke day. They critically revised the study questionnaires and participated in the spreading of the questionnaires for recruitment of participants.
## Population and geographic counties
The target population for the intervention was the general population of the Rhône county (1.8 million inhabitants) covered by two SU and one EMS organization (SAMU69). The control group was the nearby Isère county (1.2 million inhabitants) covered by two SU and one EMS organization (SAMU38). The Rhône/Isère populations had similar socio-demographic characteristics including 51.8%/50.8% of women, 58.2%/56.6% of population aged ≥65 years old, and 12.5%/11.2% unemployment rates. Citizens living in these two counties and aged above 18 years old were eligible for the survey on stroke knowledge and attitudes. No exclusion criteria applied. However those who could not read or write in French were not able to participate since the study questionnaire was only available in French.
## Outcome measures and data collection
The primary outcome, EMS calls (behaviors), was extracted from the EMS call center databases for the three periods. All calls concerning an adult aged 18 years or older, living in the studied county, and presenting FAST symptoms (Facial drooping, Arm or leg weakness, Speech difficulties) or sudden onset of neurological signs with no other immediate explanation, were included. Additionally, patients identified by the emergency dispatcher as stroke or suspected transient ischemic attack were also included. Patients' socio-demographic data, clinical signs, hour of symptom onset, hour of EMS call, and management (triage and dispatch) were collected.
Secondary outcomes related to public attitudes and knowledge were assessed at T0 and T1 using a 3-section ad-hoc questionnaire on knowledge (stroke definition, risk factors, symptoms), attitudes (urgency of the situation, response to adopt), and socio-demographic data. The questionnaire was developed by the steering committee based in available literature [bib_ref] Systematic review of mass media interventions designed to improve public recognition of..., Lecouturier [/bib_ref] [bib_ref] Stroke warning campaigns: delivering better patient outcomes? A systematic review, Mellon [/bib_ref] [bib_ref] Development of complex interventions in stroke care: a systematic review, Redfern [/bib_ref]. Its acceptability and understandability were assessed on a sample of lay users. Questions included both multiple choice questions and open-ended questions. Questions concerning stroke knowledge were open-ended; participants were asked to define stroke and which symptoms were evocative of acute stroke. To assess attitudes and the appropriateness of the response, participants had to choose between five proposals (calling EMS, calling firefighters, going to the emergency unit, visiting a general practitioner, calling a relative) and between responding immediately or waiting to see if symptoms disappeared. The questionnaire was sent by postal and electronic forms in the Rhône and Isère counties to the beneficiaries of a private health insurance, members of a neurological disease foundation, and the union for small and medium businesses (Confédération des petites et moyennes entreprises). We also used informal spreading and snowballing through Facebook and the internet websites of the aforementioned partners. In line with the French regulation for health research that applied, no consent for participation in the study was required for the type of our study.
# Data analysis
Based on the data obtained from the EMS registries, the expected number of calls for stroke suspicions was 160 per 2-month period in the Isère county and 200 per 2month period in the Rhône county. This ensured a statistical power of 95% to demonstrate a 10% absolute difference in the increase in the number of calls between the two counties over the study period. Quantitative variables were described by their mean and standard deviation or median and first and third quartiles (Q1-Q3), depending on their distribution. Qualitative variables were described using frequency and percentage.
The evolution of the main outcome (number of EMS calls) over the 3 periods was compared between the intervention and control counties using a Poisson regression analysis including as dependent variable number of EMS call, and as independent variables the county (Rhône / Isère) and the period (T0, T1 and T2). The time from symptom onset to EMS call was analyzed using a mixed linear regression model. The model included as dependent variable time from symptom onset to EMS call, and as independent variables age (< 65/ Changes in public knowledge and attitudes between T0 and T1 were compared using multivariate mixed regression models with a random effect on the intercept and on the county. We estimated the impact of the campaign on the following outcomes, included as dependent variables in the models (one model per dependent variable): knowing at least two FAST symptoms (yes/no), knowledge of the emergency of the situation (yes/no), and knowledge of the need to call EMS (yes/no). Independent variables included in each model were gender (female/male), age (age under 45 years old/[45-64]/≥65), educational level (low (less than graduate)/medium (graduate to 4 years post graduate)/high (≥5 years post graduate)), and history of stroke (stroke survivor or relative of a stroke survivor/no stroke history). P-values were two-tailed and 0.05 was considered as significant, all confidence intervals were calculated at 95% (95%CI). Analyses were conducted using SAS software 9.3 for Windows (SAS Institute Inc., Cary, NC, USA).
# Results
## Impact on behaviors: ems calls
During the 3 periods, 217,476 EMS calls were recorded in the Rhône county, of which 707 met the inclusion criteria and were included (n T0 = 214, n T1 = 234, n T3 = 259). In Isère, 74,526 EMS calls were recorded, 519 met the inclusion criteria and were included (n T0 = 186, n T1 = 174, n T3 = 159; Appendix 2).
Calls and patients' characteristics are displayed in [fig_ref] Table 1: Characteristics of EMS calls for stroke suspicion according to county and time... [/fig_ref]. Pooling the three time periods and the two [fig_ref] Table 2: Regression analysis for evolution of number of calls and time between symptom... [/fig_ref].
## Impact on public stroke knowledge and attitudes
Among the 1468 questionnaires received, 1421 could be included, 1183 in the intervention county and 238 in the control county (Appendix 3). Socio-demographic characteristics of respondents revealed differences between the 2 counties and the 2 time-periods regarding age, being a relative of a stroke survivor, and educational level [fig_ref] Table 3: Characteristics of questionnaire respondents according to county and time period [/fig_ref]. In the Rhône and Isère counties, respectively, median age was 48 years at T0 and 36 years at T1, and 57 years at T0 and 49 years at T1. Proportion of women among respondents was in the Rhône and Isère counties, respectively, 71.26% at T0 and 74.22% at T1, and 70.97% at T0 and 75.86% at T1. In the intervention county (Rhône), knowledge of at least 2 stroke symptoms significantly increased from 40.5 to 48.7% between T0 and T1 (p = 0.038, [fig_ref] Table 4: Stroke knowledge according to county and time periodAge of stroke victims a... [/fig_ref]. In this county, about 1 in 2 respondents knew a treatment existed but less than 10% mentioned thrombolysis or described such an acute phase treatment, and this was not improved after the campaign (p = 0.682). Knowledge that people of all ages could be stroke victims significantly improved after the campaign (20.1% at T0 to 38.6% at T1, p < 0.001).
In the control county (Isère), no significant changes happened between the 2 periods [fig_ref] Table 4: Stroke knowledge according to county and time periodAge of stroke victims a... [/fig_ref]. At T0 and T1, respectively 46 and 49% of respondents knew at least 2 FAST stroke symptoms. At T0 and T1, about 98% of respondents mentioned that stroke was an emergency, and 79% would call EMS (p > 0.05). Similarly to the intervention county, 48% declared that a treatment existed but very few described the existence of an acute phase treatment such as thrombolysis, with no significant difference between T0 and T1.
Multivariate analyses on knowledge of at least 2 stroke warning signs, knowledge of the emergency of the situation, and knowledge of the need to call EMS, revealed no significant difference in the knowledge and attitudes between the intervention and the control group, although there was a trend for a higher improvement in the intervention county regarding knowledge of at least 2 stroke warning signs
# Discussion
After implementation of a local campaign on stroke, the number of EMS calls for stroke suspicion significantly increased at 12 months in the intervention county compared to the control county. Nevertheless, no significant improvement in time from symptom onset to EMS call was observed, and the increase in symptom knowledge was not significantly different between intervention and control.
The present theory-informed campaign was associated to a significant improvement in behaviors, indicating that the messages conveyed might have improved individuals' skills to call EMS when faced with stroke. In a review on stroke campaign effects, Mellon et al. [bib_ref] Stroke warning campaigns: delivering better patient outcomes? A systematic review, Mellon [/bib_ref] showed that the strongest results on behaviors were obtained after repeated and sustained campaigns [bib_ref] Les campagnes d'information multimédia prolongées et répétées de la population sur les..., Nicoli [/bib_ref] [bib_ref] Sustained impact of UK FAST-test public education on response to stroke: a..., Wolters [/bib_ref] or campaigns also targeting health professionals [bib_ref] Sustained benefit of a community and professional intervention to increase acute stroke..., Morgenstern [/bib_ref]. The lack of repetition over time may explain why the present campaign failed to increase other outcomes such as time from symptom onset to EMS call. The first message of the campaign was to call EMS, hence repetition might have facilitated integration of the other messages.
In the existing literature, other national or local experiments have shown improvements in the population's knowledge [bib_ref] Comparison of stroke warning sign campaigns in Australia, England, and Canada, Trobbiani [/bib_ref] [bib_ref] Addressing stroke signs and symptoms through public education: the stroke heroes act..., Wall [/bib_ref] [bib_ref] Trends in community knowledge of the warning signs and risk factors for..., Schneider [/bib_ref] [bib_ref] Does stroke health promotion increase awareness of appropriate behavioural response? Impact of..., Hickey [/bib_ref] and/or behaviors (i.e. shorter patient delay, higher EMS or ambulance use, higher thrombolysis rate) following information campaigns [bib_ref] Stroke warning campaigns: delivering better patient outcomes? A systematic review, Mellon [/bib_ref] [bib_ref] Sustained impact of UK FAST-test public education on response to stroke: a..., Wolters [/bib_ref]. In the present study, knowledge of stroke warning signs significantly increased in the Rhône county after the campaign, however it remained low. The absolute 8% increase observed herein is similar to that observed in Canada and Australia, but the percentage of people knowing at least 2 stroke warning signs was substantially higher in the latter studies than in the current one [bib_ref] Comparison of stroke warning sign campaigns in Australia, England, and Canada, Trobbiani [/bib_ref]. Attitudes in case of acute stroke, i.e. calling EMS immediately, did not increase in the present sample but was already very high at baseline before the campaign. However, the gap between knowledge, intention, and behavior still needs to be bridged. Another interesting finding is that knowledge concerning thrombolysis appears to be highly confidential among respondents, which is in line with what has been observed in other countries [bib_ref] Does stroke health promotion increase awareness of appropriate behavioural response? Impact of..., Hickey [/bib_ref]. The spreading and efficacy of thrombolysis, and more recently endovascular thrombectomy, represent an opportunity to better communicate on the therapeutic options for stroke management. Messages of future stroke campaign should focus on treatment and emphasize the benefits of receiving them rapidly. We observed that women, people with a higher level of education, and stroke survivors or their relatives had significantly better knowledge of stroke. However, no difference was observed for respondents with risk factors. These results point out the necessity to complete mass campaigns with actions targeting specific subgroups, such as people at risk or people with low socio-educational level, to reduce inequalities. Stroke information programs with actions targeting specific groups have shown positive effects in disadvantaged counties [bib_ref] Long-term learning of stroke knowledge among children in a high-risk community, Williams [/bib_ref]. Campaigns must offer various supports specifically tailored to different levels of health literacy and use a variety of media to reach all audiences.
There are some limitations to this work. Missing data on the time of symptom onset limited the possibility to analyze the effect of the campaign on this parameter. This limitation is due to the data collection process of the EMS call registry which is integrated into usual care. This reflects a need for improving EMS dispatchers' practices, as symptom onset hour is a key factor for acute stroke management. Another limitation concerns a potential selection bias inherent to the sampling method that led to differences in the demographic characteristics of the respondents from the two counties compared with census data and between the 2 periods within each county. We took this limit into account by adjusting the analyses on potential confounding factors. The study is also limited to assessing the influence of the intervention on the intention to call EMS. However, the study did not evaluate the relationship between the outcome variables and the TPB, and the influence of the intervention on the individual constructs of the TPB as applied to this study population.. Individual repeated measures would be interesting to in-depth changes induced by the campaign at individual level. This study presents also several strengths. Indeed, few information campaigns were evaluated with a sufficient level of evidence and a control group [bib_ref] Systematic review of mass media interventions designed to improve public recognition of..., Lecouturier [/bib_ref] [bib_ref] Does stroke health promotion increase awareness of appropriate behavioural response? Impact of..., Hickey [/bib_ref]. Mellon et al. pointed out a large heterogeneity among interventions, evaluation designs, and effect sizes of the interventions between the campaigns that have been evaluated [bib_ref] Stroke warning campaigns: delivering better patient outcomes? A systematic review, Mellon [/bib_ref]. Another strength is the user-centered and collaborative approach, involving representatives of all stake-holders and the use of a theoretical model [bib_ref] The theory of planned behavior, Ajzen [/bib_ref]. The latter helped us encompass barriers such as the lack of perceived control and of perceived benefit to call EMS, previously identified [bib_ref] Why patients delay their first contact with health services after stroke? A..., Bonniec [/bib_ref]. This campaigned could easily be generalized and reproduced in different settings. Moreover, our evaluation measured outcomes related to both behaviors, attitudes, and knowledge with a 12-month perspective. While this study provided a quantitative evaluation of the campaign, associating qualitative approach might be useful to understand how the campaign impact behaviors in the light of the TBP. In France, a national campaign on stroke is launched each year, but there is no structured display with massive broadcasting covering the entire territory, leading to the development of local initiatives [bib_ref] Les campagnes d'information multimédia prolongées et répétées de la population sur les..., Nicoli [/bib_ref]. The literature advocates for campaigns that are broadcasted on television, with important financial support [bib_ref] Comparison of stroke warning sign campaigns in Australia, England, and Canada, Trobbiani [/bib_ref]. It would be relevant to propose in France a more structured, large-scale national program, displayed on mass media with large audiences, and to ensure local proximity back-up with actions targeting specific populations such as high-risk patients or underprivileged populations. This is even more crucial since a recent study showed territorial disparities in stroke incidence and mortality [bib_ref] L'accident vasculaire cérébral en France : patients hospitalisés pour AVC en, Lecoffre [/bib_ref] [bib_ref] Mortalité par accident vasculaire cérébral en France en 2013 et évolutions, Lecoffre [/bib_ref].
# Conclusions
The local stroke information campaign put in place herein increased EMS call behavior and, to a lesser extent, knowledge concerning stroke warning signs. Increase in EMS for stroke suspicion constitutes the first step of the acute phase of stroke care, this might thus lead to increase in access to reperfusion strategies. Our results advocate for the design of more sustained and actively broadcasted campaigns to rise stroke awareness on effective treatments. The revolution in acute stroke management due to the development of mechanical
[table] Table 1: Characteristics of EMS calls for stroke suspicion according to county and time period EMS Emergency Medical Services, y years, SU Stroke Unit, ED Emergency Department, ICU Intensive Care Unit [/table]
[table] Table 2: Regression analysis for evolution of number of calls and time between symptom onset and call to EMS in the two counties [/table]
[table] Table 3: Characteristics of questionnaire respondents according to county and time period [/table]
[table] Table 4: Stroke knowledge according to county and time periodAge of stroke victims a Thrombolysis or description of an agent to dissolve thrombus Results are presented as n(%). P-values after adjustment on age, gender, educational level, and history of stroke [/table]
|
Multifunctional γδ T cells and their receptors for targeted anticancer immunotherapy
[bib_ref] Early human cytomegalovirus replication after transplantation is associated with a decreased relapse..., Elmaagacli [/bib_ref] [bib_ref] Early human cytomegalovirus replication after transplantation is associated with a decreased relapse..., Elmaagacli [/bib_ref] [bib_ref] γδT cells elicited by CMV reactivation after allo-SCT cross-recognize CMV and leukemia, Scheper [/bib_ref] [bib_ref] Gammadelta T cell effector functions: a blend of innate programming and acquired..., Bonneville [/bib_ref] [bib_ref] Shared reactivity of V{delta}2(neg) {gamma}{delta} T cells against cytomegalovirus-infected cells and tumor..., Halary [/bib_ref] [bib_ref] Re-targeting T-cells against cancer by gene-transfer of tumorreactive receptors, Marcu-Malina [/bib_ref] [bib_ref] Redirecting αβ T cells against cancer cells by transfer of a broadly..., Marcu-Malina [/bib_ref] [bib_ref] γ9 and δ2CDR3 domains regulate functional avidity of T cells harboring γ9δ2TCRs, Gründer [/bib_ref]
## Disclosure of potential conflicts of interest
Clinical trials were supported by Miltenyi Biotec. This work was supported by grants of the ZonMW 43400003, VIDI-ZonMW 917.11.337, LSBR 0902, AICR 10-736, KWF UU-2010-4669. and B cells (NTR2463 and NTR3079). [bib_ref] Immune reconstitution and strategies for rebuilding the immune system after haploidentical stem..., Oevermann [/bib_ref] Intriguingly, αβ TCR/CD19-depleted, but usually not CD3/CD19-depleted, grafts reconstitute very rapidly a broad αβ T-cell repertoire (J Kuball, unpublished observations), suggesting a very broad immunoregulatory role for γδ T cells that has recently also been proposed by others. [bib_ref] Six-of-the-best: unique contributions of γδ T cells to immunology, Vantourout [/bib_ref] As an alternative, umbilical cord blood grafts can be used as a third party source of stem cells. These grafts typically contain high percentages of γδ T cells, and we have demonstrated that CMV-and leukemiareactive γδ T cells can also be obtained from such an antigen-naïve repertoire. Importantly, all good manufacturing practice-grade clinical tools for the preparation of enriched stem cells grafts are available.
of congenitally CMV-infected neonates, implying that CD8αα expression by γδ T cells represents a pathophysiologically relevant phenomenon in vivo.
The demonstration that a proportion of γδ T cells can cross-recognize CMV and a broad panel of hematological cancer cells make them particularly attractive for clinical applications, such as adoptive cell transfer-based immunotherapy. In the context of allo-SCT, a situation favoring the reactivity of re-infused cells against CMV and leukemia in the absence of graft vs. host disease (GVHD) might be achieved with stem cell grafts enriched for γδ T cells . To this end, we and others are nowadays conducting clinical trials using stem cell grafts depleted of αβ T cells . Antitumor strategies based on γδ t cells. (A) the cell preparation for an "innate allogenic stem cell transplantation" (allo-sCt) from conventional or third party donors may selectively contain or be enriched for γδ t cells to provide anti-cytomegalovirus (CMV) and antitumor protection in the absence of graft versus host disease (GVhD). (B) In a complementary "autologous engineered transplantation," t cells are isolated from cancer patients, expanded and engineered to express γδ tCrs (optimized by combinatorial-γδtCr-chain exchange) ex vivo. reprogrammed t cells are subsequently re-infused into the patient, where they specifically recognize and kill tumor cells. |
A Generalized Formula for Converting Chi-Square Tests to Effect Sizes for Meta-Analysis
The common formula used for converting a chi-square test into a correlation coefficient for use as an effect size in metaanalysis has a hidden assumption which may be violated in specific instances, leading to an overestimation of the effect size. A corrected formula is provided.
# Introduction
One of the fundamental concepts in systematic and comparative reviews such as meta-analysis is that of the effect size. An effect size is a statistical parameter that can be used to compare, on the same scale, the results of different studies in which a common effect of interest has been measured. In experimental studies, the effect size is a measurement of the response of the subjects to an experimental treatment relative to a control group. All effect size measures are a means of representing the results of primary research in a common way so that the results from individual studies can be compared and evaluated. While a number of alternate metrics have been suggested for measuring effect size, including standardized mean differences and odds ratios, historically, one of the more popular measures of effect has been the correlation coefficient [bib_ref] Meta-analyses of the association between multilocus heterozygosity and fitness, Britten [/bib_ref] [bib_ref] Habitat loss and population decline: A meta-analysis of the patch size effect, Bender [/bib_ref] [bib_ref] Bilateral symmetry and sexual-selection: A metaanalysis, Møller [/bib_ref] [bib_ref] How closely correlated are molecular and quantitative measures of genetic variation? A..., Reed [/bib_ref] [bib_ref] Meta-analysis of sources of variation in fitness costs of plant antiherbivore defenses, Koricheva [/bib_ref] [bib_ref] Interpersonal expectancy effects: The first 345 studies, Rosenthal [/bib_ref]. The correlation coefficient is widely used, easily interpretable, and has the added bonus of being easily determinable from other commonly used statistics such as z-scores, t-tests, F-statistics, and x 2 statistics [bib_ref] Parametric measures of effect size, Rosenthal [/bib_ref]. These conversions can only be performed for single, focused contrasts (e.g., cases with a single degree of freedom), but otherwise follow simple equations. For example, the equation for converting a x 2 into a correlation [bib_ref] Parametric measures of effect size, Rosenthal [/bib_ref] is:
[formula] r~ffi ffiffiffiffiffi ffi x 2 1 ½ n s ,ð1Þ [/formula]
where the x 2 value comes from a two-group contrasts (thus a single degree of freedom) and n is the total number of samples; the sign of the correlation needs to be determined from independent study of the contrast. (x 2 tests with more than one degree of freedom are unfocused omnibus tests, and require a much more complicated procedure for conversion to an effect size; seefor details). Equationhas been used to convert x 2 tests into a correlation coefficient for use as an effect size for 45 years; unfortunately, this equation has an underlying, never-stated assumption which is sometimes violated, particularly for genetics studies: it assumes that the expected values from the x 2 test are equal for both groups.
# Results and discussion
Recall that the x 2 is calculated simply as
[formula] x 2~X O i {E i ð Þ 2 E i ,ð2Þ [/formula]
where O i and E i are the observed and expected counts, respectively, for group i. Equation (1) assumes that the expected values for the two groups are equal, that is, E 1 = E 2 = n/2. For the vast majority of x 2 tests this assumption probably holds and equation (1) is correct. However, when this assumption is not met (e.g., if the expected values are for a phenotypic ratio of 3:1 from a standard Mendelian di-heterozygous dominant-recessive cross), equation (1) is incorrect and will overestimate the effect size (occasionally even producing ''correlations'' greater than one). The problem is specifically found in the denominator of the equation. Rather than the sample size, the denominator should actually be the maximum possible x 2 value obtainable for that sample size and expected values:
[formula] r~ffi ffiffiffiffiffiffiffiffi x 2 1 ½ x 2 max s :ð3Þ [/formula]
Written this way, the logic of the equation can be interpreted as r 2 being equal to the ratio of observed x 2 to the maximum possible x 2 , a not atypical way of expressing the variance explained in a sums-of-squares framework. In the common case where one expects an equal distribution of observations among the two groups (E 1 = E 2 = n/2), this maximum possible value is equal to n (see Methods), making equation (1) correct.
On the other hand, if the expected values for the two groups are not the same, the maximum possible value is larger than n. Specifically, if the ratio of expectations among the two groups is k:1, where k represents the larger group (k$1), the maximum possible x 2 value is equal to nk (see Methods). The general form of the conversion of a x 2 value to r should therefore be r~ffi ffiffiffiffiffi ffi
[formula] x 2 1 ½ nk s :ð4Þ [/formula]
As would be expected, this generalized form simplifies to the commonly used equation when the expected ratio of observations is 1:1 (i.e., k = 1). As a simple example, imagine a genetic cross of two heterozygous individuals for a pair of alleles with a simple dominant-recessive relationship. The expectation is a phenotypic ratio of 3:1, in favor of the dominant phenotype. With 22 total offspring, the expected counts are 16.5 and 5.5, respectively. The actual observation is almost the reverse, with 16 found to have the recessive phenotype and only 6 having the dominant phenotype. The x 2 value for this cross is 26.7 (some additional factor is clearly skewing the expected phenotypic ratio). Using the traditional transformation from equation (1), one would determine an effect size r = 1.10, a value outside the acceptable range of r, impossible to interpret or use in a summary analysis (which generally requires use of Fisher's z-transform). Taking into account the adjusted formula due to the assumption of unequal expectations, we instead find an effect size of r = 0.64, still quite strong, but logically interpretable and meaningful.
It is difficult to determine how often, if ever, equation (1) has been misapplied, because one could not simply go through published meta-analyses looking for x 2 conversions, but would rather have to go back to the original studies used in each metaanalysis to determine how the x 2 test was actually calculated. Fortunately, the assumption of equal expectation among groups likely holds for the majority of x 2 tests. However, those who work in fields or encounter data where this assumption is not true (e.g., heredity experiments), need to be aware of the alternate formulation to use for their specific work, particularly as metaanalysis and other statistics that make use of effect size continue to become more common.
# Methods
Derivation of maximum possible x 2 value for two groups when the expectation is an equal distribution of counts among the groups The x 2 value is calculated as
[formula] x 2~X O i {E i ð Þ 2 E i ,ð5Þ [/formula]
where O i and E i are the observed and expected values for the i th group. In a x 2 test with only two groups with n total observations, if the expectation is an equal distribution of observations among groups, the expected value of each of the two groups will be n/2.
The maximum possible x 2 value will be obtained if one group contains all of the observations and the other group contains zero observations, thus
[formula] x 2 max~n { n = 2 À Á 2 n = 2 z 0{ n = 2 À Á 2 n = 2 n = 2 À Á 2 n = 2 z n = 2 À Á 2 n = 2~n = 2 z n = 2~n :ð6Þ [/formula]
The maximum possible x 2 value for a two group case where the expected distribution of outcomes is evenly distributed among the two groups is simply the total sample size.
Derivation of maximum possible x 2 value for two groups when the expectation is not an equal distribution of counts among the groups If the expectation among the two groups is not equal, we have to make the following change. The expected ratio between the two groups will be k:1, where k is how many times larger one group is expected to be relative to the other (k$1). In this case the expected values are nk/(k+1) and n/(k+1), respectively (Note that if k = 1, we have an expected ratio of 1:1, leading both expectations to equal n/2 and the special case described above). Given these new expectations, the maximum possible x 2 value will be when the group with the ''nk'' expectation (the larger expectation) contains 0 observations and the other group (the smaller expectation) contains all n observations. In this case, the maximum possible x 2 value becomes:
[formula] x 2 max~0 { nk = kz1 ð Þ 2 nk = kz1 ð Þ z n{ n = kz1 ð Þ 2 n = kz1 ð Þ nk = kz1 ð Þ 2 nk = kz1 ð Þ z n kz1 ð Þ = kz1 ð Þ { n = kz1 ð Þ 2 n = kz1 ð Þ nk kz1 ð Þ z nkzn = kz1 ð Þ { n = kz1 ð Þ 2 n = kz1 ð Þ nk kz1 ð Þ z nk = kz1 ð Þ 2 n = kz1 ð Þ~n k kz1 ð Þ z n 2 k 2 . kz1 ð Þ 2 n = kz1 ð Þ nk kz1 ð Þ z nk 2 kz1 ð Þ~n k 2 znk kz1 ð Þ~n k(kz1) kz1 ð Þ~n kð7Þ [/formula]
Again, if the expectation is even, k = 1, and the maximum x 2 value is simply n.
April 2010 | Volume 5 | Issue 4 | e10059
PLoS ONE | www.plosone.org
AcknowledgmentsThanks to M. Lajeunesse and M. Jennions for discussion and comments on early versions of the manuscript. |
Phylogenetic evidence of HIV-1 transmission linkage between two men who have sex with men
Background: In China, an HIV-infected man (complainant; P2) alleged that another man (defendant; P1) had unlawfully infected him with HIV through unprotected homosexual contact in 2018.Methods:We employed epidemiological, serological and phylogenetic analyses to investigate the transmission linkage between two men who have sex with men (MSM). Partial segments of three HIV-1 gene regions (gag, pol, and env) were amplified and sequenced by cloning. Maximum-likelihood (ML) and Bayesian methods were used to determine the direction and estimate the timing of transmission. Local control sequences and database control sequences were also used in the phylogenetic analysis.Results: It indicated that P2 underwent HIV seroconversion after P1 was diagnosed as HIV positive. The time to the most recent common ancestor (tMRCA) estimates consistently showed that P1 most likely became HIV-1 infected at an earlier date than P2. P1 and P2 were infected with the same HIV-1 CRF01_AE subtype according to segments of all three gene regions (gag, pol, and env). All three genetic regions of P1 have been subject to more potential selective forces than those of P2, indicating a longer evolutionary history. Bayesian and ML trees showed similar paraphyleticmonophyletic topologies of gag and env, with the virus from P1 located at the root, which supported a P1-to-P2 transmission direction.Conclusions:Phylogenetic investigations can elucidate HIV transmission linkage and might empower its use in the opposition of the intentional transmission of HIV-1 as a forensic tool.
# Background
Men who have sex with men (MSM) have become the core population of the HIV/AIDS epidemic around the world, including in China. MSM are at increasing risk of HIV-1 infection due to engaging in anal intercourse, having multiple sex partners and showing low rates of condom use. In China, a high prevalence of temporary sexual behavior has been reported among MSM, in which some HIV-infected individuals conceal their infection status, putting their sexual partner at a high risk of infection.
Since the "Florida Dentist case" was reported in the 1990s, phylogenetic analysis has been recurrently used as a forensic tool in HIV transmission investigations, including cases of nosocomial transmission, outbreaks in prisons, father-to-child transmission Open Access, and sexual abuse. However, the forensic use of phylogenetic analysis to investigate the intentional transmission of HIV among MSM is scarce and is almost completely absent in China. Sanger sequencing has been used to obtain a consensus sequence representing the predominant viral strain (more than 20%) in a sample. HIV-1 is characterized by high genetic variability, due mainly to high rates of replication, mutation, and recombination. Transmitted variants may disappear or become minor variants of a viral quasispecies over time after transmission. Next-generation sequencing (NGS), molecular cloning/sequencing, and single-genome amplification (SGA)/sequencing are techniques that have been applied to study viral quasispecies, among which molecular cloning/sequencing is cost effective and easy to operate. Thus, phylogenetic analysis of the sequences of viral quasispecies provides a broad perspective regarding viral diversity and genetic evolution in an individual, which is important in the forensic context.
In the present study, we report a phylogenetic analysis of the sequences of viral quasispecies associated with alleged HIV-1 transmission between two HIV-infected MSM in China. In this case, an HIV-infected man (complainant; designated P2) alleged that another man (defendant; designated P1) had unlawfully infected him with HIV. The present study could promote the opposition of the intentional transmission of HIV-1 in China and beyond and provide molecular evidence for use in legal trials.
# Methods
## Subjects
## P1 (defendant) was diagnosed with hiv-1 infection on
March 29, 2018, when he was 22 years old. P2 (complainant) was 21 years old when he was diagnosed as HIV positive on June 14, 2018, by a municipal HIV-testing institution in Zhejiang Province (Zhoushan Municipal Center for Disease Control and Prevention, ZSCDC). P2 alleged that P1 had unlawfully infected him with HIV-1 via unprotected sexual contact without disclosing his HIV-positive status. Thereafter, a retrospective epidemiological investigation was carried out by the corresponding CDCs responsible for case follow-up. CDC staff administered a questionnaire to obtain demographic and risk behavior information after an HIV diagnosis. The questionnaire included demographic characteristics (age, gender, current residence, registered residence, migration), risk behavior characteristics (location, time range, condom use), characteristics of sexual partners (gender, number, methods of seeking sexual partners), and HIV testing history. Medical records related to HIV testing (e.g., HIV testing during invasive examination) were also collected.
Fresh anticoagulated whole-blood samples were collected from P1 and P2 under the supervision of police at the East Railway Station in Hangzhou on September 20, 2018 and September 21, 2018, respectively. P2 started ART (antiretroviral therapy) on June 28, 2018, while P1 did not receive ART at the time of sampling. Plasma samples from P1 and P2 collected within a month after diagnosis are available in the HIV-positive sample bank of the Zhejiang Provincial Center for Disease Control and Prevention (ZJCDC).
## Amplification, cloning, and sequencing
Genomic DNA was extracted from whole blood using the QIAamp DNA Blood Mini Kit (Qiagen, Valencia, CA, United States) following the manufacturer's instructions. The PCR amplification of segments of the gag (HXB2: 781-1861, encoding part of p17 and p24), pol (HXB2: 2147-3462, encoding the protease gene and the first 300 codons of the reverse transcriptase gene) and env (HXB2: 7002-7541, encoding the V3-V4 region) gene regions of HIV-1, TA cloning, and sequencing were performed as previously described. At least 20 clones per partial HIV-1 gene region were sequenced for each subject at each time point. Positive and negative controls were established through nucleic acid extraction, PCR amplification and molecular cloning. No nucleic acid cross-contamination occurred during the experiment.
# Phylogenetic analysis
The sequences were trimmed, assembled and adjusted with Sequencher v5.0 (Genecodes, Ann Arbor, MI). HIV subtyping using the Recombination Identification Program (RIP) tool (https:// www. hiv. lanl. gov/ conte nt/ seque nce/ RIP/ RIP. html) from the Los Alamos National Laboratory (LANL) HIV Sequence Database (http:// www. hiv. lanl. gov) indicated that the subtype of the query sequences of the segments of the gag, pol, and env gene regions of HIV-1 obtained from P1, P2, and P3 belonged to HIV-1 CRF01_AE. The reference sequences of subtypes A and CRF01_AE were retrieved from the LANL HIV Sequence Database (http:// www. hiv. lanl. gov). For the database controls (DBCs), the sequences with the highest similarity were retrieved from the LANL HIV Sequence Database using the HIV BLAST tool (https:// www. hiv. lanl. gov/ conte nt/ seque nce/ BASIC_ BLAST/ basic_ blast. html). Local control sequences of HIV-1 CRF01_AE circulating in the MSM population were randomly selected. The three final datasets of the segments of the gag, pol, and env gene regions of HIV-1 sequences were designated "dataset_gag", "dataset_pol", and "data-set_env", respectively. Multiple sequence alignments of the three final datasets ("dataset_gag", "dataset_pol", and "dataset_env") were performed using MAFFT v7.222, and the alignment was subsequently manually edited using BioEdit v7.2.5. The best-fit nucleotide substitution models for the three final datasets were identified according to the Bayesian information criterion (BIC) method with three (24 candidate models) or 11 (88 candidate models) substitution schemes in jModelTest v2.1.10.
Maximum-likelihood (ML) phylogenetic trees for the three final datasets were estimated using IQ-TREE v1.6.12under a Hasegawa-Kishino-Yano (HKY)nucleotide substitution model with a gamma-distributed rate variation among sites (HKY + G) model, which was identified as the best fitting model for ML inference by jModelTest v2.1.10. Support for the inferred relationships was evaluated by bootstrap analysis with 1000 replicates. To assess the robustness of the ML tree topologies, we generated posterior probabilities for each node by performing Bayesian Markov Chain Monte Carlo (BMCMC) analyses implemented in MrBayes v3.2.7.
## Reconstruction of time-scaled phylogenies
We applied a BMCMC method to estimate the time to the most recent common ancestor (tMRCA) for P1 and P2 at each time point, as implemented in BEAST v1.10.4, and the BEAGLE v2.1.2 library programwas used for computational enhancement. To allow for variation in molecular evolutionary rates among lineages over time, we used the uncorrelated lognormal relaxed molecular clock with a discretized gamma-distributed general time-reversible substitution model. We ran Bayesian phylogenetic analyses using four coalescent tree priors (i.e., constant size, expansion, exponential growth, and logistic). To ensure adequate mixing of the model parameters, MCMC chains were run for 100 million steps, with sampling every 10,000 steps from the posterior distribution. Convergence was evaluated by calculating the effective sample sizes of the parameters using Tracer v1.7.1. All parameters had an effective sample size > 200, which is indicative of sufficient sampling.
# Selective pressure analysis
Codon sites under selective pressure were analyzed in the segments of the gag, pol, and env gene regions of the HIV-1 sequences of P1 and P2 by using the HyPhy packagehosted on the Datamonkey web server. Positively selected sites were detected using the singlelikelihood ancestor counting (SLAC), mixed-effects model of evolution (MEME), fixed-effects likelihood (FEL), and relaxed-effects likelihood (REL)methods with statistical significance set at p < 0.1 (SLAC, MEME, and FEL) and a Bayes factor (BF) cutoff value of 50 (REL).
# Statistical analysis
Statistical analyses were conducted with SPSS v19.0 software (IBM, Armonk, NY). A P value < 0.05 was considered statistically significant.
# Results
## Epidemiological information and serological testing
According to the detailed epidemiological investigation, P1 (defendant) was reported to be HIV-infected on March 29, 2018. At the end of April 2018, P2 alleged that P1 had unprotected sexual contact with him without disclosing his HIV-1 positive status. Thereafter, P2 was screened as shown to be HIV-1 positive by ELISA in a preoperative examination at Fuyang People's Hospital of Anhui Province on May 3, 2018. P2 was diagnosed as HIV antibody indeterminate by the ZSCDC , and the western blot results showed only four HIV-specific antibodies (anti-gp160, anti-p66, anti-p24, and anti-p17) on May 8, 2018. P2 was reported as HIV-1 positive, and western blot results identified gp160, gp120, p66, p55, p51, gp41, p31, p24, and p17 on June 14, 2018. Therefore, P2 underwent HIV seroconversion around May 2018, suggesting a recent infection.
Through contact tracing, we found that P2 had engaged in risky contact with P3. P2 alleged that he got to know P3 during the period from October to November 2017 and they had ever engaged in oral sex behaviors. P3 was confirmed as HIV-1 infected by the HIV confirmatory laboratory of ZSCDC on December 28, 2017. However, P3 committed suicide after his diagnosis. It was not possible to draw a subsequent blood sample from P3 to confirm the genetic relationship between P2 and P3. However, we obtained plasma samples from P3 after his diagnosis at the time of sampling on December 28, 2017 from the HIV-positive sample bank of the ZJCDC.
## Transmission linkage and direction
Partial segments of the gag, pol and env gene regions of HIV-1 were amplified, TA cloned and bidirectionally sequenced. The HIV-1 quasispecies obtained from P1, P2, and P3 within the three gene regions of HIV-1 were identified by cloning and sequencing. According to Bayesian reconstruction, the three datasets ("dataset_gag", "data-set_pol", and "dataset_env") of the sequences from P1 and P2 formed a strongly supported transmission cluster (Bayesian posterior probability = 1) within HIV-1 CRF01_AE. In "dataset_gag" and "dataset_env", a subset of sequences from P2 were paraphyletic with respect to those of P1, showing a paraphyletic-monophyletic (PM) topological relationship in which P1 was inferred to be located at the root (Bayesian posterior probability = 1). In "dataset_pol", the sequences from P1 and P2 displayed a monophyletic-monophyletic (MM) topological relationship . Similar topological relationships were also inferred using the ML approach . We inferred that HIV-1 transmission had occurred in a P1-to-P2 direction. The sequences obtained from P3 were clustered in a significantly separate cluster from the sequences obtained from P1 and P2. The phylogenetic analysis conducted in the present study excluded P3 as a source of HIV-1 transmission to P2.
## Transmission time estimation
tMRCA estimation consistently showed that P1 most likely became HIV-1 infected at an earlier date than P2 according to two of the analyzed gene fragments (gag and
# Potential selection analysis
The segments of the gag, pol, and env gene regions of HIV-1 harbored more codon sites that were affected by potential selective forces in P1 than in P2. Notably, P1 also exhibited more codon sites affected by positive and negative selection in the segments of the gag, pol, and env gene regions of HIV-1 than did P2.
# Discussion
In the present study, we employed epidemiological, serological and phylogenetic analyses to investigate a potential case of HIV-1 transmission between two MSM. The defendant (P1) had unprotected sexual contact without disclosing his HIV-positive status with the complainant (P2), which was revealed by an epidemiological investigation . The defendant (P1) was reported to be HIV-positive earlier than the complainant (P2). The complainant (P2) was recently infected, as indicated by the typical phenomenon of seropositive conversion according to the progression of WB bands. As shown in, the tMRCA estimation based on segments of the gag and pol gene regions of HIV-1 consistently showed that P1 most likely became HIV-1 infected at an earlier date than P2. The gag, pol, and env gene regions of HIV-1 were subjected to more potential selective forces in P1 than in P2, indicating a longer evolutionary history of the strain in P1. Furthermore, in the comparison of the local circulating strains and the most related sequences from the HIV database (https:// www. hiv. lanl. gov/ conte nt/ index), the viral sequences from the two individuals exhibited a high level of similarity and were most closely Second half of 2017: P1 got to know P2 and had unprotected sex behaviors 6 to 7 times in Zhejiang.
October 2017: P2 got know to P3 and they had engaged in oral sex behavior.
November 24 th , 2017: P2 was tested as HIV-antibody negative during a pre-operative examination for appendicitis surgery. related to each other. Furthermore, a paraphyletic relationship was observed for the gag and env gene sequences of P1 and P2, suggesting that P1 was the source of infection from the point of view of molecular evidence. In conclusion, by combining epidemiological information and serological testing, the relatedness and transmission direction of the HIV-1 strains of the subjects were proven by phylogenetic analysis, supporting the epidemiologically identified linkage in this case. Phylogenetic analysis results should be interpreted with caution. Because the complete sampling of all relevant individuals belonging to the same transmission event is not feasible, it cannot completely exclude the possibility of underlying intermediates. This situation is worth considering because MSM in China display a high level of sexual activity, including exhibiting multiple sex partners, especially casual sexual partners. Phylogenetics can be used to exonerate individuals if the sequences of defendants cluster significantly separately from those of complainants. However, the longer the time since infection, the greater the diversity separating viruses infecting linked individuals will be. In the present study, P3 was suspected to be a potential source of infection by partner tracking. However, P3 showed clustering in a significantly separate cluster in the phylogenetic trees for all three genetic regions; thus, it is logical to exonerate P3. This finding is consistent with a supplementary epidemiological investigation indicating that P2 only engaged in oral sex with P3 without any anal sex behaviors according to the self-report of P2. The direction of transmission is an important and challenging issue in the context of forensic HIV transmission investigations. A phylogenetic analysis of transmission direction can be performed in the context of multiple clones (or samples) representing patients' viral quasispecies. In this study, the viral quasispecies of P1 exhibited a paraphyletic relationship with those from P2 in the Bayesian phylogenetic trees of the segments of the gag and env gene regions of HIV-1 as well as in the ML tree. The results not only supported a high level of similarity between P1 and P2 but also supported a P1-to-P2 transmission direction. As reported previously, phylogenetics can be used to support the investigation of HIV transmission in the context of other types of evidence. We determined the infection window period for P2 based on serological results, and the result was consistently supported by the tMRCA estimates obtained for P2 using molecular clock models.
Leitner et al.concluded that the accuracy of a reconstructed tree topology was more dependent on the amount of genetic information than on the phylogenetic reconstruction methods used. Most investigations involving HIV cases target env, sometimes combined with the analysis gag or pol. In this study, we analyzed
# Limitations
Due to loss to follow-up and lack of other types of evidence, it is still difficult to confidently estimate the infection window period for P1, although the tMRCA of P1 was estimated using molecular clock models. We failed to amplify the segment of the env gene region of HIV-1 in later samples from P1 and P2 despite many efforts; therefore, we could not estimate the tMRCA from the segment of the env gene region of HIV-1 by the Bayesian inference method. We found that P2's tMRCA estimate was surprisingly consistent with the timing of his HIV seroconversion. A recent study showed that genetic diversity calculated from NGS data enables a more accurate estimation of coalescence times. However, the reliability of such estimations requires further investigation and validation, especially in the context of court cases.
# Conclusions
In summary, the PM phylogeny, root host label, timing and selection analysis results all supported the occurrence of HIV transmission between the two MSM involved in this case. The molecular data strongly supported the epidemiological conclusion that P1 transmitted HIV-1 to P2. |
NOise reduction with DIstribution Corrected (NORDIC) PCA in dMRI with complex-valued parameter-free locally low-rank processing
## Figure s3
The effect of NORDIC and MPPCA are shown on a single slice from a diffusion weighted volume (b=3000 s/mm 2 ) across different resolutions (left three columns) and on FA maps (right three columns) for the same slice extracted for the different reconstructions. The FA maps are obtained after EDDY processing, and the reconstructed images are before EDDY processing.
For the 3 subjects with repeated acquisitions, the effect of NORDIC and MPPCA after EDDY correction is shown in for an axial slice in each subject. In this case, each single "repetition" refers to the pair of separate acquisitions with reversed phase encoding which is used for EPI corrections. TOPUP/EDDY, combines data with opposite phase-encoding directions, improving the SNR by approximately √2 compared to a true single acquisition. . Comparison of NORDIC and MPPCA processing with averaging of repetitive acquisitions to increase SNR. The left three columns are for a single acquisition across 3 different resolutions and the right three columns are after averaging 6, 5 and 3 of the repetitive acquisitions respectively. In each case, EDDY processing was applied.
From , Panel A, NORDIC processing shows improved connection strength compared to MPPCA processing, which is evident in the corpus callosum area. At 1.5mm, the difference between MPPCA and NORDIC processed data is less compared to that at higher resolutions, mainly because the number of second fibers estimated at 1.5mm is somewhat similar between these methods, but the improvement in connection strength (the width of the connections shown in yellow) is visible which is due to the lower orientation dispersion and higher number of third fibers resolved in the NORDIC processing.
In Supplemental a comparison of denoising with the VST, NORDIC and MPPCA approaches is shown. As the SNR decreases the impact of the non-zero iid thermal noise becomes more apparent for the MPCPA processed data as revealed in a loss of q-space contrast and residual high-spatial frequency modulations in Panel B. The VST algorithm confronts the difficulty of working with magnitude data using a two-step approach to resolve the Rician and the spatially varying noise and then . Comparison of connectivity distributions from the probabilistic tractography results for 0.9mm (A), 1.17mm (B), and 1.5mm (C) data, representing connectivity of the entire subject-specific posterior corona radiata . For each, the figures illustrate the standard reconstruction on a single dMRI data set of the given resolution (labeled as "standard"), standard reconstruction performed on dMRI data obtained with averaging of multiple runs to increase SNR (labeled as Avg-standard), and NORDIC and MPPCA reconstructions of the data without any averaging.
uses MPPCA for denoising. Although this approach provides improvements over simply working with magnitude data with Rician noise distribution, it also introduces spatial smoothing which the NORDIC approach does not. . Comparison of the denoised images using VST, NORDIC and MPPCA reconstruction. In Panel A, for b=3000 s/mm 2 , the images are shown for 1.17mm and 0.9mm isotropic resolution. For both VST and MPPCA in Panel A, different residual imaging artifacts are observed, while these are not present in NORDIC. In Panel B, the image difference between two q-space samples with the same b-value are correspondingly shown for VST, NORDIC and MPPCA. For the image difference in Panel B, VST exhibit spatially smoother patches whereas with MPPCA the image difference has more noise relative to both VST and NORDIC. Both VST and MPPCA has lost contrast and features that are preserved with NORDIC. VST in general outperforms MPPCA but not NORDIC. |
Palmar Fasciitis and Polyarthritis Syndrome: A Rare Paraneoplastic Syndrome in a Patient With Prostate Carcinoma
A 73-year-old patient was seen in our hospital for treatment of metastatic adenocarcinoma of the prostate (pT 1a N 0 M 1a R 0 , BRCA-2 gene mutation). Prostatectomy and regional radiotherapy were performed and goserelin, a luteinizing hormone-releasing hormone (LHRH) analog, had been started because of disease progression. Castration-resistant progressive disease developed, and enzalutamide was added. A decrease of the prostate-specific antigen (PSA) level was achieved. Before the start of enzalutamide, the patient developed bilateral pain and stiffness of both hands combined with thickening of the hands. The symptoms progressed rapidly to bilateral flexion and extension contractures. The patient became unable to tie his shoelaces and had to use adjusted cutlery to eat. Consultation of the rheumatologist, X-rays, ultrasound and palmar skin biopsy of the hands were performed. The clinical picture resembles descriptions of "palmar fasciitis and polyarthritis syndrome" (PFPAS), a rare paraneoplastic syndrome. Positive effects of immunosuppressive medication have been reported in some cases. In our patient, treatment with oral prednisone (30 mg daily) showed no effect, therefore treatment was switched to methylprednisone pulses and methotrexate. PFPAS is an uncommon paraneoplastic syndrome characterized by rapid onset of bilateral arthritis of the hands, fasciitis of the palms, progressive stiffness and contractures. The scarcity of knowledge about PFPAS makes it difficult to recognize it at an early stage. As a paraneoplastic syndrome, it has been linked to various malignancies. Thus far, PFPAS has been described in only two other cases of prostate cancer. AGB, RK, MCG and HJB contributed to the design, AGB and Articles © The authors | Journal compilation © J Med Cases and Elmer Press Inc™ | www.journalmc.org 270 PFPAS J Med Cases. 2020;11(9):267-270HJB contributed to the writing and the critical revision of the manuscript. AGB, RK, MCG and HJB gave final approval before submission of the manuscript.
# Introduction
Palmar fasciitis and polyarthritis syndrome (PFPAS) is an uncommon paraneoplastic syndrome. It is characterized by rapid onset of bilateral arthritis of the hands and fasciitis of the palms. PFPAS causes diffuse painful swollen hands and is characteristically recognized by progressive stiffness and contractures. These progressive flexion contractures can lead to a loss of hand function.
PFPAS is mostly described as a paraneoplastic syndrome in patients with ovarian carcinoma. Thus far, PFPAS has been described in only two other cases of prostate cancer.
In this case report we describe a patient with progressive prostate cancer who developed PFPAS during treatment of his disease. We summarize the clinical characteristics and possible therapies for PFPAS.
## Case report
A 73-year-old male presented for the treatment of an adenocarcinoma of the prostate with a solitary metastasis in one inguinal lymph gland (pT 1a N 0 M 1a R 0 , BRCA-2 gene mutation). A robot-assisted prostatectomy was performed followed by radiotherapy on the inguinal regions. Despite this therapy, the prostate cancer proved recurrence for which goserelin was started. One year after the prostatectomy progressive disease was noted with new retroperitoneal tumor depositions. The prostate-specific antigen (PSA) level was low at that moment, 3.8 µg/L. Enzalutamide (androgen receptor antagonist) was added to the treatment with continuation of goserelin. A decrease of PSA was achieved, 0.36 µg/L.
In the same period, just before the start of enzalutamide, the patient developed symptoms of bilateral pain and stiffness of his hands. It started in both hands at the same moment. He also experienced thickening of the dorsal skin of his hands.
## Pfpas
J Med Cases. 2020;11(9):267-270
The symptoms progressed rapidly to bilateral flexion contractures and the impossibility to fully bend or extend his fingers.
The patient could not tie his shoelaces and had to use adjusted cutlery to eat. Physical examination showed diffuse thickening of all fingers, mostly around the metacarpophalangeal (MCP) joints. Bilateral, nodules were palpable on the palmar side of the hands. Flexion and extension contractures were present in both hands. Scleroderma from fingertips to MCP joints was seen, without other features of systemic sclerosis (no Raynaud's phenomenon, telangiectasia, calcinosis or signs of organ involvement). There were no signs of arthritis in the other joints. X-ray of his hands showed a slight osteoarthritis, but no other signs of ossal pathology. Ultrasound of the hands showed normal tendon sheaths, no sign of Dupuytren's contracture and no hyperemia. Treatment with prednisone (30 mg/day) was started without effect on the symptoms.
The patient was referred for a second opinion to an academic hospital (University Medical Center Utrecht). Physical examination also revealed a positive groove sign in both arms, left more than right. This is an indentation of the skin above the superficial veins, when the hand is elevated. This sign can be seen with fasciitis. Biopsies of one finger and the palm for pathological research showed fibrosis with a proliferation of fibroblasts in the dermis and subcutis. Intracellular edema was observed in the reticular dermis. In deeper layers of the dermis, an infiltrate with mostly lymphocytes was seen around blood vessels. The findings were not typical for systemic sclerosis but did correspond with the clinical diagnosis of PFPAS.
The diagnosis of PFPAS was confirmed (based on both clinical and pathological findings). Therapy with three methylprednisone pulses of 1,000 mg in combination with methotrexate (MTX) was started. The complaints improved to slightly less swelling and pain of the hands, but the function was still impaired.
Ten months later, therapy with goserelin and enzalutamide was switched to pembrolizumab because of further progression of the disease. Radiotherapy was started because of a new bone metastasis. Meanwhile the MTX was stopped seeing the combination of MTX and radiotherapy can increase the risk of soft tissue necrosis or osteonecrosis. Pembrolizumab was switched to lutetium-prostate-specific membrane antigen (PSMA) radionuclide therapy with a good effect on the PSA-level (drop from 10.7 to 0.68 µg/L). At the same time, prednisone and MTX were restarted and the complaints of PFPAS improved to less swelling of the hands and slightly more functional capacity.
# Discussion
The first paper about PFPAS was published in 1982 by Medsger et al reporting similar symptoms of arthritis and fasciitis in women with ovarian tumors. PFPAS causes diffuse painful swelling of the hands and is characteristically recognized by progressive stiffness and contractures. These progressive flexion contractures can lead to a loss of hand function. The bilateral swelling of the hands can be accompanied by nodules of the flexor tendons in an advanced stage. These findings and induration of the subcutaneous tissue can be referred to as "woody hands" which describes the hardness and the limited function of the hands. A "groove sign", as was also seen in our case, has been described in other cases with PFPAS. This "groove sign" is more often present in eosinophilic fasciitis, a disease which is also characterized by fibrosis and inflammation of the fascia. However, in eosinophilic fasciitis the hands are not affected. In PFPAS, arthritis of other joints can also occur, the shoulders being most often involved.
Like all paraneoplastic syndromes, PFPAS is a cancer-associated disease without being directly caused by the primary tumor or its metastases. Paraneoplastic syndromes are mediated by a trigger from the immune system against tumor cells or the release of humoral factors such as hormones or cytokines. The exact pathway of development of PFPAS remains unclear. Typically, in histologic findings fibroblast proliferation is seen which suggests a role of fibroblast activity stimulators, but further research is needed to confirm that hypothesis. Another hypothesis is the influence of hormones in the pathogenesis of PFPAS. Systematic review by Manger et al showed that nearly half of the patients with PFPAS had a malignancy of the reproductive organs. Especially ovarian adenocarcinoma is often seen and described as the most frequent underlying cause of paraneoplastic PFPAS in female patients. This hypothesis could also account for our patient as he had a mutation associated with reproductive organs and hormones: a BRCA-2 gene mutation.
Besides ovarian cancer, various malignancies have been described as underlying cause of PFPAS. In a systematic review done in 2014, 100 cases of PFPAS were described and systematically reviewed. Prostate cancer accounted for only two of these cases and after this review no other patients were described with prostate cancer as the underlying cause of PF-PAS. Our patient therefore counts for the third patient described with PFPAS associated with prostate cancer.
Treatment of PFPAS starts with the treatment of the underlying malignancy, this has shown to have a positive effect on the inflammatory component in some cases. Positive effects of immunosuppressive medication have been reported in some cases. Corticosteroids are often the first choice mainly because of the belief that PFPAS is an immune-mediated disorder. The effect of treatment with corticosteroids is mostly disappointing as the arthritis rarely improves and almost never improvement is seen of the contractures. Sometimes improvement of the swelling and tenderness is seen.
In our patient, treatment with oral prednisone (30 mg daily) showed no effect, therefore treatment was switched to methylprednisone pulses and MTX.
The pulse therapy showed improvement of the swelling but unfortunately, like described in other cases, no improvement of the contractures and the function of the hands was seen. Adequate treatment of the progression of the prostate cancer showed more effect with an improvement of the swelling although the contractures remained invalidating.
Since the contractures caused by PFPAS can be highly invalidating but also unresponsive to therapy, early diagnosis of PFPAS seems to become extra important.
# Conclusions
PFPAS is an uncommon paraneoplastic syndrome characterized by rapid onset of bilateral arthritis of the hands, fasciitis of the palms, progressive stiffness and contractures. The scarcity of knowledge about PFPAS makes it difficult to recognize it at an early stage. As a paraneoplastic syndrome, it has been linked to various malignancies. Thus far, PFPAS has been described in only two other cases of prostate cancer. The most important treatment of PFPAS is the treatment of the underlying disorder, this can create a moderate improvement by a reduction of the inflammation. The fibrosis and contractures appear to be irreversible unfortunately. Treatment with corticosteroids or other anti-immune therapy often leads to disappointing results. Recognition at an early stage remains a challenge. |
Impact of surgery on quality of life of Ugandan women with symptomatic pelvic organ prolapse: a prospective cohort study
Background: Pelvic organ prolapse (POP) is a significant public health issue that negatively affects the Quality of Life (QOL) of women in both low and high-income countries. About 20% of women will undergo surgery for POP over their lifetime. However, there is a paucity of information on the effect of surgery on QOL especially in resource-limited settings. We therefore assessed the QOL among women with symptomatic POP living in rural southwestern Uganda and the impact of surgery on their quality of life.Methods:We conducted a prospective cohort study among 120 women with symptomatic POP scheduled for surgery at the urogynecology unit of Mbarara Regional Referral Hospital. The QOL at baseline and at 1 year after surgery in the domains of physical performance, social interaction, emotional state, sexual life, sleep quality, personal hygiene and urinary bladder function was determined using a King's Quality of Life questionnaire. A paired t-test was used to compare the difference in mean scores at baseline and at 1-year post-surgery.Results:Of the 120 participants that were enrolled at baseline, 117(98%) completed the follow-up period of 1 year. The baseline QOL was poor. The domains with the poorest QOL were physical, social, sexual, emotional and sleep quality. The mean QOL scores in all the domains and the overall QOL significantly improved 1 year after surgery (p < 0.001). The overall QOL improved by 38.9% after surgery (p < 0.001).Conclusions:The QOL was poor among women with symptomatic POP and surgery improved the QOL in all the domains of life. We recommend that surgery as an option for treatment of symptomatic POP should be scaled up to improve on the QOL of these women.
# Background
Pelvic Organ Prolapse (POP) is defined as an anatomic support defect of the pelvic viscera resulting from the long-term failure of their supporting and suspension mechanisms. This weakness in the supporting mechanisms leads to descent of the pelvic organs, including the bladder, uterus, rectum and or small intestines, into or outside of the vagina. POP affects more than half of parous women over the age of 50 years worldwide. Various risk factors for POP have been reported including childbirth, ageing, heavy manual labour and menopause.
POP affects many aspects of women's life including physical, psychological, social interaction, sexual function and hygiene. Women with POP present with a variety Open Access *Correspondence: [email protected] 1 Faculty of Medicine, Mbarara University of Science and Technology, P.O.BOX 1410, Mbarara, Uganda Full list of author information is available at the end of the article of symptoms including urinary, bowel, sexual, and psychological which greatly compromise their quality of life (QOL). In the advanced stages, POP presents with debilitating physical symptoms like difficulty in walking, sitting and squatting, which negatively impacts the daily economic activities of these women like farming eventually resulting in poverty. In addition, women with POP frequently report limitations to their sexual life such as: lack of sexual desire, arousal, orgasm, and pain during intercourse which ultimately leads to loss of sexual interest with some being abandoned by their husbands. A number of psychological and mental health problems have been reported among women with POP including emotional disturbances, depression, loss of self-esteem, lack of sleep, rejection and isolation.
Therefore, symptomatic POP requires management which may be surgical or nonsurgical (conservative) to improve QOL. Conservative treatment options like pessaries, lifestyle change, and pelvic floor exercises have been shown to be effective as an alternative to surgery. However, pessaries are not readily available and hence their use by clinicians in treatment of symptomatic POP in this setting is low, leaving surgery as the main mode of management of symptomatic POP.
About 20% of women will undergo surgery for POP over their lifetimewith close to 200,000 POP surgeries being performed in the United States of America (USA) annually. However, access to POP surgery in low income countries is very limited because it is expensive and the skilled personnel to perform the surgery are few. Surgery for symptomatic POP has been shown to improve the QOL in high income countries. However, in low income countries including Uganda, the impact of surgery on QOL is rarely reported yet this evidence is important for advocacy in scaling up treatment for POP in these resource limited settings where women with POP face various challenges in accessing care.
Therefore, in this study we aimed to assess the quality of life among women with symptomatic POP in rural Southwestern Uganda before and after surgery in the seven life domains; physical, social interaction, psychological, sexual activity, personal hygiene, sleep and bladder function.
# Methods
## Study setting, design and study population
We conducted a prospective cohort study at the Urogynecology unit of Mbarara Regional Referral Hospital (MRRH) from December 1, 2018 to December 31, 2020, among women diagnosed with symptomatic pelvic organ prolapse (POP) scheduled for surgery. MRRH is a tertiary Hospital located in Mbarara district in Southwestern Uganda about 250 km from the capital city of Kampala. MRRH is the main referral hospital of the entire southwestern Uganda serving over 10 districts and also gets patients from the neighboring countries of Tanzania, Rwanda, Burundi and the Eastern Democratic Republic of Congo (DRC).
## Data collection preoperative workup
The diagnosis of POP, clinical exam, staging of POP, and decision to do surgery was made by the urogynecology surgical team. Participants were diagnosed to have POP if they had any one of the following clinical findings: cystocele, urethrocele, cystourethrocele, uterine prolapse, vault prolapse, enterocele, or rectocele. Staging of POP was done using the Pelvic Organ Prolapse Quantification (POP-Q) system validated by the International Continence Society (ICS) into stages I, II, III and IV.
An interviewer guided data capture tool was administered to collect information on the baseline characteristics of the study participants. These included: age, parity, education level, marital status, occupation, smoking, alcohol use, type and severity of the prolapse. Age in years was categorized according to reproductive age groups: 18-34 (early reproductive age group), 35-49 (late reproductive), 50-59 (peri-menopausal) and ≥ 60 (post-menopausal).
The participants had their quality of life (QOL) determined at enrolment prior to surgery. The QOL was determined using the King's Health Questionnaire. This questionnaire was validated to assess the QOL among women with urinary incontinence but we used it to assess QOL among women with POP in our study as a number of women with POP have been shown to have urinary incontinence. This interviewer-based questionnaire assesses seven QOL domains that include physical/daily roles performance, social interraction, sexual function, emotional/psychological state, personal hygiene, sleep quality and bladder function. A score (%) for each of the domains was calculated. Each life domain had a score ranging from 0 to 100%. The overall QOL for each participant was obtained as an average of the total scores in the seven domains. The higher the scores the poorer the QOL. Trained research assistants who included counsellors and nurses who were not part of the surgical team conducted the interviews and completed the King's Heath Questionnaire.
## Surgery for pelvic organ prolapse
The participants underwent surgery for the management of POP. Surgery was performed in those who had symptomatic POP stage II, III and IV and was dependent on the type of prolapse. The different surgical procedures that were performed include anterior colporrhapy for cystocele, posterior colporrhapy for rectocele, vaginal hysterectomy with vaginal vault suspension (sacrospinous ligament or uterosacral vault suspension) for uterine prolapse in those who had completed child bearing and did not want uterine sparing surgery, and bilateral sacrospinous vault suspension in those with recurrent vaginal vault prolapse. Cervicopexy was done for those with uterine prolapse that hadn't completed child bearing or wanted uterine sparing surgery.
## Postoperative follow up
After discharge, the participants were invited back to the hospital one year after surgery for a follow-up visit. Participants were contacted through a phone call one week prior to their scheduled postoperative visit. The purpose of the call was to remind the participants of their scheduled follow-up visit. Participants who could not be reached on phone were traced using the contacts of their next of kin. This was to minimize loss to follow-up. At this follow-up visit, the QOL one year after surgery was again assessed by the trained research assistants using the King's Health Questionnaire.
## Power calculation
A post-hoc power calculation was done where we assumed a sample size of 120, mean QOL score before surgery of 44.5 with a standard deviation of 20.9 and a mean QOL score after surgery of 8.0 with a standard deviation of 11.6. We therefore obtained a power of 100% to detect a difference in means. This sub study was part of a larger study to determine the recurrence rate of POP following surgery.
# Statistical analysis
Data were entered into Redcap and exported to Stata 13 (StataCorp, LLP, College Station, TX, USA) for analysis. Categorical data were presented as frequencies (%). The mean QOL score and the 95% confidence intervals for each of the seven domains and the overall mean QOL at baseline and at 1 year after surgery were calculated. The overall mean QOL score before and at 1 year after surgery was compared as well as the mean scores in each of the domains. A paired t-test was used to determine if there was a significant difference in the means. A p value of < 0.05 was considered significant.
# Ethical considerations
Ethical approvals were obtained from the Mbarara University of Science and Technology (MUST) Research Ethics Committee and the Uganda National Council for Science and Technology (UNCST) number HS368ES. Written informed consent was obtained from all the study participants or from their legally authorized representatives for those who couldn't read or write. Confidentiality was observed during all the interviews. Personal identifiers such as name and in-patient number were not collected. The study participants were assigned study ID numbers.
# Results
A total of 130 women with symptomatic POP scheduled for surgery were seen during the study period. Of these, 10 declined to consent and were not included in the study. We therefore enrolled a total of 120 participants into the study. The participants were followed up for 12 months post-surgery, with 117 completing the follow-up period giving a completion rate of 97.5%. The mean age of the study participants was 55 years (SD ± 15) with the majority (n = 46, 38.3%) aged ≥ 60 years, peasant farmers (n = 106, 88.3%) and had had five or more deliveries (n = 94, 78.3%) as shown in. The majority of the participants had uterine prolapse (n = 85, 70.8%) and the commonest stage of POP was stage III (n = 56, 46.7%) as shown in.
The surgical procedures that were performed and the related complications are shown in. The most common procedure performed among the study participants was Vaginal hysterectomy with sacrospinous ligament vault fixation in 66 (55%) of the participants. Intraoperative complications were encountered in four patients and these included: rectal injuries (n = 2), hemorrhage that required blood transfusion (n = 1), and urinary bladder injury (n = 1). The main postoperative complication was vaginal cuff infection which occurred in 10 (8.3%) of the participants.
The baseline mean QOL overall and in the specific life domains is shown in There was a significant improvement in the overall quality of life as well as in the different domains 1 year after surgery for symptomatic POP as shown in. The overall mean QOL score at baseline of 45.5% decreased by 38.9-6.6% at 1 year after surgery and this change was statistically significant (p < 0.001). There was also a decrease in all the specific QOL domains at 1 year after surgery for POP. Physical activity domain decreased by 55.1% from 66.7 to 6.6%, social interaction decreased by 52.9% from 61.9 to 9%, emotional status reduced by 52.8% from 58.4 to 5.6%, sexual performance by 55.6% from 69.1 to 13.5% and sleep quality also decreased by 35.5% from 40.4 to 4.9%. Though low at baseline, the mean scores for personal hygiene and urinary bladder function also decreased at 1 year post-surgery by 19.4 and 9% respectively. All these changes were statistically significant (p < 0.001).
# Discussion
This study showed that women presenting with symptomatic POP had a poor quality of life before surgery with an overall mean QOL score of 45.5%; surgical management for the symptomatic POP significantly improved the QOL by 38.9%.
The poor QOL in this study is similar to findings from studies done in China and rural Pakistan. This is probably because POP is a complex condition which affects both physical and functional aspects of a woman including physical, social, emotional, sleep, and sexual performance ultimately leading to poor QOL. The poor physical performance observed in this study was also reported in studies done in other low income countries. A study done in South Africa bysuggested that the poor physical performance in women with POP could be due to chronic pelvic pain caused by stretching and weakening of the pelvic ligaments which makes it difficult for these women to walk, bend and work. Women with POP in our study also had a poor social life score. This is similar to what was observed in studies done in Nepal and Ethiopia. This could be due to the fact that advanced stages of POP (III and IV) are associated with foul smelling vaginal discharge and urinary incontinence which limits the interaction with the community for fear of being discriminated. The poor QOL score among women with POP in the emotional domain was similar to what was found in a study done in USA. This low emotional state in women with POP could be due to a combination of perceived decreased body image, shame, and non-disclosure for fear of discrimination or even divorce in these women. Among the women who were still sexually active, the score in their sexual life was poor and this is similar to what was reported in other studies. Women with advanced POP are likely to experience dyspareunia, lack of libido and arousal ultimately leading to poor sexual performance. The participants also had poor QOL score in the sleep domain as was found elsewhere. The poor sleep quality may be attributed to the frequent nocturia that is common in women with symptomatic POP. Additionally, the loss of sleep could be as a result of depression in these women.
One year following surgery, the overall quality of life including physical, social, emotional, sleep and sexual performance improved. This is similar to what was observed in studies done in the high and low income countries where surgery significantly improved all aspects of women's life. Bilateral sacrospinous ligament vault suspension which was performed for participants with recurrent vaginal vault prolapse in this study has been shown to improve QOL. Surgery may have improved the overall QOL of life because it relieves patients' symptoms, restores normal anatomy and function of the pelvic structures.
The improvement in physical performance after surgery is similar to that observed in a study from Ethiopia. This improvement could be due to the correction of the anatomical defect leading to a reduction in the chronic pain and discomfort which enables the women to walk and even work comfortably. The social life of the women similarly improved probably because POP surgery has been found to correct associated urinary incontinencewhich ultimately takes away the bad urine smell resulting in improvement in the social domain. The QOL in the emotional domain improved after surgery similar to that reported in a previous study among American women. The improvement may be because the women did not have any more worries about their body image perception and fear about discrimination from society after surgery. The sleep domain improved after surgery as in a study by Ghetti et al.. This improvement is probably because of the improvement in bladder symptoms such as nocturia as well as improvement in depressive symptoms which occurs after surgery. As was reported in other studies from the developed world, sexual function improved after surgery. This is probably because the restoration of the normal vaginal length after surgery leads to a reduction in the dyspareunia that used to be experienced by these women. Additionally, this improvement could be due to a related improvement in the emotional status. Therefore, in addition to surgery, women with symptomatic POP would benefit from a multidisciplinary approach for accurate management. These women should be offered psychosocial support and counselling because they have emotional, social and sexual problems.
Our study has some limitations. This study was conducted at a single regional referral hospital in southwestern Uganda and the findings may not be generalizable to all other women in Uganda. The King's Health Questionnaire was validated to assess the QOL among women with urinary incontinence but we used it to assess QOL among women with POP in our study since a significant number of women with POP also have urinary incontinence.
# Conclusions
Quality of life among women with POP was poor overall and surgery greatly improved women's quality of life. Surgery, as an option for the management of POP, should be scaled up to improve the QOL among women with symptomatic POP. In addition, women with symptomatic POP should be offered psychosocial support and counselling because they have emotional, social and sexual problems. |
Space-Time Analysis of Testicular Cancer Clusters Using Residential Histories: A Case-Control Study in Denmark
Though the etiology is largely unknown, testicular cancer incidence has seen recent significant increases in northern Europe and throughout many Western regions. The most common cancer in males under age 40, age period cohort models have posited exposures in the in utero environment or in early childhood as possible causes of increased risk of testicular cancer. Some of these factors may be tied to geography through being associated with behavioral, cultural, sociodemographic or built environment characteristics. If so, this could result in detectable geographic clusters of cases that could lead to hypotheses regarding environmental targets for intervention. Given a latency period between exposure to an environmental carcinogen and testicular cancer diagnosis, mobility histories are beneficial for spatial cluster analyses. Nearest-neighbor based Q-statistics allow for the incorporation of changes in residency in spatial disease cluster detection. Using these methods, a spacetime cluster analysis was conducted on a population-wide case-control population selected from the Danish Cancer Registry with mobility histories since 1971 extracted from the Danish Civil Registration System. Cases (N=3297) were diagnosed between 1991 and 2003, and two sets of controls (N=3297 for each set) matched on sex and date of birth were included in the study. We also examined spatial patterns in maternal residential history for those cases and controls born in 1971 or later (N= 589 case-control pairs). Several small clusters were detected when aligning individuals by year prior to diagnosis, age at diagnosis and calendar year of diagnosis. However, the largest of these clusters contained only 2 statistically significant individuals at their center, and were not replicated in SaTScan spatial-only analyses which are less susceptible to multiple testing bias. We found little evidence of local clusters in residential histories of testicular cancer cases in this Danish population.
# Introduction
Testicular cancer incidence in Western countries dramatically increased from the 1940's to the 1990's, and has become the number one cause of cancer in males under age 40 [bib_ref] Global trends in testicular cancer incidence and mortality, Rosen [/bib_ref]. The exact reasons remain unclear, though a birth cohort effect has been shown to be an important factor [bib_ref] Increasing incidence of testicular cancer-birth cohort effects, Ekbom [/bib_ref] [bib_ref] Risk of testicular cancer according to birthplace and birth cohort in Denmark, Myrup [/bib_ref] [bib_ref] Trends in the incidence of testicular germ cell tumors in the United..., Mcglynn [/bib_ref]. The increase in rates was especially heavy in northern Europe. From 1943 to 2003 the incidence rates in Denmark increased from 3.4 to 10 cases per 100,000 person-years; its rates are among the highest in the world [bib_ref] Risk of testicular cancer according to birthplace and birth cohort in Denmark, Myrup [/bib_ref] [bib_ref] Testicular germ cell cancer incidence in an immigration perspective, Schmiedel [/bib_ref]. In recent years, these rates appear to be stabilizing [bib_ref] Trends in testicular cancer incidence in the Nordic countries, focusing on the..., Jacobsen [/bib_ref].
Biologically, most testicular cancer can be categorized as either seminomas (germ cell) or nonseminomas. Seminomas typically occur later than nonseminomas (late 30's to early 50's versus younger than 30), indicating that there may be etiological differences between the two groups. Despite many potential risk factors that have been examined, the only firm determinants of testicular cancer risk remain age, family history of testicular cancer, national origin, birth year, ethnicity, and cryptochoridism (undescended testes) which is found in 10% of all cases [bib_ref] Trends in the incidence of testicular germ cell tumors in the United..., Mcglynn [/bib_ref] [bib_ref] Cryptorchidism and testicular cancer: separating fact from fiction, Wood [/bib_ref] [bib_ref] Cancer in first-degree relatives and risk of testicular cancer in Denmark, Nordsborg [/bib_ref] [bib_ref] Identification of nine new susceptibility loci for testicular cancer, including variants near..., Ruark [/bib_ref] [bib_ref] A hierarchical frailty model for familial testicular germ-cell tumors, Valberg [/bib_ref]. Further evidence suggests that maternal factors, early childhood environment, and the in utero environment may play a role since testicular cancer usually occurs in the relatively young, but specific risk factors are as yet unidentified [bib_ref] Prenatal and perinatal risk factors for testicular cancer, Brown [/bib_ref] [bib_ref] Rising risk of testicular cancer by birth cohort in the United States..., Mckiernan [/bib_ref]. Certainly there is a latency period between exposure to carcinogens and disease manifestation, which should be taken into account when conducting an epidemiologic analysis of testicular cancer.
A spatial analysis in Britain was conducted but without finding evidence of responsible factors that were geographically-differentiated [bib_ref] Spatial variation and temporal trends of testicular cancer in Great Britain, Toledano [/bib_ref]. Though this study looked at testicular cancer rates over several years, they were not able to account for human mobility and used aggregate data (electoral wards) rather than individual-level data. Exposures leading to testicular cancer may occur many years before diagnosis, therefore cluster detection based on residence at diagnosis may not accurately reflect the life history associated with etiology.
Latency is a difficult barrier in investigating spatial cancer clusters, as individuals are often highly mobile and will move residences more than once over the course of their life [bib_ref] Inflammatory breast cancer clusters: A hypothesis, Levine [/bib_ref] [bib_ref] Response to Condon et al. comments on "Cancer clusters in the USA:..., Goodman [/bib_ref] [bib_ref] Cancer clusters in the USA: what do the last twenty years of..., Condon [/bib_ref]. Several methodologies have recently been developed that help solve this difficult statistical task, with two of the most promising being generalized additive models (GAMs) with a loess smoother [bib_ref] Method for mapping population-based case-control studies: an application using generalized additive models, Webster [/bib_ref] and Q-statistics [bib_ref] local and focused geographic clustering for case-control data with residential histories, Jacquez [/bib_ref] [bib_ref] Case-control geographic clustering for residential histories accounting for risk factors and covariates, Jacquez [/bib_ref] [bib_ref] Space-Time Clustering of Non-Hodgkin Lymphoma Using Residential Histories in a Danish Case-Control..., Nordsborg [/bib_ref]. Each of these methods allows for covariate adjustment and incorporates the use of time-specific geographic coordinates. Q-statistics have recently been shown to be effective at locating simulated clusters [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref] , and are here applied on a large testicular cancer dataset from Denmark.
By systematically examining residential histories, geographical clusters of testicular cancer may appear that can yield clues to what has driven the recent rise in incidence. Here we tested this hypothesis by undertaking a space-time analysis of residential histories of Danish testicular cancer cases and 2 sets of matched controls. This analysis is the first of its kind, in that it uses a comprehensive, nation-wide longitudinal database including all residential histories in Denmark to investigate local clusters of testicular cancer over a long time period. We also selected two independent control groups, something rarely done in spatial epidemiology, but recently shown to impact findings of spatial cluster analyses [bib_ref] Space-Time Clustering of Non-Hodgkin Lymphoma Using Residential Histories in a Danish Case-Control..., Nordsborg [/bib_ref].
# Materials and methods
## Study population
The Danish Cancer Registry is a comprehensive nationwide cancer registry begun in 1943. For the purposes of this study, we selected testicular cancer cases diagnosed between January 1 st , 1991 and December 31 st , 2003 (n = 3,297) ( The remaining cases were classified as non-seminoma. There were 1871 seminoma and 1426 non-seminoma cases. Two independent control groups were randomly drawn from the Danish Civil Registration System (CRS) and matched with cases according to age in a 1:1 ratio. More specifically, each control group contained 3,297 males born in Denmark on the same date as a cancer case (or same month if there were no alive controls born on the same day) that were cancer-free (with the exception of non-melanoma skin cancer) and living in Denmark at the time of the matching cases' diagnosis. The use of two independent control groups is supported by recent work that suggests cluster results were not consistent across two control groups, leading to the conclusion that the spatial distribution of the control groups might be influencing findings of clusters [bib_ref] Space-Time Clustering of Non-Hodgkin Lymphoma Using Residential Histories in a Danish Case-Control..., Nordsborg [/bib_ref]. Residential histories of cases and controls were traced using data from the CRS from 1971 to date of diagnosis. Geocoding success and accuracy were very high, with geographic coordinates assigned to 98% of the addresses (67,244 / 68,578). For those matched to the home address (88% of residences) the address point is defined within 1 meter of the front door of the house, and the precision of the geocoding is within a few meters. Another 4% were matched at the neighbor's residence, 1% at the street level, and 5% at the municipality level. The Danish Data Protection Agency approved the study. In accordance with Danish law written consent was not obtained as the study was entirely register-based and did not involve biological samples from, or contact with study participants. IRB approval of work with geographic coordinates of residences was obtained through Western Institutional Review Board. Cases and controls were de-identified and assigned a random ID prior to analysis. Some protected health information (location and date of illness) were required to complete the study, but were not associated with patient identifiable information during the analysis. We also collected residential addresses of the mothers of cases and controls from 1971 to date of birth of the case or control. This information was available for the youngest 18% of the case-control pairs. These data were used to investigate space-time local clusters of the mothers of cases and controls during pregnancy. If one of the individuals in a case-control pair had missing information on the mother's address, the pair was excluded from the analyses. Information on mothers' addresses existed for 589 case-control pairs in the first case-control group, and 591 case-control pairs in the second group, resulting in 3862 and 3913 residential addresses respectively. Of these 99% were geo-referenced.
## Conditional logistic regression
An investigation was conducted to look for potentially important covariates using conditional logistic regression analysis on cases and each control group. The covariates were selected based on data availability in the existing registers and existing knowledge about potential risk factors for testicular cancer and included: birth weight (kg) and birth length (cm), age of mother at birth, age of father at birth, maternal marital status (married, never married, divorced) from the Danish Birth Registry for those born after 1972, time-weighted average socioeconomic status (SES) using yearly income for 271 municipalities in Denmark, and family history of testicular cancer and of any cancer among first degree relatives with a reported primary cancer diagnosis in the Cancer Registry. Results from the logistic regression analysis were used for adjustment of key covariates in the spatial cluster analysis, as described in the following sections.
# Q-statistics background
A complete discussion on the incorporation of residential histories into disease cluster detection in the form of Q-statistics can be found elsewhere [bib_ref] local and focused geographic clustering for case-control data with residential histories, Jacquez [/bib_ref] [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref] , and may be referred to for a more extensive explanation of these methods. Here we will briefly review the statistics applied to this particular study. Q-statistics are calculated according to a nearest neighbor method. Over the course of the study period, the number of geographically nearest neighbors surrounding each individual who are a case (not a control) is calculated. A new set of Q-statistics are calculated at each time step, defined as any point in time at which at least one person changed home residences. For example, if everyone maintained the same address from 2000 to 2002 and then someone moved in 2003, the time step between calculations of the statistic would be 3 years. Each statistic is a sum over a matrix of nearest-neighbor relationships. Statistical significance is determined by permutation testing. The user specifies the number of nearest neighbors (k) prior to running the calculations. Each statistic is duration-weighted so as to more heavily weight individuals who have lived in the surrounding area longer. The basic formula for Q-statistics is as follows:
[formula] Q ðkÞ i;t ¼ c i X k j¼1 Z ðkÞ i;j;t c j Equation 1 [/formula]
This is the count, at time t, of the number of k nearest neighbors of case i that are also a case. The case-control identifiers, c i and c j , , for individual i and j are binary (1 if a case, 0 if a control). When i is a control, Q ðkÞ i;t = 0. The term Z ðkÞ i;j;t is a binary spatial proximity matrix of nearest neighbors that is 1 when participant j is one of the k nearest neighbors of participant i at time t; and 0 otherwise. Since Q ðkÞ i;t is a sum of the case status of the k unique nearest neighbors of individual i, the statistic is in the range 0.k. When i is a case, low values of Q ðkÞ i;t indicate cluster avoidance (e.g., a case surrounded by controls), and large values indicate a cluster of cases.
Further Q-statistics build on this basic equation to identify clusters that occur according to certain spatiotemporal patterns:
[formula] Q ðkÞ i ¼ Z T t¼t 0 Q ðkÞ i;t dt Equation 2 Q ðkÞ t ¼ X n 1 i¼1 Q ðkÞ i;t Equation 3 Q ðkÞ ¼ X n 1 i¼1 Q ðkÞ i Equation 4 [/formula]
Equation 2 (the sum over all time points of Equation 1) is used to identify cases that are consistently centers of spatial clusters through time. Equation 3 represents the global statistic which indicates whether clustering occurs throughout the entire area at a particular moment in time. It is calculated by summing Equation 1 over all cases at that time point. Equation 4 further summarizes this statistic, and is the global case clustering of residential histories throughout the study area for the entire period. It is calculated by summing Equation 2 over all cases. This statistic considers all of the residential histories simultaneously for the entire study period, and is a measure of the persistence of global clustering. It is large when case clustering persists through time.
In this paper, Equation 2 is most frequently used in determining whether an individual is the center of a cluster. This is based on performance of Q-statistics as previously shown using simulations based in part on these data [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref]. Equation 1 (Q ðkÞ i;t ) is used to identify when and where an individual is the center of a local cluster. The simulation analysis [bib_ref] Cancer clusters in the USA: what do the last twenty years of..., Condon [/bib_ref] was undertaken to help account for multiple testing bias arising from the large number of statistical tests conducted in Q-statistics run across residential histories. In the simulation analysis we created clusters and evaluated the predictive capability of different versions of the local and global Qstatistics. Type 1 error was minimized and the ability to detect true clusters was maximized when we used the following criteria. First, candidate cluster constituents were identified using p 0.001 for Q ðkÞ i and p 0.05 for Q ðkÞ i;t . Here we used information from the statistics defined by Equations 1 and 2 simultaneously to identify possible cluster members. Second, we required at least 4 of these cluster members to be nearest neighbors of one another in order to declare a cluster statistically significant. Equations 3 and 4 were not helpful in discerning the simulated clusters.
Statistical significance is determined by randomizing the case-control identifiers over the residential histories under the null hypothesis of no association between places of residence and case-control status. Only case-control status is randomized, maintaining the integrity of the individual residential histories, which are then used to calculate the Q statistics. The randomization procedure is repeated over many iterations to build up the distributions of the Qstatistics under the null hypothesis. For adjustment of key covariates, the null hypothesis can account for them by employing the adjusted probabilities of being a case as calculated from regression [bib_ref] Case-control geographic clustering for residential histories accounting for risk factors and covariates, Jacquez [/bib_ref]. The equation for predicting the probability of being a case given the vector of covariates and risk factors for the i th individual is:
[formula] pðc i ¼ 1jx i Þ ¼ e b 0 x i 1 þ e b 0 x i Equation 5 [/formula]
Here the logit function is the natural log of the odds, and β is the vector of regression (slope) coefficients. Using the results of the conditional logistic regression equation, the coefficients for each variable along with values for each individual are used to assign individual probabilities of being a case in the adjusted analysis. Note that the range of possible p-values is determined by the number of randomizations of the null hypothesis applied. Given the computational power and time required for these analyses, 999 randomizations was the maximum reasonable number of iterations, generating a minimum p-value of 0.001. Given the problem of multiple testing in these and many other spatial analyses, our recent simulation analyses [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref] suggested a guide for examining case-control residential histories which we implement and describe below.
# Spatial cluster analysis
Spatial cluster analyses were conducted using Q-statistics in SpaceStat (BioMedware, Inc., Ann Arbor, MI). The dataset was divided into 3 groups: all cases, seminomas, and mothers of cases and their matched controls, and each group was analyzed using different measures of time (age, calendar year, year prior to diagnosis for all cases and seminomas, and calendar year and months prior to birth for the mothers of cases). The different measures of time were used in the event that different environmental effects could be responsible for local clusters, i.e., if individuals were all located in the same region at the same date or whether they were more aligned by years prior to diagnosis (indicating similar latency period from exposure to disease manifestation). Q-statistics examine each case at each time step as a possible center of cluster, which is a thorough approach but it introduces the possibility of multiple testing. Our simulation study based in part on these data produced a guide to help account for multiple testing [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref] of individual cases over time, and suggested that a possible cluster could be further evaluated if 4 or more significant cases were detected in the same area with a Q i (k) , p = 0.001 and Q it (k) , p 0.05 using k = 15 nearest neighbors. We used this guide for the analyses performed in the present study. Largest clusters were re-examined at the time slice suggested by the Q-statistics using Kulldorf's scan statistic in SaTScan (v 9.0.1) to compare with results generated by an established cluster detection method [bib_ref] Space-Time Clustering of Non-Hodgkin Lymphoma Using Residential Histories in a Danish Case-Control..., Nordsborg [/bib_ref] [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref] [bib_ref] Space-time clusters of Crohn's disease in northern France, Genin [/bib_ref] [bib_ref] Applying Spatial Analysis Tools in Public Health: An Example Using SaTScan to..., Sherman [/bib_ref]. However, this method did not account for human mobility, thus analyses were conducted on sub-sets of the original space-time data, which included only one location per individual. These time slices were selected to match statistically significant time periods identified by Q-statistics, the approach suggested in the simulation study [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref]. We used a Bernoulli model in SaTScan, and the p-value for test of significance was obtained from Monte Carlo simulations (999 replications). We analyzed circular clusters with a maximum cluster size of 50% of the total population.
# Results
There were 3297 cases of testicular cancer, 1871 of which were seminomas, and two independent sets of 3297 controls in this Danish population-wide case-control study. In the conditional logistic regression analyses, the only variable that was a statistically significant predictor for both seminomas and all cases using both control groups was that of having a family history of testicular cancer [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref]. There were 69 cases with testicular cancer in first-degree relatives, and 39 of those were seminomas; 26 controls across both control groups with testicular cancer in first degree relatives, and 13 among controls of seminomas [bib_ref] Cancer in first-degree relatives and risk of testicular cancer in Denmark, Nordsborg [/bib_ref]. For all cases, the parameter estimate was 1.72 using control group 1 (p-value = 0.001, Hazard Ratio = 5.58) and 1.75 using control group 2 (p-value = 0.001, Hazard Ratio = 5.75). Given that this family history variable may reflect a common underlying exposure, there is the possibility of over-adjustment. Unadjusted spatial cluster analyses are presented, along with adjusted analyses which include the probability of being a case given a family history of testicular cancer. The results of the unadjusted cancer cluster investigation are shown in [fig_ref] Table 2: Unadjusted Analysis. [/fig_ref]. The table lists each test, with the number of significant individuals found and the number of individuals in the largest cluster for both control group 1 and control group 2. Also listed are number of the same individuals that were significant using both control group 1 and control group 2, and the general locations of each group of significant individuals. There was no overlap in individuals identified as centers of clusters across the two control groups in any of the analyses, including all cancer, seminomas, and mothers of cases. The largest cluster identified included 2 significant cases in the center of the cluster in Aarhus, 6-8 years prior to diagnosis, using all cases and control group 1. Another cluster was also identified in Aarhus containing residence of mothers of 2 cases in 1971, again using control group 1. All other clusters included no more than 1 significant case at the center of each cluster and persisted for many years. Given propensity for detecting false positives in this space-time cluster analysis, these results did not reach Results of the unadjusted cancer cluster analysis of testicular cancer in Denmark. There were k = 15 nearest neighbors used in every analysis. The number of significant clusters, number of persons in the largest cluster, indication of whether there were significant global statistics, the location of each cluster, and whether there were individual cases which were found in clusters using each control group are listed for each analysis. All testicular cancer cases, seminomas only, and the mothers of cases were aligned according to age at diagnosis, calendar year of diagnosis, and number of years prior to diagnosis (YPD). For the two largest clusters, the timing of the clusters is indicated. our threshold of 4 or more significant cases, which was one of the recommendations from our simulation study [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref]. The results from the analyses adjusted for family history of testicular cancer are shown in [fig_ref] Table 3: Adjusted Analysis [/fig_ref] and are highly similar to the unadjusted analyses. Again there was no overlap in individuals identified as centers of clusters across the two control groups in any of the analyses. There was some similarity with the unadjusted analyses in locations of significant individuals, with several of them being from either Copenhagen or Aarhus. No cluster contained more than one significant case at its center.
The locations of significant clusters sometimes overlapped across the different analyses. This was particularly true in the Copenhagen and Aarhus regions. However, Kulldorf's scan statistic failed to confirm the two largest clusters in Aarhus which contained 2 significant cases at their center. Using control group 1 at 7 years prior to diagnosis, no significant clusters were found. The largest cluster detected using SaTScan contained 19 cases and was located in Copenhagen (RR = 2.0, p = 0.055); Q-statistics also found a possible cluster in this region of Copenhagen at p = 0.003 for Q i (k) , which failed to meet p = 0.001 that was required by our simulation study to help account for multiple testing [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref]. Using maternal residential histories and control group 1 in 1971, the largest cluster detected by SaTScan was again located in Copenhagen and not statistically significant (RR = 2.02, p = 0.51).
# Discussion
This study used a complete record of all residential histories in Denmark to investigate local clusters of testicular cancer among residents from 1971 until diagnosis in 1991-2003. While a few small clusters were detected, no cluster contained more than two significant cases at its center, short of the four significant cases required to overcome multiple testing bias as recommended by our simulation study [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref]. Further, the selection of a second control group also proved helpful to curb multiple testing bias by allowing us to examine whether the presence of a cluster remained consistent. While some clusters were found in different regions of Copenhagen using both control groups, the clusters never covered the same locations or contained the same individuals, suggesting a likelihood of being chance findings. The results of this analysis are in line with one other small area clustering study of testicular cancer rates in England [bib_ref] Spatial variation and temporal trends of testicular cancer in Great Britain, Toledano [/bib_ref] which also suggested a lack of evidence of geographic clustering of testicular cancer. Our analysis goes further than this previous study by showing little evidence of local clusters even when incorporating individual-level data with changes in residency. Our method did not find evidence of local clusters in mothers' residential histories suggesting in utero exposures.
There is no established protocol for detecting space-time clusters in mobile populations. When considering mobility one must consider that cases may spend different durations of time moving in and out of a cluster region. Nearest-neighbor Q-statistics allow us to investigate local and global clustering throughout residential histories in case-control studies, but are subject to chance findings resulting from multiple testing. In our simulation analyses [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref] based in part on these testicular cancer data, we created many different types of clusters and arrived at a rule of thumb to help distinguish true clusters from false positives. The rule of thumb, a cluster of 4 or more individuals (Q i (k) , p = 0.001 and Q it (k) p 0.05) using k = 15, was successful for distinguishing larger clusters, which were confirmed by the scan statistic in SaTScan. We followed this approach in the analyses reported here and did not find evidence of clusters. Importantly, due to the large data set of residential histories, a single analysis took up to 12 hours. Consequently we could not explore a wide range of different levels of k for the analyses, but had to rely on results from the simulation study when selecting k = 15; although a few sensitivity analyses using k = 10, 20, and 100, and combining the two control groups together did not change our conclusions. To our knowledge, this is the first examination of spatial clusters of testicular cancer using residential histories. Cases were identified in the virtually complete high-quality populationbased Danish Cancer Registry [bib_ref] The Danish cancer registry, Gjerstorff [/bib_ref] [bib_ref] The Danish Cancer Registry-history, content, quality and use, Storm [/bib_ref] , thus the study had very reliable case ascertainment. Furthermore, the Danish Civil Registration System provided an ideal frame for control selection and collection of residential addresses back to 1971 [bib_ref] The Danish civil registration system, Pedersen [/bib_ref]. We adjusted for family history of testicular cancer as this was associated with testicular cancer and may vary spatially; however the adjustment did not change our conclusions.
The primary objective of this work was to generate insights concerning the etiology of testicular cancer, a disease that showed rapid increases in incidence throughout many Western regions in the second half of the 20 th century and has few established risk factors. This study design allowed us to overcome many of the limitations commonly found in spatial analyses. We used individual-level data from the Danish Cancer Registry to examine potential clusters using all cases of testicular cancer diagnosed over a 13 year period. Though it is difficult to assess the power in these analyses, we had large sample size to look for clusters only among seminomas, a more histologically similar subset of testicular cancer with a greater likelihood of common etiologic factors. We also examined potential clusters in residential history data for mothers of 589 case-control pairs and for a few years prior to pregnancy. We examined spatial patterns using multiple measures of time including locations at different ages, calendar years, and years prior to diagnosis, along with months prior to birth for the mothers.
We presented analyses both unadjusted and adjusted for family history of testicular cancer. The ability to adjust our analysis based on relevant covariates is a strength of the Q-statistics method. It is important to understand whether any detected clusters are due to variations in known covariates, or an unknown variable associated with living in a particular location. In this analysis, neither unadjusted nor adjusted results provided compelling evidence of clusters. In the adjusted results, the number of cases per potential cluster decreased suggests = ing that family history of testicular cancer may have been partially driving the results of the unadjusted analysis.
We selected two independent control groups which was helpful for interpreting the findings. Lastly, we used Q-statistics, one of the few approaches available for examining clusters throughout case-control residential histories, an approach which has been shown to have effective performance in a simulation study in this region [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref]. The study could have been improved by including more years of follow-up of residential histories of cases, controls, and among mothers; the Civil Registration System began recording residential data in 1971, so it is not possible to include earlier years. Nonetheless, the study design allowed us to assess possible geographic clusters of testicular cancer using 20 years of residential histories.
The development and application of space-time cluster statistics that allow for multiple addresses and mobility is still in its infancy. Further performance evaluation of Q-statistics would help to demonstrate their utility, along with performance evaluations of other newly developed methods including the multiple address function available in SaTScan v. 9.3 (Mar 20, 2014). The limitation of not having a direct method for adjustment for multiple testing in Q-statistics was addressed in a previous simulation paper using simulated clusters, and is the reason for considering only those clusters with more than 4 individuals as significant using Q ðkÞ i p = 0.001 and Q ðkÞ i;t p 0.05 [bib_ref] Performance of cancer cluster Q-statistics for case-control residential histories, Sloan [/bib_ref]. Even though our null results were confirmed by time slice analysis in SaTScan, it is possible that our guideline for differentiating clusters from false positives is overly restrictive or perhaps too lenient under different situations. This is ground for future work.
Environmental influences that vary geographically either do not play a strong role in the incidence of testicular cancer in this population or our method did not detect them. The cohort effect reported in previous studies may be due to a more ubiquitous environmental factor that does not exhibit spatial variation. Additional research directions are needed to aid the pursuit of risk factors responsible for the increased risk of testicular cancer in young men.
[table] Table 1: All individuals had testicular cancer as their primary cancer diagnosis with ICD-O morphology behaviour code 3 (malignant, primary Space-Time Testicular Cancer Cluster Analysis in Denmark PLOS ONE | DOI:10.1371/journal.pone. [/table]
[table] Table 2: Unadjusted Analysis. [/table]
[table] Table 3: Adjusted Analysis. Mothers of Cases (589 group 1/591 group 2 case-control pairs)Results of the cluster analysis of testicular cancer in Denmark, adjusted for family history of testicular cancer, following the same format as used inTable 2.doi:10.1371/journal.pone.0120285.t003 [/table]
|
B.1.617.2 (Delta) Variant of SARS-CoV-2: features, transmission and potential strategies
Coronavirus disease 2019 caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has become a pandemic. With the continuous evolution of the viral genome, SARS-CoV-2 has evolved many variants. B.1.617.2, also called Delta, is one of the most concerned variants. The Delta variant was first reported in India at the end of 2020 but has spread globally, by now, to 135 countries and is not stand still. Delta shared some mutations with other variants, and owned its special mutations on spike proteins, which may be responsible for its strong transmission and increasing virulence. Under these circumstances, a systematic summary of Delta is necessary. This review will focus on the Delta variant. We will describe all the characteristics of Delta (including biological features and clinical characteristics), analyze potential reasons for its strong transmission, and provide potential protective ways for combating Delta.
# Introduction
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the causative pathogen of Coronavirus disease 2019 (COVID- [bib_ref] Inhibition of SARS-CoV-2 viral entry upon blocking N-and O-glycan elaboration, Yang [/bib_ref] , emerged in late 2019 [bib_ref] A Novel Coronavirus from Patients with Pneumonia in China, Zhu [/bib_ref]. Because of the alarming levels of spread and severity, COVID-19 was characterized as a pandemic on . Since the outbreak of the epidemic, a large number of lives have died in the process of fighting the virus. To date, the cumulative number of infection is nearly 212 million and the number of cumulative deaths is 4.4 million. COVID-19 has led to a serious public health crisis, and poses a huge threat to daily life.
The first whole genome sequence of SARS-CoV-2 was available within ten days after the occurrence of clustered cases [bib_ref] A new coronavirus associated with human respiratory disease in China, Wu [/bib_ref]. Several months later, three million genome sequences were submitted. A great quantity of mutations obtained by analyzing these sequences reveal the variability of SARS-CoV-2 genome. Actually, it was not surprising that SARS-CoV-2 has new variants, because RNA virus evolves and changes easily. The mutations we can detect were those that can survive and spread successfully [bib_ref] The origins and potential future of SARS-CoV-2 variants of concern in the..., Otto [/bib_ref]. Since the first COVID-19 genome sequence was uploaded to the database, scientists began to monitor the evolution of SARS-CoV-2 in real time.
The emergence of B.1.1.7 has attracted attention [bib_ref] Emergence and rapid transmission of SARS-CoV-2 B.1.1.7 in the United States, Washington [/bib_ref] [bib_ref] Early transmissibility assessment of the N501Y mutant strains of SARS-CoV-2 in the..., Leung [/bib_ref] , because there is no previous evidence that variation will increase the adaptability of SARS-CoV-2 genome. In general, a new variant may increase circulation frequency or the change of clinical characteristics. B.1.1.7, also known as Alpha variant, was first reported in the United Kingdom. It exhibited a significant increase in transmission [bib_ref] Emergence of SARS-CoV-2 B.1.1.7 Lineage -United States, Galloway [/bib_ref] [bib_ref] Assessing transmissibility of SARS-CoV-2 lineage B.1.1.7 in England, Volz [/bib_ref] and was listed as variant of concern (VOC) by the World Health Organization (WHO) [bib_ref] Assessing transmissibility of SARS-CoV-2 lineage B.1.1.7 in England, Volz [/bib_ref] [bib_ref] Detection of a SARS-CoV-2 variant of concern in South Africa, Tegally [/bib_ref] , P.1 (Gamma) [bib_ref] Antibody evasion by the P.1 strain of SARS-CoV-2, Dejnirattisai [/bib_ref] , B.1.617.2 (Delta) and B.1.1.529 (Omicron). The possibility of SARS-CoV-2 variants with distinct characteristics to evolve was increased due to the immense number of current cases. B.1.351 starts in South Africa, and P.1 in Brazil, the outstanding feature of these two viruses is that they have the ability to escape immunity [bib_ref] SARS-CoV-2 variants B.1.351 and P.1 escape from neutralizing antibodies, Hoffmann [/bib_ref]. In other words, they may lead to decreased vaccine efficacy and ineffective drug treatment. Delta is also a VOC. Compared to the original virus, Delta has the following four features: (1) high viral load. In terms of viral load, Delta is 1,260 times that of last year's pandemic strain;
(2) large exhaled virus concentration. It makes people more likely to be exposed to the virus; (3) highly contagious disease; (4) un-conspicuous symptoms. The two aspects mentioned later make Delta difficult to prevent. This review focused on the features and transmission of B.1.617.2 variant. The potential strategies against Delta will also be discussed. [bib_ref] Clinical and virological features of SARS-CoV-2 variants of concern: a retrospective cohort..., Ong [/bib_ref].
## Features of delta variant
## Biological characteristics of delta variant
In general, Delta virus is a variant of SARS-CoV-2, so their biological characteristics were similar to SARS-CoV-2, which is an enveloped, positive-sense single-stranded RNA virus, belonging to the beta coronavirus (β-CoV) [bib_ref] Coronavirus biology and replication: implications for SARS-CoV-2, V'kovski [/bib_ref]. The genome of SARS-CoV-2 comprises fourteen open reading frames (ORFs) [bib_ref] Structure of SARS-CoV-2 ORF8, a rapidly evolving immune evasion protein, Flower [/bib_ref] [bib_ref] SARS-Cov-2 ORF3a: Mutability and function, Bianchi [/bib_ref] , that encode sixteen non-structural proteins (NSP), nine accessory proteins and four structural proteins. NSPs participated in the formation of replicase complex and remaining parts are involved in viral entry, assembly and release [bib_ref] Structure of SARS-CoV-2 ORF8, a rapidly evolving immune evasion protein, Flower [/bib_ref] [bib_ref] Mechanisms of SARS-CoV-2 Transmission and Pathogenesis, Harrison [/bib_ref]. Spike protein (S) is critical for viral infection, which is an important target for combating SARS-CoV-2 [bib_ref] Distinct conformational states of SARS-CoV-2 spike protein, Cai [/bib_ref]. S protein has a receptor-binding domain (RBD), an S1/S2 polybasic cleavage site and 3 O-linked glycans [bib_ref] Inhibition of SARS-CoV-2 viral entry upon blocking N-and O-glycan elaboration, Yang [/bib_ref]. All specific functional structures are the product of natural evolution.
However, compared with the original virus, Delta is special. There are many mutations happened to Delta genome. Mutations in S protein are particularly abundant. According to the genome sequence analysis on GISAID, Delta lineage has 8 spike mutations (T19R, G142D, del157/158, L452R, T478K, D614G, P681R, and D950N) [bib_ref] Clinical and virological features of SARS-CoV-2 variants of concern: a retrospective cohort..., Ong [/bib_ref] [bib_ref] Global initiative on sharing all influenza data -from vision to reality, Shu [/bib_ref]. T19R, G142D, and del157/158 are located in the N-terminal domain (NTD), and L452R and T478K are in the receptor binding domain (RBD). Together with the D614G, all of the mutations mentioned above occur on the S1 subunit. P681R and D950N occur on the S2 subunit [fig_ref] Figure 1: Amino acid changes in S protein of VOCs [/fig_ref]. It is interesting that N501Y was not found in Delta, but can be found in Alpha, Beta and Gamma variant. [bib_ref] SARS-CoV-2 N501Y variants of concern and their potential transmission by mouse, Huang [/bib_ref] [bib_ref] Higher infectivity of the SARS-CoV-2 new variants is associated with K417N/T, E484K,..., Khan [/bib_ref] Neither does the E484K [bib_ref] SARS-CoV-2 variants, spike mutations and immune escape, Harvey [/bib_ref]. Although Delta lacks these well-known mutations, that are common to multiple viral strains, it has produced many distinctive variants. For example, P681R, L452R and D950N. P681R resembled P681H of Alpha variant. L452R and D950N are new ones [fig_ref] Table 1: Characteristics of SARS-CoV-2 Variants of Concern [/fig_ref]. Distinguishing features make Delta a new viral strain.
## Clinical characteristics of delta variant
The rapid spread of Delta virus is bringing a threat to us. Infection with Delta virus, causes flu-like symptoms [bib_ref] Early Transmission Dynamics in Wuhan, China, of Novel Coronavirus-Infected Pneumonia, Li [/bib_ref]. Fever, dry cough, weakness, coughing with sputum, headache, short breath, and aching pain in muscle were considered to be the common symptoms [bib_ref] An Update on Molecular Diagnostics for COVID-19, Islam [/bib_ref]. Most patients had the first manifestation of hypoesthesia or loss of smell and taste [bib_ref] Review of the Clinical Characteristics of Coronavirus Disease 2019 (COVID-19), Jiang [/bib_ref]. Some patients with severe disease often have dyspnea or hypoxemia one week after onset, others can quickly develop into acute respiratory distress syndrome (ARDS), septic shock, metabolic acidosis, coagulation dysfunction and multiple organ failure [bib_ref] Acute respiratory failure in COVID-19: is it "typical, Li [/bib_ref]. Very few patients have the manifestations of central nervous system involvement and acral ischemic necrosis [bib_ref] Olfactory transmucosal SARS-CoV-2 invasion as a port of central nervous system entry..., Meinhardt [/bib_ref]. There was no significant difference between the clinical characteristics of Delta virus infection and the symptoms caused by other strains.
## Transmission of delta variant
According to the latest report issued by the WHO (Edition 51, published 3 August 2021), the cumulative number of patients with coronavirus infection reported globally is now nearly 197 million and the number of cumulative deaths is 4.2 million. Americas and Europe were the fastest growing regions in case incidence, and the Americas and South-East Asia Regions were reported to have the high mortality rate. Such scary number of infections is partly due to the prevalence of Delta virus.
## Global spread of delta variant
Delta was first reported in India at the end of 2020 but have spread globally. By detecting throat and nose swabs from people who aged 5 years and older in England, Paul et al. found that B.1.1.7, which first appeared in the United Kingdom, was gradually replaced by Delta. And in the United States of America, since the first Delta infected person was detected in March, the dominant virus strain has also changed from B.1.1.7 to B.1.617 and P.1. Up to now, 135 countries have reported cases of the Delta variant [bib_ref] Transmission Dynamics of an Outbreak of the COVID-19 Delta Variant B.1.617.2-Guangdong Province,..., Zhang [/bib_ref]. Delta variant has strong transmission and increasing virulence.
## Possible reasons for rapid spread of delta variant
Based on current evidence, Delta variant is more transmissible than the Alpha variant. The ratio is about 40-60%, which may be related to greater risk of hospitalization. The factors driving the recent rapid growth of Delta-associated cases may be due to the followings [fig_ref] Figure 2: Possible reasons for Delta variant rapid transmission [/fig_ref] : (1) More mutations and closer synergy. D614G exists in Delta genome. As the dominant viral strain in 2020, Delta is well known for its strong transmission and replication ability [bib_ref] Spike mutation D614G alters SARS-CoV-2 fitness, Plante [/bib_ref] [bib_ref] Evaluating the Effects of SARS-CoV-2 Spike Mutation D614G on Transmissibility and Pathogenicity, Volz [/bib_ref]. P681R changed the furin cleavage sites of the virus, which accelerated cell-cell fusion [bib_ref] SARS-CoV-2 Spike Mutations, L452R, T478K, E484Q and P681R, in the Second Wave..., Cherian [/bib_ref]. L452R increased the ability of virus to invade cells [bib_ref] The Impact of Mutations in SARS-CoV-2 Spike on Viral Infectivity and Antigenicity, Li [/bib_ref]. Although the role of many other mutations has not been studied clearly, the diversity of mutations does make Delta more susceptible; (2) Immune escape. A recent study found that SARS-CoV-2 spike L452R variant, a critical variant in Delta, evades cellular immunity and increases infectivity [bib_ref] SARS-CoV-2 spike L452R variant evades cellular immunity and increases infectivity, Motozono [/bib_ref] The same conclusion was also confirmed in another study [bib_ref] Transmission, infectivity, and neutralization of a spike L452R SARS-CoV-2 variant, Deng [/bib_ref] ;
(3) Many protective measures are ineffective for Delta.
A large number of experimental results suggested that current methods used for the prevention and treatment of infection do not work on Delta, including convalescent plasma [bib_ref] Reduced neutralization of SARS-CoV-2 B.1.617 by vaccine and convalescent serum, Liu [/bib_ref] , some monoclonal antibodies [bib_ref] Reduced sensitivity of SARS-CoV-2 variant Delta to antibody neutralization, Planas [/bib_ref] , and partial vaccines [bib_ref] Reduced sensitivity of SARS-CoV-2 variant Delta to antibody neutralization, Planas [/bib_ref] ; (4) Insufficient prevention and control. Shorten generation intervals or high transmissibility are responsible for Delta pandemic [bib_ref] Analysis of SARS-CoV-2 genomic epidemiology reveals disease transmission coupled to variant emergence..., Bandoy [/bib_ref]. Rigorous control could reduce these indicators.
## Potential strategies for combating delta variant
Variations of SARS-CoV-2 is threatening, but there are still many ways to deal with it. Some direct and indirect strategies are listed below.
## Early diagnosis
Timely and effective diagnosis is necessary for the control of infectious diseases. Diagnosis of COVID-19 is the first step in the prevention and treatment. SARS-CoV-2 mutates frequently because of its unique genome structure and replication system. Reverse transcription loop-mediated isothermal amplification (RT-LAMP) is a powerful molecular tool for detecting SARS-CoV-2 RNA, including VOCs and VOIs. Alves et.al conducted a clinical validation of colorimetric RT-LAMP, they found that RT-LAMP is a fast and sensitive diagnostic tool and can be used to detect SARS-CoV-2 variants [bib_ref] A colorimetric RT-LAMP assay and LAMP-sequencing for detecting SARS-CoV-2 RNA in clinical..., Dao [/bib_ref]. Advances of CRISPR-Cas system indicates its significance in COVID-19 diagnosis, many excellent scientists build CRISPR-Cas-based system to identify infectious people. The most representative system are DETECTER [bib_ref] CRISPR-Cas12-based detection of SARS-CoV-2, Broughton [/bib_ref] and SHERLOCK [bib_ref] Clinical validation of a Cas13-based assay for the detection of SARS-CoV-2 RNA, Patchsung [/bib_ref]. CRISPR-Cas system provides us with fast and reliable detection methods [bib_ref] COVID-19 one year later: a retrospect of CRISPR-Cas system in combating COVID-19, Zhan [/bib_ref]. Early diagnosis is part and parcel of controlling Delta.
## Vaccines and monoclonal antibodies
Vaccines is a promising method to combating virus. Vaccination frequency is an important factor affecting the efficacy of mRNA vaccine. A study from NEJM shows that effectiveness after one dose of mRNA vaccine BNT162b2 was notably lower, only 30.7%. but the effectiveness of two doses has remarkable improvement, almost 88% [bib_ref] Effectiveness of Covid-19 Vaccines against the B.1.617.2 (Delta) Variant, Lopez [/bib_ref]. At the same time, an article published in the Lancet disclosed this result. This article also believes that mRNA vaccine can fight against COVID-19, but it is best to vaccinate with two doses [bib_ref] Neutralising antibody activity against SARS-CoV-2 VOCs B.1.617.2 and B.1.351 by BNT162b2 vaccination, Wall [/bib_ref]. In addition to vaccination frequency, vaccination time and vaccination method will also have an impact on vaccine effectiveness. Zhang et al. found that vaccination in the morning produces a stronger protective immune response, which may be because the human immune system is affected by circadian rhythm [bib_ref] Time of day influences immune response to an inactivated vaccine against SARS-CoV-2, Zhang [/bib_ref]. show that ChAd-SARS-CoV-2-S, an adenovirus vector vaccine, can protect against SARS-CoV-2 invasion, and intranasally administered ChAd-SARS-CoV-2-S induces durable protection in BALB/c mice [bib_ref] An intranasal vaccine durably protects against SARS-CoV-2 variants in mice, Hassan [/bib_ref]. The effectiveness of various vaccines on B.1.617.2 were summarized in [fig_ref] Table 2: Effectiveness of various vaccines on Delta variant [/fig_ref]. Monoclonal antibodies are equally important for curbing COVID-19. Although Bamlanivimab was proved ineffective, Etesivimab, Casirivimab and Imdevimab were proved to be useful [bib_ref] Reduced sensitivity of SARS-CoV-2 variant Delta to antibody neutralization, Planas [/bib_ref].
## Structural biology assists in identifying drug targets
Before the advent of high-resolution crystal structures, it was common to combine biology, chemistry and physics to uncover the feature of proteins or small molecule compounds. However, the resolution is low and error prone, and the combination of the two substances cannot be revealed. Technological innovation has removed the obstacles for scientific research. Structural biology plays an essential role in antiviral research. structural biologists have used the advanced technology, x-ray crystallography (X-ray) and cryo-electron microscopy (cryo-EM), to analyze critical information. Yao et al. reported the molecular assembly of the authentic SARS-CoV-2 virus using cryoelectron tomography (cryo-ET) and subtomogram averaging (STA), which revealed SARS-CoV-2 panorama in a delicate manner [bib_ref] Structural insights into SARS-CoV-2 proteins, Arya [/bib_ref] [bib_ref] Molecular Architecture of the SARS-CoV-2 Virus, Yao [/bib_ref]. Further refine the application of structural biology, receptor binding region [bib_ref] Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor, Lan [/bib_ref] , neutralizing antibody structure [bib_ref] SARS-CoV-2 neutralizing antibody structures inform therapeutic strategies, Barnes [/bib_ref] , spike mutation [bib_ref] Structural biology in the fight against COVID-19, Barcena [/bib_ref] are also within its scope. Using common viral strains, structural biology has helped people identify the key sites to break the virus and try to develop relevant inhibitors [bib_ref] Structure of M(pro) from SARS-CoV-2 and discovery of its inhibitors, Jin [/bib_ref] [bib_ref] Crystal structure of SARS-CoV-2 main protease provides a basis for design of..., Zhang [/bib_ref] [bib_ref] Activity profiling and crystal structures of inhibitor-bound SARS-CoV-2 papain-like protease: A framework..., Rut [/bib_ref]. Jin et.al found more than 10,000 potentially valid compounds through structure-based drug design and screening, and six of these compounds exhibited promising M-pro inhibited activity in cell-based assays [bib_ref] Structure of M(pro) from SARS-CoV-2 and discovery of its inhibitors, Jin [/bib_ref]. Beyond that, potential inhibitors are also identified by studying the complex structure of the drug binding to the virus [bib_ref] Activity profiling and crystal structures of inhibitor-bound SARS-CoV-2 papain-like protease: A framework..., Rut [/bib_ref] [bib_ref] Mechanism and inhibition of the papain-like protease, PLpro, of SARS-CoV-2, Klemm [/bib_ref] [bib_ref] The complex structure of GRL0617 and SARS-CoV-2 PLpro reveals a hot spot..., Fu [/bib_ref] [bib_ref] Uncovering Flexible Active Site Conformations of SARS-CoV-2 3CL Proteases through Protease Pharmacophore..., Pathak [/bib_ref]. Although the research on antiviral drugs is still making efforts to be closer to clinical transformation, some non-antiviral drugs, such as Branebrutinib, have been successfully entered into clinical trials based on the help of structural biology [bib_ref] Discovery of Branebrutinib (BMS-986195): A Strategy for Identifying a Highly Potent and..., Watterson [/bib_ref]. Therefore, we have reasons to trust structural biology to determine drug targets for Delta variant.
## Epidemic prevention and control
To protect ourselves from infected by Delta, we should not only rely on external measures, but also improve self-protection awareness. A large amount of evidence shows that the strength of government prevention and control, and the degree of citizens' compliance are positively correlated with the infection. Delta strain is special, but its biological characteristics have not undergone subversive changes. For this reason, wear masks, wash hands frequently, keep social distance and avoid crowded gathering are still worth advocating.
# Conclusions
After invading China, COVID-19 fights the world and defeats the defense lines of all countries, forcing WHO to declare that COVID-19 has reached the highest level of infectious disease. COVID-19 has spread worldwide over a year, and the extent and severity of the outbreak is deeply concerned. Many methods have been proposed to combat the virus, such as reduce exposure, drug treatment and vaccination, that can effectively delay the spread of the virus. While microorganisms are alive, they evolve in order to survive and reproduce, so is SARS-CoV-2. A lot of characteristic variations have been derived in the process of SARS-CoV-2 evolution. Delta is a SARS-CoV-2 variant originally found in India. It has spread to more than 135 countries in just half a year. At present, Delta has become the main variant in the world. The scientific community's understanding of Delta is still limited except its known infectious power. It is necessary to systematically expound the characteristics and global impact of Delta. In this article, we reviewed three major sections describing different aspects of Delta variant.
Firstly, we introduced features of Delta, including its biological features and clinical characteristics. Delta is a variant of SARS-CoV-2, similar to the original virus. But has its own distinguished feature. There are many remarkable mutations happened in S protein, for example, L452R and D950N. These mutations work together to make Delta virus become variant of concern.
Secondly, we described the dissemination of virus. Delta variant has become the leading variation and has involved in 135 countries. At the same time, a large number of studies show that the ultra-fast transmission speed not only increases the risk of infection, but also increases the hospitalization rate. Therefore, we summarized some reasons for this situation so that we can better understand and conquer Delta. Finally, we proposed some possible measures for the treatment and prevention of Delta. We suggest that early diagnosis is important for control the infectious disease. Although there is no specific drug, vaccination is a reassuring treatment. There is a lot of evidence that the antibodies produced after vaccination can neutralize Delta variant. In addition, the latest research suggests that chimeric spike mRNA vaccines may be the direction to overcome the virus [bib_ref] Chimeric spike mRNA vaccines protect against Sarbecovirus challenge in mice, Martinez [/bib_ref]. Using structural biology to find potential drug target is also suggested. Last but not least, improving self-protection awareness never go out of style.
In conclusion, understanding the relevant information of Delta will help us fight it better. Only by knowing ourselves and others can we defeat the virus.
## Abbreviations
SARS-CoV-2: Severe acute respiratory syndrome coronavirus 2; COVID-19: coronavirus disease 2019; VOC: variant of concern; WHO: World Health Organization; VOI: variant of interest; β-CoV: beta coronavirus; ORFs: open reading frames; NSP: non-structural protein; S: Spike protein; RBD: receptor-binding domain; NTD: N-terminal domain; ARDS: acute respiratory distress syndrome; RT-LAMP: reverse transcription loop-mediated isothermal amplification; CRISPR: clustered regularly interspaced short palindromic repeats; X-ray: X-ray crystallography; cryo-ET: cryoelectron tomography; STA: subtomogram averaging.
[fig] Figure 1: Amino acid changes in S protein of VOCs. Description of SARS-CoV-2 spike mutation in different virus strain. (A) B.1.1.7; (B) B.1.351; (C) P.1; (D) B.1.617.2. The colored columns describe the structural domain of spike protein. NTD: N-terminal domain; RBD: receptor-binding domain; HP1: heat protein 1; HP2: heat protein 2. [/fig]
[fig] Figure 2: Possible reasons for Delta variant rapid transmission. This figure shows three reasons that explain the rapid spread of Delta variant and its emergence as the dominant mutant in many countries. The first frame suggested that the increased mutations in spike protein of Delta made contributions to transmission, the second frame depicts the occurrence of immune escape in Delta variant, and the third frame showed the decreased efficacy of convalescent plasma, monoclonal antibodies and vaccines. All above were potential reasons. [/fig]
[table] Table 1: Characteristics of SARS-CoV-2 Variants of Concern [/table]
[table] Table 2: Effectiveness of various vaccines on Delta variant [/table]
|
Metabolic interactions between dynamic bacterial subpopulations
eLife's transparent reporting formWe encourage authors to provide detailed information within their submission to facilitate the interpretation and replication of experiments. Authors can upload supporting documentation to indicate the use of appropriate reporting guidelines for health-related
## Sample-size estimation
- You should state whether an appropriate sample size was computed when the study was being designed - You should state the statistical method of sample size computation and any required assumptions - If no explicit power analysis was used, you should describe how you decided what sample (replicate) size (number) to use Please outline where this information can be found within the submission (e.g., sections or figure legends), or explain why this information doesn't apply to your submission:
## Replicates
- You should report how often each experiment was performed - You should include a definition of biological versus technical replication - The data obtained should be provided and sufficient information should be provided to indicate the number of independent biological and/or technical replicates - If you encountered any outliers, you should describe how these were handled - Criteria for exclusion/inclusion of data should be clearly stated - High-throughput sequence data should be uploaded before submission, with a private link for reviewers provided (these are available from both GEO and ArrayExpress)
Please outline where this information can be found within the submission (e.g., sections or figure legends), or explain why this information doesn't apply to your submission:
No explicit power analysis was done prior to each experiment as these experiments rely on single cell data from thousands of individual cells which vary from experiment to experiment. Our use of 600 to several thousands of single cells in these experiments amounts to confidence levels 95% with 4% or better confidence interval per each replicate (much higher for most experiments which used thousands of cells). Each experiment was performed with at least 3 biological replicates.
The methods or figure legends explain that each experiment was repeated at least three times. This refers to biological replicates. Biological replicates used a different bacterial inoculum and were grown in different (distinct) culture vessels. For microbial pad and snapshot data from figures 2 panel C and figure 4 the raw data from three separate replicates is provided in the supplementary. For
## Statistical reporting
- Statistical analysis methods should be described and justified - Raw data should be presented in figures whenever informative to do so (typically when N per group is less than 10) - For each experiment, you should identify the statistical tests used, exact values of N, definitions of center, methods of multiple test correction, and dispersion and precision measures (e.g., mean, median, SD, SEM, confidence intervals; and, for the major substantive results, a measure of effect size (e.g., Pearson's r, Cohen's d) - Report exact p-values wherever possible alongside the summary statistics and 95% confidence intervals. These should be reported for all key questions and not only when the p-value is less than 0.05.
Please outline where this information can be found within the submission (e.g., sections or figure legends), or explain why this information doesn't apply to your submission:
(For large datasets, or papers with a very large number of statistical tests, you may upload a single table file with tests, Ns, etc., with reference to sections in the manuscript.)
## Group allocation
- Indicate how samples were allocated into experimental groups (in the case of clinical studies, please specify allocation to treatment method); if randomization was used, please also state if restricted randomization was applied - Indicate if masking was used during group allocation, data collection and/or data analysis Please outline where this information can be found within the submission (e.g., sections or figure legends), or explain why this information doesn't apply to your submission:
Additional data files ("source data") - We encourage you to upload relevant additional data files, such as numerical data that are represented as a graph in a figure, or as a summary table - Where provided, these should be in the most useful format, and they can be uploaded as "Source data" files linked to a main figure or table - Include model definition files including the full list of parameters used - Include code used for data analysis (e.g., R, MatLab) - Avoid stating that data files are "available upon request" Please indicate the figures or tables for which source data files have been provided:
Data is reported in the figure legends, and when numbers are reported in the main text the error and appropriate designation (mean) is included. P values for expression data, is included in supplementary tables 2 and 3.
No masking or grouping was used. |
Pseudoangiomatous Stromal Hyperplasia (PASH) of the Breast: An Uncommon Finding in an Uncommon Patient
Objective:Unknown ethiology Background:Pseudoangiomatous stromal hyperplasia (PASH) is an uncommon benign breast lesion. Case Report:PASH is reported in a young female in treatment for neurological diseases with multi-drug therapy (clonazepam, valproate and risperidone). Her menstrual cycles are irregular, and she reached menarche very late.Conclusions:The higher PASH prevalence in premenopausal woman (the majority of whom are actively taking oral contraceptive pills), in 24% to 47% of men with gynecomastia and during pregnancy supports a hormonal etiology; the interaction between clonazepam, valproate, risperidone and progesterone could increase the level of progesterone that could stimulate PASH growth.
# Background
Pseudoangiomatous stromal hyperplasia (PASH) in an uncommon benign breast lesion that generally presents as a fastgrowing palpable lesion or gynecomastia. PASH etiology is unknown, but hormonal influences are supposed. Our case report deals with a young woman with irregular menstrual cycles, delayed menarche and multi-drug therapy, who developed a breast lump increasing in volume.
## Case report
A 30-year-old female presented to our hospital with a left breast mass in February 2017. Her past medical history included intellectual disability and epilepsy from age of 2, secondary to neonatal hypoxia. She was receiving sodium valproate, phenobarbital, clonazepam and risperidone. She had no family history of breast disease or ovarian cancer. Her body weight was normal (body mass index 21 kg/m 2 ). She reached menarche at 22 years; her parents referred the use of estro-progestinic pills before the age of 22 years, unsuccessful in drug-induced menstrual flow. At PASH diagnosis, her menstrual cycle was irregular with a normal menstrual flow every 40 to 60 days. She had no pregnancies in the past.
Physical examination showed a 3 cm fixed and firm lump with a not-defined border below left nipple areola complex. Ultrasound examination showed a solid, not homogenous lesion measuring 20×10 mm in size in left Q3-5 with mild peripheral vascularization at color-doppler evaluation, irregular margins, ductal ectasia, without acoustic shadowing. There was likely bilateral reactive axillary lymphadenopathy. Fibrous mastopathy signs and fibroadenomas were found in both breasts. The ultrasound finding was suspicious for malignancy. Under ultrasonography guidance, core needle biopsy was performed, and histopathologic report showed benign fibroepithelial node, enriched in vascular space, suggestive for PASH. Considering the breast mass dimension, lumpectomy was performed. During surgery, the lump was hardly dissociable from adjacent tissue. The tumor resected from the left breast measured 25 mm. She was discharged on day 1 after surgery. Pathology confirms hamartoma-like fibroepithelial lesion with PASH features. No recurrence of disease is observed at 32 months follow-up.
# Discussion
Pseudoangiomatous stromal hyperplasia (PASH) is an uncommon breast benign stromal lesion, first described by Vuitch in 1986. Incidental microscopic PASH can be found in up to 23% of consecutive breast specimens. In contrast, PASH as a main pathological finding is a rare entity. Preoperative core biopsy fails to diagnose PASH in 35% of cases. Ultrasound imaging is not specific: generally, it appears like a hypo-echoic ovoidal mass with regular margin. Treatment is usually surgical excision, although "watch and wait" strategy can be applied if PASH diagnosis is made on core biopsy and if the lump is less than 2 cm in size. Rapidly increased lump needs surgical approach to valuate contextual presence of ductal carcinoma in situ (DCIS). Recurrence rate is 9% to 21%, probably due to the persistence of a residual mass after surgery. PASH is related to benign and malign breast lesions in up to 23% of cases but is not associated with an increased risk for malignity, rather it seems to be protectivealthough 2 case of synchronous tumoral PASH in the breast and axillary tissue are reported in literature. On the other hand, there is an increased ipsilateral breast cancer prevalence in patients treated for PASH more than 5 years after PASH biopsy.
PASH is a complex network of slit-like spaced lined by endothelial-like spindle cells surrounded by dense collagenous stroma. Fibroblasts and myofibroblasts proliferation and collagen over-secretion create a solid tissue with cystic areas resembling ectatic vessels (pseudo vascular spaces).
PASH stromal cells nuclei express high density progesterone receptors; estrogen receptors expression is more variable. Normal mammary stroma that showed no progesterone receptor staining instead. Progesterone is metabolized by cytochrome P450, that is inhibited by clonazepam; valproate and risperidone are metabolized by cytochrome P450, so it could be a competition for the cytochrome between our patient multidrug therapy and progesterone. These interactions could increase the level of progesterone that could stimulate PASH growth. A hormonal etiology is also supported by the higher PASH prevalence in premenopausal woman (the majority of whom are actively taking oral contraceptive pills), in 24% to 47% of men with gynecomastiaand by reported case of PASH in a transgender male during hormone therapyor during pregnancy. PASH affects pre-menopausal women or menopausal women in hormone replacement therapy (HRT). Anti-hormonal therapy could theoretically reduce PASH mass: a case report suggests tamoxifen use as alternative approach in management of PASH but the lack of evidence does not support anti-hormonal therapy as adjuvant therapy.
# Conclusions
PASH is an uncommon but benign lesion of the breast. Its origin is still unclear, but it seems to be related to hormone stimulation (primarily progesterone stimulation).
Efforts are necessary to recognized PASH lesion during diagnostic examinations (ultrasound, breast core-biopsy) in order to consider watch and wait management, procrastinating surgery if not necessary. A multidisciplinary approach is the best way to manage breast lesions, including PASH.
Knowledges about risk factor (related to hormonal etiology previously described) is essential to report suspected PASH lesion to the pathologist and help him to distinguish PASH from invasive cancer and support a conservative management if possible.
## Department and institution in which the work was done
Department of Obstetrics and Gynecology, Umberto I Hospital, University of Turin, Italy. |
Peptide Signaling Pathways Regulate Plant Vascular Development
Plant small peptides, including CLAVATA3/EMBRYO SURROUNDING REGION-RELATED (CLE) and Epidermal Patterning Factor-Like (EPFL) peptides, play pivotal roles in coordinating developmental processes through cell-cell communication. Recent studies have revealed that the phloem-derived CLE peptides, CLE41/44 and CLE42, promote (pro-)cambial cell proliferation and inhibit xylem cell differentiation. The endodermis-derived EPFL peptides, EPFL4 and EPFL6, modulate vascular development in the stem. Further, several other peptide ligands CLE9, CLE10, and CLE45 play crucial roles in regulating vascular development in the root. The peptide signaling pathways interact with each other and crosstalk with plant hormone signals. In this mini-review, we summtarize the recent advances on peptides function in vascular development and discuss future perspectives for the research of the CLE and EPFL peptides.
# Introduction
Intercellular communication and interaction are fundamentally important for plants to finetune growth and development in response to ever-changing environmental conditions. Secreted signaling peptides play crucial roles in regulating vascular development by short-range intercellular communication . Peptides are often referred to as proteins smaller than 100 amino acids and derived from precursor proteins or directly translated from short open reading frames embedded in longer transcripts. The CLAVATA3/EMBRYO SURROUNDING REGION-RELATED (CLE) peptides are essential for plant development and stress responses. The Arabidopsis genome contains 32 CLE genes encoding 27 unique peptides. These CLE peptides are recognized at the cell surface by transmembrane receptors and activate intracellular signal transduction, therefore regulating plant growth and development. In this mini-review, we summarize the function and interaction of peptide signaling pathways in vascular developmentand discuss the gaps in current knowledge and provide perspective for future research.
## Vascular development
Plant vasculatures transport water and nutrients to all developing organs. In the Arabidopsis stem, the vascular tissues are organized as bundles, in which the procambium or cambium resides in the middle, xylem inside, and phloem outside of the bundle. At the shoot tip, vascular bundles are well separated along the stem periphery. In the more mature stem, cambium starts developing in positions between bundles, connects vascular cambium, and forms a ring of cambial cells. At the bottom of the stem, cambium cells divide and form xylem precursor cells inward and phloem precursor cells outward of the stem. Eventually, these precursors differentiate into secondary xylem and secondary phloem, respectively. Other plant organs, such as the root and hypocotyls, also go through secondary growth depending on the activity of cambium cells. Regulation of vascular development is vital for plant secondary growth and crop production and therefore remains a critical research area in plant biology.
## Tdif-pxy signaling regulates vascular development
The proliferation of cambial stem cells is essential for vascular development and the mechanical strength of the plant stem. Peptide signals play crucial roles in regulating vascular development through cell-to-cell communication. The tracheary element differentiation inhibitory factor (TDIF) peptide is so far the best-understood peptide in this process. Tracheary element differentiation inhibitory factor is synthesized from the phloem and travels to the cambium, where TDIF binds to its receptor PHLOEM INTERCALATED WITH XYLEM/TDIF RECEPTOR (PXY/TDR) on the plasma membrane. Overexpression of TDIF coding genes CLE41 and CLE44 or exogenous application of synthetic TDIF peptide promote procambial cells proliferation and inhibit xylem cell differentiation. These experiments established the function of the TDIF-PXY module in cambial cell proliferation, xylem differentiation, and vascular patterning. Downstream of the TDIF-PXY module is two WUSCHELrelated HOMEOBOX (WOX) transcription factors, WOX4 and WOX14. WOX4 is expressed mainly in the procambium and cambium, and its expression level is upregulated by applying CLE ligand. WOX14 acts redundantly with WOX4 in regulating vascular cell division. The TDIF-PXY-WOX signaling plays a crucial role in the maintenance of the vascular meristem during secondary growth.
The crystal structure of the extracellular domain of PXY/ TDR in complex with TDIF peptide has been determined. The extracellular domain of PXY/TDR adopts a super-helical structure comprising 22 leucine-rich repeats (LRRs) and specifically recognizes TDIF by its inner concave surface. The TDIF peptide mainly adopts a "Ω"-like conformation binding the inner surface of the LRRs. The highly conserved amino acids of TDIF, including Hyp4, G6, Hyp7, and P9, are critical for the interaction with PXY/TDR. However, there are some discrepancies regarding which amino acids of PXY/TDR are essential to the interaction with TDIF. Structurebased sequence alignment showed that the TDIF-interacting motifs are conserved among plant species, suggesting a conserved TDIF-PXY recognition mechanism.
The TDIF-PXY-WOX4 signaling is conserved in plants. The TDIF peptide was first discovered from an analysis in zinnia. Later, TDIF peptide was found in Arabidopsisand the genomes of Euphyllophytes. The function of TDIF-PXY-WOX4 signaling has been investigated in poplar. Transgenic studies showed that the tissue-specific expression of PttPXY and PttCLE41 genes resulted in enhanced secondary growth in hybrid aspen, indicating that TDIF-PXY signaling is conserved in poplar. Consistent with this study, PttWOX4 gene is specifically expressed in the cambial region and controls cell division activity in the vascular cambium.
In a recent study, a regulatory network downstream of TDIF-PXY was identified using an enhanced yeast one-hybrid system. Among the transcription factorpromoter interactions, WOX14 positively regulated the expression of TARGET OF MONOPTEROS 6 (TMO6), which encodes a Dof-type transcription factor. In addition, WOX4 and TMO6 both positively regulate the expression of a third transcription factor gene, LATERAL ORGAN BOUNDARIES DOMAIN4. These data support a mechanism that this feed-forward loop controls vascular stem cell proliferation and helps to define the function of PXY signaling in determining the shape of the vascular bundle.
The TDIF-PXY signaling module regulates cell division in the (pro-)cambium and controls vascular organization via interacting with other factors. The SOMATIC EMBRYOGENESIS RECEPTOR KINASE (SERK) family regulates plant growth and development, as well as many other biological processes. The SERK genes are expressed in procambial cells, suggesting that the SERK family may have a function in regulating vascular development. Indeed, the TDIF receptor PXY forms heterodimers with SERKs, including SERK1, SERK2, and SERK3 (BAK1), in regulating procambial cell proliferation. A NAC domain transcription factor XVP is localized on the plasma membrane and forms a complex with TDIF co-receptor PXY-BAK1. The XVP is expressed in the cambium and regulates xylem differentiation and vascular patterning. Further study showed that XVP regulates xylem differentiation through a crucial factor VASCULAR-RELATED
## Epfl4/6-er signaling in the regulation of vascular cambium
The EPIDERMAL PATTERNING FACTOR-LIKE 4 (EPFL4) and EPFL6/CHALLAH (CHAL) secretory peptides were first identified as regulators of inflorescence development through activating the ER signaling. Two additional ER paralogs, ER-LIKE 1 (ERL1) and ERL2, together with ER, play crucial roles in regulating inflorescence architecture, organ shape, and epidermal stomatal patterning in Arabidopsis. In vascular development, the endodermis-produced EPFL4 and EPFL6 peptides function as ligands for phloem-located ER in regulating procambial cell division and stem elongation. Therefore, the cell-to-cell communication between the endodermis and the phloem plays a crucial role in procambial development . Both ER and ERL1 are expressed in inflorescence stem and function redundantly in regulating procambial development and xylem differentiation. The mpk3/mpk6 and mkk4/mkk5 mutants showed short pedicels and a corymb-like inflorescence similar to the phenotypes of the er mutant. Further evidence showed that the MAPK cascade functions downstream of the ER receptor in regulating pedicel length and inflorescence architecture. Therefore, the MAPK cascade, involving YDA-MKK4/MKK5-MPK3/MPK6, functions downstream of ER in regulating cell proliferation and plant organ development. It would be interesting to investigate if this MAPK pathway also regulates procambium development in the stem and hypocotyl.
## Other cle peptide signaling pathways in vascular development
Besides CLE41/CLE44, EPFL4, and EPFL6, other peptides, including CLE9/10, CLE25, and CLE45, also play critical roles in vascular development. The CLE9 and CLE10 genes are preferentially expressed in xylem precursor cells of the root vasculature. Interestingly, the peptide CLE9/10 forms a complex with the receptor kinase HAESA-LIKE 1 (HSL1) to regulate stomatal lineage cell division. In the root xylem development, CLE9/10 is perceived by BARELY ANY MERISTEM (BAM) class receptor kinases in regulating periclinal cell division;. BARELY ANY MERISTEM 1 and its homologs, BAM2 and BAM3, are class XI LRR receptor-like kinases similar to CLV1 and PXY. The bam1 bam2 and bam1 bam3 mutants exhibited increased xylem files similarly to the cle9 mutant. Indeed, physical interaction between CLE9/10 and BAM1 was demonstrated by three independent approaches, confirming the BAMs as CLE9 receptors in root xylem development. The CLE45 peptide forms a complex with BARELY ANY MERISTEM 3 (BAM3) to regulate Arabidopsis protophloem differentiation. Protophloem consists of narrow thin-walled cells and is developed from procambium. Studies revealed that CLE45 is expressed from protophloem sieve elements in root meristems and perceived fully by CLAVATA 2 (CLV2) and the CORYNE (CRN) receptor proteins. CRN is required for BAM3-mediated CLE45 signaling via stabilizing BAM3 expression. In addition, a protophloem-specific CRN expression rescues the resistance of the crn mutant to root-active CLE peptides. Collectively, these data support a vital function of the CLE45 and CLV/BAM receptors in vascular development. Downstream of BAM3, a MEMBRANE-ASSOCIATED KINASE REGULATOR protein, positively regulates CLE45 signaling. Further, OCTOPUS (OPS), a polarly localized membrane-associated protein, promotes root protophloem differentiation by antagonizing CLE45 signaling through direct interference with BAM3 and CLV2-CRN;.
The CLE25 and CLE26 play crucial roles in regulating root protophloem development. CLE25 expression is tightly associated with phloem initiation, and the loss-of-function mutant cle25 shows a delayed protophloem differentiation phenotype. Genetic screening for suppressors identified CLV2 as a potential receptor for CLE25. Further peptide assay and protein interaction experiment found that CLE-RESISTANT RECEPTOR KINASE (CLERK) interacts with CLV2 and possibly forms a receptor complex for CLE25. However, direct binding between CLE25 and the CLERK-CLV2 receptor complex remains to be established experimentally.
The CLE45 and CLE25/26 peptides are all expressed in root and repress phloem development as shown by expression analyses and synthetic peptides treatment experiments. CLE25 may be involved in the formative cell division during protophloem differentiation. It is still unclear which division step is regulated by CLE25 in the two successive formative cell division steps during protophloem development. In addition, there are some differences in terms of perception of these peptides. The CLE25 and CLE26 peptides could be perceived by CLERK, while CLE45 is perceived by BAM3. Further, CLE25 may be involved in response to detrimental environmental signals. Further investigation is needed to clarify the functional difference of these peptides in protophloem development.
## The crosstalk between tdif-tdr signaling and other signaling pathways
The development of vascular tissue involves interactions between TDIF-PXY and other peptide signaling pathways. The receptor kinases PXY and ER genetically interact to coordinate cell division and organization. Gene expression analysis suggested that PXL1 and PXL2 expression is upregulated in er mutant stems. The expression of EPFL4 and EPFL6 is altered in pxf (mutation of the PXY family, including pxy, pxl1, and pxl2) and er pxy mutants. Further study showed that the mutant of er pxf had considerably fewer cells per vascular bundle than either pxf, er, or pxy er, suggesting that PXL1 and PXL2 are functionally redundant with ER to regulate vascular proliferation in the stem. These results demonstrate that PXY and ER coordinately regulate vascular proliferation and organization via inter-tissue signaling.
TDFI-PXY signaling also interacts with plant hormonal signals, such as brassinosteroids (BR), auxin, and ethylene signaling. The GLYCOGEN SYNTHASE KINASE 3 (GSK3) family members, including BRASSINOSTEROID-INSENSITIVE 2 (BIN2), are crucial downstream components of the TDIF signaling pathway in suppressing xylem differentiation. The PXY/TDR receptor interacts with GSK3s on the plasma membrane and activates GSK3s in a TDIF-dependent fashion. The interaction between PXY/ TDR and BIN2 inhibits the transcription factor BES1 to prevent procambial cell differentiation into the xylem. Recently, an additional study showed that the GSK3 protein BIN2-LIKE 1 (BIL1) integrates the PXY/TDR signaling through phosphorylating the auxin response factor MP/ARF5, which limits the effect of MP/ARF5 on ARR7 and ARR15 expression, thus enhancing vascular cambial activity. Together, these results suggest that BIL1 mediated the crosstalk between peptide signaling and hormone signaling to maintain the cambial activity;. Ethylene promotes cambium activityand interacts with TDIF-PXY signaling. The interaction between TDIF-PXY and plant hormone may help plants better perceive and respond to longdistance developmental and environmental signals.
## The crosstalk between epfl4/6-er signaling and other signaling pathways
The LRR receptor-like kinase ER is a stem growth regulator and involves in diverse developmental processes, including epidermal and mesophyll development, stomatal density, and vascular development. For example, hydrogen peroxide induces cortex proliferation, while SPINDLY, an O-linked glucosamine acetyltransferase, functions downstream of ER to fine-tune cortex proliferation, indicating a connection between ER signaling and cellular redox homeostasis. Another study showed that an LRR receptor-like kinase SUPPRESSOR OF BIR1-1, previously known to control plant immune responses and abscission, appears downstream of BREVIPEDICELLUS (BP) and ER to regulate the precocious differentiation of xylem fiber. In addition, ER signaling pathway and the SWR1 chromatin remodeling complex jointly regulate inflorescence architecture by promoting the expression of the PACLOBUTRAZOL RESISTANCE (PRE) gene family. Further study showed that a growth-related transcription factor HOMOLOG OF BEE2 INTERACTING WITH IBH1 (HBI1) functions downstream of PRE1 to regulate inflorescence architecture by promoting pedicel elongation. These studies indicate potential crosstalks between ERFL4/6 and other pathways in regulating vascular development.
# Conclusion
Significant progress has been made in our understanding of peptide signaling, including the TDIF-PXY/TDR module and the EPFL4/6-ER/ERL1 module, in regulating vascular development. While the TDIF-PXY signaling in vascular development has been the subject of much research since the discovery of TDIF, many questions remain. For instance, compared to the wild type, pxy mutants showed more fiber cells indicating that PXY has a potential function in regulating xylem development. Moreover, environmental factors, including light, carbon dioxide concentration, and drought, all affect the expression pattern of several EPFL family genes, indicating that ER signaling could respond to external environmental stimuli. Therefore, studying the crosstalk between environmental factors and peptide signaling in vascular development is an exciting field in the future.
Further, the EPFL4/6-ER/ERL1 module regulates procambial maintenance in inflorescence stems, and a MAPK signaling module YDA-MKK4/5-MPK3/6 functions downstream of ER. Identification of the substrates of MPK3/6 would be crucial to understand the molecular mechanism of the EPFL-ER signaling. The crosstalk between TDIF-PXY and EPFL-ER signaling has been proposed, but the mechanistic interaction has not been established. It is possible that a common downstream signaling hub may connect both pathways in regulating cambial activity.
In addition, the BAM1 receptor can potentially bind both CLV3 and CLE25 peptides, indicating that the same receptors can perceive different CLE peptides in various biological contexts. In contrast, the CLE9/10 peptide ligand is recognized by two different receptor kinases HSL1 and BAM3 in regulating stomatal lineage cell division and periclinal cell division of xylem precursor cells, respectively. Therefore, the same receptor can perceive different CLE peptides, while the same peptide can bind to different receptors, indicating extremely complex peptide signal transduction networks in plants. The complexity may be necessary for peptide signaling to have developmental and organ-specific functions. Further dissection of the peptide signaling pathways will provide new insight into our understanding of vascular development.
# Author contributions
BY and HW wrote the review. All authors contributed to the article and approved the submitted version.
# Funding
This work is supported by the National Science Foundation (IOS-1453048) and in part, by USDA NIFA Hatch project #CONS00990 to HW. |
HIV Reservoir Decay and CD4 Recovery Associated With High CD8 Counts in Immune Restored Patients on Long-Term ART
Background: Whether varying CD8 counts influence the human immunodeficiency virus (HIV) reservoir and CD4 restoration in patients with CD4 counts ≥ 500 cells/µL after long-term antiretroviral therapy (ART) remains unknown. In this study, we analyzed relationships between CD8 levels and viral reservoir decay or CD4 recovery in immune restored patients on long-term ART.Methods: Chronic HIV-infected patients who received 5 years of ART with CD4 counts ≥ 500 cells/µL were grouped according to CD8 counts: CD8 <500 (Group 1), 500-1,000 (Group 2), and ≥1,000 cells/µL (Group 3). CD4 recovery, viral decay, CD8 T-cell function, and their correlations were analyzed during ART among the three groups.Results: Dynamics of viral decay and CD4 recovery were different among the three groups. Both viral decay and CD4 recovery were higher in Group 3 than the other two groups after 5 years of ART, mainly during years 3-5 of ART. Higher expression levels of Ki67 while PD-1 levels were lower on CD8 T-cells in Group 3 compared with the other groups, and Group 3 showed stronger CD8 T-cells functional capacity after 3 years of ART. Reduced HIV DNA levels and increased CD4 counts between years 3 and 5 of ART were positively correlated with CD8 counts and function.Conclusions: High CD8 counts are beneficial for persistent viral decay and CD4 recovery in immune restored patients during long-term ART.
# Introduction
Antiretroviral therapy (ART) has markedly increased CD4 counts in HIV-infected patients and reduced AIDS-related morbidity and mortality. In the past, significantly more attention has been paid to CD4 counts while the CD8 T-cell compartment was relatively underappreciated. CD4 recovery to above 500 cells/µL is frequently observed in the patients under effective ART, but CD8 counts are consistently elevated even after long-term treatment. Elevation of CD8 counts is associated with increased immune anergy and risks of non-AIDS-related clinical events in HIV-infected patients during ART [bib_ref] CD4/CD8 ratio normalisation and non-AIDS-related events in individuals with HIV who achieve..., Mussini [/bib_ref].
The viral reservoir size and CD4 counts are two important measurements to determine the effect of ART after plasma viral loads remain at non-detectable levels. Previous studies showed that CD8 counts were positively correlated with the viral reservoir [bib_ref] Antiretroviral therapy initiated within 6 months of HIV infection is associated with..., Jain [/bib_ref] [bib_ref] Plasma indoleamine 2,3-dioxygenase activity is associated with the size of the human..., Chen [/bib_ref] , and high CD8 counts were associated with a poor increase in CD4 T-cells during ART [bib_ref] Low-level viremia and proviral DNA impede immune reconstitution in HIV-1-infected patients receiving..., Ostrowski [/bib_ref]. These results were partly attributed to CD8 T-cell exhaustion and a decrease in functional capacity [bib_ref] Elevation and persistence of CD8 T-cells in HIV infection: the Achilles heel..., Cao [/bib_ref]. During HIV infection, poor CD8 T-cell response to control virus may be due to absence of CD4 T-cell help. However, CD8 T-cells play an important role in the control of HIV replication [bib_ref] Mechanisms of CD8(+) T cellmediated suppression of HIV/SIV replication, Mcbrien [/bib_ref] and HIV reservoir [bib_ref] CD8(+) lymphocytes are required for maintaining viral suppression in SIV-infected macaques treated..., Cartwright [/bib_ref] [bib_ref] CTL-mediated immunotherapy can suppress SHIV rebound in ART-free macaques, Fan [/bib_ref]. Previous studies have focused on the impact of different levels of CD8 T-cells on virus and CD4 recovery but did not consider if CD4 counts rebounded to normal levels [bib_ref] Antiretroviral therapy initiated within 6 months of HIV infection is associated with..., Jain [/bib_ref] [bib_ref] Low-level viremia and proviral DNA impede immune reconstitution in HIV-1-infected patients receiving..., Ostrowski [/bib_ref]. CD4 recovery to >500 cells/µL is a standard marker for immune restoration [bib_ref] Combining CD4 recovery and CD4: CD8 ratio restoration as an indicator for..., Lee [/bib_ref]. For patients with CD4 recovery above 500 cells/µL after long-term ART, the impact of different CD8 levels on viral reservoir decay and CD4 recovery is largely unknown.
In this study, we enrolled patients with immune restoration and subdivided them into three groups according to CD8 counts after 5-years of ART. We observed different patterns of viral decay and CD4 recovery among these three groups. Counter to the tendencies observed during the first year of ART, the CD8 high expressing group demonstrated persistent viral reservoir decay and an additional increase in CD4 counts compared to the other two groups during years 3-5 of ART, and this phenomenon is associated with increased CD8 counts and functional capacity. These findings suggest that high CD8 counts may sustain viral reservoir decay and CD4 recovery in immune restored patients during long-term effective ART.
# Methods
## Study design and participant
HIV/AIDS patients who underwent 5-years of ART during the period of January 2010 to May 2019 in Fifth Medical Center of Chinese PLA General Hospital were enrolled in this study. We excluded patients with missing CD4 data, death, drug resistance, irregular medication, and HIV-related opportunistic infections. The remaining 203 patients were divided into groups according to different CD4 and CD8 levels after 5-years of ART [fig_ref] FIGURE 1 |: Flow chart of participants enrolled in this study [/fig_ref]. Participants were eligible for our study if they achieved immune restoration (undetectable plasma HIV RNA levels within 6 months of ART and two CD4 counts >500 cells/µL after 5-years of ART). Unlike the uniform standard for CD4 T-cell stratification in HIV/AIDS, there is no uniform standard for the categorization of CD8 T-cells at present. In our study, CD8 counts were stratified based on relevant clinical studies [bib_ref] Elevation and persistence of CD8 T-cells in HIV infection: the Achilles heel..., Cao [/bib_ref] [bib_ref] CD8+ lymphocyte phenotype and cytokine production in long-term nonprogressor and in progressor..., Zanussi [/bib_ref] [bib_ref] CD4:CD8 ratio and cd8 count as prognostic markers for mortality in human..., Trickey [/bib_ref]. Following these selection criteria, 81 eligible participants were grouped on the basis of three different categories of two CD8 counts after 5-years of ART: 18 in Group 1 (CD8 counts <500 cells/µL, G1), 45 in Group 2 (CD8 counts from 500 to < 1000 cells/µL, G2), and 18 in Group 3 (CD8 counts ≥1,000 cells/µL, G3). There were 15 (4 without baseline samples), 19 (2 without baseline samples), and 12 (3 without baseline samples) participants with serial samples in G1, G2, and G3, respectively [fig_ref] FIGURE 1 |: Flow chart of participants enrolled in this study [/fig_ref].
This study was approved by the institutional review boards of Fifth Medical Center of Chinese PLA General Hospital and the study subjects gave informed consent in line with the Declaration of Helsinki.
## Data and sample collection
Data were collected as follows: sex, age at ART initiation, route of infection, CD4 counts, CD8 counts, viral load, and ART regiment. At the same time, blood samples were obtained pre-ART (if available), and at years 1, 3, and 5 on treatment. Peripheral blood mononuclear cells (PBMCs) were stored in liquid nitrogen before use. Plasma samples were frozen at −80 - C until use.
Plasma HIV RNA, CD4, and CD8 Counts Determination CD4 and CD8 counts were determined by flow cytometry (FACS Caliber, BD Biosciences, New Jersey, USA), and HIV RNA load was quantified using the COBAS AmpliPrep/TaqMan real-time RT-PCR Test (Roche, CA, USA).
## Hiv dna quantification
Extraction of total cellular DNA from thawed PBMCs was measured by QIAamp DNA Mini Kit (Qiagen, Valencia, California). HIV DNA quantitative detection kit (SUPBIO, Guangzhou, China) was used to amplify and quantify the cellassociated HIV DNA. The lower limit of detection was 20 copies/1 × 10 6 cells, and the quantification range was 50∼1 × 10 6 copies/1 × 10 6 cells.
# Flow cytometric analysis
The following antibodies were used in three panels: CD3 BV421 (clone SK7), CD3 BV510 (clone HIT3a), CD8 APC-H7 (clone SK1), CD4 APC-H7 (clone RPA-T4), CD45RA BV510 (clone HI100), CCR7 PE-Cy7 (clone 3D12), HLA-DR BV421 (clone G46-6), and Ki67 BV421 (clone B56) were obtained from BD Bioscience. CD4 FITC (clone SK3) and programmed death-1 (PD-1) PE (clone EH12.2H7) were obtained from BioLegend (San Diego, CA, USA). CD3 APC (clone UCHT1) and CD38 FITC (clone HIT2) were obtained from QuantoBio (Beijing, China). Cryopreserved PBMCs had more than 85% viability after thawing, then were washed and surface stained for 20 min at room temperature. Cells were then washed and intracellular stained after being permeabilized and fixed using the Permeabilization/Fixation Kit (eBioscience, Waltham, MA, USA).
To evaluate total and HIV-specific CD8 T-cells responses, thawed PBMCs were stimulated with anti-CD3 antibody (clone OKT3, 1 µg/mL, BioLegend) or overlapping peptides comprised by HIV env, gag, and pol (1 µg/mL, respectively, JPT, Berlin, Germany) for 6 h at 37 - C in a humidified incubator with 5% CO 2 . The costimulatory antibodies CD28 and CD49d (clone CD28.2 and 9F10, 1 µg/mL, respectively, BioLegend), along with the transport inhibitor Brefeldin A (3 µg/mL, eBioscience)
## Hiv antibody measurement
Specific antibody responses were measured with 20 µL of thawed plasma by enzyme immunoassay according to the manufacturer's instructions (Western blot [WB] kit 2.2, MP Diagnostics, Singapore), against 10 HIV viral proteins (gp160, gp120, p66, p55, p51, gp41, p39, p31, p24, and p17). Methods of analyzing the WB strips have been previously published [bib_ref] Human immunodeficiency virus (HIV)-antibody repertoire estimates reservoir size and time of antiretroviral..., Rocca [/bib_ref]. When the gray-scale value of each antigen was ≥50, 10-49, or 0-9% of those calculated in the strong positive control for the same antigen, the corresponding value was assigned a score 1, 0.5, or 0, respectively [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref] , left panel). Finally, a total WB score of each strip was assigned in the sum of the number of positive and weak responses (from 0 to 10).
## Statistical analyses
Statistical analyses were performed using SPSS 23.0 software (Chicago, IL, USA). The non-parametric Kruskal-Wallis test for continuous variables and the Fisher's exact test for categorical variables were used for multiple comparisons between groups. The Mann-Whitney U test was used for comparison of variables between two groups. The slopes of CD4 counts and HIV DNA over time were estimated by fitting participant-specific linear regressions [bib_ref] HIV-1 DNA decay dynamics in blood during more than a decade of..., Besson [/bib_ref]. Correlations were analyzed using the Spearman test. Two-sided P < 0.05 was considered statistically significant.
# Results
## Study participant characteristics
Baseline characteristics were similar among the three groups, except that participants in G3 had higher pre-ART CD8 counts compared with the other two groups (all P < 0.05, [fig_ref] TABLE 1 |: Patient characteristics in the three groups analyzed [/fig_ref]. Similar to that in [fig_ref] TABLE 1 |: Patient characteristics in the three groups analyzed [/fig_ref] , the characteristics of participants with serial samples in the three groups are shown in [fig_ref] TABLE 1 |: Patient characteristics in the three groups analyzed [/fig_ref]. Plasma viral loads of all participants had reduced less than the lower limit of detection within 6 months of ART and remained undetectable during the follow-up periods.
## Longitudinal changes of cd4 counts during art
To investigate changes of circulating CD4 and CD8 counts among the three groups of the participants during ART, we found that almost all participants had a substantial increase in CD4 counts after 5 years of ART (median CD4 counts at baseline, G1 vs. G2 vs. G3, 366 vs. 334 vs. 351 cells/µL; median CD4 counts at year 5 of ART, G1 vs. G2 vs. G3, 611 vs. 672 vs. 746 cells/µL, [fig_ref] FIGURE 2 |: Longitudinal changes in CD4 counts and CD8 counts during ART [/fig_ref]. However, the trajectories of CD4 counts were different among the three groups. The median increased CD4 counts were 209, 214, and 122 cells/µL in G1, G2, and G3 in the first year of ART, respectively; however, the median increased CD4 counts of each group were 58, 64, and 101 cells/µL between years 3 and 5 of ART, respectively [fig_ref] FIGURE 2 |: Longitudinal changes in CD4 counts and CD8 counts during ART [/fig_ref]. There was a lowest increase in CD4 counts in G3 within 1 year of ART, while G3 patients had a significant increase of CD4 counts during years 3-5 of ART and had higher CD4 counts after 5 years of ART than the other groups [fig_ref] FIGURE 2 |: Longitudinal changes in CD4 counts and CD8 counts during ART [/fig_ref].
Furthermore, [fig_ref] TABLE 2 |: Slopes of CD4 counts and HIV DNA over time [/fig_ref] shows the slopes of CD4 counts during years 0-1, 1-3, and 3-5 of ART, respectively. We found that within 1 year of ART, the slopes of CD4 counts in G3 were significantly lower compared to G1 and G2 (P = 0.046, P = 0.013); however, the slopes of CD4 counts in G3 were significantly higher compared to G1 and G2 after 3 years of ART (all P < 0.05). These results were consistent with abovementioned results about variation trends of CD4 counts over time among the three groups.
There was significantly higher baseline CD8 counts in G3 than the other two groups. Notably, mean CD8 counts were consistently above 1,000 cells/µL in G3 during the following 5 years of follow-up check [fig_ref] FIGURE 2 |: Longitudinal changes in CD4 counts and CD8 counts during ART [/fig_ref].
## Dynamics of viral reservoir during art
To examine the viral reservoir decay dynamics of the three groups, we assessed levels of cell-associated HIV DNA [bib_ref] Total HIV-1 DNA, a marker of viral reservoir dynamics with clinical implications, Avettand-Fenoel [/bib_ref] [bib_ref] HIV reservoirs: what, where and how to target them, Churchill [/bib_ref] , HIV specific antibodies [bib_ref] HIV antibody characterization as a method to quantify reservoir size during curative..., Burbelo [/bib_ref] [bib_ref] Qualitative and quantitative HIV antibodies and viral reservoir size characterization in vertically..., Brice [/bib_ref] , and PD-1 expression on CD4 T-cells [bib_ref] Cell-based measures of viral persistence are associated with immune activation and programmed..., Hatano [/bib_ref] [bib_ref] Targeting immune checkpoint molecules to eliminate latent HIV, Boyer [/bib_ref] [bib_ref] Programmed cell death-1 contributes to the establishment and maintenance of HIV-1 latency, Evans [/bib_ref] [bib_ref] PD-1 blockade potentiates HIV latency reversal ex vivo in CD4(+) T cells..., Fromentin [/bib_ref] in this study. Cell-associated HIV DNA level is commonly used to quantitatively assess the HIV reservoir [bib_ref] Total HIV-1 DNA, a marker of viral reservoir dynamics with clinical implications, Avettand-Fenoel [/bib_ref] [bib_ref] HIV reservoirs: what, where and how to target them, Churchill [/bib_ref]. The levels of cellassociated HIV DNA decayed rapidly in the three groups within the first year of ART initiation [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref]. At year 5 of ART, the level of HIV DNA tended to be lowest in G3 [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref] , and a significantly reduced HIV DNA level only occurred in G3 after 3 years of ART [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref]. In addition, the decay slope of HIV DNA was significantly higher in G3 compared to G1 during years 3-5 of ART (P = 0.01, [fig_ref] TABLE 2 |: Slopes of CD4 counts and HIV DNA over time [/fig_ref]. The percentages of patients with reduced HIV DNA between years 3 and 5 of ART were 26.7% (4/15), 73.7% , and 75% (9/12) in G1, G2, and G3, respectively [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref].
HIV specific antibodies could also be used to estimate viral reservoir size [bib_ref] HIV antibody characterization as a method to quantify reservoir size during curative..., Burbelo [/bib_ref] [bib_ref] Qualitative and quantitative HIV antibodies and viral reservoir size characterization in vertically..., Brice [/bib_ref]. Based on the WB score [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref] , left panel), we observed the dynamics of WB score in the three groups during ART [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref] , and WB score decreased most in G3 compared with the other groups from baseline to year 5 of ART (P < 0.05 and P < 0.01, respectively, [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref] , right panel).
As PD-1 + CD4 T-cells is a preferential reservoir and the percentage of PD-1 expression on CD4 T-cells is positively correlated with the HIV reservoir [bib_ref] Cell-based measures of viral persistence are associated with immune activation and programmed..., Hatano [/bib_ref] [bib_ref] Targeting immune checkpoint molecules to eliminate latent HIV, Boyer [/bib_ref] [bib_ref] Programmed cell death-1 contributes to the establishment and maintenance of HIV-1 latency, Evans [/bib_ref] [bib_ref] PD-1 blockade potentiates HIV latency reversal ex vivo in CD4(+) T cells..., Fromentin [/bib_ref] , we also investigated PD-1 expression on CD4 T-cells [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref]. We observed that the mean percentage of PD-1 + CD4 T-cells was lowest in group 3 at year 5 of ART [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref] , and the percentages of PD-1 + CD4
## Figure 3 | on cd4 t-cells. (e) comparisons of the percentages of pd-1 expressions on cd4 t-cells in three groups during art. (f) changes in percentage of pd-1 expression on
CD4 T-cells were compared in three different groups from baseline to year 5, from baseline to year 1, from years 1 to 3, and from years 3 to 5 (right panel). Group 1 (G1, n = 15 participants, 4 missing data at baseline), Group 2 (G2, n = 19 participants, 2 missing data at baseline), Group 3 (G3, n = 12 participants, 3 missing data at baseline), HC (health controls, n = 12 participants). In (C,E), data are mean ± SEM. In (right panel of A,F), boxes show median and IQR, and whiskers are minimum and maximum. *P < 0.05, **P < 0.01. CI, confidence interval; SEM, standard error of the mean; cps, copies.
T-cells significantly dropped in G3 particularly during years 3-5 of ART [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref]. These data suggest that G3 patients showed the most significant decrease in viral reservoir during 5 years of ART, especially in years 3-5 of ART. Therefore, patients with high CD8 counts may experience sustained viral reservoir decay in immune restored patients.
## Ki67 and pd-1 expression on cd8 t-cells during art
To determine the mechanism of CD8 counts elevation, the expressions of Ki67 and PD-1 on CD8 T-cells were measured to evaluate cellular proliferation and exhaustion status, respectively. After initiation of ART, both Ki67 and PD-1 expression on CD8 T-cells dropped significantly among three groups during the first year [fig_ref] FIGURE 4 |: Ki67 and PD-1 expression on CD8 T-cell during ART [/fig_ref]. The percentages of Ki67 + CD8 Tcells maintained the highest levels in G3 before and after ART [fig_ref] FIGURE 4 |: Ki67 and PD-1 expression on CD8 T-cell during ART [/fig_ref]. There were no differences in PD-1 expression on CD8 T-cells between groups at baseline; however, the mean percentage of PD-1 + CD8 T-cells continued to be lowest in G3 after 1 years of ART [fig_ref] FIGURE 4 |: Ki67 and PD-1 expression on CD8 T-cell during ART [/fig_ref].
We also evaluated CD8 T-cell subsets [fig_ref] FIGURE 2 |: Longitudinal changes in CD4 counts and CD8 counts during ART [/fig_ref] , HLA-DR, and CD38 co-expression on CD8 T-cells [fig_ref] FIGURE 3 |: Analysis of viral decay among the three groups during ART [/fig_ref] during ART. We found that HLA-DR and CD38 co-expression on CD8 T-cells was higher in G3 than the other groups during ART.
This data set indicates CD8 T-cells with a higher cellular proliferation activity and a lower immune exhaustion in G3 during long-term ART, and this may be the reason for a sustained high-level of CD8 T-cells in G3.
## Total and hiv-specific cd8 responses during art
To assess total CD8 T-cells functionalities, PBMCs were stimulated with anti-CD3 and analyzed for granzyme-B, perforin, CD107a, IFN-γ, and TNF-α expression on CD8 Tcells [fig_ref] FIGURE 4 |: Ki67 and PD-1 expression on CD8 T-cell during ART [/fig_ref]. No differences in any functional marker between groups at baseline were observed, with the following exception: the mean percentage of perforin + CD8 T-cells was higher in G3 than that in G1 and G2 [fig_ref] FIGURE 5 |: Total and HIV-specific CD8 T-cell functionalities during ART [/fig_ref]. At years 3 and 5 of ART, the proportions of each functional marker expression on CD8 T-cells were highest in G3, compared with the other groups [fig_ref] FIGURE 5 |: Total and HIV-specific CD8 T-cell functionalities during ART [/fig_ref].
At the same time, PBMCs were stimulated with HIV overlapping peptides to assess HIV-specific CD8 T-cells responses [fig_ref] FIGURE 5 |: Total and HIV-specific CD8 T-cell functionalities during ART [/fig_ref]. The proportions of CD8 T-cells producing CD107a, IFN-γ, and TNF-α decreased markedly in all groups during the first year of ART [fig_ref] FIGURE 5 |: Total and HIV-specific CD8 T-cell functionalities during ART [/fig_ref]. Similar to total CD8 T-cells functions reported above, the proportions of each functional marker expression on HIV-responsive CD8 T-cells were highest in G3 at years 3 and 5 of ART [fig_ref] FIGURE 5 |: Total and HIV-specific CD8 T-cell functionalities during ART [/fig_ref]. Taken together, CD8 T-cells in G3 have significantly higher functional capacity when compared with the other two groups.
## Correlation of reduced hiv dna and increased cd4 counts during years 3-5 of art with cd8 function and count
Finally, correlation analyses were conducted between reduced HIV DNA, increased CD4 counts, and CD8 counts including functional analysis during ART in all patients.shows that reduced HIV DNA levels during years 3-5 were positively correlated with HIV-specific CD8 T-cells responses, including TNF-α expression level at year 5 (r = 0.294, P = 0.048), CD107a expression levels at years 3 and 5 (r = 0.476, P = 0.001; r = 0.375, P = 0.010, respectively). Participants with higher CD8 counts at years 1, 3, and 5 of ART had more reduced HIV DNA levels during years 3-5 of ART (r = 0.350, P = 0.017; r = 0.294, P = 0.047; r = 0.393, P = 0.007).
As shown in, increased CD4 counts during years 3-5 were positively correlated with CD8 functions between years 3-5 of ART as follows: IFN-γ expression level at years 3 and 5 (r = 0.318, P = 0.031; r = 0.371, P = 0.011, respectively), TNF-α expression level at year 3 (r = 0.326, P = 0.027), CD107a expression levels at years 3 and 5 (r = 0.322, P = 0.029; r = 0.297, P = 0.045, respectively), granzyme B expression levels at year 5 (r = 0.351, P = 0.017), and perforin expression levels at year 5 (r = 0.314, P = 0.034). However, similar associations were only observed between increased CD4 counts during years 3-5 and CD8 counts at year 5 of ART (r = 0.301, P = 0.042).
# Discussion
The aim of this study is to investigate the effect of different levels of CD8 T-cells on CD4 recovery and viral reservoir decay under the condition of CD4 recovery above 500 cells/µL after 5 years of ART. We found that both CD4 counts increased and viral reservoir decay were highest in the persistently high CD8 counts group (G3) compared to the other groups during 5 years of ART. Rate of CD4 counts gain and rate of HIV DNA levels decay were rapid within the first year of ART in all the three groups. These results were consistent with previous reports that CD4 recovery and HIV DNA decline occurred mainly in the first 2 years of ART [bib_ref] HIV-1 DNA decay dynamics in blood during more than a decade of..., Besson [/bib_ref] [bib_ref] The extent of HIV-1-related immunodeficiency and age predict the long-term CD4 T..., Kaufmann [/bib_ref]. This may be the result of productively infected cell death and prevention of new cell infections [bib_ref] Integrated HIV DNA accumulates prior to treatment while episomal HIV DNA records..., Murray [/bib_ref].
Our results also showed that an increase in CD4 counts in the first year of ART was significantly lower in G3 than that Comparisons of mean (SEM) of the percentages of Ki67 (C) and PD-1 (D) expressions on CD8 T-cells between groups during ART. Group 1 (G1, n = 15 participants, 4 missing data at baseline), Group 2 (G2, n = 19 participants, 2 missing data at baseline), Group 3 (G3, n = 12 participants, 3 missing data at baseline), HC (health controls, n = 12 participants). *P < 0.05, **P < 0.01, ****P < 0.0001.
in the other two groups. This result was consistent with the report that elevated CD8 counts at ART initiation associated with poor CD4 recovery (1). However, within 5 years of ART especially during years 3-5 of ART, the patients in G3 had the most significant increase in CD4 counts. This result is different from the report that high CD8 levels were associated with poor CD4 recovery during long-term ART (1, 6), which may be due to the fact that this study enrolled patients without any CD4 level restrictions during 2 years of ART. We also found elevated CD8 counts associated with poor CD4 increase in patients with poor CD4 reconstitution within 3 years of ART (unpublished date). In addition, a previous study has shown that a higher CD4 counts was associated with some slight benefit for HIVinfected individuals with CD4 counts above 500 cells/µL [bib_ref] CD4 cell count and the risk of AIDS or death in HIV-Infected..., Young [/bib_ref]. We presumed that a higher CD4 count in G3 may be more beneficial for immune restored patients.
We found that HIV DNA was significantly reduced in G3 patients during years 3-5 of ART which was different from G1 and G2. More importantly, we also found that the HIV reservoir size decreased significantly in G3 and tended to be the smallest in G3 compared with the other groups after 5 years of ART. Therefore, this study provides evidence that persistent viral reservoir decay is likely to occur in immune restored patients with higher CD8 counts after long-term ART. Viard et al. reported that HIV DNA levels did not significantly decrease after 3 years of ART [bib_ref] Impact of 5 years of maximally successful highly active antiretroviral therapy on..., Viard [/bib_ref] ; however, other studies reported that some individuals had continuous viral reservoir decay after long-term ART [bib_ref] Determinants of HIV-1 reservoir size and long-term dynamics during suppressive ART, Bachmann [/bib_ref]. Chun et al. reported that the HIV-1 proviral DNA load in CD4 T-cells showed no correlation with CD8 counts in treated patients, but an inverse correlation with CD4:CD8 ratio [bib_ref] Relationship between the size of the human immunodeficiency virus type 1 (HIV-1)..., Chun [/bib_ref]. In our study, patients with CD4 recovery above 500 cells/µL after 5 years of ART were enrolled, which was different from previous [bib_ref] CD4:CD8 ratio and cd8 count as prognostic markers for mortality in human..., Trickey [/bib_ref]. The above-mentioned opposite outcomes may be attributed to discrepancies in study subjects.
Owing to the HIV reservoir, it has been estimated that achieving a "sterilizing" cure would need more than 60 years of ART alone [bib_ref] Latent infection of CD4+ T cells provides a mechanism for lifelong persistence..., Finzi [/bib_ref]. The smaller the reservoir size, the longer time the virus would need in order to rebound [bib_ref] HIV-1 DNA predicts disease progression and post-treatment virological control, Williams [/bib_ref]. Whether immune restored patients with high CD8 counts are more likely to accelerate decay of the HIV reservoir and achieve a functional cure requires further investigation.
In order to explore the mechanism further, we compared the dynamic changes of CD8 T-cells function in the three groups. High PD-1 expression on CD8 T-cells were linked to impaired proliferation and dysfunction [bib_ref] Exhaustion of activated CD8 T cells predicts disease progression in primary HIV-1..., Hoffmann [/bib_ref] and associated with HIV persistence [bib_ref] PD-1 expression in HIV-Specific CD8+ T cells before antiretroviral therapy is associated..., Ghiglione [/bib_ref] , and Ki67 is a common marker for proliferation. Consistent with previous reports [bib_ref] HIV-specific CD8(+) T cells and HIV eradication, Jones [/bib_ref] , our results illustrated that HIV-specific CD8 cytotoxic T-cells (CTL), the expression of PD-1, and ki67 on CD8 T-cells all decreased rapidly during the first year of ART. However, after the first year of ART, G3 patients had remarkably higher frequencies of functional CD8 T-cells than the other two groups. The expression of PD-1 on CD8 T-cells maintained at the lowest levels in G3 after 1 year of ART. As blocking PD-1 could reverse the CTL response (36), the persistence low levels of PD-1 on CD8 T-cells coincided with the most potent CD8 T-cell function in G3 after 1 year of ART.
As increased CD4 counts and reduced viral levels during years 3-5 of ART were both higher in G3 than the other groups, we then analyzed the relationships between the abovementioned two indictors and CD8 counts and function. We found that increased CD4 counts and reduced viral levels were associated with CD8 counts and function at years 3 and 5 of ART. These results are consistent with the report that HIVspecific CD8 cytotoxic T-cells (CTL) [bib_ref] Stimulation of HIV-1-specific cytolytic T lymphocytes facilitates elimination of latent viral reservoir..., Shan [/bib_ref] and strong CD8 T-cells responses [bib_ref] HIV-specific CD8(+) T cells and HIV eradication, Jones [/bib_ref] [bib_ref] Post-treatment control of HIV infection, Conway [/bib_ref] were beneficial to HIV reservoir decay. Dysfunctional CD4 T-cells could contribute to the exhaustion of the CD8 T-cells [bib_ref] Quantifying residual HIV-1 replication in patients receiving combination antiretroviral therapy, Zhang [/bib_ref] , and in this study we enrolled patients with CD4 recovery above 500 cells/µL after long-term ART. Therefore, whether CD4 T-cell immune recovery helps restoration of CD8 T-cells function is worth further study. Inherent characteristics may play an important role between the three groups .
The mechanisms of sustained CD8 elevation remain largely unclear, which may include CMV co-infection [bib_ref] CD8 T-cell expansion and inflammation linked to CMV coinfection in ART-treated HIV..., Freeman [/bib_ref] [bib_ref] Asymptomatic CMV replication during early human immunodeficiency virus (HIV) infection is associated..., Smith [/bib_ref] , bystander activation [bib_ref] CD8+ T cells specific for EBV, cytomegalovirus, and influenza virus are activated..., Doisne [/bib_ref] , and a shifted compartment distribution from lymphoid tissues [bib_ref] Mechanisms of CD8(+) T cellmediated suppression of HIV/SIV replication, Mcbrien [/bib_ref]. In addition, there are some limitations to this study. Firstly, as male-to-male sex is the main route of HIV transmission in major cities of China (43) at present, most HIV-infected individuals in our center are men. Participants were overwhelmingly male in this study, thus the results do not represent the situation of HIV-infected women. Secondly, the latent reservoir is mainly located in lymphoid tissues rather than in the peripheral blood [bib_ref] Severe CD4+ T-cell depletion in gut lymphoid tissue during primary human immunodeficiency..., Guadalupe [/bib_ref] [bib_ref] Massive infection and loss of memory CD4+ T cells in multiple tissues..., Mattapallil [/bib_ref]. Unfortunately, our study only obtained peripheral blood samples, without corresponding lymphoid tissues. Besides, patients with longer-term (e.g., >5years) effective ART were not enrolled in our study. The decay slopes of the latent reservoir are variant among patients in different term of ART, which is greatest within 5 years after initiation of ART. This decay tends to be stable thereafter [bib_ref] Determinants of HIV-1 reservoir size and long-term dynamics during suppressive ART, Bachmann [/bib_ref]. Finally, the case number was small, and studies with a larger sample size and a longer follow-up period would be better to make sound conclusions.
In conclusion, our findings showed that CD4 recovery and viral reservoir decay were associated with persistently high CD8 counts in immune restored patients during long-term ART. The information provided herein will contribute to a better understanding of CD8 persistence on viral reservoir decay and immune recovery under ART.
# Data availability statement
The datasets generated for this study are available on request to the corresponding author.
# Ethics statement
The studies involving human participants were reviewed and approved by the institutional review boards of Fifth Medical Center of Chinese PLA General Hospital. The patients/participants provided their written informed consent in line with the Declaration of Helsinki to participate in this study.
# Author contributions
L-XZ, Y-MJ, CZ, J-WS, and F-SW conceived the study, designed the experiments, and analyzed the data. L-XZ and Y-MJ performed the experiments. XF, R-NX, H-HH, J-YZ, L-FW, C-BZ, LJ, and MS contributed to reagents and materials. L-XZ, Y-MJ, and F-SW wrote the article. All authors read and approved the final manuscript.
[fig] FIGURE 1 |: Flow chart of participants enrolled in this study. [/fig]
[fig] FIGURE 2 |: Longitudinal changes in CD4 counts and CD8 counts during ART. (A) Time trajectories of CD8 counts and CD4 counts over 5 years in participants receiving ART. Individual trajectories are shown with red lines, mean values are indicated by thick black lines. (B) Statistical analysis of CD4 counts in the three groups at baseline, years 1, 3, and 5 on ART. (C) Comparisons of increased CD4 counts from three different groups are assessed from baseline to year 5, from baseline to year 1, from years 1 to 3, and from years 3 to 5. Group 1 (G1, n = 18 participants), Group 2 (G2, n = 45 participants), Group 3 (G3, n = 18 participants). In (B,C), boxes show median and IQR, and whiskers are minimum and maximum. *P < 0.05, **P < 0.01. [/fig]
[fig] FIGURE 4 |: Ki67 and PD-1 expression on CD8 T-cell during ART. Representative flow cytometric data of Ki67 (A) and PD-1 (B) staining gated on CD8 T-cells. [/fig]
[fig] FIGURE 5 |: Total and HIV-specific CD8 T-cell functionalities during ART. Proportions of granzyme B, perforin, CD107a, IFN-γ, and TNF-α expression on CD8 T-cells (A) after stimulation with anti-CD3, and percentages of granzyme B, perforin, CD107a, IFN-γ and TNF-α expression on CD8 T-cells (B) after stimulation with HIV overlapping peptides show as mean ± SEM. Statistical analysis of functional markers mentioned above in three groups during ART, respectively. Group 1 (G1, n = 15 participants, 4 missing data at baseline), Group 2 (G2, n = 19 participants, 2 missing data at baseline), Group 3 (G3, n = 12 participants, 3 missing data at baseline), HC (health controls, n = 12 participants). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. [/fig]
[fig] FUNDING: This work was supported by National Science and Technology Major Project of the Ministry of Science and Technology of China (grant numbers 2017ZX10202102-004-002 and 2018ZX10302104-002). [/fig]
[table] TABLE 1 |: Patient characteristics in the three groups analyzed. Data are n (%) or median (IQR). ART, antiretroviral therapy; HIV, human immunodeficiency virus; IQR, interquartile range; NRTIs, nucleoside reverse transcriptase inhibitors; NNRTIs, non-nucleoside reverse transcriptase inhibitors; PIs, protease inhibitor. ∆ P < 0.001 (G1 vs. G3), P = 0.002 (G2 vs. G3). in Immunology | www.frontiersin.org [/table]
[table] TABLE 2 |: Slopes of CD4 counts and HIV DNA over time. [/table]
[table] TABLE 3 |: (A) Reduced HIV DNA levels during years 3-5 of ART correlated to levels of CD8 function and counts over the course of ART.studies. Long-term non-progressors with low HIV reservoir have high levels of CD8 counts(13), which suggests that patients in G3 maybe have a similar immune status. Our results suggest high CD8 counts are conducive to viral decay in patients with immune restoration during long-term ART. It may explain why individuals with high CD8 counts have a reduced risk of AIDSrelated events during ART (1). Nevertheless, Trickey et al. found that high CD8 counts were associated with AIDS mortality in this population with high CD4 counts and suppressed viral loads [/table]
|
Estimating Cardiorespiratory Fitness Without Exercise Testing or Physical Activity Status in Healthy Adults: Regression Model Development and Validation
Background: Low cardiorespiratory fitness (CRF) is an independent predictor of morbidity and mortality. Most health care settings use some type of electronic health record (EHR) system. However, many EHRs do not have CRF or physical activity data collected, thereby limiting the types of investigations and analyses that can be done.Objective: This study aims to develop a nonexercise equation to estimate and classify CRF (in metabolic equivalent tasks) using variables commonly available in EHRs.Methods: Participants were 42,676 healthy adults (female participants: n=9146, 21.4%) from the Aerobics Center Longitudinal Study examined from 1974 to 2005. The nonexercise estimated CRF was based on sex, age, measured BMI, measured resting heart rate, measured resting blood pressure, and smoking status. A maximal treadmill test measured CRF.Results:After conducting nonlinear feature augmentation, separate linear regression models were used for male and female participants to calculate correlation and regression coefficients. Cross-classification of actual and estimated CRF was performed using low CRF categories (lowest quintile, lowest quartile, and lowest tertile). The multiple correlation coefficient (R) was 0.70 (mean deviation 1.33) for male participants and 0.65 (mean deviation 1.23) for female participants. The models explained 48.4% (SE estimate 1.70) and 41.9% (SE estimate 1.56) of the variance in CRF for male and female participants, respectively. Correct category classification for low CRF (lowest tertile) was found in 77.2% (n=25,885) of male participants and 74.9% (n=6,850) of female participants.Conclusions:The regression models developed in this study provided useful estimation and classification of CRF in a large population of male and female participants. The models may provide a practical method for estimating CRF derived from EHRs for population health research.
# Introduction
# Background
The use of data from electronic health records (EHRs) beyond day-to-day medical management is rapidly emerging in the fields of digital health, public health, and epidemiology [bib_ref] Machine learning in epidemiology and health outcomes research, Wiemken [/bib_ref] [bib_ref] Big data, health informatics, and the future of cardiovascular medicine, Kim [/bib_ref]. However, access to cardiorespiratory fitness (CRF), a valuable health metric, is limited. This limitation is primarily due to the medical service (cardiopulmonary stress test) being costly, time-consuming, and generally focused on cardiac patients [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref] [bib_ref] Using machine learning on cardiorespiratory fitness data for predicting hypertension: The Henry..., Sakr [/bib_ref] [bib_ref] Rationale and design of the Henry Ford Exercise Testing Project (the FIT..., Al-Mallah [/bib_ref]. CRF is a comprehensive measure of one's functional capacity (mL O 2 · kg 1 · min 1 ) driven by the combination of heart, lung, and muscle function. It is an important marker of health status in the general adult population [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref]. Further demonstrating the importance of CRF, the American Heart Association released a scientific statement proposing that CRF be considered a clinical vital sign. The scientific rationale behind the statement is driven by the voluminous evidence that demonstrates that low CRF is a strong independent predictor of adverse health outcomes (ie, all-cause mortality, cancer, stroke, heart disease, and diabetes incidence) [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref].
## Prior work
To help increase accessibility to CRF data, researchers have developed an array of nonexercise estimated CRF (NEECRF) equations to estimate CRF [bib_ref] The role of body habitus in predicting cardiorespiratory fitness: The FRIEND Registry, Baynard [/bib_ref] [bib_ref] An overview of non-exercise estimated cardiorespiratory fitness: estimation equations, cross-validation and application, Wang [/bib_ref] [bib_ref] Accuracy of exercise-based equations for estimating cardiorespiratory fitness, Peterman [/bib_ref]. NEECRF equations commonly include age, gender, resting heart rate, smoking status, BMI, and self-reported physical activity status (PAS) [bib_ref] An overview of non-exercise estimated cardiorespiratory fitness: estimation equations, cross-validation and application, Wang [/bib_ref] [bib_ref] Longitudinal cardiorespiratory fitness algorithms for clinical settings, Jackson [/bib_ref] [bib_ref] Changes in aerobic power of women, ages 20-64 yr, Jackson [/bib_ref] [bib_ref] Assessing cardiorespiratory fitness without performing exercise testing, Jurca [/bib_ref]. Studies have shown NEECRF to predict all-cause and cardiovascular disease mortality on par with measured CRF [bib_ref] Longitudinal algorithms to estimate cardiorespiratory fitness: associations with nonfatal cardiovascular disease and..., Artero [/bib_ref] [bib_ref] A non-exercise testing method for estimating cardiorespiratory fitness: associations with all-cause and..., Stamatakis [/bib_ref]. However, the assessment of PAS required to calculate NEECRF in patients is not typically conducted or documented in health care settings [bib_ref] The extent to which family physicians record their patients' exercise in medical..., Lindeman [/bib_ref]. Therefore, using an NEECRF model without PAS (non-PAS) may be more feasible.
In a 2019 comprehensive NEECRF review, a few peer-reviewed non-PAS NEECRF equations were identified in adult populations reporting correlations and SE estimates [bib_ref] An overview of non-exercise estimated cardiorespiratory fitness: estimation equations, cross-validation and application, Wang [/bib_ref]. Most of the equations were developed using only age, height, and weight combinations, and some used variables not commonly found in EHRs (eg, waist girth, predicted/ideal weight, or exercise mode).
Though correlations were moderate to high, the sample populations were too small to determine the classification accuracy of low CRF. Accurate classification of low CRF is essential for large-scale investigations [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref] [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref]. Low CRF (lowest tertile) classification was recently investigated by Peterman et al [bib_ref] Accuracy of exercise-based equations for estimating cardiorespiratory fitness, Peterman [/bib_ref] , who found Baynard's [bib_ref] The role of body habitus in predicting cardiorespiratory fitness: The FRIEND Registry, Baynard [/bib_ref] simplified non-PAS NEECRF equation to have poor classification ability in a sizable (n=4871) adult population. Other researchers have also suggested that simple non-PAS NEECRF formulas have low validity, and nuanced approaches are warranted [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref] [bib_ref] Validity of VO2max equations for aerobically trained males and females, Malek [/bib_ref]. Investigators have also noted that valid non-PAS NEECRF models may have broad applications for public health, epidemiology, surveillance, practice, and research [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref] [bib_ref] An overview of non-exercise estimated cardiorespiratory fitness: estimation equations, cross-validation and application, Wang [/bib_ref] [bib_ref] Accuracy of exercise-based equations for estimating cardiorespiratory fitness, Peterman [/bib_ref] [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref] [bib_ref] Cardiorespiratory fitness, coronary artery calcium, and cardiovascular disease events in a cohort..., Radford [/bib_ref].
## Goal of this study
Because it is standard practice to assess and document resting heart rate, blood pressure, BMI, and smoking status during a typical clinic visit, the primary aim of this study was to develop new models for NEECRF that could potentially be used in large-scale population health investigations using variables commonly found in EHRs. To accomplish this, we compared a non-PAS NEECRF equation to clinically measured CRF and evaluated its ability to estimate and classify CRF.
# Methods
## Study sample
The Aerobics Center Longitudinal Study (ACLS) is a prospective epidemiological investigation of participants that began in 1970 [bib_ref] Physical fitness and all-cause mortality. A prospective study of healthy men and..., Blair [/bib_ref]. The original data set for this investigation included 43,257 healthy adults who voluntarily participated in a comprehensive preventive medical examination at the Cooper Clinic in Dallas, Texas between 1974 and 2005. At baseline, all participants were free of diabetes, heart disease, stroke, cancer, positive electrocardiograms, and completed a maximal graded exercise test. Each participant gave informed consent to join the longitudinal study. The research population demographic primarily consists of Caucasian college-educated adults of middle to high socioeconomic status with an average age at baseline of 43.5 (range 20-79) years.
# Ethical approval
The study was reviewed and approved annually by the Cooper Institute Institutional Review Board, and all participants provided written informed consent.
## Measurements
Predictor variables were assessed during a preventive health examination that included objective measurements of age, BMI, resting heart rate, systolic blood pressure, diastolic blood pressure, and self-reported smoking status between 1974 and 2005 at the Cooper Clinic. Age was verified at the time of the examination. Height and weight were measured on a calibrated scale using US customary units and converted to metric scales for this investigation. BMI was calculated from measured height and weight as kg/m 2 . Manual auscultation was used to measure resting blood pressure while seated. Resting heart rate was calculated using the R-R interval on an electrocardiogram while seated. CRF was expressed as absolute metabolic equivalent tasks (METs; 1 MET = 3.5 mL O 2 · kg 1 · min 1 ) based on the total duration of a symptom-limited maximal Balke graded exercise test. Following American College for Sports Medicine Guidelines, patients were encouraged to give maximal effort, and the test end point was volitional exhaustion or termination by the physician for medical reasons. METs were calculated based on the final treadmill speed and grade. The Balke graded exercise test is highly correlated (r=0.94) with maximal graded cardiopulmonary exercise testing [bib_ref] An experimental study of physical fitness of Air Force personnel, Balke [/bib_ref] [bib_ref] Comparative analysis of physiologic responses to three different maximal graded exercise test..., Pollock [/bib_ref]. A standardized medical questionnaire was used to ascertain demographic information, lifestyle habits, and chronic disease status. More detailed information on the preventive health examination is available in prior ACLS publications [bib_ref] Longitudinal cardiorespiratory fitness algorithms for clinical settings, Jackson [/bib_ref] [bib_ref] Longitudinal algorithms to estimate cardiorespiratory fitness: associations with nonfatal cardiovascular disease and..., Artero [/bib_ref].
# Statistical analysis
## Data exclusion
Data were first examined for outliers and skewness. We removed the outliers from the data set by removing participants with predictor variables (continuous) beyond the ±3 σ interval. To do this, we calculated the mean and SD of each predictor variable, excluding smoking (categorical). Any participant with at least one predictor variable above 3 SDs or below 3 SDs was flagged as an outlier. After removing outliers and incomplete entries (n=581), the data set comprised 33,530 male participants and 9146 female participants, 98.7% of the original 43,257 participants.
## Regression
The main analysis was based on apparently healthy adults at baseline. Using a supervised machine learning technique, we conducted separate linear regression analyses for men and women to predict non-PAS NEECRF based on nonlinear augmentation of the predictor variables [bib_ref] Machine learning in epidemiology and health outcomes research, Wiemken [/bib_ref]. We also considered advanced machine learning models but did not find them advantageous. The male and female non-PAS NEECRF prediction equations were formulated to minimize the average mean squared error, where N is the number of samples in our data set:
The prediction equation used age, height (Ht), weight (Wt), BMI, resting heart rate (rHR), systolic blood pressure (SBP), diastolic blood pressure (DBP), and smoking. All variables were continuous except for smoking status (nonsmoker=0, current smoker=1). Data were standardized by subtracting the mean and dividing by the SD for each variable. Next, separate models were then trained for male and female participants. We augmented the original 8 variables with second order and interaction terms, and regressed them linearly to the dependent variable for training. In this way, the nonlinearity was transferred from the regressor to the independent variables, while the model's overall interpretability was maintained. The augmentation procedure added the following 28 second order and interaction terms: Wt 2 , Wt × Ht, Wt × Age, Wt × rHR, Wt × SBP, Wt × DBP, Wt × BMI, Ht 2 , Ht × Age, Ht × rHR, Ht × SBP, Ht × DBP, Ht × BMI, Age 2 , Age × rHR, Age × SBP, Age × DBP, Age × BMI, rHR 2 , Ht × SBP, Ht × DBP, rHR × BMI, SBP 2 , SBP × DBP, SBP × BMI, DBP 2 , and DBP × BMI (for a total of 36 variables). Because smoking status was a categorical variable, it was not used to create the additional variables. The augmented data set was input into an elastic net linear regressor and trained and evaluated via 10-fold cross-validation [bib_ref] Regularization and variable selection via the elastic net, Zou [/bib_ref]. Optimal model hyperparameters were calculated for the male (α=.001, λ=1.0) and female (α=.004, λ=1.0) data sets through the cross-validation procedure. Pearson correlation coefficients were then calculated using the non-PAS NEECRF equations for the male and female data sets (shown in Multimedia Appendix 1). Lastly, for comparison, we cross-validated Baynard's [bib_ref] The role of body habitus in predicting cardiorespiratory fitness: The FRIEND Registry, Baynard [/bib_ref] simplified non-PAS NEECRF equation (77.96 − 10.35 (sex; M=0, F=1) − 0.92 (BMI) − 0.32 (age)) [bib_ref] Accuracy of exercise-based equations for estimating cardiorespiratory fitness, Peterman [/bib_ref].
## Classification accuracy
We cross-classified non-PAS NEECRF and CRF for three specified cut points (lowest quintile, lowest quartile, and lowest tertile) commonly used in epidemiological investigations to define low CRF [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref]. Next, CRF distributions were then grouped by males and females where α served as the value of the α-th percentile using the calculation non-PAS NEECRF>α. After classification, we determined the non-PAS NEECRF accuracy, sensitivity, positive predictive value, and F 1 score. The reference standard was the measured CRF, and low CRF was defined as a positive test. All analyses were performed in scikit-learn version 0.22.2 (NumFOCUS).
# Results
Descriptive statistics are provided in [fig_ref] Table 1: Baseline characteristics of participants [/fig_ref]. Correlation coefficients between each independent variable and CRF are presented in Multimedia Appendix 1. The multiple Rs and mean deviations for non-PAS NEECRF (in METs) were high at 0.70 (mean deviation 1.33) for male participants and moderate at 0.65 (mean deviation 1.23) for female participants. The models explained 48.4% (SE estimate 1.70, 95% CI 0.05-3.97) of the variance in CRF for male participants and 41.9% (SE estimate 1.56, 95% CI 0.05-3.48) for female participants. Multimedia Appendix 1 provides a simple independent variable input Google Sheet for researchers and data scientists to easily calculate NEECRF. [fig_ref] Table 2: Predictive accuracy of nonexercise estimated CRF classification of lowest cardiovascular fitness against... [/fig_ref] provides the findings regarding the accuracy, positive predictive, and sensitivity values using the lowest quintile, quartile, and tertile to classify low CRF for male and female participants. While overall classification accuracy was meaningful for a nondiagnostic test across the three models, the optimal model was the lowest tertile. Combined male and female positive predictive value were 0.60, sensitivity 0.67, and F 1 score 0.63 [bib_ref] Population and age-based cardiorespiratory fitness level investigation and automatic prediction, Xiang [/bib_ref] [bib_ref] Sensitivity, specificity, and predictive values: foundations, pliabilities, and pitfalls in research and..., Trevethan [/bib_ref]. The F 1 score is the best practice summary metric consisting of the harmonic mean of positive predictive value and sensitivity for classification (0=low, 1=high) [bib_ref] Population and age-based cardiorespiratory fitness level investigation and automatic prediction, Xiang [/bib_ref]. Assuming a balanced data set (n=2529) by Peterman et al [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref] , we calculated the F 1 scores from their reported findings for equations applicable to EHRs [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref]. We found F 1 scores ranging from 0.04 to 0.56. Based on a residual plot (Multimedia Appendix 1), we found the model was most accurate for CRF MET values in the 7.5 to 12.5 METs range but tended to underestimate MET values >12.5 METs and overestimate MET values <7.5 METs. Notably, this is a common finding in non-PAS NEECRF studies [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref]. Similar to Peterman et al [bib_ref] Accuracy of exercise-based equations for estimating cardiorespiratory fitness, Peterman [/bib_ref] , we cross-validated Baynard's [bib_ref] The role of body habitus in predicting cardiorespiratory fitness: The FRIEND Registry, Baynard [/bib_ref]
# Discussion
## Principal findings
Using a greater combination of clinical measures commonly found in EHRs, this study compared non-PAS NEECRF with objectively measured CRF in the largest population to date [bib_ref] Longitudinal cardiorespiratory fitness algorithms for clinical settings, Jackson [/bib_ref] [bib_ref] Changes in aerobic power of women, ages 20-64 yr, Jackson [/bib_ref] [bib_ref] Correlates and determinants of cardiorespiratory fitness in adults: a systematic review, Zeiher [/bib_ref] [bib_ref] Comparison of predicted exercise capacity equations and the effect of actual versus..., Ahmadian [/bib_ref] [bib_ref] Degrees of freedom at the start of the second 100 years: a..., Rodgers [/bib_ref]. Overall, our model may provide a more applicable method for estimating and classifying CRF than previous methods [bib_ref] An overview of non-exercise estimated cardiorespiratory fitness: estimation equations, cross-validation and application, Wang [/bib_ref] [bib_ref] Accuracy of exercise-based equations for estimating cardiorespiratory fitness, Peterman [/bib_ref] [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref]. Moreover, because the vital signs and medical information used to calculate non-PAS NEECRF are routinely captured during health care visits, our approach places nominal demand on health care staff and patients for collecting data. From a public health perspective, a moderate positive predictive value is practical, given that non-PAS NEECRF is a nondiagnostic test that is no cost and easily accessible [bib_ref] Sensitivity, specificity, and predictive values: foundations, pliabilities, and pitfalls in research and..., Trevethan [/bib_ref]. Likely, some individuals classified with low fitness may be at the lower end of the fit spectrum and benefit from health promotion [bib_ref] Sensitivity, specificity, and predictive values: foundations, pliabilities, and pitfalls in research and..., Trevethan [/bib_ref]. From a clinical perspective and considering moderate sensitivity, we concur with previous investigators that, while estimation equations are applicable for epidemiological investigations, they should not replace clinical exercise testing for patient diagnosis and management [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref] [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref]. Our findings show that the ACLS non-PAS NEECRF may provide a useful assessment of CRF to conduct population health research.
## Comparison to prior work
Comparatively, PAS-based NEECRF models have demonstrated higher positive correlation values (0.71-0.93) along with a higher degree (~90%) of correct classification accuracy for low CRF than non-PAS NEECRF models [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref]. Because physical activity is a key contributor to CRF, including a PAS variable in an NEECRF model improves accuracy [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref] [bib_ref] An overview of non-exercise estimated cardiorespiratory fitness: estimation equations, cross-validation and application, Wang [/bib_ref] [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref]. However, a recent review of distinct EHRs across 20 countries found that only 18.8% of family practice clinics had structured PAS questionaries embedded within the EHR, with documented PAS in the EHR ranging from 10% to 86% [bib_ref] The extent to which family physicians record their patients' exercise in medical..., Lindeman [/bib_ref]. Notably, no validated questionaries designed for PAS-based NEECRF calculations were used. Therefore, the ability to conduct large-scale studies aggregating existing EHR data across local, domestic, or international systems to predict CRF is unlikely. Conversely, our model may provide a global approach to aggregating EHR data across systems to predict CRF and conduct analyses.
In 2019, Wang et al [bib_ref] An overview of non-exercise estimated cardiorespiratory fitness: estimation equations, cross-validation and application, Wang [/bib_ref] provided a comprehensive list of peer-reviewed non-PAS NEECRF models that used some combination of age, BMI, or gender to predict CRF [bib_ref] An overview of non-exercise estimated cardiorespiratory fitness: estimation equations, cross-validation and application, Wang [/bib_ref]. Samples were from small populations, and moderate to high correlations were reported generally without SE estimates [bib_ref] An overview of non-exercise estimated cardiorespiratory fitness: estimation equations, cross-validation and application, Wang [/bib_ref]. Notably, the validity and usefulness of these simplified types of equations have been called into question [bib_ref] An overview of non-exercise estimated cardiorespiratory fitness: estimation equations, cross-validation and application, Wang [/bib_ref] [bib_ref] Accuracy of exercise-based equations for estimating cardiorespiratory fitness, Peterman [/bib_ref] [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref] [bib_ref] Validity of VO2max equations for aerobically trained males and females, Malek [/bib_ref]. We found that the findings from such investigations may be limited because the studies only reported correlations and lacked sufficient sample sizes to calculate prediction values for low CRF. It is also important to note that high correlation values do not necessarily result in a more accurate classification of low CRF [bib_ref] Accuracy of exercise-based equations for estimating cardiorespiratory fitness, Peterman [/bib_ref] [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref].
Recently, Peterman et al [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref] determined the ability of 7 non-PAS NEECRF equations to accurately classify low NEECRF (tertile) compared to measured CRF in a demographically comparable cohort (n=2529) to ACLS. Only 3 of the 7 equations apply because of variables not commonly found in EHRs. On balance, the classification accuracy of low CRF (tertile) appeared to be better for ACLS. In a separate investigation, Peterman et al [bib_ref] Accuracy of exercise-based equations for estimating cardiorespiratory fitness, Peterman [/bib_ref] also tested Baynard's [bib_ref] The role of body habitus in predicting cardiorespiratory fitness: The FRIEND Registry, Baynard [/bib_ref] simplified non-PAS NEECRF using the Ball State cohort (n=4871) to assess classification accuracy. The equation had an r of 0.76; however, there was poor accuracy for detecting individuals positive for low CRF (37%). We also cross-validated Baynard's [bib_ref] The role of body habitus in predicting cardiorespiratory fitness: The FRIEND Registry, Baynard [/bib_ref] equation with our data set and found low correlations, thus did not attempt to determine accuracy [bib_ref] Accuracy of exercise-based equations for estimating cardiorespiratory fitness, Peterman [/bib_ref]. Although the ACLS and Ball State cohorts are demographically similar, our finding is expected to some degree because the ACLS equation is specifically trained from the ACLS data set [bib_ref] Accuracy of exercise-based equations for estimating cardiorespiratory fitness, Peterman [/bib_ref] [bib_ref] Comparison of non-exercise cardiorespiratory fitness prediction equations in apparently healthy adults, Peterman [/bib_ref]. Moreover, these equations need to be tested in epidemiological investigations with EHR data to see how well they predict health outcomes.
# Limitations
This study is not without limitations. Our study's primary limitation regarding correlation was that the measured reference CRF was conducted using a Balke graded maximal exercise test that estimates absolute METs. This testing strongly correlates with adults' maximal graded cardiopulmonary exercise testing and is routinely used for clinical and epidemiological purposes [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref] [bib_ref] An experimental study of physical fitness of Air Force personnel, Balke [/bib_ref] [bib_ref] Comparative analysis of physiologic responses to three different maximal graded exercise test..., Pollock [/bib_ref]. Though the ACLS data set provides the largest known healthy female data set for clinically measured CRF, a larger female sample size may have provided a slightly better predictive model [bib_ref] Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case..., Ross [/bib_ref] [bib_ref] Degrees of freedom at the start of the second 100 years: a..., Rodgers [/bib_ref]. Nonetheless, our analyses demonstrated reasonable correlation and classification. Our analyses are based on a large predominantly Caucasian cohort; it is unknown if the results generalize to other ethnic groups. Notably, the homogeneity of the ACLS cohort may have strengthened the internal validity of our results by limiting possible confounders. The main strength of this investigation is that it was conducted on the largest cohort to date with a larger number of objectively measured predictive variables to estimate non-PAS NEECRF.
# Conclusions
The ACLS non-PAS NEECRF equation may provide a useful population health metric for CRF. More work should be conducted regarding diverse populations, the incidence of chronic conditions, and longitudinal repeated measures analyses toward improving public health and surveillance capability.
[table] Table 1: Baseline characteristics of participants. [/table]
[table] Table 2: Predictive accuracy of nonexercise estimated CRF classification of lowest cardiovascular fitness against the reference CRF. [/table]
|
Nature-based activities improve the well-being of older adults
Current literature shows that interaction with urban greenery can have a wide range of positive health outcomes. Targeted nature-based programs, such as therapeutic horticulture, have been shown to result in multiple health benefits for older adults residing in temperate environments, but much less research has been carried out on populations with different phenotypes, such as older Asian adults in the tropics. The current study investigated the effects of a 24-session therapeutic horticulture program on 47 older participants in Singapore, with an experimental pretest posttest design. We found that participants maintained healthy sleep patterns and psychological health, as well as showed reduced anxiety and improved cognitive functioning (p < 0.05). In addition, they reported an increase in mean happiness score after each session. This study provides new evidence using a comprehensive set of indicators across the affective, cognitive, functional, psychosocial and physical domains, supporting current literature on the benefits of nature programs, with a novel focus on tropical environments. It provides evidence that the nature-based intervention has the potential to be translated to programs to benefit older adults in the tropics.Highly urbanised environments offer productivity and convenience, but they also contain more stressors that affect health negatively, such as higher heat load, carbon emmissions and pollution, compared to the rural environments 1 . The positive association between contact with urban nature and health has been well established in literature. Back in the 1860s, city planners were already pronouncing that nature both tranquilizes and enlivens the mind; reinvigorating one's whole system 2 . The number of scientific manuscripts published on the subject has also increased exponentially in the recent years, reporting wide-ranging benefits, from lower mortality 3 and morbidity 4 to improvements in physical and mental health 5 . Findings from the meta-analyses show statistically significant associations with a wide range of health benefits, including reduced diastolic blood pressure, heart rate, salivary cortisol, incidence of type II diabetes and stroke, cardiovascular and all-cause mortality, as well as health-denoting associations with pregnancy outcomes, HRV, and HDL cholesterol, and self-reported improvements in health 6 . However, most of the research was conducted in temperate regions, offering very limited understanding on the health effects of nature exposure in regions with different climatic and environmental factors, such as the tropics 7 .Although the benefits of nature exposure on health are well established, the specific mechanisms and pathways are still not well understood. One of the leading theories is the Biophilia Hypothesis, which proposes that love for nature and ability to thrive in a natural environment are inherent to humans, dictated by evolutionary drives and instincts 8 . In contrast, biophobia, the affiliation with technology and human artifacts, is deemed to be culturally acquired rather than inborn, and as such unnatural 9 . Consequently, the healing properties of nature exposure are thought to be due to returning to our basic instincts and natural predispositions. Another commonly cited theory is the Attention Restoration Theory (ART), which proposes that contact with nature provides humans a softly fascinating environment that allows the mind to restore attentional fatigue. Nature enables people to be engaged in an "effortless" state of mind, as compared to other physical environments, which demand directed attention and hence exhaust our cognitive resources 10 .The benefits of contact with nature are particularly well documented for older adults 11 . One nature-based activity that is known to be popular with older adults is gardening. It has been shown to promote overall health and quality of life, physical strength, fitness and flexibility, cognitive ability, and socialization 12 . A treatment related to gardening is therapeutic horticulture (TH). This is a facilitated process through which participants OPEN
enhance their well-being by being involved in plant and plant-related activities. Due to its documented benefits, TH is gaining importance and TH programs are found in a wide variety of healthcare, rehabilitative, and residential settings in North America and a few other temperate countries such as the United Kingdom and Japan. A systematic review on the effects of TH showed that there was pre-post improvement in well-being, anxiety and depression, social relations, cognitive and functional outcomes [bib_ref] The effectiveness of horticultural therapy on older adults: a systematic review, Nicholas [/bib_ref]. Of the 20 studies included in the review, [bib_ref] Age and education correction of mini-mental state examination for English and Spanish-speaking..., Mungas [/bib_ref] were conducted on participants from long-term residential care facilities, and as such the results cannot be generalized to communities or day-care settings [bib_ref] The effectiveness of horticultural therapy on older adults: a systematic review, Nicholas [/bib_ref]. Only one study was conducted in the tropical environment, where the study population was healthy older adults with good cognitive functioning and good mobility [bib_ref] Effects of horticultural therapy on asian older adults: a randomized controlled trial, Ng [/bib_ref]. Therefore, whether TH is effective for the general older adult population residing in the tropics, and its impact on a range of physical and cognitive status measures requires further investigation. Moreover, long-term outcomes of TH on physical and mental health are largely unexplored.
The current study investigated the long-term effects of a 24 session (6-month) TH program on older adults (n = 47) from three senior day-care centers. The study aimed to provide a better understanding of TH programming in the tropical environment and its health outcomes among older adults in Singapore. The TH program aims to promote enhanced health and wellbeing. We hypothesized that participants would, over time, show improvements in perceived physical health, sleep hygiene, depression, anxiety, cognitive performance (orientation, attention, memory, language, spatial visualization and motor skill), and social connectedness. We also hypothesized that the program would improve momentary positive affect immediately after each session.
# Methods
Participants. Elderly participants (n = 47, 33 females) with a mean age of 77.5 years (SD = 7.8, range 60-95) were recruited progressively from three senior day care centers (SAC) in Singapore, between May 2017 to December 2018. 12 participants were first recruited from SAC 1 and assigned to group 1. They received the intervention from May to October 2017. This was followed by group 2 (intervention from November 2017 to May 2018) which comprised another 7 participants from SAC 1, and group 3 (intervention from June to December 2018) which comprised 11 participants from SAC 2, and group 4 (intervention from January to July 2019) which comprised 9 participants from SAC 3, and lastly group 5 (intervention from January to July 2019 and carried out on days different from group 4) which comprised 8 participants from SAC 3. All participants from SAC 1 (n = 19, 14 females), SAC 2 (n = 11, 8 females) and SAC 3 (n = 17, 11 females) gave consent to participate in the Mini Mental State Exam (MMSE) [bib_ref] Mini-mental state": a practical method for grading the cognitive state of patients..., Folstein [/bib_ref] to test for signs of dementia. An age and education adjusted formula MMSEadj = Raw MMSE − (0.471 × [education − 12]) + (0.131 × [age − 70]) was used to adjust the MMSE scores for those older than 75 years old [bib_ref] Age and education correction of mini-mental state examination for English and Spanish-speaking..., Mungas [/bib_ref]. We used the cut-off score of 17 and lower as an exclusion criterion, as an indicator of severe cognitive impairment [bib_ref] Self-, collateral-and clinician assessment of depression in persons with cognitive impairment, Chopra [/bib_ref]. Those excluded were deemed to benefit from a more individualized horticultural therapy treatment program. The distribution of age and years of education was balanced in the three SACs. The study was approved by the National Healthcare Group Domain Specific Review Board (DSRB), with the reference L2016/00949. The research was performed in accordance with relevant guidelines and regulations. All participants provided informed consent.
Intervention. The intervention was a 24-session weekly TH program comprising an equal mix of horticultural-based and nature art activities. The program was developed with the input of a horticultural therapy expert registered with the American Horticultural Therapy Association (AHTA). In addition, we incorporated knowledge of local plants and related activities. The one-hour sessions followed a planned order, beginning with a series of introductory horticultural-based activities-growing pea sprouts, setting up planters, growing vegetables from different modes. This was alternated with a few sessions of simple nature-art activities (e.g. making sun-catchers, leaf sketching) before another few sessions of horticultural-based activities. The program is outlined in . All the sessions are designed to stimulate engagement of participants' senses through . Activities in the 24-week therapeutic horticulture intervention. "H" refers to horticultural-based activity and "A" refers to nature-art activity.
Week Activity Type Week Activity Type www.nature.com/scientificreports/ touching, seeing and smelling the plant materials. While nature-art activities promote a sense of creativity and achievement, offering immediate gratification, horticultural-based activities encourage a waiting and nurturing attitude, offering a deep sense of purpose in the longer term. Hence, the combination of the two types of activities enabled the participants to experience complementary benefits throughout the program. The team of 4 who facilitated the program comprised personnel with experience in conducting urban farming workshops for older adults and they have received training by the AHTA registered horticultural therapy expert.
The sessions were carried out in a public garden, the Therapeutic Garden @HortPark, in five separate groups, with a mean group size of eight seniors (SD = 2.4, range 7-12). The garden has trees that provide shade, water features that soothe and calm, and plants that stimulate the sense of hearing, sight, touch and smell. Overall, the site was designed to bring a feeling of peace and wellbeing to its users. In addition, there are features such as customized raised planters to facilitate gardening by elderly participants. Throughout the intervention, the participants were provided transport in a minibus from their respective SAC to the Therapeutic Garden @Hort-Park. Each session followed an organized structure, comprising simple stretching exercises, introduction to the activity, hands-on activity and a conclusion phase where the participants shared their thoughts of the session.
Outcome measures. Affective outcome. To evaluate affect before and after the intervention throughout the 24-week program, we used the simple Visual Analogue Scale (VAS), originally developed by Hayes and Patterson. We used VAS to measure momentary positive affect as its ease of administration did not further burden the respondents. VAS before and after each session helped capture the momentary change in affect in the 24-month long study, where external changes surrounding participants may act as confounding factors. The happiness level was measured by asking participants to place a single vertical mark on a horizontal line at a point corresponding to their current subjective degree of happiness. The VAS, with the word "unhappiest" indicated on the left end and on the word "happiest" on the right end, is a simple and useful method to evaluate degree of subjective happiness in the elderly people [bib_ref] Evaluation of subjective happiness in the elderly using a visual analogue scale..., Matsubayashi [/bib_ref]. Higher scores indicate a happier state. The scale has been tested to be a valid and reliable single item happiness instrument that is suitable for a wide range of respondents, including those illiterate [bib_ref] Does it matter how happiness is measured? Evidence from a randomized controlled..., Studer [/bib_ref]. VAS was introduced in the study after the first group (n = 12) completed intervention on the remaining 35 participants.
Cognitive outcomes. To evaluate the course of cognitive changes in participants over time, as well as to document their response to the treatment program, we used the MMSE [bib_ref] Mental State Examination (MMSE) probably one of the most cited papers in..., Nilsson [/bib_ref]. As the majority of participants were above 70 years old, the Clinical Dementia Rating (CDR), which characterizes cognitive and functional performance in the domains of memory, orientation, judgment & problem solving, community affairs, home & hobbies, and personal care, was used to provide an assessment of whether any of them developed dementia during the 24 month period. We had planned to notify the center manager concerned for follow up should the CDR score indicates mild dementia, but all participants maintained CDR scores within the healthy range throughout.
Functional outcomes. Sleep hygiene was assessed using Pittsburgh Sleep Quality Index (PSQI) [bib_ref] The Pittsburgh sleep quality index: a new instrument for psychiatric practice and..., Buysse [/bib_ref]. The PSQI, a self-rated questionnaire, is a valid and reliable instrument for assessing sleep quality and disturbances in the non-clinical population [bib_ref] Criterion validity of the Pittsburgh Sleep Quality Index: investigation in a non-clinical..., Grandner [/bib_ref]. It has 19 items in seven "component" scores: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication, and daytime dysfunction.
To measure participants' functional ability in their basic activities of daily living (ADL), we used the Barthel Index (BI) [bib_ref] Functional evaluation: the Barthel Index: a simple index of independence useful in..., Mahoney [/bib_ref] , which is commonly used to assess behaviours relating to everyday activities of post-stroke patients or those with other disabilities. Participants rated their ability on ten ADL functions, including feeding, personal hygiene and toilet use. Lower scores indicate less independence, whereas higher scores indicate greater independence. The BI is a valid and reliable instrument that has been recommended as the standard approach to assess physical disability [bib_ref] The Barthel ADL index: a standard measure of physical disability, Wade [/bib_ref].
Psychosocial and Physical outcomes. To evaluate participants' psychosocial health, we used the Zung Self-Rating Depression Scale (SDS) [bib_ref] A self-rating depression scale, Zung [/bib_ref] and the Zung Self-Rating Depression Scale (SAS) [bib_ref] A rating instrument for anxiety disorders, Zung [/bib_ref] respectively. The SDS is a 20-item questionnaire that examines the affective, psychological and somatic symptoms associated with depression. The SAS questionnaire, on the other hand, has 20 items that measure anxiety levels, detecting cognitive, autonomic, motor and central nervous system symptoms. Each item is scored on a Likert scale ranging from 1 to 4, and the total is termed the raw score. Both the SDS and SAS have been shown to be valid and reliable instruments for screening anxiety and depression [bib_ref] Screening for anxiety and depression: reassessing the utility of the Zung scales, Dunstan [/bib_ref].
To evaluate if participants were socially isolated, we used the five-item Friendship Scale (FS). The FS has been shown to be a valid and reliable tool to assess degree of social connectedness and loneliness [bib_ref] Measuring social isolation in older adults: development and initial validation of the..., Hawthorne [/bib_ref]. Subjects who are very socially isolated will obtain scores in the range 0-11, those who are isolated or having little social support will obtain scores of 12-15, those with some social support will obtain score of 16-18, and those who are socially connected range will obtain 19-21. The very socially connected will score within the range 22-24 [bib_ref] Measuring social isolation in older adults: development and initial validation of the..., Hawthorne [/bib_ref].
Participants' self-perceived health states were evaluated using the EQ-5D-3L visual analogue scale (EQ VAS). They rated their health on the vertical scale where the endpoints are labelled 'Best imaginable health state' and 'Worst imaginable health state' . The EQ VAS has been found to be good scale for assessing older adults, with adequate acceptability and feasibility [bib_ref] EQ-5D-3L for assessing quality of life in older nursing home residents with..., Pérez-Ros [/bib_ref]. To compare the findings with previous studies from other populations, the results were additionally analysed across different age groups.
Time-points of measurement. To understand the changes in participants' functional, psycho-social and physical domains before, during and after the TH intervention, the respective outcome variables measured at baseline Scientific Reports | (2020) 10:18178 | https://doi.org/10.1038/s41598-020-74828-w www.nature.com/scientificreports/ (up to two weeks before intervention) were compared with outcomes at 3 months (mid-intervention), 6 months (within one week post-intervention), 9 months (3 months post intervention) and 12 months (6 months post intervention). The specific time-points were chosen to assess dose-response relationship with TH exposure (3 months), and the long-term effects (9 and 12 months) to inform future public health recommendations. As cognitive impairments progress slowly, it is important to measure outcomes after a prolonged period, here 12 months. To understand the effect of each activity on participants' momentary affect, the VAS was administered one hour before the start of each session at the respective SAC, and immediately after the session ended in the garden. The time points for the various measurements are illustrated in [fig_ref] Figure 1: Outcome measures at different time points [/fig_ref].
# Statistical analysis.
We used the ANOVA repeated measures to compare the group means at the different time points, and Greenhouse-Geisser correction was applied where homogeneity of variance assumption was violated. Bonferroni corrected means were reported for the post-hoc comparisons. The paired-samples t-test was used to determine whether there was a statistically significant mean difference in individual VAS scores of happiness level before and after each session. All analyses were performed using IBM SPSS Statistics version 26. Alpha level of 0.05 was used as a level of statistical significance. Due to follow-up issues after the completion of the intervention, the responses at timepoints 3 (9-month) and 4 (12-month) dropped to 38 and 44 respectively. The cases with missing values were omitted in the pairwise comparisons.
# Results
Sample characteristics. The general baseline characteristics of the participants are summarized in [fig_ref] Table 2: Baseline characteristics of the study participants [/fig_ref].
The majority were females above 70 years old, of Chinese ethnicity and retired. The sample varied in socio-economic status, as indicated by housing type and years of schooling. Six participants required assistance to move about during the intervention, either in the form of wheelchair or walking aids.
## Effects of the intervention on participants' momentary affect.
We measured participants' happiness level before and after each session using the VAS. 753 sets of pre-post responses from 35 participants were collected and analyzed. A score of "10" represents the happiest whilst "1" represents most unhappy. Post-session
## Effects of the intervention on outcomes over time.
We measured outcomes in the cognitive, functional psychosocial and physical domains at the timepoints illustrated in [fig_ref] Figure 1: Outcome measures at different time points [/fig_ref]. Mean values at each time point and the results of statistical analyses for all the outcomes are summarised in [fig_ref] Table 3: Summary of outcomes at the different timepoints and changes in outcomes over... [/fig_ref] MMSE. The highest possible score for MMSE is 30, representing maximum cognitive performance. The results showed that there was statistically significant difference in mean MMSE score among the three time points measured. The TH treatment elicited a statistically significant increase of 0.87 in participants' mean MMSE score at time point 2 compared to baseline, p = 0.01. At time-point 4, six months post treatment, there was a marginal drop of MMSE score by 0.07, but this was not statistically significant.
CDR. The scoring of CDR ranges between 1 and 5. A score close to 0 indicates no dementia and a score of 1 indicates mild dementia. A higher score corresponds to an increasing severity in one's cognitive decline. The CDR was administered to assess if any participants had risks of dementia during the 24-month period. We had the intention of referring them for follow-up if needed. The results showed that there was a mild but statistically significant increase in the CDR scores over time, but overall, the scores were within the healthy range of close to 0. whereas the lowest score of 0 represents a totally dependent state [bib_ref] A self-rating depression scale, Zung [/bib_ref]. The results showed that the mean scores in BI were close to 100 at all the five timepoints. There was no significant difference in the scores among the timepoints.
PSQI. All 7 questions are scored on a scale of 0-3, whereby 3 reflects the negative extreme. The maximum global score for PSQI is 21. A score of 5 or greater indicates a poor sleeper [bib_ref] Evaluating sleep quality in older adults: the Pittsburgh sleep quality Index can..., Smyth [/bib_ref]. The results showed that the participants maintained healthy sleep quality at all the timepoints measured. They showed marginally better PSQI global scores compared to baseline during the intervention period at timepoints 2 and 3, but this was not significant (p > 0.05).
## Sds and sas.
Participants' depression and anxiety states were assessed using SDS and SAS. The mean SDS and SAS scores at time points 0, 1, 2 3 and 4 are summarised in [fig_ref] Table 3: Summary of outcomes at the different timepoints and changes in outcomes over... [/fig_ref]. They are below the cut-off points [bib_ref] Norms for Zung's self-rating anxiety scale, Dunstan [/bib_ref] , which are 50 and 40 respectively. Higher scores indicate presence of depressive and anxiety disorders. The results showed that participants' SDS and SAS scores were within the healthy range at all the five timepoints. There was marginal improvements (reduction) in mean SDS scores at the different timepoints compared to baseline but this was not significant. On the other hand, there was significant improvement in SAS scores across the time points. Compared to baseline, the TH intervention elicited significant improvements in mean SAS scores at timepoints 2 and 3, by − 2.18 ± 0.90 (p < 0.05) and − 1.87 ± 0.60 (p = 0.03).
## Fs.
Participants' sense of social connectedness was assessed using the FS. There was no significant difference in mean FS scores among the different timepoints, and all the scores were equivalent to being socially connected.
## Eq vas.
There were no significant differences in the mean EQ VAS scores across the different time points.
The mean scores for participants aged 65-74 and 75 and above are summarised in .
# Discussion
The current study provided evidence in support of the existing literature on the effectiveness of TH in enhancing well-being of older adults, filling the information gap for the Asian phenotype in the tropical environment and providing novel information on the long-term effectiveness. The intervention resulted in several health benefits for the participants, namely more positive momentary affect, improved cognitive function and reduced anxiety. The effects of the latter two areas were sustained for up to 6 months after the intervention was completed. Throughout and six months after the intervention, the participants sustained their functional performance, maintaining healthy BI and PSQI scores. This is deemed to be a positive outcome, considering majority of them were aged 70 years and above, and would normally deteriorate in functional health in the span of one year. Similarly, they maintained a non-depressive state and a healthy ED VAS score. The mean EQ VAS scores for participants who were aged 65-74 years (n = 15) and 75 years and above (n = 29) during and after the intervention were consistently better that the population norm in the same age group in Thailand 36 and the United Kingdom 37 . Overall, this study has demonstrated the positive effects of the 24-week TH on older adults in the tropics, in improving their momentary positive affect, cognitive performance and reduced anxiety over time. Urban nature supports health by functioning as "salutogens", factors in the environment that support health and well-being, to promote positive experiences and behavior. Its benefits on the health and wellbeing of older . Mean EQ VAS scores in participants of two age groups during and after intervention. www.nature.com/scientificreports/ adults have economic importance, considering the societal expenditures on health and related services for the group. Singapore, with 100% urban population, is expected to have more than 20% of its population over the age of 65 years by 2030. According to a recent survey "Understanding the Quality of Life of Seniors in Singapore", older adults in the country reported a lower quality of life than the general population in areas such as recreation, leisure and affect. We think that Singapore is well placed to introduce nature-based solutions to improve the quality of life of the older adult population, considering the profuse greenery and accessible green spaces managed throughout the entire island. Moreover, older adults have more hours of leisure time, and stand to derive deeper benefits by taking part in TH as a form of nature-based recreation and occupation. There are several limitations to the study. For ethical reasons, a pretest posttest methodology was selected over a randomized control trial as the recruitment centres raised concerns about withholding the treatment from the potential control group. Additionally, budget constraints precluded adopting a waitlist randomized control trial approach. We were not able to include biological makers as objective assessments, as we had done in our past study [bib_ref] Effects of horticultural therapy on asian older adults: a randomized controlled trial, Ng [/bib_ref] , for technical reasons linked to sample storage. The strength of this study was the long-term assessment of a lifestyle nature-based programme that measured several cognitive, executive function, lifestyle and behavioural outcomes in the often-overlooked population of older adults, with a relatively low drop-out rate.
# Conclusion
The current study evaluated the effects of a 24-week TH intervention programme on older adults conducted in the Therapeutic Garden @HortPark (Singapore), which has been specifically designed to support mental and physical health. The therapeutic programme was centred around nature exposure and caring for plants, a nurturing process that has been shown to promote physical and mental health, as well as development of a personal relationship with nature. The TH program promoted understanding and exploration of nature and cultivated a sense of purpose in the participants. In addition, we found that the participants maintained healthy sleep, psychological health and showed improved cognitive functioning over time. This study provides evidence that structured nature-based programmes play a positive role in maintaining and enhancing the physical and mental health and wellbeing in older populations. We conclude that the 24-week TH program is a promising intervention to be extended in other SACs in Singapore, to improve the wellbeing of older adults. We recommend that TH be introduced as a part of healthy aging recommendations in tropical climates and beyond.
## Data availability
The data that support the findings of this study are available from the corresponding author, A.S., upon reasonable request.
Received: 17 April 2020; Accepted: 5 October 2020
[fig] Figure 1: Outcome measures at different time points. The maximum score of 100 represents a fully independent subject capable of performing basic ADL [/fig]
[table] Table 2: Baseline characteristics of the study participants. In Singapore, higher socio-economic status is linked with living in Landed property or Condominium (private estate). Other housing options are indicative of government housing. [/table]
[table] Table 3: Summary of outcomes at the different timepoints and changes in outcomes over time. ANOVA was used to analyse differences over time (significant differences are in bold) MMSE, Mini Mental State Exam, CDR, Clinical Dementia Rating, BI, Barthel Index, PSQI, Pittsburgh Sleep Quality Index, SDS, Zung Self-Rating Depression Scale, SAS, Zung Self-Rating Depression Scale, FS, Friendship Scale, EQ VAS, EQ-5D-3L Visual Analogue Scale. [/table]
|
The relative efficacy of bona fide psychotherapies for post-traumatic stress disorder: a meta-analytical evaluation of randomized controlled trials
Background: In the treatment of PTSD, meta-analyses suggest comparable efficacy of cognitive behavioural therapies and various trauma focused treatments, but results for other treatments are inconsistent. One meta-analysis found no differences for bona fide therapies, but was critizised for overgeneralization and a biased study sample and relied on an omnibus test of overall effect size heterogeneity that is not widely used. Methods: We present an updated meta-analysis on bona fide psychotherapies for PTSD, contrasting an improved application of the omnibus test of overall effect size heterogeneity with conventional random-effects meta-analyses of specified treatment types against all others. Twenty-two studies were eligible, reporting 24 head-to-head comparisons in randomized controlled trials of 1694 patients. Results: Head-to-head comparison between trauma focused and non-trauma focused treatments revealed a small relative advantage for trauma focused treatments at post-treatment (Hedges' g = 0.14) and at two follow-ups (g = 0.17, g = 0.23) regarding PTSD symptom severity. Controlling and adjusting for influential studies and publication bias, prolonged exposure and exposure therapies (g = 0.19) were slightly more efficacious than other therapies regarding PTSD symptom severity at post-treatment; prolonged exposure had also higher recovery rates (RR = 1.26). Present-centered therapies were slightly less efficacious regarding symptom severity at post-treatment (g = −0.20) and at follow-up (g = −0.17), but equally efficacious as available comparison treatments with regards to secondary outcomes. The improved omnibus test confirmed overall effect size heterogeneity. Conclusions: Trauma focused treatments, prolonged exposure and exposure therapies were slightly more efficacious than other therapies in the treatment of PTSD. However, treatment differences were at most small and far below proposed thresholds of clinically meaningful differences. Previous null findings may have stemmed from not clearly differentiating primary and secondary outcomes, but also from a specific use of the omnibus test of overall effect size heterogeneity that appears to be prone to error. However, more high-quality studies using ITT analyses are still needed to draw firm conclusions. Moreover, the PTSD treatment field may need to move beyond a focus primarily on efficacy so as to address other important issues such as public health issues and the requirements of highly vulnerable populations.
# Background
There is an on-going debate on whether psychotherapies differ in efficacy in the treatment of post-traumatic stress disorder (PTSD). In the broader context of psychotherapy research this is a well-known topic since Rosenzweig [bib_ref] Some implicit common factors in diverse methods of psychotherapy, Rosenzweig [/bib_ref] introduced the assumption that all psychotherapies are equally effective because of underlying common factors. This topic still remains mostly controversial and unsolved [bib_ref] The Dodo Bird Verdict -controversial, inevitable and important: A commentary on 30..., Budd [/bib_ref] , but most notably owing to progress in meta-analytical methodology, some important contributions from an evidence-based point of view could be made for which methodological aspects appeared crucial as well. Whereas early extensive meta-analyses did not make precise distinctions between disorders and outcomes [bib_ref] Meta-analysis of psychotherapy outcome studies, Smith [/bib_ref] [bib_ref] A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, Wampold [/bib_ref] , but included different psychotherapies for different disorders, critics warned against overgeneralization, undifferentiated methodology, and confounded results [bib_ref] Beware the Dodo bird: The dangers of overgeneralization, Chambless [/bib_ref] [bib_ref] A conceptual and methodological analysis of the nonspecifics argument, Derubeis [/bib_ref] [bib_ref] Mediators and moderators in meta-analysis: There's a reason we don't let dodo..., Shadish [/bib_ref]. Empirical justification for this critique was provided, for example, in the treatment of anxiety disorders: there, an advantage of cognitive behavioural therapy (CBT) over relaxation treatment pertained in panic disorder but not generalized anxiety disorder, and was evident in primary, symptomoriented, outcomes, but not in secondary, more general, outcomes [bib_ref] Specificity of treatment effects: Cognitive therapy and relaxation for generalized anxiety and..., Siev [/bib_ref] [bib_ref] The dodo bird, treatment technique, and disseminating empirically supported treatments, Siev [/bib_ref].
With regards to PTSD, existent meta-analytical reviews suggest that CBT, trauma focused CBT (TFCBT), exposure therapies, and eye movement desensitization reprocessing (EMDR) are equally effective, but results of non-trauma focused therapies like hypnotherapy, psychodynamic therapy or supportive therapy are heterogeneous and inconsistent, meaning either that they are less effective or were yet not sufficiently examined to prove their efficacy [bib_ref] Psychological treatment of post-traumatic stress disorder (PTSD), Bisson [/bib_ref] [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] [bib_ref] A multidimensional metaanalysis of psychotherapy for PTSD, Bradley [/bib_ref] [bib_ref] Meta-analysis of psychological treatments for posttraumatic stress disorder in adult survivors of..., Ehring [/bib_ref] [bib_ref] Efficacy of exposure versus cognitive therapy in anxiety disorders: Systematic review and..., Ougrin [/bib_ref] [bib_ref] A meta-analytic review of prolonged exposure for posttraumatic stress disorder, Powers [/bib_ref] [bib_ref] Comparing the efficacy of EMDR and traumafocused cognitive-behavioral therapy in the treatment..., Seidler [/bib_ref] [bib_ref] Effects of psychotherapeutic treatments for PTSD: A metaanalysis of controlled clinical trials, Sherman [/bib_ref] [bib_ref] Efficacy of group treatment for posttraumatic stress disorder symptoms: A meta-analysis, Sloan [/bib_ref] [bib_ref] Comparative efficacy of treatments for post-traumatic stress disorder: A meta-analysis, Van Etten [/bib_ref] [bib_ref] Metaanalysis of the efficacy of treatments for posttraumatic stress disorder, Watts [/bib_ref]. Especially the status of supportive therapies appears currently unclear, leaving the question open whether specific techniques really make any difference over common factors [bib_ref] The dodo bird, treatment technique, and disseminating empirically supported treatments, Siev [/bib_ref] [bib_ref] Is cognitive-behavioral therapy more effective than other therapies? A meta-analytic review, Tolin [/bib_ref].
Meta-analytical evidence presented by Benish, Imel, and Wampold [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] fuelled this debate, indicating that any differences in efficacy of psychotherapies for PTSD vanish once the analysis is restricted to head-to-head comparisons, thus controlling for some possible confounders, like differences in diagnostic criteria or length of treatment [bib_ref] Mediators and moderators in meta-analysis: There's a reason we don't let dodo..., Shadish [/bib_ref] , and includes only 'bona fide' psychotherapies, i.e., treatments that are intended to be therapeutic. Bona fide criteria (see the section on Eligibility below) were utilized by this research group also in a number of other meta-analyses [bib_ref] Distinctions without a difference: Direct comparisons of psychotherapies for alcohol use disorders, Imel [/bib_ref] [bib_ref] Direct comparisons of treatment modalities for youth disorders: A meta-analysis, Miller [/bib_ref] [bib_ref] A meta-(re)analysis of the effects of cognitive therapy versus 'other therapies' for..., Wampold [/bib_ref] , but were utilized since by some independent other researchers as well [bib_ref] Is cognitive-behavioral therapy more effective than other therapies? A meta-analytic review, Tolin [/bib_ref] [bib_ref] Comparing bona fide psychotherapies of depression in adults with two meta-analytical approaches, Braun [/bib_ref] [bib_ref] Psychotherapy versus secondgeneration antidepressants in the treatment of depression: A meta-analysis, Spielmans [/bib_ref]. The intention of using bona fide criteria is to restrict the analysis to head-to-head comparisons of active treatments to raise its validity. Including studies with mere 'intent-to-fail' control conditions introduces heterogeneity and bias, which may result in spurious differences of treatment efficacy in meta-analysis [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] [bib_ref] Determining what works in the treatment of PTSD, Wampold [/bib_ref].
The Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] meta-analysis has been critizised for a biased selection of the available evidence and overgeneralization [bib_ref] Do all psychological treatments really work the same in posttraumatic stress disorder?, Ehlers [/bib_ref]. For example, Cloitre [bib_ref] Effective psychotherapies for posttraumatic stress disorder: A review and critique, Cloitre [/bib_ref] included 44 comparisons from 27 studies, whereas Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] included only 17 comparisons from 15 studies. Specifically, Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] were critizised for including present-centered therapies and therapies whose efficacy appeared debatable, but excluding supportive therapies [bib_ref] Do all psychological treatments really work the same in posttraumatic stress disorder?, Ehlers [/bib_ref]. Moreover, supportive therapies have been categorized in other meta-analyses either as active treatments [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] [bib_ref] Effects of psychotherapy with people who have been sexually assaulted: A meta-analysis, Taylor [/bib_ref] or mere control conditions [bib_ref] A multidimensional metaanalysis of psychotherapy for PTSD, Bradley [/bib_ref] [bib_ref] A meta-analytic review of prolonged exposure for posttraumatic stress disorder, Powers [/bib_ref]. Application of an operational definition of what constitutes an 'active treatment' is therefore needed (e.g., the bona fide criteria), in order to make sense of differing results and conclusions reached by extant reviews and meta-analyses.
However, as of yet, there have been no comprehensive attempts to update and re-evaluate these issues and results empirically. One further meta-analysis [bib_ref] Efficacy of group treatment for posttraumatic stress disorder symptoms: A meta-analysis, Sloan [/bib_ref] examined group treatments for PTSD and found no indication of differences in treatment efficacy; however, this result was based on a sample of ten studies wherein treatments were not necessarily bona fide. Ougrin [bib_ref] Efficacy of exposure versus cognitive therapy in anxiety disorders: Systematic review and..., Ougrin [/bib_ref] found no differences between cognitive therapy and exposure therapy, aggregating the evidence of only five studies. Munder et al. [bib_ref] Is the allegiance effect an epiphenomenon of true efficacy differences between treatments?..., Munder [/bib_ref] examined differences between various trauma focused treatments, but only with regards to their dependency on researcher allegiance. Watts et al. [bib_ref] Metaanalysis of the efficacy of treatments for posttraumatic stress disorder, Watts [/bib_ref] presented a comprehensive metaanalysis on the efficacy of treatments for PTSD, including also pharmacotherapies, but did not strictly rely on headto-head comparisons and psychotherapies were not specifically bona fide. Tolin [bib_ref] Is cognitive-behavioral therapy more effective than other therapies? A meta-analytic review, Tolin [/bib_ref] , focusing specifically on CBT, reported an advantage of CBT over other psychotherapies in the treatment of anxiety and depressive disorders, but not for PTSD. A recently published meta-analysis [bib_ref] Meta-analysis of psychological treatments for posttraumatic stress disorder in adult survivors of..., Ehring [/bib_ref] , that was not restricted to head-to-head comparisons, reported for adult survivors of childhood abuse best effects for trauma focused treatments in an individual setting. Therefore, the studies conducted in this area have been mixed in their findings and fairly heterogeneous in their approach.
There is yet one further important aspect of the Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] meta-analysis that did not receive a re-evaluation. Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] used a specific method (see Methods) which was proposed by Wampold et al. [bib_ref] A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, Wampold [/bib_ref] in order to circumvent the need for a priori treatment categorizations. Instead of meta-analytically comparing individual classes of treatment against all others, as in other metaanalyses on relative efficacy, Wampold et al. [bib_ref] A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, Wampold [/bib_ref] proposed an omnibus test of overall effect size heterogeneity that does not rely on a priori treatment categorizations, but on the random assignment of positive (+) or negative signs (−) to the effect sizes of the head-to-head comparisons. The results of the first meta-analysis which used this approach [bib_ref] A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, Wampold [/bib_ref] had a huge impact on the debate on whether psychotherapies really differ in efficacy. However, this test cannot be considered a standard procedure, because it is not commonly used in meta-analytical research on relative efficacy. In fact, it seems that it is used only in metaanalyses of the same researchgroup [bib_ref] A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, Wampold [/bib_ref] [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] [bib_ref] Distinctions without a difference: Direct comparisons of psychotherapies for alcohol use disorders, Imel [/bib_ref] [bib_ref] Direct comparisons of treatment modalities for youth disorders: A meta-analysis, Miller [/bib_ref] [bib_ref] Is the allegiance effect an epiphenomenon of true efficacy differences between treatments?..., Munder [/bib_ref]. Therefore, it remains currently unclear whether this method provides results that are consistent with conventional methods. Specifically, low statistical power and the risk of capitalizing on chance may threaten the statistical conclusion validity of this test (see Methods).
The main aim of this meta-analysis was to make a contribution to both the overall field of research on relative treatment efficacy with regards to methodology, and to the specific field of research on the relative efficacy of treatments for PTSD, utilizing advanced meta-analytical methodology. The current study re-addressed the question whether there are no differences in the efficacy of treatments for PTSD, utilizing the bona fide criteria for an operational definition of 'active treatments' , updating the available evidence, and evaluating the study selection and methodology used by Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] (see Methods section). Specifically, we improved the application of the proposed statistical method and contrasted its results with conventional meta-analysis of direct comparisons, and also explored whether trauma focused treatments were in direct head-to-head comparisons more efficacious than treatments that were non-trauma focused. Moreover, we also investigated effects of patient and treatment variables on relative efficacy and controlled for study quality.
# Methods
The methodology of this study adhered to PRISMA guidelines.
## Eligibility
Eligibility criteria for studies to be included in this metaanalysis were (see [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] : (1) a randomized and controlled study design (RCT), investigating the relative efficacy of (2) at least two bona fide psychotherapies; (3) therapies needed to be conducted in two or more sessions; (4) participants were adults that were (5) diagnosed with PTSD according to the valid edition of the DSM at the time of the respective study (DSM-III or DSM-IV); (6) PTSD symptom severity was assessed with self-report or clinician rating. Studies that provided insufficient information to compute effect sizes were excluded as were component, dismantling, and parameter studies. Moreover, we also excluded studies that contained mere additional analyses on previously published data, ensuring that primary data entered analysis only once.
Target group, treatment format and study quality did not serve as eligibility criteria in the present study. In order to examine the possible effect of various patient and treatment characteristics as well as the influence of study quality, moderator analyses were conducted (see the section on Additional Analyses, below). Taylor and Harvey [bib_ref] Effects of psychotherapy with people who have been sexually assaulted: A meta-analysis, Taylor [/bib_ref] examined the effects of psychotherapy specifically for people who have been sexually assaulted and Ehring et al. [bib_ref] Meta-analysis of psychological treatments for posttraumatic stress disorder in adult survivors of..., Ehring [/bib_ref] for adult survivors of childhood abuse. Sloan et al. [bib_ref] Efficacy of group treatment for posttraumatic stress disorder symptoms: A meta-analysis, Sloan [/bib_ref] and Bisson et al. [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] examined, amongst others, group therapies. However, other metaanalyses did not limit themselves to specific target groups or treatment formats .
To qualify as an 'active treatment' , the bona fide definition [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] [bib_ref] Determining what works in the treatment of PTSD, Wampold [/bib_ref] required that treatments had to be delivered by a trained therapist who adapted the treatment to patients on the basis of a therapeutic relationship (i.e., no delivery of a non-modifiable standard protocol, e.g., progressive muscle relaxation); treatments also needed to be conducted personally and face-to-face (i.e., no online treatments or treatments conducted with, e.g., audio material). Moreover, at least two of the following four criteria had to be fulfilled with regards to their descriptions in the studies: (a) a citation to an established school or approach to psychotherapy; (b) a description of the therapy that contained a reference to a psychological process (e.g., operant conditioning); (c) a reference to a treatment manual that was used to guide the delivery of the treatment; (d) the identification of active ingredients of the treatment and citations for these ingredients.
## Information sources and search
Primary studies were identified in two ways. We obtained the 15 studies investigated by Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] and performed a literature search in the databases MEDLINE, Cinahl Health, Psychology and Behavioral Sciences Collection, PsycINFO, PsycArticles, PubMed, Social Sciences Fulltext, and Web of Knowledge, using the keywords PTSD, posttraumatic, post-traumatic, post traumatic in combination with psychotherapy and treatment, and also in combination with relative efficacy, outcome study and comparative efficacy. We refined the search results by restricting it to 'adults' , 'humans' and 'randomized controlled trials' in some databases (see [fig_ref] Figure 1: Flow of studies through the selection process [/fig_ref]. In addition, we searched reference sections of previous meta-analyses of PTSD treatment, controlled studies of psychotherapy outcomes for PTSD, and literature reviews of PTSD treatment [bib_ref] Effective psychotherapies for posttraumatic stress disorder: A review and critique, Cloitre [/bib_ref] [bib_ref] Empirically supported psychological treatments for adult acute stress disorder and posttraumatic stress..., Ponniah [/bib_ref].
Moreover, we meticulously re-assessed the bona fide status of all treatments included in previous meta-analyses and of the treatments of the 27 studies included in the review of Cloitre [bib_ref] Effective psychotherapies for posttraumatic stress disorder: A review and critique, Cloitre [/bib_ref]. This review was specifically mentioned by Ehlers et al. [bib_ref] Do all psychological treatments really work the same in posttraumatic stress disorder?, Ehlers [/bib_ref] in their criticism with regards to a biased study sample of Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref].
## Study selection and data collection
Prior to study selection, descriptions of 112 treatments classified as bona fide in previous meta-analyses on PTSD [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] , alcohol abuse [bib_ref] Distinctions without a difference: Direct comparisons of psychotherapies for alcohol use disorders, Imel [/bib_ref] , and depression [bib_ref] A meta-(re)analysis of the effects of cognitive therapy versus 'other therapies' for..., Wampold [/bib_ref] as well as descriptions of treatments not classified as bona fide in these study samples were screened to ensure that we applied the bona fide criteria correctly. Moreover, we closely adhered to the descriptions and examples provided by [bib_ref] Determining what works in the treatment of PTSD, Wampold [/bib_ref].
The second author extracted the data and coded the studies. Cases where coding appeared unclear were resolved by discussion and consensus between both authors. Moreover, coding of included studies was compared afterwards with coding provided in other meta-analyses for the same studies, where available [bib_ref] A multidimensional metaanalysis of psychotherapy for PTSD, Bradley [/bib_ref] [bib_ref] A meta-analytic review of prolonged exposure for posttraumatic stress disorder, Powers [/bib_ref] [bib_ref] Comparing the efficacy of EMDR and traumafocused cognitive-behavioral therapy in the treatment..., Seidler [/bib_ref] [bib_ref] Effects of psychotherapy with people who have been sexually assaulted: A meta-analysis, Taylor [/bib_ref] ; in these cases, no differences in coding were observed. Wherever possible, intention-to-treat (ITT) analyses were used in meta-analytical computations.
## Data items
Coded and extracted data items pertained to (a) study characteristics: publication year, type of publication, country, type of analysis (ITT or completer only); (b) patient characteristics: total number, number per treatment arm, percentage of women, mean age, type of trauma, mean time since trauma, PTSD symptom severity prior to treatment, percentage of comorbid disorders (depression, anxiety, substance use disorder), site of patient recruitment (community, clinical, community and clinical mixed, other), use of psychotropic drugs; (c) treatment characteristics: type of treatment, treatment format (individual vs. group), treatment length, total number of sessions, session length; and (d) all relevant data with regards to the primary and secondary outcomes, clinically significant change, and dropout (see the section on Summary Measures, below).
## Study quality
Risk of bias was evaluated on the basis of the seven criteria of the Practice Guidelines from the International Society for Traumatic Stress Studies [bib_ref] Effective treatments of PTSD, Foa [/bib_ref] , complemented by standards of high-quality studies proposed by Cuijpers et al. [bib_ref] Psychotherapy for depression in adults: A meta-analysis of comparative outcome studies, Cuijpers [/bib_ref]. Study quality was evaluated with respect to whether (1) randomization was conducted adequately and independently from researchers and therapists (by using, e.g., computer software); (2) assessment was blinded; (3) studies reported reasons for dropout or analyses on differences between dropouts and completers; (4) therapists were specifically trained in the provided treatment; (5) adherence to treatment manuals or protocols was ensured with adequate measures or checked empirically; 6) ITT-analyses were conducted and adequately reported; (7) a treatment manual or protocol existed. Items were coded 0 = no, 1 = yes, and summed, resulting in an omnibus measure of study quality that ranged from 0 to 7. Additionally, we also investigated criteria (1), (2), and (3) separately. These criteria agree closely with the Jadad et al. [bib_ref] Assessing the quality of reports of randomized clinical trials: Is blinding necessary?, Jadad [/bib_ref] criteria 1 that are widely-used in medical research and that were already also used in psychotherapy research [bib_ref] A meta-analytic review of prolonged exposure for posttraumatic stress disorder, Powers [/bib_ref] [bib_ref] Is cognitive-behavioral therapy more effective than other therapies? A meta-analytic review, Tolin [/bib_ref] ; with regards to blinding, we considered only blinded assessment; double-blind trials, as proposed by Jadad et al. [bib_ref] Assessing the quality of reports of randomized clinical trials: Is blinding necessary?, Jadad [/bib_ref] are adequate in medical research, but are generally not feasible in psychotherapy research.
## Summary measures
First, between-group differences in primary and secondary outcomes (see below) were meta-analysed at posttreatment and, where available, at follow-up. Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] restricted their analyses to post-treatment differences. All direct measures of PTSD symptom severity (e.g., Clinician Administered PTSD Scale [CAPS], Impact of Events Scale [IES], PTSD Symptom Scale Self-Report [PSS-SR]) were considered primary outcomes. This was in contrast to Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] who considered measures of anxiety (e.g., State-Trait Anxiety Inventory [STAI]) primary outcomes as well. Anxiety is one of the key symptoms of PTSD and anxiety disorders are among the most frequent comorbid disorders of PTSD [bib_ref] Effective treatments of PTSD, Foa [/bib_ref]. However, measures of anxiety do not address the full range of PTSD symptoms. Hence, they were not considered primary outcomes in the present study. On the other hand, Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] computed a second effect size, aggregating all post-treatment outcome measures that were reported in individual studies. As these included primary outcomes as well, there was a substantial overlap of this second effect measure with the primary outcome measure. Instead of following such an approach, we considered all measures of symptom severity of comorbid disorders (i.e., anxiety, depression, substance use disorder), trauma-related symptoms (e.g., traumarelated guilt), general symptom distress, social functioning, and quality of life as secondary outcomes, thus having a true second measure for a more differentiated analysis, see [bib_ref] Specificity of treatment effects: Cognitive therapy and relaxation for generalized anxiety and..., Siev [/bib_ref] [bib_ref] The dodo bird, treatment technique, and disseminating empirically supported treatments, Siev [/bib_ref]. Following [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] , we also distinguished primary outcomes with regards to self-reports and clinician ratings. While the meta-analysis of Van Etten and Taylor [bib_ref] Comparative efficacy of treatments for post-traumatic stress disorder: A meta-analysis, Van Etten [/bib_ref] did not find any differences in efficacy with regards to assessment, others did [bib_ref] Effects of psychotherapy with people who have been sexually assaulted: A meta-analysis, Taylor [/bib_ref] [bib_ref] Self-reported versus clinicianrated symptoms of depression as outcome measures in psychotherapy research..., Cuijpers [/bib_ref]. A complete list of primary and secondary outcome measures is displayed in Appendix (Tables 5 and 6).
Second, additional outcomes pertained to treatment differences in percentages of clinically significant changes and dropouts at post-treatment and, where available, at follow-up. Where available, recovery rates were used to assess clinically significant change. Otherwise, rates of participants improved were utilized. As in the analysis of primary and secondary outcomes, data of ITT analyses were used, where available. Dropout rates were calculated either on patients dropping out of treatment or on the basis of ITT participant numbers (i.e., after randomization).
## Categorization of treatments
Based on descriptions and references given in the respective studies, treatments were categorized into specific types that were in line with previous meta-analyses [bib_ref] A multidimensional metaanalysis of psychotherapy for PTSD, Bradley [/bib_ref] [bib_ref] A meta-analytic review of prolonged exposure for posttraumatic stress disorder, Powers [/bib_ref] [bib_ref] Comparing the efficacy of EMDR and traumafocused cognitive-behavioral therapy in the treatment..., Seidler [/bib_ref] ] and a review [bib_ref] Effective psychotherapies for posttraumatic stress disorder: A review and critique, Cloitre [/bib_ref]. Treatment categorization in these studies followed a bottom-up strategy, sorting and clustering treatments according to their provided descriptions consistently into groups. The present metaanalysis also adhered to this bottom-up treatment categorization scheme, adhering as closely as possible to treatment descriptions in studies for the categorization of treatments. Similar to previous meta-analyses of PTSD comparing several treatment categories, categories resulting from this process were sometimes narrower and sometimes broader, depending on whether therapies followed a specified treatment manual (e.g., eye movement desensitization and reprocessing [EMDR], or were combined because of a similar theoretical approach (e.g., exposure therapies). In order to maximize statistical power, we were careful not to narrow categories down too much, as this would have yielded a large number of categories and only few studies per category. To compensate for this broadening of categories, we conducted also meta-analyses for distinct subgroups (see below) and controlled for outliers.
Types of treatment included in the analyses were: Exposure therapies (EX), prolonged exposure (PE [bib_ref] Treatment of posttraumatic stress disorder in rape victims: A comparison between cognitive-behavioral..., Foa [/bib_ref] , and both combined (EXPE), EMDR, cognitive behavioural therapies (CBT), and present-centered therapies (PCT). PCT included bona fide psychotherapies that were mostly linked to elements of Roger's clientcentered psychotherapyand encompassed supportive, present-centered, problemsolving, and psychoeducative elements. Treatments coded as CBT included some therapies that utilized, but did not primarily rely on, elements of other therapies (e.g., exposure [bib_ref] Randomized trial of cognitive-behavioral therapy for chronic posttraumatic stress disorder in adult..., Mcdonagh [/bib_ref] [bib_ref] A controlled comparison of eye movement desensiti-zation and reprocessing versus exposure plus..., Power [/bib_ref] , and treatments such as Seeking Safety [bib_ref] Promising treatments of women with comorbid PTSD and substance abuse, Hien [/bib_ref] or Image Rehearsal [bib_ref] Imagery Rehearsal for posttraumatic nightmares: a randomized controlled trial, Cook [/bib_ref].
All treatments were coded whether they were trauma focused (TF) or not (NTF). TF treatments were defined as focusing on the memory for the traumatic event and/ or on trauma-related appraisals [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref]. Najavits and Hien [bib_ref] Helping vulnerable populations: A comprehensive review of the treatment outcome literature on..., Najavits [/bib_ref] recommended to name non-trauma focused therapies for what they are providing and not just to qualify them for what they are not providing. Nevertheless, we adhered to TF/NTF distinction as it is frequently used in this field of research. The TF/NTF distinction served to strictly compare two categories on the same level of abstraction and to address the question of whether trauma focused treatments were more efficacious than non-trauma focused ones. Again, treatment descriptions in studies were utilized as closely as possible to code whether treatments were trauma focused or not. Within the CBT category we also examined the efficacy specifically of trauma focused treatments that included elements of exposure (TFCBT + EX).
# Synthesis of results
For primary and secondary outcomes, Hedges' g was used as effect size measure, similar to previous meta-analyses [bib_ref] A meta-analytic review of prolonged exposure for posttraumatic stress disorder, Powers [/bib_ref] [bib_ref] Comparing the efficacy of EMDR and traumafocused cognitive-behavioral therapy in the treatment..., Seidler [/bib_ref] [bib_ref] Effects of psychotherapeutic treatments for PTSD: A metaanalysis of controlled clinical trials, Sherman [/bib_ref] [bib_ref] Effects of psychotherapy with people who have been sexually assaulted: A meta-analysis, Taylor [/bib_ref] , but in contrast to Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] who used Cohen's d. Hedges' g is widely recommended in the metaanalytical literature to control for the overestimation of effects with Cohen's d in small samples and its use is also recommended in the Practice Guidelines from the International Society for Traumatic Stress Studies [bib_ref] Effective treatments of PTSD, Foa [/bib_ref]. In cases where studies reported more than one outcome (primary or secondary), effect sizes were obtained for each outcome and then averaged within outcomes; dependency between outcomes was accounted for by assuming a correlation of .50 in the computation of the variance of the averaged effect size as in Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref]. In cases where both full-scale and subscale scores of a measure were reported, full-scale scores were used. In cases where only subscale, but not full-scale, scores were reported, published coefficients on the intercorrelation of subscales was used instead of assuming a uniform correlation of .50 in computing the averaged effect size variance (e.g., for the IES-revised form, coefficients were taken from [bib_ref] Psychometric properties of the Impact of Event Scale -Revised, Creamer [/bib_ref] , p. 1491). For additional outcomes, percentages of clinically significant changes and dropouts, risk ratios (RR) were computed and used in analysis.
For all analyses, random-effects models were utilized, assuming that treatment differences were not represented by a single population effect. With regards to analyses themselves, two independent approaches were utilized: (1) conventional meta-analysis of direct comparisons, contrasting specified types of treatment against all other available treatments, restricted to types with five or more comparisons, see [bib_ref] Psychotherapy for depression in adults: A meta-analysis of comparative outcome studies, Cuijpers [/bib_ref] , contrasting also trauma focused versus non-trauma focused treatments, and contrasting further all exposure-based treatments (EXPE) against all other treatments. The conventional analyses comprised comparisons regarding primary and secondary outcomes, clinical significance, and dropout as well as analyses of publication bias, sensitivity analyses, and moderator analyses; (2) the Wampold homogeneity test (WHT) with regards to primary and secondary outcomes at post-treatment [bib_ref] A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, Wampold [/bib_ref] [bib_ref] Meta-analytic methods to test relative efficacy, Wampold [/bib_ref].
For conventional meta-analyses, the software metafor [bib_ref] Conducting meta-analyses in R with the metafor package, Viechtbauer [/bib_ref] in R was used. Heterogeneity was tested with the Q test and assessed with the I 2 statistic that describes the ratio of true heterogeneity to the total observed dispersion with a range of 0 % to 100 %, reporting also 95 % confidence intervals. Heterogeneity was assumed small for I 2 = 25 %, moderate for I 2 = 50 %, and high for I 2 = 75 % [bib_ref] Quantifying heterogeneity in a meta-analysis, Higgins [/bib_ref].
With regards to the interpretation of standardized mean differences, i.e., Hedge's g, we adhered to benchmarks of Coheninterpreting g = 0.20 as small, .50 as medium, and 0.80 as large effects. For the interpretation of whether effects are likely clinically meaningful, we adhered to Ferguson [bib_ref] An effect size primer: A guide for clinicians and researchers, Ferguson [/bib_ref] and the NICE Guideline Development Group [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] who proposed thresholds of g = .41 and g = .50, respectively. With regards to the interpretation of risk ratios pertaining to recovery/improvement rates and dropout, we adhered to thresholds devised by the NICE Guideline Development Group [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] , regarding risk ratios ≤ 0.80 or ≥ 1.25 as clinically meaningful. Where 95 % confidence intervals of risk ratios included these thresholds, results were interpreted as 'limited evidence' of differential efficacy. Significance was set to p < .05.
The omnibus hypothesis, whether there were any differences in treatment efficacy overall, regardless of treatment categorization, was investigated with the WHT. The WHT requires in a first step the random assignment of a positive (+) or negative sign (−), in equal proportion [bib_ref] A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, Wampold [/bib_ref] , to the effect sizes of the head-to-head comparisons. The aggregated effect of the randomly aligned treatment differences will approach zero if k, the number of comparisons, is large, but this is not of direct interest in the WHT. In the second, crucial, step, the homogeneity of the obtained effect size distribution is investigated via the Q test. If the null hypothesis of no overall treatment differences is not true, then there should be a disproportionate number of observations in the tails of the distribution [bib_ref] A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, Wampold [/bib_ref]. Therefore, if homogeneity has to be rejected, the null hypothesis of no overall treatment differences must also be rejected. This null hypothesis has been tested with the WHT in a number of previous analyses [bib_ref] A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, Wampold [/bib_ref] [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] [bib_ref] Distinctions without a difference: Direct comparisons of psychotherapies for alcohol use disorders, Imel [/bib_ref] [bib_ref] Direct comparisons of treatment modalities for youth disorders: A meta-analysis, Miller [/bib_ref] and has never been rejected.
However, two aspects may compromise the validity of the WHT: (1) the Q test is known for its low test power when k is small [bib_ref] Quantifying heterogeneity in a meta-analysis, Higgins [/bib_ref] ; in the meta-analysis of Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] k was 17, heterogeneity may thus have been overlooked in previous research; (2) the WHT relies on only one run of random sign assignment. As stated above, random assignment need not result in an aggregated effect size of null; the probability of this event approaches unity only for k → ∞. Calculation of the Q statistic depends on the aggregated effect size. Random fluctuations of the aggregated effect size thus affect also the Q test of the WHT. Performing the WHT only once for a given dataset may therefore capitalize on random fluctuations whose possible effect is greater for small k. In conclusion, the Q test of the WHT cannot be considered an exact test, and Q values, as well as I 2 values derived from this Q statistic, cannot be considered exact statistics of effect size heterogeneity.
In the present study, we sought to overcome this problem by applying the WHT repeatedly on 30 independent runs of random sign assignments, which can be thought of as a Monte Carlo permutation test. Monte Carlo permutation tests are akin and asymptotically equivalent to (exact) permutation tests (i.e., Fisher's exact test) which obtain the distribution of a test statistic under the null hypothesis by calculating all possible values of this statistic by rearranging (i.e., permuting) the observed data in all possible ways. P values are obtained by computing the proportion of test statistic values that are greater than the test statistic of the observed data. Monte Carlo permutation tests sample only a number of permutations and do not perform all possible permutations. In the present case and with regards to the WHT, no clearly defined test statistic could be computed for the observed data: as explained above, the Q statistic of the WHT is itself not exact, but intrinsically fuzzy and subject to random fluctuations. However, the principles of Monte Carlo permutation tests could be applied to sample the distribution of Q values of the WHT and to obtain the number of significant Q tests for a given data set. Instead of relying on the p value of a single WHT, we counted the number of p values that were less than .05 among 30 replications of the WHT. Assuming the null hypothesis of no overall treatment differences is true, we expected 1.5 significant results (5 % of 30) with a nominal α of .05. To determine whether the null hypothesis of the WHT had to be rejected, the obtained number of significant results among the 30 replications was compared to this expected value, using a binomial test. This test had a power of more than 95 % to reject the null hypothesis of the WHT of no overall treatment differences, if the true number of significant Q values was at least 7.5 (i.e., 25 % of Q tests were significant).
## Publication bias
Publication bias in meta-analysis has to be investigated in order to check whether studies in the analysis are representative of the overall population of completed studies. A systematic bias could occur either because of one-sided search-strategies or because small studies or studies without significant results were not published on a regular basis. Publication bias was investigated with the trim-andfill procedure [bib_ref] Trim and fill: A simple funnel-plot-based method of testing and adjusting for..., Duval [/bib_ref] with regards to all meta-analyses of direct comparisons, using metafor in R.
## Additional analyses dependencies in the data
One of the studies (see Results) contained three comparisons. In order to test and correct for this dependence, all meta-analyses of direct comparisons involving these three comparisons were repeated, keeping only the headto-head comparison with the largest effect. Three other studies (see Results) compared treatments of the same type. Meta-analyses of direct comparisons involving these studies were repeated as well, excluding these studies.
## Sensitivity analyses
Furthermore, overall stability of meta-analyses of direct comparisons was examined with sensitivity analysis using the leave-one-out method, investigating whether there were influential outliers in the data. All of these analyses were used to investigate the robustness of obtained results and to adjust for possible outliers that may otherwise bias and distort results.
## Moderator analyses
Moderator analyses were carried out for meta-analyses of direct comparisons to explain unobserved heterogeneity, using mixed-effects models in cases where more than ten comparisons were available [bib_ref] Psychotherapy for depression in adults: A meta-analysis of comparative outcome studies, Cuijpers [/bib_ref] , testing for effects of study quality [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] [bib_ref] A meta-analytic review of prolonged exposure for posttraumatic stress disorder, Powers [/bib_ref] , type of analysis (ITT vs. completer only) [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] , participant sex (male vs. female vs. mixed) and proportion of women [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] [bib_ref] Efficacy of group treatment for posttraumatic stress disorder symptoms: A meta-analysis, Sloan [/bib_ref] [bib_ref] Metaanalysis of the efficacy of treatments for posttraumatic stress disorder, Watts [/bib_ref] , type of trauma [bib_ref] Efficacy of group treatment for posttraumatic stress disorder symptoms: A meta-analysis, Sloan [/bib_ref] [bib_ref] Metaanalysis of the efficacy of treatments for posttraumatic stress disorder, Watts [/bib_ref] , time since trauma [bib_ref] A meta-analytic review of prolonged exposure for posttraumatic stress disorder, Powers [/bib_ref] , PTSD symptom severity prior to treatment [bib_ref] Predicting treatment outcome on three measures for post-traumatic stress disorder, Karatzias [/bib_ref] (severe vs. extreme), and treatment length and total number of sessions [bib_ref] A meta-analytic review of prolonged exposure for posttraumatic stress disorder, Powers [/bib_ref] [bib_ref] Effects of psychotherapy with people who have been sexually assaulted: A meta-analysis, Taylor [/bib_ref].
# Results
## Study selection and study characteristics
In addition to the 15 studies included in Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] , we retrieved seven further studies that fulfilled the inclusion criteria, resulting in a total sample of 22 studies [bib_ref] Treatment of posttraumatic stress disorder in rape victims: A comparison between cognitive-behavioral..., Foa [/bib_ref] [bib_ref] Randomized trial of cognitive-behavioral therapy for chronic posttraumatic stress disorder in adult..., Mcdonagh [/bib_ref] [bib_ref] A controlled comparison of eye movement desensiti-zation and reprocessing versus exposure plus..., Power [/bib_ref] [bib_ref] Promising treatments of women with comorbid PTSD and substance abuse, Hien [/bib_ref] [bib_ref] Imagery Rehearsal for posttraumatic nightmares: a randomized controlled trial, Cook [/bib_ref] [bib_ref] Brief psychotherapy for posttraumatic stress disorders, Brom [/bib_ref] [bib_ref] A randomized controlled trial of exposure therapy and cognitive restructuring for posttraumatic..., Bryant [/bib_ref] [bib_ref] Randomized controlled comparison of cognitive behavior therapy with rogerian supportive therapy in..., Cottraux [/bib_ref] [bib_ref] The relative efficacy and treatment distress of EMDR and a cognitive-behavior trauma..., Devilly [/bib_ref] [bib_ref] A comparison of exposure therapy, stress inoculation training, and their combination for..., Foa [/bib_ref] [bib_ref] Comparison of two treatments for traumatic stress: A community-based study of EMDR..., Ironson [/bib_ref] [bib_ref] Treatment of PTSD: Stress inoculation training with prolonged exposure compared to EMDR, Lee [/bib_ref] [bib_ref] Treatment of posttraumatic stress disorder by exposure and/or cognitive restructuring: A controlled..., Marks [/bib_ref] [bib_ref] A comparison of narrative exposure therapy, supportive counseling, and psychoeducation for treating..., Neuner [/bib_ref] [bib_ref] Cognitive-behavior therapy vs. exposure therapy in the treatment of PTSD in refugees, Paunovic [/bib_ref] [bib_ref] Comparing virtual reality exposure therapy to present-centered therapy with 11 U.S. vietnam..., Ready [/bib_ref] [bib_ref] A comparison of cognitive-processing therapy with prolonged exposure and a waiting condition..., Resick [/bib_ref] [bib_ref] Prolonged exposure versus eye movement desensitization and reprocessing (EMDR) for PTSD rape..., Rothbaum [/bib_ref] [bib_ref] Randomized trial of trauma-focused group therapy for posttraumatic stress disorder, Schnurr [/bib_ref] [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref] [bib_ref] A randomized trial of cognitive therapy and imaginal exposure in the treatment..., Tarrier [/bib_ref] [bib_ref] Comparative efficacy, speed, and adverse effects of three PTSD treatments: Exposure therapy,..., Taylor [/bib_ref] (as the re-assessment of the study selection is an important topic of interest, see Additional file 1 for full references of the excluded studies). In sum, 161 abstracts were thoroughly screened and 46 studies assessed in full-text for eligibility, see [fig_ref] Figure 1: Flow of studies through the selection process [/fig_ref].
Studies were excluded because they (reference numbers pertain to the reference list in Additional file 1: (a) compared treatments that were not bona fide [bib_ref] Some implicit common factors in diverse methods of psychotherapy, Rosenzweig [/bib_ref] [bib_ref] The Dodo Bird Verdict -controversial, inevitable and important: A commentary on 30..., Budd [/bib_ref] [bib_ref] Meta-analysis of psychotherapy outcome studies, Smith [/bib_ref] [bib_ref] A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, Wampold [/bib_ref] [bib_ref] Beware the Dodo bird: The dangers of overgeneralization, Chambless [/bib_ref]
[formula] 2 ; (b) [/formula]
were component or dismantling studies [bib_ref] A conceptual and methodological analysis of the nonspecifics argument, Derubeis [/bib_ref] [bib_ref] Mediators and moderators in meta-analysis: There's a reason we don't let dodo..., Shadish [/bib_ref] [bib_ref] Specificity of treatment effects: Cognitive therapy and relaxation for generalized anxiety and..., Siev [/bib_ref] [bib_ref] The dodo bird, treatment technique, and disseminating empirically supported treatments, Siev [/bib_ref] [bib_ref] Psychological treatment of post-traumatic stress disorder (PTSD), Bisson [/bib_ref] [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] 3 ; (c) reported preliminary analyses of studies that were not included in the present meta-analysis themselves [bib_ref] A multidimensional metaanalysis of psychotherapy for PTSD, Bradley [/bib_ref] ; (d) reported additional analyses on data of studies already included in the present meta-analysis [bib_ref] Meta-analysis of psychological treatments for posttraumatic stress disorder in adult survivors of..., Ehring [/bib_ref] [bib_ref] Efficacy of exposure versus cognitive therapy in anxiety disorders: Systematic review and..., Ougrin [/bib_ref] [bib_ref] A meta-analytic review of prolonged exposure for posttraumatic stress disorder, Powers [/bib_ref] [bib_ref] Comparing the efficacy of EMDR and traumafocused cognitive-behavioral therapy in the treatment..., Seidler [/bib_ref] ; (e) did not examine adults and/or did not ascertain a PTSD diagnosis according to DSM-III or DSM-IV [bib_ref] Effects of psychotherapeutic treatments for PTSD: A metaanalysis of controlled clinical trials, Sherman [/bib_ref] [bib_ref] Efficacy of group treatment for posttraumatic stress disorder symptoms: A meta-analysis, Sloan [/bib_ref] [bib_ref] Comparative efficacy of treatments for post-traumatic stress disorder: A meta-analysis, Van Etten [/bib_ref] ; (f) included treatments that were delivered in a standard protocol that could not be adapted to individual patients [bib_ref] Metaanalysis of the efficacy of treatments for posttraumatic stress disorder, Watts [/bib_ref] [bib_ref] Is cognitive-behavioral therapy more effective than other therapies? A meta-analytic review, Tolin [/bib_ref] [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] ; (g) examined only two participants in one treatment arm [bib_ref] Distinctions without a difference: Direct comparisons of psychotherapies for alcohol use disorders, Imel [/bib_ref]. This study [bib_ref] Distinctions without a difference: Direct comparisons of psychotherapies for alcohol use disorders, Imel [/bib_ref] was excluded on statistical grounds as estimates of population variance, on which Hedges' g depend, are undefined when n < 3.
From the studies included in the review of Cloitre [bib_ref] Effective psychotherapies for posttraumatic stress disorder: A review and critique, Cloitre [/bib_ref] , only 3 studies remained which fully met the inclusion criteria and therefore were also included in the present meta-analysis [bib_ref] Promising treatments of women with comorbid PTSD and substance abuse, Hien [/bib_ref] [bib_ref] A comparison of narrative exposure therapy, supportive counseling, and psychoeducation for treating..., Neuner [/bib_ref] [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref]. Deviating from Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] , the Schnurr et al. [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref] study was included. This study investigated a PCT about which Wampold et al. [bib_ref] Determining what works in the treatment of PTSD, Wampold [/bib_ref] argued it was not bona fide because it bore little resemblance to PCT commonly applied in practice. However, according to the information provided in the study itself, the treatment fulfilled enough criteria (b, c, and d; see the section on Eligibility, above) to be considered bona fide. Moreover, Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] included another study of the same authors [bib_ref] Randomized trial of trauma-focused group therapy for posttraumatic stress disorder, Schnurr [/bib_ref] that contained basically the same treatment, only delivered in a group format. On these premises, the Schnurr et al. [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref] study was therefore included.
Basic characteristics of the 22 included studies, quality ratings, and individual effect sizes of primary and secondary outcomes are displayed in [fig_ref] Table 1: Characteristics of included studies [/fig_ref]. Overall, 1694 patients participated in the studies. Studies were published between 1989 and 2010 and 18 (82 %) were conducted in English-speaking countries, 11 of which in the United States. Patients' mean age ranged from 21.5 to 59.4 years, seven studies contained only women (total n = 664), three studies only men (total n = 460), in the remaining 12 studies (total n = 570) on average 56.0 % of patients were female. Twelve studies included patients with a specific trauma history or otherwise specific patient characteristics (total n = 1175): Childhood sexual abuse and adult rape victims (five studies, total n = 305), war veterans (four studies, total n = 744), fugitives (two studies, n = 51), comorbid substance use disorder (one study, n = 75). Fourteen studies provided information on time since trauma, which ranged from 1.2 to 22.9 years, with a mean of 7.1 years. With regards to PTSD symptom severity prior to treatment, eleven studies included severely traumatized patients (total n = 902; CAPS: 60-79, IES > 43), five studies included extremely traumatized patients (total n = 556; CAPS > 79). The remaining six studies did not report on severity or used instruments that provide no cut-offs with regards to severity (i.e., PDS, PSS-I, PSS-SR, PCL).
Details regarding the categorization of treatments in the present meta-analysis and in previous meta-analyses are reported in Appendix [fig_ref] Table 7: Treatment categorizations of included studies in the present meta-analysis compared to previous... [/fig_ref]. The psychodynamic therapy investigated in Brom et al. [bib_ref] Brief psychotherapy for posttraumatic stress disorders, Brom [/bib_ref] was neither categorized with regards to type of treatment, as it did not fit with any type, nor included in direct head-to-heads comparisons of TF and NTF treatments as it remained unclear whether or not it was trauma focused [bib_ref] Determining what works in the treatment of PTSD, Wampold [/bib_ref] ; it served only as an available comparator in meta-analyses of direct comparisons of specific treatment types. The hypnotherapy investigated in Brom et al. [bib_ref] Brief psychotherapy for posttraumatic stress disorders, Brom [/bib_ref] was categorized as EX, as the description in the study explicitly stated that "The emphasis of the hypnotherapists in our study was on behavioral therapy. The goal was to bring the patient in contact with the reality of the traumatic event and to bring about a decrease in the conditioned responses triggered by the event" (p. 607) and that hypnosis was used to facilitate this goal (meta-analyses [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] [bib_ref] A multidimensional metaanalysis of psychotherapy for PTSD, Bradley [/bib_ref] coded this treatment as 'other' treatment). The Hien et al. [bib_ref] Promising treatments of women with comorbid PTSD and substance abuse, Hien [/bib_ref] study investigated the efficacy of Seeking Safety, a manualized CBT that addresses both PTSD and substance use disorder, comparing it with another CBT (relapse prevention) that addresses only substance use disorder. Both treatments were coded as CBT, but in the CBT related meta-analyses only Seeking Safety was pooled with the other CBT treatments for PTSD (and contrasted with relapse prevention), as relapse prevention did not directly address the PTSD symptomatology.
Twenty-four head-to-head comparisons were reported in the 22 studies; one study compared three treatments. There were 15 comparisons involving CBT (of these, 10 were TFCBT + EX), seven involving EX, seven involving PE, six involving EMDR, and six involving PCT. Therapies were conducted in 12 sessions that lasted 90 min on average. Only four (8.9 %) treatments were carried out in a group format. Follow-up assessments were reported in all (100 %) studies with regards to at least one of the defined outcomes; time of follow-up ranged from 1 to 15 months after end of treatment, 4.7 months on average. Nine (40.9 %) studies reported a second follow-up, ranging 6 to 24 months after end of treatment, 10.0 months on average.
## Study quality
Ratings of study quality are presented in [fig_ref] Table 1: Characteristics of included studies [/fig_ref]. Most studies reported reasons for dropout or reported analyses of differences between completers and dropouts, ensured adherence to treatment manuals or protocols, and provided a treatment manual or protocol (18 [81.8 %] studies each). Fifteen (68.2 %) studies each reported blinded assessment and whether therapists were specifically trained in the provided treatment. Only eight (36.4 %) studies reported adequate randomization, whereas only seven (31.8 %) intention-to-treat analyses. Jadad ratings ranged between 0 and 4 points, with 2.2 points on average.
# Synthesis of results
Primary and secondary outcomes: Conventional meta-analysis of direct comparisons [fig_ref] Table 2: Meta-analyses of the efficacy of different types of treatment at post-treatment regarding... [/fig_ref] displays meta-analyses with regards to primary and secondary outcomes at post-treatment. No specific type of treatment had a higher efficacy than other available comparison treatments. However, PCTs were less efficacious with regards to primary, but not secondary, outcomes than their available comparison treatments. Trauma focused treatments were in direct head-to-head Note. a Not categorized with regards to type of treatment; b categorized according to description given in study; c used for comparison of TF versus NTF therapies. Figures of study quality correspond to ratings (+: present; 0: absent) whether (1) randomization was adequate, (2) assessment was blinded, (3) reasons for dropout or analyses on differences between dropouts and completers were reported, (4) therapists were trained in the provided treatment, (5) adherence to treatment manuals or protocols was ensured with adequate measures or checked empirically, (6) ITT-analyses were conducted and adequately reported, (7) a treatment manual or protocol existed (in this order); Jadad scores (min = 0, max = 4), based on these ratings, are reported after the colon. Outcomes correspond to estimates of Hedges' g along with 95 % confidence intervals; g > 0 signifies that effects at post-treatment were larger for the first of the two compared treatments, g < 0 signifies that effects were smaller comparisons to non-trauma focused treatments (TF vs. NTF) significantly more efficacious in primary outcomes. Heterogeneity appeared low for exposure therapies, for the head-to-head comparisons of TF vs. NTF and PCTs, but medium-to-high in both outcomes, and mostly significant according to the Q test, for CBT, prolonged exposure, and EMDR. Distinguishing between self-reports and clinician ratings did not reveal deviating results for any specific type of treatment (details omitted for brevity), but for the head-to-head comparisons of TF vs. NTF treatments: At post-treatment, TF treatments were more efficacious than NTF treatments in self-reports, k = 5, g = 0.21 [0.05, 0.37], p = .008, Q = 3.00, p = .557, I 2 = 13 %. Meta-analyses with regards to primary and secondary outcomes at follow-up (first and second follow-ups) are presented in Appendix [fig_ref] Table 8: Meta-analyses of the efficacy of different types of treatment at follow-up regarding... [/fig_ref]. TF treatments were again more efficacious than NTF treatments at both follow-ups. Pointing into the same direction, TFCBT + EX treatments appeared to be more efficacious with regards to primary outcomes at both follow-ups, but these results closely missed significance (ps = .053 and .076). PCTs were again less efficacious with regards to primary outcomes at first follow-up. Heterogeneity was again low for exposure therapies, TF vs. NTF, and PCTs, but medium-to-high for CBT, TFCBT + EX, prolonged exposure, and EMDR.
[formula] + [/formula]
Heterogeneity was also high for CBT at second follow-up. TF treatments were more efficacious than NTF treatments at first follow-up both in self-reports (k = 5, g = 0.15 [0.01, 0.29], p = .037, Q = 1.84, p = .766, I 2 = 0 %) and in clinician ratings (k = 6, g = 0.18 [0.03, 0.33], p = .016, Q = 2.58, p = .765, I 2 = 0 %).
## Primary and secondary outcomes: the wampold homogeneity test (wht)
Primary and secondary outcomes at post-treatment were also investigated with the WHT. In 30 independent runs of random sign assignment with regards to primary outcomes, the Q test was only twice not significant (p > .05) and yielded in 28 replications a significant value (p < .05); the null hypothesis of no differences in treatment efficacy had to be rejected with p < .001 (binomial test). On average, I 2 was 40 % across all replications, with a range of 28 % to 43 %. With regards to secondary outcomes, Q tests were significant in all of another 30 independent runs (p < .001); on average, I 2 was 43 %, ranging from 35 % to 46 %. Restricting the study sample to studies that were also included by Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] , we obtained 20 significant Q tests out of another 30 independent runs for primary outcomes (p < .001); on average, I 2 was 38 %, ranging from 27 % to 43 %. In contrast, Benish et Note. k = number of comparisons; Q = statistic of effect size heterogeneity. Values of Hedges' g and I 2 are presented alongside their 95 % confidence intervals. Hedges' g > 0 signifies a higher efficacy for the type of treatment of interest compared to all other available treatments, Hedges' g < 0 signifies a lower efficacy. Per type of treatment, values in the first line pertain to primary outcomes (i.e., PTSD symptom severity), values in the second line to secondary outcomes (i.e., symptom severity of comorbid disorders, trauma-related symptoms, general symptom distress, social functioning, and quality of life)
al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] reported a not significant I 2 of 9.39 % for this sample of studies.
Clinical significance and dropout displays meta-analyses with regards to clinical significance and dropout at post-treatment. Additionally, the thresholds for clinically meaningful results [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] of RR ≤ 0.80 or RR ≥ 1.25 were considered for interpretation (see the Synthesis of results subsection in the Methods section): Where 95 % confidence intervals of risk ratios included these thresholds, results were interpreted as limited evidence of differential efficacy. Clinical significance could not be examined for exposure therapies and PCTs because of less than five comparisons available each. Treatments did not differ with regards to clinical significance or dropout. Trauma focused treatments appeared to have a higher efficacy than non-trauma focused treatments, but this result closely missed significance (p = .095). Heterogeneity was low for most treatments in clinical significance, but medium-to-high for prolonged exposure and EMDR. In contrast, heterogeneity was medium-to-high in dropout for most treatments, with the exception of EMDR, where heterogeneity was low.
With regards to the required five or more available comparisons, 13 studies reported rates of recovery or improvement at a first follow-up, while only four reported respective rates for two follow-ups [bib_ref] Randomized trial of cognitive-behavioral therapy for chronic posttraumatic stress disorder in adult..., Mcdonagh [/bib_ref] [bib_ref] Randomized controlled comparison of cognitive behavior therapy with rogerian supportive therapy in..., Cottraux [/bib_ref] [bib_ref] A comparison of cognitive-processing therapy with prolonged exposure and a waiting condition..., Resick [/bib_ref] [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref]. One overall follow-up was defined and investigated, using data from the second follow-up in studies that reported two follow-ups. PCTs could not be investigated at posttreatment; at follow-up PCTs appeared to be less efficacious, k = 5, RR = 0.79 [0.61, 1.02], p = .076, Q = 6.84, p = .144, I 2 = 35 %, but this result closely missed significance. Trauma focused therapies appeared to be more efficacious at follow-up than non-trauma focued treatments, k = 6, RR = 1.21 [0.98, 1.49], p = .076, Q = 7.00, p = .221, I 2 = 15 %, but this result also closely missed significance. The only other treatment type, for which five or more comparisons were available, was CBT. CBT did neither differ from other treatments concerning rates of recovery or improvement at follow-up, k = 9, RR = 1.12 [0.96, 1.30], p = .153, Q = 10.03, p = .263, I 2 = 0 %; nor was the subgroup TFCBT + EX significantly more efficacious than comparison treatments, k = 7, RR = 1.14 [0.96, 1.34], p = .126, Q = 9.73, p = .136, I 2 = 0 %.
## Publication bias
Mostly, there was no indication that publication bias influenced results substantially (details omitted for brevity). However, with regards to primary outcomes, trim-and-fill added one missing study in comparisons of PCTs. This resulted in an overall treatment effect of g = −0.20 Meta-analyses of the efficacy of different types of treatment at post-treatment regarding clinical significance and dropout
## Dependencies in the data
Excluding studies in cases where more than one comparison was reported [bib_ref] Brief psychotherapy for posttraumatic stress disorders, Brom [/bib_ref] or where treatments of the same type were compared [bib_ref] Promising treatments of women with comorbid PTSD and substance abuse, Hien [/bib_ref] [bib_ref] Imagery Rehearsal for posttraumatic nightmares: a randomized controlled trial, Cook [/bib_ref] [bib_ref] A randomized controlled trial of exposure therapy and cognitive restructuring for posttraumatic..., Bryant [/bib_ref] did not change results substantially for any of the above reported analyses (details omitted for brevity). Thus, categorizing the hypnotherapy investigated in Brom et al. [bib_ref] Brief psychotherapy for posttraumatic stress disorders, Brom [/bib_ref] as EX, which may appear debatable, did not influence overall results.
## Sensitivity analyses
A number of studies were identified as influential cases in sensitivity analyses. Some influential studies were of higher quality and examined larger samples than the remaining studies [bib_ref] A controlled comparison of eye movement desensiti-zation and reprocessing versus exposure plus..., Power [/bib_ref] [bib_ref] Randomized controlled comparison of cognitive behavior therapy with rogerian supportive therapy in..., Cottraux [/bib_ref] [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref] , see [fig_ref] Table 1: Characteristics of included studies [/fig_ref]. Hence, their exclusion may not be considered warranted in every comparison where they appeared to be influential. Others were of a comparably lower quality [bib_ref] Comparison of two treatments for traumatic stress: A community-based study of EMDR..., Ironson [/bib_ref] [bib_ref] A comparison of cognitive-processing therapy with prolonged exposure and a waiting condition..., Resick [/bib_ref] , indicating that exclusion of these studies might be considered warranted in comparisons where they appeared influential. With regards to primary outcomes at post-treatment, studies of Resick et al. [bib_ref] A comparison of cognitive-processing therapy with prolonged exposure and a waiting condition..., Resick [/bib_ref] and Schnurr et al. [bib_ref] Randomized trial of trauma-focused group therapy for posttraumatic stress disorder, Schnurr [/bib_ref] [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref] appeared influential in the PE, EXPE and PCT categories. Excluding either the Resick et al. [bib_ref] A comparison of cognitive-processing therapy with prolonged exposure and a waiting condition..., Resick [/bib_ref] or the Schnurr et al. [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref] study rendered heterogeneity in comparisons of prolonged exposure not significant, I 2 = 27 % and 29 %, respectively. Moreover, excluding the Resick et al. [bib_ref] A comparison of cognitive-processing therapy with prolonged exposure and a waiting condition..., Resick [/bib_ref] study from comparisons of the combined EXPE category indicated a higher efficacy of treatments of this type, g = 0.19 [0.06, 0.33], p = .005, and diminished also heterogeneity to I 2 = 2 %. Notably, the Resick et al. [bib_ref] A comparison of cognitive-processing therapy with prolonged exposure and a waiting condition..., Resick [/bib_ref] study was of comparably low quality; hence, exclusion may be considered warranted, suggesting thus a true advantage of EXPE treatments over other treatments. Excluding either the Schnurr et al. [bib_ref] Randomized trial of trauma-focused group therapy for posttraumatic stress disorder, Schnurr [/bib_ref] or the Schnurr et al. [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref] study from comparisons of PCT rendered the overall treatment effect not significant, g = 0.16 [−0.10, 0.42], p = .220, and g = 0.09 [−0.08, 0.26], p = .290, respectively.
With regards to secondary outcomes at post-treatment, studies of Foa et al. [bib_ref] A comparison of exposure therapy, stress inoculation training, and their combination for..., Foa [/bib_ref] , Power et al. [bib_ref] A controlled comparison of eye movement desensiti-zation and reprocessing versus exposure plus..., Power [/bib_ref] , and, again, Resick et al. [bib_ref] A comparison of cognitive-processing therapy with prolonged exposure and a waiting condition..., Resick [/bib_ref] were influential. Excluding either of these studies rendered heterogeneity in comparisons of CBT not significant, I 2 = 46 %, 39 %, and 0 %, respectively.
With regards to clinical significance at post-treatment, studies of Ironson et al. [bib_ref] Comparison of two treatments for traumatic stress: A community-based study of EMDR..., Ironson [/bib_ref] , and, again, Schnurr et al. [bib_ref] Randomized trial of trauma-focused group therapy for posttraumatic stress disorder, Schnurr [/bib_ref] and Schnurr et al. [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref] were influential. Excluding the Schnurr et al. [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref] study rendered heterogeneity in prolonged exposure and in the combined EXPE category not significant, I 2 = 19 % and 14 %, respectively. Excluding the Ironson et al. [bib_ref] Comparison of two treatments for traumatic stress: A community-based study of EMDR..., Ironson [/bib_ref] study rendered heterogeneity there also not significant, I 2 = 39 % and 37 %, respectively, and resulted in limited evidence for a higher efficacy of these treatments, RR = 1.26 [1.02, 1.57], p = .035 (PE), and RR = 1.22 [1.01, 1.48], p = .043 (EXPE), respectively. As the Ironson et al. [bib_ref] Comparison of two treatments for traumatic stress: A community-based study of EMDR..., Ironson [/bib_ref] study was of low quality, exclusion may be considered warranted, suggesting thus a true advantage of prolonged exposure over other treatments (the advantage of EXPE appeared to be driven by prolonged exposure). Furthermore, exclusion of the Schnurr et al. [bib_ref] Randomized trial of trauma-focused group therapy for posttraumatic stress disorder, Schnurr [/bib_ref] study resulted in limited evidence for a higher efficacy of trauma focused treatments as well, RR = 1.46 [1.03, 2.08], p = .034. Finally, exclusion of the Cottraux et al. [bib_ref] Randomized controlled comparison of cognitive behavior therapy with rogerian supportive therapy in..., Cottraux [/bib_ref] study resulted in limited evidence for a higher risk of dropout for CBT, RR = 1.38 , p = .005, and for TFCBT + EX, RR = 1.46 [1.14, 1.88], p = .003 and a lower risk of dropout for PCT, RR = 0.57 [0.43, 0.75], p < .001.
## Moderator analyses
Moderator analyses revealed effects of study quality (Jadad rating) on primary outcomes at follow-up and of secondary outcomes at post-treatment in comparisons of CBT with other treatments. Studies with lower quality reported larger treatment differences, see [fig_ref] Table 4: Moderator analyses [/fig_ref]. This Note. k = number of included studies. Values of Hedges' g are presented alongside their 95 % confidence intervals. Hedges' g > 0 signifies a higher efficacy for CBT compared to all other available treatments, Hedges' g < 0 signifies a lower efficacy. Heterogeneity explained by quality rating in primary outcomes: QM (5) = 14.42, p = .013; residual heterogeneity: QE(9) = 10.88, p = .284; in secondary outcomes: QM(5) = 16.14, p < .001; QE(10) = 10.13, p = .429. a Quality rating according to Jadad et al. [bib_ref] Assessing the quality of reports of randomized clinical trials: Is blinding necessary?, Jadad [/bib_ref] pattern was also similar for secondary outcomes at post-treatment in comparisons of TFCBT + EX with other treatments
# Discussion
## Summary of evidence
In head-to-head comparisons between trauma focused treatments and non-trauma focused treatments, this meta-analysis revealed a significant difference in efficacy. Trauma focused treatments were slightly more efficacious with regards to PTSD symptom severity at post-treatment (g = 0.14), at the first (g = 0.17) and the second follow-up (g = 0.23). This result corresponds with results of other meta-analyses [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] [bib_ref] Meta-analysis of psychological treatments for posttraumatic stress disorder in adult survivors of..., Ehring [/bib_ref] , but this time based strictly on head-to-head comparisons of bona fide therapies only. The results of this meta-analysis with regards to the specific types of therapy corroborate at a first glance that bona fide psychotherapies of PTSD appear mostly similar with regards to their efficacy [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref]. There was no strong indication of cognitive behavioural therapy, exposure therapies, prolonged exposure, EMDR, or of trauma focused cognitive behavioural therapy to be more efficacious than any of the other available therapies. However, sensitivity analyses revealed that some of these null findings could be traced to low quality studies. Excluding these studies, we obtained evidence of an advantage of prolonged exposure and exposure therapies (g = 0.19) over other treatments considering PTSD symptom severity at post-treatment, and limited evidence of a higher efficacy of prolonged exposure (RR = 1.26) regarding recovery rates at post-treatment, cf. [bib_ref] A meta-analytic review of prolonged exposure for posttraumatic stress disorder, Powers [/bib_ref]. PCTs appeared somewhat less efficacious than other treatments with regards to PTSD symptom severity at post-treatment (g = −0.20; adjusting for publication bias) and at follow-up (g = −0.17), mirroring recent findings with regards to a somewhat diminished efficacy of bona fide supportive therapies also in the treatment of depression [bib_ref] Comparing bona fide psychotherapies of depression in adults with two meta-analytical approaches, Braun [/bib_ref]. However, PCTs were equally effective with regards to secondary outcomes as their available comparison treatments. No significant and clinically meaningful differences with regards to dropout were observed. All differential treatment effects were at most only small in sizeand far below proposed thresholds for clinically meaningful differences [bib_ref] Psychological treatments for chronic post-traumatic stress disorder, Bisson [/bib_ref] [bib_ref] An effect size primer: A guide for clinicians and researchers, Ferguson [/bib_ref].
Application of the omnibus test of overall effect size heterogeneity (see Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] indicated that the null hypothesis of no treatment differences had to be rejected in the full study sample; effect size heterogeneity was larger than expected by chance. Similarly, this null hypothesis had to be rejected in the study sample of Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref]. Our analysis was based on an improved application of this analytic method, using a Monte Carlo permutation approach and basing our conclusions on 30 independent runs of the Wampold homogeneity test, and differed also by utilizing a stricter definition regarding primary outcomes. Despite these differences, the obtained results cast doubt on the results of the Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] meta-analysis, but also on other meta-analyses that used this method [bib_ref] A meta-analysis of outcome studies comparing bona fide psychotherapies: Empirically, Wampold [/bib_ref] [bib_ref] Distinctions without a difference: Direct comparisons of psychotherapies for alcohol use disorders, Imel [/bib_ref] [bib_ref] Direct comparisons of treatment modalities for youth disorders: A meta-analysis, Miller [/bib_ref] [bib_ref] Is the allegiance effect an epiphenomenon of true efficacy differences between treatments?..., Munder [/bib_ref]. We thus caution on the inconsiderate application of the omnibus test of overall effect size heterogeneity and recommend a Monte Carlo permutation approach, as used in the present study, also in future applications.
Substantial effect size heterogeneity was also apparent in conventional meta-analyses of specific treatment types and sensitivity analyses revealed that it could be traced to a number of influential studies. The Schnurr et al. [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref] study was identified as influential, being subject to controversy before [bib_ref] Determining what works in the treatment of PTSD, Wampold [/bib_ref] [bib_ref] Do all psychological treatments really work the same in posttraumatic stress disorder?, Ehlers [/bib_ref] , regarding the bona fide status of the PCT investigated in one arm of this study. Yet, we also identified a number of other, apparently less controversial, studies. Some of these [bib_ref] A controlled comparison of eye movement desensiti-zation and reprocessing versus exposure plus..., Power [/bib_ref] [bib_ref] Randomized controlled comparison of cognitive behavior therapy with rogerian supportive therapy in..., Cottraux [/bib_ref] [bib_ref] Randomized trial of trauma-focused group therapy for posttraumatic stress disorder, Schnurr [/bib_ref] had a higher study quality and often larger patient samples than remaining studies, indicating that their exclusion may not be considered warranted. Studies of Ironson et al. [bib_ref] Comparison of two treatments for traumatic stress: A community-based study of EMDR..., Ironson [/bib_ref] and Resick et al. [bib_ref] A comparison of cognitive-processing therapy with prolonged exposure and a waiting condition..., Resick [/bib_ref] were of comparably lower quality, suggesting that their exclusion may be considered warranted.
Excluding influential studies, exposure therapies and prolonged exposure appeared more efficacious than other treatments with regards to PTSD symptom severity and recovery rates, whereas PCTs appeared equally efficacious as other treatments with regards to PTSD symptom severity. As the latter finding depended on the exclusion of large and high-quality studies [bib_ref] Randomized trial of trauma-focused group therapy for posttraumatic stress disorder, Schnurr [/bib_ref] [bib_ref] Cognitive behavioral therapy for posttraumatic stress disorder in women. A randomized controlled..., Schnurr [/bib_ref] , we deem them likely biased. Moreover, systematic effects of study quality were also observed in comparisons involving cognitive behavioural therapy; differences at post-treatment and follow-up were larger in studies with lower quality. Lastly, we obtained evidence of publication bias, that, when adjusted for, revealed as its most significant result that PCTs were robustly less efficacious than other treatments with regards to PTSD symptom severity.
Study quality appeared thus influential with regards to aggregated tests of differential efficacy of bona fide psychotherapies of PTSD. However, based on the available evidence, obtained results appeared fickle and need to be interpreted with caution. Even though another seven studies over-and-above the 15 studies investigated by Benish et al. [bib_ref] The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder:..., Benish [/bib_ref] were identified and included in the present meta-analysis, as much as nine (41 %) of all included studies examined less than 20 patients within one or both treatment arms, and another five (23 %) less than 30 patients; study quality appeared, overall, only mediocre. In order to be able to draw firmer conclusions, more high-quality studies using ITT analyses are still needed in this field of research.
As the current debate on the relative efficacy of PTSD treatments focused a lot on the topic of whether individual studies should be included or excluded in meta-analysis [bib_ref] Determining what works in the treatment of PTSD, Wampold [/bib_ref] [bib_ref] Do all psychological treatments really work the same in posttraumatic stress disorder?, Ehlers [/bib_ref] , the sensitivity analyses reported here show that this factor could actually be decisive in terms of the overall meta-analytical result regarding equal or different efficacy of compared treatments. Concerning the consequences with regards to the dodo bird verdict and the question of what works in psychotherapy, we thus recommend a refinement and clarification of the bona fide definition which should ideally result in a clear decision rule with regards to the inclusion and exclusion of especially supportive and presentcentered therapies in analysis. Otherwise, the selection of study samples may overly depend on the theoretical position of the study authors, or meta-analyses from authors with different theoretical positions might not be adequately comparable.
# Limitations
Using additional terms in the literature search, like the names of specific psychotherapies (e.g., "CBT", "exposure therapy" or "EMDR"), might have provided additional studies that were overlooked using the present search terms.
Treatment categorizations adhered to the descriptions provided in the studies themselves, which may have introduced spurious heterogeneity with regards to categorization and may have impacted on our results. Moreover, there were too few studies for most of the specified treatment types to systematically test effect moderation with moderator analysis. Thus, effect size heterogeneity that could not be traced to influential studies could mostly not be examined with regards to proposed moderator variables. This affected specifically meta-analyses on EMDR, where heterogeneity was substantial and could not be traced to single influential studies.
The assessment of eligibiliy criteria, in particular the bona fide status of treatments, is still a major point of discussion. Hence, independent codings of a second rater, in addition to comparing codings of the present study with codings provided in other meta-analyses, might have benefitted the reliability of our decisions to include or exclude studies in the present meta-analysis.
Studies were very variable with regards to time of follow-up. Bradley et al. [bib_ref] A multidimensional metaanalysis of psychotherapy for PTSD, Bradley [/bib_ref] suggested limiting comparisons at follow-up to assessments that took place six or more months following end of treatment. In the present study, all available follow-ups were used for analysis, regardless of exact time of follow-up, which may have influenced results. Specifically with regards to longer follow-up periods results could be confounded, because patients might have sought additional, external, treatment. This needs to be controlled both in primary studies and metaanalytical investigations. However, not enough information was available in primary studies to investigate this topic in the present meta-analysis.
Distinguishing different types of recovery and improvement might accentuate differences in efficacy, particularly because improvement is defined by study authors in different ways [bib_ref] A multidimensional metaanalysis of psychotherapy for PTSD, Bradley [/bib_ref] [bib_ref] Empirically supported complexity. Rethinking evidence-based practice in psychotherapy, Westen [/bib_ref]. As only nine studies reported recovery rates at post-treatment, combining recovery and improvement rates allowed the assessment of clinical significance, but may have influenced results.
Effects of researcher allegiance were neither examined nor controlled for in the present study. Researcher allegiance may affect differences in treatment efficacy and was reported to explain 12 % of variation in outcome regarding various trauma focused treatments [bib_ref] Is the allegiance effect an epiphenomenon of true efficacy differences between treatments?..., Munder [/bib_ref]. However, allegiance failed to explain differences of efficacy in a number of other meta-analyses [bib_ref] Is cognitive-behavioral therapy more effective than other therapies? A meta-analytic review, Tolin [/bib_ref] [bib_ref] Comparing bona fide psychotherapies of depression in adults with two meta-analytical approaches, Braun [/bib_ref] [bib_ref] Researcher allegiance and metaanalysis: The case of cognitive-therapy for depression, Gaffan [/bib_ref] , and it is not clear whether allegiance ratings reliably reflect true differences in allegiance [bib_ref] Investigator allegiance and the evaluation of psychotherapy outcome research, Berman [/bib_ref] or how well meta-analytical reviews may adjust for it.
Considering public health issues, a variety of topics not considered here may still have to be evaluated with regards to the comparative efficacy of treatments [bib_ref] Helping vulnerable populations: A comprehensive review of the treatment outcome literature on..., Najavits [/bib_ref]. Cost of treatments, the level of education and the intensity of training of therapists, the preferences of patients and the applicability of treatments for a wide range of patients may influence what works best for patients as well as what is affordable to implement in public health care. For example, present-centered therapies could be easier to deliver at a lower cost to more varied patients with a less skilled workforce, and these advantages might compensate for a lower efficacy with regards to health outcome. More research is needed here.
More research on traumatized populations with severe comorbidities, like psychosis (mostly an exclusion criteria in PTSD studies), as well as on patients with additional and multiple life burdens and comorbidities, such as homelessness, unemployment, suicidality or with a criminal justice involvement, is also needed [bib_ref] Helping vulnerable populations: A comprehensive review of the treatment outcome literature on..., Najavits [/bib_ref]. Information on the efficacy of PTSD treatments in such highly vulnerable populations is currently lacking.
# Conclusions
The bona fide distinction is an interesting and promising research heuristic that is increasingly and independently used in a number of meta-analyses [bib_ref] Is cognitive-behavioral therapy more effective than other therapies? A meta-analytic review, Tolin [/bib_ref] [bib_ref] Comparing bona fide psychotherapies of depression in adults with two meta-analytical approaches, Braun [/bib_ref] [bib_ref] Psychotherapy versus secondgeneration antidepressants in the treatment of depression: A meta-analysis, Spielmans [/bib_ref] , even though its definition and criteria may need to be extended or specified more clearly [bib_ref] The dodo bird, treatment technique, and disseminating empirically supported treatments, Siev [/bib_ref] [bib_ref] Do all psychological treatments really work the same in posttraumatic stress disorder?, Ehlers [/bib_ref]. The stated goal of the present study was to apply this distinction, but not to improve it. On these premises, evidence for at most small differences in efficacy between bona fide psychotherapies for PTSD was obtained. Differences were far below proposed thresholds of clinically meaningful differences, but corroborate that trauma focused treatments and prolonged exposure and exposure therapies are slightly more efficacious than other therapies. While the lower efficacy of present-centered therapies mirrors recent findings in the field of treatment of depression, more high-quality research is still needed to draw firm conclusions. Factors that may affect and that may broaden the scope of treatment efficacy, such as public health issues and the requirements of highly vulnerable populations, are understudied in available research and need to be considered in future research. Endnotes 1 (a) Was the study described as randomized? (b) Was the study described as double blind? (c) Was there a description of withdrawals and dropouts? For further information on the appropriate fulfillment of each criteria and scoring, see [bib_ref] Assessing the quality of reports of randomized clinical trials: Is blinding necessary?, Jadad [/bib_ref]. [bib_ref] The Dodo Bird Verdict -controversial, inevitable and important: A commentary on 30..., Budd [/bib_ref] The study of Blanchard et al. [bib_ref] Some implicit common factors in diverse methods of psychotherapy, Rosenzweig [/bib_ref] is of special interest concerning the validity of the bona fide criteria for the identification of active treatments. The authors describe both treatments as active treatments giving reasonable descriptions. Both treatments proved effective, but one of them did not mention the necessary two of the four operational bona fide criteria (citation of approach, reference to psychological process, treatment manual, active ingredients with citation). This study is therefore a good example as to why it may be necessary to develop more valid operational criteria (see Discussion). [bib_ref] Meta-analysis of psychotherapy outcome studies, Smith [/bib_ref] If a study presented comparisons between different active treatments and between treatment components, only the comparisons between the different active treatments were included (see [fig_ref] Table 1: Characteristics of included studies [/fig_ref] ; comparisons between treatment components were excluded. The mentioned excluded studies contained only comparisons between treatment components. Note. k = number of comparisons; Q = statistic of effect size heterogeneity. Values of Hedges' g and I 2 are presented alongside their 95 % confidence intervals. Hedges' g > 0 signifies a higher efficacy for the type of treatment of interest compared to all other available treatments, Hedges' g < 0 signifies a lower efficacy. Per type of treatment, values in the first line pertain to primary outcomes (i.e., PTSD symptom severity), values in the second line to secondary outcomes (i.e., symptom severity of comorbid disorders, trauma-related symptoms, general symptom distress, social functioning, and quality of life)
[fig] Figure 1: Flow of studies through the selection process [/fig]
[table] Table 1: Characteristics of included studies [/table]
[table] Table 2: Meta-analyses of the efficacy of different types of treatment at post-treatment regarding primary and secondary outcomes [/table]
[table] Table 4: Moderator analyses: efficacy of CBT according to study quality [/table]
[table] Table 5: Primary outcome measures [/table]
[table] Table 6: Secondary outcome measures # Studies Used in studies Measures of depression, anxiety, substance abuse, and of trauma-related guilt, anger, dissociation, and cognitions Measures of quality of life, general symptom distress, and social functioning Note. Listed are references to the original instruments, which may differ from references provided in the studies themselves [/table]
[table] Table 7: Treatment categorizations of included studies in the present meta-analysis compared to previous meta-analysesBrom et al.[57], p. 607: "The emphasis of the hypnotherapists in our study was on behavioral therapy. The goal was to bring the patient in contact with the reality of the traumatic event and to bring about a decrease in the conditioned responses triggered by the event." breathing retraining, role playing, covert modeling, guided self-dialogue, thought stopping (seeRothbaum et al., 2000). No instructions for exposure were included, see[42].Foa et al.[61] Stress inoculation training (SIT) [/table]
[table] Table 8: Meta-analyses of the efficacy of different types of treatment at follow-up regarding primary and secondary outcomes [/table]
|
Acceptability and Feasibility of an Isometric Resistance Exercise Program for Abdominal Cancer Surgery: An Embedded Qualitative Study
Although it is recognized in the early stages of cancer recovery that changes in lifestyle including increases in physical activity improves physical function, there are no clear findings whether low versus moderate intensity activity or home or gym exercise offer optimal benefit. Isometric-resistance exercises can be carried out with very little equipment and space and can be performed while patients are bed-bound in hospital or at home. This embedded qualitative study, based in an English hospital trust providing specialist cancer care, was undertaken as a component of a feasibility trial to evaluate the acceptability and feasibility of an isometric-resistance exercise program and explore the suitability of functional assessments by drawing from the experiences of abdominal cancer patients following surgery. Telephone interviews were undertaken with 7 participants in the intervention group, and 8 interviews with the usual care group (n ¼ 15). The gender composition consisted of 11 females and 4 males. Participants' ages ranged from 27 to 84 (M ¼ 60.07, SD ¼ 15.40). Interviews were conducted between August 2017 and May 2018, with audio files digitally recorded and data coded using thematic framework analysis. Our results show that blinding to intervention or usual care was a challenge, participants felt the intervention was safe and suitable aided by the assistance of a research nurse, yet, found the self-completion questionnaire tools hard to complete. Our study provides an insight of trial processes, participants' adherence and completion of exercise interventions, and informs the design and conduct of larger RCTs based on the experiences of abdominal cancer surgery patients.
# Introduction
Surgery is one of the main types of treatment for abdominal cancer, with a high risk of post-operative complications and a notable decrease in physical function. Colorectal is the third most common cancer with 1.8 million cases worldwide in 2018 (10.6% of all cancers). Following colorectal cancer, the most common types of abdominal cancers also include stomach (6.1%), liver (5.0%), followed by cervical (3.3%), pancreatic (2.7%), kidney (2.4%) and ovarian (1.7%).Although it is recognized in the early stages of cancer recovery that changes in lifestyle including increasing physical activity can help to improve overall well-being, there are is very little clinical evidence in terms of how different modes of physical function and modes of exercise can be incorporated following active cancer treatment. [bib_ref] Effects of preoperative and postoperative resistance exercise interventions on recovery of physical..., Stephensen [/bib_ref] [bib_ref] Exercise interventions on health-related quality of life for cancer survivors, Mishra [/bib_ref] Data from the United Kingdom suggests that a third of people living with and beyond cancer are completely inactive, [bib_ref] Implementing the exercise guidelines for cancer survivors, Wolin [/bib_ref] [bib_ref] Exercise is medicine in oncology: engaging clinicians to help patients move through..., Schmitz [/bib_ref] and 20% reported moderate or severe difficulties with mobility or usual activities. [bib_ref] Patient-reported outcomes of cancer survivors in England 1-5 years after diagnosis: a..., Glaser [/bib_ref] [bib_ref] Physical activity in cancer prevention and survival: a systematic review, Mctiernan [/bib_ref] The World Cancer Research Fund also found strong evidence that being physically active decreased the risk of one of the main types of abdominal cancer-that of colorectal cancer (WCRF Colon Cancer Report 2018). Evidence on physical activity for people recovering from cancer has been shown to improve physical function without increases in fatigue associated with exercise. However, there is little evidence to identify the optimum mode, frequency, intensity and duration of activity required for beneficial effects in cancer populations. [bib_ref] Exercise interventions for cancer patients: systematic review of controlled trials, Stevinson [/bib_ref] In addition, there are no clear findings to compare whether low versus moderate intensity activity, or home-based versus gym exercise offer the optimal benefit. The rationale to determine whether home or gym exercises offer optimal benefit is based on previous trials evidence, which showed similar outcomes between patients regarding exercise maintenance and adherence. [bib_ref] Gym-based exercise and home-based exercise with telephone support have similar outcomes when..., Jansons [/bib_ref] Isometricexercises can be performed in a confined space such as in the home without requiring access to exercise equipment. Homebased isometric-resistance exercises therefore have the potential to have a positive effect on patients undergoing elective abdominal cancer surgery, as previous studies have shown that it can preserve and optimize their physical condition. [bib_ref] Preoperative therapeutic programme for elderly patients scheduled for elective abdominal oncological surgery:..., Dronkers [/bib_ref] While there is recognition of the benefits of aerobic physical activity, [bib_ref] Exercise in the rehabilitation of breast cancer survivors, Pinto [/bib_ref] [bib_ref] Exercise in cancer recovery: an overview of the evidence, Courneya [/bib_ref] [bib_ref] Effects of exercise on cancer-related fatigue, Dimeo [/bib_ref] [bib_ref] Physical exercise and immune system function in cancer survivors: a comprehensive review..., Fairey [/bib_ref] randomized or quasi-randomized studies with a clear resistence muscle strengthening component has seen low numbers of participants to allow for any statistically significant evidence to be used for or against this type of exercise program. The low numbers of participants has been due to recruitment taking place in single study centers with exclusion of participants with colon surgery. The methodological quality of these randomized studies are moderate, with unclear bias, difficulty in blinding trial participants and therapists, and in respect to description of the intervention, some information lacking in terms of equipment and methodology with regard to aerobic and functional activity components. [bib_ref] Effects of preoperative and postoperative resistance exercise interventions on recovery of physical..., Stephensen [/bib_ref] Resistance exercise (or strength) training could help to facilitate recovery of muscle function. [bib_ref] Effect of isometric strength training on mechanical, electrical, and metabolic aspects of..., Komi [/bib_ref] In particular isometric resistance (or static) training has been used in the rehabilitation of weak or atrophied muscle following surgery; [bib_ref] The effect of training intensity on voluntary isometric strength improvement, Szeto [/bib_ref] important factors to be considered in establishing an effective training regime include the training intensity, the number and durations of voluntary contractions, and the number and frequency of training sessions. [bib_ref] The effect of training intensity on voluntary isometric strength improvement, Szeto [/bib_ref] Isometricresistance exercises can be carried out with very little equipment and space, and can be performed while patients are bed-bound in hospital or at home. [bib_ref] Effects of preoperative and postoperative resistance exercise interventions on recovery of physical..., Stephensen [/bib_ref] The overarching aim of the Exercise Peri-Operative Programme (EPOP) trial was to develop an isometric resistance exercise program intended to improve the physical function of patients undergoing elective abdominal surgery for cancer, which aimed at expediting their return to normal physical function. The aims of the qualitative study were to qualitatively evaluate the acceptability and feasibility of an isometric resistance exercise program, and explore the suitability of assessments for physical function by drawing from the experiences of abdominal cancer surgery patients involved in the EPOP study.
Understanding the experiences of the patients is vital to consider which elements of the exercise program were relevant and useful, as well as which parts were less attractive. Exploring how different patients viewed and engaged with the intervention is helpful in ascertaining how it can be made more feasible to implement in the larger study. [bib_ref] The personbased approach to enhancing the acceptability and feasibility of interventions, Yardley [/bib_ref]
# Materials and methods
## Overall study design
The study comprised 2 main phases. Phase 1 involved a systematic review followed by a development study aimed at designing, developing and refining the intervention with healthcare professionals and patients (further details have been described in a forthcoming publication). Phase 2 involved the delivery of the intervention with the intervention group receiving a 12 week isometric-resistance program (see [fig_ref] Table 1: Isometric-Resistance Exercise Program Developed for the EPOP Study [/fig_ref]. The isometric resistance exercise program consisted of a series of 10 static exercises, and utilized the abdominal, back, neck, arm, hand, leg and foot muscles. It was planned to be completed within a 12 week period following surgery. The usual care group were encouraged to walk when they felt able but did not receive physiotherapy advice on specific exercises. The qualitative study was a planned component of the trial and was incorporated as part of Phase 2 and comprised postintervention follow-up exit interviews. The rationale of the qualitative study focused on assessing whether it was feasible to recruit patients with elective abdominal surgery, and second, exploring the adherence of patients to perform the exercises, finish the program and complete the functional assessments provided as questionnaires in the form of a booklet.
The Phase 2 trial data collected included self-reported and objective measures of assessments. Self-reported physical function assessment measures were collected using the Short Physical Performance Battery Test (SPPB) and the "timed-upand-go' test (TUG). The outcome measures were collected at baseline, 2, 6 and 12 weeks. It also included the collection of qualitative data to capture patient experiences of the study. For the purposes of this paper, we report upon the findings from the embedded qualitative study in Phase 2.
## Participant recruitment and consenting procedures
All patients were recruited from a large general hospital trust in England providing specialist cancer care who were undergoing laparotomies and laparoscopies for stomach, colorectal, and gynecological malignancies. Initially, potential participants were approached face-to-face by a member of their clinical team on behalf of the lead clinician by way of a signed letter, which was accompanied by a patient information sheet to assist the participant to make a decision whether to take part in the study. Participants provided written informed consent to take part in both the intervention and a follow-up telephone in"terview.
## Randomization
The trial was a 2-arm, parallel group trial with participants randomly allocated with a ratio of 1:1 prior to surgery. Randomization was performed using a secure, centralized and independent online randomization service (http://sealedenvelope. com/). The clinical trials administrator was issued with a password to access the service to randomize participants. Participants were randomized individually and in sequence within 24 hours of enrollment (i.e. screened as eligible and have provided informed consent). There was no participant allocation to specific surgical groups based on approach or primary diagnoses.
## Sample selection
Following surgery after the 12-week intervention period had ended and all outcome measures and follow-up appointments had taken place, participants from both arms of the study were invited to take part in a telephone interview. Nineteen of the 23 participants were followed up for interview (the other 4 could not be called due to changes in contact details). Out of the 19 who were contacted by telephone, 4 did not participate in an interview: one participant was away on holiday; another wanted to wait until her chemotherapy had ended, and 2 did not respond to voice messages left to arrange interviews. In order to ensure the views of all participants were captured, purposive sampling techniques were employed to attain a maximum variation across the intervention and usual care groups, and to reflect the variation of the overall trial sample (n ¼ 23). Participants were purposively selected to help construct a sample to achieve maximum variation, which included age and gender, date of surgery and the date of commencement and completion of the exercise program. Within a small sample, there were experiences where elements of the program were shared (Patton 2002, 2nd ed.), but there were also differences such as perceived baseline fitness, domestic environment and time commitments. From the exit interviews, it was not apparent whether the type of cancer influenced participants' responses on performing the exercises or completing them.
From the embedded qualitative study, 7 interviews were undertaken with participants in the intervention group, and 8 interviews in the usual care group. The gender composition consisted of 11 females and 4 males. Participants' ages ranged from 27 to 84 (M ¼ 60.07, SD ¼ 15.40). No males who were interviewed were allocated into the intervention arm, therefore no qualitative data relates to males undertaking the exercises. For the intervention itself, the sample consisted of 17 females and 6 males, which indicates that the participants invited for the embedded qualitative study was similar in terms of representative gender proportions [fig_ref] Table 2: Characteristics of the Patients of the Overall Trial Sample [/fig_ref].
## Trunk and legs
Lie on your back with your arms by your sides and your legs straight. Flex your feet toward you and press your knees down against the bed. Your ankles may raise off the ground. Hold for 10 seconds, then relax. Repeat twice.
## Hand and arm
Lie on your back with your arm by your side, elbow bent at a right angle close to your body. Hold your wrist with your other hand, across your body. Try to move your hand inwards while resisting the movement by pushing with your other hand. Hold for ___ seconds. Repeat twice with each arm. 5 Foot and Lower Leg Sit with your legs stretched out in front. Put a looped exercise band around your feet. Turn the soles of your feet toward each other. Then turn the soles of your feet away from each other. Keep your knees straight, and facing the ceiling. Repeat twice. 6 Abdominal muscles Lie on your back with your arms by your sides.
Push your shoulders and heels toward the floor, lifting your pelvis off the floor. Hold for 10 seconds. Repeat twice. 7 Arms and Shoulders Sit on a chair with your hands behind your neck. Place your elbows as close together in front of you as you can. Move your elbows apart as far as you can to the sides, then move them back together again as close as they will go. Repeat 5 times 8 Trunk and Legs
Sit on a chair, keeping your back straight. Lift your leg off the seat while keeping your knee bent, then return to the starting position. Repeat 5 times on each leg.
## Hand and arm
Stand or sit. Stretch one arm to the opposite shoulder, assisting the movement by pushing your elbow with your other hand. Hold the position for 10 seconds, then relax. Repeat twice on both sides.
## Data collection
One off semi-structured telephone interviews were undertaken with patients to explore the experience of taking part in the EPOP feasibility trial (n ¼ 15).Each interview lasted approximately 10 to 15 minutes. No other people were present during the interviews besides the participants and researcher.
No other repeated interviews were carried out at a later time point. Interviews took place by appointment within 3 days to around 4 weeks following the end of the 12 week intervention. The topic guides were developed by the research team and piloted with the project's Patient and Public Involvement (PPI) representatives (or lay members) who were abdominal cancer patient survivors. Two topic guides were developed-one for the intervention group and the other for the usual care group. The participants in the intervention group were asked about how they found the exercises, what they thought about the questionnaires, their thoughts about trial processes and their overall experiences of taking part in the study. The usual care group were asked the same question prompts but with omissions to questions relating directly to the exercise program and functional assessments. The interviewer used a topic guide to provide prompts for discussion, but participants were encouraged to take the conversation in directions they believed were important. All interviews were conducted between August 2017 and May 2018, with audio files being digitally recorded. Transcripts were not returned to participants for comment but were available on request if the participant wished to view it. No further field-notes were made during or after the interviews. Every interview was conducted by FH (female) who is an experienced qualitative researcher (at the time was Research Fellow based at the University) with a 15 year research portfolio working on patient experiences in health and social care and was part of the trial team. The research nurses informed the participants that a member of the trial team (FH) would contact them following the end of the study to explore their experiences and to discuss any problems they may have encountered. The participants were also told that the researcher was based at the University and wanted to hear their views before progressing to the larger trial. Despite limitations with conducting follow-up telephone interviews (discussed further below), participants were asked at the end of the conversation if they wanted to share any experiences that had not emerged during the discussion and participants took the opportunity to explore new themes that were not included on the interview guide. Data saturation was achieved when interview data had reached "information redundancy" as no new ideas or themes were being generated. [bib_ref] Saturation in qualitative research: exploring its conceptualization and operationalization, Saunders [/bib_ref]
# Data analysis
Interview data was coded by FH using a thematic framework approach, which took part in 4 stages. 1) Familiarization involving data immersing, gaining an oversight of the content and identifying topics of interest; 2) constructing an initial thematic analysis involving refining and coding the data into nodes and sub-nodes; 3) reviewing data abstracts which involved organizing the data to identify any coherent data groups; and 4) summarizing and writing up the data using summaries for each group of data appearing under each theme.All analysis was derived from the data. Although participants did not provide input directly onto the findings, the lay representatives had the opportunity to discuss them during trial meetings. This analysis was aided by the use of a qualitative software analysis program (NVIVO Pro 12) (see .Results
## Summary findings of the trial
We recruited 23 patients to our intervention study. Eleven patients were able to successfully complete the intervention program safely and easily. Twelve patients completed the control arm of the study. There were no adverse incidents related to the exercise program itself, which required withdrawal from the program. There were 6 incidents related to surgical complications. The intervention measured physical function using the Short Physical Performance Test Battery (SPPB) and the Timed-up-and-go (TUG) test (full results to be published on a forthcoming paper).
## Finding of qualitative study
The emergent themes that were identified from the qualitative interviews were: (i) safety and suitability of the exercise program; (ii) timing and setting for commencing the intervention; (iii) exercise guidance needs to aid adherence; (iv) completion of functional assessments; and lastly, (iv) issues around blinding and randomization [fig_ref] Figure 1: Themes regarding the experiences of patients on the EPOP isometric resistance exercise... [/fig_ref].
## Intervention acceptability
Safety and suitability of the exercise program. From the 7 participants who were allocated to the intervention group, their . EPOP Phase 2 exit interviews coding frame.
## Name description
Adherence (parent node) Patients' views performing the exercises and completing the programme Advice and support (parent node)
Training requirements, information received on exercises, advice sought (face-to-face or telephone), reassurance when performing exercises at home Exercise advice from nurses (face-to-face or telephone) (child node)
Patients' views on the advice received from the nurses, confidence in advice given, feeling able to raise questions and returning to the nurses for remote advice Exercise information guidance (child node) Patients' impressions of the induction to the programme, whether their questions were addressed during the session and views on the guidance booklet on the exercises Appropriateness of programme (parent node)
## How s exercises to help recovery
Exercise programme (child node) How patients felt about performing the exercises -easy, moderate, hard Functional assessments (questionnaires) (child node)
What patients thought about the questionnaires / booklets they were given to complete Optimal timing and setting (parent node) For commencement and continuation of the exercise programme Hospital and_or home (child node) Responses from patients of their preferred setting to commence the exercises Timing following surgery (child node)
Patients' views on when they felt able to commence the exercises Other factors identified encouraging participation (parent node)
External factors such as domestic environment, motivations for maintaining fitness, having a recovery plan, feeling in control Randomisation and blinding (parent node) Control & usual care groups' views on trial processes Recruitment (parent node) Who recruited them, what explanations were given to them to help decide whether or not to take part Retention (parent node)
Issues around retention and views from intervention or control groups about dropping out / attrition Safety and suitability of exercises (parent node)
Patients' views on how exercises felt, impact of surgery, adverse reactions from cancer treatment, any injuries I'm actually helping myself to recover, to build my strength, so and you've got a lot of time during the day. You've got that time to
## P09, female, intervention group
Another participant explained that the exercises were suitable and safe at the beginning, and did not have any concerns with them:
No, I just think generally everything . . . I think the exercises felt quite gentle up until the last . . . yeah . . .
## P04, female, intervention group
One participant reported that she could not do some of the exercises due to having an open wound:
. . . well the problem I had which I don't know whether [research nurse] actually explained to you, was that I had the wound that I had hadn't healed properly . . . yes I couldn't do the pelvic exercises . . . but the other, the top of body I could do. . . . oh yes, yes especially you know the upper body, I mean as I say with the pelvic area obviously I was a bit put back by that because obviously once I knew that I had a problem with the stitching and I had, well I don't know whether she's put it down it was discharged and I had to have the wound opened and packed and everything, so obviously I didn't want to put a strain on that so the exercises involved lying down and sitting up and all that I put on hold obviously because I didn't want to put anything under strain. But the rest of them I found were fine.
## P10, female, intervention group
This participant's statement indicates that having the exercises aimed at the lower body and pelvic area were not possible for her to complete given that she had an open wound that had not healed. She spoke about being able to put these specific exercises "on hold" so that she did not put a strain on that part of her body when performing the exercises, but at the same time, found it possible to be able to continue with the other exercises.
Timing and setting for commencing the intervention. The participants from the intervention group were asked if they found the exercises acceptable to carry them out after surgery. Some of the participants reported that they were able to complete the exercises and felt self-motivated to continue to do them after surgery, while others noted they did not want to commence them until after they arrived home. Participant P03 reported that she did not have any concerns about carrying out the exercises while still in hospital: Yeah, fine. It was good. It felt nice, controlled and . . . you know one of the things you feel when you wake up from that sort of operation you realize you can't go at things like a bull n a china shop anyway, it was nice to begin with something partly because it's so very boring being in hospital anyway.
## P03, female, intervention group
Another participant did not feel motivated to do the exercises until she came home: I think it's possibly a little bit unrealistic because you're groggy from the effects of the anesthetic and you probably feel a little bit self conscious as well in hospital doing them, possibly. So I think really, realistically for me I didn't start until I came back home, out again.
## P04, female, intervention group
Yet, when she arrived home, once she started the exercises, she felt that she was making good progress to keep motivated:
. . . and maybe that's because when you suddenly look at doing 30 and maybe it's a bit of a mind-set that instead of thinking about doing thirty all in one go it's that sort of mind-set of maybe the first sort of week P04, female, intervention group Another participant (P14, female, intervention group) mentioned that she found performing the exercises a little difficult, but she did find the time to do them on a regular basis:
. . . no no, I thought if it's gonna-I didn't expect it to be easy, that's just you want to get better don't you so you try and do everything you can really so, yeah . . . I tended to do them first thing in the morning and then some of them you had to do twice a day, I would find you know afternoon or something to do them . . .
## P14, female, intervention group
## Appropriateness of exercise guidance & functional assessments
Exercise guidance needs to aid adherence. The participants were asked to comment on the instructions on performing the exercises. One participant (P14, Female, Intervention Group) struggled to understand the instructions on how to do the exercises:
. . . well I just couldn't really understand it to be honest because you had to do, I had quite a few different ones to do and I couldn't because of my shoulder I couldn't do that and stuff so I tended to just do ones I could do and she narrowed it down so I was just sort of doing like 3 or 4 instead of well more P14, female, intervention group Completion of functional assessments. The participants who were interviewed in the intervention group had a variation of responses when commenting upon the functional assessments. Two respondents (P10 & P09) commented that the functional assessments were appropriate. One of the participants stated that:
. . . yeah no problem with those again, it's quite nice because you can sort of get a feeling yourself of how you're coming along . . . I could see the difference, then when I went back to the next one I could again tell the difference that I was stronger and I was able to do it better so it gave me a little bit of a measure myself.
## P09, female, intervention group
There were 4 participants (P14, P03, P04 & P13) who struggled with and disliked completing the functional assessments. One participant (P13, Female, Intervention Group) commented that:
. . . mmm I don't know, well I don't know, I don't think they did no . . . I tried to do my best I could and explained to [research nurse] cos I said you know I can't keep up with filling these forms and she said "oh no" you need to tell them that.
## P13, female, intervention group
Yet, another participant spoke about the functional assessments being a useful tool to measure progress by using them to self-reflect upon any improvements in her exercise performance. Once this participant had worked out how to record her responses, she thought the functional assessments were appropriate:
Yeah definitely, because when I looked at it I could see that, you know, it had gone from feeling hard . . . well not very hard but quite difficult to being actually this is quite easy. And that sort of fitted in with how my actual progress was going generally.
## P04, female, intervention group
## Acceptability and feasibility of blinding and randomization
Issues around blinding and randomization. Blinding was a challenge in the study, as the interview data suggests, the majority of participants were able to identify which group they were randomized into:
. . . well the ones that, well there were 2 groups weren't there? I was the one that was doing the extra exercises-yes . . .
## P10, female, intervention
Yeah, I had sort of said at the beginning I'd do it if I got allocated that and they said that you can't chose so I was just lucky it came up but it's what I wanted.
## P03, female, intervention
It would appear from the quotations above that maintaining blinding was difficult in an exercise-related intervention. Out of the 15 participants who were interviewed (female ¼ 11; male ¼ 4), 11 participants commented that they knew which group they were randomized into, and 4 said that they did not know. Even though patients were given an explanation about being allocated into the usual care or intervention group, one respondent stated that he was still unaware that he was allocated into any specific group:
Interviewer: So do you know which group you were allocated to in the study?
Respondent: I don't know that at all, even if I was allocated to a group.
## P05, male, usual care
The participants were asked if they knew that by agreeing to take part in the study, this would involve being randomized and they accepted being allocated into groups: I was hopeful that by being part of the [trial] group that I was going to be benefit from-or hopefully benefit from-the physio exercise program that . . . that would benefit me as well-Well obviously you have to have a control group . . . Well I mean I understood the concept of a control group in, you know, medical trials . . . so that was fine. P01, male, usual care . . . yeah it was the non-exercise . . . I didn't mind because it was a random computer generated thing wasn't it? P08, female, usual care
The first male participant above explained that although he was disappointed not to be randomized into the intervention group, he understood that as he was taking part in a trial, he accepted the process of being allocated into the usual care group which he saw a part of trial logistical requirements. The second female participant also understood that she was randomized into the usual care group through a computer generated program that she also accepted was part of the trial process.
# Discussion
The EPOP feasibility study on isometric resistance exercise interventions following abdominal cancer surgery has helped to inform the design and conduct of the larger RCT by exploring the acceptability of the intervention and adherence to the program, safety and suitability, appropriateness of functional assessments, and trial processes around randomization and blinding. The feasibility of blinding in physical activity interventions, as demonstrated in previous studies, was shown to be a challenge as most participants were able to identify which group they had been allocated to. In addition, it was found that the functional assessments enabled participants to self-reflect to measure their own progress, yet, the self-completion questionnaire tools were reported to be hard to complete.
Participants who were in the intervention group had the ability and willingness to understand and adhere to the exercise program, as indicated by the 7 participants who were interviewed from this group. They commented about how they felt the intervention was safe and suitable, aided by the assistance of a research nurse, who was available to provide advice and help with adapting the exercises in relation to their post-surgery mobility requirements. Our findings concur with a study by de Almeida et. al (2017) of an early mobilization program in abdominal cancer patients after major surgery. The program based on core stability, aerobic and resistance training, orthostatic and gait training, was found to be safe and feasible. [bib_ref] Early mobilization programme improves functional capacity after major abdominal cancer surgery: a..., De Almeida [/bib_ref] De Almeida notes the importance of a multi-professional approach including oncologists, surgeons, physiotherapists, nurses and psychiatrists, which contributed to high levels of adherence. The potential of a multi-professional approach in an exercise program is an important element to help with adaptations to the program and maintain adherence, and requires consideration for a future study.
One participant reported that due to an open wound she was unable to undertake any exercises in the pelvic area. Despite this, she continued with exercises to the upper body under clinical supervision. In a study by on an early mobilization program in colorectal surgery patients, it was found that patients were willing and capable of participating in a light to moderate intensity resistance-exercise program, by taking into consideration their specific post-operative status and stratified to reflect the patient's individual needs. [bib_ref] In-hospital resistance training to encourage early mobilization for enhanced recovery programs after..., Schram [/bib_ref] Although the patient in the EPOP study experienced mobility impediments due to an open wound, her response to continue with the exercises should also be noted, and that in discussion with the nurse she was able to carry on with clinical support. found in their study on the long-term experiences of patients following esophageal or gastric cancer surgery, the changes in physical status resulted in patients feeling that they had lost control over their lives. Being involved in a supportive cancer recovery program allowed the patient in our study to take back control over her life. [bib_ref] Long-term experiences after oesophagectomy/gastrectomy for cancer-a focus group study, Malmström [/bib_ref] The importance of post-surgery cancer supportive care programs, at a wider level, demonstrates a need for patients to have active involvement in their recovery plans to enable them to feel in control of their lives. [bib_ref] Patients' experiences of supportive care from a long-term perspective after oesophageal cancer..., Malmström [/bib_ref] Participants commented that the functional assessments were useful for them to measure progress and set personal goals. [bib_ref] Prehabilitation in cancer care: patients' ability to prepare for major abdominal surgery, Beck [/bib_ref] found in their study on a prehabilitation intervention for patients preparing for abdominal cancer surgery showed that patients found it important to write the "dose" of exercise they had performed, which provided them a "personal" competition to be able to complete the activities and tick a box on a leaflet. The psychological impact of recording functional change in an important message from our study. [bib_ref] Prehabilitation in cancer care: patients' ability to prepare for major abdominal surgery, Beck [/bib_ref] However, it is also noted from the qualitative data collected from the EPOP study that the participants reported some common problems with completing the functional assessments, and highlights the need to explore reformatting tools in a far more "easy read" and accessible lay-out. [bib_ref] Developing an easy read version of the adult social care outcomes toolkit..., Turnpenny [/bib_ref] have noted that there are ways of adapting outcome measurement tools to make them more accessible through "easy read" materials, which are characterized by plain language, simple layout and format, and using images to illustrate key messages in the text. [bib_ref] Developing an easy read version of the adult social care outcomes toolkit..., Turnpenny [/bib_ref] Although there are no common standards for producing "easy read" materials, there are national and international guidelines available to create accessible tools.The guidance will be consulted to re-design the functional assessments for the larger trial and any modified questionnaires will require re-validation.
The timing of the introduction of the exercise program was also a factor that may have impacted on execution and completion of the exercises. Although the evidence on the ideal time to promote physical activity is mixed, [bib_ref] Promoting lifestyle change among cancer survivors: when is the teachable moment?, Rabin [/bib_ref] [bib_ref] The feasibility of the prostate cancer: Evidence of Exercise and Nutrition Trial..., Shingler [/bib_ref] noted in their study on prostate cancer survivors, that following surgery men in the study "felt able to embark on physical activity or nutrition interventions 6 weeks aftermaking this an acceptable timing for future interventions." [bib_ref] The feasibility of the prostate cancer: Evidence of Exercise and Nutrition Trial..., Shingler [/bib_ref] Shingler et al. stress the importance of identifying an optimum time to introduce lifestyle behavior changes for cancer survivors. [bib_ref] The feasibility of the prostate cancer: Evidence of Exercise and Nutrition Trial..., Shingler [/bib_ref] In the EPOP study, drawing from the participants' responses, some respondents noted that they would be able to start the exercises immediately after surgery while still in hospital, while other participants indicated not until they had been discharged home. The consideration of when to commence the exercise program is inter-linked with the issue of exercise setting and whether the setting encouraged patients to commence the exercises, or created a barrier to adherence. What this means for our study is that identifying an optimal time and setting for introducing exercise programs following surgery should not fit a "one size fits all" approach, but should be individually tailored should exercise support be offered. Karlsson (2019) suggests that exercise support should be considered with respect to an individual's current physical activity and attitudes toward physical exercise. [bib_ref] Older patients' attitudes towards, and perceptions of, preoperative physical activity and exercise..., Karlsson [/bib_ref] For the EPOP study, we found that the home-based setting of the intervention enabled the exercises to be carried out without any timetabling restrictions or transportation considerations, which the participants noted could be fitted in and around their daily routines. One of the key advantages of the intervention was that it was individualized rather than groupbased which meant that participants were able to set personal goals for achievement, while noting benefits such as gains in mobility and strength, and recognizable improvements in physical performance. [bib_ref] Feasibility, safety and preliminary evidence of the effectiveness of a home-based exercise..., Suttanon [/bib_ref] Further supervision and intermittent advice by way of a home visit or regular telephone call giving additional support could help identifying what is required for adherence, what the expected optimal achievement could be, as well as providing strategies to mitigate against some of the mobility issues encountered after cancer surgery. Therefore for a future trial, consideration of regular supervision with well-trained professionals could also enhance adherence, and improve participant experience. [bib_ref] Pilot and feasibility studies in exercise, physical activity, or rehabilitation research, El-Kotob [/bib_ref] The lack of blinding noted in the responses from the participants in the EPOP study is a common challenge noted by El-Kotob and Giangregorio (2018) in RCTs of exercise, rehabilitation or physical activity interventions. They acknowledge that while a comparison or "usual care" group is an ideal to the intervention group, however, blinding in physical activity interventions are almost impossible resulting in the potential for bias. In our study, we found that the participants were at first attracted to taking part in the study as there was the potential to be in the exercise group. When finding out they were in the usual care group, they vocalized an element of disappointment. Although the EPOP participants declared feeling dissatisfied, this did not impact significantly on attrition or drop-out rates, unlike for example Barker et al.'s study (2016) whose participants indicated losing interest after being allocated to the control group and was the reason given for withdrawing from the study. [bib_ref] Feasibility of pilates exercise to decrease falls risk: a pilot randomized controlled..., Barker [/bib_ref] Blinding in physical activity-based controlled studies has been shown to be a persistence problem, and requires further reflection for a future study.
The participants in the EPOP study found randomization an acceptable trial process, and acknowledged that they understood the logic of being allocated to either an intervention or usual care group. Unlike the , our participants had reconciled the fact that agreeing to taking part in the feasibility study would not necessary provide an opportunity to receive the exercise program. Considerations around determining the acceptance of randomization may be linked to the EPOP study having modest recruitment figures with 23 consenting in total, with 11 in the intervention and 12 in the usual care group, which enabled recruiters (research nurses) to offer an adequate explanation over what randomization involved.
We accept that there were certain study limitations. While we acknowledge the importance of capturing the views of the research nurses and physiotherapists who were involved in the delivery of the intervention, it was not possible to arrange these interviews due to time constraints and ongoing staff changes, which made it difficult to carry out this data collection. Future research might gain more insights into intervention delivery processes from the position of the healthcare professionals. Although the researcher was a member of the trial team, she did not meet the participants face-to-face who took part in the study, which may have impacted on being able to collect indepth qualitative data. Although there were no explicit refusals by participants who were invited to take part in the qualitative interviews (n ¼ 4), it would have been useful to identify whether these non-responders were unwilling or unable to take part. In addition, conducting interviews with those who refused to take part in the feasibility RCT would be insightful to understand whether changes in recruitment processes may be required for the larger RCT.
# Conclusion
In this embedded qualitative study, we examined abdominal cancer surgery patients' perspectives of an isometric-resistance exercise program with respect to recruiting participants to an RCT, randomization and blinding, and exploring adherence of participants to perform and finish the program, and their experiences of completing the function assessments. Results indicated that participants had the ability and willingness to adhere to the program indicating they found it suitable and appropriate. One patient required supervised modifications to the program who experienced mobility impediments due to surgery. Some participants found the functional assessments difficult to complete with respondents declaring a distinct dislike of them. The timing and setting of the program was a factor that may have impacted on the commencement of exercises. The home-based setting alongside regular supervised advice was found to be very positive. The participants found randomization and recruitment an acceptable trial process.
[fig] 10: Foot and Lower Leg Lie on your back with one leg bent, foot flat on the floor, and the other leg straight. Flex your toes up to the ceiling and lift your straight leg up to about 20 cm off the bed. Make sure to keep your knee straight. Hold for 10 seconds. Repeat 5 times on both sides. [/fig]
[fig] Figure 1: Themes regarding the experiences of patients on the EPOP isometric resistance exercise study. [/fig]
[table] Table 1: Isometric-Resistance Exercise Program Developed for the EPOP Study. [/table]
[table] Table 2: Characteristics of the Patients of the Overall Trial Sample. [/table]
|
Survey of ex vivo drug combination effects in chronic lymphocytic leukemia reveals synergistic drug effects and genetic dependencies
Drug combinations that target critical pathways are a mainstay of cancer care. To improve current approaches to combination treatment of chronic lymphocytic leukemia (CLL) and gain insights into the underlying biology, we studied the effect of 352 drug combination pairs in multiple concentrations by analysing ex vivo drug response of 52 primary CLL samples, which were characterized by "omics" profiling. Known synergistic interactions were confirmed for B-cell receptor (BCR) inhibitors with Bcl-2 inhibitors and with chemotherapeutic drugs, suggesting that this approach can identify clinically useful combinations. Moreover, we uncovered synergistic interactions between BCR inhibitors and afatinib, which we attribute to BCR activation by afatinib through BLK upstream of BTK and PI3K. Combinations of multiple inhibitors of BCR components (e.g., BTK, PI3K, SYK) had effects similar to the single agents. While PI3K and BTK inhibitors produced overall similar effects in combinations with other drugs, we uncovered a larger response heterogeneity of combinations including PI3K inhibitors, predominantly in CLL with mutated IGHV, which we attribute to the target's position within the BCR-signaling pathway. Taken together, our study shows that drug combination effects can be effectively queried in primary cancer cells, which could aid discovery, triage and clinical development of drug combinations.These authors contributed equally:
# Introduction
Abnormal B-cell receptor (BCR) signaling is a key mechanism of disease development and progression in chronic lymphocytic leukemia (CLL), and this is the basis for the clinical success of therapies targeting different downstream kinases. Ibrutinib, for example, is a covalent inhibitor of Bruton's tyrosine kinase (BTK). BTK is essential for the activation of the AKT/ERK and the NF-κB pathway through the BCRand involved in B-cell adhesion and chemokine-mediated homing. Ibrutinib has shown activity in several hematological malignanciesand has led to high response rates in CLLand mantle cell lymphoma. Another example is idelalisib, an inhibitor of the delta isoform of phosphatidylinositol 3-kinase (PI3Kδ). PI3Kδ is a key component of the BCR-signaling cascadethat mediates effects on CLL cell proliferation, migration, and survival. Although inhibitors of the BCR (BCRi) have transformed the treatment of CLL, drug resistance emerges, and BCRi in the vast majority of cases do not lead to complete remission. A promising approach to overcome such limitations are combinatorial approaches, as they may increase response rates and durabilities by circumventing potential resistance mechanisms. However, given the striking single agent activity of targeted agents (e.g., BTK inhibitors, PI3K inhibitors, and BH3 mimeticsand the advent of effective combination regimens, the clinical development of new combination regimes may be held back due to a lack of opportunities for testing.
Most systematic combinatorial drug studies to date have been based on cell line models. In diffuse large B-cell lymphoma (DLBCL) cell lines, inhibitors of mTOR, PI3K, Bcl-2, and chemotherapeutic agents were reported to show synergistic activity when combined with ibrutinib. Also, ibrutinib cooperates with lenalidomide by downregulating IRF4. While cell lines are valuable disease models, the use of primary patient material has potential advantages including the absence of subclone selection and the ability to better represent the natural molecular heterogeneity of the disease. CLL cells in the body reside in multiple compartments with different microenvironments that might alter pathway sensitivity. Here, we limit our study to primary patient material from the peripheral blood (PB), which has the advantage of being easily accessible. This implies a need for further validation of the results with respect to microenvironment-dependent modulation, but it also facilitates the application of such assays in precision medicine approaches for the donating patients.
Hypothesis-driven studies that investigated particular combinations of drugs on primary CLL cells showed synergistic effects for the combination of the Mdm2 inhibitor nutlin-3and of the XPO1 inhibitor selinexor with ibrutinib. Furthermore, ibrutinib and idelalisib combined with the Bcl-2 inhibitor venetoclax yielded synergistic effects in CLL. Such studies provide potential starting points for clinical development, as shown by successful trials for combination of ibrutinib and venetoclax. Combination of BCRi with immunotherapy has entered clinical care, and combination of ibrutinib with chemoimmunotherapy (bendamustine/rituximab, fludarabine/cyclophosphamide/rituximab) showed the promising results in a phase II and III studies. A recent study combined epigenome with single-cell chemosensitivity profiling in patient samples collected before and during ibrutinib therapy and observed preferential sensitivity to proteasome, PLK1, and mTOR inhibitors during ibrutinib treatment. Despite the success of some new combination regimens including ibrutinib in clinical studies, there has been a lack of larger, more systematic approaches that characterize a range of combinatorial drug treatments across representative cohorts of primary cancer cells, to search for optimal, new combinations. Here, we report an assay in which we systematically investigated drug combination effects on a set of primary CLL samples. In light of the clinical success of BCR targeting, our study focuses on combinations involving BCR inhibitors, as these are already a part of clinically relevant combination approaches in CLL and other B-cell lymphomas. We uncovered molecular subgroup-specific effects and found differences between the combinatorial effects of PI3K and BTK inhibitors specifically in CLL with mutated IGHV.
# Material and methods
## Patient samples
PB samples with high lymphocyte fractions (median 95%, ranging from 78-99%) from 52 patients of the University Hospital of Heidelberg fulfilling standard diagnostic criteria for CLL were obtained. Samples from two additional patients were used in a first experiment but not considered for the analysis due to the lack of a suitable control plate. Patients provided written informed consent in accordance with the Declaration of Helsinki and local ethical approval. Mononuclear cells were isolated by centrifugation over Ficoll-Paque Premium (GE healthcare, Freiburg) and cryopreserved.
## Cell culture and in vitro treatments
Cells were thawed in RPMI-1640 supplemented with 5% FBS, 1% penicillin/streptomycin, and 1% glutamine. We used 40,000 primary cells per well, cultured in 50 µl RPMI supplemented with 10% pooled and heat inactivated ABtype human serum (RPMI-HS, Biomedicals) at 37°C, 5% CO 2 . The drugs and cells were dispensed using an electronic pipette (Millipore) into 384 flat bottom Greiner plates. The plates were covered with breathe easy sheets and Greiner cell culture lids to prevent evaporation and incubated at 37°C with 5% CO 2 for 48 hours.
Compounds and combinatorial drug screen Drugs (n = 34, Supplementary Table 1) were obtained from commercial suppliers and kept at a concentration of 10 mM in DMSO. Each of a set of 32 drugs (termed library drugs in the following) was combined with each of a set of 11 drugs (termed combination drugs) ; two combination drugs were not part of the library, thus a total of 34 drugs were used. The drugs were chosen from clinically used drugs (including fludarabine, ibrutinib, venetoclax) and other inhibitors of pathways of importance in CLL. We prepared a master plate by diluting each library drug using the same doses as described by Dietrich et al.. In short, all 32 library drugs were tested at 5 concentrations, i.e., with fourfold serial dilutions starting at 1 µM for navitoclax, venetoclax, SNS-032, YM155, and doxorubicine, 4 µM for arsentrioxide, 7 µM for chaetoglobosin A and 10 µM for all other drugs. For each experiment, drugs from the master plate were diluted with 10% pooled and heat-inactivated AB-type human serum (RPMI-HS) and transferred to 384well plates (Greiner) using electronic pipettes (Millipore). We then added to these wells each of 11 combination drugs at 1 or 2 concentrations (100 nM for ibrutinib, idelalisib, spebrutinib, R406, afatinib, fludarabine, pomalidomide, everolimus, and encorafenib; 10 nM for YM155 and ibrutinib, and 5 nM for venetoclax) and tested each combination on the samples from 6 to 52 patients. The highest number of CLL samples were analyzed for the combination drugs ibrutinib (n = 52 patient samples), idelalisib (n = 30), afatinib (n = 30), and spebrutinib (n = 16) . Based on similarity of observed responses for combinations with ibrutinib 10 and 100 nM in the first 16 patient samples, only combination with ibrutinib 100 nM was tested in the further 36 patient samples .
## Validation experiments
Measurements of 10 × 10 concentrations were assayed with selected patient samples. Navitoclax, venetoclax, and afatinib were tested in combination with ibrutinib with twofold serial dilutions starting at 2 µM (two patient samples) and 0.2 µM (three samples) for navitoclax, at 0.1 µM for venetoclax, at 5 µM for afatinib (three samples), and at 20 µM for afatinib (two samples) and ibrutinib.
## Cytotoxicity assay
To assess cell viability in the primary screen and in the validation experiments, the cells were incubated with the drugs for 48 h (based on the results of Dietrich et al., and then 12 µl of CellTiter Glo luminescent cell viability assay reagent (Promega) was added using an electronic pipette (Millipore). Plates were incubated for 15 min at room temperature. Luminescence was measured using a microplate reader (Tecan) with an integration time of 200 ms. 14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0. 14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0. 14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0. 14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0. .8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=1. .8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=1. .8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=1.
## Molecular characterization
To explore molecular underpinnings of drug response heterogeneity, we used whole exome sequencing for paired tumor and normal samples (n = 52), targeted 454 sequencing (n = 52), assessment of IGHV status (n = 52), assessment of structural variants (n = 52), genome-wide DNA methylation profiles (450k microarrays) (n = 52), and RNA sequencing (n = 52), as previously reported. Supplementarygives an overview of the patients' somatic mutation landscape. IGHV mutation status was determined as described by Ghia et al.. The variant allele frequency (VAF) of mutations was determined, but not included in the analysis. The minimum VAF of the TP53 mutant samples in our study was 5% (see.
## Pamgene tyrosine kinase array
Kinase activities of primary CLL samples were measured using the PamStation® 12 system in combination with PamChip® Tyrosine Kinase Array Chips according to the manufacturer's instructions (www.pamgene.com). Briefly, primary CLL cells were lysed in M-PER Mammalian Extraction Buffer (Pierce). Cells were treated with 1.67 µM afatinib before or after cell lysis. 2 μg of cleared cellular lysate was mixed with 4 μl 10× protein tyrosine kinase reaction buffer (PK), 0.4 μl 1 M DTT, 0.4 μl 100× BSA, 1 μl 4 mM ATP, and 0.3 μl 1 mg/ml monoclonal antiphosphotyrosine FITC conjugate (clone PY20) adjusted to 40 μl with distilled H 2 O. All chemicals were provided by Pam-Gene International BV. Each array was blocked with 0.2% BSA and washed with PK solution. Afterwards, kinase reaction was carried out at 30°C. The reaction mix was pulsed back and forth through the porous material of the PamChip. Every fifth cycle a picture with a built-in CCD camera and after 60 cycles five additional pictures with increasing exposure time were taken. Spot intensities were normalized to local background signal by subtracting the median background signal from the median spot intensity, using the BioNavigator (Pam-Gene) software. To increase the dynamic range of the measurements, the five pictures taken after 60 cycles were summarized into a single value by determining the slope of these signals as a function of exposure time, scaled by a factor of 100. The resulting values of the three biological replicates were logarithm transformed and averaged. The upstream tyrosine kinase responsible for on-chip peptide phosphorylation was predicted by using the "PamApp for PTK upstream Kinase Analysis" within the BioNavigator software (PamGene). This analysis is based on documented kinase-substrate relationships and in silico predictions from the phosphoNET database.
# Statistical analysis
The luminescence values per well were divided by the median of the DMSO-controls of each plate, resulting in values that indicate relative viability compared with the control; in the following, we term these viability values. As multiple wells on each plate contained the combination drug as a single agent we used the median of their viability values as a measure of the single agent effect for the plated sample. For the library drugs, every plate contained one well per concentration, which provided a viability value for the effect in combination or as a single agent in the plated sample. For four patient samples, we acquired replicate measurements; these were highly similar and the average values were used for subsequent analyses. Measurements of viability values above 1.4 (corresponding to the 99.9% quantile of the data) were excluded from the analysis (i.e., treated as missing values), as these are likely outliers due to technical reasons. To map the landscape of combinatorial effects of drugs and to identify synergistic combinations, the effects of combinations were assessed by comparison to an independent effect model, as follows. For each combination of two drugs A and B, we compared the measured viability v AB for the combination to the expected viability value given by the product of the corresponding single drug effects
[formula] v à AB ¼ v A v B . [/formula]
We defined a synergy index (SI) as the average of the difference, v à AB c ð Þ À v AB c ð Þ, taken across the five concentrations c of the library drug for each patient sample. In particular, viability measured lower than predicted resulted in SI > 0, indicating synergy. To summarize drug combinations across patient samples, the median SI was used. This approach to assess synergistic effects is strongly related to the Bliss-independence model, and it shares its basic assumption of effect independency. In contrast to the Bliss model, we also considered combination effects with drugs that had a prosurvival effect, i.e., we included viability values above 100%. While the Bliss Index is often defined as the quotient of the expected and measured combination effect, here we used instead their difference to make the SI more numerically stable for small values of v AB (i.e., drug combinations with strong effects). To test significance of synergistic combinations we used a one-sided paired t-test on the viabilities from the independence model and the measured viability from the drug combination across patient samples, separately at each concentration.
In addition, drug combination effects were also quantified with the highest single agent (HSA) approach.
Here, we thresholded all data to lie between 0 and 100% and defined the effect of a drug A based on the viability values v A as e A ¼ 1 À min 1; v A ð Þ. The HSA combination index is then given by:
[formula] CI HSA ¼ e AB À max e A ; e B ð Þ; [/formula]
with a value greater than 0 indicating that the combination has a stronger effect than each single agent on its own. To test significance of synergistic combinations we used a onesided paired t-test of the maximal single effect and the combined effect at each concentration across patient samples.
The p values resulting from the t-tests were adjusted for multiple testing across all drug-drug combinations and concentrations (m = 1920 tests) using the Benjamini-Hochberg procedure, separately for the two types of combination index (i.e., SI and HSA).
Both approaches above are based solely on drug effects at single concentrations and, whilst providing useful screening criteria for potentially interesting combinations, they might also identify less interesting cases. In particular, strongly nonlinear single agent dose-response relationships in the considered concentration range (such as for example observed with YM155, afatinib, venetoclax, and fludarabine) can result in high values of SI and/or CI for noninteresting combinations, including combination of the drug with itself. To account for this limitation and to complement the effect-based combination scores at single concentrations we used the 10 × 10 validation experiments to assess the synergy for individual patient samples using the ZIP score(as implemented in the R package synergyfinder), which compares the potency of the dose-response curves between single drugs and their combination.
To assess associations of genetic features (e.g., IGHV status and TP53 mutation) with drug responses, Student's t test with equal variance between groups was performed between the corresponding groups for each single drug or drug combination that had at least three samples in each group (i.e., M-CLL/U-CLL or TP53wt/mut). Resulting p values were adjusted for multiple testing using the Benjamini-Hochberg procedureacross all tested genetic features, tested set of drugs (i.e., all drug combinations or all library compounds) and concentrations leading to a total of m = 2226 or m = 320 tests.
# Results
## Landscape of drug response and modulation by genetic features
First, we evaluated the effect of 32 library drugs applied as single agents in 52 primary CLL samples ex vivo. For this, we created dose-response curves based on 5 concentrations using ATP-based viability assessment 48 h after drug exposure, resulting in 8320 (32 × 5 × 52) measurements ("Material and methods"). To explore associations of recurrent genetic features (TP53 mutations and IGHV status) with drug sensitivities, we tested for differential drug responses between the sample groups with and without each of these mutations (t-test, FDR < 5%,. This analysis confirmed previously reported associations, including increased sensitivity of TP53 wild-type samples to fludarabine, doxorubicin, and nutlin-3and increased sensitivity of CLL samples with unmutated IGHV status (U-CLL) to BCRi, e.g., spebrutinib, R406, and ibrutinib, as well as two checkpoint kinase (CHEK) inhibitorsthat we recently reported to target the BCR. In addition, we tested for associations of drug responses with the DNA methylation status of the samples according to Oakes et al.. We found that a low-programmed DNA methylation status of the samples was associated with increased sensitivity toward inhibitors of BCR and BCR downstream targets as well as of CHK, a finding consistent with earlier studies(FDR < 5%, , while the high-programmed methylation cluster displayed increased sensitivity toward Bcl-2 inhibitors and doxorubicin. These findings were also recovered by unsupervised clustering of samples by their response profiles to the 32 agents across concentrations: we observed a kinase inhibitor-based gradient with increased sensitivity toward kinase inhibition for samples that were largely U-CLL. Conversely, we considered the (dis)similarities of the drugs by computing for each pair of drugs the correlation coefficient across the 52 CLL samples. Drugs with similar functional mechanisms clustered together with a dominant cluster of kinase inhibitors targeting BCR, the downstream targets MEK or AKT or interfering with the BCR-signalosome (e.g., AZD7762). Likewise, navitoclax and venetoclax, both targeting Bcl-2, or fludarabine and nutlin-3, both TP53 dependent drugs, clustered together. Altogether, these findings indicate that our assay platform produces biologically relevant data that are consistent with the previous results and known drug mechanisms.
## Combinatorial drug screen on primary cll cells
To study heterogeneity of drug responses in the combinatorial screen we first clustered the samples by their response profiles to the 32 agents across concentrations in combination with ibrutinib (n = 52). In line with the findings for ibrutinib as a single agent, IGHV status emerged as a dominant factor in the clustering, via several kinase inhibitors (including dasatinib, AZD7762, PF477736). However, using the combinatorial screen we uncovered further heterogeneity of responses. We identified a subgroup of patient samples with high sensitivity to all combinations, which was driven by exquisite sensitivity to ibrutinib. These were mainly IGHV unmutated samples (6/7). To further investigate the influence of IGHV status as well as TP53 mutation on sensitivities toward drug combinations, we tested for such associations using marginal testing (t-test) for all drug-drug combinations that had at least three samples in each group (i.e., M-CLL/U-CLL or TP53wt/mut, respectively). As shown in, the strength of response to a large set of drug combinations with ibrutinib were associated ** ** ** * **** **** *** *** ** ** *** *** **** *** *** ** **** **** **** **** *** ** *** *** ** * **** **** **** ** * Navitoclax Venetoclax . For each drug, the five concentrations were tested separately and the most significant concentration is shown. The x-axis shows the mean difference in viability. Positive difference indicates higher sensitivity of U-CLL/TP53 wild type compared with M-CLL/TP53 mutated. The y-axis shows the logarithms of p values obtained from a Student's t test and adjusted for multiple testing across both settings and all tested drugs and concentrations. The dashed line indicates a FDR-threshold of 5%. Significant differences for IGHV were evident for combinations with BCR-pathway inhibitors (Spebrutinib, R406), CHK inhibitors (PF477736, AZD7762), and dasatinib. Significant differences for TP53 were revealed for fludarabine and nutlin-3. b Drug response curves of p53-interfering drugs fludarabine, nutlin-3, and doxorubicine. Patient samples are grouped according to TP53 mutational status and shown is the mean viability within each group. Error bars denote two standard errors. The drugs are significantly more active in TP53 wild-type samples at multiple concentrations. *p < = 0.05, **p < = 0.01, ***p < = 0.001,****p < = 0.0001. c Drug response curves as in (b) for selected drugs stratified by IGHV mutational status. The three BCR inhibitors, dasatinib, and two CHK inhibitors in the upper and middle rows are significantly more active in U-CLL samples at multiple concentrations. For the two Bcl-2 inhibitors venetoclax and navitoclax higher sensitivity is observed in M-CLL samples. d Heatmap showing the response to the library drugs (columns) across all patient samples (n = 52, rows). Color code indicates viability values. Samples are annotated by IGHV and TP53 mutational status (black = mut/M-CLL, white = wt/U-CLL). Unsupervised clustering identified a BCR/CHK inhibitor-based gradient, which separates M-and U-CLL. e Pearson correlations for each pair of drug responses across all 52 CLL patient samples are shown. Each matrix element corresponds to a correlation coefficient, red indicating positive and blue negative values. Based on the clustering, patterns of high correlation emerged for drugs targeting the same pathways (e.g., Bcl-2 inhibitors and BCRi/CHK inhibitors).
with IGHV mutation status (FDR < 5%), including BCRpathway inhibitors (spebrutinib, R406), CHK inhibitors (PF477736, AZD7762), and dasatinib. Likewise, IGHV status also had a strong impact on drug response phenotypes for combinations with idelalisib . These findings are in line with the critical and Viability effect −log10(pval.adj) **** **** *** **** **** ** *** *** **** **** ** *** **** **** **** differential activity of BCR signaling in these molecular groups.
The influence of TP53 mutations on the response to fludarabine and nutlin-3 as single agents was found to be preserved in combinations with ibrutinib. In addition, we observed increased effects in the lowprogrammed methylation cluster for combinations with ibrutinib (n = 16) and idelalisib (n = 10) with inhibitors of BCR and downstream targets as well as CHK inhibitors as well as for afatinib in combination with CHK inhibitors .
## Landscape of drug combination effects in cll
To gain an overview of combination effects, we clustered the 32 library drugs based on their median synergy indices ("Material and methods") across patient samples with the 11 combination drugs, including both concentrations for ibrutinib . We identified drugs with multiple synergies (e.g., fludarabine (n = 5 out of 12), nutlin-3 (n = 3 out of 12), navitoclax (n = 7 out of 12), venetoclax (n = 4 out of 12), and afatinib (n = 5 out of 12)) and found striking similarities for combinations involving a BCR inhibitor as one of the partners, which suggests strong functional convergence . For instance, both ibrutinib and idelalisib had similar, strong patterns of synergy with the Bcl-2 inhibitor navitoclax. In contrast, combinations of different BCRi or of BCRi and the CHK inhibitors did not display synergy (white, SI ⩽ 0). While it might be hypothesized that simultaneous inhibition of multiple BCR kinases (BTK, PI3K, SYK) could provide synergy compared with single target inhibition, e.g., due to increased efficacy of BCR inhibition, our results indicate that this is not the case. Simultaneous inhibition of downstream pathway members (AKT, mTOR) did not result in synergy according to the synergy index but led to an additional viability decrease compared with the effect of the single drugs (Supplementary . In particular, combination of the mTOR inhibitor everolimus with ibrutinib resulted in decreased viabilities compared with each single drug (78-81% mean viability of control across 5 concentrations for combination vs. 86% mean viability for ibrutinib 100 nM vs. 91-94% mean viability for everolimus) . Similarly, combination of ibrutinib with the AKT-inhibitor MK-2206 showed decreased viability compared with the single drugs: mean viability under the combination ranged from 58-85%, compared with 63-101% for MK-2206 alone and 86% for ibrutinib 100 nM alone . These findings imply that AKT and mTOR are activated in part in a BCR independent manner in CLL, and thus are relevant targets even in the presence of BCRi.
## Bcri and bcl-2 inhibitors show synergy in cll
Navitoclax is a Bcl-2/Bcl-X L /Bcl-w inhibitor with clinical activity in CLLand recent clinical studiessuggest that combinations of BH3 mimetics and inhibitors of BCR signaling are effective in CLL. Indeed, we observed strong cooperative activity of navitoclax in combination with ibrutinib (100 nM), predominantly at intermediate concentrations of navitoclax [2.5 µM-15.6 nM] .
To understand inter-individual differences between such synergistic effects, we considered the individual patient samples' SIs . Interactions were synergistic (SI > 0) in almost all patient samples (49/52), with a median of 0.08 . The SI was correlated with sensitivity to navitoclax as single agent with high synergy indices for less sensitive samples. Using a 10 × 10 combination matrix of concentration series ("Material and methods", , we confirmed synergistic effects across concentration ranges (strongest for navitoclax 6.2-50 nM). At higher concentrations of navitoclax no synergy could be observed due to complete cell death already under the single agent.
As thrombocytopenia caused by Bcl-X L inhibition limited its clinical use, venetoclax, a selective Bcl-2 inhibitor, was developed. Venetoclax showed stronger viability effects than navitoclax at equimolar dose[1 µM-250 nM]. The x-axis shows the mean difference in the viability values. Positive difference indicates higher sensitivity of U-CLL/TP53 wild type. For each drug, the five concentrations were tested separately and the most significant concentration is shown. The y-axis indicates the logarithms of p values obtained from a Student's t test and adjusted for multiple testing across both settings and all tested drugs and concentrations. Significant differences were evident for combinations with core BCR-pathway inhibitors (Spebrutinib, R406), CHK inhibitors (PF477736, AZD7762), and dasatinib. Significant differences for TP53 were revealed for fludarabine and nutlin-3 as combination drug. c Drug response curves of CLL samples to CHK inhibitors (AZD7762, PF477736) and dasatinib in combination with ibrutinib 100 nM. Shown is the mean viability within each group (M-CLL/U-CLL) and error bars denote two standard errors. U-CLL samples are significantly more sensitive toward CHK inhibitors and dasatinib. p values were assessed using Student's t test. d Drug response curves as in (c) stratified by TP53 mutational status. TP53 wild-type samples are more sensitive toward fludarabine and nutlin-3. . Therefore, synergistic activity in combination with ibrutinib 100 nM was observed for doses below 63 nM. We confirmed the synergistic effects of venetoclax and ibrutinib across concentration ranges (strongest for ibrutinib 78 nM-1250 M, venetoclax 1.56-6 nM) using a 10 × 10 concentration matrix . 14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0 14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0 14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0 14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0 .8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2. p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2. p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2. p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.8e−08 p=2.and current trial activity (NCT02756897, NCT03580928, NCT03868722).
## Combinations of bcri with fludarabine or nutlin-3
Fludarabine-based combinations are standard of care in CLL. Previous work suggested combination activity of ibrutinib and chemotherapeutic agents such as the purine analog fludarabine. Our screen uncovered a number of potentially synergistic combinations for fludarabine including BCR inhibitors . Ibrutinib and fludarabine showed synergistic effects across multiple concentrations (median SI: 0.08). This effect was observed in 48/52 patient samples. Based on the association of p53 status and IGHV on the effect of both drugs, we asked how TP53 and IGHV status influenced synergy. While unmutated IGHV positively influenced synergy, TP53 status did not influence synergy.
We observed synergistic combination activity of the Mdm2 inhibitor nutlin-3 with ibrutinib, idelalisib, spebrutinib, and R406 . This includes the combination of ibrutinib 100 nM with higher concentrations of nutlin-3 (10-2.5 µM), with a median SI of 0.05 and SI > 0 in 42/52 patient samples, in line with a finding of Voltan et al..
## Combinations of bcri and afatinib display synergy in cll
Afatinib is an EGFR inhibitor in clinical use for nonsmall cellular lung cancer harboring activating EGFR mutations. A striking number of synergistic interactions were observed for afatinib in combination with BCRi . For example, afatinib in combination with ibrutinib or spebrutinib showed the highest SI among all combinations with afatinib (0.10 for ibrutinib 100 nM; 0.12 for spebrutinib). To understand this unexpected finding, we assessed the effect of afatinib across concentrations and observed that afatinib increased viability at low concentrations . Ibrutinib completely antagonized this effect, suggesting that the effect is mediated by interference with upstream parts of BCR signaling. The synergistic effect was most pronounced at intermediate afatinib concentrations 5 and 625 nM] . Synergy indices calculated individually for each patient sample revealed synergy for all patient samples with a tendency of higher synergy in U-CLL samples .
To expand on these findings, we included afatinib as a combination drug and assayed it against the 32 drug library for 30 samples. Synergy with BCRi including ibrutinib, idelalisib, and R406 and the abolition of afatinib's prosurvival effects by ibrutinib were confirmed. The synergistic effect of combinations of ibrutinib with afatinib at most concentrations (strongest for afatinib 625 nM-1.25 µM) was also verified using 10 × 10 concentration matrices starting at 5 µM (three patient samples) and 20 µM (two patient samples) .
To further confirm the effects exerted by afatinib we investigated the viability effects by FACS (Annexin/ 7AAD). We observed increased viability after 24 and 48 h, an effect blocked by addition of ibrutinib (data not shown). Afatinib was designed to inhibit EGFR at high specificity. Its effect in our study, however, is unlikely to be mediated Synergistic interactions in CLL. a Heatmap summarizing all synergistic combination effects as determined by the independent effect model ("Material and methods"). Colors indicate for each combination the median synergy index (SI) across all assayed patient samples (numbers depending on the combination drug as in , with red representing synergy (median SI > 0) and white lack of synergy or antagonism (median SI ⩽ 0). Rows and columns are ordered according to a hierarchical clustering dendrogram. Drug combinations with a star showed a significant synergy in at least one concentration at a FDR of 5%. The heatmap revealed combination partners with multiple synergies (synergy index > 0, e.g., navitoclax, fludarabine, afatinib) and showed similarities for different BCR inhi- . e Combination activity for navitoclax and ibrutinib in five primary CLL samples as measured in a 10 × 10 matrix of combinational concentration series and assessed using the ZIP score ("Material and methods"). The heatmap indicates the viability values (left) and ZIP scores (right) at each pair of concentrations. ZIP scores shown in the heatmap were smoothed, using the mean ZIP score across the five samples in a running 3 × 3 concentration window. A positive ZIP value (red) denotes synergistic effects, a negative value lack of synergy (blue). Synergistic combination activity is observed in certain concentration ranges.
via EGFR, as EGFR is not expressed in CLL cells. We also found no evidence for the expression of EGFR in CLL cells at RNA or protein level (data not shown). Further binding partners of afatinib include ERBB2, ERBB4, GAK, BLK, IRAK1. To explore the relevant target(s) in CLL, we analyzed RNA expression data, which showed that . e Combination activity for afatinib and ibrutinib as visualized by a 10 × 10 matrix of combinational concentration series as in .
BLK, GAK, IRAK1a, and MAPK9/14 are expressed in CLL; thus it is possible that the activity of afatinib in CLL is mediated via one or more of those components (Supplementary . Irreversible kinase inhibitors such as afatinib and ibrutinib, which target a cysteine residue in the ATP binding site of kinases such as EGFR and BTK, often bind both, EGFR and BTK, albeit often with higher potency for the intended target. When testing other EGFR inhibitors, we found for some of them (e.g., canertinib) similar dose-response curves including a prosurvival effect at lower or intermediate concentrations . Previous work showed that EGFR inhibitors in AML cells can target Syk within the BCR pathway and induce cell differentiation and cell death. We used the PamGene array ("Material and methods") to study the effect of afatinib on kinase activity in three primary CLL samples sensitive to afatinib treatment. By adding afatinib before or after lysis of the cells we identified multiple members of the Src-Kinase family (BLK, SRC, YES) as well as the Syk kinase as direct and downstream targets of the drug . When testing the combination activity of ibrutinib with a Syk inhibitor (R406) we did not find synergy. In summary, inhibition of Syk in CLL does not fully explain the prosurvival effect of afatinib and abolition of this effect when adding ibrutinib. This indicates that afatinib effects in CLL might be mediated by BLK functioning upstream of BCR-signaling components such as BTK.
## Synergy of ibrutinib and mek-inhibitor selumetinib
The effectiveness of MEK inhibition in CLL has been demonstrated in ex vivo studies. Our screen uncovered a synergistic combination activity of ibrutinib and selumetinib, predominantly at higher (>625 nM) selumetinib concentrations. While the overall effect was weak, it was consistently observed in 39/52 patient samples (median SI: 0.02), suggesting that a part of MEK activation in CLL is independent of the BCR.
## Combination of bcr inhibitors lack cooperative activity in cll
To characterize potential cooperative effects among different inhibitors of the BCR pathway, we included inhibitors of BTK (ibrutinib, spebrutinib), PI3K (idelalisib), and Syk (R406) in the drug library. Combination of these BCRi with one another did not result in synergistic effects. As shown in, the BCRi combinations showed small or no synergistic effects. For example, combination of ibrutinib with spebrutinib, and of idelalisib with duvelisib showed the lowest SI, in line with complete target inhibition by either drug. The use of BCR and CHK inhibitors or dasatinib lacked cooperative combination activity (51-89% median viability in combination vs. 53-86% for AZD7762 alone, 50-83% in combination vs. 48-81% for dasatinib alone, 46-102% in combination vs. 42-109% for PF477736 alone), suggesting that the activity of all these inhibitors in CLL derives from already maximal BCR component inhibition.
## Combinations of bcr inhibitors with other drugs
Given the overall similarity of combination effects of BTK and PI3K inhibitors , we asked whether there may also be differences between them. We exploited the parallel design of our study to compare combinations involving idelalisib and combinations involving ibrutinib. In line with co-targeting of the BCR we found highly correlated patterns of interactions with the drug library (r = 0.94). However, we observed drug-specific differences of the combination effects, suggesting that the effects of inhibition of BTK and PI3K are not identical (FDR < 5%, paired t-test for each concentration,. In particular, we observed differential sensitivities for the combination with BTK inhibitor spebrutinib, where combinations with the PI3K inhibitor idelalisib were significantly more effective, and for the combination with afatinib, where combinations with ibrutinib 100 nM were significantly more effective. These differences were unlikely to be caused by dosage, as ibrutinib and idelalisib had similar cytotoxicity at 100 nM (median viability across all combinations: 84.3% for ibrutinib 100 nM vs. 83.8% for idelalisib 100 nM, p = 0.68). Taking a global view across all library drugs, we found that combination with ibrutinib resulted in a stronger correlation between the response profiles of the different library drugs compared with idelalisib, red cluster). These differences were more pronounced in M-CLL than in U-CLL.
# Discussion
While inhibitors of the BCR have revolutionized treatment options, addressing resistance and relapse requires systematic approaches for identifying effective drug combinations. Here we present an unbiased screen to test drug combinations in CLL and use it to identify synergistic drug combinations from a panel of 32 × 11 drugs. We identified novel auspicious therapeutic options by co-targeting the BCR pathway with other pathway dependencies and observed striking combination activity of BCRi with a range of mechanistically diverse compound classes including Bcl-2 inhibitors, chemotherapeutic agents, and afatinib. Bcl-2 inhibitors were previously suggested by ex vivo and in vivo studies to show beneficial combination effects with ibrutinib in CLL, DLBCL, and MCL. This synergistic effect was confirmed in our screen for two Bcl-2 inhibitors, navitoclax, and venetoclax, and can be explained by simultaneous interruption of two pathways representing salvage pathways under monotherapy. The doses of ibrutinib, idelalisib, venetoclax, and navitoclax used in our experiments are attainable in patients' blood plasma, so translation of such combinations into the clinic appears feasible. Indeed, the first clinical studies in patients undergoing combination therapy showed an acceptable tolerance and a similar adverse-event profile compared with monotherapy with ibrutinib or venetoclax. samples. Comparable cytotoxic effects were observed across concentrations for the library drugs as single agents and in combination. c The heatmap shows the correlations of the 32 library drugs in combination with ibrutinib (lower triangle) and idelalisib (upper triangle) based on the viability values across 30 patient samples. Clusters of drugs with high correlation appear as red squares. Combining the library drugs with ibrutinib lead to overall increased correlation between pairs of library drugs than for the library drugs alone, for idelalisib this effect was weaker. d As in (c) shown separately for U-and M-CLL. Correlation of ibrutinib and idelalisib was similar within U-CLL. For M-CLL, differences specific to idelalisib were observed.
Furthermore, our screen recovered synergistic activity of ibrutinib and idelalisib in combination with fludarabine. This is in line with previous preclinical work that suggested combination activity of chemotherapeutic agents with ibrutinib in DLBCL by disruption of NF-κB signaling and decreased expression of antiapoptotic proteins by ibrutinib re-sensibilizing B-cells to undergo apoptosis. In CLL, a sensitization of idelalisib treated CLL cells to fludarabine was described. In line with preclinical data, clinical studies suggested beneficial combination effects of chemotherapeutic agents with ibrutinib in CLL. In addition, we observed cooperative combination activity for MEK-inhibitor selumetinib and ibrutinib. This complements previous studies in DLBCL and MCL cell lines and a DLBCL mouse model that suggested potentiation of ibrutinib's viability decreasing effects by downregulation of p-ERK-1/2 through inhibition of the MEK/ERK/AKT pathway.
Interestingly, synergistic effects were also observed for EGFR inhibitor afatinib in combination with BCRi as well as with agents targeting downstream (e.g., TAE684, NSC74859, MK-2206). Previous work demonstrated that the activity of EGFR inhibitor gefitinib in U-CLL by inhibition of BCR signaling via reduced phosphorylation of Syk/Zap-70/ERK/AKT and decreasing prosurvival proteins such as Mcl-1. Likewise, in AML Syk inhibition through gefitinib suggests an EGFR independent mode of action in leukemia cells. The role of Syk is supported by the PamGene kinase assay that we performed in CLL cells with afatinib, where we found Syk among the top ten targets of afatinib. For afatinib, synergistic effects resulted in part from abolition of its prosurvival effects at low concentrations. Here, we identified BLK as a possible afatinib target upstream of the BCR. This connection to the BCR might explain the synergistic effects.
Although previous studies demonstrated that the efficacy of combinations of BCRi in DLBCL cell lines, CLL and MCL, we observed no evidence of beneficial combination effect for ibrutinib with other BCRi in our study. Nevertheless, targeting downstream elements in the BCR pathway could result in increased cytotoxicity in vivo as BCRi induce mobilization of CLL cells from the lymph nodes, their survival promoting niches in vivo. While the combination of 32 drugs with ibrutinib or idelalisib suggested overall similarity based on targeting the BCR, we uncovered pathway-specific differences. These were most pronounced in M-CLL where BCR signaling is less crucial and several other survival pathways are important. Here, the effect of idelalisib on correlation of drug combinations was less pronounced compared with ibrutinib. This might be due to idelalisib targeting downstream of ibrutinib and PI3K being involved in many other pathways compared with BTK.
While this screening platform offers a practical systematic approach for the identification of new drug combinations in primary leukemia cells ex vivo, it also has limitations. Cell death was assessed using an ATP-based viability assay (CellTiter Glo), therefore agents with negligible cytotoxicity cannot be judged well. CLL cells taken from PB do not proliferate, so the effect of drugs impeding proliferation cannot be gauged. Preclinical ex vivo models can also not model organismal toxicity, as evidenced by unexpected severe organ toxicities of combination approaches in clinical trials. Due to a limited sample size, our screen lacked the power to uncover further heterogeneity of drug effects due to molecular heterogeneity of tumors. While we did not incorporate VAFs into the analysis, this information could be used to further increase the power to detect response differences associated with somatic mutations.
Despite these limitations, this study identified potentially promising combination approaches for CLL that may warrant translation into clinical trials. In particular, this study, supported by the prior results, provides a rationale for further trialing the addition of Bcl-2 inhibitors or fludarabine to ibrutinib/idelalisib therapy. Second, given the striking synergistic interaction of BCRi with the EGFR inhibitor afatinib, activity of afatinib alone and in combination with BCRi in CLL should be further assessed.
To foster further research in this area, we provide a Shiny web resource as a resource for the community to explore the results from our screen (http://mozi.embl.de/ public/combiScreen).
## Data availability
The data of the molecular profiling were obtained from Dietrich et al.. The raw data generated by the combinatorial drug screen experiments are available from the EMBL-EBI BioStudies repository (accession number S-BSST381). The processed data can be explored using our Shiny App http://mozi.embl.de/public/combiScreen.
## Code availability
The code for the analysis contained in this paper is available at https://github.com/bv2/drug_combi_screen_CLL. Generation Drug Response Profiling for Personalized Cancer Care" and the "Krebsliga".
Author contributions TZ conceptualized the study. ML, LS, and TZ designed the experiments. ML, TW, LW, CM, LS, AJ, JH, BW, SD, KT, and HB performed experiments. BV, ML, MO, JL, WH, and TZ analyzed data. ML, BV, WH, and TZ wrote the original draft of the paper. All authors reviewed and edited the paper.
## Compliance with ethical standards
Conflict of interest LS is currently a full-time employee of Takeda Pharma Vertrieb GmbH & Co. KG.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/. |
Bifidobacterium and enteral feeding in preterm infants: Cluster-randomized trial
Background: This study evaluated the benefit of Bifidobacterium bifidum OLB6378 (B. bifidum) in very lowbirthweight (VLBW) infants (birthweight <1500 g) for the acceleration of enteral feeding. Methods: A cluster-randomized, double-blind, placebo-controlled trial was conducted in 19 hospitals, divided into two groups: the B group (n = 10 hospitals; B. bifidum given to infants within 48 h of birth) and the P group (n = 9 hospitals; infants received a placebo). The primary outcome was establishment of enteral feeding after birth, defined as the postnatal day at which enteral feeding exceeded 100 mL/(kg/day). Secondary outcomes were defined as incidence of morbidity and somatic growth before discharge. Results: Overall, 283 VLBW infants were enrolled in the study: B group, n = 153; and P group, n = 130. Enteral feeding was established within 21 days after birth in 233 infants, of whom 119 received B. bifidum and 114 received placebo until their bodyweight reached 2000 g. Enteral feeding was established significantly earlier in the B group, at 11.0 ± 3.6 days versus 12.1 ± 3.8days in P group (P < 0.05). Infant growth during the stay in the neonatal intensive care unit was not different between groups, but the incidence of late-onset sepsis among all enrolled infants was significantly lower in the B group (3.9%, 6/153) than in the P group (10.0%, 13/130; P < 0.05). No differences were observed in the incidence of other adverse outcomes including mortality. Conclusions: B. bifidum in VLBW infants accelerated the establishment of enteral feeding after birth without increasing the incidence of adverse effects.
Establishment of enteral feeding after birth is often delayed in preterm infants because of their immature intestinal function. These infants also have difficulty maintaining normal gut flora, which further prevents functional maturation of the intestine. [bib_ref] Bacterial colonization and gut development in preterm neonates, Cilieborg [/bib_ref] A relatively high incidence of cesarean delivery and prophylactic use of antibiotics after birth among preterm infants inhibits the activity of beneficial bacteria such as Lactobacilli and Bifidobacteria. [bib_ref] The intestinal bacterial colonisation in preterm infants: A review of the literature, Westerbeek [/bib_ref] Bacterial translocation from the gastrointestinal tract is an important pathway in the initiation of late-onset sepsis and necrotizing enterocolitis (NEC) in very low-birthweight (VLBW) infants. [bib_ref] New concepts of microbial translocation in the neonatal intestine: Mechanisms and prevention, Sherman [/bib_ref] The emerging intestinal microbiota, nascent intestinal epithelia, naive immunity, and suboptimal nutrition play a role in facilitating this bacterial translocation.
Probiotics have been shown, in many clinical trials, to promote establishment of normal gut flora and prevent diseases in preterm infants, [bib_ref] Probiotics for prevention of necrotizing enterocolitis in preterm infants, Alfaleh [/bib_ref] but few studies have examined the effect of probiotics on other benefits, such as support of enteral feeding and growth and prevention of sepsis. Here, we carried out a small-scale pilot study to evaluate the efficacy and safety of Bifidobacterium bifidum OLB6378 (B. bifidum) for VLBW infants prior to initiating a larger multicenter clinical trial, and found that the daily bodyweight gain was significantly higher in infants who received B. bifidum supplementation within 48 h of birth. [bib_ref] Effect of Bifidobacterium administration on very-low-birthweight infants, Yamasaki [/bib_ref] We subsequently conducted the present study on a larger group of preterm infants to determine the effect of B. bifidum within 48 h after birth on the establishment of enteral feeding in VLBW infants.
# Methods
## Design
This study was conducted as a cluster-randomized multicenter, double-blind, placebo-controlled trial. The institutional review board of Tokyo Women's Medical University approved the study protocol, under clinical trial registration number UMIN000002543.
## Patients
Nineteen neonatal intensive care units (NICU) in Japan, none of which had previously used probiotics, participated in the trial. In order to avoid cross contamination and infant-to-infant dissemination of B. bifidum within the same NICU, the 19 NICUs were divided into two groups. All participating hospitals were stratified according to patient volume in order to balance the number of patients between the groups. Pair-matched randomization of clusters was done using computer-generated random numbers, designated B and P. In the 10 B group NICUs, infants received B. bifidum, while infants in the nine P group NICUs received a placebo.
The study group consisted of VLBW infants (birthweight <1500 g) who were born or transferred within 24 h to the participating hospitals during the period from January 2010 to March 2011. Exclusion criteria were lack of parental consent or presence of major congenital malformation, systemic infection, or failure to give B. bifidum or placebo within 48 h due to a clinical condition that precluded oral feeding. Data from all infants who received at least one dose of B. bifidum or placebo were analyzed. Any infant who died or was transferred before the establishment of enteral feeding was designated as a failure of enteral feeding.
All clinical variables were registered through the website at each participating hospital into the neonatal research network database. [bib_ref] Morbidity and mortality of infants with very low birth weight in Japan:..., Kusuda [/bib_ref] No additional parameters specific to this study were added to the database.
## Interventions
Probiotic supplement, containing approximately 2.5 × 10 9 viable cells of B. bifidum/500 mg, was supplied as a freeze-dried powder in dextrin (Meiji, Tokyo, Japan), as previously reported. [bib_ref] Effect of Bifidobacterium administration on very-low-birthweight infants, Yamasaki [/bib_ref] The probiotic was approximately divided in half, and each portion was suspended in 0.5 mL warm water, breast milk, or infant formula. B. bifidum was then given to infants in B group through an enteral nutrition catheter within 48 h after birth. After the first dose, B. bifidum was given to infants twice a day until the bodyweight reached 2000 g. Placebo consisting of 500 mg dextrin was supplied, prepared, given to infants of P group in the same manner as described for the probiotic. Treatment and nutrition were otherwise given according to conventional practice in each NICU.
## Outcome measures
The primary outcome was establishment of enteral feeding, defined as the postnatal day at which the amount of enteral feeding exceeded 100 mL/(kg/day). Using the neonatal research network database value of 14 ± 4 postnatal days for mean age at establishment of enteral feeding in VLBW infants in 2003, [bib_ref] Morbidity and mortality of infants with very low birth weight in Japan:..., Kusuda [/bib_ref] we considered inability to reach an enteral feeding volume of 100 mL/(kg/day) within 21 days (mean + 2SD) as failure to achieve enteral feeding. Once enteral feeding was established, VLBW infants received either B. bifidum or placebo until the bodyweight reached 2000 g.
Secondary outcomes specifically evaluated in this study were length of hospital stay, bodyweight at discharge, bodyweight gain per day (g/day), head circumference at discharge (cm), and increase in head circumference/hospital days (cm).
Morbidity as recorded in the database was also analyzed as a secondary outcome in all enrolled infants as previously defined in the neonatal research network data base. [bib_ref] Morbidity and mortality of infants with very low birth weight in Japan:..., Kusuda [/bib_ref] Especially, NEC was diagnosed in cases of stage ≥I disease according to the Bell classification. Late-onset sepsis was defined as sepsis occurring ≥1 week after birth, proven on positive blood culture.
## Sample size calculation
The intracluster correlation coefficient was calculated using the mean age and standard deviation at which enteral feeding was established, expressed as the day feeding volume reached 100 mL/(kg/day) and was set at 0.02. [bib_ref] Morbidity and mortality of infants with very low birth weight in Japan:..., Kusuda [/bib_ref] We set a sample size of 19 hospitals in each of the two study arms before starting the trial, in order to achieve a statistical power of 80% (two sided = 0.05) to detect a decrease of 2 days to reach full feeding in the proportion of infants receiving B. bifidum during the stay in NICU. The resulting sample size was calculated as 77 infants for each arm.
# Statistical analysis
Student's t-test was used for all normally distributed data with equal variance, and results are expressed as mean ± SD. For data not conforming to normal distribution or equal variance, medians were compared using the Mann-Whitney test. The day at which enteral feeding was established in both groups was compared using the Kaplan-Meier method. Frequency data for mortality and morbidity were analyzed on chi-squared test. All statistical analysis was carried out using StatView (version 5.0; SAS Institute, Cary, NC, USA). P was adjusted using a design effect determined by average cluster size. Statistical significance was set at the 0.05 level.
# Results
# Background characteristics
Overall, 585 VLBW infants were admitted during the study period, of whom 298 were assigned to B group and 287 to P group. Of these, 145 and 157 infants were excluded from the B and P groups, respectively, due to lack of parental consent, congenital anomalies, systemic infection or failure to give B. bifidum or placebo within 48 h after birth for clinical reasons. Of the infants not enrolled in the study, almost 80% were not enrolled due to lack of parental consent within 48 h after birth. Finally, 283 VLBW infants were enrolled in the study: 153 in B group and 130 in P group. Among the enrolled infants, 19 in B group and 11 in P group died during their stay in NICU or were transferred to another hospital before establishment of enteral feeding, and were designated as failure to establish enteral feeding. Of the remaining infants, 233 (119 in B group and 114 in P group) began enteral feeding within 21 days after birth and received B. bifidum or placebo until their bodyweight reached 2000 g. These results are illustrated as a flow chart in . summarizes the background characteristics and outcomes of the enrolled infants. Because this study randomized the groups at the hospital level rather than the infant level, the power to detect a difference between groups was affected by the average number of infants in the cluster, known as the design effect. Thus, a design effect of 1.21 in this study was calculated according to an average sample size of 11.5 infants per hospital as described in the previous section. Further calculation of P was done by dividing both the number of participants and the number of events by this design effect.
No significant difference in gestational age or birthweight was detected between groups, but there was a significant difference between the groups in the rates of antenatal steroid and cesarean section. [fig_ref] Table 2: Outcome of establishment of enteral feeding before 21 days after birth Student's... [/fig_ref] summarizes clinical outcome in both study groups. The postnatal day at which the amount of enteral feeding exceeded 100 mL/(kg/day) was significantly earlier in B group infants compared with P group infants (11.0 ± 3.6 days vs [bib_ref] Up-regulation of polymeric immunoglobulin receptor expression by the heat-inactivated potential probiotic Bifidobacterium..., Nakamura [/bib_ref] [fig_ref] Fig 2: Day at which infants reached enteral feeding at 100 mL/kg/ day [/fig_ref] plots days at which infants in each group reached enteral feeding at a volume of 100 mL/(kg/day). The difference between the groups was significant (P < 0.05), most evident at approximately 10 days after birth.
## Primary outcome
## Secondary outcomes
No significant differences were observed between groups in terms of length of hospital stay, bodyweight at discharge, bodyweight gain per day, head circumference at discharge, or head circumference increase.
No difference was observed in the mortality and morbidities among the infants enrolled, except the incidence of late-onset sepsis . The incidence of sepsis ≥1 week after birth was significantly lower in B group (6/153, 3.9%) than in P group (13/130, 10.0%; P < 0.05).
# Subgroup analysis
The frequency of use of parenteral hyperalimentation was greater in P group (88.6%) than in B group (77.3%; P < 0.05). Because this difference might influence outcome, the VLBW infants who received parenteral hyperalimentation were re-evaluated. The postnatal day at the establishment of enteral feeding was signifi-cantly earlier in B group infants after re-evaluation (B group, 11.0 ± 3.4 days; P group, 12.4 ± 3.8 days; P < 0.05).
# Discussion
This study clearly shows the benefit of early use of B. bifidum on enteral feeding among VLBW infants. Early introduction of preterm infants to enteral feeding results in earlier achievement of full feeding without apparent increase in risk of morbidity including NEC. [bib_ref] Early or delayed enteral feeding for preterm growth-restricted infants: A randomized trial, Leaf [/bib_ref] Probiotics have been tested in clinical trials as a way to improve feeding tolerance in preterm infants. [bib_ref] Probiotics-supplemented feeding in extremely low-birth-weight infants, Al-Hosni [/bib_ref] [bib_ref] Early administration of Bifidobacterium breve to preterm infants: Randomised controlled trial, Kitajima [/bib_ref] [bib_ref] Effects of probiotics on enteric flora and feeding tolerance in preterm infants, Lee [/bib_ref] [bib_ref] Lack of effects of oral probiotics on growth and neurodevelopmental outcomes in..., Chou [/bib_ref] [bib_ref] A randomized placebo-controlled comparison of 2 prebiotic/probiotic combinations in preterm infants: Impact..., Underwood [/bib_ref] Some studies report that probiotics have beneficial effects in preterm infants, such as acceleration of enteral feeding and growth improvement. [bib_ref] Probiotics-supplemented feeding in extremely low-birth-weight infants, Al-Hosni [/bib_ref] [bib_ref] Early administration of Bifidobacterium breve to preterm infants: Randomised controlled trial, Kitajima [/bib_ref] [bib_ref] Effects of probiotics on enteric flora and feeding tolerance in preterm infants, Lee [/bib_ref] This trial was performed as a multicenter clinical trial, following a previous single-center study that showed enhancement of enteral feeding and growth in VLBW infants treated with B. bifidum. [bib_ref] Effect of Bifidobacterium administration on very-low-birthweight infants, Yamasaki [/bib_ref] The pilot study showed that probiotics given within 48 h after birth can very effectively colonize the immature bowel without increasing morbidity, [bib_ref] Effect of Bifidobacterium administration on very-low-birthweight infants, Yamasaki [/bib_ref] and supported the benefit of the early use of B. bifidum in enteral feeding. The pilot study was also performed in one hospital only, with all infants treated in a similar manner according to hospital practice. In contrast, the present study was performed in 19 hospitals, and treatment and nutrition protocols conformed to the methods of each NICU. This multicenter study was also performed as a cluster-randomized trial, taking into account how growth of the infant seemed to be affected by the nutrition method. Given that enrolled infants were followed up, a long-term growth benefit might be able to be observed.
Probiotics protect the immature intestine from NEC, an effect associated with reduction of inflammatory reaction in the ileum, regulation of the main components of the mucus layer, and improvement of intestinal integrity. [bib_ref] Bifidobacterium bifidum improves intestinal integrity in a rat model of necrotizing enterocolitis, Khailova [/bib_ref] The ability of probiotics to downregulate apoptosis in the rat NEC model and in intestinal epithelial cell line-6 cells appear to be a cyclo-oxygenase-2mediated phenomenon. [bib_ref] Bifidobacterium bifidum reduces apoptosis in the intestinal epithelium in necrotizing enterocolitis, Khailova [/bib_ref] Probiotics have also been shown to maintain integrity of the mucosal barrier, reducing its permeability, and to strengthen intestinal cell tight junctions. [bib_ref] Probiotic bacteria enhance murine and human intestinal epithelial barrier function, Madsen [/bib_ref] [bib_ref] The effect of a bifidobacter supplemented bovine milk on intestinal permeability of..., Stratiki [/bib_ref] B. bifidum may also influence the mucosal barrier, and support early enteral feeding, but this is not well studied in Japan due to the low incidence of NEC. [bib_ref] Trends in morbidity and mortality among very low birth weight infants from..., Kusuda [/bib_ref] Because we encountered no cases of NEC, the preventive effect of probiotics on this disease was not evaluated in the present study. There was a significant difference in the frequency of use of parenteral hyperalimentation in this study. Infants who receive hyperalimentation may be more likely to experience delayed enteral feeding, and might be more susceptible to any intestinal effects of B. bifidum. For this reason, we re-evaluated the VLBW infants who received parenteral hyperalimentation separately.
Bifidobacterium bifidum resulted in a significant reduction in late-onset sepsis, possibly related to its influence on the mucosal barrier. More importantly, an in vitro study showing upregulation of IgA and pIgR expression suggests that the probiotic may increase intestinal soluble IgA (sIgA). [bib_ref] Effect of Bifidobacterium bifidum OLB6377 and Bifidobacterium bifidum OLB6378 on expression of..., Nakamura [/bib_ref] [bib_ref] Up-regulation of polymeric immunoglobulin receptor expression by the heat-inactivated potential probiotic Bifidobacterium..., Nakamura [/bib_ref] sIgA was detectable in the early postnatal period in the saliva of preterm infants. [bib_ref] Detection of secretory immunoglobulin A (SIgA) in saliva of ventilated and non-ventilated..., Hayes [/bib_ref] [bib_ref] Immunoglobulins in saliva of preterm and full-term infants, Wan [/bib_ref] In addition, use of probiotics for preterm infants has been shown to upregulate transforming growth factor-β, which in turn upregulates mucosal IgA expression. [bib_ref] Bifidobacterium breve enhances transforming growth factor beta1 signaling by regulating Smad7 expression..., Fujii [/bib_ref] Dietary supplementation of preterm infants with probiotics starting early after birth has been shown to produce an increase in fecal IgA. [bib_ref] A fermented formula in pre-term infants: Clinical tolerance, gut microbiota, downregulation of..., Campeotto [/bib_ref] [bib_ref] Effects of Bifidobacterium lactis Bb12 supplementation on body weight, fecal pH, acetate,..., Mohan [/bib_ref] The increase in sIgA by B. bifidum may explain the observed decrease in late-onset sepsis. Reduction in bacterial translocation is an alternative mechanism by which probiotics help to maintain the integrity of the mucosal barrier.
The comparatively high number of B group infants failing to establish enteral feeding is a potential limitation of this study. Because we observed no adverse effect related directly to the use of B. bifidum or placebo, clinical conditions appear to be the only factors correlated with failed enteral feeding. In addition, mortality did not differ between the study groups, and was similar to what was previously reported. [bib_ref] Trends in morbidity and mortality among very low birth weight infants from..., Kusuda [/bib_ref] These findings suggest that B. bifidum may be used safely for VLBW infants.
Difference in antenatal steroid use was another concern of this study, because this might favor reduced morbidity in the B group. Given that this was a multicenter clustered clinical trial, adjustment of antenatal steroid usage equally between the groups was difficult. Nevertheless, the incidence of morbidity was similar in B and P group infants, suggesting that any such effect was not clinically significant.
# Conclusion
Use of B. bifidum in VLBW infants accelerated the establishment of enteral feeding without increasing morbidity. A significant decrease in late-onset sepsis was observed among all enrolled infants.
[fig] Fig 2: Day at which infants reached enteral feeding at 100 mL/kg/ day. (•), Bifidobacterium bifidum group (n = 119); (○) placebo group (n = 114). [/fig]
[table] Table 2: Outcome of establishment of enteral feeding before 21 days after birth Student's t-test and Kaplan-Meier method, ‡ Student's t-test; § Mann-Whitney test. B, Bifidobacterium bifidum. [/table]
|
Human entorhinal cortex represents visual space using a boundary-anchored grid
When participants performed a visual search task, fMRI responses in entorhinal cortex (EC) exhibited a 6-fold periodic modulation by gaze movement direction. The orientation of this modulation was determined by the shape and orientation of the bounded search space. These results indicate that human EC represents visual space using a boundary-anchored grid, analogous to that used to represent navigable space in rodents.
## Main text
During spatial navigation in rodents, grid cells fire when the body of the animal occupies a hexagonal lattice of spatial locations tiling the floor of the environment 1 . These cells are believed to support a metric for navigational space that is anchored to environmental boundaries [bib_ref] Grid cell symmetry is shaped by environmental geometry, Krupic [/bib_ref] [bib_ref] Shearing-induced asymmetry in entorhinal grid cells, Stensola [/bib_ref]. Recent work with monkeys has expanded the variety of spaces that might be represented by grid cells, by demonstrating the existence of neurons in EC that fire in a hexagonal lattice of positions on a screen while animals explore visual space [bib_ref] A map of visual space in the primate entorhinal cortex, Killian [/bib_ref]. However, it is currently unknown whether a similar grid-like coding of visual space exists in humans, or whether putative grid representations of visual space obey the same boundary-anchoring principles as grid representations of navigational space.
To address these issues, we used fMRI methods previously developed for identifying grid signals in humans during virtual navigation [bib_ref] Evidence for grid cells in a human memory network, Doeller [/bib_ref]. These methods are motivated by the observation that firing patterns for grid cells within an individual tend to have the same orientation [bib_ref] Microstructure of a spatial map in the entorhinal cortex, Hafting [/bib_ref] [bib_ref] Evidence for grid cells in a human memory network, Doeller [/bib_ref]. Because of this common orientation, movements along a shared grid axis will yield a stronger grid-driven fMRI signal in EC than movements between grid axes, resulting in 60° periodic modulation by movement direction. We reasoned that if grid cells represent visual space in humans, then we should observe a similar 60° periodic fMRI signal as a function of gaze movement direction while participants visually explored the environment.
To test this idea, participants (n=36) were scanned with fMRI and had their gaze tracked while they performed an unconstrained visual search task in which they had to find a target letter ('L') among numerous distractors letters ('T's) [fig_ref] Figure 1: Visual grid-like representation in human entorhinal cortex [/fig_ref]. A square border surrounded the search display for half the participants (n=18) and a rectangular border surrounded the display for the other half (n=18).
For each participant, we split the fMRI data into halves, identified the orientation of the 60° periodic signal as a function of gaze movement direction within EC in one half of the data, and tested the reliability of this visual grid orientation in the independent second half [fig_ref] Figure 1: Visual grid-like representation in human entorhinal cortex [/fig_ref]. Consistent with our prediction, we observed significant reliable 6fold modulation of the fMRI signal as a function of gaze movement direction bilaterally in EC [fig_ref] Figure 1: Visual grid-like representation in human entorhinal cortex [/fig_ref]. This result reflects greater fMRI response when gaze movement directions were aligned with the three grid axes than misaligned [fig_ref] Figure 1: Visual grid-like representation in human entorhinal cortex [/fig_ref]. Conducting the same analyses for other rotational symmetries, we found no evidence of reliable 90° or 45° periodic signals across independent halves of the data in EC [fig_ref] Figure 1: Visual grid-like representation in human entorhinal cortex [/fig_ref]. Notably, across participants, the magnitude of the 60° periodic EC signal significantly correlated with self-reported navigational ability, suggesting that the same population of grid cells might support both vision and navigation [fig_ref] Figure 2: Visual grid orientation is anchored to the search display geometry [/fig_ref]. All gaze movement directions were sampled during the visual search task, and we detected no 6-fold biases in gaze behavior that could explain the presence of a 6-fold symmetric fMRI signal [fig_ref] Figure 3: Visual grid orientation rotates in concert with rotation of the search display [/fig_ref]. Thus these results are evidence of a grid representation in human EC that codes for locations in visual space, complementing previous findings of grid representations in navigable space [bib_ref] Evidence for grid cells in a human memory network, Doeller [/bib_ref] [bib_ref] Direct recordings of grid-like neuronal activity in human spatial navigation, Jacobs [/bib_ref]. Grid-like coding of visual space was also observed in a medial prefrontal region of interest previously reported to exhibit a grid-like response during navigation 5 .
We next explored the coordinate system that EC uses to encode visual space. For grid cells to provide useful information about environmental locations, grid cell firing patterns must be stably anchored to features of the external world, such as environmental boundaries. In previous work examining grid cells tiling visual space in monkey EC, the head of the animal was fixed relative to the visual display, making it difficult to determine whether these cells coded locations in egocentric (head-centered) or allocentric (world-centered) coordinates. To address this issue, we tested whether EC grid representations of visual space exhibit two signatures of boundary-anchored coding previously observed in rodent grid cells.
First, we asked whether grid orientations are reliably aligned by search display shape. When rodents explore square environments, the grid lattice aligns to ±7.5° from the cardinal axes of the borders 2, 3 [fig_ref] Figure 2: Visual grid orientation is anchored to the search display geometry [/fig_ref]. We looked for a similar effect in the participants who searched square displays [fig_ref] Figure 2: Visual grid orientation is anchored to the search display geometry [/fig_ref]. Across these participants, the average EC visual grid orientations were significantly clustered around ±7.5° offset from the cardinal axes of the square display border [fig_ref] Figure 2: Visual grid orientation is anchored to the search display geometry [/fig_ref]. Moreover, of the 14/18 participants who showed significant clustering of grid angles across voxels in bilateral EC, 12/14 had grid angles significantly clustered across voxels 6°-9° offset from the display borders [fig_ref] Figure 2: Visual grid orientation is anchored to the search display geometry [/fig_ref]. By contrast, visual grid orientations for the rectangular display participants were not clustered around ±7.5° from the rectangular borders across participants [fig_ref] Figure 2: Visual grid orientation is anchored to the search display geometry [/fig_ref]. Indeed, grid orientations were closer to 7.5° offset from the display borders in the square display participants than in the rectangular display participants (t-test, t(34)=2.26, p=0.015, one-tailed). Because the shape of the display was the only stable environmental feature that differed between these participants, these results confirm that visual grid orientations were affected by the geometry of the visual environment.
Second, we examined whether rotation of the search display would induce a corresponding rotation of the visual grid. To address this question, each participant who performed the search task with the upright rectangular search displays also completed two additional scan runs in which the displays were rotated 30° clockwise [fig_ref] Figure 3: Visual grid orientation rotates in concert with rotation of the search display [/fig_ref]. If the visual grid code is anchored to the borders of the search display, then rotation of the search display should yield a corresponding 30° rotation of the visual grid orientation, as observed in navigating rodents when chamber boundaries are rotated 2 . We found that the fMRI signal in EC during rotated runs was better predicted by a grid angle that was rotated 30° relative to the upright-displayfit grid orientation than it was by a grid angle that was not rotated [fig_ref] Figure 3: Visual grid orientation rotates in concert with rotation of the search display [/fig_ref] ; . Furthermore, the average grid orientation during rotated runs was offset 28.33°± 2.87° (mean angle ± s.e.m.) relative to the grid orientation during upright runs [fig_ref] Figure 3: Visual grid orientation rotates in concert with rotation of the search display [/fig_ref]. Interesting, 6 participants showed little grid angle rotation [fig_ref] Figure 3: Visual grid orientation rotates in concert with rotation of the search display [/fig_ref]. Surprisingly, these nonrotating participants were faster at finding the target letter during the rotated-display runs than the participants whose grid orientations rotated [fig_ref] Figure 3: Visual grid orientation rotates in concert with rotation of the search display [/fig_ref]. Thus, although visual grids were anchored to the borders of the search display on average, there were individual differences in which external reference frame was selected, and these differences had consequences for search behavior.
In sum, we report the first evidence that human EC represents locations in visual space using a grid code. This visual grid code exhibited two signatures of boundary-anchoring previously observed in rodent grid cells-alignment to boundaries based on the shape of the environment, and rotation when the environmental borders are rotated-indicating that similar computational principles anchor primate and rodent grid cells to the external world, even across different spatial domains. These results may illuminate a longstanding controversy over the representation of visual space. Previous evidence suggests that the mammalian visual system represents space in retinotopic coordinates [bib_ref] Evidence against visual integration across saccadic eye movements, Irwin [/bib_ref] [bib_ref] Maps of visual space in human occipital cortex are retinotopic, not spatiotopic, Gardner [/bib_ref] [bib_ref] Higher level visual cortex represents retinotopic, not spatiotopic, object location, Golomb [/bib_ref] , which are updated before each eye movement based on information about the intended direction of the upcoming saccade [bib_ref] The updating of the representation of visual space in parietal cortex by..., Duhamel [/bib_ref]. Although non-retinotopic spatial codes are observed under some circumstances [bib_ref] Parietal neurons encoding spatial locations in craniotopic coordinates, Galletti [/bib_ref] [bib_ref] Separate body-and world-referenced representations of visual space in parietal cortex, Snyder [/bib_ref] , it is often unclear whether these codes are egocentric (head-centered) or allocentric (world-centered), and evidence for an allocentric map that represents where a viewer is looking relative to stable visual environmental cues has remained sparse (although see refs. [bib_ref] Allocentric spatial referencing of neuronal activity in macaque posterior cingulate cortex, Dean [/bib_ref] [bib_ref] Spatial view cells and the representation of place in the primate hippocampus, Rolls [/bib_ref]. The current results provide evidence for such a map and suggest a mechanism by which it might be generated. During navigation, grid cells are thought to perform path integration by using self-motion inputs 1, 15 to update allocentric representations of location 2, 3 . We hypothesize that visual grid cells may use a similar path integration mechanism to update an allocentric representation of the current gaze position based on eye motion signals present in the hippocampus and EC [bib_ref] Saccade direction encoding in the primate entorhinal cortex during visual exploration, Killian [/bib_ref]. Beyond navigation, recent work has also shown that a grid-like code is used to represent both imagined and conceptual spaces [bib_ref] Grid-like processing of imagined navigation, Horner [/bib_ref] [bib_ref] Organizing conceptual knowledge in humans with a gridlike code, Constantinescu [/bib_ref] [bib_ref] Grid-cell representations in mental simulation, Bellmund [/bib_ref]. Our data add to this growing body of work by showing that grid cells may provide the mechanism by which locations in visual space are coded, thus allowing us to form durable visuospatial representations that are stable across eye movements.
# Online methods
Participants 36 participants (14 male) took part in this experiment (mean age: 23; range: 18-32). All participants gave written consent and were paid for participating, in compliance with procedures approved by the University of Pennsylvania Internal Review Board. All had normal vision and reported to be in good health with no history of neurological disease. Data from 7 additional participants were collected but discarded before analysis of fMRI data due to poor eye tracking quality (6 because of inaccurate gaze reconstructions; one because of poor sampling of all gaze angles). Data from one additional participant was discarded due to excessive head motion during scanning (>3 mm average absolute head motion). Following scanning, each participant completed the Santa Barbara Sense of Direction (SBSOD) questionnaire [bib_ref] Development of a self-report measure of environmental spatial ability, Hegarty [/bib_ref] , which provides a standardized measure of self-reported navigational ability.
## Visual search task
Participants completed a series of 6.5 min fMRI scan runs during which they performed a visual search task. Square display participants completed four runs and rectangular display participants completed six runs. Participants were randomly assigned to display shape groups. During each run, participants viewed visual search displays consisting of a single target letter 'L' shown amongst distractor letters 'T's (letter height = 0.74°). Participants were instructed to use their eyes to search for the target, and to press a button when they found the target letter. Each trial was self-paced, and lasted an average of 7.50±0.58 seconds (mean±s.e.m.). Stimuli were presented using Matlab (2016a, The MathWorks Inc., Massachusetts) and the Psychtoolbox 22 (Version 3.0.11). A pseudo-random search display was generated on each trial, such that all letters had a random location within the borders of the search display shape, subject to the constraint that only partial overlap between the letters was permitted, and a random orientation. Note that this meant that the shape implied by the array of letters was the same as the shape defined by the drawn border. Each search display had one of three possible densities or [81, 100, 121] letters total in the square and rectangle conditions, respectively). The search display density was randomly selected on each trial, with the constraint that each of the three possible densities was presented once before repeating. Search displays subtended a visual angle of 17.0° × 17.0° (square participants) or 11.0° × 17.0° (rectangular participants), and the search display border line thickness was 0.21°. There was a variable inter-trial interval of 2-6 seconds, randomly selected on each trial, during which participants fixated on a centrally located fixation cross. The onset of each trial was time-locked to the onset of an fMRI acquisition.
For the rectangular display participants, four scan runs consisted of upright rectangular displays, and two runs consisted of rectangular displays rotated by 30° clockwise. For these participants, the presentation order of the displays was URUURU, where U and R correspond to upright and rotated displays, respectively. This ordering ensured that any effect of display rotation could not be due to general drift across runs.
## Eye tracking methods and preprocessing
Participant's gaze position during scanning was monitored and recorded using a LiveTrack AV MR-compatible eye tracking camera (Cambridge Research Systems, Rochester England). The gaze position of the right eye was recorded at 30 Hz. Prior to each scan run, gaze position was calibrated using a series of nine fixation points evenly spaced between −8° and +8° in the horizontal and vertical dimensions relative to screen center. The average calibration error across all runs was 0.332° ± 0.018° (mean ± s.e.m.). In order to separate periods of gaze movements from periods of fixations, periods of gaze movement were defined by a movement-velocity-thresholding procedure, as follows. To reduce misattribution of gaze movements to eye tracking noise, the gaze position time course was first temporally smoothed with a boxcar filter (half width = 0.185 seconds). Gaze movements were then identified based on a median split of the smoothed gaze movement instantaneous velocity. Gaze position measurements in the bottom half of gaze movement velocities were treated as no movement, as were samples during which participants blinked. Gaze movements with velocities in the upper median half tended to be long saccades relative to the size of the search displays, with an average ballistic gaze trajectory length of 1.60° ± 0.60° (mean ± standard deviation). Note that this velocity-thresholding procedure is conservative in that it excludes short gaze movements during which we would not expect to observe a strong grid-like fMRI signal, based on previous fMRI studies of human navigation [bib_ref] Evidence for grid cells in a human memory network, Doeller [/bib_ref] [bib_ref] Grid-like processing of imagined navigation, Horner [/bib_ref]. Based on this method of classifying gaze movements, 7.1% ± 0.57% (mean ± s.e.m.) of all fMRI acquisitions contained no gaze movements whatsoever for the entire duration of the acquisition, which served as the implicit baseline relative to which fMRI signal change was measured.
## Fmri acquisition
Scanning was performed at the Center for Functional Imaging at the University of Pennsylvania using a 3T Siemens Prisma scanner equipped with a 64-channel head coil. High-resolution T1-weighted images for anatomical localization were acquired using a 3dimensional magnetization-prepared rapid-acquisition gradient-echo pulse sequence (repetition time [TR], 1620 ms; echo time [TE], 3.09 ms; inversion time, 950 ms; voxel size, 1x1x1 mm; matrix size, 192x256x160). T2*-weighted images sensitive to blood oxygenation level-dependent contrasts were acquired using a gradient-echo echoplanar pulse sequence (TR, 1000 ms; TE, 25 ms; flip angle 45°; voxel size, 2x2x2 mm; field of view, 192; matrix size, 96x96x78; multiband acceleration factor of 4). Ten additional fMRI volumes were also collected at the start of each scan run that were excluded from data analysis to account for signal steady-state transition. Visual stimuli were displayed at the rear bore face on an InVivo SensaVue Flat Panel Screen at 1920 × 1080 pixel resolution (diag = 80.0 cm, w × h = 69.7 × 39.2 cm). Participants viewed the stimuli through a mirror attached to the head coil. Behavioral responses were collected using a fiber-optic button box. fMRI analysis -preprocessing FMRI data analysis was carried out using FSL FEAT (FMRIB's Software Library, version 6.00, www.fmrib.ox.ac.uk/fsl). The following standard data preprocessing was performed: motion correction using MCFLIRT 23 , non-brain removal using BET [bib_ref] Fast robust automated brain extraction, Smith [/bib_ref] ; spatial smoothing using a Gaussian kernel of FWHM 8mm; grand-mean intensity normalization of the 4D dataset by a single multiplicative factor for each scan run; highpass temporal filtering (Gaussian-weighted least-squares straight line fitting, with sigma=50.0s). For second-level group analyses, EPI images were registered to the high-resolution anatomical image using boundary-based reconstruction and then normalized into standard space (MNI305) using non-linear registration. All data normalization was performed using Freesurfer (version 5.3.0, http://surfer.nmr.mgh.harvard.edu/).
## Fmri analysis -identifying grid-like coding of visual space
We performed a split-half analysis to estimate the orientation of the visual grid code during periods of gaze movement, following procedures used previously to identify grid-like codes during virtual navigation [bib_ref] Evidence for grid cells in a human memory network, Doeller [/bib_ref] [bib_ref] Grid-like processing of imagined navigation, Horner [/bib_ref] [bib_ref] Reduced grid-cell-like representations in adults at genetic risk for Alzheimer's disease, Kunz [/bib_ref] [bib_ref] The GridCAT: A toolbox for automated analysis of human grid cell codes..., Stangl [/bib_ref] [fig_ref] Figure 1: Visual grid-like representation in human entorhinal cortex [/fig_ref]. Data were first split into halves by run (runs [bib_ref] Grid cell symmetry is shaped by environmental geometry, Krupic [/bib_ref] [bib_ref] A map of visual space in the primate entorhinal cortex, Killian [/bib_ref] and [bib_ref] Microstructure of a spatial map in the entorhinal cortex, Hafting [/bib_ref] [bib_ref] Shearing-induced asymmetry in entorhinal grid cells, Stensola [/bib_ref] for square participants; runs [bib_ref] Microstructure of a spatial map in the entorhinal cortex, Hafting [/bib_ref] [bib_ref] A map of visual space in the primate entorhinal cortex, Killian [/bib_ref] and [bib_ref] Shearing-induced asymmetry in entorhinal grid cells, Stensola [/bib_ref] [bib_ref] Direct recordings of grid-like neuronal activity in human spatial navigation, Jacobs [/bib_ref] for rectangle participants, so that only the upright rectangle runs were used to identify grid-like coding in this initial analysis). For each half of the data, we identified the angular orientation of the putative visual grid axes in each participant's bilateral EC. The grid orientation thus obtained was then subsequently used to predict a grid signal during the other independent half of the runs.
To fit the orientation of the 6-fold gaze movement direction-modulated signal within EC, we constructed a general linear model (GLM) with two parametric modulators (PMs) for periods of gaze movement. These two PMs were cos(6a(t)) and sin(6a(t)), where a(t) is the gaze movement direction sampled at time t (30 Hz). Each PM was down-sampled to the TR (1 Hz) by summing the values of the PM within each TR. The weights (b1 and b2) on these PMs were fitted to the fMRI time series for each voxel within the anatomically defined bilateral EC ROI. This EC ROI was constructed uniquely for each participant based on the automatic anatomical parcellation of the EC derived from FreeSurfer structural reconstruction. We then calculated the orientation of the 6-fold gaze movement directionmodulation from the mean weights across all voxels in the EC ROI as φ = [arctan(<b2>/ <b1>)]/6, separately for each run, where arctan was mapped into 360° space, varying between 0° and 60°, according to signs of b2 and b1. Finally we computed the circular average orientation across runs for each separate half of the data. Grid orientations were quantitatively similar if we averaged the beta weights voxel-wise across runs before calculating the orientation instead of averaging the orientations across runs (circular correlation across participants: c=0.84, p<10 −5 ).
To test whether the fit orientations predicted the analogous 6-fold periodic signal in the other independent half of the data, we constructed a GLM with a PM modeling the effect of gaze movement direction on the fMRI signal. The value of this PM at each timepoint was the cosine of gaze movement direction at that timepoint aligned to the orientation predicted by the first half of the data, cos(6(a(t)-φ)), where a(t) is the gaze movement direction sampled at time t (30 Hz). This PM was down sampled to the TR by summing the values of the PM within each TR. Each beta from this analysis reflects the extent of reliable split-half φoriented 6-fold gaze movement direction-modulated fMRI signal (which we call "orientation consistency"). The beta weights for this PM were averaged across all scan runs within each participant. The group-level test of the significance of these weights was small-volume FWE-corrected (SVC) within a group-level bilateral EC ROI, which was defined as the union of all individual-participant anatomically-defined EC ROIs projected into MNI space. To confirm that the gaze movement direction-modulated signal in EC exhibited a specifically 60° periodicity, we conducted this same split-half analysis for 90° (i.e., 4-fold) and 45° (i.e., 8-fold) periodicities.
Analyses were performed in FSL using FILM with local autocorrelation correction [bib_ref] Temporal autocorrelation in univariate linear modeling of FMRI data, Woolrich [/bib_ref]. Included in all GLMs was a binary boxcar regressor of no interest corresponding to periods of visual search and its temporal derivative, as well as six nuisance PMs to account for head motion-related artifacts. All regressors were convolved with double gamma hemodynamic response function and filtered by the same high pass filter as the fMRI data before entry into the GLM.
## Fmri analysis -reliable offset of the grid-like representation orientation from the search display shape
To test whether grid orientations across participants consistently cluster around an offset of ±7.5° from the cardinal axes of the search display borders, we first computed the average of the grid orientations across all EC voxels and runs within each participant. We then folded the grid orientations of all participants by φ mod 15°, which would align all hypothesisconsistent alignments to 7.5° in a circular 0° to 15° space. Next we performed a V-test for nonuniformity centered around 7.5°. The V-test is similar to Rayleigh's test for circular uniformity with the difference that under the alternative hypothesis the distribution is nonuniform centered at a particular hypothesized angle (in this case, 7.5°) [bib_ref] CircStat: a MATLAB toolbox for circular statistics, Berens [/bib_ref].
To test whether grid orientations in voxels within individual participants clustered around 7.5° offset from cardinal axes, we first evaluated whether EC voxels in each participant showed orientation clustering around any angle. To do so, we averaged the grid angle derived from each voxel across runs, yielding a distribution of voxel-wise grid orientations. Next we tested these voxel-wise grid orientations for non-uniformity using Rayleigh's test for circular data. Note that because grid orientations were averaged voxel-wise across fMRI runs for this analysis, significant orientation clustering also requires temporal stability across runs of the grid orientation across voxels. This analysis identified participants who had significant nonuniformity, i.e. orientation clustering, of grid angles in EC (p<0.05, accounting for spatial smoothing). Finally, we tested whether the voxel-wise grid orientations in participants with significant clustering were specifically clustered around 6.0°-9.0° degrees in 0.5° increments (via folding and V-test, as described above; p-values were Bonferonni corrected for the seven grid angles tested).
## Fmri analysis -rotation of the grid-like representation orientation with rotation of the search display
To test whether the visual grid orientations of rectangular-display participants rotated in concert with the rotated displays, we first computed the circular average of the grid orientation derived from each upright-display run. Next we rotated this average grid orientation, φ, by 30° and used this rotated orientation to predict the fMRI signal during rotated-display runs with a GLM. Specifically, a single PM to was used to model the effect of gaze movement direction on the fMRI signal during the rotated rectangle runs: a cosine of gaze movement angle aligned to the 30° rotated grid orientation, cos(6(a(t)-(φ+30°))). Positive weights from this analysis indicate that the 60° periodic fMRI signal is better predicted when the orientation of the grid axes is rotated 30° from φ during rotated-display runs, whereas negative weights indicate that rotated display runs are better predicted by the original grid orientation φ without rotation. The weights for this PM were first combined across both rotated-display runs in each participant, and then tested across participants with small-volume FWE-correction within the group-level bilateral EC ROI.
To examine the distribution of rotation effects across EC voxels, we first identified the grid orientation for each voxel during the rotated scan runs in the same fashion as we did for the upright scan runs. We then compared the distribution of grid orientations across all EC voxels when the display was upright to the distribution when the display was rotated. Specifically, for each participant, we subtracted φ from each EC voxel's grid orientation, separately for the upright and rotated display runs, so that the average grid orientation across voxels were aligned relative to φ in each participant. We then calculated the distribution of voxels with grid orientations occurring from 0°-60° in 2° increments separately for the upright and rotated scan runs, and subtracted the upright distribution from the rotated distribution. If grid orientations across voxels rotate in concert with rotation of the search display, then there should be a higher percentage of voxels with grid angles around φ+30° when the display is rotated than upright.
## Statistics
No statistical methods were used to pre-determine sample sizes, but our sample size was similar to those reported in previous publications [bib_ref] Grid-like processing of imagined navigation, Horner [/bib_ref] [bib_ref] Organizing conceptual knowledge in humans with a gridlike code, Constantinescu [/bib_ref] [bib_ref] Grid-cell representations in mental simulation, Bellmund [/bib_ref]. Parametric t-tests and non-parametric sign-tests were used throughout the paper. For each parametric test, unless otherwise noted, data values met normality assumptions (Lilliefors test, p>0.05). If data did not meet normality assumptions, only sign-tests are reported. Rayleigh's tests and V-tests were also used, as described in detail in the two preceding fMRI analysis methods sections. Data collection and analysis were not performed blind to the conditions of the experiment. A Life Sciences Reporting Summary is available.
## Code availability
The code that support the findings of this study are available from the corresponding authors upon request.
## Data availability
The data that support the findings of this study are available from the corresponding authors upon request.
# Supplementary material
Refer to Web version on PubMed Central for supplementary material. a) Left: example square visual search display (for display purposes, example display has fewer letters than actual displays and relative letter size is increased). Right: schematic of the scene visible during scanning. b) Reliable grid-like coding of visual space was observed in bilateral EC (t-test, p<0.05, SVC in bilateral EC; peak MNI coordinates: 40/−4/−38, peak Z=3.09). c) fMRI response in a 2mm sphere centered on the peak EC voxel from (b) for periods of gaze movement aligned to grid orientation φ (within ±15° of a φ axis) and misaligned (more than ± 15° from all φ axes) (aligned: t-test, t(35)=1.95, p=0.030, sign-test p=0.033; misaligned: t-test, t(35)=−2.60, p=0.014, sign-test p=0.029, two-tailed tests). d) Split-half orientation consistency (beta weight) in the spherical EC voxel ROI from (c) for 90° and 45° periodicities (magnitude of 60° plotted for scale). Neither 90° nor 45° showed significant orientation consistency (90°: t-test, t(35)=−1.15, p=0.87, sign-test p=0.56; 45°: ttest, t(35)=−1.02, p=0.84, sign-test p=0.93). Note that these null effects were not specific to the EC ROI based on the 60° periodicity analysis, as we saw no effect for 90° or 45° in the entire EC at p<0.05 (SVC). Throughout the figure, all statistical tests are one-tailed unless otherwise noted, and n=36 participants; error bars show ±1 SEM; *p<0.05; ***p<0.001; ns, not significant. a) The grid orientations of rodents navigating through square environments align to an offset of 7.5° perpendicular to an environment wall (example cell shown is adapted with permission from ref. [bib_ref] Shearing-induced asymmetry in entorhinal grid cells, Stensola [/bib_ref]. b) We tested whether the visual grid orientation φ was similarly offset 7.5° from the square display borders. Specifically, because the possible range of φ is between 0°-60°, we examined whether grid orientations cluster around 4 possible angles, each 7.5° from one of the two cardinal axes of the display borders. c) Grid orientations of the square display participants. Left: average grid angle in each participant (blue squares), on the range of 0°-60°; Middle: histogram of average grid orientations across participants modulo 15°, showing clustering around 7.5° (n=18; V-test, V=5.18, p=0.0421); Right: average percentage of grid orientations modulo 15° ±1 SEM across all bilateral EC voxels. The average grid angle and standard error reported above the middle histogram were computed in circular space. d) Polar histograms of all EC voxel grid orientations for two exemplary square display participants. Note clustering of grid orientations around ±7.5° from the display border (n=285, 289 voxels; V-test, V=180.42, 205.54, p=3x10 −7 , 5x10 −9 , Bonferonni corrected). e) Grid orientations of the rectangular display participants (red rectangles), organized as in (c). Across rectangular display participants, grid orientations were not clustered around 7.5° (n=18; V-test, V=−2.48, p=0.796). a) For the rectangular display participants, the search displays were rotated 30° clockwise during two scan runs. b) For the rotated-display-runs, we tested whether a 30° rotation of the grid orientation φ identified in the upright-display-runs (φ+30°) better predicted the fMRI signal than the original grid orientation (φ). In left EC, 60° periodic sinusoidal modulation of the fMRI signal was significantly greater when gaze movements were aligned to φ+30° than when they were aligned to φ (n=18 participants; t-test, p < 0.05 two-tailed, SVC in bilateral EC; peak MNI coordinates: −24/−5/−44, peak Z score = 3.32). No right hemisphere EC voxels survived SVC, and no EC voxels showed greater modulation for φ than for φ+30°. c) The percent difference in the distributions of grid orientations across all bilateral EC voxels during the rotated-display-runs compared to the upright-display-runs in each participant (bottom row) and on average ±1 SEM (top row). Across participants (n=18), there was a greater percentage of voxels with grid orientations around φ+30° when the display was rotated than upright (t-test, t(17)=2.89, p=0.005, one-tailed). d) Difference in average grid angles between the upright-display-runs and the rotated-display-runs in each participant. e) Comparison of response latency between subjects whose grid angles rotated with the display (Rot; n=12; dark rectangles in (d); >15° absolute upright vs. rotated difference) and those subjects whose grid angles remained fixed relative to an alternate reference frame (Nonrot; n=6; light rectangles in (d); <15° absolute difference) (t-test, t(16)=3.81, p=0.002, twotailed). **p<0.01.
[fig] Figure 1: Visual grid-like representation in human entorhinal cortex (EC) [/fig]
[fig] Figure 2: Visual grid orientation is anchored to the search display geometry [/fig]
[fig] Figure 3: Visual grid orientation rotates in concert with rotation of the search display [/fig]
|
An overview of graphene-based hydroxyapatite composites for orthopedic applications
a b s t r a c tHydroxyapatite (HA) is an attractive bioceramic for hard tissue repair and regeneration due to its physicochemical similarities to natural apatite. However, its low fracture toughness, poor tensile strength and weak wear resistance become major obstacles for potential clinical applications. One promising method to tackle with these problems is exploiting graphene and its derivatives (graphene oxide and reduced graphene oxide) as nanoscale reinforcement fillers to fabricate graphene-based hydroxyapatite composites in the form of powders, coatings and scaffolds. The last few years witnessed increasing numbers of studies on the preparation, mechanical and biological evaluations of these novel materials. Herein, various preparation techniques, mechanical behaviors and toughen mechanism, the in vitro/ in vivo biocompatible analysis, antibacterial properties of the graphene-based HA composites are presented in this review.
# Introduction
Bone tissues consist of organic and inorganic components, with self-healing ability and great capability to withstand mechanical loading. Fixation of bone fractures and non-unions, correction of spinal deformities and replacement of arthritic joints [bib_ref] The future of biologic coatings for orthopaedic implants, Goodman [/bib_ref] are major unmet clinical needs. Traditionally, biological approaches for bone repair involve using autografts and allografts of cancellous bone [bib_ref] Biomaterial developments for bone tissue engineering, Burg [/bib_ref]. Nowadays, calcium phosphate ceramics and bioactive glasses are introduced as promising osteoinductive and osteoconductive substitutes for large orthopedic defect remolding or regeneration [bib_ref] Bone reconstruction: from bioceramics to tissue engineering, El-Ghannam [/bib_ref]. In addition, these bioceramics are also utilized as coating on metallic implants to provide long-term performance of the devices and to minimize micromotion between bones and implants during physiologic loading [bib_ref] The future of biologic coatings for orthopaedic implants, Goodman [/bib_ref] [bib_ref] Functional coatings or films for hard-tissue applications, Wang [/bib_ref].
Hydroxyapatite (HA, Ca 10 (OH) 2 (PO 4 ) [bib_ref] Comparative investigation on the adhesion of hydroxyapatite coating on Tie6Ale4V implant: a..., Mohseni [/bib_ref] possesses chemical and crystallographic similarities to inorganic components of the bone matrix and the teeth [bib_ref] Nanoscale hydroxyapatite particles for bone tissue engineering, Zhou [/bib_ref] with excellent osteoconductivity and osteoinductivity. It has been clinically used as bioactive coatings on dental and orthopedic implants, enabling the adhesion and proliferation of osteoblast cells on the prosthetic surface, and resulting in biological fixation between bone tissues and the implant [bib_ref] Comparative investigation on the adhesion of hydroxyapatite coating on Tie6Ale4V implant: a..., Mohseni [/bib_ref]. However, one primary limitation, when used under major load bearing, is its poor mechanical properties, such as low fracture toughness and tensile strength [bib_ref] Nanoscale hydroxyapatite particles for bone tissue engineering, Zhou [/bib_ref]. To address this problem, specific reinforcing materials, such as carbon nanotubes [bib_ref] Carbon nanotube toughened hydroxyapatite by spark plasma sintering: microstructural evolution and multiscale..., Lahiri [/bib_ref] , polyethylene [bib_ref] Processing and mechanical properties of HA/ UHMWPE nanocomposites, Fang [/bib_ref] , Al 2 O 3 and TiO 2 [bib_ref] The influence of titania/hydroxyapatite composite coatings on in vitro osteoblasts behaviour, Ramires [/bib_ref] , are typically used to prepare HA composites with increased mechanical properties, but these materials result in significantly less bioactivity than that of pure HA.
Graphene is rapidly rising as a promising material for biomedical applications, featuring two-dimensional nanosheet of hexagonally bonded carbon atoms, with large surface area, high conductivity, strong mechanical properties and good
## Preparation of composites
Various preparation techniques could be exploited to make this interesting composites, which is summarized in [fig_ref] Figure 2: The summary of different preparation methods of the graphene-based composites [/fig_ref]. In most cases, the composite prepared under high temperature or high pressures have high crystallinity and mechanical properties, such as hydrothermal synthesis, spark plasma sintering and hot isostatic sintering. However, thermal spraying techniques usually lower the crystallinity of the HA coating. HA could be synthesized onto graphene and its derivatives and be directly mixed with these nanofillers by ultrasonic dispersion and ball milling.
2.1. Graphene/HA composite powder 2.1.1. In situ synthesis Nano HA particles are successfully fabricated on GO [bib_ref] In situ synthesis and biocompatibility of nano hydroxyapatite on pristine and chitosan..., Li [/bib_ref] , chitosan functionalized GO [bib_ref] In situ synthesis and biocompatibility of nano hydroxyapatite on pristine and chitosan..., Li [/bib_ref] and rGOsurfaces using in situ synthesis methods. Usually, as shown in, graphene-based powders are first dissolved and exfoliated in DI water by ultrasonic dispersion to obtain a uniform solution; then Ca(NO 3 ) 2 is added into the graphene-based solutions by stirring for a desired time; afterwards, the pH of the suspension is adjusted to 9e10 using ammonia water, and (NH 4 ) 2 HPO 4 was added dropwise into the mixture [bib_ref] In situ synthesis and biocompatibility of nano hydroxyapatite on pristine and chitosan..., Li [/bib_ref]. The resulting composite solutions are recommended to be aged for days to ensure the fully transformation of apatite into hydroxyapatite with good phase purity and well crystallinity. During the synthesis, the oxygen-containing functional groups on GO surfaces behave as receptor sites for Ca 2þ through electrostatic interactions; these anchored cations can in situ react with the phosphate ions to obtain apatite nanoparticles. The underlying reaction mechanism has been proposed and discussed by Li et al. [bib_ref] In situ synthesis and biocompatibility of nano hydroxyapatite on pristine and chitosan..., Li [/bib_ref] ; the distribution and the microstructures of HA on graphene are mainly influenced by (1) the amounts and types of the oxygenous groups on the graphene-based templates and (2) the concentration of the reagents (Ca 2þ and HPO 4 2À ), solution pH values and so on. Besides, Ca(OH) 2 and H 3 PO 4 are also utilized by Gururaj et al. to in situ deposit HA on rGO nanosheets.
Composite, prepared in this method, is expected to increase the interfacial bonding strength between graphene and HA, facilitating the stress transfer from the matrix to the graphene-based nanofillers. This facile approach is economical and can be industrially mass-produced.
## Biomimetic mineralization
Biomimetic mineralization is a facile and environmental friendly method to synthesis bone-like apatite under ambient conditions in aqueous environments. Usually graphene and its derivatives are immersed in a supersaturated or unstable solution with calcium ions and phosphate ions concentrations similar to simulated physiological condition, and apatite will nucleate and precipitate on the surface of those graphene-based materials. During the mineralization process, GO greatly enhance the nucleation and crystallization of HA, resulting in a hybrid homogeneous GO/HA coatings with dense and fine flake-like HA nanocrystalline [bib_ref] Biomimetic synthesis and characterization of hydroxyapatite/graphene oxide hybrid coating on Mg alloy..., Gao [/bib_ref]. Usually, graphene and its derivatives are surfacefunctionalized by bioactive materials to endow the composite with novel properties and facilitate the biomimetic deposition of HA. The GO can be modified by gelatin to mimic the charged proteins in extracellular matrix for regulating bone formation, and the presence of gelatin improves the attraction of calcium ions and promotes the nucleation of HA [bib_ref] Gelatin functionalized graphene oxide for mineralization of hydroxyapatite: biomimetic and in vitro..., Liu [/bib_ref]. Besides GO can be also biofunctionalized by polydopamine [bib_ref] Simultaneous reduction and surface functionalization of graphene oxide for hydroxyapatite mineralization, Liu [/bib_ref] , casein phosphopeptide [bib_ref] Casein phosphopeptide-biofunctionalized graphene biocomposite for hydroxyapatite biomimetic mineralization, Fan [/bib_ref] , carrageenan [bib_ref] Biomimetic and cellmediated mineralization of hydroxyapatite by carrageenan functionalized graphene oxide, Liu [/bib_ref] , chitosan [bib_ref] Biomimetic mineralized hierarchical graphene oxide/chitosan scaffolds with adsorbability for immobilization of nanoparticles..., Xie [/bib_ref] [bib_ref] The synergistic effect of a hybrid graphene oxide-chitosan system and biomimetic mineralization..., Depan [/bib_ref] , fibrinogen [bib_ref] Alternate layer-by-layer assembly of graphene oxide nanosheets and fibrinogen nanofibers on a..., Wang [/bib_ref] or peptide [bib_ref] Self-assembled peptide nanofibers on graphene oxide as a novel nanohybrid for biomimetic..., Wang [/bib_ref] to improve the mineralization process.
## Hydrothermal synthesis
Hydrothermal synthesis of graphene/HA composite involves of dispersing graphene or GO into aqueous solutions containing calcium and phosphate ions, and crystallizing HA nanoparticles at high reaction temperatures and vapor pressures. Rod-like HA, with an average length of 55 nm and diameter of 13 nm, has been successfully synthesized on both sides of graphene nanosheets by using the convenient one-pot hydrothermal synthesis strategy [bib_ref] One-pot synthesis of graphene/hydroxyapatite nanorod composite for tissue engineering, Fan [/bib_ref]. This technique can improve the crystallinity of HA and partially reduce GO to rGO [bib_ref] Enzymatic electrochemical glucose biosensors by mesoporous 1D hydroxyapatite-on-2D reduced graphene oxide, Bharath [/bib_ref]. By using a mixed solvent system of ethylene glycol, N,N-dimethylformamide (DMF) and water, Baradaran et al. [bib_ref] Alias, Mechanical properties and biomedical applications of a nanotube hydroxyapatite-reduced graphene oxide..., Baradaran [/bib_ref] synthesized HA nanotubes on rGO without using reducing agents. This technique is suitable for large-scale production of graphene/HA composites with good crystallinity and wellcontrolled stoichiometric composition, but possesses the disadvantages of using expensive autoclaves, and the inability to observe the crystal as it grows.
## Chemical vapor deposition
Chemical vapor deposition is a low cost and scalable technique to prepare graphene films [bib_ref] Review of chemical vapor deposition of graphene and related applications, Zhang [/bib_ref]. Novel multicomponent and biocompatible graphene/HA/Au nanocomposites are prepared by using radio-frequency chemical vapor deposition (rf-CVD), with acetylene and methane as the carbon sources [bib_ref] Novel multicomponent and biocompatible nanocomposite materials based on few-layer graphenes synthesized on..., Biris [/bib_ref]. During the deposition process, Au nanoclusters are uniformly dispersed over HA particles with diameters of 2 nm~7 nm and act as catalyst for graphene synthesis [bib_ref] Novel multicomponent and biocompatible nanocomposite materials based on few-layer graphenes synthesized on..., Biris [/bib_ref]. This research indicates that longer rf-CVD time can result in few-layers graphene with larger dimensions [bib_ref] Novel multicomponent and biocompatible nanocomposite materials based on few-layer graphenes synthesized on..., Biris [/bib_ref] [bib_ref] In vitro study of biocompatibility of a graphene composite with gold nanoparticles..., Crisan [/bib_ref]. The applications of bulk HA for hard tissue implants are limited by the low mechanical strength of consolidated HA. During the conventional sintering process, HA will dissociate into tricalcium phosphate and tetracalcium phosphate at 1000 Ce1300 C, and usually the high temperatures and long sintering time can cause grain coarsening behavior, which may deteriorate the mechanical properties of HA. As an alternative method, spark plasma sintering (SPS) is an effective approach for preparing novel nanoceramics at low temperatures for short periods of time, with the advantages of retaining fine grain sizes. Graphene/HA composites are successfully fabricated by SPS. The starting powders used for SPS can be (i) prepared by mixing HA powders/nanoparticles and graphene sheets together using mechanical milling [bib_ref] Sz epv€ olgyi, Z. K aroly, Spark plasma sintering of graphene reinforced..., Ebert [/bib_ref] and ultrasonic dispersion [bib_ref] Spark plasma sintered hydroxyapatite/graphite nanosheet and hydroxyapatite/multiwalled carbon nanotube composites: mechanical and..., Zhu [/bib_ref] [bib_ref] Directional property evaluation of spark plasma sintered GNPs-reinforced hydroxyapatite composites, Bajpai [/bib_ref] and (ii) synthesized by a liquid precipitation method [bib_ref] Synthesis of hydroxyapatiteereduced graphite oxide nanocomposites for biomedical applications: oriented nucleation and..., Liu [/bib_ref]. Graphene with diameters of several micrometers are uniformly dispersed and embedded within the HA marix and located between the HA crystal grain boundaries without agglomerations [bib_ref] Spark plasma sintered hydroxyapatite/graphite nanosheet and hydroxyapatite/multiwalled carbon nanotube composites: mechanical and..., Zhu [/bib_ref]. Typical SEM images of the SPS samples were shown in [fig_ref] Figure 4: SEM images of the surfaces of [/fig_ref].
## Hot isostatic pressing
Hot isostatic pressing (HIP) is a traditional technique to densify presintered components, consolidate powders and increase interfacial bonding [bib_ref] Fundamental aspects of hot isostatic pressing: an overview, Atkinson [/bib_ref]. It can be exploited to make HA ceramics with ultrafine microstructures and significantly improved mechanical properties [bib_ref] Processing and properties of transparent hydroxyapatite and beta tricalcium phosphate obtained by..., Boilet [/bib_ref]. Recently, graphene is introduced into this system as an effective additive for toughening ceramics/composites; novel graphene/biphasic calcium phosphate composite [bib_ref] Microstructure and anisotropic mechanical properties of graphene nanoplatelet toughened biphasic calcium phosphate..., Zhao [/bib_ref] , graphene/nickel-doped biphasic calcium phosphate composite [bib_ref] Alias, Characterization of nickel-doped biphasic calcium phosphate/graphene nanoplatelet composites for biomedical application, Baradaran [/bib_ref] , and reduced graphene oxide/nanotube hydroxyapatite composite [bib_ref] Alias, Mechanical properties and biomedical applications of a nanotube hydroxyapatite-reduced graphene oxide..., Baradaran [/bib_ref] have been successfully fabricated by using HIP technique.
## Graphene/ha composite coating
## Electrophoretic deposition
Electrophoretic deposition is a well-developed and aqueousbased colloidal process to deposit charged nano/micro particles onto conductive substrates under DC electric field. It has been widely used to prepare functional bioactive coatings with advanced nanostructures for biomedical applications, such as HA, bioglass and other bioceramic coatings [bib_ref] Electrophoretic deposition of biomaterials, Boccaccini [/bib_ref]. EPD also enables the fabrication of graphene-based novel composite coatings [bib_ref] Applications of graphene electrophoretic deposition. A review, Chavez-Valdez [/bib_ref]. In our previous study, GO/HA was successfully prepared on Ti substrate using cathodic electrophoretic deposition [bib_ref] Graphene oxide/ hydroxyapatite composite coatings fabricated by electrophoretic nanotechnology for biological applications, Li [/bib_ref] , as shown in(a). During the coating process, GO can be considered as amphiphilic macromolecules with an edge-to-center distribution of hydrophilic (oxygenated regions and sheet edges) and hydrophobic (graphenic sp 2 hybridized carbon plane) domains [bib_ref] Graphene oxide as surfactant sheets, Cote [/bib_ref] to enhance the uniform distribution of HA in the deposition suspensions(b)). The 2D novel nanostructure of GO can increase the interlocking of HA nanoparticles and decrease the cracking and delamination the coatings [bib_ref] Graphene oxide/ hydroxyapatite composite coatings fabricated by electrophoretic nanotechnology for biological applications, Li [/bib_ref]. EPD of HA is usually followed by post-heating treatment to increase the coating/substrate bonding strength and the addition of GO can effectively increase this cohesive strength. The microstructure of the GO/HA composite coating is proposed to be combination of an inner compact layer and outer relatively porous and unsealed layer [bib_ref] Graphene oxide/ hydroxyapatite composite coatings fabricated by electrophoretic nanotechnology for biological applications, Li [/bib_ref].
## Electrochemical deposition
Electrochemical deposition of HA involves dissolving calcium and phosphate ions in a buffer solution with controlled pH values and temperature under varying electrical current [bib_ref] The structural and biological properties of hydroxyapatite-modified titanate nanowire scaffolds, Zhao [/bib_ref]. When the voltage is applied, Ca 2þ will migrate onto the cathode surface due to electrostatic attraction and react with the OH À therein produced by the electrolysis of water, resulting in the in situ nucleation and growth of HA on the cathode surface [bib_ref] Electrochemical growth of composite hydroxyapatite coatings for controlled release, Fu [/bib_ref]. Zeng et al. [bib_ref] Graphene oxide/hydroxyapatite composite coatings fabricated by electrochemical deposition, Zeng [/bib_ref] fabricate GO/HA coatings on Ti by using this technique; GO was dispersed and mixed with electrolyte for deposition which consist of Ca(NO 3 ) 2 , NH 4 H 2 PO 4 , NaNO 3 (to improve the ionic strength of the solution) and H 2 O 2 (to restrict the formation of hydrogen gas at cathode). The resulting pure HA coating exhibits a rough morphology with shell-like flakes and the GO/HA composite coating shows uniform and porous topography. The increase of GO contents in the electrolyte can improve the HA crystallinity and bonding strength of the coatings.
## Thermal spray
Thermal sprayed HA and HA-based composite coatings have been successfully used on commercially available Ti-based orthopedic implants, having the advantage of high deposition rate, good bonding strength and variable coating thickness. This process involves heating the HA powders to melting state at high temperature, which may cause the decomposition of HA and exhibit detrimental effects on the coating biocompatibilities. Therefore, Liu et al. takes vacuum cold spraying as an alternative to prepare graphene/HA nanostructured coatings at room temperature [bib_ref] Hydroxyapatite/graphene-nanosheet composite coatings deposited by vacuum cold spraying for biomedical applications: inherited..., Liu [/bib_ref]. The graphene/HA composite powder is prepared by wet chemical approach, and the sprayed coatings have tailorable thickness and display competitive adhesive strength and fracture toughness, with graphene evenly embedded in HA matrix [bib_ref] Hydroxyapatite/graphene-nanosheet composite coatings deposited by vacuum cold spraying for biomedical applications: inherited..., Liu [/bib_ref]. The FESEM images of the coating were shown in [fig_ref] Figure 6: FESEM views of the as-deposited nanostructured coatings, [/fig_ref] [bib_ref] Hydroxyapatite/graphene-nanosheet composite coatings deposited by vacuum cold spraying for biomedical applications: inherited..., Liu [/bib_ref].
## Graphene/ha composite scaffolds
Pure graphene 3D porous structures can be achieved by chemical vapor deposition method with Ni foam as template [bib_ref] Three-dimensional graphene foam as a biocompatible and conductive scaffold for neural stem..., Li [/bib_ref] and hydrothermal approach [bib_ref] Fabrication and characterization of graphene hydrogel via hydrothermal approach as a scaffold..., Lim [/bib_ref] for tissue engineering applications. Moreover, graphene can be utilized as reinforcements for porous nanocomposites and this promising nanocomposite scaffolds can be produced by using lyophilisation or electrospinning.
## Lyophilisation
Lyophilisation is a frequently used technique to prepare porous scaffolds by freezing the solution and then reducing the surrounding pressure to enable the sublimation of frozen water. A scaffold with desirable mechanical and biological properties is obtained by lyophilizing the GO, HA and sodium alginate mixtures [bib_ref] Novel porous graphene oxide and hydroxyapatite nanosheets-reinforced sodium alginate hybrid nanocomposites for..., Xiong [/bib_ref]. This novel nanocomposite scaffolds possess a porosity over 85% and average pore size larger than 150 mm; compared with HA 20 wt%-sodium alginate hybrid scaffolds, the addition of 1 wt% GO can improve their compressive strength and modulus by 23.2% and 28.3% respectively [bib_ref] Novel porous graphene oxide and hydroxyapatite nanosheets-reinforced sodium alginate hybrid nanocomposites for..., Xiong [/bib_ref]. Nair et al. [bib_ref] Graphene oxide nanoflakes incorporated gelatin-hydroxyapatite scaffolds enhance osteogenic differentiation of human mesenchymal..., Nair [/bib_ref] incorporated the GO nanoflakes into gelatin-HA matrix by using freeze drying method and its morphology was shown in [fig_ref] Figure 7: Scanning electron micrographs of the composite scaffolds [/fig_ref].
## Electrospinning
Electrospinning employs an electrical field produced under high voltage to force out the polymeric liquid from the spinneret, resulting in a polymeric fibrous and porous scaffolds on the collectors [bib_ref] The significance of electrospinning as a method to create fibrous scaffolds for..., Repanas [/bib_ref]. Liang et al. prepared a composite nanofiber scaffold consisting of gelatin/chitosan/HA/GO by electrospinning [bib_ref] Electrospinning gelatin/ chitosan/hydroxyapatite/graphene oxide composite nanofibers with antibacterial properties, Hong-Pei [/bib_ref] and the effects of the solution composition on fiber morphology were investigated; the addition of GO can obtain nanofibers with uniform and smooth microstructures and endows the fibrous scaffold with good antibacterial effect against both Staphylococcus aureus and Escherichia coli. Ma et al. [bib_ref] Preparation and cytocompatibility of polylactic acid/hydroxyapatite/graphene oxide nanocomposite fibrous membrane, Ma [/bib_ref] prepared a porous polylactic acid (PLA)/HA/GO scaffold using electrospinning method and SEM images of the composite were shown in [fig_ref] Figure 8: SEM images of PLA/HA/GO nanofibers with high magnification [/fig_ref].
## 3d printing
Three dimensional printing is a superior additive manufacturing technique to print scaffold with customized shape, controlled chemistry and porosities and shows great potential for its application in bone tissue engineering [bib_ref] Bone tissue engineering using 3D printing, Bose [/bib_ref]. Although bone has selfhealing abilities, the large bone loss or damage cannot be healed completely and spontaneously. A scaffold or matrix materials should be incorporated to assist this healing process. Wu et al. [bib_ref] Grapheneoxide-modified b-tricalcium phosphate bioceramics stimulate in vitro and in vivo osteogenesis, Wu [/bib_ref] prepared GO surface modified b-tricalcium phosphate (b-TCP) scaffolds by first using 3D printing method and then soaking the b-TCP scaffold into GO/water suspension as shown in [fig_ref] Figure 9: Scheme illustration for GO modification of b-TCP bioceramics stimulates the in vivo... [/fig_ref].
## Self-assembling
Self-assembly of GO provides a facile and efficient method to produce graphene-based macrostructures. As shown in [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] , GO and HA nanoparticles (nHA) were ultrasonically mixed in ice bath resulting in a homogeneous suspension, and then the mixture were heated at 200 C for 3 h to induce self-assembly [bib_ref] Threedimensional porous scaffold by self-assembly of reduced graphene oxide and nano-hydroxyapatite composites..., Nie [/bib_ref]. This technique reduced GO to rGO without using reductant and organic solvent, which could minimize cytotoxicity of the composite.
## Mechanical properties of composites
## Mechanical properties of graphene and hydroxyapatite
As a biologically active calcium phosphate ceramic, its poor mechanical properties, such as low fracture toughness and tensile strength, limit its applications in major load-bearing scenario. Therefore, the main goal to incorporate graphene and its derivatives into HA is to improve the overall mechanical properties of the composite, in order to better fulfill its biological functions.
The atomically perfect monolayer graphene displayed a Young's modulus of 1.0 TPa and a fracture strength of 130 GPa, which was predicted to be stronger than any other materials [bib_ref] Measurement of the elastic properties and intrinsic strength of monolayer graphene, Lee [/bib_ref]. The mechanical properties of graphene are influenced by the structural defects (monatomic vacancies and Stone-Wales dislocations) [bib_ref] Mechanical and thermal transport properties of graphene with defects, Hao [/bib_ref] and doping or functionalization defects [bib_ref] The role of defects and doping in 2D graphene sheets and 1D..., Terrones [/bib_ref]. As the oxygenated derivatives of graphene, graphene oxide contains reactive oxygenic groups, rendering it higher chemical activities for surface modification [bib_ref] The chemistry of graphene oxide, Dreyer [/bib_ref] but with lower effective Young's modulus (207.6 ± 23.4 GPa) [bib_ref] Mechanical properties of monolayer graphene oxide, Suk [/bib_ref]. GO could be reduced to reduced graphene oxide (rGO) with higher Young's modulus of 0.25 TPa [bib_ref] Elastic properties of chemically derived single graphene sheets, Omez-Navarro [/bib_ref]. Compared with pristine graphene, its oxygenated derivatives (GO and rGO) displayed higher stabilization behavior in aqueous media and could be attractive and promising nanoscale reinforcement fillers in biocomposites [bib_ref] Graphene oxide/ hydroxyapatite composite coatings fabricated by electrophoretic nanotechnology for biological applications, Li [/bib_ref]. These nanofillers could enhance the interfacial bonding within the components, and facilitate stress transfer in the composites [bib_ref] Graphene-based composite materials, Stankovich [/bib_ref].
## Mechanical properties of composites
Composites are expected to exhibit improved properties than their individual components. The addition of GO could efficiently increase the adhesion strength of the HA coatings. The binding strength between the GO/HA coating and Ti substrate were evaluated according to ASTM F1044-99 in our previous research [bib_ref] In situ synthesis and biocompatibility of nano hydroxyapatite on pristine and chitosan..., Li [/bib_ref] , and compared with pure HA coating, the adhesion strength of 5 wt %GO/HA coating increased from 1.55 ± 0.39 MPa to 3.3 ± 0.25 MPa.
The rGO/HA composites were prepared by hydrothermal approach and consolidated by hot isostatic pressing technique; and then their mechanical properties were assessed using the indentation method as shown in [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] [bib_ref] Fast preparation of nano-hydroxyapatite/ superhydrophilic reduced graphene oxide composites for bioactive applications, Zanin [/bib_ref]. According to the investigation of S. Baradaran et al., even low addition of rGO has a great effect on the bulk mechanical properties, and the composite containing 1.5 wt% rGO displayed a maximum fracture toughness and elastic modules, around 86% and 40% higher than that of the pure HA [bib_ref] Fast preparation of nano-hydroxyapatite/ superhydrophilic reduced graphene oxide composites for bioactive applications, Zanin [/bib_ref].
## Toughening mechanism
The mechanical efficiency of the reinforcement fillers in the composites were mainly determined by several factors:(1) the intrinsic mechanical properties of the fillers, (2) the inherent mechanical behavior of the matrix, (3) the fillers volume fraction, (4) the preparation method of the composites, (5) the interfacial bonding strength between the filler and matrix, and (6) the distribution level of the fillers within the matrix [bib_ref] Alias, Mechanical properties and biomedical applications of a nanotube hydroxyapatite-reduced graphene oxide..., Baradaran [/bib_ref] [bib_ref] Mechanical properties of graphene platelet-reinforced alumina ceramic composites, Liu [/bib_ref].
The toughening mechanism for the graphene-based HA composites, according to previous investigations [bib_ref] Microstructure and anisotropic mechanical properties of graphene nanoplatelet toughened biphasic calcium phosphate..., Zhao [/bib_ref] [bib_ref] Synthesis of hydroxyapatiteereduced graphite oxide nanocomposites for biomedical applications: oriented nucleation and..., Liu [/bib_ref] [bib_ref] Alias, Mechanical properties and biomedical applications of a nanotube hydroxyapatite-reduced graphene oxide..., Baradaran [/bib_ref] , are crack branching, crack bridging, pull out, crack deflection and other reinforcing mechanisms. To be specific:
(1) The addition of the 2D fillers can result in the formation of HA with smaller sizes [bib_ref] Synthesis of hydroxyapatiteereduced graphite oxide nanocomposites for biomedical applications: oriented nucleation and..., Liu [/bib_ref] [bib_ref] Synthesis of hydroxyapatite-reduced graphite oxide nanocomposites for biomedical applications: oriented nucleation and..., Liu [/bib_ref] ；and the GO/rGO may functionalized as nucleus for the crystallization of HA crystals [bib_ref] In situ synthesis and biocompatibility of nano hydroxyapatite on pristine and chitosan..., Li [/bib_ref] , which in turn may restrain the growth of HA nanoparticles. It is reported that the fine grain strengthening mechanism could be a possible cause for the improved mechanical properties of the HA composites [bib_ref] Synthesis of hydroxyapatiteereduced graphite oxide nanocomposites for biomedical applications: oriented nucleation and..., Liu [/bib_ref]. This inhibitory effect on grain growth was well studied by Liu et al. [bib_ref] Inhibited grain growth in hydroxyapatiteegraphene nanocomposites during high temperature treatment and their..., Liu [/bib_ref] and illustrated in [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] (2) Moreover, the calculated pullout energy of graphene from HA is around 3e40 J m À2 , which is much higher than the fracture energy (1 J m À2 ) of bulk HA [bib_ref] Synthesis of hydroxyapatiteereduced graphite oxide nanocomposites for biomedical applications: oriented nucleation and..., Liu [/bib_ref]. Therefore, the indentation-induced cracks, as shown in [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] , preferably propagate within the HA matrix, instead of along the HAgraphene interface. The stress transfer characteristics in graphene-based nanocomposites were evaluated from a computational model using a multi-scale finite element approach [bib_ref] Finite element prediction of stress transfer in graphene nanocomposites: the interface effect, Spanos [/bib_ref].
(3) Graphene-based nanosheets have high specific surface area enabling increased contact area with the matrix. And due to its high Young's modulus and flexibility, graphene may be located around the grain boundaries and aligned with the grain shape [bib_ref] Microstructure and anisotropic mechanical properties of graphene nanoplatelet toughened biphasic calcium phosphate..., Zhao [/bib_ref] , which is expected to enhance the mechanical interlocking and adhesion strength within the matrix [bib_ref] In situ synthesis and biocompatibility of nano hydroxyapatite on pristine and chitosan..., Li [/bib_ref]. The graphene-based 2D reinforcement fillers can inhibit the propagation of the crack. Liu et al. [bib_ref] Hydroxyapatite/graphene-nanosheet composite coatings deposited by vacuum cold spraying for biomedical applications: inherited..., Liu [/bib_ref] prepared graphene/HA coatings by vacuum cold spraying and the authors incubated the samples in culture media for one month without cells, and the coating adhesion values showed no remarkable changes.
## Biocompatibility of the composites
## In vitro biocompatibility
From a mechanical point of view, this graphene-based HA nanostructural composites have proven to be promising candidates for hard tissue repairing or regeneration. Furthermore, these materials should be biocompatible with high stability in biological environment. Their toxicity was well analyzed and studied.
## Cytocompatibility
Usually the graphene-based HA composites were co-cultured with (1) embryonic fibroblast cell line (NIH 3T3 cells) [bib_ref] Bacterial cellulose-reinforced hydroxyapatite functionalized graphene oxide: a potential osteoinductive composite, Ramani [/bib_ref] , mouse fibroblast cell line (L929 cells) [bib_ref] Simultaneous reduction and surface functionalization of graphene oxide for hydroxyapatite mineralization, Liu [/bib_ref] [bib_ref] In situ synthesis and biocompatibility of nano hydroxyapatite on pristine and chitosan..., Li [/bib_ref] [bib_ref] Alternate layer-by-layer assembly of graphene oxide nanosheets and fibrinogen nanofibers on a..., Wang [/bib_ref] [bib_ref] Graphene oxide/ hydroxyapatite composite coatings fabricated by electrophoretic nanotechnology for biological applications, Li [/bib_ref] , (3) stem cells: mesenchymal stem cells (MSC) [bib_ref] Fast preparation of nano-hydroxyapatite/ superhydrophilic reduced graphene oxide composites for bioactive applications, Zanin [/bib_ref] and (4) other cells: PBMC (human peripheral blood mononuclear cells) [bib_ref] Bioactive hydroxyapatite/ graphene composite coating and its corrosion stability in simulated body..., Jankovi C [/bib_ref] [bib_ref] Graphene-based antibacterial composite coatings electrodeposited on titanium for biomedical applications, Jankovi C [/bib_ref]. Fan et al. [bib_ref] One-pot synthesis of graphene/hydroxyapatite nanorod composite for tissue engineering, Fan [/bib_ref] seeded MC3T3-E1 cells onto graphene/HA composites with various HA contents and the cells displayed flatter morphology than on the GO and HA coating, and the composite containing 40 wt% showed higher bone cellular activities. In order to further increase the biocompatibility of the graphene/HA composites, bioactive polymers were added into the system, such as carrageenan [bib_ref] Biomimetic and cellmediated mineralization of hydroxyapatite by carrageenan functionalized graphene oxide, Liu [/bib_ref] , gelatin [bib_ref] Gelatin functionalized graphene oxide for mineralization of hydroxyapatite: biomimetic and in vitro..., Liu [/bib_ref] and polydopamine [bib_ref] MC3T3-E1 preosteoblast cell-mediated mineralization of hydroxyapatite by poly-dopamine-functionalized graphene oxide, Cheng [/bib_ref]. Combined with the graphene-based fillers, these polymers could facilitate the HA mineralization process which promoted the osteogenic differentiation of MC3T3-E1 cells. Lee et al. [bib_ref] Enhanced osteogenesis by reduced graphene oxide/hydroxyapatite nanocomposites, Lee [/bib_ref] cultured the MC3T3-E1 cells (2 Â 10 5 cells/mL) with colloidal dispersion of HA particles, rGO nanosheets and rGO/HA nanocomposites (10 mg/mL) for 1e21days and the cells were analyzed by Alizarin red staining (ARS). The results were shown in [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref].
Oyefusi [bib_ref] Hydroxyapatite grafted carbon nanotubes and graphene nanosheets: promising bone implant materials, Oyefusi [/bib_ref] investigated the hFOB 1.19 cell proliferation and differentiation by using total protein assays and Western blot analysis of osteocalcin expression with promising results. Liu et al. [bib_ref] Hydroxyapatite/graphene-nanosheet composite coatings deposited by vacuum cold spraying for biomedical applications: inherited..., Liu [/bib_ref] cultured the osteoblast cells on the composite coatings showing higher proliferation rate and better stretching behavior on the HA-based coatings than on bare Ti. The random curly incorporated graphene sheets possessed more active sites for cell binding and could adsorb the key serum proteins (fibronectin) to further enhance the cell attachment [bib_ref] Hydroxyapatite/graphene-nanosheet composite coatings deposited by vacuum cold spraying for biomedical applications: inherited..., Liu [/bib_ref].
The graphene/HA composites also showed good cytocompatibility to MG63 cells. The HA coated GO exhibited comparable biocompatibility to HA minerals [bib_ref] Biomimetic graphene oxide-hydroxyapatite composites via in situ mineralization and hierarchical assembly, Li [/bib_ref]. However, high GO contents in the composites might inhibit cell viabilities [bib_ref] Graphene oxide/ hydroxyapatite composite coatings fabricated by electrophoretic nanotechnology for biological applications, Li [/bib_ref]. Compared with the GO/HA composites, the addition of chitosan could increase its bone-forming abilities [bib_ref] In situ synthesis and biocompatibility of nano hydroxyapatite on pristine and chitosan..., Li [/bib_ref]. Ramani et al. [bib_ref] Bacterial cellulose-reinforced hydroxyapatite functionalized graphene oxide: a potential osteoinductive composite, Ramani [/bib_ref] prepared bacterial cellulose/GO/HA composites with high osteoinductive and the MG63 cells cultured with 50 mg/mL composites had higher ALP activities.
Liu et al. [bib_ref] Simultaneous reduction and surface functionalization of graphene oxide for hydroxyapatite mineralization, Liu [/bib_ref] evaluated the cytotoxicity of rGO/HA on L929 cells using MTT assay and the cell viability was more than 95% in comparison with the control; moreover, the concentration of this composite in the culture media (0, 1, 5, 10, and 20 mg/mL) had no significant influences on the mitochondrial activities within the L929 cells. Wang et al. [bib_ref] Alternate layer-by-layer assembly of graphene oxide nanosheets and fibrinogen nanofibers on a..., Wang [/bib_ref] prepared a GO/fibrinogen nanofiber scaffold by using layer-by-layer method and then incubated it in SBF for biomineralization resulting in 3D GO/HA scaffolds, which had no obvious inhibitory effects on the in vitro cell proliferation of L929 cells. In our previous study [bib_ref] In situ synthesis and biocompatibility of nano hydroxyapatite on pristine and chitosan..., Li [/bib_ref] [bib_ref] Graphene oxide/ hydroxyapatite composite coatings fabricated by electrophoretic nanotechnology for biological applications, Li [/bib_ref] , the in vitro cytotoxicity of the GO/HA and chitosan modified GO/HA did not exhibit obvious concentration-dependent characteristics but showed a clear positive time dependence [bib_ref] In situ synthesis and biocompatibility of nano hydroxyapatite on pristine and chitosan..., Li [/bib_ref] ; and the cells displayed a healthy round shape with extended pseudopodium [bib_ref] Graphene oxide/ hydroxyapatite composite coatings fabricated by electrophoretic nanotechnology for biological applications, Li [/bib_ref].
Zanin et al. [bib_ref] Fast preparation of nano-hydroxyapatite/ superhydrophilic reduced graphene oxide composites for bioactive applications, Zanin [/bib_ref] prepared globular nano-HA onto rGO by electrodeposition and evaluated the composite's biocompatibility by culturing with MSC. The cells adhered well on the composites with a flat roughly circular morphology, presenting active formation of membrane projections. PBMC consists of lymphocytes and monocytes, as the main representatives of human immune cells [bib_ref] Bioactive hydroxyapatite/ graphene composite coating and its corrosion stability in simulated body..., Jankovi C [/bib_ref]. Compared with the control group, the cells viability rate on the composites showed mild decrease and the composite could be considered non-toxic.
## Hemocompatibility
Nair et al. [bib_ref] Graphene oxide nanoflakes incorporated gelatin-hydroxyapatite scaffolds enhance osteogenic differentiation of human mesenchymal..., Nair [/bib_ref] incorporated the GO nanoflakes into gelatin-HA matrix by using a freeze drying method. None of the prepared scaffolds showed hemolysis with comparable activity to the saline control groups [bib_ref] Graphene oxide nanoflakes incorporated gelatin-hydroxyapatite scaffolds enhance osteogenic differentiation of human mesenchymal..., Nair [/bib_ref]. In our previous study [bib_ref] Electrophoretic-deposited novel ternary silk fibroin/graphene oxide/hydroxyapatite nanocomposite coatings on titanium substrate for..., Li [/bib_ref] , silk fibroin/GO/HA coatings were electrophoretic-deposited on Ti substrates and the hemocompatibility of samples were evaluated according to ASTM F756-08 standard [bib_ref] Standard Practice for Assessment of Hemolytic Properties of Materials, Astm [/bib_ref]. All the samples exhibited good hemocompatibility with hemolysis rate lower than 5%, and platelet adhesion tests showed that only few platelets were observed on the composite coatings at the inactivated stage with round shape [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref]. The GO and graphene nanosheets showed a dosedependent hemolytic activity on red blood cells [bib_ref] Cytotoxicity of graphene oxide and graphene in human erythrocytes and skin fibroblasts, Liao [/bib_ref] and covering GO sheets with HA could alleviate their hemolytic activity.
## Mineralization abilities
Simulated body fluid (SBF) contains ion concentrations almost equal to those of human blood plasma. Usually the bone-forming or bone-bonding abilities of a material could be evaluated by immersing it into SBF to examine apatite formation abilities on its surface, which is a useful method to predict the in vivo osteogenic activity of a material [bib_ref] How useful is SBF in predicting in vivo bone bioactivity, Kokubo [/bib_ref]. However, some researchers suggested that this method needs further elaboration [bib_ref] Can bioactivity be tested in vitro with SBF solution?, Bohner [/bib_ref].
The addition of graphene-based nanofillers into HA matrix could accelerate the formation of apatite on its surface [bib_ref] In vitro comparative study of pure hydroxyapatite nanorods and novel polyethylene glycol/graphene..., Mohandes [/bib_ref] [bib_ref] One-pot synthesis of graphene/hydroxyapatite nanorod composite for tissue engineering, Fan [/bib_ref] [bib_ref] Bioactive hydroxyapatite/ graphene composite coating and its corrosion stability in simulated body..., Jankovi C [/bib_ref] [bib_ref] Self-assembled peptide nanofibers on graphene oxide as a novel nanohybrid for biomimetic..., Wang [/bib_ref]. Zhang et al.prepared graphene/HA composites using spark plasma sintering technique. The samples were immersed in SBF for 7 days, and, compared with the HA, the bone- like apatite layer formed on 1.0 wt% graphene/HA was much thicker than that on the pure HA substrate, probably suggesting a higher osteogenic activity of the graphene-reinforced HA. The proposed mineralization process were depicted in [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref]. Compared with pure HA, the graphene/HA composites exhibited lower stability. The incorporation of graphene into HA could result in smaller grain size and more specific area, leading to fast dissolution of calcium and higher negative surface charge on the composites [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref]. Therefore, more calcium could be attracted on its surface [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] and the Ca-rich layer could generate thicker bone-like layer [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] [bib_ref] Bioactive hydroxyapatite/ graphene composite coating and its corrosion stability in simulated body..., Jankovi C [/bib_ref]. In order to further increase the boneforming ability of the composites, the nanosheets could be coated by bioactive polymers such as polyethylene glycol [bib_ref] In vitro comparative study of pure hydroxyapatite nanorods and novel polyethylene glycol/graphene..., Mohandes [/bib_ref] and selfassembled peptide nanofibers [bib_ref] Self-assembled peptide nanofibers on graphene oxide as a novel nanohybrid for biomimetic..., Wang [/bib_ref].
## In vivo biocompatibility
Bioactivity of the graphene-based HA composites in the context of osteogenesis by using in vitro cell models have been extensively investigated. However, there were few in vivo animal studies using these novel material.
Lee et al. [bib_ref] Enhanced osteogenesis by reduced graphene oxide/hydroxyapatite nanocomposites, Lee [/bib_ref] prepared the HA grafts and the rGO/HA grafts. The grafts were filled into bone defects (6 mm in diameter and 2.5 mm in depth) that were trephined in the parietal bone of 12e13 weekold male New Zealand rabbits [bib_ref] Enhanced osteogenesis by reduced graphene oxide/hydroxyapatite nanocomposites, Lee [/bib_ref]. Four weeks after the surgery, the non-treated control defects were filled with thin and loose connective tissues without many new bones [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] , and dense connective tissue and small particles were observed in the one with HA grafts [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref]. For the groups implanted by rGO/HA grafts [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] , several newly formed bone were detected, indicating the accelerated bone remodeling process. From the Masson's trichrome staining analysis, the amount of the newly formed bone in the rGO/HA treated defects [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] were substantially higher than that in the HA groups [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref]. The histometric evaluations [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] showed that the rGO/HA grafts displayed significantly greater new bone density than the control/ HA groups.
Wu et al. [bib_ref] Grapheneoxide-modified b-tricalcium phosphate bioceramics stimulate in vitro and in vivo osteogenesis, Wu [/bib_ref] prepared GO modified b-tricalcium phosphate (b- TCP) scaffolds by first using 3D printing and then soaking the b-TCP into GO/water suspension. The in vivo bone formation of the scaffold was evaluated by implanting it into critical-sized calvarial defects in New Zealand white male rabbits. Compared with b-TCP scaffolds, the GO modified scaffolds had greater bone formation abilities in the defects both at 4 and 8 weeks post implantation. And the new bone was visible in the periphery and center of the defect in the GO-b-TCP group, which was both quantitatively and qualitatively better than that in the control group as it is shown in [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref].
Both the in vitro and in vivo biocompatible analysis suggested that the graphene-based HA composites showed great potential applications for bone reconstruction, due to its beneficial effects on the adhesion, proliferation and differentiation of osteoblast-related cells, fibroblast-related cells or stem cells. Graphene and its derivatives displayed strong non-covalent binding abilities on osteogenic inducers (dexamethasone and b-glycerolphosphate) [bib_ref] Origin of enhanced stem cell growth and differentiation on graphene and graphene..., Lee [/bib_ref] which accelerated the osteogenic differentiation of stem cells.
## Antibacterial property
The infections of the implantation site after surgery [bib_ref] Graphene-based antibacterial composite coatings electrodeposited on titanium for biomedical applications, Jankovi C [/bib_ref] and biomedical devices associated infections (especially with the formation of biofilm) [bib_ref] Electrophoretic deposition of graphene oxide reinforced chitosan-hydroxyapatite nanocomposite coatings on Ti substrate, Shi [/bib_ref] [bib_ref] Bioinspired anchoring AgNPs onto micro-nanoporous TiO2 orthopedic coatings: trap-killing of bacteria, surface-regulated..., Jia [/bib_ref] are the major cause for delayed healing, implant failure and repeated surgeries [bib_ref] Bioinspired anchoring AgNPs onto micro-nanoporous TiO2 orthopedic coatings: trap-killing of bacteria, surface-regulated..., Jia [/bib_ref]. Therefore, it is necessary to develop novel composite materials with combined bioactivity and antibacterial properties.
In addition to the bone-forming ability, HA shows favorable affinity for bacterial adhesion [bib_ref] Electrophoretic deposition of graphene oxide reinforced chitosan-hydroxyapatite nanocomposite coatings on Ti substrate, Shi [/bib_ref]. When HA was exploited as coating materials on Ti implants, the increased HA contents therein resulted in enhanced attachment of the bacteria, which could deteriorate its osteointegration property and further weaken the biological fixation ability of HA-coated implants in hard tissues [bib_ref] Deposition, characterization, and enhanced adherence of escherichia coli bacteria on flame-sprayed photocatalytic..., Liu [/bib_ref]. Accordingly, it is imperative to endow HA with antibacterial or bacteriostatic effects by preparing HA composites. Graphenebased nanomaterials exhibit excellent cytotoxicity to bacteria. Liu et al. [bib_ref] Antibacterial activity of graphite, graphite oxide, graphene oxide, and reduced graphene oxide:..., Liu [/bib_ref] investigated the antibacterial activity of four types of graphene-based materials (graphite, graphite oxide, GO and rGO) toward an Escherichia coli and the GO dispersion displayed the highest antibacterial activity. The membrane stress induced by the nanosheets warping [bib_ref] Antibacterial activity of graphite, graphite oxide, graphene oxide, and reduced graphene oxide:..., Liu [/bib_ref] , the cell membrane damage [bib_ref] Toxicity of graphene and graphene oxide nanowalls against bacteria, Akhavan [/bib_ref] caused by direct contact with sharp edges of the nanosheets and the oxidative stress generated by the reactive oxygen species (ROS) production [bib_ref] Antibacterial activity of graphite, graphite oxide, graphene oxide, and reduced graphene oxide:..., Liu [/bib_ref] were the major mechanisms in the bacterial inactivation.
Jankovi c et al. [bib_ref] Bioactive hydroxyapatite/ graphene composite coating and its corrosion stability in simulated body..., Jankovi C [/bib_ref] investigated the antibacterial efficiency of the electrophoretic graphene/HA coatings against the Gram-positive pathogenic bacteria strain Staphylococcus aureus and the Gramnegative bacteria strain Escherichia coli. However no antibacterial effects were observed in that study [bib_ref] Bioactive hydroxyapatite/ graphene composite coating and its corrosion stability in simulated body..., Jankovi C [/bib_ref] which might be ascribed to the low contents of the incorporated graphene. As an alternative method, Ag was introduced into the coatings, and the graphene/Ag/ HA coatings showed strong antibacterial activity only after 3 h coculturing with the S. aureus and E. coli, inhibiting harmful biofilm formation [bib_ref] Graphene-based antibacterial composite coatings electrodeposited on titanium for biomedical applications, Jankovi C [/bib_ref]. In our previous research GO reinforced chitosan/HA coatings were deposited onto Ti and the antibacterial adhesion assay indicated that the amount of the adherent bacterial cells decreased greatly on the composite coatings compared with pure HA coatings [bib_ref] Electrophoretic deposition of graphene oxide reinforced chitosan-hydroxyapatite nanocomposite coatings on Ti substrate, Shi [/bib_ref]. As shown in [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] , compare with Ti and HA coating, the number of the adherent bacterial cells is significantly decreased on the chitosan/HA and GO/chitosan/HA coatings, with the potential antibacterial mechanism displayed in [fig_ref] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 [/fig_ref] [bib_ref] Electrophoretic deposition of graphene oxide reinforced chitosan-hydroxyapatite nanocomposite coatings on Ti substrate, Shi [/bib_ref].
The novel graphene-based HA composites were competitive candidates for hard tissue repairing or regeneration with increased osteogenic activity, well hemocompatibility, and promising antibacterial properties. However, in some cases the resulting composites exhibited less bioactivity [bib_ref] Simultaneous reduction and surface functionalization of graphene oxide for hydroxyapatite mineralization, Liu [/bib_ref] [bib_ref] Graphene oxide/ hydroxyapatite composite coatings fabricated by electrophoretic nanotechnology for biological applications, Li [/bib_ref] [bib_ref] Bioactive hydroxyapatite/ graphene composite coating and its corrosion stability in simulated body..., Jankovi C [/bib_ref]. This discrepancy may originated from the preparation techniques of graphene, the size and distribution of this 2D nanoscale filler. In the scenario of graphene oxide, the types of oxygen-containing functional groups and the oxygen content could also affect the bioactivity of the composites. Therefore, more researches should be conducted to shed light on this issues.
# Conclusions
This review of literature presented various preparation techniques, mechanical behaviors and toughen mechanism, in vitro/ in vivo biocompatible analysis, antibacterial properties of the graphene-based HA composites for orthopedic applications. By combining the superior mechanical properties of graphene (and its derivatives) and the high bioactivities of HA, the graphene-based HA composites show great potentials in hard tissue repair and regeneration. The major problem in production of this novel composite is the difficulty in homogeneous distribution of the 2D reinforcement fillers within the matrix. Evaluation of the composites mechanical properties has been the focus of the research. Biocompatibility of graphene-based HA composites has been studied mostly in terms of in vitro cytocompatibility, and further in vivo analysis should be conducted before being considered for clinical applications.
[fig] Figure 1: The number of publications on Graphene-based HA composites from year 2009e2016 (2009 [16], 2011 [17e19], 2012 [20e24], 2013 [25e32], 2014 [33e49], 2015 [50e82], 2016 [83e112]). This figure does not include the papers published in 2017 [113e130]. [/fig]
[fig] Figure 2: The summary of different preparation methods of the graphene-based composites. [/fig]
[fig] Figure 3: (a) The proposed in situ synthesis mechanism of HA on pristine GO sheets. The SEM and TEM images of GOeHA (bed) composites, the insets of insets (d) show the selected area electron diffraction (SAED) patterns of the corresponding composites. The black arrows of (b) point to the wrinkles of the GO sheets. These figures were adapted from Ref.[30]. [/fig]
[fig] Figure 4: SEM images of the surfaces of (a) HA, (b) 0.5 wt% graphene/HA composite and (c) 1.0 wt% graphene/HA composite[25]. [/fig]
[fig] Figure 5: (a) Schematic illustration of the EPD process (b) The TEM images of the HA nanoparticles with 2 wt% GO sheets (c) The SEM images of composite coatings containing (c) 0 wt %, (d) 2 wt% GO and (e) 5 wt% GO sheets, and corresponding optical photos. These figures were adapted from Ref.[38]. [/fig]
[fig] Figure 6: FESEM views of the as-deposited nanostructured coatings, (a) the pure HA coating, (b) the HA-0.1 wt% graphene coating, and (c) the HA-1.0 wt% graphene coating. À1: surface view, À2: magnified surface view, and À3: cross-sectional view. The white arrow head points to graphene located on the surfaces of the coatings, and magnified views of typical areas from the cross-section of the HA-graphene coating showing clearly the presence of GN in the coating and at the coating/substrate interface (c-3). Graphene -induced layered structure is clearly seen for the HA-graphene coatings. [/fig]
[fig] Figure 7: Scanning electron micrographs of the composite scaffolds. The inset demonstrates the photograph of the scaffolds. Note: GHA: gelatin-HA scaffold, GOGHA 0.5: gelatin-HA scaffold with 0.5 wt% GO, GOGHA1.0: gelatin-HA scaffold with1 wt% GO[68]. [/fig]
[fig] Figure 8: SEM images of PLA/HA/GO nanofibers with high magnification (a) and low magnification (c), electrospun PLA with high magnification (b) and low magnification (d)[23]. [/fig]
[fig] Figure 9: Scheme illustration for GO modification of b-TCP bioceramics stimulates the in vivo osteogenesis[69]. [/fig]
|
Monitoring NAD(H) and NADP(H) dynamics during organismal development with genetically encoded fluorescent biosensors
Cell metabolism plays vital roles in organismal development, but it has been much less studied than transcriptional and epigenetic control of developmental programs. The difficulty might be largely attributed to the lack of in situ metabolite assays. Genetically encoded fluorescent sensors are powerful tools for noninvasive metabolic monitoring in living cells and in vivo by highly spatiotemporal visualization. Among all living organisms, the NAD(H) and NADP(H) pools are essential for maintaining redox homeostasis and for modulating cellular metabolism. Here, we introduce NAD(H) and NADP(H) biosensors, present example assays in developing organisms, and describe promising prospects for how sensors contribute to developmental biology research.
# Background
Environmental nutrients and cellular metabolism are crucial for organismal development [bib_ref] Tracking placental development in health and disease, Aplin [/bib_ref]. During embryonic development, cellular metabolism is dynamically programmed; for example, pyruvate is the main nutrient supporting development from the zygote to morula stage, and glucose plays a central role in nutrition during the ensuing morula-to-blastocyst transition [bib_ref] The roles of pyruvate, lactate and glucose during preimplantation development of embryos..., Brown [/bib_ref]. The nutritional functions of pyruvate and glucose are not interchangeable, although pyruvate is just the glycolytic product of glucose, thereby highlighting the highly programmable requirements of nutrients and metabolism for development. The importance of developmental metabolism is clearly evidenced by broad developmental defect spectra in babies with inborn error of metabolism [bib_ref] Inborn errors of metabolism: from preconception to adulthood, Kruszka [/bib_ref].
The human metabolome consists of more than 4100 metabolites [bib_ref] Recon3D enables a three-dimensional view of gene variation in human metabolism, Brunk [/bib_ref]. Among them, core coenzymes I and II, i.e., nicotinamide adenine dinucleotides (NAD + and NADH) and their phosphorylated forms (NADP + and NADPH), are involved in a large number of metabolic reactions and thus have a systematic impact on the entire metabolic network [fig_ref] Figure 1: NAD [/fig_ref]. In eukaryotic cells, there are two main pyridine dinucleotide pools: cytosol and mitochondria. The cytosolic free NAD + concentration is approximately 50-110 μM and is usually 60-1000fold higher than NADH [bib_ref] Biosensor reveals multiple sources for mitochondrial NAD(+), Cambronne [/bib_ref] [bib_ref] Semisynthetic biosensors for mapping cellular concentrations of nicotinamide adenine dinucleotides, Sallin [/bib_ref]. Mitochondrial free NAD + is approximately 230 μM, and the ratio of NADH/NAD + reaches approximately 0.1 to 1 [bib_ref] Biosensor reveals multiple sources for mitochondrial NAD(+), Cambronne [/bib_ref] [bib_ref] Semisynthetic biosensors for mapping cellular concentrations of nicotinamide adenine dinucleotides, Sallin [/bib_ref]. The free NADPH concentration is approximately 3 μM in the cytosol and 37 μM in mitochondria, and NADPH is dominant over NADP + , with an NADPH/NADP + ratio ranging from 15 to 333 [bib_ref] mouse pancreatic islets, and nutrient-induced insulin secretion, Hedeskov [/bib_ref] [bib_ref] Genetically encoded fluorescent sensors reveal dynamic regulation of NADPH metabolism, Tao [/bib_ref] [bib_ref] The redox state of free nicotinamideadenine dinucleotide phosphate in the cytoplasm of..., Veech [/bib_ref]. The cytosolic and mitochondrial NAD(H) and NADP(H) pools are relatively independent of each other. However, NADH can be transported between these compartments via the malate-aspartate or Open Access *Correspondence: [email protected]; [email protected] † Ting Li and Yejun Zou contributed equally to this work. Research Unit of New Techniques for Live-cell Metabolic Imaging, Chinese Academy of Medical Sciences, Beijing, China Full list of author information is available at the end of the article glycerol phosphate shuttle, and the mitochondrial NAD + transporter SLC25A51 has recently been identified [bib_ref] MCART1/ SLC25A51 is required for mitochondrial NAD transport, Kory [/bib_ref] [bib_ref] SLC25A51 is a mammalian mitochondrial NAD(+) transporter, Luongo [/bib_ref].
NAD(H) plays a fundamental role in redox homeostasis and energy metabolism. Since NAD(H) participates in more than one hundred redox reactions as electron carriers, the ratio of NAD + to NADH is a commonly used biochemical index for cellular redox status. Cytosolic NADH is primarily generated by glycolysis, while it can be consumed either during the reduction of pyruvate to lactate or transport into mitochondria [fig_ref] Figure 1: NAD [/fig_ref]. In mitochondria, NADH is generated via oxidation of pyruvate and fatty acids and the tricarboxylic acid (TCA) cycle, and it is consumed for ATP production via mitochondrial respiration. As NADH oxidized form, NAD + is mainly synthesized through three pathways, including the salvage pathway from the precursor nicotinamide, nicotinamide mononucleotide, nicotinamide riboside or dihydronicotinamide riboside [bib_ref] NAD(+) in aging, metabolism, and neurodegeneration, Verdin [/bib_ref] ; the Preiss-Handler pathway from nicotinic acid [bib_ref] NAD(+) in aging, metabolism, and neurodegeneration, Verdin [/bib_ref] ; or the de novo pathway from tryptophan [bib_ref] NAD(+) in aging, metabolism, and neurodegeneration, Verdin [/bib_ref] , aspartate [bib_ref] The biosynthesis of nicotinamide adenine dinucleotides in bacteria, Begley [/bib_ref] , and chorismate [bib_ref] Construction of an alternative NAD(+) De novo biosynthesis pathway, Ding [/bib_ref]. In addition, NAD + is consumed by a few regulatory enzymes, including SIRT, PARP, ART, CD38, and SARM1 [bib_ref] NAD(+) metabolism and its roles in cellular processes during ageing, Covarrubias [/bib_ref] , and is also involved in the 5′ cap modification of mRNAs [bib_ref] 5′-end NAD(+) capping and DeNADding, Kiledjian [/bib_ref]. Therefore, NAD + plays important roles in cell metabolism, cell signaling, and gene expression regulation [bib_ref] Evolving concepts in NAD(+) metabolism, Chini [/bib_ref].
Although structurally similar to NAD(H), NADP(H) has distinct biochemical functions. NADP(H) provides fundamental reducing power for biosynthetic reactions and antioxidant functions. Cytosolic NADPH is mainly produced via the pentose phosphate pathway in mammalian cells. In addition, a few enzymes, including isocitrate dehydrogenases, malic enzymes, and methylenetetrahydrofolate dehydrogenases, contribute to the production of NADPH in both the cytosol and mitochondria [bib_ref] Quantitative flux analysis reveals folate-dependent NADPH production, Fan [/bib_ref] [fig_ref] Figure 1: NAD [/fig_ref]. NADPH maintains cellular redox homeostasis via two enzymatic antioxidant defense systems: the glutathione system (GSH/GSSG) and the thioredoxin system (Trx-SH/Trx-SS); NADPH consumption also occurs in the biosynthetic process of fatty acids, deoxyribonucleotides and some amino acids. Moreover, NADPH is used as a substrate of NADPH oxidase that produces radical oxygen species on the plasma membrane [bib_ref] The NOX family of ROS-generating NADPH oxidases: physiology and pathophysiology, Bedard [/bib_ref]. One direct link between NAD(H) and the NADP(H) pool is phosphorylation of NAD + into NADP + mediated by NAD + kinase in the cytosol or mitochondria [bib_ref] Analysis of redox landscapes and dynamics in living cells and in vivo..., Zou [/bib_ref] , and the other occurs via transdehydrogenase in mitochondria, which converts NADH and NADP + to NAD + and NADPH under a proton gradient [bib_ref] Cofactor balance by nicotinamide nucleotide transhydrogenase (NNT) coordinates reductive carboxylation and glucose..., Gameiro [/bib_ref] [bib_ref] NAD kinase levels control the NADPH concentration in human cells, Pollak [/bib_ref]. Cytosolic NADH is produced during glycolysis, which is either consumed during conversion of pyruvate into lactate or transported into mitochondria by reducing equivalent shuttle pathways. Mitochondrial NADH is generated by oxidation of pyruvate and fatty acids and in the tricarboxylic acid cycle (TCA), and it is oxidized back into NAD + by the electron transfer chain (ETC). Glycolysis and mitochondrial respiration supply energy production. NADPH is mainly produced by the pentose phosphate pathway and other metabolic enzymes whose isoforms exist in both the cytosol and mitochondria, and it is consumed for the synthesis of biomass and antioxidant defense responses. The NAD(H) and NADP(H) pools connect through NAD + kinase (NADK) and nicotinamide nucleotide transhydrogenase (NNT). Abbreviations: glucose-6-P, glucose-6-phosphate; G-3-P, glyceraldehyde-3-phosphate; PEP, phosphoenolpyruvate; ribose-5-P, ribose 5-phosphate Therefore, NAD(H) and NADP(H) integrate inputs from a few metabolic pathways and exert their influences on various metabolic activities. While they both globally report the redox status of cells, NAD(H) and NADP(H) reflect the integrated activities of energy and anabolic metabolism, respectively. Since these metabolic activities are closely associated with organismal development, monitoring NAD(H) and NADP(H) levels shed light on how cellular metabolism is weaved into the developmental program. To date, however, little is known about NAD(H) and NADP(H) dynamics during developmental stages. A main hurdle may be that it lacks an ideal assay for them until the development of genetically encoded fluorescent sensors, which revolutionize in situ metabolic research.
## Methods for nad(h) and nadp(h) analysis
A number of NAD(H) and NADP(H) assays have been established. Biochemical methods, such as chromatography and mass spectrometry, are based on cellular lysis and are incompetent for living cell measurement. NADH and NADPH have autofluorescence with a 340-nm excitation maximum and a 460-nm emission maximum. Based on the optical properties, NAD(P) H autofluorescence has been studied by single-photon or multiphoton excitation to monitor metabolic states in living cells or in vivo for many years [bib_ref] Role of NADH shuttle system in glucose-induced activation of mitochondrial metabolism and..., Eto [/bib_ref] [bib_ref] Neural activity triggers neuronal oxidative metabolism followed by astrocytic glycolysis, Kasischke [/bib_ref] [bib_ref] Mitochondrial function in vivo evaluated by NADH fluorescence: from animal models to..., Mayevsky [/bib_ref]. Unfortunately, NAD(P)H autofluorescence is limited by low sensitivity and cell injury caused by ultraviolet irradiation [bib_ref] Two-photon FLIM of NAD(P)H and FAD in mesenchymal stem cells undergoing either..., Meleshina [/bib_ref]. Even worse, it is difficult to distinguish NADH and NADPH due to their similar spectra [bib_ref] The fluorometric measurement of pyridine nucleotides, Lowry [/bib_ref]. A previous study with fluorescence lifetime imaging (FLIM) reported that NADH and NADPH could be differentiated on the basis of a simple assumption that bound NADH and bound NADPH possess different fluorescence lifetimes inside the cell; however, FLIM is not technically simple for most laboratories [bib_ref] Separating NADH and NADPH fluorescence in live cells and tissues using FLIM, Blacker [/bib_ref]. It's worth noting that the twophoton FLIM techniques eliminate the problem with UV irradiation and exhibit considerable penetration depth. Unlike NADH and NADPH, their oxidized counterparts NAD + and NADP + have no intrinsic fluorescence. In situ magnetic resonance imaging technology can dynamically detect NAD + in the human brain using 31 P spectroscopy, but it is too complicated to use for broad research [bib_ref] In vivo (31) P MRS assessment of intracellular NAD metabolites and NAD(+)..., Lu [/bib_ref]. Two semisynthetic fluorescent probes based on fluorescence resonance energy transfer (FRET) have been developed to spatiotemporally monitor NAD + and NADP(H) levels in live cells, including NAD-Snifit and NADP-Snifit, which are designed by fusing human sepiapterin reductase (SPR) with two self-labeling proteins-Halo-tag and SNAP-tag-as a FRET pair [bib_ref] Semisynthetic biosensors for mapping cellular concentrations of nicotinamide adenine dinucleotides, Sallin [/bib_ref]. NAD-Snifit and NADP-Snifit do not selfsufficiently form intrinsic fluorophores and require exogenous dyes for labeling. Thus, the excess chemical dye has to be removed by a washing procedure, which is not only time intensive but also may render the analysis susceptible to artifacts for developmental research. Therefore, live-cell and in vivo assay methods with high spatiotemporal resolution are highly desirable for monitoring NAD(H) and NADP(H) dynamics during organismal development.
## Genetically encoded nad(h) and nadp(h) sensors
When expressed in living cells or in vivo, fluorescent protein-based sensors can monitor the spatiotemporal dynamics of metabolites with high specificity and sensitivity and have become a revolutionary technology for metabolic research [bib_ref] Creating new fluorescent probes for cell biology, Zhang [/bib_ref]. In the past two decades, over 200 genetically encoded fluorescent sensors have been developed for cellular metabolites, messengers, and conditions [bib_ref] Mitochondrial biosensors, De Michele [/bib_ref]. These biosensors generally consist of two basic components: substrate-binding proteins and one or two fluorescent proteins. Various transcription factors and regulatory proteins specifically sense intracellular biomolecules from bacteria to mammals. Substratesensing proteins often trigger conformational changes upon biomolecule binding, which induces fluorescence changes in fluorescent proteins. Fluorescence can be readily measured by routine instruments such as plate readers, flow cytometry or fluorescence microscopy.
To date, there are nine genetically encoded fluorescent sensors available for pyridine dinucleotides: the NAD + sensors LigA-cpVenus [bib_ref] Biosensor reveals multiple sources for mitochondrial NAD(+), Cambronne [/bib_ref] and FiNad ; the NADH sensor Frex [bib_ref] Genetically encoded fluorescent sensors for intracellular NADH detection, Zhao [/bib_ref] ; the NAD + /NADH ratio sensors Peredox [bib_ref] Imaging cytosolic NADH-NAD(+) redox state with a genetically encoded fluorescent biosensor, Hung [/bib_ref] , RexYFP [bib_ref] Genetically encoded fluorescent indicator for imaging NAD(+)/NADH ratio changes in different cellular..., Bilan [/bib_ref] , and SoNar [bib_ref] SoNar, a highly responsive NAD+/NADH sensor, allows high-throughput metabolic screening of anti-tumor..., Zhao [/bib_ref] [bib_ref] In vivo monitoring of cellular energy metabolism using SoNar, a highly responsive..., Zhao [/bib_ref] ; the NADP + sensors Apollo-NADP + [bib_ref] Apollo-NADP(+): a spectrally tunable family of genetically encoded sensors for NADP(+), Cameron [/bib_ref] and NADPsor [bib_ref] A genetically encoded biosensor for in vitro and in vivo detection of..., Zhao [/bib_ref] , and the NADPH sensor iNap [bib_ref] Genetically encoded fluorescent sensors reveal dynamic regulation of NADPH metabolism, Tao [/bib_ref] [bib_ref] Analysis of redox landscapes and dynamics in living cells and in vivo..., Zou [/bib_ref] [fig_ref] Table 1: Genetically encoded fluorescent sensors for NAD [/fig_ref]. Fluorescent biosensors must meet several criteria for live-cell and in vivo developmental studies, including high specificity, large responsiveness, appropriate affinity, strong brightness, and ratiometric readout, which allows reliable and convenient capture of subtle changes in physiological contexts. The majority of them can hardly meet these requirements due to their poor selectivity, relatively small dynamic response (i.e., RexYFP and Apollo-NADP + sensors), inappropriate affinity (i.e., Peredox and NADPsor sensors), weak fluorescence (i.e., Frex sensor), or intensiometric readout (i.e., Peredox sensor) [fig_ref] Table 1: Genetically encoded fluorescent sensors for NAD [/fig_ref].
The NADH/NAD + ratio sensor SoNar and the NADPH sensor iNap are superior with regard to the above criteria and are chosen to demonstrate how sensors are utilized to dynamically monitor metabolic states in living cells and live mice [bib_ref] SoNar, a highly responsive NAD+/NADH sensor, allows high-throughput metabolic screening of anti-tumor..., Zhao [/bib_ref]. SoNar was designed by inserting circularly permutated yellow fluorescent protein (cpYFP) into the truncated Thermus aquaticus T-Rex protein . The sensor has two excitation peaks, which enable an intrinsically ratiometric measurement. SoNar responds to the NAD + /NADH ratio but does not depend on either individual NAD + or NADH concentrations alone. It exhibits a large dynamic response up to 1500% and high brightness [bib_ref] SoNar, a highly responsive NAD+/NADH sensor, allows high-throughput metabolic screening of anti-tumor..., Zhao [/bib_ref]. The NADPH family sensor iNap was generated through rational designdirected protein engineering that switches SoNar's substrate selectivity from NADH to NADPH [bib_ref] Genetically encoded fluorescent sensors reveal dynamic regulation of NADPH metabolism, Tao [/bib_ref] , and it maintains the most superior properties of SoNar . The NADPH family sensor iNap has four members with different K d value, covering the most physiological concentration of NADPH in cells. The iNap1 member responds only to NADPH, with a rapid (less than 1 s) and large (900%) response. Due to their superior performances, SoNar and iNap sensors have been successfully used for metabolic studies on wound response in vivo [bib_ref] Genetically encoded fluorescent sensors reveal dynamic regulation of NADPH metabolism, Tao [/bib_ref] , hematopoiesis [bib_ref] MDH1-mediated malateaspartate NADH shuttle maintains the activity levels of fetal liver hematopoietic..., Gu [/bib_ref] , aging of intestinal stem cells [bib_ref] Warburg-like metabolic reprogramming in aging intestinal stem cells contributes to tissue hyperplasia, Morris [/bib_ref] , homing of leukemia-initiating cells [bib_ref] Metabolic imaging reveals a unique preference of symmetric cell division and homing..., Hao [/bib_ref] , immortalization of neural stem cells [bib_ref] Oxidative metabolism drives immortalization of neural stem cells during tumorigenesis, Bonnay [/bib_ref] , embryo development , tumorigenesis [bib_ref] The alternative activity of nuclear PHGDH contributes to tumour growth under nutrient..., Ma [/bib_ref] , ferroptosis [bib_ref] MESH1 is a cytosolic NADPH phosphatase that regulates ferroptosis, Ding [/bib_ref] , photosynthesis and photorespiration in Arabidopsis thaliana [bib_ref] In planta study of photosynthesis and photorespiration using NADPH and NADH/NAD(+) fluorescent..., Lim [/bib_ref] , and so on. These studies demonstrate that SoNar and iNap sensors are not only sensitive enough to measure the physiological changes of NADH and NADPH, but also powerful for interrogating their functions. The pH fluctuation may influence the response of the two sensors, and the effect can be corrected by normalization to iNapc, which has a similar response to pH as sensors but responds to neither NAD(H) nor NADPH .
## Real-time tracking of the cell cycle with sonar and inap1
Cells copiously proliferate through the mitotic cell cycle during organismal development. Although it is clearly known that cyclin-dependent events play central roles in the control of cell cycle progression, it is less understood how cells regulate their metabolism to meet varying demands during different mitotic cell cycling phases, such as maintenance of redox homeostasis and buildup of proteins, lipids, and deoxyribonucleic acids. Here, we demonstrate how to utilize SoNar and iNap1 to visualize NADH/NAD + and NADPH dynamics during the cell cycle progression of mammalian cells. To facilitate robust and long-term imaging, a coding DNA sequence for the SoNar or iNap sensor is delivered into HeLa cells via lentivirus infection, and HeLa cells stably expressing the sensor are then seeded into glass-bottom plates. Fluorescence imaging can be carried out either by a common automatic microscope equipped with an on-stage CO 2 incubator or a high-content imaging system. A perfect focus system helps the long-term tracking of cells of interest, and optimization of imaging parameters can minimize phototoxicity, including light intensity, exposure time, gain, time interval and use of high-sensitivity detectors. The NADH/NAD + level in HeLa cells did not show detectable changes during the cell cycle, while the NADPH level rose in the mitotic phase [fig_ref] Figure 3: Real-time tracking NADH/NAD + and NADPH dynamics during the cell cycle [/fig_ref]. The NADPH elevation is moderate and robust. Alterations in pentose phosphate pathway activities may account for the oscillating NADPH dynamics [bib_ref] APC/C (CDH1) synchronizes ribose-5-phosphate levels and DNA synthesis to cell cycle progression, Li [/bib_ref] [bib_ref] Polo-like kinase 1 coordinates biosynthesis during cell cycle progression by directly activating..., Ma [/bib_ref]. As a control, iNapc fluorescence did not significantly change during the cell cycle [fig_ref] Figure 3: Real-time tracking NADH/NAD + and NADPH dynamics during the cell cycle [/fig_ref]. This observation clearly exemplifies the power of sensors in capturing transient and subtle changes in metabolism, which can hardly be detected by other existing assays. During early development, eggs continuously cleave with pyruvate as the main nutrient until the morula stage; thus, it is intriguing to investigate whether NADPH fluctuates in this process and how it occurs. These sensors may be considered to solve these mysteries.
## Metabolic imaging of the embryonic development of zebrafish with sonar and inap1
Danio rerio is an excellent model organism for developmental research because its embryo is transparent and develops ex vivo. We chose zebrafish to establish and exemplify the application of sensors in studies of developmental metabolism. Both SoNar and iNap sensors have small molecular sizes and can be readily transferred into embryos through microinjection of sensor-coding DNA, mRNAs or purified sensor proteins. Using sensor nucleic acids allows for long-term and lineage-specific tracking, while using sensor proteins enables NAD(P)(H) monitoring immediately from fertilized eggs. For sensor proteins, less than ten nanograms of sterile protein is enough for microinjection into the animal pole of embryos at the one-or two-cell stage. For convenience of imaging, developing larvae are anesthetized with tricaine and dechorionated using tweezers. For imaging during the larval stage, melanization inhibitors such as N-phenylthiourea can be applied to prevent pigment formation. Fluorescence can be detected by a confocal laser scanning microscope system, and the use of supersensitive HyD hybrid detectors or light-sheet imaging systems should The design of NADH/NAD + sensor SoNar and NADPH sensor iNap. Both SoNar and iNap are cpYFP-based single fluorescent protein sensors. The control sensor iNapc has a similar response to pH as SoNar and iNap but respond to neither NAD(H) or NADPH. For SoNar and iNap, fluorescent protein cpYFP was inserted into a monomer of NAD + (NADH)-binding bacterial protein T-Rex. Binding of NADH (and NAD + ) and NADPH induces changes in protein conformation and fluorescence of SoNar and iNap sensors, respectively decrease phototoxicity. The majority of injected embryos develop into larvae without any noticeable abnormality. Fluorescence was relatively homogenous throughout the whole larval body, suggesting a uniform dispersion of sensor protein . Assays were carried out at either 24 h or 48 h after fertilization. The addition of 5 μM rotenone, a respiration complex I inhibitor, significantly increased the F405/F488 ratio of the SoNar sensor, whereas the treatment did not give rise to any response in the control fluorescent protein iNapC . Inhibition of mitochondrial respiration usually leads to a compensatory boosting of glycolysis and thus an elevation of cytosolic NADH/NAD + ; therefore, this result verified that SoNar is able to monitor NADH/NAD + dynamics in vivo in real time. Oxidative stress may perturb the NADPH pool by triggering cells to consume NADPH for survival. NADPH assays were also carried out at either 24 h or 48 h after fertilization [fig_ref] Figure 5: Monitoring NADPH dynamics by iNap1 in zebrafish embryos at 24 or 48... [/fig_ref]. Hydrogen peroxide at 50 mM rapidly decreased the F405/F488 ratio of iNap1 within 1-2 min and then recovered quickly [fig_ref] Figure 5: Monitoring NADPH dynamics by iNap1 in zebrafish embryos at 24 or 48... [/fig_ref] , suggesting rapid consumption and recovery of in vivo NADPH upon oxidative stress. The results show that mammalian cells have a strong tendency to maintain physiological NADPH homeostasis.
# Conclusions and perspectives
Differential gene expression is recognized as playing a central role in organismal development. Progression in sequencing technology, such as single-cell and spatial RNA sequencing, has led to an exciting flurry of epigenetic and transcriptomic studies on embryonic development [bib_ref] A singleembryo, single-cell time-resolved model for mouse gastrulation, Mittnenzweig [/bib_ref]. Metabolic After about 24 h, the culture medium was changed into phenol-red free medium for fluorescence imaging. The fluorescence microscopy system was equipped with a Nikon Eclipse Ti-E automatic microscope, monochrome-cooled digital microscope camera (model no. DS-Qi1 Mc-U2), PFS, A Plan Apo 40 × 0.95 NA objective, a highly stable Sutter Lambda XL light source, and an on-stage CO 2 incubator (Tokai Hit). For dual-excitation ratio imaging, 407-BP 17-nm or 482-BP 35-nm excitation filters (Semrock) and a 535-BP 40-nm emission filter altered by a Lambda 10-3 filter wheel (Sutter Instruments) were used. A perfect focus system was applied to stabilize the focal plane and high-sensitivity detectors were utilized to shorten the exposure time. The NADPH level exhibits a moderate and robust increase in the mitotic phase A-B, while the NADH/NAD + level remains almost constant C-D, as well as the iNapc fluorescence E-F. All fluorescence images were pseudocolored by F 407 /F 482 (R 407/482 ). Scale bar, 10 μm. For A-F, adapted from the reference [bib_ref] Analysis of redox landscapes and dynamics in living cells and in vivo..., Zou [/bib_ref] regulation of development, however, has been much less studied, which might be largely attributed to the great technological challenge of metabolite assays [bib_ref] Chemical embryology redux: metabolic control of development, Song [/bib_ref]. Namely, metabolites are not able to amplify in vitro, such as nucleic acids, and they usually demonstrate highly complicated spatiotemporal dynamics. Genetically encoded fluorescent sensors are a powerful tool for noninvasive metabolic monitoring. This technology achieves single-cell and subcellular resolution and real-time visualization of developmental processes; hence, it is much more costeffective and time-saving than single-cell sequencing analysis. The biological insights that sensors give are also unique and thought-provoking. For example, lactate and other sensors have recently discovered a glycolytic burst that is induced by cytoskeleton remodeling during endothelial contraction, and sensor-based visible evidence is comparably convincing [bib_ref] Single-cell metabolic imaging reveals a SLC2A3-dependent glycolytic burst in motile endothelial cells, Wu [/bib_ref].
Sensor-based assays have promising potential for developmental biology research. First, it can image cellular metabolism in situ during developmental and regeneration processes [bib_ref] Metabolic programming in early life in humans, Fall [/bib_ref] [fig_ref] Figure 6: Typical applications of genetically encoded metabolite sensors in developmental research [/fig_ref]. The preparation and delivery of sensors can be readily realized, and imaging is particularly convenient for model organisms with ex vivo development, such as Caenorhabditis elegans and Xenopus laevis [bib_ref] Worm-Book: the online review of Caenorhabditis elegans biology, Girard [/bib_ref]. To visualize mouse embryonic metabolism, a transgenic mouse line expressing a FRET pyruvate sensor was engineered, and embryos during the presomitic development stage were cultured ex vivo [bib_ref] Spatiotemporal analysis of a glycolytic activity gradient linked to mouse embryo mesoderm..., Bulusu [/bib_ref]. The sensor reports an increase in glycolytic activity in the posterior presomitic mesoderm compared to the anterior mesoderm. Noticing genetic differences of model organisms and environmental factors such as temperature, codon usage and experimental protocols need to be optimized to ensure the correct soluble expression and fluorophore maturation of sensor proteins. Second, Monitoring NADH/NAD + dynamics by SoNar in zebrafish embryos at 24 or 48 h post-fertilization. In vivo fluorescence imaging A-B and quantification C of zebrafish larvae (A, 24 hpf; B, 48 hpf ) expressing SoNar or iNapC in response to 5 μM rotenone (mitochondria complex I inhibitor) indicating regions of interest (white dashed line). Highly viable wild-type AB zebrafish embryos were collected at one or two-cell stage and 1 nL sensor protein in HEPES buffer was injected into the animal pole. Embryos were collected and placed in a Petri dish with egg water (60 μg/ mL Sea salts) at 28 °C. Larvae was anesthetized with 0.6 mM tricaine in E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl 2 , 0.33 mM MgSO 4 ) and then imaged via a Leica TCS SP8 SMD confocal laser scanning microscope with a HCX Plan APO CS 10/0.40 NA or HC Plan FLUOTAR 5/0.15 NA dry objective. Zebrafish larvae in 48 hpf were dechorionated under the stereomicroscope using fine tweezers. All fluorescence images were pseudocolored by F 405 /F 488 (R 405/488 ). Scale bar, 100 μm A or 200 μm B. The handling procedures were approved by the Institutional Animal Care and Use Committee of Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences. For A-C, adapted from the reference [bib_ref] Analysis of redox landscapes and dynamics in living cells and in vivo..., Zou [/bib_ref] sensors can be employed to study the metabolism of rare cell types or transient developmental stages [fig_ref] Figure 6: Typical applications of genetically encoded metabolite sensors in developmental research [/fig_ref]. It is of great challenge to acquire single-cell samples in these settings if single-cell analysis in vitro is applied [bib_ref] Single-cell (multi) omics technologies, Chappell [/bib_ref]. However, cis-acting regulatory elements such as lineage-and stage-specific promoters and enhancers may be used to drive sensor expression in a controllable manner, enabling precise metabolic monitoring of cell types and stages of interest. Third, sensors may be applied to dissect how metabolism is coordinated with cellular signaling and gene expression control to form a functional organ using organoids [fig_ref] Figure 6: Typical applications of genetically encoded metabolite sensors in developmental research [/fig_ref]. Recently, considerable progress has been made in organoid research to recapitulate the three-dimensional architectures and functions of organs [bib_ref] Human organoids: model systems for human biology and medicine, Kim [/bib_ref]. However, there are still many mysteries about the functions and underlying mechanisms of intracellular metabolites and intercellular communication. Take NAD + as an instance. In adult mice, NAD + can be synthesized from tryptophan in the liver and degraded into nicotinamide; then, nicotinamide is transported into other tissues for regeneration of NAD + . It is very intriguing to utilize organoids to investigate whether and how NAD + levels are coordinated between different developing cells. Finally, stem cells are ideal systems in which sensors may have good applications. For example, using sensors as reporters, chemical and genetic screens can be readily carried out to identify and quantitatively analyze metabolic regulators in stem cells [bib_ref] In vivo monitoring of cellular energy metabolism using SoNar, a highly responsive..., Zhao [/bib_ref] [fig_ref] Figure 6: Typical applications of genetically encoded metabolite sensors in developmental research [/fig_ref].
Given the close link of NAD + /NADH and NADP + / NADPH, it is biologically important to decipher their circadian dynamics in subcellular compartments including cytosol, nucleus and mitochondria, and in different organs relevant to rhythmic regulations such as brain, liver and endocrine glands. Recently, we have developed a novel NAD + metabolism sensor, named FiNad , and it allows us to integrate the redox toolbox (NAD + sensor FiNad, NADH sensor SoNar, NADP + sensor Apollo-NADP(+) [bib_ref] Apollo-NADP(+): a spectrally tunable family of genetically encoded sensors for NADP(+), Cameron [/bib_ref] , and NADPH sensor iNap) to characterize the circadian oscillation landscape and functions of core coenzymes. Further research is necessary to address this issue.
To date, however, sensors for important metabolites are lacking, including most amino acids, intermediate metabolites in glycolysis, pentose phosphate pathway and TCA, nucleotides, lipids and vitamins, and the development of high-performance sensors is highly desirable. A limitation for sensor-based metabolic assays is that information on only one or at most a few metabolites can be acquired at one time; thus, combination with multiomics technologies is necessary for comprehensively understanding the genetic and metabolic control of organismal development.
[fig] Figure 1: NAD(H) and NADP(H) play a vital role in cellular metabolism. NAD(P)(H) are key coenzymes for cellular metabolism. [/fig]
[fig] Figure 3: Real-time tracking NADH/NAD + and NADPH dynamics during the cell cycle. Fluorescence images A, C, E and quantification B, D, F of NADPH (iNap1) A-B, NADH (SoNar) C-D, and iNapc E-F dynamics during cell division. The entire coding sequences of iNap1, SoNar, iNapc were subcloned into the vector pLVX-IRES-Puro vector for the generation of recombinant lentivirus. HeLa cells with lentiviral infection and stable expression of sensors were then seeded into glass-bottom plates. [/fig]
[fig] Figure 5: Monitoring NADPH dynamics by iNap1 in zebrafish embryos at 24 or 48 h post-fertilization. In vivo fluorescence imaging A-B and quantification C of zebrafish larvae (A, 24 hpf; B, 48 hpf ) expressing iNap1 or iNapC in response to 50 mM H 2 O 2 indicating regions of interest (white dashed line). The Procedures were similar to that in Fig. 4. Scale bar, 100 μm A or 200 μm B. For A-C, adapted from the reference(Zou et al. 2018) [/fig]
[fig] Figure 6: Typical applications of genetically encoded metabolite sensors in developmental research. A Real-time tracking live-cell or in vivo metabolism in different developmental stages such as fertilization, egg cleavage, morphogenesis and organogenesis. B Illuminating metabolism of specific cells (i.e. rare cells in developmental stages) in situ. C Identifying functional metabolic shifts in organoids. D Screening metabolic regulators in culture cells with siRNA or sgRNA library [/fig]
[table] Table 1: Genetically encoded fluorescent sensors for NAD(H) and NADP(H)N.D., not determined a AXP denotes ADP and ATP b It is measured in the presence of 1 mM AXP c NADPH sensors iNap 1-4 have different affinities with K d values of ∼2.0 μM, ∼6.0 μM, ∼25 μM, and ∼120 μM, respectively d SoNar's and iNap's fluorescence excited at 420 nm, dynamic range, K NAD + /NADH and K NADPH are pH resistant e Venus-tagged Apollo-NADP + sensor was pH resistant in the pH range of 7.25-8, but showed a progressively decreasing dynamic range below pH 7.25 f The data come from side-by-side comparison studies(Tao et al. 2017;Zhao et al. 2015) [/table]
|
A network-assisted co-clustering algorithm to discover cancer subtypes based on gene expression
## Proof of convergence
We use the auxiliary function approach to prove the convergence of our algorithm.
Z(h, h ) is called an auxiliary function for F (h) if Z(h, h ) ≥ F (h) and Z(h, h) = F (h). Let h (t+1) = arg min h Z(h, h (t) ), then F (h (t+1) ) ≤ Z(h (t+1) , h (t) ) ≤ Z(h (t) , h (t) ) ≤ F (h (t) ).
Lemma 1. For any matrices U ∈ R k×k + , M, M ∈ R n×k + , if U is symmetric, the following inequality holds:
[formula] n i=1 k j=1 (M U ) ij M 2 ij M ij ≥ tr(M T M U ). (1) Proof. Let M ij = v ij M ij . LHS = n i=1 k j,l=1 M il U lj M ij v 2 ij = n i=1 k j,l=1 M il U lj M ij v 2 ij + v 2 il 2 . RHS = n i=1 k j,l=1 v il M il v ij M ij U jl = n i=1 k j,l=1 M il U lj M ij v ij v il . [/formula]
Therefore,
[formula] LHS − RHS = n i=1 k j,l=1 M il U lj M ij (v ij − v il ) 2 2 ≥ 0. For any real-valued matrix A, define A + = |A|+A 2 , A − = |A|−A 2 . Theorem 1. Let J(M ) = tr(−2P M T + M QM T ), [/formula]
where P ∈ R n×k and Q ∈ R k×k are fixed matrices, and Q is symmetric, M ∈ R n×n . Then
[formula] Z(M, M ) = −2 i,j P + ij M ij (1 + log M ij M ij ) + i,j P − ij M 2 ij + M ij 2 2M ij + i,j (M Q + ) ij M 2 ij M ij − i,j,l Q − jl M ij M il (1 + log M ij M il M ij M il )(2) [/formula]
is an auxiliary function of J(M ).
[formula] M ij = M ij P + ij + (M Q − ) ij P − ij + (M Q + ) ij (3) Proof. ∀x ∈ R + , 1 + log x ≤ x =⇒ P + ij M ij (1 + log M ij M ij ) ≤ P + ij M ij ; Q − jl M ij M il (1 + log M ij M il M ij M il ) ≤ Q − jl M ij M il a 2 + b 2 ≥ 2ab =⇒ P − ij M 2 ij + M ij 2 2M ij ≥ P − ij M ij . Lemma1 =⇒ (M Q + ) ij M 2 ij M ij ≥ tr(M T M Q + ). [/formula]
Therefore,
[formula] Z(M, M ) ≥ −2tr(P + M T ) + 2tr(P − M T ) + tr(M T M Q + ) − tr(M Q − M T ) = J(M ). [/formula]
To find the minimum of Z(M, M ), we take
[formula] ∂Z ∂M ij = −2P + ij M ij M ij + 2P − ij M ij M ij + 2 (M Q + ) ij M ij M ij − 2 (M Q − ) ij M ij M ij , ∂ 2 Z ∂M ij ∂M lm = δ il δ jm (2P + ij M ij M 2 ij + 2 P − ij M ij + 2 (M Q + ) ij M ij + 2 (M Q − ) ij M ij M 2 ij ) ≥ 0. Therefore, Z(M, M ) is a convex function of M . Let ∂Z ∂M ij = 0, we have M ij = M ij P + ij +(M Q − ) ij P − ij +(M Q + ) ij . [/formula]
Therefore, arg min M Z(M, M ) has entries M ij
[formula] P + ij +(M Q − ) ij P − ij +(M Q + ) ij . Let P = X T W GS = A, Q = S T GT W GS = B and M = F , we can see that updating F using F ij = F ij A + ij + (F B − ) ij A − ij + (F B + ) ij(4) [/formula]
monotonically decreases the value of the objective function J in the method part. Besides, we know that J ≥ 0, so the updating algorithm converges. Since W is a diagonal matrix with positive entries, we can similarly let P = XF S T = W −1 C, Q = SF T F S T = D and M = G, so updating G using
[formula] G ij = G ij C + ij + (W GD − ) ij C − ij + (W GD + ) ij = G ij (W −1 C) + ij + (GD − ) ij (W −1 C) − ij + (GD + ) ij(5) [/formula]
monotonically decreases the value of the objective function J, and as J ≥ 0, the updating algorithm converges.
# Supplementary method
Simulate gene expression based on network We designed a method to simulate gene expression data based on network interaction structure. We assumed that for sample i in subgroup k, gene expression x ·i ∼ N (µ k , Σ), where µ k is a column vector representing the mean expression levels for subgroup k, and Σ is used to model the network structure. To estimate Σ, we used graph Laplacian of the network. We first obtained an adjacency matrixà from the original network matrix E :à = max(E, E ), E is the transposal of E. The degree matrixD is defined as a diagonal matrix with entries equal to the sum of the corresponding rows ofÃ. The graph Laplacian is thus L =D −Ã. We estimated Σ as:
[formula] Σ = v(I −D − 1 2ÃD − 1 2 ) [/formula]
The main idea is that the expression levels of genes connected in the network structure are correlated and that the correlations are proportional to the proximities in the network. We used this technique to simulate datasets where interactions between genes can be considered. To determine a constant v, we compared the diagonal entries of matrix I −D − 1 2ÃD − 1 2 (expression variance) and the variance of gene expression levels. In our simulation, we set v = 0.5. VAMP8 VAV1 : Heatmap of GBM expression data. Rearranged according to our NCIS results. Genes listed are the 50 genes that are overlapped in the ordered ANOVA p-value list and the ordered gene weight list.
# Supplementary tables
[formula] GNAI1 SH3GL2 SH3GL3 SOX10 ARHGAP29 CASP8 CASP10 CEBPB JAK3 LRRFIP1 RELB RUNX2 SHC1 TNFRSF14 TPM4 EPHB1 CASK RAF1 CD151 CLIC1 FLNA GJA1 SSH3 ACVR2B APBA2 CSNK1E [/formula]
[fig] Furthermore: fixing M , Z(M,M') is a convex function of M and it has the global minimum at [/fig]
[fig] Figure S2, Figure S3: Kaplan-Meier survival curves of GBM data. Red for Subgroup Neural, green for Mesenchymal, blue for Proneural and purple for Classical; horizontal axis is the survival time (days) and vertical axis is the survival rate). a. NCIS (α = 0.85) defined subtypes; b. NCIS (α = 0) defined subtypes; c. Consensus clustering (k = 4) defined subtypes; d. GBM paper defined subtypes. Expression patterns of C1QA subnetwork in GBM subtypes. Genes directly connected to C1QA and genes targeting C1QA's downstream genes are included. Color of circle corresponds to gene expression level; size of circle corresponds to gene weight. a. Subtype Neural; b. Subtype Mesenchymal; c. Subtype Proneural; d. Subtype Classical. [/fig]
[table] Table S1: Cophenetic coefficients for BRCA data [/table]
[table] Table S2: Cophenetic coefficients for GBM data [/table]
[table] Table S3: P-value of the dependence test for different clinical features and GBM subtypes. For survival time, we used logrank test; for tumor necrosis percentage and tumor nuclei percentage, we used ANOVA. Although GBM paper (Verhaak et al. [/table]
|
The Functional Potential of Microbial Communities in Hydraulic Fracturing Source Water and Produced Water from Natural Gas Extraction Characterized by Metagenomic Sequencing
Microbial activity in produced water from hydraulic fracturing operations can lead to undesired environmental impacts and increase gas production costs. However, the metabolic profile of these microbial communities is not well understood. Here, for the first time, we present results from a shotgun metagenome of microbial communities in both hydraulic fracturing source water and wastewater produced by hydraulic fracturing. Taxonomic analyses showed an increase in anaerobic/ facultative anaerobic classes related to Clostridia, Gammaproteobacteria, Bacteroidia and Epsilonproteobacteria in produced water as compared to predominantly aerobic Alphaproteobacteria in the fracturing source water. The metabolic profile revealed a relative increase in genes responsible for carbohydrate metabolism, respiration, sporulation and dormancy, iron acquisition and metabolism, stress response and sulfur metabolism in the produced water samples. These results suggest that microbial communities in produced water have an increased genetic ability to handle stress, which has significant implications for produced water management, such as disinfection.
# Introduction
High-volume hydraulic fracturing operations for natural gas development from deep shale produce millions of gallons of wastewater over the lifetime of the well, [bib_ref] Evaluating the Environmental Implications of Hydraulic Fracturing in Shale Gas Reservoirs, SPE..., Arthur [/bib_ref] , [bib_ref] Water Management Challenges Associated with the Production of Shale Gas by Hydraulic..., Gregory [/bib_ref] , commonly termed as 'produced water'. This produced water contains elevated concentrations of salts, metals, hydrocarbons and radioactive elements [bib_ref] Water Management Challenges Associated with the Production of Shale Gas by Hydraulic..., Gregory [/bib_ref] , [bib_ref] Spatial and Temporal Correlation of Water Quality Parameters of Produced Waters from..., Barbot [/bib_ref] , [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref] ,, [bib_ref] Fractured Shale Gas Potential In New York, Hill [/bib_ref]. Microbial communities in produced water can utilize hydrocarbons as sources of carbon and energyand transform redox labile salts and metals. This can give rise to significant water management challenges [bib_ref] Microbial Communities in Flowback Water Impoundments from Hydraulic Fracturing for Recovery of..., Mohan [/bib_ref] and increased production costs [bib_ref] The impact of corrosion on oil and gas industry, Kermani [/bib_ref] ,. For instance, sulfidogenic and acid producing bacteria can cause corrosion of metal infrastructure, souring of natural gas, and reduced formation permeability [bib_ref] The impact of corrosion on oil and gas industry, Kermani [/bib_ref] ,, [bib_ref] Use of Microbiocides in Barnett Shale Gas Well Fracturing Fluids to Control..., Fichter [/bib_ref] ,.
Deleterious microbial activity is commonly controlled with biocides at significant cost to the driller. However, despite biocide use, microbial activity is prevalent in produced water. Previous studies have shown that biocide effectiveness may be limited by high salt concentrations, organic compounds, and long residence times in the subsurface [bib_ref] A critical assessment of the efficacy of biocides used during the hydraulic..., Struchtemeyer [/bib_ref] , [bib_ref] Bacterial communities associated with hydraulic fracturing fluids in thermogenic natural gas wells..., Struchtemeyer [/bib_ref] , [bib_ref] Deactivation of Industrial Water Treatment Biocides (Paper No. 10049), Williams [/bib_ref]. Other studies have shown that microbial communities in produced water are distinct from those in the injected fracturing fluid, and correlate well with changes in geochemical and environmental conditions [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref] , [bib_ref] Bacterial communities associated with hydraulic fracturing fluids in thermogenic natural gas wells..., Struchtemeyer [/bib_ref] , [bib_ref] Bacterial communities associated with production facilities of two newly drilled thermogenic natural..., Davis [/bib_ref]. This implies that the common practice of recycling produced water for subsequent hydraulic fracturing may introduce adapted populations into the formation [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref].
Over the past decade molecular ecology surveys based on the 16S rRNA gene have increased our knowledge about the taxonomic composition of microbial communities in reservoir environments [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref] , [bib_ref] Bacterial communities associated with hydraulic fracturing fluids in thermogenic natural gas wells..., Struchtemeyer [/bib_ref] , [bib_ref] Bacterial communities associated with production facilities of two newly drilled thermogenic natural..., Davis [/bib_ref] , [bib_ref] Microbial community structure analysis of produced water from a high-temperature North Sea..., Dahle [/bib_ref] , [bib_ref] Characterizing microbial diversity in production water from an Alaskan mesothermic petroleum reservoir..., Pham [/bib_ref] , [bib_ref] Microbial diversity in production waters of a low-temperature biodegraded oil reservoir, Grabowski [/bib_ref] , [bib_ref] Microbial diversityofan oil water processing site and its associated oil field: the..., Van Der Kraan [/bib_ref] , [bib_ref] Prokaryotic community structure and sulfate reducer activity in water from high-temperature oil..., Gittel [/bib_ref]. However, these studies offer limited insights on the metabolic capabilities of the microbial community, as they rely on taxonomic inference based on 16S rRNA gene similarity to previously isolated microorganisms. As an example of the limitations of using previously isolated microorganisms to infer metabolic capability, the 'core genome' of the well-studied Escherichia coli is typically less than 50% of the genes in the genome, and ,30% of the E. coli pan-genome [bib_ref] Order and disorder during Escherichia coli divergence, Hendrickson [/bib_ref]. On the other hand, shotgun metagenomic surveys enable access to complete genetic information within microbial genomes from uncultured, mixed consortia [bib_ref] Functional metagenomic profiling of nine biomes, Dinsdale [/bib_ref] , [bib_ref] Community genomics among stratified microbial assemblages in the ocean's interior, Delong [/bib_ref] , [bib_ref] Metagenomics: DNA sequencing of environmental samples, Tringe [/bib_ref]. These surveys have provided significant insights on the functional potential of microorganisms in diverse environments such as marine samples [bib_ref] Community genomics among stratified microbial assemblages in the ocean's interior, Delong [/bib_ref] , corals [bib_ref] Metagenomic analysis of the microbial community associated with the coral Porites astreoides, Wegley [/bib_ref] , activated sludge [bib_ref] Metagenomic and metatranscriptomic analysis of microbial community structure and gene expression of..., Yu [/bib_ref] , permafrost [bib_ref] The functional potential of high Arctic permafrost revealed by metagenomic sequencing, qPCR..., Yergeau [/bib_ref] , hydrocarbon and sandstone reservoirs [bib_ref] Metagenomics of Hydrocarbon Resource Environments Indicates Aerobic Taxa and Genes to be..., Dongshan [/bib_ref] , [bib_ref] Halomonas sulfidaeris-dominated microbial community inhabits a 1.8 km-deep subsurface Cambrian Sandstone reservoir, Dong [/bib_ref] , and swine gut [bib_ref] Comparative fecal metagenomics unveils unique functional capacity of the swine gut, Lamendella [/bib_ref]. Despite the importance of microbial activity in produced water brines from hydraulic fracturing operations, the functional potential of associated microbial communities has not yet been studied. In this study, the metagenome of fracturing source water and produced water at two different time points from a Marcellus Shale natural gas well in Westmoreland County, PA was generated using Illumina MiSeq technology. The microbial ecology from 16S rRNA surveys and chemical composition of these samples has been described in a previous publication [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref]. Sequences from each sample were assembled into contiguous sequences (contigs) and analyzed for taxonomic affiliations and functional potential of the microbial communities.
# Materials and methods
## Sampling
Samples of hydraulic fracturing source water, and produced water on days 1 and 9 were collected from a horizontally drilled Marcellus Shale natural gas well in Westmoreland County, Pennsylvania, U.S.A in October 2011. The source water used for fracturing was a mix of fresh reservoir water (,80%) and produced water (,20%) from previous fracturing operations. Fracturing additives amended to the source water included proppant (silica sand), scale inhibitor (ammonium chloride), biocide (mixture of tributyl tetradecyl phosphonium chloride, methanol and proprietary chemicals), hydrochloric acid, gel (paraffinic solvent), breaker (sodium persulfate) and friction reducer (hydrotreated petroleum distillate). Details regarding the sampling procedure and chemical additives used in the fracturing process are described elsewhere [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref]. The aqueous geochemical characteristics of these samples were described previously [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref].
## Dna extraction, library preparation and illumina sequencing
Unfiltered water samples were centrifuged at 6,000 g for 30 min in an Avanti J-E centrifuge (Beckman Coulter, Brea, CA) to pellet cells. DNA was extracted from 0.25 g of cell pellet using MO BIO power soil DNA isolation kit (MO BIO, Carlsbad, CA) according to the manufacturer's instructions. DNA was prepared using Nextera XT DNA sample preparation kit (Illumina, San Diego, CA) according to manifacturer's instructions at Genewiz (South Plainfield, NJ). DNA for sequencing was quantified using qPCR prior to clustering, and sequenced using the Illumina MiSeq (Illumina, San Diego, CA) with a 26250 PE configuration at Genewiz, NJ. Sequencing demultiplexing was performed on the Illumina MiSeq instrument using samplespecific barcodes.
## Bioinformatic analyses
The raw unpaired sequences were checked for sequencing tags and adapters using the predict function implemented within the TagCleaner program [bib_ref] TagCleaner: Identification and removal of tag sequences from genomic and metagenomic datasets, Schmieder [/bib_ref]. No sequencing tags or adapters were identified. Sequences were then subjected to quality control using the FastX toolkit within the Galaxy platform [bib_ref] Galaxy: A comprehensive approach for supporting accessible, reproducible, and transparent computational research..., Goecks [/bib_ref] with a minimum length 100 and minimum quality score 20. The velvet assembler [bib_ref] Using the Velvet de novo assembler for short-read sequencing technologies, Zerbino [/bib_ref] was used to assemble sequences that passed quality control into contiguous sequences. The assembly parameters were empirically optimized for the dataset prior to assembly [fig_ref] Table S2: Assembly optimization statistics [/fig_ref] ; the dataset was processed using a kmer length of 77. Generated contigs .500 bp in length were uploaded to the MG-RAST server [bib_ref] The metagenomics RAST server -a public resource for the automatic phylogenetic and..., Meyer [/bib_ref] with associated metadata files for taxonomic affiliations and functional annotations. Sequence similarity searches in MG-RAST was performed using the BLAT tool [bib_ref] Sul SJ, MGTAXA-A free software for taxonomic classification of metagenomic sequences with..., Kent [/bib_ref]. The metagenomes from fracturing source water, day 1 produced water, and day 9 produced water are available in the MG-RAST server [bib_ref] The metagenomics RAST server -a public resource for the automatic phylogenetic and..., Meyer [/bib_ref] [bib_ref] Classification with interpolated markov models, Brady [/bib_ref] , on the Galaxy bioinformatics workbench [bib_ref] Galaxy: A platform for interactive large scale genome analysis, Giardine [/bib_ref] , [bib_ref] Galaxy: A comprehensive approach for supporting accessible, reproducible, and transparent computational research..., Goecks [/bib_ref] , using default parameters and taxonomy as defined by the NCBI taxonomic tree. Data is for contig abundance and does not reflect read mapping.
As an additional assembly-independent analysis, sequence data was mapped against reference genomes downloaded from NCBI with CLC Genomics Workbench (Version 6.5.1, CLC Bio, Aarhus, Denmark)using default parameters and no masking. Reference genomes were selected based upon taxonomic observations in MG-RAST annotation and a previous microbial ecology investigation [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref]. Prior to mapping, sequencing data was trimmed to a minimum length of 100 bp and minimum quality score of 20. Furthremore, sequences for the sulfite reductase subunits A and B (dsrA/dsrB) and the suflur metabolism gene adenylyl sulfate reductase subunit A (apsA) were downloaded from NCBI and mapped against the trimmed sequencing data using CLC Genomics Workbench (Version 6.5.1, CLC Bio, Arhus, Denmark).
# Results and discussion
A total of 10 002, 17 055 and 16 661 contigs from the fracturing source water, produced water day 1 and day 9 samples, respectively, were uploaded to MG-RAST for downstream analyses. All uploaded contigs passed MG-RAST quality control and de-replication filters. The metagenomics sequence statistics are summarized in .
## Taxonomic composition
Taxonomic affiliations were assigned to contigs with predicted proteins and rRNA genes based on comparison with the M5NR database. Alpha diversity (predicted phylotypes) for the fracturing source water, produced water day 1 and day 9 samples were 90, 79 and 88, respectively [fig_ref] Figure 1: Class level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref]. Rarefaction curves for each of the samples were asymptotic suggesting that the majority of taxonomic diversity was recovered from the samples [fig_ref] Figure 1: Class level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref]. Alpha diversity values and rarefaction curves were obtained using the MG-RAST tool.
Bacteria constituted the dominant domain (97-99% of the total community) in all samples. However, a shift in bacterial community composition was detected between the samples at the class and order levels [fig_ref] Figure 1: Class level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref] , 2). Contigs affiliated to the class Alphaproteobacteria constituted the majority of the community in the fracturing source water (81%) and produced water day 1 (67%) samples [fig_ref] Figure 1: Class level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref]. Within Alphaproteobacteria, the dominant order detected was Rhodobacterales (68-88% of the Alphaproteobacteria; 55-59% of the total community) in both the source water and produced water day 1 samples [fig_ref] Figure 2: Order level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref]. The relative abundance of Alphaproteobacteria decreased to ,2% of the community in the produced water day 9 sample. Previous qPCR analysis of these samples suggests that that the total bacterial population remained constant at 10 6 -10 7 copies of 16S RNA gene/ml [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref].
An increase in the number of contigs associated with the class Clostridia was observed in the produced water day 1 sample (17%) as compared to the fracturing source water (1%). However, the relative abundance of Clostridia decreased to 3% in the produced water day 9 sample. The majority of the Clostridia in the produced water day 1 sample were affiliated to the order Thermoanaerobacterales (94% of Clostridia; 16% of the total community) [fig_ref] Figure 2: Order level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref]. Gammaproteobacteria sequences constituted a minor fraction (6%) of the total community in the fracturing source water and produced water day 1 samples but increased in relative abundance to constitute the dominant class (52%) in the produced water day 9 sample. Within the Gammaproteobacteria of the produced water day 9 sample, dominant orders included Vibrionales (67% of Gammaproteobacteria) and Alteromonadales (23% of Gammaproteobacteria) [fig_ref] Figure 2: Order level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref]. The day 9 samples also showed an increase in relative abundance of Epsilonproteobacteria (16%) and Bacteroidia (10%) classes as compared to the other samples (,2% of the total community). The major bacterial phyla, classes and orders identified in this study were consistent with previous 16S rRNA gene based clone library and pyrosequencing surveys of these samples [fig_ref] Figure 2: Order level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref] [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref]. These results indicate a shift towards facultative anaerobic/anaerobic and halophilic communities in the produced water samples as compared to a predominantly aerobic community in the fracturing source water. At the class level, in each of the samples less that 3% of the total sequences did not affiliate to any taxonomic group.
A minor fraction of the total community was represented by contigs affiliated to Archaea (0.1-0.4%), Viruses (0.3-1%) and Eukaryota (0.4-1.4%) domains. These domains were not analyzed for in the previous 16S rRNA gene survey of these samples [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref] , and were not considered in more detailed functional classification of the metagenomes. . Metagenomic sequence statistics of fracturing source water (SW), produced water day 1 (PW day 1) and produced water day 9 (PW day 9). Mapping results
Metagenomic reads were mapped against a diverse set of reference genomes to confirm MG-RAST taxonomic results and only reference genomes with good mapping results are discussed in this section. Reference genome mapping results confirmed taxonomic MG-RAST contig analysis. The best mapping results for source water were obtained when sequences were mapped against reference genomes of Alphaproteabacteria, specifically of the order Rhodobacterales [fig_ref] Figure 3: Fraction of genome coverage for source water [/fig_ref]. Similarly, produced water day 1 sample mapping results suggest that it was dominated by bacteria of the orders Rhodobacterales and Thermoanaerobacterales [fig_ref] Figure 3: Fraction of genome coverage for source water [/fig_ref]. A distinct shift in bacterial community was observed between produced water day 1 samples and produced water day 9 samples based on mapping results. Best mapping results for produced water day 9 samples were obtained for reference genomes in the order Campylobacterales and Alteromondales further supporting the MG-RAST results [fig_ref] Figure 3: Fraction of genome coverage for source water [/fig_ref]. Produced water samples demonstrated a distinctive signature with reads mapping best to few select reference genomes, while source water sample reads were distributed more evenly throughout all included reference genomes. For four reference genomes (Thermoanaerobacter sp. X514, Thermoanaerobacter pseudethanolicus, Thermoanaerobacter mathranii in produced water day 1 samples and Marinobacter hydrocarbonoclasticus DSM 7299 in produced water day 9 sample) more than 80% coverage was achieved suggesting that these species could play important roles in the microbial community of the representative sample [fig_ref] Figure 3: Fraction of genome coverage for source water [/fig_ref]. Highest observed reference genome coverage for source water sample sequences were 79% for Roseovarius sp. 217, 40% for Ruegeria pomeroyi and 38% for Rhodobacter sphaeroides [fig_ref] Figure 3: Fraction of genome coverage for source water [/fig_ref]. For produced water day 1 samples, about 10% of all trimmed sequencing reads mapped against the three Thermanaerobacter genomes included in the analysis and 8-13% of reads mapped successfully against Roseovarius sp. 217 and Roseovarius nubinhibens genomes. 7.7% of produced water day 1 reads mapped against the Ruegeria pomeroyi genome. 4-6% of reads for produced water day 9 samples mapped against two different Marinobacter and Arcobacter reference genomes and one Vibrio reference genome. Almost 16% of all reads from source water samples mapped against Roseovarius sp. 217 and approximately 4-6% of reads for source water sample mapped against each Dinoroseobacter shibae, Ruegeria pomeroyi, Rhodobacter sphaeroides and Rhodobacter capsulatus genomes. All mapping results are summarized in . The high number of reads form source water and produced water day 1 samples mapping against Roseovarius species is in agreement with previous 16S rRNA gene sequencing [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref] , implying the Roseovarius species might be of importance in these waters. Roseovarius sp. was previously identified in natural gas brines from the Marcellus shale and its potential implications are discussed elsewhere [bib_ref] Microbial Communities in Flowback Water Impoundments from Hydraulic Fracturing for Recovery of..., Mohan [/bib_ref].
The goal of this analysis was to provide an independent confirmation of MG-RAST results. Mapping results depend on the reference genomes selected and these reference genomes might . Read distribution for source water (SW), produced water day 1 (PW day 1) and produced water day 9 (PW day 9) samples. Reads were mapped against reference genomes using CLC Genomic workbench version 6.5.1 using default parameters. Shown are percentages of reads mapped against each reference genome included in the analysis for all three samples. doi:10.1371/journal.pone.0107682.g004 not be the same isolates found in the environment. While reference genomes for uncultured microorganisms from oil/gas environments are limited, the positive results achieved by this mapping analysis confirm the initial taxonomic assessment.
## Sulfur metabolism gene mapping results
Very few reads in all three samples were successfully mapped against the sulfur metabolism genes dsrA and dsrB. 7 reads of produced water day 1 sample and 55 reads of produced water day 9 sample were successfully mapped against the dsrA/dsrB gene of Desulfovibrio desulfuricans with a coverage of 28% and 78% respectively . In addition 10 reads of produced water day 9 sample were successfully mapped against the dsrA/dsrB gene of Desulfotignum balticum with a coverage of 19% . For aspA genes, the produced water day 9 sample showed best results with 16, 11, 9 and 6 reads successfully mapped against aspA genes of Desulfovirbio alaskensis, Desulfococcus mulitvorans, Desulfotignum balcticum and Desulfobacterium autotorphicum with a coverage of 94%, 46%, 33% and 31% respectively . Very few source water and produced water day 1 reads were mapped successfully against the aspA genes included in the analysis . These results suggest that sulfur metabolism could play a more important role in produced water day 9 sample due the higher abundance of genes associated with sulfur metabolism. Organisms that can metabolize sulfur compounds to sulfide are of interest in oil and gas environments because of their potential role in infrastructure corrosion, gas souring, worker safety as well as environmental health concerns.
## Functional classification of metagenomes
The SEED subsystems database [bib_ref] The subsystems approach to genome annotation and its use in the project..., Overbeek [/bib_ref] , was used to predict the metabolic potential of fracturing source water and produced water samples. Level 1 indicates the broadest set of functional categories to which sequences are assigned, and Level 2 refers to more specific functional assignments within Level 1 categories. The abundance of contigs designated to Level 1 functional categories is illustrated in [fig_ref] Figure 5: Actual abundance of contigs belonging to Level 1 functional categories in source... [/fig_ref]. The metabolic potential (based on Level 1 and Level 2 functional categories) between the samples was compared in a normalized manner [fig_ref] Figure 6: Normalized abundance [/fig_ref] to account for differences in community structure, size of the library, gene content between samples and to effectively compare low abundance functional categories [bib_ref] Integrated metatranscriptomic and metagenomic analyses of stratified microbial assemblages in the open..., Shi [/bib_ref]. Read normalization was performed within the MG-RAST analysis pipeline, in accordance with standards for metagenomic analysis.
The five most abundant Level 1 functional categories in all three samples were found to be clustering-based subsystems (e.g. genes where functional coupling is evident but function is unknown;,14%), carbohydrate metabolism (7-9%), amino acids and derivatives (7-8%), miscellaneous (eg: genes associated with iron sulfur cluster assembly and Niacine-Choline transport and metabolism; 8-9%), protein metabolism (6-8%), suggesting the dominant role of these functional categories in all samples [fig_ref] Figure 5: Actual abundance of contigs belonging to Level 1 functional categories in source... [/fig_ref]. These functional categories were similarly identified as dominant in previous studies of soil [bib_ref] Structure, fluctuation and magnitude of a natural grassland soil metagenome, Delmont [/bib_ref] , [bib_ref] Simultaneous assessment of soil microbial community structure and function through analysis of..., Urich [/bib_ref] , marine samples [bib_ref] Functional metagenomic profiling of nine biomes, Dinsdale [/bib_ref] , [bib_ref] Detection of large numbers of novel sequences in the metatranscriptomes of complex..., Gilbert [/bib_ref] , activated sludge [bib_ref] Functional metagenomic profiling of nine biomes, Dinsdale [/bib_ref] , freshwater [bib_ref] Functional metagenomic profiling of nine biomes, Dinsdale [/bib_ref] and hypersaline environments [bib_ref] Functional metagenomic profiling of nine biomes, Dinsdale [/bib_ref]. Normalization of gene abundance data shows a relative increase in each of the above functional categories in the produced water samples as compared to the fracturing source water [fig_ref] Figure 6: Normalized abundance [/fig_ref] implying that core systems necessary for survival are enriched in the produced water community.
While comparison of gene abundance affiliated with the dominant broad Level 1 categories suggests similar functional profiles across samples, analysis of more specific Level 2 functional categories shows sample specific differences in metabolic capabilities [fig_ref] Figure 7: Normalized abundance [/fig_ref]. Differences in metabolic potential indicate a selective pressure exerted in the subsurface for microbes with particular metabolic capabilities. For instance, within the Level 1 carbohydrate metabolism category, sequences related to Level 2 functional categories such as mono-, di-, oligo-and polysaccharides, and aminosugar metabolism were present in higher relative abundance in the produced water samples [fig_ref] Figure 7: Normalized abundance [/fig_ref]. This finding correlates well with the expected higher content of carbohydrates in produced water samples [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref]. Carbohydrates and polysaccharide compounds added during hydraulic fracturing can serve as carbon and energy sources for microbial activity. Within the Level 1 protein metabolism category, sequences affiliated with the Level 2 selenoprotein category were detected only in the produced water samples [fig_ref] Figure 7: Normalized abundance [/fig_ref]. One possible explanation is the role of selenoproteins in combating oxidative stress, which may arise from elevated concentrations of organic or inorganic dissolved constituents in produced water [bib_ref] Molecular biomarkers of oxidative stress in aquatic organisms in relation to toxic..., Valavanidis [/bib_ref]. Results showed that Rhodobacterales were the dominant population involved in oxidative stress response in source water and produced water day 1 samples [fig_ref] Figure 8: Taxonomic classification of oxidative stress contigs for each analyzed water sample as... [/fig_ref]. However, Alteromonadales and Vibrionales were the dominant orders involved in oxidative stress response in produced water day 9 sample [fig_ref] Figure 8: Taxonomic classification of oxidative stress contigs for each analyzed water sample as... [/fig_ref]. Within the Level 1 clustering subsystem, genes affiliated with the Level 2 carbohydrate metabolism show a relative increase in the produced water samples as compared to fracturing source water [fig_ref] Figure 7: Normalized abundance [/fig_ref]. An increase in the relative abundance of genes related to carbohydrate metabolism in produced water compared to fracturing source water suggests the potential for utilization of hydrocarbons added either as fracturing fluid amendments or those derived from the shale formation and an overall shift to a more heterotrophic microbial community. Less abundant Level 1 functional categories showing an increase in normalized abundance in produced water samples [fig_ref] Figure 6: Normalized abundance [/fig_ref] included genes affiliated with stress response (3%), respiration (3-4%), iron acquisition and metabolism (1%), sulfur metabolism (1%), and dormancy and sporulation (0.2-1%). Analysis of Level 2 functional categories within these Level 1 domains identified differences in metabolic potential between these samples [fig_ref] Figure 7: Normalized abundance [/fig_ref]. Within the Level 1 stress response domain, produced water samples showed a greater relative abundance of sequences affiliated with Level 2 categories such as acid stress, heat shock, periplasmic stress and osmotic stress [fig_ref] Figure 7: Normalized abundance [/fig_ref]. The increase in the relative abundance of these genes suggests a response to external stress experienced by the produced water microbial community. Results suggest that produced water day 1 population involved in osmotic stress response was dominated by the order Rhodobacterales and produced water day 9 population involved osmotic stress response was dominated by the orders Vibrionales and Alteromonadales [fig_ref] Figure 9: Taxonomic classification of osmotic stress contigs for each analyzed water sample as... [/fig_ref]. Subsurface stresses can include increased subsurface temperatures (.40uC), addition of HCl and biocides to fracturing fluid, and higher concentrations of dissolved salts [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref]. Within the Level 1 respiration category, sequences affiliated to the Level 2 category of sodium ion coupled energetics were undetected in fracturing source water (Na + 2.9 g/ L) but increased in relative abundance with time in produced water samples (Na + concentrations in PW day 1 and day 9 were 13.9 and 43 g/L) [fig_ref] Figure 7: Normalized abundance [/fig_ref]. This suggests that the produced water microbial community could use sodium ion coupled energetics for their energy needs, consistent with previous observations in saline environments [bib_ref] Bioenergetics of marine bacteria, Kogure [/bib_ref]. In the Level 1 domain of sulfur metabolism, the relative abundance of genes affiliated with Level 2 functional categories of inorganic and organic sulfur assimilation increased in produced water samples as compared to fracturing source water [fig_ref] Figure 7: Normalized abundance [/fig_ref]. Genes recovered from produced water day 1 show that populations involved in sulfur metabolism were dominated by the orders Rhodobacterales and Thermoanaerobacterales [fig_ref] Figure 1: Class level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref]. However, sulfur metabolism in produced water day 9 samples was dominated by the orders Vibrionales and Bacteroidales [fig_ref] Figure 1: Class level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref]. Within the Level 1 domain of iron metabolism, sequences affiliated with siderophores, undetected in the fracturing source water, increased with time in produced water samples [fig_ref] Figure 7: Normalized abundance [/fig_ref]. Siderophores are strong chelators of ferric iron secreted and are utilized by bacteria for iron metabolism [bib_ref] Microbial Iron Acquisition: Marine and Terrestrial Siderophores, Sandy [/bib_ref]. Relative increase in siderophore affiliated genes correlates with an increase in total iron concentrations with time in produced water (4.2-81.6 mg/L) . Within the Level 1 dormancy and sporulation category, high relative abundance of Level 2 spore DNA protection related sequences in produced water day 1 sample [fig_ref] Figure 7: Normalized abundance [/fig_ref] suggests the potential for long term dormancy of cells through DNA protection [bib_ref] Mini Review: I Will Survive: Protecting and Repairing Spore DNA, Setlow [/bib_ref]. BLAT analysis [bib_ref] Sul SJ, MGTAXA-A free software for taxonomic classification of metagenomic sequences with..., Kent [/bib_ref] showed that these genes were similar to those present in Thermoanaerobacter, a bacterial order that constituted 16% of the total community in this sample [fig_ref] Figure 2: Order level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref]. An increase in the relative abundance of spore forming bacteria and genes affiliated with sporulation and dormancy is an important consideration in biocide application, and may provide an explanation for the previously observed limited efficacy of biocides [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref].
Concluding Remarks. This study is the first shotgun metagenomic analysis of produced water from hydraulic fracturing for natural gas production and provides novel insights on taxonomic and functional potential of this pertinent yet unexplored environment. Taxonomic analysis showed that Bacteria constituted the dominant (.98%) domain in both fracturing source water and produced water samples. Results demonstrated the emergence of distinct bacterial classes and orders in the produced water samples and fracturing source water samples. These bacterial taxa were consistent with results from a previous 16S rRNA gene based survey of these samples [bib_ref] Microbial Community Changes in Hydraulic Fracturing Fluids and Produced Water from Shale..., Mohan [/bib_ref]. The metabolic profile showed both a relative increase and functional changes in genes responsible for carbohydrate metabolism, respiration, sporulation and dormancy, iron acquisition and metabolism, stress response and sulfur metabolism in the produced water samples as compared to the fracturing source water sample. These results suggest that the microbial community is responsive to changes in hydrocarbon content, induced stresses such as increase in temperature, addition of biocides, and an increase in concentration of dissolved salts such as iron and sulfur. The detection of genes affiliated with sodium ion coupled energetics exclusively in the produced water samples suggests the use of sodium ion based energetics by microorganisms in these sodium rich environments. Understanding the evolving metabolic capabilities of microbial communities in produced water will help the industry and its regulators improve environmental and economic sustainability of oil and gas extraction through more informed water management decisions. [fig_ref] Figure 1: Class level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref] Plot of refraction curves with associated Alpha diversity in fracturing source water (SW), produced water day 1 (PW day 1) and produced water day 9 (PW day 9). (TIF) [fig_ref] Figure 2: Order level affiliations assigned to contigs with predicted proteins and rRNA genes... [/fig_ref] Sequences affiliated to major bacterial phyla in source water, Produced water day 1 and Produced water day 9 using 16S rRNA gene pyrosequencing and metagenomics.
## Supporting information
## (tif)
Table S1 Chemical composition of source water and produced water (PW) samples days 1, 9 and 187. (TIF) Mapping results for source water, produced water day 1 and produced water day 9 sequencing data against selected bacteria species reference genomes. Mapping analysis was performed using CLC Genomics Workbench version 6.5.1 with default parameters. (TIF) Mapping results, (A), for source water, produced water day 1 and produced water day 9 sequencing data against the genome sequences of the dsrA/dsrB gene of selected microbial organisms. Mapping analysis was performed using CLC Genomics Workbench version 6.5.1 with default parameters. (B) Mapping results for source water, produced water day 1 and produced water day 9 sequencing data against the genome sequences of the apsA gene of selected microbial organisms. Mapping analysis was performed using CLC Genomics Workbench version 6.5.1 with default parameters. Author Contributions
[fig] Figure 1: Class level affiliations assigned to contigs with predicted proteins and rRNA genes in source water (SW), produced water day 1 (PW day 1) and produced water day 9 (PW day 9). Total community includes Bacteria, Archaea, Viruses and Eukaryota. doi:10.1371/journal.pone.0107682.g001 [/fig]
[fig] Figure 2: Order level affiliations assigned to contigs with predicted proteins and rRNA genes in source water (SW), produced water day 1 (PW day 1) and produced water day 9 (PW day 9). Total community includes Bacteria, Archaea, Viruses and Eukaryota. Only orders representing .2% of the total community are shown in the figure. doi:10.1371/journal.pone.0107682.g002 [/fig]
[fig] Figure 3: Fraction of genome coverage for source water (SW), produced water day 1 (PW day 1) and produced water day 9 (PW day 9) samples. Reads were mapped against reference genomes using CLC Genomic workbench version 6.5.1 using default parameters. Shown are fractions of reads mapped against each reference genome included in the analysis for all three samples. doi:10.1371/journal.pone.0107682.g003 [/fig]
[fig] Figure 4: Read distribution for source water (SW), produced water day 1 (PW day 1) and produced water day 9 (PW day 9) samples. Reads were mapped against reference genomes using CLC Genomic workbench version 6.5.1 using default parameters. Shown are percentages of reads mapped against each reference genome included in the analysis for all three samples. doi:10.1371/journal.pone.0107682.g004 [/fig]
[fig] Figure 5: Actual abundance of contigs belonging to Level 1 functional categories in source water (SW), produced water day 1 (PW day 1) and produced water day 9 (PW day 9). Functional annotations were assigned based on the Subsystems database. doi:10.1371/journal.pone.0107682.g005 [/fig]
[fig] Figure 6: Normalized abundance (values of 0-1) of contigs belonging to Level 1 functional categories in source water (SW), produced water day 1 (PW day 1) and produced water day 9 (PW day 9). Functional annotations were assigned based on the Subsystems database. doi:10.1371/journal.pone.0107682.g006 [/fig]
[fig] Figure 7: Normalized abundance (values of 0-1) of contigs belonging to selected Level 2 functional categories within associated Level 1 categories in source water (SW), produced water day 1 (PW day 1) and produced water day 9 (PW day 9).Functional annotations were assigned based on the Subsystems database. The affiliations of Level 2 categories to Level 1 categories are coded as follows CS-Clustering based subsystems; C-Carbohydrates; PM-Protein metabolism; DNA-DNA metabolism; R-Respiration; SR-Stress response; SM-Sulfur metabolism; DS-Dormancy and sporulation; IAM-Iron acquisition and metabolism. doi:10.1371/journal.pone.0107682.g007 [/fig]
[fig] Figure 8: Taxonomic classification of oxidative stress contigs for each analyzed water sample as assigned by MGTAXA. SW-Source water; D1-Produced water day 1; D9-Produced water day 9. Only the top six bacterial orders to which most contigs were assigned to are shown in the figure. The less abundant bacterial orders are grouped as ''other''. doi:10.1371/journal.pone.0107682.g008 [/fig]
[fig] Figure 9: Taxonomic classification of osmotic stress contigs for each analyzed water sample as assigned by MGTAXA. SW-Source water; D1-Produced water day 1; D9-Produced water day 9. Only the top four bacterial orders to which most contigs were assigned to are shown in the figure. The less abundant bacterial orders are grouped as ''other''. doi:10.1371/journal.pone.0107682.g009 [/fig]
[table] Table S2: Assembly optimization statistics. Velvet 1.2.08 was used to optimize assembly of Source Water derived sequences. (TIF) [/table]
|
Role of ACTH and Other Hormones in the Regulation of Aldosterone Production in Primary Aldosteronism
The major physiological regulators of aldosterone production from the adrenal zona glomerulosa are potassium and angiotensin II; other acute regulators include adrenocorticotropic hormone (ACTH) and serotonin. Their interactions with G-protein coupled hormone receptors activate cAMP/PKA pathway thereby regulating intracellular calcium flux and CYP11B2 transcription, which is the specific steroidogenic enzyme of aldosterone synthesis. In primary aldosteronism (PA), the increased production of aldosterone and resultant relative hypervolemia inhibits the renin and angiotensin system; aldosterone secretion is mostly independent from the suppressed renin-angiotensin system, but is not autonomous, as it is regulated by a diversity of other ligands of various eutopic or ectopic receptors, in addition to activation of calcium flux resulting from mutations of various ion channels. Among the abnormalities in various hormone receptors, an overexpression of the melanocortin type 2 receptor (MC2R) could be responsible for aldosterone hypersecretion in aldosteronomas. An exaggerated increase in plasma aldosterone concentration (PAC) is found in patients with PA secondary either to unilateral aldosteronomas or bilateral adrenal hyperplasia (BAH) following acute ACTH administration compared to normal individuals. A diurnal increase in PAC in early morning and its suppression by dexamethasone confirms the increased role of endogenous ACTH as an important aldosterone secretagogue in PA. Screening using a combination of dexamethasone and fludrocortisone test reveals a higher prevalence of PA in hypertensive populations compared to the aldosterone to renin ratio. The variable level of MC2R overexpression in each aldosteronomas or in the adjacent zona glomerulosa hyperplasia may explain the inconsistent results of adrenal vein sampling between basal levels and post ACTH
The major physiological regulators of aldosterone production from the adrenal zona glomerulosa are potassium and angiotensin II; other acute regulators include adrenocorticotropic hormone (ACTH) and serotonin. Their interactions with G-protein coupled hormone receptors activate cAMP/PKA pathway thereby regulating intracellular calcium flux and CYP11B2 transcription, which is the specific steroidogenic enzyme of aldosterone synthesis. In primary aldosteronism (PA), the increased production of aldosterone and resultant relative hypervolemia inhibits the renin and angiotensin system; aldosterone secretion is mostly independent from the suppressed renin-angiotensin system, but is not autonomous, as it is regulated by a diversity of other ligands of various eutopic or ectopic receptors, in addition to activation of calcium flux resulting from mutations of various ion channels. Among the abnormalities in various hormone receptors, an overexpression of the melanocortin type 2 receptor (MC2R) could be responsible for aldosterone hypersecretion in aldosteronomas. An exaggerated increase in plasma aldosterone concentration (PAC) is found in patients with PA secondary either to unilateral aldosteronomas or bilateral adrenal hyperplasia (BAH) following acute ACTH administration compared to normal individuals. A diurnal increase in PAC in early morning and its suppression by dexamethasone confirms the increased role of endogenous ACTH as an important aldosterone secretagogue in PA. Screening using a combination of dexamethasone and fludrocortisone test reveals a higher prevalence of PA in hypertensive populations compared to the aldosterone to renin ratio. The variable level of MC2R overexpression in each aldosteronomas or in the adjacent zona glomerulosa hyperplasia may explain the inconsistent results of adrenal vein sampling between basal levels and post ACTH Abbreviations: β-AR, β-adrenergic receptors; 5-HT4R, serotonin receptor; AC, adenylate cyclase; ACTH, adrenocorticotropin hormone; APA, aldosterone-producing adenoma; APA, aldosterone-producing adenoma; APCCs, aldosterone-producing cell clusters; AT-1 receptor, angiotensin II type 1 receptor; ATP, adenosine triphosphate; AVS, adrenal venous sampling; BAH, bilateral adrenal hyperplasia; cAMP, cyclic adenosine monophosphate; CREB, cAMP response element binding protein; CYP11B1, cytochrome P450 family 11 subfamily B member 1 encodes 11-beta hydroxylase; CYP11B2, cytochrome P450 family 11 subfamily B member 2 encodes aldosterone synthetase; ET, endothelin-1; FH, familial hyperaldosteronism; FST, fludrocortisone suppression test; GIP, glucose-dependent insulinotropic peptide; GPCR, G-protein-coupled receptor; GRA, glucocorticoid-remediable aldosteronism; Gs-α, G stimulatory α subunit; LH-hCG, luteinizing hormone/human chorionic gonadotropin; MC2R, melanocortin type 2 receptor; mRNA, messenger ribonucleic acid; PA, primary aldosteronsim; PAC, plasma aldosterone; PKA, protein kinase; RAS, rennin-angiotensin system; RT-PCR, reverese transcriptase-polymerase chain reaction; StAR, steroidogenic acute regulatory protein; TRH, thyrotropin releasing hormone; UA, unilateral adrenalectomy; V1-AVPR, vasopressin receptor type 1; ZF, zona fasciculata; ZG, zona glomerulosa.
iNTRODUCTiON Primary aldosteronism (PA) was first described in patients with unilateral aldosterone-producing adenomas [bib_ref] Presidential address. I. Painting background. II. Primary aldosteronism, a new clinical syndrome, Conn [/bib_ref]. It is characterized by increased aldosterone secretion causing salt retention, increased urinary potassium excretion, relative hypervolemia, suppressed plasma renin activity (PRA), and hypertension. PA is the most common curable form of secondary hypertension as it affects 4.3% of the general hypertensive population, 9.5% of patients referred to hypertension clinics [bib_ref] Prevalence of primary aldosteronism in patient's cohorts and in population-based studies -a..., Hannemann [/bib_ref] , and up to 20% of those with resistant hypertension [bib_ref] Is there an unrecognized epidemic of primary aldosteronism? Pro, Calhoun [/bib_ref]. PA is most often secondary to bilateral adrenal hyperplasia (BAH; 50-70% of cases) or to an aldosterone-producing adenoma (APA; 30-50% of cases) (4). The classical concept that a unique unilateral aldosteronoma is the causative lesion responsible for a high proportion of this surgically curable form of PA was recently challenged by the identification of zona glomerulosa (ZG) hyperplasia and nodulation adjacent to aldosteronomas when resected adrenals are examined carefully [bib_ref] Adrenocortical zonation in humans under normal and pathological conditions, Nishimoto [/bib_ref] [bib_ref] Clinicopathologic correlates of primary aldosteronism, Duan [/bib_ref].
In order to prevent cardiovascular, metabolic and renal morbidities, early diagnosis and management of PA are mandatory [bib_ref] Long-term cardio-and cerebrovascular events in patients with primary aldosteronism, Mulatero [/bib_ref] [bib_ref] Cardiovascular complications associated with primary aldosteronism: a controlled cross-sectional study, Savard [/bib_ref]. Unilateral adrenalectomy (UA) provides superior benefit compared to medical therapy in lateralized PA in terms of cardiovascular outcomes [bib_ref] Long-term cardiac effects of adrenalectomy or mineralocorticoid antagonists in patients with primary..., Catena [/bib_ref] [bib_ref] Quality of life in patients with bilateral primary aldosteronism before and during..., Ahmed [/bib_ref] , quality of life [bib_ref] Quality of life in patients with bilateral primary aldosteronism before and during..., Ahmed [/bib_ref] , and all-cause mortality [bib_ref] Observational study mortality in treated primary aldosteronism: the German Conn's registry, Reincke [/bib_ref] ; however, in BAH, pharmacological blockade of aldosterone excess using mineralocorticoid receptor antagonists, such as spironolactone or eplerenone, is the recommended treatment (4). Therefore, subtyping of PA is required to direct patients to surgical vs. medical therapy [bib_ref] Case detection, diagnosis, and treatment of patients with primary aldosteronism: an endocrine..., Funder [/bib_ref]. To date, adrenal vein sampling (AVS) is the gold standard to differentiate lateralized from bilateral sources of PA (4) because adrenal imaging provides poor specificity in detecting lateralized cases [bib_ref] Comparison of adrenal vein sampling and computed tomography in the differentiation of..., Magill [/bib_ref] except in patients <35 years old [bib_ref] Accuracy of adrenal imaging and adrenal venous sampling in predicting surgical cure..., Lim [/bib_ref].
In PA excess, plasma aldosterone concentration (PAC), despite suppressed renin activity, is not really autonomous, as frequently stated. It could be autonomous if it was solely or constitutively regulated by somatic and germline mutations of various ion channels genes regulating intracellular ionic homeostasis and cell membrane potential as reviewed elsewhere [bib_ref] Local control of aldosterone production and primary aldosteronism, Lalli [/bib_ref]. In fact, several autocrine/paracrine hormones and regulatory mechanisms [bib_ref] Autocrine/paracrine regulatory mechanisms in adrenocortical neoplasms responsible for primary adrenal hypercorticism, Lefebvre [/bib_ref] activate variable levels of aberrant eutopic or ectopic receptors [bib_ref] Screening for membrane hormone receptor expression in primary aldosteronism, Zwermann [/bib_ref] , which regulate aldosterone secretion either in unilateral adenomas or in BAH. In this review, we will focus on the role of one of these hormones, the adrenocorticotropic hormone (ACTH), in stimulating aldosterone secretion in normal and pathologic conditions and briefly mention others which play similar roles in PA.
## Normal physiology of the renin-angiotensin system
Renin is an enzyme produced primarily by the juxtaglomerular apparatus of the kidney and its release is the rate-limiting step in the regulation of the RAS [bib_ref] Historical perspective of the renin-angiotensin system, Hall [/bib_ref] [bib_ref] Control of aldosterone secretion: a model for convergence in cellular signaling pathways, Spat [/bib_ref] ; it is controlled by four factors: (1) the macula densa comprises chemoreceptors for monitoring the sodium and chloride loads present in the distal tubule, (2) juxtaglomerular cells acting as pressure transducers that sense stretch of the afferent arteriolar wall and thus renal perfusion pressure, (3) the sympathetic nervous system (SNS), which increases the release of renin, particularly in response to upright posture, in addition to (4) inhibiting factors, including K + , Ca ++ , angiotensin II, and atrial natriuretic peptides [bib_ref] Historical perspective of the renin-angiotensin system, Hall [/bib_ref].
The action of renin on angiotensinogen, synthesized in the liver, generates angiotensin I [bib_ref] Historical perspective of the renin-angiotensin system, Hall [/bib_ref]. Angiotensin-converting enzyme (ACE), localized in cell membranes particularly of the lung, cleaves angiotensin I into angiotensin II, which is the main biologically active angiotensin [bib_ref] Historical perspective of the renin-angiotensin system, Hall [/bib_ref]. Angiotensin II functions through the AT-1 receptor (AT1R) to maintain normal extracellular volume and blood pressure by increasing aldosterone secretion from the ZG via increased transcription of CYP11B2 (aldosterone synthase) [fig_ref] FiGURe 1 |: Mechanisms responsible for aldosterone synthesis in zona glomerulosa cells under normal physiological... [/fig_ref] as well as constricting vascular smooth muscle, releasing norepinephrine, and epinephrine from the adrenal medulla, enhancing the activity of the SNS and finally promoting the release of vasopressin [bib_ref] Historical perspective of the renin-angiotensin system, Hall [/bib_ref].
Zona glomerulosa cells are organized in rosette structures that spontaneously generate periodic depolarizing changes in membrane potential that are modulated in frequency by angiotensin II and extracellular K + [bib_ref] Zona glomerulosa cells of the mouse adrenal cortex are intrinsic electrical oscillators, Hu [/bib_ref] [bib_ref] Role of voltage-gated calcium channels in the regulation of aldosterone production from..., Barrett [/bib_ref] ; Angiotensin II induces cell membrane depolarization most probably due to a Gi-mediated shift in the voltage dependence of channel activation toward more negative potentials thereby increasing intracellular Ca 2+ signal, which stimulates hormone-sensitive lipase and steroidogenic acute regulatory protein (StAR). Another mechanism by which angiotensin II binding to AT1R stimulates aldosterone secretion implicates activating the phospholipase C/inositol 1,4,5-trisphosphate pathway, releasing Ca 2+ stores from the endoplasmic reticulum, and activation of T-type voltage-gated Ca 2+ channels (23) [fig_ref] FiGURe 1 |: Mechanisms responsible for aldosterone synthesis in zona glomerulosa cells under normal physiological... [/fig_ref].
Dopamine, atrial natriuretic peptide, and heparin inhibit aldosterone secretion. The secretion of aldosterone is restricted administration in the determination of source of aldosterone excess. In the rare cases of glucocorticoid remediable aldosteronism, a chimeric CYP11B2 becomes regulated by ACTH activating its chimeric CYP11B1 promoter of aldosterone synthase in bilateral adrenal fasciculate-like hyperplasia. This review will focus on the role of ACTH on excess aldosterone secretion in PA with particular focus on the aberrant expression of MC2R in comparison with other aberrant ligands and their GPCRs in this frequent pathology. The strongly negative resting membrane potential of zona glomerulosa (ZG) cells under resting physiological conditions is maintained by the concentration gradient of K + between the intracellular and extracellular space, which is generated by the activity of the Na + , K + -ATPase. Angiotensin II and increased K + lead to cell membrane depolarization, which opens voltage-dependent Ca 2+ channels. Furthermore, Angiostensin II acts through the Angiotensin II type 1 receptor (AT1R) inducing Ca 2+ release from the endoplasmic reticulum. Consequently, the increase in intracellular Ca 2+ concentration activates the calcium signaling pathway, which triggers activation of CYP11B2 transcription. The role for ACTH in the regulation of aldosterone secretion whether in normal physiology or in PA is in part determined by the level of expression of ACTH receptors (MC2R) in ZG cells. MC2R which is a GPCR coupled to the stimulatory Gsα subunit may induce an increase of intracellular cAMP concentration which activates protein kinase A thereby increasing CREB phosphorylation and CYP11B2 transcription. Aberrant expression of other GPCR may also be responsible for aldosterone excess despite a suppressed renin angiotensin system: eutopic GPCR include those for serotonin (5-HT4R); ectopic GPCR include those for glucose-dependent insulinotropic peptide (GIPR), luteinizing hormone/human chorionic gonadotropin (LH-hCG R), β-adrenergic receptors (β-AR), vasopressin (V1-AVPR) glucagon (glucagon receptor), TRH (TRH R), and Endothelin-1 ETA and ETB receptors. Other mechanisms implicated in PA involve somatic and germline mutations in ion channels genes regulating intracellular ionic homeostasis and cell membrane potentials: increase intracellular Na + concentrations and cell membrane depolarization result from KCNJ5 gain-of-function mutations affecting GIRK4 and ATP1A1 mutations of the Na + , K + -ATPase. Direct increase of intracellular Ca 2+ concentrations could also result from mutations in ATP2B3 encoding for the plasma membrane Ca 2+ -ATPase, mutations in CACNA1D affecting the Cav1.3 subunit of the L-type voltage-gated calcium channel or CACNA1H affecting the Cav3.2 subunit of the voltage-gated calcium channel. Finally dysregulation in cellular proliferation/apoptosis accelerating adenoma formation could be due either to activation of the Wnt/β-catenin pathway or to gene mutations such as ARMC5 although the mechanism of the latter mutation is not fully elucidated.
to the ZG because of zone-specific expression of aldosterone synthase (CYP11B2), which is regulated by the activation of calcium signaling [bib_ref] Adrenal corticosteroid biosynthesis, metabolism, and action, Arlt [/bib_ref].
## Pathophysiology of primary aldosteronism
The binding of free aldosterone to the mineralocorticoid receptor in the cytosol of epithelial cells [bib_ref] Adrenal corticosteroid biosynthesis, metabolism, and action, Arlt [/bib_ref] , principally in the kidney, controls potassium homeostasis and maintains normal intravascular volume by increasing intestinal and renal Na + and Cl − absorption and reabsorption, respectively. Increased production of aldosterone in PA results in sodium retention, hypertension, and can also result in hypokalemia [bib_ref] Control of aldosterone secretion: a model for convergence in cellular signaling pathways, Spat [/bib_ref]. In addition to the two most common subtypes of PA (BAH in 50-70% of the cases and APA in 30-50%), less frequent causes include primary (unilateral) adrenal hyperplasia (5%), aldosterone-producing adrenocortical carcinoma (<1%), familial hyperaldosteronism (1%), and ectopic aldosterone-producing adenoma or carcinoma (<0.1%). The mechanisms implicated in the pathophysiology of PA are not fully elucidated. Somatic and germline mutations in ion channels genes regulating intracellular ionic homeostasis and cell membrane potential were described in sporadic APA and type-III familial PA (25-28) [fig_ref] FiGURe 1 |: Mechanisms responsible for aldosterone synthesis in zona glomerulosa cells under normal physiological... [/fig_ref]. Somatic mutations in the potassium channel gene KCNJ5 are found in almost 30-40% of aldosteronomas and alter channel selectivity allowing enhanced Na + conductance. Na + influx results in cell depolarization, the activation of voltage-gated Ca 2+ channels, aldosterone production, and cell proliferation [bib_ref] K+ channel mutations in adrenal aldosterone-producing adenomas and hereditary hypertension, Choi [/bib_ref] [bib_ref] Prevalence, clinical, and molecular correlates of KCNJ5 mutations in primary aldosteronism, Boulkroun [/bib_ref]. Somatic and germline mutations in CACNA1D gene encoding a voltage-gated calcium channel result in channel activation and less depolarized potentials causing increased Ca 2+ influx, aldosterone production and cell proliferation in affected ZG cells [bib_ref] Somatic and germline CACNA1D calcium channel mutations in aldosterone-producing adenomas and primary..., Scholl [/bib_ref] [bib_ref] Hypertension with or without adrenal hyperplasia due to different inherited mutations in..., Scholl [/bib_ref]. Mutations in ATP1A1 gene (encoding the Na + /K + ATPase α subunit) and ATP2B3 gene (encoding the plasma membrane Ca 2+ ATPase) were identified in 5.2 and 1.6%, respectively of patients in a series of APA [bib_ref] Somatic mutations in ATP1A1 and ATP2B3 lead to aldosterone-producing adenomas and secondary..., Beuschlein [/bib_ref]. Mutations in CACNA1H gene, which encodes a voltage-gated calcium channel (Cav3.2) were discovered in children with PA; they result in impaired channel inactivation and activation at more hyperpolarized potentials, producing increased intracellular Ca 2+ and aldosterone excess [bib_ref] Recurrent gain of function mutation in calcium channel CACNA1H causes early-onset hypertension..., Scholl [/bib_ref]. Different mutations in the genes described above are found in different aldosterone-producing nodules from the same adrenal, suggesting that somatic mutations are independent events [bib_ref] 27: genetic abnormalities in lateralized multinodular primary aldosteronism, Fernandes-Rosa [/bib_ref] [bib_ref] Different somatic mutations in multinodular adrenals with aldosterone-producing adenoma, Fernandes-Rosa [/bib_ref].
No mutations of any of the above ion channel genes were found in BAH or in ZG hyperplasia adjacent to the dominant aldosteronomas [bib_ref] Somatic mutations in ATP1A1 and ATP2B3 lead to aldosterone-producing adenomas and secondary..., Beuschlein [/bib_ref] [bib_ref] Prevalence, clinical, and molecular correlates of KCNJ5 mutations in primary aldosteronism, Boulkroun [/bib_ref] [bib_ref] 27: genetic abnormalities in lateralized multinodular primary aldosteronism, Fernandes-Rosa [/bib_ref] [bib_ref] Different somatic mutations in multinodular adrenals with aldosterone-producing adenoma, Fernandes-Rosa [/bib_ref] ; these findings suggest that nodule formation and excess aldosterone production are two dissociated events, implying a two-hit hypothesis for APA formation [bib_ref] Local control of aldosterone production and primary aldosteronism, Lalli [/bib_ref]. The first hit causing a unilateral aldosteronoma or a dominant nodule adjacent to ZG hyperplasia may result from a somatic mutation in one of the genes described above, at least in approximately 60% of cases. Possible causes of the second hit that results in dysregulation in cellular proliferation/apoptosis accelerating adenoma formation could be due either to activation of the Wnt/β-catenin pathway [bib_ref] Aldosterone-producing adenoma formation in the adrenal cortex involves expression of stem/progenitor cell..., Boulkroun [/bib_ref] [bib_ref] The Wnt/beta-catenin pathway in adrenocortical development and cancer, El Wakil [/bib_ref] , PKA pathway, or to gene mutations such as ARMC5 (37) [fig_ref] FiGURe 1 |: Mechanisms responsible for aldosterone synthesis in zona glomerulosa cells under normal physiological... [/fig_ref]. However, the pathophysiology of progression from normal adrenal to APA and the causes of diffuse bilateral hyperplasias, either as BAH or in mild form adjacent to the dominant aldosteronoma, are still unknown. Aldosterone-producing cell clusters (APCCs), which have increased expression of CYP11B2, are nests of cells below the adrenal capsule. They protrude into cortisol-producing cells that are usually negative for CYP11B2 expression. Nishimoto et al. found that APCCs are common in normal adrenals, and they harbor a different mutational spectrum compared to APA suggesting that APCCs could be a precursor for APA [bib_ref] Aldosterone-stimulating somatic gene mutations are common in normal adrenal glands, Nishimoto [/bib_ref]. In addition, several hormones activating variable levels of eutopic or ectopic aberrant receptors (18) [fig_ref] FiGURe 1 |: Mechanisms responsible for aldosterone synthesis in zona glomerulosa cells under normal physiological... [/fig_ref] as well as autocrine and paracrine regulatory mechanisms [bib_ref] Autocrine/paracrine regulatory mechanisms in adrenocortical neoplasms responsible for primary adrenal hypercorticism, Lefebvre [/bib_ref] can increase aldosterone secretion in PA (either APA or BAH) independently from the suppressed RAS (see later section).
## Role of acth in aldosterone production in normal physiology
Adrenocorticotropic hormone can stimulate aldosterone secretion acutely and transiently under normal conditions, but to a lesser extent than angiotensin II and potassium. ACTH is a 39-amino-acid peptide, which results from the cleavage of its proopiomelanocortin (POMC) precursor by prohormone convertases PC1/3 and may be further cleaved by PC2 to generate α-melanocyte-stimulating hormone (α-MSH) [bib_ref] Proopiomelanocortin, a polypeptide precursor with multiple functions: from physiology to pathological conditions, Raffin-Sanson [/bib_ref] [bib_ref] Molecular evolution of GPCRs: melanocortin/melanocortin receptors, Dores [/bib_ref]. It is mainly produced in the anterior pituitary corticotropes, but is also produced in brain, adrenal medulla, skin, and placenta [bib_ref] Ectopic adrenocorticotropin (ACTH) and corticotropin-releasing hormone (CRH) production in the adrenal gland:..., Vrezas [/bib_ref] [bib_ref] The tissue-specific processing of pro-opiomelanocortin, Bicknell [/bib_ref] [bib_ref] Functional melanocortin-2 receptors are expressed by mouse aorta-derived mesenchymal progenitor cells, Evans [/bib_ref]. ACTH can induce aldosterone production at lower doses than the ones needed for cortisol and DHEA production [bib_ref] Responses of plasma adrenocortical steroids to low dose ACTH in normal subjects, Daidoh [/bib_ref]. Furthermore, ACTH stimulates aldosterone production acutely and sometimes chronically.
## Acute effects of acth
The initial binding of ACTH to its specific melanocortin type 2 receptor (MC2R) stimulates both cortisol and aldosterone secretion. MC2R (45) is a seven transmembrane domain receptor that belongs to the family of melanocortin receptors (MCRs) [bib_ref] The cloning of a family of genes that encode the melanocortin receptors, Mountjoy [/bib_ref].
Five MCRs constitute a distinct family of G-protein coupled hormone receptors (GPCR); MC2R is the smallest MCR and GPCR [bib_ref] The cloning of a family of genes that encode the melanocortin receptors, Mountjoy [/bib_ref] [bib_ref] Cloning and functional characterization of a family of receptors for the melanotropic..., Cone [/bib_ref]. MC2R is expressed in zona fasciculata (ZF) and ZG cells [bib_ref] Effects of melanocortins on adrenal gland physiology, Chan [/bib_ref]. The binding of ACTH to its MC2R induces the dissociation of Gs-α subunit and activation of adenylate cyclase (AC) that generates cAMP from ATP (48) [fig_ref] FiGURe 1 |: Mechanisms responsible for aldosterone synthesis in zona glomerulosa cells under normal physiological... [/fig_ref]. cAMP molecules bind to specific domains of the regulatory subunits of protein kinase A (PKA) thereby dissociating the tetramer and releasing the catalytic subunit (PRKACA) from its inactivating regulatory subunits. Activated PRKACA phosphorylates and activates steroidogenic acute regulatory protein (StAR) as well as cAMP response element binding protein (CREB), thereby increasing StAR expression. On the other hand, activation of the PKA pathway induces a slow but sustained calcium influx through the L-type calcium channels. The subsequent increase in intracellular calcium activates calcium/calmodulin-dependent protein kinase and steroidogenesis [bib_ref] Acute and chronic regulation of aldosterone production, Hattangady [/bib_ref] [bib_ref] Adrenal and extra-adrenal functions of ACTH, Gallo-Payet [/bib_ref].
## Chronic effects of acth
In contrast to in vivo studies that suggest that ACTH is a shortterm stimulator of aldosterone production, in vitro studies showed that ACTH can act as a major stimulus of aldosterone secretion. Continuous intravenous infusion of ACTH leads to a transient stimulation of aldosterone secretion, whereas its pulsatile administration leads to a sustained stimulation of aldosterone up to 72 h (51). Moreover, chronic exposure to ACTH (2 days or more) leads to transformation of ZG cells into ZF-like cells with elongated mitochondria with lamellar and tubular cristae becoming round with ovoid cristae; at the functional level, the synthesis of angiotensin II receptors, steroidogenic enzymes, and their products is altered (52-56).
## Role of acth in excess aldosterone secretion in pa diurnal rhythmicity of aldosterone
In recumbent normal subjects on a regular diet, the circadian rhythm of PAC is regulated by the activity of plasma renin independently of ACTH [bib_ref] Studies of the control of plasma aldosterone concentration in normal man. 3...., Williams [/bib_ref]. In contrast, patients with PA have a circadian rhythm of PAC mediated by changes in ACTH rather than by the suppressed plasma renin-angiotensin II levels (58). Several groups described that PAC falls following overnight sleep when ACTH levels are low despite upright posture or angiotensin II infusion. Similarly, they noted a marked increase in PAC shortly after ACTH administration (59-63), which was higher compared to normal controls or patients with essential hypertension [bib_ref] Plasma aldosterone response to ACTH in primary aldosteronism and in patients with..., Kem [/bib_ref] [bib_ref] Responses of aldosterone-producing adenomas to ACTH and angiotensins, Saruta [/bib_ref]. Furthermore, abolition of diurnal rhythm by dexamethasone in PA demonstrates the impact of ACTH on adrenal steroidogenesis [bib_ref] Role of endogenous ACTH on circadian aldosterone rhythm in patients with primary..., Sonoyama [/bib_ref]. Administration of dexamethasone 0.75-2.0 mg per day for 2 days decreased aldosterone levels by a mean of 49% in a group of 15 patients with aldosteronomas; in 33%, the suppression was greater than 80% (67).
## Acth role in familial hyperaldosteronism
Familial hyperaldosteronism (FH) type-1, previously known as glucocorticoid-remediable aldosteronism (GRA), was first described as a form of hyperaldosteronism relieved by dexamethasone [bib_ref] Hypertension, increased aldosterone secretion and low plasma renin activity relieved by dexamethasone, Sutherland [/bib_ref]. It is suspected in young PA patients whose relatives suffer from cerebrovascular accidents. It is an autosomal dominant disease whereby the promoter of the chimeric 11β-hydroxylase/aldosterone synthase gene belongs to the 5′ end of CYP11 B1 (11β hydroxylase) and drives the expression of the 3′ end of CYP11 B2 (aldosterone synthase) ectopically in ZF cells under the main regulation by ACTH (69); in these patients, dexamethasone usually decreases aldosterone secretion by more than 80% or to <4 ng/dL [bib_ref] Evaluation of the dexamethasone suppression test for the diagnosis of glucocorticoid-remediable aldosteronism, Litchfield [/bib_ref] , but the diagnosis is now performed using genetic analysis. In contrast to FH type-1, FH type-2 is defined as PA in a patient with a first-degree relative (parent/sibling/offspring) with established PA but without FH type-1 gene rearrangement. Linkage analysis has mapped FH type-2 to chromosome 7p22 but no responsible gene has been identified yet [bib_ref] Further evidence for linkage of familial hyperaldosteronism type II at chromosome 7p22..., Sukor [/bib_ref]. The prevalence of FH type-2 in PA is higher (1.2-6%) than FH type-1 (≤1%). The FH type-3 and -4 are not regulated by ACTH stimulation, but they are caused by germline mutations in KCNJ5 (71) and CACNA1D/CACNA1H [bib_ref] Hypertension with or without adrenal hyperplasia due to different inherited mutations in..., Scholl [/bib_ref] [bib_ref] Recurrent gain of function mutation in calcium channel CACNA1H causes early-onset hypertension..., Scholl [/bib_ref] genes, respectively.
## Acth suppression or stimulation tests can reveal the presence of pa
Based on the rationale that ACTH plays a more important role in PA than in normal subjects or those with essential hypertension, investigators in Athens compared the classical saline infusion test (SIT) to postdexamethasone SIT in 151 patients with single adrenal adenomas and detected almost double rate of aldosterone hypersecretion following dexamethasone administration (24 vs. 12%) [bib_ref] High prevalence of autonomous cortisol and aldosterone secretion from adrenal adenomas, Piaditis [/bib_ref]. Similarly, they used a combined fludrocortisone-dexamethasone suppression test (FDST), which is a modification of the classic confirmatory fludrocortisone suppression test (FST) for the diagnosis of PA; it involves the administration of dexamethasone to hypertensives patients at midnight of the last day of the FST in order to eliminate the stimulatory effect of ACTH on aldosterone secretion. They demonstrated that the prevalence of PA rises from 5 to 13% with the usual diagnostic tests to 28.7-31% when using FDST; mineralocorticoid receptor blockade resulted in significant improvement in blood pressure in these patients [bib_ref] High prevalence of autonomous aldosterone secretion among patients with essential hypertension, Gouli [/bib_ref] [bib_ref] Evidence of primary aldosteronism in a predominantly female cohort of normotensive individuals:..., Markou [/bib_ref] [bib_ref] Primary aldosteronism in hypertensive patients: clinical implications and target therapy, Papanastasiou [/bib_ref]. The same group administered an ultralow-dose (0.03 μg) ACTH to 113 hypertensives without PA: the 30 patients (27%) who exhibited an aldosterone hyperresponse had significantly higher PAC, ARR, and PAC/ACTH ratio in the treadmill test; normalization of blood pressure by mineralocorticoid antagonists in these patients was also evident compared to the group of hypertensive not sensitive to ACTH/stress [bib_ref] Stress-induced aldosterone hyper-secretion in a substantial subset of patients with essential hypertension, Markou [/bib_ref]. Therefore, the benefit of mineralocorticoid blockade could extend even to hypertensive patients without confirmed PA who present an aldosterone hyperreseponse to ACTH/stress, this category of hypertenives harboring a mild form of BAH. In contrast, another group examining the diagnostic accuracy of ACTH test in 158 hypertensive patients found that it was not very effective in differentiating between APA patients and non-PA patients [bib_ref] Evaluation of the (1-24) adrenocorticotropin stimulation test for the diagnosis of primary..., Terui [/bib_ref].
## Use of acth to identify the source of aldosterone excess
Many efforts were conducted to find an easier and cheaper test than adrenal vein sampling (AVS), which is available only in tertiary care center with experienced angioradiologists to distinguish between lateralized and bilateral sources of aldosterone. Differential increase in PAC during upright posture was suggested to be a valuable tool to distinguish APA from BAH, but further studies showed that several APA and BAH had similar rise in PAC to upright posture [bib_ref] Angiotensin-responsive aldosterone-producing adenomas: postoperative disappearance of aldosterone response to angiotensin, Tunny [/bib_ref]. APA whether responsive or not to angiotensin II was found to be more sensitive to ACTH stimulation resulting in larger increase of PAC than in patients with BAH or essential hypertension [bib_ref] Aldosterone regulation in primary aldosteronism: differences between adenoma and bilateral hyperplasia, Mantero [/bib_ref] [bib_ref] Plasma aldosterone response to ACTH in subtypes of primary aldosteronism, Stowasser [/bib_ref] [bib_ref] Significance of adrenocorticotropin stimulation test in the diagnosis of an aldosterone-producing adenoma, Sonoyama [/bib_ref]. BAH patients also displayed increased response of PAC to ACTH administration than normal subjects or patients with essential hypertension [bib_ref] Identification and implications of new types of mineralocorticoid hypertension, Biglieri [/bib_ref]. PAC increased more after ACTH bolus in the APA group compared with BAH group, which had an intermediate increase compared to normal controls [bib_ref] Screening for membrane hormone receptor expression in primary aldosteronism, Zwermann [/bib_ref] [bib_ref] Role of endogenous ACTH on circadian aldosterone rhythm in patients with primary..., Sonoyama [/bib_ref]. A study in which patients received dexamethasone (1 mg) the evening before receiving i.v. injection of 50 IU of ACTH showed that the exaggerated PAC response was higher after 120 min in patients with APA than in BAH [bib_ref] Diagnostic value of ACTH stimulation test in determining the subtypes of primary..., Jiang [/bib_ref]. It was suggested that this could be used for identifying the etiology subtype; however, significant overlap was present between APA, unilateral hyperplasia, and BAH cases and using an optimal cutoff value of the aldosterone >78 ng/dL for APA, provided a sensitivity of 76.8% and a specificity of 87.2% (83).
Kline et al. studied 65 patients with confirmed PA who were divided by histology into confirmed lateralized and nonlateralized; PAC in inferior vena cava (IVC) sampled during AVS before and after cosyntropin infusion was analyzed. Baseline and peak IVC aldosterone was higher in lateralized patients (APA) but incremental aldosterone rise was much greater in subjects with bilateral hyperplasia [bib_ref] A marked proportional rise in IVC aldosterone following cosyntropin administration during AVS..., Kline [/bib_ref]. This shows that ACTH can regulate APA as well as BAH, but that the effects are more pronounced in APA.
## Role of acth stimulation during adrenal venous sampling
The usefulness of ACTH stimulation in the conduct of AVS procedure is controversial and remains a matter of debate because of conflicting results. Some centers use cosyntropin infusion or bolus in order to minimize stress-induced or spontaneous fluctuations in aldosterone secretion when performing sequential non-simultaneous AVS, to maximize the gradient of cortisol from the adrenal vein to the inferior vena cava, and to maximize aldosterone secretion from an APA [bib_ref] What are the keys to successful adrenal venous sampling (AVS) in patients..., Young [/bib_ref]. In contrast, other groups found that ACTH-stimulation of aldosterone production from the contralateral gland or adjacent hyperplasia may reduce the gradient of aldosterone production resulting in incorrect lateralization [bib_ref] Adrenocorticotropic hormone stimulation during adrenal vein sampling for identifying surgically curable subtypes..., Seccia [/bib_ref] [bib_ref] Basal and post ACTH aldosterone and its ratios are useful during adrenal..., El Ghorayeb [/bib_ref]. The effect of both continuous ACTH infusion and bolus on the performance and interpretation of AVS in confirmed PA patients was investigated [bib_ref] Effect of adrenocorticotropic hormone stimulation during adrenal vein sampling in primary aldosteronism, Monticone [/bib_ref]. Both methods lead to a significant increase in selectivity index for the right adrenal vein and ACTH bolus for the left adrenal vein. Lateralization index was not significantly affected after continuous ACTH infusion and i.v. bolus. In 88 and 78% of the patients, the diagnosis obtained was the same before and after ACTH infusion and i.v. bolus, respectively [bib_ref] Effect of adrenocorticotropic hormone stimulation during adrenal vein sampling in primary aldosteronism, Monticone [/bib_ref]. Recently, our group demonstrated that ACTH increased selectivity on both sides from 66.7% in basal samples to 91.8% poststimulation. A discordance of lateralization between basal and post-ACTH values was observed in 28% of cases, mostly lateralized cases basally that became bilateral post ACTH [bib_ref] Basal and post ACTH aldosterone and its ratios are useful during adrenal..., El Ghorayeb [/bib_ref]. The variation in the response to ACTH stimulation could be due to the variable expression of MC2R in APA (see later) [bib_ref] Screening for membrane hormone receptor expression in primary aldosteronism, Zwermann [/bib_ref]. Careful examination of the levels of aldosterone in the adrenal vein contralateral to the dominant or lateralized APA and pathology confirmed the frequent presence of bilateral background hyperplasia and this could predict less favorable post-operative outcome with residual hypertension [bib_ref] Aldosterone-producing adenoma formation in the adrenal cortex involves expression of stem/progenitor cell..., Boulkroun [/bib_ref] [bib_ref] Basal and post ACTH aldosterone and its ratios are useful during adrenal..., El Ghorayeb [/bib_ref].
## Increased but variable expression of mc2r in pa
The explanation for the increased role of ACTH in the regulation of aldosterone in PA may be secondary to the overexpression and function of MC2R in this condition. The expression of MC2R mRNA was shown to be upregulated in human adrenocortical neoplasms specifically in functional adenomas in contrast to nonfunctioning adenomas and carcinomas [bib_ref] Expression of adrenocorticotrophic hormone receptor mRNA in human adrenocortical neoplasms: correlation with..., Reincke [/bib_ref]. More specifically, a few studies have demonstrated increased eutopic expression of MC2R assessed by RT-PCR or transcriptome studies in resected aldosteronomas as compared to cortisol-secreting adenomas, non-functional adenomas, and adrenocortical carcinomas [bib_ref] ACTH receptor mRNA in human adrenocortical tumors: overexpression in aldosteronomas, Arnaldi [/bib_ref] [bib_ref] Angiotensin II type 1 receptor and ACTH receptor expression in human adrenocortical..., Schubert [/bib_ref] [bib_ref] G-protein-coupled receptors in aldosterone-producing adenomas: a potential cause of hyperaldosteronism, Ye [/bib_ref]. A particularly pertinent informative study included 15 adrenal tumors (14 APA and 1 BAH); MC2R mRNA levels were increased by a mean of 3.9-fold in those tissues compared to normal adrenal [bib_ref] Screening for membrane hormone receptor expression in primary aldosteronism, Zwermann [/bib_ref]. However great variability existed in the level of expression in each tumor as 4 had lower levels than normal (0.3-fold to 0.7-fold), while those with increased expression varied between 1.4-and 20.6-fold compared to normal. The data are limited to mRNA expression without available measurements at the protein levels (no good specific MC2R antibody), but correlated well with the in vivo increased response to ACTH administration. There is almost no data on MC2R expression in BAH as those patients are usually not operated, but in the only case with BAH studied by this group MC2R was 20-fold increased. These data appear to be compatible with the findings that the majority of patients stimulated with ACTH during AVS will have concordant results before and after ACTH as the majority overexpress MC2R; however, the 28% of discordant results we found (87) may be explained by cases where MC2R are relatively decreased in the dominant adenoma but is present in adjacent hyperplasia. This hypothesis remains to be validated in prospective studies.
## Other hormones and aberrant receptors regulating aldosterone secretion in primary aldosteronism
Adrenocorticotropin hormone is not the exclusive trigger of aldosterone secretion since several other hormones have a role in the pathophysiology of PA in addition to the ion channels mutations. Serotonin plays a significant role in aldosterone synthesis in normal physiological and in PA. The administration of serotonin 5-HT4 agonists such as metoclopramide, cisapride, and tegaserod resulted in higher stimulation of aldosterone in PA as compared to the physiological moderate increase in normal individuals [bib_ref] Screening for membrane hormone receptor expression in primary aldosteronism, Zwermann [/bib_ref] [bib_ref] Characterization of serotonin(4) receptors in adrenocortical aldosterone-producing adenomas: in vivo and in..., Lefebvre [/bib_ref] [bib_ref] Expression profile of serotonin4 (5-HT4) receptors in adrenocortical aldosterone-producing adenomas, Cartier [/bib_ref]. Whereas non-specific inhibitors of 5-HT such as cyproheptadine and ketanserin produced only minor and transient inhibition of aldosterone secretion in aldosteronomas [bib_ref] Suppression of aldosterone by cyproheptadine in idiopathic aldosteronism, Gross [/bib_ref] [bib_ref] Effect of ketanserin in primary aldosteronism, Mantero [/bib_ref] , specific 5-HT4R antagonists such as GR113808 were potent inhibitors of basal-and cisapride-induced aldosterone secretion [bib_ref] Characterization of serotonin(4) receptors in adrenocortical aldosterone-producing adenomas: in vivo and in..., Lefebvre [/bib_ref]. Chromaffin cells, endothelial cells, nerve terminals, and cells of the immune system are localized in the immediate vicinity of ZG cells and can secrete various factors to control aldosterone secretion [bib_ref] Intraadrenal interactions in the regulation of adrenocortical steroidogenesis, Ehrhart-Bornstein [/bib_ref]. Local release of 5-HT by perivascular mast cells (MC) can activate 5-HT4R expressed in ZG cells and consequently stimulate aldosterone production [bib_ref] Production and metabolism of serotonin (5-HT) by the human adrenal cortex: paracrine..., Lefebvre [/bib_ref]. A role of MC in tumorigenesis was proposed [bib_ref] Mast cells are required for angiogenesis and macroscopic expansion of Myc-induced pancreatic..., Soucek [/bib_ref] [bib_ref] Mast cells in tumor growth: angiogenesis, tissue remodelling and immune-modulation, Maltby [/bib_ref]. The density of MC was found to be increased in APA tissues compared with normal adrenals [bib_ref] Mast cell hyperplasia is associated with aldosterone hypersecretion in a subset of..., Duparc [/bib_ref]. As the 5-HT4R have been found to be overexpressed in the majority of APA (but variable as MC2R) [bib_ref] Autocrine/paracrine regulatory mechanisms in adrenocortical neoplasms responsible for primary adrenal hypercorticism, Lefebvre [/bib_ref] [bib_ref] Screening for membrane hormone receptor expression in primary aldosteronism, Zwermann [/bib_ref] [bib_ref] Regulation of aldosterone secretion by several aberrant receptors including for glucose-dependent insulinotropic..., Lampron [/bib_ref] and the ligand may be locally overexpressed also, a paracrine loop of regulation of aldosterone production appears to be present.
The compelling evidence supporting that various aberrant GPCR are frequently expressed in bilateral macronodular adrenal hyperplasia and Cushing's syndrome (103) led many researchers to investigate the presence of aberrant GPCR in PA. Adrenal production of aldosterone in APA and BAH was found to be under the influence of aberrant GPCR and their ligands, as demonstrated by in vivo and in vitro studies [bib_ref] Aberrant G-protein coupled receptor expression in relation to adrenocortical overfunction, Lacroix [/bib_ref] [bib_ref] Aberrant hormone receptors in primary aldosteronism, Mazzuco [/bib_ref]. The expression of ectopic receptors, which are usually not expressed at significant levels in normal ZG cells include those for glucose-dependent insulinotropic peptide (GIPR) (106), luteinizing hormone/ human chorionic gonadotropin (LH-hCG R) [bib_ref] Screening for membrane hormone receptor expression in primary aldosteronism, Zwermann [/bib_ref] [bib_ref] Increase in gamma-globin mRNA content in human erythroid cells treated with angelicin..., Lampronti [/bib_ref] [bib_ref] Estradiol-and progesterone-related increases in the renin-aldosterone system: studies during ovarian stimulation and..., Sealey [/bib_ref] [bib_ref] Variation in the renin angiotensin system throughout the normal menstrual cycle, Chidambaram [/bib_ref] [bib_ref] Elevated expression of luteinizing hormone receptor in aldosteroneproducing adenomas, Saner-Amigh [/bib_ref] [bib_ref] Relationship between aldosterone and progesterone in the human menstrual cycle, Szmuilowicz [/bib_ref] [bib_ref] A case of primary aldosteronism in pregnancy: do LH and GNRH receptors..., Albiger [/bib_ref] [bib_ref] LH, progesterone, and TSH can stimulate aldosterone in vitro: a study on..., Nicolini [/bib_ref] , β-adrenergic receptors (β-AR) [bib_ref] Propranolol therapy for ectopic beta-adrenergic receptors in adrenal Cushing's syndrome, Lacroix [/bib_ref] , vasopressin (V1-AVPR) [bib_ref] Screening for membrane hormone receptor expression in primary aldosteronism, Zwermann [/bib_ref] [bib_ref] Increase in gamma-globin mRNA content in human erythroid cells treated with angelicin..., Lampronti [/bib_ref] [bib_ref] Eutopic overexpression of vasopressin v1a receptor in adrenocorticotropin-independent macronodular adrenal hyperplasia, Mune [/bib_ref] [bib_ref] Evidence for a role of vasopressin in the control of aldosterone secretion..., Perraudin [/bib_ref] , glucagon (glucagon receptor), TRH (TRH R) [bib_ref] Screening for membrane hormone receptor expression in primary aldosteronism, Zwermann [/bib_ref] [bib_ref] LH, progesterone, and TSH can stimulate aldosterone in vitro: a study on..., Nicolini [/bib_ref] [bib_ref] The role of thyroid hormone in blood pressure homeostasis: evidence from short-term..., Fommei [/bib_ref] and Endothelin-1 ETA and ETB receptors [bib_ref] Endothelin receptor blockade lowers plasma aldosterone levels via different mechanisms in primary..., Rossi [/bib_ref]. Using a microarray approach in 10 aldosteronomas compared with five normal adrenals and 13 cortisol-secreting adenomas, the six GCPRs with highest increase in expression included LHCGR, 5-HT4R, GnRHR, glutamate receptor metabotropic 3, endothelin receptor ETB receptors, and MC2R (92). [fig_ref] TABLe 1 |: Types of GPCR involved in aldosterone hypersecretion in patients with PA [/fig_ref] summarizes the different types of aberrant eutopic or ectopic GPCR involved in aldosterone excess in PA. Co-expression of multiple aberrant GPCR was also reported; renin-independent stimulation of aldosterone secretion was observed in vivo following mixed meal, oral glucose, or administration of GIP, vasopressin, and tegaserod in a patient with unilateral source of PA [bib_ref] Increase in gamma-globin mRNA content in human erythroid cells treated with angelicin..., Lampronti [/bib_ref]. On the other hand, co-secretion of aldosterone and cortisol due to aberrant expression of GPCR was noted; in a patient with BMAH and β-AR-aberrant expression, isoproterenol stimulated both cortisol and aldosterone production [bib_ref] Propranolol therapy for ectopic beta-adrenergic receptors in adrenal Cushing's syndrome, Lacroix [/bib_ref].
Activating somatic CTNNB1 (β-catenin) mutations have now been identified in tumors of three women with APAs, two who presented during pregnancy and one after menopause [bib_ref] Pregnancy, primary aldosteronism, and adrenal CTNNB1 mutations, Teo [/bib_ref]. All three had heterozygous activating mutations of CTNNB1 and expressed aberrant LHCG and GNRH receptors at levels 100-fold higher than in other APAs. It was shown that the CTNNB1 mutation led to activation of the WNT pathway; it was suggested that could be the cause of dedifferentiation of gonadal progenitor cells present in the adrenal tissues with increased expression of gonadal receptors. It is thought that the high levels of endogenous human chorionic gonadotropin (hCG) during pregnancy and of gonadotropin-releasing hormone (GnRH) and luteinizing hormone (LH) after menopause led to the identification of APAs in these patients [bib_ref] Pregnancy, primary aldosteronism, and adrenal CTNNB1 mutations, Teo [/bib_ref]. It is currently unclear whether these aberrant regulatory secretory mechanisms by ACTH and other hormones and the overexpression of their GPCR in PA are secondary to unknown proliferative mechanisms or are primary and at least partially responsible for the abnormal proliferation, the initiation of diffuse BAH. However, they clearly play a role in aldosterone secretion which is not autonomous.
# Conclusion
Our understanding of the increased occurrence and complexity of molecular etiology and unique signature in each case of PA has progressed greatly in recent years. The increased role of ACTH, of the variable expression of MC2R, and of other aberrant GPCR in PA should receive further attention in the future. The development of effective antagonists to MC2R and other aberrant GPCR could eventually offer interesting alternatives in patients with bilateral sources of excess aldosterone in combination with better antagonists of the mineralocorticoid receptor.
# Author contributions
NG, IB, and AL contributed to the conception and design of the manuscript as well as to drafting the review article and they all provided final approval of the version to be published.
# Funding
This work is supported by grant 201209NMD from the Canadian Institutes of Health Research to AL and IB. NG fellowship support was also provided in part by Novartis Canada. IB receives a salary Grant from FRQS (Fonds de la recherche du Québec-Santé).
## References
[fig] FiGURe 1 |: Mechanisms responsible for aldosterone synthesis in zona glomerulosa cells under normal physiological conditions and excess production in primary aldosteronism. [/fig]
[table] TABLe 1 |: Types of GPCR involved in aldosterone hypersecretion in patients with PA. Insulin-induced hypoglycemia Exercise/ stress test hyperaldosteronism Isoproterenol (β1-agonist) GnRH receptor, LH/hCG receptor (ectopic) (18, 102, 107-112) Luteal phase of ovarian cycle/Pregnancy (transient) El Ghorayeb et al. Role of ACTH in Primary Aldosteronism Frontiers in Endocrinology | www.frontiersin.org June 2016 | Volume 7 | Article 72 [/table]
|
Late emergence of macular sparing in a stroke patient
Rationale: Occlusive cerebrovascular disease is the most common cause of homonymous hemianopia (HH) with macular sparing.Patient concerns: A 61-year-old man came to our ophthalmology clinic complaining of right-side hemianopia. Ophthalmic examination, visual field (VF) examination, and brain magnetic resonance imaging (MRI) were performed.Diagnoses: He had right HH without macular sparing on the initial VF test. And brain MRI 6 days after the visual symptoms began revealed a left occipital infarction.Interventions and Outcomes: Thirty-seven days after the onset, his follow-up 24-2 VF examination showed HH with bilateral macular sparing, which was not apparent in the initial VF examination. About 4 months after the stroke, his central 10-2 VF examination also showed HH with bilateral macular sparing.Lessons: We report a case of HH with a dramatic improvement in central vision several days after an occipital infarction. To our knowledge, this is the first case to show macular sparing developing after several days.Abbreviations: HH = homonymous hemianopia, MCA = middle cerebral arteries, MRA = magnetic resonance angiography, MRI = magnetic resonance imaging, PCA = posterior cerebral arteries, VF = visual field.
# Introduction
Occlusive cerebrovascular disease is the most common cause of homonymous hemianopia (HH) with macular sparing. [bib_ref] Homonymous hemianopias: clinicalanatomic correlations in 904 cases, Zhang [/bib_ref] A paper published in 1951 mentioned that the central visual field (VF) can be bisected and the visual field of the macular area spared. [bib_ref] Homonymous hemianopia, bilateral with macular sparing, of vascular origin, Steiner [/bib_ref] We report a case of right HH with macular sparing that was not seen at the initial examination after a left occipital infarction, and became obvious several days later.
# Material and methods
The patient received a complete ophthalmologic and neurologic examination, including a review of medical and ocular histories, measurement of best-corrected visual acuity, slit-lamp biomicroscopy, visual field examination and magnetic resonance imaging (MRI).
# Results
A 61-year-old man came to our ophthalmology clinic complaining of right-side hemianopia. He experienced dizziness and a headache immediately after lifting a heavy object 4 days ago. Subsequently, he could not see the right side of his visual field and had a flashing light sensation on the right side. He had right HH on the initial VF test [fig_ref] Figure 2: Humphrey visual field examination The Humphrey 24-2 visual field examination showed [/fig_ref]. We referred him to the neurology clinic because an occipital lobe infarction was suspected. MRI 6 days after the visual symptoms began revealed only the patchy acute infarction involving the left occipital cortex [fig_ref] Figure 1: Magnetic resonance imaging brain magnetic resonance imaging performed 6 days after an... [/fig_ref]. In MRA, there were no suspicious lesions in large intracranial and cervical arteries. A systemic work-up revealed that the occipital infarction was the result of untreated atrial fibrillation. He began medical treatment including antithrombotic therapy for atrial fibrillation. He experienced a subjective symptomatic improvement in the central visual field over time after the onset of symptoms. Thirty-seven days after the onset, his follow-up 24-2 VF examination showed HH with bilateral macular sparing which was not apparent in the initial VF examination of both the eyes [fig_ref] Figure 2: Humphrey visual field examination The Humphrey 24-2 visual field examination showed [/fig_ref]. About 4 months after the stroke, his central 10-2 VF examination also showed HH with bilateral macular sparing [fig_ref] Figure 2: Humphrey visual field examination The Humphrey 24-2 visual field examination showed [/fig_ref]. affected side. HH with macular sparing has been reported in ischemic stroke patients. [bib_ref] Homonymous hemianopias: clinicalanatomic correlations in 904 cases, Zhang [/bib_ref] [bib_ref] Homonymous hemianopia, bilateral with macular sparing, of vascular origin, Steiner [/bib_ref] [bib_ref] A case of bilateral homonymous hemianopsia with macular sparing, Dufresne [/bib_ref] [bib_ref] A historical review of the representation of the visual field in primary..., Leff [/bib_ref] However, the mechanism of macular sparing is controversial. One hypothesis is that the visual cortex has a dual blood supply, from the middle (MCA) and posterior (PCA) cerebral arteries or the posterior temporal and calcarine arteries. [bib_ref] A case of bilateral homonymous hemianopsia with macular sparing, Dufresne [/bib_ref] [bib_ref] A historical review of the representation of the visual field in primary..., Leff [/bib_ref] The occlusion of 1 artery might result in macular sparing if another artery still provides blood to the visual cortex that represents macular vision. However, Smith and Richardson [bib_ref] The course and distribution of the arteries supplying the visual (striate) cortex, Smith [/bib_ref] reported a variety of patterns by which the 4 branches of the PCA supply the visual cortex. Therefore, another hypothesis, incomplete damage to the visual cortex, might explain this phenomenon.
Our patient noticed improvement in the central VF after seeing flashing light for several days, and his follow-up 24-2 VF examination 37 days after the occipital infarction showed HH with bilateral macular sparing, which was not apparent in the initial 24-2 VF examination. The macular sparing was observed in a subsequent 10-2 VF examination after 114 days. This might result from increased collateral circulation to the PCA territory. We postulate that a PCA infarction caused the temporary HH without macular sparing and resumed blood supply to the tip of occipital lobe led to the improvement in the central VF.
Without knowing the exact mechanism of macular sparing, we report a case of HH with a dramatic improvement in central vision several days after an occipital infarction. This phenomenon might be more common than has been previously thought in HH with macular sparing. As our patient visited the ophthalmology clinic initially, an initial VF examination was performed at an early stage of infarction. If initial VF examinations were performed at an early stage in patients with occipital lobe infarctions, we could determine whether this phenomenon is rare or common. Although we do not understand the mechanism of this phenomenon, to our knowledge, this is the first case to show macular sparing developing after several days.
The patient has consented to the submission of the article to the journal.
[fig] Figure 2: Humphrey visual field examination The Humphrey 24-2 visual field examination showed (A) right homonymous hemianopia on day 4 and (B) right homonymous hemianopia with macular sparing on day 37. C, The Humphrey 10-2 visual field examination 114 days after the occipital infarction showed right homonymous hemianopia with macular sparing. [/fig]
[fig] Figure 1: Magnetic resonance imaging brain magnetic resonance imaging performed 6 days after an occipital infarction showed a hyperintense lesion in the left occipital lobe in (A) diffusion-weighted images and (B) fluid-attenuated inversion recovery images. [/fig]
|
Continuous positive airway pressure improves nocturnal polyuria in ischemic stroke patients with obstructive sleep apnea
Objectives:To assess the effect of continuous positive airway pressure (CPAP) on nocturia in ischemic stroke patients with obstructive sleep apnea (OSA). Methods: This was a prospective and non-randomized controlled study in which ischemic stroke patients with OSA being treated in a rehabilitation ward were enrolled. The participants who tolerated CPAP were classified as the CPAP group, while those who refused or could not tolerate CPAP were classified as the control group. The percentage change of nocturia before and after 2 weeks of CPAP therapy between the two groups were compared. Results: A total of 44 participants were enrolled in and 35 participants (mean age= 59.8±11.7 years old; mean apnea hypopnea index=42.9±16.7/h) completed the study (control group: 14, CPAP group: 21). Overall, 69% of the participants had nocturnal polyuria and 69% of them had more than one nocturia episode per night. The baseline and initial nocturia characteristics did not differ significantly between the two groups. As compared to the control group, CPAP therapy significantly decreased the nocturnal polyuria index (mean percentage change: 9% vs -21% (P=0.005)) and nocturnal urine output (mean percentage change: 6% vs -26% (P=0.04)), but not the nocturia episodes or 24-hours total urine output. Conclusion: Nocturia due to nocturnal polyuria is very common in post-stroke patients with OSA. Treating OSA by CPAP significantly reduces nocturnal polyuria, but not nocturia frequency, in ischemic stroke patients.
# Introduction
Obstructive sleep apnea (OSA) is a common chronic disorder, with an estimated prevalence of ~4%-7% among the adult general population 1 and an independent risk factor for ischemic stroke and cardiovascular death. [bib_ref] Obstructive sleep apnea as a risk factor for stroke and death, Yaggi [/bib_ref] Stroke is the second leading cause of death worldwide, and a major cause of mental and physical impairment. [bib_ref] Mortality by cause for eight regions of the world: global burden of..., Murray [/bib_ref] The prevalence of OSA is ~60% in stroke patients, [bib_ref] Sleep-related breathing and sleep-wake disturbances in ischemic stroke, Hermann [/bib_ref] which is substantially higher than the prevalence among the general population. Previously published data revealed that patients who survive a stroke and have OSA have an increased rate of mortality, especially mortality caused by cardiovascular events, compared to stroke survivors who do not have OSA. [bib_ref] Obstructive sleep apnea as a risk factor for stroke and death, Yaggi [/bib_ref] The OSA patients also have worse neurological and functional statuses and longer periods of hospitalization than the patients without OSA. [bib_ref] Obstructive sleep apnea is related to impaired cognitive and functional status after..., Aaronson [/bib_ref] Overall, OSA negatively affects short-term and long-term stroke outcomes, length of hospitalization, and recurrence risk. [bib_ref] Worse outcome after stroke in patients with obstructive sleep apnea: an observational..., Mansukhani [/bib_ref] These effects may occur via a number of pathogenic pathways influenced by intermittent hypoxemia, sympathetic stimulation, and hypertension that in turn influence cerebral vasoregulation and atrial fibrillation. [bib_ref] Obstructive sleep apnea in acute stroke: a role for systemic inflammation, Ifergane [/bib_ref] Nocturia is a common problem among the elderly. More than one nocturia episode per night is considered pathological, and nearly 60%-80% of elderly people have more than one episode per night. [bib_ref] The prevalence and causes of nocturia, Bosch [/bib_ref] Nocturia has been found to be an important risk factor for falls in ambulatory elders, [bib_ref] Nocturia: a risk factor for falls in the elderly, Stewart [/bib_ref] and elderly individuals who void three or more times per night have a greater mortality rate than those who void less often. [bib_ref] Mortality in the elderly in relation to nocturnal micturition, Asplund [/bib_ref] Global polyuria is defined as a 24-hour urinary output that exceeds 40 mL/kg body weight and results in increased 24-hour urinary frequency. Nocturnal polyuria is defined as .20% of daily urine output occurring at night in young patients and .33% occurring at night in elderly patients. [bib_ref] The standardisation of terminology in nocturia: report from the standardisation Sub-committee of..., Van Kerrebroeck [/bib_ref] Fitzgerald's studies showed that nocturia is a common symptom in OSA, with the prevalence of nocturia in OSA patients being positively correlated with OSA severity. [bib_ref] Nocturic frequency is related to severity of obstructive sleep apnea, improves with..., Fitzgerald [/bib_ref] Another study also found that nocturia is an independent predictor for severe OSA after adjusting for age, gender, and neck circumference, and suggested that nocturia might be used to screen for OSA in patients with ischemic stroke. [bib_ref] Nocturia is an independent predictor of severe obstructive sleep apnea in patients..., Chen [/bib_ref] Continuous positive airway pressure (CPAP) is the standard of therapy for OSA. CPAP therapy can reduce cardiovascular mortality, improve cognitive function, and improve quality-of-life. 14 Moreover, long-term CPAP treatment for OSA in ischemic stroke patients has been found to be associated with a reduction in the excess risk of mortality and improved functional neurologic outcomes. [bib_ref] Continuous positive airway pressure treatment reduces mortality in patients with ischemic stroke..., Martínez-García [/bib_ref] Another study also found that CPAP significantly reduces the frequency of nocturia in non-stroke OSA patients; [bib_ref] Continuous positive airway pressure reduces nocturia in patients with obstructive sleep apnea, Margel [/bib_ref] however, the effects of CPAP on nocturia in ischemic stroke patients with OSA have yet to be well investigated. As such, the purpose of this study was to investigate the effects of CPAP on nocturia in ischemic stroke patients with OSA.
## Materials and methods study subjects
This study was conducted in the rehabilitation wards of a teaching hospital from October 2013 to October 2016. All of the participants had been admitted to the wards for recent ischemic stroke (that is, a stroke that occurred more than 4 weeks before, but within the last 12 months) and were diagnosed with OSA. The diagnosis of ischemic stroke for each participant was made on the basis of a full clinical assessment with detailed neurological examinations and neuroimaging studies. OSA was defined by an apnea hypopnea index (AHI) of .15/hour with .50% of respiratory events being of the obstructive or mixed type. Participants with central sleep apnea (.50% of respiratory events being of the central type), sleep hypoventilation (arterial PCO 2 .55 mmHg for 10 minutes or increased arterial PCO 2 .10 mmHg in comparison to when awake with arterial PCO 2 .50 mmHg for 10 minutes), chronic obstructive airway disease, orofacial abnormality, urinary retention or incontinence, as well as patients who received other therapies like an oral appliance or soft palate surgery, were excluded. In addition, those patients who had an adjustment in their diuretic or medication for benign prostate hyperplasia (BPH) during the study period were also excluded, as were any potential participants who had a neurogenic bladder or active urinary tract infection during the study period. Meanwhile, those participants who refused CPAP therapy or could not tolerate it were classified as the control group. This study was approved by the institutional review board of Chang Gung Memorial Hospital (105-2690C), and all the subjects signed an informed consent form.
## Risk factors survey and clinical presentation recording
All of the participants were screened for hypertension (a history of hypertension or systolic blood pressure .140 mmHg or diastolic blood pressure .90 mmHg on the second day of admission) and diabetes mellitus (a history of diabetes mellitus or fasting blood glucose .126 mg/dL). The functional status (Barthel index), 17 polysomnography parameters, body weight, age, sex, and serum creatinine level of each participant was recorded. The presence of significant renal dysfunction (serum creatinine .1.4 mg/dL), congestive heart failure, poorly controlled diabetes mellitus, other genitourinary system diseases, or global polyuria (24-hour urine output of more than 40 mL/kg) was also recorded and reported because these factors might interfere with nocturia and make the association between nocturia and OSA more complicated. Medications for BPH, hypertension, and diabetes mellitus were considered confounding factors. Other medications, including hypnotics, anti-depressants, and lipid-lowering medications, may also influence nocturia. Thus, those medications were monitored during the urine frequency and volume recording period at both baseline and second-time measurements. If the medications were changed between the baseline and second-time measurements, then those changes were also recorded and reported.
## Polysomnography
Standard overnight polysomnograms were performed using a computerized polysomnography system (N7000 Embla, Broomfield, CO). The parameters recorded were electroencephalograms, bilateral electrooculograms, submental and bilateral anterior tibialis electromyograms, electrocardiograms, nasal and oronasal airflow (nasal pressure and thermistor),
## 243
Yu et al arterial oxygen saturation by finger probe pulse oximetry, chest and abdominal movements (inductance plethysmography), body position, and sound intensity. Sleep stages were scored manually in 30-second epochs. Obstructive apnea was defined as the absence of airflow for at least 10 seconds in the presence of respiratory effort, while central apnea was defined as the absence of airflow without concurrent respiratory effort. Hypopnea was defined as any decrease in airflow .30% lasting .10 seconds and followed by either an oxygen desaturation of at least 3% or EEG arousal. The oxyhemoglobin desaturation index, defined as the number of desaturations per hour, was also calculated. All of the sleep stage and respiratory event scoring was completed according to the criteria of the America Academy of Sleep Medicine version 2012. [bib_ref] Rules for scoring respiratory events in sleep: update of the 2007 AASM..., Berry [/bib_ref]
## Cpap intervention protocol
The CPAP efficacy and compliance of the participants affected the study results. Therefore, a standard CPAP protocol was arranged to minimize the factors that influence CPAP compliance and to optimize the CPAP efficacy. Before receiving CPAP therapy, all the participants underwent education about the consequences of OSA and the benefits of CPAP treatment. Mask selection and mask wearing training were done before the initial night of CPAP therapy. Most of the subjects started with automated CPAP (AutoSet Spirit S8, ResMed, Sydney, Australia). The CPAP pressure generally started with a maximum pressure of 10-16 cm H 2 O and a minimum pressure of 4-6 cm H 2 O, after which the maximum and minimum pressures were adjusted according to the given patient's residual AHI record and subjective complaints. Because the residual AHI and duration of CPAP usage are important, regular checks were conducted to ensure that the CPAP recording included the residual AHI, average hours of daily use, and leak value. CPAP tracking was performed on days 3, 7, 14, and 28 to ensure good mask fitting and adequate CPAP usage. Moreover, any CPAP-related side-effects were immediately addressed in order to optimize CPAP compliance.
## Measurements of nocturia before and after cpap therapy
The urinary frequency and volume of each subject were recorded at two time periods, with measurements over 3 consecutive days for each period. The baseline urine recording was conducted before the CPAP trial. The second urine recording for the CPAP group was only conducted for those subjects who could adequately use CPAP. Specifically, if a participant met the criteria of using CPAP longer than 4 hours every day and for .70% of the days in 1 week, then the urinary frequency and volume of that participant were measured. For the control group, if a participant did not take part in any CPAP trial, his or her urinary frequency and volume were measured for a second time beginning 2 weeks after the baseline measurement. The reason why the measurements were conducted 2 weeks apart was that it generally takes 1-2 weeks to evaluate CPAP intolerance or adequate usage. Meanwhile, if subjects in the control group had tried CPAP but found it intolerable, the second set of nocturia measurements was conducted 2 weeks after the patient stopped using CPAP, in order to minimize any effect from the CPAP. All of the participants were instructed to record the time and volume of each void, as well as their waking time and bedtime. Nocturia episodes are voids that occur between an individual's bedtime and waking time. The first voiding of urine by a participant after waking in the morning without going back to sleep was regarded as nocturnal urine. The nocturnal polyuria index (NPI) was defined as the ratio of nocturnal urine production to 24-hours total urine production. [bib_ref] Terminology, epidemiology, etiology, and pathophysiology of nocturia, Van Kerrebroeck [/bib_ref] The frequency of nocturia, 24-hours total urine output, nocturnal urine output, and NPI were calculated by averaging the results of the 3 days of recordings. The percentage changes of nocturia episodes, 24-hours total urine output, nocturnal urine output, and NPI between the baseline and the second set of measurements were also calculated.
## Statistics
The basic characteristics of the control and CPAP groups were compared with independent t-tests, and underlying diseases and medication usage were compared with the chi-squared test and Fisher's exact probability test. The primary end-point consisted of comparing the improvement in nocturia (in terms of nocturia episodes, nocturnal urine output, 24-hours total urine output, and NPI) between those two groups. Those parameters were compared with independent t-tests. Statistical evaluations were performed using the SPSS 16.0 Software (SPSS Inc., Chicago, IL). Data were presented as mean±SD. P,0.05 was considered statistically significant.
## Ethics approval and informed consent
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional review board of Chang Gung Memorial Hospital (105-2690C) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study.
# Results
Forty-four participants were initially enrolled in our study, and the study process is shown in [fig_ref] Figure 1: Flow diagram of the study [/fig_ref]. There were seven participants who dropped out of the study because they were discharged from the hospital early and, thus, could not complete the study. One participant suffered from an episode of delirium during the study period and refused to finish the study. Another subject was excluded because the medication he was receiving for benign prostate hyperplasia was adjusted during the study period. As such, a total of 35 subjects (mean age=59.8±11.7 years old) completed the study. All of them were able to follow our study protocol. Most of the participants had moderate-to-severe OSA (mean AHI=42.9±16.7/hour). There were four participants who refused CPAP without taking part in any CPAP trial, and 10 participants who took part in a CPAP trial but could not tolerate CPAP. and medication adjustments of these two groups were not statistically different. There were no participants who had a genitourinary system disease or global polyuria. The average AHI of the CPAP group was decreased from 43.0±17.3/hour to 6.5±4.9/hour after CPAP therapy. The average time of CPAP usage was 7.3±1.0 hours per night, and all of the participants used it for more than 70% of the days during the study period. Overall, 69% (control n=10, CPAP n=14) of the subjects had nocturnal polyuria according to the definition of a NPI .33%. Otherwise, there were 24 (69%; control n=9, CPAP n=15) patients who had more than one nocturia episode per night and 11 (31%; control n=4; CPAP n=7) patients with three or more nocturia episodes per night. There were no significant differences between the two groups in terms of baseline nocturia episodes, 24-hours total urine output, nocturnal urine output, and NPI [fig_ref] Table 2: Baseline nocturia of the control and CPAP groups [/fig_ref]. As compared to the control group, CPAP therapy significantly decreased NPI (mean percentage change=9% vs -21% (P=0.005)) and nocturnal urine output (mean percentage change=6% vs -26% (P=0.04)), but not the nocturia episodes or 24-hours total urine output [fig_ref] Table 3: Percentage changes of nocturia between the control and CPAP groups after CPAP... [/fig_ref].
# Discussion
Nocturia due to nocturnal polyuria is very common in poststroke patients with OSA. To the best of our knowledge, this is the first study to report that CPAP therapy can significantly reduce nocturnal polyuria but not nocturia frequency in ischemic stroke patients with OSA. Due to multiple comorbidities, CPAP therapy does not significantly improve nocturia frequency in ischemic stroke patients with OSA. Oztura et al's 20 study revealed that the prevalence of nocturia is 52% in patients with primary snoring, 57.2% in patients with mild OSA, 64.3% in patients with moderate OSA, and 76.9% in patients with severe OSA. Their results show that the prevalence of nocturia increases with increasing OSA severity. In addition, several previous studies found that CPAP could effectively reduce the frequency of nocturia and night time urine volume, while also improving the quality-of-life in subjects without ischemic stroke. [bib_ref] Continuous positive airway pressure reduces nocturia in patients with obstructive sleep apnea, Margel [/bib_ref] [bib_ref] Effect of the continuous positive airway pressure on the nocturnal urine volume..., Miyauchi [/bib_ref] [bib_ref] Effect of continuous positive airway pressure on nocturnal urine production in patients..., Miyazato [/bib_ref] The reasons why the nocturia frequency of the stroke patients
## 246
Yu et al in this study was not improved by CPAP use are not clearly understood, but we speculate that ischemic stroke patients are generally elderly individuals with multiple comorbidities, such as insomnia, BPH, bladder dysfunction, sleep disorder, and heart failure, who also usually use several kinds of medications. Those are all possible factors influencing nocturia and could result in large variations in nocturia frequency. A previous study also found that younger OSA patients experienced better nocturia improvements than elderly patients when receiving CPAP. [bib_ref] Obstructive sleep apnea syndrome should be considered as a cause of nocturia..., Maeda [/bib_ref] The results of this study support that finding and suggest that age-related urinary diseases and voiding dysfunction may mask the influence of CPAP on nocturia in ischemic stroke patients. Therefore, the effects of CPAP on nocturia in stroke patients are less obvious than in non-stroke and younger patients. Hajduk et al's 24 study reported that nocturnal polyuria was found in 47.8% of non-ischemic OSA patients, a rate which is considerably lower than the 69% rate found in our study. The older ages, ischemic strokes, and multiple comorbidities of the patients in the present study may all have played a role in that difference. Miyazato et al's 22 study proved that CPAP could effectively reduce NPI and nocturnal urine output in non-stroke OSA patients, but no previous study had investigated the effects of CPAP on nocturia in ischemic stroke patients with OSA and multiple comorbidities. Our study is, thus, the first investigation of that type. OSA is caused by upper airway obstruction during sleep. The obstructed site is usually on the tongue or soft palate, and the obstruction leads to intermittent hypoxemia, increasing sympathetic activity, and altering hormone regulation. [bib_ref] Abnormal vasoactive hormones and 24-hour blood pressure in obstructive sleep apnea, Møller [/bib_ref] In addition, airway obstruction following continuous breathing generates an abnormal airway pressure change. This results in the development of an extremely high negative airway pressure level caused by sucking through an obstructed airway. This in turn causes the heart to receive a false signal of volume overload. The hormonal response to this signal is increased atrial natriuretic peptide (ANP) secretion. The normal physiological response to atrial stretch is to excrete ANP, which is a natriuretic, diuretic, and vasorelaxant cardiac hormone. [bib_ref] Obstructive sleep apnea, nocturia and polyuria in older adults, Umlauf [/bib_ref] ANP can also inhibit the secretion of arginine vasopressin (AVP), rennin-angiotensin, and aldosterone, while increasing glomerular filtration and further increasing nocturnal urine output. [bib_ref] Natriuretic peptide system: physiology and clinical utility, Suttner [/bib_ref] AVP is the major hormone responsible for the regulation of urine formation. In healthy adults, the diurnal release of AVP into plasma has its peak during night time. The net effect of this AVP secretion is diminished urine production during the sleep interval. [bib_ref] Nocturnal polyuria and nocturnal arginine vasopressin (AVP): a key factor in the..., Abdelfatah [/bib_ref] Sakakibara et al's 29 study revealed that post-stroke patients had nocturnal polyuria with abnormal circadian rhythm in terms of plasma AVP secretion, and Umlauf et al's 26 study showed that plasma ANP levels were significantly higher among subjects with higher AHI levels. The possible mechanism of OSA in increasing nocturnal urine output could be that the interrupted sleep and increased ANP secretion cause diminished AVP levels. Meanwhile, CPAP could eliminate the respiratory events and reduce the airway pressure swings seen in OSA. Thus, the mechanism by which CPAP reduces nocturnal polyuria would be by abolishing the negative pressure swings and intermittent hypoxia of OSA by reducing ANP and normalizing AVP levels. As nocturnal urine output is less affected by insomnia, BPH, and other genitourinary system diseases, CPAP could thus significantly decrease nocturnal polyuria, but not the frequency of nocturia in stroke patients.
There are several limitations to this study. First, it recruited patients in a rehabilitation ward. Most of the patients were survivors of a recent ischemic stroke. However, some transient ischemic attack or minor stroke patients are not admitted for rehabilitation. Thus, such patients with minor strokes were not included in this study. In addition, patients who are comatose, too drowsy, and have an unstable clinical condition are not transferred to the rehabilitation ward. Therefore, the present study may have had a selection bias and may not represent the entire ischemic stroke population, and thus the results of the study may not be generalizable to minor or severe ischemic stroke patients. The second limitation was the lack of upper airway evaluation. Previous studies showed that the upper airway status, such as nasal obstruction and a floppy epiglottis, might influence OSA severity and CPAP compliance. [bib_ref] What is the role of the larynx in adult obstructive sleep apnea?, Dedhia [/bib_ref] Because the upper airway was not routinely evaluated in our study, the impact of any upper airway abnormality on group allocation could not be assessed. Further studies are recommended to clarify this issue. The third limitation of this study was its relatively small sample size. This may have influenced the statistical power of the study, especially with respect to nocturia episodes. It is, thus, suggested that a large randomized controlled study be conducted to address this limitation.
# Conclusion
In conclusion, CPAP treatment can significantly reduce nocturnal polyuria, but not nocturia frequency, in ischemic stroke patients with OSA. Treating comorbidities associated with nocturia while also providing CPAP is suggested as a way to substantially reduce nocturia in ischemic stroke patients with OSA.
[fig] Figure 1: Flow diagram of the study. Abbreviations: AhI, apnea hypopnea index; BPh, benign prostate hyperplasia; CPAP, continuous positive airway pressure; OsA, obstructive sleep apnea. [/fig]
[table] 10: participants tried CPAP only for a few days. These 14 participants were classified as the control group. The other 21 participants who could tolerate and accept CPAP were classified as the CPAP group. All of participants in the CPAP group used a nasal mask in the study period. The basic characteristics, underlying diseases, [/table]
[table] Table 2: Baseline nocturia of the control and CPAP groups [/table]
[table] Table 3: Percentage changes of nocturia between the control and CPAP groups after CPAP therapy [/table]
|
Structures of the Middle East respiratory syndrome coronavirus 3C‐like protease reveal insights into substrate specificity
Middle East respiratory syndrome coronavirus (MERS-CoV) is a highly pathogenic virus that causes severe respiratory illness accompanied by multiorgan dysfunction, resulting in a case fatality rate of approximately 40%. As found in other coronaviruses, the majority of the positive-stranded RNA MERS-CoV genome is translated into two polyproteins, one created by a ribosomal frameshift, that are cleaved at three sites by a papain-like protease and at 11 sites by a 3C-like protease (3CL pro ). Since 3CL pro is essential for viral replication, it is a leading candidate for therapeutic intervention. To accelerate the development of 3CL pro inhibitors, three crystal structures of a catalytically inactive variant (C148A) of the MERS-CoV 3CL pro enzyme were determined. The aim was to co-crystallize the inactive enzyme with a peptide substrate. Fortuitously, however, in two of the structures the C-terminus of one protomer is bound in the active site of a neighboring molecule, providing a snapshot of an enzyme-product complex. In the third structure, two of the three protomers in the asymmetric unit form a homodimer similar to that of SARS-CoV 3CL pro ; however, the third protomer adopts a radically different conformation that is likely to correspond to a crystallographic monomer, indicative of substantial structural plasticity in the enzyme. The results presented here provide a foundation for the structure-based design of small-molecule inhibitors of the MERS-CoV 3CL pro enzyme. Figure 6(a) Stereoview of the superimposed structures of MERS-CoV 3CL pro crystal form III protomer A (green ribbons) and protomer C (magenta ribbons). (b, c) Surface representations of protomer A (b) and protomer C (c) with domains I and II colored gray, the linker loop (residues 188-204) cyan, domain III magenta, the oxyanion loop (residues 143-148) blue and the S1 binding pocket green.
# Introduction
Middle East respiratory syndrome coronavirus (MERS-CoV) was first reported in 2012 following isolation from a patient in Saudi Arabia . MERS-CoV causes severe pneumonia [bib_ref] PLoS Pathog. 10, e1004250, Falzarano [/bib_ref] reminiscent of the severe acute respiratory syndrome (SARS) outbreak of 2003, but cases of MERS-CoV exhibit a higher mortality rate than those of SARS-CoV (approximately 40% versus 10%). Although the number of new cases peaked in early 2014 (http://www.who.int/csr/disease/coronavirus_ infections/archive_updates/en/;, the outbreak continues. The severity and rapid spread of MERS and SARS illustrate the need for the development of new therapeutics to combat known and emerging coronaviruses.
MERS-CoV belongs to the genus Betacoronavirus, which is divided into four clades: a-d. The clade b SARS coronavirus (SARS-CoV) is thought to have its reservoir in bats [bib_ref] PLoS Pathog. 10, e1004250, Falzarano [/bib_ref] , with civets as an intermediate host facilitating human infection. MERS-CoV belongs to Betacoronavirus clade c, along with the closely related bat coronaviruses HKU4 (BatCoV-HKU4) and HKU5. A conspecific virus that shares 85% genome sequence identity with MERS-CoV has been isolated from the Neoromica capensis bat. Recent work showed that introduction of a clinical isolate of MERS-CoV into dromedary camels resulted in mild respiratory illness followed by persistent shedding of infectious virus from the upper respiratory tract. Taken together, these results suggest that MERS-CoV originated in bats, with camels serving as the carrier for human infection.
Coronaviruses, including MERS-CoV, SARS-CoV and the usually milder human coronaviruses (HCoV) HCoV-229E, HCoV-NL63 and HCoV-OC43, share a common organization of their polycistronic positive-strand RNA genomes. On the 5 0 end of the MERS-CoV genome are the two large open reading frames (ORF1a and ORF1b) encoding nonstructural proteins (nsps), followed by genes encoding the spike, envelope, membrane and nucleocapsid structural proteins. The genomic mRNA of ORF1a is translated into the polyprotein pp1a. A longer polyprotein (pp1ab) is the product of a ribosomal frameshift that joins ORF1a together with ORF1b (van. ORF1a encodes two proteases: a papain-like protease (PL pro ) and a 3C-like 'main' protease (3CL pro ). The 3CL pro , which in its essential role in viral replication is also called the 'main protease' (M pro ), processes the polyprotein at 11 cleavage sites (consensus: LQ#A/S), including those flanking it. The essential function and conservation among 3CL pro s from different coronaviruses make the main protease an attractive drug target for currently known and future emerging coronaviruses . In contrast, the structural and accessory genes encoded towards the 3 0 end of coronavirus genomes exhibit too much variability to serve as targets for broad anti-coronaviral agents . Coronaviral 3CL pro s are chymotrypsin-like proteases except that they use cysteine as the nucleophile in a catalytic dyad instead of serine in a catalytic triad. SARS-CoV 3CL pro exists in a monomer-dimer equilibrium in solution, but the homodimer is the enzymatically active form. Each monomer consists of three structural domains: domains I and II contain the catalytic site and chymotrypsin-like scaffold and are connected to a third C-terminal domain via a long loop. In this study, we report the structure of a catalytically inactive variant (C148A) of MERS-CoV 3CL pro in three different crystal forms, each providing distinct biological insights.
# Materials and methods
2.1. Cloning, expression and protein purification Expression vectors were constructed by Gateway recombinational cloning (Life Technologies, Grand Island, New York, USA). The 3CL pro gene was amplified by polymerase chain reaction (PCR) from a cDNA clone constructed using total RNA isolated from MERS-CoV Jordan (primers: 5 0 -CAC CAG CGG TTT GGT GAA AAT GTC ACA TCC C-3 0 and 5 0 -TTA CTA CTG CAT AAC CAC ACC CAT AAT CTG C-3 0 ).
To construct the catalytically inactive C148A variant, a MERS-CoV 3C-like protease amplicon was first used as a PCR template with primers PE2635 (5 0 -GGC TCG GAG AAC CTG TAC TTC CAG AGC GGT TTG GTG AAA ATG TCA CAT-3 0 ) and PE2636 (5 0 -GGG GAC CAC TTT GTA CAA GAA AGC TGG GTT ATT ACT GCA TAA CCA CAC CCA TAA TCT GC-3 0 ), which added nucleotides encoding a tobacco etch virus (TEV) protease recognition site to the 5 0 end of the MERS-CoV 3CL pro sequence. The product of the reaction was amplified in a second PCR with primers PE277 (5 0 -GGGG ACA AGT TTG TAC AAA AAA GCA GGC TCG GAG AAC CTG TAC TTC CAG-3 0 ) and PE2636 to produce a product competent for Gateway cloning. The PCR product was recombined into donor vector pDONR221 to produce the entry vector pDN2482. The active-site cysteine (Cys148) was changed to an alanine with the QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent, Santa Clara, California, USA) using primers PE2732 (5 0 -ACC AAC ACT ACC AGC AGA ACC ACA CAG AAA GGA ACC CTT A-3 0 ) and PE2733 to produce the entry vector pDN2544. pDN2544 was recombined into the destination vector pDEST-527 (Protein Expression Laboratory, Leidos Biomedical Research Inc., Frederick, Maryland, USA) to produce pDN2551, an expression vector encoding a TEV protease-cleavable hexahistidine tag preceding MERS-CoV 3CL pro (residues 1-306; C148A). The protein was produced in Escherichia coli strain Rosetta 2(DE3) (EMD Millipore, Billerica, Massachusetts, USA). Cells were grown to mid-log phase at 310 K in LB broth containing 100 mg ml À1 ampicillin, 30 mg ml À1 chloramphenicol and 0.2% glucose. Overproduction of the fusion protein was induced with IPTG at a final concentration of 1 mM for 4 h at 303 K. The cells were pelleted by centrifugation and stored at 193 K.
For protein purification, all procedures were performed at 277-281 K. 5 g of E. coli cell paste were suspended in 150 ml buffer A (50 mM Tris, 200 mM NaCl, 25 mM imidazole pH 7.2). The cells were lysed with an APV-1000 homogenizer (Invensys APV Products, Albertslund, Denmark) at 69 MPa and centrifuged at 30 000g for 30 min. The supernatant was filtered through a 0.2 mm polyethersulfone membrane and applied onto a 5 ml HisTrap FF column (GE Healthcare Life Sciences, Pittsburgh, Pennsylvania, USA) equilibrated with buffer A. The column was washed to baseline with buffer A and eluted with a linear gradient of imidazole to 500 mM in buffer A. Fractions containing recombinant protein were pooled, concentrated using an Amicon YM10 membrane (EMD Millipore, Billerica, Massachusetts, USA), diluted to an imidazole concentration of about 25 mM with 50 mM Tris pH 7.2, 200 mM NaCl buffer and digested overnight at 277 K with His 6 -tagged TEV protease. TEV protease digestion, which removed the His 6 affinity tag and amino acids encoded by sequences that facilitate Gateway cloning, resulted in a native protein product devoid of cloning artifacts. The digest was applied onto a 5 ml HisTrap FF column equilibrated in buffer A and recombinant protein emerged in the column effluent. The effluent was incubated overnight at 277 K with 10 mM dithiothreitol, concentrated using an Amicon YM10 membrane and applied onto a HiPrep 26/60 Sephacryl S-200 HR column (GE Healthcare Bio-Sciences Corporation) equilibrated with 25 mM Tris pH 7.2, 150 mM NaCl, 2 mM tris(2-carboxyethyl)phosphine buffer. The peak fractions were pooled and concentrated to about 20 mg ml À1 (as estimated at 280 nm using a molar extinction coefficient of 43 890 M À1 cm À1 derived using the ExPASy ProtParam tool. Aliquots were flashfrozen with liquid nitrogen and stored at 193 K. The molecular weight of the product was confirmed by electrospray ionization mass spectroscopy.
## Protein crystallization
Catalytically inactive (C148A) MERS-CoV 3CL pro (20.3 mg ml À1 ) was subjected to various crystallization screens including the MCSG Suite (Microlytic, Burlington, Massachusetts, USA) and Morpheus; Molecular Dimensions, Altamonte Springs, Florida, USA) using the sitting-drop vapor-diffusion method and a Gryphon crystallization robot (Art Robbins Inc., Sunnyvale, California, USA). Further optimization of the initial crystallization hits was performed by the hanging-drop vapor-diffusion method. Three different crystal forms were obtained. Crystal form I appeared from condition E10 of Morpheus by mixing 2 ml protein (20.3 mg ml À1 ) with 2 ml well solution [0.1 M Tris-Bicine pH 8.5, 0.03 M diethylene glycol, 0.03 M triethylene glycol, 0.03 M tetraethylene glycol, 0.03 M pentaethylene glycol, 10%(w/v) PEG 8000, 20%(v/v) ethylene glycol] and sealing the drop over 500 ml well solution. Crystal form II appeared under condition H10 from Morpheus [0.1 M Tris-Bicine pH 8.5, 0.02 M sodium l-glutamate, 0.02 M dl-alanine, 0.02 M glycine, 0.02 M dl-lysine-HCl, 0.02 M dl-serine, 10%(w/v) PEG 8000, 20%(v/v) ethylene glycol]. All stock reagents for crystallization conditions from the Morpheus Screen were obtained from Molecular Dimensions. Crystal form III was initially obtained from condition H1 of the MCSG 3 screen and was optimized by mixing 2 ml protein solution (20.3 mg ml À1 ) with 2 ml well solution [0.1 M HEPES pH 7.5, 0.2 M proline, 10%(w/v) polyethylene glycol 3350] and sealing over 500 ml well solution. All crystallization plates were incubated at 292 K and crystals generally appeared within 1-5 d. For data collection, crystal forms I and II were retrieved directly from the crystallization drop using a LithoLoop (Molecular Dimensions) and flash-cooled by plunging into liquid nitrogen without the need for additional cryoprotectant. Crystal form III was cryoprotected by transferring a crystal into a new drop consisting of well solution supplemented with 20%(v/v) polyethylene glycol 200, soaking for 1 min and flash-cooling by plunging into liquid nitrogen.
## X-ray data collection, structure solution and refinement
All X-ray diffraction data for crystal forms I and III were collected using a MAR345 detector mounted on a Rigaku MicroMax-007 HF high-intensity microfocus generator equipped with VariMax HF optics (Rigaku, The Woodlands, Texas, USA) and operated at 40 kV and 30 mA ( = 1.5418 A). Crystals were held at 93 K. For crystal form I, 525 diffraction images were collected with an exposure time of 600 s per research papers X-ray diffraction data-collection and refinement statistics.
Values in parentheses are for the highest resolution shell. image, an oscillation angle of 0.5 and a crystal-to-detector distance of 200 mm. For crystal form III, 360 images were collected with an exposure time of 180 s per image, an oscillation angle of 0.5 and a crystal-to-detector distance of 150 mm. Diffraction data from crystal form II were collected remotely on the SER-CAT beamline 22-BM at the Advanced Photon Source, Argonne National Laboratory, Lemont, Illinois, USA. Using an X-ray wavelength of 1.0 Å and a MAR CCD 225 detector, 360 images were collected with an exposure time of 6 s per image, an oscillation angle of 0.5 and a crystal-to-detector distance of 125 mm. All X-ray diffraction data were integrated and scaled using HKL-3000.
## Mers-cov
Firstly, the structure of MERS-CoV 3CL pro crystal form III was solved by molecular replacement using chain A of the main protease of coronavirus HKU4 (PDB entry 2yna; 81% sequence identity; Q. Ma, Y. Xiao & R. Hilgenfeld, unpublished work) as a search model, after stripping away all nonprotein atoms and changing non-identical residues to alanines. Molecular replacement was performed with MOLREP from the CCP4 suite. Two molecules (chains A and B) were located in the asymmetric unit using data to 2.5 Å resolution. The sequence for chains A and B could be fitted completely into the electron-density maps. A third molecule (chain C) was also found, but only residues 11-190 fitted well into the electron-density maps. Inspection of the initial electrondensity maps after rigid-body refinement with REFMAC5 (Murshudov et al., 2011) revealed a large region of well defined 2mF o À DF c and mF o À DF c electron-density features for protein residues adjacent to residues 11-190 of chain C. This indicated that residues 191-306 of chain C, corresponding to domain III of MERS-CoV 3CL pro , had undergone a large rigid-body movement. Therefore, another round of molecular replacement was performed with MOLREP by fixing the positions of chains A, B and residues 11-190 of chain C and then using residues 200-306 of chain C as a search model. Inspection of the new electron-density maps revealed a good fit of residues 200-306, confirming the alternate conformation of this region of the protein in chain C. The model was refined after several rounds of manual rebuilding and inspection with Coot, refinement with REFMAC5 and addition of water and other solvent molecules.
The structures of crystal forms I and II were subsequently solved by molecular replacement with MOLREP from the CCP4 suite of programs using chain A of crystal form III as a search model. Refinements for crystal form I were completed using PHENIXand Coot, while the structures of crystal forms II and III were refined using REFMAC5. All structure validations were performed with MolProbity . Secondary-structure elements were assigned using phenix.ksdssp. The catalytically inactive MERS-CoV 3CL pro C148A homodimer as found in crystal form I. Protomer A is colored green and protomer B red. The residues forming the catalytic dyad are depicted as blue spheres.
# Results and discussion
## Overall structure of mers-cov 3cl pro
The three different crystal forms (I, II and III) of catalytically inactive (C148A) MERS-CoV 3CL pro provide a structural view of three distinct states of the enzyme. Data-collection and refinement statistics for all three crystal forms are reported in . In all crystal forms a biological homodimer was observed that is similar to other 3CL pro enzymes such as those encoded by TGEV, HCoV-229E , SARS-CoV, IBV-CoV and HCoV-HKU1The two molecules of the homodimer are approximately perpendicular to one another. Each monomer is composed of a core chymotrypsin-like fold that is formed by two domains (domains I and II, residues 1-187), a connecting loop (residues 188-204) and a C-terminal -helical domain (referred to as domain III; residues 205-306). The C-terminal domain mediates dimerization; it has been demonstrated to play a key role in controlling the dimer-monomer equilibrium in other 3CL pro family members.
Crystals of forms I, II and III belonged to space groups C222 1 , C2 and P2 1 2 1 2 1 , respectively. There are three protomers in the asymmetric unit of crystal form I. Two of them form a canonical homodimer (protomers A and B), while the third forms an analogous homodimer with a symmetry mate (protomers C and C 0 ). There are no intermolecular interactions that mimic the binding of a peptide product in this crystal form. On the other hand, in both crystal forms II and III there is unambiguous electron density in the active site of protomer A that corresponds to the intercalated C-terminal tail residues of a neighboring protomer [fig_ref] Figure 2: Figures were prepared with PyMOL [/fig_ref]. The C-terminal residues Met301-Gln306 correspond to the P6-P1 sites of the autoprocessed product of the mature enzyme and therefore represent an enzyme-product complex. Surprisingly, in crystal form III, a significant shift in the orientation of domain III in protomer C, which inserts its C-terminal tail into the active site of protomer A, is observed (discussed below). Analysis of the crystal packing environment suggests that protomer C in crystal form III represents a crystallographic monomer, as it does not form a homodimer with any symmetry mate.
## Comparison with structural homologs
The coordinates of MERS-Cov 3CL pro crystal form II were submitted to the PDBeFold server to search for structural homologs. The closest match was identified as the BatCoV-HKU4 main protease. Alignment of protomer A of MERS-CoV 3CL pro with protomer A of BatCoV-HKU4 (PDB entry 2yna) yields an r.m.s.d. of 0.7 Å over 270 C -atom pairs (81% sequence identity) when superimposed using the 'super' command in PyMOL. Alignment of the MERS-CoV and BatCoV-HKU4 3CL pro homodimers yields an r.m.s.d. of 0.8 Å over 552 C -atom pairs. Superposition of MERS-CoV 3CL pro protomer A with protomer A from SARS-CoV 3CL pro (PDB entry 1uk3, 50% sequence identity;yields an r.m.s.d. of 1.9 Å over 258 C -atom pairs. When the structures of the two homodimers are aligned, the r.m.s.d. is 2.2 Å over 537 C -atom pairs. Inspection of the superimposed homodimers reveals that the chymotrypsin-like cores (domains I and II) align very closely (r.m.s.d. of 0.9 Å over 164 C -atom pairs). When the domain III structures of MERS-CoV pro and SAR-CoV 3CL pro are aligned, the r.m.s.d. is higher (1.4 Å ). The even higher r.m.s.d. that is obtained when the complete homodimers are superimposed (2.2 Å ) reflects a small shift in the orientation of domain III. There is a high degree of conservation of the residues that form the active site in the 3CL pro enzymes of MERS-CoV, BatCoV-HKU4 and SARS-CoV. The residues surrounding the P1 0 , P1 and P2 substrate-binding pockets are particularly well conserved, which may be advantageous for the design of broad-spectrum inhibitors targeting coronaviral 3CL pro enzymes
## Details of the enzyme-product interactions
The fortuitous capture of an enzyme-product complex in crystal forms II and III at high resolution (1.55 and 1.97 Å , respectively) permits a detailed analysis of the intermolecular interactions and provides structural insight into substrate specificity and catalysis, complementing studies of other 3CL pro enzymes . In crystal form II, residues Met301-Gln306 of protomer D are intercalated in the active site of protomer A. The interactions between the C-terminal peptide (product) residues and the active site are illustrated in [fig_ref] a: Stereoview of the hydrogen-bonding interactions [/fig_ref]. The S1 pocket, which is formed by residues Leu27, His41, Phe143-Ser150 and His166-Glu169, is occupied by the P1 residue Gln306, which is required for efficient processing by all coronavirus 3CL pro family members. The side chain of Gln306 is held tightly in the S1 pocket near the catalytic dyad formed by His41 and Ala148 (Cys148 in the wild-type enzyme;via hydrogen bonds between (i) the P1 Gln306 N "2 atom and the side-chain O "1 atom of Glu169 ( the Gln306 O "1 atom and the His166 N "2 atom (2.7 Å ), (iii) the backbone carbonyl O atom of Gln306 and the N "2 atom of His41 (3.0 Å ) and (iv) the Gln306 OXT atom and the backbone amide of Gly146 (3.0 Å ). Additionally, the main-chain amide N atom of Gln306 is hydrogen-bonded to the backbone carbonyl O atom of Gln167 (3.0 Å ). The Ala148 C atom is located 3.3 Å away from the backbone carbonyl C atom of Gln306, confirming that Cys148 would be appropriately positioned to act as the catalytic nucleophile in the active enzyme. Residue Ser1 from protomer B forms hydrogen bonds from its side-chain O (2.8 Å ) and backbone amide N (2.8 Å ) atoms to the carboxylate side chain of Glu169 of protomer A, an interaction that is important for the maintenance of the biological homodimer structure. Likewise, residue Ser1 from protomer A also forms analogous hydrogen-bond interactions with Glu169 in protomer B. The P2 residue, Met305, is nestled into a hydrophobic pocket formed by His41, Gln167, Met168, Asp190, Lys191 and Gln192. In addition to hydrophobic contacts with neighboring side-chain residues, the backbone amide N atom of Met305 is hydrogen-bonded to the O "1 atom of Gln192 (2.9 Å ). Modeling of additional residues into the S2 pocket suggests that this site favors bulkier hydrophobic residues, in accord with the observed preference for leucine in this position of most natural processing sites in the MERS-CoV and SARS-CoV polyproteins. The S3 site is occupied by Val304, the side-chain atoms of which occupy two alternate conformations in crystal form II. Val304 is surrounded by residues Met168, Glu169 and Gln192. Hydrogen-bonding interactions between the backbone amide N atom of Val304 and the backbone carbonyl O atom of Glu169 (3.0 Å ) and between the backbone carbonyl of Val304 and the backbone amide N atom of Glu169 (2.9 Å ) contribute additional stabilizing interactions. The P4 residue, Val303, is bound to the S4 site, which is formed by residues Gln192-Gln195, Met168, Glu169 and Leu170. The side chain of Val303 stacks against the hydrophobic side chain of Leu170. The S5 site is occupied by Gly302, which is held in place primarily by a watermediated hydrogen bond to the Gly302 amide N atom (2.9 Å ) and to the His194 N 1 atom (2.8 Å ). Met301 begins to protrude into the solvent space and does not form any significant contacts with the active-site region other than stacking against the side chain of His194. Comparison of the activesite structure between the enzyme-product complex observed in crystal form II and those of the unbound structures from crystal form I illustrates that upon substrate/ product binding, the residues forming the S1 pocket do not undergo any significant conformational shifts. Slight adjustments of the rotamers of side chains of residues His41, Gln192, Met168, Glu169 and His194 are observed, which are likely to facilitate substrate binding
## An alternate conformation of mers-cov 3cl pro
A distinguishing feature of MERS-CoV 3CL pro crystal form III is the conformational change observed in protomer C. Although protomers A and B exhibit the canonical MERS-CoV 3CL pro homodimer structure, in order to insert its C-terminal tail into the active site of protomer A, protomer C has undergone a substantial conformational change. The core chymotrypsin-like part (domains I and II) of protomer C aligns well with those of protomers A or B (r.m.s.d. of 0.6 Å over 163 C -atom pairs; residues 11-190), but when domains I and II of the three protomers are aligned then domain III of protomer C occupies a very different position than it does in protomers A or B [fig_ref] a: Stereoview of the hydrogen-bonding interactions [/fig_ref]. Conversely, if domain III of protomers A and C are superimposed then they align well (r.m.s.d. of 1.1 Å over 98 C -atom pairs; residues 200-306) but their chymotrypsin-like domains appear to have shifted relative to one another (not shown). Hence, the conformational change affects the relative orientation of the N-and C-terminal parts of the molecule but does not alter the conformations of the individual domains. The first ten residues in protomer C are disordered and the large shift in the orientation of domain III is mediated by a conformational change in the linker loop (Phe188-Ser204; residues His194-Val196 are disordered), in which it moves to cover the active site (Figs. 6b and 6c), potentially impeding access to substrates.
The four molecules found in the asymmetric unit of crystal form II exist as two canonical homodimers (AB and CD), but the C-terminal tail of protomer D is inserted into the active site of protomer A. Therefore, the distortion observed in protomer C of crystal form III is not a necessary prerequisite for the intermolecular interaction that mimics an enzymeproduct complex. The distortion of protomer C in crystal form III is probably tolerated because it does not form a canonical 3CL pro homodimer with a neighboring symmetry mate. A similar situation was observed in the crystal structure of infectious bronchitis virus IBV-CoV 3CL pro (PDB entry 2q6d; 40% sequence identity;. In the case of IBV-CoV 3CL pro three molecules were found in the asymmetric unit, with protomers A and B forming a homodimer and the C-terminal tail of protomer C inserted into the active site of protomer A. When domains I and II in protomers A and C were aligned, they were found to have very similar conformations (r.m.s.d. of 1.0 Å over 171 C -atom pairs), but substantial differences were observed in the orientation of domain III in the two molecules; namely, a 5 Å shift of domain III away from domains I and II. The authors claimed that protomer C represents a novel monomeric form of IBV-CoV 3CL pro that was induced by binding of the C-terminus in the active site of the homodimer. Structural alignment of domains I and II of MERS-CoV protomer C from crystal form III with protomer C from IBV-CoV yields an r.m.s.d. of 1.2 Å over 161 C -atom pairs (residues 1-193). However, there is a significant shift in the orientation of domain III between the two homologs [fig_ref] a: Stereoview of the hydrogen-bonding interactions [/fig_ref]. One difference is that the entire linker region in the IBV-CoV homolog could be modeled into electron density, whereas MERS-CoV 3CL pro residues 194-196 are disordered, resulting in different conformations of the linker loops in the two homologs. Additionally, we do not observe the oxyanion loop (residues 143-148) adopting a 3 10 -helix as seen the IBV-CoV 3CL pro structure. This is likely to be due to differences in the conformation of loop residues 276-293 in the two structures. The larger shift in the position of domain III in the MERS-CoV 3CL pro structure than occurs in the structure of IBV-CoV 3CL pro causes these loop residues to come into close contact with the oxyanion loop in MERS-CoV 3CL pro . As a result, a hydrogen bond is formed between the backbone carbonyl of Leu287 and the side-chain Ser142 O atom, which may prevent the formation of a 3 10helix.
Previous studies with variants of the SARS-CoV 3CL pro enzyme in which the residues involved in dimerization were altered revealed that certain amino-acid substitutions, such as G11A and R289A, cause a structural shift in 3CL pro that disrupts dimerization and gives rise to a shift in the orientation of domain III similar to what we observe in the case of protomer C in MERS-CoV 3CL pro crystal form III ;. Prior studies of monomeric forms of other 3CL pro enzymes revealed that there is very little or no activity in this state. The significant structural flexibility found in the interdomain linker loop region suggests that there may be significant structural plasticity in 3CL pro enzymes that allows the shift between dimeric and monomeric forms. Indeed, prior studies of SARS-CoV 3CL pro protease demonstrated that truncations of the linker loop between the chymotrypsin-like domain and domain III gave rise to a significant reduction in enzymatic activity, confirming that the proper orientation of the linker between domains I/II and domain III is important. Although protomer C of MERS-CoV 3CL pro crystal form III exhibits a large change in the orientation of domain III similar to what was observed in both IBV-CoV 3CL pro and engineered monomers of SARS-CoV 3CL pro , experimental insight into the enzymatic activity of this form is currently lacking. Therefore, more studies need to be conducted to determine whether this conformation is a crystallographic artifact or a monomeric form of the enzyme that is also populated in solution to some degree.
# Conclusion
In summary, we have determined three crystal structures of MERS-CoV 3CL pro representing the free enzyme, an enzymeproduct comple and a crystallographic monomer arising from a conformational change in the linker loop that results in a large shift in the orientation of domain III. The enzyme-product complex reveals the structural basis of substrate recognition by MERS-CoV 3CL pro on the N-terminal side of the scissile bond. The high degree of conservation between the active sites of coronavirus 3CL pro enzymes, particularly in their S2, S1 and S1 0 pockets, suggests that broad-spectrum coronaviral 3CL pro inhibitors can be developed. This objective will be facilitated by determining additional structures of 3CL pro enzymes alone and in complex with substrates and inhibitors.
[fig] Figure 2: Figures were prepared with PyMOL (v.1.5.0.4; Schrö dinger). Structural alignments were performed with either PyMOL or PDBeFold (Krissinel & Henrick, 2004).research papers Acta Cryst. (2015). D71, 1102-1111 Needle et al. MERS-CoV 3C-like protease 1105 [/fig]
[fig] a: Stereoview of the hydrogen-bonding interactions (within 3.2 Å ) between the C-terminal residues 301-306 of MERS-CoV 3CL pro protomer D (crystal form II, C atoms in green) and the active site of protomer A (C atoms in gray). Residue Ser1 (C atoms in yellow) is from protomer B of the homodimer. (b) Stereoview of the active-site residues from protomer A of the free enzyme form (crystal form I, C atoms in magenta) superimposed onto the active site of product-bound protomer A (crystal form II, C atoms in gray). [/fig]
|
SuperPlots: Communicating reproducibility and variability in cell biology
# Introduction
While far from perfect, the P value offers a pragmatic metric to infer whether an observed difference is reproducible and substantial relative to the noise in the measurements [bib_ref] Effect sizes and p values: what should be reported and what should..., Greenwald [/bib_ref]. The P value should be treated as a mere heuristic, interpreted as the degree of compatibility between the observed dataset and a given statistical model. A P value reports the probability that the observed data-or any more extreme values-would occur by chance (the "null hypothesis"). But a small P value does not actually tell us which assumption is incorrect, the null hypothesis or some other assumption of the statistical model (e.g., normal distribution, random sampling, equal variance, etc.). In the case of treating each cell as an n, the assumption that is violated is independent sampling, not necessarily the null hypothesis. The resulting P values are worse than useless: counting each cell as a separate n can easily result in false-positive rates of >50% [bib_ref] Multilevel analysis quantifies variation in the experimental effect while optimizing power and..., Aarts [/bib_ref]. For excellent practical guides to statistics for cell biologists, readers are referred to [bib_ref] T testing the immune system, Lamb [/bib_ref] and [bib_ref] Empowering statistical methods for cellular and molecular biologists, Pollard [/bib_ref]. In this paper, we specifically address simple ways to communicate reproducibility when performing statistical tests and plotting data.
Error bars and P values are often used to assure readers of a real and persistent difference between populations or treatments. P values are based on the difference between population means (or other summary metrics) as well as the number of measurements used to determine that difference. In general, increasing the number of measurements decreases the resulting P value. To convey experimental reproducibility, P values and standard error of the mean should be calculated using biological replicates-independent measurements of a population of interest, typically from independent samples or separate experiments [bib_ref] Pseudoreplication and the Design of Ecological Field Experiments, Hurlbert [/bib_ref] [bib_ref] The problem of pseudoreplication in neuroscientific studies: is it affecting your analysis?, Lazic [/bib_ref] [bib_ref] Replicates and repeats--what is the difference and is it significant? A brief..., Vaux [/bib_ref] [bib_ref] Multilevel analysis quantifies variation in the experimental effect while optimizing power and..., Aarts [/bib_ref] [bib_ref] Criteria for biological reproducibility: what does "n" mean?, Naegle [/bib_ref] [bib_ref] What exactly is 'N' in cell culture and animal experiments?, Lazic [/bib_ref]. Limited time and resources often constrain cell biologists to repeat any particular experiment only a handful of times, so a typical sample size n is often in the single digits. However, if authors assign n as the number of cells observed during the experiment, n may be on the order of hundreds or thousands, resulting in small P values and error bars that do not convey the experimental reproducibility or the cell-level variability.
For example, if a researcher measures the length of 20 neurons in a zebrafish and 20 neurons in a fish exposed to a toxin, the correct n for each condition is 1, because the toxin exposure was only performed once. Without repeating the treatment multiple times with multiple fish, there is no way to know whether any observed difference was from the toxin or due to natural or otherwise uncontrolled differences between those two individual fish. The reader does not care that those two particular fish are different, but that treatments result in a consistent difference across multiple fish. The P value should be calculated to reflect the latter, not the former.
Well-designed studies embrace both cellto-cell and sample-to-sample variation [bib_ref] Points of significance: Sources of variation, Altman [/bib_ref]. Repeatedly quantifying a biological parameter rarely converges on a single "true" value, due to the complexity of living cells or because many biological processes are intrinsically stochastic. Calculating standard error from thousands of cells conceals this expected variability. We have written this tutorial to help cell biologists plot data in a way that highlights both experimental robustness and cell-to-cell variability. Specifically, we propose the use of distribution-reproducibility "SuperPlots" that display the distribution of the entire dataset, and report statistics (such as means, error bars, and P values) that address the reproducibility of the findings.
## What population is being sampled?
To clarify what your sample size n should be, ask yourself: What population are you trying to sample? The choice of n determines the population being evaluated or compared [bib_ref] Criteria for biological reproducibility: what does "n" mean?, Naegle [/bib_ref] [bib_ref] What exactly is 'N' in cell culture and animal experiments?, Lazic [/bib_ref] [bib_ref] Empowering statistical methods for cellular and molecular biologists, Pollard [/bib_ref]. A typical cell biology experiment strives to draw general conclusions about an entire population of cells, so the sample selection should reflect the breadth of that population. For example, to test if a treatment changes the speed of crawling cells, you could split a flask of lymphocytes into two wells, treat one well with a drug of interest and one with a placebo, and then track individual cells in each of the two wells. If you use each cell as a sample (n = number of cells), the two populations you end up comparing are the cells in those two particular wells. Multiple observations within one well increase the precision for estimating the mean for that one sample, but do not reveal a truth about all cells in all wells. By repeating the experiment multiple times from new flasks, and using each experiment as a sample (n = number of independent experiments), you evaluate the effect of the treatment on any arbitrary flask of similar cells. (For more examples, see If you are interested only in cell-to-cell variability within a particular sample, then n could be the number of cells observed. However, making inferences beyond that sample is difficult, because the natural variability of individual cells can be overshadowed by systematic differences between biological replicates. Whether caused by passage number, confluency, or location in the incubator, cells often vary from sample to sample and day to day. For example, an entire flask of cells can be described as "unhappy." Accordingly, cells from experimental and control samples (e.g., tubes, flasks, wells, coverslips, rats, tissue samples, etc.) may differ from each other, regardless of the experimental treatment. When authors report the sample size as the number of cells, the statistical analysis cannot help the reader evaluate whether differences are due to the intended treatment or sample-to-sample variability. We are not prescribing any specific definition of n; researchers should consider what main source of variability they hope to overcome when designing experiments and statistical analyses [bib_ref] Points of significance: Sources of variation, Altman [/bib_ref].
## Statistics in cell biology typically assume independent tests of a hypothesis
Analysis becomes challenging when the experimental unit-the item that can be randomly assigned to a treatment-is different than the biological entity of interest. For example, we often care about how individual cells react to a treatment, but typically treat entire dishes of cells at a time. To test the hypothesis that two treatments or populations are different, the treatment must be applied or the populations sampled multiple times. Neighboring cells within one flask or well treated with a drug are not separate tests of the hypothesis, because the treatment was only applied once. But if individual cells are microinjected with a drug or otherwise randomly assigned to a different treatment, then each cell really can be a separate test of a hypothesis.
Finding truly independent groups and deciding what makes for a good biological replicate can be challenging [bib_ref] Replicates and repeats--what is the difference and is it significant? A brief..., Vaux [/bib_ref] [bib_ref] Points of significance: replication, Blainey [/bib_ref] [bib_ref] Criteria for biological reproducibility: what does "n" mean?, Naegle [/bib_ref] [bib_ref] What exactly is 'N' in cell culture and animal experiments?, Lazic [/bib_ref]. For example, is it acceptable to run multiple experiments from just one thawed aliquot of cells? Is it necessary to generate multiple knockout strains? Is it sufficient to test in one cell line? There's no single right answer: each researcher must balance practicality with robust experimental design. At a minimum, researchers must perform an experiment multiple times if they want to know whether the results are robust.
## Calculating p values from cell-level observations
Cell biologists often observe hundreds of cells per experiment and repeat an experiment multiple times. To leverage that work into robust statistics, one needs to take into account the hierarchy of the data. Combining the cell-level data from multiple independent experiments squanders useful information about run-to-run variability . There is ample literature about the analysis of this type of hierarchical data [bib_ref] A study of clustered data and approaches to its analysis, Galbraith [/bib_ref] , which takes into account both the variance within a sample and the clustering across multiple experimental runs [bib_ref] Multilevel analysis quantifies variation in the experimental effect while optimizing power and..., Aarts [/bib_ref] , or that propagate the error up the chain, such as a nested ANOVA . Recently, statisticians have proposed a Bayesian approach to multilevel analysis [bib_ref] A Bayesian predictive approach for dealing with pseudoreplication, Lazic [/bib_ref]. For a detailed resource on hierarchical data analysis, see.
A simple approach-which permits conventional t test or ANOVA calculations-is to pool the cell-level data from each experiment separately and compare the subsequent sample-level means [bib_ref] Statistics notes. Units of analysis, Altman [/bib_ref] [bib_ref] A study of clustered data and approaches to its analysis, Galbraith [/bib_ref] [bib_ref] The problem of pseudoreplication in neuroscientific studies: is it affecting your analysis?, Lazic [/bib_ref]. For example, if you have three biological replicates of control and treated samples, and you measure the cell diameter of 200 cells in each sample, first calculate the mean of those 200 measurements for each sample, then run a t test on those sample means (three control, three treated). When using this simplified method, it is best to keep the number of observations per sample similar, because each sample gets the same weighting in the analysis.
While pooling dependent observations together avoids false positives [bib_ref] A study of clustered data and approaches to its analysis, Galbraith [/bib_ref] [bib_ref] Multilevel analysis quantifies variation in the experimental effect while optimizing power and..., Aarts [/bib_ref] , this simple approach might fail to detect small but real differences between groups, where more advanced techniques may prove to be more powerful. However, increasing the number of biological replicates usually has a larger influence on the statistical power than measuring many more cells in each sample [bib_ref] Multilevel analysis quantifies variation in the experimental effect while optimizing power and..., Aarts [/bib_ref]. While n of 3 is often considered a pragmatic minimum in cell biology [bib_ref] Criteria for biological reproducibility: what does "n" mean?, Naegle [/bib_ref] , distinguishing more subtle observed differences will require planning for more biological replicates and/or harnessing the power of more robust statistical analyses.
## Communicating variability with superplots
After analyzing hundreds of cells across multiple rounds of experimentation, it would be useful to incorporate both the cell-level variability and experimental repeatability into a single diagram. In A, the plots have small error bars and P values, which should raise red flags given how difficult it would be to replicate a cell biology experiment with identical results and/or to repeat it hundreds of times, which such miniscule P values imply. Bar graphs are problematic because they obscure the distribution of cell-level data as well as the sample-to-sample repeatability [bib_ref] Beyond bar and line graphs: time for a new data presentation paradigm, Weissgerber [/bib_ref]. While beeswarm, box-and-whisker, and violin plots are great at conveying information about the range and distribution of the underlying data, plotting the entire dataset does not make it appropriate to treat repeated measurements on the same sample as independent experiments.
Lord et al.
Journal of Cell Biology Therefore, we suggest authors incorporate information about distribution and reproducibility by creating "SuperPlots," which superimpose summary statistics from repeated experiments on a graph of the entire cell-level dataset . SuperPlots convey more information than a conventional bar graph or beeswarm plot, and they make it clear that statistical analyses (e.g., error bars and P values) are calculated across separate experiments, not individual cellseven when each cell is represented on the plot. For example, the mean from each experiment could be listed in the caption or plotted as a larger dot on top of the many smaller dots that denote individual measurements.
When possible, it is best to link samples by run, for instance, by color-coding the dots by experiment or a line linking paired measurements together . These linkages convey the repeatability of the work: readers learn more if they know that one experiment exhibited high readings across the board than if they have to guess the trend in each sample. Linking data can . The importance of displaying reproducibility. Drastically different experimental outcomes can result in the same plots and statistics unless experiment-to-experiment variability is considered. (A) Problematic plots treat n as the number of cells, resulting in tiny error bars and P values. These plots also conceal any systematic run-to-run error, mixing it with cell-to-cell variability. (B-D) To illustrate this, we simulated three different scenarios that all have identical underlying cell-level values but are clustered differently by experiment: B shows highly repeatable, unclustered data, C shows day-to-day variability, but a consistent trend in each experiment, and D is dominated by one random run. Note that the plots in A that treat each cell as its own n fail to distinguish the three scenarios, claiming a significant difference after drug treatment, even when the experiments are not actually repeatable. To correct that, "SuperPlots" superimpose summary statistics from biological replicates consisting of independent experiments on top of data from all cells, and P values were calculated using an n of three, not 300. In this case, the cell-level values were separately pooled for each biological replicate and the mean calculated for each pool; those three means were then used to calculate the average (horizontal bar), standard error of the mean (error bars), and P value. While the dot plots in the "OK" column ensure that the P values are calculated correctly, they still fail to convey the experiment-to-experiment differences. In the SuperPlots, each biological replicate is color-coded: the averages from one experimental run are yellow dots, another independent experiment is represented by gray triangles, and a third experiment is shown as blue squares. This helps convey whether the trend is observed within each experimental run, as well as for the dataset as a whole. The beeswarm SuperPlots in the rightmost column represent each cell with a dot that is color-coded according to the biological replicate it came from. The P values represent an unpaired two-tailed t test (A) and a paired two-tailed t test (B-D). For tutorials on making SuperPlots in Prism, R, Python, and Excel, see the supporting information.
Lord et al.
Journal of Cell Biology also eliminate the need to normalize data in order to directly compare different experimental runs. Often, multiple experiments might all exhibit the same trend, but different absolute values . By encoding the biological replicate into the data, such trends can be revealed without normalizing to a control group: P values can then be calculated using statistical tests that take into account linkages among samples (e.g., a paired or ratio t test). In fact, not taking into account linkages can make the t test too conservative, yielding false negatives [bib_ref] A study of clustered data and approaches to its analysis, Galbraith [/bib_ref]. An impressive amount of information can be depicted by color-coded beeswarm SuperPlots (see , rightmost plots), where each cell-level datapoint divulges which experiment it came from [bib_ref] A study of clustered data and approaches to its analysis, Galbraith [/bib_ref] [bib_ref] Data visualization, bar naked: A free tool for creating interactive graphics, Weissgerber [/bib_ref]. This helps convey to the reader whether each experimental round gave similar results or if one run biases the conclusion . The summary statistics and P values in beeswarm SuperPlots are overlaid on the color-coded scatter. (See , and for tutorials on how to make beeswarm SuperPlots in Prism, Python, R, and Excel using Data S1.) Whatever way authors choose to display their data, it is critical to list the number of independent experiments in the figure or caption, as well as how the means and statistical tests were calculated.
## Error bars that communicate reproducibility
The choice of error bars on a SuperPlot depends on what you hope to communicate: descriptive error bars characterize the distribution of measurements (e.g., standard deviation), while inferential error bars evaluate how likely it is that the same result would occur if the experiment were to be repeated (e.g., standard error of the mean or confidence intervals; [bib_ref] Error bars in experimental biology, Cumming [/bib_ref]. To convey how repeatable an experiment is, it is appropriate to choose inferential error bars calculated using the number of independent experiments as the sample size. However, calculating standard error of the mean by inputting data from all cells individually fails in two ways: first, the natural variability we expect from biology would be better summarized with a descriptive measure, like standard deviation; and second, the inflated n produces error bars that are artificially small (due to √n in the denominator) and do not communicate the repeatability of the experiment.
The problems with calculating error bars using cell count as the sample size are illustrated by comparing the error bars in A to those in , B-D: when each cell measurement is treated as an independent sample, the standard error of the mean is always tiny, whether or not there is variability among experimental replicates. In contrast, the error bars calculated using biological replicates grow when the results vary day to day. In cases where displaying every data point is not practical, authors should consider some way of representing the cell-to-cell variability as well as the run-to-run repeatability. This could be error bars that represent the standard deviation of the entire dataset, but with P values calculated from biological replicates.
# Conclusions
When calculating your P value, take a moment to consider these questions: What variability does your P value represent? How many independent experiments have you performed, and does this match with your n? (See for practical examples of this analysis.) We encourage authors and editors to focus less on reporting satisfying yet superficial statistical tests such as P values, and more on presenting the data in a manner that conveys both the variability and the reproducibility of the work. is a tutorial for making Su-perPlots in Excel. is a tutorial for making SuperPlots in R. is a tutorial for making SuperPlots in Python. shows how the choice of n influences conclusions. Data S1 is the raw data used to generate Figs. S4 and S5. . Other plotting examples. Bar plots can be enhanced even without using beeswarm plots. (A) Bar plots that calculate P and standard error of the mean using the number of cells as n are unhelpful. (B) A bar graph can be corrected by using biological replicates to calculate P value and standard error of the mean. (C) Dot plots reveal more than a simple bar graph. (D and E) Linking each pair by the replicate conveys important information about the trend in each experiment. (F) A SuperPlot not only shows information about each replicate and the trends, but also superimposes the distribution of the cell-level data, here using a violin plot.
# Online supplemental material
# Supplemental material
Lord et al.
Journal of Cell Biology S1 If your P value looks too good to be true, it probably is https://doi.org/10.1083/jcb.202001064 . Tutorial for making SuperPlots in Prism. We describe how to make SuperPlots in GraphPad Prism 8 (version 8.1.0) graphing software. If using other graphing software, one may create a separate, different colored plot for each replicate, then overlay those plots in software like Adobe Illustrator.
(A) When adding data to the table, leave a blank row between replicates. (B) Create a new graph of this existing data; under type of graph select "Column" and "Individual values," and select "No line or error bar." (C) After formatting the universal features of plot from B (e.g., symbol size, font, axes), go back to the data table and highlight the data values that correspond to one of the replicates. Under the "Change" menu, select "Format Points" and change the color, shape, etc. of the subset of points that correspond to that replicate. (D) Repeat for the other replicates to produce a graph with each trial color coded. (E and F) To display summary statistics, take the average of the technical replicates in each biological replicate (so you will have one value for each condition from each biological replicate), and enter those averages into another data table and graph. Use this data sheet that contains only the averages to run statistical tests. (G) To make a plot that combines the full dataset with the correct summary statistics, format this graph and overlay it with the above scatter SuperPlots (in Prism, this can be done on a "Layout"). This process could be tweaked to display other overlaid, color-coded plots (e.g., violin). , enter the values for the first replicate for the first condition into column B (highlighted in yellow), the second condition into column D (highlighted in yellow), and continue to skip columns between datasets for the remaining conditions and replicates (in this example, replicate 2 is highlighted in green and replicate 3 is in orange). For example, "Treatment A" could be control cells and "Treatment B" could be drug-treated cells. Label the empty columns as "x" and, starting with column A, enter random values to generate the scatter effect by using the formula "=RANDBETWEEN(25, 100)". To create a gap between the datasets A and B, use larger X values for treatment B by entering the formula "=RANDBETWEEN(225, 300)". (B) Highlight all the data and headings. In the insert menu, expand the charts menu to open the "Insert Chart" dialog box. Select "All Charts," and choose "X Y Scatter." Select the option that has Y values corresponding to your datasets. (In Excel for Mac, there is not a separate dialog box. Instead, make a scatter plot, right click on the plot and select "Select Data," remove the "x" columns from the list, then manually select the corresponding "X values =" for each dataset.) (C) Change the general properties of the graph to your liking. In this example, we removed the chart title and the gridlines, added a black outline to the chart area, resized the graph, adjusted the x axis range to 0-325, removed the x axis labels, added a y axis title and tick marks, changed the font to Arial, and changed the font color to black. This style can be saved as a template for future use by right clicking. We recommend keeping the figure legend until the next step. (D) Next, double click the graph to open the "Format Plot Area" panel. Under "Chart Options," select your first dataset, "Series Treatment A (replicate 1)." (On a Mac, click on a datapoint from one of the replicates, right click and select "Format Data Series.") Select "Marker" and change the color and style of the data points. Repeat with the remaining datasets so that the colors, shapes, etc. correspond to the biological replicate the data points came from. Delete the chart legend and add axis labels with the text tool if desired. (E) Calculate the average for each replicate for each condition, and pair this value with the X coordinate of 62.5 for the first treatment, and 262.5 for the second treatment to center the values in the scatterplot. Then, click the graph, and under the "Chart Design" menu, click "Select Data." Under "Legend Entries (Series)," select "Add" and under series name, select the three trial names, then select all three X and Y values for first treatment condition for "Series X Values" and "Series Y Values," respectively. Repeat for the second treatment condition, and hit "OK." (F) On the chart, select the data point corresponding to the first average and double click to isolate the data point. Format the size, color, etc. and repeat for remaining data points. (G) Optional: To add an average and error bars, either generate a second graph and overlay the data, or calculate the average and standard deviation using Excel and add the data series to the graph as was done in E and F, using the "-" symbol for the data point. . Tutorial for making SuperPlots in R. Here is some simple code to help make SuperPlots in R using the ggplot2, ggpubr, dplyr, and ggbeeswarm packages. Dataset that can be renamed "combined" is included in the supporting information. Lines separated by semicolons: ReplicateAverages <-combined % >% group_by(Treatment, Replicate) %>% summarise_each(list(mean)); ggplot(combined, aes(x=Treatment,y=Speed,color=factor(Replicate))) + geom_beeswarm(cex=3) + scale_colour_brewer(palette = "Set1") + geom_beeswarm(data=ReplicateAverages, size=8) + stat_compare_means(data=ReplicateAverages, comparisons = list(c("Control", "Drug")), method="t.test", paired=TRUE) + theme(legend.position="none"). is provided online. shows how the choice of n influences conclusions.
A supplemental dataset is also available online. Data S1 is the raw data used to generate Figs. S4 and S5. . Tutorial for making SuperPlots in Python. Here is some simple code to help make SuperPlots in Python using the Matplotlib, Pandas, Numpy, Scipy, and Seaborn packages. Dataset that can be renamed "combined.csv" is included in the supporting information. Lines separated by semicolons: combined = pd.read_csv("combined.csv"); sns.set(style="whitegrid"); ReplicateAverages = combined.groupby(['Treatment','Replicate'], as_index=False).agg({'Speed': "mean"}); ReplicateAvePivot = ReplicateAverages.pivot_table(columns='Treatment', values='Speed', index="Replicate"); statistic, pvalue = scipy.stats.ttes-t_rel(ReplicateAvePivot['Control'], ReplicateAvePivot['Drug']); P_value = str(float(round(pvalue, 3))); sns.swarmplot(x="Treatment", y="Speed", hue="Replicate", data=combined); ax = sns.swarmplot(x="Treatment", y="Speed", hue="Replicate", size=15, edgecolor="k", linewidth=2, data=ReplicateAverages); ax.legend_. remove(); x1, x2 = 0, 1; y, h, col = combined['Speed'].max() + 2, 2, 'k'; plt.plot , [y, y+h, y+h, y], lw=1.5, c=col); plt.text((x1+x2)*.5, y+h*2, "P = "+P_value, ha='center', va='bottom', color=col).
[fig] Fig: S1 shows other plotting examples.. S2is a tutorial for making SuperPlots in Prism. [/fig]
|
Decision frameworks for restoration & adaptation investment–Applying lessons from asset-intensive industries to the Great Barrier Reef
Asset-intensive industries (including water and power utilities, mineral resources and energy) are those which require significant levels of capital investment in their assets in order to operate. These industries face challenges from uncertainty in resource availability and demand for end products, the intricate and complicated nature of their assets, and the complexity of the economic, ecological and social settings in which they operate. In these industries, the application of decision frameworks that account for this uncertainty and complexity in guiding asset investment and development is standard practice. Lessons from asset-intensive industries were applied during the concept feasibility phase of the Reef Restoration and Adaptation Program (RRAP) to establish the investment case for research and development into interventions to help the Great Barrier Reef (GBR) resist, adapt to, and recover from the impacts of climate change. The authors worked with RRAP partners to define a decision framework that included structured decision-making processes (SDM), a cost-benefit analysis (CBA), and a value of information (VoI) analysis, to establish the investment case for intervening on the GBR which led to success in securing Australian Government commitment for the next phase of the Program. With climate change expected to drive increased demand for significant levels of restoration and adaptation investment in large integrated social, ecological and economic assets (such as the GBR), the lessons from RRAP offer insights for the application of decision frameworks to inform public and private investment priorities.PLOS ONEPLOS ONE | https://doi.org/10.1371/journal.pone.
# Introduction and background
## The challenge facing the great barrier reef
The Great Barrier Reef (GBR) is the world's largest living structure, an ecosystem home to a wealth of marine biodiversity unmatched anywhere in the world, a global icon so exceptional that it has been inscribed on the United Nations (UN) World Heritage List since 1981 in recognition of its Outstanding Universal Value.
In 2014, the Great Barrier Reef Marine Park Authority (GBRMPA) identified deterioration in key habitats, species and ecosystem processes from the cumulative effects of declining water quality from land-based runoff, marine pests and cyclones. More recently, GBRMPA identified the risk posed by climate change, specifically increases in sea temperatures leading to coral bleaching, ocean acidification and increasingly frequent and severe weather events, as the most significant threat to the GBR's long term health. There has been very little rigorous scientific investigation of the prospects for at-scale reef restoration and adaptation. The methods that have been used at small scales are not suitable for largerscale efforts and the scale of the scientific endeavour has been inconsistent with the magnitude and complexity of the challenge. For coral reefs to survive and prosper in a likely warmer future, investment in new interventions to support reef restoration and adaptation are urgently needed.
## The reef restoration and adaptation program
The Reef Restoration and Adaptation Program (RRAP) brings together world-leading experts to help the GBR resist, adapt to, and recover from the impacts of climate change. It aims to provide reef managers with a toolkit of innovative interventions to be implemented at scale, through comprehensive research and development (R&D). An initial 18-month, AUD 6 million concept feasibility phase of RRAP was tasked with examining the overall feasibility of undertaking restoration and adaptation interventions to protect key ecological functions and economic and social values of the GBR.
## Establishing the investment case
Over the first 12 months of this concept feasibility phase, work progressed to mature understanding of the relevant issues and uncertainties in the assessment of the feasibility, costs, risks and benefits of the numerous solutions and techniques proposed. With 6 months of the concept feasibility phase remaining and a need to begin developing the investment case, the following had been established: 1) the extent of uncertainty in the resilience and response of ecological processes to climate change scenarios, across spatial and temporal scales, 2) confirmation of the myriad uncertainties that impacted intervention development, assessment and selection, including and not limited to intervention costs, maturity, efficacy and deployment scale, 3) the extent of uncertainty in understanding of values across ecological, economic and social dimensions, and 4) challenges associated with a deep and heterogeneous knowledge base across stakeholder, regulatory, engineering, ecological, social and economic dimensions. These issues and uncertainties posed a challenge to the Program in being able to establish defensible costs and benefits associated with interventions and thus successfully establish an investment case for the Program.
## Decision frameworks in asset-intensive industries
Given these challenges, the potential applicability of decision-making approaches used successfully in complex asset-investment decision contexts, to focus the activities and outputs of the concept feasibility phase towards the investment case, was established.
Asset-intensive industries (for example, water and power utilities, mineral resources and energy) are those which require significant levels of capital investment in their physical assets in order to operate, usually preceded by a high-cost research and development stage. These industries face challenges from 1) uncertainty in resource availability and market demand for end products, 2) the intricate and complicated nature of their assets, and 3) the complexity of the economic, ecological and social settings in which they typically operate. In these industries, it is standard practice to apply frameworks drawn from decision science that account for this uncertainty and complexity in guiding asset investment and development. Examples of the application of decision frameworks and structured decision-making processes can be found in the energy sector, the minerals sectorand water and power utilities, as well as in other large industries where the theory and practice is progressed (the defense sectorand the pharmaceutical industry.
The concept of decision frameworks emanates from operations research and decision science, which comprise a suite of processes and tools used to inform managerial decision-making. By focusing on decisions as the fulcrum between the potential solutions to problems and the action taken to solve them, operations research and decision science provide approaches for characterising complex problems, and for identifying, assessing and committing to solutions to those problems. For this reason, analytical processes and tools from operations research and decision science are used extensively in complex asset-intensive industries to inform decision-making. The science underpinning operations research and decision analysis has been summarised in Decision Analysis: An Overviewfrom the detail provided in multiple publications.
The common challenges that impact the establishment of investment cases in assets, in both asset-intensive industries and for large integrated social, ecological and economic assets such as the GBR include 1) complexity of the asset, 2) disparate and heterogenous information streams, 3) need for multi-functional / multi-disciplinary stakeholder input and engagement, 4) uncertainty in asset performance in the face of external pressures and threats and 5) preponderance of potential solutions.
The application of approaches from operations research and decision science to large integrated social, ecological and economic assets such as the GBR have gained momentum in recent years. In the GBR specifically, they had previously been identified as having potential utility in helping support the decision-making needs of GBR managers, and has recently been applied directly on decisions for the GBR in helping establish and prioritise monitoring and reporting efforts. There are also examples of application to environmental issues in small community contexts, and on a preliminary scale to coral reef ecosystems.
Compared to these earlier examples, the application of approaches from operations research and decision science described in this paper is novel as it 1) is focussed on a strategic decision (as opposed to tactical), 2) describes the application of multiple decision support methods and tools as part of an overall decision framework (as opposed to a single method) and 3) is a practical application (as opposed to a theoretical application) to establish an investment case that has since been funded.
The question being asked during the concept feasibility study was: Are the approaches used to inform decision-making on complex problems in asset-intensive industries useful for informing decision-making on complex problems in large, integrated, social, ecological and economic assets?
## Decision science processes and methods
The approaches were designed to test the application of decision frameworks from the assetintensive industries to investment in restoration and adaptation on the Great Barrier Reef (GBR.) A three-stage approach was used to identify, develop and execute the activities required to inform the investment case for RRAP. This three-stage approach is summarised below in
## Establish a decision framework
Decision framework. Decision frameworks are required to characterise the interaction between the people, processes and systems for decision-making about a complex asset. Decision frameworks existing in multiple settings and can be described as the principles, processes, and practices to proceed from information and desires to choices that inform actions and outcomes. The visualisation of a comprehensive decision framework in developed by the authors, draws on this literature, as well as our view of best-practice from our experience in asset-intensive industries. This is the theoretical framework through which the RRAP complications were proposed to be addressed. Each of the core elements of this decision framework are described in more detail below. The definition used in this scope for a decision framework is:
## Plos one
Decision frameworks for restoration & adaptation investment in the Great Barrier Reef "Decision Framework: The architecture of people, processes and systems used to make decisions".
Decision hierarchy. For RRAP we first established the decision hierarchy which is a key element of a coherent decision framework in large, complex assets. They are used in the early phase of large developments and in the management of operating large assets to characterise and organise problems and opportunities according to the potential value proposition of successful solutions and the complexity and uncertainty that impacts identification and selection of potential solutions. Decision hierarchies are used in asset management contexts, business case development, software development for decision-makers, and in theory of decision-making. The definition of decision hierarchy used by the authors draws on this literature, as well as our view of best-practice definition from our experience in asset-intensive industries, and is defined as below:
"Decision Hierarchy: The organisation of the decisions that need to be made, supported by who makes those decisions and how they relate to objectives, value drivers and boundaries across the decision landscape".
Objectives hierarchy. The setting of a coherent objectives hierarchy is critical to the success of any project or program that would benefit from implementing a decision framework. They are used in sustainability contexts, in contexts where asset developers are trying to balance financial and broader organisational aspirations, where public agencies are seeking to develop social infrastructure, and numerous applications are presented in the earliest decision science literature. The definition of objectives hierarchy used by the authors draws on this literature, as well as our view of best-practice definition from our experience in asset-intensive industries, and is defined as below:
"Objectives Hierarchy: The arrangement of organisational, program and / or project objectives into relevant hierarchical levels".
The objectives hierarchy should be derived directly from the strategy for the program as set by the appropriate governance body. The decision support experts serve the dual role of providing subject matter expertise to support relevant processes and activities that inform decision making, and, of providing assurance for the governance group ensuring there is an effective link between decision-making and strategy.
Decision processes. The decision processes used by the program are instrumental to successful alignment of stakeholders and thus to executing quality decisions. The importance of effective business processes is well understood in management circlesand in the decision sciences. Clearly understanding and planning decision processes, as well as identifying the relevant processes for each decision in the decision hierarchy will maximise the chance that decision-making is smooth and aligned. The definition of decision processes used by the authors is as below:
"Decision Processes: The processes by which decisions are framed, choices are identified, developed and logically analysed for their consequences and trade-offs, and commitments to action are made, across the decision landscape".
Stakeholder engagement. The importance of stakeholder engagement in a decision framework is paramount as alignment among the relevant people, using the right processes and tools, and the right knowledge, is necessary for quality decisions. A key part to this is ensuring the relevant stakeholders are prepared to undertake the discussions required to commit to action. There is strong evidence throughout decision science and management literature that stakeholder engagement and participation is key to decision-making success.
Knowledge base, management information systems and decision support systems. Decision processes require an appropriate knowledge base upon which to make the decisions. This includes accessing the right modelling, data and information, as well as the use of relevant management information systems and decision support systems where appropriate. Ensuring that the best possible knowledge is accessible to decision-makers is an important aspect in the development and management of large asset basesand is described variously in dedicated journal publications (e.g., Decision Support Systems Journal). The availability of knowledge through management information systems and the utility of decision support systems are characterised widely in operations research, decision science and information management literature. The definition of knowledge base, management information system and decision support system used by the authors are as below:
"Knowledge Base: The data, modelling and information available pertaining to the asset, including understanding of provenance and uncertainty, used to inform decision-making".
"Management Information System: The computerised system that gathers data from multiple sources and makes it available to users (including synthesis) to support quality decisionmaking".
"Decision Support System: The computerised system that gathers data from identified sources, synthesises it, and makes it available to users in accordance with specified decision processes to support quality decision-making on specific semi-structured and unstructured decision problems". Quality decisions. The comprehensive decision framework is brought together by the requirements for quality decisions, which include effectiveness, transparency, defensibility, consistency, efficiency and an understanding of certainty. Each of these requirements maps to a function within the decision framework for research or development programs as shown below in; for example, effectiveness maps to the objectives hierarchy through outcomes, and consistency maps to the decision processes through process. These requirements are described variously in theoretical texts for decision sciences.
Value of information and work breakdown. There are two decision framework functions that serve as key planning interfaces with the wider research and development program in which it sits. The decision framework requirement for efficiency is made functional through value of information (VoI) method that will help with prioritising where effort could be spent in the rest of the program to decrease uncertainty in the knowledge base-this method is detailed later in this paper. The decision framework requirement for certainty is made functional through a work breakdown of the rest of the program that focuses effort on decreasing uncertainty in the knowledge base as prioritised by the decision framework. As such, a comprehensive decision framework should help the program managers to effectively prioritise effort to efficiently increase certainty to inform higher quality decisions.
## Applying the structured decision-making process
Structured decision-making (SDM) is a general term to describe approaches used to help individuals and groups navigate through tough multi-dimensional choices characterised by uncertain science, disparate information, diverse stakeholders and difficult trade-offs. SDM is derived from the science of decision analysis, the foundations of which are summarised in Decision Analysis: An Overview. The primary purpose of SDM is to aid and inform decision-makers, rather than to prescribe a preferred solution. It is based on the concept that quality decisions are those which are based on values (understanding what's important) and
consequences (understanding what's likely to happen). It is aimed at providing consistency, transparency and defensibility to decisions. SDM is particularly helpful for complex problems, like restoration and adaptation on the GBR, where, multiple disciplines (e.g., engineering, environmental science, social science and finance) need to work together to develop solutions that are rigorous, inclusive, defensible and transparent. The factors that contribute to a decision problem being complex are presented in. Because of these factors, complex decisions can be characterised as being high stakes, complicated, having no single overarching expert and requiring justification. The decisions in RRAP meet each of these criteria.
While there are numerous variations of SDM processes used to support complex decisionmaking, almost all share the six core steps that were used in this project as summarised below.
## Implementing the relevant decision analysis methods
The cost-benefit analysis method. This section summarises the cost-benefit analysis (CBA) method, with a more detailed method available in Reef Restoration and Adaptation Program: Cost-Benefit Analysis.
CBA was identified as a potentially relevant analysis method to establish and communicate the value of intervention. CBA attempts to systematically estimate and compare the total benefits and (4) captures linkages between the proposal and other sectors of the economy, especially when proposals are large in scale such that wider economic effects are material. The CBA method that was proposed for the scope of work involved producing an economic model for assessing the benefits of environmental and social protection against the costs of the restoration and adaptation required to do so. The method is based on those presented in The Economics of Groundwater Remediation and Protection, and Environmental and Economic Sustainability, and is aligned to standard cost-benefit analysis processes used by governments globally. The approach describes and measures the case for investment in economic terms, by explicitly monetising costs and benefits to society to the extent possible and appropriate, and adding these to private costs and benefits of a proposed project or action.
In economics, the overall objective of any decision is assumed to be the maximisation of human welfare over time. To compare the different benefit and cost streams over time, the process of discounting is used and amounts over time are expressed as "present values". Economic analysis recommends the decision with the maximum net present value (NPV present value of net benefits, or benefits minus costs, over time) or the highest benefit cost ratio (ratio of the present value of benefits to the present value of costs).
In practice, only some of the benefits identified can be readily quantified and monetised, but those considered in the cost-benefit analysis must be sufficient to determine the case for investment. Those monetised are likely be several of the key private benefits (such as benefits to industry, benefits to livelihoods). External benefits (such as benefits to communities, benefits to ecosystems) are less readily monetised as there is often no market data that could be directly used for their estimation.
The value of information method. This section summarises the value of information (VoI) method, with a more detailed method available in Appendix C of Reef Restoration and Adaptation Program: Cost-Benefit Analysis.
Value of information (VoI) is a field of decision analysis used to estimate the value of acquiring additional data to reduce decision uncertainty. The origins of VoI lie in the work carried out by Raiffa and Schlaifer on applied statistical decision theory at Harvard University in 1961. It has had many applications in different decision-making domains including, medical, economic, environmental, agricultural and ecological research. Whereas basic decision analysis enables decision-makers to identify the best course of action when faced with a situation of uncertainty, VoI provides guidance on how decision-makers might invest in reducing that uncertainty before selecting a course of action.
The starting point is some objective function to be maximised, and a choice between courses of action leading to uncertain outcomes with respect to the objective function. In the case of RRAP the objective function is the intervention strategy, the course of action is implementation with or without additional research into these interventions, and the outcome is the resulting cost and efficacy of the intervention.
VoI can be described as the increase in expected value (and certainty) in the outcome of a course of action with the benefit of additional information relating to this action, compared to a course of action without the benefit of the same information. Alternatively, positive VoI occurs when the value in the outcome of a decision benefiting from additional information, minus the cost of obtaining that information, exceeds the value in the outcome of the same decision without the information.
On this basis, VoI can be used as a quantitative method to estimate the return on investment of carrying out additional research into an objective function (intervention), through some form of data collection exercise, with a view on increasing expected value (and certainty) in the outcome of the course of action chosen (implementation). The value of effort to increase certainty in current knowledge (e.g., research, expert elicitation, technology trials, field surveys, etc) can be assessed through this methodology and is referred to as pre-posterior analysis. A common way to reduce uncertainty is to gain new information, however, not all uncertainty is equally important to reduce. The most important uncertainties are those that, when reduced, will enable a more effective solution and potentially lower cost and higher benefit. New information is of little value if it does not change the efficacy or cost of the proposed action in the absence of said new information.
VOI was identified as a potentially relevant analysis method to establish and communicate the value of research and development given (1) level of uncertainty in RRAP, (2) availability of the proposed Program budget and potential Program benefits from the CBA to inform a VOI analysis and (3) stated Program desire to prioritise effort in reducing uncertainty in the following stage of RRAP.
## Findings
## Decision hierarchy elicitation enables identification of and alignment on evolving high-value decisions
Eliciting and agreeing a decision hierarchy with a broad range of Program participants and stakeholders was found to be useful for RRAP, because as the program evolved, the highest value decision on which to focus RRAP efforts evolved and could be transparently communicated and understood. The process also enabled delineation of the decisions that were within the boundaries of RRAP, on the boundaries of the Program and outside the remit of the Program.
Decision framework processes such as framing workshops proved valuable in establishing the decision hierarchy for RRAP. In multiple workshops with key RRAP members, statements of success for RRAP were determined, which enabled identification of the highest value decision in this phase of the Program which needed to be the subject of the investment case. These workshops elicited, reviewed and refined a set of nine statements completing the sentence "we know we will be successful if. . ." . These statements provided a detailed understanding of the requirements that RRAP leadership have for effectively communicating and justifying their decision-making outcomes both within and outside RRAP. These statements of success continued to be relevant throughout successive framing sessions for other high-value decisions throughout the Program.
The highest-value decision identified for the purpose of establishing the investment case was "What are the optimum Reef restoration and adaptation intervention strategies to invest in and deploy". However, it soon became clear that the high degree of uncertainty within RRAP meant that it was premature to be able to effectively characterise and assess the choices for this decision within the timeframe of the concept feasibility phase. This was a pattern throughout the study-the highest value decision as basis for the investment case evolved a further 3 times as understanding of the uncertainty in the Program increased. The timeline is presented in. The first decision is, however, likely to be the primary decision for subsequent phases of the Program as research activities decrease uncertainty.
The establishment of a decision hierarchy for RRAP was found to be an effective process in focussing decision-science effort on the highest value decisions possible to be answered given the available information (and uncertainty in information) to inform the investment case. The nature of this concept feasibility phase included many completely new areas of enquiry, large amounts of uncertainty, complex interfaces between disciplines, and evolving decision-making needs. The focus on understanding decision hierarchy at all times allowed for several effective . Nine statements of success were identified to inform the decision-making hierarchy for RRAP, as defined at the commencement of the engagement.
## Statements of success "we know we will be successful if. . ."
A Stakeholders are confident that all the relevant inputs are captured that allow us to articulate the value of the interventions in RRAP.
B We can take the outputs from ecological models and effectively convert them into expression of value; with an acceptable degree of rigour; and that all assumptions are considered reasonable / well sourced.
C Decision outcomes are accepted as capturing the range of opinions of stakeholders; and all stakeholders are accepting of the decision outcomes.
D The decision appropriately makes use of the data available at the correct level (i.e., sufficient resolution data in order to make the decision.) E The decision process is adaptable to future changes to the option set, value functions and valuations and is able to be used at various points over the RRAP process (six years).
F The decision can be informed with imperfect data (with disclaimers).
G Non-specialists and non-familiar people with reef modelling can understand how the decision accounts for uncertainty.
H We can demonstrate that, from a cost-benefit perspective, active restoration and adaptation interventions are a valid new management strategy for the Great Barrier Reef (GBR.)
I We can effectively decide the relative effort among the intervention strategies to be developed / researched over the next five years.
https://doi.org/10.1371/journal.pone.0240460.t003
pivots of the elements within the decision framework to follow adjustments in the Program due to this complexity, stimulated by the increased breadth and maturity of knowledge emerging from the Program's social-scientists, scientists, engineers and modellers.
## Sdm was the relevant process for rraps highest value decisions
It was found that structured decision-making (SDM) is a relevant process to be applied to RRAP as the attributes of complex decision problems match the attributes of RRAP. The SDM process was applied to four high-value decisions within RRAP, with two of those instances delivering value to the program objectives for the concept feasibility phase. The other two applications of SDM were found not to be effective at the concept feasibility phase as there was insufficient certainty of knowledge to commit to action; the SDM process is, however, highly likely to be applicable to these problems in later phases of the Program with greater availability of knowledge. Details of how the SDM process was applied to four high-value decisions within RRAP is summarised in .
The SDM processes were effective in aligning stakeholders, decision-makers and information sources from within the complex team to establish the investment case for the program. Two of the decisions will be detailed in the next section:
## Cost-benefit analysis is an effective method to establish and communicate the value of intervention
It was found that cost-benefit analysis (CBA) is a relevant method to communicate the value of intervention on the GBR, and by extension whether it is worth investing in RRAP. This section summarises the application of CBA during the concept feasibility phase and the relevant findings on the applicability of CBA method to RRAP. Detailed accounts of the CBA methodology are available in Reef Restoration and Adaptation Program: Cost-Benefit Analysisand Reef Restoration and Adaptation Program: Intervention Analysis and Recommendations.
The intent of the CBA for RRAP was to demonstrate, within the high degree of uncertainty inherent in the program, whether there is a strong set of intervention options and conditions within which investment in RRAP is favourable, and thus enable decision-makers to determine whether the Program receives funding to progresses to the next stage. Simply, the analysis examines whether RRAP demonstrates enough potential net benefits to continue. The SDM process was used as the framework to guide the CBA, according to the stages summarised in
## Plos one
Decision frameworks for restoration & adaptation investment in the Great Barrier Reef Key insights from the CBA. The CBA was effective in being able to establish that RRAP is an investable proposition. By taking into account the balance of costs and benefits, it was shown that there is significant potential economic upside from RRAP interventions of up to AUD 4.1B net present value (NPV) (2016 dollars, 3.5% discount rate) under base case assumptions, which is equivalent to AUD 28B undiscounted over 60 years. Taking a 10% exceedance probability for 1,000 iterations of sensitivity parameters, the potential upside net benefit from RRAP was shown to be up to AUD 14.5B NPV (2016, 3.5%). Thus, RRAP is likely an investable proposition for further R&D more often than not across a broad range of conditions and potential uncertainties. For more detailed presentation and discussion of these findings, please refer to the reports at the top of this section.
From a cost-benefit perspective, restoration and adaptation intervention is a valid new management strategy for the GBR and further R&D should be invested in. The CBA for RRAP showed that, within the high degree of uncertainty inherent in the Program, there are a strong set of conditions where active restoration and adaptation interventions are likely to deliver positive net benefits. As a result, it was recommended that the Program receive investment to progress to the next stage of R&D.
Key insights and recommendations for using CBA as a decision method for RRAP. The high degree of uncertainty within RRAP at many levels, at least one order of magnitude greater than what would typically be observed during a R&D program concept feasibility phase, means that this CBA should be considered an early attempt to examine the performance of potential program elements. That is, the assessment considered the high degrees of uncertainty within RRAP and examined, within this range, the potential futures for RRAP performance. It did not assume that engineering and scientific action will be taken that result in poor performance; it instead flags where further research will need to be undertaken to reduce uncertainty, better define potential performance, and thus ensure that only well-performing program elements are progressed.
The CBA process should be iterated as uncertainty decreases. It is one tool within RRAP that can be used to ensure only well performing program elements are advanced to the next stage. This informed high-level implications for the next stages of RRAP investment. However, it must be noted that the high degree of uncertainty within the source data means they are necessarily high-level. The SDM process that provided the framework for the CBA, however, is repeatable in subsequent RRAP stages to inform intervention selection and ensure optimised implementation of the Program.
## Value of information is an effective method to establish and communicate the value of research and development
Used in conjunction with the CBA, it was found that value of information (VoI) analysis is a relevant method to communicate the value of R&D into interventions on the GBR, and thus whether it is worth investing in the next R&D phase of RRAP. This section summarises the application of the VoI analysis during the concept feasibility phase and findings on the applicability of VoI methods to RRAP. Detailed accounts of the VoI methodology are available in Reef Restoration and Adaptation Program: Cost-Benefit Analysis.
R&D programs have inherent value, in isolation from the potential value from implementation of that R&D. To this end, we employed an SDM process to examine the inherent value of RRAP R&D. The stages are summarised inKey insights from the value of information analysis. The VoI analysis was able to establish that there is a strong case for the investment in at least one phase of R&D, despite the conservative assumptions made in recognition of the significant uncertainty in the Program. The major insights that were established through the VoI analysis were:
- Doing nothing (no R&D or implementation) is an option, but not one recommended by RRAP -the CBA established that action on the GBR has significant potential upside of between AUD 4.1B and AUD 144B net present value (NPV) (2016, 3.5 percent discount rate), and conducting the R&D phase of the Program inherently retains this as an option value.
- Attempting to deploy interventions without an R&D phase is not recommended-the limited number of interventions that might be possible (i.e., technically feasible and able to gain regulatory approval without R&D) are small-scale, likely to be expensive per unit area, poorly guided, with high ensuing ecological risk and with no guarantee they would deliver net benefits, so R&D is essential to improve these factors.
- Investing AUD 100m in an initial phase of R&D is highly likely to be worthwhile-only a 10 percent improvement in the certainty of a small-scale intervention implementation being successful is required in order to justify the costs of the phase of R&D.
- Securing an additional AUD 50m for an initial phase of R&D is highly likely to be worthwhileif success certainty increases by more than an additional 1.8 percent, the additional investment is worthwhile.
- If implementation does not proceed after five years' R&D, there remain positive benefits of the investment in R&D-the benefits include expenditure multiplier, and research and knowledge
## Plos one
benefits in the Australian economy. Quantification of the expenditure multiplier demonstrates that these benefits alone exceeds the cost of investment under reasonable assumptions.
Key insights and recommendations for using value of information as a decision method for RRAP. The successful application of VoI to establish and communicate the value proposition of investment in R&D provides a template for guiding both later phases of R&D investment, and, prioritising investment in program areas to resolve uncertainty during each phase of R&D. For the former, while a high likelihood of success of small-scale intervention after a second round of R&D is required to deem the second round R&D investment worthwhile, this decision point is four years away, allowing time for better data and information to generate more accurate insights prior to decision-making. For the latter, the ability of this approach to characterise the impact of uncertainty resolution on the ultimate net benefits of the program, enables prioritisation on uncertainty reducing activities in the program, and screening of those interventions where the cost of uncertainty reduction is incommensurate with the potential benefits.
## Implications and recommendations for the application of decision frameworks to establish the investment case for solutions to complex problems in large integrated social, ecological and economic assets
The findings from this application of decision frameworks commonly used in asset-intensive industries to establish the investment case for R&D in restoration and adaptation on the Great Barrier Reef (GBR) has implications for RRAP and for investments in large, integrated social, ecological and economic assets more broadly, especially in the context of climate change:
- Decision frameworks comprising processes and tools from decision science have high utility for generating commitment to action for large-scale restoration and adaptation research, development and intervention.
- Decision hierarchies, their determination and iteration, have high utility in R&D programs subject to high levels of uncertainty to enable rapid pivoting of the application of decision processes and tools to the highest value decisions.
- Structured decision-making (SDM) processes have high utility in R&D programs to generate quality decisions and commitment to action, especially in highly complex settings such as large social, ecological and economic assets with many stakeholders, requiring high levels of investment, and, with impacts of uncertain future scenarios such as climate change.
- The cost-benefit analysis (CBA) method has high utility in R&D programs to articulate the investment case of intervention to preserve large, integrated social, ecological and economic assets.
- The value of information (VoI) method has high utility in R&D programs to articulate the investment case for ongoing R&D into intervention options for large, integrated, social, ecological and economic assets.
Based on the outcomes of this application, and drawing upon the experience of the authors with similar applications in asset-intensive industries, the following recommendations are made for RRAP, and for other programs looking to establish the case for investment in restoration and / or adaptation R&D or intervention in similarly large, integrated, social, ecological and economic assets:
- Programs should consider implementing or upgrading to a comprehensive program-wide decision framework that enables integration of sub-program areas that feed knowledge and a characterisation of uncertainty into program planning and investment prioritisation processes.
- Programs should consider applying a coherent objective setting process, integrated with governance groups and strategic processes, that ensure alignment between program decisionmaking and desired strategic outcomes that can be transparently articulated through investment cases.
- Programs should consider establishing a decision hierarchy, ideally during conception, to map the decisions that need to be made, and by who, and how they relate to objectives, value drivers and boundaries across the program, to enable the highest value investment case to be made.
- Programs should ensure that asset knowledge (data, modelling and information) includes understanding of provenance and uncertainty, to enable investment cases to be framed according to the value proposition despite uncertainty, and, to present the investment case in terms of the value proposition of reducing uncertainty.
# Conclusions
This paper has demonstrated that the application of frameworks commonly used in assetintensive industries, drawn from decision science, were critical in establishing the investment case for R&D in restoration and adaptation on the Great Barrier Reef (GBR) in the context of climate change. The resulting commitment from investors to fund the R&D phase of RRAP has resulted in the largest-ever coordinated action to assist a globally significant social, ecological and economic asset in its fight to survive climate change.
There is significant further work required in this area, below are some proposed areas of future research that could build on the findings in this paper:
- Extend the value of information (VoI) analytical approach, conducted at the program scale, to establish an iterative VoI process within RRAP to assist with prioritising investment in R&D.
- Extend the cost-benefit analysis (CBA) conducted at the program scale, to establish an iterative CBA process within RRAP to assist with prioritising investment in implementation of restoration and adaptation options.
- Characterise a decision framework for the entirety of RRAP, in order to streamline activities towards quality decision-making (that enables efficient, effective, transparent and defensible decisions.)
- Map knowledge and uncertainty within the decision framework for RRAP to enable quality decisions to be made about investing in reduction of uncertainty / improvement of knowledge.
- Investigate more widely the application of coherent decision frameworks than just the restoration and adaptation efforts on the GBR.
- Investigate more widely the application of coherent decision frameworks for informing decision-making on complex problems in large integrated social, ecological and economic assets with many stakeholders.
There are also many interfaces with other areas of research that afford ongoing debate, including:
- Is the monetization of benefits valid enough to justify investment cases for large scale adaptation and restoration, given the uncertainties of translating potential social and environmental values to monetary benefits?
- Do decision frameworks sufficiently characterise the social and political realities of stakeholder decision-making to have utility for investment decision-making?
- How are traditional owner perspectives and value systems appropriately represented through structured decision-making (SDM) processes?
- Are there thresholds whereby the extent of underlying uncertainty in knowledge areas significantly challenges the robustness of investment case recommendations?
# Limitations
There are several limitations in the application of decision processes and tools described in this paper. These should be considered when examining the results and implications presented.
- The high degree of uncertainty within RRAP at many levels of the program means that the application of the cost-benefit analysis (CBA) method should be considered an early attempt at examining the performance of potential program elements, and not as a demonstration of what will be implemented, nor a determinant of where implementation investments will ultimately be made. The analysis provides only high-level insights into which interventions could be invested in.
- The testing of sensitivity presented in this analysis is restricted by what was carried through in source information. This analysis has used source data from other RRAP teams that contains averaged or most likely assumptions or cases, meaning that sensitivity to those assumptions or cases is not demonstrable within these results.
- Economic valuation is restricted to defendable benefits, with likely total benefits being higher. The Program was unable to fill all gaps in understanding benefits at the concept feasibility phase. Similarly, only large-scale (Great Barrier Reef-wide and regional) costs and benefits have been assessed as the Program primarily examined large-scale interventions during this phase.
their Elders past, present and emerging and we acknowledge their continuing spiritual connection to their land and sea. |
The Most Important Social Determinants of Slum Dwellers’ Health: A Scoping Review
Objectives: Given the importance of social determinants of health in promoting the health of slum residents, this study was conducted with the aim of identifying the main dimensions and components of these determinants.Methods: This scoping review study was conducted according to the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews). A comprehensive search was performed of PubMed, ProQuest, Scopus, and Web of Science for articles conducted from 2010 to the end of 2019. Studies were selected based on inclusion criteria, with a special focus on studies dealing with the social determinants of physical and mental health or illness.Results: Thirty-three articles were selected to extract information on the social determinants of health. After reviewing the articles, 7 main dimensions (housing, socioeconomic status of the family, nutrition, neighborhood characteristics, social support and social capital, occupational factors, and health behaviors) and 87 components were extracted as social determinants of health among slum dwellers.Conclusions: This framework could be used by planners, managers, and policy-makers when making decisions affecting the health of these settlements' residents due to the common characteristics of slums around the world, especially in developing countries. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/bync/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. Slum Keywords Slum, informal settlement, irregular settlements, squatter settlements, shantytown, spontaneous settlements, unauthorized settlements, low income settlements, semi-permanent settlements, unplanned settlements, uncontrolled settlements, transitional settlement, marginal settlements, autonomous settlements, favela, bustees, ghetto, poverty area MeSH terms Poverty areas Social determinants of health Keywords Health social determinant-health social determinants MeSH terms Social determinants of health AUTHOR CONTRIBUTIONS
# Introduction
According to the United Nations Center for Human Settlements, an informal settlement is a time-worn area of a city, known for low-quality housing construction, unsafe residential conditions, pollution, overcrowding, and a lack of basic facilities. International organizations have estimated that more than 1 billion people (about 32% of urban communities) live in slums. This living arrangement has grown worldwide since the 1990s, and the number of people living in slums is expected to exceed 2 billion by 2030.
Slums have some characteristics in common; they are areas where there may be proximity effects facilitating the impact of factors such as environmental pollution, massive waste, overcrowding, non-standard houses, and physical hazards (such as accidental burns and fires), on residents' health. The accumulation of problems in these areas endangers the physical, mental, and social health of these areas' residents. Studies have shown that the health status of people living in slums is much worse than that of those living in adjacent urban areas.
Addressing the health of the residents of these areas only by providing health services would not be successful. To improve the health of people living in the outskirts of cities, it is necessary to find determinants that have a major impact on health, as focusing on these determinants could improve the health status of these areas' residents. It seems possible that by improving the social and environmental conditions of slums using the analytical framework of social determinants of health, the physical, mental, and social health status of residents could be improved.
Various studies have emphasized the important role of social determinants of health in the development/eradication of diseases and in the design of necessary interventions based on these determinants. Pawar et al.noted that appropriate interventions to improve the health of slum dwellers need to be designed and implemented based on a correct understanding of the factors affecting health. Social factors are as important as physical factors in determining the health status of residents and the type of interventions that would be appropriate to propose; in other words, public policy that seeks to achieve continuous improvement in social determinants of health, such as income, education, housing, food, security, and neighborhood status, could have positive and long-lasting effects on people's health.
There are various models for social determinants of health, and the World Health Organization (WHO) has also proposed a significant model in this regard; however, despite the common features of slums and the difference in the health status of people residing in these areas and other places, no model focusing on slums has been proposed based on the available evidence and studies.
Given the importance of social determinants of health and their role in the health of slum dwellers, and also in light of the fact that in recent years, several studies have been conducted in different parts of the world to identify social determinants of health among slum dwellers, the main purpose of this study was to extract social determinants of health from various studies conducted in the last decade on people living in slums and provide a comprehensive framework of social determinants of health with dimensions and components affecting health.
This framework could be used by planners, managers, and policy-makers when making decisions affecting the health of these settlements' residents due to the common characteristics of slums around the world, especially in developing countries.
# Methods
This scoping review study was conducted according to the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews)to identify and extract the most important dimensions and components of social determinants of health from various studies conducted in this field without considering and evaluating the methodological quality of the included studies; thus, this study is considered to be a scoping review.
## Data sources and search strategy
A scoping search was performed in PubMed, ProQuest, Web of Science, and Scopus to find articles published during 2010-2019. The search was performed from November 5, 2019 to December 15, 2019. In order to search for studies conducted on social determinants of health among residents of slums, the search domain was first identified from the PubMed site, and then keywords and related words were identified and extracted from various sources. A comprehensive search was performed of the aforementioned databases using different combinations of keywords and Medical Subject Headings (MeSH) terms.summarizes the keywords and MeSH terms used in this study.
## Social determinants of health for slums
## Eligibility criteria
Given that the social determinants of health in slums could be either the determinants of a particular disease or the social determinants of health in general, the inclusion criteria for studies to be included in this survey were as follows: (1) studies published in scientific journals, (2) studies published during 2010-2019, (3) studies published in English, (4) studies investigating social determinants of physical and mental health,studies that pointed to social determinants of any type of mental or physical illness, and (6) studies whose full text was available for data extraction.
The exclusion criteria were as follows: (1) conference papers, (2) studies in languages other than English, (3) studies whose full-text was not available, and (4) studies that were conducted on health or diseases in slums but had no relation to social determinants of health.
## Data extraction
At first, the selection was based on each article's title and abstract in order to identify the articles related to the research question domain. Next, articles were selected based on inclusion criteria with a special focus on studies dealing with the social determinants of physical and mental health or illness. After these 2 steps, the full text of the remaining articles was reviewed for further investigation and extraction of determinants reported to affect health.
A special spreadsheet was designed to systematically extract data from the selected studies. The data extracted from the studies were related to the determinants of physical and mental health or physical and mental illness. The data were screened by the corresponding author (MRG) and then revised and confirmed by the second and third reviewers (FNN and SHMK).
# Synthesis and analysis
The social determinants of health or illness in slums mentioned in the articles were extracted using the data collection forms. After extracting and entering the data in the prepared forms, the data were summarized in terms of similarity and categorized based on the experience and expertise of 2 experts (FNN and SHMK).
Social determinants of health are defined as the social conditions in which people are born, grow, live, work, and grow old. They are social characteristics that are present or occur in the place of residence. The concept of social determinants of health refers to the factors that help people stay healthy until they get sick and receive services.
Social determinants of health include both the specific characteristics of a social context affecting health and the mechanisms by which social conditions affect health. Significant social determinants are those that could potentially be changed by conscious actions.
It is now believed that social factors affecting health, such as income level, education level, occupational factors, nutrition, social class, unemployment, and childhood experiences, are more important than biological factors for the occurrence of diseases, and if ignored, it will be impossible to achieve the goals of the health sector and establish justice in health.
# Ethics statement
In this study, we strived to use the data properly and without prejudice as much as possible. Ethics approval was obtained from the Ethics Committee of the University of Social Welfare and Rehabilitation Sciences (Ethical code: IR.USWR. REC.1398.165).
# Results
## Study selection
The data collected from the aforementioned databases were as follows: 515 records from PubMed, 411 records from Scopus, 1178 records from Web of Science, and 362 records from Pro-Quest. After eliminating duplicates, 2201 records were retrieved. By evaluating the titles and abstracts of the studies and by taking into account the inclusion criteria, 625 papers were selected for full-text evaluation, while 1576 papers did not meet the inclusion criteria. In the full-text evaluation stage, 592 papers were eliminated based on the exclusion criteria, and 33 studies remained for further evaluation.presents the general characteristics of the 33 selected studies. The oldest study was published in 2012, and the most recent study was published in 2019. Among the studies reviewed, 10 were cross-sectional studies, 6 were literature reviews, 5 were secondary analyses, 4 used mixed methods, and 8 studies were designed using other methods, as listed in. Eight studies were conducted in India, 7 studies were conducted in Bangladesh, 2 studies were conducted in Jordan, and the rest of the studies were carried out in various other countries.
## Characteristics of the sources of evidence
# Synthesis of results
After conducting the research steps, the social determinants of health extracted from the studies were divided into 87 components and the following 7 main dimensions: housing, socioeconomic status of the family, nutrition, neighborhood char-acteristics, social support and social capital, occupational factors, and health behaviors. In addition, the extracted components that were repeated in several articles, as well as the number of their repetitions as an indicator of their importance, were also identified.
The most important dimensions and components of social determinants of slum dwellers' health, mentioned in the studies reviewed, are as follows:
## Housing
Housing and its conditions, as mentioned in the reviewed studies, are one of the most important determinants of slum dwellers' health. The structure of housing is associated with certain health consequences, and the improvement of housing status over time affects the health of residents. The components of the housing dimension identified in some of the reviewed studies as factors affecting health are as follows:
Physical structure of housing, lack of strength, durability, and safety of buildings due to the use of brittle, non-durable, and flammable materials; Indoor air pollution and poor ventilation; The presence of small rooms in houses; Unhygienic toilets; Lack of security and fear of destruction due to illegal construction; Contamination with rodents and insects; Adverse social-psychological residential environment; The use of pathogenic and toxic substances in the construction of houses such as asbestos and lead; House seizure (rent, private property, etc.).
## Socioeconomic status of the family
Socioeconomic characteristics of the family, as one of the most important dimensions of social determinants of slum dwellers' health, encompass different components, mentioned in some of the reviewed studies as factors affecting health. These components are as follows:
Parental education level, lack of formal education; Family income, poverty in the family, income insecurity, financial pressure, dissatisfaction with the family's financial situation ; Family size, overcrowded and dense family space; Family religion; Family type (nuclear, extended, single-parent); Length of stay in the neighborhood; Violence, gender discrimination, unfair behaviors, and misbehavior against
## Social determinants of health for slums
women in the family; Violence, gender discrimination, and unfair behaviors against children in the family; Gender and age of family members; Family cultural hobbies (the use of media, reading, etc.); Family wealth status; Family race and ethnic group; Limited travel; Being covered by family support institutions; Having fights, misunderstandings, and conflicts in the family; Undesirable childhood experiences.
## Nutrition
The components of the nutrition dimension, identified in some of the reviewed studies as factors affecting health, are as follows:
Malnutrition, inadequate and poor diet, and micronutrient deficiencies; Obesity; Weak and thin family members; Food contamination; Improper selection of healthy foods; Family food insecurity; Nutritional habits (such as regular consumption of fast food, the number of times one eats vegetables, fruits, and meat); Low-birth-weight babies.
## Neighborhood characteristics
One of the most important dimensions of social determinants of health, which significantly differs in slums from other urban areas and should be given special attention, is neighborhood characteristics, which encompass many components, as follows:
Lack of access to public resources (parks, green space, museums, libraries) ]20,28]; Lack of access to spaces for sports; The presence of stray dogs and livestock in the neighborhood, density of mice; Conflict, such as being annoyed and harassed by neighbors; Lack of access to safe, high quality, and hygienic drinking water; Noise pollution; Environmental pollution of the place of residence, lack of cleanliness and poor sanitation of neighborhoods; Lack of street lighting; Existence of industrial pollution in the neighborhood; Existence of unprotected hazards (such as railways and power lines); Lack of educational facilities and schools and low quality of education; Inadequate and insufficient public health facilities located at long distances; Insecure public space due to gang crime and conflicts, neighbors' crime and conflicts; The quality of service delivery organizations in the neighborhood, such as stores, pharmacies, banks, and offices, and their long distances; Lack of private-sector health facilities; Lack of charitable and non-governmental health facilities; Lack of places for recreation; Inappropriate public transportation network and difficulty in accessing public transportation; Physical hazards such as flood, subsidence, and fire; Improper disposal of waste, environmental pollution with waste, and waste-transfer stations; Inappropriate and open sewer systems; Very high population density in the neighborhood; Air pollution; Unpaved and dirt alleys; Narrow and irregular alleys; Lack of water and electricity supply, informal use of water and electricity resources; Notoriety of the neighborhood; Lack of access to police services.
## Health behaviors or high-risk and harmful behaviors for health
The components of the health behaviors dimension, identified in some of the reviewed studies as factors affecting health, are as follows:
Not performing health screening and annual tests; Not taking care of personal hygiene such as bathing, nail trimming, and tooth-brushing; Addiction of a family member to drugs or alcohol; Parental insensitivity to family health; Excessive smoking; Insufficient physical activity; Stress; Low individual health literacy; Low environmental health literacy.
## Social support, social capital, and deprivation
Issues related to social protection, social exclusion, and social capital were also addressed in some of the reviewed studies as factors affecting the health of residents. These factors are as follows:
Social support, the number and presence of friends and relatives in the neighborhood, and having a large family; Social capital, social cohesion and solidarity, and solidarity at the neighborhood level such as closeness to neighbors; The degree of cooperation and mutual cooperation with neighbors, such as taking care of each other's property; Weakness and insignificance of traditional values; Feeling discredited due to living in an illegal place; The degree of belonging to the neighborhood; Social rejection or deprivation and family isolation; Social participation, engaging in civic activities, membership, and participation in groups and institutions such as mosques, social organizations, and non-governmental organizations; Discrimination and perceived discrimination; Low participation in religious ceremonies.
## Occupational factors
The components of the occupational factors dimension, identified in some of the reviewed studies as factors affecting health, are as follows:
Job satisfaction; Whether the mother and father are employed or unemployed; Job type (such as having informal and non-productive jobs); Workplace conditions; Long working hours with very low salary and income, especially for women and children; Job insecurity.
# Discussion
Residents of slums are faced with increased risks of disease and illness. According to the Commission on Social Determi-Social Determinants of Health for Slums nants of Health, the living environment plays an essential role in the formation and determination of health and well-being. In slums, a fully-shared physical and social environment is likely to have strong proximity effects and play a key role in the health of residents as a social determinant of health. Many studies have been conducted to determine the most important determinants affecting people's health, and different models have been proposed regardless of location. In this study, the most important determinants of health in slums, which have distinctive features from other urban areas, were collected from various studies and presented as a basis for improving the health of these settlements' residents. The main determinants were housing, socioeconomic status of the family, nutrition, neighborhood characteristics, social support and social capital, occupational factors, and health behaviors.
In the most recent model of social health determinants presented by the WHO-which deals with the dimensions of social determinants of health in general, and not social determinants of health among slum dwellers-these determinants have been divided into 2 categories: structural and intermediate. The identified determinants in this study, mostly derived from evidence-based research, are largely consistent with those included in the model presented by WHO and cover a wide range of determinants. However, a different classification was proposed for social determinants of health among residents of slums, because the model presented in this study focuses more on intermediate determinants and living conditions identified in the studies reviewed, while the WHO model focuses more on structural determinants such as governing and policy-making, perhaps because it is easier for studies to intervene on intermediate determinants. In a study by Lilford et al., a different classification was presented for social determinants of health among residents of slums compared to this study. The presence of multiple classifications could lead to the analysis of the issue from different viewpoints, facilitating the design of appropriate interventions. A comparison of the dimensions and components identified in this and other studies showed that although the identified items were mentioned in various studies, their importance and order in studies were different.
In a study by Nekoei-Moghadam et al., housing was identified as a subset of the main determinants, which is inconsistent with the present study results. Given the importance of housing and its components mentioned above, it is one of the most important factors that should be taken into account when designing health promotion interventions for slum dwellers. For example, Alaazi and Aganahnoted that a lack of housing security due to the threat of slum cleansing was a well-known social-psychological stressor affecting the physical and mental health of African slum dwellers.
According to Pawar et al., the most important determinants of diseases among slum dwellers are age, education, social class, family type, and socioeconomic status of the family. In a study by Subbaraman et al., it was noted that low family income and the problems related to living in slums were associated with mental disorders. According to the findings of this study, the family income component repeated in various sources is one of the most important and influential components affecting the health of slum dwellers. Fink et al.noted that most of the health differences observed among residents of slums could be explained by differences in maternal education, family wealth, and access to health services. Chauhan and Dharreported that the mental health status of adolescents living in an informal settlement of India was correlated with gender, social class, education level, employment status, type of work, average monthly income, type of housing, and exposure to the media. Therefore, according to the reviewed studies, the socioeconomic status of the family is considered to be one of the most important dimensions of social determinants of slum dwellers' health. The component of undesirable childhood experiences is also one of the most important social determinants of health, to the point that it could potentially be considered as a separate dimension; however, in this study, since it had common features with family status, it was included in this dimension.
Proximity effects in neighborhoods that help create a healthy environment in slums are likely to be economically affordable and increase investment returns.
In some of the reviewed studies conducted on slums, the neighborhood characteristics were addressed as an important factor affecting the health of slum dwellers. Silva et al.confirmed the presence of a significant relationship between the neighborhood characteristics and mental health of slum dwellers. In another studyconcluded that 4 important factors involved in the development of non-communicable diseases were a lack of clean water, inappropriate transportation system, inadequate physical activity, and low levels of education. Another study concluded that health promotion interventions through the creation of green spaces in slums had a significant impact on residents' health.
Nutrition, health behaviors, and occupational factors, as mentioned in the reviewed studies, are also important social determinants of health that should be considered in the health promotion programs designed for slum dwellers. The results of a study by Mondal et al.showed that behavioral factors, such as tobacco and alcohol consumption, as well as insufficient consumption of fruits and vegetables, affected the prevalence of non-communicable diseases. It was found that food insecurity and lack of food resources were associated with the spread of acquired immune deficiency syndrome in slums. In a study by Gruebner et al., job satisfaction was reported to be among the factors affecting the mental health of slum dwellers.
Pawar et al.showed that social isolation and lack of social support were significantly associated with illness among slum dwellers. In another study, social support was reported to be an important social determinant of hypertension and diabetes among residents of slums. Therefore, the dimensions and components identified in this study could be used as a framework for health promotion interventions designed for slum dwellers. This research is one of the first studies summarizing the results of various studies in this field. Given the common features of slums, there is a need for further research to identify other components and dimensions affecting the health of slum dwellers in order to compare these dimensions with those included in existing models and to clarify their differences and similarities. Since interventions based on social determinants of health require that different organizations play various roles in relation to health, it is suggested that more research should be conducted in the future to clarify the role of other organizations and institutions in promoting the health of these settlements' residents to facilitate the design and implementation of urban programs and interventions in slums based on the proposed framework of social determinants of health with identified dimensions and components.
One of the strengths of this study is the provision of an innovative integrated framework for promoting the health of slum residents, which could be useful and applicable in health promotion measures and intervention in these places. However, a limitation of this research is that only articles in English were reviewed. Moreover, the important dimensions and components of social determinants of health were classified based on their similarities, which need additional research to confirm, and it may be possible to present the classification in other forms as well.
# Conclusion
Given the importance of health as a human right, as well as the low-level of health among residents of slums or informal settlements in most parts of the world, it is recommended that interventions be designed and implemented in these areas within the framework of social determinants of health by taking into account the influential factors collected from the reviewed studies and presented herein. In addition to providing adequate and acceptable health services, special attention should be paid to social determinants of health in order to compensate for inadequacies in physical and mental health and to improve the health of these settlements' residents, as well as to establish justice in health. Eventually, this framework could be used by planners, urban managers, and policy-makers when making decisions that affect the health of these settlements' residents due to the common characteristics of slums around the world, especially in developing countries. |
Repair of Proximal Hamstring Tear Utilizing a Suture Bridge Knotless Construct
Proximal hamstring tendon injuries occur frequently in the athletic population resulting in varying degrees of functional disability depending on severity of injury. The purpose of our case vignette is to describe a surgical technique and clinical outcome for open proximal hamstring tendon repair with a confirmed biomechanically sound construct. We also describe and summarize the current literature recommendations for proximal hamstring injuries. We present a case and surgical technique report on a 27-year-old male who suffered a proximal hamstring tendon rupture. Utilizing a double row allknotless suture bridge construct with a total of four anchors and six suture limbs allowed for anatomic footprint coverage and strength. Two years of clinical follow-up was obtained evaluating hip and knee range of motion, strength, and functional ability. Our patient underwent uncomplicated open surgical repair and returned to all activity at four months following surgery. Range of motion and strength returned to preoperative levels at the four-month postoperative mark. The use of a reproducible double row all-knotless suture bridge technique provided adequate strength and stability in the setting of a proximal hamstring tendon rupture. Open and endoscopic surgical techniques performed acutely both show positive postoperative subjective outcomes as well as a high likelihood of returning to sport. Controversy remains present in regard to the repair technique as well as postoperative bracing and physical therapy recommendations.
# Introduction
The proximal hamstring complex is an important group of synergistic muscles that function to extend the hip and flex the knee. It is composed of three distinct muscles including the long head of the biceps femoris, the semitendinous muscle, and semimembranous muscle. The biceps femoris and the semitendinous muscle form a conjoint tendon that inserts medially on the ischial tuberosity, while the semimembranosus muscle inserts laterally.
During eccentric contraction, there is substantial load placed upon these muscles predisposing them to increased strain and tearing during high-intensity activities. This risk is significantly increased in sports that require running, jumping, and kicking (such as hurdling, soccer, and American football) as well as during the initial sit to stand phase of waterskiing.
Hamstring injuries represent nearly 30% of new lower extremity complaints, with the hamstring muscle complex (HMC) being the most frequently injured. The vast majority of injuries occur at the myotendinous junction. HMC injuries are typically treated conservatively with an initial period of protected weight bearing and rest, ice, compression, elevation (RICE) therapy, followed by strength and balancing exercises. Surgical intervention becomes necessary when the patient fails to improve with conservative management or meets specific surgical indications.
Patients commonly report a history of an acute, sharp pain felt just prior to the initiation of rapid acceleration or deceleration activities or the history of traumatic blow to the lower extremity while the patient's hip is fully flexed and the knee fully extended. Pain is most commonly centralized to the posterior thigh, with symptomatology over the ischial tuberosity being more suggestive of an avulsion injury, rather than a myotendinous injury. Oftentimes, an audible pop or the feeling of a snapping sensation can also be felt and, in severe cases, ecchymosis and a palpable mass may even be visualized on the posterior thigh due to proximal tendon avulsion and distal retraction.
Generally accepted indications for surgical repair include ischial tuberosity bony avulsion or tendon retraction with >2 cm displacement and more than one hamstring tendon involvement. Wood and colleaguesreviewed 72 hamstring repair cases and found that in cases of complete avulsion with hamstring retraction, a delay in surgical repair results in a more technically challenging repair, increasing the likelihood of sciatic nerve involvement, increasing the need for postoperative bracing, and reducing postoperative hamstring strength and endurance. Along with a delay in surgical repair, increased tendon retraction has been linked to worse functional outcomes. Piposar and associatesperformed a retrospective case series of 25 patients with hamstring injuries. They found that 40% of nonsurgically treated proximal hamstring avulsions of <2 cm eventually required surgical intervention. In patients with proximal hamstring injuries without significant retraction, <2 cm, an individualized treatment decision is recommended given the high propensity for need for surgical treatment in the future if the patient fails conservative therapy.
Repair of the proximal hamstring can be performed either open or endoscopically and performed with a variety of fixation constructs based on the surgeon's preference. Management of proximal hamstring avulsion injuries can be challenging from both an operative and a rehabilitation standpoint. The purpose of this case report and review is to describe our preferred approach for open surgical repair and to further reinforce the utility of a knotless double row suture bridge repair in the setting of a retracted, triple tendon, proximal hamstring avulsion.
## Case report
Our patient is a 27-year-old healthy active individual with no past medical problems (see Supplemental Video 1). He felt sharp pain located in the left buttocks area and posterior thigh during a soccer game. He had immediate pain and inability to ambulate without limping. On a physical exam, the patient had an antalgic stiff-legged gait that minimized both hip and knee motion. Marked tenderness with palpation was present along the proximal hamstring muscle belly complex and ischial tuberosity. Left hip active range of motion was limited secondary to pain. Motor strength testing revealed significant weakness in hip extension and knee flexion compared to the contralateral side. The heel drag test elicited a positive result, with pain and weakness compared to the uninjured side.
Magnetic resonance imaging (MRI) of the injured extremity demonstrated a left proximal complete tear of the left semitendinosus and biceps femoris tendons from the origin with retraction (see . There was also associated severe partial tear of the semimembranosus tendon. Plain radiographs were negative for bony fracture or other abnormality.
The patient was employed as a manual laborer and was unable to return to work following his injury. Due to his continued discomfort, inability to perform his job, and multiple tendon involvement, surgical intervention was pursued 3 weeks following the injury.
## Surgical technique
Following verbal and written consent per protocol, the patient was brought into the operative arena. Intubation and general anesthesia was initiated while the patient was supine. The patient was then placed in the prone position. All bony and soft tissue prominences were well padded. Standard sterile preparation with chloraprep and draping was then performed. Surgical timeout was completed, and a longitudinal incision was made due to the chronicity of the tear and possible need for extensive dissection (see . The superficial cutaneous nerves were protected throughout the course of the case. Once the fascia was incised, the gluteus medius was reflected superiorly. The hematoma from the hamstring rupture, approximately 6 cm distal to the ischial tuberosity, was identified and was consistent with the preoperative MRI findings. Extensive amount of scarring along the course of the sciatic nerve was noted. Neurolysis of the sciatic nerve was performed. Special care was taken to protect the sciatic nerve throughout the course of the case.
After the wound was copiously irrigated, a modified suture bridge repair was carried out. Initially, sutures were passed 1.5-2 cm into the stump of the retracted ruptured tendon to aid in visualization and lysis of tendon adhesion to surrounding tissue. The knee was bent to approximately 50 degrees in order to relieve tension of the repair. The ischial tuberosity was identified and Normal attachment Gap : Coronal MRI displaying a retracted complete tear is of the left semitendinosus and biceps femoris tendons from the origin. Associated severe partial tear of the semimembranosus tendon with associated hematoma within the posterior deep musculature measuring up to 16.5 cm.Case Reports in Orthopedics retractors were placed to protect the sciatic nerve and expose the ischial tuberosity. The footprint for the insertion of the hamstring tendons was identified and debrided in order to provide a healthy bleeding bed for improved biological healing. Two 2.6 mm FiberTak Soft Anchor (Arthrex, Naples, FL), triple loaded with three 1.3 mm SutureTapes were placed at the distal aspect of the ischial footprint (see . The anchors were placed with 8 mm of separation between each other. Once the anchors are secured, a free needle was utilized to pass the sutures sequentially through the tendon with full-thickness passes. Each free suture limb was passed from distal to proximal through the tendon with at least a 2 mm gap between each suture pass. The suture limbs were brought over proximally in order to create a suture bridge construct. The proximal row anchors were placed at the proximal aspect of the ischial footprint, one medial and one lateral, using 4.75 mm SwiveLock anchors (Arthrex, Naples, FL). A total of six suture limbs were placed into each eyelet of the Swi-veLock anchor to create a knotless construct (see. Appropriate tension was applied to the suture limbs to provide solid fixation of the tendon to its anatomic position. Probing demonstrated robust compression of the tendon back to its anatomic footprint.
The incision was then copiously irrigated. The knee was easily extended to 50 degrees without undue tension on the repair. At approximately 45 degrees of knee flexion, an increased amount of tension was noted at the repair site. Layered closure was completed in the usual fashion. The patient was placed in a locked hinged knee brace locked in 90 degrees.
Postoperatively, the patient was made non-weight bearing for a total of six weeks. He was instructed to use the hinged knee brace for the first six weeks with strict precautions on the locked position of 90 degrees for the first two weeks. After the initial two-week recovery, from postoperative weeks 2 to 4, the brace was set to allow 90 to 50 degrees of knee flexion. At four weeks following surgery, the patient gradually increased the amount of knee extension allowed by 10 to 15 degrees per week. No strengthening was allowed until eight weeks postoperatively. Concentric strengthening was permitted after 8 weeks, and eccentric strengthening was initiated 3 months postoperatively.
Two years of clinical follow-up was obtained evaluating hip and knee range of motion, strength, and functional ability. The patient demonstrated similar strength and range of motion compared to the contralateral nonoperative extremity. The patient returned to work 4 months following surgery without functional limitation.
# Discussion
Increasing activity with a focus on cross-training exercises in the general population along with the chronic highperformance demands on athletes at all levels results in injuries in the HMC being more and more common. Although ruptures and strains can occur in all three hamstring muscles, the most commonly injured is the biceps femoris, with injuries more likely to occur proximally rather than distally. Despite the most meticulous rehabilitation, many cases are recalcitrant to nonsurgical intervention, and others with more severe presentations such as avulsions or retraction require early surgical management. These factors lead to increased athlete morbidity.
Ensuring strength and stability in a repair construct is of the utmost importance to allow for adequate time for tendon healing. A recent study confirmed hamstring healing capability following proximal repair. In the study, a postoperative MRI, completed at a mean of 36+/-11.4 months following repair, demonstrated in all 12 cases, the HMC reattached to the ischial tuberosity. The hamstring repair was completed with an average of 2.5 anchors and sutures passed and tied in a Mason-Allen-type fashion.
## Case reports in orthopedics
Multiple surgical techniques, both open and endoscopic, have been reported in the literature for proximal hamstring ruptures and ischial tuberosity bony avulsions. Suture anchor fixation has provided increased variability and versatility in proximal hamstring repair with less soft tissue disruption compared to transosseous fixation methods. In a recent case report, Tetsunaga et al.describe successful surgical treatment of an avulsion fracture of the ischial tuberosity by suture anchor fixation using a suture bridge technique. Moatshe and colleaguesutilized their group's biomechanical cadaveric study to perform a five-suture anchor technique which was shown to be significantly stronger than repairs with two large or small anchors.
Minimally invasive endoscopic repair was described by Mehta et al.as another alternative method to suture anchor fixation. Different patterns of suture anchor configuration have been employed, including five anchors in an "X" pattern as well as six anchors positioned like the face of a die. Additionally, luggage tag sutures at the medial leading edge of the tendon can be utilized to improve bony contact.
Gerhardt and colleaguescompleted a biomechanical study comparing two common surgical repair techniques. They evaluated the ultimate failure load and failure mechanism of an intact hamstring tendon compared to knotless and knotted anchor configurations for hamstring repair. The all-knotless suture anchor constructs failed at the highest maximal load of the 3 groups (767.18 ± 93.50 N), including the intact tendon group (750.58 ± 172.22 N). There was a statistically significant difference in load to failure when the allknotless construct was compared with the all-knotted technique (549.56 ± 20.74 N). This demonstrated superiority of the knotless suture construct. The cadaveric study used a four suture-anchor configuration double row construct that was similar to the technique utilized in our case report. The proposed mechanism for this success is that the knotless design allows for reproducible equalization of pressure without relying on the variability of knot tying.
A double row technique allows for widely disrupted footprint pressure in efforts to improve contact with the bone tendon interface. Biomechanical studies evaluating double row constructs for rotator cuff repair have demonstrated less gap formation and improved tensile strength compared to single row repair. Although, clinical studies have shown differing results on rerupture rates and outcomes following single or double row rotator cuff repair. Double row repair provides increased tendon-to-bone interface and ultimately increased healing potential. The double row construct allows the performing surgeon to set appropriate compression and overall hamstring tension.
Outcomes for operative repair of hamstring avulsions have shown high patient satisfaction; however, decreased strength, residual pain, and decreased activity level have been reported. A cohort study performed by Shambaugh et al.showed that patients can expect to regain approximately 90% strength of the uninjured extremity with acute hamstring repair using 2-4 suture anchors. Bowman et al.found an overall satisfaction level of 94% in 58 patients treated with open or endoscopic suture anchor fixation, but runners were less satisfied. Overall, the majority of athletes (88%) returned to sports at 7.6 months, on average, with 72% returning at the same level. In comparison, the runners in the study returned to the same level only 50% of the time. Continued future research into outcomes is warranted.
Complications associated with operative HMC repair were reported as 23.17% by a recent meta-analysis, 7.99% being neurologic. Acute repairs were found to have better outcomes compared to chronic repairs. Studies have shown more difficult exposure as well as more sciatic nerve neurolyses were performed in surgically treated chronic tears. The intimate relationship of the sciatic nerve and hamstring attachment at the ischial tuberosity makes it vulnerable to injury both at the time of hamstring avulsion and at time of operative repair. Careful dissection, identifying, and protecting the sciatic nerve as well as acute repair help reduce the risk of iatrogenic injury. In chronic cases, where distal tendon retraction is of concern, a longitudinal incision may be used to provide adequate exposure to allow for tendon mobilization and proper protection of the neurovascular bundle.
Cosmetic concerns regarding incisional scar should be discussed with the patient during the informed consent process.
Marked variability exists for the postoperative physical therapy following hamstring repair owing to the various surgical techniques utilized. In a cross-sectional study of 50 rehabilitation protocols, 34% prescribed knee bracing, 23% prescribed hip bracing, and 14% did not specify the type of brace recommended. As for weight bearing, 40% of protocols advised non-weight bearing and 46% allowed toe-touch weight bearing. Advancement to full weight bearing was allowed at a mean of 7.1 weeks. Gerhardt eliminated postoperative bracing and allowed his patients, following knotless suture anchor construct fixation, to eliminate weight bearing assistance at two weeks. Toe-touch weight bearing with a knee brace flexed to 40 degrees for 4 weeks is recommended by Arner et al. to decrease the tension on the repair and is less cumbersome than a hip brace. Arner's protocol included brace discontinuation at 4 weeks postoperatively with initiation of pain free arch of motion against gravity. Progressive return to sport protocol is recommended over a 6-month period. In this case vignette, postoperative care after a suture bridge technique included hinged knee brace for the first 6 weeks to control progressively increased range of motion.
# Conclusion
The ultimate goals of proximal hamstring repair are to provide stable maximum strength fixation, optimizing bone-to-tendon healing potential, and to allow for early motion and physical therapy. Accelerated rehabilitation protocols may be initiated given the fixation strength of the all-knotless repair double row suture bridge repair technique. Our case demonstrates clinical applicability of a reproducible biomechanically strong construct that allows for early motion progression.
## Data availability
All data available within case report manuscript (case report, single patient).Case Reports in Orthopedics
## Conflicts of interest
The authors, their immediate families, and any research foundations with which they are affiliated did not receive any financial payments or other benefits from any commercial entity related to the subject of this article. |
In vivo identification of essential nucleotides in tRNALeu to its functions by using a constructed yeast tRNALeu knockout strain
# Introduction
Aminoacyl-tRNA synthetases (aaRSs) are a family of key enzymes present in three kingdoms of life [bib_ref] Aminoacyl-tRNA synthesis, Ibba [/bib_ref]. They catalyze aminoacylation of their cognate tRNAs with amino acids to form correct aminoacyl-tRNAs, which are the substrates of protein biosynthesis in the ribosome. Generally, aminoacylation of tRNA is performed by a two-step reaction-activation of the amino acid with ATP and its subsequent transfer to the 3 0 -end of the tRNA [bib_ref] Aminoacyl-tRNA synthesis and translational quality control, Ling [/bib_ref]. Based on structural features of the amino acid activation domain, aaRSs can be divided into two classes-I and II [bib_ref] Partition of tRNA synthetases into two classes based on mutually exclusive sets..., Eriani [/bib_ref]. The fidelity of the aminoacylation reaction for certain aaRSs is threatened by a limited number of non-cognate standard amino acids. However, some aaRSs have evolved proofreading mechanisms to hydrolyze misactivated amino acids or misacylated tRNAs, known as pre-transfer editing or post-transfer editing, respectively [bib_ref] tRNA-dependent pre-transfer editing by prokaryotic leucyl-tRNA synthetase, Tan [/bib_ref].
Leucyl-tRNA synthetase (LeuRS) belongs to class Ia aaRS, which is characterized by a Rossmann fold with two signature peptides, HIGH and KMSKS. Additionally, almost all LeuRSs contain a large insertion domain called connective peptide 1 (CP1) within the catalytic Rossmann fold domain. CP1 is defined as a classic editing domain, which can hydrolyze the aminoacyl bond of mischarged aa-tRNA to ensure the fidelity of the catalytic process [bib_ref] The 2 A crystal structure of leucyl-tRNA synthetase and its complex with..., Cusack [/bib_ref].
The cognate substrate of LeuRS, tRNA Leu , has multiple unique characteristics. First, it belongs to the class II tRNAs, along with tRNA Ser and tRNA Tyr , characterized by a long variable stem and loop while other tRNAs only harbor a short variable arm [bib_ref] Compilation of tRNA sequences and sequences of tRNA genes, Steinberg [/bib_ref]. Second, in the normal genetic code, leucine is specified by six different codons and at least four tRNA Leu isoacceptors exist in vivo. Therefore, how LeuRS recognizes the isoacceptors exhibiting diverse sequences and structural heterogeneity remains an interesting question. Third, LeuRS contains both an amino acid activation domain and an editing domain. The 3 0 -end of tRNA Leu should shuttle between the two active sites during aminoacylation and editing processes. Our recent studies have shown that translocation of the tRNA Leu 3 0 -terminus is crucial for the enzymatic activity [bib_ref] Role of tRNA amino acid-accepting end in aminoacylation and its quality control, Zhou [/bib_ref].
Our laboratory has investigated the interaction between LeuRS and tRNA Leu from various species, including Escherichia coli, Aquifex aeolicus, Giardia lamblia and human (cytosolic and mitochondrial) [bib_ref] Role of tRNA amino acid-accepting end in aminoacylation and its quality control, Zhou [/bib_ref] [bib_ref] Tertiary structure base pairs between D-and TpsiC-loops of Escherichia coli tRNA(Leu) play..., Du [/bib_ref] [bib_ref] Recognition of tRNA Leu by Aquifex aeolicus leucyl-tRNA synthetase during the aminoacylation..., Yao [/bib_ref] [bib_ref] A T-stem slip in human mitochondrial tRNA Leu (CUN) governs its charging..., Hao [/bib_ref]. The above-mentioned tRNA Leu could be purified after overexpression in E. coli or transcribed by T7 RNA polymerase in vitro. Several extensive in vitro studies have been carried out to clarify the recognition elements of tRNA Leu from different species during the LeuRS-tRNA Leu interaction. A73, which is absolutely conserved in all tRNA Leu , was found to be the discriminator base. In E. coli, the amino acid-accepting end (CCA 76 ) has been shown to be essential for both aminoacylation and editing [bib_ref] Role of tRNA amino acid-accepting end in aminoacylation and its quality control, Zhou [/bib_ref]. Similarly, the tRNA folding resulting from the tertiary interactions between the D-loop and the TÉC-loop is crucial for recognition by LeuRS during both steps [bib_ref] Tertiary structure base pairs between D-and TpsiC-loops of Escherichia coli tRNA(Leu) play..., Du [/bib_ref]. In the bacterium A. aeolicus, interactions between LeuRS and tRNA Leu anticodon arm are essential for translocation of the tRNA from the synthetic to the editing site [bib_ref] Recognition of tRNA Leu by Aquifex aeolicus leucyl-tRNA synthetase during the aminoacylation..., Yao [/bib_ref] [bib_ref] A present-day aminoacyl-tRNA synthetase with ancestral editing properties, Zhu [/bib_ref]. However, nucleotides of tRNA Leu that specifically control the editing reaction are unknown. In tRNA Ile , the specific nucleotides from the D-loop are sufficient to trigger the editing response [bib_ref] Discrete determinants in transfer RNA for editing and aminoacylation, Hale [/bib_ref]. In tRNA Ala , both aminoacylation and editing have been shown to be sensitive to the G3U70 base pair in the acceptor stem [bib_ref] The C-Ala domain brings together editing and aminoacylation functions on one tRNA, Guo [/bib_ref].
Despite considerable efforts, transcripts of tRNA Leu from Saccharomyces cerevisiae (yeast tRNA Leu ) could not be synthesized in vitro with a reasonable accepting activity [bib_ref] Structural and mechanistic basis of pre-and posttransfer editing by leucyl-tRNA synthetase, Lincecum [/bib_ref]. In vitro transcription by T7 RNA polymerase is a powerful approach to study recognition elements of tRNAs; but it has some limitations. First, the point-bypoint site-directed mutagenesis method is time-consuming if an exhaustive investigation is carried out. Second, in vitro transcribed tRNAs may be inactive due to the absence of critical base modifications [bib_ref] A single methyl group prevents the mischarging of a tRNA, Pu¨tz [/bib_ref]. An alternative approach consists in using in vivo assays, which may overcome some of these limitations and approach the physiological conditions. Thus, in some conditions, it may become a useful substitute for the T7 RNA transcription approach. In vivo screening assays are often based on tRNA knockout strains. In E. coli, new approaches such as one-step inactivation of chromosomal genes and specific plasmids have been developed to delete a tRNA gene in order to investigate its structure and function in vivo [bib_ref] One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products, Datsenko [/bib_ref] [bib_ref] Plasmid systems to study RNA function in Escherichia coli, Gabriel [/bib_ref]. In eukaryotes, a yeast tRNA knockout strain was used to discriminate between active and inactive tRNA Arg molecules [bib_ref] A yeast knockout strain to discriminate between active and inactive tRNA molecules, Geslain [/bib_ref]. In mammals, a tRNA Sec knockout mouse was constructed, with the aim of studying the relationship between organic aging and selenocysteine metabolism [bib_ref] Early embryonic lethality caused by targeted disruption of the mouse selenocysteine tRNA..., Bosl [/bib_ref].
Here, we constructed two knockout yeast strains for tRNA Leu genes. In the first strain, the unique gene of tRNA Leu (GAG) was deleted, and the three genes of tRNA Leu (UAG) were deleted in the second. These strains were used to study the two tRNAs in vivo. The results showed that tRNA Leu (UAG), but not tRNA Leu (GAG), is essential for yeast growth. A genetic selection was performed to isolate lethal mutants of tRNA Leu (UAG) unable to complement the yeast knockout strain. In addition, selection on plates containing the toxic analog norvaline (Nva) led to the identification of tRNA mutants unable to be edited by LeuRS [bib_ref] Modular pathways for editing non-cognate amino acids by human cytoplasmic leucyl-tRNA synthetase, Chen [/bib_ref] [bib_ref] Unique residues crucial for optimal editing in yeast cytoplasmic Leucyl-tRNA synthetase are..., Yao [/bib_ref]. In this way, seven critical nucleotides for aminoacylation of yeast tRNA Leu (UAG) and one for the editing function of yeast LeuRS were identified.
# Materials and methods
# Materials
Ethylenediaminetetraacetic (EDTA), PEG4000, lithium acetate, polyvinylpyrrolidone (PVP), bovine serum albumin (BSA), Ficoll 400, sodium citrate, potassium acetate, glucose, magnesium chloride, Tris-HCl and maleic acid were purchased from Sango (Shanghai, China). The yeast minimal synthetic defined (SD) base and dropout (DO) supplement base were purchased from Clontech (Mountain View, CA, USA). 5-fluoroorotic acid (5-FOA), restriction endonucleases and T4 DNA ligase were obtained from Fermentas (Burlington, Canada). Amino acids were purchased from Sigma (St Louis, MO, USA). The KOD plus DNA polymerase and dNTPs were purchased from TOYOBO (Osaka, Japan). Digoxigenin blocking reagent and anti-digoxigenin-AP were purchased from Roche (Basel, Switzerland). All the radioactive amino-acids and ATP are purchased from PerkinElmer (Waltham, MA, USA). T7 RNA polymerase was prepared in our laboratory as previously described. [bib_ref] T7 RNA polymerase transcription of, Li [/bib_ref]. Streptavidin Agarose Resin was purchased from Throme-Pierce (Rockford, IL, USA). Ni 2+ -NTA superflow column was purchased from Qiagen (Dusseldorf, Germany).
Plasmids pALR5 (tR4, URA, ADE3) (24), pRS313 (HIS3, CEN6, ARSH4), pRS314 (TRP1, CEN6, ARSH4), pRS315 (LEU2, CEN6, ARSH4) and pGPD415 (LEU2, CEN6, ARSH4, GPD promoter) [bib_ref] Yeast vectors for the controlled expression of heterologous proteins in different genetic..., Mumberg [/bib_ref] [bib_ref] A system of shuttle vectors and yeast host strains designed for efficient..., Sikorski [/bib_ref] were from the Institut de Biologie Mole´culaire et Cellulaire du CNRS (Strasbourg, France); pET28-ycleuS (KAN, ycleuS) and pET28-ycleuS-D419A (KAN, ycleuS-D419A) were constructed previously and stored in our laboratory [bib_ref] Unique residues crucial for optimal editing in yeast cytoplasmic Leucyl-tRNA synthetase are..., Yao [/bib_ref] ; E. coli strain BL21(DE3) was purchased from Stratagene (La Jolla, CA, USA). pRS317 (LYS2, CEN6, ARSH4) and pRS412 (ADE2, CEN6, ARSH4), S. cerevisiae strains YPH499 (Mata ura3-52 lys2-801 Amber ade2-101 Ochre trp1-D63 his3-D200 leu2-D1) and YPH501 (ura3-52 lys2-801 Amber ade2-101 Ochre trp1-D63 his3-D200 leu2-D1) [bib_ref] A system of shuttle vectors and yeast host strains designed for efficient..., Sikorski [/bib_ref] [bib_ref] Designer deletion strains derived from Saccharomyces cerevisiae S288C: a useful set of..., Brachmann [/bib_ref] were gifts from JinQiu Zhou in our institute. Digoxigenin-labeled RNA probes were synthesized by Takara (Otsu, Shiga, Japan).
## Gene cloning and site-directed mutagenesis
The yeast tRNA Leu (UAG) gene and its upstream and downstream DNA fragments were amplified from genome extracted from wild-type (WT) yeast strain YPH501 by PCR using forward primer 5 0 -AAAGAGCT CGACAAACGACTG-3 0 (SacI site in italics) and reverse primer 5 0 -AAAGGATCCGTTGCGCCAAGA-3 0 (BamHI site in italics), which introduced restriction sites for cloning. After digestion with SacI and BamHI, the PCR product was inserted into the same sites of pALR5 to give the rescuing plasmids pALR-T [tR4, URA, ADE3, TL(UAG)]. Similarly, the digested PCR product was inserted into pRS314 to give pRS314-R [TRP1, CEN6, ARSH4, TL(UAG)] which was used as a template for the random mutagenesis. The single-point mutants of tRNA Leu (UAG) were constructed by PCR according to the protocol provided by the KOD-Plus-Mutagenesis Kit (TOYOBO, Japan) (28), using mutation-containing primers and pRS314-R as template.
An additional plasmid expressing a chimeric tRNA(GAG) was constructed to rescue the triple tRNA Leu (UAG) knockout strain. The gene TL(GAG) was amplified by PCR as a 700-bp DNA fragment containing the natural promoter. It was subsequently cloned in pRS317 and the anticodon GAG was mutated to UAG. The resulting plasmid (pRS317-C) was expressing a chimeric tRNA exhibiting the main body of tRNA Leu (GAG) and a UAG anticodon. The plasmid could efficiently complement the triple tRNA Leu (UAG) knockout strain. It was particularly useful during the northern blot analyses of tRNA Leu (UAG) mutants since it did not hybridize with the probe designed to recognize tRNA Leu (UAG).
In situ disruption of the genes encoding tRNA Leu (GAG) and tRNA Leu (UAG)
Disruption of the gene for tRNA Leu (GAG) was performed as follow. First, a DNA fragment carrying the 1.5-kb DNA fragment of HIS3 gene ligated to the flanking sequences of tRNA Leu (GAG) gene was constructed. For this, two PCR amplifications were performed by LA-Taq enzyme on yeast genomic DNA. A 1-kb DNA fragment located upstream of the tRNA gene was amplified by forward primer 5 0 -AAAGGATCCTAGT GCTGATGGTATCCC-3 0 (BamHI in italics) and reverse primer 5 0 -AAAGCGGCCGCAATTCATCTGTTGGGTA A-3 0 (NotI in italics). A second 1-kb DNA fragment located downstream was amplified by forward primer 5 0 -AAATCTAGAGAATCTACGCACTTCA AT-3 0 (XbaI in italics) and reverse primer 5 0 -AAAGGGCCCGAATC TTGGAACAGGTTG-3 0 (ApaI in italics). The amplified upstream and downstream fragments were cleaved by BamHI+NotI and by XbaI+ApaI, respectively. A 1.5kb marker gene HIS3 was amplified by PCR from pRS313 by using forward primer 5 0 -AAAGCGGCCGCTCGCGC GTTTCGGTGATG-3 0 (NotI in italic) and reverse primer 5 0 -AAATCTAGAGATTTCGGCCTATTGGTT-3 0 (XbaI in italic). The HIS3 gene was cleaved by NotI and XbaI. The three fragments above were ligated between the BamHI and ApaI of pcDNA3 to generate a recombinant plasmid (PCDNA-R) in which the recombinant DNA fragment insert [designated UHD or tl(gag)::HIS3] consisted of the HIS3 marker gene in between the upstream and downstream DNA fragments of the yeast tRNA Leu (GAG) gene (Supplementary [fig_ref] Figure 1: Strategy of construction and characterization of the tRNA Leu [/fig_ref]. The recombined fragment was excised by BamHI and ApaI from pcDNA3-R, and the linear DNA fragment was introduced into the yeast diploid strain YPH501. Yeast transformants were grown on SD-His À plates at 30 C for 2-3 days. It was expected that the selected His + colonies were exhibiting a tRNA Leu (GAG) gene replaced by a HIS3 marker gene after homologous recombination with the UHD DNA fragment. To verify this hypothesis, sporulation of His + colonies was induced on appropriate plates (1% potassium acetate, 0.1% yeast extract and 0.05% glucose). Tetrads were dissected and the resulting haploids were plated on YPDA plates (1% yeast extract, 2% peptone, 2% glucose, 0.004% adenine and appropriate agar) and further analyzed (described earlier). The yeast strain deleted for tRNA Leu (GAG) gene was named tl(gag)-D1.
We used the same method to delete the three genes of tRNA Leu (UAG), and the yeast strain was named tl(uag)-D1-3.
Construction of the random mutation library and screening by plasmid shuffling Random mutagenesis of tRNA Leu (UAG) was performed by hydroxylamine treatment. Plasmid pRS314-R (Trp + ) was incubated in an appropriate volume of mutagenesis buffer (0.8 M hydroxylamine-HCl, 50 mM sodium phosphate, pH 6, 1 mM EDTA) at 37 C for 48 h to statistically obtain only one mutation per tRNA gene [bib_ref] Dominant negative mutations in the Tn10 tet repressor: evidence for use of..., Isackson [/bib_ref]. Then, the DNA solution was extracted by phenol/chloroform, and precipitated with ethanol. The library was used to transformed tl(uag)-D1-3 strain in which the three genes for tRNA Leu (UAG) were deleted. Transformants were grown at 30 C on plates of SD medium minus tryptophan (SD À Trp À ) for 2-3 days. Then, transformants were replicated to SD À Trp À plates containing 5-FOA and uracil to allow shuffling of the pALR-T maintenance plasmid [containing URA3 gene whose product converts 5-FOA to toxic 5-fluorouridine monophosphate (5-FUMP)]. The lethal clones were selected. The DNA was extracted for sequencing of the tRNA Leu (UAG) mutated gene in pRS314-R.
Complementation assays by site-directed mutated tRNA Leu Plasmids carrying the site-directed mutated tRNA Leu (UAG) genes were introduced separately into the tRNA knockout strain and transformants were grown at 30 C and treated as above with the randomly mutated library. Transformants were subjected to droptest on 5-FOA containing plates. The growth of the yeast cells was then observed and calculated [bib_ref] A yeast knockout strain to discriminate between active and inactive tRNA molecules, Geslain [/bib_ref] [bib_ref] Unique residues crucial for optimal editing in yeast cytoplasmic Leucyl-tRNA synthetase are..., Yao [/bib_ref].
## Measurement of the growth rate
Yeast cells containing the different tRNA mutations were grown in liquid SD À Trp À medium. Each culture was diluted in liquid media to an initial A 600 = 0.1. The diluted culture was shaken at 30 C. The A 600 value was measured at various intervals, and the growth rate was calculated according to the logistic growth equation [bib_ref] Inhibited cell growth and protein functional changes from an editing-defective tRNA synthetase, Bacher [/bib_ref].
## Analysis of in vivo aminoacylation by northern blot
The yeast gene knockout strains containing different mutations were grown in liquid SD À Trp À media to A 600 &0.6, and then were harvested by centrifugation. Cell pellets were suspended in 1 ml of extraction buffer (0.3 M sodium acetate, pH 4.5, 10 mM EDTA) and extracted with one volume of phenol extraction (pH 4.5). The supernatant was then re-extracted with one volume of phenol/chloroform (pH 4.5). Three volumes of 100% ethanol were mixed with the supernatant to precipitate the RNA, and the final precipitate was dissolved in 50 ml of storage buffer (10 mM sodium acetate pH 4.5).
tRNAs were separated by acid 12% PAGE containing 8 M urea. Thirty micrograms of tRNA samples were loaded on a 0.4-mm thick gel in 0.1 M sodium acetate buffer pH 5. Electrophoresis was carried out at 4 C and constant 12 W (&400 V) for 24 h in 0.1 M pH 4.5 sodium acetate buffer. The portion of the gel containing the tRNAs was transferred to a nylon membrane (Millipore, Bedford, MA, USA) by electrophoresis at constant 100 V for 30 min in 0.5Â TBE buffer. The nylon membrane was then baked for 30 min at 80 C [bib_ref] A yeast knockout strain to discriminate between active and inactive tRNA molecules, Geslain [/bib_ref] [bib_ref] The many applications of acid urea polyacrylamide gel electrophoresis to studies of..., Ko¨hrer [/bib_ref].
Pre-hybridization was performed at 65 C for 6 h in 15 ml of solution (10Â Denhardt's solution, 6Â SSC and 0.5% SDS). Hybridization was performed at 65 C for 6 h in 15 ml of the same solution, in the presence of a tRNA Leu (UAG) probe (5 0 -AGGGA TTCGA ACCCT T GCAT CCGAA GATAT-3 0 for the mutation of the anticodon arm or 5 0 -CCUUG CAUCC GAAGA UAUCA G AGCC UAAAU-3 0 for the other region), or a 5S RNA probe (5 0 -ACCCA CTACA CTACT CGGTC AGGCT C TTAC-3 0 as an internal control) labeled with digoxigenin at each 5 0 -end. Membranes were washed at 65 C in a 2% SSPE, 0.5% SDS solution for 20 min, hybridized with anti-digoxigenin-AP and visualized with CDP-star, following standard procedures [bib_ref] A sensitive non-radioactive northern blot method to detect small RNAs, Kim [/bib_ref]. Signals were quantified using a Fuji Bioimager Bas2000.
## Preparation of proteins and trna leu s
The ycLeuRS and its mutant ycLeuRS-D419A were purified as described [bib_ref] Unique residues crucial for optimal editing in yeast cytoplasmic Leucyl-tRNA synthetase are..., Yao [/bib_ref]. Escherichia coli strain BL21(DE3) was transformed with plasmids pET28-ycleuS and pET28-ycleuS-D419A which were constructed in our laboratory [bib_ref] Unique residues crucial for optimal editing in yeast cytoplasmic Leucyl-tRNA synthetase are..., Yao [/bib_ref]. Transformants were grown in 500 ml Luria-Bertani liquid medium (1% peptone, 0.5% yeast extract and 1% NaCl) containing 100 mg/ml kanamycin and grown at 37 C until the A 260 reached 0.6. IPTG was added to a final concentration of 0.2 mM, and the incubation continued at 22 C for another 10 h. The cells were harvested by centrifugation at 4000g for 10 min. The cell pellets were then sonicated and centrifuged at 34 000g for 1 h. The resulting supernatants were applied to Ni-NTA Superflow columns and incubated for 1 h. The columns were washed with 20 ml wash buffer (50 mM sodium phosphate pH 8.0, 300 mM NaCl and 20 mM imidazole) and eluted with 5 ml elution buffer (50 mM sodium phosphate pH 8.0, 300 mM NaCl and 250 mM imidazole). ycLeuRS and ycLeuRS-D419A were then pooled and concentrated by dialysis against 10 mM potassium phosphate buffer (pH 7.5) containing 50% glycerol.
The gene-encoding yeast elongation factor 1a was PCR-amplified from yeast genomic DNA and cloned into plasmid pGPD425 after the 6 His-encoding sequence, giving an N-terminal 6 His-tagged protein [bib_ref] Evidence that eukaryotic translation elongation factor 1A (eEF1A) binds the Gcn2 protein..., Visweswaraiah [/bib_ref]. The recombinant plasmid was transformed into yeast strain YPH501. Transformed yeast cells were grown at 30 C for 24 h. His-tagged EF-1a protein was purified by Ni-NTA affinity chromatography as described above.
Yeast tRNA Leu (UAG) WT and its mutants were prepared as described [bib_ref] The anticodon loop is a major identity determinant of Saccharomyces cerevisiae tRNA(Leu), Soma [/bib_ref]. The genes encoding mutated tRNA Leu s were transformed into our constructed yeast strain tl(uag)-D1-3, respectively. The transformants were cultured in 500 ml liquid YPDA medium and grown at 30 C for 24 h. Total tRNA was extracted by phenol from the harvested yeast cells as described above. Then 0.1 mg 5 0 -biotin labeled DNA probe with the sequence 5 0 -AGGGA TTCGA ACCCT TGCAT CCGA A GATAT-3 0 , a sequence complementary to the regions from the anticodon arm to the TÉC-loop of S. cerevisiae tRNA Leu , bound to 300 ml Streptavidin Agarose Resin, and 2.5 mg of yeast total RNA were mixed and incubated at room temperature for 1 h. The mixture was washed with 1 ml of NTE buffer (5 mM Tris-HCl pH 8.0, 2.5 mM EDTA and 200 mM NaCl) at room temperature, and tRNA was eluted with 1 ml 0.1Â NTE buffer at 65 C. The tRNA was further purified by electrophoresis on a 10% polyacrylamide/8 M urea gel, followed by passive elution in 10 mM Tris-HCl (pH 7.5) containing 0.3 M NaCl and 0.5 mM EDTA, and precipitation by ethanol after phenol-chloroform extraction. Eluted tRNA Leu (UAG)s were renatured by heating at 80 C for 5 min and slow cooling to room temperature (10).
## Aminoacylation, deacylation and amp formation
The aminoacylation assay of WT yctRNA Leu (UAG) and its mutants were determined at 30 C in a reaction mixture containing 60 mM Tris-HCl (pH 7.5), 10 mM MgCl 2, 2 mM dithiothreitol, 4 mM ATP, 20 mM [ 3 H]leucine, 2 mM tRNA Leu and 1 nM ycLeuRS. Kinetics constants for WT yctRNA Leu (UAG) and its mutants were determined in the reaction mixture mentioned above, in the presence of various concentrations (from 0.05 to 2 mM) of the relevant yctRNA Leu (UAG) substrates.
To determine the effect of EF-1a on leucine charging activity, aminoacylation was performed in the presence of a mix of 0.1 mg/ml EF-1a and 4 mM GTP (pre-incubated with pyruvate kinase for 30 min). In parallel, a negative control was performed in which an equal concentration of BSA and GTP was added.
Aminoacylated [ 3 H]Leu-tRNA Leu (UAG) WT and mutants were obtained by charging the corresponding tRNAs with Leu in the presence of WT ycLeuRS. The spontaneous hydrolysis levels of Leu-tRNA Leu (UAG)s were measured in 60 mM Tris-HCl (pH 7.5), 10 mM MgCl 2, 1 mM [ 3 H]Leu-tRNA Leu (UAG)s, with 0.1 mg/ml EF-1a and 4 mM GTP or with equal concentration of BSA and GTP at 30 C. Aliquots (10 ml) of reaction solution were quenched on Whatman filter pads at different times. Radioactivity was quantified in a scintillation counter (Beckman, Pasadena, CA, USA).
Misacylated Ile-tRNA Leu (UAG) WT and mutants were obtained by charging with Ile in the presence of ycLeuRS-D419A [bib_ref] Unique residues crucial for optimal editing in yeast cytoplasmic Leucyl-tRNA synthetase are..., Yao [/bib_ref]. The editing activities of ycLeuRS toward mischarged [ 3 H]Ile-tRNA Leu and mutants were measured at 30 C in 60 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 2 mM dithiothreitol and 1 mM [ 3 H]Ile-tRNA Leu and mutants. The reactions were initiated with 1 nM ycLeuRS.
Because the editing reaction consumes ATP, it can be measured using the breakdown of ATP (to AMP and PPi) in the presence of non-cognate amino acids [bib_ref] The 2 A crystal structure of leucyl-tRNA synthetase and its complex with..., Cusack [/bib_ref]. AMP formation was measured in reaction mixtures containing 60 mM Tris-HCl (pH7.5), 10 mM MgCl 2 , 5 mM DTT, 3 mM ATP, 20 nM [a-32 P]ATP, 15 mM Nva, 5 U/ml pyrophosphatase, in the presence or absence of 5 mM tRNA. The reactions were incubated at 30 C and initiated by 1 mM ycLeuRS. Aliquots (1.5 ml) were quenched in 6 ml of 200 mM sodium acetate (pH 5.0). Quenched aliquots (1.5 ml) were dripped into a TLC plate. Nva-[ 32 P]AMP, [ 32 P]AMP and [ 32 P]ATP were separated in 0.1 M ammonium acetate and 5% acetic acid. The plates were visualized by phosphor imaging and the data were further analyzed [bib_ref] Modular pathways for editing non-cognate amino acids by human cytoplasmic leucyl-tRNA synthetase, Chen [/bib_ref].
# Results
## Deletion of the gene for trna leu (gag) does not affect cell viability
The genome of S. cerevisiae contains 21 tRNA Leu genes distributed as 10 tRNA Leu (CAA) genes, seven tRNA Leu (UAA) genes, three tRNA Leu (UAG) genes and one tRNA Leu (GAG) gene [TL(GAG)]. Based on the assumption that the single-copy TL(GAG) for tRNA Leu (GAG) should be essential according to the decoding capacity of the anticodon GAG (35), we first deleted the TL(GAG) gene in the diploid YPH501 strain. The TL(GAG) gene is located between positions 700 679 and 700 760 in chromosome VII and is flanked by a tRNA Lys (UUU) gene and two genes YGR106C and YGRWDELTA19 [fig_ref] Figure 1: Strategy of construction and characterization of the tRNA Leu [/fig_ref] [bib_ref] tRNAscan-SE: a program for improved detection of transfer RNA genes in genomic..., Lowe [/bib_ref]. By the method described above, plasmid pcDNA3-R was constructed carrying HIS3 flanked by the upstream and downstream sequences of TL(GAG) (so called UHD fragment). The linear BamHI-ApaI UHD fragment was introduced into the yeast strain YPH501. In this way, the 0.1-kb TL(GAG) gene was substituted by the 1.5-kb HIS3 gene, keeping the recombination flanking sequences intact. Transformants were grown on SD/His À plates in order to select recombination events. To confirm that the diploid yeast transformants had lost TL(GAG) as expected, we examined the corresponding region of the chromosome by PCR amplification and sequencing (Invitrogen, Shanghai, China). Two DNA fragments were PCR-amplified from primers hybridizing to the flanking sequences on one side and to HIS3 on the other side [fig_ref] Figure 1: Strategy of construction and characterization of the tRNA Leu [/fig_ref]. PCR products of 2.1-kb and 1.6-kb sizes were amplified as expected from the strains carrying the UHD gene modification. Additionally, the presence of NotI and XbaI sites resulting from the insertion of HIS3 was verified by digestion, producing 1.05-kb and 1+0.6-kb fragments, respectively [fig_ref] Figure 1: Strategy of construction and characterization of the tRNA Leu [/fig_ref]. These data confirmed that the UHD fragment was inserted into the yeast strain. However, after dissecting the tetrads obtained after incubation of the diploid strain on sporulation medium, two of the spores were His À as WT haploids, while the other two haploids which lacked TL(GAG) were exhibiting a His + phenotype. We found that the haploids knockout strain could grow perfectly on YPDA plates [fig_ref] Figure 1: Strategy of construction and characterization of the tRNA Leu [/fig_ref]. These data suggested that tRNA Leu (GAG) may not be an essential tRNA in yeast and that its decoding capacity overlaps with another tRNA Leu . To verify that tRNA Leu (GAG) was not produced any more in the knockout strain, we performed a northern blot analysis on the haploid His + strain and indeed no such tRNA could be detected [fig_ref] Figure 1: Strategy of construction and characterization of the tRNA Leu [/fig_ref]. In addition, we compared the growth rate of the knockout strain [named tl(gag)-D1] and WT YPH499 strain in different conditions. At 23 C, 30 C and 37 C, on YPDA medium, similar growth rates were observed (Supplementary [fig_ref] Figure 2: Construction of the tRNA Leu [/fig_ref]. When grown on YPDA with excess NaCl or glycerol, no difference was observed between the strains (data not shown) [bib_ref] The intron of tRNA-TrpCCA is dispensable for growth and translation of Saccharomyces..., Mori [/bib_ref]. Altogether, the data showed that the gene TL(GAG) of tRNA Leu (GAG) may be a dispensable gene and thus it is not a suitable gene to construct a strain dedicated to genetic selection of tRNA mutants.
## Construction of triple knockout strain of the trna leu (uag) genes
The triple deletion of tRNA Leu (UAG) genes was performed in the haploid strain YPH499 [which was chosen due to the simple genetic manipulations with having fewer spores to dissect [bib_ref] A U1 snRNA:pre-mRNA base pairing interaction is required early in yeast spliceosome..., Se´raphin [/bib_ref] ]. tRNA Leu (UAG) is transcribed from three TL(UAG) copies, two of them located on chromosome XII and one on chromosome X. Considering that TL(UAG) might be essential, a rescuing plasmid pALR-T [tR4, URA, ADE3, TL(UAG)] was first introduced into YPH499. The deletion of the three copies of TL(UAG) was carried out successively as described above, adding sequentially the HIS3, LEU2 and ADE2 genes. PCRand sequencing analysis confirmed that the genes were deleted and replaced by the auxotrophic markers [fig_ref] Figure 2: Construction of the tRNA Leu [/fig_ref]. The deleted strains were named tl(uag)-D1, tl(uag)-D1,2 and tl(uag)-D1-3. A plasmid-shuffling assay was performed on plates containing 5-FOA to check if the rescuing plasmid pALR-T could be shuffled from these strains. On 5-FOA plates, strains tl(uag)-D1 and tl(uag)-D1,2 were perfectly growing as the WT strain indicating that the rescuing plasmid was not essential. After deletion of the third tRNA Leu gene [tl(uag)-D1-3], the strain was unable to grow on 5-FOA-containing medium, indicating that the rescuing plasmid could not be lost (and thus 5-FOA was metabolized in 5-FUMP toxic for cells) [fig_ref] Figure 2: Construction of the tRNA Leu [/fig_ref]. This strongly suggests that the triple knockout experiment was successful and that tRNA Leu (UAG) is an essential tRNA Leu which has a decoding capacity that cannot not be covered by another tRNA Leu . Northern blot analysis further confirmed that the tl(uag)-D1-3 strain was not containing any tRNA Leu (UAG) and that the triple knockout was successful [fig_ref] Figure 2: Construction of the tRNA Leu [/fig_ref]. Altogether, the data showed that the knockout of the three tRNA Leu (UAG) genes is lethal for the yeast cells. The triple mutant can be complemented by a plasmid copy of the rescuing tRNA and a shuffle assay can be set up to select and test mutated tRNA Leu (UAG).
## Site-directed mutagenesis to test the triple trna knockout strain
We mutated several well-defined tRNA Leu identity elements to validate the use of strain tl(uag)-D1-3 for in vivo studies. The discriminator base A73, a positive identity determinant for recognition by LeuRS, and the crucial G18-É55 base pair for maintaining the tertiary structure of tRNA Leu were chosen (9,10). These two elements of tRNA Leu are necessary for aminoacylation in most species. Three mutations, A73C, G18U and É55G were constructed and the mutated genes was introduced into strain tl(uag)-D1-3. Then, the three transformants were plated on 5-FOA-containing medium and no one could grow, confirming that the mutations induced lethality of the cells. Another mutant, É55C, which allowed formation of the tertiary base pair with G18 could grow normally at the presence of the 5-FOA [fig_ref] Figure 3: Scheme and diagram showing the mutations studied and the effect on cell... [/fig_ref] and Supplementary [fig_ref] Figure 3: Scheme and diagram showing the mutations studied and the effect on cell... [/fig_ref]. These results matched those of former in vitro data and strongly supported that strain tl(uag)-D1-3 as an efficient system for in vivo investigations on tRNA Leu identity determinants.
## Essential nucleotides revealed by screening of a randomly mutated library
To identify the essential nucleotides of tRNA Leu , a randomly mutated library of plasmid pRS314-R [Trp2, TL(UAG)] was generated by chemical modification by hydroxylamine that preferentially induces G!U or C!A transversions [bib_ref] Dominant negative mutations in the Tn10 tet repressor: evidence for use of..., Isackson [/bib_ref]. The library was introduced into strain tl(uag)-D1-3 rescued by pALR-T and the transformants were grown in SD À TRP À plates. Five thousand colonies were obtained. They were replicated on 5-FOA-containing medium in order to induce pALR-T shuffle. A total of 11 clones did not grow on 5-FOA-containing medium, meaning that the shuffle could not occur because pRS314-R contained a TL(UAG) gene unable to sustain cell growth or grew slowly. Plasmids from the 11 clones were extracted. The phenotypes were confirmed by a second transformation and plating on 5-FOA, then the plasmid DNA was sequenced.
Among the 11 mutations, 9 appeared within TL(UAG) whereas the 2 others were located in another part of the plasmid outside of the tRNA gene. Among the lethal mutations were G18U in the D-loop, G36U and G37U (two occurrences) in the anticodon loop, C61A and C62A (two occurrences) in the T-stem. Among the mutants that grew slowly, one was harboring a G2U mutation in the acceptor stem (70% of WT of the growth rate) and the second one a mutation G30U in the anticodon stem (inducing a 45% growth decrease) [fig_ref] Figure 3: Scheme and diagram showing the mutations studied and the effect on cell... [/fig_ref]. We grew yeast cells containing the mutants of tRNA Leu on both solid and liquid media to further confirm the phenotype. On solid medium, lethal or slow growth phenotype was observed directly (Supplementary [fig_ref] Figure 3: Scheme and diagram showing the mutations studied and the effect on cell... [/fig_ref]. In liquid medium, the A 600 value of the culture was measured at various intervals, and the growth rate was calculated [fig_ref] Figure 3: Scheme and diagram showing the mutations studied and the effect on cell... [/fig_ref].
Surprisingly, tRNA Leu (UAG) harbor a long variable arm carrying six G or C nucleotides that could have been mutated by hydroxylamine, however, no lethal mutation was selected in the variable arm. It is known that this tRNA subdomain is a crucial element for proper tRNA-enzyme recognition, which involves more structural constraints than sequence-specific interactions with the synthetase [bib_ref] Structural dynamics of the aminoacylation and proofreading functional cycle of bacterial leucyl-tRNA..., Palencia [/bib_ref]. Therefore, single-site mutations may not disrupt or affect the LeuRS:tRNA Leu complex network of constraints.
## Effects of the mutations on the steady state and aminoacylation levels of trna leu
Mutations in the tRNA primary sequences may impact several processes, such as transcription, stability, bases modification, aminoacylation, editing and interaction with the translation factors and ribosome [bib_ref] Enter transfer RNA, Hoagland [/bib_ref]. Some tRNAs are also involved in signal transduction as cellular factors [bib_ref] Elevated tRNA(iMet) synthesis can drive cell proliferation and oncogenic transformation, Marshall [/bib_ref]. To identify the mutations that affect the steady state and aminoacylation status of tRNA Leu , we performed northern blots analyses. By extracting the tRNAs in acidic conditions, we could preserve the aminoacyl-ester bond and analyze the levels of in vivo aminoacylation [bib_ref] The many applications of acid urea polyacrylamide gel electrophoresis to studies of..., Ko¨hrer [/bib_ref].
To rescue the cells harboring a lethal tRNA Leu mutants, we complemented the cells with a chimeric tRNA Leu consisting of the body of tRNA Leu (GAG) in which the codon was changed to (UAG). This tRNA was expressed from the rescuing plasmid pRS317-C. In that way, we could rescue the cells and analyze the mutated tRNA Leu (UAG) content by northern blot without hybridizing the rescuing plasmid. Total tRNAs extracts were prepared and the tRNA Leu (UAG) levels were analyzed [bib_ref] A yeast knockout strain to discriminate between active and inactive tRNA molecules, Geslain [/bib_ref].
Almost all of the 11 mutations from site-directed and random mutagenesis affected the aminoacylation levels of tRNA Leu (UAG). Mutant A73C was totally uncharged in vivo, which is consistent with in vitro results [bib_ref] Recognition of tRNA Leu by Aquifex aeolicus leucyl-tRNA synthetase during the aminoacylation..., Yao [/bib_ref] [bib_ref] The anticodon loop is a major identity determinant of Saccharomyces cerevisiae tRNA(Leu), Soma [/bib_ref]. In vitro, the long-distance interaction between G18 and É55 is important to maintain the tertiary structure and the aminoacylation activity. The present work shows that preventing the interaction with G18U or É55G mutations led to total uncharging of the tRNA by LeuRS [fig_ref] Figure 4: Northern blot analysis of tRNA Leu samples [/fig_ref]. Mutation É55C, which theoretically allowed formation of hydrogen bonds between bases 18 and 55, restored partial aminoacylation of tRNA [fig_ref] Figure 4: Northern blot analysis of tRNA Leu samples [/fig_ref] , suggesting that the structure, the specific nucleotides or their modifications play a role during the aminoacylation of tRNA.
The two mutations of G2U and G30U of yeast tRNA Leu (UAG) exhibiting a reduced growth rate in the triple tRNA knockout strain were only partially aminoacylated in vivo [fig_ref] Figure 4: Northern blot analysis of tRNA Leu samples [/fig_ref]. The G2U mutation significantly reduced the aminoacylation level ($40% compared of WT), suggesting that the modification of the geometry of the acceptor arm by changing the original G:U wobble base pair is responsible of the aminoacylation defect. In addition, we cannot exclude that G2 is a determinant of the leucine identity. On the other hand, G30U was $70% as aminoacylated as the WT tRNA Leu , suggesting that the mutation more likely impacted the translation process [fig_ref] Figure 4: Northern blot analysis of tRNA Leu samples [/fig_ref].
The lethal mutations C61A and C62A located in the T-loop led to drastically reduced tRNA Leu (UAG) levels [fig_ref] Figure 4: Northern blot analysis of tRNA Leu samples [/fig_ref] , suggesting that the transcription process was deficient or the stability of the transcripts reduced. Indeed, tRNAs are RNA polymerase III transcripts and the ICR (internal control region) box is located into the T-loop (41), thus we cannot exclude that mutations C61A and C62A disturb the transcription process.
Although the G36U mutation changed the anticodon of tRNA Leu (UAG) from leucine anticodon UAG to serine anticodon UAU (thus switching the decoding property of the tRNA), the mutant of tRNA Leu could still be charged with an amino acid by LeuRS to a significant level [fig_ref] Figure 3: Scheme and diagram showing the mutations studied and the effect on cell... [/fig_ref]. However, the mutation was lethal as expected due to the swap of decoding properties from leucine to serine codons. Surprisingly, mutation G37U totally abolished aminoacylation, indicating that G37 residue is essential for the aminoacylation of tRNA Leu (34) [fig_ref] Figure 4: Northern blot analysis of tRNA Leu samples [/fig_ref]. It has been reported that A35 of yeast tRNA Leu (UAG) is crucial for aminoacylation of tRNA. Introducing a A35G was found to reduced the k cat and K m of tRNA Leu (UAG) dramatically in vitro [bib_ref] The anticodon loop is a major identity determinant of Saccharomyces cerevisiae tRNA(Leu), Soma [/bib_ref]. In the present investigation we did not select any mutant of A35, probably because adenine residues are poorly reactive to hydroxylamine. Thus, we constructed the A35G mutation and introduced it into the triple tRNA-knockout strain. The mutation induced a lethal phenotype on 5-FOA plate [fig_ref] Figure 3: Scheme and diagram showing the mutations studied and the effect on cell... [/fig_ref] and Supplementary [fig_ref] Figure 3: Scheme and diagram showing the mutations studied and the effect on cell... [/fig_ref] as expected since the decoding property was shifted from Leu to Pro. We analyzed the in vivo aminoacylation level of the mutant and found that the aminoacylation level was decreased by about 30% [fig_ref] Figure 4: Northern blot analysis of tRNA Leu samples [/fig_ref] , which is an effect significantly different from the in vitro reported data (total abolished aminoacylation) [bib_ref] The anticodon loop is a major identity determinant of Saccharomyces cerevisiae tRNA(Leu), Soma [/bib_ref]. Thus, the anticodon mutations affected aminoacylation slightly; indicating that A35 or G36 are not critical elements of the aminoacylation reaction, which is the universal recognition model between LeuRS and tRNA Leu from various species in many in vitro studies [bib_ref] Recognition of tRNA Leu by Aquifex aeolicus leucyl-tRNA synthetase during the aminoacylation..., Yao [/bib_ref].
## The anticodon stem is involved in editing function of leurs
LeuRS contains an editing domain that deacylates misacyl-tRNA that could alter the fidelity of the translation process. Under usual growth conditions, yeast cells are not exposed to an excess of non-cognate amino acids and the editing function is hardly observable. However, when yeast cells are exposed to an excess of non-cognate amino acids, mis-aminoacylation of tRNA occurs and editing becomes essential for the viability [bib_ref] Unique residues crucial for optimal editing in yeast cytoplasmic Leucyl-tRNA synthetase are..., Yao [/bib_ref]. To assay the editing properties, we replicated the previous 5000 transformants of the triple knockout strain tl(uag)-D1-3 on plates containing 10 mM norvaline (Nva), an analog of leucine [bib_ref] Unique residues crucial for optimal editing in yeast cytoplasmic Leucyl-tRNA synthetase are..., Yao [/bib_ref]. We identified a single clone showing a reduced growth rate on 5-FOA plates. The tRNA Leu gene was exhibiting a G42U mutation [fig_ref] Figure 5: Toxic effect of Nva observed with mutants of the anticodon stem [/fig_ref] and Supplementary [fig_ref] Figure 4: Northern blot analysis of tRNA Leu samples [/fig_ref]. A dose-dependent assay was set up in the present of various concentrations of Nva to further confirm that G42U mutant was sensitive to 5-FOA [fig_ref] Figure 5: Toxic effect of Nva observed with mutants of the anticodon stem [/fig_ref] and Supplementary [fig_ref] Figure 4: Northern blot analysis of tRNA Leu samples [/fig_ref]. In the presence of 3 mM Nva, the yeast strain harboring the G42U mutation grew slower than the WT in the same conditions (Supplementary [fig_ref] Figure 4: Northern blot analysis of tRNA Leu samples [/fig_ref]. When northern blots were performed on tRNA extracts prepared in acidic conditions, a complete aminoacylation of G42U mutant was observed [fig_ref] Figure 5: Toxic effect of Nva observed with mutants of the anticodon stem [/fig_ref]. However, Leu-tRNA Leu and Nva-tRNA Leu are very similar molecules that cannot be distinguished by PAGE. Thus, we concluded that that G42U mutation does not prevent the aminocylation by Nva or Leu but more likely affect the editing of Nva-tRNA Leu , therefore leading to growth rate defects on 5-FOA plates.
Because G42 forms a strong G:C base pair with C28, it might strengthen the stability of the anticodon stem which may be critical for the editing activity. To test this hypothesis, we further constructed a double mutant C28A/G42U exhibiting a weaker A:U base pair. The in vivo aminoacylation level of the double mutant was not affected and was similar to the single G42U mutant [fig_ref] Figure 5: Toxic effect of Nva observed with mutants of the anticodon stem [/fig_ref] while the growth effect recovered to some extent as compared with the G42U single point mutation tRNA Leu (UAG) [fig_ref] Figure 5: Toxic effect of Nva observed with mutants of the anticodon stem [/fig_ref] and Supplementary [fig_ref] Figure 4: Northern blot analysis of tRNA Leu samples [/fig_ref].
Altogether, the results suggest that the anticodon stem plays a role during the editing reaction. A perfect base pairing for a correct geometry and/or a high thermodynamic affinity seem to be needed in the anticodon stem of tRNA Leu for an optimal editing activity catalyzed by LeuRS.
Aminoacylation, deacylation and EF1-a protection assays on tRNA Leu (UAG)s Yeast tRNA Leu (UAG) WT and three representative mutants (A35G, G37U, G42U) were purified from yeast transformants containing their genes using biotin-Streptavidin Agarose Resin. The partially purified tRNA Leu (UAG)s were further separated on a denaturing gel. The final tRNAs looked homogeneous in the gel (Supplementary [fig_ref] Figure 5: Toxic effect of Nva observed with mutants of the anticodon stem [/fig_ref]. The plateau value of WT and its two mutants A35U and G42U reached 1600 pmole/A 260 , however, the mutant G37U could not be aminoacylated [fig_ref] Table 1: Kinetic constants of ycLeuRS for yctRNA Leu [/fig_ref]. This is consistent with the undetectable in vivo growth rate [fig_ref] Figure 3: Scheme and diagram showing the mutations studied and the effect on cell... [/fig_ref] and 5B). Kinetic constants of ycLeuRS for WT, A35U and G42U mutants of tRNA Leu (UAG) were assayed, which showed that the catalytic efficiency of ycLeuRS for these tRNA Leu s was not obviously changed. However, leucylation of G37U mutant was abolished completely [fig_ref] Table 1: Kinetic constants of ycLeuRS for yctRNA Leu [/fig_ref] which corroborates the in vivo results [fig_ref] Figure 3: Scheme and diagram showing the mutations studied and the effect on cell... [/fig_ref]. Because modified G37 in many tRNA Leu s is crucial to their leucylation [bib_ref] Role of a tRNA base modification and its precursors in frameshifting in..., Waas [/bib_ref] , G37U mutation must change the modification of G37, and our results further indicated the importance of modified G37 in tRNA Leu .
In vivo, the ester linkage between the amino acid and tRNA is labile and is readily hydrolyzed if the aa-tRNA is not recognized and bound by elongation factor. To eliminate the possibility that some mutated tRNAs have lost their binding capacity to elongation factor in vivo, we cloned the gene encoding yeast EF-1a (33), purified the protein and carry out charging and protection assays. In the presence or absence of EF-1a, leucylation of tRNA Leu (UAG) WT and its three mutants A35G, G37U and G42U were assayed. The data showed that their accepting activities were not improved by the presence of EF-1a [fig_ref] Figure 6: Effect of EF-1a on the aminoacylation of tRNA Leu [/fig_ref]. Moreover, we assayed the protection of aminoacylated Leu-tRNA Leu by EF-1a from hydrolysis in water in the absence of ycLeuRS. Under physiological conditions (pH 7.5), the ester linkage between Leu and tRNA Leu (UAG) was moderately hydrolyzed and the presence of EF-1a only slightly improved the stability of the ester linkage [fig_ref] Figure 6: Effect of EF-1a on the aminoacylation of tRNA Leu [/fig_ref]. No obvious difference was observed between WT tRNA Leu (UAG), A35G and G42U mutants indicating that the lethal phenotype of the mutants is not due to a loss of protection by EF-1a in vivo.
Our in vivo experiment reveals that excess Nva decreases the growth rate of yeast cells harboring the tRNA Leu (UAG)-G42U mutant. Our previous work also elucidated that the Nva could be mischarged and incorrect proteins were generated [bib_ref] Unique residues crucial for optimal editing in yeast cytoplasmic Leucyl-tRNA synthetase are..., Yao [/bib_ref]. An in vitro editing assay was performed to investigate whether Nva could be charged to tRNA Leu (UAG)-G42U. Although no Nva-tRNA Leu (UAG)-G42U was accumulated when catalyzed by ycLeuRS (data not shown); the hydrolytic activity of ycLeuRS to mischarge Ile-tRNA Leu (UAG)-G42U was reduced [fig_ref] Figure 7: In vitro editing the mischarged tRNA Leu [/fig_ref] , and in the presence of tRNA Leu (UAG)-G42U and Nva, AMP formation was decreased (Table 2, [fig_ref] Figure 7: In vitro editing the mischarged tRNA Leu [/fig_ref] and C) Additionally, Nva-AMP formation was increased [fig_ref] Figure 7: In vitro editing the mischarged tRNA Leu [/fig_ref] and C) as compared to tRNA Leu (UAG) WT. Further, tRNA Leu (UAG)-A35G revealed a slight decrease in editing activity. The in vivo and in vitro data both indicated that A35 and G42 at the anticodon stem and loop are important for editing.
# Discussion
Saccharomyces cerevisiae is using superwobbling to read leucine codons (44)-For instance, in E. coli, tRNA Leu (UAG) translates CUA, CUU, CUG, but not CUC in vitro due to a uridine-5-oxyacetic acid modification [bib_ref] Over expression of a tRNA(Leu) isoacceptor changes charging pattern of leucine tRNAs..., Sørensen [/bib_ref]. Modification of bases may also have opposite effects and expand the decoding capacity. tRNA Val (UAC) of E. coli and tRNA Pro (UGG) of Salmonella, contain a 5-oxyacetic acid modified U34 which helps to decode a degenerate codon box [bib_ref] The modified wobble nucleoside uridine-5-oxyacetic acid in tRNAPro(cmo5UGG) promotes reading of all..., Nasvall [/bib_ref]. The structural basis for the expansion of the decoding capacity of tRNA by modification of uridines was recently solved [bib_ref] Mechanism for expanding the decoding capacity of transfer RNAs by modification of..., Weixlbaumer [/bib_ref]. In addition it is now evident that unmodified U34 pairs with all four bases in third position of codons (in UNN family boxes). Thus, a single anticodon with a general formula UNN is potentially able to translate the four codons in its family box. The phenomenon, called 'superwobbling' represents one possible mechanism of reduction of in number of tRNAs [bib_ref] Superwobbling facilitates translation with reduced tRNA sets, Rogalski [/bib_ref]. This form of translation takes place in Mycoplasma and some mitochondria and chloroplasts [bib_ref] Occurrence of unmodified adenine and uracil at the first position of anticodon..., Andachi [/bib_ref]. However, if superwobbling facilitates translation with a reduced set [fig_ref] Table 2: Observed steady-state constants of ycLeuRS in AMP and Nva-AMP synthesis reaction in... [/fig_ref]. of tRNAs, it has to pay the price of a reduced translation efficiency compared to most organisms that prefer more than one anticodon per family box [bib_ref] Superwobbling facilitates translation with reduced tRNA sets, Rogalski [/bib_ref]. The data here presented showed that the gene for tRNA Leu (GAG) could be deleted without impairing yeast growth, whereas the knockout of the gene of tRNA Leu (UAG) was lethal. This suggests that CU(C/U) codons were decoded by another tRNA which should obviously be tRNA Leu (UAG). Thus, the ability of yeast tRNA Leu (UAG) to translate the four codons might imply a superwobble mechanism as observed with tRNAs carrying unmodified U34 [bib_ref] Yeast tRNA Leu (anticodon U-A-G) translates all six leucine codons in extracts..., Weissenbach [/bib_ref]. Indeed, the spectrophotometric investigations performed on yeast tRNA Leu (UAG) showed that U34 was not modified and was even able to translate the six Leu codons in vitro [bib_ref] Yeast tRNA Leu (anticodon U-A-G) translates all six leucine codons in extracts..., Weissenbach [/bib_ref]. Thus, our finding that protein synthesis occurs in the absence of tRNA Leu (GAG) demonstrates that a single tRNA species with an unmodified U in the wobble position can accurately decode triplets with all four nucleotides in the third codon position, therefore supporting the superwobble hypothesis to sustain protein biosynthesis. Based on protein synthesis defects, it was proposed that superwobbling is less efficient than standard and wobble base-pairing, presumably because codon-anticodon recognition is delayed [bib_ref] Superwobbling facilitates translation with reduced tRNA sets, Rogalski [/bib_ref]. Here, we did not observe any growth defect of the tRNA Leu (GAG) knockout cells, however, we did not thoroughly analyze growth rates in different conditions, thus, we cannot exclude that superwobbling is less efficient than standard and wobble base-pairing.
An intriguing question is whether the capacity for superwobbling of tRNA Leu (UAG) carrying an unmodified U is used spontaneously in WT yeast cells or only in tRNA Leu (GAG) deleted cells. The capacity of reading may be latent in the WT cells and revealed by the tRNA deletion.
The six Leu codons are spread in a four-codon box (CUN) and a two-codon box (UUR). In S. cerevisiae, the usage of codons CUU, CUC, CUG, CUA, UUG and UUA is 12.9%, 5.8%, 11.2%, 14.2%, 27.8% and 28.1%, respectively (51). At the tRNA gene level, we can see that four tRNA genes are in charge of decoding the four-codon box and 17 tRNA genes for the two-codon box. Of interest, we can notice that CU(C/ U) codons that are theoretically decoded by tRNA Leu (GAG) represent 18.7% (12.9+5.8%) of the total number of Leu codons, which is a high value for a tRNA transcript from a single gene copy. For instance, tRNA Leu (UAG) decoding CU(G/A) is transcribed from three genes and represents 24.4% (11.2+14.2%) of Leu codons. The extreme example is tRNA Leu (CAA) which decodes UUG codons (27.8% of total Leu codons) from 10 copies of gene. Altogether, the data indicate that the ratio of codons to read per tRNA isoacceptor gene is the highest for tRNA Leu (GAG), suggesting that superwobbling by tRNA Leu (UAG) might occur in vivo as soon as a deficit of charged tRNA Leu (GAG) appears. However, it remains unclear whether superwobbling happens in normal growth conditions or only extreme conditions. tRNA Leu (UAG) triple knockout strain as a tool to select tRNA Leu mutants in vivo After construction of the triple tRNA Leu (UAG) knockout strain, three loss-of-function mutations (A73U, G18U and É55G) were constructed to test whether the strain could be used to isolate lethal mutations. The three mutations were chosen because they modified the main identity element (A73) or the tertiary structure of the tRNA (G18, É55) that has been shown to be essential for Leu identity in almost all tRNA Leu isoacceptors tested. In our assay, the triple knockout strain transformed with the mutated tRNA genes was unable to grow or survive because the cells could not shuffle the rescuing plasmid carrying the WT tRNA Leu gene. The result of this first assay was lethality in agreement with the in vitro data that showed critical decreases of the aminoacylation efficiency of the three tRNA mutants [bib_ref] Recognition of tRNA Leu by Aquifex aeolicus leucyl-tRNA synthetase during the aminoacylation..., Yao [/bib_ref]. The result confirmed that the triple knockout strain and the rescuing plasmid could be used to detect tRNA mutations critical for aminoacylation or other functions of the tRNA.
We constructed a library of tRNA Leu genes randomly mutated by hydroxylamine and used it to isolate tRNA mutations that induced lethality of the yeast knockout strain. Five thousand transformants were plated on minimal medium supplemented by the required amino acids and then replicated on plates containing 5-FOA. Clones those were unable to grow on the later medium contained lethal tRNA Leu genes. Similarly, the 5000 clones were replicated on medium containing an excess of norvaline (Nva) in order to isolate tRNA Leu mutants inducing a toxic effect in the presence of Nva. Nva is an analog of Leu that is misactivated and mischarged on tRNA Leu by LeuRS and then proofread by the editing activity of the same enzyme [bib_ref] Unique residues crucial for optimal editing in yeast cytoplasmic Leucyl-tRNA synthetase are..., Yao [/bib_ref]. From this selection, it was expected to isolate tRNA Leu mutants that impact the tRNA-dependent editing activity of LeuRS, leading to the incorporation of higher amounts of Nva into the proteins and then to toxicity.
As a result from the selections, we isolated tRNA Leu mutants that induced effects that could be classified in three distinct groups as follows.
(i) The first group contained the two lethal mutations located in the T-arm: C61A and C62A. Northern blot analysis showed that the two tRNAs were perfectly aminoacylated, however the steady-state levels of the two tRNAs were severely decreased, to $30% and 25% of the WT level for C61A and C62A, respectively. These low levels might result from a transcription defect since transcription by RNA polymerase III (41) is using an internal promoter called internal control region (ICR) located in both A box (nt [bib_ref] Role of tRNA amino acid-accepting end in aminoacylation and its quality control, Zhou [/bib_ref] [bib_ref] Tertiary structure base pairs between D-and TpsiC-loops of Escherichia coli tRNA(Leu) play..., Du [/bib_ref] [bib_ref] Recognition of tRNA Leu by Aquifex aeolicus leucyl-tRNA synthetase during the aminoacylation..., Yao [/bib_ref] [bib_ref] A T-stem slip in human mitochondrial tRNA Leu (CUN) governs its charging..., Hao [/bib_ref] [bib_ref] A present-day aminoacyl-tRNA synthetase with ancestral editing properties, Zhu [/bib_ref] [bib_ref] Discrete determinants in transfer RNA for editing and aminoacylation, Hale [/bib_ref] [bib_ref] The C-Ala domain brings together editing and aminoacylation functions on one tRNA, Guo [/bib_ref] [bib_ref] Structural and mechanistic basis of pre-and posttransfer editing by leucyl-tRNA synthetase, Lincecum [/bib_ref] [bib_ref] A single methyl group prevents the mischarging of a tRNA, Pu¨tz [/bib_ref] [bib_ref] One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products, Datsenko [/bib_ref] [bib_ref] Plasmid systems to study RNA function in Escherichia coli, Gabriel [/bib_ref] [bib_ref] A yeast knockout strain to discriminate between active and inactive tRNA molecules, Geslain [/bib_ref] [bib_ref] Early embryonic lethality caused by targeted disruption of the mouse selenocysteine tRNA..., Bosl [/bib_ref] and B box (nt 51-63) [bib_ref] Transcription by RNA polymerase III: more complex than we thought, White [/bib_ref]. Thus, lethal mutations in positions 61 and 62 might affect transcription as already proposed for equivalent mutations isolated in tRNA Arg and tRNA Ser in yeast [bib_ref] A yeast knockout strain to discriminate between active and inactive tRNA molecules, Geslain [/bib_ref] [bib_ref] Multiple mutations of the first gene of a dimeric tRNA gene abolish..., Nichols [/bib_ref]. Alternatively, the low levels of tRNAs could result from a rapid degradation due to unstable tRNA structures. Indeed, the two mutations C61A and C62A introduced mismatches in the T-arm that might destabilize the tRNA Leu structures and target them to rapid degradation.
(ii) The second group contained tRNA Leu mutations inducing decreases of the in vivo aminoacylation levels: G2U, G30U, A35G, G36U and G37U. The G2U mutation probably destabilized the structure of the acceptor arm by changing the G:U wobble base pair to a U-U. G2U mutation caused a decrease in the aminoacylation level and a decrease in growth of the yeast cells. Due to its location, G2U might influence the CCA acceptor end conformation during the aminoacylation reaction and explain the decrease of aminoacylation level. The decrease of the aminoacylation level of mutant G30U is even lower compared to G2U, however, it impacts growth rate of yeast cells as G2U mutant. Therefore, this moderate effect can hardly explain the in vivo growth defect. As G30U introduced a U 30 -C 40 mismatch in the anticodon-arm, one might speculate that the mutation impacts protein biosynthesis after aminoacylation, during codon reading on the ribosome, for instance.
Yeast cells harbor four distinct tRNA Leu isoacceptors decoding the six Leu codons. The variety of anticodons suggests that the anticodon triplet is not crucial for LeuRS recognition. Indeed, our data showed that A35G and G36U mutation caused only moderate decreases of the aminoacylation levels of $40% and 20%, respectively. This indicates that nucleotides from the anticodon loop are involved in some way in the aminoacylation process as previously proposed [bib_ref] The anticodon loop is a major identity determinant of Saccharomyces cerevisiae tRNA(Leu), Soma [/bib_ref]. However, in our case, the two mutations were lethal because they induced a loss of Leu-decoding capacity of the yeast cells and not because of the reduced aminoacylation.
Nucleotide 37 was also isolated as a lethal mutation G37U. Nucleotide 37, which is often a modified nucleotide [bib_ref] Tertiary structure checkpoint at anticodon loop modification in tRNA functional maturation, Goto-Ito [/bib_ref] , is involved in the reading frame maintenance, ensuring accurate translation on the ribosome [bib_ref] Role of a tRNA base modification and its precursors in frameshifting in..., Waas [/bib_ref] [bib_ref] Tertiary structure checkpoint at anticodon loop modification in tRNA functional maturation, Goto-Ito [/bib_ref]. For instance, a previously isolated G37U mutation in yeast tRNA Tyr was shown to be lethal despite normal transcription and aminoacylation levels (55), suggesting a defect at the translation level. Similarly, mammalian tRNA Lys lacking A37 modification caused type II diabetes. Surprisingly, our data showed that G37U mutation in yeast tRNA Leu totally abolished aminoacylation in vivo. We further purified G37U mutated tRNA from yeast and found that its aminoacylation was also abolished in vitro. Therefore, the lethal effect was first due to a defect of aminoacylation and revealed that G37 is an essential determinant for LeuRS recognition.
(iii) The third group of mutations contained G42U and derivative G42U/C28A, both located in the anticodon arm. Both induced a growth inhibition when the yeast cells were grown on medium containing high amounts of Nva, an analog of Leu, which is usually edited by LeuRS to prevent its incorporation into proteins. Editing of tRNA by LeuRS is dependent on the translocation of the 3 0 CCA end from the synthetic site to the editing site [bib_ref] Aminoacylation complex structures of leucyl-tRNA synthetase and tRNA Leu reveal two modes..., Fukunaga [/bib_ref]. Different nucleotides from the tRNA are involved in the process. In E. coli tRNA Val , a mutation of A76U specifically abolished the editing response without impairing the aminoacylation [bib_ref] Misacylation and editing by Escherichia coli valyl-tRNA synthetase: evidence for two tRNA..., Tardif [/bib_ref]. In tRNA Leu from A. aeolicus, discrete editing determinants in the anticodon arm were found to stabilize the conformation of the post-transfer editing state. In the present work, we identified a G42U mutation that reduced tolerance of yeast to excess of Nva, suggesting some toxic effect due to decrease of the editing activity of LeuRS. The data also showed that a strong C 28 :G 42 base pair is strictly required and cannot be replaced by a A:U pair. In both cases, the aminoacylation of tRNA Leu was kept intact. Here the data show that yeast growth was considerably reduced despite standard aminoacylation and tRNA levels. This strongly suggests that the toxic effect was due to the inability to clear mischarged products that appeared in the presence of elevated concentrations of Nva. This hypothesis was further confirmed by showing that the editing activity of purified G42U mutant was reduced, leading to Nva-adenylate accumulation that could be later charged on tRNA Leu and incorporated into proteins.
tRNA Leu (UAG) triple knockout strain as a tool to obtain tRNA Leu mutants Transcripts from S. cerevisiae tRNA Leu (both GAG and UAG isoacceptors) are very poor substrates for LeuRS and hardly used for biochemical investigations [bib_ref] Structural and mechanistic basis of pre-and posttransfer editing by leucyl-tRNA synthetase, Lincecum [/bib_ref]. The study on the interaction between yeast tRNA Leu and LeuRS is difficult. We could not obtain yeast tRNA Leu with high enough accepting activity from E. coli transformants containing its gene or by T7 polymerase transcription. Using the knockout strains tl(uag)-D1-3, we could not only investigate the function of tRNA Leu by the growth rate of yeast cells in vivo but also obtain any tRNA Leu (UAG) mutant with the modified bases from yeast tl(uag)-D1-3 transformants containing the mutant's gene. In vitro the studies on effect of a series of interesting individual nucleotide of yeast tRNA Leu (UAG) on its function becomes possible. We will use the yeast knockout strains to study the effect of more nucleotides of tRNA Leu (UAG) on aminoacylation and editing of ycLeuRS by site-directed mutagenesis. Further studies are in progress to find new selection tools and perform more extensive analysis of the tRNA properties.
## Supplementary data
[fig] Figure 1: Strategy of construction and characterization of the tRNA Leu (GAG) knockout strain. (A) Schematic map of the chromosomal TL(GAG) region and of the UHD DNA fragment constructed to exchange the TL(GAG) gene. Arrows symbolize the PCR primers used to verify the knockout experiment. Predicated sizes and restriction sites are indicated. (B) Agarose gel of the PCR products of YPH501 recipient strain and tl(gag)::HIS3 strain. Lane 1: 100 bp plus DNA ladder (Transgene, Beijing, China); lanes 2 and 3: PCR products amplified with tl(gag)::HIS3 genomic DNA. Lanes 4 and 5: PCR products amplified from the recipient YPH501 yeast strain. Lane 6: PCR product from lane 2 digested with NotI. Lane 7: PCR product from lane 3 digested with XbaI. (C) Growth of strains on YPDA plate. Parental strain YPH501 [a/ TL(GAG) +/+], the knockout diploid strain [a/ TL(GAG)/tl(gag)::HIS3] and the haploid knockout strain [tl(gag)::HIS3] are equally growing on YPDA. (D) Northern blot analysis on the RNA content of the yeast strains. 5S RNA was detected as an internal control. The tRNA probe was the following 5 0 -CCCGC GCCUC CGAAG AGAUC AGGAC CUCGA-3 0 . [/fig]
[fig] Figure 2: Construction of the tRNA Leu (UAG) triple knockout yeast strain. (A, C and E) Schematic map of the chromosomal regions containing the three tRNA Leu (UAG) genes and of the homologous recombination DNA fragment constructed to delete the TL(UAG) genes. Arrows symbolize the PCR primers used to verify the knockout experiment. Predicated sizes and restriction sites are indicated. (B, D and F) Agarose gels of the PCR products of YPH499 strain and tRNA Leu deleted strains. Lanes 1: DNA ladder; lanes 2 and 3: PCR products amplified with genomic DNAs from deleted strains. Lanes 4 and 5: PCR products amplified from the recipient YPH499 yeast strain. Lanes 6: PCR products from lanes 2 digested with NotI or BamHI. Lanes 7: PCR products from lanes 3 digested with XbaI, XhoI or EcoRI. (G) Parental strain (YPH499), derivative strain WT-R (transformed pALR-T into YPH499), single knockout strain tl(uag)1::HIS3, double knockout strain tl(uag)1,2::HIS3/LEU2 and triple knockout strain tl(uag)1-3::HIS3/LEU2/ADE2 were grown on a normal YPDA medium and then spread over on a plate containing 5-FOA. (H) Northern blot analysis on the RNA content of the three knockout strains. 5S RNA was detected as an internal control. The tRNA probe was the following 5 0 -CCUUG CAUCC GAAGA UAUCA GAGCC UAAAU-3 0 . [/fig]
[fig] Figure 3: Scheme and diagram showing the mutations studied and the effect on cell growth. (A) Cloverleaf structure of tRNA Leu (UAG) indicating the mutations (underlined) performed by random selected (in bold) or performed by side-directed mutagenesis (in italics). Putative impacts of the mutations are suggested. (B) Relative growth rate of yeast strains harboring a mutant tRNA Leu measuring in liquid medium. The error bars are standard deviations from three independent measurements. [/fig]
[fig] Figure 4: Northern blot analysis of tRNA Leu samples. (A) Total RNA samples from lethal mutants were extracted under acidic conditions, electrophoresed on a 6.5% polyacrylamide gel containing 8 M urea at pH 5.0, 4 C, and transferred to a Hybond membrane. The blot was probed with an oligonucleotide complementary to 5S RNA (control for RNA content) and tRNA Leu . '+' and '-' marks refer to alkaline treatment (OH À ) used to deacylate the RNA. (B) Grey scale scan and quantization of the amount of Leu-tRNA Leu . The amounts were normalized according to 5S RNA. The error bars are standard deviations from three independent measurements. [/fig]
[fig] Figure 5: Toxic effect of Nva observed with mutants of the anticodon stem. (A) The growth rate of yeast strains expressing WT tRNA Leu , G42U and G42U/C28A mutant tRNA Leu in Nva-containing liquid medium. The growth rates of the mutants are compared to the WT rate that is normalized to 100%. (B) Northern blot analysis of the WT, G42U and G42U/C28A mutants. Probes were directed against 5S RNA (internal control) and tRNA Leu . The error bars are standard deviations from three independent measurements. [/fig]
[fig] Figure 6: Effect of EF-1a on the aminoacylation of tRNA Leu (UAG)s and stability of the ester linkage. (A) Aminoacylation of tRNA Leu (UAG)s by ycLeuRS (1 nM). Reactions were performed in the presence of EF-1a (solid lines) of tRNA Leu -WT (filled circle), -A35G (filled triangle), -G37U (filled diamond) and -G42U (filled square). or in the presence of BSA (dashed lines) of tRNA Leu -WT (unfilled circle), -A35G(unfilled triangle), -G37U (unfilled diamond) and -G42U (unfilled square). (B) Spontaneous hydrolysis of Leu-tRNA Leu (UAG)s in the presence of EF-1a (solid lines): tRNA Leu -WT (filled circle), -A35G (filled triangle) and -G42U (filled square) or in the presence of BSA (dashed lines): tRNA Leu -WT (unfilled circle), -A35G(unfilled triangle) and -G42U (unfilled square). The error bars are standard deviations from three independent measurements. [/fig]
[fig] Figure 7: In vitro editing the mischarged tRNA Leu (UAG) WT and mutants by ycLeuRS and the AMP and Nva-AMP formations with the corresponding tRNAs. (A) Deacylation assays catalyzed by ycLeuRS (1 nM) of 1 mM of Ile-tRNA Leu WT (filled circle), -A35G (filled triangle), -G42U (filled square), reaction without enzyme (filled diamond) was performed as spontaneous hydrolysis control. (B) TLC plates showing the [32P]AMP and Nva-[32P]AMP formation catalyzed by ycLeuRS (1 mM). (C) Quantification of the AMP or Nva-AMP formation with tRNA Leu (UAG) -WT (filled circle), -A35G (filled triangle), -G42U (filled square) and without tRNA (filled diamond), respectively. kobs values of AMP and Nva-AMP formations are calculated from the slopes and reported in [/fig]
[table] Table 1: Kinetic constants of ycLeuRS for yctRNA Leu (UAG) WT and mutants in the aminoacylation reaction [/table]
[table] Table 2: Observed steady-state constants of ycLeuRS in AMP and Nva-AMP synthesis reaction in the presence of Nva and yctRNA Leu (UAG) WT or its mutants [/table]
|
Is the skin microbiota a modifiable risk factor for breast disease?: A systematic review
Keywords:Breast cancer Breast disease Microbiota Breast skin Antimicrobial Probiotic a b s t r a c t Purpose: High prevalence, unreliable risk discrimination and poor clinical outcomes are observed in malignant and benign breast diseases (BD). The involvement of microbial communities in the development of BD has become topical, and distal influences of microbial dysregulation in the breast have been well established. Despite advances, the role of the breast skin microbiota in BD remains unclear. Interactions between the skin microbiota and the underlying mucosal immune system are complex. In homeostasis, the skin offers a physical barrier protecting underlying breast tissue from skin commensals and noxious environmental triggers. Our review aims to illuminate the role of the skin microbiota in the development of BD. Methods: Adhering to the PRISMA protocol, a systematic review was conducted utilising the Medline and Embase search engines. Results: Through a comprehensive search of the last ten years, twenty-two studies satisfied the inclusion criteria. Proteobacteria, Firmicutes, Actinobacteria and Bacteroidetes were identified as the most prevalent phyla of both breast tissue and skin in healthy controls and BD. High abundance of skin commensals, specifically some species of Staphylococcus, have been linked in breast cancer and metastases. Similarly, dysregulated microbial abundance is also seen in inflammatory and implant-associated BD. These findings raise the hypothesis that the skin microbiota plays a role in tissue homeostasis and may contribute to a range of breast pathologies. Several mechanisms of microbial transfer to underlying tissue have been proposed, including retrograde transfer through ductal systems, breakdown of the skin barrier, and migration through nipple-aspirate fluid. Conclusion: Our review provides preliminary insights into the skin microbiota as a modifiable risk factor for BD. This raises opportunities for future studies in antimicrobials/probiotics as an adjunct to, or replacement of surgery; a diagnostic and/or prognostic tool for BD; and the possibility of conditioning the microbiota to manage BD.
## Significance of breast disease
Breast disease (BD) consists of cancerous (malignant) and noncancerous (benign) conditions. Breast cancer (BC) accounts for nearly one quarter of all cancers worldwide [bib_ref] Global cancer statistics. Ca -Cancer, Jemal [/bib_ref]. Risk discrimination remains poor, with~70 % of BC cases occurring in women of average risk [bib_ref] Model for individualized prediction of breast cancer risk after a benign breast..., Pankratz [/bib_ref]. Despite significant therapeutic advances and reduction in mortality rates, BC continues to be the major cause of cancerrelated deaths in women [bib_ref] Global cancer in women: burden and trends, Torre [/bib_ref]. Granular lobular mastitis (GLM) is an inflammatory breast disease known to clinically mimic BC [bib_ref] Clinical metagenomic analysis of bacterial communities in breast abscesses of granulomatous mastitis, Yu [/bib_ref]. Whilst showing a high rate of resolution and remission, there has been a shift away from undesirable surgical intervention [bib_ref] Granulomatous lobular mastitis, Mahlab-Guri [/bib_ref]. Furthermore, complications of breast implant surgery remain a significant burden, responsible for a great proportion of surgical revisions [bib_ref] Is banning texturized implants to prevent breast implant-associated anaplastic large cell lymphoma..., Danilla [/bib_ref].
## Microbial impact on disease
Microbial environments have shown to influence the expression of genes and cellular differentiation of the immune system [bib_ref] The effect of probiotics on immune regulation, acne, and photoaging, Kober [/bib_ref]. Dysregulation of microbial homeostasis (referred to as dysbiosis), has been recently reported in BC and other BD [bib_ref] The microbiome of aseptically collected human breast tissue in benign and malignant..., Hieken [/bib_ref]. Certain bacterial taxa are ubiquitous among studies, whilst others note the impact of more rare bacterial lineages with disease onset [bib_ref] The microbiome of aseptically collected human breast tissue in benign and malignant..., Hieken [/bib_ref]. Distal sources influencing the breast microbiota, such as the entero-mammary pathway [bib_ref] The origin of human milk bacteria: is there a bacterial enteromammary pathway..., Rodríguez [/bib_ref] , have been well established in the literature, and growing evidence has further assessed the role of urine [bib_ref] Diagnostic Kit of breast cancer via urine microbiome, An [/bib_ref] , oral [bib_ref] The origin of human milk bacteria: is there a bacterial enteromammary pathway..., Rodríguez [/bib_ref] [bib_ref] The human milk microbiota: origin and potential roles in health and disease, Fern [/bib_ref] [bib_ref] Association between breast milk bacterial communities and establishment and development of the..., Pannaraj [/bib_ref] , and vaginal [bib_ref] Microbiota of human breast tissue, Urbaniak [/bib_ref] microbes. Despite major advances, how the breast skin microbiota contributes to breast dysbiosis and the development of BD at present remains unclear.
## The skin: a symbiotic relationship
The skin is composed of an abundance of folds, invaginations and specialised niches which support a microbial ecosystem unique to its anatomical region [bib_ref] The skin microbiome, Grice [/bib_ref]. Skin commensals, such as Staphylococcus epidermidis, can establish lifelong symbiotic relationships with their host. They provide benefits through aiding in nutrition, outcompeting pathogens, and educating the innate/adaptive immune system [bib_ref] Pathogenic mechanisms and host interactions in Staphylococcus epidermidis device-related infection, Sabat E Bresc O [/bib_ref]. Influences of the skin microbiota extends beyond hygiene. Colonisation is driven by the ecology of the skin's surface, including: (i) the unique anatomical region, (ii) changes over a lifetime, and (iii) influences of endogenous and exogenous factors [bib_ref] The skin microbiome, Grice [/bib_ref] [fig_ref] Figure 1: Schematic of the breast and breast skin anatomy illustrating unique niches where... [/fig_ref]. Herein, the 'breast skin' is our primary focus. The anatomy of the nipple-areolar complex is unique: a multi celllayered structure containing numerous nerve endings, sebaceous and apocrine sweat glands [bib_ref] Multimodality approach to the nippleareolar complex: a pictorial review and diagnostic algorithm, Del Riego [/bib_ref]. Skin dysbiosis has been established with breast skin conditions (psoriasis and atopic dermatitis) [bib_ref] The skin microbiome, Grice [/bib_ref] , and skin morphological changes are seen with underlying BD (erythema, dimpling, oedema and peau d'orange) [bib_ref] Guidelines on the diagnosis and treatment of breast cancer, Zhang [/bib_ref]. Skin dysbiosis has also been linked with skin cancer through the induction of chronic self-maintaining inflammation [bib_ref] The skin microbiome and immune system: potential target for chemoprevention?, Sherwani [/bib_ref] [bib_ref] Non-Melanoma Skin Cancer: news from microbiota research, Squarzanti [/bib_ref]. However, it is still largely unknown whether skin dysbiosis has structural and/or functional influences on underlying breast tissue, and to what extent that precipitates BD development.
## Objective
Our review aims to appraise the literature and illuminate the role of the breast skin microbiota in the development of BD. This raises opportunities for modification and manipulation of the skin microbiota as a conduit to personalised medicine in BD.
# Methods
## Data sources and searches
We performed a systematic review of the literature following the standard Cochrane Collaboration methods and adhering to the Preferred Reporting Items for Systematic Reviews and Metaanalysis (PRISMA) recommendations [bib_ref] Preferred reporting Items for systematic reviews and meta-analyses: the PRISMA statement, Moher [/bib_ref].
A structured electronic literature search was conducted, using Medline and Embase databases. The search strategy consisted of the keywords and MeSH terms for the breast skin microbiota and breast tissue microbiota: 'skin', 'epidermis', 'tissue'. 'microbiota', 'microbiome', 'breast'. Date of publication was limited to the last 10 years. The following filters were applied: language (English) and species (human). A hand search of bibliographies was conducted to identify any additional articles. All titles and abstracts were independently reviewed by two of the authors. If there was disagreement between the reviewers, the inclusion of studies was determined by consensus. All study types: randomised control trials (RCTs), caseecontrol, cohort and case studies were eligible for inclusion; reviews were excluded.
The different study designs and the heterogeneity of the outcomes reported in the studies precluded the possibility of pooling data across the studies. Therefore, a narrative synthesis was conducted. Key study characteristics and methodological quality are described in [fig_ref] Table 1: Summary of relative phyla abundance across studies for breast skin [/fig_ref] 3. Results
## Literature search
The search identified 478 articles [fig_ref] Figure 2: Flowchart of breast skin/tissue literature search [/fig_ref]. After duplicates were removed, 414 papers were screened for inclusion. 347 articles were excluded following title and abstract screen and a further 45 based on study type, excluding a total of 392. Twenty-two articles were selected for data extraction. Hand search of bibliographies yielded no additional papers. Articles appraised date from 2014 to 2020.
## Key study characteristics
Developments in DNA/RNA sequencing technologies have provided the opportunity to closely monitor potential association(s) between the development of BD and the tissue's resident microbiota. Breast skin and breast tissue samples from healthy controls (HC) and BD were of interest. Variation in sequencing methodologies among the publications is shown in Supplementary Tables 1 and 2, along with patient demographics, sample size, sample sites.
## Phyla of healthy and diseased breast
Collectively, all studies identified Proteobacteria, Firmicutes, Actinobacteria and Bacteroidetes as the main residents of the breast microbiota. [fig_ref] Table 1: Summary of relative phyla abundance across studies for breast skin [/fig_ref] highlights the relative abundance of the phyla with respect to HC versus BD in both breast skin [fig_ref] Table 1: Summary of relative phyla abundance across studies for breast skin [/fig_ref] and breast tissue [fig_ref] Table 1: Summary of relative phyla abundance across studies for breast skin [/fig_ref].
Proteobacteria and Firmicutes are observed in greater abundance in comparison to other taxa across the studies appraised. Significant species diversity has been identified in Proteobacteria residing within the skin. Genealogical research has shown that pathological Proteobacteria found in the environment can be harboured in the skin [bib_ref] Proteobacteria from the human skin microbiota: species-level diversity and hypotheses, Cosseau [/bib_ref] , causing difficulty in delineating pathological microbes from commensal. Similar to what is observed in the skin samples, the breast tissue demonstrates predominance of Proteobacteria and Firmicutes [fig_ref] Table 1: Summary of relative phyla abundance across studies for breast skin [/fig_ref]. Reduced Proteobacteria abundance in BD samples is evident in studies comparing HC to BD [22e25]. Although subtle differences in Firmicutes abundance between HC and BD samples are observed, no overarching trend in predominance of these organisms can be established at the phyla level. Lack of clarity towards the prevalence of cutaneous microbes in breast tissue dysbiosis warrants closer examination of the microbiota at the genus and species level.
## A closer look into dysbiosis
Dysbiosis is the result of introduced pathogenic microbes, or loss of symbiotic microbes [fig_ref] Figure 3: Microbial shifts contrasting HC and BD tissue samples [/fig_ref]. Studies supporting the former raise questions as to whether the penetration of skin commensals to underlying tissue provides the opportunity for bacteria to become pathobiontic. Staphylococcus, is a common skin commensal that has evolved strategies for surviving on the skin, including the ability to be halotolerant, and utilise urea that is present in sweat as a nitrogen source [bib_ref] What lives on our skin: ecology, genomics and therapeutic opportunities of the..., Scharschmidt [/bib_ref]. Staphylococcus, along with Methylobacterium, Enterobacter and Streptococcus have also shown survivability in fatty-acid environments such as the breast [bib_ref] Microbiota of human breast tissue, Urbaniak [/bib_ref]. These skin microbes have not only demonstrated the ability to habitat a new environment but also metabolise fatty-acids, resulting in possible deleterious changes to the microarchitecture of the tissue.
Alternatively, a notable decrease in symbiotic microbes like Prevotella, Lactococcus, and Sphingomonas have been associated with BD [bib_ref] The microbiome of aseptically collected human breast tissue in benign and malignant..., Hieken [/bib_ref] [bib_ref] Distinct microbial communities that differ by race, stage, or breast-tumor subtype in..., Smith [/bib_ref] [bib_ref] Microbial dysbiosis is associated with human breast cancer, Xuan [/bib_ref] [bib_ref] The microbiota of breast tissue and its association with breast cancer, Urbaniak [/bib_ref]. A possible explanation pertains to the interaction between the microorganisms and activation of natural killer cells to inhibit tumour growth [bib_ref] The microbiota of breast tissue and its association with breast cancer, Urbaniak [/bib_ref]. Logically, a reduction in such microorganisms within the breast tissue would hinder any tumour-suppression and immune response. The effect of these bacteria, and several others, in the pathogenesis of BD and modulation of cellular immunity needs further investigation.
## Breast cancer and the breast skin microbiota
Heiken et al., is the only study that directly compares breast skin against breast tissue in HC and BC paired samples. The authors suggest that the microbial landscape of breast tissue is distinct from that of the overlying skin tissue, exhibiting greater species richness primarily in rare or less abundant bacterial lineages [bib_ref] The microbiome of aseptically collected human breast tissue in benign and malignant..., Hieken [/bib_ref]. The paper did not comment on whether skin dysbiosis occurred during BC, but Chan et al., concludes that there is no breast skin dysbiosis associated with BC [bib_ref] Characterization of the microbiome of nipple aspirate fluid of breast cancer survivors, Chan [/bib_ref]. No significant relationship has been established between skin dysbiosis and BC. A study conducted by Urbaniak et al., provides greater insight into the role of skin bacteria in BC. Staphylococcus, a common skin commensal, was among the most abundant genera identified in BC tissue [bib_ref] Comparative analysis of racial differences in breast tumor microbiome, Thyagarajan [/bib_ref] [bib_ref] The microbiota of breast tissue and its association with breast cancer, Urbaniak [/bib_ref] and has further been implicated in BC metastasis [bib_ref] Neoadjuvant chemotherapy shifts breast tumor microbiota populations to regulate drug responsiveness and..., Chiba [/bib_ref]. Further characterisation of the species within the Staphylococcus genus is needed to understand which are involved in the breast oncogenic process. Literature has highlighted Staphylococcus aureus as a common culprit for carcinogenesis. It has been shown that the incidence of primary cancers is greater following S. aureus bacteraemia [bib_ref] Increased risk of incident primary cancer after Staphylococcus aureus bacteremia: a matched..., Gotland [/bib_ref] , and a strong association is seen with S. aureus colonisation and squamous cell carcinoma [bib_ref] Staphylococcus aureus and squamous cell carcinoma of the skin, Kullander [/bib_ref].
Other taxa shown to have increased prevalence in BC include: Bacillus, Bacteroidetes, Brevundimonas, Comamonadaceae, Enterobacteriaceae, and Methylobacterium [bib_ref] Comparative analysis of racial differences in breast tumor microbiome, Thyagarajan [/bib_ref] [bib_ref] Microbial dysbiosis is associated with human breast cancer, Xuan [/bib_ref] [bib_ref] The microbiota of breast tissue and its association with breast cancer, Urbaniak [/bib_ref] [fig_ref] Figure 3: Microbial shifts contrasting HC and BD tissue samples [/fig_ref]. Literature has found the skin microbiota to harbour these taxa [bib_ref] The skin microbiome, Grice [/bib_ref] , speculating the potential for microbial transfer from the skin to tissue. Brevudimonas, found in the dermis, has been reported as a clinically important pathobiont in cases of immunocompromised individuals [bib_ref] Brevundimonas spp: emerging global opportunistic pathogens, Ryan [/bib_ref]. In some instances, contaminants introduced from the operator and environment to the DNA sequencer have resulted in false positives. Methylobacterium has been identified as one of the common contaminants in a study conducted by Bay et al., 2020 [bib_ref] Universal dermal microbiome in human skin, Bay [/bib_ref].
## Inflammatory breast disease and the breast skin microbiota
The role of skin microbiota can be surveyed beyond the scope of malignant disease. Three studies included in this review addressed the breast microbiota and its association with inflammatory BD disease. [bib_ref] Clinical metagenomic analysis of bacterial communities in breast abscesses of granulomatous mastitis, Yu [/bib_ref] [bib_ref] Breast abscess caused by Staphylococcus aureus in 2 adolescent girls with atopic..., Park [/bib_ref] [bib_ref] Pathogens in patients with granulomatous lobular mastitis, Wang [/bib_ref] Park et al., proposed that skin dysbiosis in atopic dermatitis (AD) was the source of breast abscess development [bib_ref] Breast abscess caused by Staphylococcus aureus in 2 adolescent girls with atopic..., Park [/bib_ref]. Patients with AD exhibit significantly higher abundance of S. aureus in diseased skin compared to healthy skin [bib_ref] The skin microbiome, Grice [/bib_ref]. Breakdown of the epidermal layer experienced in individuals with AD presents an opportunity for skin microbes to transfer and cause deep tissue infection.
Granular lobular mastitis is a rare, poorly understood inflammatory disease of the breast. Studies consistently report Corynebacterium as a dominant genus found in the taxonomic profiles of GLM patients [bib_ref] Clinical metagenomic analysis of bacterial communities in breast abscesses of granulomatous mastitis, Yu [/bib_ref] [bib_ref] Pathogens in patients with granulomatous lobular mastitis, Wang [/bib_ref]. Corynebacterium has been associated with normal skin flora as well as with acute/chronic skin and soft tissue infection [bib_ref] Non diphtheritic corynebacteria: an emerging nosocomial pathogen in skin and soft tissue..., Rudresh [/bib_ref]. Corynebacterium kroppenstedtii, among other lipophilic Corynebacterium species [bib_ref] Idiopathic granulomatous lobular mastitis, Pereira [/bib_ref] , have been implicated in the development of GLM through generation of an auto-immune response [bib_ref] Idiopathic granulomatous mastitis associated with corynebacterium sp. Infection, Stary [/bib_ref] [bib_ref] A clinicopathological review of 34 cases of inflammatory breast disease showing an..., Taylor [/bib_ref]. Other taxa seen at higher prevalence in GLM include Pseudomonas, Brevundimonas, Stenotrophomonas, Acinetobacter, and Aspergillus. The most identified species included Pseudomonas aeruginosa and Pseudomonas stutzeri. Pseudomonas aeruginosa can be commonly found on the skin, especially in the axillary and anogenital regions [bib_ref] Pseudomonas aeruginosa infections of intact skin, Agger [/bib_ref] [bib_ref] Pseudomonas skin infection, Wu [/bib_ref] ; areas sharing similar traits to that of the nipple. These cutaneous microbes possess fatty-acid metabolising capabilities, and as discussed previously, can become pathobiontic in breast tissue [bib_ref] Pseudomonas aeruginosa directly shunts boxidation degradation intermediates into de novo fatty acid..., Yuan [/bib_ref].
## Complications of breast implants and the breast skin microbiota
The role of the breast skin microbiota in the pathogenesis of breast implant complications has been discussed for breast capsular contracture (CC) [bib_ref] PCR characterization of microbiota on contracted and non-contracted breast capsules, Bachour [/bib_ref] [bib_ref] Back to basics: could the preoperative skin antiseptic agent help prevent biofilm-related..., Carvajal [/bib_ref] , and breast implant-associated anaplastic large cell lymphoma (BIA-ALCL) [bib_ref] Bacterial biofilm infection detected in breast implanteassociated anaplastic large-cell lymphoma, Hu [/bib_ref] [bib_ref] Insights into the microbiome of breast implants and periprosthetic tissue in breast..., Walker [/bib_ref]. CC and BIA-ALCL are considered significant complications in breast augmentation and implant-based breast reconstruction surgery [bib_ref] Is banning texturized implants to prevent breast implant-associated anaplastic large cell lymphoma..., Danilla [/bib_ref]. Previous literature primarily attributes the development of these complications to subclinical infection and bacterial biofilms following contamination of implants during surgical insertion [bib_ref] Analysis of subclinical infections and biofilm formation in cases of capsular contracture..., Jirawatnotai [/bib_ref].
Clear correlation between CC and the presence of Staphylococcus epidermidis has been established [bib_ref] Back to basics: could the preoperative skin antiseptic agent help prevent biofilm-related..., Carvajal [/bib_ref] [bib_ref] Bacterial biofilms and capsular contracture in patients with breast implants, Rieger [/bib_ref]. Notably, Bachour et al., identified the presence of Staphylococcus spp. on only 8% of breast capsule samples. The authors propose that bacterial transfer during implantation is highly unlikely and any resultant bacteria observed is due to contamination during removal of capsules [bib_ref] PCR characterization of microbiota on contracted and non-contracted breast capsules, Bachour [/bib_ref]. Coagulase-negative Staphylococci [bib_ref] Surgical site infections: epidemiology, microbiology and prevention, Owens [/bib_ref] , Cutibacterium acnes [bib_ref] Cutibacterium acnes in primary reverse shoulder arthroplasty: from skin to deep layers, Torrens [/bib_ref] , and Pseudomonas aeruginosa [bib_ref] Nonrandom distribution of Pseudomonas aeruginosa and Staphylococcus aureus in chronic wounds, Fazli [/bib_ref] are also commonly found in surgical site infections, as well as in healthy human skin [bib_ref] The human skin microbiome, Byrd [/bib_ref]. Transfer via surgical instruments allow the potential for opportunistic subclinical infection, and subsequent breast-implant complications.
In addition to CC, bacterial colonisation and biofilm formation has been associated with BIA-ALCL [bib_ref] Bacterial biofilm infection detected in breast implanteassociated anaplastic large-cell lymphoma, Hu [/bib_ref]. Walker et al., found Staphylococcus spp. to be the most commonly colonised microbe in both the BIA-ALCL and contralateral control breast [bib_ref] Insights into the microbiome of breast implants and periprosthetic tissue in breast..., Walker [/bib_ref]. Alternatively, Hu et al., observed significantly higher abundance of Staphylococcus spp. in non-tumour capsule specimens when compared to BIA-ALCL specimens [bib_ref] Bacterial biofilm infection detected in breast implanteassociated anaplastic large-cell lymphoma, Hu [/bib_ref]. Significantly greater proportion of Ralstonia spp. was present in BIA-ALCL specimens, but there is no evidence to suggest Ralstonia is endogenous to the skin. Uncertainty towards dysbiosis as a causative agent in the development of implant-associated disease warrants greater attention and research.
## Mechanisms of microbial transfer
Mechanisms of bacterial transfer between the skin and breast are largely unknown, although various hypotheses have been proposed. Concerning breast implant surgery, iatrogenic breakdown of the skin barrier results in underlying tissue contamination by skin commensals, this was demonstrated by in the development of CC and BIA-ALCL [bib_ref] Back to basics: could the preoperative skin antiseptic agent help prevent biofilm-related..., Carvajal [/bib_ref]. Similarly in AD, Park et al., suggested that the high bacterial density, altered skin barrier, and immune dysregulation caused resident S. aureus to penetrate deep into the tissue [bib_ref] Breast abscess caused by Staphylococcus aureus in 2 adolescent girls with atopic..., Park [/bib_ref]. Translocation of skin microbiota to tissue in BC and other BD characterised with intact skin is more uncertain. Retrograde transfer via the nipple and ductal system [bib_ref] The human milk microbiota: origin and potential roles in health and disease, Fern [/bib_ref] , altered immune responses of the skin [bib_ref] Pathogenic mechanisms and host interactions in Staphylococcus epidermidis device-related infection, Sabat E Bresc O [/bib_ref] and migration through nipple aspirate fluid [bib_ref] Microbial dysbiosis is associated with human breast cancer, Xuan [/bib_ref] have all been proposed as potential mechanisms at play [bib_ref] Skin microbiota: a source of disease or defence?, Cogen [/bib_ref].
# Biases and limitations
Next-generation sequencing has allowed for rapid and accurate developments in profiling the microbiota on body surfaces. Despite advances, there are still some biases and limitations that require consideration. The microbiota is influenced by several confounding factors; both internal and external [bib_ref] Temporal variability is a personalized feature of the human microbiome, Flores [/bib_ref]. The variability seen in the microbiota among individuals presents challenges in drawing definitive conclusions, and affects the reliability of causal inference [bib_ref] Temporal variability is a personalized feature of the human microbiome, Flores [/bib_ref]. The lack of standardised methodologies (sample collection, preservation, transportation and time to analysis) continues to be a well-established limitation when comparing studies [bib_ref] The madness of microbiome: attempting to find consensus "best practice" for 16S..., Pollock [/bib_ref] [bib_ref] Understanding and overcoming the pitfalls and biases of next-generation sequencing (NGS) methods..., Boers [/bib_ref]. The effect of bias can lead to the discovery of non-existent bacterial genera, misinformed correlations between microorganisms and disease, and missed detection of true correlations. Bias may also be observed by the exclusive utilisation of 16 S rRNA gene sequencing. It has been shown in the gut microbiota that shotgun sequencing (also known as metagenomic sequencing) has more power to identify less abundant taxa, and the less abundant genera detected are biologically more meaningful [bib_ref] Comparison between 16S rRNA and shotgun sequencing data for the taxonomic characterization..., Durazzi [/bib_ref]. Moreover, it is difficult to summarise the alpha diversity (species richness within a sample) and beta diversity (species richness between samples) due to different statistical and bioinformatics analyses being used across studies.
# Conclusion
Investigation into the breast skin microbiota as a modifiable risk factor for BD has attained promising evidence. Given the vast number of associations between commensal breast skin microbes and BD, the breast skin microbiota seems to function as a source of pathogenicity to breast tissue.
Scope for clinical application is broad. Chlorhexidine, a preoperative skin antiseptic solution, is already being implemented in modern practices to reduce implant contamination [bib_ref] Back to basics: could the preoperative skin antiseptic agent help prevent biofilm-related..., Carvajal [/bib_ref]. Antimicrobials are being used for their anti-proliferative, pro-apoptotic and anti-epithelial-mesenchymal-transition (EMT) capabilities [bib_ref] Antibiotics for cancer treatment: a double-edged sword, Gao [/bib_ref]. Although effective in infections, the interplay of antibiotics and the healthy microbiota makes broad application more complex. Studies have suggested that antibiotics could adversely impact cancer treatment by disrupting intestinal microbiota, altering normal tissue metabolism, and weakening the immune response [bib_ref] Antibiotics for cancer treatment: a double-edged sword, Gao [/bib_ref]. Alternatively, the prospect of probiotics as a therapeutic means of introducing healthy bacteria to modulate disease is gaining prominence. Promising benefits have been seen in noninfectious skin conditions including acne, rosacea and protection against aging and damage [bib_ref] The effect of probiotics on immune regulation, acne, and photoaging, Kober [/bib_ref]. The breast skin provides an opportunity for topical antimicrobial/probiotic treatment of BD. In addition, understanding the skin microbiota can also assist in early screening and diagnosis of BD, as demonstrated in preliminary studies using the urine microbiota [bib_ref] Diagnostic Kit of breast cancer via urine microbiome, An [/bib_ref]. Another use includes their role as a prognostic predictor, as preliminary studies suggest that microbiota characteristics can influence the response to chemotherapeutics or determine the risk of metastasis [bib_ref] The skin microbiome and immune system: potential target for chemoprevention?, Sherwani [/bib_ref] [bib_ref] Neoadjuvant chemotherapy shifts breast tumor microbiota populations to regulate drug responsiveness and..., Chiba [/bib_ref].
More evidence establishing the role of the skin microbiota is needed, including establishing a standardised protocol for characterising clinically significant microbes, understanding environmental influences, the cellular and molecular host-microbe interaction, and mechanism of transfer from skin to tissue. Greater understanding will provide significant opportunities for not only clinical management strategies as listed above, but also the potential in engineering/manipulating the human microbiota to manage disease.
[fig] Figure 1: Schematic of the breast and breast skin anatomy illustrating unique niches where microbes can reside, and exogenous and endogenous factors that can contribute to variation in breast microbiota over a lifetime. [/fig]
[fig] Figure 2: Flowchart of breast skin/tissue literature search. [/fig]
[fig] Figure 3: Microbial shifts contrasting HC and BD tissue samples. A: BC compared to HC. B: Inflammatory BD compared to HC. C: CC compared to HC; and D: BIA-ALCL compared to HC. [/fig]
[table] Table 1: Summary of relative phyla abundance across studies for breast skin (A.) and breast tissue (B.). Proteobacteria (P), Firmicutes (F), Actinobacteria (A), Bacteroidetes (B), Other (O).Studies included in this analysis that did not comment on phyla abundances were omitted from the table. [/table]
|
Structural Health Monitoring (SHM) and Determination of Surface Defects in Large Metallic Structures using Ultrasonic Guided Waves
Ultrasonic guided wave (UGW) is one of the most commonly used technologies for non-destructive evaluation (NDE) and structural health monitoring (SHM) of structural components. Because of its excellent long-range diagnostic capability, this method is effective in detecting cracks, material loss, and fatigue-based defects in isotropic and anisotropic structures. The shape and orientation of structural defects are critical parameters during the investigation of crack propagation, assessment of damage severity, and prediction of remaining useful life (RUL) of structures. These parameters become even more important in cases where the crack intensity is associated with the safety of men, environment, and material, such as ship's hull, aero-structures, rail tracks and subsea pipelines. This paper reviews the research literature on UGWs and their application in defect diagnosis and health monitoring of metallic structures. It has been observed that no significant research work has been convened to identify the shape and orientation of defects in plate-like structures. We also propose an experimental research work assisted by numerical simulations to investigate the response of UGWs upon interaction with cracks in different shapes and orientations. A framework for an empirical model may be considered to determine these structural flaws.Author Contributions: M.A performed the analysis and wrote the manuscript text. M.S reviewed the manuscript and made contributions to its structure.Funding: This research received no external funding.Conflicts of Interest:The authors declare no conflict of interest.
# Introduction
Mechanical structures and their reliability deteriorate continuously under the influence of operating and ambient environmental conditions. The structural degradation processes may occur in different forms of corrosion, fatigue cracks, erosion, and strength reduction [bib_ref] Condition assessment, remaining useful life prediction and life extension decision making for..., Animah [/bib_ref]. The extent and severity of these mechanisms depend on the applied static/dynamic loads, operating conditions, and material properties such as corrosion resistance, microstructural features, the crystallographic orientation and the grain-boundary structure. The inspection and maintenance of mechanical structures were performed mainly in accordance with a preventive (time-based) approach in the past. In recent years, the significance of structural health monitoring (SHM) techniques has been enhanced in the health assessment, remaining useful life (RUL) prediction, and condition-based maintenance (CBM) of degrading structures [bib_ref] Maintenance optimization and inspection planning of wind energy assets: Models, methods and..., Shafiee [/bib_ref].
The SHM approach involves collecting data from various sensors installed on structures and interpreting the findings to make reliable and cost-effective and CBM decisions for structures under different operating conditions. Numerous health assessment techniques such as bulk wave ultrasonic, X-rays, infrared thermography, and eddy current are used effectively for non-destructive evaluation (NDE) of structures. However, majority of these techniques tend to be slow and cumbersome, especially in case of large structures (pipelines, marine, ships, and aerospace). The guided waves ultrasonic testing (GWUT) potentially offers a smart alternative solution to conventional approach towards NDE and SHM [bib_ref] Practical long range guided wave inspection-Applications to pipes and rail, Cawley [/bib_ref]. The guided wave (GW) has a unique ability to confine itself inside the thin-wall structures, therefore, they can propagate over large distances with minimal attenuation and loss of energy. They are also known as long-range ultrasonic testing (LRU), introduced initially to determine the integrity assessment of large and continuous pipelines [bib_ref] Hardware and software system of the guided wave pipe testing using electromagnetic-acoustic..., Lenkov [/bib_ref]. The GWs have the potential to be used inside curved structures thus suitable for inspection of various shapes and geometries over a longer distance.
The SHM is generally performed by two methods: the passive and active SHM. The passive SHM is an examining method used in static/dynamic and rotating equipment. It requires external excitation source to produce elastic waves in structures and utilize the functioning parameters such as vibration and stress levels for health assessment. However, the passive SHM does not provide the root cause(s) of the problem and the opportunity for a thorough examination of the defect. Contrary, the active SHM is a direct health assessment method for detection and estimation of structural defects using health detection act structures actuators/sensors. Like NDE, the SHM sensors detect and quantify the structural defect but in the latter technique sensors are permanently fixed/embedded in structures; to provide health monitoring data on need basis [bib_ref] Reliability analysis of structural health monitoring systems, Etebu [/bib_ref]. A schematic view of active and passive SHM methods is shown in [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref]. ultrasonic, X-rays, infrared thermography, and eddy current are used effectively for non-destructive evaluation (NDE) of structures. However, majority of these techniques tend to be slow and cumbersome, especially in case of large structures (pipelines, marine, ships, and aerospace). The guided waves ultrasonic testing (GWUT) potentially offers a smart alternative solution to conventional approach towards NDE and SHM [bib_ref] Practical long range guided wave inspection-Applications to pipes and rail, Cawley [/bib_ref]. The guided wave (GW) has a unique ability to confine itself inside the thin-wall structures, therefore, they can propagate over large distances with minimal attenuation and loss of energy. They are also known as long-range ultrasonic testing (LRU), introduced initially to determine the integrity assessment of large and continuous pipelines [bib_ref] Hardware and software system of the guided wave pipe testing using electromagnetic-acoustic..., Lenkov [/bib_ref]. The GWs have the potential to be used inside curved structures thus suitable for inspection of various shapes and geometries over a longer distance. The SHM is generally performed by two methods: the passive and active SHM. The passive SHM is an examining method used in static/dynamic and rotating equipment. It requires external excitation source to produce elastic waves in structures and utilize the functioning parameters such as vibration and stress levels for health assessment. However, the passive SHM does not provide the root cause(s) of the problem and the opportunity for a thorough examination of the defect. Contrary, the active SHM is a direct health assessment method for detection and estimation of structural defects using health detection act structures actuators/sensors. Like NDE, the SHM sensors detect and quantify the structural defect but in the latter technique sensors are permanently fixed/embedded in structures; to provide health monitoring data on need basis [bib_ref] Reliability analysis of structural health monitoring systems, Etebu [/bib_ref]. A schematic view of active and passive SHM methods is shown in [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref]. The UGW is a rapidly emerging SHM technique for inspection of metallic/composite structures [bib_ref] Progress on ultrasonic guided waves de-icing techniques in improving aviation energy efficiency, Wang [/bib_ref]. The prominent features of GWUT such as high sensitivity to surface/subsurface flaws, long reach, efficacy in finding faults of inaccessible areas have made this approach a cost-effective and preferable choice for structural health assessment [bib_ref] Structural health monitoring of immersed structures by means of guided ultrasonic waves, Rizzo [/bib_ref]. Contrary to the conventional ultrasonic testing system which offers defect diagnosis only under immediate vicinity of a diagnostic probe, the GWs (Lamb waves in particular) are able to identify defects at larger distances by virtue of its distinct low frequencies and tendency to confine with the structural surfaces [bib_ref] Embedded ultrasonics NDE with piezoelectric wafer active sensors, Giurgiutiu [/bib_ref]. It has the potential to examine various types of structural flaws such as gouges, deep corrosion spots, cracks and notches using one or more sensors [bib_ref] Maintenance optimization and inspection planning of wind energy assets: Models, methods and..., Shafiee [/bib_ref] [bib_ref] Utilization of guided elastic waves for the characterization of circumferential cracks in..., Ditri [/bib_ref] [bib_ref] Potential and limitations of a deconvolution approach for guided wave structural health..., Cicero [/bib_ref]. GWs have been used to detect, localized the defects and corrosion under insulation (CUI) in pipelines and various metallic/non-metallic surfaces [bib_ref] The effect of dispersion on long-range inspection using ultrasonic guided waves, Wilcox [/bib_ref] [bib_ref] Guided waves for monitoring of plate structures with linear cracks of variable..., Zima [/bib_ref].
This paper aims to review research publications on structural defect detection with GW technique, particularly in metallic structures. It signifies the achievements of this technique to detect various features of defects (crack/notches) on metal surfaces. We identify certain research gaps with respect to the classification of some crucial facets (shape and orientation of cracks/notches) of structural flaws, detectable by UGW techniques. Furthermore, to the best of our knowledge, the determination of these characteristics of defects (using UGW method) in plate type structures has not been addressed in any research study to date. To overcome the research gap, this paper also proposes The UGW is a rapidly emerging SHM technique for inspection of metallic/composite structures [bib_ref] Progress on ultrasonic guided waves de-icing techniques in improving aviation energy efficiency, Wang [/bib_ref]. The prominent features of GWUT such as high sensitivity to surface/subsurface flaws, long reach, efficacy in finding faults of inaccessible areas have made this approach a cost-effective and preferable choice for structural health assessment [bib_ref] Structural health monitoring of immersed structures by means of guided ultrasonic waves, Rizzo [/bib_ref]. Contrary to the conventional ultrasonic testing system which offers defect diagnosis only under immediate vicinity of a diagnostic probe, the GWs (Lamb waves in particular) are able to identify defects at larger distances by virtue of its distinct low frequencies and tendency to confine with the structural surfaces [bib_ref] Embedded ultrasonics NDE with piezoelectric wafer active sensors, Giurgiutiu [/bib_ref]. It has the potential to examine various types of structural flaws such as gouges, deep corrosion spots, cracks and notches using one or more sensors [bib_ref] Maintenance optimization and inspection planning of wind energy assets: Models, methods and..., Shafiee [/bib_ref] [bib_ref] Utilization of guided elastic waves for the characterization of circumferential cracks in..., Ditri [/bib_ref] [bib_ref] Potential and limitations of a deconvolution approach for guided wave structural health..., Cicero [/bib_ref]. GWs have been used to detect, localized the defects and corrosion under insulation (CUI) in pipelines and various metallic/non-metallic surfaces [bib_ref] The effect of dispersion on long-range inspection using ultrasonic guided waves, Wilcox [/bib_ref] [bib_ref] Guided waves for monitoring of plate structures with linear cracks of variable..., Zima [/bib_ref].
This paper aims to review research publications on structural defect detection with GW technique, particularly in metallic structures. It signifies the achievements of this technique to detect various features of defects (crack/notches) on metal surfaces. We identify certain research gaps with respect to the classification of some crucial facets (shape and orientation of cracks/notches) of structural flaws, detectable by UGW techniques. Furthermore, to the best of our knowledge, the determination of these characteristics of defects (using UGW method) in plate type structures has not been addressed in any research study to date. To overcome the research gap, this paper also proposes an experimental research assisted by numerical simulation on thin metal plates (aluminium). It aims to determine the signal variations in excited Lamb waves (short tone burst) after interacting with the surface cracks (through-thickness and partial through-thickness) in several shapes and orientations. Based on the theoretical knowledge and previously held research work on GWs, it seems quite possible to sense the projected changes in an output signal (amplitude, reflection/transmitted coefficients, acoustic impedance and arrival time) upon interaction with these aspects of structural discontinuities.
The rest of the paper is organized as follows: Section 2 describes the types, specifications, and salient facets of guide waves for SHM. Section 3 presents the literature review, deliberating the scope, capability, application areas, and implications of this technology. Section 4 comprises discussion of the results, proposes an experimental plan to apply UGWs, especially in SHM of large marine structures. Section 5 concludes the study and discusses directions for future research.
## Ultrasonic guided waves and shm
The GWs are elastic waves transmitting in form of wave packets, and confined along the structures with distinct boundaries. The propagation of GWs depends on the material properties and slight variation may take place with variation in environmental conditions. These waves are highly dispersive, and almost infinite modes are generated by the superposition of basic longitudinal and transverse waves. The Lamb wave is one of the main types of GWs used in SHM. They are elastic perturbations normally traveling with a different phase (C p ) and group (C g ) velocity in metals. In here, the displacement of medium occurs both in the direction of wave propagation as well as in a perpendicular direction. The Shear and Rayleigh waves are the other types of GWs which are also used in SHM and NDE. These waves travel with slower speed than the Lamb waves. The vibration of Lamb waves may be observed through the entire thickness of the medium.
The GWs are able to travel over long distances without any significant loss of energy [bib_ref] Group velocity and characteristic wave curves of Lamb waves in composites: Modeling..., Wang [/bib_ref]. Although certain complications are also associated with the GW technology in SHM such as multimode excitation, dispersive behavior, boundary scattering, and difficulty in interpretation [bib_ref] Application of Wavelet Transform as a Signal Processing Method for Defect Detection..., Kordbacheh [/bib_ref]. Nevertheless, several approaches (e.g., variation in incident pulse frequency, pulse shape, short tone burst, filters, signal processing method) have been used to minimize the undesirable features of GWs [bib_ref] The Mode Conversion of a Guided Wave by a Part-Circumferential Notch in..., Lowe [/bib_ref] [bib_ref] Response of Gaussian-modulated guided wave in aluminum: An analytical, numerical, and experimental..., Khan [/bib_ref].
When the elastic waves are propagated on a metallic surface, the induced stress waves tend to produce elastic strain. The damage/discontinuities in structures cause exclusive wave scattering and mode conversions, which can be visualized on the output sensor. The SHM techniques can be used to capture these crack transitions in the form of variation in reflected/transmitted waves, which can be further utilized for identification of vicinity of defect and its quantification [bib_ref] Detection of cuts and impact damage at the aircraft wing slat by..., Senyurek [/bib_ref]. Contrary to the labour intensive, time-consuming, and expensive inspections by other techniques, GWUT can be helpful to inspect wider structural area in less time, cost and cheap equipment setup [bib_ref] Design and testing of an efficient and compact piezoelectric energy harvester, Korla [/bib_ref] [bib_ref] Angular crack monitoring of Aluminum plate by Lamb wave analysis, Demetgül [/bib_ref].
The UGWs have been used in health prognostic studies in several ways including contact method, semi contact, and non-contact probing techniques [bib_ref] Ultrasonic guided waves-based monitoring of rail head: Laboratory and field tests, Rizzo [/bib_ref]. In the contact method, various types of acoustic sensors are used which are either bonded (with epoxy) or embedded in structural surfaces. The excitation and sensing of acoustic waves can be achieved by contact transducers of several types, normally made of an electro-active and magneto-active substance. These substances have the ability to change their features when induced by an electric or magnetic signature and respond in the form of strain actuation/sensing. The piezoelectric polymers (PVDF), piezoelectric ceramics (PZT), electrostrictive ceramics (PMN), and magnetostrictive materials (Terfenol-D) are the types of transducer used in SHM.
The electromagnetic acoustic transducers (EMATs) are a non-contact technique based on the principles of Lorentz force or magnetostriction which have been used enormously in defect diagnosis studies. The EMATs are ideal to directly excite and detected S H waves in the structures, however they are quite larger in size and their utility is restricted to the conductive materials only [bib_ref] Study and comparison of different EMAT configurations for SH wave inspection, Ribichini [/bib_ref] [bib_ref] Shear horizontal piezoelectric fiber patch transducers (SH-PFP) for guided elastic wave applications, Köhler [/bib_ref] [bib_ref] A Novel Sensor Design for Generation and Detection of Shear-Horizontal Waves Based..., Schubert [/bib_ref]. The piezo wafer active sensor (PWAS) and S H -PWAS have also been used for damage detection [bib_ref] Shear horizontal wave excitation and reception with shear horizontal piezoelectric wafer active..., Kamal [/bib_ref]. These acoustic transducers can be utilized in various arrangements like wedge, square, circular, comb and phase array type [bib_ref] Improvement of SH-wave EMAT phased array inspection by new eight segment probes, Sawaragi [/bib_ref]. The PZT is more frequently used for SHM in aerospace, mechanical and civil engineering appliances. It works on the electro-mechanical (EM) impedance/admittance principle. The EM admittance signatures can be used for structural health assessment and flaw detection corresponding to the signal variations [bib_ref] Effect of external vibration on PZT impedance signature, Yang [/bib_ref]. PWAS can be used in several ways such as frequency response transfer function, E/M impedance and wave propagation. A new type of a thin, lightweight and flexible transducers is Shear-horizontal piezoelectric fiber patch (SH-PFP), which has been adopted for exciting the S H waves in light structures in aerospace [bib_ref] Shear horizontal piezoelectric fiber patch transducers (SH-PFP) for guided elastic wave applications, Köhler [/bib_ref] [bib_ref] A Novel Sensor Design for Generation and Detection of Shear-Horizontal Waves Based..., Schubert [/bib_ref] [bib_ref] A Novel Piezoelectric Fibre Patch Transducer for Shear Horizontal Wave Modes, Köhler [/bib_ref] The propagation of GWs can be divided into two broad categories: longitudinal and transverse waves. When these waves propagate in a traction-free metal plate, the longitudinal wave particles tend to travel along the direction of wave propagation (in from of compressions and rarefactions) whereas the transverse wave particles oscillate perpendicular to the direction of wave propagation. The velocity of acoustic waves depends on various properties of the medium of propagation, as shown in the following equations:
[formula] C T = E 2ρ(1 + V)(1)C L = E(1 − V) ρ(1 − 2V)(1 + V)(2) [/formula]
where the term E is Young's modulus, V is Poisson's ratio, ρ is density, and C T and C L represent transverse and longitudinal wave velocities, respectively. Readers can also refer to [bib_ref] Wave Propagation in Structures, Doyle [/bib_ref] for a detailed study on the characteristics and application of long-range UGWs in metallic and composite structures.
## Lamb guided waves
The Lamb ultrasonic guide waves, also known as plate waves, are guided between two parallel traction free surfaces, such as the lower and upper surface of a plate. The Lamb wave speed is also dependent on the frequency of propagating waves. The Lamb waves are a complicated form of GW molded by superimposition of longitudinal and Shear/transverse waves. The two basic modes of Lamb waves are symmetric and antisymmetric modes. Both modes of Lamb waves are dispersive in nature and equally used for SHM. The former mode of waves travel faster while being affected less by the external conditions. The frequency of GW and thickness of structural are deeply correlated to define various features of the GWs. The frequency-depth (f-d) product is proportional to the number of Lamb-wave modes.
Similarly, for smaller f-d product, only fundamental Lamb wave modes of symmetric and antisymmetric (i.e., S 0 and A 0 ) exist. The shape and mode of propagation of particle in both symmetric and antisymmetric Lamb waves are illustrated in . The use of antisymmetric modes of Lamb waves has been reported as a preventive technique for marine fouling deposition on underwater ship structure. In the structural plates immersed in liquid medium, the adequate selection of wave mode can significantly decrease energy leakage into the water [bib_ref] Characterization of Ultrasonic Wave Propagation in the Application of Prevention of Fouling..., Carellan [/bib_ref]. . Schematic of (a) symmetric particle displacement (b) antisymmetric particle displacement (adapted from.
## Shear guided waves
The Shear waves are also called transverse waves in which most of the movement of the medium is perpendicular to the direction of wave propagation. They exist in two forms: the Shear horizontal . Schematic of (a) symmetric particle displacement (b) antisymmetric particle displacement (adapted from. Use of Lamb waves on structural defect detection was first started in early 1960. Since then, rigorous research studies have been conducted for application of these waves in health prognostics of metallic/composite structures [bib_ref] Detection of cuts and impact damage at the aircraft wing slat by..., Senyurek [/bib_ref] [bib_ref] Modelling of in-plane wave propagation in a plate using spectral element method..., Rucka [/bib_ref] [bib_ref] Structural health monitoring using scanning laser vibrometry: II. Lamb waves for damage..., Mallet [/bib_ref] [bib_ref] Application of Guided Wave Propagation in Diagnostics of Steel Bridge Components, Rucka [/bib_ref] [bib_ref] Damage detection in composite materials using Lamb wave methods, Kessler [/bib_ref]. Various SHM based studies have been conducted on Lamb waves behavior on metallic and non-metallic structural defects [bib_ref] Detection of cuts and impact damage at the aircraft wing slat by..., Senyurek [/bib_ref] [bib_ref] Modelling of in-plane wave propagation in a plate using spectral element method..., Rucka [/bib_ref] [bib_ref] Application of Guided Wave Propagation in Diagnostics of Steel Bridge Components, Rucka [/bib_ref] [bib_ref] Structural health monitoring using scanning laser vibrometry: I. Lamb wave sensing, Staszewski [/bib_ref]. These waves have the greater tendency to propagate larger distances (up to 100 m) with little attenuation [bib_ref] Mode and transducer selection for long range lamb wave inspection, Wilcox [/bib_ref]. The real-time data of Lamb waves is also amalgamated with the external excitations, operational and environmental noises, further complicating the interpretation process analysis by shifting of the arrival time [bib_ref] Effect of external vibration on PZT impedance signature, Yang [/bib_ref] [bib_ref] Damage diagnosis and prognosis methodology to estimate safe load for aircraft structure, Seshadri [/bib_ref] [bib_ref] An intelligent methodology for railways monitoring using ultrasonic guided waves, Moustakidis [/bib_ref]. Therefore, various compensatory approaches have been introduced to minimize the signal damages. Chan and Cawley [bib_ref] Lamb waves in highly attenuative plastic plates, Chan [/bib_ref] studied the impact of attenuation on the dispersion behavior of Lamb wave in plastic plates (high-density polyethylene). The attenuation in Lamb wave was found to be five times less that of the Shear wave.
## Shear guided waves
The Shear waves are also called transverse waves in which most of the movement of the medium is perpendicular to the direction of wave propagation. They exist in two forms: the Shear horizontal (S H ) and the Shear vertical (S V ) waves. The particle vibration of S H waves is polarized parallel to the plate surface, as illustrated in [fig_ref] Figure 3: Shear wave propagation [/fig_ref]. The S H waves exist in both symmetric and antisymmetric modes and both are dispersive in nature, with the exception of very fundamental S H0 mode. The significance of S H waves has increased in recent years by the inculcation of new techniques to overcome difficulties associated with S H wave generation. Therefore, superior transducers have been introduced for using shear waves in SHM, such as magnetostrictive transducers and EMATs. . Schematic of (a) symmetric particle displacement (b) antisymmetric particle displacement (adapted from.
## Shear guided waves
The Shear waves are also called transverse waves in which most of the movement of the medium is perpendicular to the direction of wave propagation. They exist in two forms: the Shear horizontal (SH) and the Shear vertical (SV) waves. The particle vibration of SH waves is polarized parallel to the plate surface, as illustrated in [fig_ref] Figure 3: Shear wave propagation [/fig_ref]. The SH waves exist in both symmetric and antisymmetric modes and both are dispersive in nature, with the exception of very fundamental SH0 mode. The significance of SH waves has increased in recent years by the inculcation of new techniques to overcome difficulties associated with SH wave generation. Therefore, superior transducers have been introduced for using shear waves in SHM, such as magnetostrictive transducers and EMATs. The SH wave is used in ice detection and de-icing applications in various structures. They have also been preferred for NDT and health monitoring studies of underwater structures because they do not leak out from structural into the liquid medium. Therefore, the maximum propagation energy retains inside the submerged structure. The SH waves have the unique ability to propagate around curved surfaces with minimal energy loss. These waves can propagate circumferentially around the pipe structures [bib_ref] Circumferential guided wave EMAT system for pipeline screening using shear horizontal ultrasound, Clough [/bib_ref]. The SH waves are polarized in a plane parallel to a surface boundary so they can be reflected without substantial mode conversion and variation in amplitude. Likewise, SH waves have less reflection, beam steering, and attenuation compared with the SV waves; therefore they are considered as a preferred choice in defect detection in welded regions [bib_ref] Shear horizontal (SH) ultrasound wave propagation around smooth corners, Petcher [/bib_ref]. However, it has been observed that any sudden change in the surface thickness may lead to the mode conversion. The S H wave is used in ice detection and de-icing applications in various structures. They have also been preferred for NDT and health monitoring studies of underwater structures because they do not leak out from structural into the liquid medium. Therefore, the maximum propagation energy retains inside the submerged structure. The S H waves have the unique ability to propagate around curved surfaces with minimal energy loss. These waves can propagate circumferentially around the pipe structures [bib_ref] Circumferential guided wave EMAT system for pipeline screening using shear horizontal ultrasound, Clough [/bib_ref]. The S H waves are polarized in a plane parallel to a surface boundary so they can be reflected without substantial mode conversion and variation in amplitude. Likewise, S H waves have less reflection, beam steering, and attenuation compared with the S V waves; therefore they are considered as a preferred choice in defect detection in welded regions [bib_ref] Shear horizontal (SH) ultrasound wave propagation around smooth corners, Petcher [/bib_ref]. However, it has been observed that any sudden change in the surface thickness may lead to the mode conversion.
## Rayleigh waves
Rayleigh waves are also known as surface acoustic waves (SAW) and normally found in solids having free surfaces. As S V wave exists only along the thickness of the plate, it superposes with the longitudinal wave and forms a new type of wave called the Rayleigh wave. These waves travel close to the free surface with slight penetration in the depth. For this reason, Rayleigh waves are also known as surface guided waves. The Rayleigh waves are polarized in a plane normal to the surface and wave particles propagate in a way, making the path reversing ellipses in the vertical plane. These waves occur on the surface of semi-infinite elastic media and suitable for surface and subsurface flaw detection. [fig_ref] Figure 4: Rayleigh waves propagation [/fig_ref] shows Rayleigh wave propagation in metal plates.
## Rayleigh waves
Rayleigh waves are also known as surface acoustic waves (SAW) and normally found in solids having free surfaces. As SV wave exists only along the thickness of the plate, it superposes with the longitudinal wave and forms a new type of wave called the Rayleigh wave. These waves travel close to the free surface with slight penetration in the depth. For this reason, Rayleigh waves are also known as surface guided waves. The Rayleigh waves are polarized in a plane normal to the surface and wave particles propagate in a way, making the path reversing ellipses in the vertical plane. These waves occur on the surface of semi-infinite elastic media and suitable for surface and subsurface flaw detection. [fig_ref] Figure 4: Rayleigh waves propagation [/fig_ref] shows Rayleigh wave propagation in metal plates. These waves travel over the surface of a semi-infinite solid or a very thick plate, where the wavelength is very small as compared to the thickness of the plate. These waves have also been used in NDE of the structural defects, confined to surfaces [bib_ref] Thick hollow cylindrical waveguides: A theoretical, numerical and experimental study, Ziaja [/bib_ref]. They can be used for inspection of artillery shells, acoustic microscopy, and leaky Lamb wave method for composite materials [bib_ref] Maintenance optimization and inspection planning of wind energy assets: Models, methods and..., Shafiee [/bib_ref].
## Phase and group velocity
The phase velocity (Cp) of an acoustic wave is the rate at which the phase of the wave propagates in spatial dimensions. The individual harmonics travel with different Cp, but the superimposed packet travels with the group velocity (Cg). In dispersive elastic wave propagation, the packets of waves travel with a different velocity (Cg) than individual wave velocity Cp.
It has been revealed by Wilcox et al. [bib_ref] The effect of dispersion on long-range inspection using ultrasonic guided waves, Wilcox [/bib_ref] that the dispersion of GWs cause the wave-packets to disperse in space and time domains as they travel in a structure. The equations for Cp and Cg are given by:
[formula] Cp = (3) Cg =(4) [/formula]
where k is wave number and ω is angular velocity of waves. [fig_ref] Figure 5: Phase and group velocities of GWs [/fig_ref] illustrates the phase and group velocities of GWs propagating in metal structures. During the propagation of ultrasonic GWs, the phase velocity is normally a function of the wave frequency and thickness of the structure under examination, therefore, it is obvious that for a given frequency, different thickness samples have different wave velocity [bib_ref] Circumferential guided wave EMAT system for pipeline screening using shear horizontal ultrasound, Clough [/bib_ref]. These waves travel over the surface of a semi-infinite solid or a very thick plate, where the wavelength is very small as compared to the thickness of the plate. These waves have also been used in NDE of the structural defects, confined to surfaces [bib_ref] Thick hollow cylindrical waveguides: A theoretical, numerical and experimental study, Ziaja [/bib_ref]. They can be used for inspection of artillery shells, acoustic microscopy, and leaky Lamb wave method for composite materials [bib_ref] Maintenance optimization and inspection planning of wind energy assets: Models, methods and..., Shafiee [/bib_ref].
## Phase and group velocity
The phase velocity (C p ) of an acoustic wave is the rate at which the phase of the wave propagates in spatial dimensions. The individual harmonics travel with different C p , but the superimposed packet travels with the group velocity (C g ). In dispersive elastic wave propagation, the packets of waves travel with a different velocity (C g) than individual wave velocity C p.
It has been revealed by Wilcox et al. [bib_ref] The effect of dispersion on long-range inspection using ultrasonic guided waves, Wilcox [/bib_ref] that the dispersion of GWs cause the wave-packets to disperse in space and time domains as they travel in a structure. The equations for C p and C g are given by:
[formula] C p = ω k (3) C g = ∂ω ∂k (4) [/formula]
where k is wave number and ω is angular velocity of waves. [fig_ref] Figure 5: Phase and group velocities of GWs [/fig_ref] illustrates the phase and group velocities of GWs propagating in metal structures. During the propagation of ultrasonic GWs, the phase velocity is normally a function of the wave frequency and thickness of the structure under examination, therefore, it is obvious that for a given frequency, different thickness samples have different wave velocity [bib_ref] Circumferential guided wave EMAT system for pipeline screening using shear horizontal ultrasound, Clough [/bib_ref].
## Attenuation in ultrasonic guided waves
The combined effect of scattering and absorption is called attenuation. In ultrasonic GWs, the attenuation is the rate of decay of the waves as they propagate through any material. The energy of GWs dissipates gradually with distance which manifests the steady drop in magnitude of wave signals. Attenuation in UGWs has been reported as a function of wave modes, physical properties of the material, and frequency. The presence of any structural damage or inhomogeneity such as welds, joints, stiffeners or fasteners increases the dissipation factor. As compared to other forms of GWs, the Lamb wave and SH wave can travel longer distances with minimal loss in amplitude [bib_ref] Shear horizontal (SH) ultrasound wave propagation around smooth corners, Petcher [/bib_ref] [bib_ref] The use of Lamb waves for the long range inspection of large..., Cawley [/bib_ref]. The application of surface coatings, intricacies in structural geometry and immersion/embedded conditions often result in high attenuation of acoustic waves. These factors restrict the range of GWs and consequently degrade its quality of the defect identification [bib_ref] Active health monitoring of an aircraft wing with an embedded piezoelectric sensor/actuator..., Zhao [/bib_ref]. The low excitation frequency often results in a phase-lead and substantial attenuation of the signal. During a research investigation on phase velocity and attenuation of GWs in viscoelasticity of human tissues, Nenadic et al. [bib_ref] Phase velocities and attenuations of shear, Lamb, and Rayleigh waves in plate-like..., Nenadic [/bib_ref] reported that the attenuation of the Shear wave was found to be 20% lower than that of Lamb-Rayleigh waves.
## Sensor placement methods for gwut
The pitch-catch and pulse-echo are the most commonly used methods in defect detection. In a pitch-catch, GWs are excited by the actuator and only the waves transmitted across the damage region are captured/monitored by the sensor placed opposite to the exciter, in a catching position. This method normally uses one or more sensors. Any change in signal intensity/energy, velocity shift, arrival time, amplitude, and impedance can signify the characteristics of the damaged area.
In pulse-echo, both the exciter and receptor sensors are located on the same side and the sensor receives the echoed wave (i.e., reflected part of the signal only) signal from any structural discontinuities or boundaries [bib_ref] Structural health monitoring of aircraft structures, Rocha [/bib_ref]. Figures 6 and 7 exemplify both arrangements. In the pulse-echo method, the defect is detected in the form of additional echoes, whereas in the pitch-catch method, the principles of wave dispersion and attenuation are used as the defect indicators. The pitchcatch configuration is also known as through transmission. Because of the wide coverage capability of the pitch-catch method, it has been proposed by few researchers for inspecting the floor of fuel storage tanks, using lamb waves [bib_ref] Long-range Ultrasonic Non-destructive Testing of Fuel Tanks, Mažeika [/bib_ref].
## Attenuation in ultrasonic guided waves
The combined effect of scattering and absorption is called attenuation. In ultrasonic GWs, the attenuation is the rate of decay of the waves as they propagate through any material. The energy of GWs dissipates gradually with distance which manifests the steady drop in magnitude of wave signals. Attenuation in UGWs has been reported as a function of wave modes, physical properties of the material, and frequency. The presence of any structural damage or inhomogeneity such as welds, joints, stiffeners or fasteners increases the dissipation factor. As compared to other forms of GWs, the Lamb wave and S H wave can travel longer distances with minimal loss in amplitude [bib_ref] Shear horizontal (SH) ultrasound wave propagation around smooth corners, Petcher [/bib_ref] [bib_ref] The use of Lamb waves for the long range inspection of large..., Cawley [/bib_ref]. The application of surface coatings, intricacies in structural geometry and immersion/embedded conditions often result in high attenuation of acoustic waves. These factors restrict the range of GWs and consequently degrade its quality of the defect identification [bib_ref] Active health monitoring of an aircraft wing with an embedded piezoelectric sensor/actuator..., Zhao [/bib_ref]. The low excitation frequency often results in a phase-lead and substantial attenuation of the signal. During a research investigation on phase velocity and attenuation of GWs in viscoelasticity of human tissues, Nenadic et al. [bib_ref] Phase velocities and attenuations of shear, Lamb, and Rayleigh waves in plate-like..., Nenadic [/bib_ref] reported that the attenuation of the Shear wave was found to be 20% lower than that of Lamb-Rayleigh waves.
## Sensor placement methods for gwut
The pitch-catch and pulse-echo are the most commonly used methods in defect detection. In a pitch-catch, GWs are excited by the actuator and only the waves transmitted across the damage region are captured/monitored by the sensor placed opposite to the exciter, in a catching position. This method normally uses one or more sensors. Any change in signal intensity/energy, velocity shift, arrival time, amplitude, and impedance can signify the characteristics of the damaged area.
In pulse-echo, both the exciter and receptor sensors are located on the same side and the sensor receives the echoed wave (i.e., reflected part of the signal only) signal from any structural discontinuities or boundaries [bib_ref] Structural health monitoring of aircraft structures, Rocha [/bib_ref]. Figures 6 and 7 exemplify both arrangements. In the pulse-echo method, the defect is detected in the form of additional echoes, whereas in the pitch-catch method, the principles of wave dispersion and attenuation are used as the defect indicators. The pitch-catch configuration is also known as through transmission. Because of the wide coverage capability of the pitch-catch method, it has been proposed by few researchers for inspecting the floor of fuel storage tanks, using lamb waves [bib_ref] Long-range Ultrasonic Non-destructive Testing of Fuel Tanks, Mažeika [/bib_ref]. The output signal received on the sensor is the exclusive source of information about the nature of the defect. It may encompass several changes in the characteristics of the induced signal after interaction with structural discontinuities. The error signal is obtained by taking the difference between undamaged and damaged signal. The highly advanced oscilloscopes (as shown in [fig_ref] Figure 6: Pitch-catch method. [/fig_ref] have the capability to carry out the basic system validation using advanced triggers, fast viewing modes, measurement parameters, or serial decodes. Additionally, the multi-domain analysis such as Fast Fourier transformation (FFT) and short-time Fourier transform (STFT) convert useful signal information from time domain into the easily interpretable frequency domain. These vigorous signal transfiguring functions are incorporated in these oscilloscopes (such as Lecroy wave-runners ® Xi series) for ease of signal processing, analysis, and interpretation.
## Research literature review on various aspects of ugws technology
Over the years, the acoustic waves in form of ultrasonic, acoustic emission (AE), laser vibrometry, and UGWs have been used immensely for SHM and NDE of metallic and composite The output signal received on the sensor is the exclusive source of information about the nature of the defect. It may encompass several changes in the characteristics of the induced signal after interaction with structural discontinuities. The error signal is obtained by taking the difference between undamaged and damaged signal. The highly advanced oscilloscopes (as shown in [fig_ref] Figure 6: Pitch-catch method. [/fig_ref] have the capability to carry out the basic system validation using advanced triggers, fast viewing modes, measurement parameters, or serial decodes. Additionally, the multi-domain analysis such as Fast Fourier transformation (FFT) and short-time Fourier transform (STFT) convert useful signal information from time domain into the easily interpretable frequency domain. These vigorous signal transfiguring functions are incorporated in these oscilloscopes (such as Lecroy wave-runners ® Xi series) for ease of signal processing, analysis, and interpretation.
## Research literature review on various aspects of ugws technology
Over the years, the acoustic waves in form of ultrasonic, acoustic emission (AE), laser vibrometry, and UGWs have been used immensely for SHM and NDE of metallic and composite The output signal received on the sensor is the exclusive source of information about the nature of the defect. It may encompass several changes in the characteristics of the induced signal after interaction with structural discontinuities. The error signal is obtained by taking the difference between undamaged and damaged signal. The highly advanced oscilloscopes (as shown in [fig_ref] Figure 6: Pitch-catch method. [/fig_ref] have the capability to carry out the basic system validation using advanced triggers, fast viewing modes, measurement parameters, or serial decodes. Additionally, the multi-domain analysis such as Fast Fourier transformation (FFT) and short-time Fourier transform (STFT) convert useful signal information from time domain into the easily interpretable frequency domain. These vigorous signal transfiguring functions are incorporated in these oscilloscopes (such as Lecroy wave-runners ® Xi series) for ease of signal processing, analysis, and interpretation.
## Research literature review on various aspects of ugws technology
Over the years, the acoustic waves in form of ultrasonic, acoustic emission (AE), laser vibrometry, and UGWs have been used immensely for SHM and NDE of metallic and composite structures [bib_ref] Structural health monitoring using scanning laser vibrometry: I. Lamb wave sensing, Staszewski [/bib_ref] [bib_ref] The use of Lamb waves for the long range inspection of large..., Cawley [/bib_ref] [bib_ref] Analysis of axis symmetric circular crested elastic wave generated during crack propagation..., Haider [/bib_ref] [bib_ref] Defect sensitivity of piezoelectric strip transducers based on planar Lamb waves, Soorgee [/bib_ref]. Apart from their applications in health prognostic studies, the GWs have been used for various other functions such as ice detection, de-icing, pigging, and antifouling operations in various isotropic and anisotropic structures [bib_ref] Progress on ultrasonic guided waves de-icing techniques in improving aviation energy efficiency, Wang [/bib_ref] [bib_ref] Characterization of Ultrasonic Wave Propagation in the Application of Prevention of Fouling..., Carellan [/bib_ref]. Although, under some specific circumstances, the reach and efficacy of GWs is confined by the contribution of environmental factors such as temperature, moisture and humidity, and external excitations of running machinery [bib_ref] Thermal sensitivity of Lamb waves for structural health monitoring applications, Dodson [/bib_ref] [bib_ref] Feature Extraction and Sensor Fusion for Ultrasonic Structural Health Monitoring Under Changing..., Lu [/bib_ref] [bib_ref] Ultrasonic guided waves for health monitoring of high-pressure composite tanks, Castaings [/bib_ref].
The GWs scatter at various angles after interaction with structural defects and this scattering phenomena is proportional to the severity/size of discontinuity. It is difficult to collect most of the dispersed signals by limited number of sensors. Therefore, the use of multiple detached transducers in various array configurations has become instrumental to investigate large structural areas with better accuracy [bib_ref] Embedded-ultrasonics structural radar for in situ structural health monitoring of thin-wall structures, Giurgiutiu [/bib_ref]. When a guided ultrasonic wave is stimulated with a crack, the excited input signal transforms and subsequently results in a difference of signal amplitude and phase shift between signal response of damaged and undamaged structures [bib_ref] Characterization of Aircraft Structural Damage using Guided Wave Based Finite Element Analysis..., Seshadri [/bib_ref]. The sensitivity of GWs towards different defects in the thickness of the structure depends on the type and location of the damage as well [bib_ref] Lamb wave based structural health monitoring, Raghavan [/bib_ref]. The structural discontinuities and damage in mechanical structures can be examined by quantifying the various changes in output signal, including E/M impedance signatures [bib_ref] Experimental investigation of E/M impedance health monitoring for spot-welded structural joints, Giurgiutiu [/bib_ref]. The precise collection of the time of flight (TOF) has been considered as an important step to localize the damaged region in a structure. In an experimental study, Dai and Herevealed that the performance of a time-frequency method called Wigner-Ville Distribution (WVD) was found superior to the other signal transforming methods (Hilbert envelope and Gabor wavelet transform) in estimating the precise TOF of excited signal based on energy distribution.
The GWUs have shown enormous potential in SHM of aviation based metallic and composite structures. These structures are prone to abrupt variation in ambient conditions of humidity level, ambient temperature, stress loading, and pressure that can instigate the corrosion and fatigue based structural cracks [bib_ref] Damage diagnosis and prognosis methodology to estimate safe load for aircraft structure, Seshadri [/bib_ref]. The traditional SHM techniques are labour-intensive and require considerable de-striping of assemblies and removal of surface protections for inspection. The frequent de-striping of machinery components may yield maintenance induced damages (MIDs). Therefore, the need for a more efficient, reliable, and optimized health monitoring system like GWUT is always a desirable element for operators and maintenance crew so that they can inspect a relatively large area with minimal disruption [bib_ref] Damage localization in plates using mode conversion characteristics of ultrasonic guided waves, Hosseini [/bib_ref].
The ice deposition on structures may compromise the efficacy of NDE and remote monitoring systems by warning for a false damage or missing the detection of actual defect failure. More recently, some selective modes of guided waves have been used successfully for detection of ice on various metallic/composite structures such as wind turbine blades, aero-structures, helicopter rotors and ship's hull. The icing phenomena is a critical issue in the lube oil/fuel circuits of aircrafts and their aerodynamic efficiency and maneuverability. Some researchers used GW modes to differentiate water and ice content as well as deicing on structures in cold climates. The guided waves produced by a piezoelectric transducer (bonded/embedded on a surface) can be used to induce shear stresses to delaminate/fracture ice deposits on various structural surfaces [bib_ref] Development of a de-icing system for aerodynamic surfaces based on ultrasonic waves, Maio [/bib_ref]. Displacement controlling of in-plane/out-of-plane wave modes excited on the outer surface of structure can be helpful to detect water or ice layers.
In an ice sensing experiment, Zhao and Rose [bib_ref] Ultrasonic guided wave tomography for ice detection, Zhao [/bib_ref] deliberated the behaviour of S 0 mode of Lamb wave to differentiate between the water and ice layers on aluminium plate (1.6-mm thick). A circular array of 16 sensors were used with an exciting frequency of 350 KHz. The S 0 mode having dominant in-plane displacement on metal surface was found to be insensitive to the water layer. However, it remained very sensitive towards the ice loading. Later, the probability-based reconstruction (PRA) algorithm was used to construct the ice images in other types of material such as titanium and carbon-fiber-reinforced titanium plates. In case, ships maneuvering in the polar region, the amalgamated ice layers and water disturbances may cause variation in the signal response of GWs and the ice detection may be misleading the damage detection process. Memmolo et al. [bib_ref] Engineering, I. Guided wave propagation and interaction with ice layers in marine..., Memmolo [/bib_ref] conducted a numerical based study (finite element analysis with ABAQUS ® ) for ice detection on ship structures using both pitch-catch and pulse-echo methods. It was revealed that the back-scattering in the pulse-echo signal and forward scattering in the pitch-catch configuration are highly dependent on the length of the ice layer. This study proposed that it is quite much possible to quantify the dimensions of the ice layer.
In the cold countryside, icing of wind turbine structures (especially blades) is a common problem, which may hinder its operation. The ice layers on the surface of the blades can change the propagation characteristics of the acoustic waves used for condition monitoring. Wang et al. [bib_ref] Ice monitoring of a full-scale wind turbine blade using ultrasonic guided waves..., Wang [/bib_ref] investigated icing and temperature effect on wind turbine blades under extremely cold climates by a simulation carried out in a frozen tunnel. Three parameters, including amplitude, group velocity, and spectrum value of Lamb wave were used as indicators. The amplitude and spectrum of S 0 wave mode were found to be more sensitive to the variations in the thickness of the ice loading than the temperature factor, whereas the group velocity of S 0 mode was deeply influenced by the changes in both the temperature and the thickness of ice layer. Complexities in the estimation of ice formation and temperature variation were addressed effectively using a principal component analysis (PCA) based ice monitoring method.
During a research study on the use of antisymmetric (flexural) cylindrical GWs in the underwater pipelines, Na and Kunduused a new transducer coupler and found it to be effective in identifying the presence of defects in butt welds, anti-corrosion covers, insulation and modest portions of soil/concrete embedded structures. However, some limitations were observed when allowed to pass through flanges and long embedded pipes. In the health monitoring of pipelines, the detection and quantification of wall thinning process is mainly focused which is caused by the internal or external corrosion. The external corrosion is generally found in areas where water is entrapped, which damages surface coatings, wall thinning and pitting corrosion [bib_ref] Best practice for the assessment of defects in pipelines-Corrosion, Cosham [/bib_ref]. The sensitivity of GWs towards corrosion defect would be different than that of a crack damage. The reflected part of the signal generated in case of corrosion defect will be comparatively smaller but the TOF shift would be different depending upon the severity of the corrosion loss and variation in the group velocity of the propagated wave. The research study of Gao and Rose [bib_ref] Goodness dispersion curves for ultrasonic guided wave based SHM: A sample problem..., Gao [/bib_ref] revealed that the variation in TOF is a function of corroded section on the structure. The reflected and transmitted coefficients of GW signal has been used to detect accumulations of sludge content in pipelines [bib_ref] Feasibility study of sludge and blockage detection inside pipes using guided torsional..., Ma [/bib_ref].
Apart from use of GWUT in structures, Tse and Rostami [bib_ref] Advanced signal processing methods applied to guided waves for wire rope defect..., Tse [/bib_ref] demonstrated the application of this technique to defect assessment in wire rope cross-section, using magnetostrictive UGWs. The output signals were investigated using a Short Time Fourier Transform (STFT) as well as Wavelet analysis, and the experimental results were able to locate and identify the severity of imperfections. For the ease of signal interpretation, modulated pulses of GWs with short tone burst are generally preferred in experimental studies [bib_ref] Structural health monitoring using scanning laser vibrometry: I. Lamb wave sensing, Staszewski [/bib_ref]. Some of these modulations such as rectangular and square modulated signal may not be feasible as they can hamper the signal processing by excitation of the larger spectrum and additional modes. The Gaussian-modulated waves have been preferred for signal excitation by some researchers [bib_ref] Response of Gaussian-modulated guided wave in aluminum: An analytical, numerical, and experimental..., Khan [/bib_ref] [bib_ref] Piezoelectric transducer parameter selection for exciting a single mode from multiple modes..., Zhang [/bib_ref]. Cawley and Alleyne [bib_ref] The use of Lamb waves for the long range inspection of large..., Cawley [/bib_ref] reported that in order to generate a pure mode of propagation, it is essential to regulate the bandwidth of excitation in both the frequency and wavenumber domains, which can be best achieved by using a short tone burst of Gaussian or Hanning window, in case of frequency. A limited cycle sinusoidal tone burst reduced undesired reflections between energy packets [bib_ref] Damage diagnosis and prognosis methodology to estimate safe load for aircraft structure, Seshadri [/bib_ref].
Time-shift average algorithms were used by Michaels and Michaels [bib_ref] Guided wave signal processing and image fusion for in situ damage localization..., Michaels [/bib_ref] in differential filtered signals to acquire numerous images of the same structure. These images were then merged to enhance the resolution of defect localization and signal to noise ratio (SNR). The increasing distance along the plate, however, results in rapid reduction of SNR. The structural cuts and impact damage on the trailing edge of the aluminium slat were investigated by Senyurek [bib_ref] Detection of cuts and impact damage at the aircraft wing slat by..., Senyurek [/bib_ref] using Hilbert transform. It was revealed that the cuts created new boundaries and the reflected waves were found to be helpful in estimating the location of the defects. The intricacy of reflected or transmitted signals may be moderated by using digital bandpass filtering hence any variation in filtered signals can be examined to identify and localize damages area of structures [bib_ref] Guided wave signal processing and image fusion for in situ damage localization..., Michaels [/bib_ref].
The advancement of sensor equipment over the years has benefitted SHM/NDE techniques and assisted in the optimization of maintenance/RUL prediction for structures [bib_ref] Piezoelectric wafer active sensors in lamb wave-based structural health monitoring, Yu [/bib_ref]. The importance of CBM has enhanced in the modern-day maintenance due to developments in sensor technology. Conventional preventive, proactive and reliability-based maintenance strategies have been revamped with the addition of health monitoring data in overall decision-making for future maintenance actions and inspection intervals [bib_ref] A review on machinery diagnostics and prognostics implementing condition-based maintenance, Jardine [/bib_ref]. The GWUT can be an extremely useful addition in the CBM for structural health inspection and overall maintenance planning of the mega-structures. The variations in amplitude, and travel time of GWs have been used by various researchers for identification and location of structural cracks in aging structures and found it as an effective method for structural integrity inspections the [bib_ref] Detection of cracks at rivet holes using guided waves, Fromme [/bib_ref] [bib_ref] Scattering of Lamb waves from a rivet hole with edge cracks, Chang [/bib_ref]. The study by Alleyne and Cawley [bib_ref] The Interaction of Lamb Waves with Defects, Alleyne [/bib_ref] revealed that the elastic propagation of Lamb waves on surface cracks of the metallic plate is reliant on the wave mode (symmetric or antisymmetric), its mode order, frequency-depth (f-d) product and the crack geometry. In a similar research study, Hall et al. [bib_ref] Guided wave damage characterization via minimum variance imaging with a distributed array..., Hall [/bib_ref] reported that the interaction of GWs with a surface defect is different from a sub-surface defect and it is primarily dependent on the defect size, angle of incidence, and excitation frequency.
Yu et al. [bib_ref] Crack imaging and quantification in aluminum plates with guided wave wavenumber analysis..., Yu [/bib_ref] used UGWs for quantification of cracks in aluminium structures using an advanced wave field analysis methods. The A 0 mode exhibited strong out of phase motion than S 0 mode. It was noticed that the transmission index was found stronger than the reflection index in half-through-thickness cracks than full through-thickness cracks. This effect was attributed to the fact that the metal portion underneath the half-thickness crack allows transmission of forwarding wave. The material impedance is another factor for deciding transmission and reflection of the acoustic signal. The effect of defect orientation (metal surfaces) on reflection and transmission coefficients of GWs has been described in. It has been revealed that the reflection coefficient reduces whereas the transmission coefficient rises with an increase in the angle (θ) of the incidence wave with respect to notch, hence depicting a strong angular dependence of the scattered Lamb waves. Conversely, with the same angle θ, the transmission coefficient declines with the notch length, while the reflection coefficient increases up to 60 mm of the notch length. Zima and Rucka [bib_ref] Guided waves for monitoring of plate structures with linear cracks of variable..., Zima [/bib_ref] conducted experiment-based research assisted by numerical simulation for quantification of various sizes of cracks on a steel plate. A sine modulated Hanning widow was used for excitation of Lamb wave and a binary damage imaging algorithm was applied to identify the linearly crafted cracks with a greater accuracy.
The GWUT has been able to diagnose cracks or CUIs in the bends/elbows of pipelines. Only a limited portion of the wrapped material is required to be removed for placement of sensors and defect diagnosis process is conducted without removal of entire material [bib_ref] Long-range ultrasonic testing (LRUT) of pipelines and piping, Mudge [/bib_ref]. The axial cracks in carbon steel pipelines were examined by Qi et al.using Shear wave. A good agreement is reported between experimental work and finite-element modeling (FEM) analyses. Highest detection sensitivity with fewer mode conversions was recorded in the middle area of elbow than in the intrados and extrados regions. The flexural mode of guided waves have been preferred for use in pipeline inspections, as the scattering pattern of these waves can be vital to differentiate between circular holes, cracks located at different orientations and slanting angles [bib_ref] Scattering of torsional flexural guided waves from circular holes and crack-like defects..., Zhang [/bib_ref]. The authors of this paper believe that the research of Zhang et al. [bib_ref] Opportunistic maintenance for wind turbines considering imperfect, reliability-based maintenance, Zhang [/bib_ref] can be extended for identification of shape and orientation of structural defects in thin plate type structures using Lamb GWs. In this regard, the research studies of Yu et al. [bib_ref] Crack imaging and quantification in aluminum plates with guided wave wavenumber analysis..., Yu [/bib_ref] and Su and Yeprovided good grounds for identification of structural defects placed at various angles, shapes, and orientations in plate-like metallic structures.
In the oil and gas industry, the underground and subsea pipelines are more prone to general corrosion, fatigue, stress cracking and pitting [bib_ref] Corrosion induced failure analysis of subsea pipelines, Yang [/bib_ref]. Most of the times, these huge pipe networks carry petroleum fluids at high pressure/temperature and are exposed to extreme environmental sea conditions from the external side [bib_ref] Risk-based maintenance planning of subsea pipelines through fatigue crack growth monitoring, Arzaghi [/bib_ref] [bib_ref] Condition monitoring of subsea pipelines considering stress observation and structural deterioration, Chen [/bib_ref]. Apart from piping material properties, the degradation of these marine based structures depends on various physical, chemical and biological compositions of seawater [bib_ref] Internal corrosion hazard assessment of oil & gas pipelines using Bayesian belief..., Shabarchin [/bib_ref]. Generally, flaw detection of underground or subsea pipes is carried out by in-line inspection (ILI) tools, also referred to as smart or intelligent pigs, working on the principles of magnetic flux leakage (MFL) or conventional ultrasonic inspection technique. Various limitations are involved in the use of these methods. The former technique has limitations to discriminate among various severity levels of defect and their position (external or internal surface). The latter technique can be extremely cumbersome and expensive in large pipe networks and difficult positions. Therefore, use of GWs has been proposed by Ángela et al. [bib_ref] The benefits of long range ultrasonic sensors for the efficient InLine inspection..., Angulo [/bib_ref] for the pigging system, which has more reliable, longer area coverage with limit sensors, permanent and highly efficient. The standard practice ASTM E2775and Clough et al. [bib_ref] Circumferential guided wave EMAT system for pipeline screening using shear horizontal ultrasound, Clough [/bib_ref] have also supported the plan to use GWUT for screening of defects in pipelines. [fig_ref] Figure 8: GWs inspection of a metallic pipeline structures[96] [/fig_ref] depicts GWs propagation in metallic pipes.
Sensors 2018, 18, x FOR PEER REVIEW 12 of 26 more reliable, longer area coverage with limit sensors, permanent and highly efficient. The standard practice ASTM E2775and Clough et al. [bib_ref] Circumferential guided wave EMAT system for pipeline screening using shear horizontal ultrasound, Clough [/bib_ref] have also supported the plan to use GWUT for screening of defects in pipelines. [fig_ref] Figure 8: GWs inspection of a metallic pipeline structures[96] [/fig_ref] depicts GWs propagation in metallic pipes. In a non-contact approach (laser-based), a pulsed laser is used to generate elastic waves and a laser Doppler vibrometer or interferometers are used for collection and detection of structural discontinuities [bib_ref] Laser Doppler velocimetry and PZT sensing for the study of guided waves..., Maio [/bib_ref]. The Fabry-Pérot and heterodyne interferometers are often used for contactless acquisition of guided waves. More recently, various defect-imaging techniques have been developed and the characteristics of low-frequency GWs have been used to obtain images of defects. The non-contact methods can also be used in both plate type and tubular structures. Defect imaging of the thinning process on the aluminium pipeline has been described in [bib_ref] Non-contact imaging of pipe thinning using elastic guided waves generated and detected..., Hayashi [/bib_ref]. The high-power fiberlaser and a laser Doppler vibrometer (LDV) were used for generation and detection of flexural elastic waves (modulated frequency 22 kHz to 36 kHz). A semi-analytical finite element method (SAFEM) was used for phase velocity dispersion curves and the frequency image averaging (FIA) to minimize the undesirable circumferential resonance patterns. A fast imaging method (E-camera) was used for imaging of metal loss at all positions of a straight and branched pipes located at a distance of 2.6-6 m from E-camera.
In comparison with non-contact methods, some researchers believe that the contact based techniques have some limitations as follows [bib_ref] Complete noncontact laser ultrasonic imaging for automated crack visualization in a plate, An [/bib_ref] :
- Guided wave excitation and sensing is limited (by contact method) to some discrete points, therefore it is hard to achieve a high resolution to sense minor emerging defects; - High cost for transducer, assemblies, installation, and labour cost. The vulnerability for equipment damage is high, leading to high maintenance budgets; - Limitation for use of contact transducers under harsh environmental conditions and application to convoluted/high speed rotating surfaces such as high temperature/pressure, radioactive surfaces, rail wheels.
The laser ultrasonic excitation can be performed by using pulse lasers, continuous lasers, or laser interferometry. The non-contact method has been preferably used in crack/corrosion detection, quantification, and localization of structural defects in various areas [bib_ref] Noncontact monitoring of immersed plates by means of laser-induced ultrasounds, Pistone [/bib_ref] [bib_ref] Fully noncontact wave propagation imaging in an immersed metallic plate with a..., Lee [/bib_ref]. It can provide high spatial resolution hence damage diagnosis can be achieved more easily. Further, non-contact techniques are reported to be economical, easily applicable to harsh environments with little maintenance. The laser-based ultrasonic propagation imager (UPI) with piezoelectric sensors have been used for signal excitation and reception by Park and Lee [bib_ref] Application of the ultrasonic propagation imaging system to an immersed metallic structure..., Park [/bib_ref] during inspection of submerged metallic structures, with a 2-mm crack. The employed method was successful in quantifying the defects on fully submerged surfaces in stationary/oscillating waters and non-submerged structures. It has been revealed that the measured crack size for both submerged cases increased slightly than actual crack size and the crack locations were observed to be slightly tilted due to refraction/change in impedance. The presence of liquid on one or both sides of the steel plate was investigated by Wilcox et al. [bib_ref] Omnidirectional guided wave inspection of large metallic plate structures using an EMAT..., Wilcox [/bib_ref] using EMAT in which slight attenuation (1 dB/m) of the S0 wave mode was observed, Actuators and sensors for tube inspection GWs transmitting in either direction In a non-contact approach (laser-based), a pulsed laser is used to generate elastic waves and a laser Doppler vibrometer or interferometers are used for collection and detection of structural discontinuities [bib_ref] Laser Doppler velocimetry and PZT sensing for the study of guided waves..., Maio [/bib_ref]. The Fabry-Pérot and heterodyne interferometers are often used for contactless acquisition of guided waves. More recently, various defect-imaging techniques have been developed and the characteristics of low-frequency GWs have been used to obtain images of defects. The non-contact methods can also be used in both plate type and tubular structures. Defect imaging of the thinning process on the aluminium pipeline has been described in [bib_ref] Non-contact imaging of pipe thinning using elastic guided waves generated and detected..., Hayashi [/bib_ref]. The high-power fiber-laser and a laser Doppler vibrometer (LDV) were used for generation and detection of flexural elastic waves (modulated frequency 22 kHz to 36 kHz). A semi-analytical finite element method (SAFEM) was used for phase velocity dispersion curves and the frequency image averaging (FIA) to minimize the undesirable circumferential resonance patterns. A fast imaging method (E-camera) was used for imaging of metal loss at all positions of a straight and branched pipes located at a distance of 2.6-6 m from E-camera.
In comparison with non-contact methods, some researchers believe that the contact based techniques have some limitations as follows [bib_ref] Complete noncontact laser ultrasonic imaging for automated crack visualization in a plate, An [/bib_ref] :
## -
Guided wave excitation and sensing is limited (by contact method) to some discrete points, therefore it is hard to achieve a high resolution to sense minor emerging defects; - High cost for transducer, assemblies, installation, and labour cost. The vulnerability for equipment damage is high, leading to high maintenance budgets; - Limitation for use of contact transducers under harsh environmental conditions and application to convoluted/high speed rotating surfaces such as high temperature/pressure, radioactive surfaces, rail wheels.
The laser ultrasonic excitation can be performed by using pulse lasers, continuous lasers, or laser interferometry. The non-contact method has been preferably used in crack/corrosion detection, quantification, and localization of structural defects in various areas [bib_ref] Noncontact monitoring of immersed plates by means of laser-induced ultrasounds, Pistone [/bib_ref] [bib_ref] Fully noncontact wave propagation imaging in an immersed metallic plate with a..., Lee [/bib_ref]. It can provide high spatial resolution hence damage diagnosis can be achieved more easily. Further, non-contact techniques are reported to be economical, easily applicable to harsh environments with little maintenance. The laser-based ultrasonic propagation imager (UPI) with piezoelectric sensors have been used for signal excitation and reception by Park and Lee [bib_ref] Application of the ultrasonic propagation imaging system to an immersed metallic structure..., Park [/bib_ref] during inspection of submerged metallic structures, with a 2-mm crack. The employed method was successful in quantifying the defects on fully submerged surfaces in stationary/oscillating waters and non-submerged structures. It has been revealed that the measured crack size for both submerged cases increased slightly than actual crack size and the crack locations were observed to be slightly tilted due to refraction/change in impedance. The presence of liquid on one or both sides of the steel plate was investigated by Wilcox et al. [bib_ref] Omnidirectional guided wave inspection of large metallic plate structures using an EMAT..., Wilcox [/bib_ref] using EMAT in which slight attenuation (1 dB/m) of the S 0 wave mode was observed, whereas the attenuation of A 0 mode found to be around 57 dB/m. In a similar research, An et al. [bib_ref] Complete noncontact laser ultrasonic imaging for automated crack visualization in a plate, An [/bib_ref] used laser ultrasonic scanning system and a novel image processing technique to visualize crack on aluminium plates and found it a reliable defect characterization method.
Rao et al.reported that the flaw detection in submerged structures using UGW is rather a more challenging because of the leaking wave energy into the liquid. In an experiment based research study on corrosion monitoring and prediction of wall thickness loss, Fromme [bib_ref] Corrosion monitoring using high-frequency guided waves, Fromme [/bib_ref] used transducer based GWs excitation and output signal response was measured through a laser interferometer. A good agreement between practical and theoretical predictions was achieved. The crack detection in submerged structure with a fully non-contact method has been reported by Lee et al. [bib_ref] Fully noncontact wave propagation imaging in an immersed metallic plate with a..., Lee [/bib_ref] in which image blurriness was found to be more dominant factor, in submerged condition. A schematic diagram of a complete non-contact laser ultrasonic scanning system (excitation and sensing) is illustrated in [fig_ref] Figure 9: A fully non-contact laser ultrasonic scanning system [/fig_ref]. whereas the attenuation of A0 mode found to be around 57 dB/m. In a similar research, An et al. [bib_ref] Complete noncontact laser ultrasonic imaging for automated crack visualization in a plate, An [/bib_ref] used laser ultrasonic scanning system and a novel image processing technique to visualize crack on aluminium plates and found it a reliable defect characterization method. Rao et al.reported that the flaw detection in submerged structures using UGW is rather a more challenging because of the leaking wave energy into the liquid. In an experiment based research study on corrosion monitoring and prediction of wall thickness loss, Fromme [bib_ref] Corrosion monitoring using high-frequency guided waves, Fromme [/bib_ref] used transducer based GWs excitation and output signal response was measured through a laser interferometer. A good agreement between practical and theoretical predictions was achieved. The crack detection in submerged structure with a fully non-contact method has been reported by Lee et al. [bib_ref] Fully noncontact wave propagation imaging in an immersed metallic plate with a..., Lee [/bib_ref] in which image blurriness was found to be more dominant factor, in submerged condition. A schematic diagram of a complete non-contact laser ultrasonic scanning system (excitation and sensing) is illustrated in [fig_ref] Figure 9: A fully non-contact laser ultrasonic scanning system [/fig_ref]. The pulse-echo technique was used by Djili et al. [bib_ref] Notch detection in copper tubes immersed in water by leaky compressional guided..., Djili [/bib_ref] for detection of artificially crafted notches on submerged copper tubes. A single transducer was used for excitation and reception of leaky GWs and notches of various small sizes were detected with a greater accuracy. The response of GWs in aluminium-based aero structures was studied by Zhao et al. [bib_ref] Practical long range guided wave inspection-Applications to pipes and rail, Cawley [/bib_ref] and several complications were observed in wave propagation over a longer distance due to strong attenuation from surface paint and riveting. A novel algorithm and correlation analysis technique called Reconstruction Algorithm for Probabilistic Inspection of Defects (RAPID) was adopted for collection of output signal from rivet cracks and corrosion patches. This technique was found efficient in detection, size estimation and localization of defects.
A predictive numerical based study was conducted in a steel plate using acoustic emission (AE) method. It was revealed that peak amplitudes of antisymmetric (A0) signal is greater than that of symmetric (S0) signal [bib_ref] A Helmholtz Potential Approach to the Analysis of Guided Wave Generation During..., Haider [/bib_ref]. They also observed that the peak amplitude of bulk wave was found to be less significant than that of the GWs peak amplitudes. Because of some special characteristics of A0 mode of GWs (i.e., out-off-the-plate displacements) they have been used as a preventive technique The pulse-echo technique was used by Djili et al. [bib_ref] Notch detection in copper tubes immersed in water by leaky compressional guided..., Djili [/bib_ref] for detection of artificially crafted notches on submerged copper tubes. A single transducer was used for excitation and reception of leaky GWs and notches of various small sizes were detected with a greater accuracy. The response of GWs in aluminium-based aero structures was studied by Zhao et al. [bib_ref] Practical long range guided wave inspection-Applications to pipes and rail, Cawley [/bib_ref] and several complications were observed in wave propagation over a longer distance due to strong attenuation from surface paint and riveting. A novel algorithm and correlation analysis technique called Reconstruction Algorithm for Probabilistic Inspection of Defects (RAPID) was adopted for collection of output signal from rivet cracks and corrosion patches. This technique was found efficient in detection, size estimation and localization of defects.
A predictive numerical based study was conducted in a steel plate using acoustic emission (AE) method. It was revealed that peak amplitudes of antisymmetric (A 0 ) signal is greater than that of symmetric (S 0 ) signal [bib_ref] A Helmholtz Potential Approach to the Analysis of Guided Wave Generation During..., Haider [/bib_ref]. They also observed that the peak amplitude of bulk wave was found to be less significant than that of the GWs peak amplitudes. Because of some special characteristics of A 0 mode of GWs (i.e., out-off-the-plate displacements) they have been used as a preventive technique for bio-fouling on ship structures [bib_ref] Characterization of Ultrasonic Wave Propagation in the Application of Prevention of Fouling..., Carellan [/bib_ref]. Carvalho et al. [bib_ref] On the reliability of an automated ultrasonic system for hull inspection in..., Carvalho [/bib_ref] used the ultrasounic automated system to inspect corrosion induced superficial defects in marine/ship structures. The artificial neural networks (ANN) has been used by some researchers as a reliable method in automatic ultrasonic testing and defect detection in pipelines [bib_ref] Reliability of non-destructive test techniques in the inspection of pipelines used in..., Carvalho [/bib_ref].
Recently, Martinez et al.presented a numerical modelling approach to study the combined effect of geometry complexity and residual stresses produced during the fatigue crack growth (FCG), using the lamb waves. Successful interaction of GWs with a crack size of 2.1 mm on aluminium structure has been reported. The drop in amplitude (due to attenuation) was recorded to be more significant in A 0 wave mode than the S 0 mode on the higher frequency. Fromme and Sayir [bib_ref] Detection of cracks at rivet holes using guided waves, Fromme [/bib_ref] observed the scattering effect of A 0 wave mode around rivet holes, notch cut and fatigue cracks in an aluminium plate. The piezo transducers were used for excitation of antisymmetric Lamb GWs and its scatter measure was taken with the help of laser interferometer. The field results depicted significant changes with the presence of a notch or a crack, which were much smaller than the wavelength of excited GWs.
In addition to the uses of Lamb guided waves, the review of the literature revealed that the S H and Rayleigh waves are also used extensively in structural diagnostic studies of plate type and tubular shaped metallic/composite structures [bib_ref] Shear horizontal wave excitation and reception with shear horizontal piezoelectric wafer active..., Kamal [/bib_ref]. As during propagation in an elastic medium, S H wave is polarized parallel to a surface, so they reflect without mode conversion. Additionally, during this process the amplitude changes are insignificant, therefore, S H wave is considered to be superlative for crack detection in welds [bib_ref] Improvement of SH-wave EMAT phased array inspection by new eight segment probes, Sawaragi [/bib_ref] [bib_ref] Ultrasonic inspection of stainless steel butt welds using horizontally polarized shear waves, Fortunko [/bib_ref] [bib_ref] EMAT pipe inspection with guided waves, Salzburger [/bib_ref] [bib_ref] Three-Dimensional Scattering of an Incident Plane Shear Horizontal Guided Wave by a..., Zhu [/bib_ref]. It has distinct capability of travelling across bended surfaces with minimal energy losses [bib_ref] Shear horizontal (SH) ultrasound wave propagation around smooth corners, Petcher [/bib_ref]. The S H wave with EMAT was found effective in defect identification in a 7 mm thick carbon steel sample. It was found to be highly reliable in defect detection, axial sizing and approximation of circumferential positioning of a defect for both artificial cracks and corrosion defect, even with a thick paint coating [bib_ref] Circumferential guided wave EMAT system for pipeline screening using shear horizontal ultrasound, Clough [/bib_ref]. The reduction of wall thickness of corroded pipelines has been quantified by Andruschak et al. [bib_ref] An NDT guided wave technique for the identification of corrosion defects at..., Andruschak [/bib_ref] using S H waves. The defect location and quantification for corrosion loss was clearly established from high amplitude reduction and delayed arrival timings of received signal.
In ships and offshore marine platforms, accelerated rates of corrosion based deterioration has been reported, primarily due to the harsh seawater climatic conditions. The general corrosion and pitting occur more frequently. Apart from these, crevice, galvanic, fretting, groove, fatigue, edge and stress corrosion are also encountered in various loaded and dynamic loaded structures [bib_ref] Ultimate strength analysis of aged steel-plated structures exposed to marine corrosion damage:..., Wang [/bib_ref]. Fromme et al. [bib_ref] High-frequency guided ultrasonic waves to monitor corrosion thickness loss, Fromme [/bib_ref] studied the corrosion based thickness loss of marine steel plates, partly immersed in salt water. A partially noncontact method was used in which the Lamb waves (A 0 and S 0 ) were simultaneously excited using wedge transducer (at a center frequency of 450 KHz) and the reflected signal was collected with a laser interferometer. The change in the response signal due to the wave mode interference was measured and the wall thinning was monitored. A good agreement was found between experimental results and theoretical estimation. In partial non-contact method, the contact less mechanism can be used for either excitation or sensing. This experimental setup is shown in [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref]. The SHM methods face various impediments to distinguish the variations due to structural defects and those occur by transitions in environmental and operational conditions. The active SHM method of UGWs is built on the principle of the difference between the most recent time-trace (with damage) and a baseline time-trace recorded in absence of structural damage. The residual signal after the subtraction of a baseline time-trace is referred to as a remaining time-trace which can be effected by the significant changes in environmental factors, particularly the temperature, and surface conditions [bib_ref] Quantification of environmental compensation strategies for guided wave structural health monitoring, Croxford [/bib_ref] [bib_ref] A methodology for structural health monitoring with diffuse ultrasonic waves in the..., Lu [/bib_ref]. These variations may adversely affect the efficiency of the subtraction process and the residual time-trace may be amalgamated with the coherent noise signal because of the shift in arrival timing of wave-packets from structural topographies. This degradation occurs due to the change in the velocity of acoustic waves and thermal expansion of structure [bib_ref] Reliability analysis of structural health monitoring systems, Etebu [/bib_ref]. In order to minimize the effect of temperature variations on the performance of GWUT, some researchers have proposed compensatory strategies such as optimal baseline selection (OBS), baseline signal stretch (BSS) and a combination of both [bib_ref] Reliability analysis of structural health monitoring systems, Etebu [/bib_ref] [bib_ref] A methodology for structural health monitoring with diffuse ultrasonic waves in the..., Lu [/bib_ref] [bib_ref] Guided wave health monitoring of complex structures by sparse array systems: Influence..., Clarke [/bib_ref] [bib_ref] Guided waves for autonomous online identification of structural defects under ambient temperature..., Moll [/bib_ref]. In [bib_ref] Reliability analysis of structural health monitoring systems, Etebu [/bib_ref] , the authors recommended that these compensatory techniques are also valid for numerous other climatic variations, provided they tend to result in a homogenous and isotropic change in the velocity of the guided wave. Some researchers have reported that the minor temperature change (up to 40 °C) does not affect the performance of GWs therefore special compensation techniques may not be required, except in cases where the structural damage to be sensed is very small in size. The research work of Croxford et al. [bib_ref] Strategies for overcoming the effect of temperature on guided wave structural health..., Croxford [/bib_ref] revealed that signal fluctuations even due to minor changes in temperature is enough to generate significant coherent noise (−25 dB), rendering the use of compensatory measures. The variation in TOF of guided waves with the increase in temperatures (20-150 °C) has been monitored in [bib_ref] Effects of elevated temperature on guided-wave structural health monitoring, Raghavan [/bib_ref]. A small error in location of defect was recorded till 80 °C, however, a significant decrease in sensitivity was recorded beyond this temperature.
During a research study on SHM of rail tracks, Moustakidis et al. [bib_ref] An intelligent methodology for railways monitoring using ultrasonic guided waves, Moustakidis [/bib_ref] presented that the variation in ambient temperature and humidity can mask changes in the output signal due to structural damage. It was also noticed that certain factors such as rain and high temperature play a significant role in deteriorating the damage detection capability of GWs used with contact method. [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref] shows the experimental setup of this research. Some researchers have reported that the minor temperature change (up to 40 - C) does not affect the performance of GWs therefore special compensation techniques may not be required, except in cases where the structural damage to be sensed is very small in size. The research work of Croxford et al. [bib_ref] Strategies for overcoming the effect of temperature on guided wave structural health..., Croxford [/bib_ref] revealed that signal fluctuations even due to minor changes in temperature is enough to generate significant coherent noise (−25 dB), rendering the use of compensatory measures. The variation in TOF of guided waves with the increase in temperatures (20-150 - C) has been monitored in [bib_ref] Effects of elevated temperature on guided-wave structural health monitoring, Raghavan [/bib_ref]. A small error in location of defect was recorded till 80 - C, however, a significant decrease in sensitivity was recorded beyond this temperature.
During a research study on SHM of rail tracks, Moustakidis et al. [bib_ref] An intelligent methodology for railways monitoring using ultrasonic guided waves, Moustakidis [/bib_ref] presented that the variation in ambient temperature and humidity can mask changes in the output signal due to structural damage. It was also noticed that certain factors such as rain and high temperature play a significant role in deteriorating the damage detection capability of GWs used with contact method. [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref] shows the experimental setup of this research. decrease in sensitivity was recorded beyond this temperature.
During a research study on SHM of rail tracks, Moustakidis et al. [bib_ref] An intelligent methodology for railways monitoring using ultrasonic guided waves, Moustakidis [/bib_ref] presented that the variation in ambient temperature and humidity can mask changes in the output signal due to structural damage. It was also noticed that certain factors such as rain and high temperature play a significant role in deteriorating the damage detection capability of GWs used with contact method. [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref] shows the experimental setup of this research. [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref]. GWUT transducer mounting for rail foot inspections (adapted from [bib_ref] An intelligent methodology for railways monitoring using ultrasonic guided waves, Moustakidis [/bib_ref].
Performance of transducers is crucial in the GWUT especially under variant operating conditions. This performance is highly influenced by several external factors such as environmental dynamics, high operating temperatures in particular. As the general purpose piezoelectric sensors [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref]. GWUT transducer mounting for rail foot inspections (adapted from [bib_ref] An intelligent methodology for railways monitoring using ultrasonic guided waves, Moustakidis [/bib_ref].
Performance of transducers is crucial in the GWUT especially under variant operating conditions. This performance is highly influenced by several external factors such as environmental dynamics, high operating temperatures in particular. As the general purpose piezoelectric sensors used in UGWS have low curie temperatures, therefore their signal reception performance deteriorates significantly in higher temperatures. The intrinsic E/M impedance, admittance signatures of PWAS degrades when exposed to higher temperatures. In these operating conditions, designed piezoelectric transducers (lithium niobate) are preferred, especially in high temperature operating conditions such as nuclear reactors, boilers and gas turbines [bib_ref] Characterization of Aircraft Structural Damage using Guided Wave Based Finite Element Analysis..., Seshadri [/bib_ref].
In some cases, non-contact GWs methods have also been preferred for SHM and NDE at elevated temperatures [bib_ref] Non-contact imaging of pipe thinning using elastic guided waves generated and detected..., Hayashi [/bib_ref]. Several researchers have established the passive SHM concept on rail tracks, where the dynamic excitation signature is triggered by the external source (i.e., rolling wheels of a train) and non-contact air-coupled ultrasonic receivers may be placed underneath a testing railcar for rail examination. An experimental study to assess the performance of the passive ultrasonic inspection of rail has been demonstrated in [bib_ref] Passive Extraction of Dynamic Transfer Function From Arbitrary Ambient Excitations: Application to..., Lanza [/bib_ref] , at train speeds up to 80 mph. The problems such as highly variable energy excitation from rotating train wheels were resolved using the deconvolution operator and Green's function. Upon successful completion of field tests, this innovative concept is expected to perform efficient health inspections of rail over longer lengths, with a minimal or no disruption of regular traffic. A passive inspection mechanism is illustrated in [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref]. used in UGWS have low curie temperatures, therefore their signal reception performance deteriorates significantly in higher temperatures. The intrinsic E/M impedance, admittance signatures of PWAS degrades when exposed to higher temperatures. In these operating conditions, designed piezoelectric transducers (lithium niobate) are preferred, especially in high temperature operating conditions such as nuclear reactors, boilers and gas turbines [bib_ref] Characterization of Aircraft Structural Damage using Guided Wave Based Finite Element Analysis..., Seshadri [/bib_ref].
In some cases, non-contact GWs methods have also been preferred for SHM and NDE at elevated temperatures [bib_ref] Non-contact imaging of pipe thinning using elastic guided waves generated and detected..., Hayashi [/bib_ref]. Several researchers have established the passive SHM concept on rail tracks, where the dynamic excitation signature is triggered by the external source (i.e., rolling wheels of a train) and non-contact air-coupled ultrasonic receivers may be placed underneath a testing railcar for rail examination. An experimental study to assess the performance of the passive ultrasonic inspection of rail has been demonstrated in [bib_ref] Passive Extraction of Dynamic Transfer Function From Arbitrary Ambient Excitations: Application to..., Lanza [/bib_ref] , at train speeds up to 80 mph. The problems such as highly variable energy excitation from rotating train wheels were resolved using the deconvolution operator and Green's function. Upon successful completion of field tests, this innovative concept is expected to perform efficient health inspections of rail over longer lengths, with a minimal or no disruption of regular traffic. A passive inspection mechanism is illustrated in [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref].
## Discussion and proposed methodology for future research
The role of health monitoring in operational availability, upkeep, and maintenance of equipment and structures has increased exponentially in recent years. Conventional maintenance procedures such as corrective and preventive are transforming rapidly towards technology-driven upkeep methods (such as CBM, SHM, etc.) due to an ever-increasing reliance on progressive technological tools. The role of these technologies is more vital in equipment health prognostic studies, RUL assessment and through-life engineering support (TES) of structures, especially in cases when they are operating away from base maintenance facilities, e.g., aircraft, ships, subsea pipelines, and offshore energy structures.
The GWUT is a powerful and rapidly evolving technology with enormous potentials to be used as NDT and SHM of various platforms. Its long-range defect diagnostic capability with minimal preparatory costs (i.e., without paint, insulation and lagging removal) for inspection is the biggest Passive receivers Wheel excitations [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref]. The passive inspection prototype for the field tests using non-contact air receivers.
## Discussion and proposed methodology for future research
The role of health monitoring in operational availability, upkeep, and maintenance of equipment and structures has increased exponentially in recent years. Conventional maintenance procedures such as corrective and preventive are transforming rapidly towards technology-driven upkeep methods (such as CBM, SHM, etc.) due to an ever-increasing reliance on progressive technological tools. The role of these technologies is more vital in equipment health prognostic studies, RUL assessment and through-life engineering support (TES) of structures, especially in cases when they are operating away from base maintenance facilities, e.g., aircraft, ships, subsea pipelines, and offshore energy structures.
The GWUT is a powerful and rapidly evolving technology with enormous potentials to be used as NDT and SHM of various platforms. Its long-range defect diagnostic capability with minimal preparatory costs (i.e., without paint, insulation and lagging removal) for inspection is the biggest competitive edge over contemporary diagnostic techniques. Contrary to the limitations of other types of defect diagnostic technologies, guided waves have a lower level of operational barriers for usage in any type of material and environmental conditions. They have been used enormously in magnetic/non-magnetic, metals/composites isotropic/anisotropic structures of various shapes. The flexibility of use with contact or non-contact methods has enhanced its viability and future scope. The latter is relatively a more sophisticated technique, which has frequently been used in recent days for investigation of defects in complex, sensitive, dynamically loaded structures, especially in high temperatures where the use of simple contact type sensors may not be appropriate for use. Nevertheless, the contact transducer technique is still the most widely used method in UGWs because of its low initial cost, the flexibility of use, and portability.
In the reviewed literature, the efficacy of GWUT in different materials, operating and environmental conditions has been studied through experimental and numerical methods. This technique has been successfully used to locate and quantify the defects in various types of structures by classification of variations in numerous parameters of output signals, like frequency, amplitude, reflected/transmitted coefficient, velocity, TOF, and impedance. Certain limitations of UGWs have also been deliberated in the literature especially in bends/joints, complex geometries, submerge and thickly painted structures. These performance degradations have been attributed to the highly dispersion nature, scattering effects and attenuation of some acoustic guided waves under specific conditions. In the real world situations, the performance of UGWs can be affected by intermingling/interference of external acoustic signatures of running machinery (in surrounding) with the of GWs signals. More improvements in the use of GWs methodology is still needed to overcome the highlighted impediments, increase reliability, effectiveness, and diagnostic ability for identification of various defect features.
The Lamb wave interaction with damaged surface influences their propagation characteristics significantly, tempting various distinctive scattering phenomena (transmission, reflection, mode conversion); dependent on locations and severity of damaged area. The omnidirectional scattering (around the hole) of GWs has been reported in literature, when interacted with a through-thickness hole in the composite surface. The positioning of the sensor with respect to crack location becomes crucial in cases where the cracks are oriented in specific dimension/directions, shapes and angles. Therefore, the development of meticulous damage identification approach with an appropriate number of sensors would be required to capture the scattered waves optimally.
The literature has reported that the angle of crack orientation and length of crack has an important relationship with the reflection and transmission coefficients of incident elastic waves in plate-like metal surface. It has been revealed by some researchers that the reflection coefficient decreased whereas the transmission coefficient increased with the crack angle. This relationship was found to be even more complex in case of change of length of crack and angle. However, it is suggested that the reflection coefficient alone may not be able to define the length of the crack/notch exclusively. Moreover, the notch width can also influence the reflection and transmission ratio of GWs. Similarly, some research studies have reported that the transmission coefficient is higher in half through-thickness crack than in the full through-thickness, because of the fact that the GWs travel in the remaining half through-thickness part of the metallic surface.
The estimation of a crack depth has a great significance in determining the severity of crack and pit to crack transition. In absence of near-surface boundaries in metallic plates, the reduction in transmitted signal between actuator and sensor can be attributed to the quantitative measure of the defect. Nevertheless, in real-world situations the performance of active SHM with GWs can be adversely influenced by external excitation signatures; hence, defect investigation process can be further complicated. In the applications where external excitations are significant, passive SHM methods (such as acoustic emission) have been used by some researchers with contact/contactless methods. For example, the contact between wheels of train and rail track generates high excitation acoustic emissions which have been used for health inspection of rail tracks using non-contact methods.
Marine structures are often subjected to corrosive seawater conditions and enormous stress levels due to internal and external loading conditions. Therefore, the rate of various degradation mechanisms and failure probability are quite phenomenal. In research literature, the crack identification on water submerged metal structures has been deliberated at length. It has been reported that the defects were detected successfully with minor variations in the location of actual defects. This variation may be accredited to refraction phenomena in the denser medium. Behaviour of various modes of guided waves have been investigated to control antifouling process, detection of ice layers and de-icing in the ship's underwater hull, aerospace, and wind turbine structures. Many experiment/numerical based research studies have presented that the GWs can be used effectively to differentiate between the ice layer, water content and structural defects on external surfaces of engineering structures that are operating under dynamic climatic conditions.
The GWUT can be extremely useful for SHM, fracture diagnostics in obscure locations and subsequent CBM planning of dynamically loaded marine structures. It has been reported that presently the structural integrity inspection of ships and associated harbour installations (floating docks) is carried out using either conventional NDT methods like bulk ultrasonic waves, which require complete removal of surface protective paints or destructive methods like sand or grit blasting. These inspections can only be performed in dedicated graving/floating docks, which make inspection tasks expensive, labour intensive and time-consuming. The use of autonomous vehicles to facilitate video imaging of underwater structures and inspection of difficult to access areas (to identify the severity of corrosion, damage and anode wear) has already been reported in the literature [bib_ref] Modelling of pitting corrosion in marine and offshore steel structures-A technical review, Bhandari [/bib_ref] [bib_ref] Performing Detailed Level 1 Pipeline Inspection in Deep Water with a Remotely..., Kros [/bib_ref]. A comparatively new method with integration of GWUT (contact/contactless methods) and remotely operated vehicles (ROV) or autonomous underwater vehicles (AUV) can be introduced as future health inspection/prognostic technique for ships and other marine structures.
The literature review disclosed that the decision-making for the sensor(s) placement methods (pulse-echo, pulse-catch and a number of sensors) is extremely important. The precise identification/ quantification of defect features (e.g., orientation, size and relative position of cracks) is very much reliant on the selection of crack detection approach and the positioning of the sensors. The experimental results on metallic and composite plates presented in [bib_ref] Crack imaging and quantification in aluminum plates with guided wave wavenumber analysis..., Yu [/bib_ref] would be instrumental to develop a relationship between the directionality of defect features (in-line or offset position with respect to the direction of incident GWs), and the reflection/transmission ratio of induced guided waves. Although, the variation in the reflection/transmission coefficient with the orientation of defect in a particular direction has been elaborated. However, the aspects of identification/quantification of defect orientation and shape have not been discussed, probably because of nonlinearity/uncertainty of scattering data with the variation in the length of notch/crack. Therefore further research work can be conducted to explore a more realistic relationships and methodologies.
Various breakthroughs have been reported on defect characterization in metallic structures. However, to the best of knowledge of the authors of this paper, the research publications for determination of shape and orientation of structural defects in plate type metal structures are still obscure. The prominence of these two parameters (shape and orientation) is extremely vital in the purview of fatigue and fracture mechanisms, especially for the structures working under extreme static/dynamic loading conditions. The fatigue growth propagation and pit to crack transition processes are highly dependent on shape and orientation of the developed cracks in metallic structures, specifically those operating under complex loading and corrosive conditions. The identification of these prominent features of structural defects can be even more interesting than just detection or location of the defect. Advanced working methodologies, statistical/optimization tools and experimental equipment would be desirable for interpretation and empirical formulation of output (reflected/transmitted) signals. Therefore, the use of advanced and high-resolution equipment (Lecory wave-runner oscilloscope-Xi series), actuators/sensors (PAC's integral preamp-WDI-AMP) would be instrumental to minimize the equipment driven uncertainties.
This paper proposed an experimental-based research study to determine the shape and orientation of structural defect. The cracks/notches in several shapes and orientations can be artificially crafted on thin plate type aluminum plates. Formulation of an empirical model is possible based on variations in output signal parameters on various sensors placed in different configurations, as illustrated in [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref]. From the literature review, it is asserted that the reflection and transmission coefficient varies nonlinearly with the change in orientation, shape and size of the structural damage. Therefore, the proposed sensor arrangement (in [fig_ref] Figure 1: Active and passive SHM methods [/fig_ref] will be helpful to detect the optimized reflected/transmitted coefficient patterns, with variation of defect features. Similarly, the change in the actuator position with respect to the crack would also affect the scattering pattern of GWs. This information (identification of scattering pattern vis-a-vis defect characteristics) is extremely beneficial in development of the empirical model. The acquired experimental results can be validated numerically by simulation software such as COMSOL Multiphysics or ABAQUS.
## Conclusions and future works
The UGWs have immense potential in various types of structures for CM, SHM, and NDE. They have the propensity to be used without disrupting the ongoing operations in the structure under investigation, be it a pipeline embedded in soil/insulation applied exhaust/air conditioning ducts, rail track, aircraft, or a marine structure. The use of various modes of guided waves have been reported for ice detection, de-icing and differentiation between actual defect, ice layer and pure liquid phases in structural applications.
The flexibility of using GWs technology with several methods (contact, non-contact, semi contact and passive SHM) provide greater opportunities to extend its scope and applications in various other fields. To overcome the limitation of GWs in variant environments, various methodologies (OBS, BSS) have been adopted as compensatory strategies to optimize the performance. The potential of Lamb guided wave modes has been accepted in health monitoring of marine structures, aircraft, wind turbine blades, pigging of pipelines and antifouling on ship structures. Researchers have recognized the potential of UGWs to localized defects in various geometries and complex structures; hence, it provides the opportunity to analyze the severity of the defect, optimize inspection/maintenance interval and predict failure probability and RUL of valuable assets.
During this review, it was established that no research work has been published yet suggesting a methodology or an empirical method to determine the shape, widths, depths, and orientation of defects in thin plate structures. In the proposed research methodology, cracks/notches on thin metal plates will be produced artificially in various geometries (shapes and orientation). In the proposed method, the defects may be investigated at various angles with respect to the position of transducers. Upon successful identification of defect geometries, an empirical model would be designed to predict the defect features and quantification. The numerical study can be conducted for validation and
## Conclusions and future works
The UGWs have immense potential in various types of structures for CM, SHM, and NDE. They have the propensity to be used without disrupting the ongoing operations in the structure under investigation, be it a pipeline embedded in soil/insulation applied exhaust/air conditioning ducts, rail track, aircraft, or a marine structure. The use of various modes of guided waves have been reported for ice detection, de-icing and differentiation between actual defect, ice layer and pure liquid phases in structural applications.
The flexibility of using GWs technology with several methods (contact, non-contact, semi contact and passive SHM) provide greater opportunities to extend its scope and applications in various other fields. To overcome the limitation of GWs in variant environments, various methodologies (OBS, BSS) have been adopted as compensatory strategies to optimize the performance. The potential of Lamb guided wave modes has been accepted in health monitoring of marine structures, aircraft, wind turbine blades, pigging of pipelines and antifouling on ship structures. Researchers have recognized the potential of UGWs to localized defects in various geometries and complex structures; hence, it provides the opportunity to analyze the severity of the defect, optimize inspection/maintenance interval and predict failure probability and RUL of valuable assets.
During this review, it was established that no research work has been published yet suggesting a methodology or an empirical method to determine the shape, widths, depths, and orientation of defects in thin plate structures. In the proposed research methodology, cracks/notches on thin metal plates will be produced artificially in various geometries (shapes and orientation). In the proposed method, the defects may be investigated at various angles with respect to the position of transducers. Upon successful identification of defect geometries, an empirical model would be designed to predict the defect features and quantification. The numerical study can be conducted for validation and optimization of designed experiment and its methodology.
The outcome of experimental results based on this proposed plan can be of highly valuable to bring further improvement in GWUT based SHM/NDE in various field applications, particularly for identification of damage tolerance of structural defects and subsequent prediction of RUL and life extension (LE). It can be made as a part of future structural maintenance/inspection strategy of offshore installations and ships, which normally operates with limited man-power and away from base maintenance facilities. The big-data of SHM acquired from large marine structures (using ROV/AUV integrated GWs methods) can be continuously provided to the remote maintenance experts, through wireless transmission or internet of things (IOT). In presence of ground maintenance teams, this data can be interpreted and the defect (if any) can be investigated through comparison of data (existing versus historical SHM data). Subsequently, this priori information on structural health can be used to update/optimize the forthcoming maintenance/inspection schedules, using probabilistic inference methods such as Bayesian inference. Determination of characteristics and directional properties of defects (shape, depth, orientation, the relative position of crack) through GW technologies can be extremely helpful to predict the future growth rate and defect intensity. Notwithstanding, further research studies are essential to minimize the glitches with the use of guided wave technology for health monitoring of assets in practical field as well as to increase its reliability and effectiveness, under vibrant operating conditions Although some researchers have developed the experimental-based compensatory methods to uplift the performance of GWUT affected by the environmental (mostly temperature) variations. However, implementation of these benign methods in the field experiments is not free from numerous complications due to the abrupt transient behavior of a wide range of climatic factors which ca be triggered simultaneously. The ship's hull is consistently under static/dynamic loading conditions from internally (machinery) or external factors (momentum/wave action) which aggravate the challenges for effective SHM, exponentially. Contrary to the lab experiments, the GWUT based health inspection of these giant structures involves several factors at the same time, which have the tendency to adversely affect the defect diagnostic capability. A few prominent factors are the non-uniform thermal gradients, wetting/icing of external surfaces, exterior acoustic signatures, and the surface protective measures. In the future research studies, field-oriented experimentation under marine weather conditions is necessitated to explore the capability of various GWs (Shear, Lamb, etc.) on large and valuable marine assets. The FEM approaches can also be used for the validation process, which is a cost-effectiveness way to evaluate the proficiency of GWs for characterization of salient defect features.
[fig] Figure 1: Active and passive SHM methods. [/fig]
[fig] Figure 3: Shear wave propagation (adapted from[40]). [/fig]
[fig] Figure 4: Rayleigh waves propagation (adapted from[10]). [/fig]
[fig] Figure 5: Phase and group velocities of GWs. [/fig]
[fig] Figure 6: Pitch-catch method. [/fig]
[fig] Figure 7: Pulse-echo method. [/fig]
[fig] Figure 8: GWs inspection of a metallic pipeline structures[96]. [/fig]
[fig] Figure 9: A fully non-contact laser ultrasonic scanning system. [/fig]
[fig] Sensors 2018 ,: 18, x FOR PEER REVIEW 16 of 26 [/fig]
|
Leiomyosarcoma of the Ovarian Vein: a Case Report with Radiological Findings
[bib_ref] Leiomyosarcoma of large arteries and veins, Kevorkian [/bib_ref] [bib_ref] Primary venous leiomyosarcoma: a rare but lethal disease, Dzsinich [/bib_ref] [bib_ref] Primary venous leiomyosarcoma: a rare but lethal disease, Dzsinich [/bib_ref] [bib_ref] Leiomyosarcoma of the ovarian vein: an unusual cause of severe abdominal and..., Kawai [/bib_ref] [bib_ref] Leiomyosarcoma of the ovarian vein: report of a case, Iannelli [/bib_ref] [bib_ref] Leiomyosarcoma of the ovarian vein: report of a case, Iannelli [/bib_ref] [bib_ref] Leiomyosarcoma of the ovarian vein: report of a case, Iannelli [/bib_ref]
## Case report
A 39-year-old woman visited the hospital and presented with a hard, firm and palpable mass in the right upper abdomen. The medical and family histories were not remarkable. Serum levels of the tumor markers, including CA 19-9, carcinoembryonic antigens and CA 125, as well as laboratory findings, were all normal.
Excretory urography revealed lateral displacement of the lower pole of the right kidney and ureteropelvic junctional obstruction caused by the space-occupying lesion. An abdominal ultrasonogram (Acuson Sequoia 512; Siemens Medical System, Erlangen, Germany) showed the presence of a solid medium-echoic retroperitoneal Hyeon Je Cho, MD 1 Ho Kyun Kim, MD 1 Jung Ho Suh, MD 1 Ghi Jai Lee, MD 1 Jae-Chan Shim, MD [bib_ref] Leiomyosarcoma of large arteries and veins, Kevorkian [/bib_ref] Young Hwa Kim, MD 1 Hyuck-Sang Lee, MD 2 Yun Kyung Kang, MD 3 mass with partial invasion of the inferior vena cava (Figs. 1A C). An examination with 16-channel multi-detector CT (Mx8000 IDT; Philips Medical System, Eindhoven, the Netherlands) revealed the presence of a lobulated and well-defined right retroperitoneal mass measuring 5.1 5.5 6.9 cm with extensive cystic degenerations and highly enhanced solid portions on contrast-enhanced scans [fig_ref] Figure 1: Radiological and intraoperative findings of ovarian vein leiomyosarcoma in 39-year-old woman [/fig_ref]. Coronal images definitely confirmed the vertically elongated arm-like structures contiguous with the mass, extending upwards to the right renal vein with partial protrusion into the inferior vena cava [fig_ref] Figure 1: Radiological and intraoperative findings of ovarian vein leiomyosarcoma in 39-year-old woman [/fig_ref]. In the downward direction, the mass was connected to the normal right ovarian vein in the pelvis [fig_ref] Figure 1: Radiological and intraoperative findings of ovarian vein leiomyosarcoma in 39-year-old woman [/fig_ref]. The mass seemed to be separated from the adjacent right kidney and small bowels. Both renal and iliac angiograms showed the tumor vessels originating from the inferior segmental branch of the right renal artery.
During surgery, the mass was found to be located in the retroperitoneal space inferomedial side of the right kidney. The upper portion of the right ovarian vein was distended with tumor thrombi [fig_ref] Figure 1: Radiological and intraoperative findings of ovarian vein leiomyosarcoma in 39-year-old woman [/fig_ref]. There was no invasion of the adjacent organs and no enlarged lymph nodes.
Crosscut sections of the tumor showed mainly solid, and partly hemorrhagic and necrotic appearances with a Adjuvant radiotherapy was performed over a six-week period. Follow-up abdominopelvic CTs after two and three months showed the presence of a 1.0 cm low attenuation nodule in the liver, that had newly developed and enlarged in size over the short-term intervals. We assumed the presence of a liver metastasis, but the patient was lost to further planned follow-up.
# Discussion
Leiomyosarcomas of the vein arise from the smooth muscle cells of the tunica media of the vessel wall. They grow slowly and bilaterally along the wall of the vessel.
Retroperitoneal leiomyosarcomas exhibit themselves as completely extravascular (62%), completely intravascular (5%) or with both extra-and intraluminal growth patterns (33%) [bib_ref] Leiomyosarcoma of the retroperitoneum and inferior vena cava: radiologic-pathologic correlation, Hartman [/bib_ref]. As the leiomyosarcomas of the inferior vena cava grow, they eventually invade adjacent tissues and metastasize distantly; however, they tend to expand along the tissue plans of least resistance. There are no reported cases of tissue invasion of the adrenals, kidneys or bowels for leiomyosarcomas of the inferior vena cava [bib_ref] Computed tomography of leiomyosarcoma of the inferior vena cava, Young [/bib_ref]. Like other retroperitoneal tumors, these tumors seldom present with symptoms until they progress into a huge mass. When the tumors are discovered at this later stage, there is a poor outlook for long-term survival.
As the cases are rare, previously reported radiological findings of a ovarian vein leiomyosarcoma are few in number. The lesions have been merely described as a heterogeneous retroperitoneal mass with ipsilateral hydronephrosis by conventional CT and excretory urography [bib_ref] Leiomyosarcoma of the ovarian vein: an unusual cause of severe abdominal and..., Kawai [/bib_ref] [bib_ref] Leiomyosarcoma of the ovarian vein: report of a case, Iannelli [/bib_ref].
The findings of excretory urography, representing the effect of mass, are not specific for the diagnosis of the leiomyosarcomas of vascular origin.
The general findings of gray-scale ultrasonography of an inferior vena cava leiomyosarcoma are solid but often show cystic necrosis with irregular walls [bib_ref] Inferior vena caval leiomyosarcoma: diagnosis and biopsy with color Doppler sonography, Singh-Panghaal [/bib_ref]. In our case, Cho et al.
## S16
Korean J Radiol 9(Suppl), July 2008
## E
F G S17 the sonographic findings were compatible with those of the CT scans, i.e., the pedicular structure was a tumor-invading ovarian vein. However, sonography had some limitations in that it failed to demonstrate the whole features of the tumor and the relationship between the tumor and the normal ovarian vein in the pelvis. These limitations may be due to the overlying bowel, the extent of the mass, or limitations in the depth resolution of the ultrasound modality [bib_ref] Leiomyosarcoma of the retroperitoneum and inferior vena cava: radiologic-pathologic correlation, Hartman [/bib_ref]. The use of CT is superior to the use of sonography to reveal the highly enhancing and necrotic characteristics of the tumor, which is specific for the diagnosis of leiomyosarcoma [bib_ref] Computed tomography of leiomyosarcoma of the inferior vena cava, Young [/bib_ref]. When CT shows a solid or necrotic extraluminal retroperitoneal mass not originating from a retroperitoneal organ, and possessing a contiguous intravascular enhancing appearance, it is accepted as a pathognomonic finding of vascular leiomyosarcoma [bib_ref] Leiomyosarcoma of the retroperitoneum and inferior vena cava: radiologic-pathologic correlation, Hartman [/bib_ref]. Low-attenuation areas representing necrosis, or rarely, a cystic mass due to diffuse necrosis, may be seen [bib_ref] Preoperative CT and MR imaging of inferior vena cava Leiomyosarcoma, Blum [/bib_ref]. In our case, the eccentric cystic area is visible on sonography and CT; however, the extensive central necrosis is only visible on CT.
With great advances in digital imaging and high spatial resolution of multi-detector CT over the last few decades, scrolling through multi-stacked images and conducting multiplanar reconstruction is possible. These methods are useful in demonstrating the relationship between a mass and vessels, as well as detecting the tumor or venous thrombi. Thus, it was not difficult to presume that the mass originated in the right ovarian vein based on anatomical conceptions, as the tumor possessed vertically elongated bi-armed structures, the typical site of the right ovarian vein.
The angiographic findings of vascular leiomyosarcomas are generally accepted as nonspecific for the diagnosis. Angiography may have a role for demonstrating the primary vessels supplying the tumors and determining the degree of vascularity [bib_ref] Preoperative CT and MR imaging of inferior vena cava Leiomyosarcoma, Blum [/bib_ref].
Metastases will occur eventually in most survival cases of other venous or retroperitoneal leiomyosarcomas [bib_ref] Primary venous leiomyosarcoma: a rare but lethal disease, Dzsinich [/bib_ref] [bib_ref] Leiomyosarcoma of the retroperitoneum and inferior vena cava: radiologic-pathologic correlation, Hartman [/bib_ref]. Hematological metastasis occurs more commonly than lymphatic metastasis. The frequent sites for metastases are the liver, lung, brain, and peritoneum [bib_ref] Primary venous leiomyosarcoma: a rare but lethal disease, Dzsinich [/bib_ref] [bib_ref] Leiomyosarcoma of the retroperitoneum and inferior vena cava: radiologic-pathologic correlation, Hartman [/bib_ref] [bib_ref] Inferior vena caval leiomyosarcoma: diagnosis and biopsy with color Doppler sonography, Singh-Panghaal [/bib_ref]. Local recurrence occurs occasionally in 40% to 77% of cases, even though a complete excision was made [bib_ref] Leiomyosarcoma of the retroperitoneum and inferior vena cava: radiologic-pathologic correlation, Hartman [/bib_ref]. Our case also showed the CT findings of a liver metastasis, although a confirmation was not obtained.
In conclusion, the general imaging findings of our case were coincidental with those of other venous or retroperitoneal leiomyosarcomas. Sonography seemed to be helpful for the diagnosis of ovarian vein leiomyosarcoma, but the use of multi-detector CT was the tool of choice to confirm the diagnosis. Scrolling through multi-stacked images and multiplanar reconstruction of the CT images have great benefits for the preoperative predictions of the location, extent, and growth patterns of the tumor, as well as the relationship of the tumor and vessels. We expect that the primary diagnostic tool for vascular leiomyosarcomas will be the use of multi-detector CT and it may expel the other unnecessary, cost consuming or invasive diagnostic procedures.
[fig] Figure 1: Radiological and intraoperative findings of ovarian vein leiomyosarcoma in 39-year-old woman. A. Transverse sonography reveals well-defined, spherical, medium-echoic and solid mass with eccentric cystic portion (arrow) at right paravertebral region. B, C. Longitudinal and transverse sonography shows vascular structure, filled with solid components (arrowheads) and connected to main mass (circle), of which superior end shows round protrusion (arrows) in lumen of junctional area of inferior vena cava and right renal vein. D. Just below level of aortic bifurcation, axial contrast-enhanced CT scan shows lobulated and well-defined lobulated retroperitoneal mass with highly enhanced solid components and extensive cystic degenerations (asterisk) at right paravertebral region.C D yellow-tinged gray color. Microscopically, the tumor showed a typical pattern of intersecting fascicles of spindle cells with hyperchromatic nuclei and eosinophilic cytoplasms, representing the smooth muscle cells in the media. The tumor cells frequently showed highly pleomorphic nuclei, indicative of dedifferentiation. Immunohistochemically, the tumor cells were strongly positive for both smooth muscle actin and desmin. The cells showed positive expression of CD34. Expression of CD117 (KIT), usually expressed in gastrointestinal stromal tumors and normally negative in leiomyosarcoma, was negative. [/fig]
|
Dengue Human Infection Models Supporting Drug Development
Dengue is a arboviral infection that represents a major global health burden. There is an unmet need for effective dengue therapeutics to reduce symptoms, duration of illness and incidence of severe complications. Here, we consider the merits of a dengue human infection model (DHIM) for drug development. A DHIM could allow experimentally controlled studies of candidate therapeutics in preselected susceptible volunteers, potentially using smaller sample sizes than trials that recruited patients with dengue in an endemic country. In addition, the DHIM would assist the conduct of intensive pharmacokinetic and basic research investigations and aid in determining optimal drug dosage. Furthermore, a DHIM could help establish proof of concept that chemoprophylaxis against dengue is feasible. The key challenge in developing the DHIM for drug development is to ensure the model reliably replicates the typical clinical and laboratory features of naturally acquired, symptomatic dengue.
The 4 serotypes of dengue virus (DENV-1 to DENV-4) are the most important arboviral pathogens of humans. The global burden of dengue, recently estimated at approximately 100 million cases per year, remains unchallenged by licensed vaccines or sustainable vector control strategies. The scale of the dengue burden in the endemic countries of Asia and Latin America has a negative economic impact and places a significant clinical burden on often fragile health care systems [bib_ref] Economic impact of dengue illness in the Americas, Shepard [/bib_ref] [bib_ref] Use of multiple data sources to estimate the economic cost of dengue..., Shepard [/bib_ref] [bib_ref] Total economic cost and burden of dengue in Nicaragua: 1996-2010, Wettstein [/bib_ref].
Dengue is an acute systemic febrile illness that manifests with abrupt onset as an undifferentiated fever that is difficult to distinguish from other infections without laboratory diagnostic tests. Headache, malaise, myalgia, retro-orbital pain are common symptoms in adults. The clinical features differ by age group, with symptoms such as cough, vomiting and abdominal pain more common in children [bib_ref] Clinical features and differences between child and adult dengue infections in Rayong..., Hanafusa [/bib_ref]. Leukopenia and thrombocytopenia are common laboratory features.
In most cases, fever and symptoms resolve uneventfully after 3-7 days with no long-term sequelae. In a subset of cases, a transient vascular leakage syndrome develops after 3-4 days of illness, which, when severe, can lead to life-threatening hypovolemic shock (called dengue shock syndrome) and/or hemorrhage. The vascular leak syndrome may be associated with complement activation and a coagulopathy that can contribute to the risk of hemorrhage [bib_ref] The potential pathogenic role of complement in dengue hemorrhagic shock syndrome, Bokisch [/bib_ref] [bib_ref] Hemostatic changes in Vietnamese children with mild dengue correlate with the severity..., Wills [/bib_ref]. Signs of vascular leakage include an increasing hematocrit, microalbuminuria, or the accumulation of fluids at serosal surfaces (eg, pleural effusions).
Supportive care and the careful titration of minimum volumes of parenteral crystalloid fluids to maintain stable cardiac output during the 1-3-day period of vascular permeability are critical elements in dengue case management. Vasopressors are commonly used in cases with prolonged shock. Clinically significant hemorrhage requiring use of blood products is more common in adults than in children [bib_ref] Clinical features of dengue in a large Vietnamese cohort: intrinsically lower platelet..., Trung [/bib_ref].
Improvements in the management of severe dengue cases have seen the case mortality rate decline over the last decade in many but not all endemic settings [bib_ref] Epidemiological factors associated with dengue shock syndrome and mortality in hospitalized dengue..., Anders [/bib_ref] [bib_ref] Dengue mortality in patients under 18 years old: an analysis from the..., Lumbiganon [/bib_ref] [bib_ref] Determinants of mortality from severe dengue in Brazil: a population-based case-control study, Moraes [/bib_ref]. At present there is no validated way of identifying which patients will progress to more severe disease, meaning that health facilities in endemic areas are often overwhelmed with patients with dengue who require ongoing observation as outpatients or are admitted for inpatient observation.
## The cases for dengue therapeutic and prophylactic agents
## Antivirals
A drug that can shorten the duration of illness and reduce the risk of disease progression would be a significant advance for both patients and for overstretched health systems in dengueendemic countries. An obvious approach is to target virus replication with a small molecule antiviral drug that given early in the course of illness has the potential to shorten the duration of infection. Moreover, because high early DENV viremia levels are a risk factor for the development of more severe disease, an antiviral drug might also reduce the incidence of severe complications [bib_ref] High circulating levels of the dengue virus nonstructural protein NS1 early in..., Libraty [/bib_ref] [bib_ref] Dengue viremia titer, antibody response pattern, and virus serotype correlate with disease..., Vaughn [/bib_ref]. In addition, a drug that reduces the magnitude and duration of viremia has the potential to reduce onward DENV transmission by reducing the length of time an individual patient is infectious to mosquitoes [bib_ref] Therapeutics for dengue: recommendations for design and conduct of early-phase clinical trials, Simmons [/bib_ref]. Alternatives to small molecule antiviral drugs include therapeutic monoclonal antibodies; however, high costs and requirement for parenteral administration are currently barriers to their use.
## Disease modulators
Both host and virus factors are important in influencing the outcome of DENV infection. A range of host factors, such as age, sex, genotype, and flavivirus infection history, influence the disease outcome [bib_ref] Epidemiological factors associated with dengue shock syndrome and mortality in hospitalized dengue..., Anders [/bib_ref] [bib_ref] Dengue haemorrhagic fever/dengue shock syndrome: lessons from the Cuban epidemic, Kouri [/bib_ref] [bib_ref] Genome-wide association study identifies susceptibility loci for dengue shock syndrome at MICB..., Khor [/bib_ref]. Because the manifestations of severe dengue are typically observed in secondary DENV infections and when innate and adaptive immune responses have driven steep declines in viremia levels, it is widely held that the immune system plays an important role in disease pathogenesis [bib_ref] Immunopathological mechanisms in dengue and dengue hemorrhagic fever, Green [/bib_ref]. This represents the rationale for use of an immunomodulatory drug that can favorably modulate the poorly defined immunopathogenic mechanisms postulated to underlie severe dengue. It is possible that the dichotomous approach of targeting either the virus or the immune response is an oversimplified strategy to the development of dengue therapeutics. Other physiological systems are involved in the control of endothelial integrity, but their role in dengue remains poorly understood [bib_ref] The control of vascular integrity by endothelial cell junctions: molecular basis and..., Dejana [/bib_ref].
## Prophylaxis
A vaccine would be the most effective prophylactic public health intervention for control of dengue, but the development of an efficacious tetravalent dengue vaccine is proving to be a challenge [bib_ref] Dengue vaccine development: a 75% solution, Halstead [/bib_ref] [bib_ref] Protective efficacy of the recombinant, live-attenuated, CYD tetravalent dengue vaccine in Thai..., Sabchareon [/bib_ref]. An alternative or adjunctive approach is the use of drug chemoprophylaxis in the community. Chemoprophylaxis using oseltamivir has been advocated for controlling the spread of influenza within households and was commonly used in some settings during the 2009 H1N1 influenza pandemic, creating a precedent for such an approach for an acute viral infection [bib_ref] Management of influenza in households: a prospective, randomized comparison of oseltamivir treatment..., Hayden [/bib_ref] [bib_ref] Oseltamivir ring prophylaxis for containment of 2009 H1N1 influenza outbreaks, Lee [/bib_ref]. The safety profile of a prophylactic dengue drug would have to be exemplary, and there are major questions of whether this approach could be cost-effective and sustainable, given the temporally and spatially heterogenous patterns of DENV transmission in endemic regions [bib_ref] Travelling waves in the occurrence of dengue haemorrhagic fever in Thailand, Cummings [/bib_ref] [bib_ref] Trends in patterns of dengue transmission over 4 years in a pediatric..., Balmaseda [/bib_ref]. Moreover, in most dengue-endemic regions a significant fraction of the population is already immune and will not benefit from chemoprophylaxis. Consequently, in endemic areas the number needed to treat, and the duration they need to be treated for, may be large for the prevention of a single symptomatic case and very large for the prevention of severe cases. There is a marginally stronger argument for chemoprophylaxis in populations with no immunity to DENV. Examples of this include denguenaive travelers or military personnel who are transiently present in areas of high DENV transmission. Similarly, chemoprophylaxis could be indicated for communities experiencing "virgin soil" dengue outbreaks because of importation by viremic travelers or infected mosquitoes.
## Previous clinical trials of therapeutics for dengue and relevance to the dengue human infection model
For a disease with an estimated global burden prevalence of 100 million cases per year, remarkably few clinical trials of treatment approaches have been performed. Early trials tested fluid resuscitation approaches in patients with hypovolemic shock [bib_ref] Acute management of dengue shock syndrome: a randomized double-blind comparison of 4..., Ngo [/bib_ref] [bib_ref] Fluid replacement in dengue shock syndrome: a randomized, double-blind comparison of four..., Dung [/bib_ref] [bib_ref] Comparison of three fluid solutions for resuscitation in dengue shock syndrome, Wills [/bib_ref]. More-recent trials have tested the candidate antiviral agents chloroquine and balapiravir or have targeted the immune response by use of oral corticosteroids [bib_ref] A randomized controlled trial of chloroquine for the treatment of dengue in..., Tricou [/bib_ref] [bib_ref] Effects of short-course oral corticosteroid therapy in early dengue infection in Vietnamese..., Tam [/bib_ref] [bib_ref] A randomized, double-blind placebo controlled trial of balapiravir, a polymerase inhibitor, in..., Nguyen [/bib_ref]. Although none of these trials have demonstrated therapeutic efficacy, they have provided a framework for the conduct of such trials and has led to the publication of recommendations for a standardized approach to dengue clinical trial design and conduct [bib_ref] Therapeutics for dengue: recommendations for design and conduct of early-phase clinical trials, Simmons [/bib_ref]. In addition, a trial of lovastatin in dengue is currently underway and further trials are planned [bib_ref] Therapeutics for dengue: recommendations for design and conduct of early-phase clinical trials, Simmons [/bib_ref] [bib_ref] Lovastatin for adult patients with dengue: protocol for a randomised controlled trial, Whitehorn [/bib_ref]. The virological findings in these trials, consistent with previous literature, have underscored the need for antiviral therapy to commence early in the course of illness, because illness resolves and patients are no longer infectious to Aedes aegypti mosquitoes by 5-7 days after illness onset [bib_ref] Dengue viremia titer, antibody response pattern, and virus serotype correlate with disease..., Vaughn [/bib_ref] [bib_ref] Therapeutics for dengue: recommendations for design and conduct of early-phase clinical trials, Simmons [/bib_ref] [bib_ref] Kinetics of viremia and NS1 antigenemia are shaped by immune status and..., Tricou [/bib_ref] [bib_ref] Differing influences of virus burden and immune activation on disease severity in..., Libraty [/bib_ref] [bib_ref] Host and viral features of human dengue cases shape the population of..., Nguyet [/bib_ref].
## Potential roles for a human infection model in dengue drug development
In principal, a dengue human infection model (DHIM) provides opportunities for fast-tracking dengue drug development, particularly therapeutic and prophylactic antivirals. The overall strength of the DHIM is that it allows for controlled prophylactic or therapeutic studies of DENV infection in susceptible participants who have been preselected for a set of desirable characteristics. In contrast, prophylactic or therapeutic studies of DENV infection in an endemic country population are subject to the vagaries of ethnic diversity of the patient population, fluctuating serotype and genotype prevalence, variable viremia levels, and prior flavivirus infection histories. Collectively, these sources of variation in an endemic country population probably boost the required patient sample size required for detection of therapeutic or clinical safety signals.
## Critical hurdles for a dhim
Central to the usefulness of a DHIM is whether it mimics the main features of a symptomatic dengue case. This is a fundamental and critical hurdle in the development of a successful DHIM for use in dengue therapeutic development. Thus, the kinetics of clinical, virological, hematological, and biochemical changes that are typically present in a naturally acquired and clinically uncomplicated adult dengue case should also manifest in the adult DHIM. In the beginning, at least, we assume that the DHIM will involve participants who are flavivirus naive, and thus all experimental infections will be primary infections. In the context of dengue drug development, the characteristics of a symptomatic, naturally acquired DENV infection that a DHIM would need to replicate include acquisition of infection and certain clinical, virological, and laboratory features.
## Acquisition of infection
A DHIM could use DENV-infected mosquitoes to elicit infection in human volunteers. Although this approach may be optimal in replicating the natural history of DENV infection because, for example, the immunoregulatory effects of mosquito saliva would be replicated, it may present additional regulatory and logistical hurdles. Alternatively, a DHIM could use subcutaneous inoculation of laboratory-cultured virus, as used in the challenge experiments by Sun and colleagues [bib_ref] Experimental dengue virus challenge of human subjects previously vaccinated with live attenuated..., Sun [/bib_ref]. This approach bypasses issues surrounding use of mosquitoes, although it remains to be determined if there are material differences between needledelivered and mosquito-delivered DENV infection.
## Clinical features
Typical signs and symptoms of dengue that a DHIM would need to replicate in otherwise healthy adults include fever, headache, myalgia, musculoskeletal pain, and nausea or vomiting [bib_ref] Early clinical and laboratory indicators of acute dengue illness, Kalayanarooj [/bib_ref]. Minor hemorrhagic features may occur between days 3 and 6 of illness (eg, petechiae) [bib_ref] Early clinical and laboratory indicators of acute dengue illness, Kalayanarooj [/bib_ref]. Prevention or rapid amelioration of these self-limiting symptoms will be a target of dengue therapeutic drug development. Primary infections in otherwise healthy adults are rarely associated with severe clinical complications, but it would be prudent to exclude the elderly and those with comorbid conditions from the volunteer pool, because these groups are at elevated risk of severe complications [bib_ref] Severe dengue virus infection in travelers: risk factors and laboratory indicators, Wichmann [/bib_ref] [bib_ref] Clinical features of 62 imported cases of dengue fever in Japan, Itoda [/bib_ref].
## Virological features
The intrinsic incubation period of DENV infection has been estimated to be between 4-10 days. Work that reanalyzed historical mosquito exposure experiments estimated that the intrinsic incubation period in primary infection was approximately 6 days [bib_ref] Natural history of dengue virus (DENV)-1 and DENV-4 infections: reanalysis of classic..., Nishiura [/bib_ref]. The duration of viremia varies by serotype and DENV infection history. The peak viremia level is often not observed in naturally occurring infections because it precedes the time point when patients attend health care facilities, but findings suggest that the peak level occurs later in primary infections as compared to secondary [bib_ref] Kinetics of viremia and NS1 antigenemia are shaped by immune status and..., Tricou [/bib_ref]. Studies suggest that primary infections with DENV-1 are associated with a longer viremic period and higher viremia levels [bib_ref] Kinetics of viremia and NS1 antigenemia are shaped by immune status and..., Tricou [/bib_ref] [bib_ref] Kinetics of plasma viremia and soluble nonstructural protein 1 concentrations in dengue:..., Duyen [/bib_ref]. An optimal DHIM would replicate the virological features of a naturally acquired DENV infection to enable confident predictions of therapeutic or prophylactic drug efficacy in the field.
## Laboratory features
A DHIM should elicit thrombocytopenia and leukopenia, both typical features in patients with primary or secondary dengue. In addition, many patients have minor elevations in transaminase and creatine kinase levels [bib_ref] Clinical features of dengue in a large Vietnamese cohort: intrinsically lower platelet..., Trung [/bib_ref] [bib_ref] Liver involvement associated with dengue infection in adults in Vietnam, Trung [/bib_ref]. Clotting abnormalities are also often observed in severe cases with elevated activated partial thromboplastin times and reduced levels of fibrinogen; however, these occur less frequently in primary infections [bib_ref] Coagulation abnormalities in dengue hemorrhagic fever: serial investigations in 167 Vietnamese children..., Wills [/bib_ref].
## Strengths and weaknesses of a dhim for drug development
The strengths and weaknesses of a DHIM for performing early-phase clinical studies compared to performing such studies in naturally acquired cases in an endemic country are summarized in [fig_ref] Table 1: Early-Phase Clinical Studies of Antiviral Drugs [/fig_ref]. In short, the great advantage of the DHIM for early-phase drug development is the ability to perform controlled prophylactic or early therapeutic studies in a setting amenable to intensive pharmacological and basic research investigations.
## Focus areas in which the dhim could substantially advance clinical drug development
A DHIM that mimics natural infection would be particularly well suited to early-phase studies of chemoprophylaxis and accompanying pharmacokinetic studies, in which the timing of treatment and infection can be controlled. In contrast, in a dengue-endemic country, early-phase chemoprophylaxis trials would require large and lengthy community-based cohort studies, as has been used to derive evidence for oseltamivir in the chemoprophylaxis of influenza [bib_ref] Management of influenza in households: a prospective, randomized comparison of oseltamivir treatment..., Hayden [/bib_ref]. The DHIM would thus be a more cost-effective and time-efficient route to acquiring proof of concept evidence that chemoprophylaxis for dengue is possible. For a given antiviral drug, a DHIM would also be well-suited for studies that accurately define the therapeutic window in symptomatic cases, since the time of infection, onset of symptoms and treatment will be prospectively documented. In contrast, in endemic countries enrollment of study participants is dependent on patients seeking health care, who will have various durations of illness history at the time of study enrollment. A DHIM can also aid the development of alternative therapeutic approaches, such as immunomodulatory agents. In addition, a DHIM may aid the development of noninvasive physiological monitoring techniques within clinical trials, perhaps through the identification of appropriate physiological measurements that could be used as end points [bib_ref] Therapeutics for dengue: recommendations for design and conduct of early-phase clinical trials, Simmons [/bib_ref].
In conclusion, the work of Sun and colleagues [bib_ref] Experimental dengue virus challenge of human subjects previously vaccinated with live attenuated..., Sun [/bib_ref] has shown that a DHIM is possible. This is a significant advance in the field with the potential to contribute to the development of successful dengue therapeutics and much else. The future use of the DHIM in dengue clinical research will require a significant investment from multiple stakeholders. Its contribution in drug development may aid in the selection of appropriate candidate drugs for further development and assist pharmacokinetic studies. The potential significance of these contributions suggests that investing in a DHIM is both appropriate and timely.
## Notes
Financial support. This work was supported by fellowships from the Wellcome Trust of the United Kingdom (grants 097430/Z/11/Z to J.W. and 084368/Z/07/Z to C.P.S.).
Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
[table] Table 1: Early-Phase Clinical Studies of Antiviral Drugs: Strengths and Weaknesses of the DHIM vs Trial Conducted in a Dengue-Endemic Setting [/table]
|
Home environment: respiratory and allergic phenotypes from birth to age six in the PELAGIE cohort
# Introduction
The prevalence of childhood asthma and allergies varies greatly worldwide: 1 at age 6-7, 2.8-37.6% of children have had asthma symptoms in the past year and 2-22.3% eczema symptoms. Asthma is the most frequent chronic disease in childhood and represents a major burden for patients, their families, and society.Asthma and allergic disorders commonly appear in infancy, but their pathogenesis and interplay remain incompletely understood.To date, only three studies have collected longitudinal data for multiple 12-month profiles of respiratory and allergic symptoms.Rancière et al.simultaneously observed profiles of wheezing, dry night cough, eczema, and rhinitis symptoms from ages 0 to 4 in 2522 children from the PARIS (Pollution and Asthma Risk: an Infant Study) birth cohort. Two transient (rhinitis or wheeze) and two persistent (cough/rhinitis or dermatitis) phenotypes were identified, with a low symptoms reference group. Panico et al.analyzed wheezing and proxies for atopy (eczema and/or hay fever ever) in 11,632 children from the Millennium Cohort Study at ages 3, 5, and 7 and obtained a 4-trajectory clustering: a group with low symptom levels, another with low wheeze and high atopic symptoms, a third with high levels of both wheeze and atopic symptoms, and finally a group with high levels of wheeze only. Belgrave et al.studied wheeze, eczema, and rhinitis at ages 1, 3, 5, 7, and 11 in 9801 children from the ALSPAC (The Avon Longitudinal Study of Parents and Children) and MAAS (Manchester Asthma and Allergy Study) birth cohorts and identified eight developmental profiles: no disease, atopic march, persistent eczema and wheeze, persistent eczema with later-onset rhinitis, persistent wheeze with later-onset rhinitis, transient wheeze, eczema only, and rhinitis only.
The development and expression of asthma/allergic disorders depend on gene-environment interactions.Worldwide variations in prevalence rates of asthma and allergies call attention to the potential role of the environment, as genetic differences alone cannot explain this heterogeneity.The literature has identified numerous avoidable determinants for respiratory and allergic disorders in childhood, in particular, the indoor environment,because young children usually spend most of their time at home.The objective of this study was to describe the joint trajectory of asthma, eczema, and rhinitis symptoms at three time points, at ages 1, 2, and 6 years, in a population-based cohort conducted in Brittany (France) from 2002 to 2006. We explored the associations of these phenotypes with some avoidable indoor environmental factors in early life.
# Results
## Participants and descriptive data
The cluster analysis, included 935 of the 3323 eligible mother-child couples from the PELAGIE (Perturbateurs Endocriniens: Étude Longitudinale sur les Anomalies de la Grossesse, l'Infertilité et l'Enfance) mother-child cohort .
Compared with the nonrespondents, respondents were older (p < 0.001), had higher education level (p < 0.001), and were less often nulliparous (p = 0.008) or a smoker at inclusion (p < 0.001).
The participating mothers had a mean age of 30.8 years (SD 4.1), and 69.8% had at least some post-secondary education. At inclusion, 22.0% smoked, and 14.7% drank some alcohol. Most of the children were full-term (96.8%) and born by vaginal delivery (83.2%); 51.3% were boys. There was a family history of allergy in 249 (26.3%) mother-child pairs. Fewer than half of the children had been exclusively breastfed (41.1%) during the first 3 months. At age 2, 34.8% had been exposed to environmental tobacco smoke (ETS), 44.4% to bleach, 45.0% to glue, and 48.5% to alkyd/acrylic paints used for renovation activities in the dwellings. Half of the children had been exposed to pets at age 2, 40.5% slept on a mattress bought at least 3 years ago, and 41.2% in a bedroom aired less than once a day. Overall, 55.0% of the dwellings used individual heating devices fueled by gas/fuel oil/ wood or chimneys.
## Outcome data
The best partition for the cluster analysis was a five-cluster solution . The prevalence of respiratory/allergic symptoms (wheeze, rash, rhinitis, and cough) at age 1, 2, and 6 years of the child (except cough, recorded only at age 2 and 6 years) were used to characterize and label each cluster obtained (supplementary. At time point 1, 4% of the subjects were assigned to a cluster we labeled "reference" phenotype. The prevalence of respiratory/allergic symptoms for this phenotype was quite low and ranged from 0.0 to 13.7%, except for cough, which was prevalent in 24.4% of the children at 6 years of age only. The "transient cough" phenotype (36.4%) had cough in the first 2 years (100%) but had recovered by age 6 to reach 25.2%, almost identical to the "reference" at that age. In the "eczema/cough" phenotype (12.3%), 94.8.% of the children experienced itchy rash in early life, which decreased by age 6 to 33.0%. Cough was more frequent than in the reference group during the first 2 years and was still reported in 34.2% of these children at age 6. The "wheeze/cough" phenotype showed a very substantial decrease in the prevalence of wheezing, from 77.3% at age 2 to 22.7% at age 6. Cough also decreased in this phenotype, from 95.5% at age 2 to 29.1% at age 6. The itchy rash prevalence in this phenotype was Flow chart of the PELAGIE mother-child cohort. The absence to follow-up at different time points figures on the left (Inclusion, 2-year follow-up) and the right (6-year follow-up) side of the flow chart. Complete data for this analysis were available for 935 mother-child couples K. Apel et al..3 at age 1 and 2 and 20.0% at age 6. In the "mixed" phenotype, all respiratory and allergic symptoms exceeded prevalence rates from the reference group. Eczema symptoms were more frequent than in any but the eczema/cough phenotype, and wheezing more frequent over time than in any but the wheezing/cough phenotype. Cough at age 6 was more prevalent (42.0%) than in either the wheeze/cough or transient cough phenotype.
The description of other respiratory and allergic health outcomes is available in Supplementary. At age 2, 64.9 and 72.9% children in the eczema/cough and wheeze/cough phenotypes, respectively, had had otitis media. Most of the children who had ever been awakened by breathlessness belonged to the wheeze/cough phenotype (15.5%), and most of those who had been awakened by respiratory discomfort belonged to the mixed phenotype (41.9%). The mixed phenotype also contained the most children with hay fever (27.0%) and other allergic rhinitis (24.3%).summarizes the associations between physiciandiagnosed diseases and phenotypes. At the age 6 of the child, 107 (11.4%), 212 (22.6%), and 57 (6.1%) parents declared that a doctor had already diagnosed their child with asthma, eczema, or food allergy, respectively. The highest proportion of the 107 children with physician-diagnosed asthma belonged to the wheeze/cough phenotype (29%). Comparably 29.3% of the 212 children diagnosed with eczema were allocated to the eczema/cough phenotype, but 23.1% were defined to the reference and the transient cough phenotypes. At time point 2, 8 (1%) of children diagnosed with food allergy had been attributed to the eczema/ cough phenotype. Logistic regression models showed that the wheeze/cough phenotype was associated only with physiciandiagnosed asthma. Both the eczema/cough and mixed phenotypes were associated with all three physician-diagnosed conditions of interest (asthma, eczema, and food allergies). The transient cough phenotype, however, did not differ significantly from the reference group for any of these conditions. Other respiratory and/or allergic symptoms reported by parents are shown in.
Main results Factors associated with phenotypes are shown in. The risk of transient cough and eczema/cough phenotypes were lower in children born to mothers aged 35 years and older and in those with exclusive breastfeeding for 3 months. Family related factors associated with the wheeze/cough phenotype were multiparity, maternal smoking at inclusion, and family history of allergy; the latter was also associated with the mixed phenotype. Children born to mothers aged at least 35 years had a reduced risk of belonging to the wheeze/cough phenotype. An environmental determinant associated with this phenotype was exposure to glues for indoor renovation activities in early life. Children with the wheeze/cough phenotype were at reduced risk at age 2 of sleeping on a mattress bought at least 3 years earlier. The only environmental determinant associated with the mixed phenotype was renovation-related use of alkyd/acrylic paints in early life. The sensitivity analysis (for preterm birth >37 weeks of gestation and birth weight < 10th percentile) did not modify the adjusted model.
# Discussion
This study identified five phenotypes of respiratory and allergic symptoms over the first 6 years of life in a population-based mother-child cohort. Phenotypes differed in their physiciandiagnosed asthma/allergic diseases and parent-reported comorbidities. We observed an association between some avoidable domestic environmental determinants with respiratory/allergic phenotypes from birth to age 6, specifically, early postnatal exposure to glues and paints.
One of this study's strengths is its use of an unsupervised statistical method to describe the coevolution of a set of respiratory and allergic symptoms. Unsupervised clustering techniques are particularly appropriate to describe patterns of evolution in a set of variables measured repeatedly like in birth Description of the five-cluster solution obtained in the cluster analysis. Each graph corresponds to one cluster. Asthma and allergy symptoms (rash, wheezing, cough, and rhinitis symptoms) figure as prevalence at each time point (1, 2, and 6 years) and serve to characterize each cluster. The "reference group" had overall low asthma/allergy symptoms cohorts. KmL3D allows clustering individuals on several trajectory variables jointly and takes into account their possible coevolution while liberating the observer from clinical a priori. Meanwhile, some subjective choices had to be made, basically the number of clusters tested, which was of necessity due to our sample size. We were able to explore symptoms in very early and late preschoolers over a relatively long period and past age 5, the threshold when wheezing starts to be an indicator of asthma. The use of binary data in the k-means algorithm might have led to clustering errors in the construction of phenotypes, but these errors are likely to be nonsystematic.
The most important limitation involves the measurement of symptoms on which phenotypes were built. Wheezing was assessed with items derived from ISSAC (The International Study of Asthma and Allergies in Childhood) questionnaire, 3 validated on 7 year old children. It is uncertain what kinds of respiratory noises parents qualify as wheezing.The questionnaire is nonetheless widely used in epidemiologic studies in younger childrenand was used exclusively in two 4,6 of the three 4-6 articles to which we compare this study. Eczema symptoms might be overestimated. Because we did not assess typical localizations of itchy rash, and did not dispose of any eczema-medication data, noneczema symptoms such as xerodermia might have been assessed and could explain a certain degree of misclustering of children with eczema allocated to the reference group and the transient cough phenotype. Different cough measures were used for the follow up of the 2 and 6 year olds. Since we did not dispose of any specific item on cough in the 2 first years of life, we supposed that parents who declared having had a child with bronchiolitis/bronchitis, had had a child with cough. This could be source of misclustering. Data on asthma/allergic symptoms were not collected between ages 2 and 6, so that we lack data about relevant time periods in symptom evolution, in particular, age 3.Correct asthma medication might lead to less symptoms what for our symptom trajectories might be imprecise. But our phenotypes were consistent with most of the physician-diagnosed conditions and other respiratory and allergic symptoms, which reinforces their likelihood. Asthma medication was also consistent with phenotypes since it was more frequent in the wheeze/cough and mixed phenotypes. Misclustering due to under-declared of correctly treated symptoms are unlikely to be differential and would, as the case may be, let the association observed tend towards zero. We, therefore, do not believe that it affected associations observed.
Another limitation is attrition. The questionnaires were selfadministered, sent by the postal service, and lengthy (>1 h for completion) due to the broad research scope covered by the cohort. Each of these factors may have discouraged participation and induced bias in association estimates. Because the outcomes measured in this paper do not concern behavioral variables, we believe, in accordance with Greene et al.that follow-up bias is likely to be small.
Three previous studies have used unsupervised techniques to study multiple longitudinal respiratory-and allergy-related symptoms and to identify phenotypes.Comparison appears thorny; not all studies used only the ISAAC questionnaire, 5 they considered different variables for phenotype identification, and health outcomes were not measured at the same time points. As Pinart et al.recently underlined, efforts to standardize the reporting of allergic disease phenotypes in cohorts are needed. However, similarities are apparent, especially for phenotypes. Our wheeze/cough phenotype was exclusively associated with physician-diagnosed asthma and similar to the transient wheeze phenotype identified by Rancière et al.Likewise, two of the studies 4,5 identified a group with high levels of eczema symptoms and low levels of other disorders, strongly associated with physician-diagnosed eczema.Similarities also exist between the persistent sensitization-associated cough/rhinitis phenotype in Rancière et al.and our mixed phenotype: cough and rhinitis were predominant, moderate wheeze and eczema symptoms remained, and it was associated with all the physician-diagnosed conditions tested (asthma, eczema, and food allergy).
The prenatal and early postnatal period is an important window of vulnerability for developing immune and respiratory systems, which are sensitive to environmental chemicals.Our wheeze/ cough phenotype was associated with early-life exposure to glue, and Rancière et al.observed associations between transient wheezing and postnatal exposure to new particle board furniture. Our mixed phenotype showed associations with early exposure to alkyd/acrylic paints used for interior home redecoration. Glues, particle board furniture, and paints can emit chemical compounds that, inhaled or by direct skin contact, can induce adverse health effects. Early exposure to chemicals used for indoor renovation activity is suspected of harmful effects, including the development of asthma 11,21 and eczema.We did not, however, measure these compounds, and their effects are not clearly understood.
The causal relation between postnatal exposure to ETS and respiratory outcomes in children is well established.It was studied in two of the papers cited above 4,6 and observed by Rancière et al.in two of their phenotypes, rhinitis and dermatitis. A lack of statistical power due to our smaller sample might explain the absence of association in our study.
Exposure to bleach was not associated with any phenotype. Studies of household cleaning products and their effect on asthma/allergy in children are scarce and conflicting. Maternal exposure to a set of domestic chemicals including bleach has been associated with persistent wheezing in offspring at age 3.More recent focus on the specific use of bleach during pregnancy has shown no association with wheezing at age 1, 25 while a crosssectional study found 10-13-year-old children living in homes cleaned with bleach at least weekly less likely to have asthma, eczema, or sensitization to aeroallergens.No phenotypes were associated with exposure to pets at home, whose respiratory/ allergic health effects on children are currently the subject of debate. 9,27 Surprisingly, children from our wheeze/cough phenotype were less likely to sleep on a mattress supposed to be laden with house dust mite (HDM) than the reference group. Rancière et al.observed a positive association between their dermatitis phenotype and the presence of a used mattress (older than 3 years at birth). Reverse causation due to the purchase of a new mattress in symptomatic children explaining our surprising association is improbable because it concerned only 19 children. HDM are thought to be implicated in the development of allergic asthma in children at high risk,with a nonlinear relation: both very high and very low exposure to HDM are probably protective against asthma and HDM sensitization at age 5.We treated the age of the mattress as a surrogate for HDM exposure in early life, but we cannot quantify the actual exposure, nor compare our sample to a population at high risk for asthma/allergy.
We do not believe that phenotypes identified represent clinical entities that clinicians should relate to. The development and expression of asthma/allergic disorders depend on geneenvironment interactions. From the point of view of the primary care giver, if one focuses on potential prevention targets, to date, genetic predisposition to respiratory/allergic diseases is inaccessible to intervention. This paper aimed to describing respiratory/ allergic symptoms in the first 6 years of life. It was quite correctly in line with the currently admitted concept of allergic march. We generated hypotheses on the potentially harmful effect on respiratory and allergic health in children aged 0-6 of some chemical exposures that are directly related to parental behavior (renovation and redecoration) and thus avoidable. If confirmed in observational studies and interventional home-based trials, environmental counseling by primary care givers, because of their close relationship to families and potential knowledge of domestic behavior, could be particularly useful in prevention strategies for respiratory and allergic diseases in childhood.
# Methods
Study design, participants, and settings PELAGIE, described previously, 30,31 is a population-based mother-child cohort that enrolled 3421 women in pregnancy (before 19 weeks' gestation) in Brittany between 2002 and 2006. Data were collected at birth and when the child was 2 and 6 years old. We studied a subgroup of 935 mother-child pairs for whom respiratory and allergic data were available at both follow ups . Participants provided written informed consent for data collection, and a national INSERM (Institut National de la Santé Et de la Recherche Médicale) ethics committees approved the study procedures.
Methods of data collection Data were collected by a self-administered questionnaire filled out mostly by the mother at inclusion and when the child was aged 2 and 6 years, as well as medical reports at delivery (obstetric data collected by midwives and physical examination of the newborn by a pediatrician).
## Variables
The ISAAC questionnaire 32 assesses four respiratory/allergic symptoms: wheezing, eczema, rhinitis, and cough. Wheezing was measured at ages 1, 2 (2 years questionnaire), and 6 (6 years questionnaire). At age 1 and 2, wheezing was defined as at least one episode of whistling in the chest or K. Apel et al.
an asthma attack. Children not reported to wheeze but with physiciandiagnosed asthma were also considered wheezers. Wheezing or whistling in the child's chest and/or a wheezy sound there during or after exercise during the last 12 months defined wheezing at age 6. Eczema was defined at ages 1, 2, and 6 by a positive answer to the question: "Has your child ever had an itchy rash which was coming and going in the last twelve months?" Rhinitis symptoms were assessed for the first 2 and the sixth years of life and defined by positive answers to ISAAC questions: "Has your child ever had a problem with sneezing, or a runny, or a blocked nose when he/she DID NOT have a cold or the flu?" and "In the past 12 months, has this nose problem been accompanied by itchy-watery eyes?" The PELAGIE questionnaire did not include any specific item to assess child's cough at the 2-year follow-up. Cough during the first 2 years of life was defined by a positive answer to the question: "Since birth, has your child ever had bronchitis or bronchiolitis?" At age 6, cough was the occurrence of dry night cough in the past 12 months.
To confirm phenotypes, we assessed parent-reported comorbidities such as otitis media at age 2 or ever awakening due to breathlessness or respiratory discomfort and hay fever or other allergic rhinitis at age 6. They were also described against physician-diagnosed asthma, eczema, and food allergies as at the 6-year follow-up we also asked the parents whether a physician had ever diagnosed those disorders.
We assessed indoor environmental factors known or suspected to be involved with respiratory and allergic symptoms.
At age 2, we assessed: ETS, defined as maternal and/or paternal consumption of at least one cigarette per day at home since birth (yes/ no); whether child spends time in rooms cleaned with bleach (yes/no); renovation activities in the dwelling since birth that used glues or either alkyd or acrylic paints (yes/no); usual home heating, defined by individual heating by at least one of the following: gas/fuel oil/wood or chimneys (yes/no).
At age 2, we assessed: pets at home, that is, at least one of the following: cat/dog/rodent/bird/farm animal (yes/no) since birth; daily airing of the child's bedroom (yes/no); and HDM exposure (yes/no), which was evaluated indirectly by the age of the child's mattress considered old if purchased at least 3 years earlier.
# Statistical methods
Phenotype identification relied on the joint temporal evolution of wheezing, cough, eczema, and rhinitis symptoms. We used the KmL3D unsupervised clustering method 33 designed to study the joint longitudinal evolution of several variables over time. This method applies the k-means algorithm on longitudinal data (it divides the original sample in clusters of individuals with similar symptom trajectories). Moreover, several outcomes/symptom trajectories are considered simultaneously to compute interindividual distances. Each individual is represented by a matrix with his/her levels of different symptoms (lines) at different time points (columns). The multidimensional Euclidean metrics was used to calculate interindividual distances (Euclidean norm of the distance matrices). As the k-means algorithm requires predefining a number of clusters, we ran the kmeans algorithm 20 times to obtain, respectively 3, 4, and 5 clusters. These numbers of clusters, based on an a priori choice were reasonable regarding our sample size and expected interpretable clinical phenotypes.The choice of the optimal clustering was based on the Davies Bouldin cluster validity index.For each clustering (different numbers of clusters), a proximity indicator reflecting the average within-cluster distance/betweencluster distance is computed. The optimal clustering minimizes this proximity indicator. The resulting clusters were then interpreted as clinical phenotypes based on the mean symptom trajectories observed in each cluster and our medical expertize. The complete cluster analysis was performed with the R software package KmL3D.Associations between phenotypes and covariates were estimated with univariate polytomous logistic regression models. We used single random imputation by the mean or median when missing values for covariates did not exceed 5.0%. The final regression model included the lifestyle factors associated with phenotypes in the univariate analysis and all the environmental factors, regardless of the strength of their association with phenotypes. Multivariate regression models were adjusted for maternal body mass index (BMI ≥ 30 kg/m 2 : yes/no) and age at inclusion, parity, maternal smoking habits at inclusion (at least one cigarette per day: yes/ no), family history of allergy (mothers' history of allergy and/or maternal grandparent's history of asthma: yes/no), child's gender, vaginal delivery (yes/no), and exclusive breastfeeding for at least 3 months (yes/no). A sensitivity analysis, included preterm children and those with a birth weight <10th centile expected for gestational age. The prevalence of these two characteristics was low and thus not appropriate for use in the adjusted regression model. Logistic regression analyses were performed with SAS 9.4, and results were expressed as odds ratios and their 95% confidence intervals (95% CI). The level of statistical significance was 0.05. Our data were analyzed in 2017.
## Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
## Data availability
The data from the PELAGIE cohort are not freely available as they are the property of our research institute. Access might be asked by contacting the corresponding author.
## Code availability
Codes used for the cluster analysis can be accessed by contacting the corresponding author. |
Clinical Observations on Functional Nervous Disorders
London, 1864. pp. 585. 2. Practical and Pathological Researches on the Various Forms of ParalysisLondon, 1864. pp. 585. 2. Practical and Pathological Researches on the Various Forms of Paralysis
THE BRITISH AND FOREIGN
the minute histological investigations of Beale, relative to the nervous centres ai:>d the nerves, are rich in materials for the use of the practical physician. The two works now before us are instances of the application of the results and hypotheses arrived at by those and other Neurologists in the interpretation of the principal lesions of the Nervous system, and afford gratifying evidence of'progress made in the diagnosis, prognosis, and rational treatment of such lesions.Br. Handfield Jones restricts his observations to such nervous disorders as are called functional, omitting the results of manifest organic lesion. It seems to him " a vain dispute, whether in strict accuracy there are or are not any such disorders. The probability is that there are not?that in all morbid action the cells and fibres of the organs Undergo some molecular change from their perfectly normal condition. ^ is, however, perfectly certain that there are very grave disorders in ^vhich the most careful scrutiny fails to detect any actual change, in "which complete recovery is perfectly possible, and in which the ' juvantia' are such as operate more in modifying the power of the organs than their texture." These opinions generally most physicians will be ready to endorse. -There are undoubted functional derangements of organs whose histology ls IUuch better understood and more readily demonstrated, whether in 68-xxxiv. *1
Brown-Sequard, Claude Bernard, and Lockhart Clarke, and the minute histological investigations of Beale, relative to the nervous centres ai:>d the nerves, are rich in materials for the use of the practical physician. The two works now before us are instances of the application of the results and hypotheses arrived at by those and other Neurologists in the interpretation of the principal lesions of the Nervous system, and afford gratifying evidence of'progress made in the diagnosis, prognosis, and rational treatment of such lesions.
Br. Handfield Jones restricts his observations to such nervous disorders as are called functional, omitting the results of manifest organic lesion. It seems to him " a vain dispute, whether in strict accuracy there are or are not any such disorders. The probability is that there are not?that in all morbid action the cells and fibres of the organs Undergo some molecular change from their perfectly normal condition. ^ is, however, perfectly certain that there are very grave disorders in ^vhich the most careful scrutiny fails to detect any actual change, in "which complete recovery is perfectly possible, and in which the juvantia' are such as operate more in modifying the power of the organs than their texture." These opinions generally most physicians will be ready to endorse. -There are undoubted functional derangements of organs whose histology ls IUuch better understood and more readily demonstrated, whether in 68-xxxiv. *1 Reviews.
[Oct.
health or disease, than the cerebro-spinal system. There are abnormal states of circulation, and therewith, as concomitant conditions, alterations of innervation without any structural changes of organs.
And, in the case of the nervous system, such functional variations are even more conceivable than in other viscera, whether regard be had to its structural peculiarities or to the results of function?the nerveforce generated. The tendency of modern research is to assimilate nervous action to electricity, and the movement in this dii'ection will derive increased impetus in the minds of perhaps many from the observations of Dr. Beale1 of the non-existence of the presumed a-polar and "uni-polar nerve-cells, and of the constant presence of two fibres to each nerve-cell to form a complete circuit. Assuming, therefore, the homology of the nervous with the electric force, we may, cl priori, argue in favour of the occurrence of disturbances and variations in that force?i.e., of the existence of nervous disoi'der, arising from inaction and overexertion, and particularly from alterations in the velocity, pressure, quantity, and quality of the blood supplied to the nerve elements. Ln derangements of the sort in question, "the most careful scrutiny fails (and we believe must ever fail) to detect any actual change"
?at least, until the disturbed balance of forces has eventuated in alterations of the mechanism involved.
After what lapse of time disordered action will induce organic change, is not determinable, but experience indicates that it may go on for a lengthened period, and yet ultimately yield to such " juvantia as operate more in modifying the power of the organs than their texture." Thus, for example, mental disorder may exist for months, or even for years, and suddenly vanish under the influence of strong emotion or of coincident bodily disease; or otherwise may, as an intermittent malady, decline and disappear repeatedly, to be as often renewed.
Nevertheless, though compelled to admit the occurrence of functional disorder without organic alterations of nerve-tissue, it would imply an indifference to truth to assume in any given case the absence of such changes. The improved methods of research by chemistry and the microscope now applied to the examination of nerve-substance, show clearly enough that serious lesion may subsist which the unaided eye, and even at times also the microscope, without the assistance of chemical reagents, fail to detect. The careful observations of diseases of the spinal cord made by Mr. Lockhart Clarke may be cited in illustration of this fact, and we are disappointed that those observations, and the lessons deducible from them, are not made use of by Dr. Meryon in his history of spinal paralysis. Without microchemical investigation of the nerve-tissue, it is impossible to pronounce with certainty on the absence of organic lesion in many cases of nervous disorder, where no morbid changes reveal themselves to unassisted vision.
The difficulty of determining the presence or absence of actual lesion of the nerve-tissue, so great even after death, and when that tissue can 1 Philosophical Transactions, 1863, p. 543-571. We refer the reader to a fuller notice of Dr. Beale's paper, at a subsequent part of this number.
1864.] he subjected to minute scrutiny, must be still greater during life, owing to the frequent obscurity of diagnosis, the prevalence of similar symptoms under varied morbid states, the want of an intimate knowledge of the function of various parts, and the untangible character of nervous products. Dr. Jones's task, therefore, of separately treating of nervous disorders subsisting without organic lesion, is one beset with difficulty, and its performance must partake of imperfection in proportion as the diagnosis of his illustrative cases fails to discover their actual pathological condition. To aid in the discrimination of simple disturbance ?f function from the consequences of positive alteration in structure, he refers especially to the effects of remedies employed, or, in his own vvords, to " juvantia," which " operate more in modifying the power of the organs than their texture," such as galvanism, strychnia, and, in general, "nerve-tonics." Between a case of functional paralysis and ?ne of paralysis from organic lesion there is, Dr. Jones argues, a wide interval. Strychnia and galvanism, which will probably benefit the former malady, will in all probability make the latter worse; certainly will not cure it. Yet while contending for this wide interval in the typical cases assumed in illustration of his subject, he is prepared to admit that "numerous instances of more or less mixed character intervene. Inflammatory disease is, from one point of view, organic ] from another functional. It commences essentially as the latter j it ends as t!Je former." This is mucli the same thing as saying that the distinction between functional and organic nervous maladies is of a shadowy character, and has no definite existence in nature, cognisable by human observation. Notwithstanding, however, this inability of distinguishing with precision the one condition from the other, practice presents o us numerous nervous disorders amenable to treatment deserving the *nost attentive study. It is such disorders that Dr. H. Jones makes the subject of his treatise; and he finds no lack of matter, as nearly ^ printed pages testify. Indeed, some of the maladies?as, for instance, epilepsy?furnish of themselves, as the 'Publishers' Circular' . eiQonstrates, ample material for learned and long dissertations, and, Jn a general work of the sort under notice, claim, by their importance, space which we should have preferred to see occupied with a more Copious history of nervous disorders less written about specially. The disorders considered by Dr. Jones are viewed, as the title of the w?fk intimates, from a clinical point of view, and therefore are well exemplified by cases occurring in the author's own practice, or placed on record by other physicians. The arrangement adopted " is merely topographical .... commencing with the encephalon, including the cerebral emispheres, mesocephale, cerebellum, and medulla oblongata, we shall face, the nose especially, became of a deep red, and all the vessels of the face much congested. There was extreme photophobia. It was half an hour before the hyperemia began to subside. The irritation in this case evidently was reflected from the branches of the first division of the fifth nerve supplying the lids, on to the vaso-motor nerves of the arteries supplying the skin of the face, which, in consequence of the morbid impression became dilated?not contracted, as they normally should according to the law of reflex action. This was a marked example of inhibitory action. . . . The pathology of these cases is no doubt the same; a nerve of special sense, a muscuto-motor or a vaso-motor being paralysed according to the direction which the irritation happens to take." (p. 13.) The result of the nervous stimulus?i.e., whether paralysing or exciting ? also depends very much on the condition of the nervous centre which is affected.
We have quoted at length these illustrations of what is meant by inhibitory influence, inasmuch as it is a recent hypothesis, and in the main agrees with what has been taught under the term reflex paralysis. The latter mode of explanation involved the admission that the paralysis was due to ansemia of the nervous centre from which the affected nerves started, "produced by the reflection of the original irritation on the vasa-motor nerves supplying the arteries.' To this mode of interpretation Dr. Handfield Jones objects that:
''(1st.) It is difficult to suppose that a spasm of reflex origin should be limited to such a very small extent of vessels as would be involved in some instances?e.g., palsy of one of the sixth nerves, ptosis of one eye. (2nd.) It ls. almost impossible to believe that a contraction of vessels should be so persistent as the hypothesis requires. Can we suppose, in the case of amaurosis above cited, that the arteria centralis retina; was spasmodically occluded for thirteen months ? (3rd.) It has been found by Gull that irritation of the renal nerves does not cause contraction of the vessels of the spinal cord, nor paralysis of the lower limbs, as Brown-Sequard stated in explanation of the paraplegia trom renal disease. (4t.li.) In some cases of paralysis from exposure to wet and cold, the paralysis continues long after the exciting cause lias ceased, and ls removed by stimuli applied to the sensory cutaneous surface. Here the paralysis must be non-organic ; and yet it can scarcely be supposed to depend ?n anaemia of the centres resulting: from arterial spasm. On the other hand, *t is intelligible that the nerve-cells might be thrown into a state of enfeebled action by the cold, &c., from which they could not easily recover. These cases, though not typically inhibitory, seem to me very illustrative of the nature of the morbid action." (p. 14.)
There is a certain weight in these objections generally to Dr. Brown-Sequard's hypothesis. The fourth one nevertheless has, per se, little importance; for if spasm of the arteries of a nerve-centre can produce anaemia and consequent paralysis of its efferent nerves, it does not necessarily follow that the paralysis should disappear forthwith when the arterial spasm is removed and the healthy nutrition of the centre restored, for this reason, that the prolonged anaemia will in all probability have induced an altered state of the nerve tissue incompatible ^ith normal function, which may persist for some length of time. If this be so, then the only distinction Dr. Jones can discover between inhibitory and reflex paralysis?viz., that in the former the " paretic" state may sometimes " persist for an indefinite time after the cessation Reviews.
[Oct." of the cause which has morbidly affected it," whilst in the latter it increases or diminishes with the irritation and ceases with it?becomes, as ifc appears to us, futile, (p. 13.) Further on, at page 139, Dr. Jones reopens the question, and states that reflex paralysis is a rare affection, and that he has never met with an instance.
He would restrict the term to " Cases where the paralysis is evidently dependent on a persistent irritation, increasing when this is increased, and vice versa. It is characteristic of true reflex paralysis, that removal of the irritation proves curative, while all other means fail. It is certain that no true case of reflex paralysis would be benefited by strychnia or galvanic excitation of the nerves or muscles of the affected part, nor by stimulating applications to the cutaneous surface. These are appropriate to the paretic state, and form by their success a good test of its presence."
"Without discussing the correctness of the latter assertion respecting the inutility of remedial agents such as those mentioned in cases of reflex paralysis, we will only remark that the assigned characteristic of true reflex paralysis applies with equal force to the cases Dr. Jones has adduced as examples of inhibitory paralysis; as, for instance, that of paralysis of the external rectus from whitlow of the thumb, and that of amaurosis from dental irritation.
So far, therefore, as the author's arguments and illustrations are concerned, we find no sufficient basis of distinction between inhibitory and reflex paralysis; and we prefer, on the whole, to retain the latter term to express the symptomatic condition, whether due to anaemia of the nerve centre, as supposed by Brown-Sequard, or to a " paretic" (enfeebled) state of its nerve-cells, as Dr. Jones imagines, produced by irritation of an incident nerve. What difference subsists is only between the hypotheses projiounded in explanation of the resultant phenomenon?the paralysis?and it remains to be proved whether this may not at one time be due to anaemia from vascular constriction, and at another to a paretic state of the nerve-cells of the irritated centre. Iu either case the term reflex paralysis appears applicable.
The author shows a predilection which we do not equally share, for foreign words and expressions, most, if not all of which, supplant genuine English ones, or could be replaced by English derivatives. The advance of science no doubt necessitates new terms, whilst the character of our language renders compound words less possible than does the German ; we are, therefore, compelled to coin many words from the Greek and Latin tongues, and usage familiarizes us with many other such for which the plea of necessity cannot be urged. Consequently, a critic may be indulgent when these classic derivatives are not too thickly strewn in the pages under notice; but it behoves him to dissuade from the excessive use of such, as in the work now before us.
The author employs the word paresis and the adjective paretic largely throughout his work. It is a pure Greek word, signifying relaxation or weakness, and, so far as we are aware, was first used in modei-n times by Dr. Ernst Salomon, a Swedish physician, to designate the peculiar form of paralysis among the insane, mostly known as "general paralysis," but in which actual motor paralysis, particularly in the earlier stages, is so slightly marked as to be overlooked by those little acquainted with the disease. In this general paresis of the insane the mental symptoms are more marked than the paralysis of motion and sensation, the paralysis being, in matter of degree, incomplete or imperfect, an apparent state of weakened nervous power.
In commencing his chapter on " Cerebral Paresis," Dr. Handfield Jones adopts the word pretty much in this signification, remarking: " By this term I mean a state in which, without demonstrable orgauic change, there is greater or less enfeeblement of the functional power of the brain." But paresis, as a term, is not limited to conditions in which organic change is absent} for in the general paresis of the insane such change is often very pronounced in the nervous centres; and in various parts of his book the author has used it, when speaking of paralysis due to organic lesion. Indeed, he has loosely applied the word, * substituting it, we think, without good reason, for paralysis, and vice versd. Even if rightly restricted in its use, it expresses nothing more, except in the case of general paresis of the insane, than cerebral or nervous exhaustion 01* debility. The subjects treated by Dr. Handfield Jones in the work before us ai'e necessarily very numerous, for functional nervous disorder may have its seat in any organ or part of the body. Hence we find chapters on cerebral and spinal hypersemia aud aussmia, on cerebral and spinal paresis, on cerebral excitement and delirium tremens, on tetanus, catalepsy, and epilepsy, on spasmodic and neuralgic affections and neuroses of numerous parts, 011 hysteria and malarioid disorder, &c. The chapter on the last-named subject will be read with much interest and instruction. Thereis an almost endless variety of disorders generated by malaria through the medium of the nervous system. In connexion with these, the author passes under review many " of the forms in which malarious disease appears-in this country maladies which are scarcely recognised in ordinary text-books."
?Tlie last chapter of the book is occupied with a discussion of the lerapeutical properties of those drugs which are employed in nervous 'isorders. It is one of value to the practitioner on account of the amount of information brought together in its pages. ^ The final essay of the author bears the quaint title of the " Yindeiniatio, the harvest or vintage, implying the summary, or the essence ol his teachings.
The scope of Dr. Meryon's work is much more limited. He processes to deal only with the various forms of paralysis, and accordingly Proceeds, after a good introductory sketch of the anatomy and physiology of the nerve-centres and nerves, to pass under review the several lesions of the spinal cord and brain generally recognised as causes of paralysis. The two chapters on these lesions are followed, ty others 011 " Paralysis from Blood-poisoning, " Paralysis Irom Reflex Action," and on " Progressive Forms of Paralysis. But though the author has produced a readable and instructive book, it has not quite t^c qualities we look for in a special treatise. Its matter might, in "$92
Reviews.
[Oct.
"our judgment, be better arranged, and the history of most of tlie lesions considered might be more complete. Paralysis from dislocation and fracture of the spinal column has a place, but the more common form consequent on caries receives no attention. Cerebral apoplexy and red and white softening are described, but abscess of the brain escapes discussion. Apoplexy constitutes an important chapter, but the account of it is wanting in precision and clearness. He comprises under the title cases in which there is cerebral haemorrhage, and others in which there is only congestion. He narrates the case (Case 31) of an American gentleman, F. B., aged sixty-seven, who got an apoplectic attack, with sudden head-ache, vomiting, and loss of consciousness for a time, after having pi*eviously suffered " much bodily fatigue and very great anxiety concerning his country.*' The attack was followed by slight drawing of the mouth to the left side, drooping of the right eyelid, deafness of the right ear, disabled lips, a constant dribbling of saliva, considerable difficulty in swallowing, and distressing paroxysms of dyspnoea on falling asleep. We will quote most of the author's comments on this case, premising that we cannot see in it, as he does, an example of reflex paralysis, and that, as far as We can apprehend, his meaning is obscurely expressed :
"As an instance of diffused influence of emotional impression, this case would perhaps be more correctly placed under the head of reflex paralysis; for this latter may have had a centric origin and been attributable to a source within the cerebrum, just as the opposite phenomenon of excited muscular action of automatic form and cerebral origin is often induced by insanity, but inasmuch as the paralysis resulted from effusion or mechanical pressure, probably in the medulla oblongata near to the nucleus of the facial nerve, 1 have referred to it here. " The case is furthermore interesting as touching a remark by Dr. Laycock in a paper On the Reflex Function of the Brain,' .... that emotions of the mind act principally on the excito-motory system by relaxing the sphincters, and inducing vomiting, micturition, and defecation. In my patient, however, I am disposed to think that the nucleus of the pneumogastric nerve was affected; and it is a matter of some importance, seeing that emetics are sometimes given with a view to relieve a disordered condition of stomach. Three emetics were, in fact, given to F. B., but without arresting the disposition to sickness."
The following case, quoted from Trousseau, of a youth, aged seventeen, who got facial paralysis after lying asleep, on a cold day, upon a heap of stones, having been previously in a state of perspiration, is instructive in itself, but we do not understand why it should be inserted in a chapter on apoplexy. It may be urged that it is given in order to illustrate the diagnosis between such a form of idiopathic paralysis and paralysis from cerebral lesion; but if so, this purpose is very indifferently fulfilled by it; for after the description of the symptoms of the case is concluded, Dr. Meryon subjoins these remarks: "These ?cases of idiopathic facial paralysis generally terminate favourably, and the more rapidly they come on, the quicker do they get welland then proceeds thus: " The power over the orbicularis palpebrarum, therefore, is the only perceptible difference; and Dr. Cazalis, of the Salpetriere, has described the imperfectly palsied lid as a diagnostic sign of facial paralysis dependent 011 mtra-cranial haemorrhage, and the perfectly-fixed lid as indicative of affections of the nerve only."
In the chapters on red and white softening, the details of diagnosis ttight be more complete and better arranged. In the middle of his dissertation on red softening, he enters (p. 114) into a discussion "whether softening of the cerebral commissures between the two cerebral hemispheres may not be a cause of intellectual disorder, and treats the question without regard to the nature of the softening.
The minute pathology of red softening is all comprised in the following sentences: " The pathological change which is known as red softening is essentially a Capillary haemorrhage into the surrounding substance of the brain, and the 0I% difference between it and the blood-clot in apoplexy is that the extravasated blood is infiltrated into the cerebral substance, with which it is intermixed in the one case, whilst it forms to itself a cavity, and exists as a mass 0r clot in the other."
There is no attempt here or elsewhere to discriminate the varieties ?f cerebral softening in their minute and intrinsic details; indeed, all the varieties are not even enumerated. The description of red soften-*ng above quoted applies rather to the consequences of cerebral contusion, and fails to convey an accurate notion of the results of cerebritis or inflammatory softening. The fallacious notion that red softening consists only of extravasated blood is again repeated (p. 117), where jts treatment is discussed, and it is stated that " the principal indica-J?n is to reduce the violence of arterial action, and thereby to prevent the further escape of blood."
The section on Induration of the Brain contains no notice of the researches of Virchow, Albers, and others, relative to the mode and md of deposit of abnormal histological elements within the cerebral lssue, or, to borrow their term, the parenchymatous infarction of the fain.
The whole morbid anatomy of induration presented to Us is comprehended in the following quotation, the clearness of yhich is unfortunately beclouded by the sentence marked by ourselves italics: " This form of induration (no form, by the way, has been previously mentioned [Rev.]) is generally the result of chronic inflammation; but an opposite condition of the intra-cranial substance is recognised by pathologists as resulting from great and exhausting exudations in other parts of the body?
Such, for example, as frequently occurs in the course of typhoid fever, or after Scarlatina, when the brain-substance, like other tissues, may be drained of its element, and become hard and tough." (p. 125.) In conclusion, we would remark that, in the remaining sections of book, observations occur deserving consideration ; and the illustrate cases, derived from the author's own experience or from the Writings of others, will always possess value to the student; but we ^niiot conscientiously say that the work satisfactorily represents the Reviews.
[Oct;; present state of pathological knowledge respecting the several forms of paralysis.
The treatise of Dr. Handfield Jones will no doubt find its way into the libraries of medical men by reason of the true practical character of its contents, and the importance and the frequently perplexing nature of the disorders of which it treats. The chapter on the genei'al pathology of nervous diseases is worthy of attentive study, and need be well digested before the subsequent portions of the work are read. The several morbid conditions discussed are pourtrayed by the record of cases, the symptoms, course, and treatment of which are fully detailed and largely commented upon.
Proceed to the spinal cord, and thence to the several nerves or nervous districts, which are found by experience to be most prone to disorder." (P-4.) Such a topographical arrangement may be practically convenient, but has no other claim to attention, for many of the morbid conditions discussed as local affections, such as cerebral excitement, headache, vertigo, chorea, &c. &c., are often or mostly the consequences general bodily illness. We are, indeed, convinced that Dr. Jones
Reviews.[Oct, himself assigns no importance to the arrangement, but has used it simply from the inability of medical science to supply the basis of a classification founded on the exact lesions of the nervous system, of which the maladies recognised provisionally as morbid entities, are nothing more than the outward manifestations.After an introductory chapter, the author devotes a second to the general pathology of the nervous system, in which he particularly elucidates the nature of "inhibitory" nerve-phenomena, and submits to criticism some parts of the hypothesis of reflex paralysis, as advocated by Dr. Brown-Sequard. He coincides generally with Lister in asserting that inhibitory action is not restricted, as Pfliiger taught, to a certain set or system of nerve-fibres, but " that one and the same afferent nerve may, according as it is operating mildly or energetical^, either exalt or depress the functions of the nervous centre on which it acts." This doctrine, however, Dr. Jones would modify, "so far as to believe that it is not the energetic operation of an afferent nerve that causes inhibitory action, but its being injuriously affected by some impression made upon it. The enfeebled state of the nerve itself, or of the centre to which it proceeds, or the severity or malignity of the impression, may give rise to the familiar effect." (p. 10.)This so-called inhibitory influence plays an important part in physiological and pathological phenomena, and as an hypothesis elucidates numerous anomalous cases of disease.What is meant by it will best appear from some examples adduced by Dr. Jones:?"O. J., aged thirty-seven, got a whitlow on the last phalanx of the left thumb. The lymphatics were inflamed, and the axillary glands swollen; the whole arm was very painful. While the limb was in this state, one morning he found that he saw double, and had a squint in the left eye. At the Ophthalmic Hospital it was found that the external rectus muscle was completely paralysed, and lie had circumorbital pain. It was supposed that there was periosteal inflammation about the orbit, and iodide of potassium was given, the whitlow was poulticed, and the arm fomented. After a month of this treatment there was 110 improvement of the eye, but the arm inflammation had quite subsided. A piece of dead bone was now removed from the seat of the whitlow ; soon after which the squint disappeared, as well as the pain in the arm, and about the orbit. The external rectus had quite recovered its power.In this instance pain in sensory nerves about the orbit and paralysis of a single motor nerve were co-results of the morbid impression conveyed from the diseased finger to the centre. Dr. Watson refers to the production of amaurosis without visible change in the eye, in consequence apparently of irritation of the dental nerves, the blindness ceasing after the extraction of some teeth which had grown irregularly. He quotes from Mr. Lawrence an interesting case, in which the extraction of a carious tooth, with a splinter of wood projecting from one of its fangs, procured the restoration of the sight of the eye of the same side, which had been entirely lost for thirteen months. In such cases the paralysis of the retina or of the optic tubercles may fairly be designated inhibitory. . . . Some while ago I had a gentleman under my care with acute rosacea of the face and head, and chronic corneitis, with vascular development on the cornea. I applied on one occasion some liq. plumbi diacet., diluted with an equal amount of water, to the upper lids (everted), which 1 found very red. This caused excessive irritation; the eyes became greatly congested, watered extremely, and were very pain(ul; while the skin of the
Jones & Meryon on Nervous Disorders.
Jones & Meryon oh Nervous Disorders. |
Chromosomal imbalance letter: Phenotypic consequences of combined deletion 8pter and duplication 15qter
Exact breakpoint determination by oligonucleotide array-CGH has improved the analysis of genotype-phenotype correlations in cases with chromosome aberrations allowing a more accurate definition of relevant genes, particularly their isolated or combined impact on the phenotype in an unbalanced state. Chromosomal imbalances have been identified as one of the major causes of mental retardation and/or malformation syndromes and they are observed in~2-5% of the cases. Here we report a female child born to non-consanguineous parents and having multiple congenital anomalies such as atrial septal defect and multiple ventricular septal defects, convergent strabismus, micropthalmia, seizures and mental retardation, with her head circumference and stature normal for her age. Cytogenetic study suggested 46,XX,add(8)(p23). Further analysis by array-CGH using 44K oligonucleotide probe confirmed deletion on 8p23.3p23.1 of 7.1 Mb and duplication involving 15q23q26.3 of 30 Mb size leading to 46,XX, der(8)t(8;15)(p23.3;q23)pat.arr 8p23.3p23.1 (191,530-7,303,237)x1,15q23q26.3(72,338,961-102,35,195)x3. The unique phenotypic presentation in our case may have resulted from either loss or gain of a series of contiguous genes which may have resulted in a direct phenotypic effect and/or caused a genetic regulatory disturbance. Double segmental aberrations may have conferred phenotypic variability, as in our case, making it difficult to predict the characteristics that evolved as a result of the global gene imbalance, caused by the concomitant deletion and duplication.
# Introduction
Chromosomal rearrangements are frequently observed in patients with multiple congenital anomalies (MCA), dysmorphism and developmental delay with/without mental retardation [bib_ref] Whole-genome array-CGH identifies novel contiguous gene deletions and duplications associated with developmental..., Aradhya [/bib_ref] [bib_ref] Diagnostic utility of array-based comparative genomic hybridization in a clinical setting, Baris [/bib_ref] [bib_ref] CGH-array study and its utility in children for detection of constitutional and..., Andrieux [/bib_ref]. Inheritance of a balanced translocation from either or both parents is often responsible for structural chromosomal defects leading to segmental duplication or deletion of the chromosome pair in an affected individual. The paradigm shift in diagnostics with the implementation of next generation in silico softwares and array based comparative genomic hybridization (aCGH) technology, chromosome breakpoint determination, analyses of critical regions involved in genetic disorders and copy number evaluation has helped to correlate chromosomal region alteration and the resulting phenotype [bib_ref] Highthroughput analysis of subtelomeric chromosome rearrangements by Use of array-based comparative genomic..., Veltman [/bib_ref] [bib_ref] Chauffaille ML: 7q36 deletion and 9q22 duplication: effects of a double imbalance, Pelegrino [/bib_ref].
Segmental deletion of chromosome 8p [bib_ref] Deletion of a 5-cM region at chromosome 8p23 is associated with a..., Giglio [/bib_ref] [bib_ref] Distal 8p deletion (8p23.1--8pter): a common deletion?, Hutchinson [/bib_ref] and duplication involving chromosome 15q [bib_ref] Partial duplication of the long arm of chromosome 15: confirmation of a..., Zollino [/bib_ref] [bib_ref] Duplication of the distal long of chromosome 15: report of three patients..., Roggenbuck [/bib_ref] independently are well characterized and accurately compared with the clinical features seen in the affected individuals. However, segmental aneusomies simultaneously covering large regions on both the chromosomes, leading to a phenotypic presentation, have rarely been described.
Partial deletion of 8p23 is a relatively frequent deletion syndrome characterized by major congenital anomalies, especially congenital heart defects, seizures, behavioral abnormalities and postnatal growth deficiency [bib_ref] Distal 8p deletion (8p23.1--8pter): a common deletion?, Hutchinson [/bib_ref]. Facial dysmorphism may be subtle and mental retardation less severe than in those with deletions associated with more proximal breakpoints [bib_ref] Distal 8p deletion (8p23.1--8pter): a common deletion?, Hutchinson [/bib_ref].
Duplication of the distal long arm of chromosome 15 has been reported in various studies, which mainly highlighted it as an overgrowth syndrome characterized by prenatal overgrowth, macrocephaly, tall stature and craniosynostosis [bib_ref] Partial duplication of the long arm of chromosome 15: confirmation of a..., Zollino [/bib_ref]. However, such cases exclusively involve the distal most segment i.e. 15q25. It has been proposed by [bib_ref] Duplication of the distal long of chromosome 15: report of three patients..., Roggenbuck [/bib_ref] , that the aforementioned distal 15q trisomy syndrome may be the result of the disruption of the gene linked to 15q25 region, rather than partial trisomy for the region. Their study reported 3 cases sharing features like ptosis, small size and developmental delay [bib_ref] Duplication of the distal long of chromosome 15: report of three patients..., Roggenbuck [/bib_ref].
Here we present a female child with deletion 8p23. 3p23.1 as well as duplication 15q23q26.3. The aberrant chromosome 8 was inherited from the phenotypically normal father who was the carrier of a balanced translocation 46,XY,t(8;15)(p23;q23). We describe her phenotype, at birth, at 4 years and 8 years of age. In addition, we describe the phenotypic consequences of the concomitant effect of segmental deletion and duplication of the same.
## Clinical presentation
A 4 year-old female child was referred for congenital heart defects, dysmorphic facial features and developmental delay evident since her birth. She was the first child born to a non-consanguineous couple. During pregnancy, her mother had developed oligohydramnios and pre-eclampsia in the third trimester. She had a history of two spontaneous 1 st trimester abortions, one before and one after the birth of proband. The baby was born full term by caesarian section. Her birth weight was 3.5 kg and she did not cry at birth (Apgar score 4-6). She received oxygen for respiratory distress in the NICU for 6 days. Initially, she had feeding difficulties but eventually started breastfeeding at 1½ month. Loud holosystolic murmur was evident from the 2 nd day of life, and the 2D echocardiograph showed atypical muscular atrial septal defect (ASD) and multiple ventricular septal defects (VSD), with no patent ductus arteriosus (PDA). TORCH investigations were normal.
At the age of 4 years, the head circumference was 47 cm (10 th percentile), height 94.5 cm (15 th percentile) and weight 14.5 kg (25 th percentile) [bib_ref] IAP growth monitoring guidelines for children from birth to 18 years, Khadilkar [/bib_ref]. The prominent dysmorphic features were ptosis, downward slanted palpebral fissure, microphthalmia, left convergent strabismus, wide nasal base, long philtrum, open mouth, low set ears, short neck, micrognathia, puffy cheeks, short fingers, bilateral 1 st incurved finger and absence of thenar eminence [ [fig_ref] Figure 1: Clinical presentation of the proband at 8 years of age [/fig_ref] ]. Developmental delay was evident from the first year as social smile was absent even at 6 months. She learnt sitting at 10 months and walking at 20 months. Speech was absent and her developmental quotient at presentation was of a 2 year old child. She had 3 attacks of partial seizures during the first 20 months and was kept on carbamazepine 150 mg/day and sodium valproate 100 mg/day till she was 5 years old. Since then, she has been free of epileptic symptoms. Sequential complete hemogram was suggestive of chronic iron deficiency anemia with hemoglobin levels of 8-8.5 gm%. 2D echocardiograph again at 3 years showed multiple small muscular VSDs and a single ASD of 4.2 × 7.0 mm.
On re-examination at the age of 8 years, the head circumference was 49.5 cm (10 th percentile), height 120 cm (25 th percentile) and weight 28 kg (75 th percentile) [bib_ref] IAP growth monitoring guidelines for children from birth to 18 years, Khadilkar [/bib_ref]. Mental retardation was evident. She could not speak but responded to commands. The bladder and bowel control had not yet been achieved. Apart from these, no other developmental abnormalities were noted. She could carry out routine activities independently. It is to be noted that she had normal for age head circumference, height and weight.
# Results
# Metaphase chromosome analysis
After obtaining institutional ethical committee approval and informed written consent form, metaphase chromosome preparations were obtained from PHA stimulated lymphocyte cultures according to the standard procedure with slight modifications [bib_ref] Chromosome preparations of leukocytes cultured from human peripheral blood, Moorhead [/bib_ref]. Chromosome analysis was carried out by GTG-banding at 550-band level according to ISCN 2013 nomenclature in both patient and parents (50 metaphases, each). Patient's karyotype pattern showed additional genetic material on the short arm of #8p i.e.46,XX,add(8)(p23). Parental chromosomal investigation revealed that the father was a carrier of a balanced translocation 46,XY,t(8;15)(p23;q23) [ [fig_ref] Figure 2: Partial karyotype of the father showing breakpoints on both derivative chromosomes [/fig_ref] ]. This suggests that the child had inherited der(8) from the father thus leading to an unbalanced genetic makeup.
## Array-cgh
Genomic DNA was extracted from peripheral blood lymphocytes using standard SDS-proteinase K extraction method [bib_ref] A simple salting out procedure for extracting DNA from human nucleated cells, Miller [/bib_ref]. Extracted genomic DNA concentration was determined with NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Berlin, Germany). Evaluation of gene copy number was performed by 44k oligonucleotide array-Comparative Genomic Hybridization (aCGH) by following manufacturer's recommendations (Human Genome CGH microarray 44B kit, Agilent Technologies Inc., Santa Clara, CA, USA). Female genomic DNA (Promega Corporation, Madison, WI, USA) was used as a sex-matched reference, which was analyzed with the CGH-analysis software v3.4 (Agilent Technologies Inc., Santa Clara, CA, USA) by applying Z-score segmentation algorithm with a window size of 10 points to identify chromosome aberrations. Analysis was performed using 3-points filter and 0.2 variation which lead to confirmation of partial deletion 8p region of 7. [fig_ref] Figure 3: Breakpoint characterization by 44K oligonucleotide array-CGH [/fig_ref] ,b].
## Fluorescence in situ hybridization (fish)
FISH analysis was performed using BAC clones RP11-139L10 covering 8p23 → pter, RP11-95F11 and RP11-100A1 spanning 15q23 → qter. Nick Translation Kit (Vysis, Abbott Molecular, USA) was used for labeling. The two clones -RP11-95F11 and RP11-100A1 were labeled with Spectrum Orange and Spectrum Green respectively whereas, third BAC clone was labeled using both fluorochromes in 1:1 ratio labeling. FISH signals were observed using Olympus BX-51 microscope (Olympus, Germany) and pseudo-coloring was carried out using Adobe Photoshop [ [fig_ref] Figure 4: Metaphase showing signals of both BAC clones RP11-95F11 and RP11-100A1 spanning 15q25 [/fig_ref] ].
# Discussion
This case report presents a 4 year old girl who was diagnosed with double segmental chromosomal aberrations 46,XX,der(8)t(8;15)(p23;q23)pat. The derivative #8 was inherited from the father who was a carrier of balanced 8;15 chromosomal translocation which led to partial deletion of 8p23.3p23.1 and duplication of 15q23q26.3 in the child. Inheriting unbalanced chromosomal rearrangements that originated from a translocation event in either parent are often associated with mental retardation and/or congenital malformations [bib_ref] Highthroughput analysis of subtelomeric chromosome rearrangements by Use of array-based comparative genomic..., Veltman [/bib_ref].
The distal 8p deletion is of a relatively frequent occurrence with distinct clinical features [bib_ref] Duplication of the distal long of chromosome 15: report of three patients..., Roggenbuck [/bib_ref]. Congenital heart defects, behavioral problems, mild to moderate mental retardation/developmental delay, strabismus and mild facial and digital anomalies are frequent phenotypic characteristics seen in these patients [bib_ref] Partial duplication of the long arm of chromosome 15: confirmation of a..., Zollino [/bib_ref] [bib_ref] Two patients with atypical interstitial deletions of 8p23.1: Mapping of phenotypical traits, Páez [/bib_ref] [bib_ref] Delineation of the critical deletion region for congenital heart defects on chromosome..., Devriendt [/bib_ref]. The severity of mental retardation, pre and postnatal growth deficiency including microcephaly and facial dysmorphism becomes more pronounced as the deletion site involves more proximal regions [bib_ref] Delineation of the critical deletion region for congenital heart defects on chromosome..., Devriendt [/bib_ref].
Cytogenetic evidence suggests that the haploinsufficiency of ≥1 gene located in 8p23 behaves as a dominant mutation, thus impairing heart differentiation and leading to a wide spectrum of congenital heart defects (CHDs) [bib_ref] Distal 8p deletion (8p23.1--8pter): a common deletion?, Hutchinson [/bib_ref]. The gene responsible for the heart defects in this syndrome has been identified to be GATA4 on 8p23.1 [bib_ref] Two patients with atypical interstitial deletions of 8p23.1: Mapping of phenotypical traits, Páez [/bib_ref]. Haploinsufficiency of GATA4 is thought to play a critical role in the development of these birth defects [bib_ref] Chromosome 8p23.1 deletions as a cause of complex congenital heart defects and..., Wat [/bib_ref] [bib_ref] A duplication including GATA4 does not co-segregate with congenital heart defects, Joziasse [/bib_ref]. This data is in concordance with the findings in our patient who has atrial septal defect and multiple ventricular septal defects since birth.
Microcephaly is another feature frequently seen in patients with distal 8p deletion [bib_ref] Delineation of the critical deletion region for congenital heart defects on chromosome..., Devriendt [/bib_ref]. The MCPH1 gene, mapped to chromosome 8p23, has been implicated to be a candidate gene for primary microcephaly [bib_ref] Molecular evolution of microcephalin, a gene determining human brain size, Wang [/bib_ref] [bib_ref] Identification of microcephalin, a protein implicated in determining the size of the..., Jackson [/bib_ref]. Although, in our patient, microcephaly was not noted, this finding has also been reported by other studies [bib_ref] Duplication of the distal long of chromosome 15: report of three patients..., Roggenbuck [/bib_ref] [bib_ref] Van Den Berghe K, Van Den Berghe H: The fetal phenotype of..., Fryns [/bib_ref]. This gene has been shown to have a high penetrance with large deletion, as has been detected in the present case. However, the possibility of non-penetrance and/or nonexpressivity as the cause behind the absence of microcephaly cannot be entirely excluded [bib_ref] Delineation of the critical deletion region for congenital heart defects on chromosome..., Devriendt [/bib_ref]. The concomitant presence of duplication of distal 15q leading to a normal for age head circumference could also be a possibility.
Our proband had mild mental retardation with very little speech development and facial dysmorphic features like wide nasal base, puffy cheeks, low set ears, micrognathia are in concordance with other cases of 8p deletion syndrome [bib_ref] Partial duplication of the long arm of chromosome 15: confirmation of a..., Zollino [/bib_ref]. But majority of her facial features, including those mentioned above, like ptosis, down slanted palpebral fissures, long philtrum, open mouth, mid crease in the lower lip are similar to the distinguishable facial characteristics seen in cases of distal duplication of 15q [bib_ref] Distal 15q trisomy: phenotypic comparison of nine cases in an extended family, Schnatterly [/bib_ref].
Since the first report of 15q22 → qter duplication by , at least 71 other cases of similar or smaller distal 15q imbalances have been described [bib_ref] A girl with 15q overgrowth syndrome and dup(15)(q24q26.3) That included telomeric sequences, Gutiérrez-Franco [/bib_ref]. The phenotype-genotype correlations observed in these cases resulted in the delineation of the 15q overgrowth syndrome, which is caused by the increased dosage of the genes that are present between 15q25-q26.3. The findings of overgrowth have reliably been associated with the extra copy of the IGF1R (insulin-like growth factor 1 receptor) gene located at 15q26.3 [bib_ref] A girl with 15q overgrowth syndrome and dup(15)(q24q26.3) That included telomeric sequences, Gutiérrez-Franco [/bib_ref]. However, in our patient, no findings suggestive of overgrowth were observed. This can be explained by the fact that this girl has an associated duplication extending proximally till 15q23. [bib_ref] Duplication of the distal long of chromosome 15: report of three patients..., Roggenbuck [/bib_ref] [bib_ref] Partial duplication of the long arm of chromosome 15: confirmation of a..., Zollino [/bib_ref] described cases with duplication from 15q24-q26.3 and 15q25.1-qter respectively, which showed postnatal growth retardation and developmental delay [bib_ref] Partial duplication of the long arm of chromosome 15: confirmation of a..., Zollino [/bib_ref] [bib_ref] Duplication of the distal long of chromosome 15: report of three patients..., Roggenbuck [/bib_ref]. Roggenbuck and co-workers in their study also proposed that the features of overgrowth observed in patients with distal duplication of 15q may not be specifically related to increased dosage of genes in this chromosomal region [bib_ref] Duplication of the distal long of chromosome 15: report of three patients..., Roggenbuck [/bib_ref]. This could be the reason why the proband exhibits normal growth. In summary, the phenotypic effects of 8p deletion were dominant over 15q duplication as depicted in [fig_ref] Table 1: Comparison of phenotypes having partial 8p deletion and partial 15q duplication and... [/fig_ref]. Both microcephaly and postnatal growth retardation reported in patients of 8p deletion syndrome and distal 15q duplication syndrome respectively were not observed in our patient. Combined chromosomal aberrations, as the present one, may confer phenotypic variability, thereby making it difficult to perform genotype-phenotype correlations [bib_ref] Deletion of a 5-cM region at chromosome 8p23 is associated with a..., Giglio [/bib_ref]. As for all complex genetic features or disorders, phenotype is influenced by additional genetic and environmental factors. Also haploinsufficiency of the deleted gene(s) alone cannot explain this clinical variability, other modifying factors like single base pair mutations and/or polymorphisms of uninvolved loci may contribute [bib_ref] Chromosomal phenotypes and submicroscopic abnormalities, Devriendt [/bib_ref].
To conclude, such correlations of defined phenotypic manifestations with the deletion or duplication of specific genes provides an opportunity to analyze gene-gene interactions and help to further unravel the intricacies of the human genome.
## Consent
Written informed consent was obtained from the parents for publication of this case report and accompanying images of the child. A copy of the written consent is available for review by the Editor-in-Chief of this journal.
[fig] Figure 1: Clinical presentation of the proband at 8 years of age. [/fig]
[fig] Figure 2: Partial karyotype of the father showing breakpoints on both derivative chromosomes. [/fig]
[fig] Figure 3: Breakpoint characterization by 44K oligonucleotide array-CGH. a: 7.1 Mb deletion at 8p [arr 8p23.3p23.1(191,530-7,303,237)x1] and b: 30 Mb duplication at 15q [arr 15q23q26.3(chr15:72,338,961-102,351,195)x3]. [/fig]
[fig] Figure 4: Metaphase showing signals of both BAC clones RP11-95F11 and RP11-100A1 spanning 15q25.3 to 15q26.3 region on two normal #15 and on derivative #8p [thin arrow]. Whereas, clone RP11-139L10 covering 8p23.3 is seen only on normal #8p [broad arrow]. [/fig]
[table] 1: Mb [arr cgh 8p23.3p23.1(191,530-7,303,237)(hg19-NCBI build37)x1] and partial 15q duplication of 30 Mb [arr cgh 15q23q26.3 (chr15:72,338,961-102,351,195)x3], i.e. 46,XX,der(8)t(8;15) (p23.3;q23)pat.arr 8p23.3p23.1(191,530-7,303,237)x1,15q23 q26.3(72,338,961-102,351,195)x3 [ [/table]
|
The Architect Who Lost the Ability to Imagine: The Cerebral Basis of Visual Imagery
While the loss of mental imagery following brain lesions was first described more than a century ago, the key cerebral areas involved remain elusive. Here we report neuropsychological data from an architect (PL518) who lost his ability for visual imagery following a bilateral posterior cerebral artery (PCA) stroke. We compare his profile to three other patients with bilateral PCA stroke and another architect with a large PCA lesion confined to the right hemisphere. We also compare structural images of their lesions, aiming to delineate cerebral areas selectively lesioned in acquired aphantasia. When comparing the neuropsychological profile and structural magnetic resonance imaging (MRI) for the aphantasic architect PL518 to patients with either a comparable background (an architect) or bilateral PCA lesions, we find: (1) there is a large overlap of cognitive deficits between patients, with the very notable exception of aphantasia which only occurs in PL518, and (2) there is large overlap of the patients' lesions. The only areas of selective lesion in PL518 is a small patch in the left fusiform gyrus as well as part of the right lingual gyrus. We suggest that these areas, and perhaps in particular the region in the left fusiform gyrus, play an important role in the cerebral network involved in visual imagery.
# Introduction
Thinking of a concept, whether it is a flower or a cat or even a unicorn, can bring up vivid, image-like experiences without external visual input. This is generally referred to as visual imagery or mental imagery, although the latter can extend to other senses (e.g., sound, smell, or touch). The basis of mental imagery has long been debated [bib_ref] What the mind's eye tells the mind's brain: A critique of mental..., Pylyshyn [/bib_ref] [bib_ref] The imagery debate: Analogue media versus tacit knowledge, Pylyshyn [/bib_ref] [bib_ref] The medium and the message in mental imagery: A theory, Kosslyn [/bib_ref] and there is still uncertainty about its neural underpinnings.
Zeman and colleagues [bib_ref] Loss of imagery phenomenology with intact visuo-spatial task performance: A case of..., Zeman [/bib_ref] gave the inability to generate mental imagery a name, aphantasia, and described individuals with congenital aphantasia who never had this ability. The loss of mental imagery following brain injury-acquired aphantasia-in individuals who had normal imagery before their injury is also well documented, dating back at least to Charcot and Bernard [bib_ref] Un cas de suppression brusque et isolée de la vision mentale des..., Charcot [/bib_ref] (but see [bib_ref] Is the Charcot and Bernard case (1883) of loss of visual imagery..., Zago [/bib_ref]. However, as noted by Farah [bib_ref] The neurological basis of mental imagery: A componential analysis, Farah [/bib_ref] , cases of acquired imagery deficits can be associated with a wide range of lesions (occipital, temporal, or parietal) in either hemisphere, and no other functional deficits consistently co-occurred with imagery loss with the exception of loss of (visual) dreaming. One plausible reason for this heterogeneity is that mental imagery is not a single phenomenon but can be divided into relatively distinct components, with different underlying anatomy. Some distinguish between a generation process, long-term visual memory, and an inspection process [bib_ref] The neurological basis of mental imagery: A componential analysis, Farah [/bib_ref] , or subsystems such as appearance-based (e.g., shape/color judgment) vs. spatial (e.g., mental navigation/scanning) imagery [bib_ref] Visual and spatial mental imagery: Dissociable systems of representation, Farah [/bib_ref] [bib_ref] Two visual systems in mental imagery: Dissociation of "what" and "where" in..., Levine [/bib_ref] (see also [bib_ref] Seeing and imagining in the cerebral hemispheres: A computational approach, Kosslyn [/bib_ref]. Supporting this, a meta-analysis of imaging studies showed that while several regions were coactivated during appearance-based and spatial imagery, the former mapped onto the ventral visual stream while the latter evoked specific activity in the dorsal stream [bib_ref] A PET meta-analysis of object and spatial mental imagery, Mazard [/bib_ref].
It has been argued that the primary visual cortex (V1) plays a significant role in visual mental imagery [bib_ref] Neural foundations of imagery, Kosslyn [/bib_ref] [bib_ref] When is early visual cortex activated during visual mental imagery?, Kosslyn [/bib_ref]. Several studies have shown cortical activation in V1 during imagery tasks (e.g., [bib_ref] Human primary visual cortex and lateral geniculate nucleus activation during visual imagery, Chen [/bib_ref] [bib_ref] The role of area 17 in visual imagery: Convergent evidence from PET..., Kosslyn [/bib_ref] [bib_ref] Transient activity in the human calcarine cortex during visual-mental imagery: An event-related..., Klein [/bib_ref] [bib_ref] Brain areas underlying visual mental imagery and visual perception: An fMRI study, Ganis [/bib_ref] [bib_ref] Visual mental imagery induces retinotopically organized activation of early visual areas, Slotnick [/bib_ref] and rTMS (repetitive transcranial magnetic stimulation) targeting V1 can disrupt visual imagery [bib_ref] The role of area 17 in visual imagery: Convergent evidence from PET..., Kosslyn [/bib_ref]. In addition, individual differences in mental imagery capability covary with differences in V1 surface area [bib_ref] Smaller primary visual cortex is associated with stronger, but less precise mental..., Bergmann [/bib_ref] , V1 functional connectivity [bib_ref] Distinct top-down and bottom-up brain connectivity during visual perception and imagery, Dijkstra [/bib_ref] , and representational overlap between visual imagery and perception in the retinotopic cortex [bib_ref] Disentangling visual imagery and perception of real-world objects, Lee [/bib_ref]. However, while patients with intact V1 can have severe impairments in mental imagery [bib_ref] Selective deficit of mental visual imagery with intact primary visual cortex and..., Moro [/bib_ref] , seemingly intact imagery without a functioning V1 has also been reported [bib_ref] Visual activation of extra-striate cortex in the absence of V1 activation, Bridge [/bib_ref] [bib_ref] Vivid visual mental imagery in the absence of the primary visual cortex, Bridge [/bib_ref] (see also [bib_ref] Cortical blindness and visual imagery, Chatterjee [/bib_ref].
Thus, damage to V1 appears neither necessary nor sufficient for inducing imagery deficits. A review [bib_ref] The relationship between visual perception and visual mental imagery: A reappraisal of..., Bartolomeo [/bib_ref] of case studies suggested that extensive left temporal damage is necessary for a visual imagery deficit for object form or color (see also [bib_ref] A PET meta-analysis of object and spatial mental imagery, Mazard [/bib_ref] , and more generally that high-level visual areas in the temporal lobe might be particularly important for visual imagery. The fact that patients have been reported to have both high-level visual deficits and selective imagery loss in the same domain (e.g., severe problems in visual recognition and revisualization of faces, [bib_ref] Further observations on the nature of prosopagnosia, Shuttleworth [/bib_ref] , and that actual viewing and visual imagery for particular objects or object categories can evoke a similar pattern of activity in high-level ventral stream regions [bib_ref] Disentangling visual imagery and perception of real-world objects, Lee [/bib_ref] [bib_ref] Mental imagery of faces and places activates corresponding stimulus-specific brain regions, O Craven [/bib_ref] [bib_ref] Reading the mind's eye: Decoding category information during mental imagery, Reddy [/bib_ref] , is in alignment with the general idea of shared mechanisms between visual imagery and visual perception (for recent reviews, see [bib_ref] Shared neural mechanisms of visual perception and imagery, Dijkstra [/bib_ref] [bib_ref] The heterogeneity of mental representation: Ending the imagery debate, Pearson [/bib_ref].
Visual imagery and perception however cannot share all mechanisms as there are patients on record with seemingly preserved mental imagery but impaired visual perception [bib_ref] A case of integrative visual agnosia, Riddoch [/bib_ref] [bib_ref] Routes to object constancy: Implications from neurological impairments of object constancy, Humphreys [/bib_ref] [bib_ref] Memories are made of this: The effects of time on stored visual..., Riddoch [/bib_ref] [bib_ref] Dissociation between mental imagery and object recognition in a brain-damaged patient, Behrmann [/bib_ref] [bib_ref] Intact visual imagery and impaired visual perception in a patient with visual..., Behrmann [/bib_ref]. For example, case H.J.A. [bib_ref] A case of integrative visual agnosia, Riddoch [/bib_ref] suffered from visual agnosia, achromatopsia, prosopagnosia, alexia without agraphia and topographical impairments. Despite these deficits, H.J.A.'s mental imagery was relatively-albeit not completely-spared. The opposite pattern, impaired visual mental imagery but relatively normal visual perception, has also been reported [bib_ref] A case study of mental imagery deficit, Farah [/bib_ref] [bib_ref] Loss of visual imagery and loss of visual knowledge-A case study, Goldenberg [/bib_ref]. An example is a patient who had suffered a left occipital and medial temporal infarct. While his visual recognition abilities were generally good, he showed apparent problems in mental imagery such as describing an elephant as having a "tiny waist" and having trouble with verifying sentences that required visual imagery (e.g., "A grapefruit is larger than an orange") [bib_ref] A case study of mental imagery deficit, Farah [/bib_ref].
Here we present patient PL518, an architect who reported almost complete loss of visual mental imagery following bilateral stroke in the areas supplied by the posterior cerebral artery (PCA). His responses on the Vividness of Visual Imagery Questionnaire (VVIQ, ad modum [bib_ref] Lives without imagery-Congenital aphantasia, Zeman [/bib_ref] as well as a range of visuoperceptual tests are compared to three other patients with bilateral PCA stroke, as well as another architect with a large unilateral PCA stroke in the right hemisphere. We also compare the structural images of their lesions. The aim of the study is to: (a) describe the correspondence between the perceptual and neuropsychological profile of PL518 compared to the other patients, and (b) to delineate cerebral areas that are uniquely affected in the aphantasic patient and could thus play a fundamental role in the generation of visual imagery.
# Materials and methods
## Participants
Patient PL518 and four other patients participated in this specific study. All were recruited as part of a larger study of PCA stroke (the Back of the Brain (BoB) project, described in [bib_ref] The Neuropsychology of Stroke in the Back of the Brain: Clinical and..., Robotham [/bib_ref]. 46 controls were included in the BoB project. All participants provided written, informed consent, and the project was approved by the ethical committees of Manchester (North West Research Ethics Committee; MREC 01/8/094) and UCL (London Queen Square Research Ethics Committee, UCL; 16/EM/0348). See for demographics and background data for the included patients and controls. Additional background data as well as raw scores on the perceptual and neuropsychological tests can be found in . . Demographic and lesion information for the five included patients and controls, and scores on basic tests. Handedness was measured by the short form of the Edinburgh Handedness Inventory (EHI) [bib_ref] Edinburgh handedness inventory-short form: A revised version based on confirmatory factor analysis...., Veale [/bib_ref] ; depression was measured with the short version of the Geriatric Depression Scale (GDS-15 [bib_ref] Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version, Yesavage [/bib_ref]. General cognition was screened with the Oxford Cognitive Screen (OCS) [bib_ref] The Oxford Cognitive Screen (OCS): Validation of a stroke-specific short cognitive screening..., Demeyere [/bib_ref] , and the number of impaired subtests are listed. Digit span forward and backward was measured with the WAIS-IV UKand total scores are listed. Basic motor reaction time (RT) was measured by responding to a bar of light presented horizontally on a screen (test described in [bib_ref] The Neuropsychology of Stroke in the Back of the Brain: Clinical and..., Robotham [/bib_ref] PL518 suffered a bilateral PCA stroke 35 months before the current investigation. At that time, he had corrected to normal visual acuity and a slight visual field defect primarily affecting the parafovea of the upper left quadrant (see for acuity and visual field data for all participants). He reported problems with seeing colors following his stroke and scored outside the normal range on a formal test of color perception (Farnsworth D-15 [bib_ref] The Farnsworth panel D-15 test, Linksz [/bib_ref]. His intermediate vision (assessed with subtests from the L-POST [bib_ref] The Leuven Perceptual Organization Screening Test. (L-POST), an online test to assess..., Torfs [/bib_ref] was largely uncompromised, except for difficulties with figure-ground segmentation. His basic response time (RT) to visual stimuli (test described in [bib_ref] The Neuropsychology of Stroke in the Back of the Brain: Clinical and..., Robotham [/bib_ref] was unaffected, and his auditory digit span forwards and backwards (WAIS-IV UKwere within the normal range. PL518 reported severe problems in face recognition following his stroke and volunteered that he had problems recognizing his own face in the mirror. He also reported increased problems in finding his way around. Neuropsychological testing showed a clear deficit in face recognition affecting learning of new faces as well as judgment of familiarity and recognition of famous faces. In contrast, he performed within the normal range on several tests of object recognition, including perceptually challenging tests, with the notable exception of a memory test for houses (Cambridge House Memory Test, [bib_ref] Visual agnosia and posterior cerebral artery infarcts: An anatomical-clinical study, Martinaud [/bib_ref] designed as an equivalent to the Cambridge Face Memory Test [bib_ref] The Cambridge Face Memory Test: Results for neurologically intact individuals and an..., Duchaine [/bib_ref]. His word recognition accuracy was well within normal range, while response times in reading out loud and the effect of word length on RT was slightly but significantly elevated compared to controls.
In the context of the BoB project, PL518 spontaneously reported that his visual imagery was "gone" following his stroke, which led us to contact four additional patients either with similar lesions (bilateral PCA stroke) or a similar background (architect) for a follow-up interview about their visual imagery. The lesion location, size, and time since injury for these patients, as well as other background characteristics, are presented in along with summary data from the control group.
## Visual imagery
As PL 518 was the focus of the study, a long and in-depth interview was also carried out about his visual imagery before and after his stroke. In order to get more information about his ability to store visual information in his mind, he was asked to carry out the Rey-Osterrieth Complex [bib_ref] examen psychologique dans les cas d'encéphalopathie traumatique.(Les problems.), Rey [/bib_ref] at the end of the interview. The other patients did not complete this test.
All five patients were asked to complete a version of the Vividness of Visual Imagery Questionnaire (VVIQ-modified, [bib_ref] Lives without imagery-Congenital aphantasia, Zeman [/bib_ref] , that is a modified version of the VVIQ [bib_ref] Visual imagery differences in the recall of pictures, Marks [/bib_ref]. VVIQ-modified has 16 items where participants are to imagine various scenarios (e.g., a relative or friend, a rising sun) and rate the vividness of their visual image on a five-point Likert scale where 1 indicates no image at all and 5 indicates that the image is perfectly clear and vivid. Scores on the VVIQ-modified can range from 16 to 80, with 16 representing the lowest possible imagery score and 80 the highest possible imagery score. Various versions of the VVIQ have been validated, and the questionnaire has been shown to be a valid psychometric tool for measuring the vividness of visual imagery with both high construct validity and internal consistency reliability [bib_ref] Psychometric quality of a revised version Vividness of Visual Imagery Questionnaire, Campos [/bib_ref] [bib_ref] Internal consistency and construct validity of two versions of the Revised Vividness..., Campos [/bib_ref] [bib_ref] Imagery vividness, creativity and the visual arts, Morrison [/bib_ref]. PL518 and PM024 performed the questionnaire in the lab while the other three patients were interviewed over the telephone. The four control patients were also asked four general questions about their visual imagery, to compare with the interview of PL518. They were asked to answer the following questions with yes, no, or don't know: (1) can you imagine things visually in your mind? (if no: do you sometimes experience brief flashes of imagery?); (2) would you say that your memories have a visual aspect to them in your mind?; (3) do you see visual images in dreams?; and (4) has your visual imagery changed following your stroke? The normal controls did not perform the visual imagery questionnaire.
## Neuropsychological and experimental tests
The BoB project is a comprehensive neuropsychological and imaging project investigating perceptual deficits following posterior brain injury [bib_ref] The Neuropsychology of Stroke in the Back of the Brain: Clinical and..., Robotham [/bib_ref]. A main aim of the overall project is to compare patient performance with faces, objects, and words. The main findings of the project are not yet published (paper in preparation,. Here we report data from the five included patients and controls on tests and experiments selected to be comparable across categories for faces, objects, and words, and these are briefly described below. These experiments, as well as all other tests included in the project, are described in full in [bib_ref] The Neuropsychology of Stroke in the Back of the Brain: Clinical and..., Robotham [/bib_ref]. The experimental tests were run on laptop computers with screen resolution of 1366 × 768, or on desktop computers with a screen resolution of 1920 × 1080.
## Delayed matching and surprise recognition of words, objects and faces-the wof test
This novel paradigm is designed to test immediate and delayed memory for words, objects, and faces (WOF). In the first part (delayed matching), participants were asked to decide whether two sequentially presented images varying in size are the same or not. There were 48 trials for each stimulus type and both accuracy and RTs are measured. The second part (surprise recognition) followed after a short break (where participants performed an unrelated task, the Farnsworth D-15). Here, participants were asked to decide whether they saw the presented stimuli in the delayed matching task or not. There were 12 trials, and accuracy and RTs were measured. In total, 12 measures were derived from this task: 2 metrics (accuracy and RT) * 3 stimulus types (words, objects, faces) * 2 paradigms (delayed matching and surprise recognition). See [bib_ref] The Neuropsychology of Stroke in the Back of the Brain: Clinical and..., Robotham [/bib_ref] for a more detailed description.
## Familiarity decisions
Familiarity decision tasks were run for faces, objects, and words. For faces, participants were asked to decide whether a presented face was famous or not (80 trials in total). For objects, we used a 72-trial version of a well-studied object decision task [bib_ref] Normal and abnormal category-effects in visual object recognition: A legacy of Glyn..., Gerlach [/bib_ref] , presenting line drawings of real objects and chimeric non-objects. Participants were asked to decide if the picture represents a real object or a non-object. For words, we used a lexical decision task with 60 trials, where participants were asked to decide whether the presented letter string represented a real word or a pseudoword. For all three familiarity decision tests, both accuracy and RTs for correctly categorized familiar items (famous faces, real objects, real words) were analyzed.
## Naming of familiar items
Tests of picture naming (line drawings), face naming (famous faces) and word reading (regular words). For pictures and words, both accuracy and RTs for correctly named items are analyzed (a voice key was used for RT measurement). For famous faces, only accuracy is recorded as measuring RTs in face naming tasks is complicated by participants making other verbal responses than names (e.g., "it's that guy from the Parliament . . . ").
## Structural mri: lesions
Structural brain imaging data were acquired from all subjects. Structural scans were acquired on two 3T Phillips Achieva scanners with 32-channel head-coils and a SENSE factor of 2.5 in London and Manchester. A high-resolution T1 weighted structural scan was acquired for spatial normalization, including 260 slices covering the whole brain with TR = 8.4 ms, TE = 3.9 ms, flip angle = 8 degrees, FOV = 240 × 191 mm 2 , resolution matrix = 256 × 206 and voxels size = 0.9 × 1.7 × 0.9 mm 3 . Automated outlines of the area affected by stroke were generated using Seghier et al.'s modified segmentation-normalization procedure [bib_ref] Lesion identification using unified segmentation-normalisation models and fuzzy clustering, Seghier [/bib_ref]. Segmented images were smoothed with an 8mm full-width half maximum Gaussian kernel and submitted to the automated routines for lesion identification and definition modules using the default parameters. The automated method involves initial segmentation and normalizing into tissue classes of grey matter, white matter, cerebro-spinal fluid (CSF), and an extra tissue class for the presence of a lesion. After smoothing, voxels that emerge as outliers relative to the normal population are identified and the union of these outliers provides the "fuzzy lesion map", from which the lesion outline is derived. The generated images were used to create the lesion overlap maps.
# Results
## Visual imagery
In the clinical interview, PL518 reported an almost complete absence of visual imagery following his stroke. This was in stark contrast to his (in his own opinion) above average ability for visual imagery before his stroke that he had relied upon in his work as an architect. He said: "Before, my visualization abilities were pretty impressive. At my work, I could visualize and remember things that most people had not thought about. I would be sitting there and I would say, well, you can't do X, Y and Z, because you've got this happening here and there. Now I have to look at the drawing and work my way through it." During the interview, he also described how it had felt to do a mental rotation task: "I cannot do it as quickly or the same way as I would have done before my stroke. Before, bang, I would just know the answer. Now it is a much more conscious process. It's almost as though I physically am trying to move things inside my head." He was then asked whether his difficulty with mental rotation affected his ability to work as an architect, to which he responded: "Well I just do everything on the computer. That is one of the advantages of us using computers for these sorts of thing nowadays. You can see the stuff happen." He also described how he is just about able to imagine very simple shapes, but this is done using something akin to motor or spatial imagery and he struggles to imagine more than one shape at a time: "If I tried to visualize shapes like a square, pyramid or sphere lined up next to each other, and I try and focus with a kind of spotlight on the corner of one shape, I can mentally trace a line around the shape. But as soon as I focus on one shape, the others disappear." When asked if he could imagine an elephant, he seemed to mostly think of the abstract concept of an elephant: "I can think of elephants, iconic elephants like Babar or Elmer, but I can only visualize bits of them. It's almost painful." When asked to describe the place he stayed during his last holiday and its surroundings he provided few very vague details about a couple of the bars from the street they had lived on, and he apparently did not visually imagine himself there: "I am recalling almost like a list. I do the same when going somewhere. I have to remember a list".
PL518 s copy and retention of the Rey figure are shown in [fig_ref] Figure 1: PL518′s performance on the Rey Complex Figure Test [/fig_ref]. The drawings were scored for accuracy according to the Taylor's (1969) method described in Spreen and Straus (1991). 35/36 points were given for the copy and 18/36 for the three-minute recall. While these scores are within the normal range, one could have expected patient PL518, with his background as an architect, to have adopted a more structured approach to drawing the figure in the recall condition. This drawing not only lacks many details but also some of the core elements. Also, some of the included elements are placed incorrectly.
Brain Sci. 2020, 10, [bib_ref] Three-dimensional probabilistic maps of mesial temporal lobe structures in children and adolescents'..., Bouyeure [/bib_ref] 6 of 15 the bars from the street they had lived on, and he apparently did not visually imagine himself there: "I am recalling almost like a list. I do the same when going somewhere. I have to remember a list". PL518′s copy and retention of the Rey figure are shown in [fig_ref] Figure 1: PL518′s performance on the Rey Complex Figure Test [/fig_ref]. The drawings were scored for accuracy according to the Taylor's (1969) method described in Spreen and Straus (1991). 35/36 points were given for the copy and 18/36 for the three-minute recall. While these scores are within the normal range, one could have expected patient PL518, with his background as an architect, to have adopted a more structured approach to drawing the figure in the recall condition. This drawing not only lacks many details but also some of the core elements. Also, some of the included elements are placed incorrectly. PL518′s complaints regarding his visual imagery were also clearly reflected in his responses on the VVIQ-modified where he scored 18 (i.e., a mean score of 1.13 per item), corresponding to minimal imagery [bib_ref] Lives without imagery-Congenital aphantasia, Zeman [/bib_ref]. None of the other patients reported any changes in the nature or vividness of their visual imagery following their strokes-neither in the VVIQ-modified nor the general questions; they all responded yes to the first three general questions about being able to see images in their minds, and no when asked if their visual imagery had changed following their stroke. Their respective scores on the VVIQ-modified were: PL502: 49 (mean: 3.06); PL545: 53 (mean: 3.31); PM006: 72 (mean: 4.5); PM024: 76 (mean: 4.75). See Appendix A for the patients' responses to the individual questions.
## Neuropsychological and experimental tests
For the accuracy measures, PL518 is clearly impaired with faces, and shows a deficit (performing more than two standard deviations (SDs) from the control mean) on most individual face measures. He performs within the low-normal range on the object tests and is clearly on level with controls in the tests with word stimuli. For the RT measures, PL518 shows a deficit on most face measures (note that his RTs in the surprise recognition test may not be a good indicator of severity, as his accuracy in this test was very low). He responds with latencies within the normal range on the object tests but shows elevated RTs in the lexical decision and word reading tests.
Comparing the neuropsychological profile of PL518 to the other included patients, we find that one or more of them show deficits on the same tests/measures and in the cognitive domain(s) as PL518 (see [fig_ref] Figure 2: Radar plots showing the results of PL518 [/fig_ref] for an illustration of their cognitive profiles on the selected tests, and for an overview of test results). A comparison of the neuropsychological profile of the two architects (PL518 and PM024) shows that PM024 (with no aphantasia) shows the same pattern of performance as PL518 on most tests, including measures of face recognition, object recognition and word reading. PL518 s complaints regarding his visual imagery were also clearly reflected in his responses on the VVIQ-modified where he scored 18 (i.e., a mean score of 1.13 per item), corresponding to minimal imagery [bib_ref] Lives without imagery-Congenital aphantasia, Zeman [/bib_ref]. None of the other patients reported any changes in the nature or vividness of their visual imagery following their strokes-neither in the VVIQ-modified nor the general questions; they all responded yes to the first three general questions about being able to see images in their minds, and no when asked if their visual imagery had changed following their stroke. Their respective scores on the VVIQ-modified were: PL502: 49 (mean: 3.06); PL545: 53 (mean: 3.31); PM006: 72 (mean: 4.5); PM024: 76 (mean: 4.75). See Appendix A for the patients' responses to the individual questions.
## Neuropsychological and experimental tests
For the accuracy measures, PL518 is clearly impaired with faces, and shows a deficit (performing more than two standard deviations (SDs) from the control mean) on most individual face measures. He performs within the low-normal range on the object tests and is clearly on level with controls in the tests with word stimuli. For the RT measures, PL518 shows a deficit on most face measures (note that his RTs in the surprise recognition test may not be a good indicator of severity, as his accuracy in this test was very low). He responds with latencies within the normal range on the object tests but shows elevated RTs in the lexical decision and word reading tests.
Comparing the neuropsychological profile of PL518 to the other included patients, we find that one or more of them show deficits on the same tests/measures and in the cognitive domain(s) as PL518 (see [fig_ref] Figure 2: Radar plots showing the results of PL518 [/fig_ref] for an illustration of their cognitive profiles on the selected tests, and for an overview of test results). A comparison of the neuropsychological profile of the two architects (PL518 and PM024) shows that PM024 (with no aphantasia) shows the same pattern of performance as PL518 on most tests, including measures of face recognition, object recognition and word reading. Indeed, there is no measure on which PL518 shows a clear deficit, where PM024 is clearly within the normal range (see [fig_ref] Figure 2: Radar plots showing the results of PL518 [/fig_ref]. The key difference between the two patients, then, is in the measure of their visual imagery. Comparing PL518 to the three other bilateral patients [fig_ref] Figure 2: Radar plots showing the results of PL518 [/fig_ref] and [fig_ref] Figure 1: PL518′s performance on the Rey Complex Figure Test [/fig_ref] , again there is no domain where PL518 is clearly impaired where the other patients are consistently within the normal range. In comparison to the three bilateral patients too, then, the key difference is in visual imagery.
Indeed, there is no measure on which PL518 shows a clear deficit, where PM024 is clearly within the normal range (see [fig_ref] Figure 2: Radar plots showing the results of PL518 [/fig_ref]. The key difference between the two patients, then, is in the measure of their visual imagery. Comparing PL518 to the three other bilateral patients [fig_ref] Figure 2: Radar plots showing the results of PL518 [/fig_ref] and [fig_ref] Figure 1: PL518′s performance on the Rey Complex Figure Test [/fig_ref] , again there is no domain where PL518 is clearly impaired where the other patients are consistently within the normal range. In comparison to the three bilateral patients too, then, the key difference is in visual imagery.
## Lesion localisation
PL518′s lesion is most extensive on the right side, including damage to the occipital pole, the lingual gyrus, the whole fusiform gyrus and extending anteriorly to the parahippocampal region. On the left side, the lesion affects only the medial fusiform gyrus and lingual gyrus, while the left occipital pole, and lateral portions of the fusiform gyrus are spared. Seeand for comparisons of lesion localization for PL518 and the other patients.
First, comparing the lesions of PL518 to the architect without aphantasia (PM024) shows that PL518 has selective left hemisphere posterior medial fusiform damage extending medially and anteriorly along the collateral sulcus, and selective right hemisphere damage to the superior medial lingual gyrus. Second, comparing the lesion of PL518 to the three patients with bilateral strokes but no aphantasia shows that PL518 has selective damage in the right fusiform gyrus and a portion of the right lingual gyrus, and additional smaller areas of selective damage in PL518 are found in the left fusiform gyrus. Combined, these comparisons reveal only small areas of selective damage in PL518 in the right lingual gyrus and left posterior medial fusiform gyrus.
## Lesion localisation
PL518 s lesion is most extensive on the right side, including damage to the occipital pole, the lingual gyrus, the whole fusiform gyrus and extending anteriorly to the parahippocampal region. On the left side, the lesion affects only the medial fusiform gyrus and lingual gyrus, while the left occipital pole, and lateral portions of the fusiform gyrus are spared. Seeand for comparisons of lesion localization for PL518 and the other patients.
First, comparing the lesions of PL518 to the architect without aphantasia (PM024) shows that PL518 has selective left hemisphere posterior medial fusiform damage extending medially and anteriorly along the collateral sulcus, and selective right hemisphere damage to the superior medial lingual gyrus. Second, comparing the lesion of PL518 to the three patients with bilateral strokes but no aphantasia shows that PL518 has selective damage in the right fusiform gyrus and a portion of the right lingual gyrus, and additional smaller areas of selective damage in PL518 are found in the left fusiform gyrus. Combined, these comparisons reveal only small areas of selective damage in PL518 in the right lingual gyrus and left posterior medial fusiform gyrus. . Comparison of regions of interest within the occipital and temporal lobes affected in PL518 compared to other patients. The fusiform gyrus (FG) was segmented into four regions (FG1-4: corresponding to posterior medial, posterior lateral, anterior medial and anterior lateral, respectively) according to Lorenz and colleagues [bib_ref] Two new cytoarchitectonic areas on the human mid-fusiform gyrus, Lorenz [/bib_ref]. The occipital pole and the lingual gyrus were defined using a conventional atlas, and the parahippocampal region was identified using the images from Bouyeure and colleagues [bib_ref] Three-dimensional probabilistic maps of mesial temporal lobe structures in children and adolescents'..., Bouyeure [/bib_ref]. An x indicates that at least 10% of the corresponding region of interest was affected by a patient's stroke. Occipital Pole
## Patient
[formula] x x x x FG 1 x x FG 2 x x FG 3 x x x FG 4 x x x Lingual Gyrus x x x x Parahipp. Gyrus x x x [/formula]
# Discussion
The present study reports case PL518, an architect who lost his ability for visual imagery following a bilateral PCA stroke 35 months prior to this investigation. We compare his performance across a range of perceptual and cognitive tests and a visual imagery questionnaire with four other PCA stroke patients, an architect with a large right hemisphere lesion and three bilateral cases. PL518′s profile on the perceptual and cognitive tests was similar to other cases with the exception that PL518 reported severe visual imagery problems following his stroke. Lesion profiles were also comparable with the exception that PL518 showed selective damage in the right lingual gyrus and . Comparison of regions of interest within the occipital and temporal lobes affected in PL518 compared to other patients. The fusiform gyrus (FG) was segmented into four regions (FG1-4: corresponding to posterior medial, posterior lateral, anterior medial and anterior lateral, respectively) according to Lorenz and colleagues [bib_ref] Two new cytoarchitectonic areas on the human mid-fusiform gyrus, Lorenz [/bib_ref]. The occipital pole and the lingual gyrus were defined using a conventional atlas, and the parahippocampal region was identified using the images from Bouyeure and colleagues [bib_ref] Three-dimensional probabilistic maps of mesial temporal lobe structures in children and adolescents'..., Bouyeure [/bib_ref]. An x indicates that at least 10% of the corresponding region of interest was affected by a patient's stroke.
## Patient
# Discussion
The present study reports case PL518, an architect who lost his ability for visual imagery following a bilateral PCA stroke 35 months prior to this investigation. We compare his performance across a range of perceptual and cognitive tests and a visual imagery questionnaire with four other PCA stroke patients, an architect with a large right hemisphere lesion and three bilateral cases. PL518 s profile on the perceptual and cognitive tests was similar to other cases with the exception that PL518 reported severe visual imagery problems following his stroke. Lesion profiles were also comparable with the exception that PL518 showed selective damage in the right lingual gyrus and left medial posterior fusiform gyrus. It is tempting to suggest that these are both candidate regions for specific involvement in visual imagery.
However, Bogousslavsky and colleagues [bib_ref] Lingual and fusiform gyri in visual processing: A clinico-pathologic study of superior..., Bogousslavsky [/bib_ref] described a man whose lingual gyrus was destroyed in both hemispheres, while only the middle third of the fusiform gyrus on the left side was affected. His visual imagery was intact for colors, faces (human and animal) and places (streets). The authors concluded that the fusiform gyrus and underlying white matter, rather than the lingual gyrus, was a principal structure for color integration, face recognition, visuo-verbal processing, and corresponding visual imagery. The fact that the current primary case, PL518, had selective damage to the left fusiform gyrus is also more in alignment with other research indicating that left hemisphere regions are more consistently implicated in generating mental imagery than corresponding right hemisphere regions [bib_ref] Loss of imagery phenomenology with intact visuo-spatial task performance: A case of..., Zeman [/bib_ref] [bib_ref] The neurological basis of mental imagery: A componential analysis, Farah [/bib_ref] [bib_ref] Selective deficit of mental visual imagery with intact primary visual cortex and..., Moro [/bib_ref] [bib_ref] Vivid visual mental imagery in the absence of the primary visual cortex, Bridge [/bib_ref] [bib_ref] The neural correlates of visual mental imagery: An ongoing debate, Bartolomeo [/bib_ref] [bib_ref] A functional MRI study of mental image generation, D'esposito [/bib_ref] [bib_ref] The laterality of mental image generation: A test with normal subjects, Farah [/bib_ref] [bib_ref] The neural correlates of visual imagery vividness-An fMRI study and literature review, Fulford [/bib_ref] [bib_ref] A computational analysis of mental image generation: Evidence from functional dissociations in..., Kosslyn [/bib_ref] [bib_ref] A critical review of mental imagery defects, Trojano [/bib_ref] [bib_ref] The neural correlates of visual imagery: A co-ordinate-based meta-analysis, Winlove [/bib_ref].
A seeming counterexample comes from de Gelder and colleagues [bib_ref] Visual imagery influences brain responses to visual stimulation in bilateral cortical blindness, De Gelder [/bib_ref]. They described patient TN who had bilateral cortical blindness due to lesions in the primary visual cortices in both hemispheres. The lesion also reached some high-level visual ventral areas, including parts of the left posterior fusiform gyrus. Despite this damage, de Gelder and colleagues [bib_ref] Visual imagery influences brain responses to visual stimulation in bilateral cortical blindness, De Gelder [/bib_ref] argued that TN was able to generate visual mental imagery. However, judging from the lesion reconstruction (their [fig_ref] Figure 2: Radar plots showing the results of PL518 [/fig_ref] , the left medial posterior fusiform might have been at least partially spared in this patient. Also, the imagery tasks used involved a significant motor or action component, and correspondingly TN's functional activation pattern in the imagery conditions was primarily fronto-parietal.
Fitting with a role of the left fusiform gyrus in visual imagery, some developmental prosopagnosics appear to have functional abnormalities in this region [bib_ref] Left hemisphere abnormalities in developmental prosopagnosia when looking at faces but not..., Gerlach [/bib_ref] [bib_ref] Congenital prosopagnosia: Multistage anatomical and functional deficits in face processing circuitry, Dinkelacker [/bib_ref] [bib_ref] Early left-hemispheric dysfunction of face processing in congenital prosopagnosia: An MEG study, Dobel [/bib_ref] as well as reduced or absent mental imagery, not only for faces but also for objects and scenes [bib_ref] Visual mental imagery in congenital prosopagnosia, Grüter [/bib_ref]. examined face imagery in prosopagnosics for featural imagery (questions regarding facial features, e.g., "Who has a wider mouth: Sophia Loren or Ingrid Bergman?") as well as facial configurations (questions on overall face shape or configuration, e.g., "Who has the more angular face: George Washington or Abraham Lincoln?"). In acquired prosopagnosics, they found that right-sided occipito-temporal lesions affected imagery for facial configuration but not for facial features, while bilateral lesions additionally impaired imagery for facial features [bib_ref] Face imagery and its relation to perception and covert recognition in prosopagnosia, Barton [/bib_ref]. This fits well with the left fusiform gyrus responding more to facial features while the right fusiform gyrus is more involved in configural processing [bib_ref] Hemispheric asymmetries for whole-based and part-based face processing in the human fusiform..., Rossion [/bib_ref]. It is possible that the generation of mental imagery heavily relies on the assembly of separately stored visual features or parts, and that this generation of multipart images specifically taxes left hemisphere regions [bib_ref] A case study of mental imagery deficit, Farah [/bib_ref] [bib_ref] A computational analysis of mental image generation: Evidence from functional dissociations in..., Kosslyn [/bib_ref]. This is consistent with PL518 s description of the fragmented minimal visual imagery that he possibly still has (e.g., visualizing bits of elephants).
Compared to before his stroke, PL518 seems to make greater use of verbal strategies (e.g., recalling a list). If PL518 still has some mental imagery, it nonetheless mostly seems to be based on an altered strategy which could be described as motor, action-based, or spatial, such as mentally tracing a line around a shape or doing mental rotation by physically trying to move things inside the head. This is reminiscent of patient MX [bib_ref] Loss of imagery phenomenology with intact visuo-spatial task performance: A case of..., Zeman [/bib_ref] who also reported the loss of the experience of visual imagery as well as an unusual or altered strategy when attempting a mental rotation task, where he needed to match individual blocks and angles perceptually when making his decision.
The two architects, PL518 and PM024, had similar functional deficits, including prosopagnosia, but described vastly different visual imagery (minimal vs. very clear and lively). It is tempting to speculate, therefore, that the additional left hemisphere affection in PL518 contributes significantly to his disruption of imagery. In particular, the small patch in the medial left fusiform gyrus where PL518 has unique damage compared to all the four other patients presents as a good candidate for playing a critical part in the generation of visual mental imagery. While our findings indeed suggest that this region is an important node in the cerebral network underlying visual imagery, other areas, including right hemisphere ventral occipito-temporal areas, left hemisphere areas further anterior in the temporal lobe (see e.g., [bib_ref] Face imagery and its relation to perception and covert recognition in prosopagnosia, Barton [/bib_ref] , more posterior areas in the left occipital lobe, and regions outside of the ventral visual stream are also likely to partake in at least some aspects of visual imagery. For example, while mental imagery generation might mainly depend on structures in the posterior left hemisphere, right parietal regions have been found to be important for spatial comparisons of the contents of visual imagery [bib_ref] The dynamics of interhemispheric compensatory processes in mental imagery, Sack [/bib_ref] , see also [bib_ref] Environment and object mental images in patients with representational neglect: Two case..., Palermo [/bib_ref]. The right hemisphere could also have some ability to generate visual imagery for overall shape [bib_ref] A computational analysis of mental image generation: Evidence from functional dissociations in..., Kosslyn [/bib_ref] , and had we included sensitive measures of configural processing deficits in mental imagery in addition to the VVIQ, it is possible that subtle deficits in PM024 could have been discovered. It is also worth noting that the aphantasic architect PL518 had bilateral damage, while mental imagery generation could possibly be taken over by the right hemisphere in cases of unilateral left hemisphere disruption [bib_ref] The dynamics of interhemispheric compensatory processes in mental imagery, Sack [/bib_ref].
The most commonly used questionnaire to measure mental imagery is various versions of the VVIQ [bib_ref] Lives without imagery-Congenital aphantasia, Zeman [/bib_ref] [bib_ref] Visual imagery differences in the recall of pictures, Marks [/bib_ref]. The VVIQ has good psychometric qualities and vividness correlates with some other behavioral and neural measures of visual imagery [bib_ref] Imagery vividness, creativity and the visual arts, Morrison [/bib_ref] [bib_ref] Vividness of mental imagery: Individual variability can be measured objectively, Cui [/bib_ref] [bib_ref] Vividness of visual imagery depends on the neural overlap with perception in..., Dijkstra [/bib_ref]. These questionnaires do have their limitations, though, as they rely on self-reporting and only measure overall vividness of visual mental imagery. Mental imagery is, however, of a multimodal nature [bib_ref] Multimodal mental imagery, Nanay [/bib_ref] and includes for example smell, touch, sound and taste. Also, there are several different aspects of visual imagery, and in order to capture this more completely, a measure would need to include items specifically for spatial imagery, as well as imagery for colors, objects, places, faces, and even subsets of these such as featural vs. configural face imagery. More fine-grained mental imagery questionnaires and additional behavioral measures that likely rely on mental imagery, such as the clock task [bib_ref] On the different roles of the cerebral hemispheres in mental imagery: The..., Grossi [/bib_ref] [bib_ref] Comparisons of mental clocks, Paivio [/bib_ref] [bib_ref] Mental imagery and autobiographical memory in Alzheimer's disease, El Haj [/bib_ref] , the taller/wider task [bib_ref] A computational analysis of mental image generation: Evidence from functional dissociations in..., Kosslyn [/bib_ref] [bib_ref] Mental imagery and autobiographical memory in Alzheimer's disease, El Haj [/bib_ref] , or mental letter construction [bib_ref] Visually-and motor-based knowledge of letters: Evidence from a pure alexic patient, Bartolomeo [/bib_ref] , animal tails test [bib_ref] Visual and spatial mental imagery: Dissociable systems of representation, Farah [/bib_ref] , drawing objects from memory [bib_ref] Visual object agnosia, prosopagnosia, achromatopsia, loss of visual imagery, and autobiographical amnesia..., Ogden [/bib_ref] [bib_ref] Visual memory loss and autobiographical amnesia: A case study, Greenberg [/bib_ref] and the binocular-rivalry technique [bib_ref] The functional impact of mental imagery on conscious perception, Pearson [/bib_ref] [bib_ref] New directions in mental-imagery research, Pearson [/bib_ref] , would provide further insights into whether mental imagery deficits are due to a loss of all imagery across modalities, specific loss of visual imagery, or specific loss of subcomponents of visual imagery. Such specific aspects of mental imagery were not directly assessed in the present study.
It is still debated whether imagery and perception may be dissociated, or whether they depend on common networks. In one sense, the current results support the former as some patients with heavy damage to ventral stream areas and associated problems with visual cognition nonetheless appear to have intact visual imagery. Our neuropsychological approach suggests that some ventral stream regions might not be necessary for visual imagery despite containing information on imagined objects [bib_ref] Disentangling visual imagery and perception of real-world objects, Lee [/bib_ref] [bib_ref] Mental imagery of faces and places activates corresponding stimulus-specific brain regions, O Craven [/bib_ref] [bib_ref] Reading the mind's eye: Decoding category information during mental imagery, Reddy [/bib_ref] [bib_ref] Neural dynamics of perceptual inference and its reversal during imagery, Dijkstra [/bib_ref]. On the other hand, the areas specifically associated with PL518 s visual imagery loss are better known for their role in visual perception. A key difference between imagery and perception could however lie in their different network dynamics where imagery is dominated by top-down feedback [bib_ref] Disentangling visual imagery and perception of real-world objects, Lee [/bib_ref] [bib_ref] Neural dynamics of perceptual inference and its reversal during imagery, Dijkstra [/bib_ref] ; this could even map onto different cortical layers within the same region [bib_ref] Laminar fMRI: Applications for cognitive neuroscience, Lawrence [/bib_ref] [bib_ref] Two distinct feedback codes in V1 for 'real'and 'imaginary'internal experiences, Bergmann [/bib_ref]. Even if a region serves both perception and imagery, is it still possible that distinct computations and separable subpopulations of neurons are involved.
It should finally be noted that individual differences in premorbid ability for imagery might play a role in the effects of stroke on these abilities. PL518 reported that his abilities for visual mental imagery had been above average before his stroke. These abilities had enabled him to visualize the spatial and visual attributes of buildings and rooms in rich details and contributed greatly to his achievements as an architect. This fits a general pattern noted by Farah [bib_ref] The neurological basis of mental imagery: A componential analysis, Farah [/bib_ref] where many cases of acquired deficits in visual imagery involved people whose day-to-day activities had likely demanded visualization. As the normal variability in visual imagery from congenital aphantasia to hyperphantasia becomes better understood, this factor may perhaps help explain variability in the effect of brain injury on visual imagery.
# Conclusions
While several brain regions in both hemispheres are involved in different aspects of mental imagery, our results indicate that the right lingual gyrus and especially the left posterior medial fusiform gyrus are candidate regions for specific involvement in visual imagery. These regions were only affected in the aphantasic architect PL518 compared to non-aphantasic patients with comparable cognitive and perceptual deficits.
# Supplementary materials:
The following are available online at http://www.mdpi.com/2076-3425/10/2/59/s1, . Individual neuropsychological patient raw data. [fig_ref] Figure 1: PL518′s performance on the Rey Complex Figure Test [/fig_ref]. Radar plots comparing individual bilateral patients' neuropsychological profile to PL518. Acknowledgments: We wish to thank Matthew A. Lambon Ralph for contributing to the BoB-project and thus making the present work possible, and Fakutsi for support during the preparation of this manuscript.
## 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. . Individual responses on the Vividness of Imagery Questionnaire (VVIQ)-modified [bib_ref] Lives without imagery-Congenital aphantasia, Zeman [/bib_ref]. Total and mean scores in bold.
# Appendix a
## Vviq
[fig] Figure 1: PL518′s performance on the Rey Complex Figure Test. (A): Copy. (B): three-minute recall. [/fig]
[fig] Figure 2: Radar plots showing the results of PL518 (in red) and the other patients on all the included measures of object, word, and face recognition. Numbers denote z-scores based on the control means and SDs for the respective tests. Impaired performance (>−2 SDs from the control mean) is marked by the dotted grey line, and scores closer to the center are more impaired (represents lower accuracy and slower RTs). Left panel (A,C) shows accuracy, right panel (B,D) shows RTs. Upper panel (A,B) shows PL518 vs. PM024 (architect with right hemisphere lesion). Lower panel (C,D) shows PL518 vs. the other bilateral patients. See individual radar plots comparing PL518 individually to bilateral patients inFigure S1 [/fig]
[fig] Figure 3: (A) shows lesion overlap between PL518 and the other bilateral patients. (B) shows lesion overlap between PL518 and PM024. Left hemisphere depicted on the left. The lesions are in MNI space overlaid on the MNI152 template. See Figure S2 for individual comparisons of PL518 and the bilateral patients. [/fig]
|
Interhemispheric connectivity of primary sensory cortex is associated with motor impairment after stroke
Neuroimaging-derived markers are used to model post-stroke impairment. Among these, lesion size, corticospinal-tract lesion-load (CST-LL) and resting-state functional-connectivity (rs-FC) have been correlated with impairment. It has been shown that the sensory cortex (S1) is associated with motor learning and is essential for post-stroke recovery; yet stroke-induced changes in S1 connectivity alone are yet to be investigated. We aim to determine whether interhemispheric rs-FC could be used to refine imaging models of stroke-related impairment. Subjects' post-stroke and age-matched controls underwent rs-fMRI. Stroke-related disability was correlated with lesion size, CST-LL and interhemispheric S1 and M1 rs-FC as independent seeds. Regression analyses were performed to assess the contribution of these markers in stroke-related deficits. Post-stroke subjects showed an asymmetrical pattern of rs-FC in which affected hemisphere S1 and M1 were mostly connected with ipsi-lesional regions. Correlations between rs-FC and stroke-severity were found. Adding rs-FC of S1 to the regression model of impairment decreased the variance 31% compared to lesion size only. After a stroke, S1 interhemispheric connectivity is decreased, with S1 only connected with ipsi-lesional regions. This asymmetry correlates with neurological and motor impairment. Furthermore, when combined with lesion anatomical measures, S1 connectivity might be an important marker in explaining stroke outcome.Despite recent advances in the diagnosis and treatment of cerebrovascular disease, stroke remains the leading cause of adult impairment worldwide 1 . Identifying a way to accurately determine post-stroke outcome is crucial to achieve the best results in rehabilitation. Neuroimaging techniques have been used to study anatomical and functional markers with respect to neurological and motor impairment. Results of anatomical imaging studies suggest that both lesion size and corticospinal tract lesion load (CST-LL, i.e., overlap of the lesion with the CST) correlate with post-stroke neurological 2 , upper-3,4 and lower-limb 5,6 motor deficit.Anatomical markers indicate the extent of damage caused by stroke but do not however, provide key information on the amount of cerebral reorganization that can take place in intact and peri-infarct tissue 7 that may significantly impact behavioral outcome. In fact, it has been shown that regions distant to the infarct play an important role in post-stroke neuroplasticity and functional recovery 8 .Functional markers such as resting-state functional magnetic resonance imaging (rs-fMRI) allow for the assessment of whole-brain functional connectivity (FC) at rest by measuring spontaneous modulation in blood oxygenation level-dependent (BOLD) signal without any task, providing a measure of temporal correlation of blood oxygenation fluctuations between brain regions 9 as a surrogate marker of synchronized neuronal activity at rest. Although typically unilateral, stroke induces interhemispheric changes. Not only does it reduce activity in the affected hemisphere but causes, via transcallosal inhibitory fibers, a release of activity in the unaffected hemisphere (transcallosal disinhibition). In patients with stroke, this increased activity in the non-lesioned hemisphere
has repeatedly been demonstrated [bib_ref] Influence of interhemispheric interactions on motor function in chronic stroke, Murase [/bib_ref] [bib_ref] Somatotopy and movement representation sites following cortical stroke, Cramer [/bib_ref] and this lack of inhibition from the lesioned hemisphere likely interferes with recovery [bib_ref] Influence of interhemispheric interactions on motor function in chronic stroke, Murase [/bib_ref] [bib_ref] Technology insight: noninvasive brain stimulation in neurology-perspectives on the therapeutic potential of..., Fregni [/bib_ref]. Consequently, studies in chronic stroke combining anatomical and rs-FC of interhemispheric connectivity found that patients with less interhemispheric M1 rs-FC and lower fractional anisotropy are more impaired [bib_ref] Resting state interhemispheric motor connectivity and white matter integrity correlate with motor..., Chen [/bib_ref].
Measures of FC might therefore be good candidates to further define behavioral outcomes post-stroke, in addition to anatomical markers. Most studies have focused on connectivity of the motor cortex region [bib_ref] Resting state interhemispheric motor connectivity and white matter integrity correlate with motor..., Chen [/bib_ref] [bib_ref] Resting interhemispheric functional magnetic resonance imaging connectivity predicts performance after stroke, Carter [/bib_ref] [bib_ref] Contribution of the resting-state functional connectivity of the contralesional primary sensorimotor cortex..., Xu [/bib_ref] ; however, changes in FC in other regions may provide additional information on cerebral reorganization. For instance, changes in somatosensory cortex activity [bib_ref] Modulation dynamics in the orofacial sensorimotor cortex during motor skill acquisition, Arce-Mcshane [/bib_ref] [bib_ref] Sensorimotor adaptation changes the neural coding of somatosensory stimuli, Nasir [/bib_ref] [bib_ref] Functional imaging of perceptual learning in human primary and secondary somatosensory cortex, Pleger [/bib_ref] [bib_ref] Somatosensory evoked potentials show plastic changes following a novel motor training task..., Andrew [/bib_ref] and sensory networks 20 are known to accompany motor learning therefore, integrity of sensory cortex connectivity might be an important marker of motor function post-stroke. Rs-FC of the sensorimotor cortex (M1 and S1 together) have shown similar results as the motor cortex alone, namely that interhemispheric asymmetry is reduced with chronicity and that it correlates with motor impairment [bib_ref] Resting interhemispheric functional magnetic resonance imaging connectivity predicts performance after stroke, Carter [/bib_ref] [bib_ref] Contribution of the resting-state functional connectivity of the contralesional primary sensorimotor cortex..., Xu [/bib_ref]. It remains unclear however, what the contribution of the rs-FC of the primary sensory cortex (S1) alone to neurological or motor impairment really is.
The purpose of the present study is thus to identify stroke-induced changes in rs-FC of the primary sensory cortex (S1). In contrast to previous studies, we extended our analysis to identify inter and intra-hemispheric connectivity independently to both M1 and S1 regions on both affected and non-affected hemispheres. In the same participants, we also assessed lesion size and CST-LL to determine if changes in interhemispheric rs-FC could, added with anatomical markers, better describe neurological and/or motor impairment post-stroke.
# Results
Rs-FC in healthy controls. Positioning the seed on either dominant or non-dominant S1 showed both intra-and inter-hemispheric rs-FC between bilateral S1, occipital and premotor cortices [fig_ref] Figure 1: rs-FC in relation to S1-seed [/fig_ref]. In addition, parietal areas were also connected bilaterally, however, S1-parietal connectivity was mostly intra-hemispheric (BA5 and 7) when the seed was positioned over the non-dominant hemisphere. S1 connectivity was intra-hemispheric for the somatosensory areas (BA40 and 43) but also inter-hemispheric when the seed was positioned on the dominant side (BA40) only.
Positioning the seed over the dominant or non-dominant M1 in healthy controls revealed a symmetrical pattern of interhemispheric rs-FC in homotopic regions of the pre-and post-central gyri, parietal and occipital lobes and [fig_ref] Figure 2: rs-FC in relation to M1-seed [/fig_ref]. Specifically, the seed was functionally connected to itself and to interhemispheric homologous region. There was intra-hemispheric connectivity of the M1-seed with ipsilateral parietal operculum in addition to inter-and intra-hemispheric connectivity observed with the sensory, parietal and occipital cortices. M1 intra-hemispheric, not inter-hemispheric connectivity was observed with the secondary somatosensory area (BA43).
## Rs-fc changes post-stroke.
In subjects post-stroke, the pattern of interhemispheric rs-FC was asymmetrical. Using S1 as seed in stroke revealed the loss of most of the interhemispheric connectivity between the seed region and contra-lesional S1 (BA2 only; [fig_ref] Figure 1: rs-FC in relation to S1-seed [/fig_ref] , bottom-left). The seed instead remained connected intra-hemispherically with itself. S1 connectivity with somatosensory areas was intra-hemispheric and only inter-hemispheric when the seed was on the non-affected S1. The S1-parietal connectivity was purely intra-hemispheric, in contrast with controls who did show inter-hemispheric connectivity between S1 and parietal areas. There was rs-FC between the S1 seed positioned over the affected or non-affected side with the premotor cortex, similar to controls however, the inter-hemispheric connectivity between affected S1 and contra-lesional premotor cortex was not present anymore. When the affected M1 was used as the seed region [fig_ref] Figure 2: rs-FC in relation to M1-seed [/fig_ref] , bottom-left), most of the sensory regions connected to the affected M1 were only intra-hemispheric , Seed location A M1). M1 connectivity was only intra-hemispheric and on the affected side only (BA40). Furthermore, the interhemispheric M1-parietal connectivity observed in controls was not observed. Instead, M1-parietal connectivity was only present within the affected hemisphere (BA5 and 7). When the non-affected hemisphere was used as the seed region, a similar pattern was observed where connectivity between M1 (seed) and the parietal regions remained mainly within the non-affected hemisphere, where the seed was positioned. Connection of either S1 or M1 to the visual cortex was lost post-stroke.
Anatomical and function markers in relation to stroke severity. Stroke-related neurological deficit, as measured with the National Institutes of Health Stroke Scale (NIHSS) score, strongly and significantly correlated with measures of laterality quantified by a laterality index (LI) of rs-FC with the affected S1 or M1 as the seed. It also correlated with the weighted CST-LL (wCST-LL) and lesion size, as shown in . Upper-limb impairment quantified by the Chedoke-McMaster arm and hand components also strongly and significantly correlated with LI of rs-FC with the affected S1 or M1 as seed, wCST-LL, and lesion size. For the lower-limb, only lesion size correlated strongly, but not significantly, with the Chedoke-McMaster foot and leg components. Weaker correlations were found for the foot component and rs-FC LI when the affected S1 or M1 were used as seeds, and wCST-LL. No correlations were found with the leg component.
To further evaluate the added value of rs-FC in determining post-stroke impairment, a hierarchical multiple regression analysis was performed. The lesion size, as the anatomical marker, was entered in the first block of the regression as it best correlated with NIHSS and all components of the Chedoke-McMaster than wCST-LL. For the second block, LI rs-FC with affected S1 as the seed was entered as the FC marker. LI rs-FC with affected S1 seed was selected for the model over LI rs-FC with M1 as both are highly and statistically correlated with each other (r = 0.852, p = 0.007) and with the NIHSS score. A very low level of multicollinearity was present (variance inflation factor <5 for all independent outcomes). Results of the hierarchical regression analysis with NIHSS as the dependent variable showed a significant regression equation with the lesion size explaining 66.1% of the variance (F = 11.69, p = 0.014). Adding the FC variable in the second block of the regression significantly improved the relationship (R 2 change = 0.308 F = 79.028, p = 0.000), explaining 96.9% of the variance.
Results of the second hierarchical regression analysis with stroke impairment, as measured by the Chedoke-McMaster, showed that lesion size was a significant factor of arm (86%, p = 0.001), hand (60.3%, p = 0.023) and foot (68.5%, p = 0.022) impairment but not leg. The addition of LI rs-FC with the S1 seed in the second block increased arm (R 2 change = 0.103, F = 76.55, p = 0.000) from 86% to 96.8% and hand impairment regression (R 2 change = 0.268. F = 16.855, p = 0.006) from 60.3% to 87.1% and had no effect on the other components.
# Discussion
This study shows that after stroke there is a disruption of interhemispheric rs-FC between sensory areas. In contrast with previous studies that focused mainly on M1 alone or combined with S1 connectivity, we extended our analysis to identify inter and intrahemispheric connectivity to S1 and M1 regions separately. Our results suggest that after a stroke, there is a significant loss of S1 interhemispheric connectivity. Furthermore, S1 showed increased intra-hemispheric connectivity (mainly with premotor and parietal regions). Our results also demonstrate that after stroke a decreased connectivity between sensorimotor regions and the occipital cortex is present. Additionally, we showed that rs-FC between the primary sensory region may, when combined with anatomical measures of stroke lesion, be an important marker in describing stroke outcome. Moreover, lower-limb impairments are more difficult to describe in contrast to upper-limb impairment and overall neurological deficit.
The symmetrical interhemispheric rs-FC observed in control subjects resembles the sensorimotor network 21 involving the precentral, postcentral gyrus and supplementary motor areas. As with other studies, we observed that the affected hemisphere S1 or M1 was mostly connected within that same hemisphere primary somatosensory regions [bib_ref] Resting state interhemispheric motor connectivity and white matter integrity correlate with motor..., Chen [/bib_ref] [bib_ref] Resting interhemispheric functional magnetic resonance imaging connectivity predicts performance after stroke, Carter [/bib_ref] [bib_ref] Longitudinal changes of resting-state functional connectivity during motor recovery after stroke, Park [/bib_ref]. Our findings support those that showed that a bilateral and symmetrical connectivity between hemispheres is crucial to achieve normal motor performance. In the current study, decreased rs-FC between ipsi-lesional S1 or M1 and contra-lesional sensorimotor areas and parietal region was demonstrated post-stroke. On the other hand, increased intra-hemispheric connectivity between sensorimotor areas and parietal cortex was observed. Instead, functional connectivity of either S1 or M1 with the occipital cortex was lost in stroke patients. These findings coincide with previous studies that reported changes in connectivity of the ipsi-lesional M1 with non-primary sensorimotor regions, such as the fronto-parietal cortex [bib_ref] Increased cognitive load during simple and complex motor tasks in acute stage..., Puh [/bib_ref] , middle frontal gyrus, thalamus, cerebellum and non-motor brain areas such as the occipital cortex [bib_ref] Longitudinal changes of resting-state functional connectivity during motor recovery after stroke, Park [/bib_ref]. Our findings suggest that disruption of the symmetrical interhemispheric rs-FC may lead to changes in connectivity of ipsi-lesional brain regions to compensate for the loss of interhemispheric connection.
In contrast to previous studies, we additionally examined interhemispheric rs-FC when positioning the seed over the non-affected hemisphere S1 or M1 demonstrating a symmetrical and bilateral pattern of rs-FC. These results indicate that connectivity between homologous brain regions is not reciprocal but is rather lesion-dependent. This could be explained by the transcallosal disinhibition hypothesis, which proposes that after a stroke, cortical excitability is decreased due to the infarct in the affected hemisphere, therefore the affected hemisphere can no longer exerts its normal inhibitory influence on the non-affected hemisphere [bib_ref] Influence of interhemispheric interactions on motor function in chronic stroke, Murase [/bib_ref] , and/or the non-affected hemisphere exerts too much influence over the already affected one [bib_ref] Transcranial direct current stimulation in stroke recovery, Schlaug [/bib_ref]. Thus, stroke can affect regions distant to the infarct compromising the functional network involved in sensorimotor performance and changes in inter and intra-hemispheric rs-FC of the sensorimotor network which might reflect changes in response to post-stroke damage.
Upper-limb motor performance has been correlated with interhemispheric rs-FC between sensorimotor regions as measured with the upper-extremity Fugl-Meyer 13,25 assessment and Action Research Arm Test score [bib_ref] Resting interhemispheric functional magnetic resonance imaging connectivity predicts performance after stroke, Carter [/bib_ref].
In addition, wCST-LL and lesion size have also shown correlations with upper-limb motor impairment as assessed by the NIHSS arm motor score 3 and UE-FM [bib_ref] Corticospinal tract lesion load: An imaging biomarker for stroke motor outcomes, Feng [/bib_ref]. In our sample, both wCST-LL and lesion size were good markers of upper-limb impairment. Lesion size has been known to correlate with motor and neurological impairment 2 while CST damage has been shown to contribute mostly to upper-limb impairment in patients with subcortical infarct 3 , but not in those with combined cortical and subcortical lesions [bib_ref] The contribution of lesion location to upper limb deficit after stroke, Park [/bib_ref]. Thus, the stronger correlations observed in our sample with lesion size is likely attributable to the presence of both upper and lower limb deficits as well as both cortical and subcortical infarcts. The addition of rs-FC when the seed was positioned over the affected S1 further decreased the variance of the model by 31% for the upper-limb impairment. However, the current analysis showed that neither anatomical markers (wCST-LL or lesion size) were good indicators of lower-limb impairment. Adding a functional marker did not improve the relationship. Recent studies have demonstrated that post-stroke patients were able to walk even after damage of the CST suggesting that lower-limb performance is dependent on other brainstem tracts such as the reticulospinal-tract 27 , vestibulospinal-tract, rubrospinal-tract and projections from the contralesional cortex 28 which might play a major role in lower-limb impairment or spinal motor programs may play and additional role. Thus, lower-limb impairment should be considered differently than the upper-limb when evaluating potential for functional recovery. Findings in our study suggest that wCST-LL is a unique imaging surrogate marker only for upper-limb post-stroke motor outcome. Lower-limb impairments are more difficult to model -based on the integrity of a single tract-compared to upper-limb impairment and overall neurological deficit. Despite its novel results, our relatively small number of subjects requires more investigation to generalize our results and to assess whether data from the acute state could predict outcome earlier in the recovery. The seeds and masks were generated using the Jülich atlas, therefore creating reproducible masks in MNI space that are 90% inclusive for both S1 and M1. Because these masks include neurons that cross the theoretical boundaries of a given area, there exists an unavoidable overlap between masks. Although we observe a difference between S1 and M1 connectivity, suggesting that the differences we observe are region specific, a residual uncertainty as to localization remains, which is inherent to cytoarchitectonic probability maps. Functional connectivity studies have been conducted in stroke yet, the changes in tissue composition at the site of the lesion is at various stages of necrosis and gliosis, affecting the BOLD signal [bib_ref] Removal of artifacts from resting-state fMRI data in stroke, Yourganov [/bib_ref]. Although we focused our analysis in regions remote to the infarct, it cannot be excluded that the change in tissue composition might have an effect on functional connectivity. In addition, it cannot be excluded that the difference in echo-planar imaging resolution might be an additional source of variability in our group comparisons. Adding longitudinal data would further help determine the predictive value of rs-FC markers for motor impairment.
Our results demonstrate that stroke can affect the functional connectivity of regions distant to the infarct, specifically S1, potentially further compromising motor performance. In addition, we showed that asymmetry in primary sensory rs-FC, when combined with anatomical measures of stroke lesion size, can be an important marker defining post-stroke outcome. Determining the amount of impairment and especially potential for recovery, as suggested by the amount of S1 symmetry, could potentially contribute to the optimization of clinical decisions and treatment post-stroke. This study highlights the important contribution of neuroimaging techniques may play to understand the brain's reorganization after a stroke.
# Methods
Subjects. Eight subjects (mean age: 67 SD 8 years; 5 males, [fig_ref] Table 3: Stroke patient characteristics [/fig_ref] in the early chronic phase (>3 months) after a first ischemic cortical or subcortical stroke resulting in upper or lower-limb motor impairment (Chedoke-McMaster 30 <7), and eight healthy age-matched subjects (mean age: 67 SD 9 years; 5 males) participated in this study. Healthy subjects were recruited from the community and had no history of other neurological, cardiovascular or psychiatric diseases. Handedness was assessed with the Edinburgh Handedness Inventory. All participants had scores >0.75 or <−0.75, hence no mixed dominance (1 control was left handed). Prior to their participation, all subjects gave their informed consent to this study that was approved by the McGill Faculty of Medicine Institutional Review Board regulations for human subjects' studies. All methods were carried out in accordance with approved institutional guidelines and regulations.
# Mri analysis. lesion.
Prior to the analysis, images of the left handed control and patients who had left hemisphere lesions were mirrored along the Z-plane so that all dominant hemispheres and all lesions would be on the same side for each group. Stroke lesions were manually outlined in each subject T1-weighted image [fig_ref] Figure 3: Lesion location [/fig_ref]. Final infarct segmentation was independently reviewed by a stroke neurologist (AT) and segmentation was adjusted if necessary. To avoid the presence of artifacts due to the loss of neural activity at the tissue of the infarction, voxels identified as part of the stroke volume were excluded from the resting-state analysis.
Lesion Load. To assess damage to the WM fibers forming the CST, a lesion load (LL) of the affected hemisphere CST was calculated. The overlap volume of the lesion with a probabilistic CST from the Jülich atlas 31 was computed according to the wCST-LL method 3 . The probabilistic CST was linearly transformed into each individual's native MRI space. For each subject, a weighting factor was obtained from the probability of each voxel in each lesion slice intersecting with the probabilistic CST. Then, the wCST-LL was calculated according to the following equation (1).
[formula] = Σ ⋅ ⋅ ⋅ - w n n CST LL 1 10 I( ) (voxel volume) weighting factor (1) max [/formula]
where n max is the total number of intersecting voxels between the lesion volume and fiber tract and I(n) is the probability of the n th voxel (as represented in the probabilistic tract).
Resting-State image preprocessing. Data analysis was performed using a resting-state pipeline developed by the Center for Research on Brain, Language and Music (www.crblm.ca) relying on FSL 5.0.8 (FMRIB Software Library, Oxford, UK) [bib_ref] Bayesian analysis of neuroimaging data in FSL, Woolrich [/bib_ref] and Matlab software (www.mathworks.com). Image preprocessing consisted of: (1) Discarding the first 5 volumes in each scan series to allow steady-state imaging, (2) slice-time correction (using Fourier-space time-series phase shifting), (3) brain extraction using FSL Brain Extraction Tool 33 , (4) motion correction, using a 6-parameter affine transformation built-in in FLIRT (FMRIB Linear Image Registration Tool) 33 , (5) global intensity normalization, (6) spatial smoothing with a Gaussian kernel of FWHM 6 mm (7) temporal high-pass filtering (Gaussian-weighted least-square straight-line fitting with σ = 100.0 s). Two transformations were performed: (1) from functional image to T1-weighted anatomical image (using a 7-Degrees of Freedom (DOF) transformation), (2) from MRI native to MNI space (using a 12-DOF linear affine transformation; voxel size 2 × 2 × 2 mm).
## Seed masks.
A seed-based connectivity analysis was used to identify voxels temporally correlated with BOLD signal fluctuations with bilateral M1 or S1. Masks were created in MNI space using a combination of the probabilistic Jüelich atlas 34 BA4a and BA4p for M1 and BA1, BA2, BA3a and BA3b for S1 [fig_ref] Figure 4: Schematic representation of the seed regions [/fig_ref]. The Jülich atlas is a cytoarchitectonic probability map, which includes S1 or M1 neurons that lay outside the typical Brodmann area boundaries. The resulting masks were thresholded at a voxel probability of 10% (ensuring that most voxels associated with the ROI were included), binarized, and linearly transformed into MNI space. The resulting M1 seed thus included 99% of BA4a and BA4p on both right and left sides but also intersected with BA6 (36% left and 38% right) and BA3b (81% left and 60% right). The resulting S1 seed comprised 99% of BA3b on the left and 98% on the right, 96% of BA2 on the left and 93% right, 92% of BA1 on the left and 97% on the right, but also overlapped with BA4a (63% left and 66% right). We accepted this overlap with adjacent regions because we wanted to ensure that >90% of the target region was used as seed.
Resting-state FC map computation. The BOLD signal from the preprocessing step was entered in a regression analysis, using a general linear model with the seed region time-series and the nuisance variables time-series as dependent variables, with FEAT (FMRIB Expert Analysis Tool) [bib_ref] General multilevel linear modeling for group analysis in FMRI, Beckmann [/bib_ref]. Physiological noise removal improves the results of the FC analysis at rest, therefore, the time-series for nuisance variables was calculated and excluded from the analysis as follows: first, the time series for Cerebrospinal Fluid (CSF), WM, and grey matter (GM) were extracted. The resulting segmented CSF, WM and GM images were thresholded (ensuring 80% tissue type probability) and transformed into functional space. The mean time-series was calculated by averaging the times-series from all voxels within the seed region. A total of nine nuisance regressors were used: CSF, WM, global signal, and six motion parameters (x, y, and z translations and rotations obtained from the motion-correction step). Relative motion was similar between groups (stroke: 0.06 SD 0.03 mm vs. controls 0.07 SD 0.05 mm, t = −0.457, p = 0.654). For each individual, a separate multiple regression analysis was performed on the time-series of the nuisance signal using FEAT. This way, the nuisance signals are removed and the residual image represents the temporal correlation of the corrected BOLD signal in the seed region with each voxels forming rs-FC maps with cluster thresholding at Z > 2.3, p = 0.05 corrected using a Gaussian Random Field (GRF) theory.
Lateralization Index. To quantify changes in inter-hemispheric FC, a LI was calculated. All voxels functionally connected to M1 on each hemisphere were calculated as a mean difference ratio for the seed positioned over the affected M1. The same procedure was repeated for the seed position over the affected S1 but for the mean difference ratio of bilateral S1. The same masks used to generate the seed region were used to calculate LI: (ΣFC A/D − ΣFC NA/ND )/(ΣFC A/D + ΣFC NA/ND ) where ΣFC A/D denotes the number of voxels functionally connected to the seed inside the affected (patients) or dominant (controls) sensory or motor area (intrahemispheric connectivity) and ΣFC NA/ND is the number of voxels functionally connected to the seed inside the non-affected/ non-dominant hemisphere (interhemispheric connectivity). LI closer to 1 thus indicates connectivity within the affected (or dominant) hemisphere, i.e., more intra-hemispheric connectivity whereas an LI closer to −1 within the non-affected (or non-dominant), would indicate mainly inter-hemispheric connectivity (i.e., seed to opposite hemisphere, not to itself). Values close to 0 indicate symmetry, where intra-and inter-hemispheric connectivity is similar.
Statistical Analyses. Z-statistic images from the individual-level analysis were entered into to a group-level analysis GLM using FEAT mixed-effects model with a cluster thresholding Z ≥ 3, p ≤ 0.05 corrected using a GRF theory. Significant areas connected to the seed and their locations in MNI coordinates were identified. Pearson's correlation was used to determine whether a subject's anatomical and functional markers correlated with their NIHSS and Chedoke-McMaster upper and lower-limb scores. Hierarchical regression analyses were performed using wCST-LL, lesion size and LI values as independent variables to model post-stroke impairment.
Data availability. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
[fig] Figure 1: rs-FC in relation to S1-seed. Top panel shows control group and bottom panel shows post-stroke group connectivity maps when the seed is positioned over S1.SCIENtIfIC REPORtS | (2018) 8:12601 | DOI:10.1038/s41598-018-29751-6 [/fig]
[fig] Figure 2: rs-FC in relation to M1-seed. Top panel shows control group and bottom panel shows post-stroke group connectivity maps when the seed is positioned over M1. [/fig]
[fig] Figure 3: Lesion location. (A) Affected; NA: Non-affected. Location of the lesion superimposed in each individual T1-weighted image. The slice with the largest lesion size is presented. [/fig]
[fig] Figure 4: Schematic representation of the seed regions. Seeds were created using a combination of the probabilistic Jüelich atlas. [/fig]
[table] Table 3: Stroke patient characteristics. R: Right; L: Left; S: subcortical; C: cortical; MCA: Middle Cerebral Artery; *Stroke size also represents the volume excluded for the resting-state analysis. * * National Institute of Health Stroke Scale (0 = normal); † Chedoke McMaster Stroke Assessment (7 = normal); md: missing data; participant could not be tested due to an acute injury suffered between image acquisition and clinical assessmen. [/table]
|
CRL4BDCAF11 E3 ligase targets p21 for degradation to control cell cycle progression in human osteosarcoma cells
Cell cycle progression in mammals is strictly controlled by a number of cyclin-dependent kinases (CDKs) and CDK inhibitors (CKIs), the expression of which is often dysregulated in cancer cells. Our previous work revealed that Cullin 4B (CUL4B), a critical component of the Cullin4B-RING E3 ligase complex (CRL4B), is overexpressed in human osteosarcoma cells through an unknown mechanism.Here, we demonstrated that CUL4B forms an E3 ligase with RBX1 (RING-box 1), DDB1 (DNA damage binding protein 1), and DCAF11 (DDB1 and CUL4 associated factor 11) in human osteosarcoma cells. In vitro and in vivo ubiquitination analyses indicated that CRL4B DCAF11 E3 ligase was able to specifically ubiquitinate a CDK inhibitor-p21 Cip1 at K16, K154, K161 and K163 but not at K75 and K141. Knocking down any component of the CRL4B DCAF11 complex, including CUL4B, DDB1 or DCAF11, using short hairpin RNAs (shRNAs) attenuated the ubiquitination level of p21 Cip1 , inhibited osteosarcoma cell proliferation, led to cell cycle arrest at S phase, and decreased colony formation rate. Taken together, our data suggest that the CRL4B DCAF11 complex represents a unique E3 ligase that promotes the ubiquitination of p21 Cip1 and regulates cell cycle progression in human osteosarcoma cells.Both prokaryotic and eukaryotic cells are controlled by an ordered series of events known as the cell cycle, which includes the G 0 , G 1 , S, G 2 and M phases 1, 2 . The cell cycle is strictly controlled by a number of regulatory partner pairs: the cyclins and the cyclin-dependent kinases (CDKs) 34.] Of these regulatory partners, Cyclin A-CDK2 mainly functions in S phase; Cyclin D-CDK4, Cyclin D-CDK6 and Cyclin E-CDK2 regulate the transition from G 1 to S phase; and Cyclin B-CDK1 regulates progression from G 2 to M phase 3, 4 . Cell cycle progression from one phase to the next is controlled by checkpoints, including the G 1 , G 2 /M and metaphase checkpoints 5, 6 . In addition, an effector protein family known as CDK inhibitors (CKIs) also plays important roles in regulating cell cycle progression by suppressing CDK functions 3, 7 . Two families of CKIs, including CDK interacting protein/kinase inhibitory protein (Cip/Kip) and inhibitor of kinase 4/alternative reading frame (INK4a/ARF), are able to disrupt cell cycle progression by affecting different CDKs 8, 9 . For example, members of the Cip/Kip family, including p21, p27 and p57, can suppress CDK2 activity, while members of the INK4a/ARF family, such as INK4A (p16), INK4B (p15), INK4C (p18) and INK4D (p19), are able to inhibit the activities of CDK4 and CDK6 8-10 . Dysregulation of either CDKs or CKIs can disrupt cell cycle progression, thereby resulting in the pathogenesis of a number of diseases, including cancer 10 . Expression of these CDKs and CKIs can be regulated at both the transcriptional and post-transcriptional levels. One example of post-transcriptional regulation is ubiquitination of p21 Cip1 and p27 Kip by different E3 ligases, such as SCF Skp2 and CRL4 Cdt2 11-14 .Eukaryotic organisms contain a family of hydrophobic proteins known as Cullins, which mainly function as scaffolds and which combine with RING proteins and adaptor proteins to form ubiquitin E3 ligase-Cullin-RING ligases (CRLs) 12, 14, 15 . The CRLs recognize different substrates and affect a wide variety of cellular processes, including tumourigenesis 12 . Of particular interest in our studies are the CRL4 E3 ligases, which are formed by Published: xx xx xxxx OPEN www.nature.com/scientificreports/ 2 Scientific RepoRts | 7: 1175 |Cullin 4 (CUL4), RING-box protein 1 (RBX1), the adaptor protein-damaged DNA binding protein 1 (DDB1), and the DDB1 and CUL4-associated factors (DCAFs)12,14,15. All of the CRL4s in different organisms share a similar core architecture, in which E3 ligase activity is determined by CUL4-RBX1 and substrate specificity is controlled by DCAFs 12, 14-17 . More than 100 DCAFs have been identified based on characteristic motifs, including WD40 repeats, WDxR motifs, and DDB boxes 18 . The human genome encodes two CUL4 proteins, CUL4A and CUL4B, which share 82% protein sequence identity without showing obvious functional redundancy 17, 18 . CUL4A overexpression is widely reported in different cancers, including breast cancer 19 , ovarian cancer 20 , hepatocellular carcinomas 21 , adrenocortical carcinomas 22 , and childhood medulloblastoma 23 , by targeting different substrates such as DDB2, p12, CDT1, STAT1, Chk1 and p21 Cip1 18-23 . In recent years, several studies have determined that CUL4B is also overexpressed in some cancer types, such as oesophageal carcinomas and HeLa cells, by targeting H2AK119 and Cyclin E, respectively 24, 25 . Our previous work also identified CUL4B overexpression in osteosarcoma cells through an unknown molecular mechanism 26 .To illuminate the molecular function of CUL4B, especially to determine interacting proteins and to identify the substrates of CRL4B E3 ligase in osteosarcoma cells, we first confirmed interactions between CUL4B and RBX1 or DDB1 in vivo and in vitro. By examining the protein levels of several DCAFs in osteosarcoma cells, we found that DCAF11 was specifically upregulated, whereas other DCAFs, such as DCAF1, 4, 8, and 15, were not, and we then determined the interactions between DCAF11 and DDB1. By tagging DCAF11 with Flag, followed by immunoprecipitation (IP) and mass spectrometry (MS), we identified proteins that associate with DCAF11. We then identified a specific protein, p21 Cip1 , which interacts with DCAF11. Our in vivo and in vitro studies support a model in which CRL4B DCAF11 E3 targets p21 Cip1 for ubiquitination to control cell cycle progression in human osteosarcoma cells.
# Results
CUL4B is upregulated at both the transcriptional and the post-transcriptional levels in human osteosarcoma cells. The human genome encodes seven Cullins, CUL1, 2, 3, 4A, 4B, 5, and 7, which function as scaffolds to facilitate the assembly of E3 ligase complexes and transfer ubiquitin from E2 to substrates [bib_ref] Cullin family proteins and tumorigenesis: genetic association and molecular mechanisms, Chen [/bib_ref]. Dysregulation of these Cullin members has been broadly reported to contribute to tumourigenesis through diverse mechanisms such as their involvement in DNA damage and repair, cell cycle progression, and the ubiquitination of oncoproteins or tumour suppressors [bib_ref] Cullin family proteins and tumorigenesis: genetic association and molecular mechanisms, Chen [/bib_ref]. Our previous work revealed that the CUL4B gene was overexpressed in the osteosarcoma cell line Saos-2 [bib_ref] CUL4B promotes proliferation and inhibits apoptosis of human osteosarcoma cells, Chen [/bib_ref]. However, we did not assess whether only CUL4B expression, and not the expression of the other six Cullin genes, was upregulated in osteosarcoma cells. To this end, we examined the mRNA and protein levels of these seven Cullins in four human osteosarcoma cell lines, U2OS, MG63, Saos-2 and HOS, versus the osteoblast cell line hFOB1. [bib_ref] The human homologue for the Caenorhabditis elegans cul-4 gene is amplified and..., Chen [/bib_ref] , a non-cancerous cell line that is commonly used as a control in osteosarcoma studies. Quantitative RT-PCR (qRT-PCR) analysis indicated that CUL4B mRNA levels were specifically upregulated (~3-4-fold induction) in these four osteosarcoma cell lines compared with hFOB1.19 cells [fig_ref] Figure 1: Human osteosarcoma cell lines have elevated CUL4B expression [/fig_ref]. Among the other six Cullins, only CUL4A was slightly overexpressed (~1.4-1.7-fold induction) in osteosarcoma cells, but not the other five Cullins, including CUL1, CUL2, CUL3, CUL5 and CUL7 [fig_ref] Figure 1: Human osteosarcoma cell lines have elevated CUL4B expression [/fig_ref]. Consistently, western blot (WB) analysis also indicated that the CUL4B protein was significantly induced (~3-4fold induction) in all four osteosarcoma cell lines and that CUL4A was slightly overexpressed (~1.3-1.6-fold induction), whereas other Cullin family proteins were not [fig_ref] Figure 1: Human osteosarcoma cell lines have elevated CUL4B expression [/fig_ref]. These results suggested that CUL4B is specifically upregulated at both the transcriptional and the post-transcriptional levels in human osteosarcoma cells, implying an important role in the pathogenesis of human osteosarcoma. Given the similar CUL4B expression levels in these four osteosarcoma cell lines, we performed the following experiments in only U2OS and/or Saos-2 cells.
CUL4B interacts with RBX1, DDB1 and DCAF11 to form an E3 complex in human osteosarcoma cells. As previously described 2, 14-17 , CUL4B can associate with RBX1, DDB1 and DCAFs to form complexes, and DCAFs further recognize and ubiquitinate different substrates in a variety of cancer types. To determine whether CUL4B can interact with RBX1 and DDB1 in osteosarcoma cells, we first constructed the pCDNA3-CUL4B-Flag vector, which we then transfected into hFOB1. [bib_ref] The human homologue for the Caenorhabditis elegans cul-4 gene is amplified and..., Chen [/bib_ref] , U2OS and Saos-2 cells. After IP experiments with anti-Flag-agarose, we detected whether CUL4B pulled down DDB1 and RBX1. As shown in [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref] , the CUL4B-Flag immunocomplex from hFOB1.19, U2OS or Saos-2 cells precipitated not only DDB1, but also RBX1. However, in U2OS cells transfected with the empty vector pCDNA3-Flag (negative control), RBX1 and DDB1 failed to immunoprecipitate [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref]. In addition, the same amount of CUL4B-Flag protein pulled down much more DDB1 and RBX1 protein in U2OS and Saos-2 osteosarcoma cells than in hFOB1.19 cells [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref] , suggesting that DDB1 and RBX1 are also upregulated in osteosarcoma cells. To further verify the interactions between CUL4B-DDB1 and CUL4B-RBX1, we constructed and transformed the following combinations of plasmids into AH109 yeast cells: The pCDNA3-CUL4B-Flag or pCDNA3-Flag empty vector was transformed into hFOB1.19, U2OS or Saos-2 cells. After 48 h incubation, Flag-tagged proteins were immunoprecipitated with anti-Flag agarose. The pull-down products were analysed by immunoblotting with anti-Flag, anti-RBX1, or anti-DDB1 antibodies. (B) Interactions between CUL4B/DDB1 and CUL4B/ RBX1 in yeast. The pGBKT7-CUL4B plasmid was co-transformed with pGADT7-DDB1 or pGADT7-RBX1 into the yeast strain AH109. Growth of the transformed yeast was assayed on media lacking Trp and Leu (top panel) or lacking Trp, Leu, and His (bottom panel). Columns in each panel represent serial 10-fold dilutions. (C) Flag-tagged CUL4B associated with DCAF11 in vivo. The pCDNA3-CUL4B-Flag plasmid was transformed into hFOB1. [bib_ref] The human homologue for the Caenorhabditis elegans cul-4 gene is amplified and..., Chen [/bib_ref] (1), U2OS (2) or Saos-2 (3) cells. After 48 h incubation, Flag-tagged proteins were immunoprecipitated with either protein A agarose or anti-Flag-agarose. The cell lysate and pull-down products were analysed by immunoblotting with anti-CUL4B, anti-DCAF1, anti-DCAF4, anti-DCAF8, anti-DCAF11 or anti-DCAF15 antibodies. (D) DCAF11 was overexpressed in human osteosarcoma cell lines. Cell lysates were applied to WB analyses with specific antibodies against CUL4B, DDB1, RBX1, DCAF11, DCAF4, or β-Actin.
[formula] pGBKT7−CUL4B + pGADT7−DDB1, pGBKT7 + pGADT7−DDB1, [/formula]
Scientific RepoRts | 7: 1175 | DOI:10.1038/s41598-017-01344-9 pGBKT7−CUL4B + pGADT7, pGBKT7−CUL4B + pGADT7−RBX1, and pGBKT7 + pGADT7−RBX1. The transformed yeast cells were then plated onto synthetic dropout medium, and positive colonies were selected and assayed to determine these interactions. The yeast two-hybrid (Y2H) results showed that CUL4B also interacts with DDB1 and RBX1 in yeast cells [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref]. However, under the same growth conditions, CUL4B, DDB1 and RBX1 failed to interact with the corresponding empty vector, including pGADT7 and pGBKT7 [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref]. Thus, the in vivo Co-IP and in vitro Y2H studies indicated that CUL4B forms a complex with DDB1 and RBX1.
Thus far, the complex still lacks a DCAF substrate recognition receptor to form a CRL4B E3 ligase. Several DCAF proteins, such as DCAF1, 4, 8, 11 and 15, have been reported to interact with DDB1 in different cancer cells [bib_ref] EZH2 generates a methyl degron that is recognized by the DCAF1/DDB1/CUL4 E3..., Lee [/bib_ref] [bib_ref] The CUL4-DDB1 ubiquitin ligase complex controls adult and embryonic stem cell differentiation..., Gao [/bib_ref] [bib_ref] Molecular architecture and assembly of the DDB1-CUL4A ubiquitin ligase machinery, Angers [/bib_ref] [bib_ref] A genome-scale CRISPR-Cas9 screening method for protein stability reveals novel regulators of..., Wu [/bib_ref]. To determine whether any of these proteins could interact with DDB1 in human osteosarcoma cells, the pCDNA3-CUL4B-Flag plasmid was transformed into hFOB1.19, U2OS or Saos-2 cells. We then performed IP experiments with anti-Flag-agarose and detected whether CUL4B pulled down DCAF1, 4, 8, 11 and 15. As shown in [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref] , the CUL4B-Flag immunocomplex from hFOB1.19, U2OS or Saos-2 cells pulled down DCAF4 and DCAF11 [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref] , but not DCAF1, 8 and 15, indicating that these two DCAFs form complexes with CUL4 and DDB1 in osteosarcoma cells. Comparing the DCAF4 and DCAF11 protein levels in cell lysates and in the CUL4B-Flag immunocomplex revealed that DCAF11 was upregulated in the cell lysates of U2OS and Saos-2 compared with hFOB1.19 cells [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref] , and the same amount of CUL4B-Flag immunocomplex was able to pull down much more DCAF11 protein in osteosarcoma cells than in hFOB1.19 cells [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref]. However, DCAF4 protein levels were not obviously altered in these cells [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref]. These results suggested that DCAF11 might play a major role because of its stronger interaction with DDB1 compared with DCAF4 in osteosarcoma cells. To confirm this hypothesis, we examined DCAF4 and DCAF11 protein levels in different osteosarcoma cell lines. As shown in [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref] , DCAF11, instead of DCAF4, was significantly induced in U2OS, MG63, Saos-2 and HOS osteosarcoma cell lines compared with hFOB1.19 cells. These results clearly demonstrated that CUL4B forms an E3 complex with RBX11, DDB1 and DCAF11, termed CRL4B DCAF11 , in osteosarcoma cells, and more importantly, that the components of this complex were highly expressed in osteosarcoma cells.
CRL4B E3 ligase recognizes p21 as a substrate in human osteosarcoma cells. CRL4 E3 ligases have been reported to recognize many substrates through different DCAFs in different cancer types [bib_ref] Cullin family proteins and tumorigenesis: genetic association and molecular mechanisms, Chen [/bib_ref] [bib_ref] PCNA-dependent regulation of p21 ubiquitylation and degradation via the CRL4Cdt2 ubiquitin ligase..., Abbas [/bib_ref] [bib_ref] Targeting Cullin-RING E3 ubiquitin ligases for drug discovery: structure, assembly and small-molecule..., Bulatov [/bib_ref] [bib_ref] RING-type E3 ligases: master manipulators of E2 ubiquitinconjugating enzymes and ubiquitination, Metzger [/bib_ref] [bib_ref] CRL4(CDT2) targets CHK1 for PCNA-independent destruction, Huh [/bib_ref] [bib_ref] EZH2 generates a methyl degron that is recognized by the DCAF1/DDB1/CUL4 E3..., Lee [/bib_ref] [bib_ref] The CUL4-DDB1 ubiquitin ligase complex controls adult and embryonic stem cell differentiation..., Gao [/bib_ref] [bib_ref] Molecular architecture and assembly of the DDB1-CUL4A ubiquitin ligase machinery, Angers [/bib_ref].
To identify the substrates of CRL4B DCAF11 E3 ligase in osteosarcoma cells, we constructed the pCD-NA3-Flag-HA-DCAF11 vector, a two-epitope tagged vector designed to minimize unspecific binding during purification, which we transfected into hFOB1.19, U2OS and Saos-2 cells. Accordingly, the Flag-HA-DCAF11-associated complex was first immunoprecipitated with anti-Flag agarose, and the Flag epitope was then cleaved off. The resulting supernatant was then immunoprecipitated with anti-HA agarose, followed by elution using an HA peptide. After this two-step purification, HA-DCAF11-containing complexes were enriched and subjected to liquid chromatography tandem-mass spectrometry (LC-MS/MS) analysis. A total of 113 proteins were identified, and we listed the top ten candidates that demonstrated higher MASCOT scores in Supplementary Table 1. One putative DCAF11-interacting protein, p21, exhibited higher tandem MS read values and was previously reported to be a substrate of CRL4A E3 ligase [bib_ref] PCNA-dependent regulation of p21 ubiquitylation and degradation via the CRL4Cdt2 ubiquitin ligase..., Abbas [/bib_ref]. To verify the interaction between DCAF11 and p21, the pCDNA3-p21-Flag vector was constructed and transfected into U2OS cells, p21 was immunoprecipitated with anti-Flag agarose, and DCAF11 was detected using an immunoblot assay. The results revealed that DCAF11, but not DCAF4, was immunoprecipitated by p21 [fig_ref] Figure 3: p21 is a target of CRL4B DCAF11 E3 ligase in human osteosarcoma... [/fig_ref]. Moreover, we also observed a significant decrease in p21 protein level in U2OS and Saos-2 cell lysates compared with hFOB1.19 lysates [fig_ref] Figure 3: p21 is a target of CRL4B DCAF11 E3 ligase in human osteosarcoma... [/fig_ref]. To further verify the interaction between p21 and DCAF11, Y2H analyses were also performed using the following combinations of plasmids: pGADT7−DCAF11 + pGBKT7−p21, pGADT7−DCAF11 + pGBKT7, pGADT7 + pGBKT7−p21, pGADT7−DCAF4 + pGBKT7−p21, and pGADT7−DCAF4 + pGBKT7. Accordingly, plasmids were co-transformed into AH109 yeast cells, and the positive colonies were then selected and plated on synthetic dropout medium to determine the interactions between DCAF11 and p21. Our results indicated that DCAF11 interacted with p21, whereas DCAF4 did not [fig_ref] Figure 3: p21 is a target of CRL4B DCAF11 E3 ligase in human osteosarcoma... [/fig_ref]. These results suggested that p21 might be a substrate of the CRL4B DCAF11 E3 ligase.
The fact that components of the CRL4B DCAF11 complex were upregulated in osteosarcoma cells suggested that this E3 ligase might be more active and might cause the degradation of its specific substrates. As such, we would expect p21 protein levels to be down-regulated if it was indeed a substrate of CRL4B DCAF11 E3 ligase. To verify this hypothesis, we next examined p21 protein levels in different osteosarcoma cell lines. As shown in [fig_ref] Figure 3: p21 is a target of CRL4B DCAF11 E3 ligase in human osteosarcoma... [/fig_ref] , p21 protein levels were significantly lower in the osteosarcoma cell lines compared with hFOB1.19 cells. In addition, we also detected the expression of p27, a target of CRL4A E3 ligase, under the same conditions. No obvious changes were observed in p27 protein levels [fig_ref] Figure 3: p21 is a target of CRL4B DCAF11 E3 ligase in human osteosarcoma... [/fig_ref]. These results implied that CRL4B DCAF11 E3 ligase might directly target p21 for degradation in osteosarcoma cells. To determine whether CRL4B DCAF11 E3 ligase ubiquitinates p21 in vitro, we reconstituted the ubiquitination reaction using combinations of E1, E2, immunoprecipitated CUL4B-Flag, and purified recombinant His-p21 protein. p21 ubiquitination was detected by immunoblotting for p21. Incubating p21 with E1, E2 and CUL4B-Flag produced species with higher molecular weights than p21, which were absent in reactions lacking E2 or CUL4B-Flag [fig_ref] Figure 3: p21 is a target of CRL4B DCAF11 E3 ligase in human osteosarcoma... [/fig_ref] , indicating that CRL4B DCAF11 E3 ligase can ubiquitinate p21 in vitro.
K16, K154, K161 and K163 are required for p21 ubiquitination. The amino acid sequence of p21 contains six lysine (K) residues, K16, K75, K141, K154, K161 and K163 [fig_ref] Figure 4: CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 [/fig_ref]. To identify the ubiquitination site(s) of p21, we mutated each of these residues to arginine (R) and evaluated the ubiquitination of these mutants using an in vitro ubiquitination assay. As shown in [fig_ref] Figure 4: CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 [/fig_ref] , a slightly reduced ubiquitination pattern was observed for K16R, K154R, K161R and K163R in reactions with E1, E2 and CUL4B-Flag, but not for K75R and K141R. Furthermore, we created different p21 mutations, including p21 K154R K161R K163R , p21 K16R K161R K163R , p21 K16R K154R Scientific RepoRts | 7: 1175 | DOI:10.1038/s41598-017-01344-9 K163R , p21 K16R K154R K161R and p21 K16R K154R K161R K163R , to analyse ubiquitination patterns. As shown in [fig_ref] Figure 4: CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 [/fig_ref] , only faint ubiquitination bands were observed for p21 K154R K161R K163R , p21 K16R K161R K163R , p21 K16R K154R K163R and p21 K16R K154R K161R in reactions with E1, E2 and CUL4B-Flag. Importantly, ubiquitination was completely abolished in the p21 K16R K154R K161R K163R mutant [fig_ref] Figure 4: CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 [/fig_ref]. Based on the p21 ubiquitination pattern, our results indicated that CRL4B DCAF11 E3 ligase catalyses the multi-monoubiquitination rather than the poly-ubiquitination of p21.
To determine whether CRL4B DCAF11 E3 ligase was able to ubiquitinate p21 in vivo, we first examined p21 ubiquitination with and without suppression of CUL4B by shRNA. As shown in [fig_ref] Figure 4: CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 [/fig_ref] , inhibiting CUL4B expression significantly reduced p21 ubiquitination, suggesting that CRL4B DCAF11 E3 ligase contributed to p21 ubiquitination in vivo. To examine whether K16, K154, K161 and K163 were also important for p21 ubiquitination in vivo, we first constructed p21 K75R , p21 K141R , p21 K154R K161R K163R (K16, in which only K16 was present), p21 K16R K161R K163R (K154, in which only K154 was present), p21 K16R K154R K163R (K161, in which only K161 was present), p21 K16R K154R K161R (K163, in which only K163 was present), and p21 K16R K154R K161R K163R (4M, in which all four sites were mutated) in the pCDNA3-Flag vector and co-transfected them with HA-Ub. Consistent with the in vitro results, the in vivo ubiquitination analysis also indicated that K16, K154, K161 and K163 were required for p21 ubiquitination, whereas K75 and K141 were not [fig_ref] Figure 4: CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 [/fig_ref]. Ubiquitination was almost completely absent in the p21 K16R K154R K161R K163R mutant [fig_ref] Figure 4: CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 [/fig_ref].
Mutating K16R, K154R, K161R and K163R enhances p21 stability and decreases osteosarcoma cell proliferation. The in vitro and in vivo results revealed that K16, K154, K161 and K163 were required for p21 ubiquitination. To determine whether mutating these four sites affected cell proliferation, we first constructed p21 K16R , p21 K75R , p21 K141R , p21 K154R , p21 K161R , p21 K163R , p21 K154R K161R K163R , p21 K16R K161R K163R , p21 K16R K154R K163R , p21 K16R K154R K161R and p21 K16R K154R K161R K163R in the pCDNA3-Flag vector and then transfected each individual plasmid into U2OS cells. After 48 h, the levels of these mutated proteins were determined, and the results showed that the p21 K16R -Flag, p21 K54R -Flag, p21 K161R -Flag and p21 K163R -Flag proteins were much more stable than the p21 K75R -Flag and p21 K141R -Flag proteins [fig_ref] Figure 5: p21 stability affects osteosarcoma cell proliferation [/fig_ref] , which further suggested that mutating K16R, K154R, K161R or K163R might attenuate ubiquitination levels, thereby causing the accumulation of these proteins. Next, cell proliferation rates in U2OS cells expressing these single-mutated proteins were also determined using an MTT kit. Interestingly, cells harbouring K16R, K154R, K161R or K163R mutations showed slight reductions in cell proliferation compared with cells transfected with p21 WT -Flag, p21 K75R -Flag or p21 K141R -Flag [fig_ref] Figure 5: p21 stability affects osteosarcoma cell proliferation [/fig_ref]. In addition, The purified His-p21 protein was incubated with E1, with or without E2, and with or without CUL4B-Flag in reaction buffer, followed by immunoblotting for p21.
Scientific RepoRts | 7: 1175 | DOI:10.1038/s41598-017-01344-9
we also determined the stability of other triple and quadruple mutants, and p21 K154R K161R K163R -Flag, p21 K16R K161R K163R -Flag, p21 K16R K154R K163R -Flag, p21 K16R K154R K161R -Flag and p21 K16R K154R K161R K163R -Flag were much more stable than p21 WT [fig_ref] Figure 4: CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 [/fig_ref]. Of these triple and quadruple mutants, the quadruple mutant showed the highest abundance [fig_ref] Figure 4: CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 [/fig_ref]. Consistent with these results, cells transfected with these mutants exhibited significantly reduced proliferation rates compared with cells transfected with p21 WT [fig_ref] Figure 4: CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 [/fig_ref] , suggesting the importance of K16, K154, K161 and K163 in maintaining cell proliferation.
Because mutating these four critical sites increased p21 stability, we next sought to determine whether these mutants had longer half-lives. Accordingly, we transfected p21 WT -Flag, p21 K16R -Flag, p21 K75R -Flag, and p21 K16R K154R K161R K163R -Flag into U2OS cells. To ensure that the cells expressed similar levels of Flag-tagged p21, we treated U2OS cells immediately after transfection with the proteasome inhibitor MG132 and by changing the medium every 12 h. After 48 h, cells were incubated in fresh medium without MG132, and samples were collected every 20 min. Then, Flag-tagged p21 were detected by immunoblotting, which indicated that the half-lives of p21 WT -Flag and p21 K75R -Flag were approximately 80 min [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref] , whereas the half-life of p21 K16R -Flag was nearly 100 min (Supplementary [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref]. Interestingly, p21 K16R K154R K161R K163R -Flag was very stable, and we did not observe obvious degradation even at the 100-min time point [fig_ref] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in... [/fig_ref]. These results further suggested that the K16R, K154R, K161R and K163R sites are important for p21 ubiquitination.
Knocking down components of the CRL4B DCAF11 complex inhibits osteosarcoma cell proliferation, induces S phase cell cycle arrest, and decreases colony formation rates. To investigate whether the components of the CRL4B DCAF11 complex, including CUL4B, DDB1 and DCAF11, were able to regulate p21 protein levels, we used shRNAs to knock down the expression of these three genes. As shown in [fig_ref] Figure 6: Knocking down CRL4B DCAF11 complex components inhibits osteosarcoma cell growth [/fig_ref] , shRNA-transfected cells exhibited significant reductions in the levels of the target proteins and obvious increases in p21 levels. Given that p21 is a well-known inhibitor of CDK2, a kinase required for the transition from G1 to S phase, we next sought to determine whether CDK2 protein levels were down-regulated following the induction of p21 in cells with CUL4B, DDB1 or DCAF11 knock down. As expected, we observed a significant reduction in CDK2 [fig_ref] Figure 6: Knocking down CRL4B DCAF11 complex components inhibits osteosarcoma cell growth [/fig_ref]. In addition, cell proliferation assays were also performed to evaluate the down-regulation of components of the CRL4B DCAF11 E3 ligase, and as shown in [fig_ref] Figure 6: Knocking down CRL4B DCAF11 complex components inhibits osteosarcoma cell growth [/fig_ref] , knocking down CUL4B, DDB1 or DCAF11 inhibited cell proliferation.
Given that the down-regulation of CRL4B DCAF11 components led to the accumulation of p21 and to the reduction of CDK2, two key regulators of cell cycle progression, we proposed that the down-regulation of these three members should affect cell cycle progression. To verify this hypothesis, we subjected cells with reduced CUL4B, DDB1 or DCAF11 expression to flow cytometry analyses. As expected, osteosarcoma cells expressing lower levels of CUL4B, DDB1 or DCAF11 exhibited dramatically higher percentages of cells in S phase [fig_ref] Figure 6: Knocking down CRL4B DCAF11 complex components inhibits osteosarcoma cell growth [/fig_ref]. These results suggested that S phase cell cycle arrest might contribute to the cell proliferation inhibition observed upon the down-regulation of CRL4B DCAF11 E3 ligase components. In addition, colony formation rates significantly decreased after the down-regulation of CRL4B DCAF11 E3 ligase components [fig_ref] Figure 6: Knocking down CRL4B DCAF11 complex components inhibits osteosarcoma cell growth [/fig_ref].
# Discussion
Accumulating evidence suggests critical roles for Cullin family proteins and their associated CRL E3 ligases in tumourigenesis [bib_ref] Cullin family proteins and tumorigenesis: genetic association and molecular mechanisms, Chen [/bib_ref] [bib_ref] PCNA-dependent regulation of p21 ubiquitylation and degradation via the CRL4Cdt2 ubiquitin ligase..., Abbas [/bib_ref] [bib_ref] Targeting Cullin-RING E3 ubiquitin ligases for drug discovery: structure, assembly and small-molecule..., Bulatov [/bib_ref] [bib_ref] RING-type E3 ligases: master manipulators of E2 ubiquitinconjugating enzymes and ubiquitination, Metzger [/bib_ref] [bib_ref] Arabidopsis cockayne syndrome A-like proteins 1A and 1B form a complex with..., Zhang [/bib_ref] [bib_ref] EZH2 generates a methyl degron that is recognized by the DCAF1/DDB1/CUL4 E3..., Lee [/bib_ref] [bib_ref] The CUL4-DDB1 ubiquitin ligase complex controls adult and embryonic stem cell differentiation..., Gao [/bib_ref] [bib_ref] Molecular architecture and assembly of the DDB1-CUL4A ubiquitin ligase machinery, Angers [/bib_ref]. In the present study, we discovered that CUL4B interacted with DDB1, RBX1 and DCAF11 to form an E3 ligase known as CRL4B DCAF11 . We also determined that CRL4B DCAF11 directly targeted p21 for multi-monoubiquitination at K16, K154, K161 and K163 through in vitro and in vivo studies. Moreover, knocking down any component of the CRL4B DCAF11 complex by shRNA resulted in p21 accumulation, cell growth inhibition, S phase cell cycle arrest, and reduced colony formation rates. Our results provide evidence for a potential new DCAF11 pathway in which DCAF11 associates with CUL4B and DDB1 to target p21 for proteolysis in human osteosarcoma cells [fig_ref] Figure 7: Schematic diagram of the CRL4B DCAF11 ubiquitin complex in human osteosarcoma cells [/fig_ref].
CUL4A and CUL4B, two paralogs in mammalian cells, share high amino acid sequence similarity and have partially redundant functions in many biological processes, such as nucleotide excision repair (NER) [bib_ref] CUL4A abrogation augments DNA damage response and protection against skin carcinogenesis, Liu [/bib_ref] [bib_ref] The cullin 4B-based UV-damaged DNA-binding protein ligase binds to UV-damaged chromatin and..., Guerrero-Santoro [/bib_ref] [bib_ref] Ubiquitin proteasome system (UPS): what can chromatin do for you?, O'connell [/bib_ref]. Our previous studies have indicated that CUL4B is upregulated in human osteosarcoma cells [bib_ref] CUL4B promotes proliferation and inhibits apoptosis of human osteosarcoma cells, Chen [/bib_ref]. However, we did not examine the expression of other Cullin members to determine if only CUL4B was specifically upregulated. As such, in the present study, we first detected the expression of all seven Cullin members in different human osteosarcoma cell lines. CUL4B mRNA and protein were both specifically overexpressed in these human osteosarcoma cell lines, although CUL4A also exhibited a slight induction under the same conditions. Thus, CUL4B was considered to be specifically overexpressed in human osteosarcoma cells, possibly through an unknown transcriptional mechanism, which also indicated that CUL4B has a unique role in human osteosarcoma cells. serves as a scaffold that binds to adaptor protein DDB1 and the substrate receptor protein DCAF11 at the N terminus and interacts with RING protein RBX1 at the C terminus, thereby forming the CRL4B E3 ligase. The CRL4B E3 ligase can promote ubiquitin transfer from RBX1-bound E2 to the specific substrate p21 in osteosarcoma cells, leading to p21 ubiquitination. The degradation of p21 diminishes the inhibition of CDK2, thereby resulting in cell cycle arrest at S phase.
Scientific RepoRts | 7: 1175 | DOI:10.1038/s41598-017-01344-9
In addition, CUL4B has been demonstrated to interact with DDB1 and RBX1 in many studies [bib_ref] Cullin family proteins and tumorigenesis: genetic association and molecular mechanisms, Chen [/bib_ref] [bib_ref] DCAFs, the missing link of the CUL4-DDB1 ubiquitin ligase, Lee [/bib_ref]. Similarly, we also examined and confirmed these interactions through in vivo Co-IP and in vitro Y2H studies. The RBX1-CUL4B-DDB1 complex interacts with a number of substrate receptors known as DCAFs (e.g., DDB2, CSA, COP1 and DCAF1-19) to recognize thousands of substrates [bib_ref] Cullin family proteins and tumorigenesis: genetic association and molecular mechanisms, Chen [/bib_ref] [bib_ref] DCAFs, the missing link of the CUL4-DDB1 ubiquitin ligase, Lee [/bib_ref]. One limitation of our studies is that we did not screen for DCAFs that associate with CUL4B in osteosarcoma cells; instead, we only examined the expression of DCAF1, 4, 8, 11 and 15, and not the expression of other DCAFs. Fortunately, we found that DCAF11 had a similar expression pattern to CUL4B and was upregulated in different osteosarcoma cells. To determine whether DCAF11 was the only DCAF protein that interacted with DDB1 in osteosarcoma cells, it is necessary to utilize DDB1 as the bait to screen associating proteins and then select candidate DCAFs to examine expression in osteosarcoma cells.
We screened proteins that interacted with DCAF11 and identified a target of CRL4B DCAF11 E3 ligase known as p21. p21 protein levels were significantly decreased in osteosarcoma cells. Surprisingly, although p21 and p27 function in the regulation of cell cycle progression, we observed no changes in p27 levels in osteosarcoma cells. A number of studies have shown that p21 can be ubiquitinated by several E3 ligases, such as CRL4 Cdt2 and SCF . Human CRL4 Cdt2 targets and ubiquitinates p21 during S phase in a proliferating cell nuclear antigen (PCNA)-dependent manner [bib_ref] PCNA-dependent regulation of p21 ubiquitylation and degradation via the CRL4Cdt2 ubiquitin ligase..., Abbas [/bib_ref]. Inactivation of CRL4 Cdt2 E3 ligase prevents the nuclear export of CDC6 during S phase, and this block of nuclear export depends on p21 [bib_ref] The CRL4Cdt2 ubiquitin ligase targets the degradation of p21Cip1 to control replication..., Kim [/bib_ref]. SCF Skp2 E3 ligase participates in the degradation of p21 in S phase. Knocking out Skp2 in mouse embryo fibroblasts results in higher p21 levels than in wild-type fibroblasts in S phase because of the deficient p21 ubiquitination [bib_ref] Role of the SCFSkp2 ubiquitin ligase in the degradation of p21Cip1 in..., Bornstein [/bib_ref]. Our in vivo and in vitro ubiquitination studies indicated that CRL4B DCAF11 E3 ligase was able to ubiquitinate p21 at four different sites and failed to ubiquitinate the p21 K16R K154R K161R K163R mutant in vitro and in vivo. Interestingly, cells expressing p21 K16R , p21 K154R , p21 K161R , p21 K163R , p21 K154R K161R K163R , p21 K16R K161R K163R , p21 K16R K154R K163R , p21 K16R K154R K161R or p21 K16R K154R K161R K163R exhibited lower cell proliferation rates. Our results strongly indicate that p21 multi-monoubiquitination is important for maintaining p21 function. By comparing the ubiquitination pattern of p21 with previous publications, we found a major difference: some studies found that even when all lysines in p21 were mutated, p21 was still ubiquitinated at the N terminus [bib_ref] Proteasome-mediated degradation of p21 via N-terminal ubiquitinylation, Bloom [/bib_ref] [bib_ref] Terminal ubiquitination of extracellular signal-regulated kinase 3 and p21 directs their degradation..., Coulombe [/bib_ref] [bib_ref] UV Induces p21 rapid turnover independently of ubiquitin and Skp2, Lee [/bib_ref]. In our case, only K16, K154, K161 and K163 were required for p21 ubiquitination, and the quadruple mutant lacking all four of these sites could not be ubiquitinated. One possible explanation for these different ubiquitination patterns is that p21 can be modified by different E3 ligases in different cell types and that different E3 ligases might recognize different lysines. In our study, we did not detect ubiquitination at the p21 N terminus because we detected no p21 ubiquitination after mutating K16, K154, K161 and K163. Thus, it was not necessary to further detect ubiquitination at the p21 N terminus. On the other hand, these different ubiquitination patterns provided a new view on the modification of p21. In addition, in cells depleted of CUL4B, DDB1 or DCAF11, p21 degradation was impaired, which resulted in cell cycle arrest at S phase, cell growth inhibition and a reduced colony formation rate. This provides a strategy to target different components of this E3 ligase in osteosarcoma therapy.
In our LC-MS/MS results , we found that CDK2 was pulled down by CUL4B. We also observed a significant decrease in CDK2 levels after down-regulation of CUL4B, DDB1 or DCAF11 in osteosarcoma cells [fig_ref] Figure 6: Knocking down CRL4B DCAF11 complex components inhibits osteosarcoma cell growth [/fig_ref] , which raised the question of whether DCAF11 is responsible for the degradation of CDK2. To answer this question, we performed in vivo Co-IP and in vitro Y2H assays to determine whether DCAF11 interacts with CDK2. Unfortunately, we did not detect an interaction between these two proteins (Supplementary [fig_ref] Figure 3: p21 is a target of CRL4B DCAF11 E3 ligase in human osteosarcoma... [/fig_ref]. Thus, CDK2 might be only a CUL4B-associated protein, and its down-regulation in osteosarcoma cells expressing lower CUL4B, DDB1 or DCAF11 might not be regulated by DCAF11 but might rather be regulated by p21 induction, as p21 is an inhibitor of CDK2.
In summary, our results demonstrate that CUL4B forms an E3 ligase with RBX1, DDB1 and DCAF11 to recognize p21 as a substrate in human osteosarcoma cells. Our study offers insights into how CUL4B specifically functions in osteosarcoma cells and demonstrates that osteosarcoma cells have a unique CRL4B DCAF11 E3 ligase system that functions in the regulation of cell cycle progression.
# Materials and methods
Cell lines and growth conditions. The hFOB1.19 osteoblast cell line and the four osteosarcoma cell lines U2OS, MG63, HOS and Saos-2 were purchased from the American Type Culture Collection (ATCC, USA) and maintained in Dulbecco's modified Eagle's medium (DMEM) (Gibco, USA) supplemented with 10% heat-inactivated foetal bovine serum (FBS) (Gibco, USA) and penicillin-streptomycin (100 U/ml each). All cell lines were cultured in an incubator at 37 °C with 95% air and 5% CO 2 . Cells were split twice per week when they were approximately 80% confluent. Real-time PCR analysis. Total RNA was isolated using TRIzol reagent (Invitrogen, USA), and cDNA was synthesized using the SuperScript First-Strand Kit (Invitrogen, USA) after the removal of genomic DNA using RNase-free DNase I (Takara, Japan). The resulting cDNAs were applied to quantitative RT-PCR analysis using the specific primers listed in Supplementary Table 2. The PCR procedure was carried out as follows: 95 °C for 1 min, followed by 45 cycles of 95 °C for 10 sec and 68 °C for 30 sec, and then 4 °C for 10 min. All experiments were performed in triplicate, and gene expression was assessed using the 2 −ΔCt method by normalizing to β-Actin, the internal control, as previously described Immunoprecipitation and mass spectrometry analysis. Cells transfected with pCD-NA3-Flag-HA-DCAF11 or pCDNA3-Flag-HA were lysed with Pierce IP lysis buffer (Thermo Fisher Scientific, USA) according to the manufacturer's instructions and supplemented with 1x protease inhibitors (Roche, USA). Lysates were subjected to IP analysis using anti-Flag-agarose (Sigma, USA). Briefly, lysates were incubated with beads for 4 h at 4 °C. Then, beads were washed five times with lysis buffer, and proteins were eluted with Flag peptide (Sigma, USA) for 2 h at room temperature. HA-tagged DCAF11 protein was visualized on an SDS gel and stained with Coomassie Brilliant Blue R 250 as previously described [bib_ref] Inhibition of the Arabidopsis salt overly sensitive pathway by 14-3-3 proteins, Zhou [/bib_ref]. The stained protein bands corresponding to the DCAF11 protein were digested with trypsin and analysed by LC-MS/MS. The MS/MS spectra were searched against the NCBI database using a local MASCOT search engine (V.2.3).
Yeast two-hybrid assay. To determine the interactions between CUL4B and DDB1, CUL4B and RBX1, DCAF11 and p21, and DCAF4 and p21 in yeast cells, the coding sequence of either CUL4B or p21 was cloned into the pGBKT7 vector between the NdeI and the BamHI sites. The coding sequence of DDB1, RBX1, DCAF4 or DCAF11 was cloned into the pGADT7 vector between the EcoRI and the BamHI sites. The following combinations of plasmids, including pGBKT7−CUL4B + pGADT7−DDB1, pGBKT7−CUL4B + pGADT7, pGADT7-DDB1 + pGBKT7, pGBKT7−CUL4B + pGADT7−RBX1, pGADT7−RBX1 + pGBKT7, pGBKT7− p21 + pGADT7−DCAF11, pGBKT7−DCAF11 + pGBKT7, pGBKT7−p21 + pGADT7, pGBKT7− p21 + pGADT7−DCAF4, and pGBKT7−DCAF4 + pGBKT7, were co-transformed into AH109 yeast cells. The transformed yeast cells were selected on synthetic complete medium without Trp and Leu (SC-T/L). Interactions were determined on synthetic complete medium lacking Trp, Leu, and His (SC-T/L/H).
Plasmid and shRNA transfection. For plasmid transfection, cells were transfected using Lipofectamine 2000 (Invitrogen, USA) in suspension with 100 ng of plasmid. Cells were immediately transferred to 0.5 ml of DMEM medium with FBS in 24-well plates and incubated at 37 °C for 48 h.
Specific knockdown of CUL4B, DDB1 and DCAF1 expression in U2OS cells was carried out with a lentiviral system. The shRNAs, including shCUL4B (TRCN0000006532), shDDB1 (TRCN0000082856), and shDCAF11 (TRCN0000138902), were purchased from Sigma (USA). Briefly, the lentiviruses containing either negative control shRNA or shCUL4B, shDDB1 and shDCAF11 were transfected into U2OS cells following standard procedures. After transfection for 24 h, the virus-infected cells were selected with puromycin (1 μg/ml) for 48 h and then used for the subsequent experiments.
Cell cycle analysis by flow cytometry. Cells that were approximately 80% confluent were washed twice with ice-cold 1x PBS and then treated with 0.25% trypsin-EDTA. The cell suspension was fixed with 70% ethanol at 4 °C for 24 h. Cells were then washed with 1% bovine serum albumin (BSA) in 1x PBS and then stained for total DNA content with a solution containing 50 μg/mL propidium iodide (PI), 50 μg/mL RNase, 0.1% Triton X-100, and 0.1 mM EDTA in PBS for 30 min at 37 °C. Cell cycle distribution was analysed by flow cytometry (BD Biosciences, USA) and analysed with CellQuest software (BD Biosciences).
Cell proliferation and colony formation assays. Cell proliferation assays were performed as previously described [bib_ref] Heterotrimerization of the growth factor receptors erbB2, erbB3, and insulin-like growth factor-i..., Huang [/bib_ref]. Briefly, cells were transfected with different p21 plasmids, or shCUL4B, shDDB1 or shDCAF11 were plated onto 96-well plates for 0 h, 24 h, 48 h, 72 h, 96 h or 120 h. Then, a cell proliferation kit (Thermo Fisher Scientific, USA) was used to determine cell viability at 490 nm according to the manufacturer's instructions.
For colony formation assays, cells at approximately 80% confluence were diluted and seeded at approximately 1000 cells per well. After a 12-h incubation in a 37 °C humidified atmosphere containing 5% CO 2 , cells were transfected with shCUL4B, shDDB1 or shDCAF11 and then continuously cultured with DMEM medium for 14 days; the medium was changed every 3 days. Cell colonies were stained for 15 min with a solution containing 0.5% crystal violet and 25% methanol, followed by three rinses with ddH 2 O to remove excess dye. The colony numbers were counted by a gel documentation system (Bio-Rad Laboratories, Inc.).
## In vitro and in vivo ubiquitination analysis.
For the in vitro p21 ubiquitination assay, immunoprecipitated CUL4B-Flag was sequentially washed with lysis buffer (20 mM Tris-HCl, pH 8.0, 0.5% NP-40, 1 mM EDTA, and 150 mM NaCl), LiCl buffer (50 mM Tris-HCl, pH 8.0, and 0.5 M LiCl), and ubiquitination buffer (25 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , and 2 mM NaF) and then eluted with Flag peptide in ubiquitination buffer. The pET28a-His-p21 vector was transformed into Escherichia coli strain BL21 (DE3), and the recombinant His-tagged p21 protein was purified with Ni-NTA magnetic agarose beads (Clontech, USA) according to the manufacturer's protocol. Ubiquitination reactions were carried out with His-p21 (substrate), CRL4B complex, 0.1 μM E1 (Sigma, USA), 0.4 μM E2 (Sigma, USA), and 10 μg of bovine ubiquitin (Sigma, USA) in ubiquitination buffer supplemented with 2 mM ATP and 0.6 mM DTT at 37 °C for 75 min, and reactions were terminated by boiling in 2x SDS-PAGE loading buffer. Ubiquitination was analysed by immunoblotting with the anti-p21 antibody.
For the in vivo ubiquitination assay, Flag-tagged p21 or p21 mutants were co-expressed in U2OS cells with HA-ubiquitin for 48 h. Lysates were prepared by boiling cells with buffer containing 2% SDS, 150 mM NaCl, and 10 mM Tris-HCl, pH 8.0, and then diluted 10-fold with buffer composed of 10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% NP-40, 1 mM DTT, 2 mM EDTA, and 1x protease inhibitor (cocktail). Cell lysates were incubated with Flag beads for 2 h at 4 °C with rotation, and beads were washed five times with buffer containing 10 mM Tris-HCl, pH 8.0, 1 M NaCl, 1 mM EDTA, and 1% NP-40. The resin was boiled in 2x SDS-PAGE loading buffer, and ubiquitination was analysed by immunoblotting with an anti-HA antibody.
# Statistical analysis.
All experiments were independently replicated at least three times. Statistical analyses were performed using two-sided Student's t tests. Significance was set at P < 0.05.
[fig] Figure 1: Human osteosarcoma cell lines have elevated CUL4B expression. (A) Cullin mRNA levels, including CUL1, CUL2, CUL3, CUL4A, CUL4B, CUL5 and CUL7, were determined by qRT-PCR. Expression was normalized to β-Actin in each cell line, and the resulting ratios in hFOB1.19 cells were arbitrarily defined as 1-fold. Representative data from three independent experiments are shown. *P < 0.05 and **P < 0.001. (B) WB analyses with specific antibodies against CUL1, CUL2, CUL3, CUL4A, CUL4B, CUL5, CUL7, or β-Actin. [/fig]
[fig] Figure 2: CUL4B interacted with DDB1, RXB1 and DCAF11 to form a complex in vivo and in vitro. (A) Flagtagged CUL4B interacted with RBX1 and DDB1 in vivo. [/fig]
[fig] Figure 3: p21 is a target of CRL4B DCAF11 E3 ligase in human osteosarcoma cells. (A) Flag-tagged p21 associated with DCAF11 in vivo. The pCDNA3-p21-Flag plasmid was transformed into hFOB1.19 (1), U2OS (2) or Saos-2 (3) cells. After 48 h incubation, Flag-tagged proteins were immunoprecipitated with either protein A agarose or anti-Flag-agarose. The cell lysate and pull-down products were analysed by immunoblotting with anti-p21, anti-DCAF1, anti-DCAF4, anti-DCAF11 or β-Actin antibodies. (B) p21 interacted with DCAF11 in yeast. The pGBKT7-p21 plasmid was co-transformed with pGADT7-DCAF11 or pGADT7-DCAF4 into the yeast strain AH109. Growth of the transformed yeast was assayed on media lacking Trp and Leu (top panel) or lacking Trp, Leu, and His (bottom panel). Columns in each panel represent serial 10-fold dilutions. (C) p21 was significantly down-regulated in human osteosarcoma cell lines. Cell lysates were applied to WB analyses with specific antibodies against p21, CUL4B, p27, or β-Actin. (D) p21 was ubiquitinated by CRL4B DCAF11 E3 ligase in vitro. [/fig]
[fig] Figure 4: CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163. (A) Possible ubiquitination sites of p21. The amino acid sequence of p21 contains six lysines. (B,C) CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 in vitro. Purified His-p21 and p21 mutants, including p21 K16R , p21 K75R , p21 K141R , p21 K154R , p21 K161R , p21 K163R , p21 K154R K161R K163R , p21 K16R K161R K163R , p21 K16R K154R K163R , p21 K16R K154R K161R and p21 K16R K154R K161R K163R, were incubated with E1, E2, and CUL4B-Flag in ubiquitination reaction buffer, followed by immunoblotting for p21. (D) CRL4B DCAF11 E3 ligase ubiquitinated p21 in vivo. U2OS cells were transfected with p21-Flag and HA-Ubiquitin, followed by retroviral infection of CUL4B shRNA. After 48 h, cells were lysed, immunoprecipitated with Flag antibody, and then probed with anti-HA antibody for ubiquitinated p21. Right, demonstrated knockdown of CUL4B by shRNA. (E) CRL4B DCAF11 E3 ligase ubiquitinated p21 at K16, K154, K161 and K163 in vivo. U2OS cells were transfected with HA-Ubiquitin and p21-Flag or mutations, including p21 K75R -Flag, p21 K141R -Flag, p21 K154R K161R K163R -Flag (K16), p21 K16R K161R K163R -Flag (K154), p21 K16R K154R K163R -Flag (K161), p21 K16R K154R K161R -Flag (K163) and p21 K16R K154R K161R K163R -Flag (4 M). After 48 h, cells were lysed, immunoprecipitated with Flag antibody, and then probed with anti-HA antibody for ubiquitinated p21. [/fig]
[fig] Figure 5: p21 stability affects osteosarcoma cell proliferation. (A-C) Mutations of ubiquitination sites resulted in p21 accumulation in vivo. U2OS cells were transfected with p21-Flag or p21 mutants, including p21 K16R -Flag, p21 K75R -Flag, p21 K141R -Flag, p21 K154R -Flag, p21 K61R -Flag, p21 K163R -Flag, p21 K154R K161R K163R -Flag, p21 K16R K161R K163R -Flag, p21 K16R K154R K163R -Flag, p21 K16R K154R K161R -Flag, and p21 K16R K154R K161R K163R -Flag. After 48 h incubation, cells were lysed and immunoprecipitated with anti-p21, anti-CUL4B or β-Actin antibody. (B-D) Mutations of ubiquitination sites decreased cell proliferation. Cells used in (A and C) were applied to MTT assays to evaluate cell proliferation, and cell viability was determined at 490 nm. [/fig]
[fig] Figure 6: Knocking down CRL4B DCAF11 complex components inhibits osteosarcoma cell growth. (A) Knocking down CRL4B DCAF11 complex components leads to the accumulation of p21. U2OS cells transfected with Con-shRNA, CUL4B-shRNA, DDB1-shRNA, or DCAF11-shRNA. After 48 h incubation, cells were lysed and probed with anti-CUL4B, DDB1, DCAF11, p21, CDK2, or β-Actin antibody. (B) Knocking down CRL4B DCAF11 complex components inhibits osteosarcoma cell proliferation. Cells used in A were applied to the MTT assay to evaluate cell proliferation, and cell viability was determined at 490 nm. (C) Knocking down CRL4B DCAF11 complex components induces cell cycle arrest at S phase. Cells used in A were subjected to flow cytometry analysis of cell cycle distribution. (D,E) Knocking down CRL4B DCAF11 complex components decreased colony formation rates. Cells (1 × 10 3 ) used in A were seeded onto 12-well plates, and cells were cultured with 0.1 ml of fresh medium containing 0.5% FBS for two weeks. Then, cells were stained with 0.5% crystal violet (D), and colony numbers were counted by a gel documentation system (E). [/fig]
[fig] Figure 7: Schematic diagram of the CRL4B DCAF11 ubiquitin complex in human osteosarcoma cells. CUL4B [/fig]
|
Mechanisms of in vivo binding site selection of the hematopoietic master transcription factor PU.1
# Supplementary methods
## Mass spectrometry analysis of bisulfite-converted dna -genomic regions for massarray
analyses were chosen that either indicated an epigenetic transition from HPC to monocytes (induction of H3K4me1) or the induced binding of transcription factors in combination with de novo H3K4me1 appearance during monocyte to macrophage differentiation. Primer design, sodium bisulfite conversion, amplification and MALDI-TOF mass spectrometry (MassARRAY Compact MALDI-TOF, Sequenom, San Diego, CA) were done as described. Methylation was quantified from mass spectra using the Epityper software (Sequenom, San Diego, CA). The following primers were used to generate amplicons from bisulfite treated DNA:
[formula] Primer name Sequence 5'-3' ADAP1_F aggaagagagGGTGGAGAGGGAAGTGAATTTTTATAATT ADAP1_R cagtaatacgactcactatagggagaaggctTCCCTAATCCAAAACCAAAAACAAC SMAD7_F aggaagagagGAGGTTGGAGGTTATGAAGAGGTTT SMAD7_R cagtaatacgactcactatagggagaaggctTACTAACCACCCAATACAAACCCAC RAB4A_F aggaagagagGGGAGATTTTAGAGGGTTTTTAGGATT RAB4A_R cagtaatacgactcactatagggagaaggctAAAATAAACCTCCCCATTCCACTCT NCRNA00111_F aggaagagagAAAATTTGTAGGGTTGATTTGAAGT NCRNA00111_R cagtaatacgactcactatagggagaaggctATAAATCACAACAACACACTTTCCC PIM3_F aggaagagagGTTTGGTAGGTAGGGGTTTTGTTTT PIM3_R cagtaatacgactcactatagggagaaggctCCAAACTAAATCCTTCTCAACTCCTC KCNE2_F aggaagagagTGTGTGGTATTTTATTATGGTAGTTTTAGT KCNE2_R cagtaatacgactcactatagggagaaggctCACCCAAATAATCAATCCTCAAACC VRK2_F aggaagagagTTTTGGGAATTATGAGTGATAGATTTAGAA VRK2_R cagtaatacgactcactatagggagaaggctTCAAAATAAATTATACTTTCCTCCAATTT TSSC1_F aggaagagagTTTTTATGAGTGGGAGGTGTGAGTT TSSC1_R cagtaatacgactcactatagggagaaggctCCTCAAAATCCTACAAACTACCCAAAC RSRC1_F aggaagagagGGTTTTTTGTTATGTGAGAAAGATTTGTAG RSRC1_R cagtaatacgactcactatagggagaaggctCACACAACATTTCAAAACAACTCTC CACNA1C_F aggaagagagATGGAAGTTGAGAATTGAATTAATGAATGT CACNA1C_R cagtaatacgactcactatagggagaaggctCAAAACCAAAAAACATCCACAACAA CORO1C_F aggaagagagTTTTAGTGATAAGGGTTGGGTGTTG CORO1C_R cagtaatacgactcactatagggagaaggctCCCAATAAAAATTTTCCCAAAAAAAA PLEKHA6_F aggaagagagGGGTTTTTTTTGGGAGTTTTTTTT PLEKHA6_R cagtaatacgactcactatagggagaaggctACAATTAAACCATTAAACACCACAAC OXSM_F aggaagagagTTATTTTTAGGAATGAAAGGGGAAGG OXSM_R cagtaatacgactcactatagggagaaggctAAACAAACAACTCTTTCTCCAAATAAAACT ST18_F aggaagagagTTTTTTTGGATTTTGTTTAGGTAAG ST18_R cagtaatacgactcactatagggagaaggctAAACATTTCCCTACATCTTTATTTCAC OSR1_F aggaagagagTTTTGGTTTTAATTTAATGTTGTTATGTGG OSR1_R cagtaatacgactcactatagggagaaggctAAAAAAACCACAAATCTTTTTACAATTCC [/formula]
Microscale Thermophoresis -The sequence of the full-length hPU.1 was amplified by PCR from pORF9-hSPI1 (InvivoGen San Diego, USA) and recombined into a modified pDM8 vector, encoding an N-terminal His-tag, using the Gateway technology (life technologies). The protein was expressed in Rosetta2(DE)pLysS (Novagen) and purified by Nickel affinity chromatography (Qiagen). Doublestranded DNA molecules were annealed from single-stranded, HPLC-purified oligonucleotides (Sigma-Aldrich). The annealing reaction (10 µl) was performed in 1x annealing buffer (20 mM Tris-HCl pH 7.4, 2 mM MgCl2, 50 mM NaCl) and comprised 20 µM of the Cy3-labeled oligonucleotide and 20.8 µM of the unlabeled oligonucleotide (lower strand). The annealing reaction was incubated for 15 min at 95°C in a thermoblock (peQLab) and afterwards allowed to slowly cool down to room Pham et al.
temperature over night. The annealing reaction was checked on an 8% native polyacrylamide gel which was analyzed on a fluorescence imager (FLA-5000, Fujifilm).
The binding assay was carried out using the Nanotemper Monolith NT.115 (initial settings: LED power:
90%, IR-laser power: 80%, 25°C). For each motif affinity measurement 16 reactions were prepared on ice (MST-buffer: 20 mM Tris-HCl pH 7.6, 1.5 mM MgCl2, 0,5 mM EGTA, 10% glycerol, 300 mM KCl, 10 mM DTT). The Cy3-labeled dsDNA oligo was always kept at a constant concentration of 50 nM.
The unlabeled protein was titrated in a 1:1-dilution series starting with a concentration of 23 µM. Every binding assay comprised one control reaction without any protein. After loading the binding reactions into standard capillaries (NT.115) the mixture was incubated for 15 min at 25 °C in the Nanotemper before starting the measurement. The data was analyzed using the NT-analysis acquisition software
# Figure methods
Examples of thermophoresis curves obtained for four representative motifs. The top motif represents the motif with the highest PWM log-odds score, the bottom motif (mutated) contains a three nucleotide exchange in the core recognition site of PU.1 and shows no detectable binding. Pham et al.
## Chip-seq peak finding and annotation -
Analysis of mapped ChIP-seq tags was performed using HOMER, which is freely available at http://biowhat.ucsd.edu/homer/. ChIP-Seq quality control and transcription factor peak finding, TSS annotation (based on GENCODE V13) and motif analysis were done essentially as described . Genome Ontology annotation and ChIP-seq tag annotation of peak sets, or motif-centered regions was done using scripts provided by HOMER. Next generation sequencing data (either published or generated in this study) that were used in this study are listed below:
(A) Published sequencing data used in this study De novo motif searches -Motif enrichment in transcription factor peak sets was done using HOMER by comparing sequences of cell type-specific peaks (+/-100 bp) to 50,000 randomly selected genomic fragments of the same size, matched for GC content and autonormalized to remove bias from lowerorder oligo sequences. Due to the numerous enrichment tests made during the motif discovery procedure and the vast search space, corrections for multiple hypothesis testing must be carried out empirically by randomizing the target and background assignments and repeating the motif discovery procedure. One hundred randomizations (which were performed for each individual motif search) failed to yield motifs with enrichment P-values less than 1e-19, implying the false discovery rate for motifs with a P-value less than 1e-19 reported in this study is < 1%. Motif enrichment around bound motifs (+/-100 bp) was done by comparing motif-centered regions with non-overlapping, GC-matched, and autonormalized regions centered on non-bound motifs. In either case, motif enrichment is calculated using the cumulative hypergeometric distribution by considering the total number of target and background sequence regions containing at least one instance of the motif. Reannotation of the PWM to the human genome (hg19, either total or repeat-masked) was done using the scanMotifGenomeWide.pl script contained in the HOMER suite. HPC, MO, and MAC PU.1 ChIPseq tags were counted around all motif instances (+/-100 bp) across the repeat-masked human genome to determine non-bound PU.1 motifs (no ChIP-seq tag within the 200-bp window) across the non-repetitive genome. To determine the total set of bound PU.1 motifs, all bound PU.1 motif instances from HPC, MO and MAC were merged using bedTools. Extraction of log-odds scores for individual motifs or peaks was done using the annotatePeaks.pl program (provided by HOMER), which returns the highest scoring motif position as well as log-odds scores for each peak/region. Sequences were extracted using homerTools (provided by HOMER). [fig_ref] Table S1: Microscale thermophoresis-derived K D values for selected PU [/fig_ref] Complete results of the microscale thermophoresis measurements for 75 selected sequences included in the PU.1 PWM.
## Known motif analyses -
## Analyses of published
## Summary of supplementary tables and figures
## Table s2
Characteristics of chromatin domain categories.
## Figure s1
Dynamics of PU.1 binding during HPC to monocyte differentiation.
## Figure s2
Genome Ontology enrichment analysis for bound and non-bound PU.1 motifs
## Figure s3
Distribution of epigenetic marks at bound and non-bound PU.1 consensus sequences.
## Figure s4
Relationship between transcription factor binding and DNA methylation.
## Figure s5
Motif composition around bound vs. non-bound PU.1 consensus sites or around PU.1 peaks not recognized by the consensus PU.1 motif.
## Figure s6
Distribution of motif log-odds scores at PU.1 bound regions for three cell stages and alternative PWM.
## Figure s7
Comparison of motif log-odds scores with signal intensity Z scores from protein binding microarray (PBM) experiments.
## Figure s8
Characterization of homotypic PU.1 motif clusters.
## Figure s9
Bound PU.1 motifs with differentiation-dependent DNase I accessibility changes.
## Figure s10
PU.1 motifs in gene deserts.
## Figure s11
Expression correlation in CTCF-flanked domains contingent of their H3K4me1 level.
## Figure s12
Motif distribution in CTCF-flanked domains contingent of their H3K4me1 level.
## Figure s13
Distribution of motif-associated PU.1 tag counts in CTCF-flanked domains contingent of their H3K4me1 level.
## Figure s14
Motif score distribution and DNAse I accessibility in CTCF-flanked domains contingent of their H3K4me1 level.
## Figure s15
Cell type-specific domain activities in MO, MAC and HPC.
## Figure s16
Motif analyses in domains with differential activity between MO and liver.
## Figure s17
Enrichment of PU.1 co-associated transcription factor consensus sites in domains showing cell type-specific activity
## Figure s18
Distribution of PU.1 motifs across domain categories. , the remaining data was generated by the ENCODE or the Roadmap Epigenomics projects (high-throughput sequencing data sets used in this study are listed in [fig_ref] Table S1: Microscale thermophoresis-derived K D values for selected PU [/fig_ref]. Heat maps depict the methylation status of individual CpGs from red (100%) over blue (50%) to yellow (0%) with each box representing a single CpG. Data of at least three independent donors were averaged. Motif log odds score ***
## Figure s6 distribution of motif log-odds scores at pu.1 bound regions for three cell stages and alternative pwm. (a-d)
Combined bean & box plots showing the distribution of motif log-odds scores of annotated PU.1 motifs (white boxes) or best scoring motifs within total (blue boxes) or cell type-specific peaks (light blue boxes). (A) corresponds to the motif de novo extracted from human HPC PU.1 peaks, (B) corresponds to a motif de novo extracted from mouse peritoneal macrophage PU.1 peaks. In (C,D), motifs de novo derived motifs from human macrophage and monocyte PU.1 peaks were used, which were generated using normalized motif frequencies to correct for the depletion of CpG containing motifs. Solid bars of boxes display the interquartile ranges (25-75%) with an intersection as the median; whiskers, max/min values. Significantly different motif score distributions in pairwise comparisons are indicated (*** P < 0.001, Mann-Whitney U test, two-sided). The log odds score representing the motif detection threshold is indicated by the horizontal dotted line. The motif logos are shown on top of each plot along with the fraction of PU.1 bound regions (200 bp) containing at least one motif instance, the expected frequency of the motif in random sequences (in parentheses) as well as P values (hypergeometric) for the overrepresentation
## Figure s7
Comparison of motif log-odds scores with signal intensity Z scores from protein binding microarray (PBM) experiments. Published PBM data for the DNA binding PU.1 ETS domain and all possible 8-mers was used to compare PBM signal intensity Z scores, which represent a measure of protein affinity, with motif log-odds scores.
To adjust for size differences between both measures, we focussed on the central 8-mers in our PWM (NNGGAANN). If several 12-mers of the original PWM overlapped a core PWM 8-mer, the highest log-odds score was assigned to it. The scatterplot shows a good agreement between both measures (coefficient of determination R 2 =0.59).
## Figure s13
Distribution of motif-associated PU.1 tag counts in CTCF-flanked domains contingent of their H3K4me1 level. The PU.1 PWM was mapped across the masked human genome and all recognized sites were binned contingent on their motif log-odds scores and their location in CTCF-flanked domains. Bean plots show the distribution of PU.1 ChIP-seq tag counts (TC) associated with motifs. High score motifs show the highest PU.1 tag counts and are almost always bound in domains with high activity. The dotted line represents a TC of 12, which was used to define binding events.
## Figure s17
Enrichment of PU.1 co-associated transcription factor consensus sites in domains showing cell type-specific activity. Hierarchical clustering (Pearson correlation uncentered, average linkage) of enrichment values for co-association of the indicated PU.1-bound consensus motifs (within +/-100-bp) in liver-or osteoblast (OB)-specific domains and MO-specific domains (compared to either liver or OB). P values for motif co-enrichment were calculated using the hypergeometric test relative to the distribution in the total repeat-masked set. Data are presented as a heatmap where red (blue) coloring indicates a significant enrichment (depletion) of motif co-occurrence. Numbers in boxes represent corresponding relative changes in motif co-enrichment.
[fig] 1. 2: 229), which plots a binding curve using the normalized fluorescence of the labeled dsDNA at different concentrations of the unlabeled protein. Each binding assay was performed twice and the mean value was calculated. For every single binding assay a maximum of three outlier values were eliminated. Four representative examples for thermophoresis curves are shown below. [/fig]
[fig] Figure S2, Figure S3: of PU.1 binding during HPC to monocyte differentiation. (A) PU.1 ChIP-Seq tag counts for peak regions are compared between human CD133+ hematopoietic progenitor cells (HPC) and monocytes (MO) in a hexbin density plot. The colors represent the relative density of peaks in each location within the density plot. (B) Genomic distribution of total and cell stage-specific (at least four-fold different) PU.1 bound regions relative to GENCODE Genes V13. (C) De novo identified sequence motifs associated with PU.1 peak regions in HPC or MO. The fraction of PU.1 bound regions (200 bp) containing at least one motif instance, the expected frequency of the motif in random sequences (in parentheses) as well as P values (hypergeometric) for the overrepresentation are given below each motif. At the bottom of each panel, the best matching known motif is given along with its Genome ontology enrichment analysis for bound and non-bound PU.1 motifs. The bar chart shows log 2 -ratios of observed versus expected annotation frequencies (based on RefSeq) across the human genome. Adjusted enrichment P-values (hypergeometric, log 10 ) for comparisons between bound and non-bound motifs are given on the right. Distribution of epigenetic marks at bound and non-bound PU.1 consensus sequences. Histograms for genomic distance distributions of the indicated epigenetic data sets centered across bound or unbound PU.1 consensus sites across a 4-kb (or 1-kb) genomic interval. The top panel includes data from(Pham et al., 2012) [/fig]
[fig] Figure S5: 3% (2.5%), P <10 -115 [PB0058.1_Sfpi1_1, 0.68] 2.0% (0.1%), P <10 -113 [MA0139.1_CTCF, 0.87] 3.5% (0.7%), P <10 -105 [MA0060.1_NFYA, 0.87] 1.3% (0.1%), P <10 -82 [MF0095.1_YY1, 0.73] 3.6% (0.9%), P <10 -85 [GM12878-Usf1-ChIP-Seq, 0.91] 2.3% (0.4%), P <10 -72[-] Motif composition around bound vs. non-bound PU.1 consensus sites or around PU.1 peaks not recognized by the consensus PU.1 motif. De novo derived sequence motifs associated with (A) bound peak regions (motif-centered) in HPC, MO and MAK compared to non-bound, motif-centered regions, (B) non-canonical PU.1 peak regions (not recognized by the consensus PWM) compared to non-bound, motif-centered regions and (C) non-bound, motif-centered regions compared to all bound peak regions (motif-centered) in HPC, MO and MAK. The fraction of regions (200 bp windows) containing at least one motif instance, the expected frequency of the motif in bckground sequences (in parentheses) as well as P values (hypergeometric) for the overrepresentation of each motif are provided below each motif. At the bottom of each panel, the best matching known motif is given along with its similarity score. In (C), only motifs are shown that show elevated levels of vertebrate conservation compared to neighboring sequences. [/fig]
[fig] Figure S8, Figure S9 BoundFigure: Characterization of homotypic PU.1 motif clusters. (A) Non-repetitive PU.1 motif occurences were segregated according to neibouring motif frequency into single PU.1 motifs (1 motif), pairs of two neighboring motifs (2 motifs), and clusters of three or more motifs (≥3 motifs) seperated by less than 100bp. (B) Bar chart showing the fractions of bound and non-bound motifs, motif pairs or motif clusters. Clusters and pairs are more frequently bound than single motifs (*** P < 10 -300 , hypergeometric test). (C) Combined bean & box plot showing the distribution of motif log-odds scores of single motifs (black), as well as the highest scoring motifs of pairs (blue) and clusters (red). Solid bars of boxes display the interquartile ranges (25-75%) with an intersection as the median; whiskers represent max/min values. Motif score distributions in pairwise comparisons are highly significant (*** P < 0.001, Mann-Whitney U test, two-sided). The detection threshold is indicated by the dotted line. (D) Heatmap illustrating the relative and absolute frequencies of motif log-odds score combinations in motif pairs. Motif 1 represents the motif with the higher score. Numbers in bold indicate the absolute number of bound pairs of a given combination. The corresponding number of non-bound pairs is given below. Asterixs indicate that the bound fraction is either significantly enriched or depleted relative to non-bound (*** P < 0.001, ** P < 0.01, hypergeometric test).Enrichment ratios are indicated by coloring. (E) Histogram for the distribution of DNase I cleavage sites in MO centered across single motifs (black), as well as the highest scoring motifs of pairs (blue) and clusters (red) across a 500 bp genomic interval. (F) Histograms showing average per-nucleotide vertebrate conservation (PhastCons & PhyloP) surrounding single motifs (black), as well as the highest scoring motifs of pairs (blue) and clusters (red). (G) Histograms for the distribution of PU.1 Chip-seq tag counts (TC) in HPC, MO and MAC centered across single motifs (black), as well as the highest scoring motifs of pairs (blue) and clusters (red) across a 1kb genomic interval. (H) Histograms showing the distribution of indicated consensus motifs around PU.1 motifs as a function of motif PU.1 motifs with differentiation-dependent DNase I accessibility changes. (A) Bean plots showing the distribution of DNase I cleavage frequency around HPC PU.1 bound (consensus site: colored filling, non-consensus sites: black filling) and non-bound motifs (gray filling) depending on motif score classes. DNase I cleavage events (at nucleotide resolution, tag counts normalized to 10 7 ) were counted in a 200-bp window around each motif. Horizontal bars mark the median of each distribution. DNase I cleavage data (representing four independent donors) were originally generated by the ENCODE or the Roadmap Epigenomics projects (for accession nos. see the Supplementary Methods).(B) Combined bean & box plot showing the distribution of motif log-odds scores of PU.1 motifs in that are bound and accessible in HPC (purple), bound and non-accessible in HPC (rose) or bound and non-accessible in HPC but accessible in MO (red). Solid bars of boxes display the interquartile ranges (25-75%) with an intersection as the median; whiskers, max/min values. Significantly different motif score distributions in pairwise comparisons are indicated (*** P < 0.001, Mann-Whitney U test, two-sided). (C) Histograms for genomic distance distributions of the indicated sequencing data sets centered across PU.1 consensus sites that are bound and non-accessible in HPC but accessible in MO across a 4-kb (or 1-kb) genomic interval. HPC data is in purple and MO data in dark red. (D) Histogram showing average per-nucleotide vertebrate conservation (PhastCons) surrounding PU.1 motifs in that are bound and accessible in HPC (purple line), bound and non-accessible in HPC (rose line) or bound and non-accessible in HPC but accessible in MO (red line)S10 PU.1 motifs in gene deserts. (A) Histogram showing the average per-nucleotide vertebrate conservation (PhastCons) surrounding PU.1 motifs in gene deserts or across the entire genome. (B) Bean plots showing the distribution of normalized PU.1 ChIP-seq tag counts across PU.1 motifs in motif score classes. (C) Combined bean & box plot showing the distribution of motif log-odds scores of non-bound or bound PU.1 motifs across the whole genome or in gene deserts only. Solid bars of boxes display the interquartile ranges (25-75%) with an intersection as the median; whiskers, max/min values. Significantly different motif score distributions in pairwise comparisons are indicated (*** P < 0.001, Mann-Whitney U test, two-sided [/fig]
[fig] Figure S11, Figure S12: Distributions of gene expression levels in CTCF-flanked domains contingent of their H3K4me1 level. Combined bean & box plot showing the distribution of mRNA expression levels (based on published microarray data) of domain-associated genes in MO (A), MAC (B) and CD4 T cells (C). Solid bars of boxes display the interquartile ranges (25-75%) with an intersection as the median; whiskers, max/min values. Motif distribution in CTCF-flanked domains contingent of their H3K4me1 level. Bar charts of motif frequencies across domain activity bins for MAC (A) and T cells (B). CTCF-flanked regions were binned according to their H3K4me1 levels in MAC or T cells. In both cases, binding events in MAC were counted (since T cells don't express PU.1). [/fig]
[fig] Figure S15: Cell type-specific domain activities in MO, MAC and HPC. Tag count per bp ratios for MO vs. HPC (A) or MO vs. MAC (B) are plotted against average tag counts for CTCF-flanked domains (MvA-plots). The correlation coefficients for direct comparisons of log-transformed tag counts (TC) per bp are given above each diagram. A B Figure S16 Motif analyses in domains with differential activity between MO and liver. (A) Tag count per bp ratios for MO vs. liver are plotted against average tag counts for CTCF-flanked domains (MvA-plot). The correlation coefficient for the direct comparison of tag counts (TC) per bp are given above the diagram. (B) Distribution of normalized PU.1 ChIP-seq tag counts around motifs contingent on motif score classes in domains showing cell type-specific activity. The horizontal bar indicates the median of each distribution. The dotted line indicates the tag threshold for peaks considered bound. (C) Combined bean & box plot showing the distribution of motif log-odds scores for all (total) motifs or motifs bound in MO in domains showing cell type-specific activity. Solid bars of boxes display the interquartile ranges (25-75%) with an intersection as the median; whiskers represent min/max. Coloring indicates the type of domains tested. Significantly different motif score distributions in pair-wise comparisons are indicated (*** P < 0.001, Mann-Whitney U test, two-sided). (D) Bar chart of total and bound motif frequencies in OB-or MO-specific domains. The additional boxed chart shows frequencies of bound motifs overlapping with DNase I accessible sites in MO. [/fig]
[fig] Figure S18: Distribution of PU.1 motifs across domain categories. PU.1 motifs were collected that had at least one of the indicated enriched (red coloring) or depleted motif (blue coloring) in close neighborhood (+/-100-bp). The combined bean bean & box plots indicate the frequency distributions (motifs/10-kb) of all PU.1 motifs (top panel) or PU.1 motifs that are associated with at least one of the enriched (middle panel) or depleted motifs (bottom panel) in the indicated domain categories. Gene deserts, inactive domains (no H3K4me1 in MO) and domains with specific activity in unrelated cell types (Liver-and OB-specific) are generally characterized by lower densities of PU.1 motifs and fewer PU.1 motifs with coassociated, enriched transcription factor motifs. [/fig]
[table] Table S1: Microscale thermophoresis-derived K D values for selected PU.1 motifs [/table]
[table] Table S2: Characteristics of chromatin domain categories [/table]
[table] 0 1 100: Normalized PU.1 TC (± 100-bp around motifs) Domain activity bins (average H3K4me1 tag count per bp in MO) Figure S14 Motif score distribution and DNAse I accessibility in CTCF-flanked domains contingent of their H3K4me1 level. (A) Bean plots showing the distribution of motif log-odds scores for MO PU.1 bound (red filling) and non-bound motifs (gray filling) in CTCF-flanked regions depending on domain activity. Horizontal bars represent medians of motif score distributions. (B) Frequency of bound and non-bound PU.1 motifs overlapping DNase I accessible sites in MO as a function of domain activity. [/table]
|
The ethics of elective psychopharmacology
Pharmacological cognitive enhancers (PCEs) are used to improve cognitive functions, such as attention, learning, memory and planning in patients with impairments in cognition resulting from traumatic brain injury (TBI) or from neuropsychiatric disorders such as Alzheimer's disease (AD), mild cognitive impairment, schizophrenia, and attention deficit hyperactivity disorder (ADHD). Moreover, PCEs have been shown to improve cognition in healthy volunteers with no psychiatric disorders. This article describes the rationale behind the need for their use in neuropsychiatric patients and illustrates how PCEs can ameliorate cognitive impairments, improve quality of life and wellbeing, and therefore reduce the economic burden associated with these disorders. We also describe evidence that PCEs are being used as cognitive enhancers by healthy people. Crucially, as the lifestyle use of these drugs becomes very popular in the healthy population, a final aim is to present an overview of the current and future neuroethical considerations of enhancing the healthy brain. As information regarding their actual use, benefits and harms in various healthy populations is currently lacking, we propose research that aims to obtain relevant empirical data, monitor the short-and long-term effectiveness and side-effects, and initiate accurate surveys to determine current patterns and quantity of usage of PCE drugs by healthy people. Furthermore, in order to instigate a dialogue between neuroethics and neuropsychopharmacology, we urge scientists to explore and communicate the social and ethical implications of their research to the public. Finally, we discuss and highlight other means of enhancing cognition in both patients and healthy adults, including education and physical exercise. This article is based on the plenary lecture at the CINP conference in Hong Kong (2010).The online version of this article is published within an Open Access environment subject to the conditions of the Creative Commons Attribution-NonCommercial-ShareAlike licence <http://creativecommons.org/licenses/by-nc-sa/2.5/>. The written permission of Cambridge University Press must be obtained for commercial re-use.
## Pharmacological cognitive enhancers (pces)
Cognitive impairment is a core deficit of a number of neuropsychiatric disorders [bib_ref] Cognitive impairment in patients with schizophrenia displaying preserved and compromised intellect, Weickert [/bib_ref] and drugs that improve facets of cognition such as attention, learning, memory and executive functions are known as PCEs [bib_ref] Foresight Mental Capital and Wellbeing Project. The Government Office for, Morein-Zamir [/bib_ref]. These drugs alter neurotransmitter modulation of cognition leading to improvements in cognitive deficits in patients with traumatic brain injury (TBI) [bib_ref] Off-label uses of Modafinil, Teitelman [/bib_ref] , depression, addiction [bib_ref] A double-blind, placebo-controlled trial of modafinil (200 mg/day) for methamphetamine dependence, Shearer [/bib_ref] , multiple sclerosis [bib_ref] Modafinil for fatigue in MS : a randomized placebo-controlled double-blind study, Kraft [/bib_ref] , Parkinson's disease [bib_ref] Treatment of excessive daytime sleepiness in patients with Parkinson's disease with modafinil, Nieves [/bib_ref] , and those suffering from Alzheimer's disease (AD), schizophrenia, and attention deficit hyperactivity disorder (ADHD). The goal for their use is to ameliorate impaired functional outcomes. There is substantial opportunity for meeting the challenge of improving cognition and mental wellbeing in those with mental health problems and for reducing substantially the factors that contribute to the loss of mental, social and economic capital [bib_ref] The mental wealth of nations, Beddington [/bib_ref].
A good illustration is AD, which is a neurodegenerative disorder characterized by a decline in cognitive and behavioural functioning. It is the commonest cause of dementia and one of the most disabling and burdensome health conditions worldwide [bib_ref] Global prevalence of dementia : a Delphi consensus study, Ferri [/bib_ref]. There are currently 820 000 people with dementia in the UK, which costs £23 billion per year . Globally, if no effective prevention strategies or neuroprotective medications are developed, 81.1 million people will suffer from dementia by 2040 [bib_ref] Global prevalence of dementia : a Delphi consensus study, Ferri [/bib_ref]. However, a treatment that would reduce severe cognitive impairment in older people by just 1 % a year would cancel out all estimated increases in the long-term care costs due to ageing population [bib_ref] Cognitive impairment in older people : future demand for long-term care services..., Colantonio [/bib_ref].
Cognitive enhancers hold significant benefits in ameliorating these cognitive impairments in AD and countering these economic and social burdens [bib_ref] Cholinesterase inhibitors for patients with Alzheimer's disease : systematic review of randomised..., Kaduszkiewicz [/bib_ref]. For example, cholinesterase inhibitors (ChEI), such as donepezil, that inhibit centrally active acetylcholinesterase (AchE) and boost acetylcholine in the brain, compensate for the degeneration of neurons in the neocortex that regulate attention and memory, and are effective in the treatment of mild and moderate AD .
Selective ChEIs release growth factors, interfere with amyloid deposition, or modulate nicotinic receptors [bib_ref] Cholinesterase inhibitors and beyond, Pepeu [/bib_ref] while future drugs may exert their beneficial effects by activating various neurotransmitters including noradrenaline (NA), dopamine (DA), serotonin (5-HT), GABA and glutamate [bib_ref] Fluoxetine protects against amyloid-beta toxicity, in part via daf-16 mediated cell signalling..., Aboukhatwa [/bib_ref]. Furthermore, drugs that activate synaptic NMDA receptors work synergistically with AMPA receptors to produce long-lasting changes in the synaptic functioning and enable the encoding of new memories. This promotes the phosphorylation of CREB that slows down the pathological changes observed in AD. However, these drugs only minimize the neural damage caused by glutamate's neurotoxic effects and evidence is needed of improved episodic memory in AD patients. Importantly, some drugs have significant side-effects. For example, donepezil is contraindicated for people with liver problems [bib_ref] Alzheimer's disease : progress or profit, Mount [/bib_ref] while others have modest clinical efficacy in early [bib_ref] Efficacy and safety of cholinesterase inhibitors in Alzheimer's Disease. A meta-analysis, Lanctot [/bib_ref] and advanced stages . Furthermore, about 30-40 % of patients with AD may not respond to ChEI, and approximately 29 % of patients treated with ChEI leave clinical trials because of adverse events [bib_ref] Cholinesterase inhibitors for Alzheimer's disease, Birks [/bib_ref]. Therefore, it is important to develop novel and effective neuroprotective agents that selectively target the underlying neuropathology associated with amnestic mild cognitive impairment and AD. Developing these drugs could be advantageous to the individual and to society, particularly given the significant ageing population in the UK and the USA. It is not within the scope of this review to cover in detail the range and action of all the current and novel PCEs and therefore the reader is referred toand The Academy of Medical Sciences (2008).
Notwithstanding, PCEs might also improve the quality of life in patients with TBI, which is the most common cause of disability in young people [bib_ref] Cognitive impairment in older people : future demand for long-term care services..., Colantonio [/bib_ref] [bib_ref] Emotional distress and quality of life in relatives of patients with severe..., Norup [/bib_ref]. For example, survivors of TBI often suffer from chronic cognitive deficits [bib_ref] Changes over time in cognitive and structural profiles of head injury survivors, Salmond [/bib_ref] in areas such as sustained attention and learning which implicate impaired cholinergic function [bib_ref] ERPs and behavioural indices of long-term preattentive and attentive deficits after closed..., Polo [/bib_ref] [bib_ref] Cognitive outcome in traumatic brain injury survivors, Salmond [/bib_ref]. Furthermore, voxel-based morphometry studies reveal structurally reduced greymatter density and changes in hippocampus and neocortex in TBI patients. This is further consistent with the cholinergic dysfunction account that commonly contributes to the development of TBI-induced cognitive impairments . Consequently, the use of ChEIs that increase cholinergic function may be of benefit to TBI patients [bib_ref] A randomized controlled trial of rivastigmine for chronic sequels of traumatic brain..., Tenovuo [/bib_ref]. However, whether PCEs improve apathy, which is often considerably disabling and detrimental to rehabilitative efforts in TBI patients, needs to be determined in future research [bib_ref] Effects of methylphenidate on cognition and apathy in normal pressure hydrocephalus :..., Keenan [/bib_ref] [bib_ref] Modafinil therapy for apathy in an elderly patient, Padala [/bib_ref].
Equally, patients with schizophrenia can also benefit from PCEs through improvements in executive functions . Schizophrenia is a complex, lifelong disorder that significantly impairs cognitive and motivational function in approximately 1 % of the world's population. Although psychotic symptoms, such as hallucinations and delusions, can be managed with antipsychotic treatment [bib_ref] Randomized, double-blind, placebo-controlled study of paliperidone extended-release and quetiapine in inpatients with..., Canuso [/bib_ref] , patients continue to suffer dysfunctions in cognition, affect and motivation, which account for substantial decrements in social and occupational functioning [bib_ref] Measuring changes in functional status among patients with schizophrenia : the link..., Matza [/bib_ref] [bib_ref] Use of environmental supports among patients with schizophrenia, Velligan [/bib_ref]. Specifically, evidence indicates that these patients are substantially impaired in a wide range of neuropsychological task performances [bib_ref] Neurocognitive deficit in schizophrenia : a quantitative review of the evidence, Heinrichs [/bib_ref] [bib_ref] Neuropsychological impairments in schizophrenia : integration of performance-based and brain imaging findings, Reichenberg [/bib_ref] , and these impairments often impede everyday function and quality of life for many patients . PCEs may prove beneficial as an add-on to antipsychotic medication, as it has been suggested that, in patients with schizophrenia, even small improvements in cognitive functions, such as enhancing the ability to adapt efficiently to new situations and to plan effectively, could help them make the transition to independent living [bib_ref] Designing outcome studies to determine efficacy and safety of antipsychotics for '..., Altamura [/bib_ref] [bib_ref] Cognitive impairment as a target for pharmacological treatment in schizophrenia, Davison [/bib_ref]. Enhancing patients ' cognition will not only improve their quality of life, but will also enable them to access jobs and integrate with society. In consequence, governments are relieved from the cost burden of ongoing care for these patients [bib_ref] Burden of schizophrenia in recently diagnosed patients : healthcare utilisation and cost..., Nicholl [/bib_ref]. In keeping with this, we have shown that PCE drugs improve cognitive performance in patients with schizophrenia , and that modafinil (Provigil 1 ), a wake-promoting drug licensed for narcolepsy, can improve cognitive flexibility as measured by extra-dimensional attentional set-shifting in patients with chronic schizophrenia [bib_ref] Modafinil improves cognition and attentional set shifting in patients with chronic schizophrenia, Turner [/bib_ref]. Recently, modafinil has been shown to enhance some aspects of social cognition such as emotional facial recognition in patients with first episode of psychosis [bib_ref] The effects of modafinil on cognitive and emotional functions in the first..., Scoriels [/bib_ref]. However, many of the studies cited above are acute proof of concept studies and therefore one must be cautious about inferred long-term clinical significance, which still requires confirmation in experimental studies.
Nevertheless, children with ADHD can benefit from PCEs. ADHD is a heritable and disabling disorder characterized by core cognitive and behavioural symptoms of impulsivity, hyperactivity, and inattention. It is the most prevalent neuropsychiatric childhood disorder which affects around 3-7 % of children worldwide [bib_ref] The worldwide prevalence of ADHD : a systematic review and metaregression analysis, Polanczyk [/bib_ref]. Structural abnormalities in fronto-striato-circuitry [bib_ref] Differential patterns of striatal activation in young children with and without ADHD, Durston [/bib_ref] and dysfunction in catecholamine neurotransmission, specifically in NA and DA pathways in the prefrontal cortex (PFC) have been implicated in ADHD [bib_ref] A meta-analysis of association studies between the 10-repeat allele of a VNTR..., Yang [/bib_ref]. This leads to inefficient information processing and hypo-activation in the frontal lobes. As a result, ADHD patients have significant impairments in performing working memory (WM) and executive function tasks [bib_ref] Atomoxetine improved response inhibition in adults with attention deficit/hyperactivity disorder, Chamberlain [/bib_ref] [bib_ref] Impaired spatial working memory in adults with attention-deficit/hyperactivity disorder : comparisons with..., Dowson [/bib_ref]. However, if these impairments are not treated early, they can lead to significant negative life events such as drop out from education, job dismissal, criminal activities, substance abuse, and driving accidents. PCEs such as methylphenidate (Ritalin 1 ) [bib_ref] Methylphenidate has differential effects on blood oxygenation level-dependent signal related to cognitive..., Dodds [/bib_ref] , selective NA reuptake inhibitor (SNRI) atomoxetine (Strattera 1 ) [bib_ref] Atomoxetine improved response inhibition in adults with attention deficit/hyperactivity disorder, Chamberlain [/bib_ref] [bib_ref] Safety and tolerability of atomoxetine over 3 to 4 years in children..., Donnelly [/bib_ref] and modafinil [bib_ref] Cognitive enhancement effects of modafinil in healthy volunteers, Turner [/bib_ref] , increase DA and NA levels in the PFC [bib_ref] Attention deficit hyperactivity disorder and substance use disorders, Wilens [/bib_ref] and alleviate cognitive impairments in ADHD patients [bib_ref] The effects of methylphenidate on decision making in attention-deficit/hyperactivity disorder, Devito [/bib_ref] [bib_ref] Modafinil improves cognition and response inhibition in adult ADHD, Turner [/bib_ref]. Indeed, studies in our laboratory that use double-blind placebo-controlled designs showed that methylphenidate improves WM, cognitive flexibility, attention and response inhibition in both children and adults with [bib_ref] Modafinil improves cognition and response inhibition in adult ADHD, Turner [/bib_ref]. For example, [bib_ref] The effects of methylphenidate on decision making in attention-deficit/hyperactivity disorder, Devito [/bib_ref] used the Cambridge Gambling Task, a test of decision making and risk taking, and showed that a single 0.5 mg/kg dose of methylphenidate reduced large bets on this task in ADHD boys, who performed similarly to healthy boys without medication. In consequence, PCEs can effectively improve core symptoms, abnormal behaviours, self-esteem, cognition, social and family function in ADHD patients . However, methylphenidate is successful in treating only about 60-70 % of ADHD children, meaning that 30 % of patients with ADHD either do not respond to treatment or the drug causes adverse side-effects, such as headache, stomach pain, loss of appetite, trouble sleeping, dizziness and nausea, which precludes the use of methylphenidate. Atomoxetine may be a more acceptable treatment due to its low abuse liability and minimal adverse sideeffects [bib_ref] Comparison of the subjective, physiological, and psychomotor effects of atomoxetine and methylphenidate..., Heil [/bib_ref]. Still, unlike methylphenidate which improves spatial WM (SWM), sustained attention, and response time in ADHD patients and healthy volunteers [bib_ref] Effects of methylphenidate on spatial working memory and planning in healthy young..., Elliott [/bib_ref] [bib_ref] Modafinil improves cognition and response inhibition in adult ADHD, Turner [/bib_ref] , atomoxetine only improves response inhibition possibly due to its selective NA modulation [bib_ref] Neurochemical modulation of response inhibition and probabilistic learning in humans, Chamberlain [/bib_ref] [bib_ref] Atomoxetine improved response inhibition in adults with attention deficit/hyperactivity disorder, Chamberlain [/bib_ref] , and is likely to be less effective in treating the range of cognitive deficits associated with ADHD. Therefore, there is a need for medication with improved efficacy and reduced side-effects for ADHD.
Furthermore, studies employing the same methodology show that modafinil also significantly improves short-term memory span, visual memory, spatial planning, and stop-signal motor inhibition in ADHD adults [bib_ref] Modafinil improves cognition and response inhibition in adult ADHD, Turner [/bib_ref]. These improvements are consistent with randomized, double-blind placebocontrolled clinical trials with modafinil that demonstrate symptom reduction in ADHD children and adolescents [bib_ref] Efficacy and safety of modafinil film-coated tablets in children and adolescents with..., Biederman [/bib_ref] [bib_ref] A randomized, double-blind, placebo-controlled study of modafinil film-coated tablets in children and..., Greenhill [/bib_ref] [bib_ref] Modafinil film-coated tablets in children and adolescents with attention-deficit/hyperactivity disorder : results..., Swanson [/bib_ref]. Hence, in psychiatry ethics, developing novel PCEs that improve the wellbeing and quality of life for these patients meets the ' right to receive effective treatment that would offer them a reasonable opportunity to improve their mental condition ' .
## Pharmacogenomics
Another important argument for developing novel PCEs relates to the rise of pharmacogenomics and individualized medicine that aim to combat neuropsychiatric and neurodegenerative diseases. Pharmacogenomics is the discipline behind how genes influence the body's response to drugs. There is growing evidence that key gene variants can change activity within specific neuronal circuits and, as a result, influence particular cognitive-affective phenomena. For example, the catecholamine-Omethyltrasferase (COMT) gene has been shown to affect responses to COMT inhibitors and to predict WM performance whereas the Val 108/158 polymorphism exerts a significant effect on enzyme activity and affects DA-regulated PFC activity during WM tasks, and also modifies the effect of dopaminergic drugs (e.g. the COMT enzyme inhibitor tolcapone) in the PFC [bib_ref] Genetic variation in catechol-omethyltransferase : effects on working memory in schizophrenic patients,..., Diaz-Asper [/bib_ref] [bib_ref] Executive subprocesses in working memory : relationship to catechol-O-ethyltransferase Val158Met genotype and..., Goldberg [/bib_ref]. Similarly, the therapeutic response in AD appears to be genotypespecific, with APOE-4/4 carriers being the worst responders to conventional treatments [bib_ref] Pharmacogenomics and therapeutic prospects in Alzheimer's disease, Cacabelos [/bib_ref]. However, although behavioural phenotypes and action of PCEs are generally complex, both reflecting the action of multiple genes and neurotransmitters respectively, it is possible that using pharmacogenomics to develop targeted PCEs for particular subgroups and individual responsiveness will lead to greater efficacy and reduced sideeffects.
## Pces improve cognition in healthy individuals
PCEs also improve cognition in non-sleep-deprived healthy adults. For example, [bib_ref] Cognitive enhancement effects of modafinil in healthy volunteers, Turner [/bib_ref] showed that a single oral dose of modafinil (100 mg or 200 mg) significantly improved performance on tests of digit span, visual pattern recognition memory, spatial planning, and stop signal reaction time (SSRT) task, or response inhibition, in healthy volunteers. Modafinil also improved the response time in tests of decision making, delayed matching to sample, and spatial planning [bib_ref] Effects of modafinil on working memory processes in humans, Mü Ller [/bib_ref]. More recently, it improved accuracy in an attention-shifting task, without reaction time trade-off [bib_ref] Modafinil improves rapid shifts of attention, Marchant [/bib_ref] , especially when participants ' resources were most challenged. Consistent with this, Mü ller et al. demonstrated that modafinil significantly reduced error rates in a long-delay visuo-spatial task and manipulation conditions, without speed-accuracy trade-off.
Similarly, an acute dose of atomoxetine can improve response inhibition in healthy adults. [bib_ref] Atomoxetine modulates right inferior frontal activation during inhibitory control : a pharmacological..., Chamberlain [/bib_ref] used functional magnetic resonance imaging and examined the brain mechanism by which atomoxetine exerts its cognitive enhancing effects in healthy volunteers. They found that atomoxetine led to increased activation in the right inferior frontal gyrus (RIFG) when participants attempted to inhibit their responses in the SSRT task. The RIFG has previously been shown to be activated during inhibitory motor control . It is thought that atomoxetine improves response inhibition via noradrenergic mechanisms. In contrast, methylphenidate has been shown to enhance SWM performance in healthy adults [bib_ref] Effects of methylphenidate on spatial working memory and planning in healthy young..., Elliott [/bib_ref]. In a study employing sophisticated neuropsychological tests and brainimaging measures (see Supplementary online material), [bib_ref] Methylphenidate enhances working memory by modulating discrete frontal and parietal lobe regions..., Mehta [/bib_ref] showed that methylphenidate improves both performance and efficiency in the SWM neural network, which includes the dorsolateral PFC and posterior parietal cortex in healthy volunteers [bib_ref] Evidence for a two-stage model of spatial working memory processing within the..., Owen [/bib_ref]. These areas have been significantly associated with WM and executive functions. Moreover, studies using positron emission tomography and contrasting [ 11 C]raclopride binding, with the participants either on or off methylphenidate, have further indicated that methylphenidate influences dopaminergic function, particularly in the striatum [bib_ref] Reproducibility of repeated measures of endogenous dopamine competition with [ 11 C]..., Wang [/bib_ref]. Other DA agonists improve WM and performance of executive tasks in healthy individuals [bib_ref] Dopaminergic enhancement of cognitive function, Mehta [/bib_ref] [bib_ref] Differential dopaminergic modulation of executive control in healthy subjects, Roesch-Ely [/bib_ref]. There is evidence for enhancement of other forms of memory by PCEs. For example, evidence from healthy volunteers show that ampakine CX516 and ChEIs also lead to moderate improvements in recall and short-term memory [bib_ref] Acute effects of the ampakine farampator on memory and information processing in..., Wezenberg [/bib_ref]. In particular, pilots who took donepezil just before learning specific manoeuvres in a flight simulator outperformed a control group on tests of performance conducted 1 month later [bib_ref] Donepezil and flight simulator performance : effects on retention of complex skills, Yesavage [/bib_ref]. However, the mechanism of action for improvement in attention, memory and executive function of PCEs still remains to be determined in many cases. For example, to exert its cognitive enhancing effects, modafinil has been shown to elevate numerous neurotransmitters including NA, DA and glutamate [bib_ref] Modafinil : a review of neurochemical actions and effects on cognition, Minzenberg [/bib_ref] [bib_ref] Effects of modafinil on dopamine and dopamine transporters in the male human..., Volkow [/bib_ref].
## Neurotransmitter modulation of cognition
Evidently, as these improvements relate to neurotransmitter modulation and function, the effects of some PCEs might follow the Yerkes-Dodson law, which explains the relationship between arousal and performance. This principle might be translated to several neurotransmitter systems where cognitive function often follows an inverted U-shaped curve, with deviations from the optimal level in either direction producing suboptimal performance [bib_ref] Paradoxical ' effects of psychomotor stimulant drugs in hyperactive children from the..., Robbins [/bib_ref] [bib_ref] Acute tryptophan depletion evokes negative mood in healthy females who have previously..., Robinson [/bib_ref]. For instance, low levels of NA release engage a 2 -adrenergic receptors and improve executive function whereas higher levels of NA release engage a 1 -adrenergic receptors which cause significant stress in humans and animals [bib_ref] Through the looking glass : differential noradrenergic modulation of prefrontal cortical function, Arnsten [/bib_ref] [bib_ref] Increased dopamine and norepinephrine release in medial prefrontal cortex induced by acute..., Finlay [/bib_ref] and impair prefrontal functionality [bib_ref] Adrenergic pharmacology and cognition : focus on the prefrontal cortex, Ramos [/bib_ref]. A similar U-inverted relationship is evidenced between DA and WM function [bib_ref] Inverted-U dopamine D1 receptor actions on prefrontal neurons engaged in working memory, Vijayraghavan [/bib_ref] as both marked increases and decreases of DA in the PFC have been associated with sub-optimal performance . Consistent with this hypothesis, methylphenidate improves cognitive performance in individuals with greater impairment [bib_ref] Differential effects of methylphenidate on attentional functions in children with attention-deficit/ hyperactivity..., Konrad [/bib_ref] [bib_ref] Methylphenidate enhances working memory by modulating discrete frontal and parietal lobe regions..., Mehta [/bib_ref] while guanfacine (Tenex 1 ), an a 2 -adrenergic receptor agonist, has beneficial effects on WM and attentional functions in patients with ADHD, but does not improve WM or executive functions in healthy male volunteers [bib_ref] Lack of effects of guanfacine on executive and memory functions in healthy..., Mü Ller [/bib_ref]. Thus, the effects of pharmacological substances on cognition are complex as cognition is a multifaceted construct encompassing numerous mental functions including both cold cognition (such as attention, planning, problem solving, and response inhibition) and hot cognition (such as risky decision making ; [bib_ref] Neuropsychology of mood disorders, Roiser [/bib_ref]. For instance, PCEs may further modulate important human virtues such as creativity. A recent study by [bib_ref] When we enhance cognition with adderall, do we sacrifice creativity? A preliminary..., Farah [/bib_ref] showed that the mixed amphetamine salts, adderall, licensed for the treatment of ADHD, enhanced performance on convergent tasks of creativity for lowerperforming individuals and either impaired or did not change it for higher-performing individuals. These results on improvement and impairment on higher cognitive function with PCEs raises the issue of what we mean by a general term 'enhancement '. As healthy adults fall into a wide spectrum of normality, some individuals may be improved by a PCE drug while others remain unchanged or are even impaired [bib_ref] Cognitive effects of modafinil in student volunteers may depend on IQ, Randall [/bib_ref] [bib_ref] Paradoxical ' effects of psychomotor stimulant drugs in hyperactive children from the..., Robbins [/bib_ref]. Furthermore, there is as yet no robust empirical research to demonstrate that PCEs have effects on divergent thinking in healthy people.
## Lifestyle use of pces by healthy individuals
The above results demonstrate the potential of PCEs to enhance certain cognitive domains in healthy adults. Therefore, attitudes towards their use by the general population need to be considered. In the next section, we focus our discussion on current and future trends of the use of PCEs by healthy people.
In the past few years there has been an unprecedented rise in the use of PCEs among healthy individuals for cognitive enhancement. Cognitive enhancement can be defined as the amplification or extension of core capacities of the mind through improvement or augmentation of internal and external information processing systems . Healthy university students [bib_ref] Adderall is definitely not a drug' : justifications for the illegal use..., Desantis [/bib_ref] and academics have been using PCEs to improve their cognitive function. More specifically, students are taking PCEs to improve academic performance [bib_ref] Motives and perceived consequences of nonmedical ADHD medication use by college students..., Rabiner [/bib_ref] and are framing their actions as both physically harmless and morally acceptable [bib_ref] Adderall is definitely not a drug' : justifications for the illegal use..., Desantis [/bib_ref]. For example, in the USA, 16 % of college students [bib_ref] Student perceptions of methylphenidate abuse at a public liberal arts college, Babcock [/bib_ref] and 8 % of undergraduates reported having illicitly obtained and used prescription stimulants [bib_ref] Illicit use of prescribed stimulant medication among college students, Hall [/bib_ref] [bib_ref] Nonmedical use of prescription opioids and stimulants among student pharmacists, Lord [/bib_ref] [bib_ref] Prevalence and motives for illicit use of prescription stimulants in an undergraduate..., Teter [/bib_ref]. Furthermore, a 2005 survey by the US National Institute on Drug Abuse [bib_ref] The Basics of commonly abused drugs, Nida [/bib_ref] found that 2.5 % of 13-to 14-yr-olds, 3.4 % of 15-to 16-yr-olds and 5.1 % of 17-to 18-yr-olds abused methylphenidate. In 2009, the figures for these groups were 1.8 %, 3.6 % and 2.1 %, respectively (NIDA, 2009). Presumably, these young people are obtaining stimulant drugs from others who have prescriptions or purchasing them via the internet or street dealers.
Currently, the global market share of modafinil is more than US$700 million per year [bib_ref] Neuroenhancment : status quo and perspectives, Norman [/bib_ref]. Consistent with [bib_ref] The social effects of advances in neuroscience : legal problems, legal perspectives, Greely [/bib_ref] claim, that healthy physicians on call, students and academics are increasingly using PCEs to enhance cognitive abilities, it is estimated that around 90 % of modafinil is predominantly used off-label by healthy, non-sleepdeprived individuals [bib_ref] Effects of modafinil on cognitive and meta-cognitive performance, Baranski [/bib_ref] [bib_ref] Poised to challenge need for sleep, ' wakefulness enhancer ' rouses concerns, Vastag [/bib_ref]. In contrast, beta-blockers that are prescribed to reduce anxiety in clinical patients have been used by musicians to dampen physiological tremors in order to improve their performances on stage.
In the UK, a newspaper survey of 1000 students showed that 1 in 10 were taking prescription drugs for cognitive enhancement. In England, prescription rates of stimulants have been rising steadily from 220000 in 1998 to 418300 in 2004 [bib_ref] Reshaping the Human Condition : Exploring Human Enhancement, Niyadurupola [/bib_ref]. In 2008, the journal Nature conducted a poll about the use of PCEs by healthy academics, in which 1400 scientists from 60 different countries responded [bib_ref] Poll results : look who is doping, Maher [/bib_ref]. One in five respondents used drugs for cognitive enhancement, with 52 % of them obtaining the drug by prescription, while 34 % obtained the drug via the internet and 14 % through their pharmacy. The most popular drug was methylphenidate, with 62 % of users ; 44 % reported taking modafinil mainly to improve concentration, and 15 % reported taking beta-blockers for anxiety. Of all respondents, 96 % thought that people with neuropsychiatric disorders should be given cognitive enhancing drugs. In contrast, 86 % of respondents thought that healthy children under the age of 16 yr should be restricted from taking PCE drugs.
Although some of these data were not rigorously collected, they nonetheless suggest the increased use of PCEs among healthy individuals. Their widespread use is not surprising given that small percentage increments in performance can lead to significant improvements in functional outcome. Indeed, a 10 % improvement in memory score could lead to an improvement in an A-level grade or degree class .
## Neuroethical issues in cognitive enhancement
Nevertheless, the increase in lifestyle use of PCEs by healthy people raises numerous ethical issues that inform the growing field of neuroethics. Neuroethics is the study of the ethical, legal and social questions that arise when scientific findings about the brain are carried into medical practice, legal interpretations and health and social policy. As such, modifying our inherent self, character and individuality through PCE drugs has important implications for society. For these reasons, their lifestyle use has prompted a significant interest both in the media and the public [bib_ref] Modafinil in the media : metaphors, medicalisation and the body, Coveney [/bib_ref] [bib_ref] Turbo charging the brain, Stix [/bib_ref]. There is a concern that PCEs will threaten our notion of personhood and will dampen essential characteristics of what it means to be human [bib_ref] Neurocognitive enhancement : what can we do and what should we do?, Farah [/bib_ref].
As a consequence, enhancing the brain and higher cognitive processes demands strong ethical considerations and a practical policy framework. To address this, we have argued elsewhere that before PCE drugs are prescribed to healthy people, their longterm safety, side-effects and their effectiveness must be tested to provide important facts necessary for further decision making about their regulation . Moreover, we have engaged with the media about the need to establish regulations for the use of PCE drugs by healthy people [bib_ref] It's time to establish rules for the use of cognitive-enhancing drugs by..., Mohamed [/bib_ref]. In Nature, we emphasized the need to ensure their safe use by healthy people [bib_ref] Towards responsible use of cognitive-enhancing drugs by the healthy, Greely [/bib_ref] while in Science we advocated that ethical considerations in regard to societal issues associated with the use of PCEs by healthy people should be part of neuroethical training within university neuroscience programmes . But, what are the advantages and disadvantages of healthy people using PCEs? Since PCEs improve those with low cognitive performance [bib_ref] Paradoxical ' effects of psychomotor stimulant drugs in hyperactive children from the..., Robbins [/bib_ref] , it might be possible to mitigate the adverse environmental effects, such as poverty, on the brain and cognition through their use. This might contribute to removing disparity in society. It may also be that some ' healthy' people actually have undiagnosed attentional or other problems and are actually selfmedicating with drugs such as ritalin. Furthermore, even healthy adults, who normally function well, are not always performing optimally due to sleep deprivation, jet lag or other stressors, and some might need to perform at their best possible level on every occasion (e.g. surgeons, air traffic controllers). In addition, PCEs might enable us to perform better in other competitive or life threatening situations. For instance, psychostimulants have been employed to boost cognition in soldiers in combat [bib_ref] A double-blind, placebo-controlled investigation of the efficacy of modafinil for sustaining the..., Caldwell [/bib_ref] [bib_ref] Psychostimulants and military operations, Moran [/bib_ref] [bib_ref] Ethical use of cogniceuticals in the militaries of democratic nations, Russo [/bib_ref]. PCEs have also been demonstrated to improve performance in shift workers [bib_ref] A systematic review of modafinil : potential clinical uses and mechanisms of..., Ballon [/bib_ref] , pilots [bib_ref] Efficacy of stimulants for fatigue management : the effects of provigil 1..., Caldwell [/bib_ref] and school pupils with ADHD [bib_ref] A review of non-medication interventions to improve the academic performance of children..., Trout [/bib_ref]. Recently, the US Defence Advanced Research Project Agency (DARPA, 2007) introduced the Augmented Cognition Programme to enhance soldiers ' memory and cognition through technology when under conditions of interrupted sleep and stress. However, if proven to be safe, PCEs may be preferred for cost-effectiveness when compared to other methods of enhancement such as expensive technology.
In contrast, the disadvantages of using PCEs include the potential harms and long-term side-effects that they might have in healthy people, particularly in adolescents where the brain is still in development. There are strong safety concerns, especially in the absence of informative data, for healthy individuals as the risk of adverse side-effects might outweigh the beneficial effects of PCE drugs. The abuse liability of some of the PCEs such as methylphenidate is also a concern. A recent study showed that modafinil blocked DA transporters and increased DA in the caudate, putamen and nucleus accumbens in healthy human brain [bib_ref] Effects of modafinil on dopamine and dopamine transporters in the male human..., Volkow [/bib_ref] , which are areas in a network known to be involved in drug-seeking behaviour and addiction [bib_ref] Drug addiction : the neurobiology of behaviour gone awry, Volkow [/bib_ref]. This indicates the need for awareness about the risks involved in PCE use among healthy people and shows that a full ethical consideration of their use is required. To date, there have been no randomized psychopharmacological trials investigating the long-term effects of PCE drugs on healthy people.
There is still a further safety concern about the risks of purchasing substances advertised as PCEs over the internet . As these drugs are not prescribed by a qualified doctor, they might not be suitable for some people. For instance, contraindications of atomoxetine and modafinil include heart problems and hepatic impairments. Additionally, if one is taking other medication there might be serious drug-drug interactions which could be dangerous in some cases.
With regard to personal autonomy, there are ethical concerns about healthy people being coerced or even forced into using a PCE. Society might force people to take psychoactive agents in order to perform better or to be in a particular mental state. For example, authorities in the USA ordered a mentally ill inmate in criminal proceedings to take psychotropic medication to improve his competence to stand trial and be executed. There is also a considerable potential for indirect coercion resulting from a highly demanding 24/7 society where people feel compelled to take PCEs in order to meet social or workplace demands. Healthy people may resort to self-medication for inadequate sleep or overexertion at work. For example, 33 % of respondents in [bib_ref] Poll results : look who is doping, Maher [/bib_ref] poll indicated that they would feel pressure to give PCE drugs to their children if other children at school were taking them. However, the use of PCEs to enhance cognition is one solution to improving the individual and society. Indeed, we have argued elsewhere that there are other methods of boosting cognition, including education and exercise . For instance, physical exercise can improve learning and memory [bib_ref] Running enhances spatial pattern separation in mice, Creer [/bib_ref] [bib_ref] Science and society : be smart, exercise your heart : exercise effects..., Hillman [/bib_ref]. Through these non-pharmacological means we might be able to effectively and safely enhance cognition and well-being in society.
Another argument against their use is that they might further exacerbate the ever-growing disparity and inequality in society, especially if only the wealthy can access them. Equally, is it morally justified to use PCEs during exams, and does it give the user unfair advantage over those who are equally capable but not cognitively enhanced with drugs? Many universities as yet have no formal policy about the use of PCEs during exams. If PCEs become easily accessible in the future, will society consider their use as cheating or will they equate them to having a caffeine boost from coffee? Is it possible that once PCE drugs are widely available we might run the risk of becoming a homogeneous society? Could our perception of ourselves change from being human to being mechanistic beings with a modicum of emotion? Are we going to be over-enhanced only to be plagued by unwanted memories? Will using PCEs outdate important human virtues such as hard work and reflection and make us unable to take credit for our minds ' achievements?
The advantages and disadvantages of using PCEs have to be evaluated carefully. With regards to fairness, enhancing cognition might lead to dramatic social benefits by reducing natural inequality and promoting social justice [bib_ref] Justice, fairness, and enhancement, Savulescu [/bib_ref]. This is because increasing cognitive ability on an individual level could have dramatic and positive effects on society and the economy as a whole. For instance, a 3 % population-wide increase in IQ would reduce poverty rates by 25 % [bib_ref] A risk assessment perspective on the neurobehavioral toxicity of endocrine disruptors, Weiss [/bib_ref] , and would lead to an annual economic gain of US$165-195 billion and up to 1.5 % GDP growth [bib_ref] Updated estimates of earnings benefits from reduced exposure of children to environmental..., Salkever [/bib_ref] [bib_ref] Low-level lead exposure and children's IQ : a meta-analysis and search for..., Schwartz [/bib_ref].
## Public engagement in neuroscience
However, if healthy people take PCEs to gain a competitive edge but fail to see any difference in the long-term or notice possible impairments observed in high-functioning adultsit could spark controversy and outcry in the public. Hence, determining who can use PCEs and under what circumstances involves complex decision making and ethical judgements. Thus, how neuroscientific discoveries impact on society has given rise to an enormous interest in the field of neuroethics, including the foundation of the Neuroethics Society (http:// www.neuroethicssociety.org) which advocates further research on ethical questions that are yet to be answered. For instance, what are the possible longterm harms of using PCE drugs in healthy people, particularly in the developing brain? What are the implications of developments in pharmacogenomics? Without formal regulation of their use, healthy people can purchase PCEs via the internet, with all the inherent dangers in doing so. How would such easy access affect widespread use of these drugs by healthy young and elderly people and also impact on society? How would we, as neuropsychopharmacologists, react if we discover that our colleagues or our children's friends are taking PCEs? How should governments react? These questions in neuropsychopharmacology and neuroethics merit further rigorous research. They also clearly indicate the need to engage in discussion with the public about the social and ethical implications of the use of PCEs by healthy individuals [bib_ref] Scientists talking to the public : is there anyone out there, Ringach [/bib_ref]. For this to happen, neuropsychopharmacologists need to integrate experimental results within a neuroethical framework. In order to do this, they need to innovatively work together with social scientists, philosophers, and ethicists (Morein-Zamir & . This increases neuroscientists ' role and responsibility in society [bib_ref] Neurocognitive enhancement : what can we do and what should we do?, Farah [/bib_ref] and puts them in a leading position to engage policy makers and a broad group of stakeholders, including the general public. This will ensure that technological advances in neuroscience are put to maximal benefit and minimal harm.
# Conclusions
PCEs have the potential to ameliorate cognitive dysfunction and to provide important clinical benefits for patients. Further development of more effective PCEs with fewer side-effects, in addition to neuroprotective agents for patients with neurodegenerative diseases such as AD, is clearly worthy of pursuit. Pharmacogenomics will make it possible to target individuals with safe and effective drugs. PCEs can also improve cognitive function such as memory and attention in healthy individuals. However, their longterm cognitive enhancing potential as well as their side-effects in healthy people needs to be rigorously determined. Currently, the unprecedented rise of PCE use among the healthy raises numerous ethical issues. Scientists need to work together with social scientists, philosophers, ethicists, policy makers, and teachers to actively discuss the ethical consequences of PCE usage. This will ensure maximal benefit and minimal harm in the advances in neuroscience. Finally, the use of PCEs to enhance cognition is one solution to improving the individual and society. However, this does not preclude other means of enhancing cognition such as education and exercise [bib_ref] The mental wealth of nations, Beddington [/bib_ref].
## Note
Supplementary material accompanies this paper on the Journal's website (http://journals.cambridge.org/ pnp). Acknowledgements B.J.S. is partly funded by a Wellcome Trust Grant Award (program grant 019407) and a consortium joint award from the MRC and Wellcome Trust (BCNI grant G0001354). A.D.M. is a Wellcome Trust-funded Ph.D. student at Clare Hall, Cambridge and is funded by The Oxford Centre for Neuroethics at the University of Oxford (Grant : 086041/Z/08/Z).
We thank Dr Ulrich Mü ller, Christopher Lewis and Charlotte Housden for discussion.
# Statement of interest
B.J.S. consults for Cambridge Cognition Ltd. She has consulted for Novartis, Shire, GlaxoSmithKline, Lilly and Boehringer-Ingelheim. She has also received honoraria for Grand Rounds in Psychiatry at Massachusetts General Hospital (CME credits) |
An additional challenge for head and neck radiologists: anatomic variants posing a surgical risk – a pictorial review
Anatomic variants in the head and neck are quite numerous and occur frequently: a minority of them increase the risk of complications during surgical procedures and may be visualized on cross-sectional images. As some of these complications are potentially fatal, awareness (and accurate reporting) of such variants is a basic responsibility of radiologists, particularly when surgery in the pertinent anatomic area is under consideration.
# Background
Anatomic variants in the head and neck are quite numerous and occur frequently, particularly in the sinonasal region. Some of them, mainly in the paranasal region, are known to predispose to pathology, whereas most bear little (if any) clinical significance. Only a minority of variants increase the risk of complications and iatrogenic damage during surgical procedures. As some of these complications are fatal, awareness and accurate reporting of such variants is a basic responsibility of radiologists. In this pictorial review, anatomic variants posing a surgical risk will be classified under four main categories: abnormal bone pneumatization, bone dehiscence and asymmetry, anomalous vessel course, and anomalous nerve course.
## Abnormal bone pneumatization
In the paranasal area, the Onodi cell is probably the most alarming variant. This cell is the extension of an ethmoid cell above and/or lateral to the sphenoid sinus; hence, it is also referred to as a sphenoethmoid cell. Given its location, it may have an intimate relationship with the optic nerve canal.
In fact, a more restrictive definition of the Onodi cell includes an optic nerve canal protrusion or dehiscence. During endoscopic sinus surgery, the transgression of the walls of an overlooked Onodi cell may result in irreversible optic nerve injury and/or profuse hemorrhage. The prevalence of this anomaly is variable: Shin et al.found an incidence of~30% with a good correlation between multidetector CT (MDCT) and intraoperative findings. However, literature data on its prevalence are inconsistent, with incidence ranging between 10.9%and 65%.
On cross-sectional scans, the Onodi cell can be best appreciated in coronal plane images, when a horizontal septum is seen crossing the sphenoid sinus lumen. Then, axial and sagittal reconstructions should be carefully scrutinized to detect the sphenoethmoid recess (the reference point of the actual location of the sphenoid sinus) and confirm the location of the Onodi cell on top of the sphenoid sinus.
Another point of concern for optic nerve injury during surgery is the anterior clinoid process pneumatization. In this case, the risk of iatrogenic damage is related to the thickness of the bony walls of the process and to the degree of pneumatization surrounding the nerve. The prevalence of this variant is low (6-13%).
The infraorbital (or Haller's) cell is an extension of ethmoid pneumatization to the orbital wall, inferolateral to the ethmoid bulla . This cell may obstruct the ethmoid infundibulum, thereby predisposing the patient to maxillary sinusitis and may increase the hazard of orbital penetration during endoscopic sinus surgery; while opening such a cell, in fact, the surgeon may dangerously leverage on the inferomedial orbital wall.
Both clinoid pneumatization and Haller's cell are readily detected on coronal MDCT scans or on cone beam CT (CBCT) scans.
A unilaterally shrunk maxillary sinus may indicate the collapse of the uncinate process along the inferomedial orbital wall. This anomalous position of the uncinate process enhances the risk of accidental orbital penetration during uncinectomy, a frequent procedure in the early stages of endoscopic sinus surgery. The pathophysiology in this anatomic variant is triggered by adhesion between the uncinate process and the orbital wall; in fact, this induces chronic hypoventilation and negative pressure within the maxillary sinus resulting in shrinkage of the lumen and inflammatory thickening of the mucosa. Negative sinus pressure also induces partial collapse of the orbital floor and increased vertical diameter of the orbit. All of these signs are referred to as silent sinus syndromeand can be easily detected on (CB) MDCT or magnetic resonance imaging (MRI) scans.
The anatomy of the frontal sinus is largely conditioned by the degree of pneumatization of surrounding air cells (mostly agger nasi cells) and by the cranial attachment of the uncinate process. Some configurations (i.e., supraorbital ethmoid air cell and deep olfactory fossa) significantly influence the complexity of the procedure and may increase the relative risk of complications.
## Bone dehiscence and asymmetry
Focal dehiscence of the orbital wall may occur as a result of trauma; when no history of trauma is reported by the patient, it is a matter of debate whether such an anomaly should be classified as congenital or secondary to minor, clinically overlooked, traumatic events. The risk of iatrogenic orbital injury during endoscopic sinus surgery is quite obvious.. On MDCT scans displayed with bone-windowing, herniation of the orbital content through gaps in the medial or inferior wall may be easily concealed when the ethmoid or maxillary mucosa is thickened; therefore, whenever gaps are seen on scans with bone windowing, soft tissue reconstructions should be obtained.
The anatomic configuration of the ethmoid roof is quite variable. The depth of the cribriform plate is a key point for endoscopic sinus surgery planning; Keros classified it in three groups, according to the length of the vertical lamella of the ethmoid: type I indicates less than 3 mm depth, type III more than 7 mm, type II ranging from 3 mm to 7 mm [8]. Pre-operative assessment of the depth of the olfactory fossa is crucial because, during surgery, the forces applied to the concha media may lead to breakage of the vertical lamella and to cerebrospinal fluid leak. The Keros type III configuration bears an increased risk of such complication. Meyers and Valvassoriproposed a more practical classification: a horizontal line is drawn connecting the cribriform plate to the lateral orbital wall and the depth is defined based on whether the line crosses the upper third of the orbit, the midline, or further below. Furthermore, asymmetry between the two sides, not an infrequent condition, portends an increased surgical risk and should therefore be accurately reported.
When endoscopic surgery of the sphenoid sinus is planned, intraluminal protrusion or focal dehiscence of the vertical tract of the internal carotid artery (ICA) canal increases the risk of surgical injury . Moreover, bone septa may act as leverage on the carotid canal.
In the temporal bone, the tympanic segment of the facial nerve may protrude into the middle ear cavity through a dehiscent bony canal. This is better appreciated on coronal reformations, showing the nerve hanging in the middle ear cavity, strictly contiguous to the oval window ; however, the variant may be obscured by middle ear disease. Bone dehiscence of the facial nerve canal may also be at the second genu; in a review of 202 patients treated surgically for chronic ear disease, dehiscence of the facial nerve canal (overall seen in 8.9%) was equally frequent in the tympanic segment and second genu.
In addition, the sigmoid plate covering the jugular bulb may be dehiscent; large gaps will result in protrusion of the vein in the mesotympanum, often manifesting with tinnitus and a vascular tympanic membrane. Occasionally, multiple tegmen defects (honeycomb tegmen)may increase the risk of middle cranial fossa penetration during middle ear surgery.
## Anomalous vessel course
The high-riding truncus brachiocephalicus is upwardly shifted in the lower neck, with the bifurcation lying close to the thyroid gland. This variant probably develops as a consequence of anomalous regression of the IV arch .
The caudal part of the thyroid gland can be supplied by a thyroid ima artery, a variant often associated with the absence of inferior thyroid arteries. A thyroid ima artery may arise from the aortic arch, the Carotid canal variants. a Both canals protrude into the sphenoid sinus cavities; the bony wall is thinned on the right side (arrows). b Incomplete bony septa stem from the bony walls of the canals (arrows): during endoscopic surgery, such septa may act as a leverage and facilitate injury of the ICA Facial nerve canal dehiscence. Coronal CBCT reconstruction shows dehiscence of the canal and minimal protrusion of the facial nerve (arrow) within the middle ear cavity truncus brachiocephalicus, right common carotid artery, or internal thoracic artery and reaches the thyroid bed coursing along the anterior surface of the trachea.
Both of these variant vessels may pose a surgical threat, principally when tracheostomy or thyroid/parathyroid surgery is planned; identification on crosssectional imaging requires careful assessment of vascular structures in the peritracheal soft tissues.
A retropharyngeal carotid artery relates to the relatively common medial shift of the ICA (less frequently the common or external carotid artery); it can be bilateral, a condition referred to as "kissing carotids" and is prevalently seen at the oropharyngeal and hypopharyngeal level.
The incidence of this variant is linearly related to patient age; thus, it is possibly explained by increased tortuosity and atherosclerotic changes or by hypertension. Severe complications may be generated by retropharyngeal carotid artery injury, even during routine surgery such as tonsillectomy or peritonsillar abscess drainage. Interestingly, a change in position (from and to retropharyngeal) on MDCT examinations acquired at different time points has been described in 6.3% of cases.
In the temporal bone, the vertical portion of the petrous internal carotid artery may be undeveloped and bypassed by hypertrophied inferior tympanic and caroticotympanic arteries, coursing in the hypotympanum: this condition is named an aberrant ICA. It may manifest as pulsatile tinnitus and mimic a vascular mass on otoscopy or may remain asymptomatic. CBCT/ MDCT may indicate the absence of the vertical portion of the ICA, presence of a hypotympanic soft tissue mass, enlargement of the inferior canaliculus, and absence of bone coverage on the intratympanic segment of the vessel. As the hypotympanic soft tissue mass may be obscured by diffuse inflammatory opacification of the middle ear, awareness of this condition is crucial. On MRI, time-of-flight (TOF) angiography shows a pinched contour at the intersection of the vertical and horizontal segments of the ICA.
Rarely, hemorrhage during middle ear surgery may be produced by injury to a persistent stapedial artery. During fetal life, the stapedial artery provides a connection between branches of the external and internal carotid artery; in about 0.05% of cases, the vessel does not regress and may be seen in its entire course, arising from the petrous ICA. The persistent stapedial artery crosses the antero-medial hypotympanum, courses between the crura of the stapes to reach the facial nerve canal, and follows retrogradely a short segment of its tympanic High-riding truncus brachiocephalicus. MIP (a) and 3D volume rendering (b) reconstruction in an oblique frontal orientation depicting a high-riding truncus brachiocephalicus (arrows) in the lower neck coursing transverse in front of the trachea, high above the sternal manubrium portion up to the geniculate ganglion, where it enters the extradural space in the middle cranial fossa. When the stapedial artery persists, the middle meningeal artery arises from it, and thus MDCT and MRI images show the absence of the foramen spinosum; in addition, a small vascular canal may be seen along the cochlear promontory and the facial nerve canal will have an abnormally large diameter. High-resolution submillimetric (i.e., 0.9 mm or less isotropic voxel) MRI sequences with gadolinium will show the vessel, along with an abnormal enhancement along the second segment of the petrous facial nerve.
## Anomalous nerve course
In the sinonasal area, anomalous nerve course is often the consequence of abnormal pneumatization and/or dehiscence of the bony walls of their canals. As a result, cranial nerves in the maxillofacial area and temporal bone may be seen coursing within air cavities. The maxillary and vidian nerve may protrude into the sphenoid sinus when the pterygoid root is pneumatized thus creating a lateral recess. Less commonly, the infraorbital nerve may protrude into or hang in the maxillary sinus.
Surgical risk is amplified if the bony laminae surrounding such nerves are dehiscent; an intrasinusal infraorbital nerve is at risk during endoscopic surgery when hidden by or coursing within the laminae of an infraorbital (Haller's) cell.
Two nerve course variants in the mandible demand extreme caution during tooth extraction: the mandibular nerve canal may be completely encircled by molar teeth roots; additionally, a retromolar canal is seen in~25% of cases branching from the most proximal part of the mandibular canal to reach the retromolar fossa: nerves supplying molar teeth may be damaged during extraction of included elements. Both conditions are exquisitely demonstrated by CBCT images. A retrospective review in 136 patients with 257 impacted third molars identified several risk factors for iatrogenic nerve damage: contact between tooth roots and nerve, absent cortication of the nerve canal, nerve shape (teardrop and dumbbell shape), and nerve position relative to dental roots (lingual and interradicular).
In the lower neck, relevant nerve course variants may be detected only during surgery, such as the extralaryngeal ramification of the recurrent nerve andThyroid ima artery. Axial CT scan: a thin artery arises from the right common carotid artery (arrow in c), bends inferiorly (arrow in d) and then turns upwards (arrowheads in b and c) to reach the lower pole of the left thyroid gland lobe (asterisk in a) the numerous variations of the course of the spinal accessory nerve.
Though not directly demonstrable on crosssectional images, the presence of a non-recurrent right laryngeal nerve may be heralded by an indirect finding, namely the aberrant right subclavian artery. During embryologic life, the inferior laryngeal nerves supply V-VI branchial arches; on the right side, as these arches disappear, the nerve course retracts cranially being finally "trapped" around the right subclavian artery. Anomalous regression of the right IV arch results in an independent origin of the subclavian artery from the arch and a non-recurrent inferior laryngeal nerve unrestrained by the subclavian artery retracts cranially in the neck. A non-recurrent course of the right laryngeal nerve increases the risk of iatrogenic nerve injury during thyroid and parathyroid surgery.
# Conclusion
Some anatomic variants in the supra-and infrahyoid neck may increase surgical risk and, if overlooked, have the potential to generate serious complications. Often such variants produce subtle findings on crosssectional images. It is the responsibility of the radiologist reporting a head and neck scan to carefully scrutinize the regional anatomy in search of variants and to report on them, particularly when surgery is under consideration. . A thin arterial structure (arrows) is seen close to the ICA (c) along the promontory (b) and along the tympanic segment of the facial nerve (a). CT of the brain (d,e): coronal multiplanar reformation (MPR) and native axial acquisition. A linear structure runs along the cochlear promontory (arrows), and the left foramen spinosum is absent (black arrowhead points to the right foramen, normally present). Findings are consistent with a persistent stapedial artery |
Helicobacter pylori‐derived outer membrane vesicles contribute to Alzheimer's disease pathogenesis via C3‐C3aR signalling
The gut microbiota represents a diverse and dynamic population of microorganisms that can influence the health of the host. Increasing evidence supports the role of the gut microbiota as a key player in the pathogenesis of neurodegenerative diseases, including Alzheimer's disease (AD). Unfortunately, the mechanisms behind the interplay between gut pathogens and AD are still elusive. It is known that bacteria-derived outer membrane vesicles (OMVs) act as natural carriers of virulence factors that are central players in the pathogenesis of the bacteria. Helicobacter pylori (H. pylori) is a common gastric pathogen and H. pylori infection has been associated with an increased risk to develop AD. Here, we are the first to shed light on the role of OMVs derived from H. pylori on the brain in healthy conditions and on disease pathology in the case of AD. Our results reveal that H. pylori OMVs can cross the biological barriers, eventually reaching the brain. Once in the brain, these OMVs are taken up by astrocytes, which induce activation of glial cells and neuronal dysfunction, ultimately leading to exacerbated amyloid-β pathology and cognitive decline. Mechanistically, we identified a critical role for the complement component 3 (C3)-C3a receptor (C3aR) signalling in mediating the interaction between astrocytes, microglia and neurons upon the presence of gut H. pylori OMVs. Taken together, our study reveals that H. pylori has a detrimental effect on brain functionality and accelerates AD development via OMVs and C3-C3aR signalling.K E Y WO R D SAlzheimer's disease, gut-brain axis, Helicobacter pylori, bacterial extracellular vesicles (bEVs), outer membrane vesicles (OMVs), C3, complement This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. of XIE et al. of We reveal that H. pylori OMVs exacerbate Aβ pathology and induce cognitive impairment via regulation of glial cell activation and neuronal dysfunction. Moreover, we identify a critical role for complement component 3 (C3)-C3a receptor (C3aR) signalling in mediating the interactions of astrocyte-microglia-neuron in the context of the H. pylori OMVs challenge. This study provides new crucial insights into the role of H. pylori-derived OMVs in brain function and AD development. METHODS. Key resourcesKey resources were provided inTable S1.. Bacteria culture and outer membrane vesicles isolationH. pylori 26695 (ATCC 700392) was cultured at 37 - C in a microaerobic atmosphere (85% N 2 , 10% CO 2 , 5% O 2 ). H. pylori 26695 was grown in Brucella broth (10 g meat peptone, 10 g casein peptone, 2 g yeast extract, 5 g sodium chloride, 1 g glucose, 0.1 g sodium bisulphite in 1 L) supplemented with 10% exosome-depleted foetal bovine serum (FBS) (Characterized inFigure S1a-d) for 72 h until the late -log phase was reached. Culture supernatant (2 L per batch) was harvested at an optical density at 600 nm (OD 600nm ) of 1.0-1.4 in an UV-VIS spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). An OD 600nm of 0.05 corresponds to an inoculum of ∼6 × 10 6 CFU/ml as determined by viable count in horse blood agar incubated at 37 - C in the microaerophilic atmosphere for 72 h. To isolate and purify OMVs, the bacterial debris and large contaminants were removed by two centrifugation steps of (8000 × g, 4 - C, 15 min). The supernatants were filtered sequentially using 0.45 and 0.22 μm membrane filters and concentrated using a 10 kDa cut-off centrifugal filter to a volume of ∼30 ml. The purified OMVs were subsequently isolated by SEC using commercially available qEVoriginal/35 nm columns. Fractions #1-30 of 0.5 ml were collected. Finally, fractions 7-11 were pooled and OMVs in these fractions were concentrated again using a 10 kDa cut-off centrifugal filter until a volume of 1 ml. The number of OMV particles was measured by nanoparticle tracking analysis (NTA) using a Zetaview system (Particle Metrix, Germany). Examination of 30 μl OMV and control samples by transmission electron microscopy (TEM) did not indicate the presence of contaminating bacterial cells. The OMV protein concentrations were measured by a nanodrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). The yield of H. pylori 26695 OMVs was ∼1 mg protein (∼2 × 10 12 particles) per 1 L culture supernatant. OMV samples were pooled per batch (2 L per batch) and a total of three batches were collected for this study. Each batch was characterized as shown inFigure S2(a-d). OMV samples were aliquoted and stored in PBS at −80 - C for less than 6 months. As shown inFigure S1(a-d) andFigure S2(a-d), there were still some residual particles in exosome-depleted the FBS. To exclude that the observed effect was due to non-H. pylori derived particles present in the bacterial culture medium, the OMV isolation protocol was in parallel also applied to culture medium. Briefly, same volume (i.e., 2 L per batch) of 10% exosome-depleted FBS containing culture medium was filtered sequentially using 0.45 and 0.22 μm membrane filters. Next, a 10 kDa cut-off centrifugal filter was used to concentrate the samples to a volume of ∼30 ml. The same SEC fractions 7-11 were then collected and concentrated again by using a 10 kDa cut-off centrifugal filter to a volume of 1 ml. These samples were used as control in all our experiments.. TLR activation analysisLipopolysaccharide (LPS) activity levels were measured using the HEK-Blue mTLR4 assay (InvivoGen) according to the manufacturer's protocol. Briefly, HEK-Blue mTLR4 cells were seeded at 25,000 cells per well in a 96-well plate in detection medium. The fractions (#1-30) from SEC were added to the reporter cell line and incubated with HEK-Blue detection medium for 12 h, followed by absorption measurement of the culture medium at 655 nm by iMark Microplate Absorbance Reader (Bio-Rad) and relative TLR4 activation was calculated.S U P P O R T I N G I N F O R M AT I O NAdditional supporting information can be found online in the Supporting Information section at the end of this article.How to cite this article: Xie,
# Introduction
Alzheimer's disease (AD) is a devastating age-related neurodegenerative disorder with an alarming increasing prevalence [bib_ref] Mitochondrial dysfunction caused by outer membrane vesicles from Gram-negative bacteria activates intrinsic..., Cummings [/bib_ref]. It is already well-known for decades that the deposition of amyloid-beta (Aβ) protein in senile plaques outside neurons and the formation of neurofibrillary tangles (NFTs) composed of hyperphosphorylated Tau (p-Tau) protein inside neurons result in the loss of synapses and neurodegeneration which ultimately leads to symptoms associated with AD (DeTure and Dickson, 2019). More recently, it became increasingly clear that also inflammation, both peripheral and central, are important (and often early) events in AD progression [bib_ref] Systemic inflammation impairs microglial Abeta clearance through NLRP3 inflammasome, Tejera [/bib_ref] [bib_ref] Low-grade peripheral inflammation affects brain pathology in the App NL-G-F mouse model..., Xie [/bib_ref]. Except for the recently FDAapproved Aducanumab of which the therapeutic effect is not yet conclusively established [bib_ref] Aducanumab: A new phase in therapeutic development for Alzheimer's disease?, Lalli [/bib_ref] , current Alzheimer's drugs address only disease symptoms [bib_ref] Mitochondrial dysfunction caused by outer membrane vesicles from Gram-negative bacteria activates intrinsic..., Cummings [/bib_ref]. The lack of effective treatments may be attributed by the inadequate understanding of the mechanism underlying AD pathology. Over the past few decades, the three most common proposed hypotheses are the Aβ cascade, the Tau and the neuroinflammation hypothesis. However, the process inducing Aβ peptide deposit, Tau protein degeneration, neuroinflammation and ultimately neuronal cell loss, remains to be elucidated [bib_ref] Alzheimer's disease hypothesis and related therapies, Du [/bib_ref]. Due to the absence of disease-modifying therapeutics and the disturbing multiplying prevalence worldwide, more controversial emerging hypotheses are gaining increased attention. For example, recent epidemiological and experimental evidence provides support for an infectious hypothesis proposing that pathogens may be a common underlying cause of AD [bib_ref] Porphyromonas gingivalis in Alzheimer's disease brains: Evidence for disease causation and treatment..., Dominy [/bib_ref] [bib_ref] Infectious hypothesis of Alzheimer disease, Seaks [/bib_ref]. According to the infectious theory, pathogens or pathogen-derived components are believed to trigger inflammation, which can lead to the accumulation of Aβ and NFTs and neuronal dysfunction, resulting in AD [bib_ref] Infectious hypothesis of Alzheimer disease, Seaks [/bib_ref] [bib_ref] The impact of systemic inflammation on Alzheimer's disease pathology, Xie [/bib_ref].
Outer membrane vesicles (OMVs) are nano-sized particles surrounded by a proteolipid bilayer and are naturally released from Gram-negative bacteria [bib_ref] Outer-membrane vesicles from Gram-negative bacteria: Biogenesis and functions, Schwechheimer [/bib_ref]. They carry various biological molecules from their parental bacteria, including microbe-associated molecular patterns (MAMPs), enzymes and toxins, that can act as virulence factors in the host [bib_ref] The tremendous biomedical potential of bacterial extracellular vesicles, Xie [/bib_ref]. OMVs have gained tremendous interest in a wide range of biomedical fields due to their inherent ability to deliver their cargo both locally and systematically to recipient cells [bib_ref] The tremendous biomedical potential of bacterial extracellular vesicles, Xie [/bib_ref]. This cell-free intercellular communication plays an important role in the physiology and pathogenesis of the bacteria [bib_ref] Host cell interactions of outer membrane vesicle-associated virulence factors of enterohemorrhagic Escherichia..., Bielaszewska [/bib_ref]. For example, in mice colonized with Escherichia coli (E. coli), the E. coli OMVs were reported to spread to a wide range of host tissues, including the heart, liver, kidney, spleen, as well as the brain [bib_ref] Visualizing transfer of microbial biomolecules by outer membrane vesicles in microbe-hostcommunication in..., Bittel [/bib_ref]. As OMVs contain various MAMPs, they can induce pro-inflammatory responses in the periphery but also in the brain by crossing the blood-brain barrier (BBB) [bib_ref] Inflammasome activation by bacterial outer membrane vesicles requires guanylate binding proteins, Fang [/bib_ref] [bib_ref] Delivery of periodontopathogenic extracellular vesicles to brain monocytes and microglial IL-6 promotion..., Ha [/bib_ref] [bib_ref] Extracellular RNAs in periodontopathogenic outer membrane vesicles promote TNF-alpha production in human..., Han [/bib_ref]. Moreover, OMVs from Porphyromonas gingivalis (P. gingivalis) and Paenalcaligenes hominis (P. hominis) cultures, isolated via protein precipitation and sucrose gradient centrifugation, respectively, were shown to be involved in the pathogenesis of AD in mice [bib_ref] Outer membrane vesicles of Porphyromonas gingivalis trigger NLRP3 inflammasome and induce neuroinflammation,..., Gong [/bib_ref]. Administration of these OMVs by oral gavage decreased the tight junctions (TJs) expression (e.g., at the BBB and entered the brain parenchyma, thereby inducing neuroinflammation, tau phosphorylation and memory dysfunction in mice [bib_ref] Outer membrane vesicles of Porphyromonas gingivalis trigger NLRP3 inflammasome and induce neuroinflammation,..., Gong [/bib_ref]. In addition, P. hominis OMVs have also been shown to penetrate the brain via the vagus nerve and cause cognitive impairment . However, the OMV isolation methods used in these studies may also co-purify other components (e.g., proteins, lipids and protein aggregates), which may interfere with the interpretation of these functional studies [bib_ref] The tremendous biomedical potential of bacterial extracellular vesicles, Xie [/bib_ref]. In addition, the organic dyes (DiO and FITC) used to label OMVs may also label co-purified components, resulting in false positive signals in OMV tracking studies [bib_ref] Effective visualization and easy tracking of extracellular vesicles in glioma cells, Mondal [/bib_ref].
Helicobacter pylori (H. pylori) is a pathogen which infects more than half of the world's population and is associated with the development of gastroduodenal diseases. In addition, Numerous studies have shown that H. pylori infection may also increase the risk of various extragastric diseases such as cardiovascular diseases (OR = 1.58, 95% CI: 1.34-1.87) [bib_ref] An updated meta-analysis of the relationship between Helicobacter pylori infection and the..., Tong [/bib_ref] , hypertension (OR = 1.34, 95% CI: 1.10-1.63) [bib_ref] Inflammasome activation by bacterial outer membrane vesicles requires guanylate binding proteins, Fang [/bib_ref] , stroke (OR = 1.43, 95% CI:1.25-1.46) [bib_ref] Association between Helicobacter pylori infection and stroke: a meta-analysis of 273,135 patients, Doheim [/bib_ref] and diabetes (OR = 1.27, 95% CI: 1.11-1.45) [bib_ref] Helicobacter pylori infection as a risk factor for diabetes: a meta-analysis of..., Mansori [/bib_ref] , all of which are risk factors for AD [bib_ref] The impact of systemic inflammation on Alzheimer's disease pathology, Xie [/bib_ref]. A recent meta-analysis also identified direct evidence of a significant positive association between H. pylori infection and AD development (OR = 1.40, 95% CI: 1.12-1.76) [bib_ref] Association of intestinal disorders with Parkinson's disease and Alzheimer's disease: A systematic..., Fu [/bib_ref]. OMVs derived from H. pylori exert immunomodulatory effects by inducing the production of proinflammatory cytokines such as IL-6 and TNF, and by promoting apoptosis of gastric epithelial cells and immune cells. These inflammatory responses may further accelerate the development of H. pylori infection [bib_ref] Helicobacter pylori outer membrane vesicles involvement in the infection development and Helicobacter..., Chmiela [/bib_ref] [bib_ref] Role of outer membrane vesicles from Helicobacter pylori in atherosclerosis, Wang [/bib_ref]. In addition, H. pylori OMVs are also found to induce apoptosis of human umbilical vein endothelial cells, which may promote the formation of atherosclerotic plaques [bib_ref] Role of outer membrane vesicles from Helicobacter pylori in atherosclerosis, Wang [/bib_ref]. However, no study has attempted to determine their role in AD development and progression.
In this study, we investigated the impact of H. pylori-derived OMVs on brain functions and AD pathology using wild-type (WT) mice and App NL-G-F AD mice. The App NL-G-F strain is a second-generation mouse model of AD in which the amyloid precursor protein (APP) is not overexpressed like in most other existing transgenic mouse models. Instead, this model uses a knock-in approach to express human mutated APP at WT. The combined effect of three mutations associated with familial AD results in elevated levels of pathogenic Aβ [bib_ref] APP mouse models for Alzheimer's disease preclinical studies, Sasaguri [/bib_ref]. To study the mechanism by which H. pylori OMVs affect brain functions, WT mice were used as the elevated Aβ production in App NL-G-F mice already impairs brain functions [bib_ref] Low-grade peripheral inflammation affects brain pathology in the App NL-G-F mouse model..., Xie [/bib_ref].
## . electron microscopy
Purified OMVs were visualized by negative staining TEM as described previously . In short, samples were spotted on a parafilm sheet. Next, formvar/C-coated hexagonal copper grids (EMS G200H-Cu), which were glow discharged for 10 s, were placed on top of the droplet for 1 min with the coated side of the grid down. The grids were washed five times in droplets of Milli-Q water, stained with 1% (w/v) uranyl acetate for 10 s and air dried for 24 h before imaging. Visualization of the samples was done using a JEM 1400plus TEM (JEOL, Tokyo, Japan) operating at 80 kV.
## . animals and manipulations
Wild-type C57BL/6J, App NL-G-F (carrying Arctic, Swedish, and Beyreuther/Iberian mutations) [bib_ref] Single app knock-in mouse models of Alzheimer's disease, Saito [/bib_ref] and B6.Cg-Gt(ROSA)Sor tm(CAG-tdTomato)Hze /J (Rosa.tdTomato) reporter mice were bred at a conventional animal facility. Mice were kept in individually ventilated cages under a 14-h dark/10-h light cycle and received food and water ad libitum. Both male and female mice were used. The App NL-G-F mice, C57BL/6J and Rosa.tdTomato mice were 24-28 weeks, 12-16 weeks and 10-12 weeks of age, respectively, at the start of the experiment. The animals were randomly allocated to experimental groups. For short-term experiments, mice were orally gavaged with H. pylori OMVs (20 μg protein, 4 × 10 10 particles) or an equivalent volume of control sample alone or simultaneously i.p. injected with 0.5% DMSO/C3aRA (1 mg/kg) once per day for 5 days and sacrificed at day 7. For long-term experiments, mice were orally gavaged with H. pylori OMVs (20 μg protein, 4 × 10 10 particles) or an equivalent volume of control sample alone or simultaneously i.p. injected with 0.5% DMSO/C3aRA (1 mg/kg) three times per week (Monday, Wednesday and Friday) for 3 weeks. These mice were sacrificed at the end of treatment or cognition tests. All animal studies were conducted in compliance with governmental and EU guidelines for the care and use of laboratory animals and were approved by the ethical committee of the Faculty of Sciences, Ghent University, Belgium (EC2021-025 and EC2022-109).
## . cre-recombinase loading and in vivo biodistribution analysis of h. pylori omvs
For Cre-recombinase loading, H. pylori OMVs and Cre-recombinase were incubated with 0.1 mg/ml saponin at room temperature with shaking (400 rpm) for 2 h. To exclude residual Cre-recombinase, the Cre-recombinase-loaded H. pylori OMV solutions were subjected to qEV isolation. The loading efficiency was validated via Cre western blot analysis and b). The fractions containing OMVs were collected and concentrated using a 10-kDa centrifugal filter. Rosa.tdTomato mice were intragastrically administered Cre-recombinase-loaded H. pylori OMVs (4 × 10 10 particles) once per day for 5 days. At day 7, mice were sacrificed, and the brains were collected and analysed by performing immunostaining as described in Section 2.1. The unloaded H. pylori OMVs were used as control in this experiment.
## . quantification of the gastrointestinal barrier, bbb and blood-csf barrier permeability
Gastrointestinal barrier permeability was determined as previously described [bib_ref] Pro-inflammatory effects of matrix metalloproteinase 7 in acute inflammation, Vandenbroucke [/bib_ref]. Briefly, 4 kDa FITCdextran was intragastrically administered 4 h before collection of blood. Plasma was isolated and gastrointestinal leakage was determined by measuring fluorescence using the FLUOstar Omega reader (BMG LABTECH) at λ ex /λ em = 485/520 nm. BBB and blood-CSF barrier leakage was analysed as described previously [bib_ref] Matrix metalloprotease 8-dependent extracellular matrix cleavage at the blood-CSF barrier contributes to..., Vandenbroucke [/bib_ref]. In brief, 4 kDa FITCdextran was i.v. injected 15 min before CSF isolation by cisterna magna puncture. For blood-CSF barrier analysis, 1 μl of CSF was diluted in 99 μl PBS and fluorescence was measured using the FLUOstar Omega reader (λ ex /λ em = 485/520 nm). For the analysis of BBB integrity, mice were transcardially perfused with PBS/heparin and brain was isolated. Next, perfused brain samples were incubated in formamide overnight, samples were centrifuged for 15 min at 20,000 g in 4 - C, and 100 μl supernatant was used for fluorescence measurement using the FLUOstar Omega reader (λ ex /λ em = 485/520 nm).
## . western blotting
Hippocampus and bacterial extracts were prepared in 0.5% CHAPS buffer containing complete protease inhibitor and centrifuged for 15 min at 20,000 g and 4 - C, and protein concentration was determined using a Pierce BCA Protein Assay Kit (Thermo Fisher Scientific). Hippocampus and bacterial extraction supernatants, OMV and FBS samples were denatured in 6×Laemmli buffer, separated by SDS-PAGE gel electrophoresis, and transferred to a nitrocellulose membrane. Following blocking with Odyssey blocking buffer (LI-COR Biosciences), the membrane was incubated with primary antibodies, washed with PBS containing 0.1% Tween 20 and subsequently incubated with fluorophore-or HRP-conjugated secondary antibodies. Protein bands were visualized by Odyssey Fc Imaging System (LI-COR Biosciences) or WesternBright Quantum HRP substrate (advansta) in Amersham Imager 600 (GE Healthcare), and quantification was done in Image Studio (LI-COR Biosciences) and Amersham Imager 600 integrated analysis software (GE Healthcare), respectively.
## . cytokine measurements
Hippocampus and stomach lysates were prepared as described in Section 2.8 and IL-1β, and TNF levels of plasma and tissue lysates were measured using the Bio-Plex cytokine assay (Bio-Rad) according to the manufacturer's instructions.
## . rna extraction and rt-qpcr analysis
Total RNA was isolated from the tissue using TRIzol reagent. After homogenizing the tissue in a tube containing zirconium oxide beads on a TissueLyser (QIAGEN), chloroform was added and the homogenate was separated into three phases by centrifugation at 20,000 g in a microcentrifuge at 4 - C for 15 min. Total RNA was extracted from the upper aqueous phase using Aurum total RNA kit (Bio-Rad) according to the manufacturer's instructions. The concentration of total RNA was determined by the Nanodrop and total RNA was reverse-transcribed into cDNA with SensiFAST™ cDNA Synthesis Kit (Bioline). qPCR was performed on the Roche LightCycler 480 System (Applied Biosystems) using SensiFAST™ SYBR ® No-ROX Kit (Bioline). Results are given as relative expression values normalized to the geometric mean of reference genes, determined using GeNorm. The sequences of the primers are depicted in .
## . aβ elisa
Aβ was extracted and measured using a standard protocol as described previously [bib_ref] Counteracting the effects of TNF receptor-1 has therapeutic potential in Alzheimer's disease, Steeland [/bib_ref]. Specifically, hippocampus tissue was homogenized in Tissue Protein Extraction Buffer containing complete protease and phosphatase inhibitors using a Precellys (Bertin Technologies) and subsequently centrifuged at 5000 g for 5 min at 4 - C. Supernatant was collected and centrifuged at 4 - C for 1 h at 100,000 g (TLA-100 Rotor; Beckman Coulter). Supernatant containing soluble Aβ was collected and the pellet was further processed in GuHCl solution containing complete protease inhibitor, sonicated, vortexed, incubated for 60 min at 25 - C and centrifuged at 70,000 g for 20 min at 4 - C. Supernatant containing insoluble Aβ was 12 times diluted with GuHCl diluent. The levels of Aβ 1-42 and Aβ 1-40 in hippocampus lysates were determined using a sandwich ELISA assay. Briefly, the sample-detection mixtures were added to the ELISA plate coated with anti-Aβ 1-42 (1.5 μg/ml; JRF/cAb 42/26 ) or anti-Aβ 1-40 antibody (1.5 μg/ml; JRF/cAb 40/28 ) and incubated overnight at 4 - C with slow shaking. Absorption at 450 nm was measured after adding substrate solution (BD Biosciences OptEIA™) followed by stopping buffer (1 M H 2 SO 4 ). The amount of Aβ was determined with GraphPad Prism 8.0 using a nonlinear regression model.
## . histology
To evaluate gastric inflammation, 5 μm paraffin sections of the stomach were stained with haematoxylin (Merck) and eosin (VWR) (H&E). Scoring of immune cell infiltration in H&E-stained stomach was done blinded. Images (20× magnification) were acquired with a BX51 discussion microscope (Olympus) using an objective lens with 3 N.A. 1.0. The images were analysed in Image J software.
## . immunohistochemistry
For immunostainings on the mouse brain, the sections (one section per mouse for each staining) were cut depending on the used processing method, namely, 5 μm for paraffin, 20 μm for cryo, and 50 μm for vibratome sections. Paraffin sections were de-paraffinized in xylene and ethanol, boiled in citrate buffer for 20 min, and followed by blocking with 5% goat serum in PBS-T (PBS containing 0.3% Triton X-100) solution for 1 h at room temperature. The sections were then stained with primary antibodies (anti-GFAP, anti-IBA1, anti-Ki-67, anti-C3, anti-C3aR, anti-DCX, anti-PSD-95, anti-SYP, anti-NeuN, anti-OCLN, anti-E-cadherin (CDH1), and anti-RFP) in blocking buffer at 4 - C overnight. After washing with PBS, sections were stained with appropriate fluorophore-conjugated secondary antibodies in PBS or PBS containing 0.1% Triton X-100 for 1-2 h before washing and mounting. Cryosections for anti-ZO-1 staining were boiled directly in citrate buffer and followed the same steps as above for paraffin staining. Vibratome sections for anti-GFAP, anti-IBA1, anti-C3, anti-CD68, anti-PSD-95 and anti-SYP were treated with blocking buffer directly and followed the same steps as above for paraffin staining. A Zeiss LSM780 confocal microscope or Zeiss Axioscan Z.1 was used for imaging. Images were processed using Image J and the intensity of overlapping signals were quantified with Colocalization analysis for Image J.
## . glial cell morphology quantification
To quantify the glial cell morphology, IBA1-positive microglia and GFAP-positive astrocytes were imaged with a 40×/63× oil objective using the confocal microscope with the z-stack model. Images were analysed using filament tracing algorithm from Imaris software (Bitplane).
## . synaptic imaging and quantification
For synaptic puncta colocalization analysis, brain paraffin sections were co-immunostained with the anti-SYP and anti-PSD-95 antibodies and imaged with the 63× oil objective with the 3× zoom using confocal microscopy. Images were processed using Image J and the number of colocalized puncta was quantified with the Synapse Counter plugin for Image J [bib_ref] Colocalization of synapse marker proteins evaluated by STED-microscopy reveals patterns of neuronal..., Dzyubenko [/bib_ref]. For quantification of PSD-95 and SYP inside CD68-positive phagosomes and IBA-positive microglia, brain vibratome sections were co-immunostained with anti-IBA1, anti-CD68, and anti-SYP or anti-PSD-95 antibodies. Sections were imaged with the 40×/63× oil objective using confocal microscope with the z-stack model. Images were analysed using the surface function from Imaris software.
## . electrophysiological field recordings on acute mouse hippocampal slices
Electrophysiology was performed at the VIB-KU Leuven Center for Brain and Disease Research Electrophysiology Expertise Unit as described previously [bib_ref] Lowering synaptogyrin-3 expression rescues Tau-induced memory defects and synaptic loss in the..., Largo-Barrientos [/bib_ref]. Specifically, mice were anesthetized with isoflurane and brains were isolated into ice-cold cutting solution (87 mM NaCl, 2.5 mM KCl, 1.25 mM NaH 2 PO 4 , 10 mM glucose, 25 mM NaHCO 3 , 0.5 mM CaCl 2 , 7 mM MgCl 2 , 75 mM sucrose, 1 mM kynurenic acid, 5 mM ascorbic acid and 3 mM pyruvic acid) at pH 7.4 and 5% CO 2 /95% O 2 . 300 μm-thick parasagittal hippocampal sections were cut in cold cutting solution using a vibratome (VT1200 Leica), incubated at 34 - C for 35 min to recover and then stored in cutting solution at room temperature. For recordings, slices were placed onto a multielectrode array (MEA 2100, Multichannel Systems) and continuously perfused with artificial cerebrospinal fluid (aCSF) solution (119 mM NaCl, 2.5 mM KCl, 1 mM NaH 2 PO 4 , 11 mM glucose, 26 mM NaHCO 3 , 4 mM MgCl 2 and 4 mM CaCl 2 ) at pH 7.4 and 5% CO 2 /95% O 2 . Field excitatory post-synaptic potentials (fEPSPs) were recorded from mossy fiber-CA3 synapses or Schaffer collateral-CA1 synapses by stimulating and recording from the appropriate electrodes. First, inputoutput curves were established for each individual slice by applying single-stimuli ranging from 500 to 2750 mV with 250 mV increments. Stimulus strength that corresponds to 35% of maximal response in the input-output curve was used for the consecutive recordings. For long-term potentiation (LTP) experiments, stable fEPSPs were recorded for 30 min to establish a baseline. Afterwards, three trains of high-frequency stimulation (100 stimuli at 100 Hz) with 5 min intervals were applied to induce LTP and fEPSPs were recorded for 65 additional minutes. Post-LTP fEPSPs were measured every 5 min (average of three consecutive stimulations (15 s apart)). Recordings were analysed and processed using Multi Channel Experimenter software (Multichannel Systems).
## . y-maze test
The Y-maze test was performed with a slight modification as previously described [bib_ref] Single app knock-in mouse models of Alzheimer's disease, Saito [/bib_ref]. All the experiments were conducted in the light phase (9:00-18:00). Y maze apparatus (O'Hara & Co), made of grey plastic, consisted of three compartments (3 cm (W) bottom and 10 cm (W) top, 40 cm (L) and 12 cm (H)) radiating out from the centre platform (3 × 3 × 3 cm triangle), and positioned 60 cm above the floor. In this test, each mouse was placed in the centre of the maze facing toward one of the arms and then allowed to explore freely for 5 min. We recorded and analysed the spontaneous behavioural alternations of the mice using EthoVision tracking system (Noldus). A reduction in the behavioural alternations corresponds to memory impairment [bib_ref] Impaired hippocampal-dependent learning and functional abnormalities in the hippocampus in mice lacking..., Sarnyai [/bib_ref].
## . novel object recognition test
The NOR test was performed with a minor modification as previously described . All the experiments were conducted in the light phase (9:00-18:00). The mice were placed in a clear, acrylic box (40 × 40 × 40 cm). During the training trial, mice were given two sessions of 5 min; during the first session, mice were allowed to explore the arena without objects (habituation), and in the second session, two identical objects were placed at two opposite positions within the box at the same distance from the nearest corner. The mice were returned to their home cage for 15 min (short-term memory) after the training phase. Mice were placed back in the same box for 5 min after replacing one of the familiar objects with a novel one. Exploration of the objects by the mice was defined as touching the object with their nose or mouth. All trials were videotaped and scored manually using EthoVision tracking system (Noldus). The percentage preference for the novel object was calculated using the following formula: percentage novel object preference = (novel object exploration time/(novel object exploration time + familiar object exploration time)) × 100%. A value >50% indicates preference for the novel object. To exclude the existence of olfactory cues, the box and the objects were thoroughly cleaned with 20% ethanol after each trial.
# statistical analysis
Graphical data were analysed with GraphPad Prism 8.0 and presented as mean ± SEM. Unpaired, two-tailed Student's t-tests were used to determine statistical significance between two independent groups. One-way analysis of variance (ANOVA) with Bonferroni's multiple comparison test was used to determine statistical significance among multiple groups with one independent variable. Two-way ANOVA with Bonferroni's multiple comparison test was used to determine statistical significance among multiple groups with two independent variables. A p value of <0.05 was considered statistically significant.
# results
## . isolation and characterization of h. pylori omvs
H. pylori OMVs were purified from bacterial cultures by a combination of ultrafiltration and size exclusion chromatography (SEC) . After SEC isolation, the particle number and protein concentration of each fraction was determined by NTA and spectrophotometric analysis, respectively . LPS is the major outer membrane component of Gram-negative bacteria and one of the most abundant components of OMVs [bib_ref] Bacterial outer membrane vesicles mediate cytosolic localization of LPS and caspase-11 activation, Vanaja [/bib_ref]. TLR4 activation analysis demonstrated a noticeable increased LPS activity in the particle enriched fractions . Next, the particle-enriched fractions were pooled and subjected to OMV preparation by ultrafiltration using 10 kDa centrifugal filters . The resulting purified H. pylori OMVs were physically intact when observed under TEM using negative staining, with an average size of 176.1 ± 4.7 nm (two peaks centred at 135 and 225 nm) in diameter as determined by NTA . H. pylori OMV purification was further confirmed by enrichment of flagellin A (FlaA), the major component of flagellar filaments that is also present in OMVs [bib_ref] Helicobacter pylori growth stage determines the size, protein composition, and preferential cargo..., Zavan [/bib_ref] , in the purified OMV preparation and the OMV fractions #7-11 .
## . biodistribution of h. pylori omvs
To examine whether gastric H. pylori-derived OMVs can reach and deliver their content to the periphery, we loaded H. pylori OMVs with Cre-recombinase , which were consequently administered to Rosa.tdTomato mice by oral gavage, as depicted treatment in the schedule in . In this mouse model, a loxP-flanked STOP cassette followed by a downstream red fluorescent protein variant (tdTomato) is inserted in the ROSA locus. The loxP-flanked STOP cassette prevents transcription of the tdTomato. However, the Cre-recombinase enzyme can delete the STOP cassette, leading to robust tdTomato expression [bib_ref] A robust and high-throughput cre reporting and characterization system for the whole..., Madisen [/bib_ref]. By loading the OMVs with Cre-recombinase enzyme, the cells to which the OMV delivered their cargo, namely, the Cre-recombinase, will turn red due to the removal of the STOP cassette and the subsequent expression of the tdTomato fluorescent protein. Next, H. pylori OMV biodistribution and Cre-recombinase delivery were tracked by detection of the tdTomato expression in the stomach, colon, lung, heart, liver, kidney, spleen, and brain. A clear TdTomato signal was observed in the stomach, while absent in colon sections at the end of the gastrointestinal tract. Also, spleen, lung and heart did not show a pronounced increase in TdTomato + cells. In contrast, we did observe Cre-mediated recombination in the liver, kidney and brain (e.g., hippocampus and cortex) and . Co-stainings revealed that the majority of tdTomato + cells in the brain are astrocytes as visualized using the astrocyte specific marker GFAP but also a small amount of tdTomato + NeuN + double-positive neurons could be observed ,e and . In the stomach, tdTomato signal was detected in CD45 + cells . These data implicate that H. pylori OMVs can travel across various biological barriers and eventually even migrate to the brain. To investigate if the H. pylori OMV translocation to the brain is caused by OMV-induced disruption of the gastrointestinal barrier and central nervous system (CNS) barriers, WT mice were followed the same treatment schedule and the integrity of gastrointestinal barrier, BBB and blood-CSF barrier was analysed using a 4 kDa FITC-conjugated dextran tracer . The comparable FITC signal indicated that the H. pylori OMVs treatment had no apparent effect on the integrity of these barriers and . In agreement with this, we did not observe differences in mRNA levels of the TJs Zo and Ocln in hippocampus and choroid plexus upon H. pylori OMV treatment . Also, the protein localization of ZO-1 and OCLN in hippocampal endothelial and choroid plexus epithelial cells showed comparable near-continuous staining at the apical cell border upon H. pylori OMV oral gavage as seen in the control conditions . In the stomach, we even detected higher mRNA expression Ocln and Cdh, and increased protein expression of CDH1 in H. pylori OMVs-treated mice compared to the control mice . Collectively, these results indicate that H. pylori OMVs do not disrupt the gastrointestinal and blood-brain interfaces but may actively cross brain barriers through transcellular pathways.
## . h. pylori omvs promote ad pathology in app nl-g-f mice
To examine the effect of H. pylori OMVs on AD pathology, the OMVs were administered to App NL-G-F mice by oral gavage following the treatment scheme shown in . While OMV treatment for 1 week did not induce significant changes in Aβ pathology between H. pylori OMV-treated and control App NL-G-F mice based on the Aβ immunostaining , 3 weeks of H. pylori OMV treatment did have pronounced effects . More specifically, we observed that the amount of Aβ plaques significantly increased in the hippocampus of App NL-G-F mice upon H. pylori OMV treatment compared to control treated mice . Aβ plaque load was mostly increased in the CA1 and CA2/3 regions, but not in the dentate gyrus (DG) . In agreement with the Aβ plaque analysis, Aβ ELISA also showed a significant increase in both soluble and insoluble Aβ 1-40 and Aβ 1-42 in the hippocampus of H. pylori OMV-treated App NL-G-F mice . Additionally, morphometric analysis of Aβ stained brain sections revealed an increase in both medium (<250-600 μm 2 ) and large (>600 μm 2 ) plaques in the hippocampus of H. pylori OMV-treated App NL-G-F mice compared to control treated mice. In contrast, no difference in the number of small plaques (<250 μm 2 ) could be observed . Microglia are brain resident macrophages that play an important role in mediating Aβ degradation and clearance [bib_ref] Systemic inflammation impairs microglial Abeta clearance through NLRP3 inflammasome, Tejera [/bib_ref]. We observed a decrease in microglial internalized Aβ and a lower number of microglia, labelled via the microglial marker IBA1 adjacent to the large plaques in App NL-G-F mice upon OMV oral gavage . Consistent with our Aβ size distribution analysis , we observed that Aβ plaques were diffuse in App NL-G-F control mice, whereas they were centrally localized in H. pylori OMV-treated App NL-G-F mice . Taken together, these findings suggest that H. pylori (j) Plaques were divided into small (< 250 μm 2 ), medium (250-600 μm 2 ), and large (> 600 μm 2 ) plaques, and the number of microglia per plaque was quantified. 3-5 plaques are analysed per mouse and 5 mice per group. The graphs are shown as the mean ± SEM and the datapoints are biological replicates. Images are representative for 5 (b and h) biological replicates. Statistical significance was determined by two-tailed Student's t-test. *p < 0.05, **p < 0.01, ***p < 0.001. OMV treatment accelerated the typical hallmark of AD pathology, namely, Aβ plaque formation, which might be due to impaired microglial Aβ clearance.
## . h. pylori omvs activate c-car signalling in wt and app nl-g-f mice
As OMVs contain a certain amount of LPS [bib_ref] Bacterial outer membrane vesicles mediate cytosolic localization of LPS and caspase-11 activation, Vanaja [/bib_ref] and as our previous study showed that LPS-induced systemic inflammation affects brain inflammation and AD pathology [bib_ref] Low-grade peripheral inflammation affects brain pathology in the App NL-G-F mouse model..., Xie [/bib_ref] , we next evaluated the impact of orally gavaged H. pylori OMVs on peripheral and central inflammation in WT mice . We observed an increased infiltration of mononuclear and polymorphonuclear cells in the stomach submucosa of H. pylori OMV-treated mice and a significant increased gastric inflammation score based on the Updated Sydney System . On the contrary, the proinflammatory cytokines interleukin (IL-)1β and tumour necrosis factor (TNF) showed no differences neither on mRNA level nor on protein level in the stomach. Plasma levels of IL-1β, but not of TNF, were elevated , while Ilβ and Tnf gene expression in the hippocampus was not increased in H. pylori OMV-treated mice compared to control mice . These results show that H. pylori OMVs do not induce severe peripheral and central inflammation in WT mice. H. pylori is complement sensitive and activates mainly the classic complement pathway [bib_ref] Complement activation directly induced by Helicobacter pylori, Berstad [/bib_ref]. Indeed, we found that genes involved in the classical pathway activation (C1q through C3) are significantly upregulated in WT mice treated with H. pylori OMVs while genes more specific to the lectin, alternative, or terminal pathways of the complement are not affected ,c and . This was accompanied by an enhanced C3 protein expression and a higher degree of colocalization of C3 with GFAP positive cells in the hippocampus of H. pylori OMV-treated mice compared to WT controls . Also, western blot analysis showed an increased trend of C3 expression in the hippocampus of H. pylori OMV-treated WT mice .
In contrast, C3aR expression was unaltered on protein and mRNA levels in hippocampus tissues of WT mice . In addition, co-staining documented a high degree of colocalization of C3aR and IBA1 in these mice . Next, we studied C3-C3aR signalling in App NL-G-F mice after 3 weeks of H. pylori OMVs treatment. In contrast to the WT mice, both the mRNA and protein levels of C3 showed a significant increase in the hippocampus of H. pylori OMV-treated App NL-G-F mice compared to controls as shown by qPCR and immunostaining , while only the protein expression level of C3aR showed a trend towards upregulation after H. pylori OMV treatment in App NL-G-F mice . Co-immunostaining of C3, Aβ and GFAP also showed a high degree of colocalization of C3 and Aβ/GFAP . In addition, a quantitative increase of C3 intensity in this immunostaining was consistently detected in H. pylori OMV-treated App NL-G-F mice . Although the H. pylori OMV treatment did not affect the C3aR expression in WT mice , its protein level was upregulated in H. pylori OMV-treated App NL-G-F mice . Similarly, co-staining C3aR and IBA1 in the hippocampus showed colocalization and an increased trend of C3aR expression in H. pylori OMV-treated mice compared to controls . Altogether, these results indicate H. pylori OMV treatment can induce or promote C3 activation.
## . h. pylori omvs increase glial reactivity via c-car signalling in wt mice
Astrocytes and microglia are the major glial cells in the CNS that maintain and support homeostasis of the neuronal functions [bib_ref] Astrocytes and microglia: In sickness and in health, Vainchtein [/bib_ref]. C3-C3aR signalling plays a prominent role in immune regulation in the CNS by mediating glial reactivity [bib_ref] Complement C3aR inactivation attenuates tau pathology and reverses an immune network deregulated..., Litvinchuk [/bib_ref]. Next, we assessed the effect of H. pylori OMV treatment on glial reactivity and the role of the increased C3-C3aR signalling in this process. The WT mice were simultaneously treated with H. pylori OMVs and/or C3aRA via oral gavage and intraperitoneal injection, respectively . Although no significant astrogliosis was observed in the whole hippocampus of WT mice after H. pylori OMV treatment , a significant increase in GFAP signal was observed specifically in the hippocampal CA3 region . Increased expression of GFAP represents astroglial activation and/or astrogliosis [bib_ref] Induction of glial fibrillary acidic protein expression in astrocytes by nitric oxide, Brahmachari [/bib_ref]. Moreover, quantitative morphometric 3D analysis of GFAP + astrocytes in the CA3 region showed that H. pylori OMV-treated WT mice have reactive astrocytes, characterized by significantly longer processes and increased number of segments, branching, and terminal points relative to astrocytes from control treated WT mice . Indeed, blocking C3 signalling using C3aRA treatment did not block the H. pylori OMV effects on astrogliosis and astrocytic reactivity -e and . In parallel, we observed a significant decrease in the number of IBA + microglia in the CA3 hippocampal region and the whole hippocampus of H. pylori OMV-treated WT mice compared to control treated WT mice ,g and . However, no differences were visible in these mice on mRNA expression of microgliosis marker Aif or total Ki-67 + /Ki-67 + GFAP + /Ki-67 + IBA1 + cells . Quantitative analysis of microglial morphology showed decreased segments, branching, and terminal points in H. pylori OMV-treated WT mice, and these phenotypic effects were blocked by C3aRA treatment . To further characterize the microglial activation, we performed IBA1 and CD68 co-staining to identify CD68 + phagocytic microglia . Quantification of CD68 immunoreactivity revealed that H. pylori OMV treatment in WT mice significantly increased CD68 expression in microglia and that C3aRA treatment prevented the H. pylori OMV-induced increase . Given the specific expression of C3 and C3aR in astrocytes and microglia, respectively [bib_ref] Complement C3aR inactivation attenuates tau pathology and reverses an immune network deregulated..., Litvinchuk [/bib_ref] , we investigated whether there is an astrocyte-microglia interaction via C3-C3aR signalling after H. pylori OMV administration. We show a remarkable increase in IBA1 + microglial cells that overlap with GFAP + astrocytes in the hippocampus of H. pylori OMV-treated WT mice compared to the control treated mice and C3aRA/H. pylori OMV-treated WT mice . 3D reconstruction reveals a close spatial interaction of astrocytes, microglia and C3, suggesting the involvement of C3-C3aR signalling in this glial cell interaction .
## . h. pylori omvs induce neuronal dysfunction via c-car signalling in wt mice
We further analysed the effects of H. pylori OMV treatment on synaptic and neuronal properties and the role of C3-C3aR signalling in this process . Indeed, a reduced immunostaining intensity of presynaptic protein synaptophysin (SYP) and (i) Quantification of the percentage of PSD-95 + volume inside CD68 + phagosome (upper) and IBA1 + microglia (lower). (j) Schematic diagram indicating the general experimental set-up of LTP measurement. (k-l) fEPSP amplitude recordings over time (k) and the average of the fEPSP amplitude recordings (l) after LTP induction in CA3 mossy fiber synapses. The graphs are shown as the mean ± SEM and the datapoints are biological replicates. Images are representative for 4 or 5 (b and d) biological replicates. Statistical significance was determined by two-way ANOVA Bonferroni's multiple comparisons. *p < 0.05, **p < 0.01 postsynaptic density protein 95 (PSD-95) was observed in the CA3 area of the hippocampus of H. pylori OMV-treated WT mice compared to WT controls . Moreover, C3aRA treatment on H. pylori OMV-treated WT mice resulted in a significant elevation of PSD-95 immunoreactivity . Similarly, high-resolution images of co-immunostaining of SYP and PSD-95 revealed reduced pre-, post-, and double-positive synaptic puncta in H. pylori OMV-treated WT mice, while C3aRA treatment did prevent to a certain level (not significant trend) synapse loss . These results were further confirmed by immunoblot analysis in the hippocampus using anti-SYP and anti-PSD-95 antibodies . These findings suggest that H. pylori OMVs are potent inducers of synaptic deficits and C3-C3aR signalling plays a crucial role in this process. As our earlier analysis showed a higher intensity of CD68 + phagocytic microglia in H. pylori OMV-treated WT mice , the reduced synaptic puncta were seen in H. pylori OMV-treated WT mice may be due to increased microglial engulfment. We found that the volume of PSD-95-positive or SYP-positive puncta within IBA1 + microglia are increased in H. pylori OMV-treated WT mice and this effect was largely prevented by simultaneous C3aRA treatment. As the excessive synaptic pruning may result in imbalanced synaptic transmission and subsequent synaptic dysfunction, we evaluated the contribution of H. pylori OMVs to the synaptic function of mossy fiber-CA3 synapses by assessing the electrophysiological properties of these synapses in acute hippocampal slices of WT mice . The LTP induction was significantly decreased in H. pylori OMV-treated WT mice compared to control treated WT mice, and this phenotype was substantially improved when combined with C3aRA treatment ,l and . Furthermore, we tested whether this effect could also be observed in other neuronal networks, and recorded LTP of Schaffer collateral-CA1 synapses . The LTP recordings also showed Schaffer collateral-CA1 synapses were affected by H. pylori OMV treatment, but not as pronounced as mossy fiber-CA3 synapses . When comparing the efficacy of C3aRA treatment in CA1-and CA3-pathways, similar protective effects against H. pylori OMV-induced synaptic defects could be observed ,l and . These findings suggest that the detrimental effect of H. pylori OMV on synaptic plasticity is at least in part due to increased synapse elimination by activated microglia.
The abnormal activation of microglia also actively participates in the phagocytosis of live neurons [bib_ref] Primary phagocytosis of viable neurons by microglia activated with LPS or Abeta..., Fricker [/bib_ref]. In this context, we performed NeuN staining and the quantification showed reduced neuronal numbers in the CA3 region of the hippocampus in H. pylori OMV-treated WT mice . C3aRA treatment in H. pylori OMV-treated WT mice resulted in a powerful blockade of neurodegeneration . The reduced neuronal density in the hippocampus by H. pylori OMV treatment may at least partly be due to reduced proliferation of neuronal progenitor cells as demonstrated by DCX and Ki-67 immunostaining . In addition, incorrect microglia-neuron interactions are known to lead to microglial phagocytosis of live neurons and excessive neuronal loss [bib_ref] Induction of glial fibrillary acidic protein expression in astrocytes by nitric oxide, Brahmachari [/bib_ref]. We then quantified the CD68 + phagosome in the CA3 stratum pyramidale and revealed an increased CD68 immunoreactivity in H. pylori OMV-treated WT mice. Again, C3aRA treatment prevented this OMV-induced increase .
Combined, these data demonstrate that H. pylori OMV can induce synaptic deficits and neurodegeneration, and that enhanced microglial activation phagocytosis may contribute to this process through C3-C3aR signalling.
## . blocking c-car signalling attenuates ad pathology and memory loss induced by h. pylori omvs in app nl-g-f mice
Finally, we investigated the effect of H. pylori OMVs on memory functions and the potential of blockage of C3-C3aR signalling in preventing memory impairments. Therefore, App NL-G-F mice were simultaneously treated with H. pylori OMVs and/or C3aRA via oral gavage and intraperitoneal injection, respectively . After 3 weeks of treatment, we examined the effects on cognitive function using Y-maze and novel object recognition (NOR) tests. The H. pylori OMV-treated App NL-G-F mice showed cognitive impairment based on significant decreased spontaneous alternation in the Y-maze task and performance to the new object in the NOR test . Interestingly, C3aRA treatment prevented the occurrence of cognitive impairment caused by H. pylori OMVs in App NL-G-F , as shown in the NOR test, while this was less pronounced in the Y-maze . Consistent with the cognitive results, H. pylori OMV treatment worsen Aβ pathology and again this effect was largely blocked by C3aRA treatment . In line with this, C3aRA treatment protected the microglial Aβ clearance from the influence of H. pylori OMVs, and showed the same levels of microglia internalized Aβ as App NL-G-F controls . Lastly, quantification of microglia around Aβ plaques also demonstrated that C3aRA treatment inhibited H. pylori OMV-induced clustering decrease in App NL-G-F mice compared with control App NL-G-F mice . In addition, LTP recordings revealed that Schaffer collateral-CA1 synapses in App NL-G-F mice were significantly impaired upon oral H. pylori OMV administration and this effect was blocked by C3aRA treatment -n and . Together, these findings suggest that H. pylori OMVs cause cognitive impairment in App NL-G-F mice and that by blocking the C3-C3aR signalling AD pathology and memory loss induced by H. pylori OMV can be attenuated. and the average of the fEPSP amplitude recordings (n) after LTP induction in CA1 Schaffer collateral synapses. The graphs are shown as the mean ± SEM and the datapoints are biological replicates. Images are representative for 5 (c, e and i) or 3-4 (m and n) biological replicates. Statistical significance was determined by two-way ANOVA Bonferroni's multiple comparisons. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.
# discussion
The gut microbiota is a complex community of microorganisms that resides in our gastrointestinal ecosystem and of which a disbalance not only implies various gut disorders but also CNS disorders. Growing evidence supports the concept that the gut microbiota affects host brain function including cognitive behaviour and that a microbiota dysbiosis may mediate or affect AD pathogenesis [bib_ref] Alzheimer's disease and gut microbiota, Hu [/bib_ref]. However, the mechanisms of this microbiota-gut-brain axis are not yet fully understood. It is well established that microbiota-derived molecules such as LPS, peptidoglycan, and short-chain fatty acids directly modulate neuroinflammation and Aβ accumulation [bib_ref] Systemic immune challenges trigger and drive Alzheimer-like neuropathology in mice, Krstic [/bib_ref] [bib_ref] Low-grade peripheral inflammation affects brain pathology in the App NL-G-F mouse model..., Xie [/bib_ref]. Several of these bioactive molecules are also present in microbiota-derived OMVs, that is, nano-sized spherical buds of the outer membrane produced by both pathogenic and non-pathogenic Gram-negative bacteria [bib_ref] The tremendous biomedical potential of bacterial extracellular vesicles, Xie [/bib_ref]. OMVs can transport their cargo to proximal and distal cells to exert physiological and pathological effects [bib_ref] Inflammasome activation by bacterial outer membrane vesicles requires guanylate binding proteins, Fang [/bib_ref] [bib_ref] Delivery of periodontopathogenic extracellular vesicles to brain monocytes and microglial IL-6 promotion..., Ha [/bib_ref] [bib_ref] Extracellular RNAs in periodontopathogenic outer membrane vesicles promote TNF-alpha production in human..., Han [/bib_ref]. Recently, microbial-derived OMVs have also been demonstrated to be involved in several pathologies such as mitochondrial dysfunction [bib_ref] Mitochondrial dysfunction caused by outer membrane vesicles from Gram-negative bacteria activates intrinsic..., Cummings [/bib_ref] [bib_ref] Outer membrane vesicles from Neisseria gonorrhoeae target PorB to mitochondria and induce..., Deo [/bib_ref] , systemic bone loss [bib_ref] Extracellular vesicles derived from the periodontal pathogen Filifactor alocis induce systemic bone..., Kim [/bib_ref] , pulmonary fibrosis [bib_ref] Dysregulated lung commensal bacteria drive interleukin-17B production to promote pulmonary fibrosis through..., Yang [/bib_ref] and cognitive impairment [bib_ref] Outer membrane vesicles enhance tau phosphorylation and contribute to cognitive impairment, Wei [/bib_ref].
H. pylori is a major human pathogen that induces inflammation of the stomach and is etiologically related to various gastric and extragastric diseases, like AD. Next to the induction of inflammation, H. pylori infection can also increase the production of AD risks, such as reactive oxygen species, lipid peroxides and homocysteine. Here, we investigated whether of H. pylori-derived OMVs might be key players in the previously reported H. pylori effects on AD pathology. Our results indicate that OMVs can travel from the gut to the brain and eventually affect brain functions. Notably, we observed that both the gastrointestinal and brain barriers remained intact after H. pylori OMV administration, suggesting that H. pylori OMVs may enter the brain via a transcellular pathway crossing the barriers or indirectly via the vagus nerve, as was shown for P. hominis OMVs . In addition, we could not rule out the possibility that cargo released from disrupted OMVs might somehow reach brain barriers and eventually enter the brain. The intact barriers may limit the entry of the OMVs into the bloodstream and the brain, which may explain the lack of very pronounced peripheral and central inflammation that we observed in our study. In addition, a recent study reported that H. pylori OMVs induced approximately 5-fold higher production of pro-inflammatory cytokines in the spleen upon intraperitoneal injection compared to oral administration [bib_ref] Helicobacter pylori-derived extracellular vesicles increased in the gastric juices of gastric adenocarcinoma..., Choi [/bib_ref] , indicating a positive correlation between peripheral inflammation levels and OMV amount in peripheral tissues. Taken together, peripheral inflammation is probably not the underlying mechanism by which H. pylori OMVs potentially affect brain function.
Our study reveals that H. pylori OMVs promote Aβ plaque formation in App NL-G-F mice. Interestingly, the effects on Aβ pathology can only be detected after 3 weeks of oral gavage with H. pylori OMVs, suggesting that a chronic more robust H. pylori infection of the gut is needed, while a transient, acute infection does not affect the brain [bib_ref] Alzheimer's disease and Helicobacter pylori infection: Inflammation from stomach to brain, Albaret [/bib_ref]. Of note, the Aβ plaques are most enriched in the CA1 and especially in the CA2/3 regions of the hippocampus upon H. pylori OMV administration. This difference in Aβ plaque distribution is probably due to the distribution of APP-expressing interneurons in the different hippocampal regions as demonstrated by [bib_ref] Contribution of GABAergic interneurons to amyloid-beta plaque pathology in an APP knock-in..., Rice [/bib_ref]. In addition to the amount of Aβ aggregation, the H. pylori OMV treatment also increased the compactness and surface area of Aβ plaques, important pathological features of AD [bib_ref] TREM2 haplodeficiency in mice and humans impairs the microglia barrier function leading..., Yuan [/bib_ref].
Astrocytes and microglia are resident cells of the CNS, and they contribute not only to the formation of the BBB but also to the homeostasis and immune functions of the CNS [bib_ref] Glial influence on the blood brain barrier, Alvarez [/bib_ref] [bib_ref] Astrocytes and microglia: In sickness and in health, Vainchtein [/bib_ref]. Recent evidence has shown that both cell types can take up and internalize microbial-derived OMVs and induce an inflammatory response [bib_ref] Visualizing transfer of microbial biomolecules by outer membrane vesicles in microbe-hostcommunication in..., Bittel [/bib_ref] [bib_ref] Outer membrane vesicles enhance tau phosphorylation and contribute to cognitive impairment, Wei [/bib_ref]. In line with this observation, we detected H. pylori OMV uptake by astrocytes and induction of activation of glial cells, reflected by the increase (astrocytes) and reduction (microglia) in dendrite length, the number of segments, branch points and terminal points, respectively. In H. pylori OMV-treated mice, we also observed increased interaction between astrocytemicroglia and microglia-neuron, and ultimately the accumulation of Aβ and cognitive impairment. As the principal immune effector and phagocytic cells in CNS, microglia play an important role in Aβ clearance by a variety of phagocytic and digestive mechanisms [bib_ref] Systemic inflammation impairs microglial Abeta clearance through NLRP3 inflammasome, Tejera [/bib_ref] [bib_ref] Clearance of cerebral Abeta in Alzheimer's disease: reassessing the role of microglia..., Zuroff [/bib_ref]. We found that H. pylori OMV treatment resulted in a decrease in the number of Aβ phagocytosing microglia. In addition, we found that H. pylori OMV administration induced neurodegeneration and synaptic deficits, which were at least partially modulated by phagocytosis of reactive microglia. Indeed, a recent study has also shown that reactive microglia are involved in neurodegeneration and excessive synaptic pruning [bib_ref] Complement C3aR inactivation attenuates tau pathology and reverses an immune network deregulated..., Litvinchuk [/bib_ref]. However, we cannot exclude the additional effects of signals released by neuronal activity to instruct phagocytosis of microglia [bib_ref] Microglia modulate neurodegeneration in Alzheimer's and Parkinson's diseases, Bartels [/bib_ref]. In addition, the abnormal neuronal activity can directly contribute to the Aβ aggregation as we observed in our previous study [bib_ref] Low-grade peripheral inflammation affects brain pathology in the App NL-G-F mouse model..., Xie [/bib_ref] and aggregation of Aβ into compact plaques in turn affects local neuronal dysregulation such as disruption of synaptic integration [bib_ref] Cortical synaptic integration in vivo is disrupted by amyloid-beta plaques, Stern [/bib_ref]. In addition to Aβ accumulation, synapse loss is also closely associated with cognitive decline [bib_ref] The clinical promise of biomarkers of synapse damage or loss in Alzheimer's..., Colom-Cadena [/bib_ref]. In this respect, we speculate that the H. pylori OMV-exacerbated Aβ pathology and synaptic dysfunction are the main contributors to the observed memory defects.
Mechanistically, we identified the C3-C3aR signalling pathway as an important regulator in the H. pylori OMV-mediated detrimental effects on the brain. The complement pathway is a critical regulator of innate immunity and the complement pathway molecules are expressed in the CNS where they have been implicated in AD processes [bib_ref] Complement and microglia mediate early synapse loss in Alzheimer mouse models, Hong [/bib_ref]. H. pylori is complement sensitive and activates mainly the classic pathway [bib_ref] Complement activation directly induced by Helicobacter pylori, Berstad [/bib_ref]. The classical complement activation requires the cleavage of the central complement factor C3 to C3a and C3b, which elicit downstream events through binding to their receptors C3aR and CR3, respectively [bib_ref] The complement system: an unexpected role in synaptic pruning during development and..., Stephan [/bib_ref]. Combining the predominant expression of astroglial C3 and microglial C3aR observed in our study and a previous study [bib_ref] Neurotoxic reactive astrocytes are induced by activated microglia, Liddelow [/bib_ref] , we speculate that H. pylori OMV increases astrocyte-microglia crosstalk via C3-C3aR and microglia activation trigger the neuronal dysfunction, Aβ deposition and ultimately cognitive impairment. Our results are in line with previously reported that C3-C3aR signalling is important in the astrocyte-microglia and neuron-immune crosstalk and that it influences network function and Aβ pathology [bib_ref] Astrocyte-microglia cross talk through complement activation modulates amyloid pathology in mouse models..., Lian [/bib_ref] [bib_ref] NFkappaB-activated astroglial release of complement C3 compromises neuronal morphology and function associated..., Lian [/bib_ref]. Blocking this pathway with pharmacological inhibition of C3aR curtailed H. pylori OMV induced dysfunction of glial cells (proliferation, activation and phagocytosis) and neurons (neurodegeneration and synaptic deficits), and almost completely rescued the Aβ pathology and memory loss, suggesting C3aR inhibition may represent a novel therapeutic target that may ameliorate AD pathology induced by H. pylori infection. Importantly, we observed neither the beneficial effects of C3aRA treatment alone on Aβ pathology nor cognitive functions in App NL-G-F mice despite previously reported that C3aR has a detrimental effect in promoting viral-induced synapse loss and cognitive decline [bib_ref] A complement-microglial axis drives synapse loss during virus-induced memory impairment, Vasek [/bib_ref] and that C3aRA treatment relieves Aβ pathology [bib_ref] Astrocyte-microglia cross talk through complement activation modulates amyloid pathology in mouse models..., Lian [/bib_ref]. Whether the varied outcomes are due to the difference in mouse models, the time of treatment administration or other effects remains to be seen.
It is important to note that in addition to H. pylori, other pathogens such as periodontal bacteria (e.g., P. gingivalis and Fusobacterium nucleatum), Borrelia burgdorferi and Chlamydia pneumoniae have also been reported to be associated with the pathogenesis of AD [bib_ref] Multi-pathogen infections and Alzheimer's disease, Vigasova [/bib_ref]. In addition, a recent study showed that non-pathogenic E. coli OMVs are also able to deliver their cargo from the gut to the brain [bib_ref] Visualizing transfer of microbial biomolecules by outer membrane vesicles in microbe-hostcommunication in..., Bittel [/bib_ref] , suggesting potential effects on brain functions. Our analysis was restricted to the effects of H. pylori-derived OMV on the CNS, so it is uncertain whether these effects are specific to H. pylori OMVs or more broadly to OMVs in general. Therefore, further studies comparing OMVs from different bacteria are needed to decipher whether the observed effects are specific to H. pylori-derived OMVs.
Taken together, we are the first to describe a novel aspect of the gut-brain interplay, namely, that H. pylori-derived OMVs modulate physiological functions of glial cells and neurons, and can worsen AD pathology. We further identified C3-C3aR signalling as an essential regulator in the impact of H. pylori OMVs on brain functionality. Indeed, pharmacological inhibition of C3aR prevents H. pylori OMV induced activation of glial cells, synaptic impairment and neuronal loss, thereby rescuing Aβ pathology and memory functions. These findings not only provide new insights supporting the infectious hypothesis behind AD pathogenesis, but also identify OMVs as important players in the gut-brain axis.
## Au t h o r c o n t r i b u t i o n s
## A c k n o w l e d g e m e n t s
We thank M. Mercken (Johnson & Johnson Pharmaceuticals Research and Development, Beerse, Belgium) for providing the antibodies against Aβ (JRF/ABN/24, JRF/cAB40/28, and JRF/cAB42/26) for the Aβ ELISA, Femke Baeke and Riet De Rycke for performing transmission electron microscopy, Sofie De Bruyckere for technical assistance related to bacterial culture. We also want to thank the VIB Bioimaging Core for training, support and access to the instrument park. This work was supported by Ghent University Special Research fund (BOF) (BOF-STA29-17 and 01G03121), Research Foundation-Flanders (FWO) , Chinese Scholarship Council (CSC) (201808360194), Foundation for Alzheimer's Research Belgium (SAO-FRA) and Baillet Latour Fund. The illustrations in this manuscript were partially created with BioRender.com.
## C o n f l i c t o f i n t e r e s t
The authors have declared that no conflict of interest exists.
O R C I D Junhua Xie https://orcid.org/0000-0003-0684-8515 Ine Vlaeminck https://orcid.org/0000-0001-8934-6179 Keimpe Wierda https://orcid.org/0000-0002-8784-9490
[fig] F: I G U R E Characterization of Helicobacter pylori OMVs. (a) Graphical illustration of H. pylori OMV isolation from bacterial cultures. (b) Particle and protein concentration in the different SEC fractions (0.5 ml per fraction) were determined using nanoparticle tracking analysis (NTA; red) and Nanodrop (brown), respectively. Data show all 30 fractions. (c) Limulus amebocyte lysate (LAL) assay quantified LPS activity levels in the different SEC fractions. (d) Representative negative staining TEM images of purified H. pylori OMVs from SEC fractions #7-11. Scale bars: 1 μm (left) and 0.5 μm (right). (e) Size distributions of purified H. pylori OMVs from SEC fractions #7-11 analysed by NTA. (f) Western blot analysis of laA in bacterial cell lysate and purified H.pylori OMVs from SEC fractions #7-11 (left; 10 μg total protein was loaded) and all fractions separately (right; 20 μl per fraction was loaded). The graphs are shown as mean ± SEM. Data in b, c and d are shown for at least three biological replicates. [/fig]
|
Subsets and Splits